1990-2017
Inventory of
U.S. Greenhouse Gas
Emissions and Sinks
SEPA
United States
Environmental Protection
Agency
EPA 430-R-19-001
-------
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 2017, inclusive, will be made available for the
final report published on April 11, 2019 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, CH4, and N20 emissions was directed by John Steller. 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 Iovinelli 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 N20 fluxes associated with forest land.
We thank the Department of Agriculture's Agricultural Research Service (Stephen Del Grosso) and the Natural
Resource Ecology Laboratory 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, N20
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, CH4 emissions, and N20 emissions from
aquaculture in coastal wetlands.
We would also like to thank Marian Martin Van Pelt, Leslie Chinery, Alexander Lataille, Sabrina Andrews 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, Rani Murali, Emily Kent, Drew Stilson, Cara
Blumenthal, Louise Bruning, Emily Peterson, Helena Caswell, Katie O'Malley, Howard Marano, Neha Vaingankar,
Terrance Glover, Diana Jaramillo, Megha Kedia, and Ursula Jongebloed 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, and
Amie Aguiar support the development of emissions estimates for wastewater. Cortney Itle, Amie Aguiar, Kara
Edquist, Amber Allen, and Spencer Sauter support the inventories for Manure Management, Enteric Fermentation,
Wetlands Remaining Wetlands, and Landfilled Yard Trimmings and Food Scraps (included in Settlements
Remaining Settlements). Casey Pickering, Brandon Long, Gopi Manne, Marty Wolf, Colin Peirce, Bryan Lange, and
Aylin Sertkaya develop estimates for Natural Gas and Petroleum Systems. Cortney Itle, Gopi Manne, 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, 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 14, 2019 and comments received are posted to the EPA web site.
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 Recent Trends in U.S. Greenhouse Gas Emissions and Sinks ES-4
ES.3 Overview of Sector Emissions and Trends ES-18
ES.4 Other Information ES-23
1. INTRODUCTION 1-1
1.1 Background Information 1-3
1.2 National Inventory Arrangements 1-10
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 1-22
1.8 Completeness 1-24
1.9 Organization of Report 1-25
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 2-23
2.3 Precursor Greenhouse Gas Emissions (CO, NOx, NMVOCs, and SO2) 2-34
3. ENERGY 3-1
3.1 Fossil Fuel Combustion (CRF Source Category 1A) 3-5
3.2 Carbon Emitted from Non-Energy Uses of Fossil Fuels (CRF Source Category 1A5) 3-44
3.3 Incineration of Waste (CRF Source Category 1A5) 3-51
3.4 Coal Mining (CRF Source Category lBla) 3-55
3.5 Abandoned Underground Coal Mines (CRF Source Category lBla) 3-60
3.6 Petroleum Systems (CRF Source Category lB2a) 3-64
vi Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
3.7 Natural Gas Systems (CRF Source Category lB2b) 3-80
3.8 Abandoned Oil and Gas Wells (CRF Source Categories lB2a and lB2b) 3-99
3.9 Energy Sources of Precursor Greenhouse Gas Emissions 3-103
3.10 International Bunker Fuels (CRF Source Category 1: Memo Items) 3-104
3.11 WoodBiomass andBiofuels Consumption (CRF Source Category 1A) 3-109
4. INDUSTRIAL PROCESSES AND PRODUCT USE 4-1
4.1 Cement Production (CRF Source Category 2A1) 4-8
4.2 Lime Production (CRF Source Category 2A2) 4-12
4.3 Glass Production (CRF Source Category 2A3) 4-17
4.4 Other Process Uses of Carbonates (CRF Source Category 2A4) 4-20
4.5 Ammonia Production (CRF Source Category 2B1) 4-24
4.6 Urea Consumption for Non-Agricultural Purposes 4-28
4.7 Nitric Acid Production (CRF Source Category 2B2) 4-31
4.8 Adipic Acid Production (CRF Source Category 2B3) 4-35
4.9 Caprolactam, Glyoxal and Glyoxylic Acid Production (CRF Source Category 2B4) 4-39
4.10 Silicon Carbide Production and Consumption (CRF Source Category 2B5) 4-42
4.11 Titanium Dioxide Production (CRF Source Category 2B6) 4-45
4.12 Soda Ash Production (CRF Source Category 2B7) 4-48
4.13 Petrochemical Production (CRF Source Category 2B8) 4-51
4.14 HCFC-22 Production (CRF Source Category 2B9a) 4-58
4.15 Carbon Dioxide Consumption (CRF Source Category 2B10) 4-61
4.16 Phosphoric Acid Production (CRF Source Category 2B10) 4-64
4.17 Iron and Steel Production (CRF Source Category 2C1) and Metallurgical Coke Production 4-68
4.18 Ferroalloy Production (CRF Source Category 2C2) 4-77
4.19 Aluminum Production (CRF Source Category 2C3) 4-81
4.20 Magnesium Production and Processing (CRF Source Category 2C4) 4-86
4.21 Lead Production (CRF Source Category 2C5) 4-91
4.22 Zinc Production (CRF Source Category 2C6) 4-94
4.23 Semiconductor Manufacture (CRF Source Category 2E1) 4-99
4.24 Substitution of Ozone Depleting Substances (CRF Source Category 2F) 4-112
4.25 Electrical Transmission and Distribution (CRF Source Category 2G1) 4-121
4.26 Nitrous Oxide from Product Uses (CRF Source Category 2G3) 4-128
4.27 Industrial Processes and Product Use Sources of Precursor Gases 4-131
5. AGRICULTURE 5-1
5.1 Enteric Fermentation (CRF Source Category 3A) 5-3
5.2 Manure Management (CRF Source Category 3B) 5-9
5.3 Rice Cultivation (CRF Source Category 3C) 5-17
vii
-------
5.4 Agricultural Soil Management (CRF Source Category 3D) 5-23
5.5 Liming (CRF Source Category 3G) 5-42
5.6 Urea Fertilization (CRF Source Category 3H) 5-45
5.7 Field Burning of Agricultural Residues (CRF Source Category 3F) 5-47
6. LAND USE, LAND-USE CHANGE, AND FORESTRY 6-1
6.1 Representation of the U.S. Land Base 6-8
6.2 Forest Land Remaining Forest Land (CRF Category 4 Al) 6-22
6.3 Land Converted to Forest Land (CRF Category 4A2) 6-43
6.4 Cropland Remaining Cropland (CRF Category 4B1) 6-50
6.5 Land Converted to Cropland (CRF Category 4B2) 6-59
6.6 Grassland Remaining Grassland (CRF Category 4C1) 6-65
6.7 Land Converted to Grassland (CRF Category 4C2) 6-74
6.8 Wetlands Remaining Wetlands (CRF Category 4D1) 6-80
6.9 Land Converted to Wetlands (CRF Category 4D2) 6-98
6.10 Settlements Remaining Settlements (CRF Category 4E1) 6-101
6.11 Land Converted to Settlements (CRF Category 4E2) 6-120
6.12 Other Land Remaining Other Land (CRF Category 4F1) 6-125
6.13 Land Converted to Other Land (CRF Category 4F2) 6-126
7. WASTE 7-1
7.1 Landfills (CRF Source Category 5Al) 7-3
7.2 Wastewater Treatment (CRF Source Category 5D) 7-19
7.3 Composting (CRF Source Category 5B1) 7-35
7.4 Waste Incineration (CRF Source Category 5C1) 7-38
7.5 Waste Sources of Precursor Greenhouse Gases 7-38
8. OTHER 8-1
9. RECALCULATIONS AND IMPROVEMENTS 9-1
10. REFERENCES 10-1
viii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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-6
Table ES-3: CO2 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.) ES-12
Table ES-4: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2 Eq.)
ES-18
Table ES-5: U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and
Forestry (MMT CO Eq.) ES-22
Table ES-6: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.) ES-24
Table ES-7: U.S. Greenhouse Gas Emissions by Economic Sector with Electricity-Related Emissions Distributed
(MMTCO2 Eq.) ES-25
Table ES-8: Recent Trends in Various U.S. Data (Index 1990 = 100) ES-26
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 2017) 1-16
Table 1-5: Estimated Overall Inventory Quantitative Uncertainty (MMT CO2 Eq. and Percent) 1-23
Table 1-6: IPCC Sector Descriptions 1-25
Table 1-7: List of Annexes 1-25
Table 2-1: Recent Trends 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) 2-5
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.) 2-10
Table 2-5: CO2 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.) 2-12
Table 2-6: Emissions from Industrial Processes and Product Use (MMT CO2 Eq.) 2-16
Table 2-7: Emissions from Agriculture (MMT CO2 Eq.) 2-18
Table 2-8: U.S. Greenhouse Gas Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and
Forestry (MMT CO2 Eq.) 2-20
Table 2-9: Emissions from Waste (MMT CO2 Eq.) 2-23
Table 2-10: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq. and Percent of Total in
2017) 2-24
Table 2-11: Electric Power-Related Greenhouse Gas Emissions (MMT CO2 Eq.) 2-26
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 2017 2-28
Table 2-13: Transportation-Related Greenhouse Gas Emissions (MMT CO2 Eq.) 2-30
ix
-------
Table 2-14: Recent Trends in Various U.S. Data (Index 1990 = 100) 2-33
Table 2-15: Emissions of NO,. CO, NMVOCs, and SO -(kt) 2-35
Table 3-1: CO2, CH4, and N20 Emissions from Energy (MMT CO2 Eq.) 3-2
Table 3-2: CO2, CH4, and N20 Emissions from Energy (kt) 3-3
Table 3-3: CO2, CH4, and N20 Emissions from Fossil Fuel Combustion (MMT CO2 Eq.) 3-5
Table 3-4: CO2, CH4, and N20 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.) 3-6
Table 3-6: Annual Change in CO2 Emissions and Total 2017 CO2 Emissions from Fossil Fuel Combustion for
Selected Fuels and Sectors (MMT CO2 Eq. and Percent) 3-7
Table 3-7: CO2, CH4, andN20 Emissions from Fossil Fuel Combustion by Sector (MMT CO2 Eq.) 3-11
Table 3-8: CO2, CH4, andN20 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.).... 3-12
Table 3-9: CO2 Emissions from Stationary Fossil Fuel Combustion (MMT CO2 Eq.) 3-13
Table 3-10: CH4 Emissions from Stationary Combustion (MMT CO2 Eq.) 3-13
Table 3-11: N20 Emissions from Stationary Combustion (MMT CO2 Eq.) 3-14
Table 3-12: Electric Power Generation by Fuel Type (Percent) 3-15
Table 3-13: CO2 Emissions from Fossil Fuel Combustion in Transportation End-Use Sector (MMT C02Eq.)... 3-23
Table 3-14: CH4 Emissions from Mobile Combustion (MMT CO2 Eq.) 3-26
Table 3-15: N2O Emissions from Mobile Combustion (MMT CO2 Eq.) 3-27
Table 3-16: Carbon Intensity from Direct Fossil Fuel Combustion by Sector (MMT CO2 Eq./QBtu) 3-32
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) 3-34
Table 3-18: Approach 2 Quantitative Uncertainty Estimates for CH4 and N20 Emissions from Energy-Related
Stationary Combustion, Including Bio mass (MMT CO2 Eq. and Percent) 3-39
Table 3-19: Approach 2 Quantitative Uncertainty Estimates for CH4 and N20 Emissions from Mobile Sources
(MMT CO2 Eq. and Percent) 3-42
Table 3-20: CO2 Emissions from Non-Energy Use Fossil Fuel Consumption (MMT CO2 Eq. and Percent) 3-45
Table 3-21: Adjusted Consumption of Fossil Fuels for Non-Energy Uses (TBtu) 3-46
Table 3-22: 2017 Adjusted Non-Energy Use Fossil Fuel Consumption, Storage, and Emissions 3-46
Table 3-23: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Non-Energy Uses of Fossil
Fuels (MMT CO2 Eq. and Percent) 3-48
Table 3-24: Approach 2 Quantitative Uncertainty Estimates for Storage Factors of Non-Energy Uses of Fossil Fuels
(Percent) 3-48
Table 3-25: CO2, CH4, andN20 Emissions from the Incineration of Waste (MMT CO2 Eq.) 3-52
Table 3-26: CO2, CH4, and N20 Emissions from the Incineration of Waste (kt) 3-52
Table 3-27: Municipal Solid Waste Generation (Metric Tons) and Percent Combusted (BioCycle dataset) 3-53
Table 3-28: Approach 2 Quantitative Uncertainty Estimates for CO2 and N20 from the Incineration of Waste (MMT
CO2 Eq. and Percent) 3-54
Table 3-29: Coal Production (kt) 3-55
Table 3-30: CH4 Emissions from Coal Mining (MMT CO2 Eq.) 3-56
x Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 3-31: CH4 Emissions from Coal Mining (kt) 3-56
Table 3-32: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Coal Mining (MMT CO2 Eq.
and Percent) 3-59
Table 3-33: CH4 Emissions from Abandoned Coal Mines (MMT CO2 Eq.) 3-60
Table 3-34: CH4 Emissions from Abandoned Coal Mines (kt) 3-61
Table 3-35: Number of Gassy Abandoned Mines Present in U.S. Basins in 2017, Grouped by Class According to
Post-Abandonment State 3-62
Table 3-36: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Abandoned Underground Coal
Mines (MMT CO2 Eq. and Percent) 3-63
Table 3-37: CH4 Emissions from Petroleum Systems (MMT CO2 Eq.) 3-65
Table 3-38: CH4 Emissions from Petroleum Systems (kt CH4) 3-66
Table 3-39: CO2 Emissions from Petroleum Systems (MMT CO2) 3-66
Table 3-40: CO2 Emissions from Petroleum Systems (kt CO2) 3-66
Table 3-41: N20 Emissions from Petroleum Systems (metric tons CO2 Eq.) 3-66
Table 3-42: N20 Emissions from Petroleum Systems (metric tons N20) 3-66
Table 3-43: Approach 2 Quantitative Uncertainty Estimates for CH4 and CO2 Emissions from Petroleum Systems
(MMT CO2 Eq. and Percent) 3-69
Table 3-44: Recalculations of CO2 in Petroleum Systems (MMT CO2) 3-71
Table 3-45: Recalculations of CH4 in Petroleum Systems (MMT CO2 Eq.) 3-71
Table 3-46: HF Oil Well Completions National CH4 Emissions (Metric Tons CH4) 3-72
Table 3-47: HF Oil Well Completions National CO2 Emissions (kt CO2) 3-72
Table 3-48: Count of Oil Wells Drilled 3-72
Table 3-49: HF Oil Well Workovers National CH4 Emissions (Metric Tons CH4) 3-73
Table 3-50: HF Oil Well Workovers National CO2 Emissions (kt CO2) 3-73
Table 3-51: Production Storage Tank National CH4 Emissions (Metric Tons CH4) 3-74
Table 3-52: Production Storage Tank National CO2 Emissions (kt CO2) 3-74
Table 3-53: Pneumatic Controller National CH4 Emissions (Metric Tons CH4) 3-74
Table 3-54: Associated Gas Venting and Flaring National CO2 Emissions (kt CO2) 3-75
Table 3-55: Miscellaneous Production Flaring National CO2 Emissions (kt CO2) 3-75
Table 3-56: Chemical Injection Pump National CH4 Emissions (Metric Tons CH4) 3-75
Table 3-57: Heater National CH4 Emissions (Metric Tons CH4) 3-76
Table 3-58: Producing Oil Well Count Data 3-76
Table 3-59: Oil Production Data (Million Barrels) 3-77
Table 3-60: Crude Oil Transportation National CO2 Emissions (kt CO2) 3-77
Table 3-61: N2O National Emissions (Metric Tons N2O) 3-77
Table 3-62: Quantity of CO2 Captured and Extracted for EOR Operations (MMT CO2) 3-79
Table 3-63: Quantity of CO2 Captured and Extracted for EOR Operations (kt) 3-80
Table 3-64: CH4 Emissions from Natural Gas Systems (MMT CO2 Eq.)a 3-82
xi
-------
Table 3-65: CH4 Emissions from Natural Gas Systems (kt)a 3-82
Table 3-66: Calculated Potential CH4 and Captured/Combusted CH4 from Natural Gas Systems (MMT CO2 Eq.). 3-
83
Table 3-67: Non-combustion CO2 Emissions from Natural Gas Systems (MMT) 3-83
Table 3-68: Non-combustion CO2 Emissions from Natural Gas Systems (kt) 3-83
Table 3-69: N20 Emissions from Natural Gas Systems (Metric Tons CO2 Eq.) 3-84
Table 3-70: N20 Emissions from Natural Gas Systems (Metric Tons N20) 3-84
Table 3-71: Approach 2 Quantitative Uncertainty Estimates for CH4 and Non-combustion CO2 Emissions from
Natural Gas Systems (MMT CO2 Eq. and Percent) 3-86
Table 3-72: Recalculations of CO2 in Natural Gas Systems (MMT CO2) 3-88
Table 3-73: Recalculations of CH4 in Natural Gas Systems (MMT CO2 Eq.) 3-89
Table 3-74: Count of Gas Wells Drilled 3-89
Table 3-75: HF Gas Well Completions National CO2 Emissions (kt CO2) 3-90
Table 3-76: Gathering Pipelines National CH4 Emissions (Metric Tons CH4) 3-90
Table 3-77: Gathering Stations National CH4 Emissions (Metric Tons CH4) 3-90
Table 3-78: Miscellaneous Production Flaring National Emissions (kt CO2) 3-91
Table 3-79: Production Segment Gas Engines National Emissions (Metric Tons CH4) 3-91
Table 3-80: Production Segment Pneumatic Controller National Emissions (Metric Tons CH4) 3-91
Table 3-81: Liquids Unloading National Emissions (Metric Tons CH4) 3-92
Table 3-82: Production Segment Storage Tanks National Emissions (kt CO2) 3-92
Table 3-83: HF Gas Well Workovers National Emissions (Metric Tons CH4) 3-92
Table 3-84: HF Gas Well Workovers National Emissions (kt CO2) 3-93
Table 3-85: Producing Gas Well Count Data 3-93
Table 3-86: AGR National CO2 Emissions (kt CO2) 3-94
Table 3-87: Processing Segment Flares National CO2 Emissions (kt CO2) 3-94
Table 3-88: Processing Segment Gas Engines National Emissions (Metric Tons CH4) 3-94
Table 3-89: Transmission Pipeline Blowdowns National CH4 Emissions (Metric Tons CH4) 3-95
Table 3-90: Transmission Pipeline Blowdowns National CO2 Emissions (kt CO2) 3-95
Table 3-91: LNG Storage Station National CH4 Emissions (Metric Tons CH4) 3-95
Table 3-92: LNG Storage Station National CO2 Emissions (kt CO2) 3-95
Table 3-93: LNG Import/Export Terminal National CH4 Emissions (Metric Tons CH4) 3-96
Table 3-94: LNG Import/Export Terminal National CO2 Emissions (kt CO2) 3-96
Table 3-95: N2O National Emissions (Metric Tons N2O) 3-97
Table 3-96: CH4 Emissions from Abandoned Oil and Gas Wells (MMT CO2 Eq.) 3-100
Table 3-97: CH4 Emissions from Abandoned Oil and Gas Wells (kt) 3-100
Table 3-98: CO2 Emissions from Abandoned Oil and Gas Wells (MMT CO2) 3-100
Table 3-99: CO2 Emissions from Abandoned Oil and Gas Wells (kt) 3-100
Table 3-100: Abandoned Oil Wells Activity Data, CH4 and CO2 Emissions (Metric Tons) 3-101
xii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 3-101: Abandoned Gas Wells Activity Data, CH4 and CO2 Emissions (Metric Tons) 3-101
Table 3-102: Approach 2 Quantitative Uncertainty Estimates for CH4 and CO2 Emissions from Petroleum and
Natural Gas Systems (MMT CO2 Eq. and Percent) 3-102
Table 3-103: NOx, CO, and NMVOC Emissions from Energy-Related Activities (kt) 3-103
Table 3-104: CO2, CH4, and N2O Emissions from International Bunker Fuels (MMT CO2 Eq.) 3-105
Table 3-105: CO2, CH4, and N2O Emissions from International Bunker Fuels (kt) 3-106
Table 3-106: Aviation Jet Fuel Consumption for International Transport (Million Gallons) 3-107
Table 3-107: Marine Fuel Consumption for International Transport (Million Gallons) 3-107
Table 3-108: CO2 Emissions from Wood Consumption by End-Use Sector (MMT CO2 Eq.) 3-109
Table 3-109: CO2 Emissions from Wood Consumption by End-Use Sector (kt) 3-109
Table 3-110: CO2 Emissions from Ethanol Consumption (MMT CO2 Eq.) 3-110
Table 3-111: CO2 Emissions from Ethanol Consumption (kt) 3-110
Table 3-112: CO2 Emissions from Biodiesel Consumption (MMT CO2 Eq.) 3-110
Table 3-113: CO2 Emissions from Biodiesel Consumption (kt) 3-111
Table 3-114: Woody Biomass Consumption by Sector (Trillion Btu) 3-111
Table 3-115: Ethanol Consumption by Sector (TrillionBtu) 3-111
Table 3-116: Biodiesel Consumption by Sector (TrillionBtu) 3-112
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) 4-10
Table 4-5: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Cement Production (MMT CO2
Eq. and Percent) 4-11
Table 4-6: CO2 Emissions from Lime Production (MMT CO2 Eq. and kt) 4-13
Table 4-7: Potential, Recovered, and Net CO2 Emissions from Lime Production (kt) 4-13
Table 4-8: High-Calcium- and Dolomitic-Quicklime, High-Calcium- and Dolomitic-Hydrated, and Dead-Burned-
Dolomite Lime Production (kt) 4-14
Table 4-9: Adjusted Lime Production (kt) 4-14
Table 4-10: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lime Production (MMT CO2
Eq. and Percent) 4-16
Table 4-11: CO2 Emissions from Glass Production (MMT CO2 Eq. and kt) 4-18
Table 4-12: Limestone, Dolomite, and Soda Ash Consumption Used in Glass Production (kt) 4-18
Table 4-13: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Glass Production (MMT CO2
Eq. and Percent) 4-19
Table 4-14: CO2 Emissions from Other Process Uses of Carbonates (MMT CO2 Eq.) 4-21
Table 4-15: CO2 Emissions from Other Process Uses of Carbonates (kt) 4-21
Table 4-16: Limestone and Dolomite Consumption (kt) 4-22
Table 4-17: Soda Ash Consumption Not Associated with Glass Manufacturing (kt) 4-23
xiii
-------
Table 4-18: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Other Process Uses of
Carbonates (MMT CO2 Eq. and Percent) 4-23
Table 4-19: CO2 Emissions from Ammonia Production (MMT CO2 Eq.) 4-25
Table 4-20: CO2 Emissions from Ammonia Production (kt) 4-25
Table 4-21: Ammonia Production, Recovered CO2 Consumed for Urea Production, and Urea Production (kt)... 4-26
Table 4-22: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Ammonia Production (MMT
CO2 Eq. and Percent) 4-27
Table 4-23: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (MMT CO2 Eq.) 4-29
Table 4-24: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (kt) 4-29
Table 4-25: Urea Production, Urea Applied as Fertilizer, Urea Imports, and Urea Exports (kt) 4-30
Table 4-26: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Urea Consumption for Non-
Agricultural Purposes (MMT CO2 Eq. and Percent) 4-31
Table 4-27: N2O Emissions from Nitric Acid Production (MMT CO2 Eq. and kt N2O) 4-32
Table 4-28: Nitric Acid Production (kt) 4-34
Table 4-29: Approach 2 Quantitative Uncertainty Estimates for N20 Emissions from Nitric Acid Production (MMT
CO2 Eq. and Percent) 4-35
Table 4-30: N2O Emissions from Adipic Acid Production (MMT CO2 Eq. and kt N2O) 4-36
Table 4-31: Adipic Acid Production (kt) 4-38
Table 4-32: Approach 2 Quantitative Uncertainty Estimates for N20 Emissions from Adipic Acid Production
(MMT CO2 Eq. and Percent) 4-38
Table 4-33: N20 Emissions from Caprolactam Production (MMT CO2 Eq. and kt N20) 4-40
Table 4-34: Caprolactam Production (kt) 4-41
Table 4-35: Approach 2 Quantitative Uncertainty Estimates for N20 Emissions from Caprolactam, Glyoxal and
Glyoxylic Acid Production (MMT CO2 Eq. and Percent) 4-41
Table 4-36: CO2 and CH4 Emissions from Silicon Carbide Production and Consumption (MMT CO2 Eq.) 4-43
Table 4-37: CO2 and CH4 Emissions from Silicon Carbide Production and Consumption (kt) 4-43
Table 4-38: Production and Consumption of Silicon Carbide (Metric Tons) 4-44
Table 4-39: Approach 2 Quantitative Uncertainty Estimates for CH4 and CO2 Emissions from Silicon Carbide
Production and Consumption (MMT CO2 Eq. and Percent) 4-45
Table 4-40: CO2 Emissions from Titanium Dioxide (MMT CO2 Eq. and kt) 4-46
Table 4-41: Titanium Dioxide Production (kt) 4-47
Table 4-42: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Titanium Dioxide Production
(MMT CO2 Eq. and Percent) 4-47
Table 4-43: CO2 Emissions from Soda Ash Production (MMT CO2 Eq. and kt CO2) 4-49
Table 4-44: Soda Ash Production (kt) 4-50
Table 4-45: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Soda Ash Production (MMT
CO2 Eq. and Percent) 4-50
Table 4-46: CO2 and CH4 Emissions from Petrochemical Production (MMT CO2 Eq.) 4-52
Table 4-47: CO2 and CH4 Emissions from Petrochemical Production (kt) 4-53
Table 4-48: Production of Selected Petrochemicals (kt) 4-55
xiv Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-49: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Petrochemical Production and
CO2 Emissions from Petrochemical Production (MMT CO2 Eq. and Percent) 4-56
Table 4-50: HFC-23 Emissions from HCFC-22 Production (MMT C02 Eq. and kt HFC-23) 4-59
Table 4-51: HCFC-22 Production (kt) 4-59
Table 4-52: Approach 2 Quantitative Uncertainty Estimates for HFC-23 Emissions from HCFC-22 Production
(MMT CO2 Eq. and Percent) 4-60
Table 4-53: CO2 Emissions from CO2 Consumption (MMT CO2 Eq. and kt) 4-61
Table 4-54: CO2 Production (kt CO2) and the Percent Used for Non-EOR Applications 4-63
Table 4-55: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from CO2 Consumption (MMT CO2
Eq. and Percent) 4-64
Table 4-56: CO2 Emissions from Phosphoric Acid Production (MMT CO2 Eq. and kt) 4-65
Table 4-57: Phosphate Rock Domestic Consumption, Exports, and Imports (kt) 4-66
Table 4-58: Chemical Composition of Phosphate Rock (Percent by Weight) 4-66
Table 4-59: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Phosphoric Acid Production
(MMT CO2 Eq. and Percent) 4-67
Table 4-60: CO2 Emissions from Metallurgical Coke Production (MMT CO2 Eq.) 4-69
Table 4-61: CO2 Emissions from Metallurgical Coke Production (kt) 4-69
Table 4-62: CO2 Emissions from Iron and Steel Production (MMT CO2 Eq.) 4-69
Table 4-63: CO2 Emissions from Iron and Steel Production (kt) 4-70
Table 4-64: CH4 Emissions from Iron and Steel Production (MMT CO2 Eq.) 4-70
Table 4-65: CH4 Emissions from Iron and Steel Production (kt) 4-70
Table 4-66: Material Carbon Contents for Metallurgical Coke Production 4-71
Table 4-67: Production and Consumption Data for the Calculation of CO2 Emissions from Metallurgical Coke
Production (Thousand Metric Tons) 4-72
Table 4-68: Production and Consumption Data for the Calculation of CO2 Emissions from Metallurgical Coke
Production (Million ft3) 4-72
Table 4-69: Material Carbon Contents for Iron and Steel Production 4-73
Table 4-70: CH4 Emission Factors for Sinter and Pig Iron Production 4-73
Table 4-71: CO2 Emission Factors for Sinter Production, Direct Reduced Iron Production and Pellet Production 4-74
Table 4-72: Production and Consumption Data for the Calculation of CO2 and CH4 Emissions from Iron and Steel
Production (Thousand Metric Tons) 4-74
Table 4-73: Production and Consumption Data for the Calculation of CO2 Emissions from Iron and Steel
Production (Million ft3 unless otherwise specified) 4-75
Table 4-74: Approach 2 Quantitative Uncertainty Estimates for CO2 and CH4 Emissions from Iron and Steel
Production and Metallurgical Coke Production (MMT CO2 Eq. and Percent) 4-76
Table 4-75: CO2 and CH4 Emissions from Ferroalloy Production (MMT CO2 Eq.) 4-78
Table 4-76: CO2 and CH4 Emissions from Ferroalloy Production (kt) 4-78
Table 4-77: Production of Ferroalloys (Metric Tons) 4-79
Table 4-78: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Ferroalloy Production (MMT
CO2 Eq. and Percent) 4-80
xv
-------
Table 4-79: CO2 Emissions from Aluminum Production (MMT CO2 Eq. and kt) 4-81
Table 4-80: PFC Emissions from Aluminum Production (MMT CO2 Eq.) 4-82
Table 4-81: PFC Emissions from Aluminum Production (kt) 4-82
Table 4-82: Production of Primary Aluminum (kt) 4-85
Table 4-83: Approach 2 Quantitative Uncertainty Estimates for CO2 and PFC Emissions from Aluminum
Production (MMT CO2 Eq. and Percent) 4-86
Table 4-84: SF6, HFC-134a, FK 5-1-12 and CO2 Emissions from Magnesium Production and Processing (MMT
C02 Eq.) 4-86
Table 4-85: SF6, HFC-134a, FK 5-1-12 and CO2 Emissions from Magnesium Production and Processing (kt)... 4-87
Table 4-86: SF6 Emission Factors (kg SF6 per metric ton of magnesium) 4-89
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) 4-90
Table 4-88: CO2 Emissions from Lead Production (MMT CO2 Eq. and kt) 4-92
Table 4-89: Lead Production (Metric Tons) 4-93
Table 4-90: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lead Production (MMT CO2
Eq. and Percent) 4-93
Table 4-91: CO2 Emissions from Zinc Production (MMT CO2 Eq. and kt) 4-95
Table 4-92: Zinc Production (Metric Tons) 4-95
Table 4-93: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Zinc Production (MMT CO2
Eq. and Percent) 4-98
Table 4-94: PFC, HFC, SF6, NF3, and N20 Emissions from Semiconductor Manufacture (MMT CO2 Eq.) 4-100
Table 4-95: PFC, HFC, SF6, NF3, and N20 Emissions from Semiconductor Manufacture (kt) 4-101
Table 4-96: F-HTF Emissions Based on GHGRP Reporting (MMT CO2 Eq.) 4-101
Table 4-97: Top 10 F-HTF Compounds with Largest Emissions Based on GHGRP Reporting (tons) 4-101
Table 4-98: Approach 2 Quantitative Uncertainty Estimates for HFC, PFC, SF6, NF3 and N20 Emissions from
Semiconductor Manufacture (MMT CO2 Eq. and Percent)3 4-110
Table 4-99: Emissions of HFCs and PFCs from ODS Substitutes (MMT CO2 Eq.) 4-112
Table 4-100: Emissions of HFCs and PFCs from ODS Substitution (Metric Tons) 4-112
Table 4-101: Emissions of HFCs and PFCs from ODS Substitutes (MMT CO2 Eq.) by Sector 4-113
Table 4-102: Approach 2 Quantitative Uncertainty Estimates for HFC and PFC Emissions from ODS Substitutes
(MMT CO2 Eq. and Percent) 4-115
Table 4-103: U.S. HFC Supply (MMT CO Eq.) 4-117
Table 4-104: Averaged U.S. HFC Demand (MMT CO Eq.) 4-119
Table 4-105: SF6 Emissions from Electric Power Systems and Electrical Equipment Manufacturers (MMT CO2 Eq.)
4-121
Table 4-106: SF6 Emissions from Electric Power Systems and Electrical Equipment Manufacturers (kt) 4-121
Table 4-107: Transmission Mile Coverage (Percent) and Regression Coefficients (kg per mile) 4-125
Table 4-108: Approach 2 Quantitative Uncertainty Estimates for SF6 Emissions from Electrical Transmission and
Distribution (MMT CO2 Eq. and Percent) 4-126
Table 4-109: N2O Production (kt) 4-128
xvi Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-110: N20 Emissions from N20 Product Usage (MMT CO2 Eq. and kt) 4-129
Table 4-111: Approach 2 Quantitative Uncertainty Estimates for N20 Emissions from N20 Product Usage (MMT
CO2 Eq. and Percent) 4-130
Table 4-112: NOx, CO, and NMVOC Emissions from Industrial Processes and Product Use (kt) 4-132
Table 5-1: Emissions from Agriculture (MMT CO2 Eq.) 5-2
Table 5-2: Emissions from Agriculture (kt) 5-2
Table 5-3: CH4 Emissions from Enteric Fermentation (MMT CO2 Eq.) 5-3
Table 5-4: CH4 Emissions from Enteric Fermentation (kt) 5-3
Table 5-5: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Enteric Fermentation (MMT
CO2 Eq. and Percent) 5-7
Table 5-6: CH4 and N2O Emissions from Manure Management (MMT CO2 Eq.) 5-10
Table 5-7: CH4 and N20 Emissions from Manure Management (kt) 5-11
Table 5-8: Approach 2 Quantitative Uncertainty Estimates for CH4 and N20 (Direct and Indirect) Emissions from
Manure Management (MMT CO2 Eq. and Percent) 5-15
Table 5-9: IPCC (2006) Implied Emission Factor Default Values Compared with Calculated Values for CH4 from
Manure Management (kg/head/year) 5-15
Table 5-10: CH4 Emissions from Rice Cultivation (MMT CO2 Eq.) 5-18
Table 5-11: CH4 Emissions from Rice Cultivation (kt) 5-18
Table 5-12: Rice Area Harvested (1,000 Hectares) 5-20
Table 5-13: Average Ratooned Area as Percent of Primary Growth Area (Percent) 5-21
Table 5-14: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Rice Cultivation (MMT CO2
Eq. and Percent) 5-22
Table 5-15: N20 Emissions from Agricultural Soils (MMT CO2 Eq.) 5-25
Table 5-16: N20 Emissions from Agricultural Soils (kt) 5-25
Table 5-17: Direct N20 Emissions from Agricultural Soils by Land Use Type and N Input Type (MMT CO2 Eq.) 5-
25
Table 5-18: Indirect N20 Emissions from Agricultural Soils (MMT CO2 Eq.) 5-26
Table 5-19: Quantitative Uncertainty Estimates of N20 Emissions from Agricultural Soil Management in 2017
(MMT CO2 Eq. and Percent) 5-40
Table 5-20: Emissions from Liming (MMT CO2 Eq.) 5-42
Table 5-21: Emissions from Liming (MMT C) 5-42
Table 5-22: Applied Minerals (MMT) 5-44
Table 5-23: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Liming (MMT CO2 Eq. and
Percent) 5-44
Table 5-24: CO2 Emissions from Urea Fertilization (MMT CO2 Eq.) 5-45
Table 5-25: CO2 Emissions from Urea Fertilization (MMT C) 5-45
Table 5-26: Applied Urea (MMT) 5-45
Table 5-27: Quantitative Uncertainty Estimates for CO2 Emissions from Urea Fertilization (MMT CO2 Eq. and
Percent) 5-46
Table 5-28: CH4 and N20 Emissions from Field Burning of Agricultural Residues (MMT CO2 Eq.) 5-47
xvii
-------
Table 5-29: CH4, N20, CO, and NOx Emissions from Field Burning of Agricultural Residues (kt) 5-48
Table 5-30: Agricultural Crop Production (kt of Product) 5-51
Table 5-31: U.S. Average Percent Crop Area Burned by Crop (Percent) 5-52
Table 5-32: Parameters for Estimating Emissions from Field Burning of Agricultural Residues 5-52
Table 5-33: Greenhouse Gas Emission Ratios and Conversion Factors 5-53
Table 5-34: Approach 2 Quantitative Uncertainty Estimates for CH4 and N20 Emissions from Field Burning of
Agricultural Residues (MMT CO2 Eq. and Percent) 5-54
Table 6-1: Net CO2 Flux from Land Use, Land-Use Change, and Forestry (MMT CO2 Eq.) 6-2
Table 6-2: Emissions from Land Use, Land-Use Change, and Forestry by Gas (MMT CO2 Eq.) 6-3
Table 6-3: Emissions and Removals (Net Flux) from Land Use, Land-Use Change, and Forestry (MMT CO2 Eq.) 6-
4
Table 6-4: Emissions and Removals from Land Use, Land-Use Change, and Forestry (MMT CO2 Eq.) 6-5
Table 6-5: Emissions and Removals from Land Use, Land-Use Change, and Forestry (kt) 6-6
Table 6-6: Managed and Unmanaged Land Area by Land-Use Categories for All 50 States (Thousands of Hectares)
6-9
Table 6-7: Land Use and Land-Use Change for the U.S. Managed Land Base for All 50 States (Thousands of
Hectares) 6-10
Table 6-8: Data Sources Used to Determine Land Use and Land Area for the Conterminous United States, Hawaii,
and Alaska 6-16
Table 6-9: Total Land Area (Hectares) by Land-Use Category for U.S. Territories 6-22
Table 6-10: Net CO2 Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land and Harvested
Wood Pools (MMT CO Eq.) 6-26
Table 6-11: Net C Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land and Harvested Wood
Pools (MMT C) 6-26
Table 6-12: Forest Area (1,000 ha) and C Stocks in Forest Land Remaining Forest Land and Harvested Wood
Pools (MMT C) 6-27
Table 6-13: Estimates of CO2 (MMT per Year) Emissions from Forest Fires in the Conterminous 48 States and
Alaska3 6-29
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) 6-33
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) 6-35
Table 6-16: Recalculations of Net C Flux from Forest Ecosystem Pools in Forest Land Remaining Forest Land and
Harvested Wood Pools (MMT C) 6-35
Table 6-17: Non-C02 Emissions from Forest Fires (MMT CO2 Eq.)a 6-37
Table 6-18: Non-C02 Emissions from Forest Fires (kt)a 6-37
Table 6-19: Quantitative Uncertainty Estimates of Non-C02 Emissions from Forest Fires (MMT CO2 Eq. and
Percent)3 6-37
Table 6-20: N2O Fluxes from Soils in Forest Land Remaining Forest Land and Land Converted to Forest Land
(MMT C02 Eq. and kt N O) 6-39
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) 6-40
xviii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 6-22: N011-CO2 Emissions from Drained Organic Forest Soilsa,b (MMT CO2 Eq.) 6-41
Table 6-23: Non-CCh Emissions from Drained Organic Forest Soilsa b (kt) 6-41
Table 6-24: States identified as having Drained Organic Soils, Area of Forest on Drained Organic Soils, and
Sampling Error 6-42
Table 6-25: Quantitative Uncertainty Estimates for Non-C02 Emissions on Drained Organic Forest Soils (MMT
CO2 Eq. and Percent)3 6-43
Table 6-26: Net CO2 Flux from Forest C Pools in Land Converted to Forest Land by Land Use Change Category
(MMT C02 Eq.) 6-44
Table 6-27: Net C Flux from Forest C Pools in Land Converted to Forest Land by Land Use Change Category
(MMT C) 6-45
Table 6-28: Quantitative Uncertainty Estimates for Forest C Pool Stock Changes (MMT CO2 Eq. per Year) in 2017
from Land Converted to Forest Land by Land Use Change 6-47
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) 6-49
Table 6-30: Net CO2 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT CO2 Eq.) 6-51
Table 6-31: Net CO2 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT C) 6-51
Table 6-32: Approach 2 Quantitative Uncertainty Estimates for Soil C Stock Changes occurring within Cropland
Remaining Cropland (MMT CO2 Eq. and Percent) 6-58
Table 6-33: 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.) 6-60
Table 6-34: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in Land Converted to
Cropland (MMT C) 6-61
Table 6-35: 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) 6-63
Table 6-36: Net CO2 Flux from Soil C Stock Changes in Grassland Remaining Grassland (MMT CO2 Eq.) 6-66
Table 6-37: Net CO2 Flux from Soil C Stock Changes in Grassland Remaining Grassland (MMT C) 6-66
Table 6-38: Approach 2 Quantitative Uncertainty Estimates for C Stock Changes Occurring Within Grassland
Remaining Grassland (MMT CO2 Eq. and Percent) 6-70
Table 6-39: CH4 and N20 Emissions from Biomass Burning in Grassland (MMT CO2 Eq.) 6-71
Table 6-40: CH4, N20, CO, and NOx Emissions from Biomass Burning in Grassland (kt) 6-71
Table 6-41: Thousands of Grassland Hectares Burned Annually 6-72
Table 6-42: Uncertainty Estimates for Non-C02 Greenhouse Gas Emissions from Biomass Burning in Grassland
(MMT CO2 Eq. and Percent) 6-73
Table 6-43: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Grassland (MMT CO2 Eq.) 6-74
Table 6-44: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Grassland (MMT C) 6-75
Table 6-45: 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) 6-78
Table 6-46: Emissions from PeatlandsRemaining Peatlands (MMT CO2 Eq.) 6-81
Table 6-47: Emissions from Peatlands Remaining Peatlands (kt) 6-81
Table 6-48: Peat Production of Lower 48 States (kt) 6-83
xix
-------
Table 6-49: Peat Production of Alaska (Thousand Cubic Meters) 6-83
Table 6-50: Peat Production Area (Hectares) 6-83
Table 6-51: Approach 2 Quantitative Uncertainty Estimates for CO2, CH4, and N20 Emissions from Peatlands
Remaining Peatlands (MMT CO2 Eq. and Percent) 6-85
Table 6-52: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands (MMT CO Eq.) 6-87
Table 6-53: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands (MMT C) 6-88
Table 6-54: CH4 Emissions from Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands (MMT CO2
Eq. andkt CII ) 6-88
Table 6-55: 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) 6-89
Table 6-56: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to Unvegetated Open Water
Coastal Wetlands (MMT CO2 Eq.) 6-91
Table 6-57: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to Unvegetated Open Water
Coastal Wetlands (MMT C) 6-91
Table 6-58: 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) 6-92
Table 6-59: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands Converted to
Vegetated Coastal Wetlands (MMT CO2 Eq.) 6-94
Table 6-60: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands Converted to
Vegetated Coastal Wetlands (MMT C) 6-94
Table 6-61: 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) 6-95
Table 6-62: N20 Emissions from Aquaculture in Coastal Wetlands (MMT CO2 Eq.) 6-96
Table 6-63: Approach 1 Quantitative Uncertainty Estimates for N20 Emissions for Aquaculture Production in
Coastal Wetlands (MMT CO2 Eq. and Percent) 6-97
Table 6-64: CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal Wetlands (MMT CO2 Eq.) 6-
98
Table 6-65: CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal Wetlands (MMT C) 6-98
Table 6-66: CH4 Emissions from Land Converted to Vegetated Coastal Wetlands (MMT CO2 Eq. and kt CH4). 6-98
Table 6-67: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes occurring within Land Converted
to Vegetated Coastal Wetlands (MMT CO2 Eq. and Percent) 6-100
Table 6-68: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements (MMT CO2 Eq.). 6-102
Table 6-69: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements (MMT C) 6-102
Table 6-70: Thousands of Hectares of Drained Organic Soils in Settlements Remaining Settlements 6-102
Table 6-71: Uncertainty Estimates for CO2 Emissions from Drained Organic Soils in Settlements Remaining
Settlements (MMT CO2 Eq. and Percent) 6-103
Table 6-72: Net C Flux from Settlement Trees (MMT CO2 Eq. and MMT C) 6-104
Table 6-73: 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) 6-106
xx Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 6-74: 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 (2017) 6-108
Table 6-75: Approach 2 Quantitative Uncertainty Estimates for Net C Flux from Changes in C Stocks in Settlement
Trees (MMT CO2 Eq. and Percent) 6-110
Table 6-76: Comparison of Settlement, Developed and Urban Land Area for Conterminous United States 6-110
Table 6-77: N2O Emissions from Soils in Settlements Remaining Settlements (MMT CO2 Eq. and kt N2O) 6-112
Table 6-78: Quantitative Uncertainty Estimates of N2O Emissions from Soils in Settlements Remaining Settlements
(MMT CO2 Eq. and Percent) 6-114
Table 6-79: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills (MMT CO2 Eq.) 6-116
Table 6-80: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills (MMT C) 6-116
Table 6-81: Moisture Contents, C Storage Factors (Proportions of Initial C Sequestered), Initial C Contents, and
Decay Rates for Yard Trimmings and Food Scraps in Landfills 6-118
Table 6-82: C Stocks in Yard Trimmings and Food Scraps in Landfills (MMT C) 6-119
Table 6-83: Approach 2 Quantitative Uncertainty Estimates for CO2 Flux from Yard Trimmings and Food Scraps in
Landfills (MMT CO2 Eq. and Percent) 6-119
Table 6-84: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Settlements (MMT CO2 Eq.) 6-121
Table 6-85: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for Land Converted to
Settlements (MMT C) 6-122
Table 6-86: 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) 6-124
Table 7-1: Emissions from Waste (MMT CO2 Eq.) 7-1
Table 7-2: Emissions from Waste (kt) 7-2
Table 7-3: CH4 Emissions from Landfills (MMT CO2 Eq.) 7-5
Table 7-4: CH4 Emissions from Landfills (kt) 7-5
Table 7-5: Approach 2 Quantitative Uncertainty Estimates for CH4 Emissions from Landfills (MMT CO2 Eq. and
Percent) 7-13
Table 7-6: Materials Discarded3 in the Municipal Waste Stream by Waste Type from 1990 to 2015 (Percent)b .. 7-17
Table 7-7: CH4 and N20 Emissions from Domestic and Industrial Wastewater Treatment (MMT CO2 Eq.) 7-20
Table 7-8: CH4 and N20 Emissions from Domestic and Industrial Wastewater Treatment (kt) 7-20
Table 7-9: U.S. Population (Millions) and Domestic Wastewater BOD5 Produced (kt) 7-23
Table 7-10: Domestic Wastewater CH4 Emissions from Septic and Centralized Systems (2017, MMT CO2 Eq. and
Percent) 7-23
Table 7-11: Industrial Wastewater CH4 Emissions by Sector (2017, 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 7-29
xxi
-------
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) 7-32
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 N20 Emissions from Composting (MMT CO2 Eq.) 7-36
Table 7-19: CH4 and N20 Emissions from Composting (kt) 7-36
Table 7-20: U.S. Waste Composted (kt) 7-36
Table 7-21: Tier 1 Quantitative Uncertainty Estimates for Emissions from Composting (MMT CO2 Eq. and Percent)
7-37
Table 7-22: Emissions of NOx, CO, and NMVOC from Waste (kt) 7-38
Table 9-1: Revisions to U.S. Greenhouse Gas Emissions (MMT CO2 Eq.) 9-3
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.) 9-5
Figure ES-1: Gross U.S. Greenhouse Gas Emissions by Gas (MMT CO2 Eq.) ES-4
Figure ES-2: Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the Previous Year ..ES-5
Figure ES-3: Cumulative Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to 1990 (1990=0,
MM I CO- Eq.) ES-5
Figure ES-4: 2017 U.S. Greenhouse Gas Emissions by Gas (Percentages based on MMT CO2 Eq.) ES-9
Figure ES-5: 2017 Sources of CO2 Emissions (MMT CO2 Eq.) ES-10
Figure ES-6: 2017 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT C02Eq.) ES-11
Figure ES-7: 2017 End-Use Sector Emissions of CO2 from Fossil Fuel Combustion (MMT CO2 Eq.) ES-11
Figure ES-8: Electric Power Generation (Billion kWh) and Emissions (MMT CO2 Eq.) ES-13
Figure ES-9: 2017 Sources of CH4 Emissions (MMT CO2 Eq.) ES-15
Figure ES-10: 2017 Sources of N20 Emissions (MMT CO2 Eq.) ES-16
Figure ES-11: 2017 Sources of HFCs, PFCs, SF6, and NF3 Emissions (MMT CO2 Eq.) ES-17
Figure ES-12: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2 Eq.) ES-18
Figure ES-13: 2017 U.S. Energy Consumption by Energy Source (Percent) ES-20
Figure ES-14: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.) ES-23
Figure ES-15: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed to Economic Sectors
(MMTCO2 Eq.) ES-25
Figure ES-16: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic Product (GDP)..ES-26
Figure ES-17: 2017 Key Categories (MMT CO2 Eq.)a ES-28
Figure 1-1: National Inventory Arrangements Diagram Inventory Process Inventory Process 1-12
Figure 1-2: U.S. QA/QC Plan Summary 1-22
Figure 2-1: Gross U.S. Greenhouse Gas Emissions by Gas (MMT CO2 Eq.) 2-1
Figure 2-2: Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the Previous Year 2-2
xxii Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 2-3: Cumulative Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to 1990 (1990=0, MMT
CO Eq.) 2-2
Figure 2-4: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2 Eq.) 2-7
Figure 2-5: 2017 Energy Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 2-9
Figure 2-6: 2017 U.S. Fossil Carbon Flows (MMT CO2 Eq.) 2-10
Figure 2-7: 2017 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT CO2 Eq.) 2-13
Figure 2-8: 2017 End-Use Sector Emissions of CO2 from Fossil Fuel Combustion (MMT CO2 Eq.) 2-13
Figure 2-9: Electric Power Generation (Billion kWh) and Emissions (MMT CO2 Eq.) 2-14
Figure 2-10: 2017 Industrial Processes and Product Use Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 2-16
Figure 2-11: 2017 Agriculture Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 2-18
Figure 2-12: 2017 LULUCF Chapter Greenhouse Gas Sources and Sinks (MMT CO2 Eq.) 2-20
Figure 2-13: 2017 Waste Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 2-22
Figure 2-14: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.) 2-24
Figure 2-15: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed to Economic Sectors
(MMT C02 Eq.) 2-27
Figure 2-16: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic Product 2-34
Figure 3-1: 2017 Energy Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 3-1
Figure 3-2: 2017 U.S. Fossil Carbon Flows (MMT CO Eq.) 3-2
Figure 3-3: 2017 U.S. Energy Use by Energy Source (Percent) 3-8
Figure 3-4: U.S. Energy Use (Quadrillion Btu) 3-8
Figure 3-5: 2017 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT CO2 Eq.) 3-9
Figure 3-6: Annual Deviations from Normal Heating Degree Days for the United States (1950-2017, Index Normal
= 100) 3-10
Figure 3-7: Annual Deviations from Normal Cooling Degree Days for the United States (1950-2017, Index Normal
= 100) 3-10
Figure 3-8: Fuels Used in Electric Power Generation (TBtu) and Total Electric Power Sector CO2 Emissions.... 3-16
Figure 3-9: Electric Power Retail Sales by End-Use Sector (Billion kWh) 3-16
Figure 3-10: Industrial Production Indices (Index 2012=100) 3-18
Figure 3-11: Fuels Used in Residential and Commercial Sectors (TBtu), Heating Degree Days, and Total Sector
CO2 Emissions 3-19
Figure 3-12: Fuels Used in Transportation Sector (TBtu), Onroad VMT, and Total Sector CO2 Emissions 3-21
Figure 3-13: Sales-Weighted Fuel Economy of New Passenger Cars and Light-Duty Trucks, 1990-2017
(miles/gallon) 3-22
Figure 3-14: Sales of New Passenger Cars and Light-Duty Trucks, 1990-2017 (Percent) 3-23
Figure 3-15: Mobile Source CH4 and N20 Emissions (MMT CO2 Eq.) 3-26
Figure 3-16: U.S. Energy Consumption and Energy-Related CO2 Emissions Per Capita and Per Dollar GDP 3-33
Figure 4-1: 2017 Industrial Processes and Product Use Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 4-2
Figure 4-2: U.S. HFC Consumption (MMT CO Eq.) 4-118
Figure 5-1: 2017 Agriculture Chapter Greenhouse Gas Emission Sources (MMT CO2 Eq.) 5-1
xxiii
-------
Figure 5-2: Annual CH4 Emissions from Rice Cultivation, 2012 (MMT CO2 Eq./Year) 5-19
Figure 5-3: Sources and Pathways of N that Result in N20 Emissions from Agricultural Soil Management 5-24
Figure 5-4: Crops, 2012 Annual Direct N2O Emissions Estimated Using the Tier 3 DAYCENT Model (MMT CO2
Eq./year) 5-27
Figure 5-5: Grasslands, 2012 Annual Direct N2O Emissions Estimated Using the Tier 3 DAYCENT Model (MMT
CO2 Eq./year) 5-28
Figure 5-6: Crops, 2012 Annual Indirect N2O Emissions from Volatilization Using the Tier 3 DAYCENT Model
(MMT C02 Eq./year) 5-29
Figure 5-7: Grasslands, 2012 Annual Indirect N2O Emissions from Volatilization Using the Tier 3 DAYCENT
Model (MMT CO Eq./year) 5-30
Figure 5-8: Crops, 2012 Annual Indirect N2O Emissions from Leaching and Runoff Using the Tier 3 DAYCENT
Model (MMT CO Eq./year) 5-31
Figure 5-9: Grasslands, 2012 Annual Indirect N20 Emissions from Leaching and Runoff Using the Tier 3
DAYCENT Model (MMT CO Eq./year) 5-32
Figure 5-10: Comparison of Measured Emissions at Field Sites and Modeled Emissions Using the DAYCENT
Simulation Model and IPCC Tier 1 Approach (kg N20 per ha per year) 5-41
Figure 6-1: 2017 LULUCF Chapter Greenhouse Gas Sources and Sinks (MMT CO2 Eq.) 6-4
Figure 6-2: Percent of Total Land Area for Each State in the General Land-Use Categories for 2017 6-12
Figure 6-3: Changes in Forest Area by Region for Forest Land Remaining Forest Land in the conterminous United
States and Alaska (1990-2017, Million Hectares) 6-25
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 (1990-2017, MMT C per Year) 6-28
Figure 6-5: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural Management within
States, 2012, Cropland Remaining Cropland 6-52
Figure 6-6: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural Management within
States, 2012, Cropland Remaining Cropland 6-53
Figure 6-7: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural Management within
States, 2012, Grassland Remaining Grassland 6-67
Figure 6-8: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural Management within
States, 2012, Grassland Remaining Grassland 6-67
Figure 7-1: 2017 Waste Chapter Greenhouse Gas Sources (MMT CO2 Eq.) 7-1
Figure 7-2: Management of Municipal Solid Waste in the United States, 2015 7-16
Figure 7-3: MSW Management Trends from 1990 to 2015 7-17
Figure 7-4: Percent of Degradable Materials Diverted from Landfills from 1990 to 2015 (Percent) 7-18
Box ES-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals ES-1
Box ES-2: EPA's Greenhouse Gas Reporting Program ES-2
Box ES-3: Improvements and Recalculations Relative to the Previous Inventory ES-5
Box ES-4: Use of Ambient Measurements Systems for Validation of Emission Inventories ES-14
Box ES-5: Recent Trends in Various U.S. Greenhouse Gas Emissions-Related Data ES-25
xxiv Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Box 1-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals 1-2
Box 1-2: The IPCC Fifth Assessment Report and Global Warming Potentials 1-9
Box 1-3: IPCC Reference Approach 1-16
Box 2-1: Methodology for Aggregating Emissions by Economic Sector 2-32
Box 2-2: Recent Trends in Various U.S. Greenhouse Gas Emissions-Related Data 2-33
Box 2-3: Sources and Effects of Sulfur Dioxide 2-36
Box 3-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals 3-4
Box 3-2: Energy Data from EPA's Greenhouse Gas Reporting Program 3-4
Box 3-3: Weather and Non-Fossil Energy Effects on CO2 from Fossil Fuel Combustion Trends 3-9
Box 3-4: Uses of Greenhouse Gas Reporting Program Data and Improvements in Reporting Emissions from
Industrial Sector Fossil Fuel Combustion 3-31
Box 3-5: Carbon Intensity of U.S. Energy Consumption 3-31
Box 3-6: Reporting of Lubricants, Waxes, and Asphalt and Road Oil Product Use in Energy Sector 3-50
Box 3-7: Carbon Dioxide Transport, Injection, and Geological Storage 3-79
Box 4-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals 4-6
Box 4-2: Industrial Processes Data from EPA's Greenhouse Gas Reporting Program 4-7
Box 5-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals 5-2
Box 5-2: Surrogate Data Method 5-21
Box 5-3: Tier 1 vs. Tier 3 Approach for Estimating N2O Emissions 5-33
Box 5-4: Surrogate Data Method 5-34
Box 5-5: Comparison of the Tier 2 U.S. Inventory Approach and IPCC (2006) Default Approach 5-43
Box 5-6: Comparison of Tier 2 U.S. Inventory Approach and IPCC (2006) Default Approach 5-50
Box 6-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals 6-7
Box 6-2: Preliminary Estimates of Land Use in U.S. Territories 6-21
Box 6-3: CO2 Emissions from Forest Fires 6-28
Box 6-4: Surrogate Data Method 6-54
Box 6-5: Tier 3 Approach for Soil C Stocks Compared to Tier 1 or 2 Approaches 6-55
Box 6-6: Grassland Woody Biomass Analysis 6-71
Box 7-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals 7-2
Box 7-2: Waste Data from EPA's Greenhouse Gas Reporting Program 7-2
Box 7-3: Nationwide Municipal Solid Waste Data Sources 7-15
Box 7-4: Overview of the Waste Sector 7-16
Box 7-5: Description of a Modern, Managed Landfill 7-18
xxv
-------
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 2017. 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
BoxES-1.4
Box ES-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals
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
1 Hie 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 UNEPAVMO Information Unit on Climate
Change. See .
3 Article 4(1 )(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
-------
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.
Box ES-2: EPA's Greenhouse Gas Reporting Program
On October 30, 2009, the U.S. Enviromnental Protection Agency (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). The
rule 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.5 Annual reporting is at the facility level,
except for certain suppliers of fossil fuels and industrial greenhouse gases.
EPA's GHGRP dataset and the data presented in this Inventory report 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 categories to improve the national estimates presented in this
Inventory consistent with IPCC guidance.6
ES.l Background Information
Greenhouse gases absorb infrared radiation, thereby trapping heat 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 other fluorine-containing halogenated substances. Although CO2, CH4, and N20 occur naturally
in the atmosphere, human activities have changed their atmospheric concentrations. From the pre-industrial era (i.e.,
ending about 1750) to 2017, concentrations of these greenhouse gases have increased globally by 45, 164, and 22
percent, respectively (IPCC 2013; NOAA/ESRL 2018a, 2018b, 2018c). This annual report estimates the total
national greenhouse gas emissions and removals associated with human activities across the United States.
Global Warming Potentials
Gases in the atmosphere can contribute to climate change both directly and indirectly. Direct effects occur when the
gas itself absorbs radiation. Indirect radiative forcing occurs when chemical transformations of the substance
produce other greenhouse gases, when a gas influences the atmospheric lifetimes of other gases, and/or when a gas
5 See and .
6 See .
ES-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 2014). The
reference gas used is CO2, and therefore GWP-weighted emissions can be 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
CO2
1
CH4a
25
N2O
298
HFC-23
14,800
HFC-32
675
HFC-125
3,500
HFC-134a
1,430
HFC-143a
4,470
HFC-152a
124
HFC-227ea
3,220
HFC-236fa
9,810
HFC-4310mee
1,640
CF4
7,390
C2F6
12,200
C4F10
8,860
C6Fl4
9,300
SFe
22,800
NF3
17,200
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 CO2 is not included. See Annex 6
for additional information.
Source: IPCC (2007)
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
-------
ES.2 Recent Trends in U.S. Greenhouse Gas
Emissions and Sinks
In 2017, total gross U.S. greenhouse gas emissions were 6,456.7 MMT, or million metric tons, of carbon dioxide
(CO2) Eq.11 Total U.S. emissions have increased by 1.3 percent from 1990 to 2017, and emissions decreased from
2016 to 2017 by 0.5 percent (35.5 MMT CO2 Eq.). The decrease in total greenhouse gas emissions between 2016
and 2017 was driven in part by a decrease in CO2 emissions from fossil fuel combustion. The decrease in CO2
emissions from fossil fuel combustion was a result of multiple factors, including a continued shift from coal to
natural gas and increased use of renewable energy in the electric power sector, and milder weather that contributed
to less overall electricity use.
Relative to 1990, the baseline for this Inventory, gross emissions in 2017 are higher by 1.3 percent, down from a
high of 15.7 percent above 1990 levels in 2007. Overall, net emissions in 2017 were 13.0 percent below 2005 levels
as shown in Table ES-2. Figure ES-1 through Figure ES-3 illustrate the overall trends in total U.S. emissions by gas,
annual changes, and absolute change since 1990, and Table ES-2 provides a detailed summary of gross U.S.
greenhouse gas emissions and sinks for 1990 through 2017. 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 (see Section ES.3 Overview of Sector Emissions and Trends).
Figure ES-1: Gross U.S. Greenhouse Gas Emissions by Gas (MMT CO2 Eq.)
9,000
8,000
7,000
6,000
cr
LU
O 5,000
U
I-
2
Z 4,000
3,000
2,000
1,000
0
o*HrNm^-mvoi^GooiOTHrsiro^-Lr)^0rvcoCT»o*HrMm^-mvO[v
O^C^CTIO^C^CTIO^C^CTIO^OOOOOOOOOOtHtHtHtHtHtHtHtH
CTCTOICTCTCTIO'CTICTIO'OOOOOOOOOOOOOOOOOO
THTHTHTHTHTHTHTHTHTHrMfNrslrMfNrslrMfMrslrMfMrslrMCNlrslrMCNlrsl
HFCs, PFCs, SFs and NF3 Subtotal
Nitrous Oxide
I Methane
Carbon Dioxide
^ ro - \r>
10
11 Hie gross emissions total presented in this report for the United States excludes emissions and removals from Land Use, Land-
Use Change, and Forestry (LULUCF). Lhe 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: 1990-2017
-------
Figure ES-2: Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the
Previous Year
1—irsiro^-Lft^rvoocnoiHrsiro^rLnvorvooCTJOi—irMro^rLnvor**
0"ia*CT»CT»a^O^CT*CT»CT»OOOOOOOOOOi—It-It-I i-HIt—It-It—IiH
0">CT»CT»Cr>CT»CT>CT»000000000000000000
t-h t-h t—i t—i t-h t-h t-h t—i tH rvj rsj rsj rsj r\j rM rsj rsj r\j rM rM rsj r\j rsl f"M rM rv| rM
Figure ES-3: Cumulative Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to
1990 (1990=0, MMT COz Eq.)
1,200
T-irMrO'd-miŁ>r>.ooCT»OTHrMro^-Ln^orvooCT»OT-irvin^i-Lrjvor,«.
CT»CT»O^CT>^CT»CT»CT>CT»OOOOOOOOOOt—It—It—It—ItHtHtHtH
CTia^CT»CTlC7>Cr»CTiCTlCTiOOOOOOOOOOOOOOOOOO
t—i -t—i T—i t—i t—i -t—i t—i t~i t"H rsj rM rM fM f\i rsj rM fM rsj rM rM rsj rsj rM rsi (M rsi
Box ES-3: Improvements and Recalculations Relative to the Previous Inventory
Each year, some emission and sink estimates in the Inventory are recalculated and revised with improved methods
and/or data. In general, recalculations are made to the U.S. greenhouse gas emission estimates either to incorporate
new methodologies or, most commonly, to update recent historical data. These improvements are implemented
consistently across the previous Inventory's time series (i.e., 1990 to 2016) to ensure that the trend is accurate.
Below are categories with recalculations resulting in an average change over the time series of greater than 10 MMT
CO2 Eq. For more information on specific methodological updates, please see the Energy chapter (Chapter 3), the
LULUCF chapter (Chapter 6), and the Recalculations and Improvements chapter (Chapter 9).
• Forest Land Remaining Forest Land: Changes in Forest Carbon Stocks (CO2)
• Land Converted to Cropland: Changes in all Ecosystem Carbon Stocks (CO2)
• Settlements Remaining Settlements: Changes in Settlement Tree Carbon Stocks (CO2)
Executive Summary ES-5
-------
• Land Converted to Forest Land: Changes in Forest Carbon Stocks (CO2)
• Land Converted to Settlements: Changes in Settlement Soil Carbon Stocks (CO2)
• Fossil Fuel Combustion: Changes in Stationary Combustion (N2O)
• Land Converted to Grassland: Changes in all Ecosystem Carbon Stocks (CO2)
In implementing improvements, the United States follows the 2006 LPCC 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."
In each Inventory, the results of all methodological changes and historical data updates are presented in the
Recalculations and Improvements chapter; and detailed descriptions of each recalculation including references for
data, are provided within each source or sink's description in the report. Changes in historical data are generally the
result of changes in statistical data supplied by other agencies.
Table ES-2: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
5,121.2
6,130.6
5,522.9
5,572.1
5,423.0
5,306.7
5,270.7
Fossil Fuel Combustion
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
Transportation
1,469.1
1,857.0
1,682.7
1,721.6
1,734.0
1,779.0
1,800.6
Electric Power Sector
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
Industrial
857.5
853.4
840.0
819.6
807.9
807.6
810.7
Residential
338.2
357.9
329.3
346.8
317.8
292.9
294.5
Commercial
226.5
226.8
224.6
232.9
245.5
232.1
232.9
U.S. Territories
27.6
49.7
42.5
41.4
41.4
41.4
41.4
Non-Energy Use of Fuels
119.6
139.6
123.5
119.9
126.9
113.7
123.2
Iron and Steel Production &
Metallurgical Coke Production
101.6
68.2
53.5
58.4
47.8
42.3
41.8
Cement Production
33.5
46.2
36.4
39.4
39.9
39.4
40.3
Petrochemical Production
21.2
26.8
26.4
26.5
28.1
28.1
28.2
Natural Gas Systems
30.0
22.6
25.1
25.5
25.1
25.5
26.3
Petroleum Systems
9.0
11.6
25.1
29.6
31.7
22.2
23.3
Ammonia Production
13.0
9.2
9.5
9.4
10.6
10.8
13.2
Lime Production
11.7
14.6
14.0
14.2
13.3
12.9
13.1
Incineration of Waste
8.0
12.5
10.3
10.4
10.7
10.8
10.8
Other Process Uses of Carbonates
6.3
7.6
11.5
13.0
12.2
11.0
10.1
Urea Fertilization
2.4
3.5
4.4
4.5
4.7
4.9
5.1
Urea Consumption for Non-
Agricultural Purposes
3.8
3.7
4.6
1.8
4.6
5.1
5.0
Carbon Dioxide Consumption
1.5
1.4
4.2
4.5
4.5
4.5
4.5
Liming
4.7
4.3
3.9
3.6
3.7
3.2
3.2
Ferroalloy Production
2.2
1.4
1.8
1.9
2.0
1.8
2.0
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.7
1.8
Titanium Dioxide Production
1.2
1.8
1.7
1.7
1.6
1.7
1.7
Glass Production
1.5
1.9
1.3
1.3
1.3
1.2
1.3
Aluminum Production
6.8
4.1
3.3
2.8
2.8
1.3
1.2
Phosphoric Acid Production
1.5
1.3
1.1
1.0
1.0
1.0
1.0
Zinc Production
0.6
1.0
1.4
1.0
0.9
0.9
1.0
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Silicon Carbide Production and
Consumption
0.4
0.2
0.2
0.2
0.2
0.2
0.2
ES-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Abandoned Oil and Gas Wells
Magnesium Production and Processing
Wood Biomass, Ethanol, and Biodiesel
Consumption"
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
Petrochemical Production
Field Burning of Agricultural Residues
Ferroalloy Production
Silicon 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
Nitric Acid Production
Adipic Acid Production
Wastewater Treatment
N2O from Product Uses
Composting
Caprolactam, Glyoxal, and Glyoxylic
Acid Production
Incineration of Waste
Semiconductor Manufacture
Field Burning of Agricultural Residues
Petroleum Systems
Natural Gas Systems
International Bunker Fuelsb
HFCs
Substitution of Ozone Depleting
Substances'1
HCFC-22 Production
Semiconductor Manufacture
Magnesium Production and Processing
PFCs
Semiconductor Manufacture
+
+
219.4
+
+
230.7
+
+
315.5
103.5
113.1
99.8
779.8
691.4
663.0
164.2
168.9
165.5
193.1
171.4
165.6
179.6
131.4
112.9
37.1
53.7
58.1
96.5
64.1
64.6
42.1
36.7
41.6
15.3
15.4
14.3
16.0
16.7
11.5
8.6
7.8
8.7
6.6
6.9
7.0
7.2
6.6
6.2
12.9
9.6
4.5
0.4
1.9
2.0
0.2
0.1
0.1
0.1
+
+
0.2
+
+
0.2
+
+
+
+
0.2
+
+
0.1
+
+
0.1
370.3
375.8
365.4
251.7
254.5
265.2
25.1
34.3
32.7
14.0
16.5
17.4
42.0
39.0
22.1
12.1
11.3
10.7
15.2
7.1
3.9
3.4
4.4
4.7
4.2
4.2
4.2
0.3
1.7
1.8
1.7
2.1
2.0
0.5
0.4
0.3
+
0.1
0.2
+
+
0.1
+
0.1
+
+
0.9
+
1.0
+
0.9
46.6
122.3
146.1
0.3
102.1
141.7
46.1
20.0
4.1
0.2
0.2
0.3
0.0
0.0
0.1
24.3
6.7
5.9
2.8
3.2
2.9
323.2 317.7 317.2 322.2
103.4
110.9
116.6
120.1
662.1
661.4
654.9
656.3
164.2
166.5
171.9
175.4
165.1
167.2
165.7
165.6
112.5
111.2
108.0
107.7
57.8
60.9
61.5
61.7
64.6
61.2
53.8
55.7
42.1
39.5
38.2
37.7
14.3
14.5
14.2
14.2
12.7
12.3
13.7
11.3
8.9
8.5
7.9
7.8
7.1
7.1
7.2
6.9
6.3
6.4
6.7
6.4
4.1
3.6
3.4
3.2
2.1
2.1
2.1
2.2
0.1
0.2
0.2
0.3
0.2
+
+
0.2
+
+
0.2
+
+
0.2
+
+
+
+
0.1
+
+
0.1
+
+
0.1
+
+
0.1
362.7
374.1
364.5
360.5
262.3
277.8
267.6
266.4
33.0
30.6
30.1
28.6
17.4
17.6
18.2
18.7
20.2
18.8
17.9
16.9
10.9
11.6
10.1
9.3
5.4
4.3
7.0
7.4
4.8
4.8
4.9
5.0
4.2
4.2
4.2
4.2
1.9
1.9
1.9
1.9
2.0
2.0
2.0
1.4
0.3
0.3
0.3
0.3
0.2
0.2
0.2
0.2
0.1
+
0.1
+
0.1
+
0.1
+
+
0.9
+
0.9
+
1.0
+
1.0
150.7
153.8
155.0
158.3
145.2
149.2
151.7
152.7
5.0
4.3
2.8
5.2
0.3
0.3
0.3
0.4
0.1
0.1
0.1
0.1
5.6
5.1
4.4
4.1
3.1
3.1
3.0
3.0
Executive Summary ES-7
-------
Aluminum Production
21.5
3.4
3.0
2.5
2.0
1.4
1.1
Substitution of Ozone Depleting
Substances
0.0
+
+
+
+
+
+
SF«
28.8
11.8
6.3
6.3
5.8
6.3
6.1
Electrical Transmission and
Distribution
23.1
8.3
4.4
4.6
4.1
4.4
4.3
Magnesium Production and Processing
5.2
2.7
1.3
0.9
1.0
1.1
1.1
Semiconductor Manufacture
0.5
0.7
0.7
0.7
0.7
0.9
0.7
NF3
+
0.5
0.5
0.5
0.6
0.6
0.6
Semiconductor Manufacture
+
0.5
0.5
0.5
0.6
0.6
0.6
Total Emissions
6,371.0
7,339.0
6,710.2
6,760.0
6,623.8
6,492.3
6,456.7
LULUCF Emissions0
7.8
16.0
17.5
17.7
28.3
15.5
15.5
LULUCF CH4 Emissions
5.0
9.0
9.9
10.1
16.5
8.8
8.8
LULUCF N2O Emissions
2.8
7.0
7.6
7.7
11.8
6.7
6.7
LULUCF Carbon Stock Change6
(814.8)
(756.1)
(731.0)
(687.8)
(739.4)
(738.1)
(729.6)
LULUCF Sector Net Total'
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
Net Emissions (Sources and Sinks)
5,564.0
6,599.0
5,996.8
6,090.0
5,912.7
5,769.7
5,742.6
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 CO2 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 CFL and N2O are reported separately from gross emissions totals. LULUCF emissions include the
CFL, and N2O emissions from Peatlands Remaining Peatlands\ CFL and N2O emissions reported for N011-CO2 Emissions
from Forest Fires, N011-CO2 Emissions from Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CFL
emissions from Land Converted to Coastal Wetlands; andN20 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.
f The LULUCF Sector Net Total is the net sum of all CFLi and N2O emissions to the atmosphere plus net carbon stock
changes.
Figure ES-4 illustrates the relative contribution of the direct greenhouse gases to total U.S. emissions in 2017,
weighted by global wanning potential. The primary greenhouse gas emitted by human activities in the United States
was CO2, representing approximately 81.6 percent of total greenhouse gas emissions. The largest source of CO2, and
of overall greenhouse gas emissions, was fossil fuel combustion. Methane emissions, which have decreased by 15.8
percent since 1990, resulted primarily from 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 source fuel combustion were the major sources of N20 emissions. Ozone depleting
substance substitute emissions and emissions of HFC-23 during the production of HCFC-22 were the primary
contributors to aggregate hydrofluorocarbon (HFC) emissions. Perfluorocarbon (PFC) emissions resulted from
semiconductor manufacturing and as a byproduct of primary aluminum production electrical transmission and
distribution systems accounted for most sulfur hexafluoride (SF6) emissions, and semiconductor manufacturing is
the only source of nitrogen trifluoride (NF3) emissions.
ES-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure ES-4: 2017 U.S. Greenhouse Gas Emissions by Gas (Percentages based on MMT CO2
Eq.)
2.6%
HFCs, PFCs, SFe and NF3 Subtotal
10.2%
ChU
81.6%
COz
Overall, from 1990 to 2017, total emissions of CO2 increased by 149.6 MMT CO2 Eq. (2.9 percent), while total
emissions of CH4 decreased by 123.5 MMT CO: Eq. (15.8 percent), and N:0 emissions decreased by 9.8 MMT CO2
Eq. (2.6 percent). During the same period, aggregate weighted emissions of HFCs, PFCs, SF6 and NF3 rose by 69.5
MMT CO2 Eq. (69.7 percent). From 1990 to 2017, HFCs increased by 111.7 MMT CO2 Eq. (239.9 percent), PFCs
decreased by 20.1 MMT CO2 Eq. (82.9 percent), SF6 decreased by 22.7 MMT CO2 Eq. (78.8 percent), and NF3
increased by 0.6 MMT CO2 Eq. (1,166 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 wanning 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
11.3 percent of total emissions in 2017. 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 45 percent (IPCC 2013; NOAA/ESRL 2018a), principally due to the combustion of fossil fuels for
energy. Globally, approximately 32,310 MMT of CO2 were added to the atmosphere through the combustion of
fossil fuels in 2016, of which the United States accounted for approximately 15 percent.13
12 Hie term "flux" is used to describe the net emissions of greenhouse gases accounting for both the emissions of CO2 to and the
removals of CO2 from the atmosphere. Removal of CO2 from the atmosphere is also referred to as "carbon sequestration."
13 Global CO2 emissions from fossil fuel combustion were taken from International Energy Agency CO: Emissions from Fossil
Fuels Combustion Overview (IEA 2018). The publication
has not yet been updated to include 2017 data.
Executive Summary ES-9
-------
Within the United States, fossil fuel combustion accounted for 93.2 percent of CO2 emissions in 2017. 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: 2017 Sources of CO2 Emissions (MMT CO2 Eq.)
Fossil Fuel Combustion
Non-Energy Use of Fuels
Iron and Steel Prod. & Metallurgical Coke Prod.
Cement Production
Petrochemical Production
Natural Gas Systems
Petroleum Systems
Ammonia Production
Lime Production
Incineration of Waste
Other Process Uses of Carbonates
Urea Fertilization
Urea Consumption for Non-Agricultural Purposes
Carbon Dioxide Consumption
Liming
Ferroalloy Production
Soda Ash Production
Titanium Dioxide Production
Glass Production
Aluminum Production
Phosphoric Acid Production
Zinc Production
Lead Production
Silicon Carbide Production and Consumption
Abandoned Oil and Gas Wells
Magnesium Production and Processing
4,912
CO2 as a Portion of All
Emissions
< 0.5
< 0.5
< 0.5
< 0.5
25 50 75 100
MMT CO2 Eq.
125
150
As the largest source of U.S. greenhouse gas emissions, CO2 from fossil fuel combustion has accounted for
approximately 77 percent of GWP-weighted emissions since 1990. 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.
Between 1990 and 2017, CO2 emissions from fossil fuel combustion increased from 4,738.8 MMT CO2 Eq. to
4,912.0 MMT CO2 Eq., a 3.7 percent total increase over the twenty-eight-year period. Conversely, CO2 emissions
from fossil fuel combustion decreased by 832.8 MMT CO2 Eq. from 2005 levels, a decrease of approximately 14.5
percent between 2005 and 2017. From 2016 to 2017, these emissions decreased by 49.9 MMT CO2 Eq. (1.0
percent).
Historically, changes in emissions from fossil fuel combustion have been the dominant factor affecting U.S.
emission trends. Changes in CO2 emissions from fossil fuel combustion are influenced by many long-term and
short-term factors. Long-term factors include population and economic trends, technological changes, 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"
ES-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
sectors. For the discussion 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. Note that
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. Figure ES-6, Figure ES-7, and Table ES-3 summarize CO2
emissions from fossil fuel combustion by end-use sector.
Figure ES-6: 2017 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT
COz Eq.)
Relative Contribution by Fuel Type
2,500
2,000
1,732
Petroleum
Coal
Natural Gas
29.5%
44.7%
" 1,500
25.8%
U.S. Territories
Commercial
Residential
Industrial
Electric Power Transportation
Note on Figure ES-6: Fossil Fuel Combustion for electric power also includes emissions of less than 0.5 MMT CO2 Eq. from
geothermal-based generation.
Figure ES-7: 2017 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
Executive Summary ES-11
-------
Table ES-3: CO2 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.)
End-Use Sector
1990
2005
2013
2014
2015
2016
2017
Transportation
1,472.1
1,861.7
1,686.9
1,726.0
1,738.2
1,783.2
1,804.9
Combustion
1,469.1
1,857.0
1,682.7
1,721.6
1,734.0
1,779.0
1,800.6
Electricity
3.0
4.7
4.3
4.5
4.3
4.2
4.3
Industrial
1,543.9
1,589.7
1,434.8
1,412.5
1,357.4
1,325.2
1,315.1
Combustion
857.5
853.4
840.0
819.6
807.9
807.6
810.7
Electricity
686.4
736.3
594.8
593.0
549.5
517.6
504.4
Residential
930.9
1,213.9
1,064.1
1,080.9
1,001.6
946.3
911.5
Combustion
338.2
357.9
329.3
346.8
317.8
292.9
294.5
Electricity
592.7
856.0
734.7
734.1
683.8
653.5
617.1
Commercial
764.3
1,029.7
929.1
938.5
908.5
865.8
839.1
Combustion
226.5
226.8
224.6
232.9
245.5
232.1
232.9
Electricity
537.7
803.0
704.5
705.6
663.0
633.6
606.2
U.S. Territories3
27.6
49.7
42.5
41.4
41.4
41.4
41.4
Total
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
Electric Power
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
aFuel 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.
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.
Transportation End-Use Sector. When electricity-related emissions are distributed to economic end-use sectors,
transportation activities accounted for 36.7 percent of U.S. CO2 emissions from fossil fuel combustion in 2017. The
largest sources of transportation CO2 emissions in 2017 were passenger cars (41.4 percent); medium- and heavy-
duty trucks (23.7 percent); light-duty trucks, which include sport utility vehicles, pickup trucks, and minivans (16.8
percent); commercial aircraft (7.1 percent); other aircraft (2.5 percent); rail (2.3 percent); pipelines (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, CH4, N20, and HFCs.
In terms of the overall trend, from 1990 to 2017, 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.,
passenger cars and light-duty trucks) increased 46 percent from 1990 to 2017,14 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. 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.
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 27 percent of CO2 emissions from
fossil fuel combustion in 2017. Approximately 62 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 14.8 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.
14 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 inFHWA'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 2017
time period. In absence of these method changes, light-duty VMT growth between 2006 and 2007 would likely have been higher.
ES-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Residential and Commercial End-Use Sectors. The residential and commercial end-use sectors accounted for 19 and
17 percent, respectively, of CO2 emissions from fossil fuel combustion in 2017. The residential and commercial
sectors relied heavily on electricity for meeting energy demands, with 68 and 72 percent, respectively, of their
emissions attributable to electricity use for lighting, heating, cooling, and operating appliances. The remaining
emissions were due to the consumption of natural gas and petroleum for heating and cooking. Total direct and
indirect emissions from the residential sector have decreased by 2 percent since 1990. Total direct and indirect
emissions from the commercial sector have increased by 10 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 2017. 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 2017.16 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 31 percent in 2017.17 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
28-year period to represent 31 percent of electric power sector generation in 2017.
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 emissions from the electric power sector have decreased by
approximately 4.8 percent since 1990, the carbon intensity of the electric power sector, in terms of CO2 Eq. per
QBtu input, has significantly decreased-by 11 percent-during that same time-frame. This trend away from a direct
relationship between the level of electric power generation and the resulting emissions is shown in Figure ES-8.
Figure ES-8: Electric Power Generation (Billion kWh) and Emissions (MMT CO2 Eq.)
5,000
§ 4,000
Nuclear Generation (Billion kWh)
Renewable Generation (Billion kWh)
Petroleum Generation (Billion kWh)
Natural Gas Generation (Billion kWh)
Coal Generation (Billion kWh)
I Total Emissions (MMT CO2 Eq.) [Right Axis]
3,000
u 2,000
o-i-irMPo^-LnuarvoocTi
0^0~i0^0~i0^0~i0^0~iCT10~i
O"} cti O"} cti O"} cti o"i cn CTi cn
N(NN(M(N(N(N(MfMMrMrMM(N(NN(N
15 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 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 (2019).
17 Values represent electricity net generation from the electric power sector. See Table 7.2b Electricity Net Generation: Electric
Power Sector ofEIA (2019).
Executive Summary ES-13
-------
Other significant CO2 trends included the following:
• Carbon dioxide emissions from non-energy use of fossil fuels increased by 3.7 MMT CO2 Eq. (3.1 percent)
from 1990 through 2017. Emissions from non-energy uses of fossil fuels were 123.2 MMT CO2 Eq. in
2017, which constituted 2.3 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 59.8 MMT CO2 Eq. (58.9 percent) from 1990 through 2017, 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 10.5
percent between 1990 and 2017. 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.
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.18 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 lias 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.19 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.20 This gridded inventory is responsive to 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).
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 CH4 in the atmosphere increased by 164 percent (IPCC 2013; NOAA/ESRL 2018b).
Anthropogenic sources of CH4 include natural gas and petroleum systems, agricultural activities, LULUCF, landfills
and other waste management activities, coal mining, wastewater treatment, stationary and mobile combustion, and
certain industrial processes (see Figure ES-9).
18 See .
19 See .
20 See .
ES-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure ES-9: 2017 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
Petrochemical Production
Field Burning of Agricultural Residues
Ferroalloy Production
Silicon Carbide Production and Consumption
Iron and Steel Production & Metallurgical Coke Production
Incineration of Waste
175
¦
¦
¦
I
I
I
< 0.5
< 0.5
< 0.5
< 0.5
< 0.5
< 0.5
ChU as a Portion of All
Emissions
10.2%
20 40 60 80 100 120
MMT CO2 Eq.
140 160 180
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 is the largest anthropogenic source of CH4 emissions in the United States. In 2017,
enteric fermentation CH4 emissions were 175.4 MMT CO2 Eq. (26.7 percent of total CH4 emissions),
which represents an increase of 11.3 MMT CO2 Eq. (6.9 percent) since 1990. This increase in emissions
from 1990 to 2017 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 2017 with 165.6 MMT CO2 Eq. of CH4 emitted into the atmosphere. Those emissions have
decreased by 27.5 MMT CO2 Eq. (14.2 percent) since 1990. The decrease in CH4 emissions is largely due
to decreases in emissions from distribution and 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 (107.7 MMT
CO2 Eq.), accounting for 16.4 percent of total CH4 emissions in 2017. From 1990 to 2017, CH4 emissions
from landfills decreased by 71.8 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.21 While the amount of landfill gas collected and
combusted continues to increase, the rate of increase in collection and combustion no longer exceeds the
rate of additional CH4 generation from the amount of organic MSW landfilled as the U.S. population grows
(EPA 2018b).
21 Carbon dioxide emissions from landfills are not included specifically in summing waste sector totals. Net carbon fluxes from
changes in biogenic carbon reservoirs and decay of disposed wood products are accounted for in the estimates for LULUCF.
Executive Summary ES-15
-------
Nitrous Oxide Emissions
Nitrous oxide (N20) is produced by biological processes that occur in soil and water and by a variety of
anthropogenic activities in the agricultural, energy, industrial, and waste management fields. While total N20
emissions are much lower than CO2 emissions, N20 is nearly 300 times more powerful than CO2 at trapping heat in
the atmosphere (IPCC 2007). Since 1750, the global atmospheric concentration of N20 lias risen by approximately
22 percent (IPCC 2013; NOAA/ESRL 2018c). 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 nitric acid production (see Figure ES-10).
Figure ES-10: 2017 Sources of N2O Emissions (MMT CO2 Eq.)
Agricultural Soli Management
Stationary Combustion
Manure Management
Mobile Combustion
Nitric Acid Production
Adiplc Acid Production
Wastewater Treatment
N2O from Product Uses
Composting
Caprolactam, Glyoxal, and Glyoxyllc Acid Production
Incineration of Waste
Semiconductor Manufacture
Field Burning of Agricultural Residues
Petroleum Systems
Natural Gas Systems
N2O as a Portion of All
Emissions
< 0.5
266
15 20 25
MMT CO2 Eq.
30
35
40
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 forthe largest sources of U.S. emissions of N20 include the following:
• Agricultural soils accounted for approximately 73.9 percent of N20 emissions and 4.1 percent of total
greenhouse gas emissions in the United States in 2017. Estimated emissions from this source in 2017 were
266.4 MMT CO2 Eq. Annual N20 emissions from agricultural soils fluctuated between 1990 and 2017,
although overall emissions were 5.8 percent higher in 2017 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 3.5 MMT CO2 Eq. (14.1 percent) from 1990
to 2017. 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 25.1 MMT CO2 Eq. (59.7 percent) from
1990 to 2017, primarily as a result of N20 national emission control standards and emission control
technologies for on-road vehicles.
HFC, PFC, SF6/ and NF3 Emissions
Hydrofluorocarbons (HFCs) and perfluorocarbons (PFCs) are families of 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 and PFCs do not deplete the stratospheric ozone layer,
and are therefore acceptable alternatives under the Montreal Protocol on Substances that Deplete the Ozone Layer.
ES-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
These compounds, however, along with SF6 and NF3, are potent greenhouse gases. In addition to having high global
wanning potentials, SF6 and PFCs have extremely long atmospheric lifetimes, resulting in their essentially
irreversible accumulation in the atmosphere once emitted. Sulfur hexafluoride is the most potent greenhouse gas the
IPCC lias evaluated (IPCC 2013).
Other emissive sources of these gases include HCFC-22 production semiconductor manufacturing, electrical
transmission and distribution systems, magnesium production and processing, and aluminum production (see Figure
ES-11).
Figure ES-11: 2017 Sources of HFCs, PFCs, SFe, and NF3 Emissions (MMT CO2 Eq.)
Substitution of Ozone Depleting Substances 153
HCFC-22 Production
Semiconductor Manufacture
Electrical Transmission and Distribution
Magnesium Production and Processing
Aluminum Production
HFCs, PFCs, SFe, and NF3 as
a Portion of All Emissions
8 10 12
MMT CO2 Eq.
14
16
18 20
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 152.7
MMT CO2 Eq. in 2017. 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 interim substitutes in many applications,
are themselves phased out under the provisions of the Copenhagen Amendments to the Montreal Protocol.
• Emissions from HCFC-22 production were 5.2 MMT CO2 Eq. in 2017, an 89 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) as a result of HFC-23 recovery and other manufacture
processing changes.
• GWP-weighted PFC, HFC, SF6, and NF3 emissions from semiconductor manufacturing have increased by
32.1 percent from 1990 to 2017, due to competing factors 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 4.7 MMT CO2 Eq. in 2017 (a 47.9 percent decrease relative to 1999).
• Sulfur hexafluoride emissions from electric power transmission and distribution systems decreased by 81.4
percent (18.8 MMT CO2 Eq.) from 1990 to 2017. 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 enviromnental
impact of SF6 emissions through programs such as EPA's SF6 Emission Reduction Partnership for Electric
Power Systems.
Executive Summary ES-17
-------
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-eight-year period of 1990 to 2017, total emissions from the Energy, Industrial Processes
and Product Use, and Agriculture sectors grew by 85.1 MMT CO2 Eq. (1.6 percent), 16.8 MMT CO2 Eq. (4.9
percent), and 51.8 MMT CO2 Eq. (10.6 percent), respectively. Emissions from the Waste sector decreased by 68.0
MMT CO2 Eq. (34.2 percent). Over the same period, total C sequestration in the LULUCF sector decreased by 85.2
MMT CO2 (10.5 percent decrease in total C sequestration), and CH4 and N20 emissions from the LULUCF sector
increased by 7.7 MMT CO2 Eq. (99.1 percent).
Figure ES-12:
Eq.)
U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2
7,000
6,000
5,000
CT
LU
4,000
O
u
H
3,000
Z
z
2,000
1,000
-1,000
Industrial Processes and Product Use
LULUCF (emissions)
Agriculture
Energy
Land Use, Land-Use Change and Forestry (LULUCF) (removals)
o
CT>
CTi
1-1 r\i
cr> o>
cr> a>
cr>
cr>
oi
CT»
Ol
rsl
o
o
o
o
CM fNj PM fM
in
o
o
CM PsJ f\l fN
IV
1—1
o
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
2013
2014
2015
2016
2017
Energy
5,339.8
6,308.0
5,695.0
5,736.4
5,584.7
5,465.3
5,424.8
Fossil Fuel Combustion
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
Natural Gas Systems
223.1
194.0
190.8
190.6
192.2
191.2
191.9
Non-Energy Use of Fuels
119.6
139.6
123.5
119.9
126.9
113.7
123.2
Petroleum Systems
51.0
48.3
66.8
71.7
71.2
60.4
61.0
Coal Mining
96.5
64.1
64.6
64.6
61.2
53.8
55.7
Stationary Combustion
33.7
42.2
41.5
41.9
39.0
38.0
36.4
Mobile Combustion
55.0
48.6
26.6
24.3
22.4
21.2
20.1
Incineration of Waste
8.4
12.9
10.6
10.7
11.1
11.1
11.1
Abandoned Oil and Gas Wells
6.6
6.9
7.0
7.1
7.1
7.2
6.9
Abandoned Underground Coal Mines
7.2
6.6
6.2
6.3
6.4
6.7
6.4
Industrial Processes and Product Use
342.1
358.0
353.1
365.2
360.8
354.6
358.9
Substitution of Ozone Depleting
Substances
0.3
102.1
141.7
145.3
149.2
151.8
152.7
Iron and Steel Production &
Metallurgical Coke Production
101.7
68.2
53.5
58.4
47.8
42.3
41.8
Cement Production
33.5
46.2
36.4
39.4
39.9
39.4
40.3
ES-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Petrochemical Production
21.4
26.9
26.5
26.6
28.2
28.4
28.5
Ammonia Production
13.0
9.2
9.5
9.4
10.6
10.8
13.2
Lime Production
11.7
14.6
14.0
14.2
13.3
12.9
13.1
Other Process Uses of Carbonates
6.3
7.6
11.5
13.0
12.2
11.0
10.1
Nitric Acid Production
12.1
11.3
10.7
10.9
11.6
10.1
9.3
Adipic Acid Production
15.2
7.1
3.9
5.4
4.3
7.0
7.4
HCFC-22 Production
46.1
20.0
4.1
5.0
4.3
2.8
5.2
Urea Consumption for Non-
Agricultural Purposes
3.8
3.7
4.6
1.8
4.6
5.1
5.0
Semiconductor Manufacture
3.6
4.7
4.6
4.8
4.9
5.0
5.0
Carbon Dioxide Consumption
1.5
1.4
4.2
4.5
4.5
4.5
4.5
Electrical Transmission and
Distribution
23.1
8.3
4.4
4.6
4.1
4.4
4.3
N2O from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
Aluminum Production
28.3
7.6
6.2
5.4
4.8
2.7
2.3
Ferroalloy Production
2.2
1.4
1.8
1.9
2.0
1.8
2.0
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.7
1.8
Titanium Dioxide Production
1.2
1.8
1.7
1.7
1.6
1.7
1.7
Caprolactam, Glyoxal, and Glyoxylic
Acid Production
1.7
2.1
2.0
2.0
2.0
2.0
1.4
Glass Production
1.5
1.9
1.3
1.3
1.3
1.2
1.3
Magnesium Production and Processing
5.2
2.7
1.4
1.0
1.1
1.2
1.2
Phosphoric Acid Production
1.5
1.3
1.1
1.0
1.0
1.0
1.0
Zinc Production
0.6
1.0
1.4
1.0
0.9
0.9
1.0
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Silicon Carbide Production and
Consumption
0.4
0.2
0.2
0.2
0.2
0.2
0.2
Agriculture
490.2
518.4
526.3
522.8
543.8
541.2
542.1
Agricultural Soil Management
251.7
254.5
265.2
262.3
277.8
267.6
266.4
Enteric Fermentation
164.2
168.9
165.5
164.2
166.5
171.9
175.4
Manure Management
51.1
70.2
75.5
75.2
78.5
79.7
80.4
Rice Cultivation
16.0
16.7
11.5
12.7
12.3
13.7
11.3
Urea Fertilization
2.4
3.5
4.4
4.5
4.7
4.9
5.1
Liming
4.7
4.3
3.9
3.6
3.7
3.2
3.2
Field Burning of Agricultural Residues
0.2
0.3
0.3
0.3
0.3
0.3
0.3
Waste
198.9
154.7
135.8
135.6
134.5
131.1
131.0
Landfills
179.6
131.4
112.9
112.5
111.2
108.0
107.7
Wastewater Treatment
18.7
19.8
19.0
19.1
19.3
19.1
19.2
Composting
0.7
3.5
3.9
4.0
4.0
4.0
4.1
Total Emissions3
6,371.0
7,339.0
6,710.2
6,760.0
6,623.8
6,492.3
6,456.7
Land Use, Land-Use Change, and
Forestry
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
Forest land
(788.0)
(750.2)
(726.4)
(678.6)
(744.4)
(741.0)
(733.1)
Cropland
34.6
40.1
55.6
54.7
60.4
57.4
56.6
Grassland
4.7
11.3
4.9
1.2
20.0
7.5
8.9
Wetlands
(0.5)
(2.0)
(0.7)
(0.6)
(0.7)
(0.7)
(0.7)
Settlements
(57.8)
(39.2)
(46.9)
(46.7)
(46.4)
(45.8)
(45.9)
Net Emission (Sources and Sinks)b
5,564.0
6,599.0
5,996.8
6,090.0
5,912.7
5,769.7
5,742.6
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 Total 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 2017.
Executive Summary ES-19
-------
In 2017, 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 N20 emissions (43 percent and 13 percent of total U.S.
emissions of each gas, respectively). Overall, emission sources in the Energy chapter account for a combined 84.0
percent of total U.S. greenhouse gas emissions in 2017.
Figure ES-13: 2017 U.S. Energy Consumption by Energy Source (Percent)
Nuclear Electric Power
8.6%
Renewable Energy
11.4%
Petroleum
37.0%
Coal
14.3%
Natural Gas
28.6%
Industrial Processes and Product Use
The Industrial Processes and Product Use 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, CH4, N20, 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. These industries include semiconductor
manufacture, electric power transmission and distribution, and magnesium metal production and processing. In
addition N20 is used in and emitted by semiconductor manufacturing and anesthetic and aerosol applications, and
CO2 is consumed and emitted through various end-use applications. Overall, emission sources in the Industrial
Process and Product Use chapter account for 5.6 percent of U.S. greenhouse gas emissions in 2017.
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, CH4 and N20 fluxes, which are
addressed in the Land Use, Land-Use Change, and Forestry chapter). Agricultural activities contribute directly to
emissions of greenhouse gases through a variety of processes, including the following source categories: enteric
fermentation in domestic livestock, livestock manure management, rice cultivation agricultural soil management,
liming, urea fertilization and field burning of agricultural residues.
In 2017, agricultural activities were responsible for emissions of 542.1 MMT CO2 Eq„ or 8.4 percent of total U.S.
greenhouse gas emissions. Methane, N20, and CO2 were the primary greenhouse gases emitted by agricultural
ES-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
activities. Methane emissions from enteric fermentation and manure management represented approximately 26.7
percent and 9.4 percent of total CH4 emissions from anthropogenic activities, respectively, in 2017. Agricultural soil
management activities, such as application of synthetic and organic fertilizers, deposition of livestock manure, and
growing N-fixing plants, were the largest source of U.S. N20 emissions in 2017, accounting for 73.9 percent of total
N20 emissions. Carbon dioxide emissions from the application of crushed limestone and dolomite (i.e., soil liming)
and urea fertilization represented 0.2 percent of total CO2 emissions from anthropogenic activities.
Land Use, Land-Use Change, and Forestry
The LULUCF chapter contains emissions of CH4 and N20, 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.22 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 2017 resulted in a net increase in C stocks (i.e., net CO2 removals) of 729.6 MMT CO2 Eq.
(Table ES-5).23 This represents an offset of 11.3 percent of total (i.e., gross) greenhouse gas emissions in 2017.
Emissions of CH4 and N2O from LULUCF activities in 2017 were 15.5 MMT CO2 Eq. and represent 0.2 percent of
total greenhouse gas emissions.24 Between 1990 and 2017, total C sequestration in the LULUCF sector decreased by
10.5 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 2017, totaling 4.9 MMT CO2 Eq. (194 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). Peatlands Remaining
Peatlands, Land Converted to Wetlands, and Drained Organic Soils resulted in CH4 emissions of less than 0.05
MMT CO2 Eq. each.
Forest fires were also the largest source of N20 emissions from LULUCF in 2017, totaling 3.2 MMT CO2 Eq. (11 kt
of N20). Nitrous oxide emissions from fertilizer application to settlement soils in 2017 totaled to 2.5 MMT CO2 Eq.
(8 kt of N20). Additionally, the application of synthetic fertilizers to forest soils in 2017 resulted in N20 emissions
of 0.5 MMT CO2 Eq. (2 kt of N20). Grassland fires resulted in N20 emissions of 0.3 MMT CO2 Eq. (1 kt of N20).
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 C02 Eq.
Carbon dioxide removals from C stock changes are presented in Table ES-5 along with CH4 and N20 emissions for
LULUCF source categories.
22 See .
23 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.
24 LULUCF emissions include the CH4 and N2O 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 N2O emissions from Forest Soils and Settlement Soils.
Executive Summary ES-21
-------
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
2013
2014
2015
2016
2017
Carbon Stock Change3
(814.8)
(756.1)
(731.0)
(687.8)
(739.4)
(738.1)
(729.6)
Forest Land Remaining Forest Land
(671.6)
(639.4)
(616.7)
(568.8)
(645.2)
(628.9)
(621.1)
Land Converted to Forest Land
(119.1)
(120.0)
(120.5)
(120.5)
(120.6)
(120.6)
(120.6)
Cropland Remaining Cropland
(40.9)
(26.5)
(11.4)
(12.0)
(6.3)
(9.9)
(10.3)
Land Converted to Cropland
75.6
66.7
66.9
66.7
66.7
67.3
66.9
Grassland Remaining Grassland
(4.2)
5.5
(3.7)
(7.5)
9.6
(1.6)
(0.1)
Land Converted to Grassland
8.7
5.1
8.3
7.9
9.8
8.5
8.3
Wetlands Remaining Wetlands
(4.0)
(5.7)
(4.3)
(4.3)
(4.4)
(4.4)
(4.4)
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Settlements Remaining Settlements
(122.1)
(127.8)
(135.9)
(135.8)
(135.4)
(134.7)
(134.5)
Land Converted to Settlements
62.9
86.0
86.4
86.5
86.5
86.4
86.2
CH4
5.0
9.0
9.9
10.1
16.5
8.8
8.8
Forest Land Remaining Forest Land:
Forest Firesb
1.5
5.2
6.1
6.1
12.6
4.9
4.9
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
0.2
0.4
0.3
0.3
0.3
Land Converted to Wetlands: Land
Converted to Coastal Wetlands
+
+
+
+
+
+
+
Forest Land Remaining Forest Land:
Drained Organic Soils'1
+
+
+
+
+
+
+
Wetlands Remaining Wetlands:
Peatlands Remaining Peatlands
+
+
+
+
+
+
+
N2O
2.8
7.0
7.6
7.7
11.8
6.7
6.7
Forest Land Remaining Forest Land:
Forest Firesb
1.0
3.4
4.0
4.0
8.3
3.2
3.2
Settlements Remaining Settlements:
Settlement Soils6
1.4
2.5
2.6
2.6
2.5
2.5
2.5
Forest Land Remaining Forest Land:
Forest Soilsf
0.1
0.5
0.5
0.5
0.5
0.5
0.5
Grassland Remaining Grassland:
Grassland Firesc
0.1
0.3
0.2
0.4
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 Soils'1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Wetlands Remaining Wetlands:
Peatlands Remaining Peatlands
+
+
+
+
+
+
+
LULUCF Emissions8
7.8
16.0
17.5
17.7
28.3
15.5
15.5
LULUCF Carbon Stock Change3
(814.8)
(756.1)
(731.0)
(687.8)
(739.4)
(738.1)
(729.6)
LULUCF Sector Net Total"
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
+ Absolute value does not exceed 0.05 MMT CO2 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.
g LULUCF emissions include the CH4 and N2O emissions reported for Peatlands Remaining Peatlands, Forest Fires,
Drained Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CFL emissions from Land
ES-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Converted to Coastal Wetlands; andN^O emissions from Forest Soils and Settlement Soils.
h The LULUCF Sector Net Total is the net sum of all CH4 and N2O emissions to the atmosphere plus net carbon stock
changes.
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
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, accounting for 82.3 percent of total greenhouse gas emissions from waste management
activities, and 16.4 percent of total U.S. CH4 emissions.25 Additionally, wastewater treatment accounts for 14.6
percent of total Waste sector greenhouse gas emissions, 2.2 percent of U.S. CH4 emissions, and 1.4 percent of U.S.
N2O emissions. Emissions of CH4 and N20 from composting are also accounted for in this chapter, generating
emissions of 2.2 MMT CO2 Eq. and 1.9 MMT CO2 Eq„ respectively. Overall, emission sources accounted for in the
Waste chapter generated 2.0 percent of total U.S. greenhouse gas emissions in 2017.
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.
Figure ES-14 shows the trend in emissions by economic sector from 1990 to 2017, 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.)
2,500
Electric Power Industry (Purple)
2,000
Transportation (Green)
LLI
« 1,500
O
U
Industry
1,000
Agriculture
Commercial (Orange)
500
Residential (Blue)
0
01
01
O
o o
o o
cm ro
o o
o o
25 Landfills also store carbon, due to incomplete degradation of organic materials such as harvest wood products, yard
trimmings, and food scraps, as described in the Land-Use, Land-Use Change, and Forestry chapter of the Inventory report.
Executive Summary ES-23
-------
Table ES-6: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.)
Economic Sectors
1990
2005H
2013
2014
2015
2016
2017
Transportation
1,527.1
1,976.oB
1,765.4
1,799.9
1,809.3
1,849.7
1,866.2
Electric Power Industry
1,875.5
2,455.9
2,088.7
2,088.9
1,949.5
1,857.2
1,778.3
Industry
1,628.6
1,508.4B
1,469.5
1,459.3
1,451.2
1,414.1
1,436.5
Agriculture
534.9
570.0H
572.6
569.2
585.2
581.7
582.2
Commercial
426.9
400.7
409.6
419.5
432.2
416.1
416.0
Residential
344.7
370.0H
356.3
376.6
349.7
326.9
330.9
U.S. Territories
33.3
58.1
48.1
46.6
46.6
46.6
46.6
Total Emissions
6,371.0
7,339.0
6,710.2
6,760.0
6,623.8
6,492.3
6,456.7
LULUCF Sector Net Total3
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
Net Emissions (Sources and Sinks)
5,564.0
6,599.0
5,996.8
6,090.0
5,912.7
5,769.7
5,742.6
Notes: Total emissions presented without LULUCF. Total net emissions presented with LULUCF. Totals may not
sum due to independent rounding. Parentheses indicate negative values or sequestration.
a Hie LUTUCF Sector Net Total is the net sum of all CH4 and N2O emissions to the atmosphere plus net carbon stock
changes.
Using this categorization, emissions from transportation activities, in aggregate, accounted for the largest portion
(28.9 percent) of total U.S. greenhouse gas emissions in 2017. Electric power accounted for the second largest
portion (27.5 percent) of U.S. greenhouse gas emissions in 2017, while emissions from industry accounted for the
third largest portion (22.2 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 21.3 percent of U.S. greenhouse gas emissions were contributed by, in order of magnitude, the
agriculture, commercial, and residential sectors, plus emissions from U.S. Territories. Activities related to
agriculture accounted for 9.0 percent of U.S. emissions; unlike other economic sectors, agricultural sector emissions
were dominated by N20 emissions from agricultural soil management and CH4 emissions from enteric fermentation.
An increasing amount of carbon is stored in agricultural soils each year, but this CO2 sequestration is assigned to the
LULUCF sector rather than the agriculture economic sector. The commercial and residential sectors accounted for
6.4 percent and 5.1 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 2019 and Duffield 2006).26 These source categories include CO2
from fossil fuel combustion and the use of limestone and dolomite for flue gas desulfurization, CO2 and N20 from
incineration of waste, CH4 and N20 from stationary sources, and SF6 from electrical transmission and distribution
systems.
When emissions from electricity use are distributed among these end-use sectors, industrial activities and
transportation account for the largest shares of U.S. greenhouse gas emissions (29.7 percent and 29.0 percent,
respectively) in 2017. The commercial and residential sectors contributed the next largest shares of total U.S.
greenhouse gas emissions in 2017 (16.1 and 14.9 percent, respectively). Emissions from these sectors increase
substantially when emissions from electricity use are included, due to their relatively large share of electricity use
26 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-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
for energy (e.g., lighting, cooling, appliances). In all sectors except agriculture, CO2 accounts for at least 81.3
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 2017.
Table ES-7: U.S. Greenhouse Gas Emissions by Economic Sector with Electricity-Related
Emissions Distributed (MMT CO2 Eq.)
Economic Sectors
1990
2005
2013
2014
2015
2016
2017
Industry
2,300.9
2,223.5
2,036.7
2,023.0
1,973.6
1,906.4
1,915.6
Transportation
1,530.2
1,980.8
1,769.8
1,804.5
1,813.7
1,854.1
1,870.6
Commercial
981.1
1,222.4
1,131.5
1,143.1
1,112.3
1,066.6
1,038.4
Residential
955.6
1,246.0
1,109.2
1,129.3
1,051.1
997.8
964.5
Agriculture
569.9
608.3
614.9
613.5
626.5
620.8
620.9
U.S. Territories
33.3
58.1
48.1
46.6
46.6
46.6
46.6
Total Emissions
6,371.0
7,339.0
6,710.2
6,760.0
6,623.8
6,492.3
6,456.7
LULUCF Sector Net Total3
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
Net Emissions (Sources and Sinks)
5,564.0
6,599.0
5,996.8
6,090.0
5,912.7
5,769.7
5,742.6
a Hie LULUCF Sector Net Total is the net sum of all CH4 and N2O emissions to the atmosphere plus net carbon stock
changes.
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.
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
¦d- m vo rv.
1—1 lH 1—1 t—I
0000
o
cn
cr>
t-i CM
CT> <7>
cn
-------
Table ES-8 provides data on various statistics related to U.S. greenhouse gas emissions normalized to 1990 as a
baseline year. These values represent the relative change in each statistic since 1990. Greenhouse gas emissions in
the United States have grown at an average annual rate of 0.1 percent since 1990, although changes from year to
year have been significantly larger. This growth rate is slightly slower than that for total energy use and fossil fuel
consumption, and much slower than that for electricity use, 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 1.0 percent since 2005. Total energy use and fossil fuel
consumption have also decreased at slower rates than emissions since 2005, while electricity use, GDP, and national
population continued to increase.
Table ES-8: Recent Trends in Various U.S. Data (Index 1990 = 100)
Variable
1990
2005
2013
2014
2015
2016
2017
Avg. Annual
Growth Rate
Since 1990a
Avg. Annual
Growth Rate
Since 2005a
Greenhouse Gas Emissions'5
100
115
105
106
104
102
101
0.1%
-1.0%
Energy Usec
100
118
115
117
116
116
116
0.6%
-0.1%
Fossil Fuel Consumption0
100
119
110
111
110
109
108
0.3%
-0.7%
Electricity Usec
100
134
136
138
137
138
136
1.2%
0.1%
GDPd
100
159
176
180
186
189
193
2.5%
1.6%
Population6
100
118
126
127
128
129
130
1.0%
0.8%
a Average annual growth rate
b GWP-weighted values
cEnergy content-weighted values (EIA 2019)
dGDP in chained 2009 dollars (BEA 2019)
e U.S. Census Bureau (2019)
Figure ES-16: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic
Product (GDP)
200
Real GDP
180
160
140
Population
> 100
80
Emissions per capita
60
Emissions per $GDP
40
20
0
m "tf-
1—1 7—I
o o
VO
1—1
hs
1—1
O i—'
i—l iH
o o
rsi
LD
1—1
o
1—1
o
o o
Source: BEA (2019), U.S. Census Bureau (2018), and emission estimates in this report.
ES-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Key Categories
The 2006IPCC Guidelines (IPCC 2006) defines a key category as a "[category] that is prioritized within the
national inventory system because its estimate has a significant influence on a country's total inventory of
greenhouse gases in terms of the absolute level, the trend, or the uncertainty in emissions and removals."27 A key
category analysis identifies priority source or sink categories for focusing efforts to improve overall Inventory
quality. In addition, a qualitative review of key categories and non-key categories can also help identify additional
emission and sink categories to consider for improvement efforts.
Figure ES-17 presents the key categories identified by Approach 1 and Approach 2 level assessments including the
LULUCF sector for 2017. A level assessment using Approach 1 identifies all sources and sink categories that
cumulatively account for 95 percent of total (gross) emissions in a given year when assessed in descending order of
absolute magnitude. An Approach 2 level assessment incorporates the results of the uncertainty analysis for each
category and identifies all sources and sink categories that cumulatively account for 90 percent of the sum of all
level assessments when sorted in 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.
27 See Chapter 4 "Methodological Choice and Identification of Key Categories" in IPCC (2006). See .
Executive Summary ES-27
-------
Figure ES-17: 2017 Key Categories (MMT CO2 Eq.)
CO2 Emissions from Mobile Combustion: Road
CO2 Emissions from Stationary Combustion - Coal - Electricity Generation
Net CO2 Emissions from Forest Land Remaining Forest Landb
CO2 Emissions from Stationary Combustion - Gas - Electricity Generation
CO2 Emissions from Stationary Combustion - Gas - Industrial
CO2 Emissions from Stationary Combustion - Oil - Industrial
CO2 Emissions from Stationary Combustion - Gas - Residential
Direct N2O Emissions from Agricultural Soil Management
CH4 Emissions from Enteric Fermentation
CO2 Emissions from Mobile Combustion: Aviation
CO2 Emissions from Stationary Combustion - Gas - Commercial
CH4 Emissions from Natural Gas Systems
Emissions from Substitutes for Ozone Depleting Substances
Net CO2 Emissions from Settlements Remaining Settlements'1
CO2 Emissions from Non-Energy Use of Fuels
Net CO2 Emissions from Land Converted to Forest Landb
ChU Emissions from Landfills
Net CO2 Emissions from Land Converted to Settlements'1
CO2 Emissions from Mobile Combustion: Other
Net CO2 Emissions from Land Converted to Croplandb
ChU Emissions from Manure Management
CO2 Emissions from Stationary Combustion - Oil - Commercial
Fugitive Emissions from Coal Mining
CO2 Emissions from Stationary Combustion - Coal - Industrial
CO2 Emissions from Stationary Combustion - Oil - Residential
CO2 Emissions from Iron and Steel Production & Metallurgical Coke Production
CO2 Emissions from Cement Production
CO2 Emissions from Mobile Combustion: Marine
Indirect N2O Emissions from Applied Nitrogen
CH4 Emissions from Petroleum Systems
CO2 Emissions from Stationary Combustion - Oil - U.S. Territories
CO2 Emissions from Petrochemical Production
Net CO2 Emissions from Cropland Remaining Croplandb
Net CO2 Emissions from Land Converted to Grassland11
ChU Emissions from Abandoned Oil and Gas Wells
Key Categories as a
Portion of All Emissions
0 500 1,000 1,500
MMT CO2 Eq.
a For a complete list of key categories and detailed discussion of the underlying key category analysis, see Annex 1. Blue bars
indicate key categories identified using Approach 1 and Approach 2 level assessment including the LULUCF sector.
b The absolute values of net CO2 emissions from LULUCF are presented in this figure but reported separately from gross
emissions totals. Refer to Table ES-5 for a breakout of emissions and removals for LULUCF by gas and source/sink category.
Quality Assurance and Quality Control (QA/QC)
The United States seeks to continually improve the quality, transparency, and usability of the Inventory of U.S.
Greenhouse Gas Emissions and Sinks. To assist in these efforts, the United States implemented a systematic
approach to QA/QC. The procedures followed for the Inventory have been formalized in accordance with the U.S.
Inventory QA/QC plan for the Inventory, and the UNFCCC reporting guidelines and 2006IPCC Guidelines. The
QA process includes expert and public reviews for both the Inventory estimates and the Inventory report.
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 or an incomplete understanding of how emissions are
generated 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
ES-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
2006IPCC 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 all source and sink categories. Within the discussion of each
emission source and sink, specific factors affecting the uncertainty surrounding the estimates are discussed.
Executive Summary ES-29
-------
i. Introduction
This report presents estimates by the United States government of U.S. anthropogenic greenhouse gas emissions and
sinks for the years 1990 through 2017. A summary of these estimates is provided in Table 2-1 and Table 2-2 by gas
and source category in the Trends in Greenhouse Gas Emissions chapter. The emission estimates in these tables are
presented on both a full 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
Greenhouse Gas Inventories at its Twenty-Fifth Session (Mauritius, April 2006). The 2006 IPCC Guidelines built
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(1)(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
-------
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 2006IPCC 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
In following the UNFCCC requirement under Article 4.1 to develop and submit national greenhouse gas emissions
inventories, 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. All emissions and removals are calculated using
internationally-accepted methods consistent with the IPCC Guidelines.6 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.7 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 report itself follows this standardized format and provides an explanation of the
IPCC methods used to calculate emissions and removals.
On October 30, 2009, the U.S. Enviromnental Protection Agency (EPA) published a rule for the mandatory
reporting of greenhouse gases from large greenhouse gas emissions sources in the United States. Implementation of
40 CFR Part 98 is referred to as the EPA's Greenhouse Gas Reporting Program (GHGRP). 40 CFR Part 98 applies
to direct greenhouse gas emitters, fossil fuel suppliers, industrial gas suppliers, and facilities that inject CO2
underground for sequestration or other reasons.8 Reporting is at the facility level, except for certain suppliers of
fossil fuels and industrial greenhouse gases. The GHGRP dataset and the data presented in this Inventory are
complementary.
EPA's 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. The
GHGRP will continue to enhance QA/QC procedures and assessment of uncertainties.
EPA continues to analyze the data on an annual basis to improve the national estimates presented in this Inventory
and uses that data for a number of categories consistent with IPCC guidance.9 EPA has already integrated GHGRP
information for several categories10 since 2012 and also identifies other categories11 where EPA plans to integrate
5 U.S. Territories include American Samoa, Guam, Puerto Rico, U.S. Virgin Islands, Wake Island, and other U.S. Pacific Islands.
6 See .
7 See .
8 See .
9 See .
10 Energy Sector (Coal Mining, Stationary Combustion [Industrial Combustion Disaggregation], and Oil and Gas Systems);
Industrial Processes and Product Use (Adipic Acid Production, Aluminum Production, Carbon Dioxide Consumption, Electrical
Transmission and Distribution, HCFC-22 Production, Lime Production, Magnesium Production and Processing, Substitution of
ODS, Nitric Acid Production, Petrochemical Production, Semiconductor Manufacture); and Waste (Landfills).
11 Industrial Process and Product Use (Ammonia Production, Cement Production, and Other Fluorinated Gas Production)
1-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
additional GHGRP data in the next edition of this report (see the Planned Improvement sections of those specific
categories for details).
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-20lh 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,12 the U.S. Global Change Research Program (USGCRP),13 and the
National Academies of Sciences, Engineering, and Medicine (NAS).14
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
effect is primarily a function of the concentration of water vapor, carbon dioxide (CO2), methane (CH4), nitrous
oxide (N20), and other trace gases in the atmosphere that absorb the terrestrial radiation leaving the surface of the
Earth (IPCC 2013).
Naturally occurring greenhouse gases include water vapor, CO2, CH4, N20, and ozone (O3). Several classes of
halogenated substances that contain fluorine, chlorine, or bromine are also greenhouse gases, but they are, for the
12 See .
13 See < https://science2017.globalchange.gov/>.
14 See .
Introduction 1-3
-------
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.15 Some other fluorine-containing halogenated substances—hydrofluorocarbons (HFCs),
perfluorocarbons (PFCs), sulfur hexafluoride (SF6), and nitrogen trifluoride (NF3)—do not deplete stratospheric
ozone but are potent greenhouse gases. These latter substances are addressed by the UNFCCC and accounted for in
national greenhouse gas inventories.
There are also several other substances that influence the global radiation budget but are short-lived and therefore
not well-mixed, leading to spatially variable radiative forcing effects. These substances include carbon monoxide
(CO), nitrogen dioxide (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, CH4, and N20 are continuously emitted to and removed from the atmosphere by natural processes
on Earth. Anthropogenic activities (such as fossil fuel combustion, cement production, land-use, land-use change,
and forestry, agriculture, or waste management), however, can cause additional quantities of these and other
greenhouse gases to be emitted or sequestered, thereby changing their global average atmospheric concentrations.
Natural activities such as respiration by plants or animals and seasonal cycles of plant growth and decay are
examples of processes that only cycle carbon or nitrogen between the atmosphere and organic biomass. Such
processes, except when directly or indirectly perturbed out of equilibrium by anthropogenic activities, generally do
not alter average atmospheric greenhouse gas concentrations over decadal timeframes. Climatic changes resulting
from anthropogenic activities, however, could have positive or negative feedback effects on these natural systems.
Atmospheric concentrations of these gases, along with their rates of growth and atmospheric lifetimes, are presented
in Table 1-1.
Table 1-1: Global Atmospheric Concentration, Rate of Concentration Change, and
Atmospheric Lifetime of Selected Greenhouse Gases
Atmospheric Variable
CO2
CH4
N2O
SF«
CF4
Pre-industrial atmospheric concentration
Atmospheric concentration
Rate of concentration change
Atmospheric lifetime (years)
280 ppm
407 ppma
2.2 ppm/yrf
See footnote11
0.700 ppm
1.850 ppmb
7 ppb/yrŁg
12.4'
0.270 ppm
0.330 ppmc
0.8 ppb/yrf
121'
Oppt
9.3 pptd
0.27 ppt/yrf
3,200
40 ppt
79 ppte
0.7 ppt/yrf
50,000
a The atmospheric CO2 concentration is the 2017 annual average at the MaunaLoa, HI station (NOAA/ESRL 2018a). The
concentration in 2018 at Mauna Loa was 409 ppm. The global atmospheric CO2 concentration, computed using an average of
sampling sites across the world, was 405 ppm in 2017.
b The values presented are global 2017 annual average mole fractions (NOAA/ESRL 2018b).
c The values presented are global 2017 annual average mole fractions (NOAA/ESRL 2018c).
dThe values presented are global 2017 annual average mole fractions (NOAA/ESRL 2018d).
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 CO2 and CH4 is the average rate of change between 2007 and 2017 (NOAA/ESRL 2018a).
The rate of concentration change for N2O, SF6, and CF4 is the average rate of change between 2005 and 2011 (IPCC 2013).
15 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: 1990-2017
-------
g The growth rate for atmospheric CH4 decreased from over 10 ppb/year in the 1980s to nearly zero in the early 2000s; recently,
the growth rate has been about 7 ppb/year.
h For a given amount of CO2 emitted, some fraction of the atmospheric increase in concentration is quickly absorbed by the oceans
and terrestrial vegetation, some fraction of the atmospheric increase will only slowly decrease over a number of years, and a
small portion of the increase will remain for many centuries or more.
1 This lifetime has been defined as an "adjustment time" that takes into account the indirect effect of the gas on its own residence
time.
Source: Pre-industrial atmospheric concentrations, atmospheric lifetime, and rate of concentration changes for CH4, N2O, SF6, and
CF4 are from IPCC (2013). The rate of concentration change for CO2 is an average of the rates from 2007 through 2017 and has
fluctuated between 1.5 to 3.0 ppm per year over this period (NOAA/ESRL 2017a).
A brief description of each greenhouse gas, its sources, and its role in the atmosphere is given below. The following
section then explains the concept of GWPs, which are assigned to individual gases as a measure of their relative
average global radiative forcing effect.
Water Vapor (H20). Water vapor is the largest contributor to the natural greenhouse effect. Water vapor is
fundamentally different from other greenhouse gases in that it can condense and rain out when it reaches high
concentrations, and the total amount of water vapor in the atmosphere is in part a function of the Earth's
temperature. While some human activities such as evaporation from irrigated crops or power plant cooling release
water vapor into the air, these activities have been determined to have a negligible effect on global climate (IPCC
2013). The lifetime of water vapor in the troposphere is on the order of 10 days. Water vapor can also contribute to
cloud formation, and clouds can have both warming and cooling effects by either trapping or reflecting heat.
Because of the relationship between water vapor levels and temperature, water vapor and clouds serve as a feedback
to climate change, such that for any given increase in other greenhouse gases, the total warming is greater than
would happen in the absence of water vapor. Aircraft emissions of water vapor can create contrails, which may also
develop into contrail-induced cirrus clouds, with complex regional and temporal net radiative forcing effects that
currently have a low level of scientific certainty (IPCC 2013).
Carbon Dioxide (C02). In nature, carbon is cycled between various atmospheric, oceanic, land biotic, marine biotic,
and mineral reservoirs. The largest fluxes occur between the atmosphere and terrestrial biota, and between the
atmosphere and surface water of the oceans. In the atmosphere, carbon predominantly exists in its oxidized form as
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 407 ppmv in 2017, a 45 percent
increase (IPCC 2013; NOAA/ESRL 2018a).1617 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 caused by the anthropogenic increase in greenhouse gas concentrations and other anthropogenic forcings
together," of which CChis the most important (IPCC 2013).
Methane (CH4). Methane is primarily produced through anaerobic decomposition of organic matter in biological
systems. Agricultural processes such as wetland rice cultivation, enteric fermentation in animals, and the
decomposition of animal wastes emit CH4, as does the decomposition of municipal solid wastes. Methane is also
emitted during the production and distribution of natural gas and petroleum, and is released as a byproduct of coal
mining and incomplete fossil fuel combustion. Atmospheric concentrations of CH4 have increased by about 164
percent since 1750, from a pre-industrial value of about 700 ppb to 1,849 ppb in 201718 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
16 The pre-industrial period is considered as the time preceding the year 1750 (IPCC 2013).
17 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).
18 This value is the global 2017 annual average mole fraction (NOAA/ESRL 2018b).
Introduction 1-5
-------
estimated that slightly more than half of the current CH4 flux to the atmosphere is anthropogenic, from human
activities such as agriculture, fossil fuel production and use, and waste disposal (IPCC 2007).
Methane is primarily removed from the atmosphere through a reaction with the hydroxyl radical (OH) and is
ultimately converted to CO2. Minor removal processes also include reaction with chlorine in the marine boundary
layer, a soil sink, and stratospheric reactions. Increasing emissions of CH4 reduce the concentration of OH, a
feedback that increases the atmospheric lifetime of CH4 (IPCC 2013). Methane's reactions in the atmosphere also
lead to production of tropospheric ozone and stratospheric water vapor, both of which also contribute to climate
change.
Nitrous Oxide (N20). Anthropogenic sources of N20 emissions include agricultural soils, especially production of
nitrogen-fixing crops and forages, the use of synthetic and manure fertilizers, and manure deposition by livestock;
fossil fuel combustion, especially from mobile combustion; adipic (nylon) and nitric acid production; wastewater
treatment and waste incineration; and biomass burning. The atmospheric concentration of N20 has increased by 22
percent since 1750, from a pre-industrial value of about 270 ppb to 330 ppb in 2017,19 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,20 where it shields the Earth from harmful levels of
ultraviolet radiation, and at lower concentrations in the troposphere,21 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).
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,22 beginning in 1996, and then
followed by intermediate requirements and a complete phase-out by the year 2030. While ozone depleting gases
19 This value is the global 2017 annual average (NOAA/ESRL 2018c).
20 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.
21 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.
22 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.
1-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.23 Additionally, NOx emissions are also likely to
decrease CH4 concentrations, thus having a negative radiative forcing effect (IPCC 2013). Nitrogen oxides are
created from lightning, soil microbial activity, biomass burning (both natural and anthropogenic fires) fuel
combustion, and, in the stratosphere, from the photo-degradation of N20. Concentrations of NOx are both relatively
short-lived in the atmosphere and spatially variable.
Non-methane Volatile Organic Compounds (NMVOCs). Non-methane volatile organic compounds include
substances such as propane, butane, and ethane. These compounds participate, along with NOx, in the formation of
tropospheric ozone and other photochemical oxidants. NMVOCs are emitted primarily from transportation and
industrial processes, as well as biomass burning and non-industrial consumption of organic solvents. Concentrations
of NMVOCs tend to be both short-lived in the atmosphere and spatially variable.
Aerosols. Aerosols are extremely small particles or liquid droplets found in the atmosphere that are either directly
emitted into or are created through chemical reactions in the Earth's atmosphere. Aerosols or their chemical
precursors can be emitted by natural events such as dust storms, biogenic or volcanic activity, or by anthropogenic
processes such as transportation, coal combustion, cement manufacturing, waste incineration, or biomass burning.
Various categories of aerosols exist from both natural and anthropogenic sources, such as soil dust, sea salt, biogenic
aerosols, sulfates, nitrates, volcanic aerosols, industrial dust, and carbonaceous24 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
23 NOx emissions injected higher in the stratosphere, primarily from fuel combustion emissions from high altitude supersonic
aircraft, can lead to stratospheric ozone depletion.
24 Carbonaceous aerosols are aerosols that are comprised mainly of organic substances and forms of black carbon (or soot)
(IPCC 2013).
Introduction 1-7
-------
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 20 13).25 Although because they remain in the atmosphere for
only days to weeks, their concentrations respond rapidly to changes in emissions.26 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.).27 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
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...28
25 The IPCC (2013) defines high confidence as an indication of strong scientific evidence and agreement in this statement.
26 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).
27 Carbon comprises 12/44ths of carbon dioxide by weight.
28 Framework Convention on Climate Change; Available online at: ;
31 January 2014; Report of the Conference of the Parties at its nineteenth session; held in Warsaw from 11 to 23 November
2013; Addendum; Part two: Action taken by the Conference of the Parties at its nineteenth session; Decision 24/CP.19; Revision
of the UNFCCC reporting guidelines on annual inventories for Parties included in Annex I to the Convention; p. 2. (UNFCCC
2014).
Global Warming Potentials
N , / MMT \
Eq. = (kt of gas) x (GWP) x (jooofct)
MMT CO-
where,
1-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Greenhouse gases with relatively long atmospheric lifetimes (e.g., CO2, CH4, 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 GWP
C02
See footnoteb
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-152a
1.4
124
HFC-227ea
34.2
3,220
HFC-236fa
240
9,810
HFC-4310mee
15.9
1,640
CF4
50,000
7,390
C2F6
10,000
12,200
O
y
0
2,600
8,860
CdFl4
3,200
9,300
SFo
3,200
22,800
NF3
740
17,200
a 100-year time horizon.
b For a given amount of carbon dioxide emitted, some fraction of the
atmospheric increase in concentration is quickly absorbed by the oceans
and terrestrial vegetation, some fraction of the atmospheric increase will
only slowly decrease over a number of years, and a small portion of the
increase will remain for many centuries or more.
c The GWP of CH4 includes the direct effects and those indirect effects
due to the production of tropospheric ozone and stratospheric water
vapor. The indirect effect due to the production of CO2 is not included.
Source: IPCC 2007
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.
Introduction 1-9
-------
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.29 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
Gas
SAR
AR4
AR5
AR5 with
feedbacks6
SAR
AR5 with
AR5 feedbacksb
CO2
1
1
1
1
NC
NC
NC
CH4c
21
25
28
34
(4)
3
9
N2O
310
298
265
298
12
(33)
0
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)
HFC-4310mee
1,300
1,640
1,650
1,952
(340)
10
312
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
C4F10
7,000
8,860
9,200
10,213
(1,860)
340
1,353
CdFl4
7,400
9,300
7,910
8,780
(1,900)
(1,390)
(520)
SFo
23,900
22,800
23,500
26,087
1,100
700
3,287
NF3
NA
17,200
16,100
17,885
NA
(1,100)
685
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 Hie GWP values presented here from the AR5 report include climate-carbon feedbacks for the non-
CO2 gases in order to be consistent with the approach used in calculating the CO2 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
CO2 resulting from methane oxidation would lead to an increase in AR5 methane GWP values by 2 for
fossil methane.
Note: Parentheses indicate negative values.
Source: IPCC 2013, IPCC 2007, IPCC 2001, IPCC 1996.
1.2 National Inventory Arrangements
The U.S. Enviromnental 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 govermnent authorities, research and academic institutions, industry
associations, and private consultants.
29 See .
1-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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. The staff of EPA coordinate the annual methodological choice, activity data
collection, emission calculations, and QA/QC, and improvement planning at the individual source 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 collection 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 involuntary outreach elforts with EPA. Finally,
EPA as the National Inventory Focal Point, in coordination with the U.S. Department of State, officially submits the
Inventory to the UNFCCC each April. Figure 1-1 diagrams the National Inventory Arrangements.
Introduction 1-11
-------
Figure 1-1: National Inventory Arrangements Diagram Inventory Process Inventory Process
United States National Inventory Arrangements
United Nations
Framework Convention
on Climate Change
Inventory Submission
Inventory Compilation
U.S. Environmental
Protection Agency
Inventory Compiler
Emission Calculations
U.S. Environmental
Protection Agency
Other U.S.
Government Agencies
USDA (USFS, ARS), NOAA,
DOD, FAA, USGS
Data Collection
Energy
• U.S. Department of Energy and its National Laboratories
• Energy Information Administration
• U.S. Department of Transportation
• Bureau of Transportation Statistics
• Federal Highway Administration
• Federal Aviation Administration
• U.S. Department of Defense - Defense Logistics Agency
• U.S. Department of Commerce - Bureau of the Census
• U.S. Department of Homeland Security
• U.S. Department of Labor's Mine Safety and Health Administration
• EPA Office of Transportation and Air Quality MOVES Model
• EPA Greenhouse Gas Reporting Program (GHGRP)
• EPA Acid Rain Program
• American Association of Railroads
• American Public Transportation Association
• U.S. Department of Labor - Mine Safety and Health Administration
• Data from research studies, trade publications, and
industry associations f
\ Ł
Agriculture/LULUCF
• U.S. Department of Agriculture (USDA)
o National Agricultural Statistics Service (NASS) and Agricultural Research
Service (ARS)
o Natural Resources Conservation Service (NRCS)
o Economic Research Service (ERS)
o Farm Service Agency (FSA)
o Animal and Plant Health Inspection Service (APHIS)
o U.S. Forest Service (USFS) Forest Inventory and Analysis (FIA) Program
• U.S. Geological Survey (USGS)
• National Oceanic and Atmospheric Administration (NOAA)
• U.S. Department of the Interior (DOI), Bureau of Land Management (BLM)
• U.S. Census Bureau
• EPA
o Office of Land and Emergency Management
o Office of Air and Radiation
• Alaska Department of Natural Resources
• Association of American Plant Food Control Officials (AAPFCO)
• Data from research studies, trade publications and industry j I
associations
\ /
Industrial Processes and Product Use
• EPA Greenhouse Gas Reporting Program (GHGRP)
• U.S. Geological Survey (USGS), National Minerals
Information Center
• American Chemistry Council (ACC)
• American Iron and Steel Institute (AISI)
• U.S. International Trade Commission (USITC)
• U.S. Aluminum Association
• Air-Conditioning, Heating, and Refrigeration Institute
• Data from other U.S. government, 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
/ V
1-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 and the unique characteristics of its emissions or removals profile. 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. A multi-stage process for collecting information from the individual source and sink category leads
and producing the Inventory is undertaken annually to compile all information and data.
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 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-3 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 regulations1 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.2 In the Inventory, EPA is publishing only data values that meet the GHGRP
aggregation criteria.3 Specific uses of aggregated facility-level data are described in the respective methodological
1 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 .
2 Federal Register Notice on "Greenhouse Gas Reporting Program: Publication of Aggregated Greenhouse Gas Data." See pp. 79
and 110 of notice at .
3 U.S. EPA Greenhouse Gas Reporting Program. Developments on Publication of Aggregated Greenhouse Gas Data, November
25, 2014. See .
Introduction 1-13
-------
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.
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.
1-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Expert, Public, and UNFCCC Review Periods
During 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, a 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 categories Planned Improvement sections. See those sections
for specific details. EPA publishes 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.4
Feedback from these review processes all contribute to improving inventory quality over time and are described
further in Annex 8.
Final Submittal to UNFCCC and Document Printing
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 Reporter database.
EPA as the National Inventory focal point and sends the official submission of the U.S. Inventory to the UNFCCC,
coordinating with the U.S. Department of State. The document is then formatted and posted online, available for the
public.5
1.4 Methodology and Data Sources
Emissions of greenhouse gases from various source and sink categories have been estimated using methodologies
that are consistent with the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC 2006). To a
great extent, this report makes use of published official economic and physical statistics for activity data and
emission factors. Depending on the emission source category, activity data can include fuel consumption or
deliveries, vehicle-miles traveled, raw material processed, etc. Emission factors are factors that relate quantities of
emissions to an activity. For more information on data sources see Section 1.2 above, Box 1-1 on use of GHGRP
data, and categories' methodology sections for more information on data sources. In addition to official statistics, the
report utilizes findings from academic studies, trade association surveys and statistical reports, along with expert
judgment, consistent with the 2006 IPCC Guidelines.
The methodologies provided in the 2006 IPCC Guidelines represent foundational methodologies for a variety of
source categories, and many of these methodologies continue to be improved and refined as new research and data
become available. This report uses the IPCC methodologies when applicable, and supplements them with other
available country-specific methodologies and data where possible. 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. Complete documentation is provided in the annexes on the detailed methodologies and data sources utilized
in the calculation of each source category.
4 See .
5 See .
Introduction 1-15
-------
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. 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.
1.5 Key Categories
The 2006IPCC Guidelines (IPCC 2006) defines a key category as a "[category] that is prioritized within the
national inventory system because its estimate lias 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."6 This
analysis can identify source and sink categories that diverge from the overall trend in national emissions. Finally, a
qualitative evaluation of key categories is performed to capture any categories that were not identified in any of the
quantitative analyses.
Approach 1, as defined in the 2006 IPCC Guidelines (IPCC 2006), was implemented to identify the key categories
for the United States. 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.
Approach 2, as defined in the 2006 IPCC Guidelines (IPCC 2006), was then implemented to identify any additional
key categories not already identified in Approach 1 assessment. This analysis, which includes each source
category's uncertainty assessments (or proxies) in its calculations, 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. For this inventory, no additional categories were identified using criteria recommend by
IPCC, but EPA continues to update its qualitative assessment on an annual basis.
Table 1-4: Key Categories for the United States (1990 and 2017)
CRF Source Category
Gas
Approach 1
Approach 2
Quala
2017
Emissions
(MMT
CO2 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
Energy
l.A.3.b CO2
Emissions from
Mobile Combustion:
Road
CO2
.
.
1,504.1
l.A.l CO2 Emissions
from Stationary
Combustion - Coal -
Electricity Generation
CO2
.
.
1,207.1
6 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: 1990-2017
-------
CRF Source Category
Gas
Approach 1
Approach 2
Quala
2017
Emissions
(MMT
CO2 Eq.)
505.6
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
Level Trend Level Trend
Without Without With With
LULUCF LULUCF LULUCF LULUCF
l.A.l CO2 Emissions
from Stationary
Combustion - Gas -
Electricity Generation
CO2
.
.
1 .A.2 CO2 Emissions
from Stationary
Combustion - Gas -
Industrial
CO2
.
.
484.7
1 .A.2 CO2 Emissions
from Stationary
Combustion - Oil -
Industrial
CO2
.
.
271.5
l.A.4.b CO2
Emissions from
Stationary
Combustion - Gas -
Residential
CO2
•
•
241.5
l.A.3.a CO2
Emissions from
Mobile Combustion:
Aviation
CO2
.
•
173.2
1 ,A.4.a CO2 Emissions
from Stationary
Combustion - Gas -
Commercial
CO2
.
.
173.2
1.A.5 CO2 Emissions
from Non-Energy Use
of Fuels
CO2
•
•
123.2
1 A.3.e CO2
Emissions from
Mobile Combustion:
Other
CO2
.
83.0
1 A.4.b CO2
Emissions from
Stationary
Combustion - Oil -
Commercial
CO2
.
57.7
1 A.2 CO2 Emissions
from Stationary
Combustion - Coal -
Industrial
CO2
.
.
54.4
l.A.4.b CO2
Emissions from
Stationary
Combustion - Oil -
Residential
CO2
.
•
53.0
l.A.3.d CO2
Emissions from
Mobile Combustion:
Marine
CO2
.
40.3
1.A.5 CO2 Emissions
from Stationary
Combustion - Oil -
U.S. Territories
CO2
.
34.3
Introduction 1-17
-------
CRF Source Category
Gas
Approach 1
Approach 2
Quala
2017
Emissions
(MMT
CO2 Eq.)
26.3
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.B.2 CO2 Emissions
from Natural Gas
Systems
CO2
•
1.B.2 CO2 Emissions
from Petroleum
Systems
CO2
•
.
23.3
l.A.l CO2 Emissions
from Stationary
Combustion - Oil -
Electricity Generation
CO2
.
.
18.9
1.A.5 CO2 Emissions
from Stationary
Combustion - Gas -
U.S. Territories
CO2
•
3.0
l.A.4.a CO2
Emissions from
Stationary
Combustion - Coal -
Commercial
CO2
•
2.0
l.A.4.b CO2
Emissions from
Stationary
Combustion - Coal -
Residential
CO2
•
0.0
1.B.2 CH4 Emissions
from Natural Gas
Systems
CH4
.
.
165.6
l.B.l Fugitive
Emissions from Coal
Mining
ch4
.
.
55.7
1.B.2 CH4 Emissions
from Petroleum
Systems
ch4
•
.
37.7
1.B.2 CH4 Emissions
from Abandoned Oil
and Gas Wells
ch4
•
6.9
l.A.4.b N011-CO2
Emissions from
Stationary
Combustion -
Residential
ch4
.
3.8
l.A.3.e CH4
Emissions from
Mobile Combustion:
Other
ch4
•
1.9
l.A.l N011-CO2
Emissions from
Stationary
Combustion -
Electricity Generation
n2o
•
24.8
l.A.3.b N2O
Emissions from
Mobile Combustion:
Road
n2o
.
•
12.0
1-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
CRF Source Category
Gas
Approach 1
Approach 2
Quala
2017
Emissions
(MMT
CO2 Eq.)
2.7
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 N011-CO2
Emissions from
Stationary
Combustion -
Industrial
N2O
•
Industrial Processes and Product Use
2.C.1 CO2 Emissions
from Iron and Steel
Production &
Metallurgical Coke
Production
CO2
.
.
41.8
2.A.1 CO2 Emissions
from Cement
Production
CO2
.
40.3
2.B.8 CO2 Emissions
from Petrochemical
Production
CO2
.
28.2
2.B.3 N2O Emissions
from Adipic Acid
Production
N2O
•
7.4
2.F Emissions from
Substitutes for Ozone
Depleting Substances
HiGWP
.
.
152.7
2.B.9 HFC-23
Emissions from
HCFC-22 Production
HiGWP
.
•
5.2
2.G.1 SFo Emissions
from Electrical
Transmission and
Distribution
HiGWP
.
•
4.3
2.C.3 PFC Emissions
from Aluminum
Production
HiGWP
•
•
1.1
Agriculture
3.G CO2 Emissions
from Liming
CO2
•
3.2
3 A CH4 Emissions
from Enteric
Fermentation
CH4
.
•
175.4
3.B CH4 Emissions
from Manure
Management
ch4
.
.
61.7
3.C CH4 Emissions
from Rice Cultivation
ch4
•
11.3
3.D.1 Direct N2O
Emissions from
Agricultural Soil
Management
N2O
.
.
227.7
3.D.2 Indirect N2O
Emissions from
Applied Nitrogen
N2O
•
•
38.8
Waste
Introduction 1-19
-------
CRF Source Category
Gas
Approach 1
Approach 2
Quala
2017
Emissions
(MMT
CO2 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
5.A CH4 Emissions
from Landfills
CH4
.
.
107.7
Land Use, Land Use Change, and Forestry
4.E.2 Net CO2
Emissions from Land
Converted to
Settlements
CO2
•
•
86.2
4.B.2 Net CO2
Emissions from Land
Converted to Cropland
CO2
•
•
66.9
4.C.2 Net CO2
Emissions from Land
Converted to
Grassland
CO2
•
8.3
4.B.1 Net CO2
Emissions from
Cropland Remaining
Cropland
CO2
•
•
(10.3)
4.A.2 Net CO2
Emissions from Land
Converted to Forest
Land
CO2
•
•
(120.6)
4.E.1 Net CO2
Emissions from
Settlements
Remaining
Settlements
CO2
•
•
(134.5)
4.A.1 Net CO2
Emissions from Forest
Land Remaining
Forest Land
CO2
•
•
(621.1)
Subtotal Without LULUCF
6,298.2 |
Total Emissions Without LULUCF
6,456.7
Percent of Total Without LULUCF
98%
Subtotal With LULUCF
5,528.2
Total Emissions With LULUCF
5,742.6
Percent of Total With LULUCF
96%
a Qualitative criteria.
1-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
1.6 Quality Assurance and Quality Control
(QA/QC)
As part of efforts to achieve its stated goals for inventory quality, transparency, and credibility, the United States has
developed a quality assurance and quality control plan designed to check, document and improve the quality of its
inventory over time. QA/QC activities on the Inventory are undertaken within the framework of the U.S. Quality
Assurance/Quality Control and Uncertainty Management Plan (QA/QC plan) for the U.S. Greenhouse Gas
Inventory: Procedures Manual for QA/QC and Uncertainty Analysis.
Key attributes of the QA/QC plan are summarized in Figure 1-2. These attributes include:
• Procedures and Forms: detailed and specific systems that serve to standardize the process of documenting
and archiving information, as well as to guide the implementation of QA/QC and the analysis of
uncertainty
• Implementation of Procedures: application of QA/QC procedures throughout the whole inventory
development process from initial data collection, through preparation of the emission estimates, to
publication of the Inventory
• Quality Assurance (QA): expert and public reviews for both the inventory estimates and the Inventory
report (which is the primary vehicle for disseminating the results of the inventory development process).
The expert technical review conducted by the UNFCCC supplements these QA processes, consistent with
the QA good practice and the 2006IPCC Guidelines (IPCC 2006)
• Quality Control (QC): application of General (Tier 1) and Category-specific (Tier 2) quality controls and
checks, as recommended by 2006 IPCC Guidelines (IPCC 2006), along with consideration of secondary
data and category-specific checks (additional Tier 2 QC) in parallel and coordination with the uncertainty
assessment; the development of protocols and templates, which provides for more structured
communication and integration with the suppliers of secondary information
• General (Tier 1) and Category-specific (Tier 2) Checks: quality controls and checks, as recommended by
IPCC Good Practice Guidance and 2006 IPCC Guidelines (IPCC 2006)
• Record Keeping: provisions to track which procedures have been followed, the results of the QA/QC,
uncertainty analysis, and feedback mechanisms for corrective action based on the results of the
investigations which provide for continual data quality improvement and guided research efforts
• Multi-Year Implementation', a schedule for coordinating the application of QA/QC procedures across
multiple years, especially for category-specific QC, prioritizing key categories
• Interaction and Coordination: promoting communication within the EPA, across Federal agencies and
departments, state government programs, and research institutions and consulting firms involved in
supplying data or preparing estimates for the Inventory. The QA/QC Management Plan itself is intended to
be revised and reflect new information that becomes available as the program develops, methods are
improved, or additional supporting documents become necessary.
In addition, based on the national QA/QC plan for the Inventory, some sector, subsector and category-specific
QA/QC plans have been developed. These plans follow the procedures outlined in the national QA/QC plan,
tailoring the procedures to the specific text and data files of the 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), 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.
Introduction 1-21
-------
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.
Figure 1-2: U.S. QA/QC Plan Summary
ro
c
<
>-
c
d)
>
c
ro
c
<
§
Obtain data in electronic
format (if possible)
Review spreadsheet
construction
Avoid hardwiring
• Use data validation
Protect cells
Develop automatic
checkers for:
• Outliers, negative
values, or missing
data
Variable types
match values
Time series
consistency
Maintain tracking tab for
status of gathering
efforts
Contact reports for non-
electroniccommunications
Provide cell references for
primary data elements
Obtaincopiesofall data
sources
~stand location of any
working/external
spreadsheets
Document assumptions
Clearly label parameters,
units, and conversion
factors
Review spreadsheet
integrity
• Equations
• Units
Inputs and output
Develop automated
checkers for:
• Input ranges
¦ Calculations
• Emission aggregation
Check input data for
transcription errors
Inspect automatic
checkers
Identify spreadsheet
modifications that could
provide additional
QA/QC checks
Check citations in
spreadsheetand text for
accuracy and style
Check reference docketfor
new citations
Review documentation for
any data/ methodology
changes
Reproduce calculations
Review time series
consistency
Review changes in
data/consistency with IPCC
methodology
Common starting
versions for each
inventory year
Utilize unalterable
summary tab foreach
source spreadsheet for
linkingto a master
summary spreadsheet
Follow strict version
control procedures
Document QA/QC
procedures
Data Gathering
Data Documentation CalculatingEmissions
Cross-Cutting
Coordination
1.7 Uncertainty Analysis of 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
1-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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. storage factors of non-energy uses of fossil fuels). The UNFCCC reporting
guidelines follow the recommendation in the 2006IPCC 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 sources. 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 interior Alaska) 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 CH4 and N20 emissions from stationary and mobile
combustion is highly uncertain.
• 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. Estimates of quantitative uncertainty for the total U.S. greenhouse gas
emissions are shown below, in Table 1-5.
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; further explanation is provided within the
respective source category text and in Annex 7. 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.
Table 1-5: Estimated Overall Inventory Quantitative Uncertainty (MMT CO2 Eq. and Percent)
2017 Emission
Uncertainty Range Relative to Emission
Standard
Estimate
Estimate3
Meanb
Deviationb
Gas
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
(MMT CO2 Eq.)
Lower
Upper
Lower
Upper
Bound0
Bound0
Bound
Bound
CO2
5,270.7
5,154.8
5,499.8
-2%
4%
5,326.0
88.7
CH4d
656.3
596.0
747.6
-9%
14%
670.5
38.7
N2Od
360.5
316.2
434.7
-12%
21%
368.7
30.4
PFC, HFC, SF6,andNF3d
169.1
168.9
188.2
-(+)%
11%
178.4
5.0
Total
6,456.7
6,350.6
6,742.9
-2%
4%
6,543.6
101.0
LULUCF Emissionse
15.5
12.9
18.6
-17%
20%
15.7
1.5
LULUCF Total Net Fluxf
(729.6)
(1,094.4)
(488.5)
50%
-33%
(793.4)
154.0
LULUCF Sector Totals
(714.1)
(1,078.2)
(472.8)
51%
-34%
(777.7)
154.0
Net Emissions (Sources and
Sinks)
5,742.6
5,408.2
6,130.0
-6%
7%
5,765.9
183.6
+ 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.
Introduction 1-23
-------
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, N2O and high GWP
gases used in the Inventory emission calculations for 2017.
e LULUCF emissions include the CH4 and N2O emissions reported for Non-CC>2 Emissions from Forest Fires, Emissions from
Drained Organic Soils, N2O Fluxes from Forest Soils, Non-CC>2 Emissions from Grassland Fires, N2O Fluxes from Settlement
Soils, Coastal Wetlands Remaining Coastal Wetlands, Peatlands Remaining Peatlands, and CH4 Emissions from Land
Converted to Coastal Wetlands.
f Net CO2 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 CO2 (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.
Emissions calculated for the U.S. Inventory reflect current best estimates; in some cases, however, estimates are
based on approximate methodologies, assumptions, and incomplete data. As new information becomes available in
the future, the United States will continue to improve and revise its emission estimates. See Annex 7 of this report
for further details on the U.S. process for estimating uncertainty associated with the emission estimates and for a
more detailed discussion of the limitations of the current analysis and plans for improvement. Annex 7 also includes
details on the uncertainty analysis performed for selected source categories.
1.8 Completeness
This report, along with its accompanying CRF tables, serves as a thorough assessment of the anthropogenic sources
and sinks of greenhouse gas emissions for the United States for the time series 1990 through 2017. 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 insignificant7 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 chapter of this report.
7 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 CO2 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."
1-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
1.9 Organization of Report
In accordance with the revision of the UNFCCC reporting guidelines agreed to at the nineteenth Conference of the
Parties (UNFCCC 2014), this Inventory of U.S. Greenhouse Gas Emissions and Sinks is segregated into five sector-
specific chapters consistent with the UN Common Reporting Framework, listed below in Table 1-6. 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.
Table 1-6: IPCC Sector Descriptions
Chapter/EPCC Sector
Activities Included
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.
Emissions resulting from industrial processes and product use of greenhouse
gases.
Emissions from agricultural activities except fuel combustion, which is
addressed under Energy.
Emissions and removals of CO2, and emissions of CH4, and N2O from land use,
land-use change and forestry.
Emissions from waste management activities.
Energy
Industrial Processes and
Product Use
Agriculture
Land Use, Land-Use
Change, and Forestry
Waste
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 verily the emission estimates, consistent with
the U.S. QA/QC plan, and any key findings.
Recalculations: 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 1-7.
Table 1-7: List of Annexes
ANNEX 1 Key Category Analysis
ANNEX 2 Methodology and Data for Estimating CO2 Emissions from Fossil Fuel Combustion
2.1. Methodology for Estimating Emissions of CO2 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, N2O, and Indirect Greenhouse Gases from Stationary
Introduction 1-25
-------
Combustion
3.2. Methodology for Estimating Emissions of CH4, N2O, 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 CO2 Emissions from Petroleum Systems
3.6. Methodology for Estimating CH4 Emissions from Natural Gas Systems
3.7. Methodology for Estimating CO2 and N2O 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 N2O Emissions from Manure Management
3.12. Methodology for Estimating N2O 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 CO2 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 F actors
8.4 . Responses During the Review Process
1-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
2. Trends in Greenhouse Gas Emissions
2.1 Recent Trends in U.S. Greenhouse Gas
Emissions and Sinks
In 2017, total gross U.S. greenhouse gas emissions were 6,456.7 MMT, or million metric tons, carbon dioxide (CO2)
Eq.1 Total U.S. emissions have increased by 1.3 percent from 1990 to 2017, and emissions decreased from 2016 to
2017 by 0.5 percent (35.5 MMT CO2 Eq.). The decrease in total greenhouse gas emissions between 2016 and 2017
was driven in part by a decrease in CO2 emissions from fossil fuel combustion. The decrease in CO2 emissions from
fossil fuel combustion was a result of multiple factors, including a continued shift from coal to natural gas and
increased use of renewable energy in the electric power sector, and milder weather that contributed to less overall
electricity use.
Since 1990, U.S. emissions have increased at an average annual rate of 0.1 percent. Figure 2-1 through Figure 2-3
illustrate the overall trend in total U.S. emissions by gas, annual changes, and absolute changes since 1990. Overall,
net emissions in 2017 were 13.0 percent below 2005 levels as shown in Table 2-1.
Figure 2-1: Gross U.S. Greenhouse Gas Emissions by Gas (MMT CO2 Eq.)
10,000
8,000
o 6,000
u
4,000
2,000
HFCs, PFCs, SFe and NF3 Subtotal
Nitrous Oxide
I Methane
Carbon Dioxide
1 Hie 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.
Trends 2-1
-------
Figure 2-2: Annual Percent Change in Gross U.S. Greenhouse Gas Emissions Relative to the
Previous Year
2.9%
3.4%
0.9% 0.9%
0.6% 0.6%
0.6%
-0.5%
¦0.9%
-0.9%
-1.6%
-2.0% -2.0%
-2.8%
-3.6%
2% 1.7%1.7% li4% j 30/J
0%
-2%
-4%
-6%
-8%
-6.3%
CTi CT*
Figure 2-3: Cumulative Change in Annual Gross U.S. Greenhouse Gas Emissions Relative to
1990 (1990=0, MMT COz Eq.)
1,200
962 968
2 600
-200
i-irMro^-m^or».oocr»o*HrMro^-Lr)kDrvooCT»Oi-irMro^-Lrjvo
CT* 0"»CT»0"» CT»01CT»CT»CT*000 OOO OOOO^—l 1—I 1—I 1—I tHtHtH
CT»CT»CT*CT>CT»CT»CT*0^CT>OOOOOOOOOOOOOOOOO
1—li—li—li—li—li—I1H1—li—IfMCNfNrMCNCNfMfNCMCNfMfMfMtNOIfMfM
Overall, from 1990 to 2017, total emissions of CO2 increased by 149.6 MMT CO2 Eq. (2.9 percent), while total
emissions of methane (CH4) decreased by 123.5 MMT CO: Eq. (15.8 percent), and total emissions of nitrous oxide
(N2O) decreased by 9.8 MMT CO2 Eq. (2.6 percent). During the same period, aggregate weighted emissions of
hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride (SF6), and nitrogen trifluoride (NF3) rose
by 69.5 MMT CO2 Eq. (69.7 percent). Despite being emitted in smaller quantities relative to the other principal
greenhouse gases, emissions of HFCs, PFCs, SF6, and NF3 are significant because many of them have extremely
high global wanning 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
11.3 percent of total emissions in 2017.
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.
2-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 2-1: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
5,121.2
6,130.6
5,522.9
5,572.1
5,423.0
5,306.7
5,270.7
Fossil Fuel Combustion
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
Transportation
1,469.1
1,857.0
1,682.7
1,721.6
1,734.0
1,779.0
1,800.6
Electric Power Sector
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
Industrial
857.5
853.4
840.0
819.6
807.9
807.6
810.7
Residential
338.2
357.9
329.3
346.8
317.8
292.9
294.5
Commercial
226.5
226.8
224.6
232.9
245.5
232.1
232.9
U.S. Territories
27.6
49.7
42.5
41.4
41.4
41.4
41.4
Non-Energy Use of Fuels
119.6
139.6
123.5
119.9
126.9
113.7
123.2
Iron and Steel Production &
Metallurgical Coke Production
101.6
68.2
53.5
58.4
47.8
42.3
41.8
Cement Production
33.5
46.2
36.4
39.4
39.9
39.4
40.3
Petrochemical Production
21.2
26.8
26.4
26.5
28.1
28.1
28.2
Natural Gas Systems
30.0
22.6
25.1
25.5
25.1
25.5
26.3
Petroleum Systems
9.0
11.6
25.1
29.6
31.7
22.2
23.3
Ammonia Production
13.0
9.2
9.5
9.4
10.6
10.8
13.2
Lime Production
11.7
14.6
14.0
14.2
13.3
12.9
13.1
Incineration of Waste
8.0
12.5
10.3
10.4
10.7
10.8
10.8
Other Process Uses of Carbonates
6.3
7.6
11.5
13.0
12.2
11.0
10.1
Urea Fertilization
2.4
3.5
4.4
4.5
4.7
4.9
5.1
Urea Consumption for Non-
Agricultural Purposes
3.8
3.7
4.6
1.8
4.6
5.1
5.0
Carbon Dioxide Consumption
1.5
1.4
4.2
4.5
4.5
4.5
4.5
Liming
4.7
4.3
3.9
3.6
3.7
3.2
3.2
Ferroalloy Production
2.2
1.4
1.8
1.9
2.0
1.8
2.0
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.7
1.8
Titanium Dioxide Production
1.2
1.8
1.7
1.7
1.6
1.7
1.7
Glass Production
1.5
1.9
1.3
1.3
1.3
1.2
1.3
Aluminum Production
6.8
4.1
3.3
2.8
2.8
1.3
1.2
Phosphoric Acid Production
1.5
1.3
1.1
1.0
1.0
1.0
1.0
Zinc Production
0.6
1.0
1.4
1.0
0.9
0.9
1.0
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Silicon 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 Consumption"
219.4
230.7
315.5
323.2
317.7
317.2
322.2
International Bunker Fuelsb
103.5
113.1
99.8
103.4
110.9
116.6
120.1
CH4c
779.8
691.4
663.0
662.1
661.4
654.9
656.3
Enteric Fermentation
164.2
168.9
165.5
164.2
166.5
171.9
175.4
Natural Gas Systems
193.1
171.4
165.6
165.1
167.2
165.7
165.6
Landfills
179.6
131.4
112.9
112.5
111.2
108.0
107.7
Manure Management
37.1
53.7
58.1
57.8
60.9
61.5
61.7
Coal Mining
96.5
64.1
64.6
64.6
61.2
53.8
55.7
Petroleum Systems
42.1
36.7
41.6
42.1
39.5
38.2
37.7
Wastewater Treatment
15.3
15.4
14.3
14.3
14.5
14.2
14.2
Rice Cultivation
16.0
16.7
11.5
12.7
12.3
13.7
11.3
Stationary Combustion
8.6
7.8
8.7
8.9
8.5
7.9
7.8
Abandoned Oil and Gas Wells
6.6
6.9
7.0
7.1
7.1
7.2
6.9
Abandoned Underground Coal
Mines
7.2
6.6
6.2
6.3
6.4
6.7
6.4
Mobile Combustion
12.9
9.6
4.5
4.1
3.6
3.4
3.2
Composting
0.4
1.9
2.0
2.1
2.1
2.1
2.2
Petrochemical Production
0.2
0.1
0.1
0.1
0.2
0.2
0.3
Trends 2-3
-------
Field Burning of Agricultural
Residues
0.1
0.2
0.2
0.2
0.2
0.2
0.2
Ferroalloy Production
+
+
+
+
+
+
+
Silicon 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
370.3
375.8
365.4
362.7
374.1
364.5
360.5
Agricultural Soil Management
251.7
254.5
265.2
262.3
277.8
267.6
266.4
Stationary Combustion
25.1
34.3
32.7
33.0
30.6
30.1
28.6
Manure Management
14.0
16.5
17.4
17.4
17.6
18.2
18.7
Mobile Combustion
42.0
39.0
22.1
20.2
18.8
17.9
16.9
Nitric Acid Production
12.1
11.3
10.7
10.9
11.6
10.1
9.3
Adipic Acid Production
15.2
7.1
3.9
5.4
4.3
7.0
7.4
Wastewater Treatment
3.4
4.4
4.7
4.8
4.8
4.9
5.0
N2O from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
Composting
0.3
1.7
1.8
1.9
1.9
1.9
1.9
Caprolactam, Glyoxal, and
Glyoxylic Acid Production
1.7
2.1
2.0
2.0
2.0
2.0
1.4
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Semiconductor Manufacture
+
0.1
0.2
0.2
0.2
0.2
0.2
Field Burning of Agricultural
Residues
+
0.1
0.1
0.1
0.1
0.1
0.1
Petroleum Systems
+
+
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
0.9
0.9
0.9
1.0
1.0
HFCs
46.6
122.3
146.1
150.7
153.8
155.0
158.3
Substitution of Ozone Depleting
Substances'1
0.3
102.1
141.7
145.2
149.2
151.7
152.7
HCFC-22 Production
46.1
20.0
4.1
5.0
4.3
2.8
5.2
Semiconductor Manufacture
0.2
0.2
0.3
0.3
0.3
0.3
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
5.9
5.6
5.1
4.4
4.1
Semiconductor Manufacture
2.8
3.2
2.9
3.1
3.1
3.0
3.0
Aluminum Production
21.5
3.4
3.0
2.5
2.0
1.4
1.1
Substitution of Ozone Depleting
Substances
0.0
+
+
+
+
+
+
SF«
28.8
11.8
6.3
6.3
5.8
6.3
6.1
Electrical Transmission and
Distribution
23.1
8.3
4.4
4.6
4.1
4.4
4.3
Magnesium Production and
Processing
5.2
2.7
1.3
0.9
1.0
1.1
1.1
Semiconductor Manufacture
0.5
0.7
0.7
0.7
0.7
0.9
0.7
NF3
+
0.5
0.5
0.5
0.6
0.6
0.6
Semiconductor Manufacture
+
0.5
0.5
0.5
0.6
0.6
0.6
Total Emissions
6,371.0
7,339.0
6,710.2
6,760.0
6,623.8
6,492.3
6,456.7
LULUCF Emissions0
7.8
16.0
17.5
17.7
28.3
15.5
15.5
LULUCF CH4 Emissions
5.0
9.0
9.9
10.1
16.5
8.8
8.8
LUTUCF N2O Emissions
2.8
7.0
7.6
7.7
11.8
6.7
6.7
LULUCF Carbon Stock Change6
(814.8)
(756.1)
(731.0)
(687.8)
(739.4)
(738.1)
(729.6)
LULUCF Sector Net Total'
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
Net Emissions (Sources and Sinks)
5,564.0
6,599.0
5,996.8
6,090.0
5,912.7
5,769.7
5,742.6
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 CO2 Eq.
2-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 N2O are reported separately from gross emissions totals. LULUCF emissions include
the CH4 andN^O 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 N2O emissions from Forest Soils and Settlement Soils. Refer to Lable 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.
f Hie LULUCF Sector Net Total is the net sum of all CFL and N2O 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
2013
2014
2015
2016
2017
CO2
5,121,179
6,130,552
5,522,908
5,572,106
5,422,966
5,306,662
5,270,749
Fossil Fuel Combustion
4,738,756
5,744,754
5,157,391
5,199,345
5,047,107
4,961,876
4,911,962
Transportation
1,469,090
1,856,999
1,682,675
1,721,581
1,733,961
1,779,031
1,800,567
Electric Power Sector
1,819,951
2,399,974
2,038,258
2,037,148
1,900,624
1,808,863
1,732,025
Industrial
857,463
853,423
840,023
819,556
807,876
807,599
810,651
Residential
338,170
357,893
329,328
346,835
317,816
292,885
294,463
Commercial
226,528
226,755
224,643
232,864
245,462
232,137
232,892
U.S. Territories
27,555
49,710
42,463
41,361
41,367
41,362
41,363
Non-Energy Use of Fuels
119,551
139,625
123,476
119,895
126,939
113,719
123,221
Iron and Steel Production &
Metallurgical Coke
Production
101,630
68,210
53,471
58,353
47,825
42,309
41,782
Cement Production
33,484
46,194
36,369
39,439
39,907
39,439
40,324
Petrochemical Production
21,222
26,810
26,395
26,496
28,062
28,110
28,225
Natural Gas Systems
30,048
22,638
25,148
25,518
25,071
25,488
26,327
Petroleum Systems
8,950
11,552
25,130
29,597
31,672
22,200
23,336
Ammonia Production
13,047
9,196
9,480
9,377
10,634
10,838
13,216
Lime Production
11,700
14,552
14,028
14,210
13,342
12,942
13,145
Incineration of Waste
7,950
12,469
10,333
10,429
10,742
10,765
10,790
Other Process Uses of
Carbonates
6,297
7,644
11,524
12,954
12,182
10,969
10,139
Urea Fertilization
2,417
3,504
4,443
4,515
4,728
4,877
5,051
Urea Consumption for Non-
Agricultural Purposes
3,784
3,653
4,556
1,807
4,578
5,132
4,958
Carbon Dioxide Consumption
1,472
1,375
4,188
4,471
4,471
4,471
4,471
Liming
4,667
4,349
3,907
3,609
3,737
3,206
3,182
Ferroalloy Production
2,152
1,392
1,785
1,914
1,960
1,796
1,975
Soda Ash Production
1,431
1,655
1,694
1,685
1,714
1,723
1,753
Titanium Dioxide Production
1,195
1,755
1,715
1,688
1,635
1,662
1,688
Glass Production
1,535
1,928
1,317
1,336
1,299
1,249
1,315
Aluminum Production
6,831
4,142
3,255
2,833
2,767
1,334
1,205
Phosphoric Acid Production
1,529
1,342
1,149
1,038
999
998
1,023
Zinc Production
632
1,030
1,429
956
933
925
1,009
Lead Production
516
553
546
459
473
450
455
Silicon Carbide Production
and Consumption
375
219
169
173
180
174
186
Abandoned Oil and Gas Wells
6
7
7
7
7
7
7
Trends 2-5
-------
Magnesium Production and
Processing
1
3
Wood Biomass, Ethanol, and
Biodiesel Consumption"
219,413
230,700
International Bunker Fuelsb
103,463
113,139
CH4c
31,194
27,657
Enteric Fermentation
6,566
6,755
Natural Gas Systems
7,723
6,856
Landfills
7,182
5,256
Manure Management
1,486
2,150
Coal Mining
3,860
2,565
Petroleum Systems
1,682
1,469
Wastewater Treatment
611
616
Rice Cultivation
641
667
Stationary Combustion
344
313
Abandoned Oil and Gas Wells
262
277
Abandoned Underground
Coal Mines
288
264
Mobile Combustion
518
384
Composting
15
75
Petrochemical Production
9
3
Field Burning of Agricultural
Residues
4
7
Ferroalloy Production
1
+
Silicon Carbide Production
and Consumption
1
+
Iron and Steel Production &
Metallurgical Coke
Production
1
1
Incineration of Waste
+
+
International Bunker Fuelsb
7
5
N2Oc
1,243
1,261
Agricultural Soil Management
845
854
Stationary Combustion
84
115
Manure Management
47
55
Mobile Combustion
141
131
Nitric Acid Production
41
38
Adipic Acid Production
51
24
Wastewater Treatment
11
15
N2O from Product Uses
14
14
Composting
1
6
Caprolactam, Glyoxal, and
Glyoxylic Acid Production
6
7
Incineration of Waste
2
1
Semiconductor Manufacture
+
+
Field Burning of Agricultural
Residues
+
+
Petroleum Systems
+
+
Natural Gas Systems
+
+
International Bunker Fuelsb
3
3
HFCs
M
M
Substitution of Ozone
Depleting Substances'1
M
M
HCFC-22 Production
3
1
Semiconductor Manufacture
M
M
Magnesium Production and
Processing
0
0
PFCs
M
M
315,545
99,763
26,522
6,620
6,624
4,517
2,322
2,584
1,665
572
462
350
282
323,187
103,400
26,483
6,568
6,603
4,502
2,311
2,583
1,682
572
510
355
283
3
317,742
110,887
26,456
6,661
6,686
4.448
2,435
2.449
1,579
579
493
340
285
3
317,191
116,594
26,196
6,875
6,629
4,319
2,461
2,154
1,528
568
549
318
289
3
322,225
120,107
26,253
7,018
6,624
4,309
2,467
2,227
1,506
568
454
312
277
249
253
256
268
257
181
163
143
135
128
81
84
85
85
86
3
5
7
10
10
8
+
8
1
8
1
8
1
8
1
3
3
3
4
4
1,226
1,217
1,255
1,223
1,210
890
880
932
898
894
110
111
103
101
96
58
58
59
61
63
74
68
63
60
57
36
37
39
34
31
13
18
14
23
25
16
16
16
16
17
14
14
14
14
14
6
6
6
6
6
7
7
7
7
5
1
1
1
1
1
1
+
1
+
1
+
1
+
1
+
+
+
3
+
+
3
+
+
3
+
+
3
+
+
3
M
M
M
M
M
M
M
M
M
M
_1_
1
M
1
M
1
M
1
M
1
M
M
M
M
M
M
2-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Semiconductor Manufacture
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
+
+
+
+
+
+
Magnesium Production and
Processing
+
+
+
+
+
+
+
Semiconductor Manufacture
+
+
+
+
+
+
+
NF3
+
+
+
+
+
+
+
Semiconductor Manufacture
+
+
+
+
+
+
+
+ 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 N2O are reported separately from gross emissions totals. Refer to Lable 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.
Notes: Lotals may not sum due to independent rounding. Parentheses indicate negative values or sequestration.
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-eight-
year period of 1990 to 2017, total emissions from the Energy, Industrial Processes and Product Use, and Agriculture
sectors grew by 85.1 MMT CO2 Eq. (1.6 percent), 16.8 MMT CO2 Eq. (4.9 percent), and 51.8 MMT CO2 Eq. (10.6
percent), respectively. Emissions from the Waste sector decreased by 68.0 MMT CO2 Eq. (34.2 percent). Over the
same period, total C sequestration in the Land Use, Land-Use Change, and Forestry (LULUCF) sector decreased by
85.2 MMT CO2 (10.5 percent decrease in total C sequestration), and emissions from the LULUCF sector increased
by 7.7 MMT CO2 Eq. (99.1 percent).
Figure 2-4: U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC Sector (MMT CO2
Eq.)
7,000 Waste
6,000
5,000
4,000
Industrial Processes and Product Use
Agriculture
, LULUCF (emissions)
O
u
3,000
2,000
1,000
-1,000
Energy
Land Use, Land-Use Change and Forestry (LULUCF) (removals)
O
Trends 2-7
-------
Table 2-3: Recent Trends in U.S. Greenhouse Gas Emissions and Sinks by Chapter/IPCC
Sector (MMT COz Eq.)
Chapter/IPCC Sector
1990
2005
2013
2014
2015
2016
2017
Energy
5,339.8
6,308.0
5,695.0
5,736.4
5,584.7
5,465.3
5,424.8
Fossil Fuel Combustion
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
Natural Gas Systems
223.1
194.0
190.8
190.6
192.2
191.2
191.9
Non-Energy Use of Fuels
119.6
139.6
123.5
119.9
126.9
113.7
123.2
Petroleum Systems
51.0
48.3
66.8
71.7
71.2
60.4
61.0
Coal Mining
96.5
64.1
64.6
64.6
61.2
53.8
55.7
Stationary Combustion
33.7
42.2
41.5
41.9
39.0
38.0
36.4
Mobile Combustion
55.0
48.6
26.6
24.3
22.4
21.2
20.1
Incineration of Waste
8.4
12.9
10.6
10.7
11.1
11.1
11.1
Abandoned Oil and Gas Wells
6.6
6.9
7.0
7.1
7.1
7.2
6.9
Abandoned Underground Coal Mines
7.2
6.6
6.2
6.3
6.4
6.7
6.4
Industrial Processes and Product Use
342.1
358.0
353.1
365.2
360.8
354.6
358.9
Substitution of Ozone Depleting
Substances
0.3
102.1
141.7
145.3
149.2
151.8
152.7
Iron and Steel Production &
Metallurgical Coke Production
101.7
68.2
53.5
58.4
47.8
42.3
41.8
Cement Production
33.5
46.2
36.4
39.4
39.9
39.4
40.3
Petrochemical Production
21.4
26.9
26.5
26.6
28.2
28.4
28.5
Ammonia Production
13.0
9.2
9.5
9.4
10.6
10.8
13.2
Lime Production
11.7
14.6
14.0
14.2
13.3
12.9
13.1
Other Process Uses of Carbonates
6.3
7.6
11.5
13.0
12.2
11.0
10.1
Nitric Acid Production
12.1
11.3
10.7
10.9
11.6
10.1
9.3
Adipic Acid Production
15.2
7.1
3.9
5.4
4.3
7.0
7.4
HCFC-22 Production
46.1
20.0
4.1
5.0
4.3
2.8
5.2
Urea Consumption for Non-
Agricultural Purposes
3.8
3.7
4.6
1.8
4.6
5.1
5.0
Semiconductor Manufacture
3.6
4.7
4.6
4.8
4.9
5.0
5.0
Carbon Dioxide Consumption
1.5
1.4
4.2
4.5
4.5
4.5
4.5
Electrical Transmission and
Distribution
23.1
8.3
4.4
4.6
4.1
4.4
4.3
N2O from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
Aluminum Production
28.3
7.6
6.2
5.4
4.8
2.7
2.3
Ferroalloy Production
2.2
1.4
1.8
1.9
2.0
1.8
2.0
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.7
1.8
Titanium Dioxide Production
1.2
1.8
1.7
1.7
1.6
1.7
1.7
Caprolactam, Glyoxal, and Glyoxylic
Acid Production
1.7
2.1
2.0
2.0
2.0
2.0
1.4
Glass Production
1.5
1.9
1.3
1.3
1.3
1.2
1.3
Magnesium Production and
Processing
5.2
2.7
1.4
1.0
1.1
1.2
1.2
Phosphoric Acid Production
1.5
1.3
1.1
1.0
1.0
1.0
1.0
Zinc Production
0.6
1.0
1.4
1.0
0.9
0.9
1.0
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Silicon Carbide Production and
Consumption
0.4
0.2
0.2
0.2
0.2
0.2
0.2
Agriculture
490.2
518.4
526.3
522.8
543.8
541.2
542.1
Agricultural Soil Management
251.7
254.5
265.2
262.3
277.8
267.6
266.4
Enteric Fermentation
164.2
168.9
165.5
164.2
166.5
171.9
175.4
Manure Management
51.1
70.2
75.5
75.2
78.5
79.7
80.4
Rice Cultivation
16.0
16.7
11.5
12.7
12.3
13.7
11.3
Urea Fertilization
2.4
3.5
4.4
4.5
4.7
4.9
5.1
Liming
4.7
4.3
3.9
3.6
3.7
3.2
3.2
Field Burning of Agricultural
Residues
0.2
0.3
0.3
0.3
0.3
0.3
0.3
2-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Waste
198.9
154.7
135.8
135.6
134.5
131.1
131.0
Landfills
179.6
131.4
112.9
112.5
111.2
108.0
107.7
Wastewater Treatment
18.7
19.8
19.0
19.1
19.3
19.1
19.2
Composting
0.7
3.5
3.9
4.0
4.0
4.0
4.1
Total Emissions3
6,371.0
7,339.0
6,710.2
6,760.0
6,623.8
6,492.3
6,456.7
Land Use, Land-Use Change, and
Forestry
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
Forest land
(788.0)
(750.2)
(726.4)
(678.6)
(744.4)
(741.0)
(733.1)
Cropland
34.6
40.1
55.6
54.7
60.4
57.4
56.6
Grassland
4.7
11.3
4.9
1.2
20.0
7.5
8.9
Wetlands
(0.5)
(2.0)
(0.7)
(0.6)
(0.7)
(0.7)
(0.7)
Settlements
(57.8)
(39.2)
(46.9)
(46.7)
(46.4)
(45.8)
(45.9)
Net Emission (Sources and Sinks)b
5,564.0
6,599.0
5,996.8
6,090.0
5,912.7
5,769.7
5,742.6
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 2017. 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 Figure 2-6).
In 2017, 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 in the Energy chapter. Energy-related activities are
also responsible for CH4 and N20 emissions (43 percent and 13 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: 2017 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-C02 Emissions from Stationary Combustion
Non-C02 Emissions from Mobile Combustion
Incineration of Waste
Abandoned Oil and Gas Wells
Abandoned Underground Coal Mines
I
Energy as a Portion of All
Emissions
50 100 150 200
MMT CO2 Eq.
250
300
Trends 2-9
-------
Figure 2-6: 2017 U.S. Fossil Carbon Flows (MMT CO2 Eq.)
International.
Bunkers ,
NEU Emissions 13
Fossil Fuel
Energy Exports
1,176
Coal Emissions
UB4
— NEU Emissions 5
Natural Gas Emissions
Coal
" Combustion
Emissions 1452
NEU Emissions 106
Domestic
Fossil Fuel
Production
Petroleum
1.464
Natural Gas Liquids,
liquefied Refinery Gas,
& Other Liquids
Fossil Fuel
Energy
Imports
Non-Energy Lfee
Carton Sequestered
Petroleum
1.395 „
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
Fossil Fuel
Consumption
U.S
Territories
Stock
Changes
027)
Natural Gas 165'
Coal 18"
NEU Imports
Table 2-4: Emissions from Energy (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
4,905.3
5,931.0
5,341.5
5,384.8
5,241.5
5,134.1
5,095.6
Fossil Fuel Combustion
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
Tmnsportation
1.469.1
1,857.0
1,682.7
1,721.6
1,734.0
1,779.0
1,800.6
Electricity Generation
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
Industrial
857.5
853.4
840.0
819.6
807.9
807.6
810.7
Residential
338.2
357.9
329.3
346.8
317.8
292.9
294.5
Commercial
226.5
226.8
224.6
232.9
245.5
232.1
232.9
U.S. Territories
27.6
49.7
42.5
41.4
41.4
41.4
41.4
Non-Energy Use of Fuels
119.6
139.6
123.5
119.9
126.9
113.7
123.2
Natural Gas Systems
30.0
22.6
25.1
25.5
25.1
25.5
26.3
Petroleum Systems
9.0
11.6
25.1
29.6
31.7
22.2
23.3
Incineration of Waste
8.0
12.5
10.3
10.4
10.7
10.8
10.8
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Biomass-Wood"
215.2
206.9
227.3
233.8
224.7
216.3
221.4
International Bunker Fuelsh
103.5
113.1
99.8
103.4
110.9
116.6
120.1
Biofitels-Ethanol"
4.2
22.9
74.7
76.1
78.9
81.2
82.1
Biofuels-Biodiesel"
0.0
0.9
13.5
13.3
14.1
19.6
18.7
CH4
366.9
303.2
298.4
298.1
293.5
283.0
283.3
Natural Gas Systems
193.1
171.4
165.6
165.1
167.2
165.7
165.6
Coal Mining
96.5
64.1
64.6
64.6
61.2
53.8
55.7
Petroleum Systems
42.1
36.7
41.6
42.1
39.5
38.2
37.7
Stationary Combustion
8.6
7.8
8.7
8.9
8.5
7.9
7.8
Abandoned Oil and Gas Wells
6.6
6.9
7.0
7.1
7.1
7.2
6.9
Abandoned Underground Coal
Mines
7.2
6.6
6.2
6.3
6.4
6.7
6.4
Mobile Combustion
12.9
9.6
4.5
4.1
3.6
3.4
3.2
Incineration of Waste
+
+
+
+
+
+
+
International Bunker Fuelsb
0.2
0.1
0.1
0.1
0.1
0.1
0.1
N2O
67.6
73.7
55.2
53.5
49.7
48.3
45.9
Stationary Combustion
25.1
34.3
32.7
33.0
30.6
30.1
28.6
2-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Mobile Combustion
42.0
39.0
22.1
20.2
18.8
17.9
16.9
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Petroleum Systems
+
+
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker Fuelsb
0.9
1.0
0.9
0.9
0.9
1.0
1.0
Total
5,339.8
6,308.0
5,695.0
5,736.4
5,584.7
5,465.3
5,424.8
+ Does not exceed 0.05 MMT CO2 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.
Note: Totals may not sum due to independent rounding.
CO2 Emissions from Fossil Fuel Combustion
As the largest contributor to U.S. greenhouse gas emissions, CO2 from fossil fuel combustion lias accounted for
approximately 77 percent of GWP-weighted emissions for the entire time series since 1990. Emissions from this
source category grew by 3.7 percent (173.2 MMT CO2 Eq.) from 1990 to 2017 and were responsible for most of the
increase in national emissions during this period. Conversely, CO2 emissions from fossil fuel combustion decreased
from 2005 levels by 832.8 MMT CO2 Eq., a decrease of approximately 14.5 percent between 2005 and 2017. From
2016 to 2017, these emissions decreased by 1.0 percent (49.9 MMT CO2 Eq.). Historically, changes in emissions
from fossil fuel combustion have been the dominant factor affecting U.S. emission trends.
Changes in CO2 emissions from fossil fuel combustion are influenced 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 sector, which leads to a decrease in emissions from reduced fuel consumption.
Energy-related 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-41 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 2-7). In recent years, the types of fuel consumed to produce electricity have changed. Total electric power
generation remained relatively flat over the past five years, but emissions have decreased due to a decreasing
reliance on coal used to generate electricity and increased generation from natural gas and renewable energy
sources. Carbon dioxide emissions from coal consumption for electric power generation decreased by 23.2 percent
since 2013, which can be largely attributed to a shift to the use of less-C02-intensive natural gas to generate
electricity and a rapid increase in the use of renewable energy in the electric power sector in recent years. Electricity
generation from renewable sources increased by 35.9 percent from 2013 to 2017 and natural gas generation
increased by 16.3 percent over the same time period (see Table 3-12 for more detail on electricity generation by
source). The decrease in coal-powered electricity generation and increase in natural gas and renewable energy
electricity generation have contributed to a 15.0 percent decrease in overall CO2 emissions from electric power
generation from 2013 to 2017 (see Figure 2-9).
The trends in CO2 emissions from fossil fuel combustion over the past five years also follow changes in heating
degree days. Carbon dioxide emissions from natural gas consumption in the residential and commercial sectors
decreased by 9.3 percent and 3.4 percent from 2013 to 2017, respectively. This trend can be largely attributed to a
14 percent decrease in heating degree days, which led to a decreased demand for heating fuel and electricity for heat
in these sectors. In addition an increase in energy efficiency standards and the use of energy-efficient products in
residential and commercial buildings lias resulted in an overall reduction in energy use, contributing to a decrease in
Trends 2-11
-------
C02 emissions in both of these sectors (EIA 2018). Combined residential and commercial sector CO2 emissions
decreased by 4.8 percent from 2013 to 2017.
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 2017.
Despite the overall decreasing trend in CO2 emissions from fossil fuel combustion over the past five years,
emissions from petroleum consumption for transportation (including bunkers) have increased by 7.5 percent since
2013; this trend can be primarily attributed to a 7.5 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.
Although CO2 emissions from the transportation sector have increased over the past five years, CO2 emissions from
all other sectors and U.S. Territories have decreased in recent years, contributing to a 1.0 percent decrease in total
CO2 emissions from fossil fuel combustion from 2016 to 2017 and a 4.8 percent reduction since 2013.
Carbon dioxide emissions from fossil fuel combustion are presented in Table 2-5 based on the underlying U.S.
energy consumer data collected by the U.S. Energy Information Administration (EIA). Estimates of CO2 emissions
from fossil fuel combustion are calculated from these EIA "end-use sectors" based on total fuel consumption and
appropriate fuel properties described below. (Any additional analysis and refinement of the EIA data is further
explained in the Energy chapter of this report.)
• Transportation. EIA'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. EIA'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. EIA'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. EIA's fuel consumption data for the residential sector consist of living quarters for private
households.
• Commercial. EIA'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 and Figure 2-7 summarize CO2 emissions from fossil fuel combustion by end-use sector.
Figure 2-8 further describes direct and indirect CO2 emissions from fossil fuel combustion, separated by end-use
sector.
Table 2-5: CO2 Emissions from Fossil Fuel Combustion by End-Use Sector (MMT CO2 Eq.)
End-Use Sector
1990
2005
2013
2014
2015
2016
2017
Transportation
1,472.1
1,861.7
1,686.9
1,726.0
1,738.2
1,783.2
1,804.9
Combustion
1,469.1
1,857.0
1,682.7
1,721.6
1,734.0
1,779.0
1,800.6
Electricity
3.0
4.7
4.3
4.5
4.3
4.2
4.3
Industrial
1,543.9
1,589.7
1,434.8
1,412.5
1,357.4
1,325.2
1,315.1
Combustion
857.5
853.4
840.0
819.6
807.9
807.6
810.7
Electricity
686.4
736.3
594.8
593.0
549.5
517.6
504.4
Residential
930.9
1,213.9
1,064.1
1,080.9
1,001.6
946.3
911.5
2-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Combustion
338.21
357.9
329.3
346.8
317.8
292.9
294.5
Electricity
592.7
856.0
734.7
734.1
683.8
653.5
617.1
Commercial
764.3
1,029.7
929.1
938.5
908.5
865.8
839.1
Combustion
226.5
226.8
224.6
232.9
245.5
232.1
232.9
Electricity
537.7
803.0
704.5
705.6
663.0
633.6
606.2
U.S. Territories3
27.6
49.7
42.5
41.4
41.4
41.4
41.4
Total
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
Electric Power
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
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.
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.
Figure 2-7: 2017 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT
COz Eq.)
Relative Contribution by Fuel Type
2,500
2,000
1,732
Petroleum
Coal
Natural Gas
- 1,500
1,000
U.S. Territories
Commercial
Residential
Industrial
Electric Power Transportation
Note on Figure 2-7: Fossil Fuel Combustion for electric power also includes emissions of less than 0.5 MMT CO2 Eq. from
geothermal-based generation.
Figure 2-8: 2017 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
Trends 2-13
-------
Electric power was the second largest emitter of CO2 in 2017 (surpassed by transportation); electricity generators
used 32 percent of U.S. energy from fossil fuels and emitted 35 percent of the CO2 from fossil fuel combustion in
2017. 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 4.8 percent since 1990, and the carbon intensity of the electric power sector, in terms of CO2 Eq. per
QBtu input, lias significantly decreased by 11 percent during that same timeframe. This decoupling of the level of
electric power generation and the resulting CO2 emissions is shown below in Figure 2-9.
Figure 2-9: Electric Power Generation (Billion kWh) and Emissions (MMT CO2 Eq.)
Nuclear Generation (Billion kWh)
Renewable Generation (Billion kWh)
Petroleum Generation (Billion kWh)
Natural Gas Generation (Billion kWh)
Coal Generation (Billion kWh)
I Total Emissions (MMT CO2 Eq.) [Right Axis]
3,500
3,000
2,500
2,000 S
c
o
1,500
to
1,000 O
500
cnoi-HpNjm^-LovDrv
O 1—I 1—I 1—I l-H 1—I 1—I 1—I T—I
OOOOOOOOO
Electric power CO2 emissions can also be allocated to the end-use sectors that use electricity, as presented in Table
2-5. With electricity CO2 emissions allocated to end-use sectors, the transportation end-use sector accounted for
1,804.9 MMT CO2 Eq. in 2017 or approximately 37 percent of total CO2 emissions from fossil fuel combustion. The
industrial end-use sector accounted for 27 percent of CO2 emissions from fossil fuel combustion when including
allocated electricity emissions. The residential and commercial end-use sectors accounted for 19 and 17 percent,
respectively, of CO2 emissions from fossil fuel combustion when including allocated electricity emissions. Both of
these end-use sectors were heavily reliant on electricity for meeting energy needs, with electricity use for lighting,
heating, air conditioning, and operating appliances contributing 68 and 72 percent of emissions from the residential
and commercial end-use sectors, respectively.
Other Significant Trends in Energy
Other significant trends in emissions from energy source categories over the twenty-eight-year period from 1990
through 2017 included the following:
• Methane emissions from natural gas systems and petroleum systems (combined here) decreased from 235.1
MMT CO2 Eq. in 1990 to 203.3 MMT CO2 Eq. in2017 (31.9 MMT CO2 Eq. or 13.6 percent decrease from
1990 to 2017). Natural gas systems CH4 emissions decreased by 27.5 MMT CO2 Eq. (14.2 percent) since
1990, largely due to a decrease in emissions from distribution transmission and storage, processing, and
exploration. The decrease in distribution is largely 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 leaks). Petroleum systems CH4
2-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
emissions decreased by 4.4 MMT CO2 Eq. (or 10.5 percent) since 1990. This decrease is due primarily to
decreases in tank emissions and associated gas venting. Carbon dioxide emissions from natural gas and
petroleum systems increased by 27 percent from 1990 to 2017, due to increases in flaring emissions.
• Carbon dioxide emissions from non-energy uses of fossil fuels increased by 3.7 MMT CO2 Eq. (3.1
percent) from 1990 through 2017. Emissions from non-energy uses of fossil fuels were 123.2MMT CO2
Eq. in 2017, which constituted 2.3 percent of total national CO2 emissions, approximately the same
proportion as in 1990.
• Methane emissions from coal mining decreased by 40.8 MMT CO2 Eq. (42.3 percent) from 1990 through
2017, primarily due to a decrease in the number of active mines and annual coal production over the time
period.
• Nitrous oxide emissions from stationary combustion increased by 3.5 MMT CO2 Eq. (14.1 percent) from
1990 through 2017. 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 25.1 MMT CO2 Eq. (59.7 percent) from
1990 through 2017, primarily as a result of N20 national emission control standards and emission control
technologies for on-road vehicles.
• Carbon dioxide emissions from incineration of waste (10.8 MMT CO2 Eq. in 2017) increased by 2.8 MMT
CO2 Eq. (35.7 percent) from 1990 through 2017, 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 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, CH4, N20, and fluorinated gases (e.g., HFC-23). These processes are shown
in Figure 2-10. Industrial manufacturing processes and use by end-consumers also release HFCs, PFCs, SF6, and
NF3 and other fluorinated compounds. In addition to the use of HFCs and some PFCs as substitutes for ozone
depleting substances (ODS), fluorinated compounds such as HFCs, PFCs, SF6, NF3, and others are employed and
emitted by a number of other industrial sources in the United States. These industries include semiconductor
manufacture, electric power transmission and distribution, and magnesium metal production and processing. In
addition, N20 is used in and emitted by semiconductor manufacturing and anesthetic and aerosol applications. Table
2-6 presents greenhouse gas emissions from industrial processes 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
2017.
Trends 2-15
-------
Figure 2-10: 2017 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
Other Process Uses of Carbonates
Nitric Acid Production
Adipic Acid Production
HCFC-22 Production
Urea Consumption for Non-Agricultural Purposes
Semiconductor Manufacture
Carbon Dioxide Consumption
Electrical Transmission and Distribution
N2O from Product Uses
Aluminum Production
Ferroalloy Production
Soda Ash Production
Titanium Dioxide Production
Caprolactam, Glyoxal, and Glyoxylic Acid Production
Glass Production
Magnesium Production and Processing
Phosphoric Acid Production
Zinc Production
Lead Production
Silicon Carbide Production and Consumption
153
I
Industrial Processes and
Product Use as a Portion of
All Emissions
5.6%
¦
¦
¦
¦
I
I
I
I
I
< 0.5
< 0.5
10
20
30
40
50
60
70
MMT COz Eq.
Table 2-6: Emissions from Industrial Processes and Product Use (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
208.8
191.7
173.1
179.2
173.0
164.5
166.9
Iron and Steel Production & Metallurgical Coke
Production
101.6
68.2
53.5
58.4
47.8
42.3
41.8
Iron and Steel Production
99.1
66.2
51.6
56.3
45.0
41.0
41.2
Metallurgical Coke Production
2.5
2.1
1.8
2.0
2.8
1.3
0.6
Cement Production
33.5
46.2
36.4
39.4
39.9
39.4
40.3
Petrochemical Production
21.2
26.8
26.4
26.5
28.1
28.1
28.2
Ammonia Production
13.0
9.2
9.5
9.4
10.6
10.8
13.2
Lime Production
11.7
14.6
14.0
14.2
13.3
12.9
13.1
Other Process Uses of Carbonates
6.3
7.6
11.5
13.0
12.2
11.0
10.1
Urea Consumption for Non-Agricultural
Purposes
3.8
3.7
4.6
1.8
4.6
5.1
5.0
Carbon Dioxide Consumption
1.5
1.4
4.2
4.5
4.5
4.5
4.5
Ferroalloy Production
2.2
1.4
1.8
1.9
2.0
1.8
2.0
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.7
1.8
Titanium Dioxide Production
1.2
1.8
1.7
1.7
1.6
1.7
1.7
Glass Production
1.5
1.9
1.3
1.3
1.3
1.2
1.3
Aluminum Production
6.8
4.1
3.3
2.8
2.8
1.3
1.2
Phosphoric Acid Production
1.5
1.3
1.1
1.0
1.0
1.0
1.0
Zinc Production
0.6
1.0
1.4
1.0
0.9
0.9
1.0
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Silicon 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
0.1
0.2
0.2
0.3
0.3
Petrochemical Production
0.2
0.1
0.1
0.1
0.2
0.2
0.3
Ferroalloy Production
+
+
+
+
+
+
+
2-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Silicon Carbide Production and Consumption
+
+
+
+
+
+
+
Iron and Steel Production & Metallurgical Coke
Production
+
+
+
+
+
+
+
Iron and Steel Production
+
+
+
+
+
+
+
Metallurgical Coke Production
0.0
0.0
0.0
0.0
0.0
0.0
0.0
N2O
33.3
24.9
21.0
22.8
22.3
23.6
22.6
Nitric Acid Production
12.1
11.3
10.7
10.9
11.6
10.1
9.3
Adipic Acid Production
15.2
7.1
3.9
5.4
4.3
7.0
7.4
N2O from Product Uses
4.2
4.2
4.2
4.2
4.2
4.2
4.2
Caprolactam, Glyoxal, and Glyoxylic Acid
1.7
2.1
2.0
2.0
2.0
2.0
1.4
Semiconductor Manufacturing
+
0.1
0.2
0.2
0.2
0.2
0.2
HFCs
46.6
122.3
146.1
150.7
153.8
155.0
158.3
Substitution of Ozone Depleting Substances3
0.3
102.1
141.7
145.2
149.2
151.7
152.7
HCFC-22 Production
46.1
20.0
4.1
5.0
4.3
2.8
5.2
Semiconductor Manufacturing
0.2
0.2
0.3
0.3
0.3
0.3
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
5.9
5.6
5.1
4.4
4.1
Semiconductor Manufacturing
2.8
3.2
2.9
3.1
3.1
3.0
3.0
Aluminum Production
21.5
3.4
3.0
2.5
2.0
1.4
1.1
Substitution of Ozone Depleting Substances
0.0
+
+
+
+
+
+
SF«
28.8
11.8
6.3
6.3
5.8
6.3
6.1
Electrical Transmission and Distribution
23.1
8.3
4.4
4.6
4.1
4.4
4.3
Magnesium Production and Processing
5.2
2.7
1.3
0.9
1.0
1.1
1.1
Semiconductor Manufacturing
0.5
0.7
0.7
0.7
0.7
0.9
0.7
NF3
+
0.5
0.5
0.5
0.6
0.6
0.6
Semiconductor Manufacturing
+
0.5
0.5
0.5
0.6
0.6
0.6
Total
342.1
358.0
353.1
365.2
360.8
354.6
358.9
+ Does not exceed 0.05 MMT CO2 Eq.
a Small amounts of PFC emissions also result from this source.
Note: Totals may not sum due to independent rounding.
Overall, emissions from the IPPU sector increased by 4.9 percent from 1990 to 2017. Significant trends in emissions
from IPPU source categories over the twenty-eight-year period from 1990 through 2017 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 152.7 MMT CO2 Eq. in
2017. 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 interim substitutes in many applications, are themselves
phased-out under the provisions of the Copenhagen Amendments to the Montreal Protocol.
• Combined CO2 and CH4 emissions from iron and steel production and metallurgical coke production
decreased by 1.2 percent to 41.8 MMT CO2 Eq. from 2016 to 2017, and have declined overall by 59.9
MMT CO2 Eq. (58.9 percent) from 1990 through 2017, due to restructuring of the industry, technological
improvements, and increased scrap steel utilization.
• Carbon dioxide emissions from cement production increased by 20.4 percent (6.8 MMT CO2 Eq.) from
1990 through 2017. They rose from 1990 through 2006 (with the exception of a slight decrease in 1997)
and then fell until 2009 due to a decrease in 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.).
• Carbon dioxide emissions from ammonia production (13.2 MMT CO2 Eq. in 2017) decreased by 1.3
percent (0.2 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.
• Nitrous oxide emissions from adipic acid production were 7.4 MMT CO2 Eq. in 2017, and have decreased
significantly since 1990 due to both the widespread installation of pollution control measures in the late
Trends 2-17
-------
1990s and plant idling in the late 2000s. Emissions fromadipic acid production have decreased by 51.5
percent since 1990 and by 56.3 percent since a peak in 1995.
• PFC emissions from aluminum production decreased by 94.8 percent (20.3 MMT CO2 Eq.) from 1990 to
2017, due to both industry emission reduction efforts and lower domestic aluminum production.
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, N20, and CO2 were the primary greenhouse gases emitted by agricultural activities.
In 2017, agricultural activities were responsible for emissions of 542.1 MMT CO2 Eq., or 8.4 percent of total U.S.
greenhouse gas emissions. Methane emissions from enteric fermentation and manure management represented
approximately 26.7 percent and 9.4 percent of total CH4 emissions from anthropogenic activities, respectively, in
2017. Agricultural soil management activities, such as application of synthetic and organic fertilizers, deposition of
livestock manure, and growing N-fixing plants, were the largest source of U.S. N20 emissions in 2017, accounting
for 73.9 percent. Carbon dioxide emissions from the application of crushed limestone and dolomite (i.e., soil liming)
and urea fertilization represented 0.2 percent of total CO2 emissions from anthropogenic activities. Figure 2-11 and
Table 2-7 illustrate agricultural greenhouse gas emissions by source.
Figure 2-11: 2017 Agriculture Chapter Greenhouse Gas Sources (MMT CO2 Eq.)
Agricultural Soil Management
Enteric Fermentation
Manure Management
Rice Cultivation
Urea Fertilization
Liming
Agriculture as a Portion of
All Emissions
Field Burning of Agricultural Residues < 0.5
I
266
80 100 120 140 160 180 200 220 240 260
MMT CO2 Eq.
Table 2-7: Emissions from Agriculture (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
7.1
7.9
8.4
8.1
8.5
8.1
8.2
Urea Fertilization
2.4
3.5
4.4
4.5
4.7
4.9
5.1
Liming
4.7
4.3
3.9
3.6
3.7
3.2
3.2
CH4
217.4
239.5
235.3
234.9
239.9
247.3
248.7
Enteric Fermentation
164.2
168.9
165.5
164.2
166.5
171.9
175.4
Manure Management
37.1
53.7
58.1
57.8
60.9
61.5
61.7
Rice Cultivation
16.0
16.7
11.5
12.7
12.3
13.7
11.3
Field Burning of Agricultural
Residues
0.1
0.2
0.2
0.2
0.2
0.2
0.2
N2O
265.7
271.1
282.7
279.7
295.4
285.8
285.2
Agricultural Soil Management
251.7
254.5
265.2
262.3
277.8
267.6
266.4
Manure Management
14.0
16.5
17.4
17.4
17.6
18.2
18.7
Field Burning of Agricultural
Residues
+
0.1
0.1
0.1
0.1
0.1
0.1
2-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Total 490.2 518.4 526.3 522.8 543.8 541.2 542.1
+ Does not exceed 0.05 MMT CO2 Eq.
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 N20 emissions in the United States, accounting
for approximately 73.9 percent of N20 emissions in 2017 and 4.1 percent of total emissions in the United
States in 2017. Estimated emissions from this source in 2017 were 266.4 MMT CO2 Eq. Annual N20
emissions from agricultural soils fluctuated between 1990 and 2017, although overall emissions were 5.8
percent higher in 2017 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 2017,
enteric fermentation CH4 emissions were 26.7 percent of total CH4 emissions (175.4 MMT CO2 Eq.),
which represents an increase of 11.3 MMT CO2 Eq. (6.9 percent) since 1990. This increase in emissions
from 1990 to 2017 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 2017, consistent with an
increase in beef cattle population over those same years.
• Overall, emissions from manure management increased 57.1 percent between 1990 and 2017. This
encompassed an increase of 66.0 percent for CH i. from 37.1 MMT CO2 Eq. in 1990 to 61.7 MMT CO2 Eq.
in 2017; and an increase of 33.6 percent for N2O, from 14.0 MMT CO2 Eq. in 1990 to 18.7 MMT CO2 Eq.
in 2017. The majority of the increase observed in CH4 resulted from swine and dairy cattle manure, where
emissions increased 29 and 134 percent, respectively, from 1990 to 2017. From 2016 to 2017, there was a
0.2 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 source of CO2 emissions reported in the Agriculture sector.
Estimated emissions from these sources were 3.2 and 5.1 MMT CO2 Eq., respectively. Liming emissions
decreased by 0.7 percent relative to 2016 and 31.8 percent relative to 1990, while urea fertilization
emissions increased by 3.6 percent relative to 2016 and 109.0 percent relative to 1990.
Land Use, Land-Use Change, and Forestry
When humans alter the terrestrial biosphere through land use, changes in land use, and land management practices,
they also influence the carbon (C) stock fluxes on these lands and cause emissions of CH4 and N20. Overall,
managed land is a net sink for 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.
The LULUCF sector in 2017 resulted in a net increase in C stocks (i.e., net CO2 removals) of 729.6 MMT CO2 Eq.
(Table 2-8).2 This represents an offset of approximately 11.3 percent of total (i.e., gross) greenhouse gas emissions
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,
Trends 2-19
-------
in 2017. Emissions of CH4 and N2O from LULUCF activities in 2017 were 15.5 MMT CO2 Eq. and represent 0.2
percent of total greenhouse gas emissions.3 Between 1990 and 2017, total C sequestration in the LULUCF sector
decreased by 10.5 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 2017, totaling 4.9 MMT CO2 Eq. (194 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). Peatlands Remaining
Peatlands, Land Converted to Wetlands, and Drained Organic Soils resulted in CH4 emissions of less than 0.05
MMT CO2 Eq. each.
Forest fires were also the largest source of N20 emissions from LULUCF in 2017, totaling 3.2 MMT CO2 Eq. (11 kt
of N2O). Nitrous oxide emissions from fertilizer application to settlement soils in 2017 totaled to 2.5 MMT CO2 Eq.
(8 kt of N2O). Additionally, the application of synthetic fertilizers to forest soils in 2017 resulted in N20 emissions
of 0.5 MMT CO2 Eq. (2 kt of N20). Grassland fires resulted in N20 emissions of 0.3 MMT CO2 Eq. (1 kt of N20).
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 Figure 2-12 and Table 2-8 along with CH4 and N20
emissions for LULUCF source categories.
Figure 2-12: 2017 LULUCF Chapter Greenhouse Gas Sources and Sinks (MMT CO2 Eq.)
Forest Land Remaining Forest Land
Settlements Remaining Settlements
Land Converted to Forest Land
Cropland Remaining Cropland
Wetlands Remaining Wetlands
Grassland Remaining Grassland
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
Non-C02 Emissions from Forest Fires
Land Converted to Grassland
Land Converted to Cropland
Land Converted to Settlements
(300) (250) (200) (150) (100) (50) 0 50 100
MMT CO2 Eq.
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
2013 2014 2015 2016 2017
Carbon Stock Change3 (814.8)
Forest Land Remaining Forest Land (671.6)
Land Converted to Forest Land (119.1)
(756.1)
(639.4)
(120.0)
(731.0) (687.8) (739.4) (738.1) (729.6)
(616.7) (568.8) (645.2) (628.9) (621.1)
(120.5) (120.5) (120.6) (120.6) (120.6)
(621.1)
I Non-C02 Emissions
Carbon Stock Change
l< 0.51
l< 0.5|
l< 0.5|
l< 0.51
l< 0.51
l< 0.51
L
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 N2O 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; andN^O emissions from Forest Soils and Settlement Soils.
2-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Cropland Remaining Cropland
(40.9)
(26.5)
(11.4)
(12.0)
(6.3)
(9.9)
(10.3)
Land Converted to Cropland
75.6
66.7
66.9
66.7
66.7
67.3
66.9
Grassland Remaining Grassland
(4.2)
5.5
(3.7)
(7.5)
9.6
(1.6)
(0.1)
Land Converted to Grassland
8.7
5.1
8.3
7.9
9.8
8.5
8.3
Wetlands Remaining Wetlands
(4.0)
(5.7)
(4.3)
(4.3)
(4.4)
(4.4)
(4.4)
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Settlements Remaining Settlements
(122.1)
(127.8)
(135.9)
(135.8)
(135.4)
(134.7)
(134.5)
Land Converted to Settlements
62.9
86.0
86.4
86.5
86.5
86.4
86.2
CH4
5.0
9.0
9.9
10.1
16.5
8.8
8.8
Forest Land Remaining Forest Land:
Forest Firesb
1.5
5.2
6.1
6.1
12.6
4.9
4.9
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
0.2
0.4
0.3
0.3
0.3
Land Converted to Wetlands: Land
Converted to Coastal Wetlands
+
+
+
+
+
+
+
Forest Land Remaining Forest Land:
Drained Organic Soils'1
+
+
+
+
+
+
+
Wetlands Remaining Wetlands:
Peatlands Remaining Peatlands
+
+
+
+
+
+
+
N2O
2.8
7.0
7.6
7.7
11.8
6.7
6.7
Forest Land Remaining Forest Land:
Forest Firesb
1.0
3.4
4.0
4.0
8.3
3.2
3.2
Settlements Remaining Settlements:
Settlement Soils6
1.4
2.5
2.6
2.6
2.5
2.5
2.5
Forest Land Remaining Forest Land:
Forest Soilsf
0.1
0.5
0.5
0.5
0.5
0.5
0.5
Grassland Remaining Grassland:
Grassland Firesc
0.1
0.3
0.2
0.4
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 Soils'1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Wetlands Remaining Wetlands:
Peatlands Remaining Peatlands
+
+
+
+
+
+
+
LULUCF Emissions8
7.8
16.0
17.5
17.7
28.3
15.5
15.5
LULUCF Carbon Stock Change3
(814.8)
(756.1)
(731.0)
(687.8)
(739.4)
(738.1)
(729.6)
LULUCF Sector Net Total"
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
+ Absolute value does not exceed 0.05 MMT CO2 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.
g LULUCF emissions include the CH4 and N2O emissions reported for Peatlands Remaining Peatlands, Forest Fires,
Drained Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CFL emissions from Land
Converted to Coastal Wetlands; andN^O emissions from Forest Soils and Settlement Soils.
h Lhe LULUCF Sector Net Lotal is the net sum of all CFL and N2O emissions to the atmosphere plus net carbon stock
changes.
Notes: Lotals may not sum due to independent rounding. Parentheses indicate net sequestration.
Trends 2-21
-------
Other significant trends from 1990 to 2017 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) lias decreased
by approximately 6.2 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 10.2 percent over the period from
1990 to 2017. This is primarily due to an increase in urbanized land area in the United States.
• Annual emissions from Land Converted to Settlements increased by approximately 37.0 percent from 1990
to 2017 due to losses in aboveground biomass C stocks from Forest Land Converted to Settlements and
mineral soils C stocks from Grassland Converted to Settlements.
• Nitrous oxide emissions from fertilizer application to settlement soils in 2017 totaled to 2.5 MMT CO2 Eq.
(8 kt of N2O). This represents an increase of 72.0 percent since 1990. Additionally, the application of
synthetic fertilizers to forest soils in 2017 resulted in N:0 emissions of 0.5 MMT CO2 Eq. (2 kt of N20).
Nitrous oxide emissions from fertilizer application to forest soils have increased by 455 percent since 1990,
but still account for a relatively small portion of overall emissions.
Waste
Waste management and treatment activities are sources of greenhouse gas emissions (see Figure 2-13). In 2017,
landfills were the third-largest source of U.S. anthropogenic CH4 emissions, accounting for 16.4 percent of total
U.S. CH4 emissions.4 Additionally, wastewater treatment accounts for 14.6 percent of waste emissions, 2.2 percent
of U.S. CH4 emissions, and 1.4 percent of N20 emissions. Emissions of CH4 and N20 from composting are also
accounted for in this chapter, generating emissions of 2.2 MMT CO2 Eq. and 1.9 MMT CO2 Eq., respectively. A
summary of greenhouse gas emissions from the Waste chapter is presented in Table 2-9.
Figure 2-13: 2017 Waste Chapter Greenhouse Gas Sources (MMT CO2 Eq.)
Landfills
Wastewater Treatment
Composting
108
Waste as a Portion of All
Emissions
2.0%
0 10 20 30 40 50 60 70 80 90 100 110
MMT CO2 Eq.
Overall, in 2017, waste activities generated emissions of 131.0 MMT CO2 Eq., or 2.0 percent of total U.S.
greenhouse gas emissions.
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.
2-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 2-9: Emissions from Waste (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CH4
195.2
148.7
129.3
128.9
127.8
124.3
124.1
Landfills
179.6
131.4
112.9
112.5
111.2
108.0
107.7
Wastewater Treatment
15.3
15.4
14.3
14.3
14.5
14.2
14.2
Composting
0.4
1.9
2.0
2.1
2.1
2.1
2.2
N2O
3.7
6.1
6.5
6.6
6.7
6.8
6.9
Wastewater Treatment
3.4
4.4
4.7
4.8
4.8
4.9
5.0
Composting
0.3
1.7
1.8
1.9
1.9
1.9
1.9
Total
198.9
154.7
135.8
135.6
134.5
131.1
131.0
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 2017, net CH4 emissions from landfills decreased by 71.8 MMT CO2 Eq. (40.0 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 since 1990, from 0.7 MMT
CO2 Eq. to 4.1 MMT CO2 Eq. in 2017, which represents slightly less 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 enviromnental benefits of composting; and
(4) loans or grant programs to establish or expand composting infrastructure.
• From 1990 to 2017, CH4 and N20 emissions from wastewater treatment decreased by 1.1 MMT CO2 Eq.
(7.0 percent) and increased by 1.6 MMT CO2 Eq. (46.4 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.
2.2 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.
Using this categorization, transportation activities, in aggregate, accounted for the largest portion (28.9 percent) of
total U.S. greenhouse gas emissions in 2017. Emissions from electric power accounted for the second largest portion
(27.5 percent), while emissions from industry accounted for the third largest portion (22.2 percent) of total U.S.
greenhouse gas emissions in 2017. 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 21.3 percent of U.S. greenhouse gas emissions were contributed by, in order of magnitude, the
agriculture, commercial, and residential sectors, plus emissions from U.S. Territories. Activities related to
agriculture accounted for roughly 9.0 percent of emissions; unlike other economic sectors, agricultural sector
Trends 2-23
-------
emissions were dominated by N20 emissions from agricultural soil management and CH4 emissions from enteric
fermentation rather than CO2 from fossil fuel combustion. The commercial and residential sectors accounted for
roughly 6.4 percent and 5.1 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 2-14 shows the trend in emissions by sector from 1990 to 2017.
Figure 2-14: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq.)
2,500
Electric Power Industry (Purple)
2,000
Transportation (Green)
1,500
Industry
z 1,000
Agriculture
Commercial (Orange)
500
Residential (Blue)
cn o
O rH
o o
Table 2-10: U.S. Greenhouse Gas Emissions Allocated to Economic Sectors (MMT CO2 Eq. and
Percent of Total in 2017)
Sector/Source
1990
2005
2013
2014
2015
2016
2017a
Percent3
Transportation
1,527.1
1,976.0
1,765.4
1,799.9
1,809.3
1,849.7
1,866.2
28.9%
CO2 from Fossil Fuel Combustion
1,469.1
1,857.0
1,682.7
1,721.6
1,734.0
1,779.0
1,800.6
27.9%
Substitution of Ozone Depleting
Substances
+
69.3
51.6
48.8
46.3
43.3
40.1
0.6%
Mobile Combustion
46.1
39.5
21.5
19.5
18.1
17.0
15.9
0.2%
Non-Energy Use of Fuels
11.8
10.2
9.6
10.0
11.0
10.4
9.6
0.1%
Electric Power Industry
1,875.5
2,455.9
2,088.7
2,088.9
1,949.5
1,857.2
1,778.3
27.5%
CO2 from Fossil Fuel Combustion
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
26.8%
Stationary Combustion
20.9
30.9
29.7
29.9
27.7
27.4
25.9
0.4%
Incineration of Waste
8.4
12.9
10.6
10.7
11.1
11.1
11.1
0.2%
Other Process Uses of Carbonates
3.1
3.8
5.8
6.5
6.1
5.5
5.1
0.1%
Electrical Transmission and
Distribution
23.1
8.3
4.4
4.6
4.1
4.4
4.3
0.1%
Industry
1,628.6
1,508.4
1,469.5
1,459.3
1,451.2
1,414.1
1,436.5
22.2%
CO2 from Fossil Fuel Combustion
814.1
803.0
794.7
774.0
767.2
767.9
771.2
11.9%
Natural Gas Systems
223.1
194.0
190.8
190.6
192.2
191.2
191.9
3.0%
Non-Energy Use of Fuels
102.1
121.3
108.4
104.7
110.9
98.2
108.5
1.7%
2-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Petroleum Systems
51.0
48.3
Coal Mining
96.5
64.1
Iron and Steel Production
101.7
68.2
Cement Production
33.5
46.2
Substitution of Ozone Depleting
Substances
+
7.8
Petrochemical Production
21.4
26.9
Ammonia Production
13.0
9.2
Lime Production
11.7
14.6
Nitric Acid Production
12.1
11.3
Adipic Acid Production
15.2
7.1
Abandoned Oil and Gas Wells
6.6
6.9
Abandoned Underground Coal
Mines
7.2
6.6
HCFC-22 Production
46.1
20.0
Other Process Uses of Carbonates
3.1
3.8
Urea Consumption for Non-
Agricultural Purposes
3.8
3.7
Semiconductor Manufacture
3.6
4.7
Carbon Dioxide Consumption
1.5
1.4
N2O from Product Uses
4.2
4.2
Stationary Combustion
4.8
4.6
Mobile Combustion
7.6
7.8
Aluminum Production
28.3
7.6
Ferroalloy Production
2.2
1.4
Soda Ash Production
1.4
1.7
Titanium Dioxide Production
1.2
1.8
Caprolactam, Glyoxal, and
Glyoxylic Acid Production
1.7
2.1
Glass Production
1.5
1.9
Magnesium Production and
Processing
5.2
2.7
Phosphoric Acid Production
1.5
1.3
Zinc Production
0.6
1.0
Lead Production
0.5
0.6
Silicon Carbide Production and
Consumption
0.4
0.2
Agriculture
534.9
570.0
N2O from Agricultural Soil
Management
251.7
254.5
Enteric Fermentation
164.2
168.9
Manure Management
51.1
70.2
CO2 from Fossil Fuel Combustion
43.4
50.4
Rice Cultivation
16.0
16.7
Urea Fertilization
2.4
3.5
Liming
4.7
4.3
Mobile Combustion
1.2
1.2
Field Burning of Agricultural
Residues
0.2
0.3
Stationary Combustion
0.1
+
Commercial
426.9
400.7
CO2 from Fossil Fuel Combustion
226.5
226.8
Landfills
179.6
131.4
Substitution of Ozone Depleting
Substances
+
17.8
Wastewater Treatment
15.3
15.4
Human Sewage
3.4
4.4
Composting
0.7
3.5
66.8
71.7
71.2
60.4
61.0
0.9%
64.6
64.6
61.2
53.8
55.7
0.9%
53.5
58.4
47.8
42.3
41.8
0.6%
36.4
39.4
39.9
39.4
40.3
0.6%
21.2
23.1
25.6
27.9
30.1
0.5%
26.5
26.6
28.2
28.4
28.5
0.4%
9.5
9.4
10.6
10.8
13.2
0.2%
14.0
14.2
13.3
12.9
13.1
0.2%
10.7
10.9
11.6
10.1
9.3
0.1%
3.9
5.4
4.3
7.0
7.4
0.1%
7.0
7.1
7.1
7.2
6.9
0.1%
6.2
6.3
6.4
6.7
6.4
0.1%
4.1
5.0
4.3
2.8
5.2
0.1%
5.8
6.5
6.1
5.5
5.1
0.1%
4.6
1.8
4.6
5.1
5.0
0.1%
4.6
4.8
4.9
5.0
5.0
0.1%
4.2
4.5
4.5
4.5
4.5
0.1%
4.2
4.2
4.2
4.2
4.2
0.1%
4.4
4.3
4.2
4.1
4.2
0.1%
4.3
4.0
3.7
3.6
3.6
0.1%
6.2
5.4
4.8
2.7
2.3
+%
1.8
1.9
2.0
1.8
2.0
+%
1.7
1.7
1.7
1.7
1.8
+%
1.7
1.7
1.6
1.7
1.7
+%
2.0
2.0
2.0
2.0
1.4
+%
1.3
1.3
1.3
1.2
1.3
+%
1.4
1.0
1.1
1.2
1.2
+%
1.1
1.0
1.0
1.0
1.0
+%
1.4
1.0
0.9
0.9
1.0
+%
0.5
0.5
0.5
0.5
0.5
+%
0.2
0.2
0.2
0.2
0.2
+%
572.6
569.2
585.2
581.7
582.2
9.0%
265.2
262.3
277.8
267.6
266.4
4.1%
165.5
164.2
166.5
171.9
175.4
2.7%
75.5
75.2
78.5
79.7
80.4
1.2%
45.4
45.5
40.7
39.7
39.4
0.6%
11.5
12.7
12.3
13.7
11.3
0.2%
4.4
4.5
4.7
4.9
5.1
0.1%
3.9
3.6
3.7
3.2
3.2
+%
0.8
0.8
0.6
0.6
0.6
+%
0.3
0.3
0.3
0.3
0.3
+%
0.1
0.1
0.1
0.1
0.1
+%
409.6
419.5
432.2
416.1
416.0
6.4%
224.6
232.9
245.5
232.1
232.9
3.6%
112.9
112.5
111.2
108.0
107.7
1.7%
47.8
49.6
50.7
51.3
50.6
0.8%
14.3
14.3
14.5
14.2
14.2
0.2%
4.7
4.8
4.8
4.9
5.0
0.1%
3.9
4.0
4.0
4.0
4.1
0.1%
Trends 2-25
-------
Stationary Combustion
1.5
1.4
1.4
1.4
1.6
1.5
1.5
+%
Residential
344.7
370.0
356.3
376.6
349.7
326.9
330.9
5.1%
CO2 from Fossil Fuel Combustion
338.2
357.9
329.3
346.8
317.8
292.9
294.5
4.6%
Substitution of Ozone Depleting
Substances
0.3
7.2
21.1
23.8
26.5
29.3
31.9
0.5%
Stationary Combustion
6.3
4.9
5.9
6.0
5.3
4.7
4.6
0.1%
U.S. Territories
33.3
58.1
48.1
46.6
46.6
46.6
46.6
0.7%
CO2 from Fossil Fuel Combustion
27.6
49.7
42.5
41.4
41.4
41.4
41.4
0.6%
Non-Energy Use of Fuels
5.7
8.1
5.4
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
6,371.0
7,339.0
6,710.2
6,760.0
6,623.8
6,492.3
6,456.7
100.0%
LULUCF Sector Net Total"
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1) (11.1%)
Net Emissions (Sources and
Sinks)
5,564.0
6,599.0
5,996.8
6,090.0
5,912.7
5,769.7
5,742.6
88.9%
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 CO2 Eq. or 0.05 percent.
a Percent of total (gross) emissions excluding emissions from LUTUCF for 2017.
b Hie LUTUCF Sector Net Total is the net sum of all CH4 and N2O emissions to the atmosphere plus net carbon stock changes.
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 28 percent
of total U.S. greenhouse gas emissions in 2017. Electric power-related emissions decreased by 5.2 percent since
1990 and by 4.2 percent from 2016 to 2017, primarily due to decreased CO2 emissions from fossil fuel combustion
and increased use of renewable energy. Between 2016 to 2017, the consumption of coal, natural gas, and petroleum
for electric power generation decreased by 2.9, 7.2, and 10.7 percent, respectively, while the amount of electricity
generated (inkWh) decreased by 1.1 percent.
From 2016 to 2017, electricity sales to the residential and commercial end-use sectors decreased by 2.3 percent and
1.0 percent, respectively. Electricity sales to the industrial sector increased by approximately 0.8 percent. Overall,
from 2016 to 2017, the amount of electricity retail sales (in kWh) decreased by 1.0 percent. The sales trend in the
residential and commercial sectors can largely be attributed to milder weather conditions (i.e., wanner winter
months and cooler summer months). A decrease in both heating and cooling degree days from 2016 to 2017 resulted
in less demand for electricity to power heating and cooling equipment (EIA 2018). 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
2013
2014
2015
2016
2017
CO2
1,831.0
2,416.3
2,054.4
2,054.1
1,917.5
1,825.1
1,747.9
Fossil Fuel Combustion
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
Coal
1,546.5
1,982.8
1,571.3
1,568.6
1,351.4
1,242.0
1,207.1
Natural Gas
175.4
318.9
444.2
442.9
525.2
545.0
505.6
Petroleum
97.5
97.9
22.4
25.3
23.7
21.4
18.9
Geothermal
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Incineration of Waste
8.0
12.5
10.3
10.4
10.7
10.8
10.8
Other Process Uses of
Carbonates
3.1
3.8
5.8
6.5
6.1
5.5
5.1
CH4
0.4
0.9
1.0
1.1
1.2
1.2
1.1
Stationary Sources3
0.4
0.9
1.0
1.1
1.2
1.2
1.1
Incineration of Waste
+
+
+
+
+
+
+
N2O
21.0
30.4
28.9
29.2
26.8
26.5
25.1
Stationary Sources3
20.5
30.1
28.6
28.9
26.5
26.2
24.8
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
2-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
SF«
23.1
8.3
4.4
4.6
4.1
4.4
4.3
Electrical Transmission and
Distribution
23.1
8.3
4.4
4.6
4.1
4.4
4.3
Total
1,875.5
2,455.9
2,088.7
2,088.9
1,949.5
1,857.2
1,778.3
+ Does not exceed 0.05 MMT CO2 Eq.
a Includes only stationary combustion emissions related to the generation of electricity.
Note: Totals may not sum due to independent rounding.
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 2019a; Duffield 2006).
These source categories include CO2 from Fossil Fuel Combustion, CH4 and N20 from Stationary Combustion.
Incineration of Waste, Other Process Uses of Carbonates, and SF6 from Electrical Transmission and Distribution
Systems. Note that only 50 percent of the Other Process Uses of Carbonates emissions were associated with electric
power and distributed as described; the remainder of Other Process Uses of Carbonates emissions were attributed to
the industrial processes economic end-use sector.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 (29.7 percent), followed closely by emissions
from transportation (29.0 percent). Emissions from the commercial and residential sectors also increase substantially
when emissions from electricity are included (16.1 and 14.9 percent, respectively). In all economic end-use sectors
except agriculture, CO2 accounts for more than 81.3 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 2-12 shows the trend in these emissions by sector from 1990 to 2017.
Figure 2-15: U.S. Greenhouse Gas Emissions with Electricity-Related Emissions Distributed
to Economic Sectors (MMT CO2 Eq.)
2,500
Industry
2,000
Transportation
ff 1,500
IN
O
u
Commercial (Orange)
Residential (Blue)
Agriculture
500
¦sr in vo n
1—1 lH 1—1 t—I
0000
o
cn
en
o 1—1
o o
o o
cm ro
o o
o o
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.
Trends 2-27
-------
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 2017
Sector/Gas 1990 2005 2013 2014 2015 2016 2017 Percent3
Industry 2,300.9 2,223.5 2,036.7 2,023.0 1,973.6 1,906.4 1,915.6 29.7%
Direct Emissions 1,628.6 1,508.4 1,469.5 1,459.3 1,451.2 1,414.1 1,436.5 22.2%
CO2 1,160.8 1,146.4 1,120.6 1,106.6 1,101.7 1,072.8 1,091.3 16.9%
CH4 353.8 293.6 289.3 289.1 285.1 275.3 275.9 4.3%
N2O 37.6 29.7 25.6 27.3 26.7 28.0 27.1 0.4%
HFCs, PFCs, SFo, andNF3 76.3 38.7 34.0 36.3 37.7 38.0 42.2 0.7%
Electricity-Related 672.3 715.2 567.2 563.7 522.4 492.3 479.2 7.4%
CO2 656.4 703.6 557.9 554.3 513.8 483.8 470.9 7.3%
CH4 0.2 0.3 0.3 0.3 0.3 0.3 0.3 +%
N2O 7.5 8.9 7.9 7.9 7.2 7.0 6.8 0.1%
SFd 83 24 1.2 1.2 1.1 1.2 1.2 +%
Transportation 1,530.2 1,980.8 1,769.8 1,804.5 1,813.7 1,854.1 1,870.6 29.0%
Direct Emissions 1,527.1 1,976.0 1,765.4 1,799.9 1,809.3 1,849.7 1,866.2 28.9%
CO2 1,480.9 1,867.2 1,692.3 1,731.6 1,744.9 1,789.4 1,810.1 28.0%
CH4 5.9 3.0 1.8 1.7 1.6 1.5 1.4 +%
N2O 40.2 36.5 19.7 17.8 16.5 15.5 14.5 0.2%
HFCsb + 69.3 51.6 48.8 46.3 43.3 40.1 0.6%
Electricity-Related 3.1 4.8 4.4 4.6 4.4 4.3 4.4 0.1%
CO2 3.1 4.8 4.3 4.5 4.3 4.2 4.4 0.1%
CH4 + + + + + + + +%
N2O + 0.1 0.1 0.1 0.1 0.1 0.1 +%
SFd + + + + + + + +%
Commercial 981.1 1,222.4 1,131.5 1,143.1 1,112.3 1,066.6 1,038.4 16.1%
Direct Emissions 426.9 400.7 409.6 419.5 432.2 416.1 416.0 6.4%
CO2 226.5 226.8 224.6 232.9 245.5 232.1 232.9 3.6%
CH4 196.3 149.7 130.3 130.0 129.0 125.5 125.3 1.9%
N2O 4.1 6.4 6.8 7.0 7.0 7.1 7.2 0.1%
HFCs + 17.8 47.8 49.6 50.7 51.3 50.6 0.8%
Electricity-Related 554.2 821.7 721.9 723.6 680.1 650.5 622.4 9.6%
CO2 541.0 808.4 710.0 711.5 668.9 639.3 611.8 9.5%
CH4 0.1 0.3 0.4 0.4 0.4 0.4 0.4 +%
N2O 6.2 10.2 10.0 10.1 9.4 9.3 8.8 0.1%
SFd 6^8 2^8 1.5 1.6 1.4 1.5 1.5 +%
Residential 955.6 1,246.0 1,109.2 1,129.3 1,051.1 997.8 964.5 14.9%
Direct Emissions 344.7 370.0 356.3 376.6 349.7 326.9 330.9 5.1%
CO2 338.2 357.9 329.3 346.8 317.8 292.9 294.5 4.6%
CH4 5.2 4.1 4.9 5.0 4.5 3.9 3.8 0.1%
N2O 1.0 0.9 1.0 1.0 0.9 0.8 0.8 +%
HFCs 0.3 7.2 21.1 23.8 26.5 29.3 31.9 0.5%
Electricity-Related 610.8 875.9 752.9 752.7 701.4 670.9 633.6 9.8%
CO2 596.4 861.8 740.5 740.2 689.8 659.3 622.7 9.6%
CH4 0.1 0.3 0.4 0.4 0.4 0.4 0.4 +%
N2O 6.8 10.9 10.4 10.5 9.7 9.6 8.9 0.1%
SFd 7.5 10 1.6 1.7 1.5 1.6 1.5 +%
Agriculture 569.9 608.3 614.9 613.5 626.5 620.8 620.9 9.6%
Direct Emissions 534.9 570.0 572.6 569.2 585.2 581.7 582.2 9.0%
CO2 50.5 58.2 53.7 53.6 49.2 47.8 47.6 0.7%
CH4 218.1 240.1 235.5 235.2 240.0 247.4 248.8 3.9%
N2O 266.3 271.7 283.3 280.4 296.0 286.4 285.8 4.4%
Electricity-Related 35.1 38.3 42.3 44.3 41.2 39.1 38.7 0.6%
CO2 34.2 37.7 41.6 43.6 40.6 38.4 38.1 0.6%
CH4 + + + + + + + +%
N2O 0.4 0.5 0.6 0.6 0.6 0.6 0.5 +%
SFd 04 0J 0.1 0.1 0.1 0.1 0.1 +%
U.S. Territories 33.3 58.1 48.1 46.6 46.6 46.6 46.6 0.7%
2-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Total Emissions
6,371.0
7,339.0
6,710.2
6,760.0
6,623.8
6,492.3
6,456.7
100.0%
LULUCF Sector Net Totalc
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
-11.1%
Net Emissions (Sources and
Sinks)
5,564.0
; 6,599.0
5,996.8
6,090.0
5,912.7
5,769.7
5,742.6
88.9%
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 CO2 Eq. or 0.05 percent.
a Percent of total (gross) emissions excluding emissions from LULUCF for year 2017.
b Includes primarily HFC-134a.
c The LULUCF Sector Net Total is the net sum of all CH4 and N2O emissions to the atmosphere plus net carbon stock changes.
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 29.7 percent of U.S. greenhouse
gas emissions in 2017. This end-use sector also includes emissions that are produced as a byproduct of the non-
energy-related industrial process activities. The variety of activities producing these non-energy-related emissions
includes CH4 emissions from petroleum and natural gas systems, fugitive CH4 emissions from coal mining,
byproduct CO2 emissions from cement manufacture, and HFC, PFC, SF6, and NF3 byproduct emissions from
semiconductor manufacture, to name a few.
Since 1990, industrial sector emissions have declined by 14.8 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 29.0 percent of U.S. greenhouse gas emissions in 2017. The largest sources of transportation greenhouse gas
emissions in 2017 were passenger cars (41.2 percent); freight trucks (23.3 percent); light-duty trucks, which include
sport utility vehicles, pickup trucks, and minivans (17.5 percent); commercial aircraft (6.9 percent); other aircraft
(2.4 percent); ships and boats (2.4 percent); rail (2.2 percent); and pipelines (2.2 percent). These figures include
direct CO2, CH4, and N20 emissions from fossil fuel combustion used in transportation 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 2017, 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 45.1 percent from 1990 to 2017, 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 20136 and has
since grown at a faster rate (2.5 percent from 2015 to 2016, and 1.0 percent from 2016 to 2017). Average new
vehicle fuel economy has increased almost every year since 2005, while light-duty truck market share decreased to
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 2017 time period.
In absence of these method changes, light-duty VMT growth between 2006 and 2007 would likely have been higher.
Trends 2-29
-------
about 33 percent in 2009 and has since varied from year to year between 36 and 45 percent. Light-duty truck market
share was about 42 percent of new vehicles in model year 2017 (EPA 2019).
Table 2-13 provides a detailed summary of greenhouse gas emissions from transportation-related activities with
electricity-related emissions included in the totals.
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 23 percent from 1990 to 2017.7 This rise
in CO2 emissions, combined with an increase in HFCs from close to zero emissions in 1990 to 40.1 MMT CO2 Eq.
in 2017, led to an increase in overall greenhouse gas emissions from transportation activities of 22 percent.8
Table 2-13: Transportation-Related Greenhouse Gas Emissions (MMT CO2 Eq.)
Gas/Vehicle
1990
2005
2013
2014
2015
2016
2017
Passenger Cars
639.6
693.8
747.0
761.8
762.0
772.6
770.7
CO2
612.2
643.3
717.2
734.8
736.9
749.9
750.6
CH4
3.2
1.3
0.8
0.7
0.6
0.6
0.5
N2O
24.1
17.6
11.8
10.5
9.7
8.9
8.2
HFCs
0.0
31.7
17.2
15.8
14.7
13.2
11.4
Light-Duty Trucks
326.7
540.2
315.1
335.5
324.5
333.8
327.3
CO2
312.2
491.1
283.6
306.2
297.5
309.1
304.8
CH4
1.7
0.8
0.3
0.2
0.2
0.2
0.2
N2O
12.8
15.0
4.7
4.4
3.8
3.5
3.1
HFCs
0.0
33.3
26.5
24.7
23.0
21.1
19.2
Medium- and Heavy-Duty
Trucks
230.3
400.1
395.2
407.6
415.5
423.6
436.5
CO2
229.3
395.4
389.0
401.4
409.1
417.2
430.0
CH4
0.3
0.1
0.1
0.1
0.1
0.1
0.1
N2O
0.7
1.2
0.9
0.9
0.8
0.8
0.8
HFCs
0.0
3.4
5.2
5.3
5.5
5.5
5.7
Buses
8.5
12.2
17.9
19.3
19.8
19.5
20.4
CO2
8.4
11.6
17.2
18.6
19.1
18.8
19.7
CH4
+
0.2
0.2
0.2
0.2
0.2
0.2
N2O
+
+
+
+
+
+
+
HFCs
0.0
0.3
0.4
0.4
0.4
0.4
0.4
Motorcycles
1.7
1.6
3.9
3.8
3.7
3.9
3.8
CO2
1.7
1.6
3.8
3.8
3.7
3.8
3.7
CH4
+
+
+
+
+
+
+
N2O
+
+
+
+
+
+
+
Commercial Aircraft3
110.9
134.0
115.4
116.3
120.1
121.5
129.2
CO2
109.9
132.7
114.3
115.2
119.0
120.4
128.0
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
1.1
1.2
Other Aircraftb
78.3
59.7
34.7
35.0
40.4
47.5
45.6
CO2
77.5
59.1
34.4
34.7
40.0
47.0
45.2
CH4
0.1
0.1
+
+
+
+
+
N2O
0.7
0.5
0.3
0.3
0.4
0.4
0.4
Ships and Boats0
47.4
45.7
40.1
29.4
34.1
41.3
44.4
CO2
46.3
44.2
37.2
26.4
30.8
37.5
40.3
CH4
0.6
0.5
0.3
0.3
0.3
0.3
0.3
N2O
0.6
0.6
0.5
0.3
0.4
0.5
0.5
HFCs
0.0
0.5
2.0
2.3
2.6
2.9
3.3
7 See previous footnote.
8 See previous footnote.
2-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Rail
39.0
50.9
44.8
46.3
44.2
40.8
41.9
CO2
38.5
50.3
44.2
45.7
43.6
40.2
41.2
CH4
0.1
0.1
0.1
0.1
0.1
0.1
0.1
N2O
0.3
0.4
0.4
0.4
0.4
0.3
0.3
HFCs
0.0
0.1
0.1
0.1
0.1
0.1
0.1
Other Emissions from
Electric Powerd
0.1
+
+
+
+
+
0.1
Pipelines'5
36.0
32.4
46.2
39.4
38.5
39.2
41.4
CO2
36.0
32.4
46.2
39.4
38.5
39.2
41.4
Lubricants
11.8
10.2
9.6
10.0
11.0
10.4
9.6
CO2
11.8
10.2
9.6
10.0
11.0
10.4
9.6
Total Transportation
1,530.2
1,980.8
1,769.8
1,804.5
1,813.7
1,854.1
1,870.6
International Bunker Fuel/
54.8
44.7
29.5
28.7
31.6
34.9
34.6
Ethanol CO2s
4.1
21.6
70.5
74.0
74.2
76.9
77.7
Biodiesel COf
0.0
0.9
13.5
13.3
14.1
19.6
18.7
+ Does not exceed 0.05 MMT CO2 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 CO2 estimates reflect natural gas used to power pipelines, but not electricity. While the operation of pipelines
produces CH4 and N2O, these emissions are not directly attributed to pipelines in the Inventory.
f Emissions from International Bunker Fuels include emissions from both civilian and military activities; these
emissions are not included in the transportation totals.
g Ethanol and biodiesel CO2 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 UFC-134a. Totals may not sum due to independent rounding.
Commercial
The commercial end-use sector, with electricity-related emissions distributed, accounts for 16.1 percent of U.S.
greenhouse gas emissions in 2017 and is heavily reliant on electricity for meeting energy needs, with electricity use
for lighting, heating, air conditioning, and operating appliances. The remaining emissions were largely due to the
direct consumption of natural gas and petroleum products, primarily for heating and cooking needs. Energy-related
emissions from the commercial sector have generally been increasing since 1990, and annual variations are often
correlated with short-term fluctuations in energy use caused by weather conditions, rather than prevailing economic
conditions. Decreases in energy-related emissions in the commercial sector in recent years can be largely attributed
to an overall reduction in energy use driven by a reduction in heating degree days, and increases in energy
efficiency.
Landfills and wastewater treatment are included in the commercial sector, with landfill emissions decreasing since
1990 and wastewater treatment emissions decreasing slightly.
Residential
The residential end-use sector, with electricity-related emissions distributed, accounts for 14.9 percent of U.S.
greenhouse gas emissions in 2017 and similarly, is heavily reliant on electricity for meeting energy needs, with
electricity use for lighting, heating, air conditioning, and operating appliances. The remaining emissions were
largely due to the direct consumption of natural gas and petroleum products, primarily for heating and cooking
needs. Emissions from the residential sector have generally been increasing since 1990, and annual variations are
often correlated with short-term fluctuations in energy use caused by weather conditions, rather than prevailing
Trends 2-31
-------
economic conditions. In the long term, the residential sector is also affected by population growth, migration trends
toward wanner 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 lias
also contributed to recent trends in energy demand in households (EIA 2018).
Agriculture
The agriculture end-use sector accounts for 9.6 percent of U.S. greenhouse gas emissions in 2017 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 2017, agricultural soil management
was the largest source of N20 emissions, and enteric fermentation was the largest source of CH4 emissions in the
United States. This sector also includes small amounts of CO2 emissions from fossil fuel combustion by motorized
farm equipment such as tractors.
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 N20 are also based on the EIA electric
power sector. Additional sources include CO2, CHi and N20 from waste incineration, as the majority of municipal
solid waste is combusted in plants that produce electricity. The Electric Power economic sector also includes SF6
from Electrical Transmission and Distribution and a portion of 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 N20 from mobile combustion are also
apportioned to the Transportation economic sector based on the EIA transportation fuel-consuming sector.
Substitution of Ozone Depleting Substances emissions are apportioned to the Transportation economic sector based
on emissions from refrigerated transport and motor vehicle air-conditioning systems. Finally, 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 N20
emissions from stationary and mobile combustion are also apportioned to the Industry economic sector based on the
EIA industrial fuel-consuming sector, minus emissions apportioned to the Agriculture economic sector. Substitution
of Ozone Depleting Substances emissions are apportioned based on their specific end-uses within the source
category, with most emissions falling within the Industry economic sector.
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 2018; EIA 2019b). These supplementary data are subtracted from the industrial fuel use reported by EIA to
2-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
obtain agriculture fuel use. CO2 emissions from fossil fuel combustion and CH4 and N20 emissions from stationary
and mobile combustion are then apportioned to the Agriculture economic sector based on agricultural fuel use.
The other IPCC Agriculture emission source categories apportioned to the Agriculture economic sector include N20
emissions from Agricultural Soils, CH4 from Enteric Fermentation, CH4 and N2O from Manure Management, CH4
from Rice Cultivation, CO2 emissions from Liming and Urea Application and CH4 and N20 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 N20 are also based on the EIA
residential fuel-consuming sector. Substitution of Ozone Depleting Substances are apportioned to the Residential
economic sector based on emissions from residential air-conditioning systems. Nitrous oxide emissions from the
application of fertilizers to developed land (termed "settlements" by the IPCC) are also included in the Residential
economic sector.
The Commercial economic sector includes CO2 emissions from the combustion of fossil fuels that are included in
the EIA commercial fuel-consuming sector. Emissions of CH4 and N20 from Mobile Combustion are also
apportioned to the Commercial economic sector based on the EIA commercial fuel-consuming sector. Substitution
of Ozone Depleting Substances emissions are apportioned to the Commercial economic sector based on emissions
from commercial refrigeration/air-conditioning systems. Public works sources, including direct CH4 from Landfills,
CH4 and N20 from Wastewater Treatment, and Composting, are also included in the Commercial economic sector.
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 electricity use, because the electric
power industry—utilities and non-utilities combined—was the second largest source of emissions in 2017; (4)
emissions per unit of total gross domestic product as a measure of national economic activity; and (5) emissions per
capita.
Table 2-14 provides data on various statistics related to U.S. greenhouse gas emissions normalized to 1990 as a
baseline year. These values represent the relative change in each statistic since 1990. Greenhouse gas emissions in
the United States have grown at an average annual rate of 0.1 percent since 1990. This growth rate is slightly slower
than that for total energy use and fossil fuel consumption and much slower than that for electricity use, overall gross
domestic product (GDP) and national population (see Table 2-14 and Figure 2-16). These trends vary 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 1.0 percent since 2005. Total energy use
and fossil fuel consumption have also decreased at slower rates than emissions since 2005, while electricity use,
GDP, and national population continued to increase.
Table 2-14: Recent Trends in Various U.S. Data (Index 1990 = 100)
Variable
1990
2005
2013
2014
2015
2016
2017
Avg. Annual Avg. Annual
Change Change
Since 1990a Since 2005a
Greenhouse Gas Emissions'5
100
115
105
106
104
102
101
0.1%
-1.0%
Energy Usec
100
118
115
117
116
116
116
0.6%
-0.1%
Fossil Fuel Consumption0
100
119
110
111
110
109
108
0.3%
-0.7%
Electricity Usec
100
134
136
138
137
138
136
1.2%
0.1%
GDPd
100
159
176
180
186
189
193
2.5%
1.6%
Population6
100
118
126
127
128
129
130
1.0%
0.8%
+ Does not exceed 0.05 percent.
a Average annual growth rate
b GWP-weighted values
c Energy-content-weighted values (EIA 2019a)
Trends 2-33
-------
d GDP ill chained 2009 dollars (BEA 2019)
e U.S. Census Bureau (2018)
Figure 2-16: U.S. Greenhouse Gas Emissions Per Capita and Per Dollar of Gross Domestic
Product
200
Real GDP
180
160
140
Population
> 100
Emissions per capita
Emissions per $GDP
Source: BEA (2018), U.S. Census Bureau (2018), 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 indirect greenhouse gases,
which include CO, NOx, NMVOCs, and 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 N20. Non-methane
volatile organic compounds—which include hundreds of organic compounds that participate in atmospheric
chemical reactions (i.e., propane, butane, xylene, toluene, ethane, and many others)—are emitted primarily from
transportation, industrial processes, and non-industrial consumption of organic solvents. In the United States, 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 indirect
9 See .
2-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
greenhouse gas formation into greenhouse gases is the interaction of CO with the hydroxyl radical—the major
atmospheric sink for CH4 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 2018),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
2013
2014
2015
2016
2017
NOx
21,745
17,336
11,345
10,808
10,293
9,608
9,126
Mobile Fossil Fuel Combustion
10,862
10,295
6,523
6,138
5,740
5,413
5,051
Stationary Fossil Fuel Combustion
10,023
5,858
3,487
3,319
3,042
2,882
2,761
Oil and Gas Activities
139
321
641
650
650
650
650
Industrial Processes and Product Use
592
572
427
414
414
414
414
Forest Fires
37
133
157
155
321
124
124
Waste Combustion
82
128
89
97
97
97
97
Grassland Fires
5
21
13
27
21
19
21
Agricultural Burning
4
6
6
6
6
6
6
Waste
+
2
2
2
2
2
2
CO
131,277
71,783
48,720
47,613
52,447
43,039
41,382
Mobile Fossil Fuel Combustion
119,360
58,615
35,525
34,135
33,159
30,786
29,112
Forest Fires
1,334
4,723
5,574
5,525
11,425
4,425
4,425
Stationary Fossil Fuel Combustion
5,000
4,648
3,847
3,686
3,686
3,686
3,686
Waste Combustion
978
1,403
1,518
1,776
1,776
1,776
1,776
Industrial Processes and Product Use
4,129
1,557
1,247
1,251
1,251
1,251
1,251
Oil and Gas Activities
302
318
628
637
637
637
637
Grassland Fires
84
358
217
442
356
324
345
Agricultural Burning
89
154
157
152
148
144
141
Waste
1
7
7
8
8
8
8
NMVOCs
20,930
13,154
11,332
11,130
10,965
10,719
10,513
Industrial Processes and Product Use
7,638
5,849
3,855
3,816
3,816
3,816
3,816
Mobile Fossil Fuel Combustion
10,932
5,724
4,023
3,754
3,589
3,342
3,137
Oil and Gas Activities
554
510
2,741
2,853
2,853
2,853
2,853
Stationary Fossil Fuel Combustion
912
716
532
497
497
497
497
Waste Combustion
222
241
122
143
143
143
143
Waste
673
114
58
68
68
68
68
Agricultural Burning
NA
NA
NA
NA
NA
NA
NA
SO2
20,935
13,196
4,421
4,241
3,343
2,686
2,553
Stationary Fossil Fuel Combustion
18,407
11,541
3,644
3,532
2,635
1,978
1,846
Industrial Processes and Product Use
1,307
831
548
498
498
498
498
Mobile Fossil Fuel Combustion
390
180
99
94
94
94
94
Oil and Gas Activities
793
619
106
88
87
87
87
Waste Combustion
38
25
23
27
27
27
27
Waste
+
1
1
1
1
1
1
Agricultural Burning
NA
NA
NA
NA
NA
NA
NA
+ Does not exceed 0.5 kt.
NA (Not Available)
Note: Totals may not sum due to independent rounding.
Source: (EPA 2018) except for estimates from Forest Fires, Grassland Fires, and Field Burning of Agricultural
Residues.
10 NOx and CO emission estimates from Field Burning of Agricultural Residues were estimated separately, and therefore not
taken from EPA (2018).
Trends 2-35
-------
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 49.2 percent
in 2017. 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.
2-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
3. Energy
Energy-related activities were the primary sources of U.S. anthropogenic greenhouse gas emissions, accounting for
84.0 percent of total greenhouse gas emissions on a carbon dioxide (CO2) equivalent basis in 2017.1 This included
97, 43, and 13 percent of the nation's CO2, methane (CH4), and nitrous oxide (N20) emissions, respectively. Energy-
related CO2 emissions alone constituted 81.6 percent of national emissions from all sources on a CO2 equivalent
basis, while the non-C02 emissions from energy-related activities represented a much smaller portion of total
national emissions (5.1 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 32,310 million metric tons (MMT) of CO2 were
added to the atmosphere through the combustion of fossil fuels in 2016, 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 CH4 and N20. 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.
Figure 3-1: 2017 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-C02 Emissions from Stationary Combustion
Non-C02 Emissions from Mobile Combustion
Incineration of Waste
Abandoned Oil and Gas Wells
Abandoned Underground Coal Mines
Energy as a Portion of All
Emissions
4,912
100 150 200
MMT COz Eq.
* Estimates are presented in units of million metric tons of carbon dioxide equivalent (MMT CO2 Eq.), which weight each gas by
its global warming potential, or GWP, value. See section on global warming potentials in the Executive Summary.
^ Global CO2 emissions from fossil fuel combustion were taken from International Energy Agency CO 2 Emissions from Fossil
Fuels Combustion Overview IEA (2018).
Energy 3-1
-------
Figure 3-2: 2017 U.S. Fossil Carbon Flows (MMT CO2 Eq.)
International.
Bunkers ,
NEU Emissions 13
Fossil Fuel
Energy Exports
1,176
Coal Emissions
UB4
— NEU Emissions 5
Natural Gas Emissions
Coal
" Combustion
Emissions 1452
NEU Emissions 106
Domestic
Fossil Fuel
Production
Petroleum
1.464
Natural Gas Liquids,
liquefied Refinery Gas,
& Other Liquids
Fossil Fuel
Energy
Imports
Non-Energy Lfee
Carton Sequestered
Petroleum
1.395 „
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
Fossil Fuel
Consumption
U.S
Territories
Stock
Changes
027)
Natural Gas 165'
Coal 18"
NEU Imports
Table 3-1 summarizes emissions from the Energy sector in units of MMT CO2 Eq., while unweighted gas emissions
inkilotons (kt) are provided in Table 3-2. Overall, emissions due to energy -related activities were 5,424.8 MMT
CO2 Eq. in 2017,3 an increase of 1.6 percent since 1990 and a decrease of 0.7 percent since 2016.
Table 3-1: CO2, Cm, and N2O Emissions from Energy (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
4,905.3
5,931.0
5,341.5
5,384.8
5,241.5
5,134.1
5,095.6
Fossil Fuel Combustion
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
Transportation
1,469.1
1,857.0
1,682.7
1.721.6
1,734.0
1,779.0
1,800.6
Electric Power
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
Industrial
857.5
853.4
840.0
819.6
807.9
807.6
810.7
Residential
338.2
357.9
329.3
346.8
317.8
292.9
294.5
Commercial
226.5
226.8
224.6
232.9
245.5
232.1
232.9
U.S. Territories
27.6
49.7
42.5
41.4
41.4
41.4
41.4
Non-Energy Use of Fuels
119.6
139.6
123.5
119.9
126.9
113.7
123.2
Natural Gas Systems
30.0
22.6
25.1
25.5
25.1
25.5
26.3
Petroleum Systems
9.0
11.6
25.1
29.6
31.7
22.2
23.3
Incineration of Waste
8.0
12.5
10.3
10.4
10.7
10.8
10.8
Abandoned Oil and Gas
Wells
+
+
+
+
+
+
+
Biomass-Wood"
215.2
206.9
227.3
233.8
224.7
216.3
221.4
International Bunker
Fuels11
103.5
113.1
99.8
103.4
110.9
116.6
120.1
Biofuels-Ethanol"
4.2
22.9
74.7
76.1
78.9
81.2
82.1
Biofuels-Biodiesel"
0.0
0.9
13.5
13.3
14.1
19.6
18.7
CH4
366.9
303.2
298.4
298.1
293.5
283.0
283.3
Natural Gas Systems
193.1
171.4
165.6
165.1
167.2
165.7
165.6
Coal Mining
96.5
64.1
64.6
64.6
61.2
53.8
55.7
Petroleum Systems
42.1
36.7
41.6
42.1
39.5
38.2
37.7
Stationary Combustion
8.6
7.8
8.7
8.9
8.5
7.9
7.8
3 Following the current reporting requirements under the UNFCCC, this Inventory report presents CO2 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: 1990-2017
-------
Abandoned Oil and Gas
Wells
6.6
6.9
7.0
7.1
7.1
7.2
6.9
Abandoned Underground
Coal Mines
7.2
6.6
6.2
6.3
6.4
6.7
6.4
Mobile Combustion
12.9
9.6
4.5
4.1
3.6
3.4
3.2
Incineration of Waste
+
+
+
+
+
+
+
International Bunker
Fuelsb
0.2
0.1
0.1
0.1
0.1
0.1
0.1
N2O
67.6
73.7
55.2
53.5
49.7
48.3
45.9
Stationary Combustion
25.1
34.3
32.7
33.0
30.6
30.1
28.6
Mobile Combustion
42.0
39.0
22.1
20.2
18.8
17.9
16.9
Incineration of Waste
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Petroleum Systems
+
+
+
+
+
+
+
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker
Fuelsb
0.9
1.0
0.9
0.9
0.9
1.0
1.0
Total
5,339.8
6,308.0
5,695.0
5,736.4
5,584.7
5,465.3
5,424.8
+ Does not exceed 0.05 MMT CO2 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.
Note: Totals may not sum due to independent rounding.
Table 3-2: CO2, ChU, and N2O Emissions from Energy (kt)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
4,905,262
5,931,045
5,341,485
5,384,792
5,241,538
5,134,056
5,095,643
Fossil Fuel Combustion
4,738,756
5,744,754
5,157,391
5,199,345
5,047,107
4,961,876
4,911,962
Non-Energy Use of
Fuels
119,551
139,625
123,476
119,895
126,939
113,719
123,221
Natural Gas Systems
30,048
22,638
25,148
25,518
25,071
25,488
26,327
Petroleum Systems
8,950
11,552
25,130
29,597
31,672
22,200
23,336
Incineration of Waste
7,950
12,469
10,333
10,429
10,742
10,765
10,790
Abandoned Oil and Gas
Wells
6
7
7
7
7
7
7
Biomass-Wood"
215,186
206,901
227,340
233,762
224,730
216,293
221,432
International Bunker
Fuelsb
103,463
113,139
99,763
103,400
110,887
116,594
120,107
Biofuels-Ethanol"
4,227
22,943
74,743
76,075
78,934
81,250
82,088
Biofuels-Biodiesel"
0
856
13,462
13,349
14,077
19,648
18,705
CH4
14,677
12,127
11,935
11,922
11,738
11,320
11,332
Natural Gas Systems
7,723
6,856
6,624
6,603
6,686
6,629
6,624
Coal Mining
3,860
2,565
2,584
2,583
2,449
2,154
2,227
Petroleum Systems
1,682
1,469
1,665
1,682
1,579
1,528
1,506
Stationary Combustion
344
313
350
355
340
318
312
Abandoned Oil and Gas
Wells
262
277
282
283
285
289
277
Abandoned
Underground Coal
Mines
288
264
249
253
256
268
257
Mobile Combustion
518
384
181
163
143
135
128
Incineration of Waste
+
+
+
+
+
+
+
International Bunker
Fuelsb
7
5
3
3
3
4
4
N2O
227
247
185
180
167
162
154
Stationary Combustion
84
115
110
111
103
101
96
Mobile Combustion
141
131
74
68
63
60
57
Incineration of Waste
2
1
1
1
1
1
1
Petroleum Systems
+
+
+
+
+
+
+
Energy 3-3
-------
Natural Gas Systems
+
+
+
+
+
+
+
International Bunker
Fuelsb
3
3
3
3
3
3
3
+ 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.
Note: Totals may not sum due to independent rounding.
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 2016) to ensure that the trend is accurate.
Updates to N20 emissions from Stationary Combustion in the Energy sector resulted in an average change over the
time series of greater than 10 MMT CO2 Eq. For more information on specific methodological updates, please see
the Recalculations Discussion for each category, in this chapter.
Box 3-1: 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 this Inventory do 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.
Box 3-2: Energy Data from EPA's Greenhouse Gas Reporting Program
On October 30, 2009, the U.S. Enviromnental 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). 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. Annual reporting is at the facility level, except for certain suppliers of fossil fuels and industrial
greenhouse gases. Data reporting by affected facilities includes the reporting of emissions from fuel combustion at
that affected facility. In general, the threshold for reporting is 25,000 metric tons or more of CO2 Eq. per year.
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 (see also Box 3-4).4 As indicated in the respective Planned Improvements
4 See .
3-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
sections for source categories in this chapter, EPA continues to examine the uses of facility-level GHGRP data to
improve the national estimates presented in this Inventory. 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. 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 frombiomass. Further information on the reporting categories in EPA's GHGRP and specific data caveats
associated with monitoring methods in EPA's GHGRP is provided on the GHGRP website.5
EPA presents the data collected by its GHGRP through a data publication tool that allows data to be viewed in
several formats including maps, tables, charts and graphs for individual facilities or groups of facilities.6
In addition to using GHGRP data to estimate emissions, 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. 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, the GHGRP data are used to provide a more detailed breakout of total
emissions in the industrial end-use sector within that source category.
3.1 Fossil Fuel Combustion (CRF Source
Category 1A)
Emissions from the combustion of fossil fuels for energy include the gases CO2, CH4, and N20. 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 N20 emissions from
stationary combustion and mobile combustion. Thus, three separate descriptions of methodologies, uncertainties,
recalculations, and planned improvements are provided at the end of this section. Total CO2, CH4, and N20
emissions from fossil fuel combustion are presented in Table 3-3 and Table 3-4.
Table 3-3: CO2, ChU, and N2O Emissions from Fossil Fuel Combustion (MMT CO2 Eq.)
Gas
1990
2005
2013
2014
2015
2016
2017
CO2
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
CH4
21.5
17.4
13.3
13.0
12.1
11.3
11.0
N2O
67.1
73.3
54.8
53.2
49.4
47.9
45.5
Total
4,827.4
5,835.5
5.225.5
5,265.5
5,108.6
5,021.1
4,968.5
Note: Totals may not sum due to independent rounding.
5 See
.
6 See .
Energy 3-5
-------
Table 3-4: CO2, ChU, and N2O Emissions from Fossil Fuel Combustion (kt)
Gas
1990
2005
2013
2014
2015
2016
2017
CO2
4,738,756
5,744,754
5,157,391
5,199,345
5,047,107
4,961,876
4,911,962
CH4
862
696
531
518
483
453
440
N2O
225
246
184
178
166
161
153
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
2017, CO2 emissions from fossil fuel combustion decreased by 1.0 percent relative to the previous year. The
decrease in CO2 emissions from fossil fuel combustion was a result of multiple factors, primarily a continued shift
from coal to natural gas and substitution from fossil to non-fossil energy sources in the electric power sector. In
2017, CO2 emissions from fossil fuel combustion were 4,912.0 MMT CO2 Eq., or 3.7 percent above emissions in
1990 (see Table 3-5).7
Table 3-5: CO2 Emissions from Fossil Fuel Combustion by Fuel Type and Sector (MMT CO2
Eq.)
Fuel/Sector
1990
2005
2013
2014
2015
2016
2017
Coal
1,717.3
2,111.2
1,654.1
1,652.4
1,424.7
1,307.5
1,267.5
Residential
3.0
0.8
0.0
0.0
0.0
0.0
0.0
Commercial
12.0
9.3
3.9
3.8
3.0
2.3
2.0
Industrial
155.2
115.3
76.0
76.0
66.3
59.2
54.4
Transportation
NE
NE
NE
NE
NE
NE
NE
Electric Power
1,546.5
1,982.8
1,571.3
1,568.6
1,351.4
1,242.0
1,207.1
U.S. Territories
0.6
3.0
2.8
4.0
4.0
4.0
4.0
Natural Gas
999.7
1,167.0
1,391.9
1,420.0
1,460.2
1,471.8
1,450.3
Residential
237.8
262.2
266.4
277.7
252.7
238.4
241.5
Commercial
142.0
162.9
179.2
189.2
175.4
170.5
173.2
Industrial
408.5
388.6
452.1
467.1
464.4
474.8
484.7
Transportation
36.0
33.1
47.0
40.2
39.4
40.1
42.3
Electric Power
175.4
318.9
444.2
442.9
525.2
545.0
505.6
U.S. Territories
NO
1.3
3.0
3.0
3.0
3.0
3.0
Petroleum
2,021.2
2,466.2
2,111.0
2,126.5
2,161.8
2,182.1
2,193.7
Residential
97.4
94.9
63.0
69.2
65.1
54.5
53.0
Commercial
72.6
54.6
41.5
39.9
67.1
59.3
57.7
Industrial
293.7
349.5
311.9
276.5
277.1
273.6
271.5
Transportation
1,433.1
1,823.9
1,635.6
1,681.3
1,694.6
1,739.0
1,758.3
Electric Power
97.5
97.9
22.4
25.3
23.7
21.4
18.9
U.S. Territories
26.9
45.4
36.6
34.3
34.3
34.3
34.3
Geothermal3
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Total
4,738.8
5,744.8
5,157.4
5,199.3
5,047.1
4,961.9
4,912.0
NE (Not Estimated)
NO (Not Occurring)
a Although not technically a fossil fuel, geothermal energy-related CO2 emissions are included for reporting purposes.
Note: Totals may not sum due to independent rounding.
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
7 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: 1990-2017
-------
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.8
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 2017 CO2 Emissions from Fossil Fuel
Combustion for Selected Fuels and Sectors (MMT CO2 Eq. and Percent)
Sector Fuel Type
2013 to 2014
2014 to 2015
2015 to 2016
2016 to 2017
Total 2017
Electric Power Coal
Electric Power Natural Gas
Transportation Petroleum
Residential Natural Gas
Commercial Natural Gas
Industrial Natural Gas
-2.7 -0.2%
-1.3 -0.3%
45.7 2.8%
11.3 4.3%
10.0 5.6%
15.0 3.3%
-217.2 -13.8%
82.3 18.6%
13.2 0.8%
-24.9 -9.0%
-13.8 -7.3%
-2.6 -0.6%
-109.4 -8.1%
19.8 3.8%
44.4 2.6%
-14.3 -5.7%
-4.9 -2.8%
10.4 2.2%
-34.9 -2.8%
-39.4 -7.2%
19.3 1.1%
3.1 1.3%
2.6 1.5%
9.9 2.1%
1,207.1
505.6
1,758.3
241.5
173.2
484.7
All Sectors3 All Fuels3
42.0 0.8%
-152.2 -2.9%
-85.2 -1.7%
-49.9 -1.0%
4,912.0
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 0.8 percent increase from
2013 to 2014, then a 2.9 percent decrease from 2014 to 2015, then a 1.7 percent decrease from 2015 to 2016, and a
1.0 percent decrease from 2016 to 2017. These changes contributed to a 4.8 percent decrease in CO2 emissions from
fossil fuel combustion from 2013 to 2017.
Trends in CO2 emissions from fossil fuel combustion over the past five years have been in large part 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 remained relatively
flat over the past five years, but emissions have decreased due to a decreasing reliance on coal used to generate
electricity. Carbon dioxide emissions from coal consumption for electric power generation decreased by 23.2
percent since 2013, which can be largely attributed to a shift to the use of less-CCh-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 decreased by 9.3
percent and 3.4 percent from 2013 to 2017, respectively. This trend can be largely attributed to a 14 percent decrease
in heating degree days, which led to a decreased demand for heating fuel and electricity for heat in these sectors. In
addition, an increase in energy efficiency standards and the use of energy-efficient products in residential and
commercial buildings has resulted in an overall reduction in energy use, contributing to a decrease in CO2 emissions
in both of these sectors (EIA 2018e).
Petroleum use in the transportation sector is another major driver of emissions, representing the largest source of
CO2 emissions from fossil fuel combustion in 2017. Despite the overall decreasing trend in CO2 emissions from
fossil fuel combustion over the past five years, emissions from petroleum consumption for transportation have
increased by 7.5 percent since 2013; this trend can be primarily attributed to a 7.5 percent increase in vehicle miles
traveled (VMT) over the same time period.
8 Based on national aggregate carbon content of all coal, natural gas, and petroleum fuels combusted in the United States.
Energy 3-7
-------
In the United States, 80 percent of the energy used in 2017 was produced through the combustion of fossil fuels such
as coal, natural gas, and petroleum (see Figure 3-3 and Figure 3-4). The remaining portion was supplied by nuclear
electric power (9 percent) and by a variety of renewable energy sources (11 percent), primarily hydroelectric power,
wind energy and biofuels (EIA 2019a).9 Specifically, petroleum supplied the largest share of domestic energy
demands, accounting for 37 percent of total U.S. energy used in 2017. Natural gas and coal followed in order of
energy demand importance, accounting for approximately 29 percent and 14 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).
Figure 3-3: 2017 U.S. Energy Use by Energy Source (Percent)
Nuclear Electric Power
8.6%
Renewable Energy
11.4%
Petroleum
37.0%
Coal
14.3%
Natural Gas
28.6%
Figure 3-4: U.S. Energy Use (Quadrillion Btu)
Total Energy
Fossil Fuels
Renewable & Nuclear
Oi-irsim^-Lnvoiv
en cti en c 0
CTi C7i CTi O"! <71
00
-------
Figure 3-5: 2017 CO2 Emissions from Fossil Fuel Combustion by Sector and Fuel Type (MMT
COz Eq.)
Relative Contribution by Fuel Type
2,500
2,000
1,732
Petroleum
Coal
Natural Gas
29.5%
44.7%
" 1,500
25.8%
U.S. Territories
Commercial
Residential
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 CH4, CO, and NMVOCs.10 These other C-containing non-CO; 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 3-3: Weather and Non-Fossil Energy Effects on CO2 from Fossil Fuel Combustion Trends
The United States in 2017 experienced a warmer winter overall compared to 2016, as heating degree days decreased
(1.3 percent). Wanner winter conditions compared to 2016 impacted the amount of energy required for heating, and
heating degree days in the United States were 15.4 percent below normal (see Figure 3-6). Cooling degree days
decreased by 8.4 percent compared to 2016, which decreased demand for air conditioning in the residential and
commercial sector. This led in part to an overall residential electricity demand decrease of 2.3 percent. Summer
conditions in 2017 were still wanner than nonnal, with cooling degree days 17.4 percent above normal (see Figure
3-7) (EIA 2019a).11
10 See the sections entitled Stationary Combustion and Mobile Combustion in this chapter for information on 11011-CO2 gas
emissions from fossil fuel combustion.
11 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 +12 percent and +19 percent for heating and cooling degree days, respectively (99 percent
confidence interval).
Energy 3-9
-------
Figure 3-6: Annual Deviations from Normal Heating Degree Days for the United States
(1950—2017, Index Normal = 100)
20
fO
E
E
o
S
O
o
10
-10
fD
E
x
kDvDrsrvrvrsrvoooocoooooajcnCT4cna>ooooo^-i^-i^-i»-i
en cn cti cn cn cn 01 ^ 01 o^o~io~>oo ooooo oo
rHHrHrHHrHrHrHrH iH H HrHHrHrHrHrHrHrHrHrHrHrHrH N fNj f\| fNJ fN] fNl (S f\l
Figure 3-7: Annual Deviations from Normal Cooling Degree Days for the United States
(1950—2017, Index Normal = 100)
to
Ł
E
o
2
>
.cr»i—iroLnr*-.
i-inLrjrN.o^i-ir^Lrjrxo^i-iroLor^CTii-iromiv.CTii-iroinrN.cri
Lnir)Lnir)LniŁ>^D^o^okorvrN.rvr^rvcjoooaoooaoCT>o^cr»o^a» _ _ _ _ _
CTi CI Ol CT» cn O"! CTi CT"> CT"i O"* O O O O O O O O O
i—i l—i i—i i—i i—i i—i i—i i—i i—i i—i i—i iH h i—i l—i l—i l—i l—i l—i l—i l—i i—i l—i l—i l—i oj in rsi r*J f\i fN rM rsi rsi
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)12 of nuclear power plants in 2017 remained high at 92 percent. In
12 Hie 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 (2018d).
3-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
2017, nuclear power represented 21 percent of total electricity generation. In recent years, the wind and solar power
sectors have shown strong growth such that, on the margin they are becoming relatively important electricity
sources. Between 1990 and 2017, renewable energy generation (inkWh) from solar and wind energy have increased
from 0.1 percent in 1990 to 8 percent of total electricity generation in 2017, 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, CH4 and N20 are emitted from stationary and mobile
combustion as well. Table 3-7 provides an overview of the CO2, CH4, and N20 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
2013
2014
2015
2016
2017
Transportation
1,524.1
1,905.6
1,709.3
1,745.9
1,756.4
1,800.3
1,820.7
CO2
1,469.1
1,857.0
1,682.7
1,721.6
1,734.0
1,779.0
1,800.6
CH4
12.9
9.6
4.5
4.1
3.6
3.4
3.2
N2O
42.0
39.0
22.1
20.2
18.8
17.9
16.9
Electric Power
1,840.9
2,430.9
2,067.9
2,067.1
1,928.3
1,836.2
1,757.9
CO2
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
CH4
0.4
0.9
1.0
1.1
1.2
1.2
1.1
N2O
20.5
30.1
28.6
28.9
26.5
26.2
24.8
Industrial
862.4
858.1
844.5
823.9
812.2
811.8
814.9
CO2
857.5
853.4
840.0
819.6
807.9
807.6
810.7
CH4
1.8
1.7
1.7
1.6
1.6
1.6
1.6
N2O
3.1
2.9
2.8
2.7
2.7
2.6
2.7
Residential
344.5
362.8
335.2
352.8
323.2
297.6
299.0
CO2
338.2
357.9
329.3
346.8
317.8
292.9
294.5
CH4
5.2
4.1
4.9
5.0
4.5
3.9
3.8
N2O
1.0
0.9
1.0
1.0
0.9
0.8
0.8
Commercial
228.0
228.2
226.0
234.3
247.0
233.7
234.4
CO2
226.5
226.8
224.6
232.9
245.5
232.1
232.9
CH4
1.1
1.1
1.1
1.1
1.2
1.2
1.2
N2O
0.4
0.3
0.3
0.3
0.4
0.3
0.3
U.S. Territories3
27.7
49.9
42.6
41.5
41.5
41.5
41.5
Total
4,827.4
5,835.5
5,225.5
5,265.5
5,108.6
5,021.1
4,968.5
a U.S. Territories are not apportioned by sector, and
emissions are total greenhouse gas emissions from all fuel
combustion sources.
Notes: Totals may not sum due to independent rounding.
Other than CO2, gases emitted from stationary combustion include the greenhouse gases CH4 and N20 and
greenhouse gas precursors NOx, CO, and NMVOCs.13 Methane and N20 emissions from stationary combustion
sources depend upon fuel characteristics, size and vintage, along with combustion technology, pollution control
equipment, ambient enviromnental 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 produces greenhouse gases other than CO2, including CH4, N20, and greenhouse gas precursors
including NOx, CO, and NMVOCs. As with stationary combustion, N20 and NOx emissions from mobile
combustion are closely related to fuel characteristics, air-fuel mixes, combustion temperatures, and the use of
pollution control equipment. Nitrous oxide from mobile sources, in particular, can be formed by the catalytic
13 Sulfur dioxide (SO2) emissions from stationary combustion are addressed in Annex 6.3.
Energy 3-11
-------
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 CH4 content of the motor fuel, the amount of
hydrocarbons passing uncombusted through the engine, and any post-combustion control of hydrocarbon emissions
(such as catalytic converters).
An alternative method of presenting combustion emissions is to allocate emissions associated with electric power to
the sectors in which it is used. Four end-use sectors were defined: industrial, transportation, 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 N20 from transportation.1415 Emissions
from U.S. Territories are also calculated separately due to a lack of end-use-specific consumption data.16 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
2013
2014
2015
2016
2017
Transportation
1,527.1
1,910.3
1,713.6
1,750.3
1,760.6
1,804.5
1,825.0
CO2
1,472.1
1,861.7
1,686.9
1,726.0
1,738.2
1,783.2
1,804.9
CH4
12.9
9.6
4.5
4.1
3.6
3.4
3.2
N2O
42.0
39.0
22.1
20.2
18.8
17.9
16.9
Industrial
1,556.7
1,603.9
1,447.9
1,425.6
1,369.7
1,337.2
1,326.9
CO2
1,543.9
1,589.7
1,434.8
1,412.5
1,357.4
1,325.2
1,315.1
CH4
2.0
2.0
2.0
2.0
2.0
1.9
1.9
N2O
10.8
12.2
11.1
11.1
10.4
10.1
9.9
Residential
944.0
1,229.9
1,080.6
1,097.7
1,016.9
961.0
925.4
CO2
930.9
1,213.9
1,064.1
1,080.9
1,001.6
946.3
911.5
CH4
5.4
4.4
5.3
5.4
4.9
4.3
4.2
N2O
7.7
11.6
11.3
11.4
10.5
10.3
9.6
Commercial
771.9
1,041.5
940.8
950.3
919.8
876.9
849.7
CO2
764.3
1,029.7
929.1
938.5
908.5
865.8
839.1
CH4
1.2
1.4
1.4
1.5
1.6
1.6
1.6
N2O
6.5
10.4
10.2
10.3
9.6
9.5
9.0
U.S. Territories3
27.7
49.9
42.6
41.5
41.5
41.5
41.5
Total
4,827.4
5,835.5
5,225.5
5,265.5
5,108.6
5,021.1
4,968.5
a U.S. Territories are not apportioned by sector, and emissions are total greenhouse gas emissions from all fuel
combustion sources.
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.
Stationary Combustion
The direct combustion of fuels by stationary sources in the electric power, industrial, commercial, and residential
14 Separate calculations were performed for transportation-related CH4 and N2O. The methodology used to calculate these
emissions are discussed in the Mobile Combustion section.
15 In this year's Inventory, electricity use from electric vehicle charging in commercial and residential locations was re-allocated
from the residential and commercial sectors to the transportation sector. These changes apply to the time period from 2010
through 2017. 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).
16 U.S. Territories consumption data that are obtained from EIA are only available at the aggregate level and cannot be broken
out by end-use sector. The distribution of emissions to each end-use sector for the 50 states does not apply to territories data.
3-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 N20. Table 3-10 and Table 3-11 present CH4 and N20
emissions from the combustion of fuels in stationary sources. The CH4 and N20 emission estimation methodology
utilizes facility-specific technology and fuel use data reported to EPA's Acid Rain Program (EPA 2018a) (see
Methodology section for CH4 and N20 from Stationary Combustion). Table 3-7 presents the corresponding direct
CO2, CH4, and N20 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
2013
2014
2015
2016
2017
Electric Power
1,820.0
2,400.0
2,038.3
2,037.1
1,900.6
1,808.9
1,732.0
Coal
1,546.5
1,982.8
1,571.3
1,568.6
1,351.4
1,242.0
1,207.1
Natural Gas
175.4
318.9
444.2
442.9
525.2
545.0
505.6
Fuel Oil
97.5
97.9
22.4
25.3
23.7
21.4
18.9
Geo thermal
0.5
0.5
0.4
0.4
0.4
0.4
0.4
Industrial
857.5
853.4
840.0
819.6
807.9
807.6
810.7
Coal
155.2
115.3
76.0
76.0
66.3
59.2
54.4
Natural Gas
408.5
388.6
452.1
467.1
464.4
474.8
484.7
Fuel Oil
293.7
349.5
311.9
276.5
277.1
273.6
271.5
Commercial
226.5
226.8
224.6
232.9
245.5
232.1
232.9
Coal
12.0
9.3
3.9
3.8
3.0
2.3
2.0
Natural Gas
142.0
162.9
179.2
189.2
175.4
170.5
173.2
Fuel Oil
72.6
54.6
41.5
39.9
67.1
59.3
57.7
Residential
338.2
357.9
329.3
346.8
317.8
292.9
294.5
Coal
3.0
0.8
0.0
0.0
0.0
0.0
0.0
Natural Gas
237.8
262.2
266.4
277.7
252.7
238.4
241.5
Fuel Oil
97.4
94.9
63.0
69.2
65.1
54.5
53.0
U.S. Territories
27.6
49.7
42.5
41.4
41.4
41.4
41.4
Coal
0.6
3.0
2.8
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
36.6
34.3
34.3
34.3
34.3
Total
3,269.7
3,887.8
3,474.7
3,477.8
3,313.1
3,182.8
3,111.4
NO (Not Occurring)
Note: Totals may not sum due to independent rounding.
Table 3-10: ChU Emissions from Stationary Combustion (MMT CO2 Eq.)
Sector/Fuel Type
1990
2005
2013
2014
2015
2016
2017
Electric Power
0.4
0.9
1.0
1.1
1.2
1.2
1.1
Coal
0.3
0.4
0.3
0.3
0.3
0.2
0.2
Fuel Oil
+
+
+
+
+
+
+
Natural gas
0.1
0.5
0.7
0.8
0.9
0.9
0.9
Wood
+
+
+
+
+
+
+
Industrial
1.8
1.7
1.7
1.6
1.6
1.6
1.6
Coal
0.4
0.3
0.2
0.2
0.2
0.2
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
1.1
1.0
1.1
Commercial
1.1
1.1
1.1
1.1
1.2
1.2
1.2
Coal
+
+
+
+
+
+
+
Fuel Oil
0.3
0.2
0.1
0.1
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
0.5
0.6
0.6
0.6
Energy 3-13
-------
Residential
5.2
4.1
4.9
5.0
4.5
3.9
3.8
Coal
0.2
0.1
0.0
0.0
0.0
0.0
0.0
Fuel Oil
0.3
0.3
0.2
0.3
0.2
0.2
0.2
Natural Gas
0.5
0.6
0.6
0.6
0.6
0.5
0.5
Wood
4.1
3.1
4.1
4.1
3.7
3.2
3.1
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
8.7
8.9
8.5
7.9
7.8
+ Does not exceed 0.05 MMT CO2 Eq.
NO (Not Occurring)
Note: Totals may not sum due to independent rounding.
Table 3-11: N2O Emissions from Stationary Combustion (MMT CO2 Eq.)
Sector/Fuel Type
1990
2005
2013
2014
2015
2016
2017
Electric Power
20.5
30.1
28.6
28.9
26.5
26.2
24.8
Coal
20.1
28.0
25.5
25.7
22.8
22.4
21.2
Fuel Oil
0.1
0.1
+
+
+
+
+
Natural Gas
0.3
1.9
3.1
3.1
3.7
3.8
3.6
Wood
+
+
+
+
+
+
+
Industrial
3.1
2.9
2.8
2.7
2.7
2.6
2.7
Coal
0.7
0.5
0.4
0.4
0.3
0.3
0.3
Fuel Oil
0.5
0.5
0.5
0.4
0.4
0.4
0.4
Natural Gas
0.2
0.2
0.2
0.2
0.2
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
0.3
0.4
0.3
0.3
Coal
0.1
+
+
+
+
+
+
Fuel Oil
0.2
0.1
0.1
0.1
0.2
0.2
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
1.0
0.9
0.8
0.8
Coal
+
+
0.0
0.0
0.0
0.0
0.0
Fuel Oil
0.2
0.2
0.2
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
0.7
0.5
0.6
0.7
0.6
0.5
0.5
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
+
+
+
+
+
+
Wood
NO
NO
NO
NO
NO
NO
NO
Total
25.1
34.3
32.7
33.0
30.6
30.1
28.6
+ Does not exceed 0.05 MMT CO2 Eq.
NO (Not Occurring)
Note: Totals may not sum due to independent rounding.
Electric Power Sector
The process of generating electricity is the largest stationary source of CO2 emissions in the United States,
representing 33 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 35.3 percent of CO2 emissions from fossil
fuel combustion in 2017. Methane and N20 from electric power represented 10.0 and 54.4 percent of total CH4 and
N2O emissions from fossil fuel combustion in 2017, respectively.
3-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.17
Total emissions from the electric power sector have decreased by 4.5 percent since 1990. The carbon intensity of the
electric power sector, in terms of CO2 Eq. per QBtu, lias decreased by 11 percent during that same timeframe with
the majority of the emissions and carbon intensity decreases occurring in the past decade as 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 31 percent in 2017.18 This
corresponded with an increase in natural gas generation and renewable energy generation largely from wind and
solar energy. Natural gas generation (in kWh) represented 11 percent of electric power generation in 1990, and
increased over the 28-year period to represent 31 percent of electric power sector generation in 2017 (see Table
3-12).
Table 3-12: Electric Power Generation by Fuel Type (Percent)
Fuel Type
1990
2005
2013
2014
2015
2016
2017
Coal
54.1%
51.1%
40.2%
39.9%
34.2%
31.4%
30.9%
Natural Gas
10.7%
17.5%
26.4%
26.3%
31.6%
32.7%
30.9%
Nuclear
19.9%
20.0%
20.2%
20.3%
20.4%
20.6%
20.8%
Renewables
11.3%
8.3%
12.5%
12.8%
13.0%
14.7%
16.8%
Petroleum
4.1%
3.0%
0.6%
0.7%
0.7%
0.6%
0.5%
Other Gases3
+%
0.1%
0.1%
0.1%
0.1%
0.1%
0.1%
Net Electricity Generation
(Billion kWh)b
2,905
3,902
3,901
3,936
3,917
3,917
3,877
+ Does not exceed 0.05 percent
a Other gases include blast furnace gas, propane gas, and other manufactured and waste gases derived from fossil fuels.
b Represents net electricity generation from the electric power sector. Excludes net electricity generation from
commercial and industrial combined-heat-and-power and electricity-only plants.
In 2017, CO2 emissions from the electric power sector decreased by 4.2 percent relative to 2016. This decrease in
CO2 emissions was a result of a decrease in fossil fuels consumed to produce electricity in the electric power sector.
Consumption of coal and natural gas for electric power decreased by 2.9 percent and 7.2 percent, respectively, from
2016 to 2017. There lias 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 14 percent from 2016 to 2017 (see
Table 3-12). The decrease in coal-powered electricity generation and increase in renewable energy electricity
generation contributed to a decrease in emissions from electric power generation over the time series (see Figure
3-8).
Decreases in natural gas costs and the associated increase in natural gas generation, particularly between 2005 and
2017, 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 49 percent while the cost of coal (in
$/MMBtu) increased by 78 percent (EIA 2019a). Also, between 1990 and 2017, 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 2017,
which also helped drive the decrease in electric power sector carbon intensity. This decrease in carbon intensity
17 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 EIA-defmed electric power sector, it is typically for the entity's own use.
18 Values represent electricity net generation from the electric power sector (EIA 2019a).
Energy 3-15
-------
occurred even as total electricity retail sales increased 37 percent, from 2,713 billion kWh in 1990 to 3,723 billion
kWhin 2017.
Figure 3-8: Fuels Used in Electric Power Generation (TBtu) and Total Electric Power Sector
CO2 Emissions
50,000
40,000
H 30,000
CD
tn
s 20,000
10,000
Petroleum (TBtu)
Nuclear (TBtu)
Renewable Energy Sources (TBtu)
Natural Gas (TBtu)
Coal (TBtu)
I Net Generation (Index from 1990) [Right Axis]
I Sector CO2 Emissions (Index from 1990) [Right Axis]
180
160
140
120
100
80
60
40
20
o
CT1
X
0)
T3
OT-Hrsiro^-LnuDrvcocnoT-HrMro^j-LniorvcocnO'^H
otoio-tCTatoiotoiototooooooooooiHi-i
C7icrtC7icrto^cr»o^cr»
CT1 cri CD CT1 CTi cn
OT-HrMro'd-mkor-soocnOT-irMPO^i-LrivDhs
OOOOOOOOOOt-It—li—l»-li—I tHIt—It—I
OOOOOOOOOOOOOOOOOO
CNCMfMCNCMfMCNCMrMCNCMfMrMCMfMrMCMfM
In 2017, electricity sales to the residential and commercial end-use sectors, as presented in Figure 3-9, decreased by
2.3 percent and 1.0 percent relative to 2016, respectively. Electricity sales to the industrial sector in 2017 increased
approximately 0.8 percent relative to 2016. Overall, in 2017, the amount of electricity retail sales (in kWh)
decreased by 1.0 percent relative to 2016.
3-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Industrial Sector
Industrial sector CO2, CH4, and N20, emissions accounted for 17, 15, and 6 percent of CO2, CH4, and N20,
emissions from fossil fuel combustion, respectively. Carbon dioxide, CH4, and N20 emissions resulted from the
direct consumption of fossil fuels for steam and process heat production.
The industrial end-use sector, per the underlying energy use data from EIA, includes activities such as
manufacturing, construction, mining, and agriculture. The largest of these activities in terms of energy use is
manufacturing, of which six industries—Petroleum Refineries, Chemicals, Paper, Primary Metals, Food, and
Nonmetallic Mineral Products—represent the vast majority of the energy use (EIA 2019a, 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.19 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 2016 to 2017, total industrial production and manufacturing output increased by 1.6 percent (FRB 2019). Over
this period, output increased across production indices for Food, Petroleum Refineries, Chemicals, and Nonmetallic
Mineral Products, and decreased slightly for Primary Metals and Paper (see Figure 3-10). 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 2016 to 2017, the underlying EIA data showed decreased consumption of coal, and increase of
natural gas in the industrial sector. The GHGRP data highlights that several industries contributed to these trends,
including chemical manufacturing; pulp, paper and print; food processing, beverages and tobacco; minerals
manufacturing; and agriculture-forest-fisheries.20
19 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.
20 Further details on industrial sector combustion emissions are provided by EPA's GHGRP. See
.
Energy 3-17
-------
Figure 3-
140
120
100
80
60
10:
Industrial Production Indices (Index 2012=100)
Total excluding Computers, Communications Equipment, and Semiconductors
Total Industrial
Paper
120
100
Food
80
60
Stone, Clay, and Glass Products
140
120
100
Chemicals
80
60
Primary Metals
Petroleum Refineries
CTi CTl O
-------
Figure 3-11: Fuels Used in Residential and Commercial Sectors (TBtu), Heating Degree Days,
and Total Sector CO2 Emissions
25,000
20,000
m
115,000
CD
<0 10,000
Coal (TBtu)
Renewable Energy Sources (TBtu)
Petroleum (TBtu)
Natural Gas (TBtu)
I Electricity Use (TBtu)
I Heating Degree Days (Index vs. 1990) [Right Axis]
I Sector CO2 Emissions (Index vs. 1990) [Right Axis]
5,000
180
160
140
120
100
80
60
40
20
0
i-irMm^-i/>voivooCT»o*HrNm^-m^orNcooi
cncncncncnCTicricricrtoooooooooo
CT>C*CT»0Cr»CTiCr>CTiCT»00000000000000000
In 2017 the residential and commercial sectors accounted for 6 and 5 percent of CO2 emissions from fossil fuel
combustion, respectively, 35 and 11 percent of CH4 emissions from fossil fuel combustion, respectively, and 2 and 1
percent of N20 emissions from fossil fuel combustion, respectively. Emissions from these sectors were largely due
to the direct consumption of natural gas and petroleum products, primarily for heating and cooking needs. Coal
consumption was a minor component of energy use in both of these end-use sectors. In 2017, total emissions (CO2,
CH4, and N20) from fossil fuel combustion and electricity use within the residential and commercial end-use sectors
were 925.4 MMT CO2 Eq. and 849.7 MMT CO2 Eq., respectively. Total CO2, CH4, and N20 emissions from
combined fossil fuel combustion and electricity use within the residential and commercial end-use sectors decreased
by 3.7 and 3.1 percent from 2016 to 2017, respectively, and heating degree days decreased by 1 percent over the
same time period. A decrease in heating degree days impacted demand for heating fuel and electricity for heat in the
residential and commercial sectors. In addition a shift toward energy efficient products and more stringent energy
efficiency standards for household equipment lias also contributed to a decrease in energy demand in households
(EIA 2018d), resulting in a decrease in energy-related emissions. In the long term, the residential sector is also
affected by population growth, migration trends toward wanner areas, and changes in total housing units and
building attributes (e.g., larger sizes and improved insulation).
In 2017, combustion emissions from natural gas consumption represented 82 and 74 percent of the direct fossil fuel
CO2 emissions from the residential and commercial sectors, respectively. Natural gas combustion CO2 emissions
from the residential and commercial sectors in 2017 increased by 1.3 percent and 1.5 percent from 2016 levels,
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, CH4, and N20 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.
Energy 3-19
-------
Transportation Sector and Mobile Combustion
This discussion of transportation emissions follows the alternative method of presenting combustion emissions by
allocating emissions associated with electricity generation to the transportation end-use sector, as presented in
Table 3-8. Table 3-7 presents direct CO2, CH4, and N20 emissions from all transportation sources (i.e., excluding
emissions allocated to electricity consumption in the transportation end-use sector).
The transportation end-use sector and other mobile combustion accounted for 1,825.1 MMT CO2 Eq. in 2017, which
represented 37 percent of CO2 emissions, 29 percent of CH4 emissions, and 37 percent of N20 emissions from fossil
fuel combustion, respectively.21 Fuel purchased in the United States for international aircraft and marine travel
accounted for an additional 121.2 MMT CO2 Eq. in 2017; 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 2017, transportation emissions from fossil fuel combustion rose by 20 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 45 percent from 1990 to 2017,22 as a result of a confluence
of factors including population growth, economic growth, urban sprawl, and periods of low fuel prices.
From 2016 to 2017, CO2 emissions from the transportation end-use sector increased by 1.21 percent. The small
increase in emissions is attributed to an increase in diesel fuel consumption by medium- and heavy-duty trucks and
jet fuel consumption by commercial aircraft.
Commercial aircraft emissions increased between 2016 and 2017, but have decreased 8 percent since 2007 (FAA
20 1 9).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 23 percent from 1990 to
2017. Annex 3.2 presents the total emissions from all transportation and mobile sources, including CO2, N20, CH4,
and HFCs.
21 Note that these totals include CO2, CH4 and N2O emissions from some sources in the U.S. Territories (ships and boats,
recreational boats, non-transportation mobile sources) and CH4 and N2O emissions from transportation rail electricity.
22 VMT estimates are based on data from FHWA Highway Statistics Table VM-1 (FHWA 1996 through 2017). 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 2017 time period. In absence of these method changes, light-duty VMT growth between 1990 and 2017
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.
3-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 3-12: Fuels Used in Transportation Sector (TBtu), Onroad VMT, and Total Sector CO2
Emissions
Aviation Gasoline (TBtu)
LPG (TBtu)
Residual Fuel (TBtu)
Natural Gas (TBtu)
Renewable Energy (TBtu)
Jet Fuel (TBtu)
Distillate Fuel (TBtu)
Motor Gasoline (TBtu)
Onroad VMT (Index vs. 1990) [Right Axis]
I Sector CO2 Emissions (Index vs. 1990) [Right Axis]
180
160
140
120
100
80
60
40
20
o^fMm^-Lovor^ooo>
CTi 01 CTi cn CTi CTi cx>
cn CT>
rgrMf>jrMfNirMfNirMr«jrNjr>JrNjnjrvJrMf>jr«»l<>l
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 CO 2 Emissions
Domestic transportation CO2 emissions increased by 23 percent (332.8 MMT CO2) between 1990 and 2017, an
annualized increase of 0.8 percent. Among domestic transportation sources in 2017, light-duty vehicles (including
passenger cars and light-duty trucks) represented 58 percent of CO2 emissions from fossil fuel combustion, medium-
and heavy-duty trucks and buses 25 percent, commercial aircraft 7 percent, and other sources 10 percent. See Table
3-13 for a detailed breakdown of transportation 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 from the transportation
sector lias increased from 0.7 billion gallons in 1990 to 13.4 billion gallons in 2017, while biodiesel consumption
has increased from 0.01 billion gallons in 2001 to 2.0 billion gallons in 2017. For further 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 1,055.4 MMT CO2 in 2017. This is an
increase of 14 percent (130.9 MMT CO2) from 1990 due, in large part, to increased demand for travel as fleet-wide
light-duty vehicle fuel economy was relatively stable (average new vehicle fuel economy declined slowly from 1990
through 2004 and then increased more rapidly from 2005 through 2017). 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 8 percent. The
decline in new light-duty vehicle fuel economy between 1990 and 2004 (Figure 3-13) reflected the increasing
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 3-21
-------
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 1325 and lias since grown at a faster rate (2.5 percent from 2015 to 2016, and 1.0 percent from 2016 to
2017). 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 45 percent.
Light-duty truck share is about 42 percent of new vehicles in model year 2017 (EPA 2018a). See also 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 88 percent from 1990 to 2017. This increase was largely
due to a substantial growth in medium- and heavy-duty truck VMT, which increased by 107 percent between 1990
and 20 1 7.26 Carbon dioxide from the domestic operation of commercial aircraft increased by 17 percent (18.1 MMT
CO2) from 1990 to 2017.27 Across all categories of aviation, excluding international bunkers, CO2 emissions
decreased by 8 percent (14.2 MMT CO2) between 1990 and 2017.28 This includes a 65 percent (22.8 MMT CO2)
decrease in CO2 emissions from domestic military operations.
Transportation sources also produce CH4 and N20; these emissions are included in Table 3-14 and Table 3-15 and in
the CH4 and N20 from Mobile Combustion section. Annex 3.2 presents total emissions from all transportation and
mobile sources, including CO2, CH4, N20, and HFCs.
Figure 3-13: Sales-Weighted Fuel Economy of New Passenger Cars and Light-Duty Trucks,
1990-2017 (miles/gallon)
30
28
26
24
22
20
"*1-
o
o
C\J
cn
cn
in
cn
cn
CO
cn
cn
cn
cn
cn
o
o
o
CM
ro
00
o
o
cn
o
o
o
1
CO
lH
o
LT>
T—I
T—I
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
Source: EPA (2018a).
25 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 inFHWA'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 2017
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 2017 time period, part of
the growth reflects a method change for estimating VMT starting in 2007. This change in methodology in FEIWA'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 2017 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.
3-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 3-14: Sales of New Passenger Cars and Light-Duty Trucks, 1990-2017 (Percent)
100%
90%
Passenger Cars
» 80%
^ 70%
Light-Duty Trucks
20%
10%
0%
0 tH
01 CT>
Ol CT>
in id rv
CTi
CT* O*
CTi
Source: EPA (2018a).
Table 3-13: CO2 Emissions from Fossil Fuel Combustion in Transportation End-Use Sector
(MMT COz Eq.)
Fuel/Vehicle Type
1990
2005
2013a
2014a
2015a
2016a
2017a
Gasolineb
958.9
1,153.6
1,037.4
1,077.4
1,070.0
1,095.3
1,092.3
Passenger Cars
604.3
639.1
712.9
730.2
732.0
744.9
745.3
Light-Duty Trucks
300.6
464.9
270.6
292.2
283.5
294.6
290.3
Medium- and Heavy-Duty
Trucksc
37.7
33.9
38.5
39.8
39.3
40.4
41.3
Buses
0.3
0.4
0.8
0.9
0.9
0.9
0.9
Motorcycles
1.7
1.6
3.8
3.8
3.7
3.8
3.7
Recreational Boats'1
14.3
13.8
10.8
10.6
10.6
10.7
10.7
Distillate Fuel Oil (Diesel)b
262.9
457.5
433.9
447.7
460.7
462.7
475.8
Passenger Cars
7.9
4.2
4.1
4.1
4.3
4.3
4.4
Light-Duty Trucks
11.5
25.8
12.9
13.9
13.9
14.3
14.3
Medium- and Heavy-Duty
Trucksc
190.5
360.2
350.0
361.2
369.4
376.4
388.3
Buses
8.0
10.6
15.5
16.9
17.3
17.0
17.9
Rail
35.5
45.5
40.1
41.6
39.9
36.7
37.9
Recreational Boats'1
2.7
2.8
2.6
2.5
2.7
2.8
2.8
Ships and Non-Recreational
Boats6
6.8
8.4
8.7
7.5
13.3
11.1
10.2
International Bunker Fuel/
11.7
9.4
5.6
6.1
8.4
8.7
9.0
Jet Fuel
184.2
189.3
147.1
148.4
157.6
166.0
171.8
Commercial Aircraft8
109.9
132.7
114.3
115.2
119.0
120.4
128.0
Military Aircraft
35.0
19.4
11.0
14.0
13.5
12.3
12.2
General Aviation Aircraft
39.4
37.3
21.8
19.2
25.1
33.4
31.5
International Bunker Fuel/
38.0
60.1
65.7
69.6
71.9
74.1
77.7
International Bunker Fuels
from Commercial Aviation
30.0
55.6
62.8
66.3
68.6
70.8
74.5
Energy 3-23
-------
Aviation Gasoline
3.1
2.4
1.5
1.5
1.5
1.4
1.4
General Aviation Aircraft
3.1
2.4
1.5
1.5
1.5
1.4
1.4
Residual Fuel Oil
22.6
19.3
15.1
5.8
4.2
12.9
16.5
Ships and Boats6
22.6
19.3
15.1
5.8
4.2
12.9
16.5
International Bunker Fuel/
53.7
43.6
28.5
27.7
30.6
33.8
33.4
Natural Gas J
36.0
33.1
47.0
40.2
39.4
40.1
42.3
Passenger Cars
+
+
+
+
+
+
+
Light-Duty Trucks
+
+
+
+
+
+
+
Medium- and Heavy-Duty
Trucks
+
+
+
+
+
+
+
Buses
+
0.6
0.8
0.8
0.9
0.8
0.8
Pipeline11
36.0
32.4
46.2
39.4
38.5
39.2
41.4
LPG J
1.4
1.7
0.6
0.6
0.6
0.6
0.6
Passenger Cars
+
+
+
+
0.1
+
+
Light-Duty Trucks
0.2
0.3
0.1
0.1
0.1
0.1
0.1
Medium- and Heavy-Duty
Trucksc
1.1
1.3
0.4
0.4
0.4
0.4
0.4
Buses
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Electricity1
3.0
4.7
4.3
4.5
4.3
4.2
4.3
Passenger Cars
+
+
0.2
0.4
0.5
0.7
0.8
Light-Duty Trucks
+
+
+
+
+
0.1
0.1
Buses
+
+
+
+
+
+
+
Rail
3.0
4.7
4.0
4.0
3.7
3.5
3.4
Totalk
1,472.1
1,861.7
1,686.9
1,726.0
1,738.2
1,783.2
1,804.9
Total (Including Bunkers)'
1,575.6
1,974.9
1,786.7
1,829.4
1,849.1
1,899.8
1,925.0
Biofuels-Ethanol'
4.1
21.6
70.5
74.0
74.2
76.9
77.7
Biofuels-Biodiesel' +
0.9
13.5
13.3
14.1
19.6
18.7
+ Does not exceed 0.05 MMT CO2 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 2017 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-21, MF-27, and VM-1 (FHWA 1996 through 2017). Table VM-1 fuel consumption data for 2017 has not been published
yet, therefore 2017 fuel consumption data is estimated using the percent change in VMT from 2016 to 2017. 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 2017 has not been published yet, therefore 2016 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 2017.
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.
g Commercial aircraft, as modeled in FAA's Aviation Environmental Design Tool (AEDT), consists of passenger aircraft,
cargo, and other chartered flights.
hPipelines reflect CO2 emissions from natural gas-powered pipelines transporting natural gas.
'Ethanol andbiodiesel 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.
J 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 1990-2016 Inventory and apply to
the 1990 to 2017 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, CO2 emissions from electric vehicle
3-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
charging were allocated to the residential and commercial sectors. They are now allocated to the transportation sector. These
changes were first incorporated in this year's Inventory and apply to the 2010 through 2017 time period.
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 CO2 emissions from U.S. Territories, since these are
covered in a separate chapter of the Inventory. Totals may not sum due to independent rounding.
Mobile Fossil Fuel Combustion CH4 andN2O Emissions
Mobile combustion includes emissions of CH4 and N20 from all transportation sources identified in the U.S.
Inventory with the exception of pipelines and electric locomotives;29 mobile sources also include non-transportation
sources such as construction/mining equipment, agricultural equipment, vehicles used off-road, and other sources
(e.g., snowmobiles, lawnmowers, etc.). 30 Annex 3.2 includes a summary of all emissions from both transportation
and mobile sources. Table 3-14 and Table 3-15 provide mobile fossil fuel CH4 and N20 emission estimates in MMT
C02Eq.31
Mobile combustion was responsible for a small portion of national CH4 emissions (0.5 percent) but was the fourth
largest source of U.S. N20 emissions (4.7 percent). From 1990 to 2017, mobile source CH4 emissions declined by
75 percent, to 3.2 MMT CO2 Eq. (128 kt CH4), due largely to control technologies employed in on-road vehicles
since the mid-1990s to reduce CO, NOx, NMVOC, and CH4 emissions. Mobile source emissions of N20 decreased
by 60 percent, to 16.9 MMT CO2 Eq. (57 kt N2O). Earlier generation control technologies initially resulted in higher
N20 emissions, causing a 30 percent increase in N20 emissions from mobile sources between 1990 and 1997.
Improvements in later-generation emission control technologies have reduced N20 output, resulting in a 69 percent
decrease in mobile source N20 emissions from 1997 to 2017 (Figure 3-15). Overall, CH4 and N20 emissions were
predominantly from gasoline-fueled passenger cars and light-duty trucks. See also Annex 3.2 for data by vehicle
mode and information on VMT and the share of new vehicles (in VMT).
29 Emissions of CH4 from natural gas systems are reported separately. More information on the methodology used to calculate
these emissions are included in this chapter and Annex 3.4.
30 See the methodology sub-sections of the CO2 from Fossil Fuel Combustion and CH4 and N2O from Mobile Combustion
sections of this chapter. Note that N2O and CH4 emissions are reported using different categories than CO2. CO2 emissions are
reported by end-use sector (Transportation, Industrial, Commercial, Residential, U.S. Territories), and generally adhere to a top-
down approach to estimating emissions. CO2 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 CO2 emissions
are included in the Industrial end-use sector instead of the Transportation end-use sector). CH4 and N2O emissions are reported
using the "Mobile Combustion" category, which includes non-transportation mobile sources. CH4 and N2O 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, CO2 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 2017.
Energy 3-25
-------
Figure
3-15: Mobile Source ChU and N2O Emissions (MMT CO2 Eq.)
Table 3-14: CH4 Emissions from Mobile Combustion (MMT CO2 Eq.)
Fuel Type/Vehicle Type3
1990
2005
2013
2014
2015
2016
2017
Gasoline On-Roadb
5.2
2.2
1.1
1.0
0.9
0.9
0.8
Passenger Cars
3.2
1.3
0.8
0.7
0.6
0.6
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
+
+
+
+
Motorcycles
+
+
+
+
+
+
+
Diesel On-Roadb
+
+
+
+
+
+
+
Passenger Cars
+
+
+
+
+
+
+
Light-Duty Trucks
+
+
+
+
+
+
+
Medium- and Heavy-Duty
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
3.1
2.8
2.4
2.3
2.2
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 Equipment4
0.6
0.6
0.2
0.2
0.1
0.1
0.1
C onstraction/Mining
Equipment6
0.9
1.0
0.7
0.6
0.5
0.4
0.4
Otherf
5.5
4.9
1.8
1.6
1.5
1.4
1.3
Total
12.9
9.6
4.5
4.1
3.6
3.4
3.2
+ Does not exceed 0.05 MMT CO2 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 2017). Table VM-1 fuel consumption data for 2017 has not been published yet,
therefore 2017 fuel consumption data is estimated using the percent change in VMT from 2016 to 2017. 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 2017 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 rail diesel
consumption for 2014-2017 are not available, therefore 2013 data is used as a proxy. Commuter and intercity rail
diesel consumption data for 2017 is not available yet, therefore 2016 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.
3-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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 2017 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.
Table 3-15: N2O Emissions from Mobile Combustion (MMT CO2 Eq.)
Fuel Type/Vehicle Type3
1990
2005
2013
2014
2015
2016
2017
Gasoline On-Roadb
37.5
33.5
17.2
15.4
14.0
12.8
11.7
Passenger Cars
24.1
17.5
11.8
10.5
9.7
8.9
8.2
Light-Duty Trucks
12.8
15.0
4.7
4.4
3.8
3.5
3.1
Medium- and Heavy-Duty
Trucks and Buses
0.5
0.9
0.6
0.5
0.5
0.4
0.4
Motorcycles
+
+
+
+
+
+
+
Diesel On-Roadb
0.2
0.3
0.4
0.4
0.4
0.4
0.4
Passenger Cars
+
+
+
+
+
+
+
Light-Duty Trucks
+
+
+
+
+
+
+
Medium- and Heavy-Duty
Trucks and Buses
0.2
0.3
0.4
0.4
0.4
0.4
0.4
Alternative Fuel On-Road
+
+
+
+
+
+
+
Non-Road
4.4
5.2
4.6
4.5
4.5
4.7
4.9
Ships and Boats
0.6
0.6
0.5
0.3
0.4
0.5
0.5
Railc
0.3
0.4
0.4
0.4
0.4
0.3
0.3
Aircraft
1.7
1.8
1.4
1.4
1.5
1.5
1.6
Agricultural Equipment4
0.5
0.6
0.6
0.6
0.6
0.6
0.5
C onstruction/Mining
Equipment6
0.6
1.0
0.8
0.8
0.8
0.8
0.9
Otherf
0.6
0.9
0.9
1.0
1.0
1.0
1.0
Total
42.1
39.0
22.1
20.3
18.9
17.9
17.0
+ Does not exceed 0.05 MMT CO2 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 2017). Table VM-1 fuel consumption data for 2017 has not been published yet, therefore
2017 fuel consumption data is estimated using the percent change in VMT from 2016 to 2017. These mileage
estimates are combined with estimates of fuel shares by vehicle type from DOE's TEDB Annex Tables A. 1 through
A.6 (DOE 1993 through 2017). TEDB data for 2017 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 rail diesel
consumption for 2014-2017 are not available, therefore 2013 data is used as a proxy. Commuter and intercity rail
diesel consumption data for 2017 is not available yet, therefore 2016 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.
Note: In 2011, FEIWA 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 2017 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.
Energy 3-27
-------
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, etc.), primary
fuel type (e.g., coal, petroleum, gas), and secondary fuel category (e.g., motor gasoline, distillate fuel oil,
etc.). 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). The EIA does not
include territories in its national energy statistics, so fuel consumption data for territories were collected
separately fromEIA's International Energy Statistics (EIA 2017).33
For consistency of reporting, the IPCC has recommended that countries report energy data using the
International Energy Agency (IEA) reporting convention and/or IEA data. Data in the IEA format are
presented "top down"—that is, energy consumption for fuel types and categories are estimated from energy
production data (accounting for imports, exports, stock changes, and losses). The resulting quantities are
referred to as "apparent consumption." The data collected in the United States by EIA on an annual basis
and used in this Inventory are predominantly from mid-stream or conversion energy consumers such as
refiners and electric power generators. These annual surveys are supplemented with end-use energy
consumption surveys, such as the Manufacturing Energy Consumption Survey, that are conducted on a
periodic basis (every four years). These consumption 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 (2019a, 2019b, 2018a), USAA (2008 through 2018), USGS
(1991 through 2015a), (USGS 2018b), USGS (2014 through 2018b), USGS (2014 through 2017), USGS
(1995 through 2013), USGS (1995, 1998, 2000, 2001, 2002, 2007), USGS (2018a), USGS (1991 through
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 CO2 Eq. in 2017.
34 See IPCC Reference Approach for Estimating CO2 Emissions from Fossil Fuel Combustion in Annex 4 for a comparison of
U.S. estimates using top-down and bottom-up approaches.
35 A crude convention to convert between gross and net calorific values is to multiply the heat content of solid and liquid fossil
fuels by 0.95 and gaseous fuels by 0.9 to account for the water content of the fuels. Biomass-based fuels in U.S. energy statistics,
however, are generally presented using net calorific values.
3-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
2015c), USGS (1991 through 2017), USGS (2018b), USGS (2014 through 2018a), USGS (1996 through
2013), USGS (1991 through 2015b), USGS (2018c), USGS (1991 through 2015c).36
3. Adjust for conversion offuels and exports of CO 2. Fossil fuel consumption estimates are adjusted
downward to exclude fuels created from other fossil fuels and exports of CO2.37 Synthetic natural gas is
created from industrial coal, and is currently included in EIA statistics for both coal and natural gas.
Therefore, synthetic natural gas is subtracted from energy 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 fossil fuel 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 (2018a), and data for CO2 exports were
collected from the Eastman 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 was 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 2018), EPA (2018b), and
FHWA (1996 through 2017).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 was 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 C
content).40 The Office of the Under Secretary of Defense (Installations and Environment) and the Defense
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 Energy statistics from EIA (2019a) are already adjusted downward to account for ethanol added to motor gasoline, biodiesel
added to diesel fuel, and biogas in natural gas.
38 These adjustments are explained in greater detail in Annex 2.1.
39 Bottom-up gasoline and diesel highway vehicle fuel consumption estimates are based on data from FHWA Highway Statistics
Table MF-21, MF-27, and VM-1 (FHWA 1996 through 2017).
40 See International Bunker Fuels section in this chapter for a more detailed discussion.
Energy 3-29
-------
Logistics Agency Energy (DLA Energy) of the U.S. Department of Defense (DoD) (DLA Energy 2018)
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
2018) for 1990 through 2001 and 2007 through 2017, 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 C 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 C
content coefficients used by the United States were obtained from EIA's Emissions of Greenhouse Gases in
the United States 2008 (EIA 2009a), and an EPA analysis of C content coefficients developed for the
GHGRP (EPA 2010). A discussion of the methodology used to develop the C content coefficients are
presented in Annexes 2.1 and 2.2.
8. Estimate C02 Emissions. Total CO2 emissions are the product of the adjusted energy consumption (from
the previous methodology steps 1 through 6), the C 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).
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 2017); for each vehicle category, the
percent gasoline, diesel, and other (e.g., CNG, LPG) fuel consumption are estimated using data from
DOE (1993 through 20 1 7).43'44
• For non-road vehicles, activity data were obtained from AAR (2008 through 2018), APTA (2007
through 2017), APTA (2006), BEA (2018), Benson (2002 through 2004), DOE (1993 through 2017),
DLA Energy (2018), DOC (1991 through 2018), DOT (1991 through 2017), EIA (2009a), EIA
(2019a), EIA (2018c), EIA (1991 through 2018), EPA (2018b) 45 and Gaffney (2007).
• For jet fuel used by aircraft, CO2 emissions from commercial aircraft were developed by the U. S.
41 Data for 2002 were interpolated due to inconsistencies in reported fuel consumption data.
42 For a more detailed description of the data sources used for the analysis of the transportation end use sector see the Mobile
Combustion (excluding CO2) 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. 1
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 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 2017.
3-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Federal Aviation Administration (FAA) using a Tier 3B methodology, consistent IPCC (2006) (see
Annex 3.3). Carbon dioxide emissions from other aircraft were calculated directly based on reported
consumption of fuel as reported by EIA. Allocation to domestic military uses was made using DoD
data (see Annex 3.8). General aviation jet fuel consumption is calculated as the remainder of total jet
fuel use (as determined by EIA) nets all other jet fuel use as determined by FAA and DoD. For more
information, see Annex 3.2.
Box 3-4: 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 lias provided an opportunity to better characterize the industrial sector's energy consumption and emissions
in the United States, through a disaggregation of EIA's industrial sector fuel consumption data from select
industries.
For GHGRP 2010 through 2017 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 lias provided guidance on aligning facility-level
reported fuels and fuel types published in national energy statistics, which guided this exercise.46
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.47 The efforts in reconciling fuels focus on standard, common fuel types (e.g.,
natural gas, distillate fuel oil, etc.) where the fuels inEIA's national statistics aligned well with facility-level
GHGRP data. For these reasons, the current information presented in the 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 2017 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.
Box 3-5: Carbon Intensity of U.S. Energy Consumption
The amount of C emitted from the combustion of fossil fuels is dependent upon the C content of the fuel and the
fraction of that C that is oxidized. Fossil fuels vary in their average C 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. In general, the C
content per unit of energy of fossil fuels is the highest for coal products, followed by petroleum and then natural
gas. The overall C 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 C 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 C
46 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 .
47 See .
Energy 3-31
-------
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 C intensity, which is related to the large percentage of its energy
derived from natural gas for heating. The C intensity of the commercial sector lias 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 C intensity of the transportation sector was closely related to the C content of petroleum products (e.g.,
motor gasoline and jet fuel, both around 70 MMT CO2 Eq./EJ), which were the primary sources of energy.48 Lastly,
the electric power sector had the highest C 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
2013
2014
2015
2016
2017
Residential3
57.4
56.6
55.3
55.4
55.5
55.1
55.0
Commercial3
59.6
57.7
56.0
55.7
57.2
56.8
56.6
Industrial3
64.4
64.5
62.0
61.5
61.2
60.8
60.5
Transportation3
71.1
71.4
71.4
71.5
71.5
71.5
71.5
Electric Powerb
87.3
85.8
81.3
81.2
78.1
76.8
77.3
U.S. Territories0
73.0
73.5
71.9
72.3
72.2
72.2
72.2
All Sectors0
73.0
73.5
70.9
70.7
69.7
69.2
69.2
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.
Note: Excludes non-energy fuel use emissions and consumption.
For the time period of 1990 through about 2008, the C 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 C intensity lias 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 2017, was approximately 10.9 percent below levels in 1990 (see Figure 3-16). To differentiate these
estimates from those of Table 3-16, the C 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 (BEA2018).
48 One exajoule (E.T) is equal to 1018 joules or 0.9478 QBtu.
3-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 3-16: U.S. Energy Consumption and Energy-Related CO2 Emissions Per Capita and Per
Dollar GDP
110
100
C02/Energy Consumption
Energy Consumption/capita
cn
o
C02/capita
Energy Consumption/$GDP
C 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).
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.
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
Energy 3-33
-------
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 120 input variables were modeled for CO2 from energy-related Fossil Fuel Combustion (including about 10
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.49 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.50
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).51 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 2017 were estimated to be between 4,806.4 and 5,135.3 MMT CO2 Eq. at a 95 percent
confidence level. This indicates a range of 2 percent below to 5 percent above the 2017 emission estimate of 4,912.0
MMT C02 Eq.
Table 3-17: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Energy-
Related Fossil Fuel Combustion by Fuel Type and Sector (MMT CO2 Eq. and Percent)
2017 Emission Estimate Uncertainty Range Relative to Emission Estimate3
Fuel/Sector (MMT CO2 Eq.) (MMT CO2 Eq.) (%^
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Coal"
1,267.5
1,223.5
1,386.1
-3%
9%
Residential
NE
NE
NE
NE
NE
Commercial
2.0
1.9
2.3
-5%
15%
Industrial
54.4
51.8
63.0
-5%
16%
Transportation
NE
NE
NE
NE
NE
Electric Power
1,207.1
1,159.9
1,321.9
-4%
10%
49 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.
50 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.
51 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.
3-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
U.S. Territories
4.0
3.5
4.8
-12%
19%
Natural Gasb
1,450.3
1,433.8
1,517.1
-1%
5%
Residential
241.5
234.7
258.5
-3%
7%
Commercial
173.2
168.3
185.2
-3%
7%
Industrial
484.7
470.3
519.6
-3%
7%
Transportation
42.3
41.1
45.2
-3%
7%
Electric Power
505.6
491.2
531.5
-3%
5%
U.S. Territories
3.0
2.6
3.5
-13%
17%
Petroleumb
2,193.7
2,061.2
2,324.4
-6%
6%
Residential
53.0
50.1
55.8
-6%
5%
Commercial
57.7
54.4
60.8
-6%
5%
Industrial
271.5
215.4
322.7
-21%
19%
Transportation
1,758.3
1,645.4
1,870.4
-6%
6%
Electric Power
18.9
18.0
20.5
-5%
9%
U.S. Territories
34.3
31.7
38.0
-8%
11%
Total (excluding Geothermal)b
4,911.6
4,805.9
5,134.7
-2%
5%
Geothermal
0.4
NE
NE
NE
NE
Total (including Geothermal)b'c
4,912.0
4,806.4
5,135.3
-2%
5%
NE (Not Estimated)
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.
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 CO2 emissions
from geothermal production.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2017. 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
insignificant.
QA/QC and Verification
In order to ensure the quality of the emission estimates from fossil fuel combustion CO2, 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 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.
Recalculations Discussion
The Energy Information Administration (EIA 2019a) updated energy consumption statistics across the time series
relative to the previous Inventory. EIA revised LPG consumption in the residential, commercial, industrial, and
transportation sectors, and lubricants consumption in the industrial and transportation sectors, for the years 2010
through 2016. EIA revised kerosene consumption in the commercial and industrial sectors, and jet fuel and
lubricants consumption in the transportation sector, for the year 1995. EIA revised distillate fuel consumption in the
residential, commercial, industrial, and transportation sectors for years 2014 through 2016. EIA updated heat
contents of motor gasoline, which changed motor gasoline consumption in the commercial, industrial, and
transportation sectors for the years 1990 through 2016. EIA also revised 1990 and 1993 natural gas consumption in
the residential, commercial, industrial, and transportation sectors, and 2016 natural gas consumption in all sectors. In
addition, the number of significant figures increased for industrial coking coal, industrial "other" coal, and industrial
"other" petroleum consumption data obtained from EIA, which decreased total energy consumption in the industrial
sector by less than 0.05 percent but increased the precision of the data.
Energy 3-35
-------
Annual carbon contents were updated for some fuels based on the availability of new data. Annual coal carbon
contents were updated across the time series to incorporate domestic coal production data obtained from EIA
(2018b), as well as state-specific coal sample data for Montana, Illinois, and Indiana. Annual natural gas carbon
contents were also updated across the time series to incorporate annual heat content data for natural gas obtained
from EIA (2019a). See Annex 2.2 for more detail on how coal and natural gas carbon contents are calculated.
Within the transportation sector, electricity use from electric vehicle charging in commercial and residential
locations was re-allocated from the Residential and Commercial sectors to the Transportation sector, starting in
2010. See the recalculations discussion of the CH4 and N20 from Mobile Combustion section below for more detail
on how EV electricity use was estimated.
EPA also updated the methodology used to calculate emissions from geothermal electricity production. Technology-
specific emission factors for geothermal electricity production obtained from EPA (2018c) were applied to total net
electricity generation by geotype (i.e., binary, flash steam, dry steam) obtained from EIA (2019c).
Revisions to LPG, lubricants, kerosene, jet fuel, distillate fuel, and motor gasoline consumption resulted in an
average annual decrease of 1.7 MMT CO2 Eq. (0.1 percent) in CO2 emissions from petroleum. Revisions to natural
gas consumption resulted in an average annual decrease of 0.2 MMT CO2 Eq. (less than 0.05 percent), while updates
to annually variable natural gas carbon contents resulted in an average annual increase of 0.2 MMT CO2 Eq. (less
than 0.05 percent), in CO2 emissions from natural gas. In aggregate, these changes resulted in an average annual
increase of 0.1 MMT CO2 Eq. (less than 0.05 percent) in CO2 emissions from natural gas. Revisions to annually
variable coal carbon contents resulted in an average annual decrease of 0.6 MMT CO2 Eq. (less than 0.05 percent) in
CO2 emissions from coal. Updates to the methodology for estimating emissions from geothermal electricity
production resulted in an average annual increase of 0.1 MMT CO2 Eq. (20.6 percent) in CO2 emissions from
geothermal electricity production. Overall, these changes resulted in an average annual decrease of 2.1 MMT CO2
Eq. (less than 0.05 percent) in CO2 emissions from fossil fuel combustion for the period 1990 through 2016, relative
to the previous Inventory.
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.52 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
EIA'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
52 See .
3-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.53
An ongoing planned improvement is to develop improved estimates of domestic waterborne fuel consumption. The
Inventory estimates for residual and distillate fuel used by ships and boats is based in part on data on bunker fuel use
from the U.S. Department of Commerce. Domestic fuel consumption is estimated by subtracting fuel sold for
international use from the total sold in the United States. It may be possible to more accurately estimate domestic
fuel use and emissions by using detailed data on marine ship activity. The feasibility of using domestic marine
activity data to improve the estimates will continue to be investigated.
EPA will evaluate and potentially update methods for allocating motor gasoline consumption to the transportation,
industrial, and commercial sectors. 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.54 EPA will continue to explore approaches to address this inconsistency, including using MOVES
on-road fuel consumption output to define the percentage of the FHWA consumption totals (from MF -21) that are
attributable to transportation, and applying that percentage to the EIA total. This would define gasoline consumption
from transportation, such that the remainder would be defined as consumption by the industrial and commercial
sectors.
EPA will continue to evaluate updates to the annual coal carbon content coefficients, such as continuing to integrate
new information from state-level geological surveys. EPA will also explore potential updates to annual variability in
carbon contents for petroleum fuels developed by EIA, such as potential updates to the data sources used to develop
transportation CO2 factors for motor gasoline and low-sulfur diesel fuel.
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.
Dm Stationary Combustion
Methodology
Methane and N20 emissions from stationary combustion were estimated by multiplying fossil fuel and wood
consumption data by emission factors (by sector and fuel type for industrial, residential, commercial, and U.S.
Territories; and by fuel and technology type for the electric power sector). The electric power sector utilizes a Tier 2
methodology, whereas all other sectors utilize a Tier 1 methodology. The activity data and emission factors used are
described in the following subsections.
Industrial, Residential, Commercial, and U.S. Territories
National coal, natural gas, fuel oil, and wood consumption data were grouped by sector: industrial, commercial,
residential, and U.S. Territories. For the CH4 and N20 emission estimates, consumption data for each fuel were
obtained from EIA'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 EIA's International
Energy Statistics (EIA 2017).55 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
53 See .
54 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.
55 U.S. Territories data also include combustion from mobile activities because data to allocate territories' energy use were
unavailable. For this reason, CH4 and N2O emissions from combustion by U.S. Territories are only included in the stationary
combustion totals.
Energy 3-37
-------
mobile source fuel use was obtained from EPA (2018b) and FHWA (1996 through 2017). Estimates for wood
biomass consumption for fuel combustion do not include municipal solid waste, tires, etc., that are reported as
biomass by EIA. Non-CCh emissions from combustion of the biogenic portion of municipal solid waste and tires is
included under waste incineration. Tier 1 default emission factors for the industrial, commercial, and residential end-
use sectors were provided by the 2006IPCC 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 onEPA's Acid Rain Program Dataset (EPA 2018a). 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 (2018a) data. The
combustion technology and fuel use data by facility obtained from EPA (2018a) were only available from 1996 to
2017, so the consumption estimates from 1990 to 1995 were estimated by applying the 1996 consumption ratio by
combustion technology type from EPA (2018a) 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,
natural gas-fired turbines, and combined cycle natural gas units.56
More detailed information on the methodology for calculating emissions from stationary combustion, including
emission factors and activity data, is provided in Annex 3.1.
Uncertainty and Time-Series Consistency
Methane emission estimates from stationary sources exhibit high uncertainty, primarily due to difficulties in
calculating emissions from wood combustion (i.e., fireplaces and wood stoves). The estimates of CH4 and N20
emissions presented are based on broad indicators of emissions (i.e., fuel use multiplied by an aggregate emission
factor for different sectors), rather than specific emission processes (i.e., by combustion technology and type of
emission control).
An uncertainty analysis was performed by primary fuel type for each end-use sector, using the IPCC-recommended
Approach 2 uncertainty estimation methodology, Monte Carlo Stochastic Simulation technique, with @RISK
software.
The uncertainty estimation model for this source category was developed by integrating the CH4 and N20 stationary
source inventory estimation models with the model for 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 N20 emission factors, based on the SAIC/EIA (2001) report.57 For these variables, the uncertainty
56 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.
57 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.
3-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
ranges were assigned to the input variables based on the data reported in SAIC/EIA (2001).58 However, the CH4
emission factors differ from those used by EIA. These factors and uncertainty ranges are based on IPCC default
uncertainty estimates (IPCC 2006).
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-18. Stationary
combustion CH4 emissions in 2017 (including biomass) were estimated to be between 5.2 and 17.5 MMT CO2 Eq. at
a 95 percent confidence level. This indicates a range of 33 percent below to 124 percent above the 2017 emission
estimate of 7.8 MMT CO2 Eq.59 Stationary combustion N20 emissions in 2017 (including biomass) were estimated
to be between 20.6 and 43.4 MMT CO2 Eq. at a 95 percent confidence level. This indicates a range of 28 percent
below to 52 percent above the 2017 emission estimate of 28.6 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
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Upper
Lower Upper
Bound
Bound
Bound Bound
Stationary Combustion
CH4
7.8
5.2
17.5
-33% +124%
Stationary Combustion
N2O
28.6
20.6
43.4
-28% +52%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
The uncertainties associated with the emission estimates of CH4 and N20 are greater than those associated with
estimates of CO2 from fossil fuel combustion, which mainly rely on the carbon content of the fuel combusted.
Uncertainties in both CH4 and N20 estimates are due to the fact that emissions are estimated based on emission
factors representing only a limited subset of combustion conditions. For the indirect greenhouse gases, uncertainties
are partly due to assumptions concerning combustion technology types, age of equipment, emission factors used,
and activity data projections.
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2017 as discussed below. Details on the emission trends through time are described in more detail in the
Methodology section, above. As discussed in Annex 5, data are unavailable to include estimates of CH4 and N20
emissions from biomass use in territories, but those emissions are assumed to insignificant.
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 QA/QC and Verification Procedures section in the introduction of the
IPPU Chapter.
QA/QC and Verification
In order to ensure the quality of the emission estimates from stationary combustion non-CCh, general (IPCC Tier 1)
and category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. The Tier 2 procedures that were implemented involved
checks specifically focusing on the activity data and emission factor sources and methodology used for estimating
CH4, N20, and the greenhouse gas precursors from stationary combustion in the United States. Emission totals for
the different sectors and fuels were compared and trends were investigated.
58 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.
59 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.
Energy 3-39
-------
Recalculations Discussion
Methane and N20 emissions from stationary sources (excluding CO2) across the entire time series were revised due
to revised data from EIA (2019) and EPA (2018a) relative to the previous Inventory. Most notably, EIA (2019)
updated wood biomass consumption statistics in the residential sector from 2009 to 2016 and the commercial sector
from 2014 to 2016. EPA revised the methodology for estimating CH4 and N20 electric power emissions due to
differences between total fuel consumption in the electric power sector reported by EIA (2019) and EPA (2018a). In
addition, nitrous oxide emission factors for coal wall-fired boilers used in the electric power sector were updated
from 0.5 kg/TJ to 5.8 kg/TJ to be consistent with EPA's Compilation of Air Pollutant Emission Factors, AP-42
(EPA 1997). The historical data changes and methodology updates resulted in an average annual increase of 0.1
MMT CO2 Eq. (1.2 percent) in CH4 emissions, and an average annual increase of 15.3 MMT CO2 Eq. (107.2
percent) in N20 emissions from stationary combustion for the 1990 through 2016 period.
Planned Improvements
Several items are being evaluated to improve the CH4 and N20 emission estimates from stationary combustion and
to reduce uncertainty for U.S. Territories. Efforts will be taken to work with EIA and other agencies to improve the
quality of the U.S. Territories data. Because these data are not broken out by stationary and mobile uses, further
research will be aimed at trying to allocate consumption appropriately. In addition, the uncertainty of biomass
emissions will be further investigated 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.
Fuel use was adjusted for the industrial sector to subtract out construction and agricultural use, which is reported
under mobile sources. Mobile source CH4 and N20 also include emissions from sources that may be captured as part
of the commercial sector. Future research will look into the need to adjust commercial sector fuel consumption to
account for sources included elsewhere.
CH4 and N20 from Mobile Combustion
Methodology
Estimates of CH4 and N20 emissions from mobile combustion were calculated by multiplying emission factors by
measures of activity for each fuel and vehicle type (e.g., light-duty gasoline trucks). Activity data included vehicle
miles traveled (VMT) for on-road vehicles and fuel consumption for non-road mobile sources. The activity data and
emission factors used are described in the subsections that follow. A complete discussion of the methodology used to
estimate CH4 and N20 emissions from mobile combustion and the emission factors used in the calculations is provided
in Annex 3.2.
On-Road Vehicles
Estimates of CH4 and N20 emissions from gasoline and diesel on-road vehicles are based on VMT and emission
factors 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.60
Emissions factors for N20 from newer on-road gasoline vehicles were calculated based upon a regression analysis
done by EPA (ICF 2017a). Methane emission factors were calculated based on the ratio of NMOG emission
standards for newer vehicles. Older gasoline vehicles on-road emissions factors 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
60 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.
3-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
emissions, and 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.61 Diesel on-road vehicle emission factors were developed by ICF
(2006b).
CH4 and N20 emission factors for AFVs were developed based on the 2017 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 operation only (tank-to-wheels); well-to-
tank emissions are calculated elsewhere in the Inventory.
Annual VMT data for 1990 through 2017 were obtained from the Federal Highway Administration's (FHWA)
Highway Performance Monitoring System database as reported in Highway Statistics (FHWA 1996 through
2017).62 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 2017).
VMT for AFVs were estimated based on Browning (2017 and 2018a). The age distributions of the U.S. vehicle fleet
were obtained from EPA (2018b, 2000), and the average annual age-specific vehicle mileage accumulation of U.S.
vehicles were obtained from EPA (2018b).
Control technology and standards data for on-road vehicles were obtained from EPA's Office of Transportation and
Air Quality (EPA 2017a, 2017b, 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).
Non-Road Mobile Sources
To estimate emissions from non-road mobile sources, fuel consumption data were employed as a measure of
activity, and multiplied by fuel-specific emission factors (in grams of N20 and CH4 per kilogram of fuel
consumed).63 Activity data were obtained from AAR (2008 through 2018), APTA (2007 through 2017), APTA
(2006), BEA (1991 through 2015), Benson (2002 through 2004), DHS (2008), DLA Energy (2018), DOC (1991
through 2018), DOE (1993 through 2017), DOT (1991 through 2017), EIA (2002, 2007, 2019a), EIA (2019b), EIA
(1991 through 2018), EPA (2018b), Esser (2003 through 2004), FAA (2019), FHWA (1996 through 2017),64
61 Additional information regarding the MOBILE model can be found online at .
62 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 2017 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.
63 The consumption of international bunker fuels is not included in these activity data, but is estimated separately under the
International Bunker Fuels source category.
64 This Inventory uses FHWA's Agriculture, Construction, and Commercial/Industrial MF-24 fuel volumes along with the
MOVES NONROAD model gasoline volumes to estimate non-road mobile source CH4 and N2O emissions for these categories.
For agriculture, the MF-24 gasoline volume is used directly because it includes both off-road trucks and equipment. For
Energy 3-41
-------
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 2017 estimates of CH4 and N20 emissions, incorporating
probability distribution functions associated with the major input variables. For the purposes of this analysis, the
uncertainty was modeled for the following four major sets of input variables: (1) VMT data, by on-road vehicle and
fuel type 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 - Uncertainty Analysis of Emission Estimates. However, a much higher level of uncertainty is associated with
CH4 and N20 emission factors due to limited emission test data, and because, unlike CO2 emissions, the emission
pathways of CH4 and N20 are highly complex.
Mobile combustion CH4 emissions from all mobile sources in 2017 were estimated to be between 2.9 and 4.1 MMT
CO2 Eq. at a 95 percent confidence level. This indicates a range of 8 percent below to 27 percent above the
corresponding 2017 emission estimate of 3.2 MMT CO2 Eq. Also at a 95 percent confidence level, mobile
combustion N20 emissions from mobile sources in 2017 were estimated to be between 15.5 and 19.3 MMT CO2
Eq., indicating a range of 8 percent below to 14 percent above the corresponding 2017 emission estimate of 16.9
MMT C02 Eq.
Table 3-19: Approach 2 Quantitative Uncertainty Estimates for ChU and N2O Emissions from
Mobile Sources (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate3
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Upper
Lower
Upper
Bound
Bound
Bound
Bound
Mobile Sources
CH4
3.2
2.9
4.1
-8%
+27%
Mobile Sources
N2O
16.9
15.5
19.3
-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 CH4 and N20 please refer to the Uncertainty Annex. As discussed in
Annex 5, data are unavailable to include estimates of CH4 and N20 emissions from any liquid fuel used in pipeline
transport or some biomass used in transportation sources, but those emissions are assumed to 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
construction and commercial/industrial gasoline estimates, the 2014 and older MF-24 volumes represented off-road trucks only;
therefore, the MOVES NONROAD 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
NONROAD equipment gasoline volumes in the construction and commercial/industrial categories.
3-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 the non-road CH4 and N20 emissions calculations for the current Inventory, resulting in
decreases across the time series in emissions from alternative fuel highway vehicles and non-highway sources, such
as construction and farm equipment. The collective result of these changes was a net increase in CH4 emissions and
a decrease in N20 emissions from mobile combustion relative to the previous Inventory. Methane emissions
increased by 0.2 percent. Nitrous oxide emissions decreased by 0.3 percent. Each of these changes is described
below.
Inventory calculations now reflect MOVES2014b, the latest version of EPA's MOVES model, released in the
summer of 2018. This update affected emissions from non-highway mobile sources across the time series. Amongst
some more minor updates, MOVES2014b includes updated non-road engine population growth rates, resulting in
generally lower equipment populations, fuel consumption, and emissions. For this year's Inventory, new non-road
CH4 and N20 emission factors were calculated using the updated 2006 IPCC Tier 3 guidance and EPA's
MOVES2014b model. Methane emission factors were calculated directly from MOVES. Nitrous oxide emission
factors were calculated using MOVES activity and emission factors by fuel type from the European Environment
Agency. Gasoline engines were broken out by 2- and 4-stroke engine types using MOVES2014b.
Alternative fuel vehicle CH4 and N20 emissions are estimated using VMT data developed by Browning (2017 and
2018a) and are based on Energy Information Administration (EIA) Alternative Fuel Vehicle Data. EIA recently
updated their historical data for vehicle counts and fuel consumption which decreased overall AFV emissions across
this Inventory year's 1990-2016 time series. This year's Inventory also includes updated and corrected methane
emissions factors forbiodiesel use in heavy-duty vehicles from 2007 and onwards.
An updated methodology (Browning 2018a) was used to estimate VMT from battery electric vehicles (BEVs), plug-
in hybrid vehicles (PHEVs), neighborhood electric vehicles (NEVs), and electric buses in this year's Inventory.
Monthly vehicle sales by make and model are now obtained from hybridcars.com, and fuel consumption by vehicle
type, make, and model is supplied by fueleconomy.gov. Average annual mileage estimates by vehicle type are
sourced from Federal Highway Administration (FHWA) Highway Statistics VM-l table. PHEVs use both electricity
and gasoline. Miles driven in all-electric mode depends upon the vehicle's all-electric range. A fleet utility factor
(SAE 2010) is now used to estimate the average percentage of miles that were all-electric based upon the all-electric
range (AER) in miles of a vehicle model. Similar to the approach for BEVs, PHEV energy consumption in all-
electric mode was calculated from vehicle counts, fuel consumption rates, and vehicle miles travelled.
Populations of NEVs from 2003 to 2010 are estimated from EIA data tables of total vehicle counts and the electric
vehicle fleet counts for all but low-speed vehicles. Fleet vehicle counts are subtracted from the total electric vehicle
counts to estimate the population of NEVs. These values are then extrapolated to calendar years after 2010 using a
regression analysis. Fuel consumption is estimated from EIA fleet data by dividing the energy used (in GGE) by the
number of vehicles and then multiplying by the number of vehicles estimated above. Fuel economy for NEVs is
assumed to be the same as light-duty automobile BEVs. All-electric bus vehicle population counts and fuel
consumption (in GGE) are directly supplied by EIA fleet data. More detail on the methods to account for emissions
from BEVs, PHEVs, NEVs, and electric buses is provided in Annex 3.2 Mobile Combustion and Browning (2018a).
Methodological recalculations were applied to the entire time series to ensure time-series consistency from 1990
through 2017 with one recent notable exception. An update by FHWA to the method for estimating on-road VMT
created an inconsistency in on-road CH4 and N20 for the time periods 1990 to 2006 and 2007 to 2017. Details on the
emission trends and methodological inconsistencies through time are described in the Methodology section, above.
Energy 3-43
-------
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.
• Evaluate and potentially update EPA's method for estimating motor gasoline consumption for non-road
mobile sources to improve accuracy and create a more consistent time series. As discussed in the
Methodology section above and in Annex 3.2, CH4 and N20 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.65 These method changes created a time-series inconsistency in the current Inventory between
2015 and previous years in CH4 and N20 estimates for agricultural, construction, commercial, and
industrial non-road mobile sources. In the current Inventory EPA has implemented one approach to address
this inconsistency. EPA will test other approaches including using 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. This percentage would then be applied 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.
• Explore updates to on-road diesel emissions factors for CH4 and N20 to incorporate diesel after treatment
technology for light-duty vehicles.
• 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 1MO GHG
Study 2014, continues to be explored.
• Update the methodology for estimating Class II and Class III rail diesel fuel consumption. For many years,
the American Short-line and Regional Railroad Association (ASLRRA) supplied annual data on Class II
and Class III rail diesel fuel consumption (national totals), but is no longer able to do so. One alternative
approach would be to estimate fuel use based on rail car loadings. EPA will explore potential updates to
annual variability in carbon contents for petroleum fuels developed by EIA, such as potential updates to the
data sources used to develop transportation C02 factors for motor gasoline and low-sulfur diesel fuel.
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
65 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.
.
3-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 3-6).
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
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 2017 from the non-energy uses of fossil fuels were 123.2 MMT
CO2 Eq„ which constituted approximately 2 percent of overall fossil fuel emissions. In 2017, the consumption of
fuels for non-energy uses (after the adjustments described above) was 5,017.5 TBtu (see Table 3-21). A portion of
the C in the 5,017.5 TBtu of fuels was stored (217.9 MMT CO2 Eq.), while the remaining portion was emitted
(123.2 MMT CO2 Eq.). Non-energy use emissions increased 8.4 percent from 2016 to 2017 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
2013
2014
2015
2016
2017
Potential Emissions
312.1
377.5
328.9
325.1
340.5
329.9
341.1
C Stored
192.5
237.9
205.4
205.2
213.6
216.2
217.9
Emissions as a % of Potential
38%
37%
38%
37%
37%
34%
36%
Emissions
119.6
139.6
123.5
119.9
126.9
113.7
123.2
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.66 67 Consumption of natural gas, LPG,
pentanes plus, naphthas, other oils, and special naphtha were adjusted to subtract out net exports of these products
66 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.
67 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 sector. Data integration is not feasible at this
time as feedstock data from EIA used to estimate non-energy uses of fuels are aggregated by fuel type, rather than disaggregated
by both fuel type and particular industries (e.g., petrochemical production) as currently collected through EPA's GHGRP and
used for the petrochemical production category.
Energy 3-45
-------
that are not reflected in the raw data from EI A. 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
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
2013
2014
2015
2016
2017
Industry
4,215.8
5,110.7
4,607.8
4,602.9
4,764.6
4,634.6
4,797.9
Industrial Coking Coal
NO
80.4
119.3
48.8
121.8
89.2
112.5
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
296.8
323.5
321.9
308.9
306.9
Asphalt & Road Oil
1,170.2
1,323.2
783.3
792.6
831.7
853.4
849.2
LPG
1,120.5
1,610.0
2,062.9
2,109.8
2,157.5
2,118.9
2,186.8
Lubricants
186.3
160.2
125.1
130.7
142.1
135.1
124.6
Pentanes Plus
117.6
95.5
45.4
43.5
78.4
53.1
81.5
Naphtha (<401 °F)
326.3
679.5
498.8
435.2
417.8
396.9
410.9
Other Oil (>401 °F)
662.1
499.5
209.1
236.2
216.8
204.0
241.7
Still Gas
36.7
67.7
166.7
164.5
162.2
166.1
163.8
Petroleum Coke
27.2
105.2
NO
NO
NO
NO
NO
Special Naphtha
100.9
60.9
96.6
104.5
97.0
88.7
94.9
Distillate Fuel Oil
7.0
11.7
5.8
5.8
5.8
5.8
5.8
Waxes
33.3
31.4
16.5
14.8
12.4
12.8
10.2
Miscellaneous Products
137.8
112.8
171.2
182.7
188.9
191.3
198.8
Transportation
176.0
151.3
143.4
149.4
162.8
154.4
142.3
Lubricants
176.0
151.3
143.4
149.4
162.8
154.4
142.3
U.S. Territories
85.6
123.2
82.4
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
81.4
76.2
76.2
76.2
76.2
Total
4,477.4
5,385.2
4,833.6
4,829.6
5,004.7
4,866.2
5,017.5
NO (Not Occurring)
Table 3-22: 2017 Adjusted Non-Energy Use Fossil Fuel Consumption, Storage, and Emissions
Adjusted Carbon
Non-Energy Content Potential Storage Carbon Carbon Carbon
Use3 Coefficient Carbon Factor Stored Emissions Emissions
3-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Sector/Fuel Type
(TBtu)
(MMT
C/QBtu)
(MMT C)
(MMT C)
(MMT C)
(MMT
CO2 Eq.)
Industry
4,797.9
NA
88.6
NA
59.0
29.6
108.5
Industrial Coking Coal
112.5
31.00
3.5
0.10
0.3
3.1
11.5
Industrial Other Coal
10.3
26.06
0.3
0.67
0.2
0.1
0.3
Natural Gas to
Chemical Plants
306.9
14.47
4.4
0.67
3.0
1.5
5.3
Asphalt & Road Oil
849.2
20.55
17.5
1.00
17.4
0.1
0.3
LPG
2,186.8
17.06
37.3
0.67
25.1
12.2
44.8
Lubricants
124.6
20.20
2.5
0.09
0.2
2.3
8.4
Pentanes Plus
81.5
19.10
1.6
0.67
1.0
0.5
1.9
Naphtha (<401° F)
410.9
18.55
7.6
0.67
5.1
2.5
9.2
Other Oil (>401° F)
241.7
20.17
4.9
0.67
3.3
1.6
5.9
Still Gas
163.8
17.51
2.9
0.67
1.9
0.9
3.4
Petroleum Coke
NO
27.85
NO
0.30
NO
NO
NO
Special Naphtha
94.9
19.74
1.9
0.67
1.3
0.6
2.2
Distillate Fuel Oil
5.8
20.17
0.1
0.50
0.1
0.1
0.2
Waxes
10.2
19.80
0.2
0.58
0.1
0.1
0.3
Miscellaneous Products
198.8
20.31
4.0
0.00
0.0
4.0
14.8
Transportation
142.3
NA
2.9
NA
0.3
2.6
9.6
Lubricants
142.3
20.20
2.9
0.09
0.3
2.6
9.6
U.S. Territories
77.3
NA
1.5
NA
0.2
1.4
5.1
Lubricants
1.0
20.20
0.0
0.09
+
+
0.1
Other Petroleum (Misc.
Prod.)
76.2
20.00
1.5
0.10
0.2
1.4
5.0
Total
5,017.5
93.0
59.4
33.6
123.2
+ Does not exceed 0.05 TBtu, MMT C, MMT CO2 Eq.
NA (Not Applicable)
NO (Not Occurring)
a To avoid double counting, net exports have been deducted.
Note: Totals may not sum due to independent rounding.
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 2018a), Toxics Release
Inventory, 1998 (EPA 2000b), Biennial Reporting System (EPA 2000a, 2009), Resource Conservation and Recovery
Act Information System (EPA 2013b, 2015, 2016b, 2018c), 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); Financial Planning Association (2006); INEGI (2006); the United States
International Trade Commission (1990 through 2017); Gosselin, Smith, and Hodge (1984); EPA's Municipal Solid
Waste (MSW) Facts and Figures (EPA 2013, 2014a, 2016a, 2018b); 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); Chemical and Engineering
News (C&EN 2017); 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, 2015a,
2016b, 2017b, 2018b); and the Guide to the Business of Chemistry (ACC 2012, 2015b, 2016a, 2017a, 2018a).
Specific data sources are listed in full detail in Annex 2.3.
Energy 3-47
-------
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 VvRISK software and the IPCC-recommended
Approach 2 methodology (Monte Carlo Stochastic Simulation technique), provides for the specification of
probability density functions for key variables within a computational structure that mirrors the calculation of the
inventory estimate. The results presented below provide the 95 percent confidence interval, the range of values
within which emissions are likely to fall, for this source category.
As noted above, the non-energy use analysis is based on U.S.-specific storage factors for (1) feedstock materials
(natural gas, 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 2017 was estimated to be between
94.3 and 168.5 MMT CO2 Eq. at a 95 percent confidence level. This indicates a range of 23 percent below to 37
percent above the 2017 emission estimate of 123.2 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)
Source
Gas
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Upper
Lower
Upper
Bound
Bound
Bound
Bound
Feedstocks
CO2
73.0
51.2
122.6
-30%
+68%
Asphalt
CO2
0.3
0.1
0.6
-59%
+118%
Lubricants
CO2
18.0
14.9
21.0
-18%
+16%
Waxes
CO2
0.3
0.2
0.6
-24%
+93%
Other
CO2
31.6
18.7
34.2
-41%
+8%
Total
CO2
123.2
94.3
168.5
-23%
37%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
Note: Totals may not sum due to independent rounding.
Table 3-24: Approach 2 Quantitative Uncertainty Estimates for Storage Factors of Non-
Energy Uses of Fossil Fuels (Percent)
Source
Gas
2017 Storage Factor
(%)
Uncertainty Range Relative to Emission Estimate3
(%) (% Relative)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Feedstocks
CO2
67.2%
54.8%
71.5%
-18%
+6%
Asphalt
CO2
99.6%
99.1%
99.8%
-0.5%
+0.3%
Lubricants
CO2
9.2%
3.9%
17.5%
-58%
+91%
Waxes
CO2
57.8%
47.4%
67.3%
-18%
+16%
Other
CO2
6.1%
6.0%
41.3%
-2%
+576%
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).
3-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
As shown in Table 3-24, feedstocks and asphalt contribute least to overall storage factor uncertainty on a percentage
basis. Although the feedstocks category—the largest use category in terms of total carbon flows—appears to have
tight confidence limits, this is to some extent an artifact of the way the uncertainty analysis was structured. As
discussed in Annex 2.3, the storage factor for feedstocks is based on an analysis of six fates that result in long-term
storage (e.g., plastics production), and eleven that result in emissions (e.g., volatile organic compound emissions).
Rather than modeling the total uncertainty around all of these fate processes, the current 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 2017 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 2016 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, 2013 through 2015, and 2017, but outputs
exceeded inputs in 2012 and 2016. 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
The Energy Information Administration (EIA 2019) updated energy consumption statistics across the time series
relative to the previous Inventory. EPA released updated hazardous waste incineration data (EPA 2018c), which
included minor updates to 2015 values. Overall, these changes resulted in an average annual increase of 0.4 MMT
CO2 Eq. (0.3 percent) in carbon emissions from non-energy uses of fossil fuels for the period 1990 through 2016,
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
Energy 3-49
-------
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 lias 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.
Box 3-6: 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.68 In this Inventory, C storage and C emissions from product use of
68 See for example Volume 3: Industrial Processes and Product Use, and Chapter 5: Non-Energy Products from Fuels and
Solvent Use of the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC 2006).
3-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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).69
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.70
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.
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—
69 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.
70 Data and calculations for lubricants and waxes and asphalt and road oil are in Annex 2.3 - Methodology for Estimating
Carbon Emitted from Non-Energy Uses of Fossil Fuels.
Energy 3-51
-------
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 2017 from this source so the data were proxied to the 2011 estimate. Carbon
dioxide emissions from incineration of waste increased 27 percent since 1990, to an estimated 10.8 MMT CO2
(10,790 kt) in 2017, as the volume of scrap tires and other fossil C-containing materials in waste increased (see
Table 3-25 and Table 3-26). Waste incineration is also a source of CH4 and N20 emissions (De Soete 1993; IPCC
2006). Methane emissions from the incineration of waste were estimated to be less than 0.05 MMT CO2 Eq. (less
than 0.5 kt CH4) in 2017, 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 N20) in 2017, 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
2013
2014
2015
2016
2017
CO2
8.0
12.5
10.3
10.4
10.7
10.8
10.8
Plastics
5.6
6.9
5.8
5.9
6.2
6.2
6.2
Synthetic Rubber in Tires
0.3
1.6
1.2
1.2
1.1
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
1.2
1.2
1.3
1.3
1.3
CH4
+
+
+
+
+
+
+
N2O
0.5
0.4
0.3
0.3
0.3
0.3
0.3
Total
8.4
12.9
10.6
10.7
11.1
11.1
11.1
+ Does not exceed 0.05 MMT CO2 Eq.
Table 3-26: CO2, ChU, and N2O Emissions from the Incineration of Waste (kt)
Gas/Waste Product
1990
2005
2013
2014
2015
2016
2017
CO2
7,950
12,469
10,333
10,429
10,742
10,765
10,790
Plastics
5,588
6,919
5,823
5,928
6,184
6,184
6,184
Synthetic Rubber in Tires
308
1,599
1,158
1,154
1,149
1,160
1,171
Carbon Black in Tires
385
1,958
1,412
1,406
1,401
1,415
1,430
Synthetic Rubber in MSW
854
766
692
692
703
703
703
Synthetic Fibers
816
1,227
1,247
1,249
1,305
1,303
1,303
CH4
+
+
+
+
+
+
+
N2O
2
1
1
1
1
1
1
+ Does not exceed 0.5 kt.
Methodology
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
3-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 fromMunicipal Solid H aste 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) and detailed unpublished backup data for some years not shown in the reports (Schneider 2007).
For 2012 through 2017, data on total waste incinerated were assumed to equal to the 2011 value from Shin (2014)
for 2012 through 2017. For synthetic rubber and carbon black in scrap tires, information was obtained biannually
from U.S. Scrap Tire Management Summary for 2005 through 2017 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 CH4 and N20. These emissions were calculated
as a function of the total estimated mass of waste incinerated and emission factors. As noted above, CH4 and N20
emissions are a function of total waste incinerated in each year; for 1990 through 2008, these data were derived from
the information published in BioCvcle (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 BioCvcle data set for 2012 through 2017, so these values were assumed to equal the 2011 BioCvcle
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)
Incinerated (% of
Year
Waste Discarded
Waste Incinerated
Discards)
1990
235,733,657
30,632,057
13.0%
2005
259,559,787
25,973,520
10.0%
2013
273,116,704a
20,756,879
7.6%
2014
273,116,704a
20,756,879
7.6%
2015
273,116,704a
20,756,879
7.6%
2016
273,116,704a
20,756,879
7.6%
2017
273,116,704a
20,756,879
7.6%
a Assumed equal to 2011 value.
Source: van Haaren et al. (2010)
Energy 3-53
-------
Uncertainty and Time-Series Consistency
An Approach 2 Monte Carlo analysis was performed to determine the level of uncertainty surrounding the estimates
of CO2 emissions and N20 emissions from the incineration of waste (given the very low emissions for CH4, no
uncertainty estimate was derived). IPCC Approach 2 analysis allows the specification of probability density
functions for key variables within a computational structure that mirrors the calculation of the Inventory estimate.
Uncertainty estimates and distributions for waste generation variables (i.e., plastics, synthetic rubber, and textiles
generation) were obtained through a conversation with one of the authors of the Municipal Solid Waste in the
United States reports. Statistical analyses or expert judgments of uncertainty were not available directly from the
information sources for the other variables; thus, uncertainty estimates for these variables were determined using
assumptions based on source category knowledge and the known uncertainty estimates for the waste generation
variables.
The uncertainties in the waste incineration emission estimates arise from both the assumptions applied to the data
and from the quality of the data. Key factors include MSW incineration rate; fraction oxidized; missing data on
waste composition; average C content of waste components; assumptions on the synthetic/biogenic C ratio; and
combustion conditions affecting N20 emissions. The highest levels of uncertainty surround the variables that are
based on assumptions (e.g., percent of clothing and footwear composed of synthetic rubber); the lowest levels of
uncertainty surround variables that were determined by quantitative measurements (e.g., combustion efficiency, C
content of C black).
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-28. Waste incineration
CO2 emissions in 2017 were estimated to be between 9.6 and 12.4 MMT CO2 Eq. at a 95 percent confidence level.
This indicates a range of 11 percent below to 15 percent above the 2017 emission estimate of 10.8 MMT CO2 Eq.
Also at a 95 percent confidence level, waste incineration N20 emissions in 2017 were estimated to be between 0.2
and 1.2 MMT CO2 Eq. This indicates a range of 47 percent below to 301 percent above the 2017 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)
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
Source
Gas
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Incineration of Waste
CO2
10.8
9.6 12.4
-11% +15%
Incineration of Waste
N2O
0.3
0.2 1.2
-47% +301%
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 use of activity data.
Recalculations Discussion
The total generation and recovery of textiles in MSW in 2015 was updated to reflect the tonnage reported in the
newest Advancing Sustainable Materials Management: Facts and Figures: Assessing Trends in Material
Generation, Recycling and Disposal in the United States (EPA 2018a), which impacted CO2 emissions from
synthetic fibers.
3-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Planned Improvements
The waste incineration estimates have recently relied on MSW mass flow (i.e., tonnage) data that lias not been
updated since 2011. These values previously came from BioCvcle (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.
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-C02 (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 BioCvcle percent combusted assumption
could be updated with Facts and Figures product tonnage combusted data.
Non-C02 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 2017, 237 underground coal mines
and 434 surface mines were operating in the United States (EIA 2018). In recent years the total number of active
coal mines in the United States has declined. In 2017, the United States was the third largest coal producer in the
world (702 MMT), after China (3,376 MMT) and India (730 MMT) (IEA 2018).
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
Energy 3-55
-------
2013
2014
2015
2016
2017
395
345
305
251
237
309,546
321,783
278,342
228,707
247,779
637
613
529
439
434
581,270
583,974
534,092
431,285
454,303
1,032
958
834
690
671
890,815
905,757
812,435
659,991
702,082
Underground mines liberate CH4 from ventilation systems and from degasification systems. Ventilation systems
pump air through the mine workings to dilute noxious gases and ensure worker safety; these systems can exhaust
significant amounts of CH4 to the atmosphere in low concentrations. Degasification systems are wells drilled from
the surface or boreholes drilled inside the mine that remove large, often highly concentrated volumes of CH4 before,
during, or after mining. Some mines recover and use CH4 generated from ventilation and degasification systems,
thereby reducing emissions to the atmosphere.
Surface coal mines liberate CH4 as the overburden is removed and the coal is exposed to the atmosphere. Methane
emissions are normally a function of coal rank (a classification related to the percentage of carbon in the coal) and
depth. Surface coal mines typically produce lower-rank coals and remove less than 250 feet of overburden so their
level of emissions is much lower than from underground mines.
In addition CH4 is released during post-mining activities, as the coal is processed, transported, and stored for use.
Total CH4 emissions in 2017 were estimated to be 2,227 kt (55.7 MMT CO2 Eq.), a decline of 42 percent since 1990
(see Table 3-30 and Table 3-31). In 2017, underground mines accounted for approximately 75 percent of total
emissions, surface mines accounted for 13 percent, and post-mining activities accounted for 12 percent. In 2017,
total CH4 emissions from coal mining increased by three percent relative to the previous year. This increase was due
to a modest increase in coal production and a slight reduction in CH4 recovered and used. The amount of CH4
recovered and used in 2017 decreased by approximately six percent compared to 2016 levels. This decrease is
primarily attributed to a decrease in reported CH4 recovery and use at three mines.
Table 3-30: ChU Emissions from Coal Mining (MMT CO2 Eq.)
Activity
1990
2005
2013
2014
2015
2016
2017
Underground (UG) Mining
74.2
42.0
46.2
46.1
44.9
40.7
41.5
Liberated
80.8
59.7
64.5
63.0
61.2
57.0
56.7
Recovered & Used
(6.6)
(17.7)
(18.3)
(16.9)
(16.3)
(16.3)
(15.2)
Surface Mining
10.8
11.9
9.7
9.6
8.7
6.8
7.2
Post-Mining (UG)
9.2
7.6
6.6
6.7
5.8
4.8
5.3
Post-Mining (Surface)
2.3
2.6
2.1
2.1
1.9
1.5
1.6
Total
96.5
64.1
64.6
64.6
61.2
53.8
55.7
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values.
ible 3-31: ChU Emissions from Coal Mining (kt)
Activity
1990
2005
2013
2014
2015
2016
2017
UG Mining
2,968
1,682
1,849
1,844
1,796
1,629
1,662
Liberated
3,234
2,390
2,580
2,522
2,448
2,279
2,270
Recovered & Used
(266)
(708)
(730)
(677)
(652)
(650)
(608)
Surface Mining
430
475
388
386
347
273
290
Post-Mining (UG)
368
306
263
270
231
193
213
Post-Mining (Surface)
93
103
84
84
75
59
63
Total
3,860
2,565
2,584
2,583
2,449
2,154
2,227
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values.
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.
Methodology
3-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• 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 surface
mines and post-mining activities consists of multiplying basin-specific coal production by basin-specific
gas content and an emission factor.
Step 1: Estimate CH4 Liberated and CH4 Emitted from Underground Mines
Underground mines generate CH4 from ventilation systems and degasification systems. Some mines recover and use
the liberated CH4, thereby reducing emissions to the atmosphere. Total CH4 emitted from underground mines equals
the CH4 liberated from ventilation systems, plus the CH4 liberated from degasification systems, minus the CH4
recovered and used.
Step 1.1: Estimate CH4 Liberatedfrom Ventilation Systems
To estimate CH4 liberated from ventilation systems, EPA uses data collected through its Greenhouse Gas Reporting
Program (GHGRP)71 (subpart FF, "Underground Coal Mines"), data provided by the U.S. Mine Safety and Health
Administration (MSHA), 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 1 8).72 Mines that report to EPA's GHGRP must report quarterly measurements of CH4 emissions
from ventilation systems; they have the option of recording 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.73
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 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 Liberatedfrom Degasification Systems
Particularly gassy underground mines also use degasification systems (e.g., wells or boreholes) to remove CH4
before, during, or after mining. This CH4 can then be collected for use or vented to the atmosphere. Nineteen mines
used degasification systems in 2017, and the CH4 removed through these systems was reported to EPA's GHGRP
under subpart FF (EPA 2018). 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. Twelve of the 19
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.74
Degasification data reported to EPA's GHGRP by underground coal mines is the primary source of data used to
develop estimates of CH4 liberated from degasification systems. Data reported to EPA's GHGRP were used
exclusively to estimate CH4 liberated from degasification systems at 15 of the 19 mines that used degasification
systems in 2017.
71 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).
72 Underground coal mines report to EPA under subpart FF of the GHGRP (40 CFR part 98). In 2017, 78 underground coal
mines reported to the program.
73 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.
74 Several of the mines venting CH4 from degasification systems use a small portion of the gas to fuel gob well blowers in remote
locations where electricity is not available. However, this CH4use is not considered to be a formal recovery and use project.
Energy 3-57
-------
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.75 EPA's GHGRP does not require
gas production from virgin coal seams (coalbed methane) to be reported by coal mines under subpart FF.76 Most
pre-mining wells drilled from the surface are considered coalbed methane wells prior to mine-through and
associated CH4 emissions are reported under another subpart of the GHGRP (subpart W, "Petroleum and Natural
Gas Systems"). As a result, GHGRP data must be supplemented to estimate cumulative degasification volumes that
occurred prior to well mine-through. There were five mines with degasification systems that include pre-mining
wells that were mined through in 2017. For four of these mines, GHGRP data were supplemented with historical
data from state gas well production databases (GSA 2018; WVGES 2018), 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). For
the remaining mine, data from a state gas well production database were used (DMME 2018).
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 2017. Twelve of these projects involved degasification
systems and one involved a ventilation air methane abatement project (VAM). Eleven of these mines sold the
recovered CH4 to a pipeline, including one that also used CH4 to fuel a thermal coal dryer. One mine used recovered
CH4to heat mine ventilation air.
EPA's GHGRP data was exclusively used to estimate the CH4 recovered and used from eight of the 12 mines that
deployed degasification systems in 2017. State sales data were used to estimate CH4 recovered and used from the
remaining four mines that deployed degasification systems in 2017 (DMME 2018; GSA 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. 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 2018).
Of the 13 mines with CH4 recovery in 2017, five intersected pre-mining wells in 2017. EPA's GHGRP and
supplemental data were used to estimate CH4 recovered and used at these mines. 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.
The supplemental data came from state gas production databases as well as mine-specific information on the timing
of mined-through pre-mining wells.
Step 2: Estimate CH4 Emitted from Surface Mines and Post-Mining Activities
Mine-specific data are not available for estimating CH4 emissions from surface coal mines or for post-mining
activities. For surface mines, basin-specific coal production obtained from the Energy Information Administration's
Annual Coal Report (EIA 2018) was multiplied by basin-specific CH4 contents (EPA 1996, 2005) and a 150 percent
emission factor (to account for CH4from over- and under-burden) to estimate CH4 emissions (King 1994; Saghafi
2013). For post-mining activities, basin-specific coal production was multiplied by basin-specific gas contents and a
mid-range 32.5 percent emission factor for CH4 desorption during coal transportation and storage (Creedy 1993).
Basin-specific in situ gas content data were compiled from AAPG (1984) and USBM (1986).
Uncertainty and Time-Series Consistency
A quantitative uncertainty analysis was conducted for the coal mining source category using the IPCC-
recommended Approach 2 uncertainty estimation methodology. Because emission estimates from underground
ventilation systems were based on actual measurement data from EPA's GHGRP or from MSHA, uncertainty is
relatively low. A degree of imprecision was introduced because the ventilation air measurements used were not
75 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.
3-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 CH4 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.
In 2015, 2016, and 2017 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 2017 were estimated to be between 50.5 and 66.2 MMT CO2 Eq. at a 95 percent confidence level. This
indicates a range of 9.3 percent below to 18.8 percent above the 2017 emission estimate of 55.7 MMT CO2 Eq.
Table 3-32: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from Coal
Mining (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Coal Mining
CH4
55.7
50.5 66.2 -9.3% +18.8%
a Range of emission estimates predicted by Monte Carlo stochastic simulation for a 95 percent confidence interval.
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.
Energy 3-59
-------
Recalculations Discussion
For the current Inventory, minor revisions were made to the 2016 annual coal production quantities for underground
and surface mines. These revisions to the 2016 activity data were based on the EIA 2017 Annual Coal Report (EIA
2018). The revisions to the underground coal production quantities resulted in an emission increase of
approximately 0.5 percent for the 2016 emissions from post-mining activities for underground mining. The revisions
to the surface coal production quantities resulted in an insignificant increase (0.001 percent) for the 2016 emissions
from surface mining and post-surface mining activities.
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.
Annual gross abandoned mine CH4 emissions ranged from 7.2 to 10.8 MMT CO2 Eq. from 1990 through 2017,
varying, in general, by less than one 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 CH4 emissions peaked in 1996 (10.8 MMT CO2 Eq.) due to the large
number of gassy mine77 closures from 1994 to 1996 (72 gassy mines closed during the three-year period). In spite of
this rapid rise, abandoned mine CH4 emissions have been generally on the decline since 1996. Since 2002, there
have been fewer than twelve gassy mine closures each year. There were no gassy mine closures in 2017. In 2017,
gross abandoned mine emissions decreased slightly from 9.5 to 9.2 MMT CO2 Eq. (see Table 3-33 and Table 3-34).
Gross emissions are reduced by CH4 recovered and used at 45 mines, resulting in net emissions in 2017 of 6.4 MMT
C02 Eq.
Table 3-33: ChU Emissions from Abandoned Coal Mines (MMT CO2 Eq.)
Activity
1990
2005
2013
2014
2015
2016
2017
Abandoned Underground Mines
7.2
8.4
8.8
8.7
9.0
9.5
9.2
Recovered & Used
+
(1.8)
(2.6)
(2.4)
(2.6)
(2.8)
(2.7)
Total
7.2
6.6
6.2
6.3
6.4
6.7
6.4
+ Does not exceed 0.05 MMT CO2 Eq.
77 A mine is considered a "gassy" mine if it emits more than 100 thousand cubic feet of CH4 per day (100 mcfd).
3-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values.
Table 3-34: ChU Emissions from Abandoned Coal Mines (kt)
Activity 1990
2005
2013 2014 2015 2016 2017
Abandoned Underground Mines 288
Recovered & Used +
334
(70)
353 350 359 380 367
(104) (97) (102) (112) (109)
Total 288
264
249 253 256 268 257
+ Does not exceed 0.5 kt
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values.
Methodology
Estimating CH4 emissions from an abandoned coal mine requires predicting the emissions of a mine from the time
of abandomnent 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 abandomnent reflects the gas content of the coal, the rate of coal mining, and the flow capacity
of the mine in much the same way as the initial rate of a water-free conventional gas well reflects the gas content of
the producing formation and the flow capacity of the well. A well or a mine that produces gas from a coal seam and
the surrounding strata will produce less gas through time as the reservoir of gas is depleted. Depletion of a reservoir
will follow a predictable pattern depending on the interplay of a variety of natural physical conditions imposed on
the reservoir. The depletion of a reservoir is commonly modeled by mathematical equations and mapped as a type
curve. Type curves, which are referred to as decline curves, have been developed for abandoned coal mines.
Existing data on abandoned coal mine emissions through time, although sparse, appear to fit the hyperbolic type of
decline curve used in forecasting production from natural gas wells.
In order to estimate CH4 emissions over time for a given abandoned mine, it is necessary to apply a decline function,
initiated upon abandomnent, to that mine. In the analysis, mines were grouped by coal basin with the assumption
that they will generally have the same initial pressures, permeability and isotherm. As CH4 leaves the system, the
reservoir pressure (Pr) declines as described by the isotherm's characteristics. The emission rate declines because
the mine pressure (Pw) is essentially constant at atmospheric pressure for a vented mine, and the productivity index
(PI), which is expressed as the flow rate per unit of pressure change, is essentially constant at the pressures of
interest (atmospheric to 30 psia). The CH4 flow rate is determined by the laws of gas flow through porous media,
such as Darcy's Law. A rate-time equation can be generated that can be used to predict future emissions. This
decline through time is hyperbolic in nature and can be empirically expressed as:
q = qt (1 + &A0("1/fc)
where,
q = Gas flow rate at time t in million cubic feet per day (mmcfd)
qi = Initial gas flow rate at time zero (tQ), mmcfd
b = The hyperbolic exponent, dimensionless
D = Initial decline rate, 1/year
t = Elapsed time from tQ (years)
This equation is applied to mines of various initial emission rates that have similar initial pressures, permeability and
adsorption isotherms (EPA 2004).
The decline curves created to model the gas emission rate of coal mines must account for factors that decrease the
rate of emissions after mining activities cease, such as sealing and flooding. Based on field measurement data, it was
assumed that most U.S. mines prone to flooding will become completely flooded within eight years and therefore
will no longer have any measurable CH4 emissions. Based on this assumption an average decline rate for flooded
mines was established by fitting a decline curve to emissions from field measurements. An exponential equation was
developed from emissions data measured at eight abandoned mines known to be filling with water located in two of
the five basins. Using a least squares, curve-fitting algorithm, emissions data were matched to the exponential
Energy 3-61
-------
equation shown below. There was not enough data to establish basin-specific equations as was done with the vented,
non-flooding mines (EPA 2004).
q = q^-0^
where,
q = Gas flow rate at time t in mmcfd
qi = Initial gas flow rate at time zero (tQ), mmcfd
D = Decline rate, 1/year
t = Elapsed time from tQ (years)
Seals have an inhibiting effect on the rate of flow of CH4 into the atmosphere compared to the flow rate that would
exist if the mine had an open vent. The total volume emitted will be the same, but emissions will occur over a longer
period of time. The methodology, therefore, treats the emissions prediction from a sealed mine similarly to the
emissions prediction from a vented mine, but uses a lower initial rate depending on the degree of sealing. A
computational fluid dynamics simulator was used with the conceptual abandoned mine model to predict the decline
curve for inhibited flow. The percent sealed is defined as 100 x (1 - [initial emissions from sealed mine / emission
rate at abandonment prior to sealing]). Significant differences are seen between 50 percent, 80 percent and 95
percent closure. These decline curves were therefore used as the high, middle, and low values for emissions from
sealed mines (EPA 2004).
For active coal mines, those mines producing over 100 thousand cubic feet per day (mcfd) of CH4 account for about
98 percent of all CH4 emissions. This same relationship is assumed for abandoned mines. It was determined that the
532 abandoned mines closed after 1972 produced CH4 emissions greater than 100 mcfd when active. Further, the
status of 304 of the 532 mines (or 57 percent) is known to be either: 1) vented to the atmosphere; 2) sealed to some
degree (either earthen or concrete seals); or 3) flooded (enough to inhibit CH4 flow to the atmosphere). The
remaining 43 percent of the mines whose status is unknown were placed in one of these three categories by applying
a probability distribution analysis based on the known status of other mines located in the same coal basin (EPA
2004).
Table 3-35: Number of Gassy Abandoned Mines Present in U.S. Basins in 2017, Grouped by
Class According to Post-Abandonment State
Basin
Sealed
Vented
Flooded
Total
Known
Unknown
Total Mines
Central Appl.
40
26
52
118
148
266
Illinois
34
3
14
51
31
82
Northern Appl.
47
22
16
85
39
124
Warrior Basin
0
0
16
16
0
16
Western Basins
28
4
2
34
10
44
Total
149
55
100
304
228
532
Inputs to the decline equation require the average CH4 emission rate prior to abandonment and the date of
abandonment. Generally, these data are available for mines abandoned after 1971; however, such data are largely
unknown for mines closed before 1972. Information that is readily available, such as coal production by state and
county, is helpful but does not provide enough data to directly employ the methodology used to calculate emissions
from mines abandoned before 1972. It is assumed that pre-1972 mines are governed by the same physical, geologic,
and hydrologic constraints that apply to post-1971 mines; thus, their emissions may be characterized by the same
decline curves.
During the 1970s, 78 percent of CH4 emissions from coal mining came from seventeen counties in seven states.
Mine closure dates were obtained for two of these 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 CH4 emissions rates during the 1970s (EPA 2004).
3-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Abandoned mine emission estimates are based on all closed mines known to have active mine CH4 ventilation
emission rates greater than 100 mcfd at the time of abandonment. For example, for 1990 the analysis included 145
mines closed before 1972 and 258 mines closed between 1972 and 1990. Initial emission rates are based on MSHA
reports, time of abandonment, and basin-specific decline curves (MSHA 2018). Coal mine degasification data are
not available for years prior to 1990, thus the initial emission rates used reflect ventilation emissions only for pre-
1990 closures. CH4 degasification amounts were added to the quantity of CH4 vented to determine the total CH4
liberation rate for all mines that closed between 1992 and 2017. Because the available pre-1972 mine data are
assumed to account for 78 percent of the pre-1972 abandoned mine CH4 emissions, the modeled results were
multiplied by 1.22 to account for all U.S. pre-1972 abandoned mine emissions. The post-1971 mine data are
assumed to represent 98 percent of the post-1971 abandoned mine CH4 emissions, and therefore the modeled results
were multiplied by 1.02 to account for all U.S. post-1971 abandoned mine emissions.
From 1993 through 2017, emission totals were downwardly adjusted to reflect CH4 emissions avoided from those
abandoned mines with CH4 recovery and use or destruction systems. The Inventory totals were not adjusted for
abandoned mine CH4 emission reductions from 1990 through 1992, because no data was reported for abandoned
coal mine CH4 recovery and use or destruction projects during that time.
Uncertainty and Time-Series Consistency
A quantitative uncertainty analysis was conducted to estimate the uncertainty surrounding the estimates of CH4
emissions from abandoned underground coal mines. The uncertainty analysis provides for the specification of
probability density functions for key variables within a computational structure that mirrors the calculation of the
inventory estimate. The results provide the range within which, with 95 percent certainty, emissions from this source
category are likely to fall.
As discussed above, the parameters for which values must be estimated for each mine in order to predict its decline
curve are: 1) the coal's adsorption isotherm; 2) CH4 flow capacity as expressed by permeability; and 3) pressure at
abandonment. Because these parameters are not available for each mine, a methodological approach to estimating
emissions was used that generates a probability distribution of potential outcomes based on the most likely value and
the probable range of values for each parameter. The range of values is not meant to capture the extreme values, but
rather values that represent the highest and lowest quartile of the cumulative probability density function of each
parameter. Once the low, mid, and high values are selected, they are applied to a probability density function.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 3-36. Annual abandoned
coal mine CH4 emissions in 2017 were estimated to be between 5.1 and 7.7 MMT CO2 Eq. at a 95 percent
confidence level. This indicates a range of 21 percent below to 19 percent above the 2017 emission estimate of 6.4
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 CH4 liberation rates exist.
Table 3-36: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from
Abandoned Underground Coal Mines (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Abandoned Underground
Coal Mines
CH4
6.4
5.1 7.7
-21% +19%
a Range of emission estimates predicted by Monte Carlo Simulation for a 95 percent confidence interval.
tion
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 focused on the emission factor and activity data sources and methodology used for estimating
Energy 3-63
-------
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 applied to the abandoned underground coal mine emission estimates for 1990 through 2017.
3.6 Petroleum Systems (CRF Source Category
lB2a)
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 leak emissions,
vented emissions (including emissions from operational upsets) and emissions from flaring. Carbon dioxide
emissions from petroleum systems are primarily associated with crude oil production and refining operations. Total
CH4 emissions from petroleum systems in 2017 were 37.7 MMT CO2 Eq. (1,506 kt), a decrease of 10 percent from
1990. Total CO2 emissions from petroleum systems in 2017 were 23.3 MMT CO2 Eq. (23,336 kt), an increase of
161 percent from 1990. Total N2O emissions from petroleum systems in 2017 were 0.02 MMT CO2 Eq. (0.08 kt), an
increase of 77 percent from 1990.
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 2016) to
ensure that the trend is accurate. Recalculations in petroleum systems in this year's Inventory include:
• Revised hydraulically fractured (HF) oil well completions and workovers methodology
• Newly calculated N20 emissions from flaring
• Newly calculated CO2 emissions from crude oil transportation
• 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 from petroleum systems. The predominant sources of emissions from exploration are
hydraulically fractured oil well completions and well testing. Other sources include well completions without
hydraulic fracturing, and well drilling. Since 1990, exploration CH4 emissions have decreased 88 percent, and while
the number of hydraulically fractured wells completed increased by a factor of nearly 3, there were decreases in the
fraction of such completions without reduced emissions completions (RECs) or flaring (from 90 percent in 1990 to 2
percent in 2017). Emissions of CH4 from exploration were highest in 2012, over 20 times higher than in 2017, and
lowest in 2017. Emissions of CH4 from exploration decreased 24 percent from 2016 to 2017. Exploration accounts
for 7 percent of total CO2 emissions from petroleum systems in 2017. Emissions of CO2 from exploration in 2017
increased by a factor of 4.2 from 1990, and 38 percent from 2016, due to an increase in hydraulically fractured oil
well completions with flaring (from 10 percent of completions in 1990 to 58 percent in 2017). Emissions of CO2
from exploration were highest in 2014, around 1.8 times as high as in 2017. Exploration accounts for 3 percent of
total N20 emissions from petroleum systems in 2017. Emissions of N20 from exploration in 2017 increased by a
factor of 3.4 from 1990, and 22 percent from 2016, due to an increase in hydraulically fractured oil well completions
with flaring (from 10 percent of completions in 1990 to 58 percent in 2017).
Production. Production accounts for approximately 97 percent of total CH4 emissions from petroleum systems. 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, CH4 emissions from production have decreased by
5 percent, due to decreases in emissions from tanks, hydraulically fractured oil well workovers, and offshore
platforms. Overall, production segment methane emissions decreased by 1 percent from 2016 levels. Production
emissions account for 77 percent of the total CO2 emissions from petroleum systems in 2017. The principal sources
of CO2 emissions are associated gas flaring, oil tanks with flares, and miscellaneous production flaring. These three
3-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
sources together account for 98 percent of the CO2 emissions from production. Since 1990, CO2 emissions from
production have increased by 236 percent, due to increases in flaring emissions from associated gas flaring, tanks,
and miscellaneous production flaring. Overall, production segment CO2 emissions increased by 6 percent from 2016
levels due to an increase in associated gas flaring and miscellaneous production flaring. Production emissions
account for 52 percent of the total N20 emissions from petroleum systems. The principal sources of N20 emissions
are associated gas flaring, oil tanks with flares, and miscellaneous production flaring. Since 1990, N20 emissions
from production have increased by 186 percent.
Crude Oil Transportation. Crude oil transportation activities account for less than 1 percent of total CH4 emissions
from petroleum systems. Emissions from tanks, marine loading, and truck loading operations account for 73 percent
of CH4 emissions from crude oil transportation. Since 1990, CH4 emissions from transportation have increased by 17
percent. Methane emissions from transportation in 2017 decreased 5 percent from 2016 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 73 percent of CO2 emissions from
crude oil transportation. Emissions from crude oil transportation account for a very small percentage of the total
emissions from petroleum systems and have little impact on the overall emissions.
Crude Oil Refining. Crude oil refining processes and systems account for 2 percent of total CH4 emissions from
petroleum systems. This low share is because most of the CH4 in crude oil is removed or escapes before the crude
oil is delivered to the refineries. There is an insignificant amount of CH4 in all refined products. Within refineries,
flaring accounts for 41 percent of the CH4 emissions, while uncontrolled blowdowns and process vents account for
19 and 16 percent, respectively. Methane emissions from refining of crude oil have increased by 16 percent since
1990, and increased less than 1 percent since 2016; however, similar to the transportation subcategory, this increase
has had little effect on the overall emissions of CH4. Crude oil refining processes and systems account for 16 percent
of total CO2 emissions from petroleum systems. Almost all (about 97 percent) of the CO2 from refining is from
flaring. Refinery CO2 emissions increased by 14 percent from 1990 to 2017, and decreased by 7 percent from 2016
levels. Flaring occurring at crude oil refining processes and systems accounts for 45 percent of total N20 emissions
from the oil industry. Refinery N20 emissions increased by 19 percent from 1990 to 2017, and decreased by 6
percent from 2016 levels.
Table 3-37: ChU Emissions from Petroleum Systems (MMT CO2 Eq.)
Activity
1990
2005
2013
2014
2015
2016
2017
Exploration3
3.0
4.5
6.3
5.0
2.1
0.5
0.4
Production (Total)
38.3
31.4
34.4
36.2
36.5
36.8
36.4
Pneumatic Controllers
19.3
17.5
18.6
19.4
19.6
20.5
20.9
Offshore Platforms
5.3
4.6
4.7
4.7
4.7
4.7
4.7
Equipment Leaksb
2.2
2.2
2.6
2.7
2.7
2.6
2.5
Gas Engines
2.1
1.7
2.2
2.3
2.3
2.2
2.2
Chemical Injection Pumps
1.2
1.7
2.1
2.2
2.2
2.1
2.0
Tanks
5.4
1.5
1.3
1.6
1.7
2.6
1.5
Other Sources
2.6
2.1
2.9
3.3
3.3
2.2
2.5
Crude Oil Transportation
0.2
0.1
0.2
0.2
0.2
0.2
0.2
Refining
0.6
0.7
0.7
0.7
0.7
0.7
0.7
Total
42.1
36.7
41.6
42.1
39.5
38.2
37.7
a Exploration includes well drilling, testing, and completions.
b Includes leak emissions from wellheads, separators, heaters/treaters, and headers.
Note: Totals may not sum due to independent rounding.
Energy 3-65
-------
Table 3-38: ChU Emissions from Petroleum Systems (kt ChU)
Activity
1990
2005
2013
2014
2015
2016
2017
Exploration3
121
181
254
201
84
19
14
Production (Total)
1,531
1,255
1,377
1,446
1,458
1,473
1,456
Pneumatic Controllers
774
701
743
111
786
818
837
Offshore Platforms
211
185
188
188
188
188
188
Equipment Leaks
88
86
104
109
107
104
102
Gas Engines
86
70
88
93
93
89
89
Chemical Injection Pumps
49
68
84
87
86
83
82
Tanks
218
60
53
63
68
102
61
Other Sources
105
86
118
131
131
89
98
Crude Oil Transportation
7
5
7
8
8
8
8
Refining
24
28
27
26
28
28
28
Total
1,682
1,469
1,665
1,682
1,579
1,528
1,506
a Exploration includes well drilling, testing, and completions.
Note: Totals may not sum due to independent rounding.
Table 3-39: CO2 Emissions from Petroleum Systems (MMT CO2)
Activity
1990
2005
2013
2014
2015
2016
2017
Exploration
0.3
0.3
2.5
3.0
2.2
1.2
1.7
Production
5.3
7.5
19.1
23.2
25.4
17.0
18.0
Transportation
+
+
+
+
+
+
+
Crude Refining
3.3
3.7
3.6
3.4
4.1
4.0
3.7
Total
9.0
11.6
25.1
29.6
31.7
22.2
23.3
+ Does not exceed 0.05 MMT CO2.
Note: Totals may not sum due to independent rounding.
Table 3-40: CO2 Emissions from Petroleum Systems (kt CO2)
Activity
1990
2005
2013
2014
2015
2016
2017
Exploration
321
331
2,461
2,976
2,167
1,200
1,657
Production
5,344
7,493
19,059
23,201
25,438
17,008
17,951
Transportation
0.9
0.7
1.0
1.2
1.2
1.1
1.1
Crude Refining
3,284
3,728
3,609
3,419
4,067
3,991
3,728
Total
8,950
11,552
25,130
29,597
31,672
22,200
23,336
Note: Totals may not sum due to independent rounding.
Table 3-41: N2O Emissions from Petroleum Systems (metric tons CO2 Eq.)
Activity
1990
2005
2013
2014
2015
2016
2017
Exploration
172
176
1,278
1,543
1,125
618
754
Production
4,414
5,332
12,980
15,817
17,429
12,749
12,640
Transportation
NE
NE
NE
NE
NE
NE
NE
Crude Refining
9,143
10,377
10,187
9,659
11,656
11,575
10,836
Total
13,728
15,885
24,445
27,020
30,210
24,942
24,231
NE (Not Estimated)
Note: Totals may not sum due to independent rounding.
Table 3-42: N2O Emissions from Petroleum Systems (metric tons N2O)
Activity
1990
2005
2013
2014
2015
2016
2017
Exploration
0.6
0.6
4.3
5.2
3.8
2.1
2.5
Production
14.8
17.9
43.6
53.1
58.5
42.8
42.4
Transportation
NE
NE
NE
NE
NE
NE
NE
Crude Refining
30.7
34.8
34.2
32.4
39.1
38.8
36.4
Total
46.1
53.3
82.0
90.7
101.4
83.7
81.3
3-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
NE (Not Estimated)
Note: Totals may not sum due to independent rounding.
Methodology
See Annex 3.5 for the Ml 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 Uncertainty and Time-Series Consistency, Recalculations
Discussion, and Planned Improvements.
Emission Factors. References for emission factors include Methane Emissions from the Natural Gas Industry by the
Gas Research Institute and EPA (EPA/GRI 1996), Estimates of Methane Emissions from the U.S. Oil Industry (EPA
1999), Drillinglnfo (2018), 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 2018).
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 emission factors calculated from
year 2015 GHGRP data. For miscellaneous production flaring, year-specific emission factors were developed for
years 2015 forward from GHGRP data, an emission factor of 0 was assumed for 1990 through 1992, and linear
interpolation was applied to develop emission factors for 1993 through 2014. For offshore oil production, two
emission factors were calculated using data collected for all federal offshore platforms by BOEM; one for oil
platforms in shallow water, and one for oil platforms in deep water. For most sources, emission factors were held
constant for the period 1990 through 2016, 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
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 factors by the corresponding CH4 data, as the CO2 content of gas relates to the
CH4 content of gas. Sources with CO2 emissions calculated from GHGRP data were HF completions and workovers,
associated gas venting and flaring, tanks, well testing, pneumatic controllers, chemical injection pumps, and
miscellaneous production flaring. For these sources, CO2 was calculated using the same methods as used for CH4.
Emission factors for offshore oil production (shallow and deep water) were derived using data from BOEM. 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.
Energy 3-67
-------
For the exploration and production segments, N20 emissions were estimated for flaring sources using GHGRP data.
Sources with N20 emissions in the exploration segment were well testing and HF completions with flaring. Sources
with N20 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 C02 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 N20 emissions for all major activities
starting with emissions that occurred in 2010. However, GHGRP does have provisions that refineries are not
required to report to the GHGRP if their emissions fall below certain thresholds (see Planned Improvements for
additional discussion). The reported total of CH4, CO2, and N20 emissions for each activity was used for the
emissions in each year from 2010 forward. To estimate emissions for 1990 to 2009, the 2010 to 2013 emissions data
from GHGRP along with the refinery feed data for 2010 to 2013 were used to derive emission factors (i.e., sum of
activity emissions/sum of refinery feed), which were then applied to the annual refinery feed in years 1990 to 2009.
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 (Drillinglnfo 2018), Energy Information
Administration (EIA) reports, Methane Emissions from the Natural Gas Industry by the Gas Research Institute and
EPA (EPA/GRI 1996), Estimates of Methane Emissions from the U.S. Oil Industry (EPA 1999), consensus of
industry peer review panels, BOEM reports, the Oil & Gas Journal, the Interstate Oil and Gas Compact
Commission, the United States Army Corps of Engineers, and analysis of GHGRP data (EPA 2018).
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, the previous year's data were used for domestic barges
and tankers as current year were not yet available. For offshore production, the number of platforms in shallow
water and the number of platforms in deep water are used as activity data and are taken from BOEM datasets; these
activity data have not been recently updated and 2010 activity are applied for all recent years.
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 has conducted a quantitative uncertainty analysis using the IPCC Approach 2 methodology (Monte Carlo
Simulation technique) to characterize uncertainty for petroleum systems. For more information, 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 CH4 and
CO2 emissions estimates. For the analysis, EPA focused on the four highest methane-emitting sources for the year
2017, which together emitted 79 percent of methane from petroleum systems in 2017, 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
78 See
3-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 2017, using the recommended IPCC methodology. The results of the
Approach 2 uncertainty analysis are summarized in Table 3-43. Petroleum systems CH4 emissions in 2017 were
estimated to be between 25.0 and 51.9 MMT CO2 Eq., while CO2 emissions were estimated to be between 15.5 and
32.2 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
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)b
(MMT CO2 Eq.)
(%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Petroleum Systems
CH4
37.7
25.0 51.9
-34% +38%
Petroleum Systems0
CO2
23.3
15.5 32.2
-34% +38%
a Range of emission estimates estimated by applying the 95 percent confidence intervals obtained from the Monte Carlo
Simulation analysis conducted for the year 2017 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 CO2 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 CO2
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 and 2010 or 2015 data points. Information on
time-series consistency for sources updated in this year's Inventory can be found in the Recalculations Discussion
below, with additional detail provided in supporting memos (relevant memos are cited in the Recalculations
Discussion). For information on other sources, please see the Methodology Discussion above and Annex 3.5.
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 workshop on greenhouse gas data for oil and gas in October of 2018, and
webinars in June of 2018 and February of 2019. EPA released memos detailing updates under consideration and
79 See
Energy 3-69
-------
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 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
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
June, October, and November 2018, EPA released draft memoranda that discussed changes under consideration and
requested stakeholder feedback on those changes. The EPA then created updated versions of the memoranda to
document the methodology implemented into the current Inventory.82 The EPA me mo ra ndum Inventory of U.S.
Greenhouse Gas Emissions and Sinks 1990-2017: Other Updates Considered for 2019 and Future GHGls (Apr.
2019 Other Updates memo) is cited in the Recalculations Discussion below.
EPA thoroughly evaluated relevant information available and made updates to exploration and production segment
methodologies for the Inventory, specifically: using GHGRP data to calculate emissions and activity factors for oil
well completions and workovers with hydraulic fracturing; using Drillinglnfo data (Drillinglnfo 2018) to calculate
well drilling activity; and revising the basis for calculating the number of active wells represented in GHGRP
reporting. 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 2016 to the current (recalculated)
estimate for 2016 (the emissions changes were mostly due to GHGRP data submission revisions); these sources are
discussed below and include production tanks, associated gas venting and flaring, miscellaneous production flaring,
pneumatic controllers, chemical injection pumps, heaters, and refineries.
Finally, emissions estimates were included for N20 from flaring activities in the exploration, production, and
refineries segments, and for CO2 from the crude oil transportation segment.
The combined impact of revisions to 2016 petroleum systems CH4 emissions, compared to the previous Inventory, is
a decrease from 38.6 to 38.2 MMT CO2 Eq. (0.4 MMT CO2 Eq., or 1 percent). The recalculations resulted in an
average increase in CH4 emission estimates across the 1990 through 2016 time series, compared to the previous
Inventory, of 3.3 MMT CO2 Eq, or 10 percent, with the largest increases in the estimates for 2005 to 2013 due to
the revised data on hydraulically fractured oil well completions.
80 See
81 See
82 Stakeholder materials including draft and final EPA memoranda for the current (i.e., 1990 to 2017) Inventory are available at
.
3-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
The combined impact of revisions to 2016 petroleum systems CO2 emissions, compared to the previous Inventory, is
a decrease from 22.8 to 22.2 MMT CO2 (0.6 MMT CO2, or 2 percent). The recalculations resulted in an average
increase in emission estimates across the 1990 through 2016 time series, compared to the previous Inventory, of 0.6
MMT CO2 Eq, or 4 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 2016 to
the current (recalculated) estimate for 2016. 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 2016,
Year 2016,
Year 2017,
2018 Inventory
2019 Inventory
2019 Inventory
Exploration
+
1.2
1.7
HF Oil Well Completions
+
1.2
1.6
Production
19.0
17.0
18.0
Tanks
7.4
5.9
4.4
Associated Gas Venting & Flaring
9.1
8.6
10.5
Miscellaneous Flaring
2.5
2.2
2.6
HF Oil Well Workovers
+
0.2
0.3
Transportation
NE
+
+
Refining
3.7
4.0
3.7
Petroleum Systems Total
22.8
22.2
23.3
NE (Not Estimated)
+ Does not exceed 0.05 MMT CO2.
Table 3-45: Recalculations of ChU in Petroleum Systems (MMT CO2 Eq.)
Previous Estimate
Current Estimate
Current Estimate
Year 2016,
Year 2016,
Year 2017,
2018 Inventory
2019 Inventory
2019 Inventory
Exploration
2.1
0.5
0.4
HF Oil Well Completions
2.0
0.4
0.3
Production
35.4
36.8
36.4
Pneumatic Controllers
18.5
20.5
20.9
Tanks
3.2
2.6
1.5
Heaters
0.8
0.7
0.7
Chemical Injection Pumps
2.0
2.1
2.0
HF Oil Well Workovers
+
0.1
0.1
Transportation
0.2
0.2
0.2
Refining
0.9
0.7
0.7
Petroleum Systems Total
38.6
38.2
37.7
+ Does not exceed 0.05 MMT CO2.
Exploration
HF Oil Well Completions (Methodological Update)
EPA revised the HF oil well completions methodology by establishing four control categories (non-REC with
venting, non-REC with flaring, REC with venting, and REC with flaring) and developing new activity and emission
factors for these categories. The new methodology is detailed in the Apr. 2019 Other Updates memo. The previous
factors (for controlled and uncontrolled event categories) relied on data analysis from the 2015 NSPS OOOOa
rulemaking proposal. As described above in the Methodology discussion, EPA has newly calculated year-specific
activity factors (fraction of events in each category) and emission factors for years 2016 forward using GHGRP
data. To estimate emissions over the time series, EPA applied the year 2016 emission factors for all prior years and
developed activity factors by following the existing methodology for HF gas well events combined with oil well-
Energy 3-71
-------
specific assumptions regarding when controls became prevalent. For HF oil well event activity factors, the following
assumptions are applied: (1) foryears 1990 to 2007, all completions and workovers are non-REC, and 10 percent of
events flare; (2) for the first year in which GHGRP data are available, 2016, control fractions across the four
categories are developed directly from reported GHGRP data; and (3) for intermediate years, 2008 to 2015, control
fractions are developed through linear interpolation. This approach produces activity factors across the time series
that are generally consistent with the previous assumption that oil well RECs are introduced beginning in year 2008,
during which 7 percent of completions and workovers are REC, and 10 percent of both REC and non-REC events
flare. EPA did not change the methodology of calculating total activity for this source, which relies on analyzing
Drillinglnfo data (Drillinglnfo 2018) to obtain the total HF oil well completion event count in each year of the time
series. Stakeholder feedback supported an approach of using GHGRP data to update activity and emissions factors
on an annual basis from 2016 forward.
Table 3-46: HF Oil Well Completions National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
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
110,326
171,881
227,632
171,542
60,488
7,043
2,168
360
560
2,502
2,788
1,804
1,018
1,791
0
0
4,800
6,081
4,383
2,714
2,223
0
0
7,707
9,764
7,037
4,358
6,424
110,685
20,796
172,441
31,070
242,642
109,422
190,175
120,925
73,712
78,525
15,132
78,525
12,606
NA
NA (Not Applicable)
Table 3-47: HF Oil Well Completions National CO2 Emissions (kt CO2)
2.5 5.3 4.0 1.4 0.2 0.2
79 547 610 395 223 410
Source 1990 2005 2013 2014 2015 2016 2017
HF Completions: Non-REC with
Venting
HF Completions: Non-REC with
Flaring
HF Completions: REC with
Venting
HF Completions: REC with
Flaring
Total Emissions 81.2 2,214 2,719 1,913 1,162 1,619
Previous Estimate 1.2 6.1 6.7 4.4 4.4 NA
NA (Not Applicable)
0.0 0.0 0.3 0.3 0.3 0.2 0.1
0.0 0.0 1,661 2,104 1,517 939 1,209
Well Drilling (Methodological Update)
EPA updated the methodology for estimating the number of oil wells drilled across the time series to use
Drillinglnfo data (Drillinglnfo 2018). The new methodology is detailed in the Apr. 2019 Other Updates memo. In
previous Inventories, the U.S. Department of Energy's Energy Information Administration (DOE/EIA) Monthly
Energy Review well drilling activity data set was used to develop well drilling activity inputs, but this publication
does not provide data after year 2010. EPA therefore developed a methodology of analyzing Drillinglnfo data to
estimate counts of oil wells drilled in each time series year, 1990 through 2017. These activity data for select years
are shown in Table 3-48 below.
Table 3-48: Count of Oil Wells Drilled
Source
1990
2005
2013
2014
2015
2016
2017
Oil Wells Drilled
Previous Estimate
19,919
17,234
18,216
12,053
35,671
17,774"
36,910
17,774"
17,359
17,774"
10,242
17,774"
10,242
NA
a - Year-specific data not available; the year 2010 estimate was assigned as a surrogate value.
3-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
NA (Not Applicable)
Production
HF Oil Well Workovers (Methodological Update)
EPA revised the HF oil well workovers methodology to use the same general approach as described above for HF
oil well completions. EPA revised the oil well workovers methodology by separating HF and non-HF events, then
establishing four control categories for HF events (non-REC with venting, non-REC with flaring, REC with venting,
and REC with flaring) and developing new activity and emission factors for these categories. The new methodology
is detailed in the Apr. 2019 Other Updates memo. The previous methodology did not use separate emissions or
activity assumptions for HF versus non-HF workover events. As described above in the Methodology discussion,
EPA lias newly calculated year-specific activity factors (fraction of events in each category) and emission factors for
years 2016 forward using GHGRP data. To estimate emissions over the time series, EPA applied the year 2016
emission factors for all prior years and developed activity factors by following the existing methodology for HF gas
well events combined with oil well-specific assumptions regarding when controls became prevalent. For HF oil well
event activity factors, the following assumptions are applied: (1) for years 1990 to 2007, all completions and
workovers are non-REC, and 10 percent of events flare; (2) for the first year in which GHGRP data are available,
2016, control fractions across the four categories are developed directly from reported GHGRP data; and (3) for
intermediate years, 2008-2015, control fractions are developed through linear interpolation. This approach produces
activity factors across the time series that are generally consistent with the previous assumption that oil well RECs
are introduced beginning in year 2008, during which 7 percent of completions and workovers are REC, and 10
percent of both REC and non-REC events flare. EPA also updated the methodology of calculating total activity for
this source; EPA applies the existing assumption used for HF gas wells, that 1 percent of HF wells are worked over
in a given year. Stakeholder feedback supported an approach of using GHGRP data to update activity and emissions
factors on an annual basis from 2016 forward.
Table 3-49: HF Oil Well Workovers National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
HF Workovers: Non-REC with
Venting
31,119
35,018
22,290
17,601
10,808
3,318
0
F1F Workovers: Non-REC with
Flaring
101
114
142
148
142
130
114
F1F Workovers: REC with
Venting
0
0
745
966
1,146
1,275
678
F1F Workovers: REC with
Flaring
0
0
485
629
746
830
1,229
Total Emissions
31,220
35,132
23,662
19,344
12,842
5,552
2,022
Previous Estimate"
77
65
79
82
82
78
NA
NA (Not Applicable)
a Estimate includes emissions for HF and non-HF workovers.
Table 3-50: HF Oil Well Workovers National CO2 Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
F1F Workovers: Non-REC with
Venting
0.7
0.8
0.5
0.4
0.2
0.1
0.0
F1F Workovers: Non-REC with
Flaring
22.2
25.0
31.1
32.3
31.1
28.4
26.2
F1F Workovers: REC with
Venting
0.0
0.0
0.0
0.1
0.1
0.1
0.0
F1F Workovers: REC with
Flaring
0.0
0.0
104.5
135.5
160.7
178.8
231.3
Total Emissions
22.9
25.8
136.1
168.3
192.1
207.4
257.5
Previous Estimate"
0.0
0.0
0.0
0.0
0.0
0.0
NA
NA (Not Applicable)
Energy 3-73
-------
a Estimate includes emissions for HF and non-HF workovers.
Tanks (Recalculation with Updated Data)
Production tank CH4 and CO2 emission estimates decreased in the current Inventory, compared to the previous
Inventory. This change was due to GHGRP submission revisions and updated production data (see the Oil
Production discussion below). For CO2 emissions, in general, a smaller fraction of the GHGRP tank throughput
went through tanks with flares and certain GHGRP-based emission factors were lower. For CH4, while a larger
fraction of the GHGRP tank throughput went through tanks without controls, the calculated GHGRP-based emission
factors were lower.
Table 3-51: Production Storage Tank National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Large Tanks w/ Flares
0
2,510
5,649
6,704
7,230
5,105
5,687
Large Tanks w/ VRU
0
1,133
2,550
3,026
3,263
19,180
8,963
Large Tanks w/o Control
209,643
52,011
38,001
45,093
48,631
66,448
40,056
Small Tanks w/ Flares
0
15
34
41
44
22
44
Small Tanks w/o Flares
4,246
2,041
2,992
3,551
3,830
3,358
2,248
Malfunctioning Dump Valves
3,998
2,345
3,770
4,473
4,824
8,079
4,339
Total Emissions
217,887
60,055
52,997
62,887
67,821
102,191
61,336
Previous Estimate
257,923
84,409
65,467
76,752
82,496
127,025
NA
NA (Not Applicable)
ible 3-52: Production Storage Tank National CO2 Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
Large Tanks w/ Flares
0
2,619
5,896
6,997
7,546
5,843
4,380
Large Tanks w/ VRU
0
5
11
13
14
4.6
4
Large Tanks w/o Control
23
6
4
4.9
5
7
5
Small Tanks w/ Flares
0
2
5
6
7
17
15
Small Tanks w/o Flares
6
3
4
5
5
5
3
Malfunctioning Dump Valves
17
10
16
19
20
18
15
Total Emissions
46
2,645
5,937
7,045
7,598
5,894
4,422
Previous Estimate
53
3,444
6,922
8,115
8,722
7,351
NA
NA (Not Applicable)
Pneumatic Controllers (Recalculation with Updated Data)
Pneumatic controller CH4 emission estimates increased in the current Inventory, compared to the previous
Inventory, due to GHGRP submission revisions and the use of GHGRP well counts from the facility overview table
(see the Well Counts discussion below). The well count change shifted certain controllers from being assigned to
natural gas systems to petroleum systems. Pneumatic controller CH4 emission estimates increased by an average of 5
percent across the 1990 to 2016 time series.
Table 3-53: Pneumatic Controller National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Pneumatic Controllers: High
Bleed
724,225
418,481
100,587
87,778
77,849
82,071
52,265
Pneumatic Controllers: Low
Bleed
49,429
43,906
29,291
28,589
25,341
17,415
19,162
Pneumatic Controllers: Int Bleed
Total Emissions
Previous Estimate
0
773,655
765,975
238,603
700,990
663,461
613,112
742,990
687,210
660,145
776,512
715,768
682,514
785,704
720,996
718,683
818,169
739,125
765,378
836,804
NA
NA (Not Applicable)
3-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Associated Gas Venting and Flaring (Recalculation with Updated Data)
Associated gas venting and flaring CO2 emission estimates decreased for 2016 and increased for 1990 through 2015
in the current Inventory, compared to the previous Inventory. Compared to the previous inventory, on average,
calculated CO2 emissions increased across the 1990 to 2015 time series by 20 percent, and decreased by 6 percent
for 2016. This change was due to GHGRP submission revisions and updated production data (see the Oil Production
discussion below). The emission calculations are performed at a basin-level, and the changes impacted each basin
uniquely. However, the changes in CO2 emissions were mainly driven by the Permian Basin data.
Table 3-54: Associated Gas Venting and Flaring National CO2 Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
Associated Gas Venting
21
11
11
12
13
6
19
Associated Gas Flaring
5,172
3,925
10,384
12,711
13,955
8,587
10,506
Total Emissions
5,193
3,937
10,395
12,723
13,968
8,593
10,525
Previous Estimate
4,028
3,314
9,193
11,248
12,234
9,108
NA
NA (Not Applicable)
Miscellaneous Production Flaring (Recalculation with Updated Data)
Miscellaneous production flaring CO2 emission estimates decreased for most years of the current Inventory, except
for an increase for 2015, compared to the previous Inventory. There were several underlying factors that impacted
the changes each year; GHGRP submission revisions, use of GHGRP well counts from the facility overview table
(see the Well Counts discussion below), a correction to the linear interpolation calculation for emission factors in
years 1993 through 2014, and updated production data (see the Oil Production discussion below). In addition the
emission calculations are performed at a basin-level, and the changes impacted each basin uniquely.
Table 3-55: Miscellaneous Production Flaring National CO2 Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
Misc. Flaring
0 1
800 1
2,487
3,157
3,571
2,201
2,631
Previous Estimate
0
929
2,541
3,181
3,418
2,455
NA
NA (Not Applicable)
Chemical Injection Pumps (Recalculation with Updated Data)
Chemical injection pump CH4 emission estimates increased by an average of 1.4 percent over the time series and for
certain recent years increased by approximately 3 percent for the current Inventory, compared to the previous
Inventory. The emission estimates increases are due to updated well counts (see the Well Counts discussion below);
emission factors and activity factors were not updated.
Table 3-56: Chemical Injection Pump National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Chemical Injection Pump
49,465
67,785
83,972
87,212
86,114
83,215
81,660
Previous Estimate
49,131
66,585
82,084
84,934
85,016
80,974
NA
NA (Not Applicable)
Heaters (Recalculation with Updated Data)
Combustion CH4 emission estimates from heaters decreased by an average of approximately 22 percent for each
year of the time series in the current Inventory, compared to the previous Inventory. The decrease is due to a
decrease in total oil production in each year, the applicable activity data for heaters, which was updated for the
current Inventory (see the Oil Production discussion below).
Energy 3-75
-------
Table 3-57: Heater National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Heater
23,935 1
14,038
22,570
26,782
28,883
26,504
28,051
Previous Estimate
26,944
18,991
27,350
32,065
34,465
32,446
NA
NA (Not Applicable)
Well Counts (Recalculation with Updated Data)
For total national well counts, EPA lias used a more recent version of the Drillinglnfo data set (Drillinglnfo 2018) to
update well counts data in the Inventory. EPA also updated the Drillinglnfo data processing methodology to more
accurately count wells in states with lease-level reporting (e.g., Kansas), which resulted in slight increased counts
across the time series. While this was not a significant recalculation (increases are 2 to 3 percent across the time
series), this is a key input to the Inventory, so results are highlighted here.
Table 3-58: Producing Oil Well Count Data
Oil Well Count
1990
2005
2013
2014
2015
2016
2017
Number of Oil Wells
Previous Estimate
564,090
553,899
480,482
469,632
582,769
569,670
605,259
589,450
597,635
590,017
577,515
561,964
566,726
NA
NA (Not Applicable)
In October 2018, EIA released an updated time series of national oil and gas well counts (covering 2000 through
2017). EIA estimates 991,000 total producing wells for year 2017. EPA's total well count for this year is 978,176.
EPA's well counts in recent time series years are generally 2 percent lower than EIA's. EIA'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.
For the count of wells included in GHGRP reporting (used to develop wellhead-based emissions and activity
factors), EPA previously referenced the wellhead counts contained within the reporting table for onshore production
equipment leak emissions. Due to updated reporting requirements for year 2017 forward, well counts provided as
part of the facility overview information (i.e., wells producing at the end of the calendar year plus wells removed
from production in a given year) provide more complete estimates. Therefore, EPA used well counts from the
facility overview table for source-specific methodologies that rely on GHGRP reported well counts in the current
Inventory. Comparing the GHGRP well counts from the facility overview table to the equipment leaks table: a larger
population of the wells were reported as "oil" production type in the facility overview information table, compared
to the equipment leaks table, which generally led to increased activity and emissions for petroleum systems. For
example, as discussed in the sections above, production segment emissions from pneumatic controllers and
miscellaneous production flaring increased.
Oil Production
EPA reviewed the national oil production data that were previously used in the Inventory and determined a more
appropriate dataset were available. In previous Inventories, production from the EIA's Monthly Energy Review were
used; specifically. Table 3.1 Petroleum Overview, "Total Crude Oil Field Production". However, this dataset
includes both onshore and offshore production and did not distinguish between the two. EIA provides more detailed
production data in an online database, including specifically reporting federal offshore production.83 The EIA online
database production data were used for the current Inventory and federal offshore production data were excluded.
This meant the production values decreased across the time series, but are more specific to onshore production. The
emission sources that rely on oil production as an activity driver and that were impacted the most by this change are
production tanks, associated gas venting and flaring, miscellaneous production flaring, and heaters (all of which are
83 Available at
3-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
discussed above). In addition, oil production data are activity drivers for estimating fugitive emissions from
production compressors and the sales area (loadings), and emissions due to pressure relief valve releases.
Table 3-59: Oil Production Data (Million Barrels)
Source
1990
2005
2013
2014
2015
2016
2017
Oil Production
2,385 1
1,399 1
2,249
2,668
2,878
2,641
2,795
Previous Estimate
2,685
1,892
2,725
3,195
3,434
3,233
NA
NA (Not Applicable)
Floating Roof Tanks
EPA removed the line item estimate for production segment floating roof tanks that was included in previous
Inventories. The number of floating roof tanks and their emissions were minimal in the context of the petroleum
production segment, and available data are limited; data on the number of floating roof tanks are only available for
1995, and the 1995 count is applied to all other years. EPA sought stakeholder input on whether and how to include
floating roof tank emission estimates in the production segment and did not receive objections to the removal of this
source. The emission estimate for this source in the previous Inventory was 159 metric tons CH4 in each year, or
0.01 percent of CH4 emissions in year 2016.
Crude Oil Transportation
EPA newly calculated CO2 emissions from crude oil transportation in the current Inventory. Prior Inventories did
not calculate CO2 emissions from crude oil transportation. CO2 emission factors were calculated by multiplying the
CH4 emission factors for each source by a conversion factor, which is the ratio of CO2 content and CH4 content in
whole crude post-separator. Total CO2 emissions from crude oil transportation are included in Table 3-60 below,
and emissions for each source can be found in Annex 3.5.
Table 3-60: Crude Oil Transportation National CO2 Emissions (kt CO2)
Source
1990
2005
2013 2014 2015 2016
2017
Crude Oil Transportation
0.9
0.7
1.0 1.2 1.2 1.1
1.1
Recalculations due to updated activity data for the quantity of petroleum transported by barge or tanker in the crude
oil transportation segment did not result in a change in CH4 emission estimates for 1990 to 2015. Updated activity
data for 2016 resulted in a decrease in calculated CH4 emissions of approximately 3 percent.
Refining
There are minimal changes in calculated CH4 and CO2 emissions for 1990 to 2015 for this segment (e.g., average
change is less than 0.1 percent each year). However, recalculations for 2016 resulted in CO2 emission estimates
increasing by 8 percent and CH4 emissions decreasing by 24 percent. The 2016 emissions changes are due to
GHGRP submission revisions.
N2O Emissions
EPA newly calculated N2O emissions in the current Inventory, as discussed in the Apr. 2019 Other Updates memo.
Prior Inventories did not calculate N20 emissions from petroleum systems. For each flaring emission source
calculation methodology which uses GHGRP data, the existing source-specific methodology was applied to
calculate N20 emission factors. This update was applied for flaring sources in the exploration, production, and
refining segments.
Table 3-61: N2O National Emissions (Metric Tons N2O)
Source 1990
2005
2013
2014
2015
2016
2017
Exploration 0.6
0.6 1
4.3
5.2
3.8
2.1
2.5
HF Completions with Flaring 0.1
0.2 |
3.8
4.7
3.3
2.0
2.5
Energy 3-77
-------
Non-Completion Well Testing
0.4
0.4
0.4
0.5
0.5
0.1
0.1
with Flaring
Production
14.8
17.9
43.6
53.1
58.5
42.8
42.4
Associated Gas Flaring
14.8
11.0
26.3
32.1
35.5
25.9
28.2
Storage Tanks with Flaring
NO
5.7
12.7
15.1
16.3
12.6
9.0
Misc. Production Flaring
NO
1.2
4.3
5.6
6.3
4.0
4.9
HF Workovers with Flaring
+
+
0.2
0.3
0.3
0.4
0.4
Crude Oil Transportation
NE
NE
NE
NE
NE
NE
NE
Refining
30.7
34.8
34.2
32.4
39.1
38.8
36.4
Refinery Flares
30.7
34.8
34.2
32.4
39.1
38.8
36.4
Total
46.1
53.3
82.0
90.7
101.4
83.7
81.3
NE (Not Estimated)
NO (Not Occurring)
+ less than 0.05
Planned Improvements
Offshore Platforms
EPA is considering updates to the offshore platform emissions calculation methodology, as discussed in the
memorandum Iriven lory of U.S. Greenhouse Gas Emissions and Sinks 1990-2016: Additional Revisions Under
Consideration ,84 The current emission factors were based on data from the 2011 DOI/Bureau of Ocean Energy
Management's (BOEM) dataset, and 2014 BOEM data are available. A different source for platform counts is also
being considered.
Well-Related Activity Data
EPA will continue to assess available data, including data from the GHGRP and stakeholder feedback on
considerations, to improve activity estimates for sources that rely on well-related activity data. For example, EPA
will review GHGRP data regarding reported well workover rates and seek information on other data sets that might
inform estimates of non-hydraulically fractured oil well completions and workovers.
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, such as an upcoming field
study by American Petroleum Institute (API) on pneumatic controllers and separate studies by research groups that
will examine offshore platform emissions. 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.
84 See
3-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• 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 3-7: Carbon Dioxide Transport, Injection, and Geological Storage
Carbon dioxide is produced, captured, transported, and used for Enhanced Oil Recovery (EOR) as well as
commercial and non-EOR industrial applications. 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, CO2 that is used in non-EOR industrial and commercial applications (e.g., food processing,
chemical production) is assumed to be emitted to the atmosphere during its industrial use. These emissions are
discussed in the Carbon Dioxide Consumption section. The naturally-occurring CO2 used in EOR operations is
assumed to be fully sequestered. Additionally, all anthropogenic CO2 emitted from natural gas processing and
ammonia plants is assumed to be emitted to the atmosphere, regardless of whether the CO2 is captured or not. These
emissions are currently included in the Natural Gas Systems and the Ammonia Production sections of the Inventory
report, respectively.
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 2017. However, for 2015 through 2017, data from EPA's GHGRP
(Subpart PP) were held constant from 2014 levels, due to data confidentiality reasons. EPA will continue to evaluate
the availability of additional GHGRP data and other opportunities for improving the emission 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 6.0 MMT of CO2 sequestered in subsurface geological formations,
and 9,577 metric tons of CO2 emitted from equipment leaks.
These estimates indicate that the amount of CO2 captured and extracted from natural and industrial sites for EOR
applications in 2017 is 59.3 MMT CO2 Eq. (59,318 kt) (see Table 3-62 and Table 3-63). 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 3-62: Quantity of CO2 Captured and Extracted for EOR Operations (MMT CO2)
Stage 1990
2005
2013 2014 2015 2016 2017
Capture Facilities 4.8
Extraction F acilities 20.8
6.5
28.3
12.2 13.1 13.1 13.1 13.1
47.0 46.2 46.2 46.2 46.2
Total 25.6
34.7
59.2 59.3 59.3 59.3 59.3
Note: Totals may not sum due to independent rounding.
Energy 3-79
-------
Table 3-63: Quantity of CO2 Captured and Extracted for EOR Operations (kt)
Stage 1990
2005
2013 2014 2015 2016 2017
Capture Facilities 4,832
Extraction Facilities 20,811
6,475
28,267
12,205 13,093 13,093 13,093 13,093
46,984 46,225 46,225 46,225 46,225
Total 25,643
34,742
59,189 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. Overall, natural gas systems emitted 165.6 MMT
CO2 Eq. (6,624 kt) of CH4 in 2017, a 14 percent decrease compared to 1990 emissions, and less than 1 percent
decrease compared to 2016 emissions (see Table 3-64, Table 3-65, and Table 3-66), 26.3 MMT CO2 Eq. (26,327 kt)
of non-combustion CO2 in 2017, a 12 percent decrease compared to 1990 emissions, and a 3 percent increase
compared to 2016 levels, and 0.005 MMT CO2 Eq. (0.02 kt) of N2O, a 438 percent increase compared to 1990
emissions.
The 1990 to 2017 trend in CH4 is not consistent across segments. Overall, the 1990 to 2017 decrease in CH4
emissions is due primarily to the decrease in emissions from the distribution (73 percent decrease), transmission and
storage (43 percent decrease), processing (45 percent decrease), and exploration (69 percent decrease) segments.
Over the same time period, the production segment saw increased methane emissions of 62 percent (with onshore
production emissions increasing 29 percent, offshore production emissions increasing 7 percent, and gathering and
boosting (G&B) emissions increasing 109 percent). The 1990 to 2017 decrease in CO2 is due primarily to decreases
in acid gas removal emissions in the processing segment, where acid gas removal emissions per plant have
decreased over time.
Methane and non-combustion 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. Below is a characterization of the five major stages
of the natural gas system. Each of the stages is described and the different factors affecting CH4 and non-combustion
CO2 emissions are discussed.
Emissions of N20 from flaring activities are included in the Inventory, with most of the emissions occurring in the
processing and production segments.
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 2016) to
ensure that the trend is accurate. Recalculations in natural gas systems in this year's Inventory include:
• Updated methodology for G&B pipeline emissions.
• Updated methodology for transmission pipeline blowdown emissions.
• Updated methodology for LNG estimates (emissions for both storage stations and import/export terminals)
within the transmission and storage segment.
• Added N20 emissions that were not previously reported in the Inventory.
• Updated the data source for well drilling activity.
• Recalculations due to GHGRP submission revisions.
The Recalculations Discussion section below provides more details on the updated methods.
3-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Exploration. Exploration includes well drilling, testing, and completions. Emissions from exploration account for 1
percent of CH4 emissions and 2 percent of CO2 emissions from natural gas systems in 2017. Well completions
account for most of the CH4 emissions in 2017, with well testing and drilling also contributing emissions. Flaring
emissions account for most of the non-combustion CO2 emissions. Methane emissions from exploration decreased
by 69 percent from 1990 to 2017, with the largest decreases coming from hydraulically fractured gas well
completions without reduced emissions completions (RECs) or flaring. Methane emissions increased 75 percent
from 2016 to 2017 due to increases in emissions from completions, mostly from hydraulically fractured well
completions with RECs without flaring. Methane emissions were highest from 2006 to 2008. Carbon dioxide
emissions from exploration increased by 18 percent from 1990 to 2017, and by 149 percent from 2016 to 2017 due
to increases in flaring. Carbon dioxide emissions were highest from 2006 to 2008. Nitrous oxide emissions
decreased 37 percent from 1990 to 2017, and increased 156 percent from 2016 to 2017.
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 well-site gas treatment equipment such as
dehydrators and separators. 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). Emissions from production (including gathering and boosting) account
for 65 percent of CH4 emissions and 11 percent of non-combustion CO2 emissions from natural gas systems in 2017.
Emissions from compressors, pneumatic controllers, and offshore platforms account for most of the CH4 emissions
in 2017. Flaring emissions account for most of the non-combustion CO2 emissions with the highest emissions
coming from miscellaneous production flaring, flaring to control tank emissions, and offshore flaring. National total
dry gas production in the U.S. increased by 53 percent from 1990 to 2017, and by 3 percent from 2016 to 2017.
Methane emissions from production increased by 62 percent from 1990 to 2017, 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 gathering and boosting stations. Methane emissions
increased 1 percent from 2016 to 2017 due to increases in emissions from gathering and boosting stations and
hydraulically fractured well workovers with RECs and venting. Carbon dioxide emissions from production
increased by 175 percent from 1990 to 2017 due to increases in flaring, and decreased 11 percent from 2016 to 2017
due primarily to a decrease in emissions from large tanks with flares. Nitrous oxide emissions increased 480 percent
from 1990 to 2017 and decreased 8 percent from 2016 to 2017.
Processing. In this stage, 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. Fugitive CH4 emissions from compressors,
including compressor seals, are the primary emission source from this stage. Most of the non-combustion CO2
emissions come from acid gas removal (AGR) units, which are designed to remove CO2 from natural gas.
Processing plants account for 7 percent of CH4 emissions and 85 percent of non-combustion CO2 emissions from
natural gas systems. Methane emissions from processing decreased by 45 percent from 1990 to 2017 as emissions
from compressors (leaks and venting) and equipment leaks decreased, and increased 3 percent from 2016 to 2017
due to increased emissions from centrifugal and reciprocating compressors. Carbon dioxide emissions from
processing decreased by 21 percent from 1990 to 2017, due to a decrease in acid gas removal emissions, and
increased 3 percent from 2016 to 2017 due to increased emissions from flaring. Nitrous oxide emissions increased
from 1990 to 2017, and have decreased 20 percent from 2016 to 2017.
Transmission and Storage. Natural gas transmission involves high pressure, large diameter pipelines that transport
gas long distances from field production and processing areas to distribution systems or large volume customers
such as power plants or chemical plants. Compressor station facilities are used to move the gas throughout the U.S.
transmission system. Leak CH4 emissions from these compressor stations and venting from pneumatic controllers
account for most of the emissions from this stage. Uncombusted engine exhaust and pipeline venting are also
sources of CH4 emissions from transmission. Natural gas is also injected and stored in underground formations, or
liquefied and stored in above ground tanks, during periods of low demand (e.g., summer), and withdrawn,
processed, and distributed during periods of high demand (e.g., winter). Compressors and dehydrators are the
primary contributors to emissions from storage. Emissions from LNG are also included under transmission and
storage. Methane emissions from the transmission and storage sector account for approximately 20 percent of
emissions from natural gas systems, while CO2 emissions from transmission and storage account for 2 percent of the
non-combustion CO2 emissions from natural gas systems. CH4 emissions from this source decreased by 43 percent
Energy 3-81
-------
from 1990 to 2017 due to reduced compressor station emissions (including emissions from compressors and leaks),
and decreased 6 percent from 2016 to 2017 due to reduced pipeline venting and the plugging of the Aliso Canyon
leak. CO2 emissions from transmission and storage have increased by 147 percent from 1990 to 2017, and by 45
percent from 2016 to 2017, due to increased emissions fromLNG export terminals. The quantity of LNG exported
from the U.S. increased by a factor of 12 from 1990 to 2017, and by 279 percent from 2016 to 2017. LNG
emissions are about 2 percent of CH4 and 74 percent of CO2 emissions from transmission and storage in year 2017.
Nitrous oxide emissions from transmission and storage increased by 79 percent from 1990 to 2017 and increased 22
percent from 2016 to 2017.
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 1,294,091 miles of distribution mains in 2017, an increase of nearly 350,000 miles since 1990
(PHMSA 2018). Distribution system emissions, which account for 7 percent of CH4 emissions from natural gas
systems and less than 1 percent of non-combustion CO2 emissions, result mainly from leak emissions from pipelines
and stations. An increased use of plastic piping, which has lower emissions than other pipe materials, has reduced
both CH4 and CO2 emissions from this stage, as have station upgrades at metering and regulating (M&R) stations.
Distribution system CH i emissions in 2017 were 73 percent lower than 1990 levels (changed from 43.5 MMT CO2
Eq. to 11.9 MMT CChEq.) and 1 percent lower than 2016 emissions, while distribution CO2 emissions in 2017 were
also 73 percent lower than 1990 levels and 1 percent lower than 2016 Emissions. CO2 emission from this segment
are less than 0.1 MMT CO2 Eq. across the time series.
Total CH4 emissions for the five major stages of natural gas systems are shown in MMT CO2 Eq. (Table 3-64) and
kt (Table 3-65). Table 3-66 provides additional information on how the estimates in Table 3-62 were calculated.
With recent updates to the Inventory, most emissions are calculated using a net emission approach. However, certain
sources are still calculated with a potential emission approach. Table 3-66 shows the calculated potential CH4
release (i.e., potential emissions before any controls are applied) from each stage, and the amount of CH4 that is
estimated to have been flared, captured, or otherwise controlled, and therefore not emitted to the atmosphere.
Subtracting the value for CH4 that is controlled, from the value for calculated potential release of CH4, results in the
total net emissions values. More disaggregated information on potential emissions and emissions is available in
Annex 3.6. See Methodology for Estimating CH4 and CO2 Emissions from Natural Gas Systems.
Table 3-64: ChU Emissions from Natural Gas Systems (MMT CO2 Eq.)a
Stage
1990
2005
2013
2014
2015
2016
2017
Exploration6
4.0
10.9
3.0
1.0
1.0
0.7
1.2
Production
67.0
89.5
108.5
108.5
108.8
107.1
108.4
Onshore Production
35.0
51.5
53.3
49.3
47.2
46.0
45.1
Offshore Production
3.5
4.3
3.8
3.8
3.8
3.8
3.8
Gathering and Boosting0
28.5
33.7
51.4
55.4
57.9
57.4
59.5
Processing
21.3
11.6
10.8
11.1
11.1
11.4
11.7
Transmission and Storage
57.2
36.1
31.0
32.4
34.2
34.5
32.4
Distribution
43.5
23.3
12.3
12.2
12.0
12.0
11.9
Total
193.1
171.4
165.6
165.1
167.2
165.7
165.6
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.
Note: Totals may not sum due to independent rounding.
Table 3-65: ChU Emissions from Natural Gas Systems (kt)a
Stage
1990
2005
2013
2014
2015
2016
2017
Explorationb
162
437
119
39
42
28
49
Production
2,679
3,578
4,340
4,338
4,353
4,286
4,337
Onshore Production
1,399
2,058
2,133
1,972
1,888
1,840
1,806
Offshore Production
141
173
151
151
151
151
151
Gathering and Boosting0
1,139
1,347
2,056
2,216
2,315
2,295
2,380
Processing
853
464
432
443
443
456
469
3-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Transmission and Storage 2,289
Distribution 1,741
1,444
932
1,239 1,295 1,368 1,380 1,295
494 487 481 480 475
Total 7,723
6,856
6,624 6,603 6,686 6,629 6,624
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.
Note: Totals may not sum due to independent rounding.
Table 3-66: Calculated Potential CH4 and Captu red/Com busted Cm from Natural Gas
Systems (MMT CO2 Eq.)
1990
2005
2013
2014
2015
2016
2017
Calculated Potential3
193.1
182.0
179.0
178.4
180.5
179.1
179.0
Exploration
4.0
10.9
3.0
1.0
1.0
0.7
1.2
Production
67.0
94.8
115.2
115.1
115.5
113.8
115.1
Processing
21.3
11.6
10.8
11.1
11.1
11.4
11.7
Transmission and Storage
57.2
41.4
37.7
39.1
40.9
41.2
39.1
Distribution
43.5
23.3
12.3
12.2
12.0
12.0
11.9
Captured/Combusted
0.0
10.6
13.4
13.4
13.4
13.4
13.4
Exploration
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Production
0.0
5.3
6.7
6.7
6.7
6.7
6.7
Processing
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Transmission and Storage
0.0
5.3
6.7
6.7
6.7
6.7
6.7
Distribution
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Net Emissions
193.1
171.4
165.6
165.1
167.2
165.7
165.6
Exploration
4.0
10.9
3.0
1.0
1.0
0.7
1.2
Production
67.0
89.5
108.5
108.5
108.8
107.1
108.4
Processing
21.3
11.6
10.8
11.1
11.1
11.4
11.7
Transmission and Storage
57.2
36.1
31.0
32.4
34.2
34.5
32.4
Distribution
43.5
23.3
12.3
12.2
12.0
12.0
11.9
a In this context, "potential" means the total
emissions calculated before voluntary reductions and regulatory
controls are applied.
Note: Totals may not sum due to independent rounding.
Table 3-67: Non-combustion CO2 Emissions from Natural Gas Systems (MMT)
Stage
1990
2005
2013
2014
2015
2016
2017
Exploration
0.4
1.8
1.3
0.8
0.3
0.2
0.5
Production
1.0
1.8
3.1
3.3
3.4
3.2
2.8
Processing
28.3
18.9
20.5
21.0
21.0
21.7
22.5
Transmission and Storage
0.2
0.2
0.3
0.3
0.3
0.4
0.5
Distribution
0.1
+
+
+
+
+
+
Total
30.0
22.6
25.1
25.5
25.1
25.5
26.3
+ Does not exceed 0.1 MMT CO2 Eq.
Note: Totals may not sum due to independent rounding.
Table 3-68: Non-combustion CO2 Emissions from Natural Gas Systems (kt)
Stage
1990
2005
2013
2014
2015
2016
2017
Exploration
409
1,756
1,281
843
291
194
483
Production
1,035
1,759
3,076
3,342
3,448
3,188
2,845
Processing
28,338
18,876
20,510
21,047
21,047
21,724
22,452
Transmission and Storage
216
219
267
272
271
368
533
Distribution
51
27
15
14
14
14
14
Total
30,048
22,638
25,148
25,518
25,071
25,488
26,327
Energy 3-83
-------
Note: Totals may not sum due to independent rounding.
Table 3-69: N2O Emissions from Natural Gas Systems (Metric Tons CO2 Eq.)
Stage
1990
2005
2013
2014
2015
2016
2017
Exploration
461
1,401
1,179
855
3,215
113
289
Production
162
900
2,330
1,997
2,773
1,019
937
Processing
NO
3,351
5,625
5,772
5,772
3,802
3,049
Transmission and Storage
257
309
341
344
347
377
461
Distribution
NO
NO
NO
NO
NO
NO
NO
Total
880
5,961
9,476
8,969
12,107
5,311
4,735
NO (Not Occurring)
Note: Totals may not sum due to independent rounding.
Table 3-70: N2O Emissions from Natural Gas Systems (Metric Tons N2O)
Stage
1990
2005
2013
2014
2015
2016
2017
Exploration
1.5
4.7
4.0
2.9
10.8
0.4
1.0
Production
0.5
3.0
7.8
6.7
9.3
3.4
3.1
Processing
NO
11.2
18.9
19.4
19.4
12.8
10.2
Transmission and Storage
0.9
1.0
1.1
1.2
1.2
1.3
1.5
Distribution
NO
NO
NO
NO
NO
NO
NO
Total
3.0
20.0
31.8
30.1
40.6
17.8
15.9
NO (Not Occurring)
Note: Totals may not sum due to independent rounding.
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 emission estimates in the Inventory,
which involves the calculation of CH4, CO2, and N20 emissions for over 100 emissions sources, 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 teclinology-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 reduction data
to calculate net emissions.
Emission Factors. Key references for emission factors for CH4 and non-combustion-related 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 2018), and others.
The EPA/GRI 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 EPA/GRI study was based on a combination of
process engineering studies, collection of activity data, and measurements at representative gas facilities conducted
in the early 1990s. Year-specific natural gas CH4 compositions are calculated using U.S. Department of Energy's
Energy Information Administration (EIA) annual gross production for National Energy Modeling System (NEMS)
oil and gas supply module regions in conjunction with data from the Gas Teclinology Institute (GTI, formerly GRI)
Unconventional Natural Gas and Gas Composition Databases (GTI 2001). These year-specific CH4 compositions are
applied to emission factors, which therefore may vary from year to year due to slight changes in the CH4
composition for each NEMS region.
GHGRP Subpart W data were used to develop CH4, CO2, and N20 emission factors for many sources in the
Inventory. In the exploration and production segments, GHGRP data were used to develop emission factors used for
3-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
all time series years for well testing, gas well completions and workovers with and without hydraulic fracturing,
pneumatic controllers and chemical injection pumps, condensate tanks, liquids unloading, miscellaneous flaring, and
gathering and boosting pipelines. In the processing segment, for recent years of the times series, GHGRP data were
used to develop emission factors for fugitives, compressors, flares, dehydrators, and blowdowns/venting. In the
transmission and storage segment, GHGRP data were used to develop factors for all time series years for LNG
stations and terminals and transmission pipeline blowdowns, and for pneumatic controllers for recent years of the
times series.
Other data sources used for CH4 emission factors include Zimmerle et al. (2015) for transmission and storage station
fugitives and compressors, and Lamb et al. (2015) for recent years for distribution pipelines and meter/regulator
stations.
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 2001 GTI publication were
used to adapt the CH4 emission factors into non-combustion 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 N20 emissions were estimated for flaring sources using GHGRP data.
See Annex 3.6 for more detailed information on the methodology and data used to calculate CH4 and non-
combustion CO2 and N20 emissions from natural gas systems.
Activity Data. Activity data were taken from various published data sets, as detailed in Annex 3.6. Key activity data
sources include data sets developed and maintained by EPA's GHGRP; Drillinglnfo, Inc. (Drillinglnfo 2018); U.S.
Department of the Interior's Bureau of Ocean Energy Management, Regulation and Enforcement (BOEMRE,
previously Minerals and Management Service); Federal Energy Regulatory Commission (FERC); EIA; the Natural
Gas STAR Program annual emissions savings 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 2016 data were used as a
proxy for 2017 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 certain sectors, 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 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 in the production and transmission and storage
segments.
In fall of 2015, a well in a California storage field began leaking methane at an initial average rate of around 50
metric tons (MT) of methane (CH4) an hour, and continued leaking until it was permanently sealed in February of
20 1 6.85 An emission estimate from the leak event was included for 2015 and 2016, using the estimate of the leak
85 For more information on the Aliso Canyon event, and the measurements conducted of the leak, please see Ensuring Safe and
Reliable Underground Natural Gas Storage, Final Report of the Interagency Task Force on Natural Gas Storage Safety, available
at .
Energy 3-85
-------
published by the California Air Resources Board (99,638 MT CH4 for the duration of the leak). The 2015 and 2016
emission estimates of 78,350 MT CH4 and 21,288 MT CH4, respectively, were added to the 2015 and 2016
estimates of fugitive emissions from storage wells. For more information, please see Inventory of U.S. Greenhouse
Gas Emissions and Sinks 1990-2015: Update for Storage Segment Emissions,86
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, 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).87 EPA used Microsoft Excel's @RISK add-in tool to
estimate the 95 percent confidence bound around CH4 emissions from natural gas systems for the current Inventory,
then applied the calculated bounds to both CH4 and CO2 emissions estimates. For the analysis, EPA focused on the
14 highest-emitting sources for the year 2016, which together emitted 76 percent of methane from natural gas
systems in 2017, 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 2017, using the IPCC methodology. The results of the Approach 2
uncertainty analysis are summarized in Table 3-71. Natural gas systems CH4 emissions in 2017 were estimated to be
between 141.8 and 193.3 MMT CO2 Eq. at a 95 percent confidence level. Natural gas systems CO2 emissions in
2017 were estimated to be between 22.5 and 30.7 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 3-71: Approach 2 Quantitative Uncertainty Estimates for CH4 and Non-combustion CO2
Emissions from Natural Gas Systems (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)b
(MMT CO2 Eq.)
(%)
Lower Upper
Boundb Boundb
Lower Upper
Boundb Boundb
Natural Gas Systems
CH4
165.6
141.8 193.3
-14% +17%
Natural Gas Systems0
CO2
26.3
22.5 30.7
-14% +17%
a Range of emission estimates estimated by applying the 95 percent confidence intervals obtained from the Monte Carlo
Simulation analysis conducted for the year 2017 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 3-64 and Table 3-65.
86 -
87 See < https://www.epa.gov/ghgemissions/natural-gas-and-petroleum-systems>.
3-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
c An uncertainty analysis for the CO2 emissions was not performed. The relative uncertainty estimated (expressed as a
percent) from the CH4 uncertainty analysis was applied to the point estimate of CO2 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.
QA/QC and Verification Discussion
The natural gas emission estimates in the Inventory are continually being reviewed and assessed to determine
whether emission factors and activity factors accurately reflect current industry practices. A QA/QC analysis was
performed for data gathering and input, documentation, and calculation. QA/QC checks are consistently conducted
to minimize human error in the model calculations. EPA performs a thorough review of information associated with
new studies, GHGRP data, regulations, public webcasts, and the Natural Gas STAR Program to assess whether the
assumptions in the Inventory are consistent with current industry practices. The EPA has a multi-step data
verification process for GHGRP data, including automatic checks during data-entry, statistical analyses on
completed reports, and staff review of the reported data. Based on the results of the verification process, the EPA
follows up with facilities to resolve mistakes that may have occurred.88
As in previous years, EPA conducted early engagement and communication with stakeholders on updates prior to
public review. EPA held a stakeholder workshop on greenhouse gas data for oil and gas in October of 2018, and
webinars in June of 2018 and February 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 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.89 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.90
Recalculations Discussion
88 See .
89 See .
90 See .
Energy 3-87
-------
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
June, October and November 2018, EPA released draft memoranda that discussed changes under consideration, and
requested stakeholder feedback on those changes. EPA then created updated versions of the memoranda to
document the methodology implemented into the current Inventory ,91 Memoranda cited in the Recalculations
Discussion below are: Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2017: Update for Natural Gas
Gathering & Boosting Emissions (April 2019 G&B memo), Inventory of U.S. Greenhouse Gas Emissions and Sinks
1990-2017: Update for Liquefied Natural Gas Segment Emissions (April 2019 LNG memo), and Inventory of U.S.
Greenhouse Gas Emissions and Sinks 1990-2017: Other Updates Considered for 2019 and Future GHGls (April
2019 Other Updates memo).
EPA thoroughly evaluated relevant information available and made several updates to the Inventory, including:
using GHGRP data to calculate emissions from gathering pipelines, transmission pipeline blowdowns, and LNG
storage stations and terminals; calculating new N20 emission factors for flaring sources throughout all segments
directly from GHGRP data; and updating the data source for well drilling activity. 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 2016 to the current (recalculated) estimate for 2016 (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 miscellaneous flaring, production segment pneumatic
controllers, liquids unloading, production segment storage tanks, G&B stations, acid gas removal (AGR) vents and
flares at gas processing plants, and gas engines in the production and processing segments. Lastly, for HF gas well
workovers, year 2017 emissions estimates are noticeably higher than previous years; the factors driving this increase
are described below.
The combined impact of revisions to 2016 natural gas sector CH4 emissions, compared to the previous Inventory, is
an increase from 163.5 to 165.7 MMT CO2 Eq. (2.2 MMT CO2 Eq., or 1 percent). The recalculations resulted in an
average increase in CH4 emission estimates across the 1990 through 2016 time series, compared to the previous
Inventory, of 0.6 MMT CO2 Eq., or 0.4 percent.
The combined impact of revisions to 2016 natural gas sector CO2 emissions, compared to the previous Inventory, is
minimal, with emissions of approximately 25.5 MMT CO2 in both Inventories. The recalculations resulted in an
average increase in emission estimates across the 1990 through 2016 time series, compared to the previous
Inventory, of 0.2 MMT CO2 Eq, or 0.7 percent.
In Table 3-72 and Table 3-73 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 2016 to the current (recalculated) estimate
for 2016. For more information, please see the Recalculations Discussion below.
Table 3-72: Recalculations of CO2 in Natural Gas Systems (MMT CO2)
Previous Estimate
Current Estimate
Current Estimate
Year 2016,
Year 2016,
Year 2017,
Stage and Emission Source
2018 Inventory
2019 Inventory
2019 Inventory
Exploration
0.1
0.2
0.5
HF Completions
0.1
0.2
0.5
Production
3.2
3.2
2.8
Gathering Pipelines
+
+
+
Miscellaneous Flaring
1.1
1.2
1.1
Tanks
1.2
1.1
0.6
HF Workovers
+
0.1
0.4
Processing
22.0
21.7
22.5
AGR Vents
16.6
16.5
16.7
Flares
5.4
5.2
5.7
Transmission and Storage
0.1
0.4
0.5
LNG Storage + + +
91 Stakeholder materials including draft and final EPA memoranda for the current (i.e., 1990 to 2017) Inventory are available at
.
3-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
LNG Import/Export
Terminals
Pipeline Blowdowns +
Distribution +
Total 25~J 25o 26.3
+ Does not exceed 0.05 MMT CO2.
Table 3-73: Recalculations of CH4 in Natural Gas Systems (MMT CO2 Eq.)
Previous Estimate
Current Estimate
Current Estimate
Year 2016,
Year 2016,
Year 2017,
Stage and Emission Source
2018 Inventory
2019 Inventory
2019 Inventory
Exploration
0.7
0.7
1.2
Production
106.8
107.1
108.4
G&B Stations
53.7
53.6
55.5
Gathering Pipelines
4.0
3.8
4.0
Pneumatic Controllers
26.3
26.6
26.4
Liquids Unloading
3.3
3.3
2.9
HF Workovers
0.4
0.4
0.8
Gas Engines
2.7
3.0
2.8
Processing
11.2
11.4
11.7
Gas Engines
6.1
6.3
6.4
Transmission and Storage
32.8
34.5
32.4
LNG Storage
1.8
0.2
0.3
LNG Import/Export
0.3
0.4
0.4
Terminals
Pipeline Blowdowns
4.6
6.3
4.6
Distribution
12.0
12.0
11.9
Total
163.5
165.7
165.6
0.2
+
+
0.4
+
+
Exploration
Well Drilling (Methodological Update)
EPA updated the methodology for estimating the number of gas wells drilled across the time series to use
Drillinglnfo data (Drillinglnfo 2018). The new methodology is detailed in the. Ipr. 2019 Other Updates memo. In
previous Inventories, the U.S. Department of Energy's Energy Information Administration (DOE/EIA) Monthly
Energy Review well drilling activity data set was used to develop well drilling activity inputs, but this publication
does not provide data after year 2010. EPA therefore developed a methodology of analyzing Drillinglnfo data to
estimate counts of gas wells drilled in each time series year, 1990 through 2017. These activity data for select years
are shown in Table 3-74 below.
Table 3-74: Count of Gas Wells Drilled
Activity
1990
2005
2013
2014
2015
2016
2017
Gas Wells Drilled
17,805 1
27,568
5,681
5,871
3,585
2,264
2,264
Previous Estimate
15,096
31,969
18,837"
18,837"
18,837"
18,837"
NA
a - Year-specific data not available; the year 2010 estimate was assigned as a surrogate value.
NA (Not Applicable)
HF Gas Well Completions (Recalculation with Updated Data)
HF gas well completion CO2 emission estimates increased 47 percent in the current Inventory for year 2016,
compared to the previous Inventory, due to GHGRP submission revisions. Specifically, the GHGRP submission
revisions reported higher CO2 emissions for HF reduced emission completions with flaring, which led to a larger
CO2 emission factor. Compared to the previous Inventory, for 1990 to 2015, the CO2 emission estimates increased
by an average of only 0.3 percent.
Energy 3-89
-------
Table 3-75: HF Gas Well Completions National CO2 Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
HF Completions - Non-REC with
Venting
HF Completions - Non-REC with
Flaring
HF Completions - REC with
Venting
HF Completions - REC with
Flaring
Total Emissions
10
390
27
1,316
11
324
2
327
+
58
+
12
+
37
NO
NO
3
394
2
929
+
502
1
218
+
164
1
438
400
1,741
1,265
832
277
111
475
Previous Estimate
397
1,748
1,148
844
277
120
NA
NO (Not Occurring)
NA (Not Applicable)
+ Does not exceed 0.5 kt CO2.
Production
Gathering Pipelines (Methodological Update)
EPA developed new activity data and net emission factors for gathering pipeline sources (leaks and blowdowns)
using GHGRP data, as detailed in the , l/?r. 2019 G&B memo. Accordingly, the updated methodology no longer
incorporates data on the Gas STAR reductions from pipeline leaks. Using GHGRP data to estimate gathering
pipeline emissions was supported by stakeholder feedback. Compared to the previous Inventory, gathering pipeline
CH4 emission estimates decreased for recent years due to the newly calculated emission factors from GHGRP and
increased for early years due to updated well count activity data that drives pipeline mileage estimates. Compared
with the previous Inventory, on average, CH4 emission estimates decreased by 6 percent across the 1990 to 2016
time series. Compared to the previous Inventory, gathering pipeline CO2 emission estimates decreased by 10 percent
across the 1990 to 2016 time series, or by an average of 1.7 MMT CO2. See the. \pr. 2019 G&B memo for
additional discussion.
Table 3-76: Gathering Pipelines National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
G&B Pipeline Leaks
78,425
117,182
138,669
137,477
136,513
136,776
141,577
G&B Pipeline
Blowdowns
8,436
12,605
14,917
14,788
14,685
14,713
19,777
Total Emissions
86,861
129,787
153,586
152,266
151,198
151,489
161,354
Previous Estimate
85,413
136,627
164,443
164,727
162,796
160,311
AM
NA (Not Applicable)
Gathering and Boosting Stations (Recalculation with Updated Data)
G&B station CH4 emission estimates decreased by 0.3 percent in the current Inventory for year 2016, compared to
the previous Inventory. This change was not the result of a methodological update, but due to updated data for
marketed onshore gas production, which drives the station count activity data. EPA presented approaches to use
GHGRP data to estimate G&B station emissions, but stakeholder feedback supported maintaining the current
Inventory methodology (see the April 2019 G&B memo). Additional G&B station considerations for future
Inventories, particularly for estimating CO2 and N20 emissions, are discussed in the Planned Improvements section
below.
Table 3-77: Gathering Stations National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Gathering and Boosting
Stations
956,870
1,107,208
1,730,573
1,877,554
1,968,205
1,949,925
2,018,566
G&B Station Episodic
Events
94,905
109,816
171,643
186,221
195,212
193,399
200,207
Total Emissions
1,051,775
1,217,024
1,902,216
2,063,775
2,163,417
2,143,324
2,218,773
3-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Previous Estimate 1,051,775 1,217,024 1,902,216 2,063,775 2,163,417 2,149,065 NA_
NA (Not Applicable)
Miscellaneous Production Flaring (Recalculation with Updated Data)
Miscellaneous production flaring CO2 emission estimates decreased in the current Inventory for 1990 to 2015 and
increased in the current Inventory for 2016, compared to the previous Inventory. The CO2 emissions changes are due
to GHGRP submission revisions and use of GHGRP well counts from the facility overview table (see the Well
Counts discussion below). In addition, the emission calculations are performed at a basin-level, and the changes
impacted each basin uniquely.
Table 3-78: Miscellaneous Production Flaring National Emissions (kt CO2)
Source 1990 2005 2013 2014 2015 2016 2017
Miscellaneous Flaring-Gulf Coast
Basin
Miscellaneous Flaring-Williston
Basin
Miscellaneous Flaring-Permian
Basin
Miscellaneous Flaring-Other
Basins
Total Emissions
Previous Estimate
NO (Not Occurring)
NA (Not Applicable)
+ Does not exceed 0.5 kt CO2.
Gas Engines (Recalculation with Updated Data)
Natural gas engine CH4 emission estimates increased in the current Inventory by an average of approximately 4
percent across the time series, compared to the previous Inventory. This change was due to the updated Drillinglnfo
gas wells counts (see the Well Counts discussion below).
Table 3-79: Production Segment Gas Engines National Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Gas Engines
Previous Estimate
116,539
116,508
123,210
117,852
131,262
121,827
128,812
118,818
125,437
114,774
118,462
106,423
113,758
NA
NA (Not Applicable)
NO 155 250 296 331 243 193
NO + + + + NO 10
NO 256 434 535 644 506 579
NO 118 293 319 343 438 308
NO 530 978 1,150 1,317 1,186 1,090
NO 572 1,057 1,241 1,415 1,129 NA
Pneumatic Controllers (Recalculation with Updated Data)
Pneumatic controller CH4 emission estimates increased in the current Inventory by an average of approximately 2
percent across the time series, compared to the previous Inventory. This change was caused by several factors:
GHGRP submission revisions, the use of GHGRP well counts from the facility overview table (see the Well Counts
discussion below), a correction to the linear interpolation calculation for activity factors in years 1993 through 2010,
and updated Drillinglnfo gas well counts (see the Well Counts discussion below).
Table 3-80: Production Segment Pneumatic Controller National Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Low Bleed
NO
23,541
27,554
32,330
30,455
32,646
33,944
High Bleed
297,952
450,013
177,784
129,712
101,930
107,162
107,398
Intermittent Bleed
194,302
531,907
970,065
927,297
943,216
924,261
915,961
Total Emissions
492,254
1,005,461
1,175,402
1,089,339
1,075,601
1,064,069
1,057,303
Previous Estimate
506,905
981,773
1,134,147
1,072,375
1,055,935
1,053,207
NA
NO (Not Occurring)
NA (Not Applicable)
Energy 3-91
-------
Liquids Unloading (Recalculation with Updated Data)
Liquids unloading CH4 emission estimates increased for 2015 and decreased for 2016 in the current Inventory,
compared to the previous Inventory. Compared to the previous Inventory, on average across the time series, liquids
unloading CH4 emission estimates increased by 2 percent. These changes were due to GHGRP submission revisions
and the use of GHGRP well counts from the facility overview table (see the Well Counts discussion below). In
particular, the percent of gas wells requiring liquids unloading increased for the GHGRP reporting year 2015 data
(which is applied to all prior years of the time series) and decreased for reporting year 2016.
Table 3-81: Liquids Unloading National Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Unloading with Plunger Lifts
NO
126,009
124,036
80,880
63,089
61,397
46,843
Unloading without Plunger Lifts
372,325
247,433
110,095
129,904
97,616
69,381
70,536
Total Emissions
372,325
373,442
234,132
210,784
160,706
130,778
117,379
Previous Estimated Emissions
379,837
365,310
220,990
202,745
153,975
132,871
NA
NO (Not Occurring)
NA (Not Applicable)
Tanks (Recalculation with Updated Data)
Production tank CO2 emission estimates increased by an average of approximately 30 percent across 1990 to 2015 in
the current Inventory and decreased by about 8 percent in the current Inventory for 2016, compared to the previous
Inventory. The change in production tank CO2 emission estimates is mainly driven by GHGRP submission
revisions. For example, GHGRP reporting year 2015 CO2 emission estimates increased, which led to an increase in
the calculated emission factors, and year 2015 emission factors are applied to all prior years of the time series.
Table 3-82: Production Segment Storage Tanks National Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
Large Tanks w/Flares
287
363
984
1,030
1,041
1,080
558
Large Tanks w/VRU
NO
1
3
3
3
2
+
Large Tanks w/o Control
167
90
147
154
155
1
+
Small Tanks w/Flares
NO
8
30
31
31
33
22
Small Tanks w/o Flares
6
4
9
10
10
12
5
Malfunctioning Separator Dump
-|-
-|-
-|-
-|-
-|-
-|-
1
Valves
1
Total Emissions
460
466
1,173
1,227
1,240
1,129
585
Previous Estimate
294
378
1,030
1,078
1,089
1,224
NA
NO (Not Occurring)
NA (Not Applicable)
+ Does not exceed 0.5 kt CO2.
HF Gas Well Workovers (Year 2017 Emissions)
Recalculations of HF gas well workover emissions did not result in large changes 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 seen in the tables below results from the use of GHGRP data reported for 2017, a year with a
sharp increase in reduced emission workovers with flaring.
Table 3-83: HF Gas Well Workovers National Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
HF Workovers - Non-REC with
Venting
25,823
66,053
69,935
25,517
2,518
7,878
11,795
HF Workovers - Non-REC with
Flaring
366
1,034
350
476
225
76
527
HF Workovers - REC with
Venting
NO
625
2,711
589
8,035
6,301
17,193
3-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
HF Workovers - REC with
Flaring
NO
5
281
26
2,383
1,297
4,197
Total Emissions
26,188
67,717
73,276
26,608
13,161
15,551
33,711
Previous Estimate
25,244
66,781
72,557
26,957
13,228
16,986
NA
NO (Not Occurring)
NA (Not Applicable)
Table 3-84: HF Gas Well Workovers National Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
HF Workovers - Non-REC with
Venting
2
4
8
2
+
+
2
HF Workovers - Non-REC with
Flaring
66
187
70
156
17
12
41
HF Workovers - REC with
Venting
NO
+
+
+
+
+
+
HF Workovers - REC with
Flaring
NO
1
55
5
59
47
313
Total Emissions
68
193
133
163
77
59
356
Previous Estimate
65
190
125
156
77
44
NA
NO (Not Occurring)
NA (Not Applicable)
+ Does not exceed 0.5 kt CO2.
Well Counts (Recalculation with Updated Data)
For total national well counts, EPA lias used a more recent version of the Drillinglnfo data set (Drillinglnfo 2018) to
update well counts data in the Inventory. EPA also updated the Drillinglnfo data processing methodology to more
accurately count wells in states with lease-level reporting (e.g., Kansas), which resulted in slight increased counts
across the time series. While this was not a significant recalculation (increases are 2 to 3 percent across the time
series), this is a key input to the Inventory, so results are highlighted here.
Table 3-85: Producing Gas Well Count Data
Activity
1990
2005
2013
2014
2015
2016
2017
Number of Gas Wells
Previous Estimate
193,718
197,626
346,862
348,470
428,947
427,828
424,308
431,446
420,418
425,651
419,005
416,881
411,450
NA
NA (Not Applicable)
In October 2018, EIA released an updated time series of national oil and gas well counts (covering 2000 through
2017). EIA estimates 991,000 total producing wells for year 2017. EPA's total well count for this year is 978,176.
EPA's well counts in recent time series years are generally 2 percent lower than EIA's. EIA'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.
For the count of wells included in GHGRP reporting (used to develop wellhead-based emissions and activity
factors), EPA previously referenced the wellhead counts contained within the reporting table for onshore production
equipment leak emissions. Due to updated reporting requirements for year 2017 forward, well counts provided as
part of the facility overview information (i.e., wells producing at the end of the calendar year plus wells removed
from production in a given year) provide more complete estimates. Therefore, EPA used well counts from the
facility overview table for source-specific methodologies that rely on GHGRP reported well counts in the current
Inventory. Comparing the GHGRP well counts from the facility overview table to the equipment leaks table: a larger
population of the wells were reported as "oil" production type in the facility overview information table, compared
to the equipment leaks table, which generally led to decreased activity and emissions for natural gas systems; for
example, as discussed in the sections above, production segment emissions from pneumatic controllers and
miscellaneous production flaring decreased across most of the time series.
Energy 3-93
-------
Processing
Acid Gas Removal (Recalculation with Updated Data)
Acid gas removal unit (AGR) CO2 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 less than 0.01 percent.
There was a decrease in CO2 emission estimates for 2016, comparing the current Inventory to the previous
Inventory. This decrease in CO2 emission estimates for 2016 is due to GHGRP submission revisions, where a lower
emission factor was calculated from the GHGRP data.
Table 3-86: AGR National CO2 Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
Acid Gas Removal
Previous Estimate
28,282
28,282
15,320
15,320
14,565
14,565
14,946
14,946
14,946
14,946
16,481
16,565
16,728
NA
NA (Not Applicable)
Flares (Recalculation with Updated Data)
Processing segment flare CO2 emission estimates increased by only 0.03 percent across the 1990 to 2015 time series
in the current Inventory and decreased by approximately 4 percent for 2016 in the current Inventory, compared to
the previous Inventory. This decrease in CO2 emission estimates for 2016 is due to GHGRP submission revisions,
where a lower emission factor was calculated from the GHGRP data.
Table 3-87: Processing Segment Flares National CO2 Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
Flares
NO 1
3,517 1
5,904
6,058
6,058
5,203
5,683
Previous Estimate
NO
3,516
5,902
6,056
6,056
5,404
NA
NO (Not Occurring)
NA (Not Applicable)
Gas Engines (Recalculation with Updated Data)
Gas engine CH4 emission estimates increased by approximately 0.1 percent across the 1990 to 2015 time series in
the current Inventory and increased by approximately 3 percent for 2016 in the current Inventory, compared to the
previous Inventory. This increase in CH4 emissions for 2016 is due to GHGRP submission revisions, where a higher
activity factor (MMhphr/plant) was calculated from the GHGRP data.
Table 3-88: Processing Segment Gas Engines National Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Gas Engines
Previous Estimate
137,102
137,102
169,388
169,101
228,152
227,671
234,119
233,626
234,119
233,626
250,368
242,451
255,822
NA
NA (Not Applicable)
Transmission and Storage
Transmission Pipeline Blow downs (Methodological Update)
EPA developed new CH4 and CO2 emission factors for transmission pipeline blowdowns using GHGRP data, as
detailed in the Apr. 2019 Other Updates memo. In response to stakeholder comments on the Public Review draft,
EPA applied year-specific emission factors calculated from GHGRP data for years 2016 forward and retained
historical emission factors for earlier years. As a result, compared to the previous Inventory, calculated CH4
emissions from this source increased by 37 percent for year 2016 and remained constant over the rest of the time
series, while CO2 emission estimates increased by 33 percent for year 2016. As the updated methodology uses net
emission factors, the Gas STAR reduction data from pipeline blowdowns have been removed from the calculations
3-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
(see discussion below). See the. Ipr. 2019 Other Updates memo for additional discussion and the Planned
Improvements section below for considerations for future Inventories.
Table 3-89: Transmission Pipeline Blowdowns National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
Pipeline Blowdowns
Previous Estimate
177,951
177,951
183,159
183,159
184,628
184,596
183,984
183,973
183,583
183,538
250,175
183,081
184,455
NA
NA (Not Applicable)
Table 3-90: Transmission Pipeline Blowdowns National CO2 Emissions (kt CO2)
Source
1990 2005 2013
2014
2015
2016
2017
Pipeline Blowdowns
5l 5l
5
5
7
5
Previous Estimate
5 5 5
5
5
5
NA
NA (Not Applicable)
LNG Storage (Methodological Update)
For LNG storage facilities, EPA updated the Inventory methodology to use available GHGRP data paired with
updated activity estimates, as detailed in the Apr. 2019 LNG memo. EPA developed facility-level average CH4 and
CO2 emission factors that represent emissions from station fugitives, compressor vented and fugitive sources, and
flaring using combined GHGRP data from years 2015 through 2017 and applied these emission factors across the
time series. To estimate LNG storage station CH4 and CO2 blowdown emissions, EPA maintained the current
Inventory emission factors. For activity data (storage station counts), EPA used the existing estimates for years 1990
through 2009 (although the total count of complete storage stations plus satellite stations were used, not a fraction of
the satellite stations like the previous Inventory methodology) and reviewed current PHMSA data in conjunction
with GHGRP data to obtain a count of active storage stations for years 2010 forward. For compressor exhaust CH4
emissions, EPA updated activity factors and maintained the current Inventory emission factors. EPA developed
average activity factors (i.e., MMhphr/station for each compressor driver type) using combined GHGRP data from
years 2015 through 2017 and applied these activity factors across the time series. Compared to the previous
Inventory, these updates resulted in an average decrease of 86 percent in CH4 emission estimates across the time
series and CO2 emission estimates increased by an average factor of 17 across the time series.
Table 3-91: LNG Storage Station National ChU Emissions (Metric Tons ChU)
Source 1990
2005
2013
2014
2015
2016
2017
LNG storage stations 1,138
1,396
1,411
1,411
1,425
1,382
1,396
LNG storage station blowdowns 6,571
8,060
8,144
8,144
8,228
7,976
8,060
LNG storage engines 476
584
590
590
596
578
584
LNG storage turbines 26
32
33
33
33
32
32
Total Emissions 8,212
10,072
10,177
10,177
10,282
9,967
10,072
Previous Estimate 63,258
73,124
73,124
73,124
73,124
73,124
NA
NA (Not Applicable)
ible 3-92: LNG Storage Station National CO2 Emissions (kt CO2)
Source 1990
2005
2013
2014
2015
2016
2017
LNG storage stations 36
45
45
45
45
44
45
LNG storage station blowdowns +
+
+
+
+
+
+
Total Emissions 37
45
45
45
46
44
45
Previous Estimate 2
2
2
2
2
2
NA
NA (Not Applicable)
+ Does not exceed 0.5 kt.
Energy 3-95
-------
LNG Import Export Terminals (Methodological Update)
For LNG terminals, EPA updated the Inventory methodology to use available GHGRP data paired with updated
activity estimates, as detailed in the Apr. 2019 LNG memo. This methodological update also resulted in the creation
of a new category for export terminals in the Inventory; previously, emissions were only estimated for import
terminals. EPA used GHGRP data to develop facility-level CH4 and CO2 emission factors that represent emissions
from station fugitives, blowdowns, compressor vented and fugitive sources, and flaring. EPA developed these
facility-level emission factors for two categories of facilities: import-only terminals (import terminals) and terminals
with export capability (export terminals). For import terminals, EPA calculated average CH4 and CO2 emission
factors using combined GHGRP data from years 2015 through 2017 and applied these emission factors across the
time series. For export terminals, EPA used year-specific GHGRP CH4 and CO2 data for 2015 through 2017 to
develop emission factors and applied the year 2015 emission factors to prior time series years. For import terminals
activity data, EPA used the existing Inventory import terminal counts for years 1990 through 2003 and reviewed
current DOE data in conjunction with GHGRP data to obtain a count of existing import terminals for years 2004
forward. For export terminals activity data, EPA reviewed current DOE data in conjunction with GHGRP data to
obtain a count of existing terminals with export capability across the time series. For compressor exhaust CH4
emissions, EPA updated activity factors and maintained the current Inventory emission factors. For import terminals
compressor exhaust, EPA developed average activity factors (i.e., MMhphr/station for each compressor driver type)
using combined GHGRP data from years 2015 through 2017 and applied these factors across the time series. For
export terminals compressor exhaust, EPA used year-specific GHGRP activity data for 2015 through 2017 to
develop activity factors (i.e., MMhphr/station for each compressor driver type) and applied the year 2015 activity
factors to prior time series years. These LNG terminal updates resulted in an average increase of 8 percent in CH4
emissions across the time series and CO2 emission estimates increased by an average factor of approximately 286
across the time series, when comparing the emissions from import and export terminals in the current Inventory to
emissions from import terminals in the previous Inventory.
Table 3-93: LNG Import/Export Terminal National ChU Emissions (Metric Tons ChU)
Source
1990
2005
2013
2014
2015
2016
2017
LNG Import Terminals Misc.
Sources3
114
284
625
625
625
568
568
LNG Import Terminal
Blowdowns
2,635
6,587
f4,49f
f4,49f
f4,49f
f 3, f 74
f 3,f 74
LNG Import Terminal Engines
226
566
f ,245
f ,245
f ,245
f ,f 32
f ,f 32
LNG Import Terminal Turbines
+
+
+
+
+
+
+
LNG Export Terminals Misc.
Sources3
801
80 f
80 f
80 f
80 f
350
f ,0f4
LNG Export Terminal
+
+
+
+
+
52
NO
Blowdowns
LNG Export Terminal Engines
NO
NO
NO
NO
NO
85
NO
LNG Export Terminal Turbines
ff
ff
ff
ff
ff
f
f
Total Emissions
3,787
8,249
17,174
17,174
17,174
15,363
15,889
Previous Estimated
3,341
15,445
10,902
10,190
10,801
10,741
NA
3 Equipment leaks, compressor vented and leak emissions, and flares.
b Includes emissions from LNG import terminals only.
NO (Not Occurring)
NA (Not Applicable)
+ Does not exceed 0.5 MT CH4.
Table 3-94: LNG Import/Export Terminal National CO2 Emissions (kt CO2)
Source
1990
2005
2013
2014
2015
2016
2017
LNG fmport Terminals Misc.
Sources3
15
37
80
80
80
73
73
LNG Import Terminal
Blowdowns
+
+
f
f
f
f
f
LNG Export Terminals Misc.
Sources3
+
+
+
+
+
98
278
3-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
NO NO NO + NO
81 81 81 172 352
+ + + + NA
a Equipment leaks, compressor vented and leak emissions, and flares.
a Includes emissions from LNG import terminals only.
NO (Not Occurring)
NA (Not Applicable)
+ Does not exceed 0.5 kt CO2.
LNG Export Tennmal
Blowdowns NO
NO
Total Emissions 15
37
Previous Estimate +
+
Distribution
There were no methodological updates to the distribution segment, but there were recalculations due to updated data
(e.g., GHGRP M&R station counts) that resulted in an average increase in calculated emissions over the time series
from this segment of 0.01 MMT CO2 Eq. CH4 (or 0.1 percent) and less than 0.01 MMT CO2 (or 0.1 percent).
Gas STAR Data Revisions
EPA updated its calculation of production and transmission segment Gas STAR reductions to take into account new
methods using net emission factors for certain sources. As in previous inventories, the "other" reductions scaling
factor for production is calculated as one minus the sum of emissions from sources with net approaches, divided by
the sum of all production segment emissions. The calculation was updated this year to remove reductions associated
with gathering pipeline blowdowns, as net emission factors are now used to calculate emission for that source. In
addition the line item for gathering pipeline leak Gas STAR reductions was removed. Similarly, reductions
associated with transmission pipelines blowdowns were removed from the transmission segment. As a result of the
update. Gas STAR reductions averaged 8.0 MMT CO2 Eq. over the time series, which is an average decrease across
the time series of 19 percent (or 1.7 MMT CO2 Eq.).
N2O Emissions
EPA newly calculated N2O emissions in the current Inventory, as discussed in the Apr. 2018 Other Updates memo.
Prior Inventories did not calculate N20 emissions from natural gas systems. For each flaring emission source
calculation methodology which uses GHGRP data, the existing source-specific methodology was applied to
calculate N20 emission factors. This update was applied for sources in the exploration, production processing, and
transmission and storage segments.
Table 3-95: N2O National Emissions (Metric Tons N2O)
Activity
1990
2005
2013
2014
2015
2016
2017
Exploration
1.5
4.7
4.0
2.9
10.8
0.4
1.0
Non-completion well testing -
0.8
1.4
1.7
1.7
1.7
NO
+
flared
HF Completions with Flaring
0.8
3.3
2.2
1.1
8.1
0.4
0.9
Non-HF Completions with
+
+
+
+
0.9
+
+
Flaring
Production
0.5
3.0
7.8
6.7
9.3
3.4
3.1
EtF Workovers with Flaring
0.1
0.4
0.2
0.2
2.2
0.1
0.7
Non-HF Workovers with
NO
1.7
NO
NO
Flaring
+
+
+
Misc. Onshore Production
NO
2.1
4.2
5.3
2.2
1.8
4.8
Flaring
Tanks with Flares
0.4
0.6
1.6
1.7
1.7
1.1
0.6
Processing
NO
11.2
18.9
19.4
19.4
12.8
10.2
Flares
NO
11.2
18.9
19.4
19.4
12.8
10.2
Transmission and Storage
0.9
1.0
1.1
1.2
1.2
1.3
1.5
Transmission Flaring
0.1
0.1
0.1
0.1
0.1
+
0.1
Storage Flaring
+
+
+
+
+
0.1
+
LNG Storage Flaring
0.7
0.8
0.8
0.8
0.8
0.8
0.8
Energy 3-97
-------
LNG Import Terminals Flaring
+
0.1
0.2
0.2
0.2
0.1
0.1
LNG Export Terminals Flaring
NO
NO
NO
NO
NO
0.2
0.5
Distribution
NO
NO
NO
NO
NO
NO
NO
Total Emissions
3.0
20.0
31.8
30.1
40.6
17.8
15.9
NO (Not Occurring)
+ Does not exceed 0.05 MT N2O.
Planned Improvements
EPA seeks stakeholder feedback on the improvements noted below for future Inventories.
Gathering and Boosting Stations
In the Oct. 2018 G&B memo, EPA presented approaches that rely on GHGRP data to estimate G&B station
emissions. Stakeholder feedback received in response to the Oct. 2018 G&B memo supported maintaining the
current Inventory approach. As such EPA maintained the current Inventory approach to estimate G&B station
emissions, and did not use a methodology that relies on GHGRP data. EPA will continue to review GHGRP data
and other research that becomes available to estimate G&B station emissions. EPA also requests specific feedback
on options to estimate G&B station flaring (CO2 and N20) emissions and AGR (CO2) emissions for future
Inventories. The current Inventory approach does not account for the CO2 emissions from flaring and AGR units.
EPA plans to review available data from upcoming studies and additional years of data reported to GHGRP to
improve estimated emissions from these sources. The GHGRP emissions from flaring in gathering and boosting
total 3,894 kt CO2 and 0.01 kt N2O reported for year 2017 (2,143 kt CO2 from miscellaneous flaring, 686 kt CO2
from flaring from dehydrators, 579 kt CO2 from flaring from tanks, and 486 kt CO2 from AGR units).
Transmission Pipeline Blowdowns
For the final 2019 Inventory estimate, in response to stakeholder feedback, EPA calculated year-specific emission
factors for transmission pipeline blowdowns using data from the first two years of GHGRP reporting, 2016 and
2017, and applied historical emission factors to all previous time series years. EPA is considering other approaches
for future Inventories, as additional years of GHGRP data become available. EPA requests feedback on whether an
updated methodology should be applied for earlier time series years (e.g., retain current emission factors for 1990 to
1992, then use linear interpolation to calculate emission factors for years 1993 through 2015; or develop an average
factor from 2016 through 2018 GHGRP data to apply for 1990 through 2015).
Well-Related Activity Data
As described in the Recalculations Discussion EPA lias updated the emission factors for several well-related
emission sources, including testing, completions, and workovers. EPA will continue to assess available data,
including data from the GHGRP and stakeholder feedback on considerations, to improve activity estimates for
sources that rely on well-related activity data. For example, EPA will seek information on other data sets that might
inform estimates of non-hydraulically fractured gas well completions and workovers.
Offshore Platforms
EPA is considering updates to the offshore platform emissions calculation methodology, as discussed in the 2018
Other Updates Memo. The current emission factors were based on data from the 2011 DOI/Bureau of Ocean Energy
Management's (BOEM) dataset, and 2014 BOEM data are available. A different source for platform counts is also
being considered.
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.
3-98 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
EPA continues to track studies that contain data that may be used to update the Inventory. Key studies in progress
include: DOE-funded work on vintage and new plastic pipelines (distribution segment), industrial meters
(distribution segment), and sources within the gathering and storage segments92; an API field study on pneumatic
controllers; a Pipeline Research Council International (PRCI) project in which researchers are gathering and
analyzing subpart W data on transmission compressor stations and underground storage facilities; and other studies
by research groups that will examine gathering and boosting emissions and offshore platform emissions. 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, stakeholder comments have highlighted areas where additional data that could inform the Inventory are
currently limited or unavailable:
• Tank malfunction and control efficiency data.
• Consider updating engine emission factors, including using subpart W data to the extent possible, and
considering whether and how to represent differences between rich- and lean-burn engines.
• 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 CH4 and CO2. Wells that are plugged have much lower
average emissions than wells that are unplugged (less than 1 kg CH4 per well per year, versus over 100 kg CH4 per
well per year). Around 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 224 kt CHi and 5 kt CO2 in 2017. Emissions of both gases
decreased by 1 percent from 1990, while the total population of abandoned oil wells increased 26 percent. Emissions
of both gases decreased by 4 percent between 2016 and 2017 as a result of well plugging activities.
Abandoned gas wells. Abandoned gas wells emitted 53 kt CHi and 2 kt CO2 in 2017. Emissions of both gases
increased by 47 percent from 1990, as the total population of abandoned gas wells increased 75 percent. Emissions
of both gases decreased by 4 percent between 2016 and 2017 as a result of well plugging activities.
92 See .
Energy 3-99
-------
Table 3-96: ChU Emissions from Abandoned Oil and Gas Wells (MMT CO2 Eq.)
Activity 1990
2005
2013 2014 2015 2016 2017
Abandoned Oil Wells 5.7
Abandoned Gas Wells 0.9
5.9
1.1
5.8 5.8 5.8 5.8 5.6
1.2 1.3 1.3 1.4 1.3
Total 6.6
6.9
7.0 7.1 7.1 7.2 6.9
Note: Totals may not sum due to independent rounding.
Table 3-97: ChU Emissions from Abandoned Oil and Gas Wells (kt)
Activity 1990
2005
2013
2014
2015
2016
2017
Abandoned Oil Wells 226
235
232
232
232
234
224
Abandoned Gas Wells 36
42
50
52
53
55
53
Total 262
277
282
283
285
289
277
Note: Totals may not sum due to independent rounding.
Table 3-98: CO2 Emissions from Abandoned Oil and Gas Wells (MMT CO2)
Activity 1990
2005
2013 2014 2015 2016 2017
Abandoned Oil Wells +
Abandoned Gas Wells +
+
+
+ + + + +
+ + + + +
Total +
+
+
+
+
+
+
+ Does not exceed 0.05 MMT CO2.
Table 3-99: CO2 Emissions from Abandoned Oil and Gas Wells (kt)
Activity 1990
2005
2013 2014 2015 2016 2017
Abandoned Oil Wells 5
Abandoned Gas Wells 2
5
2
5 5 5 5 5
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 CH4 emission factors using data from Kang et al. (2016) and Townsend-Small et al.
(2016). Plugged and unplugged abandoned well CH4 emission factors were developed at the national-level (emission
data from Townsend-Small et al.) and for the Appalachia region (using emission data from measurements in
Pennsylvania and Ohio conducted by Kang et al. and Townsend-Small et al., respectively). The Appalachia region
emissions factors were applied to abandoned wells in states in the Appalachian basin region and the national-level
emission factors were applied to all other abandoned wells.
EPA developed abandoned well CO2 emission factors using the CH4 emission factors and an assumed ratio of CO2-
to-CH4 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 CH4 and CO2 gas content
values (GRI/EPA 1996, GTI2001) 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 2017), 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
3-100 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
data to calculate the fraction of abandoned wells plugged in 2016 and 2017 (31 percent and 34 percent,
respectively), and applying linear interpolation between the 1950 value and 2016 value to calculate plugged fraction
for intermediate years. Seethe 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.93
Abandoned Oil Wells
Table 3-100: Abandoned Oil Wells Activity Data, ChU and CO2 Emissions (Metric Tons)
Source
1990
2005
2013
2014
2015
2016
2017
Plugged abandoned oil wells
386,145
616,421
741,135
758,150
778,912
800,330
871,069
Unplugged abandoned oil
wells
1,682,514
1,785,249
1,779,764
1,780,330
1,788,961
1,798,176
1,727,437
Total Abandoned Oil Wells
2,068,659
2,401,670
2,520,900
2,538,480
2,567,873
2,598,506
2,598,506
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)
317
476
552
563
575
591
643
CH4 from unplugged
abandoned oil wells (MT)
225,944
234,654
231,228
230,964
231,744
232,937
223,774
Total CH4 from Abandoned
oil wells (MT)
226,261
235,129
231,781
231,526
232,319
233,529
224,417
Total CO2 from Abandoned
oil wells (MT)
4,591
4,771
4,703
4,698
4,714
4,739
4,554
Abandoned Gas Wells
Table 3-101: Abandoned Gas Wells Activity Data, ChU and CO2 Emissions (Metric Tons)
Source
1990
2005
2013
2014
2015
2016
2017
Plugged abandoned gas wells
59,627
103,856
145,970
154,171
161,814
172,296
187,525
Unplugged abandoned gas
wells
259,807
300,784
350,532
362,033
371,645
387,115
371,886
Total Abandoned Gas Wells
319,434
404,640
496,501
516,203
533,459
559,411
559,411
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%
CH4 from plugged abandoned
gas wells (MT)
53
96
139
147
154
164
179
CH4 from unplugged
abandoned gas wells (MT)
35,899
42,258
49,681
51,367
52,788
54,985
52,822
Total CH4 from abandoned
gas wells (MT)
35,952
42,354
49,819
51,513
52,942
55,150
53,001
Total CO2 from abandoned
gas wells (MT)
1,576
1,856
2,183
2,258
2,320
2,417
2,323
Uncertainty and Time-Series Consistency
To characterize uncertainty surrounding estimates of abandoned well emissions, EPA conducted a quantitative
uncertainty analysis using the IPCC Approach 2 methodology (Monte Carlo simulation technique). See the 2018
Abandoned Wells Memo for details of the uncertainty analysis methods. EPA used Microsoft Excel's VvRISK add-in
93 See
Energy 3-101
-------
tool to estimate the 95 percent confidence bound around total methane emissions from abandoned oil and gas wells
in year 2017, then applied the calculated bounds to both CH4 and CO2 emissions estimates for each population. The
VvRISK add-in provides for the specification of probability density functions (PDFs) for key variables within a
computational structure that mirrors the calculation of the inventory estimate. EPA used measurement data from the
Kang et al. (2016) and Townsend-Small et al. (2016) studies to characterize the CH4 emission factor PDFs. For
activity data inputs (e.g., total count of abandoned wells, split between plugged and unplugged), EPA assigned
default uncertainty bounds of +/-10 percent based on expert judgment.
The IPCC guidance notes that in using this method, "some uncertainties that are not addressed by statistical means
may exist, including those arising from omissions or double counting, or other conceptual errors, or from incomplete
understanding of the processes that may lead to inaccuracies in estimates developed from models." As a result, the
understanding of the uncertainty of emission estimates for this category evolves and improves as the underlying
methodologies and datasets improve. The uncertainty bounds reported below only reflect those uncertainties that
EPA lias been able to quantify and do not incorporate considerations such as modeling uncertainty, data
representativeness, measurement errors, misreporting or misclassification.
The results presented below in Table 3-102 provide the 95 percent confidence bound within which actual emissions
from abandoned oil and gas wells are likely to fall for the year 2017, using the recommended IPCC methodology.
Abandoned oil well CH4 emissions in 2017 were estimated to be between 1.0 and 17.9 MMT CO2 Eq., while
abandoned gas well CH4 emissions were estimated to be between 0.2 and 4.2 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.
Table 3-102: Approach 2 Quantitative Uncertainty Estimates for ChU and CO2 Emissions from
Petroleum and Natural Gas Systems (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)b
(MMT CO2 Eq.)
(%)
Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Abandoned Oil Wells
CH4
5.6
1.0
17.9
-83% +219%
Abandoned Gas Wells
ch4
1.3
0.2
4.2
-83% +219%
Abandoned Oil Wells
C02
0.005
0.001
0.015
-83% +219%
Abandoned Gas Wells
C02
0.002
0.0004
0.007
-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 2017.
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
2017 by summing an estimate of total abandoned wells not included in recent databases, to an annual estimate of
abandoned wells in the Drillinglnfo data set (with year 2016 estimates used as surrogates for year 2017 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 and 2017 (31 percent and 34 percent, respectively), and applying linear
interpolation between the 1950 value and 2016 value to calculate plugged fraction for intermediate years. The same
emission factors were applied to the corresponding categories for each year of the time series.
QA/QC and Verification Discussion
The emission estimates in the Inventory are continually being reviewed and assessed to determine whether emission
factors and activity factors accurately reflect current industry practices. A QA/QC analysis was performed for data
gathering and input, documentation, and calculation. QA/QC checks are consistently conducted to minimize human
error in the model calculations. EPA performs a thorough review of information associated with new studies to
assess whether the assumptions in the Inventory are consistent with industry practices and whether new data is
available that could be considered for updates to the estimates. As in previous years, EPA conducted early
3-102 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
engagement and communication with stakeholders on updates prior to public review. EPA held a stakeholder
workshop on greenhouse gas data for oil and gas in October of 2018, and webinars in June of 2018 and February 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 2016 as
processed from Drillinglnfo, which led EPA to recalculate the estimate of historical wells not included in the
Drillinglnfo data set (which decreased from 1,149,618 to 1,108,648 historical wells not included in Drillinglnfo).
Compared with the previous Inventory, counts of abandoned oil and gas wells are on average 1.0 percent and 0.2
percent, respectively, higher over 1990 to 2016. The impact was largest in recent years, with abandoned oil and gas
well counts recalculated to be 1.5 percent and 1.6 percent, respectively, higher for 2016 comparing the previous
Inventory values to the current Inventory values; this change is also due to the use of year-specific data for year
2016 (as the previous Inventory used year 2015 estimates as surrogate for year 2016 per the established
methodology described above).
Planned Improvements
The abandoned wells source was added to the previous (i.e., 1990 through 2016) Inventory in 2018. Through EPA's
stakeholder process on oil and gas in the development of the 2018 Inventory, EPA received stakeholder feedback on
the abandoned wells update to the Inventory. EPA will continue to assess new data and stakeholder feedback on
considerations (such as the disaggregation of the well population into Appalachia and other regions, 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 UNFCCC94 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. Total
emissions of NOx, CO, and NMVOCs from energy-related activities from 1990 to 2017 are reported in Table 3-103.
Sulfur dioxide emissions are presented in Section 2.3 of the Trends chapter and Annex 6.3.
Table 3-103: NOx, CO, and NMVOC Emissions from Energy-Related Activities (kt)
Gas/Activity
1990
2005
2013
2014
2015
2016
2017
NOx
21,106
16,602
10,740
10,204
9,529
9,042
8,559
Mobile Fossil Fuel Combustion
10,862
10,295
6,523
6,138
5,740
5,413
5,051
Stationary Fossil Fuel Combustion
10,023
5,858
3,487
3,319
3,042
2,882
2,761
Oil and Gas Activities
139
321
641
650
650
650
650
Waste Combustion
82
128
89
97
97
97
97
International Bunker Fuels"
1,956
1,704
1,139
1,139
1,226
1,322
1,323
94 See .
Energy 3-103
-------
CO
125,640
64,985
41,519
40,234
39,258
36,885
35,211
Mobile Fossil Fuel Combustion
119,360
58,615
35,525
34,135
33,159
30,786
29,112
Stationary Fossil Fuel Combustion
5,000
4,648
3,847
3,686
3,686
3,686
3,686
Waste Combustion
978
1,403
1,518
1,776
1,776
1,776
1,776
Oil and Gas Activities
302
318
628
637
637
637
637
International Bunker Fuels"
103
133
129
135
141
146
152
NMVOCs
12,620
7,191
7,419
7,247
7,082
6,835
6,629
Mobile Fossil Fuel Combustion
10,932
5,724
4,023
3,754
3,589
3,342
3,137
Oil and Gas Activities
554
510
2,741
2,853
2,853
2,853
2,853
Stationary Fossil Fuel Combustion
912
716
532
497
497
497
497
Waste Combustion
222
241
122
143
143
143
143
International Bunker Fuels"
57
54
41
42
47
50
51
a These values are presented for informational purposes only and are not included in totals.
Note: Totals may not sum due to independent rounding.
Methodology
Emission estimates for 1990 through 2017 were obtained from data published on the National Emission Inventory
(NEI) Air Pollutant Emission Trends web site (EPA 2018), 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 2017. 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.95 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
95 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).
3-104 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
depart from their ports with fuel purchased within national boundaries and are engaged in international transport
separately from national totals (IPCC 2006).96
Two transport modes are addressed under the IPCC definition of international bunker fuels: aviation and marine.97
Greenhouse gases emitted from the combustion of international bunker fuels, like other fossil fuels, include CO2,
CH4 and N20 for marine transport modes, and CO2 and N20 for aviation transport modes. Emissions from ground
transport activities—by road vehicles and trains—even when crossing international borders are allocated to the
country where the fuel was loaded into the vehicle and, therefore, are not counted as bunker fuel emissions.
The 2006 IPCC Guidelines distinguish between different modes of air traffic. 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
aviation is kerosene-type jet fuel, while the typical fuel used for general aviation is aviation gasoline.98
Emissions of CO2 from aircraft are essentially a function of fuel use. Nitrous oxide emissions also depend upon
engine characteristics, flight conditions, and flight phase (i.e., take-off, climb, cruise, decent, and landing). Recent
data suggest that little or no CH4 is emitted by modern engines (Anderson et al. 2011), and as a result, CH4
emissions from this category are considered zero. In jet engines, N20 is primarily produced by the oxidation of
atmospheric nitrogen and the majority of emissions occur during the cruise phase. International marine bunkers
comprise emissions from fuels burned by ocean-going ships of all flags that are engaged in international transport.
Ocean-going ships are generally classified as cargo and passenger carrying, military (i.e., U.S. Navy), fishing, and
miscellaneous support ships (e.g., tugboats). For the purpose of estimating greenhouse gas emissions, international
bunker fuels are solely related to cargo and passenger carrying vessels, which is the largest of the four categories,
and military vessels. Two main types of fuels are used on sea-going 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 2017 from the combustion of international bunker fuels from both
aviation and marine activities were 121.2 MMT CO2 Eq., or 16.0 percent above emissions in 1990 (see Table 3-104
and Table 3-105). Emissions from international flights and international shipping voyages departing from the United
States have increased by 104.3 percent and decreased by 35.3 percent, respectively, since 1990. The majority of
these emissions were in the form of CO2; however, small amounts of CH4 (from marine transport modes) and N20
were also emitted.
Table 3-104: CO2, ChU, and N2O Emissions from International Bunker Fuels (MMT CO2 Eq.)
Gas/Mode
1990
2005
2013
2014
2015
2016
2017
CO2
103.5
113.1
99.8
103.4
110.9
116.6
120.1
Aviation
38.0
60.1
65.7
69.6
71.9
74.1
111
Commercial
30.0
55.6
62.8
66.3
68.6
70.8
74.5
Military
8.1
4.5
2.9
3.3
3.3
3.3
3.2
Marine
65.4
53.0
34.1
33.8
38.9
42.5
42.4
CH4
0.2
0.1
0.1
0.1
0.1
0.1
0.1
Aviation3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Marine
0.2
0.1
0.1
0.1
0.1
0.1
0.1
N2O
0.9
1.0
0.9
0.9
0.9
1.0
1.0
Aviation
0.4
0.6
0.6
0.7
0.7
0.7
0.7
96 Note that the definition of international bunker fuels used by the UNFCCC differs from that used by the International Civil
Aviation Organization.
97 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).
98 Naphtha-type jet fuel was used in the past by the military in turbojet and turboprop aircraft engines.
Energy 3-105
-------
Marine
0.5
0.4
0.2
0.2
0.3
0.3
0.3
Total
104.5
114.2
100.7
104.4
111.9
117.7
121.2
a CH4 emissions from aviation are estimated to be zero.
Notes: Totals may not sum due to independent rounding. Includes aircraft cruise altitude emissions.
Table 3-105: CO2, ChU, and N2O Emissions from International Bunker Fuels (kt)
Gas/Mode
1990
2005
2013
2014
2015
2016
2017
CO2
103,463
113,139
99,763
103,400
110,887
116,594
120,107
Aviation
38,034
60,125
65,664
69,609
71,942
74,059
77,696
Marine
65,429
53,014
34,099
33,791
38,946
42,535
42,412
CH4
7
5
3
3
3
4
4
Aviation3
0
0
0
0
0
0
0
Marine
7
5
3
3
3
4
4
N2O
3
3
3
3
3
3
3
Aviation
1
2
2
2
2
2
3
Marine
2
1
1
1
1
1
1
a CH4 emissions from aviation are estimated to be zero.
Notes: Totals may not sum due to independent rounding. Includes aircraft cruise altitude emissions.
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 CH4 and N20 were calculated by multiplying emission factors by measures of fuel
consumption by fuel type and mode. Emission factors used in the calculations of CH4 and N20 emissions were
obtained from the 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 N20 (IPCC 2006). For marine vessels
consuming either distillate diesel or residual fuel oil the following values (g/MJ), were employed: 0.32 for CH4 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 2017 as modeled with the Aviation Enviromnental 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 2017 were obtained 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 2017 was not possible for 1991 through 1999 because the radar dataset was not available for years
3-106 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 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 Enviromnent), 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 2018). Together,
the data allow the quantity of fuel used in military international operations to be estimated. Densities for each jet
fuel type were obtained from a report from the U.S. Air Force (USAF 1998). Final jet fuel consumption estimates
are presented in Table 3-106. See Annex 3.8 for additional discussion of military data.
In order to quantify the civilian international component of 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 2018) for 1990 through 2001, 2007 through 2017, 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 (2018). The total amount
of fuel provided to naval vessels was reduced by 21 percent to account for fuel used while the vessels were not-
underway (i.e., in port). Data on the percentage of steaming hours underway versus not-underway were provided by
the U.S. Navy. These fuel consumption estimates are presented in Table 3-107.
Table 3-106: Aviation Jet Fuel Consumption for International Transport (Million Gallons)
Nationality 1990
2005
2013 2014 2015 2016 2017
U.S. and Foreign Carriers 3,222
U.S. Military 862
5,983
462
6,748 7,126 7,383 7,610 8,011
294 339 341 333 326
Total 4,084
6,445
7,042 7,465 7,725 7,943 8,338
Note: Totals may not sum due to independent rounding.
Table 3-107: Marine Fuel Consumption for International Transport (Million Gallons)
Fuel Type 1990
2005
2013 2014 2015 2016 2017
Residual Fuel Oil 4,781
Distillate Diesel Fuel & Other 617
U.S. Military Naval Fuels 522
3,881
444
471
2,537 2,466 2,718 3,011 2,975
235 261 492 534 568
308 331 326 314 307
Total 5,920
4,796
3,081 3,058 3,536 3,858 3,850
Note: Totals may not sum due to independent rounding.
Uncertainty and Time-Series Consistency
Emission estimates related to the consumption of international bunker fuels are subject to the same uncertainties as
those from domestic aviation and marine mobile combustion emissions; however, additional uncertainties result
from the difficulty in collecting accurate fuel consumption activity data for international transport activities separate
Energy 3-107
-------
from domestic transport activities." For example, smaller aircraft on shorter routes often carry sufficient fuel to
complete several flight segments without refueling in order to minimize time spent at the airport gate or take
advantage of lower fuel prices at particular airports. This practice, called tankering, when done on international
flights, complicates the use of fuel sales data for estimating bunker fuel emissions. Tankering is less common with
the type of large, long-range aircraft that make many international flights from the United States, however. Similar
practices occur in the marine shipping industry where fuel costs represent a significant portion of overall operating
costs and fuel prices vary from port to port, leading to some tankering from ports with low fuel costs.
Uncertainties exist with regard to the total fuel used by military aircraft and ships, 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 CO2.100
There is also concern regarding the reliability of the existing DOC (1991 through 2018) 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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
99 See uncertainty discussions under Carbon Dioxide Emissions from Fossil Fuel Combustion.
100 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. The estimates in Mobile Combustion are also likely to include emissions from
ocean-going vessels departing from U.S. ports on international voyages.
3-108 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 N20 emissions from international bunker fuels in the United States. Emission totals for the different
sectors and fuels were compared and trends were investigated. No corrective actions were necessary.
Planned Improvements
The feasibility of including data from a broader range of domestic and international sources for bunker fuels is being
considered.
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 N20 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 (C) 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 2017, total CO2 emissions from the burning of woody biomass in the industrial, residential, commercial, and
electric power sectors were approximately 221.4 MMT CO2 Eq. (221,432 kt) (see Table 3-108 and Table 3-109). As
the largest consumer of woody biomass, the industrial sector was responsible for 65.3 percent of the CO2 emissions
from this source. The residential sector was the second largest emitter, constituting 20.2 percent of the total, while
the commercial and electric power sectors accounted for the remainder.
Table 3-108: CO2 Emissions from Wood Consumption by End-Use Sector (MMT CO2 Eq.)
End-Use Sector
1990
2005
2013
2014
2015
2016
2017
Industrial
135.3
136.3
139.8
140.3
138.5
138.3
144.5
Residential
59.8
44.3
58.9
59.7
52.9
46.2
44.6
Commercial
6.8
7.2
7.2
7.9
8.2
8.6
8.6
Electric Power
13.3
19.1
21.4
25.9
25.1
23.1
23.6
Total
215.2
206.9
227.3
233.8
224.7
216.3
221.4
Note: Totals may not sum due to independent rounding.
ibie 3-109: CO2 Emissions from Wood Consumption by End-Use Sector (kt)
End-Use Sector
1990
2005
2013
2014
2015
2016
2017
Industrial
135,348
136,269
139,769
140,331
138,537
138,339
144,502
Residential
59,808
44,340
58,947
59,657
52,872
46,180
44,649
Commercial
6,779
7,218
7,235
7,867
8,176
8,635
8,634
Energy 3-109
-------
Electric Power
13,252
19,074
21,389
25,908
25,146
23,140
23,647
Total
215,186
206,901
227,340
233,762
224,730
216,293
221,432
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 2017, the United States transportation sector consumed an estimated 1,135.2 trillion Btu of ethanol (95 percent of
total), and as a result, produced approximately 77.7 MMT CO2 Eq. (77,712 kt) (see Table 3-110 and Table 3-111) of
CO2 emissions. Smaller quantities of ethanol were also used in the industrial and commercial sectors. Ethanol fuel
production and consumption lias grown significantly since 1990 due to the favorable economics of blending ethanol
into gasoline and federal policies that have encouraged use of renewable fuels.
Table 3-110: CO2 Emissions from Ethanol Consumption (MMT CO2 Eq.)
End-Use Sector
1990
2005
2013
2014
2015
2016
2017
Transportation3
4.1
21.6
70.5
74.0
74.2
76.9
77.7
Industrial
0.1
1.2
3.7
1.6
1.9
1.8
1.8
Commercial
0.1
0.1
0.6
0.4
2.8
2.6
2.6
Total
4.2
22.9
74.7
76.1
78.9
81.2
82.1
3 See Annex 3.2, Table A-98 for additional information on transportation consumption of these fuels.
Note: Totals may not sum due to independent rounding.
Table 3-111: CO2 Emissions from Ethanol Consumption (kt)
End-Use Sector 1990
2005
2013 2014 2015 2016 2017
Transportation3 4,059
Industrial 105
Commercial 63
21,633
1,161
149
70,522 74,006 74,187 76,903 77,712
3,665 1,647 1,931 1,789 1,801
557 422 2,816 2,558 2,575
Total 4,227
22,943
74,743 76,075 78,934 81,250 82,088
3 See Annex 3.2, Table A-98 for additional information on transportation consumption of these fuels.
Note: Totals may not sum due to independent rounding.
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 2017, the United States consumed an estimated 253.3 trillion Btu of biodiesel, and as a result, produced
approximately 18.7 MMT CO2 Eq. (18,705 kt) (see Table 3-112 and Table 3-113) of CO2 emissions. Biodiesel
production and consumption lias 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).
Table 3-112: CO2 Emissions from Biodiesel Consumption (MMT CO2 Eq.)
End-Use Sector
1990 2005 2013
2014
2015
2016
2017
Transportation3
NO ¦ 0.9 ¦ 13.5
13.3
14.1
19.6
18.7
Total
NO 0.9 13.5
13.3
14.1
19.6
18.7
NO (Not Occurring)
3 See Annex 3.2, Table A-98 for additional information on transportation consumption of these fuels.
Note: Totals may not sum due to independent rounding.
3-110 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 3-113: CO2 Emissions from Biodiesel Consumption (kt)
End-Use Sector
1990
2005
2013
2014
2015
2016
2017
Transportation3
NO
856 1
13,462
13,349
14,077
19,648
18,705
Total
NO
856
13,462
13,349
14,077
19,648
18,705
NO (Not Occurring)
a See Annex 3.2, Table A-98 for additional information on transportation consumption of these fuels.
Note: Totals may not sum due to independent rounding.
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 3-114), provided in energy units for the industrial, residential,
commercial, and electric power sectors. One heat content (16.95 MMBtu/MT wood and wood waste) was applied to
the industrial sector's consumption while the other heat content (15.43 MMBtu/MT wood and wood waste) was
applied to the consumption data for the other sectors. An EIA emission factor of 0.434 MT C/MT wood (Lindstrom
2006) was then applied to the resulting quantities of woody biomass to obtain 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 3.1 Fossil
Fuel Combustion (CRF Source Category 1 A). 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 18.7 MMT C/QBtu (EPA 2010) to adjusted ethanol
consumption estimates (see Table 3-115). The emissions from biodiesel consumption were calculated by applying
an emission factor of 20.1 MMT C/QBtu (EPA 2010) to U. S. biodiesel consumption estimates that were provided in
energy units (EIA 2019a) (see Table 3-116).101
Table 3-114: Woody Biomass Consumption by Sector (Trillion Btu)
End-Use Sector
1990
2005
2013
2014
2015
2016
2017
Industrial
1,441.9
1,451.7
1,489.0
1,495.0
1,475.9
1,473.8
1,539.4
Residential
580.0
430.0
571.7
578.5
512.7
447.8
433.0
Commercial
65.7
70.0
70.2
76.3
79.3
83.7
83.7
Electric Power
128.5
185.0
207.4
251.3
243.9
224.4
229.3
Total
2,216.2
2,136.7
2,338.3
2,401.1
2,311.8
2,229.8
2,285.5
Note: Totals may not sum due to independent rounding.
Table 3-115: Ethanol Consumption by Sector (Trillion Btu)
End-Use Sector
1990
2005
2013
2014
2015
2016
2017
Transportation
59.3
316.0
1,030.2
1,081.1
1,083.7
1,123.4
1,135.2
Industrial
1.5
17.0
53.5
24.1
28.2
26.1
26.3
Commercial
0.9
2.2
8.1
6.2
41.1
37.4
37.6
Total
61.7
335.1
1,091.8
1,111.3
1,153.1
1,186.9
1,199.1
Note: Totals may not sum due to independent rounding.
101 CO2 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.
Energy 3-111
-------
Table 3-116: Biodiesel Consumption by Sector (Trillion Btu)
End-Use Sector
1990
2005
2013
2014
2015
2016
2017
Transportation
NO
11.6
182.3
180.8
190.6
266.1
253.3
Total
NO
11.ft
182.3
180.8
190.6
266.1
253.3
NO (Not Occurring)
Note: Totals may not sum due to independent rounding.
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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
Recalculations Discussion
EIA updated wood biomass consumption statistics in the residential sector from 2009 to 2016, and the commercial
sector from 2014 to 2016 (EIA 2019a). Ethanol consumption was reallocated across the Transportation, Industrial,
and Commercial sectors to match motor gasoline's sectoral distribution used to estimate fossil fuel combustion
emissions based on a bottom-up analysis of transportation fuel consumption. Revisions to wood biomass
consumption resulted in an average annual increase of 1.2 MMT CO2 Eq. (0.6 percent) in CO2 emissions from wood
consumption for the period 1990 through 2016, relative to the previous Inventory.
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.
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.102 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
102 See .
3-112 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.103
Currently emission estimates frombiomass and biomass-based fuels included in this Inventory are limited to woody
biomass, ethanol, and biodiesel. Other 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 EIA's Monthly Energy Review (EIA 2019a), whereas non-CCh 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 2018) 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 N20 from Stationary Combustion). EPA will
continue to investigate this discrepancy in order to apply a consistent approach to both CO2 and non-CCh emission
calculations for woody biomass consumption in the electric power sector.
103 See .
Energy 3-113
-------
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 (N20), and fluorinated greenhouse gases (e.g., HFC-23). The
greenhouse gas byproduct generating processes included in this chapter include iron and steel production and
metallurgical coke production, cement production, 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 (SF6), and nitrogen trifluoride
(NF3). The present contribution of HFCs, PFCs, SF6, and NF3 gases to the radiative forcing effect of all
anthropogenic greenhouse gases is small; however, because of their extremely long lifetimes, many of them will
continue to 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 2017, IPPU generated emissions of 358.9 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 166.9 MMT
CO2 Eq. (166,872 kt CO2) in 2017, or 3.2 percent of total U.S. CO2 emissions. Methane emissions from industrial
processes resulted in emissions of approximately 0.3 MMT CO2 Eq. (11 kt CH4) in 2017, which was less than 1
percent of U.S. CH4 emissions. Nitrous oxide emissions from IPPU were 22.6 MMT CO2 Eq. (76 kt N2O) in 2017,
or 6.3 percent of total U.S. N20 emissions. In 2017 combined emissions of HFCs, PFCs, SF6, and NF3 totaled 169.1
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
-------
MMT CO2 Eq. Total emissions from IPPU in 2017 were 4.9 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: 2017 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
Other Process Uses of Carbonates
Nitric Acid Production
Adipic Acid Production
HCFC-22 Production
Urea Consumption for Non-Agricultural Purposes
Semiconductor Manufacture
Carbon Dioxide Consumption
Electrical Transmission and Distribution
N2O from Product Uses
Aluminum Production
Ferroalloy Production
Soda Ash Production
Titanium Dioxide Production
Caprolactam, Glyoxal, and Glyoxylic Acid Production
Glass Production
Magnesium Production and Processing
Phosphoric Acid Production
Zinc Production
Lead Production
Silicon Carbide Production and Consumption
153
I
¦
¦
¦
¦
I
I
I
I
I
< 0.5
< 0.5
Industrial Processes and
Product Use as a Portion of
All Emissions
5.6%
10
20
30
40
50
60
70
MMT CO2 Eq.
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 CH4 emissions from many chemical production sources have either decreased or not changed significantly
since 1990, with the exception of petrochemical production which lias 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 generally to: mineral products,
chemical production metal production and emissions from the uses of HFCs, PFCs, SF6, and NF3.
2 See .
4-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-1: Emissions from Industrial Processes and Product Use (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
208.8
191.7
173.1
179.2
173.0
164.5
166.9
Iron and Steel Production &
Metallurgical Coke Production
101.6
68.2
53.5
58.4
47.8
42.3
41.8
Iron and Steel Production
99.1
66.2
51.6
56.3
45.0
41.0
41.2
Metallurgical Coke Production
2.5
2.1
1.8
2.0
2.8
1.3
0.6
Cement Production
33.5
46.2
36.4
39.4
39.9
39.4
40.3
Petrochemical Production
21.2
26.8
26.4
26.5
28.1
28.1
28.2
Ammonia Production
13.0
9.2
9.5
9.4
10.6
10.8
13.2
Lime Production
11.7
14.6
14.0
14.2
13.3
12.9
13.1
Other Process Uses of Carbonates
6.3
7.6
11.5
13.0
12.2
11.0
10.1
Urea Consumption for Non-
Agricultural Purposes
3.8
3.7
4.6
1.8
4.6
5.1
5.0
Carbon Dioxide Consumption
1.5
1.4
4.2
4.5
4.5
4.5
4.5
Ferroalloy Production
2.2
1.4
1.8
1.9
2.0
1.8
2.0
Soda Ash Production
1.4
1.7
1.7
1.7
1.7
1.7
1.8
Titanium Dioxide Production
1.2
1.8
1.7
1.7
1.6
1.7
1.7
Glass Production
1.5
1.9
1.3
1.3
1.3
1.2
1.3
Aluminum Production
6.8
4.1
3.3
2.8
2.8
1.3
1.2
Phosphoric Acid Production
1.5
1.3
1.1
1.0
1.0
1.0
1.0
Zinc Production
0.6
1.0
1.4
1.0
0.9
0.9
1.0
Lead Production
0.5
0.6
0.5
0.5
0.5
0.5
0.5
Silicon 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
0.1
0.2
0.2
0.3
0.3
Petrochemical Production
0.2
0.1
0.1
0.1
0.2
0.2
0.3
Ferroalloy Production
+
+
+
+
+
+
+
Silicon Carbide Production and
Consumption
+
+
+
+
+
+
+
Iron and Steel Production &
Metallurgical Coke Production
+
+
+
+
+
+
+
Iron and Steel Production
+
+
+
+
+
+
+
Metallurgical Coke Production
0.0
0.0
0.0
0.0
0.0
0.0
0.0
N2O
33.3
24.9
21.0
22.8
22.3
23.6
22.6
Nitric Acid Production
12.1
11.3
10.7
10.9
11.6
10.1
9.3
Adipic Acid Production
15.2
7.1
3.9
5.4
4.3
7.0
7.4
N2O 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
2.0
2.0
2.0
1.4
Semiconductor Manufacture
+
0.1
0.2
0.2
0.2
0.2
0.2
HFCs
46.6
122.3
146.1
150.7
153.8
155.0
158.3
Substitution of Ozone Depleting
Substances3
0.3
102.1
141.7
145.2
149.2
151.7
152.7
HCFC-22 Production
46.1
20.0
4.1
5.0
4.3
2.8
5.2
Semiconductor Manufacture
0.2
0.2
0.3
0.3
0.3
0.3
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
5.9
5.6
5.1
4.4
4.1
Semiconductor Manufacture
2.8
3.2
2.9
3.1
3.1
3.0
3.0
Aluminum Production
21.5
3.4
3.0
2.5
2.0
1.4
1.1
Substitution of Ozone Depleting
Substances
0.0
+
+
+
+
+
+
SF«
28.8
11.8
6.3
6.3
5.8
6.3
6.1
Electrical Transmission and
Distribution
23.1
8.3
4.4
4.6
4.1
4.4
4.3
Industrial Processes and Product Use 4-3
-------
Magnesium Production and
Processing
5.2
2.7
1.3
0.9
1.0
1.1
1.1
Semiconductor Manufacture
0.5
0.7
0.7
0.7
0.7
0.9
0.7
NF3
+
0.5
0.5
0.5
0.6
0.6
0.6
Semiconductor Manufacture
+
0.5
0.5
0.5
0.6
0.6
0.6
Total
342.1
358.0
353.1
365.2
360.8
354.6
358.9
+ Does not exceed 0.05 MMT CO2 Eq.
a Small amounts of PFC emissions also result from this source.
Note: Totals may not sum due to independent rounding.
Table 4-2: Emissions from Industrial Processes and Product Use (kt)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
208,834
191,654
173,073
179,191
172,963
164,524
166,872
Iron and Steel Production &
Metallurgical Coke Production
101,630
68,210
53,471
58,353
47,825
42,309
41,782
Iron and Steel Production
99,126
66,160
51,641
56,332
44,981
40,983
41,204
Metallurgical Coke Production
2,504
2,050
1,830
2,020
2,843
1,327
578
Cement Production
33,484
46,194
36,369
39,439
39,907
39,439
40,324
Petrochemical Production
21,222
26,810
26,395
26,496
28,062
28,110
28,225
Ammonia Production
13,047
9,196
9,480
9,377
10,634
10,838
13,216
Lime Production
11,700
14,552
14,028
14,210
13,342
12,942
13,145
Other Process Uses of Carbonates
6,297
7,644
11,524
12,954
12,182
10,969
10,139
Urea Consumption for Non-
Agricultural Purposes
3,784
3,653
4,556
1,807
4,578
5,132
4,958
Carbon Dioxide Consumption
1,472
1,375
4,188
4,471
4,471
4,471
4,471
Ferroalloy Production
2,152
1,392
1,785
1,914
1,960
1,796
1,975
Soda Ash Production
1,431
1,655
1,694
1,685
1,714
1,723
1,753
Titanium Dioxide Production
1,195
1,755
1,715
1,688
1,635
1,662
1,688
Glass Production
1,535
1,928
1,317
1,336
1,299
1,249
1,315
Aluminum Production
6,831
4,142
3,255
2,833
2,767
1,334
1,205
Phosphoric Acid Production
1,529
1,342
1,149
1,038
999
998
1,023
Zinc Production
632
1,030
1,429
956
933
925
1,009
Lead Production
516
553
546
459
473
450
455
Silicon Carbide Production and
Consumption
375
219
169
173
180
174
186
Magnesium Production and
Processing
1
3
2
2
3
3
3
CH4
12
4
4
6
9
11
11
Petrochemical Production
9
3
3
5
7
10
10
Ferroalloy Production
1
+
+
1
1
1
1
Silicon Carbide Production and
Consumption
1
+
+
+
+
+
+
Iron and Steel Production &
Metallurgical Coke Production
1
1
+
+
+
+
+
Iron and Steel Production
1
1
+
+
+
+
+
Metallurgical Coke Production
0
0
0
0
0
0
0
N2O
112
84
71
77
75
79
76
Nitric Acid Production
41
38
36
37
39
34
31
Adipic Acid Production
51
24
13
18
14
23
25
N2O from Product Uses
14
14
14
14
14
14
14
Caprolactam, Glyoxal, and
Glyoxylic Acid Production
6
7
7
7
7
7
5
Semiconductor Manufacture
+
+
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
4-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
HCFC-22 Production
3
1
+
+
+
+
+
Semiconductor Manufacture
M
M
M
M
M
M
M
Magnesium Production and
Processing
0
0
+
+
+
+
+
PFCs
M
M
M
M
M
M
M
Semiconductor Manufacture
M
M
M
M
M
M
M
Aluminum Production
M
M
M
M
M
M
M
Substitution of Ozone Depleting
Substances
0
+
+
+
+
+
+
SF«
1
1
+
+
+
+
+
Electrical Transmission and
Distribution
1
+
+
+
+
+
+
Magnesium Production and
Processing
+
+
+
+
+
+
+
Semiconductor Manufacture
+
+
+
+
+
+
+
NF3
+
+
+
+
+
+
+
Semiconductor Manufacture
+
+
+
+
+
+
+
+ Does not exceed 0.5 kt.
M (Mixture of gases)
a Small amounts of PFC emissions also result from this source.
Note: Totals may not sum due to independent rounding.
The UNFCCC incorporated the 2006IPCC Guidelines for National Greenhouse Gas Inventories (2006IPCC
Guidelines) as the standard for Annex I countries at the Nineteenth Conference of the Parties (Warsaw, November
11-23, 2013). This chapter presents emission estimates calculated in accordance with the methodological guidance
provided in these guidelines. For additional detail on IPPU sources that are not estimated in this Inventory report,
please review Annex 5, Assessment of the Sources and Sinks of Greenhouse Gas Emissions Not Included. These
sources are not estimated due to various national circumstances, such as that emissions from a source may not
currently occur in the United States, data are not currently available for those emission sources (e.g., ceramics, non-
metallurgical magnesium production, glyoxal and glyoxylic acid production CH4 from direct reduced iron
production), emissions are included elsewhere within the Inventory report, or data suggest that emissions are not
significant (e.g., various fluorinated gas emissions from the electronics industry and other produce uses).
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 3-6), 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 2006 IPCC 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 (C) 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 1 A)) 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.
Industrial Processes and Product Use 4-5
-------
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
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.
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. Geologic 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.
The uncertainty analysis performed to quantify uncertainties associated with the 2017 emission estimates from IPPU
continues a multi-year process for developing credible quantitative uncertainty estimates for these source categories
using the IPCC Tier 2 approach. As the process continues, the type and the characteristics of the actual probability
density functions underlying the input variables are identified and better characterized (resulting in development of
more reliable inputs for the model, including accurate characterization of correlation between variables), based
primarily on expert judgment. Accordingly, the quantitative uncertainty estimates reported in this section should be
considered illustrative and as iterations of ongoing efforts to produce accurate uncertainty estimates. The correlation
among data used for estimating emissions for different sources can influence the uncertainty analysis of each
individual source. While the uncertainty analysis recognizes very significant connections among sources, a more
comprehensive approach that accounts for all linkages will be identified as the uncertainty analysis moves forward.
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
4-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 do 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.
Box 4-2: Industrial Processes 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 reported for facilities subject to 40 CFR Part 98, though some source
categories first reported 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. While many
methodologies used in EPA's GHGRP are consistent with IPCC, 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. Further information on
the reporting categorizations in EPA's GHGRP and specific data caveats associated with monitoring methods in
EPA's GHGRP lias been provided on the GHGRP website.3
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, lias 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.4 Specific uses of aggregated facility-level data are
described in the respective methodological sections. For other source categories in this chapter, as indicated in the
respective planned improvements sections, 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.
As stated previously in the Introduction chapter, this year EPA has integrated GHGRP information for various
Industrial Processes and Product Use categories and also identified places where EPA plans to integrate additional
3 See .
4 U.S. EPA Greenhouse Gas Reporting Program. Developments on Publication of Aggregated Greenhouse Gas Data, November
25, 2014. See .
Industrial Processes and Product Use 4-7
-------
GHGRP data in additional categories5 (see those categories' Planned Improvements sections for details). EPA has
paid particular attention to ensuring time-series consistency for major recalculations that have occurred from the
incorporation of GHGRP data into these categories, consistent with 2006IPCC Guidelines and IPCC Technical
Bulletin on Use of Facility-Specific Data in National GHG Inventories.6
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.7 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, but due to concerns regarding CBI data, specific subpart QA/QC and
verification procedures are not available to include in this Inventory report. The GHGRP dataset is a particularly
important annual resource and will continue to be important for improving emissions estimates from IPPU in future
Inventory reports. Additionally, EPA's GHGRP has and will continue to enhance QA/QC procedures and
assessment of uncertainties within the IPPU categories (see those categories for specific QA/QC details regarding
the use of GHGRP data).
4.1 Cement Production (CRF Source Category
2A1)
Cement production is an energy- and raw material-intensive process that results in the generation of carbon dioxide
(CO2) from both the energy consumed in making the clinker precursor to cement and the chemical process itself 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 is where calcium carbonate (CaCCb), 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 one of these "sintering" reactions is highly exothermic, very
little extra heat energy is required, and these sintering reactions have few process emissions of CO2. 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.8
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 Pennsylvania were the leading cement-producing states in 2017 and accounted for almost 50 percent of
5 Ammonia Production, Glass Production, Lead Production, and Other Fluorinated Gas Production.
6 See .
7 See .
8 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.
4-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
total U.S. production (USGS 2018a). Based on both GHGRP data (EPA 2018) and USGS reported data, clinker
production in 2017 increased approximately 2 percent from 2016 levels as cement sales increased modestly (less
than 2 percent) in 2017, with imports stagnant in 2017 (USGS 2018a). In 2017, 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 MMT CO2 Eq. kt
1990 33.5 33,484
2005 46.2 46,194
2013 36.4 36,369
2014 39.4 39,439
2015 39.9 39,907
2016 39.4 39,439
201 7 403 40,324
Greenhouse gas emissions from cement production increased every year from 1991 through 2006 (with the
exception of a slight decrease in 1997) but decreased in the following years until 2009. Since 1990, emissions have
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 consumption. In 2017, emissions from cement production increased by 2
percent from 2016 levels. 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 were estimated using the Tier 2 methodology from the 2006IPCC Guidelines. 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, 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 lias confirmed that this is a reasonable
assumption for the United States (VanOss 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 CO2 emissions.
At some plants, essentially all CKD is directly returned to the kiln becoming part of the raw material feed, or is
likewise returned to the kiln after first being removed from the exhaust. In either case, the returned CKD becomes a
raw material, thus forming clinker, and the associated 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. The IPCC
recommends that these additional CKD CO2 emissions should be estimated as two percent of the CO2 emissions
calculated from clinker production (when data on CKD generation are not available). Total cement production
Industrial Processes and Product Use 4-9
-------
emissions were calculated by adding the emissions from clinker production to the emissions assigned to CKD (IPCC
2006).
Furthermore, 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. During the 1990 through 2015 Inventory report cycle,
EPA began incorporating clinker production data from EPA's GHGRP to estimate emissions in these respective
years. Clinker production values in the current Inventory report utilize GHGRP data for the years 2014 through 2017
(EPA 2018). Details on how this change compares to USGS reported data can be found in the section on
Uncertainty and Time-Series Consistency.
Table 4-4: Clinker Production (kt)
Year Clinker
1990 64,355
2005 88,783
2013 69,900
2014 75,800
2015 76,700
2016 75,800
201 7 77,500
Notes: Clinker production from 1990 through 2017
includes Puerto Rico.
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. However, see Planned Improvements described below to reassess this
assumption by conducting a review to identify recent studies that may provide information or data on reabsorption
rates of cement products.
Total U.S. clinker production is assumed to have low uncertainty. USGS takes a number of manual steps to review
clinker production reported through their voluntary surveys. EPA also continues to review reported clinker
production data required by GHGRP Subpart H facilities for current and future Inventory reports. 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.
4-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Based on the results of the verification process, the EPA follows up with facilities to resolve mistakes that may have
occurred.9 Facilities are also required to monitor and maintain records of monthly clinker production.
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, 2017 CO2 emissions from cement production were
estimated to be between 38.0 and 42.7 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.
Table 4-5: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Cement
Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate
(MMT CO2 Eq.)
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Cement Production
CO2
40.3
38.0 42.7 -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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above. More information on the consistency in clinker production data and emissions across the time series with the
use of GHGRP clinker data for 2014 through 2017 can be found in the Uncertainty and Time-Series Consistency
section.
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 QA/QC and Verification Procedures section in the introduction of the
IPPU chapter. 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. 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 (USGS 2017, 2018a,
2018b) 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. See the Uncertainty and Time- Series
Consistency section for information on how GHGRP data are verified.
Planned Improvements
In response to prior comments from the Portland Cement Association (PCA) and UNFCCC expert technical
reviews, 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. EPA held a technical meeting with PCA
in August 2016 to review Inventory methods and available data from the GHGRP data set. 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
9 See .
Industrial Processes and Product Use 4-11
-------
recommended by the 2006IPCC Guidelines.10 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 updated for this current Inventory.
During the Public Review comment period of the current (i.e., 1990 through 2017) Inventory report, EPA received
comments on the characterization of cement and clinker production processes described in this chapter. As a result,
EPA updated the description of these processes and comparison of available datasets (QA/QC and Verification
section).
Finally, in response to feedback from PCA during the Public Review comment period of a previous Inventory in
2017, 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 later in the cement product lifecycle. 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 (CRF Source Category
2A2)
Lime is an important manufactured product with many industrial, chemical, and environmental applications. Lime
production involves three main processes: stone preparation, calcination, and hydration. Carbon dioxide (CO2) is
generated during the calcination stage, when limestone—mostly calcium carbonate (CaCCh)—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.11 Emissions from fuels consumed for energy
purposes during the production of lime are accounted 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 2018b). The major uses are in steel making, flue gas 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 18,000 kilotons in 2017 (USGS
2018a). At year-end 2017, there were 74 operating primary lime plants in the United States, including Puerto Rico.12
10 See .
11 PCC is obtained from the reaction of CO2 with calcium hydroxide. It is used as a filler and/or coating in the paper, food, and
plastic industries.
12 In 2017, 75 operating primary lime facilities in the United States reported to the EPA Greenhouse Gas Reporting Program.
4-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Principal lime producing states in descending order of production are Missouri, Alabama, Ohio, Texas, and
Kentucky (USGS 2018a).
U.S. lime production resulted in estimated net CO2 emissions of 13.1 MMT CO2 Eq. (13,145 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 CO2 Eq.
1990 11.7 11,700
2005 14.6 14,552
2013 14.0 14,028
2014 14.2 14,210
2015 13.3 13,342
2016 12.9 12,942
201 7 m 13,145
Table 4-7: Potential, Recovered, and Net CO2 Emissions from Lime Production (kt)
Year Potential Recovered3 Net Emissions
1990 11,959 259 11,700
2005 15,074 522 14,552
2013 14,495 467 14,028
2014 14,715 505 14,210
2015 13,764 422 13,342
2016 13,312 370 12,942
2017 13,546 401 13,145
a For sugar refining and PCC production.
Note: Totals may not sum due to independent rounding.
In 2017, lime production increased compared to 2016 levels (increase of about 2 percent) at 18,000 kilotons, owing
primarily to an increase in hydrated lime output (USGS 2018a; USGS 2017).
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:
[(44.01 g/mole CO2) -h (56.08 g/mole CaO)] x (0.9500 CaO/lime) = 0.7455 g CCh/g lime
For dolomitic lime:
[(88.02 g/mole C02) h- (96.39 g/mole CaO)] x (0.9500 CaO/lime) = 0.8675 gC02/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 H2O to (Ca(OH)2 and [Ca(OH)2»Mg(OH)2]) (IPCC 2006). These factors
Industrial Processes and Product Use 4-13
-------
set the chemically combined water content to 24.3 percent for high-calcium hydrated lime, and 27.2 percent for
dolomitic hydrated lime.
The 2006IPCC 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 CO2 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 CO2 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. The amount of CO2 captured/recovered for
on-site process use is deducted from the total potential 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 CO2 removals (i.e., CO2 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 2017 (see Table 4-8) were obtained from the U.S. Geological Survey
(USGS) (USGS 2017 and 2018a) 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 2017
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 2018b). 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 CaOMgO 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
2013
13,800
2,850
2,050
260
200
2014
14,100
2,740
2,190
279
200
2015
13,100
2,550
2,150
279
200
2016
12,615
2,456
2,150
279
200
2017
12,866
2,505
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
4-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
2013
15,297
3,252
2014
15,699
3,135
2015
14,670
2,945
2016
14,185
2,851
2017
14,436
2,900
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 lias many different chemical,
industrial, enviromnental, 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.13 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.14 The
lime generated by these processes is included in the USGS data for commercial lime consumption. In the pulping
industry, mostly using the Kraft (sulfate) pulping process, lime is consumed in order to causticize a process liquor
(green liquor) composed of sodium carbonate and sodium sulfide. The green liquor results from the dilution of the
smelt created by combustion of the black liquor where biogenic carbon (C) is present from the wood. Kraft mills
recover the calcium carbonate "mud" after the causticizing operation and calcine it back into lime—thereby
generating 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) lias 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
13 Representatives of the National Lime Association estimate that CO2 reabsorption that occurs from the use of lime may offset
as much as a quarter of the CO2 emissions from calcination (Males 2003).
14 Some carbide producers may also regenerate lime from their calcium hydroxide byproducts, which does not result in
emissions of CO2. 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 + 2H2O —» C2H2 + Ca(OH) 2], not calcium
carbonate [CaCCb]. Thus, the calcium hydroxide is heated in the kiln to simply expel the water [Ca(OH)2 + heat —»CaO + H2O]
and no CO2 is released.
Industrial Processes and Product Use 4-15
-------
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. EPA initiated a dialogue with NLA to discuss data
needs to generate a country-specific LKD factor and is reviewing the information provided by NLA. 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 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 2017 were estimated to be between 12.9 and 13.4 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
13.1 MMT C02 Eq.
Table 4-10: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lime
Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Bound Bound
Lower
Bound
Upper
Bound
Lime Production
CO2
13.1
12.9 13.4
-2%
+2%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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.
Planned Improvements
Future improvements involve finishing a review of data to improve 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 the Public Review comment period of the previous (1990 to
2015) Inventory. In response to comments, EPA met with NLA on April 7, 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 response to this technical meeting, in January and February 2016, NLA
compiled and shared historical emissions information reported by member facilities on an annual basis under
voluntary reporting initiatives over 2002 through 2011 associated with generation of total calcined byproducts and
LKD (LKD reporting only differentiated starting in 2010). This emissions information was reported on a voluntary
basis consistent with NLA's facility-level reporting protocol also recently provided. EPA needs additional time to
review the information provided by NLA and plans to work with them to address needs for EPA's analysis, as there
is limited information across the time series. Due to limited resources and need for additional QA of information,
this planned improvement is still in process and has not been incorporated into this current Inventory report. This is
a long-term improvement. As an interim step, EPA has updated the qualitative description of uncertainty to reflect
the information provided by NLA.
In addition, EPA plans to review GHGRP emissions and activity data reported to EPA under Subpart S, and in
particular, aggregated activity data on lime production by type. Particular attention will be made to also ensuring
time-series consistency of the emissions estimates presented in future Inventory reports, consistent with IPCC and
4-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.15
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 process itself. Emissions from fuels consumed
for energy purposes during the production of glass are accounted for 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) which 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, K20). Stabilizers are used to make glass more chemically stable and to keep the
finished glass from dissolving and/or falling apart. Commonly used stabilizing agents in glass production are
limestone (CaCCh). dolomite (CaCO;,IVIgCO;,). alumina (AI2O3), magnesia (MgO), barium carbonate (BaCCb),
strontium carbonate (SrCCh), lithium carbonate (Li2C03), and zirconia (Z1O2) (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 over 1,500 companies that manufacture glass in the United States, with the largest being
Corning, Guardian Industries, Owens-Illinois, and PPG Industries.16
In 2017, 763 kilotons of limestone and 2,360 kilotons of soda ash were consumed for glass production (USGS
2016a; USGS 2017). Dolomite consumption data for glass manufacturing was reported to be zero for 2017. Use of
limestone and soda ash in glass production resulted in aggregate CO2 emissions of 1.3 MMT CO2 Eq. (1,315 kt) (see
Table 4-11). Overall, emissions have decreased 14 percent from 1990 through 2017.
Emissions in 2017 increased approximately 5 percent from 2016 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
15 See.
16 Excerpt from Glass & Glass Product Manufacturing Industry Profile, First Research. Available online at:
.
Industrial Processes and Product Use 4-17
-------
from glass containers towards lighter and more cost-effective polyethylene terephthalate (PET) based containers,
putting downward pressure on domestic consumption of soda ash (USGS 1995 through 2015b).
Table 4-11: CO2 Emissions from Glass Production (MMT CO2 Eq. and kt)
Year
MMT CO2 Eq.
kt
1990
1.5
1,535
2005
1.9
1,928
2013
1.3
1,317
2014
1.3
1,336
2015
1.3
1,299
2016
1.3
1,249
2017
1.3
1,315
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 C'02/mctric ton carbonate): limestone, 0.43971; dolomite, 0.47732; and soda ash, 0.41492.
Consumption data for 1990 through 2017 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), 2016 and 2017 preliminary data from the USGS Crushed Stone Commodity Expert (Willett 2018a, Willett
2018b), the USGS Minerals Yearbook: Soda Ash Annual Report (1995 through 2015) (USGS 1995 through 2015b),
USGS Mineral Industry Sun'evs for Soda Ash in February 2018 (USGS 2018) 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. Consumption for 1990 was estimated by applying the
1991 percentages of total limestone and dolomite use constituted by the individual limestone and dolomite uses to
1990 total use. Similarly, the 1992 consumption figures were approximated by applying an average of the 1991 and
1993 percentages of total limestone and dolomite use constituted by the individual limestone and dolomite uses to
the 1992 total.
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.
There is 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.17 For 2017, the unspecified uses of both limestone and dolomite consumption were not available at the time of
publication, so 2016 values were used to proxy these values.
Based on the 2017 reported data, the estimated distribution of soda ash consumption for glass production compared
to total domestic soda ash consumption is 48 percent (USGS 1995 through 2015b, 2018).
Table 4-12: Limestone, Dolomite, and Soda Ash Consumption Used in Glass Production (kt)
Activity
1990
2005
2013
2014
2015
2016
2017
Limestone
4301
9201
693
765
699
472
763
17 This approach was recommended by USGS.
4-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Dolomite 59
Soda Ash 3,177
541
3,050
0 0 0 0 0
2,440 2,410 2,390 2,510 2,360
Total 3,666
4,511
3,133 3,175 3,089 2,982 3,123
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 2017, there lias 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 2017, glass
production CO2 emissions were estimated to be between 1.3 and 1.4 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
MMT CO2 Eq.
Table 4-13: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Glass
Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Glass Production
CO2
1.3
1.3 1.4 -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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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.
Planned Improvements
As noted in the previous reports, 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 lias 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.
Industrial Processes and Product Use 4-19
-------
EPA has initiated review of EPA's GHGRP data and anticipates finalizing assessment for future integration of data
in the spring of 2019. This assessment will help to 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, so the assessment will consider the completeness of carbonate
consumption data for glass production in the United States. Particular attention will also 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.18 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 (CaCCh). dolomite (CaCChMgCCh).'9 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 and dolomite 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 this section and reported under their respective source
categories (e.g., Section 4.3, Glass Production). Emissions from soda ash consumption associated with glass
manufacturing are reported under Section 4.3 Glass Production (CRF Source Category 2A3). Emissions from fuels
consumed for energy purposes during these processes are accounted for in the Energy chapter.
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
Illinois, which contributed 50 percent of the total U.S. output (USGS 2018). 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 Illinois, Pennsylvania, and New York,
which currently contribute more than half of the total U.S. output (USGS 1995a through 2017).
In 2017, 19,851 kt of limestone, 2,088 kt of dolomite, and 2,550 kt of soda ash were consumed for these emissive
applications, excluding glass manufacturing (Willett 2018a). Usage of limestone, dolomite and soda ash resulted in
aggregate CO2 emissions of 10.1 MMT CO2 Eq. (10,139 kt) (see Table 4-14 and Table 4-15). While 2017 emissions
have decreased 8 percent compared to 2016, overall emissions have increased 61 percent from 1990 through 2017.
18 See .
19 Limestone and dolomite are collectively referred to as limestone by the industry, and intermediate varieties are seldom
distinguished.
4-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
2013 2.3 6.3 0.0 1.1 1.8 11.5
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 2.6 6.2 0.0 1.1 1.1 11.0
2017 2.6 5J) 00 U 0^ 10.1
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.
Note: Totals may not sum due to independent rounding.
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
2013
2,307
6,309
0
1,109
1,798
11,524
2014
2,911
7,111
0
1,143
1,790
12,954
2015
2,901
7,335
0
1,075
871
12,182
2016
2,585
6,164
0
1,082
1,137
10,969
2017
2,645
5,904
0
1,058
532
10,139
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.
Note: Totals may not sum due to independent rounding.
Methodology
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 CO:/metric ton carbonate, and dolomite: 0.47732 metric ton
CCh/mctric ton carbonate.20 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
20 2006IPCC Guidelines, Volume 3: Chapter 2, Table 2.1.
Industrial Processes and Product Use 4-21
-------
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 2017 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 2016 and 2017 from USGS Crushed Stone Commodity Expert (Willett
2018a, 2018b), American Iron and Steel Institute limestone and dolomite consumption data (AISI 2018), and the
U.S. Bureau of Mines (1991 and 1993a), which are reported to the nearest ton. For 2017, the unspecified uses of
both limestone and dolomite consumption were not available at the time of publication, so 2016 values were used to
proxy these values. The production capacity data for 1990 through 2017 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.
Consumption for 1990 was estimated by applying the 1991 percentages of total limestone and dolomite use
constituted by the individual limestone and dolomite uses to 1990 total use. Similarly, the 1992 consumption figures
were approximated by applying an average of the 1991 and 1993 percentages of total limestone and dolomite use
constituted by the individual limestone and dolomite uses to the 1992 total.
Additionally, each year the USGS withholds data on certain limestone and dolomite end-uses due to confidentiality
agreements regarding company proprietary data. For the purposes of this analysis, emissive end-uses that contained
withheld data were estimated using one of the following techniques: (1) the value for all the withheld data points for
limestone or dolomite use was distributed evenly to all withheld end-uses; (2) the average percent of total limestone
or dolomite for the withheld end-use in the preceding and succeeding years; or (3) the average fraction of total
limestone or dolomite for the end-use over the entire time period.
There is a large quantity of crushed stone reported to the USGS under the category "unspecified uses." A portion of
this consumption is believed to be limestone or dolomite used for emissive end uses. The quantity listed for
"unspecified uses" was, therefore, allocated to each reported end-use according to each end-use's fraction of total
consumption in that year.21
Table 4-16: Limestone and Dolomite Consumption (kt)
Activity
1990
2005
2013
2014
2015
2016
2017
Flux Stone
6,737
7,022
6,345
7,599
7,834
7,092
7,302
Limestone
5,804
3,165
4,380
4,243
4,590
4,118
5,214
Dolomite
933
3,857
1,965
3,356
3,244
2,973
2,088
FGD
3,258
6,761
14,347
16,171
16,680
14,019
13,427
Other Miscellaneous Uses
1,835
1,632
3,973
4,069
1,982
2,587
1,210
Total
11,830
15,415
24,665
27,839
26,496
23,698
21,939
Once produced, most soda ash is consumed in chemical production with minor amounts 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 (see Table 4-17) for 1990 to 2017 were obtained from the U.S. Geological Survey (USGS)
Minerals Yearbook for Soda Ash (1994 through 2015b) and USGS Mineral Industry Sur\>evs for Soda Ash (USGS
2017a, 2018). Soda ash consumption data22 were collected by the USGS from voluntary surveys of the U.S. soda
ash industry.
21 This approach was recommended by USGS, the data collection agency.
22 EPA has assessed feasibility of using emissions information (including activity data) from EPA's GHGRP; 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.
4-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-17: Soda Ash Consumption Not Associated with Glass Manufacturing (kt)
Activity
1990
2005
2013
2014
2015
2016
2017
Soda Asha
3,3511
3,1441
2,674
2,754
2,592
2,608
2,550
Total
3,351
3,144
2,674
2,754
2,592
2,608
2,550
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 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 2017 were estimated to be between 9.0 and 11.8 MMT CO2 Eq.
at the 95 percent confidence level. This indicates a range of approximately 12 percent below and 15 percent above
the emission estimate of 10.1 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)
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Other Process Uses
of Carbonates
C02
10.1
9.0 11.8
-12% +15%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Industrial Processes and Product Use 4-23
-------
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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.
Planned Improvements
EPA plans to continue the 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, 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.
In the United States, the majority of ammonia is produced using a natural gas feedstock; 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 2017, there were approximately 15 companies operating 32 ammonia producing facilities in 16
states. Roughly 50 percent of domestic ammonia production capacity is concentrated in the states of Louisiana,
Oklahoma, and Texas. In 2016, upgrades came online to increase ammonia capacity at one facility in the United
States and in 2017 two new ammonia facilities became operational (USGS 2018).
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 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 hydrogen gas 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
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.
4-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 ICCKNFLhl- which lias 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 enviromnent 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 2017 were 13.2 MMT CO2 Eq. (13,216 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 1.3 percent. Emissions in 2017 have increased by
approximately 22 percent from the 2016 levels. Agricultural demands continue to drive demand for nitrogen
fertilizers (USGS 2018).
Table 4-19: CO2 Emissions from Ammonia Production (MMT CO2 Eq.)
Source
1990 2005 2013
2014
2015
2016
2017
Ammonia Production
13.0 1 9.2 I 9.5
9.4
10.6
10.8
13.2
Total
13.0 1 9.2 I 9.5
9.4
10.6
10.8
13.2
Table 4-20: CO2 Emissions from Ammonia Production (kt)
Source
1990 ¦ 2005 ¦ 2013
2014
2015
2016
2017
Ammonia Production
13,047 1 9,196 1 9,480
9,377
10,634
10,838
13,216
Total
13,047 9,196 9,480
9,377
10,634
10,838
13,216
Methodology
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 2006IPCC Guidelines (IPCC 2006) Tier 1 and 2
methods. In the country-specific approach emissions are not based on total fuel requirement per the 2006 IPCC
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 C02/metric ton NH3 (EFMA 2000a) is
applied to the percent of total annual domestic ammonia production from natural gas feedstock.
Emissions of CO2 from 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 C02/urea factor of 44/60, assuming complete conversion of ammonia (NH3) and CO2 to urea (IPCC
2006; EFMA 2000b).
Industrial Processes and Product Use 4-25
-------
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/mctric ton NH3 for the petroleum coke
feedstock process (Bark 2004) is applied to the percent of total annual domestic ammonia production from
petroleum coke feedstock.
The emission factor of 1.2 metric ton CCh/mctric 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/mctric ton NH3, with 1.2 metric ton
CCh/mctric 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.
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 CO2 from Fossil Fuel
Combustion. See the Planned Improvements section on improvements of reporting fuel and feedstock CO 2
emissions utilizing EPA's GHGRP data to improve consistency with 2006IPCC Guidelines.
The total ammonia production data for 2011 through 2017 were obtained from American Chemistry Council (ACC
2018). 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, 2013, 2014,
2015, 2016, and 2017) for 2012 through 2017. 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.
For the current Inventory (i.e., 1990 through 2017), EPA began utilizing urea production data from EPA's GHGRP
Subpart G to estimate emissions. Urea production values in the current Inventory report utilize GHGRP data for the
years 2011 through 2017 (EPA 2018). Details on QA checks and how this change compares to previously used
USGS Minerals Yearbook: Nitrogen (USGS 2014 through 2016) reported data can be found in the Uncertainty and
Time-Series Consistency section, below. GHGRP urea production data for 2017 were not yet published and so 2016
data were used as a proxy.
Table 4-21: Ammonia Production, Recovered CO2 Consumed for Urea Production, and Urea
Production (kt)
Year
Ammonia Production
Total CO2 Consumption
for Urea Production
Urea Production
1990
15,425
5,463
7,450
2005
10,143
3,865
5,270
2013
10,930
4,501
6,137
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
4-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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. In addition, due to the fact that 2017 nitrogen data lias yet to be published, 2016 is used as a proxy which
may result in greater uncertainty.
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.
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 GHGRP with existing data from USGS. This
was to ensure time-series consistency of the emission estimates presented in the Inventory. For the year 2011, USGS
and GHGRP urea production data showed a difference of approximately 27 percent in 2011, 11 percent in 2012, 12
percent in 2013, 6 percent in 2014 and 2015, and 12 percent in 2016. The use of GHGRP data (EPA 2018) resulted
in an decrease of emissions compared to USGS data (USGS 2011 through 2018) ranging from 0.24 to 1.06 MMT
CO2 Eq. for 2011 through 2016.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-22. Carbon dioxide
emissions from ammonia production in 2017 were estimated to be between 12.6 and 13.8 MMT CO2 Eq. at the 95
percent confidence level. This indicates a range of approximately 5 percent below and 5 percent above the emission
estimate of 13.2 MMT CO2 Eq.
Table 4-22: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from
Ammonia Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Ammonia Production
CO2
13.2
12.6 13.8 -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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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.
Industrial Processes and Product Use 4-27
-------
Recalculations Discussion
This current Inventory (i.e., 1990 through 2017) has been updated to include estimates of CO2 emissions from urea
consumption for the years 2011 through 2017) based on urea production data available from the GHGRP. Urea
production data available from USGS was used in estimating emissions in prior inventories.
Urea production based on USGS data ranged from a low of 5,220 kt to a high of 6,610 kt over the period from 2011
through 2017. Urea production based on GHGRP data ranged from a low of 5,561 kt to a high of 7,390 kt over the
same time period. These data show that total annual urea production based on GHGRP data averaged 12 percent
higher than the total annual urea production based on USGS data during the same time period.
The estimated CO2 emissions from ammonia production in 2011 through 2017 are approximately 5 percent lower in
the current Inventory (i.e., 1990 through 2017) than the previous Inventory (i.e., 1990 through 2016).
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),23 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 2017) 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.24 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
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
32 plants producing ammonia in the United States during 2017, with two additional plants sitting idle for the entire
year (USGS 2018).
The chemical reaction that produces urea is:
2NH3+ C02 -> NH2COONH4 -> CO(NH2)2 +h2o
23 See .
24 See .
4-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 2017 were estimated to be 5.0 MMT CO2
Eq. (4,958 kt), and are summarized in Table 4-23 and Table 4-24. Net CO2 emissions from urea consumption for
non-agricultural purposes in 2017 have increased by approximately 31 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.
Table 4-23: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (MMT CO2
Eq.)
Source 1990
2005
2013 2014 2015 2016 2017
Urea C onsumption 3.8
3.7
4.6 1.8 4.6 5.1 5.0
Total 3.8
3.7
4.6 1.8 4.6 5.1 5.0
Table 4-24: CO2 Emissions from Urea Consumption for Non-Agricultural Purposes (kt)
Source 1990
2005
2013 2014 2015 2016 2017
Urea Consumption 3,784
3,653
4,556 1,807 4,578 5,132 4,958
Total 3,784
3,653
4,556 1,807 4,578 5,132 4,958
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 enviromnent 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 5-24) 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 C02/urea factor of 44/60, assuming complete conversion of NH3 and CCMo urea (IPCC 2006; EFMA
2000).
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 (2011). The
U.S. Census Bureau ceased collection of urea production statistics in 2011. For the current 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).
Details on QA checks and how this change compares to previously used USGS Minerals Yearbook: Nitrogen
Industrial Processes and Product Use 4-29
-------
(USGS 2014 through 2016) reported data can be found in the Uncertainty and Time-Series Consistency section,
below.
Urea import data for 2017 are not yet publicly available and so 2016 data have been used as proxy. Urea import data
for 2013 to 2016 were obtained from the Minerals Yearbook: Nitrogen (USGS 2016). Urea import data for 2011 and
2012 were taken from U.S. Fertilizer Import/Exports from the United States Department of Agriculture (USD A)
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 2017 are not yet publicly available and so 2016 data have been used as proxy. Urea export data
for 2013 to 2016 were obtained from the Minerals Yearbook: Nitrogen (USGS 2016). Urea export data for 1990
through 2012 were taken from U.S. Fertilizer Import/Exports from USDA Economic Research Service Data Sets
(U.S. Department of Agriculture 2012). USDA suspended updates to this data after 2012.
Table 4-25: Urea Production, Urea Applied as Fertilizer, Urea Imports, and Urea Exports (kt)
Year Urea 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
2013
6,137
6,059
6,470
335
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
6,580
321
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.
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 GHGRP with existing data from USGS. This
was to ensure time-series consistency of the emission estimates presented in the Inventory. For the year 2011, USGS
and GHGRP urea production data showed a difference of approximately 27 percent in 2011, 11 percent in 2012, 12
percent in 2013, 6 percent in 2014 and 2015, and 12 percent in 2016. The use of GHGRP data (EPA 2018) resulted
in an increase of emissions compared to USGS data (USGS 2011 through 2018) ranging from 0.24 to 1.06 MMT
CO2 Eq. for 2011 through 2016.
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 2017 were estimated to be
between 4.4 and 5.5 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 5.0 MMT CO2 Eq.
4-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Upper
Lower
Upper
Bound
Bound
Bound
Bound
Urea Consumption
for Non-Agricultural CO2
5.0
4.4
5.5
-12%
+12%
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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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
This current Inventory (i.e., 1990 through 2017) lias been updated to include estimates of CO2 emissions from urea
consumption for the years 2011 through 2017 based on urea production data available from the GHGRP. Urea
production data available from USGS was used in estimating emissions in prior inventories.
Urea production based on USGS data ranged from a low of 5,220 kt to a high of 6,610 kt over the period from 2011
through 2017. Urea production based on GHGRP data ranged from a low of 5,561 kt to a high of 7,390 kt over the
same time period. These data show that total annual urea production based on GHGRP data averaged 12 percent
higher than the total annual urea production based on USGS data during the same time period.
Taking into account the fluctuations in urea application for agricultural uses, urea imports, and urea exports, the
estimated CO2 emissions from urea consumption in 2011 through 2017 are approximately 13 percent higher in the
current Inventory (i.e., 1990 through 2017) than the previous Inventory (i.e., 1990 through 2016).
4.7 Nitric Acid Production (CRF Source
Category 2B2)
Nitrous oxide (N20) 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. The
basic process technology for producing nitric acid lias not changed significantly over time. Most U.S. plants were
built between 1960 and 2000. As of 2017, there were 31 active nitric acid production plants, including one high-
strength nitric acid production plant in the United States (EPA 2010; EPA 2018).
During this reaction, N20 is formed as a byproduct and is released from reactor vents into the atmosphere.
Emissions from fuels consumed for energy purposes during the production of nitric acid are accounted for in the
Energy chapter.
Industrial Processes and Product Use 4-31
-------
Nitric acid is made from the reaction of ammonia (NH3) with oxygen (O2) in two stages. The overall reaction is:
4NH3 + 802 -> 4HNO:i +4H20
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 N20. 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 N20) in 2017 (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 CO2 Eq.
kt N2O
1990
12.1
41
2005
11.3
38
2013
10.7
36
2014
10.9
37
2015
11.6
39
2016
10.1
34
2017
9.3
31
Methodology
Emissions of N20 were calculated using the estimation methods provided by the 2006IPCC Guidelines and
country-specific methods from EPA's GHGRP. The 2006IPCC 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 N20 emissions for 2010 through 2017.
2010 through 2017
Process N20 emissions and nitric acid production data were obtained directly from EPA's GHGRP for 2010 through
2017 by aggregating reported facility-level data (EPA 2018). In the United States, all nitric acid facilities producing
weak nitric acid (30 to 70 percent in strength) are required to report annual greenhouse gas emissions data to EPA as
per the requirements of its GHGRP. As of 2017, there were 31 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 N20 emissions using monitoring
equipment.25 The high-strength nitric acid facility also reports N20 emissions associated with weak acid production
and this may capture all relevant emissions, pending additional further EPA research. More details on the
calculation, monitoring and QA/QC methods applicable to nitric acid facilities can be found under Subpart V: Nitric
Acid Production of the regulation Part 98.26 EPA verifies annual facility-level GHGRP reports through a multi-step
25 Facilities must use standard methods, either EPA Method 320 or ASTM D6348-03 and must follow associated QA/QC
procedures consistent during these performance test consistent with category-specific QC of direct emission measurements.
26 See .
4-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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. Based on the results of the verification process, EPA
follows up with facilities to resolve mistakes that may have occurred.27
To calculate emissions from 2010 through 2017, the GHGRP nitric acid production data are utilized to develop
weighted country-specific emission factors used to calculate emissions estimates. 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.
1990 through 2009
Using GHGRP data for 2010,28 country-specific N20 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
N20/metric ton HNO3 produced at plants using abatement technologies (e.g., tertiary systems such as NSCR
systems) and 5.99 kg N20/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 N20 emissions from nitric acid production for years prior to the availability of GHGRP data
(i.e., 1990 through 2008 and 2009). A separate weighted 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 N20 abatement
technologies are not known at most facilities, but it is assumed that facilities reporting abatement technology use
have had this technology installed and operational for the duration of the time series considered in this report
(especially NSCRs).
The country-specific weighted N20 emission factors were used in conjunction with annual production to estimate
N20 emissions for 1990 through 2009, using the following equations:
Ei — Pi X EFwelg)lt:ecl l
EFweighted,i = X EFc) + (%PUnc,i X EFunc)\
where,
Ei = Annual N20 Emissions for year i (kg/yr)
Pi = Annual nitric acid production for year i (metric tons HNO3)
EF weighted,i = Weighted N20 emission factor for year i (kg N20/metric ton HNO3)
%Pc,i = Percent national production of HNO3 with N20 abatement technology (%)
EFC = N20 emission factor, with abatement technology (kg N20/metric ton HNO3)
%Punc,i = Percent national production of HNO3 without N20 abatement technology (%)
EFunc = N20 emission factor, without abatement technology (kg N20/metric ton HNO3)
i = year from 1990 through 2009
27 See .
28 National N2O 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 4-33
-------
• For 2009: Weighted N2O emission factor = 5.46 kg NiO/metric ton HNO3.
• For 1990 through 2008: Weighted N2O emission factor = 5.66 kg NiO/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
2013
7,580
2014
7,660
2015
7,210
2016
7,810
2017
7,780
Uncertainty and Time-Series Consistency
Uncertainty associated with the parameters used to estimate N20 emissions includes the share of U.S. nitric acid
production attributable to each emission abatement technology over the time series (especially prior to 2010), and
the associated emission factors applied to each abatement technology type. While some information has been
obtained through outreach with industry associations, limited information is available over the time series
(especially prior to 2010) for a variety of facility level variables, including plant-specific production levels, plant
production technology (e.g., low, high pressure, etc.), and abatement technology type, installation date of abatement
technology, and accurate destruction and removal efficiency rates. Production data prior to 2010 were obtained from
National Census Bureau, which does not provide uncertainty estimates with their data. Facilities reporting to EPA's
GHGRP must measure production using equipment and practices used for accounting purposes. At this time EPA
does not estimate uncertainty of the aggregated facility-level information. As noted in the Methodology section,
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 N20 emissions is assumed to have low uncertainty.
The results of this Approach 2 quantitative uncertainty analysis are summarized in Table 4-29. Nitrous oxide
emissions from nitric acid production were estimated to be between 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.
4-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-29: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from Nitric
Acid Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Nitric Acid Production
NO
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 2017.
To maintain consistency across the time series and with the rounding approaches taken by other data sets, a new
rounding approach was performed for the GHGRP Subpart V: Nitric Acid data. This resulted in production data
changes across the time series of 2010 to 2017, in which EPA's GHGRP data have been utilized. The results of this
update have had an insignificant impact on the emission estimates across the 2010 to 2017 time series. Details on the
emission trends through time are described in more detail in the Methodology section above.
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.
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 EPA's GHGRP facilities on the
installation date of any N20 abatement equipment, per recent revisions finalized in December 2016 to EPA's
GHGRP. This information will enable more accurate estimation of N20 emissions from nitric acid production over
the time series.
4.8 Adipic Acid Production (CRF Source
Category 2B3)
Adipic acid is produced through a two-stage process during which nitrous oxide (N20) is generated in the second
stage. Emissions from fuels consumed for energy purposes during the production of adipic acid are accounted for in
the Energy chapter. The first stage of manufacturing usually involves the oxidation of cyclohexane to form a
cyclohexanone/cyclohexanol mixture. The second stage involves oxidizing this mixture with nitric acid to produce
adipic acid. Nitrous oxide is generated as a byproduct of the nitric acid oxidation stage and is emitted in the waste
gas stream (Thiemens and Trogler 1991). The second stage is represented by the following chemical reaction:
('CH2)5CO(cyclohexanone) + (CH2)zCHOH (cyclohexanol) + wHN03
-» HOOC(CH2)4COOH(adipic acid) + xN20 + yH20
Process emissions from the production of adipic acid vary with the types of technologies and level of emission
controls employed by a facility. In 1990, two major adipic acid-producing plants had N20 abatement technologies in
place and, as of 1998, three major adipic acid production facilities had control systems in place (Reimer et al. 1999).
Industrial Processes and Product Use 4-35
-------
In 2017, catalytic reduction non-selective catalytic reduction (NSCR) and thermal reduction abatement technologies
were applied as N20 abatement measures at adipic acid facilities (EPA 2017).
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 2017,
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 2017).
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 10 percent over the period of 1990 through 2017, to
approximately 830,000 metric tons (ACC 2018). Nitrous oxide emissions from adipic acid production were
estimated to be 7.4 MMT CO2 Eq. (25 kt N2O) in 2017 (see Table 4-30). Over the period 1990 through 2017,
emissions have been reduced by 51 percent due to both the widespread installation of pollution control measures in
the late 1990s and plant idling in the late 2000s. Very little information on annual trends in the activity data exist for
adipic acid.
Table 4-30: N2O Emissions from Adipic Acid Production (MMT CO2 Eq. and kt N2O)
Year
MMT CO2 Eq.
kt N2O
1990
15.2
51
2005
7.1
24
2013
3.9
13
2014
5.4
18
2015
4.3
14
2016
7.0
23
2017
7.4
25
Methodology
Emissions are estimated using both Tier 2 and Tier 3 methods consistent with the 2006IPCC Guidelines. Due to
confidential business information (CBI), plant names are not provided in this section. Therefore, the four adipic
acid-producing facilities that have operated over the time series will be referred to as Plants 1 through 4. Overall, as
noted above, the two currently operating facilities use catalytic reduction NSCR and thermal reduction abatement
technologies.
2010 through 2017
All emission estimates for 2010 through 2017 were obtained through analysis of GHGRP data (EPA 2010 through
2013; EPA 2014 through 2016; EPA 2017), which is consistent with the 2006IPCC Guidelines Tier 3 method.
Facility-level greenhouse gas emissions data were obtained from EPA's GHGRP for the years 2010 through 2017
(EPA 2010 through 2013; EPA 2014 through 2016; EPA 2017) 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
4-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
by directly measuring N20 emissions using monitoring equipment.205 More information on the calculation,
monitoring and QA/QC methods for process N20 emissions applicable to adipic acid production facilities under
Subpart E can be found in the electronic code of federal regulations.206 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.207
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
N20 emissions from adipic acid production, metric tons
Qaa
Quantity of adipic acid produced, metric tons
EFaa
Emission factor, metric ton N20/metric ton adipic acid produced
DF
N20 destruction factor
UF
Abatement system utility factor
The adipic acid production is multiplied by an emission factor (i.e., N20 emitted per unit of adipic acid produced),
which has been estimated, based on experiments that the reaction stoichiometry for N20 production in the
preparation of adipic acid, to be approximately 0.3 metric tons of N20 per metric ton of product (IPCC 2006). The
"N20 destruction factor" in the equation represents the percentage of N20 emissions that are destroyed by the
installed abatement technology. The "abatement system utility factor" represents the percentage of time that the
abatement equipment operates during the annual production period. Plant-specific production data for Plant 4 were
obtained across the time series through personal communications (Desai 2010, 2011). The plant-specific production
data were then used for calculating emissions as described above.
For Plant 3, 2005 through 2009 emissions were obtained directly from the plant (Desai 2010, 2011). For 1990
through 2004, emissions were estimated using plant-specific production data and the IPCC factors as described
above for Plant 4. Plant-level adipic acid production for 1990 through 2003 was estimated by allocating national
adipic acid production data to the plant level using the ratio of known plant capacity to total national capacity for all
U.S. plants (ACC 2018; 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 2017 were obtained from the American
Chemistry Council (ACC 2018).
205 Facilities must use standard methods, either EPA Method 320 or ASTM D6348-03, and must follow associated QA/QC
procedures during these performance tests consistent with category-specific QC of direct emission measurements.
206 See .
207 See .
Industrial Processes and Product Use 4-37
-------
Table 4-31: Adipic Acid Production (kt)
Year
kt
1990
755
2005
865
2013
980
2014
1,025
2015
1,055
2016
860
2017
830
Uncertainty and Time-Series Consistency
Uncertainty associated with N20 emission estimates includes the methods used by companies to monitor and
estimate emissions. While some information lias 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 2017 were estimated to be between 7.0 and 7.7 MMT CO2 Eq. at the 95
percent confidence level. These values indicate a range of approximately 5 percent below to 5 percent above the
2017 emission estimate of 7.4 MMT CO2 Eq.
Table 4-32: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from Adipic
Acid Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Adipic Acid Production
N2O
7.4
7.0 7.7
-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 2017. Data presented in Table 4-31 are for informational purposes only. As previously reported in the
Methodology section adipic acid production data was obtained from EPA's GHGRP and used to estimate emissions
between 2010 and 2017. The GHGRP Subpart E adipic acid production data are CBI and therefore not presented in
this Inventory report. As a result, those using Table 4-31 values to calculate implied emission factors may occur
variable IEFs across the time-series.
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.
4-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
4.9 Caprolactam, Glyoxal and Glyoxylic Acid
Production (CRF Source Category 2B4)
Caprolactam
Caprolactam (CV,Hi iNO) 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 emissions of nitrous oxide (N20).
During the production of caprolactam, emissions of N20 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 (CeHioO). The classical route (Raschig process) and basic reaction
equations for production 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) -> C6HuN0.H2S04 (+4NH3 and H20) -> C6HuNO + 2(NH4)2S04
In 1999, there were four caprolactam production facilities in the United States. As of 2017, the United States had 3
companies with a total of 3 caprolactam production facilities: AdvanSix in Virginia (AdvanSix 2018), BASF in
Texas (BASF 2018), and Fibrant LLC in Georgia (Fibrant 2018; TechSci n.d. 2017).
Nitrous oxide emissions from caprolactam production in the United States were estimated to be 1.4 MMT CO2 Eq.
(5 ktN20) in2017 (see Table 4-33). National emissions from caprolactam production have decreased by
approximately 16 percent over the period of 1990 through 2017. Emissions in 2017 have decreased by
approximately 30 percent from the 2016 levels.
Industrial Processes and Product Use 4-39
-------
Table 4-33: N2O Emissions from Caprolactam Production (MMT CO2 Eq. and kt N2O)
Year
MMT CO2 Eq.
kt N2O
1990
1.7
6
2005
2.1
7
2013
2.0
7
2014
2.0
7
2015
2.0
7
2016
2.0
7
2017
1.4
5
Glyoxal
Glyoxal is mainly used as a crosslinking agent for acrylic resins, disinfectant, gelatin hardening agent, and textile
finishing agent etc. It is produced from oxidation of acetaldehyde with concentrated nitric acid, or from the catalytic
oxidation of ethylene glycol, and N20 is emitted in the process of oxidation of acetaldehyde.
Glyoxal (ethanedial) (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). Glyoxal is used as a crosslinking agent for vinyl acetate/acrylic resins, disinfectant, gelatin
hardening agent, textile finishing agent (pennanent-press cotton, rayon fabrics), wet-resistance additive (paper
coatings) (IPCC 2006).
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.
Methodology
Emissions of N2O 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
2017, as shown in this formula:
EN2o = EF x CP
where,
EN2o = Annual N20 Emissions (kg)
EF = N2O emission factor (default) (kg N20/metric ton caprolactam produced)
CP = Caprolactam production (metric tons)
During the caprolactam production process, N20 is generated as a byproduct of the high temperature catalytic
oxidation of ammonia (NH3), which is the first reaction in the series of reactions to produce caprolactam. The
amount of N20 emissions can be estimated based on the chemical reaction shown above. Based on this formula,
which is consistent with an IPCC Tier 1 approach, approximately 111.1 metric tons of caprolactam are required to
generate one metric ton of N20, or an emission factor of 9.0 kg N20 per metric ton of caprolactam (IPCC 2006).
When applying the Tier 1 method, the 2006 IPCC Guidelines state that it is good practice to assume that there is no
abatement of N20 emissions and to use the highest default emission factor available in the guidelines. In addition.
EPA did not find support for the use of secondary catalysts to reduce N20 emissions, like those employed at nitric
4-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
acid plants. Thus, the 525 thousand metric tons (kt) of caprolactam produced in 2017 (ACC 2018) resulted in N20
emissions of approximately 1.4 MMT CO2 Eq. (5 kt).
The activity data for caprolactam production (see Table 4-34) from 1990 to 2017 were obtained from the ACC
Guide to the Business of Chemistry report (ACC 2018). 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
2013
750
2014
755
2015
760
2016
755
2017
525
Carbon dioxide and methane (CH4) emissions may also occur from the production of caprolactam but currently the
IPCC does not have methodologies for calculating these emissions associated with caprolactam production.
Uncertainty and Time-Series Consistency
Estimation of emissions of N20 from caprolactam production can be treated as analogous to estimation of emissions
of N2O from nitric acid production. Both production processes involve an initial step of NH3 oxidation which is the
source of N20 formation and emissions (IPCC 2006). Therefore, uncertainties for the default values in the 2006
IPCC Guidelines is 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 2017 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 31 percent
below to 32 percent above the 2017 emission estimate of 1.4 MMT CO2 Eq.
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
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Caprolactam Production
N2O
1.4
1.0 1.9 -31% +32%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
Industrial Processes and Product Use 4-41
-------
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.
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 sought expert solicitation on data available for these emissions source categories. EPA did not receive
information regarding these industries during Expert Review but will continue to research alternative datasets. This
planned improvement is subject to data availability and will be implemented in the medium- to long-term.
4.10 Silicon 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 used
as an industrial abrasive. Silicon carbide is produced for abrasive, metallurgical, and other non-abrasive applications
in the United States. Production for metallurgical and other non-abrasive applications is not available and therefore
both CO2 and CH4 estimates are based solely upon production estimates of silicon carbide for abrasive applications.
Emissions from fuels consumed for energy purposes during the production of silicon carbide are accounted for in the
Energy chapter.
Carbon dioxide and CH4 are also emitted during the production of calcium carbide, a chemical used to produce
acetylene. Carbon dioxide is implicitly accounted for in the storage factor calculation for the non-energy use of
petroleum coke in the Energy chapter. However, CH4 emissions from calcium carbide production are not included as
data are not available to apply the Tier 3 methodology prescribed by the 2006 IPCC Guidelines. EPA is continuing
to investigate the inclusion of these emissions in future Inventory reports.
To produce SiC, silica sand or quartz (SiCh) is reacted with 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,
CH4, or carbon monoxide (CO). The overall reaction is shown below (but in practice it does not proceed according
to stoichiometry):
Si02 + 3C —* SiC + 2CO (+ 02 —> 2C02)
Carbon dioxide is also emitted from the consumption of SiC for metallurgical and other non-abrasive applications.
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 2017 were 0.2 MMT CO2 Eq. (186 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 2017 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 2017
emissions are about 53 percent lower than emissions in 1990.
4-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-36: CO2 and ChU Emissions from Silicon Carbide Production and Consumption (MMT
COz Eq.)
Year 1990
2005
2013 2014 2015 2016 2017
CO2 0.4
CH4 +
0.2
+
0.2 0.2 0.2 0.2 0.2
+ + + + +
Total 0.4
0.2
0.2 0.2 0.2 0.2 0.2
+ Does not exceed 0.05 MMT CO2 Eq.
Table 4-37: CO2 and ChU Emissions from Silicon Carbide Production and Consumption (kt)
Year
1990
2005
2013
2014
2015
2016
2017
CO2
375
219
169
173
180
174
186
CH4
1
+
+
+
+
+
+
+ Does not exceed 0.5
kt.
Methodology
Emissions of CO2 and CH4 from the production of SiC were calculated208 using the Tier 1 method provided by the
2006IPCC Guidelines. Annual estimates of SiC production were multiplied by the appropriate emission factor, as
shown below:
Esc,C02 = EFsc,C02 X Qsc
/I metric ton\
Esc,CH4 = EFSCiCH4 x Qsc x ^ 10QQkg )
where.
Esc,002 = CO2 emissions from production of SiC, metric tons
EFSC,cc>2 = Emission factor for production of SiC, metric ton C02/metric ton SiC
Qsc = Quantity of SiC produced, metric tons
Esc.' 'i 11 = CH4 emissions from production of SiC, metric tons
EFsc.- 'ii i = Emission factor for production of SiC, kilogram CH4/metric ton SiC
Emission factors were taken from the 2006 IPCC Guidelines:
• 2.62 metric tons C02/metric ton SiC
• 11.6 kg CH4/metric ton SiC
Emissions of CO2 from silicon carbide consumption for metallurgical uses were calculated by multiplying the
annual utilization of SiC for metallurgical uses (reported annually in the USGS Minerals Yearbook: Silicon) by the
carbon content of SiC (31.5 percent), which was determined according to the molecular weight ratio of SiC.
Emissions of CC^from silicon carbide consumption for other non-abrasive uses were calculated by multiplying the
annual SiC consumption for non-abrasive uses by the carbon content of SiC (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.
208 jlas no{ integrated, aggregated facility-level GHGRP information to inform these estimates. Hie 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.
Industrial Processes and Product Use 4-43
-------
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 (3.1 Fossil Fuel Combustion (CRF Source Category 1 A)) and Annex
2.1, Methodology for Estimating Emissions of CO2 from Fossil Fuel Combustion.
Production data for 1990 through 2013 were obtained from the Minerals Yearbook: Manufactured Abrasives (USGS
1991a through 2015). Production data for 2014 through 2016 were obtained from the Mineral Commodity
Summaries: Abrasives (Manufactured) (USGS 2018). Production data for 2017 were obtained from the Mineral
Industry Sun'eys: Manufactured Abrasives in the Second Quarter 2018 (USGS 2018b). Silicon carbide production
data obtained through the USGS National Minerals Information Center has been previously been rounded to the
nearest 5,000 metric tons to avoid disclosing company proprietary data. Silicon carbide consumption for the entire
time series is estimated using USGS production 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 2018), see Table 4-38.
Table 4-38: Production and Consumption of Silicon Carbide (Metric Tons)
Year Production Consumption
1990 105,000 172,465
2005
35,000
220,149
2013 35,000 134,055
2014 35,000 140,733
2015 35,000 153,475
2016 35,000 142,104
2017 35,000 163,492
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 would be 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 CH4, there is also uncertainty associated with the hydrogen-containing volatile
compounds in the petroleum coke (IPCC 2006). There is also uncertainty associated with the use or destruction of
CH4 generated from the process in addition to uncertainty associated with levels of production, net imports,
consumption levels, and the percent of total consumption that is attributed to metallurgical and other non-abrasive
uses.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-39. Silicon carbide
production and consumption CO2 emissions from 2017 were estimated to be between 9 percent below and 9 percent
above the emission estimate of 0.19 MMT CO2 Eq. at the 95 percent confidence level. Silicon carbide production
CH4 emissions were estimated to be between 9 percent below and 9 percent above the emission estimate of 0.01
MMT CO2 Eq. at the 95 percent confidence level.
4-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Upper
Lower
Upper
Bound
Bound
Bound
Bound
Silicon Carbide Production
and Consumption
CO2
0.19
0.17
0.20
-9%
+9%
Silicon Carbide Production
CH4
+
+
+
-9%
+9%
+ Does not exceed 0.05 MMT CO2 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 2017. Details on the emission trends through time are described in more detail in the Methodology section
above.
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.
4.11 Titanium Dioxide Production (CRF Source
Category 2B6)
Titanium dioxide (TiO:) 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:
2FeTi03 + 7Cl2 + 3C —> 2TiCl^ + 2FeCl^ + 3C02
2TiCl4 + 202 ~~* 2Ti02 -I-
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 FeTiCb (rutile ore) to form CO2. Since 2004, all TiO: 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 TiO: 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 2017, U.S. TiOi production totaled 1,260,000 metric tons
(USGS 2018). There were a total five plants producing TiO: in the United States in 2017.
Emissions of CO2 from titanium dioxide production in 2017 were estimated to be 1.7 MMT CO2 Eq. (1,688 kt CO2),
which represents an increase of 41 percent since 1990 (see Table 4-40). Compared to 2016, emissions from titanium
dioxide production increased by 2 percent in 2017 due to a 2 percent increase in production.
Industrial Processes and Product Use 4-45
-------
Table 4-40: CO2 Emissions from Titanium Dioxide (MMT CO2 Eq. and kt)
Year MMT CP2 Eg.
kt
1990
1.2
1,195
2005
1.8
1,755
2013
2014
2015
2016
2017
1.7
1.7
1.6
1.7
1.7
1,715
1,688
1,635
1,662
1,688
Methodology
Emissions of CO2 from TiO: production were calculated by multiplying annual national TiO: production by chloride
process-specific emission factors using a Tier 1 approach provided in 2006IPCC Guidelines. The Tier 1 equation is
as follows:
The petroleum coke portion of the total CO2 process emissions from TiO: 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 (3.1 Fossil Fuel Combustion (CRF Source Category 1 A)) 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
C02/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 2017 were obtained from the Minerals
Commodity Summary: Titanium and Titanium Dioxide (USGS 2018).209 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).
209 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.
Etd ~ EFtd x Qtd
where.
Etd
EFtd
Qtd
CO2 emissions from TiC>2 production metric tons
Emission factor (chloride process), metric ton C02/metric ton TiC>2
Quantity of TiC>2 produced
4-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-41: Titanium Dioxide Production (kt)
Year
kt
1990
979
2005
1,310
2013
1,280
2014
1,260
2015
1,220
2016
1,240
2017
1,260
Uncertainty and Time-Series Consistency
Each year, the USGS collects titanium industry data for titanium mineral and pigment production operations. If TiO:
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 TiO: pigment plants over the time
series.
Although some TiO: 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 CO2 per unit of TiO: produced as compared to that generated using petroleum coke in production. While
the most accurate method to estimate emissions would be to base calculations on the amount of reducing agent used
in each process rather than on the amount of TiO: 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 TiO: 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 TiO: production capacity that
was attributed to the chloride process was multiplied by total TiO: production to estimate the amount ofTiO:
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 TiO: production, literature data
were used for petroleum coke composition. Certain grades of petroleum coke are manufactured specifically for use
in the TiO: 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 2017 were estimated to be between 1.5 and 1.9 MMT CO2 Eq. at the 95 percent
confidence level. This indicates a range of approximately 13 percent below and 13 percent above the emission
estimate of 1.7 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
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Titanium Dioxide Production
CO2
1.7
1.5 1.9
-13% +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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
Industrial Processes and Product Use 4-47
-------
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.
Planned Improvements
Planned improvements include researching the significance of titanium-slag production in electric furnaces and
synthetic-rutile production using the Becher process in the United States. Significant use of these production
processes will be included in future Inventory reports. Due to resource constraints, this planned improvement is still
in development by EPA and is not included in this report. This is a long-term improvement.
EPA continues to assess the potential of integrating aggregated facility-level GHGRP information for titanium
dioxide production facilities based on criteria to shield underlying CBI from public disclosure. Pending available
resources, EPA will also evaluate use of GHGRP data to improve category-specific QC consistent with both Volume
1, Chapter 6 of 2006 IPCC Guidelines and the latest IPCC guidance on the use of facility-level data in national
inventories.210
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 sector.
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, Na2CC>3) 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. Emissions from soda ash used in
glass production are reported under Section 4.3, Glass Production (CRF Source Category 2A3). Glass production is
its own source category and historical soda ash consumption figures have been adjusted to reflect this change. After
glass manufacturing, soda ash is used primarily to manufacture many sodium-based inorganic chemicals, including
sodium bicarbonate, sodium chromates, sodium phosphates, and sodium silicates (USGS 2015a). 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 2018b). Only two states produce
natural soda ash: Wyoming and California. Of these two states, only net emissions of CO2 from Wyoming were
calculated due to specifics regarding the production processes employed in the state.211 Based on 2017 reported
210 See .
211 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 CO2 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 CO2 is generated as a byproduct, the CO2 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
4-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
data, the estimated distribution of soda ash by end-use in 2017 (excluding glass production) was chemical
production 57 percent; soap and detergent manufacturing, 12 percent; distributors, 11 percent; flue gas
desulfurization 8 percent; other uses, 7 percent; water treatment, 3 percent, and pulp and paper production, 2
percent (USGS 2018).212
U.S. natural soda ash is competitive in world markets because the majority of the world output of soda ash is made
synthetically. Although the United States continues to be a major supplier of world soda ash, China, which
surpassed the United States in soda ash production in 2003, is the world's leading producer.
In 2017, CO2 emissions from the production of soda ash from trona were approximately 1.8 MMT CO2 Eq. (1,753 kt
CO2) (see Table 4-43). Total emissions from soda ash production in 2017 increased by approximately 2 percent
from emissions in 2016, and have increased by approximately 22 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 2017 since experiencing a decline in domestic and
export sales caused by adverse global economic conditions in 2009.
Table 4-43: CO2 Emissions from Soda Ash Production (MMT CO2 Eq. and kt CO2)
Year MMT CO2 Eq. kt CO2
1990 1.4 1,431
2005 1.7 1,655
2013
1.7
1,694
2014
1.7
1,685
2015
1.7
1,714
2016
1.7
1,723
2017
1.8
1,753
Methodology
During the 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 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 are required to generate one metric ton of CO2, or an emission factor of 0.0974 metric tons CO2 per
metric ton trona (IPCC 2006). Thus, the 18.0 million metric tons of trona mined in 2017 for soda ash production
(USGS 2018) resulted in CO2 emissions of approximately 1.8 MMT CO2 Eq. (1,753 kt).
Once produced, most soda ash is consumed in chemical production with minor amounts 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 production (see Table 4-44) for 1990 to 2017 were obtained from the U.S. Geological
Survey (USGS) Minerals Yearbook for Soda Ash (1994 through 2015b) and USGS Mineral Industry Sun'eys for
bicarbonate no longer mines trona 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 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.
212 Percentages may not add up to 100 percent due to independent rounding.
Industrial Processes and Product Use 4-49
-------
Soda Ash (USGS 2017). Soda ash production213 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 fromEPA's
GHGRP to improve the emission estimates for Soda Ash Production source category consistent with IPCC214 and
UNFCCC guidelines.
Table 4-44: Soda Ash Production (kt)
Year
Production3
1990
14,700
2005
17,000
2013
17,400
2014
17,300
2015
17,600
2016
17,700
2017
18,000
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 in that reliable and
accurate data sources are available for the emission factor and activity data for trona-based soda ash production.
EPA plans to work with other entities to reassess the uncertainty of these emission factors and activity data based on
the most recent information and data. Through EPA's GHGRP, EPA is aware of one facility producing soda ash
from a liquid alkaline feedstock process. 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 2016). 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 1995).
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-45. Soda Ash Production
CO2 emissions for 2017 were estimated to be between 1.6 and 1.9 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 1.8 MMT
C02 Eq.
Table 4-45: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Soda Ash
Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CChEq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Soda Ash Production
CO2
1.8
1.6
1.9
-9%
+8%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
213 EPA has assessed 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.
214 See .
4-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Methodological approaches were applied to the entire time series to ensure consistency in emissions from 1990
through 2017.
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.
Planned Improvements
EPA plans to use GHGRP data for conducting category-specific QC of emission estimates consistent with both
Volume 1, Chapter 6 of 2006 IPCC Guidelines and the latest IPCC guidance on the use of facility-level data in
national inventories.215 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 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 CH4 emissions from the production of methanol and acrylonitrile are presented here and reported
under IPCC Source Category 2B8. The petrochemical industry uses primary fossil fuels (i.e., natural gas, coal,
petroleum, etc.) for non-fuel purposes in the production of carbon black and other petrochemicals. Emissions from
fuels and feedstocks transferred out of the system for use in energy purposes (e.g., indirect or direct process heat or
steam production) are currently accounted for in the Energy sector. The allocation and reporting of emissions from
feedstocks transferred out of the system for use in energy purposes to the Energy Chapter is consistent with 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
215 See.
Industrial Processes and Product Use 4-51
-------
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:
Hf, C2H4 + H2
Small amounts of CH4 are also generated from the steam cracking process. In addition, CO2 and CH4 emissions are
also generated from combustion units.
Ethylene dichloride (C2H4CI2) is used to produce vinyl chloride monomer, which is the precursor to polyvinyl
chloride (PVC). Ethylene dichloride was used as a fuel additive until 1996 when leaded gasoline was phased out.
Ethylene dichloride is produced from ethylene by either direct chlorination, oxychlorination, or a combination of the
two processes (i.e., the "balanced process"); most U.S. facilities use the balanced process. The direct chlorination
and oxychlorination reactions are shown below:
C2H4 + Cl2 -» C2H4Cl2 (direct chlorination)
C2H4 + -02 + 2HCI -» C2H4Cl2 + 2H20 (oxychlorination)
C2H4 + 3 02 —> 2C02 + 2 H20 (direct oxidation of ethylene during oxychlorination)
In addition to the byproduct CO2 produced from the direction oxidation of the ethylene feedstock, CO2 and CH4
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 2017 were 28.2 MMT C02Eq. (28,225 kt CO2) and
0.3 MMT CO2 Eq. (10 kt CH4), respectively (see Table 4-46 and Table 4-47). Since 1990, total CO2 emissions from
petrochemical production increased by 33 percent. Methane emissions from petrochemical (methanol and
acrylonitrile) production reached a low of 1.8 kt CH4 in 2011, given declining methanol production; however, CH4
emissions have been increasing every year since 2011 and are now 14 percent greater than in 1990 (though still less
than the peak in 1997) due to a rebound in methanol production.
Table 4-46: CO2 and ChU Emissions from Petrochemical Production (MMT CO2 Eq.)
Year 1990 2005 2013 2014 2015 2016 2017
CO2 21.2 26.8 26.4 26.5 28.1 28.1 28.2
4-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
CH4 0.2 0.1 0.1 0.1 0.2 0.2 0.3
Total 21.4 26.9 26.5 26.6 28.2 28.4 28.5
Table 4-47: CO2 and ChU Emissions from Petrochemical Production (kt)
Year Wll 2005 2013 2014 2015 2016 2017
CO2 21,222 26,810 26,395 26,496 28,062 28,110 28,225
CH4 9 3 3 5 7 10 10
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 CH4 emissions from production of acrylonitrile and methanol,216 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, 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 CO 2.
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, and these facilities are required to also report CH4 and N20 emissions from
combustion of process off-gas in flares. Preliminary analysis of aggregated annual reports shows that these flared
CH4 and N20 emissions are less than 500 kt/year. EPA's GHGRP is still reviewing this data across reported years to
facilitate update of category-specific QC documentation and EPA plans to address this more completely in future
reports.
Carbon Black, Ethylene, Ethylene Dichloride and Ethylene Oxide
2010 through 2017
Carbon dioxide emissions and national production were aggregated directly from EPA's GHGRP dataset for 2010
through 2016 (EPA 2017). The GHGRP data for 2016 were also used as a proxy for 2017 because the 2017 data
were unavailable prior to preparation of this report. However, preliminary analysis of GHGRP Subpart X shows that
emissions in 2017 were consistent, albeit slightly higher, with those reported in 2016. In 2016, data reported to the
GHGRP included CO2 emissions of 3,160,000 metric tons from carbon black production; 19,600,000 metric tons of
CChfrom ethylene production; 447,000 metric tons of CChfrom ethylene dichloride production; and 1,100,000
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
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
216 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 4-53
-------
process CO2 emissions; ethylene production facilities also have a third option. The mass balance method is used by
most facilities217 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.
More details on the greenhouse gas calculation, monitoring and QA/QC methods applicable to petrochemical
facilities can be found under Subpart X (Petrochemical Production) of the regulation (40 CFR Part 98).218 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.219
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 CO2 from Fossil Fuel Combustion (3.1 Fossil Fuel Combustion (IPCC Source
Category 1 A)) 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, and ethylene dichloride, carbon dioxide emission factors were derived from EPA's
GHGRP data by dividing annual CO2 emissions for petrochemical type "i" with annual production for petrochemical
type "i" and then averaging the derived emission factors obtained for each calendar year 2010 through 2016. For
ethylene oxide, the carbon dioxide emission factor was derived in the same manner, except that only data from
calendar years 2010 through2013 were used to develop the average emission factor because process improvements
in recent years have resulted in lower CO2 emissions that are not representative of operation in 1990 through 2009.
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:
• 2.63 metric tons CCVmetric ton carbon black produced
217 A few facilities producing ethylene dichloride used CO2 CEMS, those CO2 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 CO2 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 CO2,
these facilities are required to report emissions of CH4 and N2O 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 N2O from combustion of petrochemical process-off gases in flares. Preliminary analysis of the aggregated reported CH4
and N2O emissions from facilities using CEMS and N2O 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.
218 See .
219 See .
4-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• 0.77 metric tons CCh/mctric ton ethylene produced
• 0.041 metric tons COVmctric ton ethylene dichloride produced
• 0.46 metric tons COVmctric 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). As noted above, annual 2010 through 2016 production data for
carbon black, ethylene, ethylene dichloride, and ethylene oxide, were obtained from EPA's GHGRP, and data from
2016 were used as a proxy for 2017.
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 1 CO2 and CH4
emission factors to estimate emissions for 1990 through 2017. Emission factors used to estimate acrylonitrile
production emissions are as follows:
• 0.18 kg CH4/metric ton acrylonitrile produced
• 1.00 metric tons CCh/mctric ton acrylonitrile produced
Annual acrylonitrile production data for 1990 through 2017 were obtained from ACC's Business of Chemistry (ACC
2018).
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 1 CO2 and CH4
emission factors to estimate emissions for 1990 through 2017. Emission factors used to estimate methanol
production emissions are as follows:
• 2.3 kg CH4/metric ton methanol produced
• 0.67 metric tons CCh/mctric ton methanol produced
Annual methanol production data for 1990 through 2017 were obtained from the ACC's Business of Chemistry
(ACC 2018).
Table 4-48: Production of Selected Petrochemicals (kt)
Chemical
1990
2005
2013
2014
2015
2016
2017
Carbon Black
1,307
1,651
1,230
1,210
1,220
1,190
1,190
Ethylene
16,542
23,975
25,300
25,500
26,900
26,600
26,600
Ethylene Dichloride
6,283
11,260
11,500
11,300
11,300
11,700
11,700
Ethylene Oxide
2,429
3,220
3,150
3,140
3,240
3,210
3,210
Acrylonitrile
1,214
1,325
1,075
1,095
1,050
955
1,040
Methanol
3,750
1,225
1,235
2,105
3,065
4,250
4,295
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
Industrial Processes and Product Use 4-55
-------
molecular weight for gaseous feedstocks) for the mass balance methodology beginning in reporting year 2017 (81
FR 89260).220 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 CH4 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 2017 were estimated to be between 26.7 and 29.7 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 28.2 MMT CO2 Eq. Petrochemical production CH4 emissions from 2017 were estimated to be between 0.09 and
0.31 MMT CO2 Eq. at the 95 percent confidence level. This indicates a range of approximately 57 percent below to
45 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)
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Upper
Lower
Upper
Bound
Bound
Bound
Bound
Petrochemical
Production
CO2
28.2
26.7
29.7
-5%
+5%
Petrochemical
Production
CH4
0.34
0.09
0.31
-57%
+45%
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 2017.
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 QA/QC and Verification Procedures section in the introduction of the
IPPU Chapter.
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
220 See .
4-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
quality control measures for this category included the QA/QC requirements and verification procedures of EPA's
GHGRP.
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 2010 when the difference was 11 percent). For ethylene dichloride, the GHGRP
emissions are typically within 20 percent of the Tier 1 emissions (except in 2014 due to incorrect GHGRP emissions
that were not corrected before the most recent publication of the data). For ethylene oxide, GHGRP emissions vary
from 10 percent less than the Tier 1 emissions to 30 percent more than the Tier 1 emissions, depending on the year.
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 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 EFs and report on the comparison of Tier 1 emissions
estimates and GHGRP data are described below in the Planned Improvements section.
Recalculations Discussion
As previously noted above, 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
emissions from production of these petrochemicals in 1990 through 2009. In previous versions of the Inventory,
average emission factors were developed from all years of available GHGRP data. However, in recent years, the
emission factor for ethylene oxide has been steadily declining as a result of process efficiencies being implemented
through the industry; thus, in an effort to better characterize the emissions from 1990 through 2009, the emissions
factor for ethylene oxide in this year's Inventory is based on the GHGRP data only from 2010 through 2013. The
emission factor calculated using only these 4 years of data is 11 percent higher than the emission factor using all
data from 2010 through 2016. Thus, estimated CO2 emissions from ethylene oxide production in 1990 through 2009
are about 11 percent higher in the current Inventory (i.e., 1990 through 2017) than the previous Inventory (i.e., 1990
through 2016).
Planned Improvements
Improvements include completing category-specific QC of activity data and emission factors, along with further
assessment of CH4 and N20 emissions to enhance completeness in reporting of emissions from U.S. petrochemical
production, pending resources, significance and time-series consistency considerations. For example, EPA is
planning additional assessment of ways to use CH4 data from the GHGRP in the inventory. One possible approach
Industrial Processes and Product Use 4-57
-------
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. 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
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.221 Feedstock production, however, is
permitted to continue indefinitely.
HCFC-22 is produced by the reaction of chloroform (CHCI3) and hydrogen fluoride (HF) in the presence of a
catalyst, SbCk The reaction of the catalyst and HF produces SbClxFy, (where x + y = 5), which reacts with
chlorinated hydrocarbons to replace chlorine atoms with fluorine. The HF and chloroform are introduced by
submerged piping into a continuous-flow reactor that contains the catalyst in a hydrocarbon mixture of chloroform
and partially fluorinated intermediates. The vapors leaving the reactor contain HCFC-21 (CHCI2F), HCFC-22
(CHCIF2), HFC-23 (CHF3), HC1, 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, HC1 and residual HF. The
HC1 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 2017. Emissions of HFC-23 from this activity in 2017
were estimated to be 5.2 MMT CO2 Eq. (0.3 kt) (see Table 4-50). This quantity represents an 85 percent increase
from 2015 emissions and an 89 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). An uptick in
this rate, as well as in the quantity of HCFC-22 produced, was responsible for the increase in HFC-23 emissions
between 2016 and 2017. 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)
221 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]
4-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 factor 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 CCh Eq. and kt HFC-23)
Year
MMT CChEq.
kt HFC-23
1990
46.1
3
2005
20.0
1
2013
4.1
0.3
2014
5.0
0.3
2015
4.3
0.3
2016
2.8
0.2
2017
5.2
0.3
fvieth ado logy
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 2017 were obtained through reports submitted by U.S. HCFC-22 production facilities to EPA's
Greenhouse Gas Reporting Program (GHGRP). EPA's GHGRP mandates that all HCFC-22 production facilities
report their annual emissions of HFC-23 from HCFC-22 production processes and HFC-23 destruction processes.
Previously, data were obtained by EPA through collaboration with an industry association that received voluntarily
reported HCFC-22 production and HFC-23 emissions annually from all U.S. HCFC-22 producers from 1990
through 2009. These emissions were aggregated and reported to EPA on an annual basis.
For the other three plants, the last of which closed in 1993, methods comparable to the Tier 1 method in the 2006
IPCC Guidelines were used. Emissions from these three plants have been calculated using the recommended
emission factor for unoptimized plants operating before 1995 (0.04 kg HCFC-23/kg HCFC-22 produced).
The five plants that have operated since 1994 measure (or, for the plants that have since closed, measured)
concentrations of HFC-23 to estimate their emissions 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.
Plants that release (or historically have released) some of their byproduct HFC-23 periodically measure HFC-23
concentrations in the output stream 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 2017 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 (RTI 1997; RTI2008). 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
Industrial Processes and Product Use 4-59
-------
2012 96
2013 C
2014 C
2015 C
2016 C
201 7 C_
C (CBI)
Note: HCFC-22 production in 2013 through
2017 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 2017. 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 2017 (because one plant has closed), the plant that currently accounts for most emissions had a relative
uncertainty in its 2006 (as well as 2005) emissions estimate that was similar to the relative uncertainty for total U.S.
emissions. Thus, the closure of one plant is not likely to have a large impact on the uncertainty of the national
emission estimate.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-52. HFC-23 emissions
from HCFC-22 production were estimated to be between 4.8 and 5.7 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 5.2
MMT C02 Eq.
Table 4-52: Approach 2 Quantitative Uncertainty Estimates for HFC-23 Emissions from
HCFC-22 Production (MMT CO2 Eq. and Percent)
Source
_ 2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
HCFC-22 Production
HFC-23 5.2
4.8 5.7
-7% +10%
a Range of emissions reflects a 95 percent confidence interval.
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, as described in Annex 8. Source-specific quality control measures for the HCFC-22 Production
category included the QA/QC requirements and verification procedures of EPA's GHGRP. 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
4-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
standards and suitable methods published by a consensus standards organization, (4) calibrate gas cliroinatographs at
least monthly through analysis of certified standards, and (5) document these calibrations.
EPA verifies annual facility-level reports from HCFC-22 producers through a multi-step process (e.g., a
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.222
4.15 Carbon Dioxide Consumption (CRF Source
Category 2B10)
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). Carbon dioxide 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 3-7 titled "Carbon Dioxide Transport,
Injection, and Geological Storage."
In 2017, the amount of CO2 produced and captured for commercial applications and subsequently emitted to the
atmosphere was 4.5 MMT C02Eq. (4,471 kt) (see Table 4-53). This is consistent with 2014 through 2016 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 CO2 Eq. kt
1990 1.5 1,472
2005 1.4 1,375
2013 4.2 4,188
2014 4.5 4,471
2015 4.5 4,471
2016 4.5 4,471
201 7 45 4,471
222 See .
Industrial Processes and Product Use 4-61
-------
Methodology
Carbon dioxide emission estimates for 1990 through 2017 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.
2010 through 2017
For 2010 through 2017, 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 2017,
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 98.223 The number of facilities that reported data to
EPA's GHGRP Subpart PP (Suppliers of Carbon Dioxide) for 2010 through 2017 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 2017 due to data confidentiality reasons. As a result, the emissions estimates for 2015 through 2017 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 (ARI2006, 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
223 See .
4-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
available for 2009. The percentage of total production that was used for non-EOR applications for the Bravo Dome
and West Bravo Dome facilities for 1990 through 2009 were obtained from New Mexico Bureau of Geology and
Mineral Resources (Broadhead 2003; New Mexico Bureau of Geology and Mineral Resources 2006). Production
data for the McCallum Dome (Jackson County), Colorado facility were obtained from the Colorado Oil and Gas
Conservation Commission (COGCC) for 1999 through 2009 (COGCC 2014). Production data for 1990 to 1998 and
percentage of production used for EOR were assumed to be the same as for 1999, due to lack of publicly-available
data.
Table 4-54: CO2 Production (kt CO2) and the Percent Used for Non-EOR Applications
Year Jackson Dome, Bravo Dome, West Bravo McCallum Total CO2 %
MS NM Dome, NM CO2 Dome, CO Production Non-
CO2 Production CO2 Production Production CO2 Production from Extraction EORa
(kt) (% Non- (kt) (% Non- (kt) (% Non- (kt) (% Non- and Capture
EOR) EOR) EOR) EOR) Facilities (kt)
1990 1,344(100%) 63(1%) + 65 (100%) NA NA
2005 1,254(27%) 58(1%) + 63 (100%) NA NA
2013 NA NA NA NA 68,435b 6%
2014 NA NA NA NA 72,000b 6%
2015 NA NA NA NA 72,000b 6%
2016 NA NA NA NA 72,000b 6%
201 7 NA NA NA NA 72,000b 6%
+ Does not exceed 0.5 percent.
NA (Not Available)
a Includes only food & beverage applications.
bFor 2010 through 2017, 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 2017, values were held constant with those from
2014. Facility-level data are not publicly available from EPA's GHGRP.
Uncertainty and Time-Series Consistency
There is uncertainty associated with the data reported through EPA's GHGRP. Specifically, there is uncertainty
associated with the amount of 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.224
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-55. Carbon dioxide
consumption CO2 emissions for 2017 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 C02 Eq.
224 See .
Industrial Processes and Product Use 4-63
-------
Table 4-55: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from CO2
Consumption (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
CO2 Consumption
CO2
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 2017.
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.
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.225 In addition, EPA is also investigating the possibility of utilizing only extraction facility Subpart
PP data, while also updating the values for 2015 through 2017.
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.
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.
225 See .
4-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
The phosphoric acid production process involves chemical reaction of the calcium phosphate (Ca3(P04)2)
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 2017 was an estimated 26.7 million metric tons (USGS 2018). Total
imports of phosphate rock to the United States in 2017 were estimated to be approximately 2.1 million metric tons
(USGS 2018). Between 2013 and 2016, most of the imported phosphate rock (67 percent) came from Peru, with 32
percent being from Morocco and 1 percent from other sources (USGS 2018). All phosphate rock mining companies
are vertically integrated with fertilizer plants that produce phosphoric acid located near the mines. Some additional
phosphoric acid production facilities are located in Texas, Louisiana, and Mississippi that used imported phosphate
rock.
Over the 1990 to 2017 period, domestic production has decreased by nearly 46 percent. Total CO2 emissions from
phosphoric acid production were 1.0 MMT CO2 Eq. (1,023 kt CO2) in 2017 (see Table 4-56). Domestic
consumption of phosphate rock in 2017 was estimated to have increased 2 percent over 2016 levels (USGS 2018).
Table 4-56: CO2 Emissions from Phosphoric Acid Production (MMT CO2 Eq. and kt)
Year
MMT CO2 Eq.
kt
1990
1.5
1,529
2005
1.3
1,342
2013
1.1
1,149
2014
1.0
1,038
2015
1.0
999
2016
1.0
998
2017
1.0
1,023
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:
Epa Cpr * 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 is based on the assumption 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.
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 2017, 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,
Industrial Processes and Product Use 4-65
-------
and the amount mined in Idaho and Utah, are approximated using average share of U.S. production in those states
from 1993 to 2004 data. For the years 2005 through 2017, the same approximation method is used, but the share of
U.S. production in those states data 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 2017 were obtained from
USGS Minerals Yearbook: Phosphate Rock (USGS 1994 through 2015b), and from USGS Minerals Commodity
Summaries: Phosphate Rock (USGS 2016, 2017, 2018). From 2004 through 2017, 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 carbon. 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 (80 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 is based on the assumption 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
2013
2014
2015
2016
2017
U.S. Domestic Consumption
49,800
35,200
28,800
26,700
26,200
26,700
26,700
FLandNC
42,494
28,160
23,040
21,360
20,960
21,360
21,360
ID and UT
7,306
7,040
5,760
5,340
5,240
5,340
5,340
Exports—FL and NC
6,240
0
0
0
0
0
0
Imports
451
2,630
3,170
2,390
1,960
1,590
2,100
Total U.S. Consumption
44,011
37,830
31,970
29,090
28,160
28,290
28,800
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 CO2)
3.67
3.43
1.50
1.00
5.00
Source: FIPR (2003a).
Uncertainty and Time-Series Consistency
Phosphate rock production data used in the emission calculations were developed by the USGS through monthly and
semiannual voluntary surveys of the active phosphate rock mines during 2017. For previous years in the time series,
USGS provided the data disaggregated regionally; however, beginning in 2006, only total U.S. phosphate rock
production was reported. Regional production for 2017 was estimated based on regional production data from
previous years and multiplied by regionally-specific emission factors. There is uncertainty associated with the
degree to which the estimated 2017 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 calculation are reported by phosphate rock producers and are not considered to be a significant
4-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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; 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 approximately 7 percent of domestically-produced phosphate rock is used to
manufacture elemental phosphorus and other phosphorus-based chemicals, rather than phosphoric acid (USGS
2006). 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 is based on the assumption that phosphate rock consumption for purposes other than phosphoric acid
production results in CO2 emissions 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. 2017 phosphoric acid
production CO2 emissions were estimated to be between 0.9 and 1.3 MMT CO2 Eq. at the 95 percent confidence
level. This indicates a range of approximately 19 percent below and 21 percent above the emission estimate of 1.0
MMT CO2 Eq.
Table 4-59: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from
Phosphoric Acid Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Phosphoric Acid Production
CO2
1.0
0.9 1.3 -19% +21%
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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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.
Planned Improvements
EPA continues to evaluate potential improvements to the Inventory estimates for this source category, which include
direct integration of EPA's GHGRP data for 2010 through 2017 and the use of reported GHGRP data to update the
inorganic C content of phosphate rock for prior years. Confidentiality of data continues to be assessed, in addition to
the applicability of GHGRP data for the averaged inorganic C content data (by region) from 2010 through 2017 to
inform estimates in prior years in the required time series (i.e., 1990 through 2009). 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
Industrial Processes and Product Use 4-67
-------
national inventories will be relied upon.226 This long-term planned improvement is still in development by EPA and
have not been implemented into the current Inventory report.
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 iron227 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 CH4 emissions can also be generated from these processes, as well as from sinter, direct iron and
pellet production.
Currently, there are approximately nine integrated iron and steel steelmaking facilities that utilize BOFs to refine
and produce steel from iron. These facilities have 21 active blast furnaces between them as of 2015. Almost 100
steelmaking facilities utilize EAFs to produce steel primarily from recycled ferrous scrap (USGS 2018). 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 16 cokemaking facilities, of
which 3 facilities are co-located with integrated iron and steel facilities (ACCCI2016). In the United States, four
states - Indiana, Ohio, Michigan, and Pennsylvania - count for roughly 51 percent of total raw steel production
(USGS 2018).
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, Japan and India, accounting for approximately 4.8 percent of world production in 2017 (AISI 2004 through
2018).
226 See .
227 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.
4-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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 2017 were 0.6 MMT CO2 Eq. (578 kt CO2) (see Table 4-60
and Table 4-61). Emissions decreased significantly in 2017 by 56 percent from 2016 levels and have decreased by
77 percent (1.9 MMT CO2 Eq.) since 1990. Coke production in 2017 was 38 percent lower than in 2000 and 53
percent below 1990.
Table 4-60: CO2 Emissions from Metallurgical Coke Production (MMT CO2 Eq.)
Gas 1990
2005
2013 2014 2015 2016 2017
CO2 2.5
2.1
1.8 2.0 2.8 1.3 0.6
Total 2.5
2.1
1.8 2.0 2.8 1.3 0.6
Table 4-61: CO2 Emissions from Metallurgical Coke Production (kt)
Gas 1990
2005
2013 2014 2015 2016 2017
CO2 2,504
2,050
1,830 2,020 2,843 1,327 578
Total 2,504
2,050
1,830 2,020 2,843 1,327 578
Iron and Steel Production
Emissions of CO2 and CH4 from iron and steel production in 2017 were 41.2 MMT CO2 Eq. (41,201 kt) and 0.0073
MMT CO2 Eq. (0.3 kt CH4), respectively (see Table 4-62 through Table 4-65), totaling approximately 41.2 MMT
CO2 Eq. Emissions slightly increased in 2017 from 2016 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 2017, domestic production of pig iron did not change from 2016 levels. Overall, domestic pig iron production lias
declined since the 1990s. Pig iron production in 2017 was 53 percent lower than in 2000 and 55 percent below 1990.
Carbon dioxide emissions from iron production have decreased by 77 percent since 1990. Carbon dioxide emissions
from steel production have decreased by 16 percent (1.3 MMT CO2 Eq.) since 1990, while overall CO2 emissions
from iron and steel production have declined by 58 percent (57.9 MMT CO2 Eq.) from 1990 to 2017.
Table 4-62: CO2 Emissions from Iron and Steel Production (MMT CO2 Eq.)
Source/Activity
Data
1990 2005 2013
2014
2015
2016
2017
Sinter Production
2.4 | 1.7 | 1.1
1.1
1.0
0.9
0.9
Industrial Processes and Product Use 4-69
-------
Iron Production
45.7
17.7
12.0
18.7
11.8
9.9
10.4
Pellet Production
1.8
1.5
1.1
1.1
1.0
0.9
0.9
Steel Production
8.0
9.4
8.6
7.5
6.9
6.9
6.7
Other Activities3
41.2
35.9
28.7
27.9
24.3
22.5
22.4
Total
99.1
66.2
51.6
56.3
45.0
41.0
41.2
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.
Note: Totals may not sum due to independent rounding.
Table 4-63: CO2 Emissions from Iron and Steel Production (kt)
Source/Activity Data 1990
2005
2013 2014 2015 2016 2017
Sinter Production 2,448
Iron Production 45,704
Pellet Production 1,817
Steel Production 7,965
Other Activitiesa 41,193
1,663
17,664
1,503
9,396
35,934
1,117 1,104 1,016 877 869
12,031 18,722 11,780 9,928 10,386
1,146 1,126 964 869 867
8,638 7,469 6,941 6,858 6,691
28,709 27,911 24,280 22,451 22,390
Total 99,126
66,160
51,641 56,332 44,981 40,983 41,204
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.
Note: Totals may not sum due to independent rounding.
Table 4-64: ChU Emissions from Iron and Steel Production (MMT CO2 Eq.)
Source/Activity Data 1990
2005
2013 2014 2015 2016 2017
Sinter Production +
+
+ + + + +
Total +
+
+
+
+
+
+
+ Does not exceed 0.05 MMT CO2 Eq.
Table 4-65: ChU Emissions from Iron and Steel Production (kt)
Source/Activity Data
1990
2005 2013
2014
2015
2016
2017
Sinter Production
0.9
O
0
0.4
0.3
0.3
0.3
Total
0.9
T
O
O
0.4
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:
Em- —
^(<2a x Ca) - ^(<2fc X Cb)
44
12
where.
Ec02
a
b
Qa
Ca
Qb
Emissions from coke, pig iron, EAF steel, or BOF steel production, metric tons
Input material a
Output material b
Quantity of input material a. metric tons
Carbon content of input material a. metric tons C/metric ton material
Quantity of output material b, metric tons
4-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Cb
44/12
Carbon content of output material b, metric tons C/metric ton material
Stoichiometric ratio of CO2 to C
The Tier 1 methodology equations are as follows:
ES,P = Qs x EFs p
Ed,C02 — Qd X EFdŁQ2
Ep,co2 ~ Qp x EFp C02
where,
EFp,co2
Ed,C02
Qd
EFd,co2
QP
Emissions from sinter production process for pollutant p (CO2 or CH4), metric ton
Quantity of sinter produced, metric tons
Emission factor for pollutant p (CO2 or CH4), metric ton /Vmctric ton sinter
Emissions from DRI production process for CO2, metric ton
Quantity of DRI produced, metric tons
Emission factor for CO2, metric ton CCh/metric ton DRI
Quantity of pellets produced, metric tons
Emission factor for CO2, metric ton CCh/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.
Table 4-66: Material Carbon Contents for Metallurgical Coke Production
Material
kgC/kg
Coal Tar
0.62
Coke
0.83
Coke Breeze
0.83
Coking Coal
0.73
Material
kgC/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.
Although the 2006 IPCC Guidelines provide a Tier 1 CH4 emission factor for metallurgical coke production (i.e.,
0.1 g CH4 per metric ton of coke production), it is not appropriate to use because CO2 emissions were estimated
Industrial Processes and Product Use 4-71
-------
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 2018) (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
(AISI2004 through 2018) 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 coking coal, metallurgical coke, coal tar, coke oven gas, and blast
furnace gas were provided by the 2006IPCC Guidelines. The C content for coke breeze was assumed to equal the C
content of coke.
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
2013
2014
2015
2016
2017
Metallurgical Coke Production
Coking Coal Consumption at Coke Plants
35,269
21,259
19,481
19,321
17,879
14,955
15,910
Coke Production at Coke Plants
25,054
15,167
13,898
13,748
12,479
10,755
11,746
Coal Breeze Production
2,645
1,594
1,461
1,449
1,341
1,122
1,193
Coal Tar Production
1,058
638
584
580
536
449
477
Table 4-68: Production and Consumption Data for the Calculation of CO2 Emissions from
Metallurgical Coke Production (Million ft3)
Source/Activity Data
1990
2005
2013
2014
2015
2016
2017
Metallurgical Coke Production
Coke Oven Gas Production
250,767
114,213
108,162
102,899
84,336
74,807
74,997
Natural Gas Consumption
599
2,996
3,247
3,039
2,338
2,077
2,103
Blast Furnace Gas Consumption
24,602
4,460
4,255
4,346
4,185
3,741
3,683
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
4-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 [AISI2008]). 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
kgC/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
kgC/GJ
Coke Oven Gas
12.1
Blast Furnace Gas
70.8
Source: IPCC (2006), Table 4.3. Coke Oven Gas and
Blast Furnace Gas, Table 1.3.
The production process for sinter results in fugitive emissions of CH4, which are emitted via leaks in the production
equipment, rather than through the emission stacks or vents of the production plants. The fugitive emissions were
calculated by applying Tier 1 emission factors taken from the 2006 IPCC Guidelines for sinter production (see Table
4-70). Although the 1995 IPCC Guidelines (IPCC/UNEP/OECD/IEA 1995) provide a Tier 1 CH4 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
estimation of CH4 emissions is precluded. A preliminary analysis of facility-level emissions reported during iron
production further supports this assumption and indicates that CH4 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
CH4 through the consumption of fossil fuels (e.g., natural gas, etc.); however, these emission estimates are excluded
due to data limitations. Pending further analysis and resources, EPA may include these emissions in future reports to
enhance completeness. EPA is still assessing the possibility of including these emissions in future reports and have
not included this data in the current report.
Table 4-70: ChU Emission Factors for Sinter and Pig Iron Production
Material Produced
Factor
Unit
Sinter
0.07
kg CEU/metric ton
Source: IPCC (2006), Table 4.2.
Emissions of CO2 from sinter production, direct reduced iron production and pellet production were estimated by
multiplying total national sinter production and the total national direct reduced iron production by Tier 1 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.
Industrial Processes and Product Use 4-73
-------
Table 4-71: CO2 Emission Factors for Sinter Production, Direct Reduced Iron Production and
Pellet Production
Metric Ton
Material Produced
CCh/Metric 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 2017 were obtained from AISI's Annual
Statistical Report (AISI2004 through 2018) and through personal communications with AISI (AISI2008) (see
Table 4-72). In general, direct reduced iron (DRI) consumption data were obtained from the U.S. Geological Survey
(USGS)M'«era/s Yearbook- Iron and Steel Scrap (USGS 1991 through 2016) and personal communication with
the USGS Iron and Steel Commodity Specialist (Fenton 2015 through 2018). 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 2018) 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 2018) 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
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 2018) and through personal communications with AISI (AISI 2008). Data for EAF and BOF scrap steel, pig
iron, andDRI consumption were obtained from the USGSM'nerafa 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 2018) 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 EIA's Natural
Gas Annual (EIA 2016b). Carbon contents for direct reduced iron, EAF carbon electrodes, EAF charge carbon,
limestone, dolomite, pig iron, and steel were provided by the 2006 IPCC 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 Annua! 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 2018) 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 2013 2014 2015 2016 2017
Sinter Production
4-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Sinter Production
Direct Reduced Iron
Production
Direct Reduced Iron
Production
Pellet Production
Pellet Production
Pig Iron Production
Coke Consumption
Pig Iron Production
Direct Injection Coal
Consumption
EAF Steel Production
EAF Anode and Charge
Carbon Consumption
Scrap Steel
Consumption
Flux Consumption
EAF Steel Production
BOF Steel Production
Pig Iron Consumption
Scrap Steel
Consumption
Flux Consumption
BOF Steel Production
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
2013
2014
2015
2016
2017
Pig Iron Production
Natural Gas
Consumption
56,273
59,844
48,812
47,734
43,294
38,396
38,142
Fuel Oil Consumption
(thousand gallons)
163,397
16,170
17,468
16,674
9,326
6,124
4,352
Coke Oven Gas
Consumption
22,033
16,557
17,710
16,896
13,921
12,404
12,459
Blast Furnace Gas
Production
1,439,380
1,299,980
1,026,973
1,000,536
874,670
811,005
808,499
EAF Steel Production
Natural Gas
Consumption
15,905
19,985
10,514
9,622
8,751
3,915
8,105
BOF Steel Production
Coke Oven Gas
Consumption
3,851
524
568
524
386
367
374
Other Activities
Coke Oven Gas
Consumption
224,883
97,132
89,884
85,479
70,029
62,036
62,164
Blast Furnace Gas
Consumption
1,414,778
1,295,520
1,022,718
996,190
870,485
807,264
804,816
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.
12,239
516
60,563
24,946
49,669
1,485
67
42,691
319
33,511
47,307
14,713
576
43,973
8,315
1,303
50,096
13,832
37,222
2,573
1,127
46,600
695
52,194
34,400
11,400
582
42,705
5,583
3,350
38,198
9,308
30,309
2,675
1,122
47,300
771
52,641
29,600
7,890
454
34,238
5,521
4,790
37,538
11,136
29,375
2,425
1,062
48,873
771
55,174
23,755
5,917
454
33,000
5,079
4,790
32,146
7,969
25,436
2,275
1,072
44,000
998
49,451
20,349
4,526
454
29,396
4,385
C
28,967
7,124
22,293
1,935
1,120
C
998
52,589
C
C
408
25,888
4,347
C
28,916
7,101
22,395
2,125
1,127
C
998
55,825
C
C
408
25,788
Industrial Processes and Product Use 4-75
-------
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
consumption data. There is also some uncertainty to which materials produced are exported to Canada. As indicated
in the introduction to this section, the trend for integrated facilities has moved to more use of EAFs and fewer BOFs.
This trend may not be completely captured in the current data which also increases uncertainty. EPA currently uses
an uncertainty range of ±10 percent for the primary data inputs to calculate overall uncertainty from iron and steel
production, consistent with 2006 IPCC Guidelines. During EPA's discussion with AISI, AISI noted that an
uncertainty range of ±5 percent would be a more appropriate approximation to reflect their coverage of 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 2017 were estimated to be between 34.4 and 49.2 MMT CO2 Eq. at the 95 percent confidence level.
This indicates a range of approximately 18 percent below and 18 percent above the emission estimate of 41.8 MMT
CO2 Eq. Total CH4 emissions from metallurgical coke production and iron and steel production for 2017 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 19 percent below and 19 percent above the emission estimate of 0.007 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
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Metallurgical Coke & Iron
and Steel Production
CO2
41.8
34.4
49.2
-18%
+18%
Metallurgical Coke & Iron
and Steel Production
CH4
+
+
+
-19%
+19%
4-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
+ Does not exceed 0.05 MMT CO2 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 2017.
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.
Planned Improvements
Future improvements involve improving activity data and emission factor sources for estimating CO2 and CH4
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.228 This is a medium-term improvement and EPA estimates that earliest this
improvement could be incorporated is the 2020 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 2021 Inventory submission.
EPA also received comments during the Expert Review cycle of the 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 the previous Inventory (i.e., 1990 through 2016) and will require
some additional time to implement other substantive changes. In addition, the EPA is currently developing an iron
and steel carbon balance diagram to include in future inventory reports that will aid in the discussion of iron and
steel processes. This improvement is expected for the 2020 Inventory submission.
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
228 See .
Industrial Processes and Product Use 4-77
-------
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 + 7CO
While most of the carbon contained in the process materials is released to the atmosphere as CO2, a percentage is
also released as CH4 and other volatiles. The amount of CH4 that is released is dependent on furnace efficiency,
operation technique, and control technology.
When incorporated in alloy steels, ferroalloys are used to alter the material properties of the steel. Ferroalloys are
used primarily by the iron and steel industry, and production trends closely follow that of the iron and steel industry.
As of 2017, ten companies in the United States produce ferroalloys (USGS 2016a).
Emissions of CO2 from ferroalloy production in 2017 were 2.0 MMT CO2 Eq. (1,975 kt CO2) (see Table 4-75 and
Table 4-76), which is an 8 percent reduction since 1990. Emissions of CH4 from ferroalloy production in 2017 were
0.01 MMT CO2 Eq. (0.6 kt CH4), which is an 18 percent decrease since 1990.
Table 4-75: CO2 and ChU Emissions from Ferroalloy Production (MMT CO2 Eq.)
Gas
1'WO
2005
2013
2014
2015
2016
2017
CO2
1.4
1.8
1.9
2.0
1.8
2.0
CH4
+
+
+
+
+
+
Total
*> *>
1.4
1.8
1.9
2.0
1.8
2.0
+ Does not exceed 0.05 MMT CO2 Eq.
Table 4-76: CO2 and ChU Emissions from Ferroalloy Production (kt)
Gas
1990
2005
2013
2014
2015
2016
2017
CO2
CH4
2,152
1
1,392
+
1,785
1
1,914
1
1,960
1
1,796
1
1,975
1
+ Does not exceed 0.5 kt.
Methodology
Emissions of CO2 and CH4 from ferroalloy production were calculated229 using a Tier 1 method from the 2006
IPCC Guidelines by multiplying annual ferroalloy production by material-specific default emission factors provided
by IPCC (IPCC 2006). The Tier 1 equations for CO2 and CH4 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 CCVmetric ton specific
ferroalloy product
229 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.
4-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
ECHi = X EF^
i
where,
ECh4 = CH4 emissions, kg
MP! = Production of ferroalloy type metric tons
EFi = Generic emission factor for ferroalloy type kg CH i/mctric 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 CH4 estimates:
• Ferrosilicon, 25 to 55 percent Si and Miscellaneous Alloys, 32 to 65 percent Si - 2.5 metric tons
CCh/mctric ton of alloy produced; 1.0 kg CH i/mctric ton of alloy produced.
• Ferrosilicon, 56 to 95 percent Si - 4.0 metric tons CO;/metric ton alloy produced; 1.0 kg CH4/metric ton of
alloy produced.
• Silicon Metal - 5.0 metric tons CO:/metric ton metal produced; 1.2 kg CH4/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 2017 (see Table 4-77) were obtained from the U.S. Geological Survey
(USGS) through the Minerals Yearbook: Silicon (USGS 1996 through 2013) and the Mineral Industry Sun'eys:
Silicon (USGS 2014, 2015b, 2016b, 2017). 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 2017 (USGS 2013, 2014, 2015b, 2016b, 2017, 2018).
Table 4-77: Production of Ferroalloys (Metric Tons)
Year
Ferrosilicon
Ferrosilicon
Silicon Metal
Misc. Alloys
25%-55%
56%-95%
32-65%
1990
321,385
109,566
145,744
72,442
2005
123,000
86,100
148,000
NA
2013
164,229
144,908
158,862
NA
Industrial Processes and Product Use 4-79
-------
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
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.23" 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 CH4 emissions; however, specific furnace information was not available or included in the CH4 emission
estimates.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 4-78. Ferroalloy production
CO2 emissions from 2017 were estimated to be between 1.7 and 2.2 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 2.0
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
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper
Bound Bound
Lower
Bound
Upper
Bound
Ferroalloy Production
Ferroalloy Production
CO2
CH4
2.0
+
1.7 2.2
+ +
-12%
-12%
+12%
+12%
+ Does not exceed 0.05 MMT CO2 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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
230 Emissions and sinks of biogenic carbon are accounted for in the Land Use, Land-Use Change, and Forestry chapter.
4-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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, 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.231 This is a long-term planned improvement and
EPA is still assessing the possibility of incorporating this improvement into the national Inventory report. 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
2018). 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 (Na3AlF6). 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 1.2 MMT CO2 Eq. (1,205 kt) in 2017
(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
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 CO2 Eq.
kt
1990
6.8
6,831
2005
4.1
4,142
2013 3.3 3,255
2014 2.8 2,833
231 See .
Industrial Processes and Product Use 4-81
-------
2015
2.8
2,767
2016
1.3
1,334
2017
1.2
1,205
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 96 percent and 89 percent, respectively, to 0.7 MMT CO2
Eq. of CF4 (0.1 kt) and 0.4 MMT CO2 Eq. of C2F6 (0.03 kt) in 2017, 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 82 percent, while the combined CF4 and C2F6 emission rate (per metric
ton of aluminum produced) lias been reduced by 72 percent. Emissions decreased by approximately 18 percent
between 2016 and 2017 due to decreases in aluminum production. CF4 and C2F6 emissions per metric ton of
aluminum produced decreased between 2016 and 2017.
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
2013
2.3
0.7
3.0
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
1.1
Note: Totals may not sum due to
independent rounding.
Table 4-81: PFC Emissions from Aluminum Production (kt)
Year
CF4
C2F6
1990
2.4
0.3
2005
0.4
+
2013
0.3
0.1
2014
0.3
0.1
2015
0.2
+
2016
0.1
+
2017
0.1
+
+ Does not exceed 0.05 kt.
In 2017, U.S. primary aluminum production totaled approximately 0.7 million metric tons, a 9 percent decrease from
2016 production levels (USAA 2018). In 2017, two companies managed production at five operational primary
aluminum smelters. One smelter that previously announced a permanent shutdown changed its status to temporarily
shut down, and plans to start production again in 2018. Three smelters remained on standby throughout 2017 (USGS
2018). During 2017, monthly U.S. primary aluminum production was lower for every month in 2016 except August,
October, and December when compared to the corresponding months in 2016 (USAA 2018, 2017).
4-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
For 2018, total production for the January to August period was approximately 0.55 million metric tons compared to
0.49 million metric tons for the same period in 2017, a 10 percent increase (USAA 2018). Based on the increase in
production, process CO2 and PFC emissions are likely to be higher in 2018 compared to 2017 if there are no
significant changes in process controls at operational facilities.
Methodology
Process CO2 and PFC (i.e., CF4 and C2F6) emission estimates from primary aluminum production for 2010 through
2017 are available fromEPA's GHGRP—Subpart F (Aluminum Production) (EPA 2018). Under EPA's GHGRP,
facilities began reporting primary aluminum production process emissions (for 2010) in 2011; as a result, GHGRP
data (for 2010 through 2017) 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 Soderberg electrolysis cells, CO2 emissions from anode consumption during electrolysis in all prebake
and Soderberg 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).232 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., 2006IPCC 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.
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
2006 IPCC 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
232 Code of Federal Regulations, Title 40: Protection of Environment, Part 98: Mandatory Greenhouse Gas Reporting, Subpart
F—Aluminum Production. See .
Industrial Processes and Product Use 4-83
-------
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). ForPrebake 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 Soderberg 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 Soderberg and/or Prebake emission factors (metric ton of CO2
per metric ton of aluminum produced) from IPCC (2006).
Process PFC Emissions from Anode Effects
Smelter-specific PFC emissions from aluminum production for 2010 through 2017 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 2006 IPCC 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 = SxAE
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
4-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 2017 were obtained via USAA (USAA 2018). For 1990 through
2001, and 2006 (see Table 4-82) data were obtained from USGS A lineral Industry Surveys:. I lummum. Innual
Report (USGS 1995, 1998, 2000, 2001, 2002, 2007). For 2002 through 2005, and 2007 through 2016, national
aluminum production data were obtained from the USAA's Primary Aluminum Statistics (USAA 2004 through
2006, 2008 through 2017).
Table 4-82: Production of Primary Aluminum (kt)
Year
kt
1990
4,048
2005
2,478
2013
1,948
2014
1,710
2015
1,587
2016
818
2017
741
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 1.17 and 1.24 MMT CO2 Eq. at the 95 percent
confidence level. This indicates a range of approximately 3 percent below to 3 percent above the emission estimate
of 1.2 MMT CO2 Eq. Also, production-related CF4 emissions were estimated to be between 0.7 and 0.8 MMT CO2
Eq. at the 95 percent confidence level. This indicates a range of approximately 10 percent below to 10 percent above
the emission estimate of 0.7 MMT CO2 Eq. Finally, aluminum production-related C2F6 emissions were estimated to
be between 0.3 and 0.4 MMT CO2 Eq. at the 95 percent confidence level. This indicates a range of approximately 16
percent below to 17 percent above the emission estimate of 0.4 MMT CO2 Eq.
Industrial Processes and Product Use 4-85
-------
Table 4-83: Approach 2 Quantitative Uncertainty Estimates for CO2 and PFC Emissions from
Aluminum Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Aluminum Production
CO2
1.2
1.2
1.2
-3%
+3%
Aluminum Production
CF4
0.7
0.7
0.8
-10%
+10%
Aluminum Production
C2F6
0.4
0.3
0.4
-16%
+17%
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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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.
4.20 Magnesium Production and Processing
(CRF Source Category 2C4)
The magnesium metal production and casting industry uses sulfur hexafluoride (SF6) as a cover gas to prevent the
rapid oxidation of molten magnesium in the presence of air. Sulfur hexafluoride has been used in this application
around the world for more than thirty years. A dilute gaseous mixture of SF6 with dry air and/or carbon dioxide
(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 CChEq. (0.05 kt) of SF6, 0.1 MMT CO2 Eq. (0.1 kt) ofHFC-134a, and
0.003 MMT CO2 Eq. (3.1 kt) of CO2 in 2017. This represents a decrease of approximately 4 percent from total 2016
emissions (see Table 4-84) and a decrease in SF6 emissions by 5 percent. The decrease can be attributed to decrease
in secondary production SF6 emissions between 2016 and 2017 as reported through the GHGRP. In 2017, total
HFC-134a emissions increased from 0.096 MMT CO2 Eq. to 0.098 MMT CO2 Eq., or a 2 percent increase as
compared to 2016 emissions. This is mainly attributable to the increased use of this alternative for secondary
production. FK 5-1-12 emissions did not change substantially from 2016 levels. The emissions of the carrier gas,
CO2, increased from 2.7 kt in 2016 to 3.1 kt in 2017, or 14 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
2013
2014
2015
2016
2017
SFe
5.2
2.7
1.3
0.9
1.0
1.1
1.1
HFC-134a
0.0
0.0
0.1
0.1
0.1
0.1
0.1
CO2
+
+
+
+
+
+
+
FK 5-1-12"
0.0
0.0
+
+
+
+
+
Total
5.2
2.7
1.4
1.0
1.1
1.2
1.2
+ Does not exceed 0.05 MMT CO2 Eq.
a Emissions of FK 5-1-12 are not included in totals.
4-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-85: SF6, HFC-134a, FK 5-1-12 and CO2 Emissions from Magnesium Production and
Processing (kt)
Year
1990
2005
2013
2014
2015
2016
2017
SFo
0.2
0.1
0.1
+
+
+
+
HFC-134a
0.0
0.0
0.1
0.1
0.1
0.1
0.1
CO2
1.4
2.9
2.1
2.3
2.6
2.7
3.1
FK 5-1-12"
0.0
0.0
+
+
+
+
+
+ Does not exceed 0.05 kt
aEmissions of FK 5-1-12 are not included in totals.
Methodology
Emission estimates for the magnesium industry incorporate information provided by industry participants in EPA's
SF6 Emission Reduction Partnership for the Magnesium Industry as well as emissions data reported through subpart
T (Magnesium Production and Processing) of EPA's GHGRP. The Partnership started in 1999 and, in 2010,
participating companies represented 100 percent of U.S. primary and secondary production and 16 percent of the
casting sector production (i.e., die, sand, permanent mold, wrought, and anode casting). SF6 emissions for 1999
through 2010 from primary production secondary production (i.e., recycling), and die casting were generally
reported by Partnership participants. Partners reported their SF6 consumption which is assumed to be equivalent to
emissions. Along with SF6, some Partners 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 2017 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 2017 (EPA GHGRP). The
methodologies described below also make use of magnesium production data published by the U.S. Geological
Survey (USGS).
1990 through 1998
To estimate emissions for 1990 through 1998, industry SF6 emission factors were multiplied by the corresponding
metal production and consumption (casting) statistics from USGS. For this period, it was assumed that there was no
use of HFC-134a or FK 5-1-12 cover gases and hence emissions were not estimated for these alternatives.
Sulfur hexafluoride emission factors from 1990 through 1998 were based on a number of sources and assumptions.
Emission factors for primary production were available from U.S. primary producers for 1994 and 1995. The
primary production emission factors were 1.2 kg SF6 per metric ton for 1990 through 1993, and 1.1 kg SF6 per
metric ton for 1994 through 1997. The emission factor for secondary production from 1990 through 1998 was
assumed to be constant at the 1999 average Partner value. An emission factor for die casting of 4.1 kg SF6 per metric
ton which was available for the mid-1990s from an international survey (Gjestland and Magers 1996), was used for
years 1990 through 1996. For 1996 through 1998, the emission factor for die casting was assumed to decline linearly
to the level estimated based on Partner reports in 1999. This assumption is consistent with the trend in SF6 sales to
the magnesium sector that 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
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.
Industrial Processes and Product Use 4-87
-------
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 are 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 2008 through 2010, Partners did not account for all die casting tracked by USGS,
and, therefore, it was necessary to estimate the emissions of die casters who were not Partners. 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.
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 time-
series consistency. The emissions of carrier gases for permanent mold, wrought and anode processes are not
estimated in this Inventory.
4-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
2.71
2
1 1
a Weighted average includes all die casters, Partners and non-Partners. For
the majority of the time series (2000 through 2007), 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 SFo
per metric ton of magnesium for die casters that do not participate in the
Partnership.
2011 through 2017
For 2011 through 2017, for the primary and secondary producers and die casting, GHGRP-reported cover and
carrier gases emissions data were used. For sand casting, some emissions data was obtained through EPA's GHGRP.
The balance of the emissions for this industry segment was estimated based on previous Partner reporting (i.e., for
Partners that did not report emissions through EPA's GHGRP) or were estimated by multiplying emission factors by
the amount of metal produced or consumed. Partners who did not report through EPA's GHGRP were assumed to
have continued to emit SF6 at the last reported level, which was from 2010 in most cases, unless publicly available
sources indicated that these facilities have closed or otherwise eliminated SF6 emissions from magnesium production
(ARB 2015). 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 2017 was not yet available at the time of the analysis, so the 2016 values were held
constant through 2017 as a proxy.
Due to some GHGRP facilities originally submitting their GHGRP reports with errors, reporting their data late, or
not submitting data for 2017, some values were held constant at 2016 levels, affecting the overall calculations.
Uncertainty and Time-Series Consistency
Uncertainty surrounding the total estimated emissions in 2017 is attributed to the uncertainties around SF6, HFC-
134a, and CO2 emission estimates. To estimate the uncertainty surrounding the estimated 2017 SF6 emissions from
magnesium production and processing, the uncertainties associated with three variables were estimated: (1)
emissions reported by magnesium producers and processors for 2017 through EPA's GHGRP, (2) emissions
estimated for magnesium producers and processors that reported via the Partnership in prior years but did not report
2017 emissions through EPA's GHGRP, and (3) emissions estimated for magnesium producers and processors that
did not participate in the Partnership or report through EPA's GHGRP. An uncertainty of 5 percent was assigned to
the emissions (usage) data reported by each GHGRP reporter for all the cover and carrier gases (per the 2006IPCC
Guidelines). If facilities did not report emissions data during the current reporting year through EPA's GHGRP, SF6
emissions data were held constant at the most recent available value reported through the Partnership. The
uncertainty associated with these values was estimated to be 30 percent for each year of extrapolation. In 2017, a
higher proportion of emissions were estimated by holding values constant at the previous year's emissions as
compared to 2016, so the uncertainty of the 2017 total inventory estimate is relatively higher than it was in 2016.
Industrial Processes and Product Use 4-89
-------
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 1.05 and 1.21 MMT CO2 Eq. at
the 95 percent confidence level. This indicates a range of approximately 7 percent below to 7 percent above the
2017 emission estimate of 1.1 MMT CO2 Eq. The uncertainty estimates for 2017 are larger relative to the
uncertainty reported for 2016 in the previous Inventory. This is because, as discussed above, a larger proportion of
emissions from GHGRP reporters in 2017 were set equal to 2016 reported emissions due to late or non-verified
GHGRP reports.
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
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMTCCfcEq.) (%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Magnesium
Production
SFo, HFC-
134a, CO2
1.1
1.1 1.2
-7% +7%
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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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
One GHGRP-reported value for 2016 was revised due to a data verification issue. Additionally, the USGS revised
some of its production numbers for 2015, resulting in changes in SF6 emissions for die casting, sand casting, and
permanent mold. Lastly, based upon a review of historical activity data from various sources, EPA revised estimates
of non-Partner or non-GHGRP reporter die casting activity data to be zero metal produced from 2008 through 2017.
Planned Improvements
Cover gas research conducted over the last decade has found that SF6 used for magnesium melt protection can have
degradation rates on the order of 20 percent in die casting applications (Bartos et al. 2007). Current emission
estimates assume (per the 2006IPCC Guidelines) that all SF6 utilized is emitted to the atmosphere. Additional
research may lead to a revision of the 2006 IPCC Guidelines to reflect this phenomenon and until such time,
developments in this sector will be monitored for possible application to the Inventory methodology.
4-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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, 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 2017, 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 2017 the smelter processed no lead (USGS 2016, 2017).
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. Of
all the domestic secondary smelters operational in 2017, 11 smelters had capacities of 30,000 tons or more and were
collectively responsible for more than 95 percent of secondary lead production in 2017 (USGS 2017). 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 2017, 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 2017
(USGS 2017).
In 2017, total secondary lead production in the United States was slightly higher than that in 2016. A new secondary
lead refinery, located in Nevada, was completed in 2016 and production was expected to begin by the end of the
year. The plant was expected to produce about 80 tons per day of 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, and to
an increase in exports of spent starting-lighting-ignition lead-acid batteries that reduced the availability of scrap for
secondary smelters (USGS 2017).
As in 2016, U.S. primary lead production remained at production levels of zero for 2017, and has also decreased by
100 percent since 1990. This is due to the closure of the only domestic primary lead smelter in 2013 (year-end), as
stated previously. In 2017, U.S. secondary lead production increased from 2016 levels (increase of 7 percent), and
has increased by 16 percent since 1990 (USGS 1995 through 2017).
In 2017, U.S. lead production totaled 1,010,000 metric tons (USGS 2018). The resulting emissions of CO2 from
2017 lead production were estimated to be 0.5 MMT CO2 Eq. (455 kt) (see Table 4-88). At last reporting, the United
States was the third largest mine producer of lead in the world, behind China and Australia, accounting for
approximately 7 percent of world production in 2017 (USGS 2017).
Industrial Processes and Product Use 4-91
-------
Table 4-88: CO2 Emissions from Lead Production (MMT CO2 Eq. and kt)
Year MMT CP2 Eg.
kt
1990
0.5
516
2005
0.6
553
2013
2014
2015
2016
2017
0.5
0.5
0.5
0.5
0.5
546
459
473
450
455
After a steady increase in total emissions from 1995 to 2000, total emissions have gradually decreased since 2000
and are currently 7 percent lower than 1990 levels.
The methods used to estimate emissions for lead production233 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:
C02 Emissions = (DS x EFDS) + (5 x EFS)
For primary lead production using direct smelting, Sjardin (2003) and the IPCC (2006) provide an emission factor of
0.25 metric tons CCh/mctric ton lead. For secondary lead production, Sjardin (2003) and IPCC (2006) provide an
emission factor of 0.25 metric tons CO;/metric ton lead for direct smelting, as well as an emission factor of 0.2
metric tons CO;/metric ton lead produced for the treatment of secondary raw materials (i.e., pretreatment of lead
acid batteries). Since the secondary production of lead involves both the use of the direct smelting process and the
treatment of secondary raw materials, Sjardin recommends an additive emission factor to be used in conjunction
with the secondary lead production quantity. The direct smelting factor (0.25) and the sum of the direct smelting and
pretreatment emission factors (0.45) are multiplied by total U.S. primary and secondary lead production
respectively, to estimate 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 2017 activity data for primary and secondary lead production (see Table 4-89) were obtained from
the U.S. Geological Survey (USGS 1995 through 2018). The 2016 lead production value was also updated and is
summarized in Table 4-89 (USGS 2018).
233 EPA has not integrated aggregated facility-level Greenhouse Gas Reporting Program (GHGRP) information to inform these
estimates. Hie 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.
Methodology
where.
DS
S
EFds
EFS
Lead produced by direct smelting, metric ton
Lead produced from secondary materials
Emission factor for direct Smelting, metric tons CO;/metric ton lead product
Emission factor for secondary materials, metric tons CO;/metric ton lead product
4-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-89: Lead Production (Metric Tons)
Year Primary Secondary
1990 404,000 922,000
2005 143,000 1,150,000
2013
114,000
1,150,000
2014
1,000
1,020,000
2015
0
1,050,000
2016
0
1,000,000
2017
0
1,010,000
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 2017 were estimated to be between 0.4 and 0.5 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.
Table 4-90: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Lead
Production (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Lead Production
CO2
0.5
0.4
0.5
-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 2017. Details on the emission trends through time are described in more detail in the
Methodology section above.
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.
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
Industrial Processes and Product Use 4-93
-------
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.234
Initial review of activity data show that EPA's GHGRP Subpart R lead production data differ from those reported by
USGS by between 2 percent and 18 percent across the 2012 through 2017 time-series. Preliminary emissions
estimates differ by roughly the same percentages (ranging from 2 to 18 percent) across the same time period. EPA is
still reviewing available GHGRP data, differences in data reporting, and assessing the possibility of including this
planned improvement in future Inventory reports. Currently, GHGRP data is used for QA purposes and EPA expects
the earliest to begin incorporating GHGRP data would be the 2020 Inventory submission.
4.22 Zinc Production (CRF Source Category
2C6)
Zinc production in the United States consists of both primary and secondary processes. Of the primary and
secondary processes used in the United States, only the electrothermic and Waelz kiln secondary processes result in
non-energy carbon dioxide (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
234 See .
4-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
which was transported to their Monaca, PA facility where the products were smelted into refined zinc using
electrothennic 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 hydro metallurgical 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
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 2017 were estimated to be 1.0 MMT CO2 Eq. (1,009 kt CO2) (see Table
4-91). All 2017 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 2017, 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 MMT CO2 Eq. kt
1990 0.6 632
2005 1.0 1,030
2013
1.4
1,429
2014
1.0
956
2015
0.9
933
2016
0.9
925
2017
1.0
1,009
In 2017, United States primary and secondary refined zinc production were estimated to total 130,000 metric tons
(USGS 2018) (see Table 4-92). Domestic zinc mine production decreased by 9 percent in 2017, owing mostly to the
ongoing strike at the Lucky Friday Mine in Idaho and decreased output at the Red Dog Mine in Alaska (USGS
2018). Refined zinc production increased by 6 percent as a result of production resuming at the Middle Tennessee
Mines and increased production at the Clarksville, TN smelter (USGS 2018). Primary zinc production (primary slab
zinc) increased by five percent in 2017, while secondary zinc production in 2017 decreased by 13 percent relative to
2016.
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
2013 106,000 127,000 233,000
2014 110,000 70,000 180,000
2015 125,000 50,000 175,000
Industrial Processes and Product Use 4-95
-------
2016 111,000 15,000 126,000
2017 117,000 13,000 130,000
Methodology
The methods used to estimate non-energy CO2 emissions from zinc production235 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
EFdefeuit = 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
EF\yaelz Kiln : : : : * : : : " * '
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
EFgAp qtic? — 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
2017 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
235 EPA has not integrated aggregated facility-level Greenhouse Gas Reporting Program (GHGRP) information to inform these
estimates. The aggregated information (e.g., activity data and emissions) associated with Zinc Production did not meet criteria to
shield underlying confidential business information (CBI) from public disclosure.
4-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
values for 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'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 2017 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 2017 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 CO^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 COz/mctric 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 judgement from the USGS mineral commodity expert to assess approaches for splitting total production into
primary and secondary values. For 2016 and 2017, only one facility produced primary zinc. Primary zinc produced
from this facility was subtracted from the USGS 2016/2017 total zinc production statistic to estimate secondary zinc
production for these two 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
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
Industrial Processes and Product Use 4-97
-------
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 2017 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
C02 Eq.
Table 4-93: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Zinc
Production (MMT CO2 Eq. and Percent)
2017 Emission
Source
Gas
Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Zinc Production
CO2
1.0
0.8 1.2 -16% +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 consistency in emissions from 1990
through 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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.
Planned Improvements
Pending resources and prioritization of improvements for more significant sources, EPA will continue to evaluate
and analyze data reported under EPA's GHGRP that would be useful to improve the emission estimates and
category-specific QC for the Zinc Production source category, in particular considering completeness of reported
zinc production given the reporting threshold. Given the small number of facilities in the United States, particular
attention will be made to risks for disclosing CBI and ensuring time series consistency of the emissions estimates
presented in future Inventory reports, consistent with IPCC and UNFCCC guidelines. This is required as the facility-
level reporting data from EPA's GHGRP, with the program's initial requirements for reporting of emissions in
calendar year 2010, are not available for all inventory years (i.e., 1990 through 2009) as required for this Inventory.
In implementing improvements and integration of data from EPA's GHGRP, the latest guidance from the IPCC on
the use of facility-level data in national inventories will be relied upon.236 This is a long-term planned improvement
and EPA is still assessing the possibility of including this improvement in future Inventory reports.
236 See .
4-98 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
4.23 Semiconductor Manufacture (CRF Source
Category 2E1)
The semiconductor industry uses multiple greenhouse gases in its manufacturing processes. These include long-
lived fluorinated greenhouse gases used for plasma etching and chamber cleaning, fluorinated heat transfer fluids
used for temperature control and other applications, and nitrous oxide (N20) used to produce thin films through
chemical vapor deposition.
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 (SF6),
although other fluorinated compounds such as perfluoropropane (C3F8) and perfluorocyclobutane (c-C4F8) are also
used. The exact combination of compounds is specific to the process employed.
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 generated and emitted as a process byproduct. In some
cases, emissions of the byproduct gas can rival or even exceed emissions of the input gas, as is the case for NF3 used
in remote plasma chamber cleaning, which 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, and perfluoroalkylmorpholines. One percent or less consist of HFC, PFC 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 10,000.237
237 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.
Industrial Processes and Product Use 4-99
-------
For 2017, total GWP-weighted emissions of all fluorinated greenhouse gases and N20 from deposition, etching, and
chamber cleaning processes in the U.S. semiconductor industry were estimated to be 5.0 MMT CO2 Eq. These
emissions are presented in Table 4-94 and Table 4-95 below for the years 1990, 2005, and the period 2013 to 2017.
(Emissions of F-HTFs that are HFCs, PFCs or SF6 are presented in Table 4-94 and Table 4-95. Emissions of F-HTFs
that are not HFCs, PFCs or SF6 are presented in Table 4-95, Table 4-96, and Table 4-97 but are not included in
Inventory totals.) The rapid growth of this industry and the increasing complexity (growing number of layers)238 of
semiconductor products led to an increase in emissions of 153 percent between 1990 and 1999, when emissions
peaked at 9.1 MMTCO2 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 44 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 41 percent between 1990 and 2017.
Total emissions from semiconductor manufacture in 2017 were similar to 2016 emissions, decreasing by 1 percent.
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, N20 and F-
HTFs) from semiconductor manufacturing.239 Table 4-97 shows the emissions of the F-HTF compounds with the
highest emissions in tons based on reporting to EPA's GHGRP during years 2011 through 2017 240
Table 4-94: PFC, HFC, SFe, NF3, and N2O Emissions from Semiconductor Manufacture241
(MMT COz Eq.)
Year
1990
2005
2013
2014
2015
2016
2017
CF4
0.8
l.f
f .3
f .5
f .5
f .5
f .6
C2F6
2.0
2.0
f .5
f .4
f .3
f .2
f .2
C3Fs
+
O.f
O.f
O.f
O.f
O.f
O.f
C4Fs
0.0
O.f
O.f
O.f
O.f
O.f
O.f
HFC-23
0.2
0.2
0.3
0.3
0.3
0.3
0.4
SFo
0.5
0.7
0.7
0.7
0.7
0.8
0.7
NF3
+
0.5
0.5
0.5
0.6
0.6
0.6
Total F-GHGs
3.6
4.6
4.4
4.6
4.7
4.7
4.7
N2O +
O.f
0.2
0.2
0.2
0.2
0.2
HFC, PFC and SF6 F-
HTFs
0.0
+
+
+
+
+
+
Total
3.6
4.7
4.6
4.8
4.9
5.0
5.0
+ Does not exceed 0.05 MMT CO2 Eq.
Note: Totals may not sum due to independent rounding.
238 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.
239 Emissions data for HTFs (in tons of gas) from the semiconductor industry from 2011 through 2017 were obtained from the
EPA GHGRP annual facility emissions reports.
240 Many fluorinated heat transfer fluids consist of perfluoropolymetliylisopropyl 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 .
241 An extremely small portion of emissions from Semiconductor Manufacture are from the manufacturing of MEMs and
photovoltaic cells.
4-100 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 4-95: PFC, HFC, SFe, NF3, and N2O Emissions from Semiconductor Manufacture (kt)
Year
1990
2005
2013
2014
2015
2016
2017
CF4
0.11
0.15
0.17
0.20
0.21
0.21
0.22
C2F6
0.16
0.16
0.13
0.11
0.11
0.10
0.10
C3Fs
+
+
+
+
+
+
+
C4Fs
0.0
+
+
+
+
+
+
HFC-23
+
+
+
+
+
+
+
SFo
+
+
+
+
+
+
+
NF3
+
+
+
+
+
+
+
N2O
0.12
0.41
0.59
0.65
0.71
0.71
0.84
HFC, PFC and SF6
F-HTF s
0.00
+
+
+
+
+
+
Total
0.43
0.81
0.98
1.09
1.15
1.15
1.28
+ Does not exceed 0.05 kt.
Table 4-96: F-HTF Emissions Based on GHGRP Reporting (MMT CO2 Eq.)
Year
2011
2012
2013
2014
2015
2016
2017
HFCs
0.000
0.000
0.000
0.003
0.003
0.004
0.003
PFCs
0.002
0.002
0.000
0.003
0.003
0.004
0.003
SFo
0.000
0.000
0.000
0.021
0.013
0.012
0.017
Other F-HTFs
0.877
1.097
0.676
0.790
0.735
0.654
0.586
Total F-HTFs
0.879
1.099
0.676
0.816
0.754
0.673
0.609
Table 4-97: Top 10 F-HTF Compounds with Largest Emissions Based on GHGRP Reporting
(tons)
Fluorinated Heat Transfer
GHGRP-Reported Emissions (tons)
Fluid242
GWP
2011
2012
2013
2014
2015
2016
2017
Perfluorotripropylamine (3M™
Fluorinert™ FC-3283/FC-8270)
10,000
24.36
35.86
22.72
17.03
10.22
20.57
12.47
Perfluoroisopropylmorpholine
(3M™ Fluorinert™ FC-770)
10,000
12.27
9.27
10.09
7.16
3.13
7.35
5.11
PFPMIE fraction, BP 200 °C
(Solvay Galden™ HT-200)
10,000
5.81
20.71
9.49
2.21
1.58
6.41
2.20
3-ethoxy-l,l,l,2,3,4,4,5,5,6,6,6-
dodecafluoro-2-tritluoromethyl-
hexane (3M™ HFE-7500)
270
8.57
7.21
13.85
2.68
2.92
2.23
7.09
HFE-569sŁ2, (3M™ HFE-7200)
59
8.17
10.53
5.78
4.27
2.92
3.17
6.86
HFE-449sl (3M™ HFE-7100)
297
10.63
2.94
4.53
0.37
0.35
0.70
0.75
Perfluorotributylamine (PTBA,
3M™ Fluorinert™ FC40/FC-43)
10,000
10.52
3.77
1.45
0.80
0.25
1.35
1.38
PFPMIE fraction, BP 170 °C
(Solvay Galden™ HT-170)
10,000
3.37
6.93
0.57
0.55
0.93
2.22
1.95
PFPMIE fraction, BP 165 °C
(Solvay Galden™ D02-TS)
10,000
2.61
2.45
4.89
0.88
0.00
1.46
1.35
PFPMIE fraction, BP 110 °C
(Solvay Galden™ HT-110)
10,000
1.90
1.52
0.83
0.49
0.60
0.98
0.63
242Many 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
Industrial Processes and Product Use 4-101
-------
Additional Emissions from MEMS and PV
Similar to semiconductor manufacturing, the manufacturing of micro-electro-mechanical devices (MEMs) and
photovoltaic cells requires the use of multiple long-lived fluorinated greenhouse gases for various processes.
GHGRP-reported emissions from the manufacturing of MEMs and photovoltaic cells are available for the years
2011 to 2017. They are not included in the semiconductor manufacturing totals reported above. The emissions
reported by facilities manufacturing MEMs included emissions of C2F6, C3F8, C4F8, CF4, HFC-23, NF3, and SF6, and
were equivalent to only 0.08 percent to 0.40 percent of the total reported emissions from semiconductor
manufacturing in 2011 to 2017. These emissions ranged from 0.0038 to 0.0171 MMT CO2 Eq. from 2011 to 2017.
Similarly, emissions from manufacturing of photovoltaic cells were equivalent to only 0.23 percent and 0.15 percent
of the total reported emissions from semiconductor manufacturing in 2015 and 2016 respectively. Reported
emissions from photovoltaic cell manufacturing consisted of CF4, C2F6, C4F8, and CHF3.
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,
the emissions from MEMS manufacturing are likely being included in semiconductor totals.
Methodology
Emissions are based on data reported through Subpart I, Electronics Manufacture, of EPA's GHGRP, Partner
reported emissions data received through EPA's PFC243 Reduction/Climate Partnership, EPA's PFC Emissions
Vintage Model (PEVM)—a model that estimates industry emissions from etching and chamber cleaning processes
in the absence of emission control strategies (Burton and Beizaie 2001),244 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 2017 time series. Consequently,
fluorinated greenhouse gas (F-GHG) emissions from etching and chamber cleaning processes 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 2017. 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 2017.
Facility emissions of F-HTFs from semiconductor manufacturing are reported to EPA under its GHGRP and are
available for the years 2011 through 2017. 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 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).245 The 1990 to 1994 emissions are assumed to be uncontrolled, since reduction strategies such as
chemical substitution and abatement were yet to be developed.
243 In the context of the EPA Partnership and PEVM, PFC refers to perfluorocompounds, not perfluorocarbons.
244 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.
245 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.
4-102 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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),246 and (2) product type (discrete, memory or
logic).247 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).
To estimate N20 emissions, it is assumed the proportion of N20 emissions estimated for 1995 (discussed below)
remained constant for the period of 1990 through 1994.
1995 through 1999
For 1995 through 1999, total U.S. emissions were extrapolated from the total annual emissions reported by the
Partners (1995 through 1999). Partner-reported emissions are considered more representative (e.g., in terms of
capacity utilization in a given year) than PEVM-estimated emissions, and are used to generate total U.S. emissions
when applicable. The emissions reported by the Partners were divided by the ratio of the total capacity of the plants
operated by the Partners and the total capacity of all of the semiconductor plants in the United States; this ratio
represents the share of capacity attributable to the Partnership. This method assumes that Partners and non-Partners
have identical capacity utilizations and distributions of manufacturing technologies. Plant capacity data is contained
in the World Fab Forecast (WFF) database and its predecessors, which is updated quarterly. Gas-specific emissions
were estimated using the same method as for 1990 through 1994.
For this time period, the N20 emissions were estimated using an emission factor that was applied to the annual, total
U.S. TMLA manufactured. The emission factor was developed using a regression-through-the-origin (RTO) model:
GHGRP reported N20 emissions were regressed against the corresponding TMLA of facilities that reported no use
246 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).
247 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 4-103
-------
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.248 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
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).249 250 251
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
248 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.
249 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.
250 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.
251 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.
4-104 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.252 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 emissions
for non-Partners were estimated using the same method as for 2000 through 2006.
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.253 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 used in etch and clean
processes as well as emissions of fluorinated heat transfer fluids. (Fluorinated heat transfer fluids are used to control
process temperatures, thermally test devices, and clean substrate surfaces, among other applications.) They also
report N20 emissions from CVD and other processes. The F-GHGs and N20 were aggregated, by gas, across all
semiconductor manufacturing GHGRP reporters to calculate gas-specific emissions for the GHGRP-reporting
segment of the U.S. industry. At this time, emissions that result from heat transfer fluid use that are HFC, PFC and
SF6 are included in the total emission estimates from semiconductor manufacturing, and these GHGRP-reported
emissions have been compiled and presented in Table 4-94. F-HTF emissions resulting from other types of gases
(e.g., HFEs) are not presented in semiconductor manufacturing totals in Table 4-94 and Table 4-95 but are shown in
Table 4-96 and Table 4-97 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
252 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.
253 GaAs and Si technologies refer to the wafer on which devices are manufactured, which use the same PFCs but in different
ways.
Industrial Processes and Product Use 4-105
-------
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,254 if a site-specific DRE was indicated), and the fab-wide DREs reported in 20 1 4.255
To adjust emissions for facilities that abated emissions in 2011 through 2013, EPA first estimated the
percentage of gas passing through abatement systems for remote plasma clean in 2014 using the ratio of
emissions reported for CF4 and NF3.
• EPA then estimated the quantity of NF3 abated for remote plasma clean in 2014 using the ratio of emissions
reported for CF4 (which is not abated) and NF3. This abated quantity was then subtracted from the total
abated quantity calculated as described in the bullet above.
• To account for the resulting remaining abated quantity, EPA assumed that the percentage of gas passing
through abatement systems was the same across all remaining gas and process type combinations where
abatement was reported for 2014.
• The percentage of gas abated was then assumed to be the same in 2011 through 2013 (if the facility claimed
abatement that year) as in 2014 for each gas abated in 2014.
The revised emission factors and DREs were then applied to the estimated gas consumption for each facility by gas,
process type and wafer size.256
For the segment of the semiconductor industry that is below EPA's GHGRP reporting threshold, and for R&D
facilities, which are not covered by EPA's GHGRP, emission estimates are based on EPA-developed emission
factors for the F-GHGs and N20 and estimates of manufacturing activity. The new emission factors (in units of mass
of 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).257 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)258 were
regressed against the corresponding TMLA to estimate an aggregate F-GHG emissions factor (CO2 Eq./MSI
TMLA), and facility-reported N20 emissions were regressed against the corresponding TMLA to estimate a N20
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.
254 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.
255 If abatement information was not available for 2014 or the reported incorrectly in 2014, data from 2015 or 2016 was
substituted.
256 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.
257 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.)
258 Only seven gases were aggregated because inclusion of F-GHGs that are not reported in the Inventory results in
overestimation of emission factor that is applied to the various non-reporting subpopulations.
4-106 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
and the calculated non-reporting population emissions are summed to estimate the total emissions from
semiconductor manufacturing.
2013 and 2014
For 2013 and 2014, as for 2011 and 2012, F-GHG and N20 emissions data received through EPA's GHGRP were
aggregated, by gas, across all semiconductor-manufacturing GHGRP reporters to calculate gas-specific emissions
for the GHGRP-reporting segment of the U.S. industry. However, for these years WFF data was not available.
Therefore, an updated methodology that does not depend on the WFF derived activity data was used to estimate
emissions for the segment of the industry that are not covered by EPA's GHGRP. For the facilities that did not
report to the GHGRP (i.e., which are below EPA's GHGRP reporting threshold or are R&D facilities), emissions
were estimated based on the proportion of total U.S. emissions attributed to non-reporters for 2011 and 2012. EPA
used a simple averaging method by first estimating this proportion for both F-GHGs and N20 for 2011, 2012, and
2015 through 2017, resulting in one set of proportions for F-GHGs and one set for N20, and then applied the
average of each set to the 2013 and 2014 GHGRP reported emissions to estimate the non-reporters' emissions.
Fluorinated gas-specific, GWP-weighted emissions for non-reporters were estimated using the corresponding
reported distribution of gas-specific, GWP-weighted emissions reported through EPA's GHGRP for 2013 and 2014.
GHGRP-reported emissions in 2013 were adjusted to capture changes to the default emission factors and default
destruction or removal efficiencies used for GHGRP reporting 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 2017
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 2017, 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 2017 using TMLA from
WFF and a more comprehensive set of emissions, i.e., fabs with as well as without abatement control, as new
information about the use of abatement in GHGRP fabs and fab-wide were available. Fab-wide DREs represent total
fab C02 Eq.-weighted controlled F-GHG and N20 emissions (emissions after the use of abatement) divided by total
fab CO2 Eq.-weighted uncontrolled F-GHG and N20 emissions (emission prior to the use of abatement).
Using information about reported emissions and the use of abatement and fab-wide DREs, EPA was able to
calculate uncontrolled emissions (each total F-GHG and N20) for every GHGRP reporting fab. Using this, coupled
with TMLA estimated using methods described above (see 2011 through 2012), EPA derived emission factors by
year, gas type (F-GHG or N20), and wafer size (200 mm or 300 mm) by dividing the total annual emissions reported
by GHGRP reporters by the total TMLA estimated for those reporters. These emission factors were multiplied by
estimates of non-reporter TMLA to arrive at estimates of total F-GHG and N20 emissions for non-reporters for each
year. For each wafer size, the total F-GHG emissions were disaggregated into individual gases using the shares of
total emissions represented by those gases in the emissions reported to the GHGRP by unabated fabs producing that
wafer size.
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)
Industrial Processes and Product Use 4-107
-------
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 / 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 2006IPCC Guidelines. Partners are estimated to have accounted for between 56 and 79 percent of F-GHG
emissions from U.S. semiconductor manufacturing between 1995 and 2010, with the percentage declining in recent
years as Partners increasingly implemented abatement measures.
Estimates of operating plant capacities and characteristics for Partners and non-Partners were derived from the
Semiconductor Equipment and Materials Industry (SEMI) WFF (formerly World Fab Watch) database (1996
through 2012 and 2015) (e.g., Semiconductor Materials and Equipment Industry 2017). Actual worldwide capacity
utilizations for 2008 through 2010 were obtained from Semiconductor International Capacity Statistics (SICAS)
(SIA 2009 through 2011). Estimates of the number of layers for each linewidth was obtained from International
Technology Roadmap for Semiconductors: 2013 Edition (Burton and Beizaie 2001; ITRS 2007; ITRS 2008; ITRS
2011; ITRS 2013). PEVM utilized the WFF, SICAS, and ITRS, as well as historical silicon consumption estimates
published by VLSI. Actual quarterly U.S. capacity utilizations for 2011, 2012 and 2015 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 N20, respectively.
The uncertainty in ET presented in Table 4-98 below results from the convolution of four distributions of emissions,
each reflecting separate estimates of possible values of Erj.ghg, ER)N2o, Ekr,f-ghg, and Emr,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 ERj 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).259 The 2012 analysis did not take into account the use of
259 OnNovember 13, 2013, EPA published a final rule revising subpartl (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
4-108 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
abatement. For the industry segment that processed 200 mm wafers, estimates of uncertainties at a 95 percent CI
ranged from ±29 percent for C3F8 to ±10 percent for CF4. For the corresponding 300 mm industry segment,
estimates of the 95 percent CI ranged from ±36 percent for C4F8 to ±16 percent for CF4. These gas and wafer-
specific uncertainty estimates are applied to the total emissions 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
SF6. For facilities reporting partial abatement, the distribution of fraction of the gas fed through the abatement
device, for each gas, is assumed to be triangularly distributed as well. It is assumed that no more than 50 percent of
the gases are abated (i.e., the maximum value) and that 50 percent is the most likely value and the minimum is zero
percent. Consideration of abatement then resulted in four additional industry segments, two 200-mm wafer-
processing segments (one fully and one partially abating each gas) and two 300-mm wafer-processing segment (one
fully and the other partially abating each gas). Gas-specific emission uncertainties were estimated by convolving the
distributions of unabated emissions with the appropriate distribution of abatement efficiency for fully and partially
abated facilities using a Monte Carlo simulation.
The uncertainty in 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 (Erf-ghg).
The uncertainty in ER)N2o is obtained by assuming that the uncertainty in the emissions reported by each of the
GHGRP reporting facilities results from the uncertainty in quantity of N20 consumed and the N20 emission factor
(or utilization). Similar to analyses completed for subpart I (see Technical Support for Modifications to the
Fluorinated Greenhouse Gas Emission Estimation Method Option for Semiconductor Facilities under Subpart I,
docket EPA-HQ-OAR-2011-0028), the uncertainty of N20 consumed was assumed to be 20 percent. Consumption
of N20 for GHGRP reporting facilities was estimated by back-calculating from emissions reported and assuming no
abatement. The quantity of N20 utilized (the complement of the emission factor) was assumed to have a triangular
distribution with a minimum value of zero percent, mode of 20 percent and maximum value of 84 percent. The
minimum was selected based on physical limitations, the mode was set equivalent to the subpart I default N20
utilization rate for chemical vapor deposition, and the maximum was set equal to the maximum utilization rate found
in ISMI Analysis of Nitrous Oxide Survey Data (ISMI2009). The inputs were used to simulate emissions for each
of the GHGRP reporting, N20-emitting facilities. The uncertainty for the total reported N20 emissions was then
estimated by combining the uncertainties of each of the facilities reported emissions using Monte Carlo simulation.
The estimate of uncertainty in ENr, f-ghg and E\ r. N2o entailed developing estimates of uncertainties for the emissions
factors and the corresponding estimates of TMLA.
The uncertainty in TMLA depends on the uncertainty of two variables—an estimate of the uncertainty in the average
annual capacity utilization for each level of production of fabs (e.g., full scale or R&D production) and a
corresponding estimate of the uncertainty in the number of layers manufactured. For both variables, the distributions
of capacity utilizations and number of manufactured layers are assumed triangular for all categories of non-reporting
fabs. The most probable utilization is assumed to be 82 percent, with the highest and lowest utilization assumed to
be 89 percent, and 70 percent, respectively. For the triangular distributions that govern the number of possible layers
manufactured, it is assumed the most probable value is one layer less than reported in the ITRS; the smallest number
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 4-109
-------
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 N20 emissions from semiconductor manufacturing were
estimated to be between 4.7 and 5.2 MMT CO2 Eq. at a 95 percent confidence level. This range represents 5 percent
below to 5 percent above the 2017 emission estimate of 5.0 MMT CO2 Eq. 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
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimateb
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Bound0 Bound0
Lower Upper
Bound Bound
Semiconductor
Manufacture
HFC, PFC, SFe,
NF3, andN20
5.0
4.7 5.2
-5% +5%
a This uncertainty analysis does not include quantification of the uncertainty of emissions from heat transfer fluids.
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. While these emissions are included in the
semiconductor manufacturing F-GHG total emissions, they make up a considerably 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. In an effort to improve the uncertainty analysis for this
source category, HFC, PFC and SF6 emissions from the use of heat transfer fluids may be added in future inventory
years (see Planned Improvements section below). The emissions reported under EPA's GHGRP for 2014, 2015,
2016, and 2017, 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
4-110 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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
Emissions from 2011 through 2017 were updated to reflect updated emissions reporting in EPA's GHGRP, relative
to the previous Inventory. Additionally, as discussed above, GHGRP-reported emissions for 2011, 2012, and 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. Lastly, SF6 emissions from the use of
heat transfer fluids were separated from the "Other HTF" category and were estimated and reported independently.
Planned Improvements
This Inventory contains emissions estimates for N20 and for seven fluorinated gases emitted from etching and
chamber cleaning processes. However, other fluorinated gases (e.g., CsFx) are also emitted from etching and
chamber cleaning processes in much smaller amounts, accounting for less than 0.02 percent of emissions from these
processes. Previously, emissions data for these other fluorinated gases was not reported through the EPA
Partnership. However, through EPA's GHGRP, these data are available. Therefore, a point of consideration for
future Inventory reports is the inclusion of other fluorinated gases from etching and chamber cleaning processes.
In addition, EPA's GHGRP requires the reporting of emissions from other types of electronics manufacturing,
including MEMs, flat panel displays, and photovoltaic cells. There currently are seven MEMs manufacturers (most
of which report emissions for semiconductor and MEMs manufacturing separately), and no flat panel displays
manufacturing facilities reporting to EPA's GHGRP; one photovoltaic cell manufacturer previously reported to the
GHGRP.260 Emissions from MEMs and photovoltaic cell manufacturing could be included in totals in future
Inventory reports-currently they are not represented in Inventory emissions totals for electronics manufacturing.
These emissions could be estimated for the full time series (including prior to the GHGRP) and for MEMS and
photovoltaic cell manufacturers that are not reporting to the GHGRP; however, at this time the contribution to total
emissions is not significant enough to warrant the development of the methodologies that would be necessary to
back-cast these emissions to 1990 and estimate emissions for non-reporters for 2011 through 2017.
The Inventory methodology uses data reported through the EPA Partnership (for earlier years) and EPA's GHGRP
(for later years) to extrapolate the emissions of the non-reporting population. While these techniques are well
developed, the understanding of the relationship between the reporting and non-reporting populations is limited.
Further analysis of the reporting and non-reporting populations could aid in the accuracy of the non-reporting
population extrapolation in future years. In addition, the accuracy of the emissions estimates for the non-reporting
population could be further increased through EPA's further investigation of and improvement upon the accuracy of
estimated activity in the form of TMLA.
The Inventory uses utilization from two different sources for various time periods-SEMI to develop PEVM and to
estimate non-Partner emissions for the period 1995 to 2010 and U.S. Census Bureau for 2011 through 2014. SEMI
reported global capacity utilization for manufacturers through 2011. U. S. Census Bureau capacity utilization include
U.S. semiconductor manufacturers as well as assemblers. Further analysis on the impacts of using a new and
different source of utilization data could prove to be useful in better understanding of industry trends and impacts of
utilization data sources on historical emission estimates.
The current Inventory now includes HFC, PFC and SF6 emissions resulting the use of heat transfer fluids in the total
estimates of F-GHG emissions from semiconductor manufacturing. A point of consideration for future Inventory
260 Based upon information in the WFF, it appears that a small portion of GHGRP semiconductor reporters are manufacturing
both semiconductors and MEMS; however, these reporters are only reporting semiconductor emissions.
Industrial Processes and Product Use 4-111
-------
reports is the inclusion of the uncertainty surrounding these emissions in the source category uncertainty analysis
(see also uncertainty and time-series consistency).
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 1990.261 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 4-100.262
Table 4-99: Emissions of HFCs and PFCs from ODS Substitutes (MMT CO2 Eq.)
Gas
1990
2005
2013
2014
2015
2016
2017
HFC-23
0.0
+
+
+
+
+
+
HFC-32
0.0
0.3
2.8
3.4
3.9
4.6
5.3
HFC-125
+
9.0
36.5
40.0
43.4
47.0
50.0
HFC-134a
+
75.8
65.3
63.1
61.1
57.7
54.1
HFC-143a
+
9.4
25.7
26.9
27.6
28.3
28.0
HFC-236fa
0.0
1.2
1.4
1.4
1.3
1.2
1.2
CF4
0.0
+
+
+
+
+
+
Others3
0.3
6.5
10.0
10.4
11.7
12.9
14.0
Total
0.3
102.1
141.7
145.3
149.2
151.8
152.7
+ Does not exceed 0.05 MMT CO2 Eq.
3 Others represent an unspecified mix of HFCs and PFCs, which includes HFC-152a, HFC-227ea, EEFC-
245fa, 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 C0F14.
Note: Totals may not sum due to independent rounding.
Table 4-100: Emissions of HFCs and PFCs from ODS Substitution (Metric Tons)
Gas
1990
2005
2013
2014
2015
2016
2017
HFC-23
0
1
2
2
2
2
2
HFC-32
0
397
4,190
5,001
5,841
6,799
7,799
HFC-125
+
2,583
10,415
11,439
12,403
13,416
14,291
HFC-134a
+
52,974
45,644
44,095
42,735
40,358
37,846
HFC-143a
+
2,096
5,749
6,011
6,183
6,326
6,272
HFC-236fa
0
118
147
145
134
127
119
CF4
0
2
5
5
5
6
6
Others3
M
M
M
M
M
M
M
+ Does not exceed 0.5 MT.
M (Mixture of Gases)
261 [42 U.S.C § 7671, CAA Title VI]
262 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.
4-112 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
a Others represent an unspecified mix of HFCs and PFCs, which includes HFC-152a, HFC-227ea, HFC-
245fa, 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.
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.263 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 152.7 MMT CO2 Eq. emitted in 2017. 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 2017. The
end-use sectors that contributed the most toward emissions of HFCs and PFCs as ODS substitutes in 2017 include
refrigeration and air-conditioning (126.8 MMT CO2 Eq., or approximately 83 percent), aerosols (10.3 MMT CO2
Eq., or approximately 7 percent), and foams (11.2 MMT CO2 Eq., or approximately 7 percent). Within the
refrigeration and air-conditioning end-use sector, large retail food was the highest emitting end-use (34.3 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
2013
2014
2015
2016
2017
Refrigeration/Air C onditioning
+
89.7
120.0
122.5
124.8
126.5
126.8
Aerosols
0.3
7.6
10.5
10.8
11.0
10.7
10.3
Foams
+
2.1
7.4
7.9
9.3
10.3
11.2
Solvents
+
1.7
1.8
1.8
1.8
1.9
1.9
Fire Protection
+
1.1
2.1
2.2
2.3
2.4
2.5
Total
0.3
102.1
141.7
145.3
149.2
151.8
152.7
+ Does not exceed 0.05 MMT CO2 Eq.
Note: Totals may not sum due to independent rounding.
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,264 R-404A, and R-507A.265 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
263 R-404A contains HFC-125, HFC-143a, andHFC-134a.
264 R-410A contains HFC-32 and HFC-125.
265 R-507A, also called R-507, contains HFC-125 and HFC-143a.
Industrial Processes and Product Use 4-113
-------
atmosphere during equipment manufacture and operation (as a result of component failure, leaks, and purges), as
well as at 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). Many 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 has started to use 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.
Solvents
Chlorofluorocarbons, methyl chloroform (1,1,1 -trichloroethane or TCA), and to a lesser extent carbon tetrachloride
(CCI4) were historically used as solvents in a wide range of cleaning applications, including precision, electronics,
and metal cleaning. Since their phaseout, metal cleaning end-use applications have primarily transitioned to non-
fluorocarbon solvents and not-in-kind processes. The precision and electronics cleaning end-uses have transitioned
in part to high-GWP gases, due to their high reliability, excellent compatibility, good stability, low toxicity, and
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-
4-114 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
GWP option and 2-BTP is being considered. As fire protection equipment is tested or deployed, emissions of HFCs
occur.
Methodology
A detailed Vintaging Model of ODS-containing equipment and products was used to estimate the actual—versus
potential—emissions of various ODS substitutes, including HFCs and PFCs. The name of the model refers to the
fact that it tracks the use and emissions of various compounds for the annual "vintages" of new equipment that enter
service in each end-use. The Vintaging Model predicts ODS and ODS substitute use in the United States based on
modeled estimates of the quantity of equipment or products sold each year containing these chemicals and the
amount of the chemical required to manufacture and/or maintain equipment and products over time. Emissions for
each end-use were estimated by applying annual leak rates and release profiles, which account for the lag in
emissions from equipment as they leak over time. By aggregating the data for 67 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-Seri insistency
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 67
end-uses in the Vintaging Model. In order to calculate uncertainty, functional forms were developed to simplify
some of the complex "vintaging" aspects of some end-use sectors, especially with respect to refrigeration and air-
conditioning, and to a lesser degree, fire extinguishing. These sectors calculate emissions based on the entire lifetime
of equipment, not just equipment put into commission in the current year, thereby necessitating simplifying
equations. The functional forms used variables that included growth rates, emission factors, transition from ODSs,
change in charge size as a result of the transition, disposal quantities, disposal emission rates, and either stock for the
current year or original ODS consumption. Uncertainty was estimated around each variable within the functional
forms based on expert judgment, and a Monte Carlo analysis was performed. 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 percent of non-MDI aerosol propellant that is HFC-152a.
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 152.6 and 171.5 MMT CO2 Eq.
at the 95 percent confidence level. This indicates a range of approximately 0.04 percent below to 12.3 percent above
the emission estimate of 152.7 MMT CO2 Eq.
Table 4-102: Approach 2 Quantitative Uncertainty Estimates for HFC and PFC Emissions
from ODS Substitutes (MMT CO2 Eq. and Percent)
2017 Emission
Source
Gases
Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Substitution of Ozone
Depleting Substances
HFCs and
PFCs
152.7
152.6 171.5
-0.04% +12.3%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Industrial Processes and Product Use 4-115
-------
Methodological approaches were applied to the entire time series to ensure time-series consistency from 1990
through 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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.
Comparison of Reported Consumption to Modeled Consumption of HFCs
Data from EPA's Greenhouse Gas Reporting Program (GHGRP) 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 00—Suppliers of Industrial Greenhouse Gases and Subpart QQ—Importers and Exporters of
Fluorinated Greenhouse Gases Contained in Pre-Charged Equipment or Closed-Cell Foams were compared to the
modeled demand for new saturated HFCs (excluding HFC-23) used as ODS substitutes from the Vintaging Model.
The collection of data from suppliers of HFCs enables EPA to calculate the reporters' aggregated net supply-the
sum of the quantities of chemical produced or imported into the United States less the sum of the quantities of
chemical transformed (used as a feedstock in the production of other chemicals), destroyed, or exported from the
United States.266 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.
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 (2010 through 2016). 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 (VintagingModel 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.267 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
266 Chemical that is exported, transformed, or destroyed—unless otherwise imported back to the United States—will never be
emitted in the United States.
267 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.
4-116 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 ten 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, 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
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 2017 (U.S. EPA 2019a) and the chemical demand as
calculated by the Vintaging Model for the same time series. 2017 Subpart QQ GHGRP values are not yet publicly
available and are proxied to the average of 2010 through 2016 estimates.
Table 4-103: U.S. HFC Supply (MMT COz Eq.)
2010
2011
2012
2013
2014
2015
2016
2017
Reported Net Supply (GHGRP)
235
248
245
295
279
290
268
305
Industrial GHG Suppliers
235
241
227
278
254
264
240
285
HFCs in Products and Foams
NAa
7
18
17
25
26
28
20
Modeled Supply (Vintaging Model)
200
207
214
223
230
232
241
237
Percent Difference
-15%
-16%
-13%
-24%
-18%
-20%
-10%
-22%
NA (Not Available)
a Importers and exporters of fluorinated gases in products were not required to report 2010 data.
Industrial Processes and Product Use 4-117
-------
Figure 4-2: U.S. HFC Consumption (MMT CO2 Eq.)
Modeled Consumption
Reported Imports in Products
I Reported Bulk Supply
300
250
200
O
c_)
2010
2011
2012
2013
2014
2016
2017
150
100
As show n, the estimates from the Vintaging Model are lower than the GHGRP estimates by an average of 17
percent across the time series (i.e., 2010 through 2017). This difference is significantly greater than that reported in
the previous Inventory, due to a lower model estimate of consumption. The lower model estimates stem primarily
from changes made during a peer review of the Vintaging Model (see Recalculations Discussion below), calling into
question the accuracy and thoroughness of the changes made. Irrespective of these changes, 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.
• 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
4-118 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
required to report if either their total imports or their total exports of greenhouse gases are greater than or
equal to 25,000 metric tons of 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 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 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
Avg.
Avg.
Avg.
Avg.
Avg.
Avg.
Avg.
Reported Net
Supply (GHGRP)
Modeled Demand
(Vintaging Model)
242
204
247
211
270
219
287
227
285
231
279
236
287
239
Percent Difference
-16%
-14%
-19%
-21%
-19%
-15%
-17%
• 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, or destroyed than produced or 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 2017 emissions from that non-
modeled source (0.1 MMT CO2 Eq.) are much smaller than those for the ODS substitute sector (152.7
MMT C02 Eq.).
Industrial Processes and Product Use 4-119
-------
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 were included in response to a peer review conducted on
end-uses within the Refrigeration/Air Conditioning and Fire Protection sectors. (EPA 2018).
In the Refrigeration/Air Conditioning sector, updates included revisions to servicing leak rate assumptions for light-
duty vehicle and light-duty truck air conditioners and updates to the annual leak rate for road transport refrigeration
systems containing HFC refrigerant.
For the unitary air conditioning end-uses, charge sizes were adjusted for residential unitary systems, annual loss
rates were reduced for small and large commercial unitary AC systems, and disposal loss rates were reduced for
residential and small and large commercial unitary systems. In addition, HCFC-22 dry-shipped condensing units
were added to the residential unitary air conditioning end-use.
Within the Fire Protection sector, replacement ratios, growth rates, and annual loss rates for total flooding agents
and market transitions and lifetimes for total flooding and streaming agents were updated in response to the peer
review and comments received during the Public Review comment period for the 2017 Inventory (i.e., 1990 through
2015 report) for the Fire Protection sector.
Further improvements to the Vintaging Model were included in the current Inventory to reflect findings from
stakeholder outreach and research on the integral skin foam end-use. The integral skin foam blowing agent
transitions were revised to include HFC-245fa, hydrocarbons, oxygenated hydrocarbons, and HFOs. Current
assumptions in the domestic refrigeration foam end-use, including HFC and alternative blowing agent market
penetration and historical, current, and projected shipments of appliances, were also reviewed against recent data
from the Association of Home Appliance Manufacturers (AHAM) and EPA's Responsible Appliance Disposal
(RAD) program, but it was determined that current assumptions in the Vintaging Model were accurate and did not
require updates (EPA 2019b).
In response to comments from the UNFCCC Expert Review Team, initial (i.e., first-fill) emissions that occur during
manufacture or installation of Ref/AC equipment were modeled. Per IPCC (2006) guidance, first-fill emissions were
considered for all Ref/AC equipment that are charged with refrigerant within the United States, including those
which are produced for export, and excluding those that are imported pre-charged. (EPA 2019c)
Together, these updates decreased greenhouse gas emissions on average by 1.1 percent between 1990 and 2016.
Planned Improvements
Future improvements to the Vintaging Model are planned for the Foam Blowing and Aerosols sectors. A review of
blowing agent transition assumptions for Commercial Refrigeration Foam and the disaggregation of the rigid
polyurethane (PU): spray foam end-use into low-pressure, two-component spray foam and high-pressure, two-
component spray foam are anticipated to be completed by the 2020 submission.
The non-metered dose inhaler (non-MDI) aerosol end-use may be renamed to consumer aerosol and stock and
emission estimates will be updated to align with a recent national market characterization. In addition, a technical
aerosol end-use may be added to the aerosols sector, in order to capture a portion of the market that may not be
adequately encompassed by the current non-MDI aerosol end-use. These updates are anticipated to be completed by
the 2020 submission.
4-120 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
4.25 Electrical Transmission and Distribution
(CRF Source Category 2G1)
The largest use of sulfur hexafluoride (SF6), both in the United States and internationally, is as an electrical insulator
and interrupter in equipment that transmits and distributes electricity (RAND 2004). The gas has been employed by
the electric power industry in the United States since the 1950s because of its dielectric strength and arc-quenching
characteristics. It is used in gas-insulated substations, circuit breakers, and other switchgear. SF6 has replaced
flammable insulating oils in many applications and allows for more compact substations in dense urban areas.
Fugitive emissions of SF6 can escape from gas-insulated substations and switchgear through seals, especially from
older equipment. The gas can also be released during equipment manufacturing, installation, servicing, and disposal.
Emissions of SF6 from equipment manufacturing and from electrical transmission and distribution systems were
estimated to be 4.3 MMT CO2 Eq. (0.2 kt) in 2017. This quantity represents an 81 percent decrease from the
estimate for 1990 (see Table 4-105 and Table 4-106). There are two potential causes for this decrease: a sharp
increase in the price of SF6 during the 1990s and a growing awareness of the magnitude and enviromnental impact
of SF6 emissions through programs such as EPA's voluntary SF6 Emission Reduction Partnership for Electric Power
Systems (Partnership) and EPA's GHGRP. Utilities participating in the Partnership have lowered their emission
factor (kg SF6 emitted per kg of nameplate capacity) by more than 86 percent since the Partnership began in 1999. 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 31 percent from 2011 to 2017,268 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
22.8
0.3
23.1
2005
7.7
0.7
8.3
2013
4.0
0.4
4.4
2014
4.2
0.4
4.6
2015
3.8
0.3
4.1
2016
4.1
0.3
4.4
2017
4.0
0.3
4.3
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
1.0
2005
0.4
268 Analysis of emission trends from the GHGRP is imperfect due to an inconsistent group of reporters year to year.
Industrial Processes and Product Use 4-121
-------
2013
2014
2015
2016
2017
0.2
0.2
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 were 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 Guidelines269 (Although Equation 7.3 of the 2006IPCC Guidelines appears in the discussion of
substitutes for ozone-depleting substances, it is applicable to emissions from any long-lived pressurized equipment
that is periodically serviced during its lifetime.)
Emissions (kilograms SF6) = SF6 purchased to refill existing equipment (kilograms) + nameplate capacity of retiring
equipment (kilograms)270
Note that the above equation holds whether the gas from retiring equipment is released or recaptured; if the gas is
recaptured, it is used to refill existing equipment, thereby lowering the amount of SF6 purchased by utilities for this
purpose.
Gas purchases by utilities and equipment manufacturers from 1961 through 2003 are available from the RAND
(2004) survey. To estimate the quantity of SF6 released or recovered from retiring equipment, the nameplate
capacity of retiring equipment in a given year was assumed to equal 81.2 percent of the amount of gas purchased by
electrical equipment manufacturers 40 years previous (e.g., in 2000, the nameplate capacity of retiring equipment
was assumed to equal 81.2 percent of the gas purchased in 1960). The remaining 18.8 percent was assumed to have
been emitted at the time of manufacture. The 18.8 percent emission factor is an average of IPCC default SF6
emission rates for Europe and Japan for 1995 (IPCC 2006). The 40-year lifetime for electrical equipment is also
based on IPCC (2006). The results of the two components of the above equation were then summed to yield
estimates of global SF6 emissions from 1990 through 1999.
U.S. emissions between 1990 and 1999 are assumed to follow the same trajectory as global emissions during this
period. To estimate U.S. emissions, global emissions for each year from 1990 through 1998 were divided by the
estimated global emissions from 1999. The result was a time series of factors that express each year's global
emissions as a multiple of 1999 global emissions. Historical U.S. emissions were estimated by multiplying the factor
for each respective year by the estimated U.S. emissions of SF6 from electric power systems in 1999 (estimated to be
14.3 MMT C02 Eq.).
269 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 SFe during this time period, so it would not have been possible to
conceal sensitive sales information by aggregation.
270 Nameplate capacity is defined as the amount of SFe within fully charged electrical equipment.
4-122 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 SF6 sales data. The other factor that may affect the relationship between the RAND sales trends and actual
global emissions is the level of imports from and exports to Russia and China. SF6 production in these countries is
not included in the RAND survey and is not accounted for in any another manner by RAND. However, atmospheric
studies confirm that the downward trend in estimated global emissions between 1995 and 1998 was real (see the
Uncertainty discussion below).
1999 through 2017 Emissions from Electric Power Systems
Emissions from electric power systems from 1999 to 2017 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 2017 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 2017, Partner utilities, which for inventory purposes are defined as utilities that either
currently are or previously have been part of the Partnership,271 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 lateryears) through EPA's GHGRP (discussed
further below) rather than through the Partnership. In 2017, approximately 0.5 percent of the total emissions
attributed to Partner utilities were reported through Partnership reports. Approximately 88 percent of the total
emissions attributed to Partner utilities were reported and verified through EPA's GHGRP. Partners without verified
2017 data accounted for approximately 11 percent of the total emissions attributed to Partner utilities.272
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. Partners that are GHGRP reporters and have off-ramped (i.e., non-
reporting Partners), are still treated as Partners, and estimates are gap-filled based on the methodology as described
in this section.
271 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.
272 Only data reported as of August 20,2018 are used in the emission estimates for the prior year of reporting. For Partners that
did not report to the GHGRP, emissions were extrapolated based upon historical Partner-specific transmission mile growth rates,
and those Partners 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. In addition, EPA
manually reviewed the reported data and compared each facility's reported transmission miles with the corresponding quantity in
the UDI 2017 database (UDI 2017). In the first year of GHGRP reporting, EPA followed up with reporters where the discrepancy
between the reported miles and the miles published by UDI was greater than 10 percent, with a goal to improve data quality.
Industrial Processes and Product Use 4-123
-------
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 CO2 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 17 percent of U.S. transmission miles and 21
percent of estimated U.S. emissions from electric power system in 2017.273
GHGRP-only reporters that no longer report due to off-ramping 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.274 As noted
above, non-Partner emissions were reported to the EPA for the first time through its GHGRP in 2012 (representing
2011 emissions). This set of reported data was of particular interest because it provided insight into the emission rate
of non-Partners, which previously was assumed to be equal to the historical (1999) emission rate of Partners.
Specifically, emissions were estimated for Non-Reporters as follows:
• Non-Reporters. 1999 to 2011: First, the 2011 emission rates (per kg nameplate capacity and per
transmission mile) reported by Partners and GHGRP-Only Reporters were reviewed to determine whether
there was a statistically significant difference between these two groups. Transmission mileage data for
2011 was reported through GHGRP, with the exception of transmission mileage data for Partners that did
not report through GHGRP, which was obtained from UDI. It was determined that there is no statistically
significant difference between the emission rates of Partners and GHGRP-Only reporters; therefore, Partner
and GHGRP-Only reported data for 2011 were combined to develop regression equations to estimate the
emissions of Non-Reporters. Historical emissions from Non-Reporters were estimated by linearly
interpolating between the 1999 regression coefficient (based on 1999 Partner data) and the 2011 regression
coefficient.
• Non-Reporters. 2012 to Present: It was determined that there continued to be no statistically significant
difference between the emission rates reported by Partners and by GHGRP-Only Reporters. Therefore, the
emissions data from both groups were combined to develop regression equations for 2012. This was
repeated for 2013 through 2017 using Partner and GHGRP-Only Reporter data for each year.
o The 2017 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 66 percent of total U.S. transmission miles). The
regression equation for 2017 is:
273 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.
274 In the United States, SF6 is contained primarily in transmission equipment rated above 34.5 kV.
4-124 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Emissions (kg) = 0.226 x Transmission Miles
Table 4-107 below shows the percentage of transmission miles covered by reporters (i.e., associated with reported
data) and the regression coefficient for 1999 (the first year data was reported), and for 2011 through present (the
years with GHGRP reported data). The coefficient increased between 2015 and 2017.
Table 4-107: Transmission Mile Coverage (Percent) and Regression Coefficients (kg per
mile)
1999
2005
2013
2014
2015 2016
2017
Percentage of Miles Covered by Reporters
50%
51% 1
74%
75%
73% 68%
67%
Regression Coefficient3
0.71
0.35
0.23
0.23
0.21 0.21
0.23
a Regression coefficient for emissions is calculated utilizing transmission miles as the explanatory variable and emissions
as the response variable. Hie 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. In 2017,
one reporter was removed with abnormally high emissions as compared to the last several years.
Data on transmission miles for each Non-Reporter for the years 2000, 2003, 2006, and 2009, 2012, 2016 and 2017
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 growth rate grew to 1.7 percent from 2006 to 2009 as transmission miles
increased by more than 33,000 miles.
• The growth rate for 2009 through 2012 was calculated to be 1.2 percent as transmission miles grew yet
again by approximately 24,000 during this time period.
• The annual transmission mile growth rate for 2012 through 2017 was calculated to be 0.9 percent, as
transmission miles increased by approximately 26,000 miles.
Total Industry Emissions
As a final step, total electric power system emissions from 1999 through 2017 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 2017 Emissions from Manufacture of Electrical Equipment
Three different methods were used to estimate 1990 to 2017 emissions from original electrical equipment
manufacturers (OEMs).
• OEM emissions from 1990 through 2000 were derived by assuming that manufacturing emissions equaled
10 percent of the quantity of SF6 provided with new equipment. The 10 percent emission rate is the average
of the "ideal" and "realistic" manufacturing emission rates (4 percent and 17 percent, respectively)
identified in a paper prepared under the auspices of the International Council on Large Electric Systems
(CIGRE) in February 2002 (O'Connell et al. 2002). The quantity of SF6 provided with new equipment was
estimated based on statistics compiled by the National Electrical Manufacturers Association (NEMA).
These statistics were provided for 1990 to 2000.
• OEM emissions from 2000 through 2010 were estimated by (1) interpolating between the emission rate
estimated for 2000 (10 percent) and an emission rate estimated for 2011 based on reporting by OEMs
through the GHGRP (5.8 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 (155.48 MMT CO2 Eq. in 2010).
Specifically, the ratio of new nameplate capacity to total nameplate capacity of a subset of Partners for
Industrial Processes and Product Use 4-125
-------
which new nameplate capacity data was available from 1999 to 2010 was calculated. These ratios were
then multiplied by the total industry nameplate capacity estimate for each year to derive the amount of SF6
provided with new equipment for the entire industry. Additionally, to obtain the 2011 emission rate
(necessary for estimating 2001 through 2010 emissions), the estimated 2011 emissions (estimated using the
third methodology listed below) were divided by the estimated total quantity of SF6 provided with new
equipment in 2011. The 2011 quantity of SF6 provided with new equipment was estimated in the same way
as the 2001 through 2010 quantities.
• OEM emissions from 2011 through 2017 were estimated using the SF6 emissions from OEMs reporting to
the GHGRP, and an assumption that these reported emissions account for a conservative 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.275 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 3.7 and 5.0 MMT CO2 Eq. at the 95
percent confidence level. This indicates a range of approximately 14 percent below and 17 percent above the
emission estimate of 4.3 MMT CO2 Eq.
Table 4-108: Approach 2 Quantitative Uncertainty Estimates for SF6 Emissions from
Electrical Transmission and Distribution (MMT CO2 Eq. and Percent)
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to 2017 Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Electrical Transmission
and Distribution
SFo
4.3
3.7 5.0
-14% +17%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
In addition to the uncertainty quantified above, there is uncertainty associated with using global SF6 sales data to
estimate U.S. emission trends from 1990 through 1999. However, the trend in global emissions implied by sales of
SF6 appears to reflect the trend in global emissions implied by changing SF6 concentrations in the atmosphere. That
275 Uncertainty is assumed to be higher for the GHGRP-Only category, because 2011 is the first year that those utilities have
reported to EPA.
4-126 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 2016. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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 DiscusM^rs
The historical emissions estimated for this source category have undergone some revisions. SF6 emission estimates
for the period 1990 through 2016 were updated relative to the previous report based on revisions to interpolated and
extrapolated non-reported Partner data.276 For the current Inventory, historical estimates for the period 2011 through
2016 were also updated relative to the previous report based on revisions to reported historical data in EPA's
GHGRP.
In previous inventory years, non-reporter nameplate capacity was estimated by dividing the non-reporter emissions
by the average reporter leak rate. This reliance on calculated emission values to estimate nameplate capacity often
results in similar trends between the values. EPA reevaluated this methodology and developed a new approach that
relates nameplate capacity directly to transmission miles. Non-reporter nameplate capacity estimates were
recalculated by regressing reporter nameplate capacity and reporter transmission miles; the resulting coefficient was
applied to non-reporter transmission miles to determine non-reporter nameplate capacity.
Also in previous inventory years, a utility specific transmission miles growth rate was applied to determine
transmission miles for instances when a Partner utility did not report for a given year. However, when calculating
total transmission miles, a national annual growth rate, based on UDI data, was applied to extrapolate for a Partner
that did not report for a given year. These two separate approaches created an inconsistency with transmission mile
values used to arrive at a national total estimate and a utility-specific value. To ensure that these values did match,
EPA chose to apply the annual growth rate for all utilities to extrapolate for Partners who had not reported for a
given year.
As a result of the recalculations, SF6 emissions from electrical transmission and distribution increased by 0.66
percent for 2016 relative to the previous report, and SF6 nameplate capacity increased by 5.9 percent for 2016
relative to the previous report. On average, SF6 emission estimates for the entire time series decreased by
approximately 0.45 percent per year.
Planned Improvements
EPA is continuing research to improve the methodology for estimating non-reporter nameplate capacity, specifically
the distinction of the nameplate capacity of hermetically-sealed and non-hermetically sealed equipment. The current
methodology determines 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
276 The earlier year estimates within the time series (i.e., 1990 through 1998) were updated based on revisions to the 1999 U.S.
emission estimate because emissions for 1990 through 1998 are estimated by multiplying a series of annual factors by the
estimated U.S. emissions of SF6 from electric power systems in 1999 (see Methodology section).
Industrial Processes and Product Use 4-127
-------
hermetically sealed equipment and high-voltage equipment. This calculation is necessary for time-series consistency
as the partner-reported data from partnership in the prior years represents the end of year nameplate capacity.
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 is planning to
leverage this new reported data to apply an adjustment factor for the GHGRP-reported nameplate capacity totals for
2011 through 2016 to remove the nameplate capacity values attributed to the hermetically-sealed equipment.
Reported nameplate capacity totals prior to 2011 can be left as is, since it can be assumed that no hermetically sealed
equipment was reported in these totals by partners. This planned improvement will ensure better consistency of the
type of equipment nameplate capacity included in the time-series. Additionally, information on the type of new and
retiring equipment is expected 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.
Due to the GHGRP policy that allows reporters to "off-ramp" from the reporting program when their emissions
remain below certain levels for certain periods of time (e.g., below 25,000 MT CO2 Eq. for five years), the number
of electric power systems whose reports are used to develop regression coefficients and country-wide emissions
estimates is decreasing. While EPA continues to account for emissions from these electric power systems using the
estimation method for non-reporters, it is possible that their cessation of reporting could influence the value and/or
stability of the emission factors (per transmission mile) that are applied to non-reporters. EPA is planning to explore
whether this is the case. If so, EPA is planning to evaluate whether the current methodology for scaling emissions is
the best option.
Finally, EPA is exploring the possibility of discontinuing extrapolating emissions for Partners for which reported
estimates are not provided for a given length of time, e.g., for more than three or five consecutive years. Emissions
from these electric power systems would instead be estimated using the non-reporter methodology.
4.26 Nitrous Oxide from Product Uses (CRF
Source Category 2G3)
Nitrous oxide (N20) is a clear, colorless, oxidizing liquefied gas with a slightly sweet odor which is used in a wide
variety of specialized product uses and applications. The amount of N20 that is actually emitted depends upon the
specific product use or application.
There are a total of three N20 production facilities currently operating in the United States (Ottinger 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 N20 is as a
propellant in pressure and aerosol products, the largest application being pressure-packaged whipped cream. Small
quantities of N20 also are used in the following applications:
• Oxidizing agent and etchant used in semiconductor manufacturing;
• Oxidizing agent used, with acetylene, in atomic absorption spectrometry;
• Production of sodium azide, which is used to inflate airbags;
• Fuel oxidant in auto racing; and
• Oxidizing agent in blowtorches used by jewelers and others (Heydorn 1997).
Production of N20 in 2017 was approximately 15 kt (see Table 4-109).
Table 4-109: N2O Production (kt)
Year kt
1990 16
2005 15
4-128 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
2013 15
2014 15
2015 15
2016 15
2017 15
Nitrous oxide emissions were 4.2 MMT CO2 Eq. (14 kt N2O) in 2017 (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 N20. The use of
N2O as a propellant for whipped cream lias 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 MMT CO2 Eg. kt
Methodology
Emissions from N20 product uses were estimated using the following equation:
Epu = 2_P>XSaX ER")
a
Epu
N2O emissions from product uses, metric tons
P
Total U.S. production of N20, metric tons
a
specific application
Sa
Share of N20 usage by application a
ERa
Emission rate for application a. percent
The share of total quantity of N20 usage by end-use represents the share of national N20 produced that is used by
the specific subcategory (e.g., anesthesia, food processing). In 2017, the medical/dental industry used an estimated
86.5 percent of total N20 produced, followed by food processing propellants at 6.5 percent. All other categories
combined used the remainder of the N20 produced. This subcategory breakdown has changed only slightly over the
past decade. For instance, the small share of N20 usage in the production of sodium azide lias declined significantly
during the 1990s. Due to the lack of information on the specific time period of the phase-out in this market
subcategory, most of the N20 usage for sodium azide production is assumed to have ceased after 1996, with the
majority of its small share of the market assigned to the larger medical/dental consumption subcategory (Heydorn
1997). The N20 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 N20 in blood and other tissues, none of the N20 is assumed to be
metabolized during anesthesia and quickly leaves the body in exhaled breath. Therefore, an emission factor of 100
percent was used for this subcategory (IPCC 2006). For N20 used as a propellant in pressurized and aerosol food
Industrial Processes and Product Use 4-129
-------
products, none of the N20 is reacted during the process and all of the N20 is emitted to the atmosphere, resulting in
an emission factor of 100 percent for this subcategory (IPCC 2006). For the remaining subcategories, all of the N20
is consumed/reacted during the process, and therefore the emission rate was considered to be zero percent (Tupman
2003).
The 1990 through 1992 N20 production data were obtained from SRI Consulting's Nitrous Oxide, North America
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 N20 production to range between 13.6 and 18.1 thousand metric tons. Tupman (2003) provided a
narrower range (15.9 to 18.1 thousand metric tons) for 1996 that falls within the production bounds described by
Heydorn (1997). Tupman (2003) data are considered more industry-specific and current. Therefore, the midpoint of
the narrower production range was used to estimate N20 emissions for years 1993 through 2001 (Tupman 2003).
The 2002 and 2003 N20 production data were obtained from the Compressed Gas Association Nitrous Oxide Fact
Sheet and Nitrous Oxide Abuse Hotline (CGA 2002, 2003). These data were also provided as a range. For example,
in 2003, CGA (2003) estimates N20 production to range between 13.6 and 15.9 thousand metric tons. Due to the
unavailability of data, production estimates for years 2004 through 2017 were held constant at the 2003 value.
The 1996 share of the total quantity of N20 used by each subcategory was obtained from SRI Consulting's Nitrous
Oxide, North America report (Heydorn 1997). The 1990 through 1995 share of total quantity of N20 used by each
subcategory was kept the same as the 1996 number provided by SRI Consulting. The 1997 through 2001 share of
total quantity of N20 usage by sector was obtained from communication with a N20 industry expert (Tupman 2003).
The 2002 and 2003 share of total quantity of N20 usage by sector was obtained from CGA (2002, 2003). Due to the
unavailability of data, the share of total quantity of N20 usage data for years 2004 through 2017 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 N20 industry expert
(Tupman 2003). The emissions rate for all other subcategories was obtained from communication with a N20
industry expert (Tupman 2003). The emissions rate for the medical/dental subcategory was obtained from the 2006
IPCC Guidelines.
Uncertainty and Time-Series Consistency
The overall uncertainty associated with the 2017 N20 emission estimate from N20 product usage was calculated
using the 2006 IPCC Guidelines (2006) Approach 2 methodology. Uncertainty associated with the parameters used
to estimate N20 emissions include production data, total market share of each end use, and the emission factors
applied to each end use, respectively.
The results of this Approach 2 quantitative uncertainty analysis are summarized in Table 4-111. Nitrous oxide
emissions from N20 product usage were estimated to be between 3.2 and 5.2 MMT C02 Eq. at the 95 percent
confidence level. This indicates a range of approximately 24 percent below to 24 percent above the emission
estimate of 4.2 MMT C02 Eq.
Table 4-111: Approach 2 Quantitative Uncertainty Estimates for N2O Emissions from N2O
Product Usage (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
N2O from Product Uses
N2O
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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
4-130 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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 N20 product use subcategories that
accurately represent trends. This evaluation includes conducting a literature review of publications and research that
may provide additional details on the industry. This work is currently ongoing and thus the results have not been
incorporated into the current Inventory report.
Pending additional resources and planned improvement prioritization, EPA may also evaluate production and use
cycles, and the potential need to incorporate a time lag between production and ultimate product use and resulting
release of N20. Additionally, planned improvements include considering imports and exports of N20 for product
uses.
Finally, for future Inventories, EPA will examine data from EPA's GHGRP to improve the emission estimates for
the N20 product use subcategory. Particular attention will be made to ensure aggregated information can be
published without disclosing CBI and time-series consistency, as the facility-level reporting data from EPA's
GHGRP are not available for all inventory years as required in this Inventory. 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 UNFCCC277 request that information be provided on
precursor greenhouse gases, which include carbon monoxide (CO), nitrogen oxides (NOx), non-CH4 volatile organic
compounds (NMVOCs), and sulfur dioxide (S02). These gases are not direct greenhouse gases, but indirectly affect
terrestrial radiation absorption by influencing the formation and destruction of tropospheric and stratospheric ozone,
or, in the case of S02, by affecting the absorptive characteristics of the atmosphere. Additionally, some of these
gases may react with other chemical compounds in the atmosphere to form compounds that are greenhouse gases.
As some of industrial applications also employ thermal incineration as a control technology, combustion byproducts,
such as CO and NOx, are also reported with this source category. NMVOCs, commonly referred to as
"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
2017 are reported in Table 4-112. Sulfur dioxide emissions are presented in Section 2.3 of the Trends chapter and
Annex 6.3.
277 See .
Industrial Processes and Product Use 4-131
-------
Table 4-112: NOx, CO, and NMVOC Emissions from Industrial Processes and Product Use (kt)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
NOx
592
572
427
414
414
414
414
Industrial Processes
Other Industrial Processes3
343
437
307
300
300
300
300
Metals Processing
88
60
64
63
63
63
63
Chemical and Allied Product
Manufacturing
152
55
44
43
43
43
43
Storage and Transport
3
15
10
5
5
5
5
Miscellaneousb
5
2
3
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
CO
4,129
1,557
1,247
1,251
1,251
1,251
1,251
Industrial Processes
Metals Processing
2,395
752
600
553
553
553
553
Other Industrial Processes3
487
484
455
530
530
530
530
Chemical and Allied Product
Manufacturing
1,073
189
129
117
117
117
117
Miscellaneousb
101
32
48
42
42
42
42
Storage and Transport
69
97
13
7
7
7
7
Product Uses
Surface Coating
+
2
2
1
1
1
1
Other Industrial Processes3
4
0
0
0
0
0
0
Dry Cleaning
+
0
0
0
0
0
0
Degreasing
+
0
0
0
0
0
0
Graphic Arts
+
0
0
0
0
0
0
Non-Industrial Processes0
+
0
0
0
0
0
0
Other
NA
0
0
0
0
0
0
NMVOCs
7,638
5,849
3,855
3,816
3,816
3,816
3,816
Industrial Processes
Storage and Transport
1,352
1,308
724
613
613
613
613
Other Industrial Processes3
364
414
309
314
314
314
314
Chemical and Allied Product
Manufacturing
575
213
72
70
70
70
70
Metals Processing
111
45
28
26
26
26
26
Miscellaneousb
20
17
27
24
24
24
24
Product Uses
Surface Coating
2,289
1,578
1,104
1,134
1,134
1,134
1,134
Non-Industrial Processes0
1,724
1,446
1,012
1,039
1,039
1,039
1,039
Degreasing
675
280
196
202
202
202
202
Dry Cleaning
195
230
161
165
165
165
165
Graphic Arts
249
194
136
139
139
139
139
Other Industrial Processes3
85
88
61
63
63
63
63
Other
+
36
25
26
26
26
26
+ 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.
Note: Totals may not sum due to independent rounding.
4-132 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Methodology
Emission estimates for 1990 through 2017 were obtained from data published on the National Emission Inventory
(NEI) Air Pollutant Emission Trends web site (EPA 2018), and disaggregated based on EPA (2003). Data were
collected for emissions of CO, NOx, volatile organic compounds (VOCs), and SO2 from metals processing, chemical
manufacturing, other industrial processes, transport and storage, and miscellaneous sources. Emissions were
calculated either for individual source categories or for many categories combined, using basic activity data (e.g.,
the amount of raw material processed or the amount of solvent purchased) as an indicator of emissions. National
activity data were collected for individual categories from various agencies. Depending on the category, these basic
activity data may include data on production, fuel deliveries, raw material processed, etc.
Emissions for product use were calculated by aggregating product use data based on information relating to product
uses from different applications such as degreasing, graphic arts, etc. Emission factors for each consumption
category were then applied to the data to estimate emissions. For example, emissions from surface coatings were
mostly due to solvent evaporation as the coatings solidify. By applying the appropriate product-specific emission
factors to the amount of products used for surface coatings, an estimate of NMVOC emissions was obtained.
Emissions of CO and NOx under product use result primarily from thermal and catalytic incineration of solvent-
laden gas streams from painting booths, printing operations, and oven exhaust.
Activity data were used in conjunction with emission factors, which together relate the quantity of emissions to the
activity. Emission factors are generally available from the EPA's Compilation of Air Pollutant Emission Factors,
AP-42 (EPA 1997). The EPA currently derives the overall emission control efficiency of a source category from a
variety of information sources, including published reports, the 1985 National Acid Precipitation and Assessment
Program emissions inventory, and other EPA databases.
Uncertainty and Time-Series Consistency
Uncertainties in these estimates are partly due to the accuracy of the emission factors and activity data used. A
quantitative uncertainty analysis was not performed.
Methodological approaches were applied to the entire time series to ensure time-series consistency from 1990
through 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
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.
Industrial Processes and Product Use 4-133
-------
5. Agriculture
Agricultural activities contribute directly to emissions of greenhouse gases through a variety of processes. This
chapter provides an assessment of methane (CH4) and nitrous oxide (N20) emissions from enteric fermentation in
domestic livestock, livestock manure management, rice cultivation agricultural soil management, and field burning
of agricultural residues; as well as carbon dioxide (CO2) emissions from liming and urea fertilization (see Figure
5-1). Additional CO2, CH4 and N20 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-fann energy use are
reported in the Energy chapter.
Figure 5-1: 2017 Agriculture Chapter Greenhouse Gas Emission Sources (MMT CO2 Eq.)
Agricultural Soil Management
Enteric Fermentation
Manure Management
Rice Cultivation
Urea Fertilization
Liming
Field Burning of Agricultural Residues < 0.5
Agriculture as a Portion of
All Emissions
I
266
8.4%
25 50 75 100 125
MMT COz Eq.
150
175
200
In 2017, the Agriculture sector was responsible for emissions of 542.1 MMT CO2 Eq.,1 or 8.4 percent of total U.S.
greenhouse gas emissions.2 Methane emissions from enteric fermentation and manure management represent 26.7
percent and 9.4 percent of total CH4 emissions from anthropogenic activities, respectively. Of all domestic animal
types, beef and dairy cattle were by far the largest emitters of CH4. Rice cultivation and field burning of agricultural
residues were minor sources of CH4. Emissions of N20 by agricultural soil management through activities such as
fertilizer application and other agricultural practices that increased nitrogen availability in the soil was the largest
source of U.S. N20 emissions, accounting for 73.9 percent. Manure management and field burning of agricultural
residues were also small sources of N20 emissions. Urea fertilization and liming each accounted for 0.1 percent of
total CO2 emissions from anthropogenic activities.
1 Following the current reporting requirements under the United Nations Framework Convention on Climate Change (UNFCCC),
this Inventory report presents CO2 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, including Hawaii and Alaska; however, U.S.
Territories are not included.
Agriculture 5-1
-------
Table 5-1 and Table 5-2 present emission estimates for the Agriculture sector. Between 1990 and 2017, CO2 and
CH4 emissions from agricultural activities increased by 16.2 percent and 14.4 percent, respectively, while N20
emissions from agricultural activities fluctuated from year to year, but increased by 7.3 percent overall.
Table 5-1: Emissions from Agriculture (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
7.1
7.9
8.4
8.1
8.5
8.1
8.2
Urea Fertilization
2.4
3.5
4.4
4.5
4.7
4.9
5.1
Liming
4.7
4.3
3.9
3.6
3.7
3.2
3.2
CH4
217.4
239.5
235.3
234.9
239.9
247.3
248.7
Enteric Fermentation
164.2
168.9
165.5
164.2
166.5
171.9
175.4
Manure Management
37.1
53.7
58.1
57.8
60.9
61.5
61.7
Rice Cultivation
16.0
16.7
11.5
12.7
12.3
13.7
11.3
Field Burning of Agricultural Residues
0.1
0.2
0.2
0.2
0.2
0.2
0.2
N2O
265.7
271.1
282.7
279.7
295.4
285.8
285.2
Agricultural Soil Management
251.7
254.5
265.2
262.3
277.8
267.6
266.4
Manure Management
14.0
16.5
17.4
17.4
17.6
18.2
18.7
Field Burning of Agricultural Residues
+
0.1
0.1
0.1
0.1
0.1
0.1
Total
490.2
518.4
526.3
522.8
543.8
541.2
542.1
+ Does not exceed 0.05 MMT CO2 Eq.
Note: Totals may not sum due to independent rounding.
Table 5-2: Emissions from Agriculture (kt)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CO2
7,084
7,854
8,350
8,124
8,464
8,083
8,234
Urea Fertilization
2,417
3,504
4,443
4,515
4,728
4,877
5,051
Liming
4,667
4,349
3,907
3,609
3,737
3,206
3,182
CH4
8,697
9,579
9,412
9,397
9,597
9,892
9,946
Enteric Fermentation
6,566
6,755
6,620
6,568
6,661
6,875
7,018
Manure Management
1,486
2,150
2,322
2,311
2,435
2,461
2,467
Rice Cultivation
641
667
462
510
493
549
454
Field Burning of Agricultural Residues
4
7
8
8
8
8
8
N2O
892
910
949
939
991
959
957
Agricultural Soil Management
845
854
890
880
932
898
894
Manure Management
47
55
58
58
59
61
63
Field Burning of Agricultural Residues
+
+
+
+
+
+
+
+ Does not exceed 0.5 kt.
Note: Totals may not sum due to independent rounding.
Box 5-1: Methodological Approach for Estimating and Reporting U.S. Emissions and Removals
In following the United Nations Framework Convention on Climate Change (UNFCCC) requirement under Article
4.1 to develop and submit national greenhouse gas emission inventories, the emissions and removals presented in
this report and this chapter, are organized by source and sink categories and calculated using internationally-
accepted methods provided by the Intergovernmental Panel on Climate Change (IPCC) in the 2006IPCC Guidelines
for National Greenhouse Gas Inventories (2006 IPCC Guidelines). Additionally, the calculated emissions and
removals in a given year for the United States are presented in a common 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 do 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.
5-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
5.1 Enteric Fermentation (CRF Source Category
3A)
Methane is produced as part of normal digestive processes in animals. During digestion, microbes resident in an
animal's digestive system ferment food consumed by the animal. This microbial fermentation process, referred to as
enteric fermentation, produces CH4 as a byproduct, which can be exhaled or eructated by the animal. The amount of
CH4 produced and emitted by an individual animal depends primarily upon the animal's digestive system, and the
amount and type of feed it consumes.
Ruminant animals (e.g., cattle, buffalo, sheep, goats, and camels) are the major emitters of CH4 because of their
unique digestive system. Ruminants possess a rumen or large "fore-stomach" in which microbial fermentation
breaks down the feed they consume into products that can be absorbed and metabolized. The microbial fermentation
that occurs in the rumen enables them to digest coarse plant material that non-ruminant animals cannot. Ruminant
animals, consequently, have the highest CH4 emissions per unit of body mass among all animal types.
Non-ruminant animals (e.g., swine, horses, and mules and asses) also produce CH4 emissions through enteric
fermentation although this microbial fermentation occurs in the large intestine. These non-ruminants emit
significantly less CH4 on a per-animal-mass basis than ruminants because the capacity of the large intestine to
produce CH4 is lower.
In addition to the type of digestive system, an animal's feed quality and feed intake also affect CH4 emissions. In
general, lower feed quality and/or higher feed intake leads to higher CH4 emissions. Feed intake is positively
correlated to animal size, growth rate, level of activity and production (e.g., milk production, wool growth,
pregnancy, or work). Therefore, feed intake varies among animal types as well as among different management
practices for individual animal types (e.g., animals in feedlots or grazing on pasture).
Methane emission estimates from enteric fermentation are provided in Table 5-3 and Table 5-4. Total livestock CH4
emissions in 2017 were 175.4 MMT CO: Eq. (7,018 kt). Beef cattle remain the largest contributor of CH4 emissions
from enteric fermentation accounting for 72 percent in 2017. Emissions from dairy cattle in 2017 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
2013
2014
2015
2016
2017
Beef Cattle
119.1
125.2
118.0
116.5
118.0
123.0
126.3
Dairy Cattle
39.4
37.6
41.6
42.0
42.6
43.0
43.3
Swine
2.0
2.3
2.5
2.4
2.6
2.6
2.7
Horses
1.0
1.7
1.6
1.6
1.5
1.5
1.4
Sheep
2.3
1.2
1.1
1.0
1.1
1.1
1.1
Goats
0.3
0.4
0.3
0.3
0.3
0.3
0.3
American Bison
0.1
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
165.5
164.2
166.5
171.9
175.4
+ Does not exceed 0.05 MMT CO2 Eq.
Note: Totals may not sum due to independent rounding.
Table 5-4: ChU Emissions from Enteric Fermentation (kt)
Livestock Type
1990
2005
2013
2014
2015
2016
2017
Beef Cattle
4,763 1
5,007 1
4,722
4,660
4,722
4,919
5,052
Dairy Cattle
1,574
1,503
1,664
1,679
1,706
1,722
1,730
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.
Agriculture 5-3
-------
Swine
81
92
98
96
102
105
108
Horses
40
70
64
62
61
59
58
Sheep
91
49
43
42
42
42
42
Goats
13
14
13
12
12
11
11
American Bison
4
17
13
13
13
13
13
Mules and Asses
1
2
3
3
3
3
3
Total
6,566
6,755
6,620
6,568
6,661
6,875
7,018
Note: Totals may
not sum due to
independent rounding.
From 1990 to 2017, emissions from enteric fermentation have increased by 6.9 percent. Emissions have also
increased from 2016 to 2017 by 2.1 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, beef
cattle emissions increased 6.1 percent from 1990 to 2017, while the national total beef cattle population slightly
decreased (by 0.02 percent) from 1990 to 2017. Furthermore, while dairy cattle emissions increased 9.9 percent
over the entire time series, the population has declined by 3.2 percent, and milk production increased 44 percent
(USDA 2018). These trends indicate that while emissions per head are increasing, emissions per unit of product (i.e.,
meat, milk) are going down.
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 underwent increases 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 2017,
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 utilized by producers across the U.S.). Dairy
cattle emissions have continued to trend upward since 2007, in line with dairy cattle population increases. Regarding
trends in other animals, populations of sheep have steadily declined, with an overall decrease of 54 percent since
1990. Horse populations are 45 percent greater than they were in 1990, but their numbers have been declining by
about 2 percent annually since 2007. Goat populations increased by about 20 percent through 2007, then steadily
decreased through 2017. Swine populations have trended upward through most of the time series, increasing 34
percent from 1990 to 2017. The population of American bison more than tripled over the 1990 to 2017 time period,
while the population of mules and asses increased by nearly 5 times.
Methodology
Livestock enteric fermentation emission estimate methodologies fall into two categories: cattle and other
domesticated animals. Cattle, due to their large population, large size, and particular digestive characteristics,
account for the majority of enteric fermentation CH4 emissions from livestock in the United States. A more detailed
methodology (i.e., IPCC Tier 2) was therefore applied to estimate emissions for all cattle. Emission estimates for
other domesticated animals (horses, sheep, swine, goats, American bison, and mules and asses) were handled using a
less detailed approach (i.e., IPCC Tier 1).
While the large diversity of animal management practices cannot be precisely characterized and evaluated,
significant scientific literature exists that provides the necessary data to estimate cattle emissions using the IPCC
Tier 2 approach. The Cattle Enteric Fermentation Model (CEFM), developed by EPA and used to estimate cattle
CH4 emissions from enteric fermentation incorporates this information and other analyses of livestock population
feeding practices, and production characteristics.
Inventory Methodology for Cattle
National cattle population statistics were disaggregated into the following cattle sub-populations:
• Dairy Cattle
o Calves
5-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
o Heifer Replacements
o Cows
• Beef Cattle
o Calves
o Heifer Replacements
o Heifer and Steer Stackers
o Animals in Feedlots (Heifers and Steer)
o Cows
o Bulls
Calf birth rates, end-of-year population statistics, detailed feedlot placement information, and slaughter weight data
were used to create a transition matrix that models cohorts of individual animal types and their specific emission
profiles. The key variables tracked for each of the cattle population categories are described in Annex 3.10. These
variables include performance factors such as pregnancy and lactation as well as average weights and weight gain.
Annual cattle population data were obtained from the U.S. Department of Agriculture's (USDA) National
Agricultural Statistics Service (NASS) QuickStats database (USDA 2016).
Diet characteristics were estimated by region for dairy, grazing beef, and feedlot beef cattle. These diet
characteristics were used to calculate digestible energy (DE) values (expressed as the percent of gross energy intake
digested by the animal) and CH4 conversion rates (Ym) (expressed as the fraction of gross energy converted to CH4)
for each regional population category. The IPCC recommends Ym ranges of 3.0±1.0 percent for feedlot cattle and
6.5±1.0 percent for other well-fed cattle consuming temperate-climate feed types (IPCC 2006). Given the
availability of detailed diet information for different regions and animal types in the United States, DE and Ym
values unique to the United States were developed. The diet characterizations and estimation of DE and Ym values
were based on information from state agricultural extension specialists, a review of published forage quality studies
and scientific literature, expert opinion, and modeling of animal physiology.
The diet characteristics for dairy cattle were based on Donovan (1999) and an extensive review of nearly 20 years of
literature from 1990 through 2009. Estimates of DE were national averages based on the feed components of the
diets observed in the literature for the following year groupings: 1990 through 1993, 1994 through 1998, 1999
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
4 Due to inconsistencies in the 2003 literature values, the 2002 values were used for 2003, as well.
Agriculture 5-5
-------
of age by Soliva (2006). Estimates of Ym for 5 and 6 month old dairy calves are linearly interpolated from the values
provided for 4 and 7 months.
To estimate CH4 emissions, the population was divided into state, age, sub-type (i.e., dairy cows and replacements,
beef cows and replacements, heifer and steer stackers, heifers and steers in feedlots, bulls, beef calves 4 to 6 months,
and dairy calves 4 to 6 months), and production (i.e., pregnant, lactating) groupings to more fully capture differences
in CH4 emissions from these animal types. The transition matrix was used to simulate the age and weight structure
of each sub-type on a monthly basis in order to more accurately reflect the fluctuations that occur throughout the
year. Cattle diet characteristics were then used in conjunction with Tier 2 equations from IPCC (2006) to produce
CH4 emission factors for the following cattle types: dairy cows, beef cows, dairy replacements, beef replacements,
steer stackers, heifer stackers, 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.
Inventory Methodology for Non-Cattle Livestock
Emission estimates for other animal types were based on average emission factors representative of entire
populations of each animal type. Methane emissions from these animals accounted for a minor portion of total CH4
emissions from livestock in the United States from 1990 through 2017. 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 2017 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 2017 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 CH4 emissions from
enteric fermentation.
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). There have been no significant changes to the methodology since that time; consequently, these
uncertainty estimates were directly applied to the 2017 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.
5-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 5-5.
Based on this analysis, enteric fermentation CH4 emissions in2017 were estimated to be between 156.1 and 207.0
MMT CO2 Eq. at a 95 percent confidence level, which indicates a range of 11 percent below to 18 percent above the
2017 emission estimate of 175.4 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 USD A population estimates used for swine, goats, and
sheep. The horse populations are now from the same USD A 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 5-5: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from Enteric
Fermentation (MMT CO2 Eq. and Percent)
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3'b'c
(MMT CO2 Eq.)
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Enteric Fermentation
CH4
175.4
156.1 207.0 -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 2017 estimates.
c The overall uncertainty calculated in 2003, and applied to the 2017 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.
Details on the emission trends through time are described in more detail in the Methodology section.
QA/QC and Verification
In order to ensure the quality of the emission estimates from enteric fermentation, the General (IPCC Tier 1) and
category-specific (Tier 2) Quality Assurance/Quality Control (QA/QC) procedures were implemented consistent
with the U.S. Inventory QA/QC plan outlined in Annex 8. Category-specific or Tier 2 QA procedures included
independent review of emission estimate methodologies from previous inventories. 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
In the previous Inventory, 1990 to 2015 estimates were retained from the prior Inventory (i.e., 1990 through 2015
Inventory), and 2016 estimates were based on a simplified approach that used emission factors and extrapolated
population estimates for all animals. For the current Inventory, the CEFM was used for cattle for all years, resulting
in different estimates for 2016 than the prior Inventory. For non-cattle livestock in the current Inventory, updated
Tier 1 estimates were used for 2016, yielding different results than the simplified approach used for 2016 in the prior
Inventory.
Agriculture 5-7
-------
There were also changes to emissions resulting from activity data changes, including:
• The USDA published minor revisions in several categories that impacted emissions estimated for cattle for
2015, including the following:
o Calf birth data were revised;
o Dairy cow milk production values were revised for several states;
o Slaughter values were revised for steers and heifers.
• The USDA also revised population estimates for some categories of non-cattle animals, which impacted
emissions estimated for "other" livestock. Populations for market and breeding swine were changed for
some states for 2015.
• American Bison populations from the 2012 Census were carried over for 2013 through 2017 values instead
of using predictive estimates of the populations. This change yielded different emissions estimates for 2013
through 2016 for American Bison as compared to the previous Inventory.
These recalculations had an insignificant impact on the overall emission estimates.
Planned Improvements
Continued research and regular updates are necessary to maintain an emissions inventory that reflects the current
base of knowledge. Depending upon the outcome of ongoing investigations, future improvements for enteric
fermentation could include some of the following options (options below are medium- to long-term improvements):
• 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 and beef cow
lactation rates;
• Investigation of the availability of annual data for the DE, Ym, and crude protein values of specific diet and
feed components for grazing and feedlot animals;
• Further investigation on additional sources or methodologies for estimating DE for dairy cattle, given the
many challenges in characterizing dairy cattle diets;
• Further evaluation of the assumptions about weights and weight gains for beef cows, such that trends
beyond 2007 are updated, rather than held constant;
• Further evaluation of the estimated weight for dairy cows (i.e., 1,500 lbs) that is based solely on Holstein
cows as mature dairy cow weight is likely slightly overestimated, based on knowledge of the breeds of
dairy cows in the United States;
• 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
5-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• 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 current (i.e., 1990 through 2017) Inventory
regarding the CEFM model, data and assumptions used to calculate enteric fermentation beef cattle emissions. Many
of the comments received reflect potential planned improvement options listed above, of which EPA is investigating
and working with USD A and other experts to utilize the best available data and methods for estimating emissions.
As noted, many of these improvements are major updates and may take multiple years to implement in full, but EPA
will work to add clarity to improve the transparency of future inventories.
In future Inventory reports, the final 2019 Refinement to the 2006IPCC Guidelines [currently in draft] will be
reviewed and any changes will be incorporated, as applicable, to update the current Inventory estimation
methodologies.
5.2 Manure Management (CRF Source
Category 3B)
The treatment, storage, and transportation of livestock manure can produce anthropogenic CH4 and N20 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 CH4. Ambient temperature,
moisture, and manure storage or residency time affect the amount of CH4 produced because they influence the
growth of the bacteria responsible for CH4 formation. For non-liquid-based manure systems, moist conditions
(which are a function of rainfall and humidity) can promote CH4 production. Manure composition, which varies by
animal diet, growth rate, and animal type (particularly the different animal digestive systems), also affects the
amount of CH4 produced. In general, the greater the energy content of the feed, the greater the potential for CH4
emissions. However, some higher-energy feeds also are more digestible than lower quality forages, which can result
in less overall waste excreted from the animal.
As previously stated, N20 emissions are produced through both direct and indirect pathways. Direct N20 emissions
are produced as part of the nitrogen (N) cycle through the nitrification and denitrification of the N in livestock dung
and urine.5 There are two pathways for indirect N20 emissions. The first is the result of the volatilization of N in
manure (as NH3 and NOx) and the subsequent deposition of these gases and their products (NH4+ and 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 N20 emissions from livestock manure depends on the composition of the manure (manure
includes both feces and urine), the type of bacteria involved in the process, and the amount of oxygen and liquid in
the manure system. For direct N20 emissions to occur, the manure must first be handled aerobically where organic
N is mineralized or decomposed to NH4 which is then nitrified to NO3 (producing some N20 as a byproduct)
(nitrification). Next, the manure must be handled anaerobically where the nitrate is then denitrified to N20 and N2
(denitrification). NOx can also be produced during denitrification. (Groffman et al. 2000; Robertson and Groffman
5 Direct and indirect N2O 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 5-9
-------
2015). These emissions are most likely to occur in dry manure handling systems that have aerobic conditions, but
that also contain pockets of anaerobic conditions due to saturation. A very small portion of the total N excreted is
expected to convert to N20 in the waste management system (WMS). Indirect N20 emissions are produced when
nitrogen is lost from the system through volatilization (as NH3 or NOx) or through runoff and leaching. The vast
majority of volatilization losses from these operations are NH3. Although there are also some small losses of NOx,
there are no quantified estimates available for use, so losses due to volatilization are only based on NH3 loss factors.
Runoff losses would be expected from operations that house animals or store manure in a manner that is exposed to
weather. Runoff losses are also specific to the type of animal housed on the operation due to differences in manure
characteristics. Little information is known about leaching from manure management systems as most research
focuses on leaching from land application systems. Since leaching losses are expected to be minimal, leaching losses
are coupled with runoff losses and the runoff/leaching estimate provided in this chapter does not account for any
leaching losses.
Estimates of CH4 emissions from manure management in 2017 were 61.7 MMT CO2 Eq. (2,467 kt); in 1990,
emissions were 37.1 MMT CO2 Eq. (1,486 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 29 and 134 percent, respectively. From
2016 to 2017, there was a 0.2 percent increase in total CH4 emissions from manure management, due to an increase
in animal populations.
Although a large quantity of managed manure in the United States is handled as a solid, producing little CH4, the
general trend in manure management, particularly for dairy cattle and swine (which are both shifting towards larger
facilities), is one of increasing use of liquid systems. Also, new regulations controlling the application of manure
nutrients to land have shifted manure management practices at smaller dairies from daily spread systems to storage
and management of the manure on site. In many cases, manure management systems with the most substantial
methane emissions are those associated with confined animal management operations where manure is handled in
liquid-based systems. Nitrous oxide emissions from manure management vary significantly between the types of
management system used and can also result in indirect emissions due to other forms of nitrogen loss from the
system (IPCC 2006).
While national dairy animal populations have decreased since 1990, some states have seen increases in their dairy
cattle populations as the industry becomes more concentrated in certain areas of the country and the number of
animals contained on each facility increases. These areas of concentration, such as California, New Mexico, and
Idaho, tend to utilize more liquid-based systems to manage (flush or scrape) and store manure. Thus, the shift toward
larger dairy cattle and swine facilities since 1990 has translated into an increasing use of liquid manure management
systems, which have higher potential CH4 emissions than dry systems. This significant shift in both the dairy cattle
and swine industries was accounted for by incorporating state and WMS-specific CH4 conversion factor (MCF)
values in combination with the 1992, 1997, 2002, 2007 and 2012 farm-size distribution data reported in the U.S.
Department of Agriculture (USD A) Census of Agriculture (USDA 2016c).
In 2017, total N2O emissions from manure management were estimated to be 18.7 MMT CO2 Eq. (63 kt); in 1990,
emissions were 14.0 MMT CO2 Eq. (47 kt). These values include both direct and indirect N20 emissions from
manure management. Nitrous oxide emissions have increased since 1990. Small changes inN20 emissions from
individual animal groups exhibit the same trends as the animal group populations, with the overall net effect that
N2O emissions showed a 34 percent increase from 1990 to 2017 and a 3 percent increase from 2016 through 2017.
Overall shifts toward liquid systems have driven down the emissions per unit of nitrogen excreted as dry manure
handling systems have greater aerobic conditions that promote N20 emissions.
Table 5-6 and Table 5-7 provide estimates of CH4 and N20 emissions from manure management by animal
category.6
Table 5-6: ChU and N2O Emissions from Manure Management (MMT CO2 Eq.)
Gas/Animal Type 1990 2005 2013 2014 2015 2016 2017"
CH4a 37.1 53.7 58.1 57.8 60.9 61.5 61.7
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.
5-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Dairy Cattle
14.7
26.4
33.4
34.0
34.8
34.4
34.5
Beef Cattle
3.1
3.3
3.1
3.0
3.1
3.3
3.4
Swine
15.5
20.3
18.0
17.2
19.2
20.2
20.0
Sheep
0.2
0.1
0.1
0.1
0.1
0.1
0.1
Goats
+
+
+
+
+
+
+
Poultry
3.3
3.2
3.2
3.3
3.4
3.4
3.4
Horses
0.2
0.3
0.2
0.2
0.2
0.2
0.2
American Bison
+
+
+
+
+
+
+
Mules and Asses
+
+
+
+
+
+
+
N2Ob
14.0
16.5
17.4
17.4
17.6
18.2
18.7
Dairy Cattle
5.3
5.6
5.9
5.9
6.1
6.1
6.1
Beef Cattle
5.9
7.2
7.7
7.8
7.7
8.1
8.6
Swine
1.2
1.6
1.8
1.7
1.8
1.9
1.9
Sheep
0.1
0.3
0.3
0.3
0.3
0.3
0.3
Goats
+
+
+
+
+
+
+
Poultry
1.4
1.6
1.6
1.6
1.6
1.6
1.6
Horses
0.1
0.1
0.1
0.1
0.1
0.1
0.1
American Bison0
NA
NA
NA
NA
NA
NA
NA
Mules and Asses
+
+
+
+
+
+
+
Total
51.1
70.2
75.5
75.2
78.5
79.7
80.4
+ Does not exceed 0.05 MMT CO2 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 N2O emissions.
c There are no American bison N2O emissions from managed systems; American bison are
maintained entirely on pasture, range, and paddock.
Notes: Emissions from manure deposited on pasture are included in the Agricultural Soils
Management sector. Totals may not sum due to independent rounding.
Table 5-7: ChU and N2O Emissions from Manure Management (kt)
Gas/Animal Type
1990
2005
2013
2014
2015
2016
2017
CH4a
1,486
2,150
2,322
2,311
2,435
2,461
2,467
Dairy Cattle
590
1,057
1,338
1,360
1,390
1,374
1,381
Beef Cattle
126
133
122
120
126
131
135
Swine
622
812
721
688
770
807
802
Sheep
7
3
3
3
3
3
3
Goats
1
1
1
1
1
1
1
Poultry
131
129
129
132
136
136
137
Horses
9
12
9
9
9
9
8
American Bison
+
+
+
+
+
+
+
Mules and Asses
+
+
+
+
+
+
+
N2Ob
47
55
58
58
59
61
63
Dairy Cattle
18
19
20
20
20
21
21
Beef Cattle
20
24
26
26
26
27
29
Swine
4
5
6
6
6
6
7
Sheep
+
1
1
1
1
1
1
Goats
+
+
+
+
+
+
+
Poultry
5
5
5
5
5
5
5
Horses
+
+
+
+
+
+
+
American Bison0
NA
NA
NA
NA
NA
NA
NA
Mules and Asses
+
+
+
+
+
+
+
Agriculture 5-11
-------
+ 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 N2O emissions.
c There are no American bison N2O emissions from managed systems; American bison are
maintained entirely on pasture, range, and paddock.
Notes: Emissions from manure deposited on pasture are included in the Agricultural Soils
Management sector. Totals may not sum due to independent rounding.
Methodology
The methodologies presented in IPCC (2006) form the basis of the CH4 and N20 emission estimates for each animal
type. This section presents a summary of the methodologies used to estimate CH4 and N20 emissions from manure
management. See Annex 3.11 for more detailed information on the methodology and data used to calculate CH4 and
N20 emissions from manure management.
Methane Calculation Methods
The following inputs were used in the calculation of manure management CH4 emissions for 1990 through 2017:
• Animal population data (by animal type and state);
• Typical animal mass (TAM) data (by animal type);
• Portion of manure managed in each WMS, by state and animal type;
• Volatile solids (VS) production rate (by animal type and state or United States);
• Methane producing potential (B0) of the volatile solids (by animal type); and
• Methane conversion factors (MCF), the extent to which the CH4 producing potential is realized for each
type of WMS (by state and manure management system, including the impacts of any biogas collection
efforts).
Methane emissions were estimated by first determining activity data, including animal population, TAM, WMS
usage, and waste characteristics. The activity data sources are described below:
• Annual animal population data for 1990 through 2017 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, and 2012; horse and mule and ass population data for 1987, 1992, 1997,
2002, 2007, and 2012; and American bison population for 2002, 2007 and 2012 were obtained from the
Census of Agriculture (USDA 2014a). 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). 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.
5-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• 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 (USD A 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 2006IPCC Guidelines (IPCC 2006).
American bison VS production was assumed to be the same as NOF bulls.
• The maximum CH4-producing capacity of the VS (B0) 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 presented in the AgSTAR Digest (EPA 2000, 2003, 2006) and the
AgSTAR project database (EPA 2018). Anaerobic digester emissions were calculated based on estimated
methane production and collection and destruction efficiency assumptions (ERG 2008).
• For all cattle except for calves, the estimated amount of VS (kg per animal-year) managed in each WMS
for each animal type, state, and year were taken from the CEFM, assuming American bison VS production
to be the same as NOF bulls. For animals other than cattle, the annual amount of VS (kg per year) from
manure excreted in each WMS was calculated for each animal type, state, and year. This calculation
multiplied the animal population (head) by the VS excretion rate (kg VS per 1,000 kg animal mass per
day), the TAM (kg animal mass per head) divided by 1,000, the WMS distribution (percent), and the
number of days per year (365.25).
The estimated amount of VS managed in each WMS was used to estimate the CH4 emissions (kg CH4 per year)
from each WMS. The amount of VS (kg per year) were multiplied by the maximum CH4 producing capacity of the
VS (B0) (m3 CH4 per kg VS), the MCF for that WMS (percent), and the density of CH4 (kg CH4per m3 CH4). The
CH4 emissions for each WMS, state, and animal type were summed to determine the total U.S. CH4 emissions.
Nitrous Oxide Calculation Methods
The following inputs were used in the calculation of direct and indirect manure management N20 emissions for
1990 through 2017:
• Animal population data (by animal type and state);
• TAM data (by animal type);
• Portion of manure managed in each WMS (by state and animal type);
• Total Kjeldahl N excretion rate (Nex);
• Direct N20 emission factor (EFwms);
• Indirect N20 emission factor for volatilization (EFvoiatiiization);
• Indirect N20 emission factor for runoff and leaching (EFrllnon; k,,ch):
• Fraction of N loss from volatilization of NH3 and NOx (Fracg(ls): and
• Fraction of N loss from runoff and leaching (FraCnmoff/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
Agriculture 5-13
-------
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 N20 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
(FraCnmoff/ieach) were developed. Fracgas values were based on WMS-specific volatilization values as
estimated from EP A's National Emission Inventory - Ammonia Emissions from Animal Agriculture
Operations (EPA 2005). FraCmnoff/ieachmg values were based on regional cattle runoff data from EPA's Office
of Water (EPA 2002b; see Annex 3.11).
To estimate N20 emissions for cattle (except for calves), the estimated amount of N excreted (kg per animal-year)
that is managed in each WMS for each animal type, state, and year were taken from the CEFM. For calves and other
animals, the amount of N excreted (kg per year) in manure in each WMS for each animal type, state, and year was
calculated. The population (head) for each state and animal was multiplied by TAM (kg animal mass per head)
divided by 1,000, the nitrogen excretion rate (Nex, in kg N per 1,000 kg animal mass per day), WMS distribution
(percent), and the number of days per year.
Direct N20 emissions were calculated by multiplying the amount of N excreted (kg per year) in each WMS by the
N20 direct emission factor for that WMS (EFwms, in kg N20-N per kg N) and the conversion factor of N20-N to
N20. These emissions were summed over state, animal, and WMS to determine the total direct N20 emissions (kg of
N20 per year).
Next, indirect N20 emissions from volatilization (kg N20 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 N20 per kg N), and the conversion factor of N20-N to N20.
Indirect N20 emissions from runoff and leaching (kg N20 per year) were then calculated by multiplying the amount
of N excreted (kg per year) in each WMS by the fraction of N lost through runoff and leaching (FraCnmoff/ieach)
divided by 100, and the emission factor for runoff and leaching (EFrunoff/ieach, in kg N20 per kg N), and the
conversion factor of N20-N to N20. The indirect N20 emissions from volatilization and runoff and leaching were
summed to determine the total indirect N20 emissions.
Following these steps, direct and indirect N20 emissions were summed to determine total N20 emissions (kg N20
per year) for the years 1990 to 2017.
Uncertainty and Time-Series Consistency
An analysis (ERG 2003a) was conducted for the manure management emission estimates presented in the 1990
through 2001 Inventory (i.e., 2003 submission to the UNFCCC) to determine the uncertainty associated with
estimating CH4 and N20 emissions from livestock manure management. The quantitative uncertainty analysis for
this source category was performed in 2002 through the IPCC-recommended Approach 2 uncertainty estimation
methodology, the Monte Carlo Stochastic Simulation technique. The uncertainty analysis was developed based on
the methods used to estimate CH4 and N20 emissions from manure management systems. A normal probability
distribution was assumed for each source data category. The series of equations used were condensed into a single
equation for each animal type and state. The equations for each animal group contained four to five variables around
which the uncertainty analysis was performed for each state. These uncertainty estimates were directly applied to the
2017 emission estimates as there have not been significant changes in the methodology since that time.
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 5-8. Manure management
CH4 emissions in 2017 were estimated to be between 50.6 and 74.0 MMT C02 Eq. at a 95 percent confidence level,
which indicates a range of 18 percent below to 20 percent above the actual 2017 emission estimate of 61.7 MMT
7 The N2O 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 N2O emissions from unmanaged WMS are not included in the
Manure Management source category, there are no N2O emissions from American bison included in the Manure Management
source category.
5-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
C02 Eq. At the 95 percent confidence level, N20 emissions were estimated to be between 15.7 and 23.2 MMT CO2
Eq. (or approximately 16 percent below and 24 percent above the actual 2017 emission estimate of 18.7 MMT CO2
Eq.).
Table 5-8: Approach 2 Quantitative Uncertainty Estimates for CH4 and N2O (Direct and
Indirect) Emissions from Manure Management (MMT CO2 Eq. and Percent)
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Manure Management
CH4
61.7
50.6 74.0
-18% +20%
Manure Management
N2O
18.7
15.7 23.2
-16% +24%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
QA/QC and Verification
General (Tier 1) and category-specific (Tier 2) QA/QC activities were conducted consistent with the U.S. Inventory
QA/QC plan outlined in Annex 8. Tier 2 activities focused on comparing estimates for the previous and current
Inventories for N20 emissions from managed systems and CH4 emissions from livestock manure. All errors
identified were corrected. Order of magnitude checks were also conducted, and corrections made where needed. In
addition manure N data were checked by comparing state-level data with bottom up estimates derived at the county
level and summed to the state level. Similarly, a comparison was made by animal and WMS type for the full time
series, between national level estimates for N excreted and the sum of county estimates for the full time series.
Time-series data, including population, are validated by experts to ensure they are representative of the best
available U.S.-specific data. The U.S.-specific values for TAM, Nex, VS, B0, and MCF were also compared to the
IPCC default values and validated by experts. Although significant differences exist in some instances, these
differences are due to the use of U.S.-specific data and the differences in U.S. agriculture as compared to other
countries. The U.S. manure management emission estimates use the most reliable country-specific data, which are
more representative of U.S. animals and systems than the IPCC (2006) default values.
For additional verification of the 1990 to 2017 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 5-9 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 CH4 emissions.
Table 5-9: IPCC (2006) Implied Emission Factor Default Values Compared with Calculated
Values for ChU from Manure Management (kg/head/year)
Animal Type
IPCC Default
CH4 Emission
Factors
Implied CH4 Emission Factors (kg/head/year)
(kg/head/year)
1990
2005
2013
2014
2015
2016
2017
Dairy Cattle
48-112
30.2
59.4
72.3
73.4
73.9
72.9
73.1
Beef Cattle
1-2
1.5
1.6
1.6
1.6
1.7
1.7
1.7
Swine
10-45
11.5
13.3
11.0
10.7
11.3
11.5
11.1
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
2.5
2.6
2.6
2.6
Agriculture 5-15
-------
American Bison NA 1.8 2.0 2.0 2.0 2.1 2.1 2.1
Mules and Asses 0.76-1.14 0.9 1.0 0.9 0.9 1.0 1.0 1.0
NA (Not Applicable)
In addition, default IPCC (2006) emission factors for N20 were compared to the U.S. Inventory implied N20
emission factors. Default N20 emission factors from the 2006 IPCC Guidelines were used to estimate N20 emission
from each WMS in conjunction with U.S.-specific Nex values. The implied emission factors differed from the U.S.
Inventory values due to the use of U.S.-specific Nex values and differences in populations present in each WMS
throughout the time series.
Recalculations Discussion
The manure management emission estimates include the following recalculations relative to the previous Inventory:
• The CEFM produces population, VS and Nex data for cattle (except calves), that are used in the manure
management inventory. As a result, all changes to the CEFM described in Section 5.1 contributed to
changes in the population, VS and Nex data used for calculating CH4 and N20 cattle emissions from
manure management.
• State animal populations were updated to reflect updated USDA NASS datasets, which resulted in
population changes for:
o Poultry in 2015,
o Calves in 2015,
o Beef OF heifers and steers in 2015,
o Dairy heifers in 2015,
o American bison in 2012-2015, and
o Swine in 2015 (USDA 2018).
• WMS distribution data for swine were updated with data from the 2009 USDA Agricultural Resource
Management Survey (ARMS) of swine producers (ERG 2018). Anaerobic digestion data were also updated
for swine, using data from EPA's AgSTAR Program (EPA 2018).
• Temperature data were updated by NOAA due to an update in their computer program which affected the
precision of the output dataset (Gleason 2018). This resulted in minor temperature changes and
subsequently, MCF changes for all animals across the time series.
These changes impacted total emission estimates for 1990 through 2016, overall decreasing annual estimations from
less than 1 percent to 7.2 percent across the time series. The most significant changes were to the swine emissions
estimates, resulting primarily from the swine WMS update. Total swine annual estimations decreased throughout the
entire time series, but most significantly for 2013 through 2015 during which time they decreased by over 20
percent.
Planned Improvements
During the Public Review period of the previous Inventory report (i.e., 1990 through 2016), EPA received comment
on various aspects of the manure management inventory, including recommended improvements to clarify the scope
of the manure management sector and better align terminology with those used within the industry (e.g., clarifying
"managed" versus "unmanaged"), as well as comments to update data and methods which reiterated those
improvements already identified by EPA and listed below. EPA notes that many of these improvements, identified
below, are major updates and may take multiple years to implement in full, but will add clarity to improve the
transparency of future inventories.
Potential improvements (medium- to long-term improvements) for future Inventory years include:
• Continuing to obtain and incorporate existing data sources (such as the 2016 USDA ARMS dairy data) to
update WMS distributions.
• Revising the methodology for population distribution to states where USDA population data are withheld
due to disclosure concerns.
5-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• Revising the anaerobic digestion estimates to estimate CH4 emissions reductions due to the use of
anaerobic digesters (the Inventory currently estimates only emissions from anaerobic digestion systems).
• Investigating improved emissions estimate methodologies for swine pit systems with less than one month
of storage (the new swine WMS data included this WMS category).
• Updating the B0 data used in the Inventory, which are dated.
• Comparing CH4 and N20 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.
• 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 calculation
emissions.
• Implementing a methodology to calculate monthly emissions estimates to present data that show seasonal
changes in emissions from each WMS.
• Revising the uncertainty analysis to address changes that have been implemented to the CH4 and N20
estimates.
5.3 Rice Cultivation (CRF Source Category 3C)
Most of the world's rice is grown on flooded fields (Baicich 2013), and flooding creates anaerobic conditions that
foster CH4 production through a process known as methanogenesis. Approximately 60 to 90 percent of the CH4
produced by methanogenie bacteria is oxidized in the soil and converted to C02by methanotrophic bacteria. The
remainder is emitted to the atmosphere (Holzapfel-Pschorn et al. 1985; Sass et al. 1990) or transported as dissolved
CH4 into groundwater and waterways (Neue et al. 1997). Methane is transported to the atmosphere primarily
through the rice plants, but some CH4 also escapes via ebullition (i.e., bubbling through the water) and to a much
lesser extent by diffusion through the water (van Bodegom et al. 2001).
Water management is arguably the most important factor affecting CH4 emissions, and improved water management
has the largest potential to mitigate emissions (Yan et al. 2009). Upland rice fields are not flooded, and therefore do
not produce CH4, but large amounts of CH4 can be emitted in continuously irrigated fields, which is the most
common practices in the United States (USDA 2012). Single or multiple aeration events with drainage of a field
during the growing season can significantly reduce these emissions (Wassmann et al. 2000a), but drainage may also
increase N20 emissions. Deepwater rice fields (i.e., fields with flooding depths greater than one meter, such as
natural wetlands) tend to have less living stems reaching the soil, thus reducing the amount of CH4 transport to the
atmosphere through the plant compared to shallow-flooded systems (Sass 2001).
Other management practices also influence CH4 emissions from flooded rice fields including rice residue straw
management and application of organic amendments, in addition to cultivar selection due to differences in the
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 CH4 emissions (Wassmann et al. 2000b). Fertilization
practices also influences CH4 emissions, particularly the use of fertilizers with sulfate (Wassmann et al. 2000b;
Linquist et al. 2012), which can reduce CH4 emissions. Other environmental variables also impact the
methanogenesis process such as soil temperature and soil type. Soil temperature regulates the activity of
methanogenic bacteria which in turn affects the rate of CH4 production. Soil texture influences decomposition of soil
organic matter, but is also thought to have an impact on oxidation of CH4 in the soil (Sass et al. 1994).
Rice is currently cultivated in twelve states, including Arkansas, California, Florida, Illinois, Kentucky, Louisiana,
Minnesota, Mississippi, Missouri, New York, South Carolina, Tennessee and Texas. Soil types, rice varieties, and
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.
Agriculture 5-17
-------
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 lias been applied to estimate national emissions for 2013 -2017 in this
Inventory. National emission estimates based on surrogate data will be recalculated in the next (i.e., 1990 through
2018) Inventory submission using the Tier 1 and 3 methods.
Overall, rice cultivation is a minor source of CH4 emissions in the United States relative to other source categories
(see Table 5-10, Table 5-11, and Figure 5-2). The majority of emissions occur in Arkansas, California, Louisiana
and Texas. In 2017, CH4 emissions from rice cultivation were 11.3 MMT CO2 Eq. (454 kt). Annual emissions
fluctuate between 1990 and 2017, which is largely due to differences in the amount of rice harvested areas over
time, which lias been decreasing over the past two decades. Consequently, emissions in 2017 are 29 percent lower
than emissions in 1990.
Table 5-10: ChU Emissions from Rice Cultivation (MMT CO2 Eq.)
State
1990
2005
2012
2013
2014
2015
2016
2017
Arkansas
3.3
4.7
3.8
NE
NE
NE
NE
NE
California
2.0
2.1
2.0
NE
NE
NE
NE
NE
Florida
0.0
0.1
0.0
NE
NE
NE
NE
NE
Illinois
0.0
+
0.0
NE
NE
NE
NE
NE
Kentucky
0.0
+
0.0
NE
NE
NE
NE
NE
Louisiana
6.1
6.5
3.9
NE
NE
NE
NE
NE
Minnesota
+
+
+
NE
NE
NE
NE
NE
Mississippi
0.6
0.6
0.5
NE
NE
NE
NE
NE
Missouri
0.3
0.6
0.3
NE
NE
NE
NE
NE
New York
+
0.0
0.0
NE
NE
NE
NE
NE
South Carolina
0.0
0.0
0.0
NE
NE
NE
NE
NE
Tennessee
0.0
+
0.0
NE
NE
NE
NE
NE
Texas
3.7
2.1
0.9
NE
NE
NE
NE
NE
Total
16.0
16.7
11.3
11.5
12.7
12.3
13.7
11.3
+ Does not exceed 0.05 MMT CO2 Eq.
NE (Not Estimated). State-level emissions are not estimated for 2013 through 2017 Inventory, and national
emissions are determined using a surrogate data method.
Note: Totals may not sum due to independent rounding.
Table 5-11: ChU Emissions from Rice Cultivation (kt)
State
1990
2005
2012
2013
2014
2015
2016
2017
Arkansas
132
188
151
NE
NE
NE
NE
NE
California
81
82
81
NE
NE
NE
NE
NE
Florida
0
3
0
NE
NE
NE
NE
NE
Illinois
0
+
0
NE
NE
NE
NE
NE
Kentucky
0
+
0
NE
NE
NE
NE
NE
Louisiana
246
261
156
NE
NE
NE
NE
NE
Minnesota
1
2
1
NE
NE
NE
NE
NE
Mississippi
23
23
19
NE
NE
NE
NE
NE
Missouri
12
22
12
NE
NE
NE
NE
NE
New York
+
0
0
NE
NE
NE
NE
NE
South Carolina
0
0
0
NE
NE
NE
NE
NE
Tennessee
0
+
0
NE
NE
NE
NE
NE
Texas
146
86
34
NE
NE
NE
NE
NE
Total
641
667
453
462
510
493
549
454
+ Does not exceed 0.5 kt.
5-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
NE (Not Estimated). State-level emissions are not estimated for 2013 through 2017 Inventory, and national
emissions are determined using a surrogate data method.
Note: Totals may not sum due to independent rounding.
Figure 5-2: Annual ChU Emissions from Rice Cultivation, 2012 (MMT CO2 Eq./Year)
MT C02 Eq. ha1 yr
~ < 5
~ 5 to 10
¦ 10 to 15
¦ 15 to 20
¦ > 20
Note: Only national-scale emissions are estimated for 2013 through 2017 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 2012.
Methodology
The methodology used to estimate CH4 emissions from rice cultivation is based on a combination of IPCC Tier 1
and 3 approaches. The Tier 3 method utilizes 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 CH4, as well as CH4 transport through the plant and
via ebullition (Cheng et al. 2013). The method simulates the influence of organic amendments, rice straw
management on methanogenesis in the flooded soils, and ratooning of rice crops with a second harvest during the
growing season. In addition to CH4 emissions, DAYCENT simulates soil C stock changes and N20 emissions
(Parton et al. 1987 and 1998; Del Grosso et al. 2010), and allows for a seamless set of simulations for crop rotations
that include both rice and non-rice crops.
The Tier 1 method is applied to estimate CH4 emissions from rice when grown in rotation with crops that are not
simulated by DAYCENT, such as vegetable crops. The Tier 1 method is also used for areas converted between
agriculture (i.e., cropland and grassland) and other land uses, such as forest land, wetland, and settlements. In
addition, the Tier 1 method is used to estimate CH4 emissions from organic soils (i.e., Histosols) and from areas with
very gravelly, cobbly, or slialey soils (greater than 35 percent by volume). The Tier 3 method using DAYCENT lias
not been fully tested for estimating emissions associated with these crops and rotations, land uses, as well as organic
soils or cobbly, gravelly, and slialey mineral soils.
Agriculture 5-19
-------
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 factor represents emissions for continuously flooded fields with no
organic amendments. Scaling factors are used to adjust 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 USD A National Resources
Inventory (NRI) survey (USDA-NRCS 2015). The NRI is a statistically-based sample of all non-federal land, and
includes 380,956 survey points of which 1,588 are in locations with rice cultivation at the end of the NRI time
series. The Tier 3 method is used to estimate CH4 emissions from 1,393 of the NRI survey locations, and the
remaining 195 survey locations are estimated with the Tier 1 method. Each NRI survey point is associated with an
"expansion factor" that allows scaling of CH4 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 sample point). 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 2012 (USDA-NRCS 2015). The current Inventory only uses NRI data through
2012 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 5-12.
Table 5-12: Rice Area Harvested (1,000 Hectares)
State/Crop
1990
2005
2012
2013
2014
2015
2016
2017
Arkansas
599
796
613
NE
NE
NE
NE
NE
California
248
247
244
NE
NE
NE
NE
NE
Florida
0
11
0
NE
NE
NE
NE
NE
Illinois
0
0
0
NE
NE
NE
NE
NE
Kentucky
0
0
0
NE
NE
NE
NE
NE
Louisiana
380
402
226
NE
NE
NE
NE
NE
Minnesota
4
10
6
NE
NE
NE
NE
NE
Mississippi
119
115
92
NE
NE
NE
NE
NE
Missouri
47
93
46
NE
NE
NE
NE
NE
New York
1
0
0
NE
NE
NE
NE
NE
South Carolina
0
0
0
NE
NE
NE
NE
NE
Tennessee
0
1
0
NE
NE
NE
NE
NE
Texas
300
150
66
NE
NE
NE
NE
NE
Total
1,698
1,826
1,292
NE
NE
NE
NE
NE
NE (Not Estimated).
Notes: Totals may not sum due to independent rounding. States are included if NRI reports an area of rice production in the
state at any time between 1990 and 2012. Rice harvested area data have not been compiled for 2013 to 2017.
The Southeastern states have sufficient growing periods for a ratoon crop in some years. For example, in Arkansas,
the length of growing season is occasionally sufficient for ratoon crops on an average of 1 percent of the rice fields.
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 foryears 1993 through 2014), 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 5-13.
9 See .
5-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 5-13: Average Ratooned Area as Percent of Primary Growth Area (Percent)
State
1990-2012
Arkansas3
1%
California
0%
Floridab
49%
Louisiana0
32%
Mississippi3
1%
Missouri3
1%
Texas'1
45%
aArkansas: 1990-2000 (Slaton 1999 through2001); 2001-2011 (Wilson2002 through2007, 2009 through2012); 2012-2013
(Hardke 2013,2014).
bFlorida - Ratoon: 1990-2000 (Schuenemau 1997,1999 through2001); 2001 (Deren2002); 2002-2003 (Kirstein2003
through 2004,2006); 2004 (Cantens 2004 through 2005); 2005-2013 (Gonzalez 2007 through 2014).
cLouisiana: 1990-2013 (Linscombe 1999,2001 tlirough2014).
dTexas: 1990-2002 (Klosterboer 1997,1999 tlirough 2003); 2003-2004 (Stansel 2004 tlirough 2005); 2005 (Texas
Agricultural Experiment Station 2006); 2006-2013 (Texas Agricultural Experiment Station 2007 tlirough 2014).
While rice crop production in the United States includes a minor amount of land with mid-season drainage or
alternate wet-dry periods, the majority of rice growers use continuously flooded water management systems (Hardke
2015; UCCE 2015; Hollier 1999; Way et al. 2014). Therefore, continuous flooding was assumed in the DAYCENT
simulations and the Tier 1 method. Variation in flooding can be incorporated in future Inventories if water
management data are collected.
Winter flooding is another key practice associated with water management in rice fields, and the impact of winter
flooding on CH4 emissions is addressed in the Tier 3 and Tier 1 analyses. Flooding is used to prepare fields for the
next growing season, and to create waterfowl habitat (Young 2013; Miller et al. 2010; Fleskes et al. 2005).
Fitzgerald et al. (2000) suggests that as much as 50 percent of the annual emissions may occur during the winter
flood. 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 period.
A surrogate data method is used to estimate emissions from 2013 to 2017 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 tlirough 2012 emissions data that was
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 lias 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 2012 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+ e.
10 See .
Agriculture 5-21
-------
where Y is the response variable (e.g., soil organic carbon), X(3 contains specific surrogate data depending on the
response variable, and e is the remaining unexplained error. EPA tested models with a variety of surrogate data,
including commodity statistics, weather data, or other relevant information. Parameters are estimated from the
observed data for 1990 to 2012 using standard statistical techniques, and these estimates are used to predict the
missing emissions data for 2013 to 2017.
A critical issue in using splicing methods in general, is to adequately account for the additional uncertainty
introduced by predicting emissions with related information without compiling the full inventory. For example,
predicting CH4 emissions will increase the total variation in the emission estimates for these specific years,
compared to those years in which the full inventory is compiled. This added uncertainty is quantified within the
model framework using a Monte Carlo approach. The approach requires estimating parameters for results in each
Monte Carlo simulation for the full inventory (i.e., the surrogate data model is refit with the emissions estimated in
each Monte Carlo iteration from the full inventory analysis with data from 1990 to 2012).
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 2013 to 2017,
there is additional uncertainty propagated through the Monte Carlo Analysis associated with the surrogate data
method. (See Box 5-2 for information about propagating uncertainty with the surrogate data method.) The
uncertainties from the Tier 1 and 3 approaches are combined to produce the final CH4 emissions estimate using
simple error propagation (IPCC 2006). Additional details on the uncertainty methods are provided in Annex 3.12.
Rice cultivation CH4 emissions in 2017 were estimated to be between 8.6 and 16.9 MMT CO2 Eq. at a 95 percent
confidence level, which indicates a range of 25 percent below to 49 percent above the actual 2017 emission estimate
of 11.3 MMT C02 Eq. (see Table 5-14).
Table 5-14: Approach 2 Quantitative Uncertainty Estimates for ChU Emissions from Rice
Cultivation (MMT CO2 Eq. and Percent)
Source
Inventory
Method
Gas
2017 Emission
Estimate
Uncertainty Range Relative to Emission
Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Upper
Lower Upper
Bound
Bound
Bound Bound
Rice Cultivation
Tier 3
CH4
9.6
6.9
12.2
-27% +27%
Rice Cultivation
Tier 1
ch4
1.8
0.8
2.8
-55% +55%
Rice Cultivation
Total
ch4
11.3
8.6
16.9
-25% +49%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
QA/QC and Verification
General (Tier 1) and category-specific (Tier 2) QA/QC activities were conducted consistent with the U.S. Inventory
QA/QC plan outlined in Annex 8. Quality control measures include checking input data, model scripts, and results
to ensure data are properly handled throughout the inventory process. Inventory reporting forms and text are
reviewed and revised as needed to correct transcription errors. No errors were found in the reporting forms and text.
Model results are compared to field measurements to verily if results adequately represent CH4 emissions. The
comparisons included over 15 long-term experiments, representing about 80 combinations of management
treatments across all of 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.
5-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Planned Improvements
New land representation data and rice cultivation data were not compiled for the current Inventory. A surrogate data
method has been 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 a future Inventory will be to update the time
series for CH4 emissions from rice cultivation by compiling the latest land use data and related management
statistics.
In addition, a major improvement is underway to update the time series of management data with information from
the USDA-NRCS Conservation Effects Assessment Program (CEAP). This improvement will fill several gaps in the
management data including more specific data on fertilizer rates, updated tillage practices, water management,
organic amendments and more information on planting and harvesting dates. This improvement is expected to be
completed for the 1990 through 2018 Inventory (i.e., 2020 submission). However, 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 A number of agricultural activities increase mineral N availability in soils that lead to direct N20
emissions from nitrification and denitrification at the site of a management activity (see Figure 5-3) (Mosier et al.
1998), including N fertilization; application of managed livestock manure and other organic materials such as
biosolids (i.e., sewage sludge); deposition of manure on soils by domesticated animals in pastures, rangelands, and
paddocks (PRP) (i.e., by grazing animals and other animals whose manure is not managed); production of N-fixing
crops and forages; retention of crop residues; and drainage of organic soils (i.e., soils with a high organic matter
content, otherwise known as Histosols13) (IPCC 2006). Additionally, agricultural soil management activities,
including irrigation, drainage, tillage practices, and fallowing of land, can influence N mineralization from soil
organic matter and levels of asymbiotic N fixation by impacting moisture and temperature regimes in soils. Indirect
emissions of N20 occur when N is transported from a site and is subsequently converted to N20; there are two
pathways for indirect emissions: (1) volatilization and subsequent atmospheric deposition of applied/mineralized N,
and (2) surface runoff and leaching of applied/mineralized N into groundwater and surface water.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; N20 emissions from Forest Land and Settlements soils are found
in Sections 6.2 and 6.10, respectively).
11 Nitrification and denitrification are driven by the activity of microorganisms in soils. Nitrification is the aerobic microbial
oxidation of ammonium (NH4+) to nitrate (NO3"), 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.
13 Drainage of organic soils in former wetlands enhances mineralization of N-rich organic matter, thereby increasing N2O
emissions from these soils.
14 These processes entail volatilization of applied or mineralized N as NH3 and NOx, transformation of these gases within the
atmosphere (or upon deposition), and deposition of the N primarily in the form of particulate NH4+, nitric acid (HNO3), and NOx,
in addition to leaching and runoff of NO3" that is converted to N2O in aquatic systems, e.g., wetlands, rivers, streams and lakes.
Note: N2O emissions are not estimated for aquatic systems associated with N inputs from terrestrial systems in order to avoid
double-counting.
Agriculture 5-23
-------
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 NjO Emissions from Agricultural Soil Management
N Volatilization
FERTILIZER
SyntheticN Fertilizers
Synthetic N fertilizer appl ied tosoil
Organic
Amendments
Includes both commercial and
ercisl fertilizers (i
mal manure, compost,
sewage sludge, tankage etc.)
Urine and Dung from
Grazing Animals
Manure deposited on pasture range,
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 atmospheric N2 by bacteria
living in soils that do not have a direct
relationship with plants
N Flows:
m
N Inputs to
Managed Soils
—~
Direct NzO
Emissions
N Volatilization
and Deposition
0
Indirect N20
Emissions
" Histosol
Cultivation
This graphic illustrates the sources and pathways of nitrogen that result
in direct and indirect N20 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.
wg r
Surface
Wale
Groundwater
5-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Agricultural soils produce the majority of N20 emissions in the United States. Estimated emissions from this source
in 2017 are 266.4 MMT CO2 Eq. (894 kt) (see Table 5-15 and Table 5-16). Annual N2O emissions from agricultural
soils are 6 percent greater in the 2017 compared to 1990, but emissions fluctuated between 1990 and 2017 due to
inter-annual variability largely associated with weather patterns, synthetic fertilizer use, and crop production. From
1990 to 2017, on average, cropland accounted for 70 percent of total direct emissions, while grassland accounted for
30 percent. On average, 81 percent of indirect emissions are from croplands and 19 percent from grasslands.
Estimated direct and indirect N20 emissions by sub-source category are shown in Table 5-17 and Table 5-18.
Table 5-15: N2O Emissions from Agricultural Soils (MMT CO2 Eq.)
Activity
1990
2005
2013
2014
2015
2016
2017
Direct
212.7
218.9
225.9
223.4
239.0
228.8
227.7
Cropland
148.1
154.7
161.3
160.1
166.8
162.2
161.6
Grassland
64.6
64.2
64.5
63.4
72.2
66.6
66.1
Indirect
39.0
35.7
39.3
38.8
38.8
38.8
38.8
Cropland
31.4
28.7
32.2
31.7
31.6
31.6
31.6
Grassland
7.6
7.0
7.2
7.1
7.1
7.1
7.2
Total
251.7
254.5
265.2
262.3
277.8
267.6
266.4
Notes: Estimates after 2012 are based on a data splicing method (See Methodology section). Totals
may not sum due to independent rounding.
Table 5-16: N2O Emissions from Agricultural Soils (kt)
Activity
1990
2005
2013
2014
2015
2016
2017
Direct
714
735
758
750
802
768
764
Cropland
497
519
541
537
560
544
542
Grassland
217
216
217
213
242
224
222
Indirect
131
120
132
130
130
130
130
Cropland
105
96
108
106
106
106
106
Grassland
26
23
24
24
24
24
24
Total
845
854
890
880
932
898
894
Notes: Estimates after 2012 are based on a data splicing method (See Methodology section). Totals
may not sum due to independent rounding.
Table 5-17: Direct N2O Emissions from Agricultural Soils by Land Use Type and N Input Type
(MMT CO2 Eq.)
Activity
1990
2005
2013
2014
2015
2016
2017
Cropland
148.1
154.7
161.3
160.1
166.8
162.2
161.6
Mineral Soils
144.7
151.3
158.0
156.8
163.5
159.0
158.3
Synthetic Fertilizer
53.3
54.3
60.0
59.6
62.2
60.5
60.2
Organic Amendment3
11.5
12.6
13.4
13.3
13.5
13.4
13.3
Residue Nb
21.7
22.4
24.4
24.2
25.3
24.6
24.5
Mineralization and
Asymbiotic Fixation
58.3
62.0
60.2
59.7
62.5
60.6
60.3
Drained Organic Soils
3.3
3.3
3.3
3.2
3.2
3.2
3.2
Grassland
64.6
64.2
64.5
63.4
72.2
66.6
66.1
Mineral Soils
61.7
61.5
61.9
60.8
69.6
64.0
63.4
Synthetic Fertilizer
0.9
0.8
1.0
1.0
1.1
1.0
1.0
PRP Manure
16.3
14.0
12.5
12.4
13.4
12.8
12.7
Managed Manure0
0.1
0.1
0.1
0.1
0.2
0.1
0.1
Biosolids (i.e., Sewage
Sludge)
0.2
0.5
0.6
0.6
0.6
0.6
0.6
Residue Nd
15.9
16.6
18.8
18.4
21.4
19.5
19.3
Mineralization and
Asymbiotic Fixation
28.2
29.5
29.0
28.3
32.9
30.0
29.7
Drained Organic Soils
2.9
2.7
2.6
2.6
2.6
2.6
2.6
Total
212.7
218.9
225.9
223.4
239.0
228.8
227.7
Agriculture 5-25
-------
a Organic amendment inputs include managed manure, daily spread manure, and commercial organic
fertilizers (i.e., dried blood, dried manure, tankage, compost, and other).
b Cropland residue N inputs include N in unharvested legumes as well as crop residue N.
c Managed manure inputs include managed manure and daily spread manure amendments that are applied to
grassland soils.
d Grassland residue N inputs include N in ungrazed legumes as well as ungrazed grass residue N.
Notes: Estimates after 2012 are based on a data splicing method (See Methodology section). Totals may not
sum due to independent rounding.
Table 5-18: Indirect N2O Emissions from Agricultural Soils (MMT CO2 Eq.)
Activity
1990
2005
2013
2014
2015
2016
2017
Cropland
31.4
28.7
32.2
31.7
31.6
31.6
31.6
Volatilization & Atm.
Deposition
6.2
7.0
6.8
6.8
6.7
6.7
6.7
Surface Leaching & Run-Off
25.3
21.7
25.3
24.9
24.9
24.9
24.9
Grassland
7.6
7.0
7.2
7.1
7.1
7.1
7.2
Volatilization & Atm.
Deposition
4.3
4.5
4.2
4.2
4.2
4.2
4.2
Surface Leaching & Run-Off
3.2
2.5
2.9
2.9
2.9
2.9
2.9
Total
39.0
35.7
39.3
38.8
38.8
38.8
38.8
Notes: Estimates after 2012 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 N20 emissions. Figure 5-6 and Figure 5-7 show indirect
N20 emissions from volatilization, and Figure 5-8 and Figure 5-9 show the indirect N20 emissions from leaching
and runoff in croplands and grasslands, respectively. Annual emissions in 201215 are shown for the Tier 3 Approach
only.
15 Only national-scale emissions are estimated for 2013 to 2017 in the current Inventory using the splicing method, and therefore
the fine-scale emission patterns in these maps are based on Inventory data from 2012.
5-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 5-4: Crops, 2012 Annual Direct N2O Emissions Estimated Using the Tier 3 DAYCENT
Model (MMT CO2 Eq./year)
Note: Only national-scale emissions are estimated for 2013 to 2017 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2012.
Direct N20 emissions from croplands occur throughout all of the cropland regions but tend to be high in the
Midwestern Corn Belt Region (Illinois, Iowa Indiana, Ohio, southern Minnesota and Wisconsin, and eastern
Nebraska), where a large portion of the land is used for growing liighlv fertilized corn and N-fixing soybean crops
(see Figure 5-4). 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 also in many western states where rainfall and access to
irrigation water are limited.
Direct emissions from grasslands are highest in the southeast, particularly Kentucky and Tennessee, in addition to
areas in east Texas and Iowa where there tends to be higher rates of manure amendments on a relatively small
amount of pasture, compared to other regions of the United States. However, total emissions from grasslands tend to
be higher 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 grazing.
Agriculture 5-27
-------
Figure 5-5: Grasslands, 2012 Annual Direct N2O Emissions Estimated Using the Tier 3
DAYCENT Model (MMT CO2 Eq./year)
Note: Only national-scale emissions are estimated for 2013 to 2017 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2012.
Indirect N;0 emissions from volatilization in croplands have a similar pattern as the direct N20 emissions with high
emissions in the Midwestern Corn Belt and Lower Mississippi River Basin. Indirect N20 emissions from
volatilization in grasslands are higher in the Southeastern United States than in other regions. The higher emissions
in this region are mainly due to productive pastures that support intensive grazing, which in turn, stimulates NH3
volatilization. Indirect N20 emissions from surface runoff and leaching of applied/mineralized N is highest in the
Eastern United States for both croplands and grasslands. This region lias greater precipitation and higher levels of
leaching and runoff compared to arid to semi-arid regions in the Western United States.
5-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 5-6: Crops, 2012 Annual Indirect N2O Emissions from Volatilization Using the Tier 3
DAYCENT Model (MMT CO2 Eq./year)
MT C02 Eq. ha1 yr
~ 0.1 -0.25 ¦ 1.5 -2
¦ 0.25 - 0.5 ¦ > 2
Note: Only national-scale emissions are estimated for 2013 to 2017 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2012.
Agriculture 5-29
-------
Figure 5-7: Grasslands, 2012 Annual Indirect N2O Emissions from Volatilization Using the
Tier 3 DAYCENT Model (MMT CO2 Eq./year)
MT C02 Eq. ha1 yr1
~ <0.01 ¦ 0.25 to 0.5
~ 0.01 to 0.05 ¦ 0.5 to 1
¦ 0.1 to 0.25 ¦ > 2
Note: Only national-scale emissions are estimated for 2013 to 2017 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2012.
5-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 5-8: Crops, 2012 Annual Indirect N2O Emissions from Leaching and Runoff Using the
Tier 3 DAYCENT Model (MMT CO2 Eq./year)
Note: Only national-scale emissions are estimated for 2013 to 2017 using a splicing method, and therefore the fine-scale
emission patterns in this map are based on Inventory data from 2012.
Agriculture 5-31
-------
Figure 5-9: Grasslands, 2012 Annual Indirect N2O Emissions from Leaching and Runoff
Using the Tier 3 DAYCENT Model (MMT CO2 Eq./year)
Note: Only national-scale emissions are estimated for 2013 to 2017 using a splicing method, and therefore the fine-scale
emission patterns in this map are based 011 Inventory data from 2012.
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., sewage sludge) applications, crop residues, organic amendments, and biological
N fixation associated with planting of legumes on cropland and grassland soils; (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 N20 emissions from asymbiotic fixation16 and mineralization of soil organic matter and
litter. One recommendation from IPCC (2006) that lias 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.
16 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.
5-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Direct N2O Emissions
The methodology used to estimate direct N20 emissions from agricultural soil management in the United States is
based on a combination of IPCC Tier 1 and 3 approaches, along with application of a splicing method for latter
years in the Inventory time series (IPCC 2006; Del Grosso et al. 2010). 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 with the exception of biosolids (i.e., sewage sludge)
amendments (Del Grosso et al. 2010). The Tier 3 approach lias been specifically designed and tested to estimate
N20 emissions in the United States, accounting for more of the enviromnental 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
allows for the Inventory to address direct N20 emissions and soil C stock changes from mineral cropland soils in a
single analysis. Carbon and N dynamics are linked in plant-soil systems through biogeochemical processes of
microbial decomposition and plant production (McGill and Cole 1981). Coupling the two source categories (i.e.,
agricultural soil C and N20) in a single inventory analysis ensures that there is consistent activity data and treatment
of the processes, and interactions are taken into account between C and N cycling in soils.
The Tier 3 approach is based on the cropping and land use histories recorded in the USDA National Resources
Inventory (NRI) (USDA-NRCS 2015). The NRI is a statistically-based sample of all non-federal land,17 and
includes 363,286 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 the remaining 205,487 in the NRI survey that
are designated as cropland or grassland (discussed later in this section). Each point is associated with an "expansion
factor" that allows scaling of N20 emissions from NRI points to the entire country (i.e., each expansion factor
represents the amount of area with the same land-use/management history as the sample point). Each NRI point 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, and the annual data are currently available through 2012 (USDA-NRCS
2015).
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 N20 emissions on an input-
by-input basis. The Tier 1 approach requires a minimal amount of activity data, readily available in most countries
(e.g., total N applied to crops); calculations are simple; and the methodology is highly transparent. In contrast, the
Tier 3 approach developed fortliis Inventory employs a process-based model (i.e., DAYCENT) that represents the
interaction of N inputs, land use and management, as well as enviromnental conditions at specific locations.
Consequently, the Tier 3 approach produces more accurate estimates; it accounts more comprehensively for land-use
and management impacts and their interaction with enviromnental factors (i.e., weather patterns and soil
characteristics), which will enhance or dampen anthropogenic influences. However, the Tier 3 approach requires
more detailed activity data (e.g., crop-specific N amendment rates), additional data inputs (i.e., daily weather, soil
types, etc.), and considerable computational resources and programming expertise. The Tier 3 methodology is less
transparent, and thus it is critical to evaluate the output of Tier 3 methods against measured data in order to
demonstrate that the method is an improvement over lower tier methods for estimating emissions (IPCC 2006).
Another important difference between the Tier 1 and Tier 3 approaches relates to assumptions regarding N cycling.
Tier 1 assumes that N added to a system is subject to N20 emissions only during that year and cannot be stored in
soils and contribute to N20 emissions in subsequent years. This is a simplifying assumption that is likely to create
bias in estimated N20 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 N20 during subsequent years.
17 Hie NRI survey does include sample points on federal lands, but the program does not collect data from those sample
locations.
Agriculture 5-33
-------
DAYCENT is used to estimate N20 emissions associated with production of alfalfa hay, barley, corn, cotton, dry
beans, grass hay, grass-clover hay, lentils, oats, onions, peanuts, peas, potatoes, rice, sorghum, soybeans, sugar
beets, sunflowers, tobacco, tomatoes, and wheat, but is not applied to estimate N20 emissions from other crops or
rotations with other crops,18 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 N20 emissions associated with these crops and rotations, land uses, as well as organic
soils or cobbly, gravelly, and shaley mineral soils. In addition, federal grassland areas are not simulated with
DAYCENT due to limited activity data on land use histories. For areas that are not included in the DAYCENT
simulations, the Tier 1IPCC (2006) methodology is used to estimate (1) direct emissions from crops on mineral
soils that are not simulated by DAYCENT; (2) direct emissions from PRP on federal grasslands; and (3) direct
emissions from drained organic soils in croplands and grasslands.
A splicing method is used to estimate soil N20 emissions from 2013 to 2017 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 2012 emissions that are derived using the Tier 3 method. Surrogate data for these regression models
include corn and soybean yields from USD A-NASS statistics,19 and weather data from the PRISM Climate Group
(PRISM 2015). For the Tier 1 method, a linear-time series model is used to estimate emissions from 2013 to 2017
without surrogate data. See Box 5-4 for more information about the splicing method. Emission estimates for 2013 to
2017 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 gaps
at the end of the time series. This is mainly because the National Resources Inventory (NRI), 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 2012 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 + e,
where Y is the response variable (e.g., soil organic carbon), xp for the Tier 3 data 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, xp for the Tier 1 data
only contains year as a predictor of emission patterns over the time series, and therefore, is a linear time series
model with no surrogate data. Parameters are estimated from the emissions data for 1990 to 2012 using standard
statistical techniques, and these estimates are used in the model described above to predict the missing emissions
data for 2013 to 2017.
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, which is produced with the Tier 1 and 3 methods, is
combined with the uncertainty in the parameters from the data splicing model. The approach requires estimating
parameters for results in each Monte Carlo simulation for the full inventory (i.e., the surrogate data model is refit
18 A small proportion of the major commodity crop production, such as com 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.
19 See .
5-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
with the emissions estimated in each Monte Carlo iteration from the full inventory analysis with data from 1990 to
2012). Therefore, the data splicing method generates emissions estimates from each surrogate data model, which
are used to derive confidence intervals in the estimates for the missing emissions data from 2013 to 2017.
Tier 3 Approach for Mineral Cropland Soils
The DAYCENT biogeochemical model (Parton et al. 1998; Del Grosso et al. 2001 and 2011) is used to estimate
direct N20 emissions from mineral cropland soils that are managed for production of a wide variety of crops (see list
in previous paragraph) based on the cropping histories in the 2012 NRI (USDA-NRCS 2015). Crops simulated by
DAYCENT are grown on approximately 91 percent of total cropland area in the United States. For agricultural
systems in the central region of the United States, crop production for key crops (i.e., corn, soybeans, sorghum,
cotton, and wheat) is simulated in DAYCENT with a NASA-CASA production algorithm (Potter et al. 1993; Potter
et al. 2007) using the Moderate Resolution Imaging Spectroradiometer (MODIS) Enhanced Vegetation Index (EVI)
products, MOD13Q1 and MYD13Q1, with a pixel resolution of 250m.20
DAYCENT is used to estimate direct N20 emissions due to mineral N available from the following sources: (1) the
application of synthetic fertilizers; (2) the application of livestock manure; (3) the retention of crop residues 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 soil organic matter; and (5) asymbiotic fixation. Note
that commercial organic fertilizers (TVA 1991 through 1994; AAPFCO 1995 through 2015) are addressed with the
Tier 1 method because county-level application data would be needed to simulate applications in DAYCENT, and
currently data are only available at the national scale. The third and fourth sources are generated internally by the
DAYCENT model.
Synthetic fertilizer data are based on fertilizer use and rates by crop type for different regions of the United States,
and are obtained primarily from the USDA Economic Research Service. The data collection program was known as
the Cropping Practices Surveys through 1995 (USDA-ERS 1997), and then became the Agricultural Resource
Management Surveys (ARMS) (USDA-ERS 2015). Additional data are compiled through other sources particularly
the National Agricultural Statistics Service (NASS 1992, 1999, 2004). Frequency and rates of livestock manure
application to cropland during 1997 are estimated from data compiled by the USDA Natural Resources
Conservation Service (Edmonds et al. 2003), and then adjusted using county-level estimates of manure available for
application in other years. The adjustments are based on county-scale ratios of manure available for application to
soils in other years relative to 1997 (see Annex 3.12 for further details). Greater availability of managed manure N
relative to 1997 is assumed to increase the area amended with manure, while reduced availability of manure N
relative to 1997 is assumed to reduce the amended area. Data on the county-level N available for application is
estimated for managed manure systems based on the total amount of N excreted in manure minus N losses during
storage and transport, and including the addition of N from bedding materials. Nitrogen losses include direct N20
emissions, volatilization of ammonia and NOx, runoff and leaching, and poultry manure used as a feed supplement.
For unmanaged manure systems, it is assumed that no N losses or additions occur prior to the application of manure
to the soil. More information on livestock manure production is available in Section 5.2 Manure Management and
Annex 3.11.
The IPCC approach considers crop residue N and N mineralized from soil organic matter as activity data. However,
they are not treated as activity data in DAYCENT simulations because residue production, symbiotic N fixation
(e.g., legumes), mineralization of N from soil organic matter, and asymbiotic N fixation are internally generated by
the model as part of the simulation. In other words, DAYCENT accounts for the influence of symbiotic N fixation,
mineralization of N from soil organic matter and crop residue retained in the field, and asymbiotic N fixation on
N20 emissions, but these are not model inputs. The N20 emissions from crop residues are reduced by approximately
3 percent (the assumed average burned portion for crop residues in the United States) to avoid double-counting
associated with non-C02 greenhouse gas emissions from agricultural residue burning. The estimate of residue
burning is based on state inventory data (ILENR 1993; Oregon Department of Energy 1995; Noller 1996; Wisconsin
Department of Natural Resources 1993; Cibrowski 1996).
20 See .
Agriculture 5-35
-------
Additional sources of data are used to supplement the mineral N (USDA-ERS 1997, 2011), livestock manure
(Edmonds et al. 2003), and land-use information (USDA-NRCS 2015). The Conservation Technology Information
Center (CTIC 2004) provides annual data on tillage activity with adjustments for long-term adoption of no-till
agriculture (Towery 2001). Tillage has an influence on soil organic matter decomposition and subsequent soil N20
emissions. The time series of tillage data from CTIC began in 1989 and ended in 2004, so further changes in tillage
practices since 2004 are not currently captured in the Inventory and practices used in 2004 are assumed to apply for
subsequent years. Daily weather data are used as an input in the model simulations, based on gridded weather data at
a 4 km scale from the PRISM Climate Group (PRISM 2015). Soil attributes are obtained from the Soil Survey
Geographic Database (SSURGO) (Soil Survey Staff 2011).
Each NRI point is run 100 times as part of the uncertainty assessment, yielding a total of over 18 million simulations
for the analysis. Soil N20 emission estimates from DAYCENT are adjusted using a structural uncertainty estimator
to account for uncertainty in model algorithms and parameter values (Del Grosso et al. 2010). Soil N20 emissions
and associated 95 percent confidence intervals are estimated for each year between 1990 and 2012, but emissions
from 2013 to 2017 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 2012 (USDA-NRCS 2015), and
the Inventory time series will be updated in the future when new NRI data are released.
Nitrous oxide emissions from managed agricultural lands are the result of interactions among anthropogenic
activities (e.g., N fertilization, manure application, tillage) and other driving variables, such as weather and soil
characteristics. These factors influence key processes associated with N dynamics in the soil profile, including
immobilization of N by soil microbial organisms, decomposition of organic matter, plant uptake, leaching, runoff,
and volatilization, as well as the processes leading to N20 production (nitrification and denitrification). It is not
possible to partition N20 emissions into each anthropogenic activity directly from model outputs due to the
complexity of the interactions (e.g., N20 emissions from synthetic fertilizer applications cannot be distinguished
from those resulting from manure applications). To approximate emissions by activity, the amount of mineral N
added to the soil, or made available through decomposition of soil organic matter and plant litter, as well as
asymbiotic fixation of N from the atmosphere, is determined for each N source and then divided by the total amount
of mineral N in the soil according to the DAYCENT model simulation. The percentages are then multiplied by the
total of direct N20 emissions in order to approximate the portion attributed to N management practices. This
approach is only an approximation because it assumes that all N made available in soil has an equal probability of
being released as N20, regardless of its source, which is unlikely to be the case (Delgado et al. 2009). However, this
approach allows for further disaggregation of emissions by source of N, which is valuable for reporting purposes
and is analogous to the reporting associated with the IPCC (2006) Tier 1 method, in that it associates portions of the
total soil N20 emissions with individual sources of N.
Tier 1 Approach for Mineral Cropland Soils
The IPCC (2006) Tier 1 methodology is used to estimate direct N20 emissions for mineral cropland soils that are not
simulated by DAYCENT (e.g., DAYCENT has not been parametrized to simulate all crop types and some soil types
such as Histosols). For the Tier 1 Approach, estimates of direct N20 emissions from N applications are based on
mineral soil N that is made available from the following practices: (1) the application of synthetic commercial
fertilizers; (2) application of managed manure and non-manure commercial organic fertilizers; and (3)
decomposition and 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.21 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 not
simulated by DAYCENT. The total amount of fertilizer used on farms has been estimated at the county-
level by the USGS from sales records from 1990 to 2001 (Ruddy et al. 2006), and these data are aggregated
21 Commercial organic fertilizers include dried blood, tankage, compost, and other, but the dried manure and biosolids (i.e.,
sewage sludge) is removed from the dataset in order to avoid double counting with other datasets that are used for manure N and
biosolids.
5-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
to obtain state-level N additions to farms. For 2002 through 2012, state-level fertilizer for on-farm use is
adjusted based on annual fluctuations in total U.S. fertilizer sales (AAPFCO 1995 through 2007, 2008
through 2012). After subtracting the portion of fertilizer applied to crops and grasslands simulated by
DAYCENT (see Tier 3 Approach for Mineral Cropland Soils and Direct N20 Emissions from Grassland
Soils sections for information on data sources), the remainder of the total fertilizer used on farms is
assumed to be applied to crops that are not simulated by DAYCENT.
• Similarly, a process-of-elimination approach is used to estimate manure N additions for crops that are not
simulated by DAYCENT. The amount of manure N applied annually in the Tier 3 approach to crops and
grasslands is subtracted from total annual manure N available for land application (see Tier 3 Approach for
Mineral Cropland Soils and Direct N20 Emissions from Grassland Soils sections for information on data
sources), and 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 2012). Commercial fertilizers do include some manure and biosolids (i.e., 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., 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 2013), crop production yield statistics (USDA-
NASS 2018), dry matter fractions (IPCC 2006), linear equations to estimate above-ground biomass given
dry matter crop yields from harvest (IPCC 2006), ratios of below-to-above-ground biomass (IPCC 2006),
and N contents of the residues (IPCC 2006). N inputs from residue were reduced by 3 percent to account
for average residue burning portions in the United States.
The total increase in soil mineral N from applied fertilizers and crop residues is multiplied by the IPCC (2006)
default emission factor to derive an estimate of direct N20 emissions using the Tier 1 Approach.
Tier 1 soil N20 emissions from 2013 to 2017 are estimated using a splicing method that is described in Box 5-4,
with the exception of commercial fertilizer additions, which are estimated with a splicing method from 2015 to
2017. As with the Tier 3 method, the time series will be recalculated in future Inventory reports (see Planned
Improvements section).
Tier 1 Approach for Drainage of Organic Soils in Croplands and Grasslands
The IPCC (2006) Tier 1 methods are used to estimate direct N20 emissions due to drainage of organic soils in
croplands and grasslands at a state scale. State-scale estimates of the total area of drained organic soils are obtained
from the 2012 NRI (USDA-NRCS 2015) using soils data from the Soil Survey Geographic Database (SSURGO)
(Soil Survey Staff 2011). Temperature data from Daly et al. (1994 and 1998) are used to subdivide areas into
temperate and tropical climates using 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 2012. Consequently, emissions from 2013 to 2017 are estimated using a linear time
series model (see Box 5-4). Estimates for 2013 to 2017 will be recalculated in future Inventory reports when new
NRI data are available.
Tier 1 and 3 Approaches for Direct N2O Emissions from Grassland Soils
As with N20 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 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 interseeding legumes. DAYCENT is used to simulate N20 emissions from NRI
survey locations (USDA-NRCS 2015) 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), and synthetic fertilizer application. Other
Agriculture 5-37
-------
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 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. Managed manure N amendments to grasslands are estimated from Edmonds et al. (2003) and adjusted for
annual variation using data on the availability of managed manure N for application to soils, according to methods
described in the Manure Management section (Section 5.2) and Annex 3.11. Biological N fixation is simulated
within DAY CENT, and therefore is not an input to the model.
Manure N deposition from grazing animals in PRP systems (i.e., PRP manure) is another key input of N to
grasslands. The amounts of PRP manure N applied on non-federal grasslands for each NRI point are based on
amount of N excreted by livestock in PRP systems. 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 78 percent
of total PRP manure N in aggregate across the country. The remainder of the PRP manure N in each state is assumed
to be excreted on federal grasslands, and the N20 emissions are estimated using the IPCC (2006) Tier 1 method with
IPCC default emission factors.
Biosolids (i.e., 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 sewage sludge available for land application
application). Biosolids soil amendments are only available at the national scale, and it is not possible to associate
application with specific soil conditions and weather at NRI survey locations. Therefore, DAYCENT could not be
used to simulate the influence of biosolids amendments on N20 emissions from grassland soils, and consequently,
emissions from biosolids are estimated using the IPCC (2006) Tier 1 method.
As previously mentioned, each NRI point is simulated 100 times as part of the uncertainty assessment, yielding a
total of over 18 million simulation runs for the analysis. Soil N20 emission estimates from DAYCENT are adjusted
using a structural uncertainty estimator accounting for uncertainty in model algorithms and parameter values (Del
Grosso et al. 2010). N20 emissions for the PRP manure N deposited on federal grasslands and applied biosolids N
are estimated using the Tier 1 method by multiplying the N input by the default emission factor. Emissions from
manure N are estimated at the state level and aggregated to the entire country, but emissions from biosolids N are
calculated exclusively at the national scale.
Soil N20 emissions and 95 percent confidence intervals are estimated for each year between 1990 and 2012 based
on the Tier 1 and 3 methods, with the exception of biosolids (discussed below), and emissions from 2013 to 2017
are estimated using a splicing method as described in Box 5-4. As with croplands, estimates for 2013 to 2017 will be
recalculated in future inventories when new NRI data are available. Biosolids application data are compiled through
2017 in this Inventory, and therefore soil N20 emissions and confidence intervals are estimated using the Tier 1
method for all years in the time series without application of the splicing method.
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 N20 emissions from agricultural soil management
(see Table 5-15 and Table 5-16).
Indirect N2O Emissions
This section describes the methods used for estimating indirect soil N20 emissions from croplands and grasslands.
Indirect N20 emissions occur when mineral N made available through anthropogenic activity is transported from the
soil either in gaseous or aqueous forms and later converted into N20. There are two pathways leading to indirect
emissions. The first pathway results from volatilization of N as NOx and NH3 following application of synthetic
fertilizer, organic amendments (e.g., manure, biosolids), and deposition of PRP manure. Nitrogen made available
from mineralization of soil organic matter and residue, including N incorporated into crops and forage from
5-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
symbiotic N fixation, and input of N from asymbiotic fixation also contributes to volatilized N emissions.
Volatilized N can be returned to soils through atmospheric deposition, and a portion of the deposited N is emitted to
the atmosphere as N20. The second pathway occurs via leaching and runoff of soil N (primarily in the form of 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 N20 emissions. Regardless
of the eventual location of the indirect N20 emissions, the emissions are assigned to the original source of the N for
reporting purposes, which here includes croplands and grasslands.
Tier 1 and 3 Approaches for Indirect N2O Emissions from Atmospheric Deposition of Volatilized
N
The Tier 3 DAYCENT model and IPCC (2006) Tier 1 methods are combined to estimate the amount of N that is
volatilized and eventually emitted as N20. DAYCENT is used to estimate N volatilization for land areas whose
direct emissions are simulated with DAYCENT (i.e., most commodity and some specialty crops and most
grasslands). The N inputs included are the same as described for direct N20 emissions in the Tier 3 Approach for
Mineral Cropland Soils and Direct N20 Emissions from Grassland Soils sections. Nitrogen volatilization from all
other areas is estimated using the Tier 1 method and default IPCC fractions for N subject to volatilization (i.e., N
inputs on croplands not simulated by DAYCENT, PRP manure N excreted on federal grasslands, biosolids [i.e.,
sewage sludge] application on grasslands). For the volatilization data generated from both the DAY CENT and Tier
1 approaches, the IPCC (2006) default emission factor is used to estimate indirect N20 emissions occurring due to
re-deposition of the volatilized N (see Table 5-18).
Tier 1 and 3 Approaches for Indirect N2O Emissions from Leaching/Runoff
As with the calculations of indirect emissions from volatilized N, the Tier 3 DAYCENT model and IPCC (2006)
Tier 1 method are combined to estimate the amount of N that is subject to leaching and surface runoff into water
bodies, and eventually emitted as N20. DAYCENT is used to simulate the amount of N transported from lands in
the Tier 3 Approach. Nitrogen transport from all other areas is estimated using the Tier 1 method and the IPCC
(2006) default factor for the proportion of N subject to leaching and runoff. This N transport estimate includes 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 N20 in cropland and grassland systems in arid regions,
as discussed in IPCC (2006). In the United States, the threshold for significant nitrate leaching is based on the
potential evapotranspiration (PET) and rainfall amount, similar to IPCC (2006), and is assumed to be negligible in
regions where the amount of precipitation 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 N20 emissions that occur in groundwater and waterways (see Table 5-18).
Indirect soil N20 emissions from 2013 to 2017 are estimated using the splicing method that is described in Box 5-4.
As with the direct N20 emissions, the time series will be recalculated in a future Inventory report when new activity
data are compiled (see Planned Improvements section).
Uncertainty and Time-Series Consistency
Uncertainty is estimated for each of the following five components of N20 emissions from agricultural soil
management: (1) direct emissions simulated by DAYCENT; (2) the components of indirect emissions (N volatilized
and leached or runoff) simulated by DAYCENT; (3) direct emissions 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 N20 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 2013
to 2017, there is additional uncertainty propagated through the Monte Carlo Analysis associated with the splicing
method (See Box 5-4).
Agriculture 5-39
-------
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 N20 emissions. Uncertainty in the splicing method is also
included in the error propagation for 2013 to 2017 (see Box 5-4). Additional details on the uncertainty methods are
provided in Annex 3.12. Table 5-19 shows the combined uncertainty for direct soil N20 emissions ranged from 17
percent below to 18 percent above the 2017 emission estimate of 227.7 MMT CO2 Eq., and the combined
uncertainty for indirect soil N20 emissions range from 58 percent below to 143 percent above the 2017 estimate of
38.8 MMT C02Eq.
Table 5-19: Quantitative Uncertainty Estimates of N2O Emissions from Agricultural Soil
Management in 2017 (MMT CO2 Eq. and Percent)
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Direct Soil N2O Emissions
N2O
227.7
185.7 265.0
-17% +18%
Indirect Soil N2O Emissions
N2O
38.8
16.3 94.9
-58% +143%
Notes: Due to lack of data, uncertainties in managed manure N production, PRP manure N production, other organic
fertilizer amendments, and biosolids (i.e., sewage sludge) amendments to soils are currently treated as certain; these
sources of uncertainty will be included in future Inventory reports.
Additional uncertainty is associated with an incomplete estimation of N20 emissions from managed croplands and
grasslands in Hawaii and Alaska. The Inventory currently includes the N20 emissions from mineral fertilizer and
PRP N additions in Alaska and Hawaii, and drained organic soils in Hawaii. Land areas used for agriculture in
Alaska and Hawaii are small relative to major commodity cropping 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 2017. Details on the emission trends through time are described in more detail in the Methodology section.
QA/QC and Verification
General (Tier 1) and category-specific (Tier 2) QA/QC activities were conducted consistent with the U.S. Inventory
QA/QC plan outlined in Annex 8. DAYCENT results for N20 emissions and 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 are available for 41 sites, which
mostly occur in the United States, with five in Europe and three in Australia, representing over 200 different
combinations of fertilizer treatments and cultivation practices. Nitrate leaching data are available for four sites in the
United States, representing 10 different combinations of fertilizer amendments/tillage practices. DAYCENT
estimates of N20 emissions are closer to measured values at most sites compared to the IPCC Tier 1 estimate (see
Figure 5-10). In general, the IPCC Tier 1 methodology tends to over-estimate emissions when observed values are
low and under-estimate emissions when observed values are high, while DAYCENT estimates have less bias.
DAYCENT accounts for key site-level factors (i.e., weather, soil characteristics, and management) that are not
addressed in the IPCC Tier 1 method, and thus the model is better able to represent the variability in N20 emissions.
DAYCENT does have a tendency to under-estimate very high N20 emission rates; and estimates are adjusted using
the statistical model derived from the comparison of model estimates to measurements (see Annex 3.12 for more
information). Regardless, the comparison demonstrates that DAYCENT provides relatively high predictive
capability for N20 emissions, and is an improvement over the IPCC Tier 1 method.
5-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 5-10: Comparison of Measured Emissions at Field Sites and Modeled Emissions Using
the DAYCENT Simulation Model and IPCC Tier 1 Approach (kg N2O per ha per year)
9
8
^ 7
>• 6
Ł 5
4
9,3
s 2
1
0
C
i—
o
"O
a>
O
u
~ Measured
E3 DayCent
I IPCC
j
c
I—
o
>-
Hj,rD.,nJ
ikyjLl
I
>¦
O
C
o
Q
on
clO
O
u
ob
03
QJ
ra
it
ra
5 5
>-
c
ro
>-
o
c
1—
o
to
D
<
>¦
-
O
10
c
s—
o
c
ro
-
o
CO
c
s—
o
to
(U
c
c
10
CO
QJ
i—
l_
00
ra
Q.
0
0
ro
cCb
<_>
u
cib
<
ra
X
z
_i
Q_
-------
development. Experimental study sites will continue to be added for quantifying model structural uncertainty.
Studies that have continuous (daily) measurements of N20 (e.g., Scheer et al. 2013) will be given priority.
Improvements are underway to simulate crop residue burning in the DAYCENT model based on the amount of crop
residues burned according to the data that is used in the Field Burning of Agricultural Residues source category (see
Section 5.7). Alaska and Hawaii are not included for all sources in the current Inventory for agricultural soil
management, with the exception of N20 emissions from drained organic soils in croplands and grasslands for
Hawaii, synthetic fertilizer and PRP N amendments for grasslands in Alaska and Hawaii. A planned improvement to
add the remaining sources for these states into the Inventory analysis. There is also an improvement based on
updating the Tier 1 emission factor for N20 emissions from drained organic soils by using the revised factor in the
2013 Supplement to the 2006IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands (IPCC 2013).
There is also a planned improvement to improve the implementation of the Tier 1 method. Specifically, Soil N20
emissions will be estimated and reported for N mineralization from soil organic matter decomposition that is
accelerated with Forest Land Converted to Cropland and Grassland Converted to Cropland.
These improvements are expected to be completed for the next Inventory (i.e., 2020 submission to the UNFCCC,
1990 through 2018 Inventory). However, the time line may be extended if there are insufficient resources to fund all
or part of these planned improvements.
5.5 Liming (CRF Source Category 3G)
Crushed limestone (CaCCh) and dolomite (CaMg(C03)2) are added to soils by land managers to increase soil pH
(i.e., to reduce acidification). Carbon dioxide emissions occur as these compounds react with hydrogen ions in soils.
The rate of degradation of applied limestone and dolomite depends on the soil conditions, soil type, climate regime,
and whether limestone or dolomite is applied. Emissions from liming of soils have fluctuated over the past 25 years
in the United States, ranging from 3.2 MMT C02 Eq. to 6.0 MMT C02 Eq. In 2017, liming of soils in the United
States resulted in emissions of 3.2 MMT C02 Eq. (0.9 MMT C), representing a 32 percent decrease in emissions
since 1990 (see Table 5-20 and Table 5-21). The trend is driven by variation in the amount of limestone and
dolomite applied to soils over the time period.
Table 5-20: Emissions from Liming (MMT CO2 Eq.)
Source 1990
2005
2013 2014 2015 2016 2017
Limestone 4.1
Dolomite 0.6
3.9
0.4
3.6 3.3 3.5 2.9 2.9
0.3 0.3 0.3 0.3 0.3
H
0
¦u
-4
4.3
3.9 3.6 3.7 3.2 3.2
Note: Totals may not sum due to independent rounding.
Table 5-21: Emissions from Liming (MMT C)
Source 1990
2005
2013 2014 2015 2016 2017
Limestone 1.1
Dolomite 0.2
1.1
0.1
1.0 0.9 0.9 0.8 0.8
0.1 0.1 0.1 0.1 0.1
Total 1.3
1.2
1.1 1.0 1.0 0.9 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 5-22)
were multiplied by C02 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
5-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
inthe Minerals Yearbook and Mineral Industry Sun'eys (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).
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
CaCCb 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
2017 U.S. emission estimate from liming of soils is 3.2 MMT CO2 Eq. using the U.S.-specific factors. In contrast,
emissions would be estimated at 6.5 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 2017 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 2017 data, 2016 fractions were applied to a 2017
estimate of total crushed stone presented in the USGS Mineral Industry Sun'eys: Crushed Stone and Sand and
Gravel in the First Quarter of 2018 (USGS 2018).
The primary source for limestone and dolomite activity data is the Minerals Yearbook, published by the Bureau of
Mines through 1994 and by the USGS from 1995 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.
Agriculture 5-43
-------
Table 5-22: Applied Minerals (MMT)
Mineral 1990
2005
2013 2014 2015 2016 2017
Limestone 19.0
Dolomite 2.4
18.1
1.9
16.4 15.3 16.0 13.5 13.4
1.5 1.3 1.2 1.2 1.2
Uncertainty and Time-Series Consistency
Uncertainty regarding the amount of limestone and dolomite applied to soils was estimated at ±15 percent with
normal densities (Tepordei 2003; Willett 2013b). Analysis of the uncertainty associated with the emission factors
included the fraction of lime dissolved by nitric acid versus the fraction that reacts with carbonic acid, and the
portion of bicarbonate that leaches through the soil and is transported to the ocean. Uncertainty regarding the time
associated with leaching and transport was not addressed in this analysis, but is assumed to be a relatively small
contributor to the overall uncertainty (West 2005). The probability distribution functions for the fraction of lime
dissolved by nitric acid and the portion of bicarbonate that leaches through the soil were represented as triangular
distributions between ranges of zero and 100 percent of the estimates. The uncertainty surrounding these two
components largely drives the overall uncertainty.
A Monte Carlo (Approach 2) uncertainty analysis was applied to estimate the uncertainty in CO2 emissions from
liming. The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 5-23. Carbon
dioxide emissions from carbonate lime application to soils in 2017 were estimated to be between -0.35 and 6.01
MMT CO2 Eq. at the 95 percent confidence level. This confidence interval represents a range of 111 percent below
to 89 percent above the 2017 emission estimate of 3.2 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 5-23: Approach 2 Quantitative Uncertainty Estimates for CO2 Emissions from Liming
(MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMTCCfcEq.) (%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Liming
CO2
3.2
(0.35) 6.01
-111% +89%
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 2017.
QA/QC and Verification
A source-specific QA/QC plan for liming has been developed and implemented, consistent with the U.S. Inventory
QA/QC plan outlined in Annex 8. The quality control effort focused on the Tier 1 procedures for this Inventory. No
errors were found.
Recalculations Discussion
Adjustments were made in the current Inventory to improve the results. First, limestone and dolomite application
data for 2015 and 2016 were updated with the recently published data fromUSGS (2018), rather than being
approximated by a ratio method for 2016. With this revision in the activity data, the emissions decreased by 1.1 and
17 percent for 2015 and 2016, respectively, relative to the previous Inventory estimates.
5-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
5.6 Urea Fertilization (CRF Source Category 3H)
The use of urea (CCKNPbh) as a fertilizer leads to greenhouse gas emissions through the release of CO2 that was
fixed during the industrial production process. In the presence of water and urease enzymes, urea is converted into
ammonium (NH4+), hydroxyl ion (OH), and bicarbonate (HCO3 ). The bicarbonate then evolves into CO2 and water.
Emissions from urea fertilization in the United States totaled 5.1 MMT CO2 Eq. (1.4 MMT C) in 2017 (Table 5-24
and Table 5-25). Carbon dioxide emissions have increased by 109 percent between 1990 and 2017 due to an
increasing amount of urea that is applied to soils.
Table 5-24: CO2 Emissions from Urea Fertilization (MMT CO2 Eq.)
Source
1990
2005
2013
2014
2015
2016
2017
Urea Fertilization
2.4
3.5
4.4
4.5
4.7
4.9
5.1
Table 5-25: CO2 Emissions from Urea Fertilization (MMT C)
Source
1990
2005
2013
2014
2015
2016
2017
Urea Fertilization
0.7
1.0
1.2
1.2
1.3
1.3
1.4
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 production process of urea are
released after application. The annual amounts of urea applied to croplands (see Table 5-26) were derived from the
state-level fertilizer sales data provided in Commercial Fertilizer reports (TVA 1991, 1992, 1993, 1994; AAPFCO
1995 through 2018).22 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. Because fertilizer
sales data are reported in fertilizer years (July previous year through June current year), 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 and 2017 fertilizer years (i.e., July 2015 through June 2016 and July 2016 through
June 2017) were not available for this Inventory. Therefore, urea application in the 2016 and 2017 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 5-26: Applied Urea (MMT)
1990
2005
2013
2014
2015
2016
2017
Urea Fertilizer3
3.3 |
4.8 |
6.1
6.2
6.5
6.7
6.9
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.
22 Hie 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 5-45
-------
Uncertainty and Time-Series Consistency
Uncertainty estimates are presented in Table 5-27 for urea fertilization. An Approach 2 Monte Carlo analysis was
completed. 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 may be retained in the soil, and therefore the uncertainty range was set from 0
percent emissions to the maximum emission value of 100 percent using a triangular distribution. In addition, urea
consumption data also have uncertainty that is propagated through the emission calculation using a Monte Carlo
simulation approach as described by the IPCC (2006). Carbon dioxide emissions from urea fertilization of
agricultural soils in 2017 were estimated to be between 2.89 and 5.21 MMT CO2 Eq. at the 95 percent confidence
level. This indicates a range of 43 percent below to 3 percent above the 2017 emission estimate of 5.1 MMT CO2
Eq.
Table 5-27: Quantitative Uncertainty Estimates for CO2 Emissions from Urea Fertilization
(MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission
Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Urea Fertilization
CO2
5.1
2.89 5.21 -43% +3%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
There are additional uncertainties that are not quantified in this analysis. 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 2017. Details on the emission trends through time are described in more detail in the Introduction, above.
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 outlined in Annex 8. 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
Recalculations resulted from updated urea application estimates in a new AAPFCO report (2018). New activity data
for 2015 were applied to all states; 2016 and 2017 estimates were derived using the new data for 2011 and 2015.
This resulted in an emissions decrease for the United States of 0.5 percent in 2014, 3.3 percent in 2015, and 4.3
percent in 2016.
5-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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. In addition, the estimate of CO2 emissions is not
based on a Monte Carlo uncertainty analysis, but rather just the deterministic result based on multiplying the
emission factor by the amount of urea without addressing uncertainty. It would be more robust to use the uncertainty
analysis as the basis for the estimates. These improvements will be implemented in a future Inventory.
5.7 Field Burning of Agricultural Residues (CRF
Source Category 3F)
Crop production creates large quantities of agricultural crop residues, which fanners 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 CH4, N20, CO, and NOx, which are
released during combustion.
In the United States, field burning of agricultural residues commonly occurs in southeastern states, the Great Plains,
and the Pacific Northwest (McCarty 2011). The primary crops that are managed with residue burning include corn,
cotton, lentils, rice, soybeans, and wheat (McCarty 2009). In 2017, CH4 and N20 emissions from field burning of
agricultural residues were 0.2 MMT CO2 Eq. (8 kt) and 0.1 MMT CO2 Eq. (0.3 kt), respectively (Table 5-28 and
Table 5-29). Annual emissions of CH4 and N20 have increased from 1990 to 2017 by 82 percent and 72 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 5-28: ChU and N2O Emissions from Field Burning of Agricultural Residues (MMT CO2
Eq.)
Gas/Crop Type
1990
2005
2013
2014
2015
2016
2017
CH4
0.1
0.2
0.2
0.2
0.2
0.2
0.2
Maize
+
+
0.1
0.1
0.1
0.1
0.1
Rice
+
0.1
+
+
+
+
+
Wheat
+
+
+
+
+
+
+
Barley
+
+
+
+
+
+
+
Oats
+
+
+
+
+
+
+
Other Small Grains
+
+
+
+
+
+
+
Sorghum
+
+
+
+
+
+
+
Cotton
+
+
+
+
+
+
+
Grass Hay
+
+
+
+
+
+
+
Legume Hay
+
+
+
+
+
+
+
Peas
+
+
+
+
+
+
+
Sunflower
+
+
+
+
+
+
+
Tobacco
+
+
+
+
+
+
+
Vegetables
+
+
+
+
+
+
+
Chickpeas
+
+
+
+
+
+
+
Dry Beans
+
+
+
+
+
+
+
Lentils
+
+
+
+
+
+
+
Peanuts
+
+
+
+
+
+
+
Agriculture 5-47
-------
Soybeans
+
+
+
+
+
+
+
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
N2O
+
0.1
0.1
0.1
0.1
0.1
0.1
Maize
+
+
+
+
+
+
+
Rice
+
+
+
+
+
+
+
Wheat
+
+
+
+
+
+
+
Barley
+
+
+
+
+
+
+
Oats
+
+
+
+
+
+
+
Other Small Grains
+
+
+
+
+
+
+
Sorghum
+
+
+
+
+
+
+
Cotton
+
+
+
+
+
+
+
Grass Hay
+
+
+
+
+
+
+
Legume Hay
+
+
+
+
+
+
+
Peas
+
+
+
+
+
+
+
Sunflower
+
+
+
+
+
+
+
Tobacco
+
+
+
+
+
+
+
Vegetables
+
+
+
+
+
+
+
Chickpeas
+
+
+
+
+
+
+
Dry Beans
+
+
+
+
+
+
+
Lentils
+
+
+
+
+
+
+
Peanuts
+
+
+
+
+
+
+
Soybeans
+
+
+
+
+
+
+
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
Total
0.2
0.3
0.3
0.3
0.3
0.3
0.3
+ Does not exceed 0.05 MMT CO2 Eq.
Note: Totals may not sum due to independent rounding.
Table 5-29: ChU, N2O, CO, and NOx Emissions from Field Burning of Agricultural Residues
(kt)
Gas/Crop Type
1990
2005
2013
2014
2015
2016
2017
CH4
4
7
8
8
8
8
8
Maize
+
1
3
3
3
3
3
Rice
2
2
2
1
2
1
2
Wheat
1
2
1
1
1
1
1
Barley
+
+
+
+
+
+
+
Oats
+
+
+
+
+
+
+
Other Small Grains
+
+
+
+
+
+
+
Sorghum
+
+
+
+
+
+
+
Cotton
+
1
1
1
1
1
1
Grass Hay
+
+
+
+
+
+
+
Legume Hay
+
+
+
+
+
+
+
Peas
+
+
+
+
+
+
+
Sunflower
+
+
+
+
+
+
+
Tobacco
+
+
+
+
+
+
+
Vegetables
+
+
+
+
+
+
+
Chickpeas
+
+
+
+
+
+
+
Dry Beans
+
+
+
+
+
+
+
5-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Lentils
+
+
+
+
+
+
+
Peanuts
+
+
+
+
+
+
+
Soybeans
+
1
1
1
1
1
1
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
n2o
+
+
+
+
+
+
+
Maize
+
+
+
+
+
+
+
Rice
+
+
+
+
+
+
+
Wheat
+
+
+
+
+
+
+
Barley
+
+
+
+
+
+
+
Oats
+
+
+
+
+
+
+
Other Small Grains
+
+
+
+
+
+
+
Sorghum
+
+
+
+
+
+
+
Cotton
+
+
+
+
+
+
+
Grass Hay
+
+
+
+
+
+
+
Legume Hay
+
+
+
+
+
+
+
Peas
+
+
+
+
+
+
+
Sunflower
+
+
+
+
+
+
+
Tobacco
+
+
+
+
+
+
+
Vegetables
+
+
+
+
+
+
+
Chickpeas
+
+
+
+
+
+
+
Dry Beans
+
+
+
+
+
+
+
Lentils
+
+
+
+
+
+
+
Peanuts
+
+
+
+
+
+
+
Soybeans
+
+
+
+
+
+
+
Potatoes
+
+
+
+
+
+
+
Sugarbeets
+
+
+
+
+
+
+
CO
89
154
157
152
148
144
141
NOx
4
6
6
6
6
6
6
+ Does not exceed 0.5 kt
Note: Totals may not sum due to independent rounding.
Methodology
A U.S.-specific Tier 2 method is used to estimate greenhouse gas emissions from field burning of agricultural
residues from 1990 to 2012 (for more details comparing the U.S.-specific approach to the IPCC (2006) default
approach, see Box 5-6). 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)
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 bio mass for a crop
Amount of C or N per unit of dry matter for a crop
Agriculture 5-49
-------
Burning Efficiency (BE) = The proportion of prefire fuel biomass consumed23
Combustion Efficiency (CE) = The proportion of C or N released with respect to the total amount of C or N
available in the burned material, respectively
Crop production data are available by state and year from USD A (2017) 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.24 Crop area data are based on the 2012 National Resources Inventory (NRI) (USDA-NRCS 2015).
In order to estimate total crop production, the crop yield data from USD A Quick Stats crop yields is multiplied by
the NRI crop areas. The production data for the crop types are presented in Table 5-30. 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 CH4, CO, N20, and NOx emissions from the Field Burning of Agricultural Residues:
CH4 and CO, or N2O and NOx = 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 2012 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 Statistics25 (metric tons dry matter burnt ha ')
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 22 percent lower emissions of CH4 and 44 percent lower
emissions of N2O compared to this Inventory. In summary, the IPCC/UNEP/OECD/IEA (1997) method is
23 In IPCC/UNEP/OECD/IEA (1997), the equation for C orN released contains the variable 'fraction oxidized in burning'. This
variable is equivalent to (burning efficiency x combustion efficiency).
24 See Annex 5 regarding the burning of sugarcane or rye.
25 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.
5-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 CH4, CO, N20
and NOx, compared to IPCC (2006) approach that is based on dry matter rather than elemental composition.
Table 5-30: Agricultural Crop Production (kt of Product)
Crop 1990 2005 2011 2012
Maize 294,558 370,338 389,657 341,374
Rice 9,476 11,914 10,172 9,829
Wheat 79,920 68,919 61,609 70,867
Barley 9,087 5,042 3,778 5,293
Oats 5,972 2,632 1,625 1,631
Other Small Grains 2,639 2,007 1,237 1,681
Sorghum 23,688 13,080 9,344 10,422
Cotton 4,609 6,227 5,343 5,666
Grass Hay + + + +
Legume Hay + + + 464,050
Peas 64 706 302 534
Sunflower 992 1,397 792 1,204
Tobacco 1,151 347 264 521
Vegetables + 903 1,189 2,027
Chickpeas +7 + +
Dry Beans 638 1,084 1,079 1,159
Lentils + 119 59 121
Peanuts 1,822 2,242 1,906 2,649
Soybeans 56,613 87,164 86,839 84,805
Potatoes 18,960 19,471 20,296 20,517
Sugarbeets 25,017 26,604 28,922 28,488
+ Does not exceed 0.5 kt
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 30m Landsat imagery (LANDFIRE 2014), and from 2003 through 2012 using 1km
Moderate Resolution Imaging Spectroradiometer imagery (MODIS) Global Fire Location Product (MCD14ML)
using combined observations from Terra and Aqua satellites (Giglio et a. 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 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) existence of state laws limiting 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 of tea and avoid models that appear to be influenced by outliers due to the random draw of
survey locations for training the model. In order to address uncertainty, a Monte Carlo analysis is used to sample the
parameter estimates for the logistical regression model and produce one thousand estimates of burning for each crop
in the remaining forty-two states included in this Inventory. State-level area burned data are divided by state-level
Agriculture 5-51
-------
crop area data to estimate the percent of crop area burned by crop type for each state. Table 5-31 shows the resulting
percentage of crop residue burned at the national scale by crop type. State-level estimates are also available upon
request.
Table 5-31: U.S. Average Percent Crop Area Burned by Crop (Percent)
Crop 1990
2005
2011 2012
Maize +
+
+ +
Rice 5%
6%
2% 4%
Wheat +
+
1% 1%
Barley 1%
+
+ +
Oats +
+
+ 1%
Other Small Grains 1%
+
+ +
Sorghum +
1%
1% 1%
Cotton +
1%
1% 1%
Grass Hay +
+
+ +
Legume Hay +
+
+ +
Peas +
2%
3% 3%
Sunflower +
+
+ +
Tobacco 1%
1%
1% 2%
Vegetables +
2%
2% 2%
Chickpeas +
+
+ +
Dry Beans +
+
1% 1%
Lentils +
+
+ +
Peanuts 1%
1%
2% 2%
Soybeans +
+
+ +
Potatoes +
+
+ +
Sugarbeets +
+
+ +
+ Does not exceed 0.5 percent.
Additional parameters are needed to estimate the amount of burning, including residuexrop ratios, dry matter
fractions, carbon fractions, nitrogen fractions, burning efficiency and combustion efficiency. Residuexrop 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 5-32). 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 5-32 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 5-33).
Table 5-32: 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
Sorghum
0.780
0.60
0.47
0.01
0.93
0.88
5-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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; 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; Beans & pulses
Peanuts: IPCC 2006; Table 11.2; Root ratio and belowground N content are from Root crops, other
Sugarbeets: IPCC 2006; Table 11.2; values are for Tubers
Sunflower: IPCC 2006, Table 11.2; values are from 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:
Carrots: McPharlin et al. 1992; Gibberd et al. 2003; Reid and English 2000; Peach et al. 2000; see IPCC Tubers for
R:S andN frac
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); IPCC Grains forN frac
Melons: Valantin et al. 1999; squash for R: S; IPCC Grains for N frac
Onion: Peach et al. 2000, Halvorson et al. 2002; IPCC 2006 Tubers for N fractions
Peppers: combined sources (Costa and Gianquinto 2002; Marcussi et al. 2004; Tadesse et al. 1999; Diaz-Perez et al.
2008); IPCC Grains for N frac
Tomatoes: Scholberg et al. 2000a,b; Akintoye et al. 2005; AGR-N and BGR-N are from Grains
Table 5-33: 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).
Fortius Inventory, new activity data on crop areas are not available for 2013 to 2017 from the USDA National
Resources Inventory (USDA-NRCS 2015). To complete the emissions time series, a linear extrapolation of the trend
is applied to estimate the emissions in the last five years of the inventory. Specifically, a linear regression model
with autoregressive moving-average (ARMA) errors is used to estimate the trend in emissions over time from 1990
through 2012, and the trend is used to approximate the CH4, N20, CO and NOx for the last 5 years in the time series
Agriculture 5-53
-------
from 2013 to 2017 (Brockwell and Davis 2016). The Tier 2 method described previously will be applied to
recalculate the emissions in a future Inventory.
Uncertainty and Time-Series Consistency
Emissions are estimated using a linear regression model with autoregressive moving-average (ARMA) errors for
2017. The linear regression ARMA model produced estimates of the upper and lower bounds to quantify uncertainty
(Table 5-34), and the results are summarized in Table 5-34. Methane emissions from field burning of agricultural
residues in 2017 are between 0.10 and 0.29 MMT CO2 Eq. at a 95 percent confidence level. This indicates a range of
51 percent below and 49 percent above the 2017 emission estimate of 0.2 MMT CO2 Eq. Nitrous oxide emissions
are between 0.04 and 0.11 MMT CO2 Eq., or approximately 47 percent below and 46 percent above the 2017
emission estimate of 0.1 MMT CO2 Eq.
Table 5-34: Approach 2 Quantitative Uncertainty Estimates for ChU and N2O Emissions from
Field Burning of Agricultural Residues (MMT CO2 Eq. and Percent)
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Field Burning of Agricultural
Residues
CH4
0.2
0.10
0.29
-51%
+49%
Field Burning of Agricultural
Residues
N2O
0.1
0.04
0.11
-47%
+46%
Due to data limitations, there are additional uncertainties in agricultural residue burning, particularly the omission of
burning associated with Kentucky bluegrass and "other crop" residues.
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 and corrected in the
analysis of the remote sensing product related to scaling of the data for the entire state for Iowa and Indiana, and
also the application of the logical regression model.
Recalculations Discussion
Methodological recalculations are associated with the new analysis to estimate the area burned based on the
LANDFIRE data products (LANDFIRE 2014) and MODIS Global Fire Location Product (Giglio et al. 2006). The
emissions decreased on average across the times by 43 percent and 28 percent for CH4 and N20, respectively. The
new analysis is considered more robust with an evaluation of burned area across the entire time series using the
remote sensing products, rather than only subset of years, which was used in the previous Inventory.
Planned Improvements
The key planned improvement 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.
5-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
Fluxes are reported for four agricultural land use/land-use change categories: Cropland Remaining Cropland, Land
Converted to Cropland, Grassland Remaining Grassland, and Land Converted to Grassland. The reported
greenhouse gas fluxes from these agricultural lands include changes in soil organic C stocks in mineral and organic
soils due to land use and management, and for the subcategories of Forest Land Converted to Cropland and Forest
Land Converted to Grassland, the changes in aboveground biomass, belowground biomass, dead wood, and litter C
stocks are also reported. The greenhouse gas flux from Grassland Remaining Grassland also includes estimates of
non-C02 emissions from grassland fires.
Fluxes from Wetlands Remaining Wetlands include changes in C stocks and methane (CH4) 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 N20 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 and N2O emissions from soils, and CO2
fluxes 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 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 2017 resulted in a net increase in C stocks (i.e., net
CO2 removals) of 729.6 MMT CO2 Eq. (199.0 MMT C).2 This represents an offset of approximately 11.3 percent of
1 The term "flux" is used to describe the net emissions of greenhouse gases accounting for both the emissions of CO2 to and the
removals of CO2 from the atmosphere. Removal of CO2 from the atmosphere is also referred to as "carbon sequestration."
2 LULUCF Carbon Stock Change is the net C stock change from the following categories: Forest Land Remaining Forest Land,
Land Converted to Forest Land, Cropland Remaining Cropland, Land Converted to Cropland, Grassland Remaining Grassland,
Land Use, Land-Use Change, and Forestry 6-1
-------
total (i.e., gross) greenhouse gas emissions in 2017. Emissions of CH4 and N20 from LULUCF activities in 2017 are
15.5 MMT CO2 Eq. and represent 0.2 percent of total greenhouse gas emissions.3
Total C sequestration in the LULUCF sector decreased by approximately 10.5 percent between 1990 and 2017. 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 2017, while the net C
accumulation in Forest Land Remaining Forest Land, Cropland Remaining Cropland, and Grassland Remaining
Grassland slowed over this period. Net C accumulation remained steady from 1990 to 2017 in Land Converted to
Forest Land, Wetlands Remaining Wetlands, and Land Converted to Wetlands. Emissions from Land Converted to
Cropland and Land Converted to Grassland decreased during this period. The C stock change from LULUCF is
summarized in Table 6-1.
Table 6-1: Net CO2 Flux from Land Use, Land-Use Change, and Forestry (MMT CO2 Eq.)
Land-Use Category
1990
2005
2013
2014
2015
2016
2017
Forest Land Remaining Forest Land
(671.6)
(639.4)
(616.7)
(568.8)
(645.2)
(628.9)
(621.1)
Changes in Forest Carbon Stocks3
(671.6)
(639.4)
(6f 6.7)
(568.8)
(645.2)
(628.9)
(62 f.f)
Land Converted to Forest Land
(119.1)
(120.0)
(120.5)
(120.5)
(120.6)
(120.6)
(120.6)
Changes in Forest Carbon Stocksb
(119.1)
(120.0)
(f20.5)
(f20.5)
(f 20.6)
(f 20.6)
(f 20.6)
Cropland Remaining Cropland
(40.9)
(26.5)
(11.4)
(12.0)
(6.3)
(9.9)
(10.3)
Changes in Mineral and Organic Soil
Carbon Stocks
(40.9)
(26.5)
(ff.4)
(f2.0)
(6.3)
(9.9)
(f 0.3)
Land Converted to Cropland
75.6
66.7
66.9
66.7
66.7
67.3
66.9
Changes in all Ecosystem Carbon
Stocksc
75.6
66.7
66.9
66.7
66.7
67.3
66.9
Grassland Remaining Grassland
(4.2)
5.5
(3.7)
(7.5)
9.6
(1.6)
(0.1)
Changes in Mineral and Organic Soil
Carbon Stocks
(4.2)
5.5
(3.7)
(7.5)
9.6
(f.6)
(O.f)
Land Converted to Grassland
8.7
5.1
8.3
7.9
9.8
8.5
8.3
Changes in all Ecosystem Carbon
Stocksc
8.7
5.1
8.3
7.9
9.8
8.5
8.3
Wetlands Remaining Wetlands
(4.0)
(5.7)
(4.3)
(4.3)
(4.4)
(4.4)
(4.4)
Changes in Organic Soil Carbon Stocks
in Peatlands
f.f
f.f
0.8
0.8
0.8
0.7
0.7
Changes in Aboveground and Soil
Carbon Stocks in Coastal Wetlands
(5.1)
(6.8)
(5-f)
(5-f)
(5-f)
(5-f)
(5-f)
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Changes in Aboveground and Soil
Carbon Stocksd
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Settlements Remaining Settlements
(122.1)
(127.8)
(135.9)
(135.8)
(135.4)
(134.7)
(134.5)
Changes in Organic Soil Carbon Stocks
O.f
0.5
f .3
f .3
f .3
f .3
f .3
Changes in Settlement Lree Carbon
Stocks
(96.2)
(f f 6.8)
(f 25.6)
(f 25.0)
(f 24.5)
(f23.9)
(f 23.9)
Changes in Yard Lrimmings and Food
Scrap Carbon Stocks in Landfills
(26.0)
(ff.4)
(ff.7)
(f2.f)
(f2.3)
(f2.f)
(ff.9)
Land Converted to Settlements
62.9
86.0
86.4
86.5
86.5
86.4
86.2
Changes in all Ecosystem Carbon
Stocksc
62.9
86.0
86.4
86.5
86.5
86.4
86.2
LULUCF Carbon Stock Change
(814.8)
(756.1)
(731.0)
(687.8)
(739.4)
(738.1)
(729.6)
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 N2O 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; andN^O emissions from Forest Soils and Settlement Soils.
4 Carbon sequestration estimates are net figures. Lhe 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.
6-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
+ Absolute value does not exceed 0.05 MMT CO2 Eq.
3 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.
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
Emissions of CH4 from LULUCF activities are shown in Table 6-2. Forest fires were the largest source of CH4
emissions from LULUCF in 2017, totaling 4.9 MMT CO2 Eq. (194 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). Peatlands Remaining Peatlands, Land Converted to Wetlands, and
Drained Organic Soils on forest lands resulted in CH4 emissions of less than 0.05 MMT CO2 Eq. each.
For N2O emissions, forest fires were also the largest source from LULUCF in 2017, totaling 3.2 MMT CO2 Eq. (11
kt of N2O). Nitrous oxide emissions from fertilizer application to settlement soils in 2017 totaled to 2.5 MMT CO2
Eq. (8 kt of N2O). This represents an increase of 72.0 percent since 1990. Additionally, the application of synthetic
fertilizers to forest soils in 2017 resulted in N20 emissions of 0.5 MMT CO2 Eq. (2 kt of N20). Nitrous oxide
emissions from fertilizer application to forest soils have increased by 455.1 percent since 1990, but still account for
a relatively small portion of overall emissions. Grassland fires resulted in N20 emissions of 0.3 MMT 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 inN20
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.
Table 6-2: Emissions from Land Use, Land-Use Change, and Forestry by Gas (MMT CO2 Eq.)
Gas/Land-Use Sub-Category
1990
2005
2013
2014
2015
2016
2017
CH4
5.0
9.0
9.9
10.1
16.5
8.8
8.8
Forest Land Remaining Forest Land:
Forest Fires3
1.5
5.2
6.1
6.1
12.6
4.9
4.9
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
0.2
0.4
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
2.8
7.0
7.6
7.7
11.8
6.7
6.7
Forest Land Remaining Forest Land:
Forest Fires3
1.0
3.4
4.0
4.0
8.3
3.2
3.2
Settlements Remaining Settlements:
Settlement Soils'1
1.4
2.5
2.6
2.6
2.5
2.5
2.5
Forest Land Remaining Forest Land:
Forest Soils6
0.1
0.5
0.5
0.5
0.5
0.5
0.5
Grassland Remaining Grassland:
Grassland Firesb
0.1
0.3
0.2
0.4
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
Land Use, Land-Use Change, and Forestry 6-3
-------
Wetlands Remaining Wetlands:
Peatlands Remaining Peatlands
+
+
+
+
+
+
+
LULUCF Emissions
7.8
16.0
17.5
17.7
28.3
15.5
15.5
+ Does not exceed 0.05 MMT CO2 Eq.
a Estimates include emissions from fires on both Forest Land Remaining Forest Land ami Land Converted to Forest Land.
b Estimates include emissions from fires on both Grassland Remaining Grassland and Land Converted to Grassland.
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.
Note: Totals may not sum due to independent rounding.
Figure 6-1: 2017 LULUCF Chapter Greenhouse Gas Sources and Sinks (MMT CO2 Eq.)
Forest Land Remaining Forest Land
Settlements Remaining Settlements
Land Converted to Forest Land
Cropland Remaining Cropland
Wetlands Remaining Wetlands
Grassland Remaining Grassland
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-COz Emissions from Grassland Fires
N2O Emissions from Settlement Soils
Non-C02 Emissions from Coastal Wetlands Remaining Coastal Wetlands
Non-C02 Emissions from Forest Fires
Land Converted to Grassland
Land Converted to Cropland
Land Converted to Settlements
(300) (250) (200) (150) (100) (50) 0 50 100
MMT CO2 Eq.
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
2013
2014
2015
2016
2017
Forest Land Remaining Forest Land
(669.0)
(630.2)
(605.9)
(558.1)
(623.8)
(620.3)
(612.5)
Changes in Forest Carbon Stocks3
(671.6)
(639.4)
(616.7)
(568.8)
(645.2)
(628.9)
(621.1)
N011-CO2 Emissions from Forest Firesb
2.4
8.6
10.2
10.1
20.8
8.0
8.0
N2O Emissions from Forest Soilsc
0.1
0.5
0.5
0.5
0.5
0.5
0.5
N011-CO2 Emissions from Drained
Organic Soils'1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Land Converted to Forest Land
(119.1)
(120.0)
(120.5)
(120.5)
(120.6)
(120.6)
(120.6)
Changes in Forest Carbon Stocks6
(119.1)
(120.0)
(120.5)
(120.5)
(120.6)
(120.6)
(120.6)
Cropland Remaining Cropland
(40.9)
(26.5)
(11.4)
(12.0)
(6.3)
(9.9)
(10.3)
Changes in Mineral and Organic Soil
Carbon Stocks
(40.9)
(26.5)
(11.4)
(12.0)
(6.3)
(9.9)
(10.3)
Land Converted to Cropland
75.6
66.7
66.9
66.7
66.7
67.3
66.9
Changes in all Ecosystem Carbon Stocksf
75.6
66.7
66.9
66.7
66.7
67.3
66.9
Grassland Remaining Grassland
(4.1)
6.2
(3.3)
(6.7)
10.2
(1.0)
0.6
Changes in Mineral and Organic Soil
Carbon Stocks
(4.2)
5.5
(3.7)
(7.5)
9.6
(1.6)
(0.1)
N011-CO2 Emissions from Grassland
Firesg
0.2
0.7
0.4
0.8
0.7
0.6
0.6
(621.1)
I Non-C02 Emissions
Carbon Stock Change
l< 0.51
l< 0.51
l< 0.51
l< 0.5|
l< 0.5|
l< 0.51
k
6-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Land Converted to Grassland
8.7
5.1
8.3
7.9
9.8
8.5
8.3
Changes in all Ecosystem Carbon Stocksf
8.7
5.1
8.3
7.9
9.8
8.5
8.3
Wetlands Remaining Wetlands
(0.5)
(2.0)
(0.6)
(0.6)
(0.6)
(0.7)
(0.7)
Changes in Organic Soil Carbon Stocks
in Peatlands
1.1
1.1
0.8
0.8
0.8
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
N2O Emissions from Coastal Wetlands
Remaining Coastal Wetlands
0.1
0.2
0.1
0.1
0.1
0.1
0.1
N011-CO2 Emissions from Peatlands
Remaining Peatlands
+
+
+
+
+
+
+
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Changes in Aboveground and Soil
Carbon Stocks
(+)
(+)
(+)
(+)
(+)
(+)
(+)
CH4 Emissions from Land Converted to
Coastal Wetlands
+
+
+
+
+
+
+
Settlements Remaining Settlements
(120.7)
(125.3)
(133.3)
(133.2)
(132.9)
(132.2)
(132.1)
Changes in Organic Soil Carbon Stocks
0.1
0.5
1.3
1.3
1.3
1.3
1.3
Changes in Settlement Lree Carbon
Stocks
(96.2)
(116.8)
(125.6)
(125.0)
(124.5)
(123.9)
(123.9)
Changes in Yard Primming and Food
Scrap Carbon Stocks in Landfills
(26.0)
(11.4)
(11.7)
(12.1)
(12.3)
(12.1)
(11.9)
N2O Emissions from Settlement Soils'1
1.4
2.5
2.6
2.6
2.5
2.5
2.5
Land Converted to Settlements
62.9
86.0
86.4
86.5
86.5
86.4
86.2
Changes in all Ecosystem Carbon Stocksf
62.9
86.0
86.4
86.5
86.5
86.4
86.2
LULUCF Emissions'
7.8
16.0
17.5
17.7
28.3
15.5
15.5
LULUCF Carbon Stock Change"
(814.8)
(756.1)
(731.0)
(687.8)
(739.4)
(738.1)
(729.6)
LULUCF Sector Net Totalk
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
+ Absolute value does not exceed 0.05 MMT CO2 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 abovegroimd/belowgroimd 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 Grassland.
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.
1LULUCF emissions include the CH4 and N2O emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions fromLawrf Converted
to Coastal Wetlands; andN^O emissions from Forest Soils and Settlement Soils.
J LULUCF Carbon Stock Change includes any C stock gains and losses from all land use and land use conversion categories.
k Hie LULUCF Sector Net Lotal is the net sum of all CH4 and N2O emissions to the atmosphere plus net carbon stock
changes.
Notes: Lotals may not sum due to independent rounding. Parentheses indicate net sequestration.
Table 6-4: Emissions and Removals from Land Use, Land-Use Change, and Forestry (MMT
COz Eq.)
Gas/Land-Use Category
1990
2005
2013
2014
2015
2016
2017
Carbon Stock Change3
(814.8)
(756.1)
(731.0)
(687.8)
(739.4)
(738.1)
(729.6)
Forest Land Remaining Forest Land
(671.6)
(639.4)
(616.7)
(568.8)
(645.2)
(628.9)
(621.1)
Land Converted to Forest Land
(119.1)
(120.0)
(120.5)
(120.5)
(120.6)
(120.6)
(120.6)
Cropland Remaining Cropland
(40.9)
(26.5)
(11.4)
(12.0)
(6.3)
(9.9)
(10.3)
Land Converted to Cropland
75.6
66.7
66.9
66.7
66.7
67.3
66.9
Land Use, Land-Use Change, and Forestry 6-5
-------
Grassland Remaining Grassland
(4.2)
5.5
(3.7)
(7.5)
9.6
(1.6)
(0.1)
Land Converted to Grassland
8.7
5.1
8.3
7.9
9.8
8.5
8.3
Wetlands Remaining Wetlands
(4.0)
(5.7)
(4.3)
(4.3)
(4.4)
(4.4)
(4.4)
Land Converted to Wetlands
(+)
(+)
(+)
(+)
(+)
(+)
(+)
Settlements Remaining Settlements
(122.1)
(127.8)
(135.9)
(135.8)
(135.4)
(134.7)
(134.5)
Land Converted to Settlements
62.9
86.0
86.4
86.5
86.5
86.4
86.2
CH4
5.0
9.0
9.9
10.1
16.5
8.8
8.8
Forest Land Remaining Forest Land:
Forest Firesb
1.5
5.2
6.1
6.1
12.6
4.9
4.9
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
0.2
0.4
0.3
0.3
0.3
Land Converted to Wetlands: Land
Converted to Coastal Wetlands
+
+
+
+
+
+
+
Forest Land Remaining Forest Land:
Drained Organic Soils'1
+
+
+
+
+
+
+
Wetlands Remaining Wetlands:
Peatlands Remaining Peatlands
+
+
+
+
+
+
+
N2O
2.8
7.0
7.6
7.7
11.8
6.7
6.7
Forest Land Remaining Forest Land:
Forest Firesb
1.0
3.4
4.0
4.0
8.3
3.2
3.2
Settlements Remaining Settlements:
Settlement Soils6
1.4
2.5
2.6
2.6
2.5
2.5
2.5
Forest Land Remaining Forest Land:
Forest Soilsf
0.1
0.5
0.5
0.5
0.5
0.5
0.5
Grassland Remaining Grassland:
Grassland Firesc
0.1
0.3
0.2
0.4
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 Soils'1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Wetlands Remaining Wetlands:
Peatlands Remaining Peatlands
+
+
+
+
+
+
+
LULUCF Emissions8
7.8
16.0
17.5
17.7
28.3
15.5
15.5
LULUCF Carbon Stock Change3
(814.8)
(756.1)
(731.0)
(687.8)
(739.4)
(738.1)
(729.6)
LULUCF Sector Net Total"
(807.0)
(740.0)
(713.5)
(670.0)
(711.1)
(722.6)
(714.1)
+ Absolute value does not exceed 0.05 MMT CO2 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 ami 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.
g LULUCF emissions include the CH4 and N2O emissions reported for Peatlands Remaining Peatlands, Forest Fires, Drained
Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CFL emissions from Land Converted to
Coastal Wetlands; andN^O emissions from Forest Soils and Settlement Soils.
h Hie LULUCF Sector Net Lotal is the net sum of all CFL and N2O emissions to the atmosphere plus net carbon stock
changes.
Notes: Lotals may not sum due to independent rounding. Parentheses indicate net sequestration.
Table 6-5: Emissions and Removals from Land Use, Land-Use Change, and Forestry (kt)
Gas/Land-Use Category
1990
2005
2013
2014
2015
2016
2017
Carbon Stock Change (CC>2)a
(814,784)
(756,056)
(730,952)
(687,769)
(739,378)
(738,074)
(729,563)
6-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Forest Land Remaining Forest
Land (671,583)
Land Converted to Forest Land (119,073)
Cropland Remaining Cropland (40,940)
Land Converted to Cropland 75,580
Grassland Remaining Grassland (4,214)
Land Converted to Grassland 8,738
Wetlands Remaining Wetlands (4,050)
Land Converted to Wetlands (44)
Settlements Remaining
Settlements (122,119)
Land Converted to Settlements 62,921
CH4 199
Forest Land Remaining Forest
Land: Forest Firesb 59
Wetlands Remaining Wetlands:
Coastal Wetlands Remaining
Coastal Wetlands 137
Grassland Remaining Grassland:
Grassland Firesc 3
Land Converted to Wetlands:
Land Converted to Coastal
Wetlands 1
Forest Land Remaining Forest
Land: Drained Organic Soils'1 1
Wetlands Remaining Wetlands:
Peatlands Remaining Peatlands +
N2O 9
Forest Land Remaining Forest
Land: Forest Firesb 3
Settlements Remaining
Settlements: Settlement Soils6 5
Forest Land Remaining Forest
Land: Forest Soilsf +
Grassland Remaining Grassland:
Grassland Firesc +
Wetlands Remaining Wetlands:
Coastal Wetlands Remaining
Coastal Wetlands +
Forest Land Remaining Forest
Land: Drained Organic Soils'1 +
Wetlands Remaining Wetlands:
Peatlands Remaining Peatlands +
(639,396)
(119,951)
(26,544)
66,657
5,492
5,124
(5,689)
(32)
(127,755)
86,038
362
208
140
13
23
11
8
2
1
(616,684)
(120,451)
(11,367)
66,945
(3,745)
8,269
(4,325)
(44)
(568,768)
(120,493)
(12,018)
66,750
(7,549)
7,927
(4,329)
(44)
(135,916) (135,793)
86,366 86,548
397 402
245
142
243
143
16
(645,215)
(120,596)
(6,321)
66,709
9,596
9,786
(4,359)
(44)
(135,409)
86,474
659
502
143
13
(628,935)
(120,635)
(9,941)
67,314
(1,621)
8,500
(4,390)
(44)
(621,066)
(120,618)
(10,280)
66,865
(55)
8,347
(4,399)
(44)
(134,680) (134,524)
86,358 86,212
350 352
194
144
11
194
144
12
1
1
1
1
1
1
1
1
1
1
+
25
+
26
+
40
+
22
+
22
14
13
28
11
11
9
9
9
8
8
2
2
2
2
2
1
1
1
1
1
+ 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 ami 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.
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
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
Land Use, Land-Use Change, and Forestry 6-7
-------
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 2006IPCC
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 this Inventory do 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.
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 U.S.
land area into the thirty-six IPCC land-use and land-use change categories (Table 6-7) (IPCC 2006). The primary
databases are the U.S. Department of Agriculture (USDA) National Resources Inventory (NRI),6 the USDA Forest
Service (USFS) Forest Inventory and Analysis (FIA)7 Database, and the Multi-Resolution Land Characteristics
Consortium (MRLC) National Land Cover Dataset (NLCD).8 For this Inventory, only new FIA data were used to
update the time series of land use data in the conterminous United States and Hawaii (i.e., FIA data were not used to
update Alaska). A recompilation of activity data of the other land uses and Alaska will occur for the next (i.e., 1990
through 2018) Inventory when new NRI and NLCD data are available.
5 See.
6 NRI data are available at .
7 FIA data are available at .
8 NLCD data are available at and MRLC is a consortium of several U.S. government agencies.
6-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
The total land area included in the U.S. Inventory is 936 million hectares across the 50 states.9 Approximately 890
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 2017, the United States had a total of 279 million
hectares of managed Forest Land (0.4 percent increase compared to 1990). For Cropland, 163 million hectares are
estimated (6.5 percent decrease compared to 1990), 339 million hectares of managed Grassland (0 percent change
compared to 1990), 43 million hectares of managed Wetlands (1.7 percent decrease compared to 1990), 43 million
hectares of Settlements (30 percent increase compared to 1990), and 23 million hectares of managed Other Land
(3.8 percent compared to 1990) (Table 6-7). Wetlands are not differentiated between managed and unmanaged, and
are reported solely 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
2013a
2014a
2015a
2016a
2017a
Managed Lands
889,923
889,913
889,895
889,895
889,895
889,895
889,895
Forest Land
277,653
276,728
278,978
279,072
279,036
278,949
278,889
Croplands
174,427
165,600
163,056
163,056
163,064
163,065
163,065
Grasslands
338,955
341,233
338,881
338,818
338,875
338,970
339,042
Settlements
33,361
40,429
43,308
43,291
43,271
43,270
43,270
Wetlands
45,583
43,338
42,908
42,893
42,874
42,861
42,849
Other Land
21,945
22,585
22,764
22,765
22,775
22,780
22,780
Unmanaged Lands
46,272
46,282
46,300
46,300
46,300
46,300
46,300
Forest Land
9,515
8,474
8,601
8,601
8,601
8,601
8,601
Croplands
0
0
0
0
0
0
0
Grasslands
25,953
27,043
26,936
26,936
26,936
26,936
26,936
Settlements
0
0
0
0
0
0
0
9 Hie current land representation does not include areas from U.S. Territories, but there are planned improvements to include
these regions in future Inventories.
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 United States 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 are reported as managed. See the Planned
Improvements section of the Inventory for future refinements to the Wetland area estimates.
11 Other discrepancies between the land use areas in this section and subsequent sections in the LULUCF chapter are primarily
due to new activity data that were compiled for Forest Land Remaining Forest Land and Land Converted Forest Land for this
Inventory. These updates led to changes in the land representation data for other land uses through the process of combining FIA
data with NRI and NLCD (See section "Approach for Combining Data Sources"). However, an inventory was not compiled for
cropland, grassland and settlements in this Inventory, and so the estimates for those land uses are based on the land representation
data from the previous Inventory. Also, newly compiled data for Forest Land Remaining Forest Land in Alaska were not
harmonized with the land representation data in this section, leading to discrepancies with the areas presented for forest land in
this section and the later section with the forest land carbon stock data. In addition, 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. These discrepancies will be rectified in the next (1990 through 2018) Inventory.
12 These "managed area" discrepancies also occur in the Common Reporting Format (CRF) tables submitted to the UNFCCC.
Land Use, Land-Use Change, and Forestry 6-9
-------
Wetlands
0
0
0
0
0
0
0
Other Land
10,804
10,765
10,764
10,764
10,764
10,764
10,764
Total Land Areas
936,195
936,195
936,195
936,195
936,195
936,195
936,195
Forest Land
287,167
285,202
287,578
287,673
287,637
287,549
287,490
Croplands
174,427
165,600
163,056
163,056
163,064
163,065
163,065
Grasslands
364,908
368,276
365,817
365,754
365,810
365,906
365,978
Settlements
33,361
40,429
43,308
43,291
43,271
43,270
43,270
Wetlands
43,583
43,338
42,908
42,893
42,874
42,861
42,849
Other Land
32,749
33,350
33,528
33,529
33,539
33,544
33,544
Hie land use data for 2013 to 2017 were only partially updated based
on new Forest Inventory and Analysis (FIA) dat
addition, there were no new data incorporated for Alaska. New activity data for the National Resources Inventory (NRI) and
National Land Cover Dataset (NLCD) will be incorporated in the next Inventory.
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
2013"
2014"
2015"
2016"
2017"
Total Forest Land
277,653
276,728
278,978
279,072
279,036
278,949
278,889
FF
276,298
275,267
277,444
277,575
277,736
277,674
277,615
CF
205
180
164
162
150
149
149
GF
1,066
1,105
1,172
1,144
973
953
953
WF
15
47
49
48
47
47
47
SF
10
11
16
15
15
14
14
OF
58
117
133
128
117
111
111
Total Cropland
174,427
165,600
163,056
163,056
163,064
163,065
163,065
CC
162,058
150,596
149,723
149,725
149,737
149,737
149,737
FC
197
83
75
73
69
69
69
GC
11,754
14,418
12,827
12,827
12,827
12,827
12,827
WC
150
176
128
128
128
128
128
SC
76
85
91
91
91
91
91
oc
192
243
213
213
213
213
213
Total Grassland
338,955
341,233
338,881
338,818
338,875
338,970
339,042
GG
329,268
319,686
317,739
317,684
317,750
317,850
317,922
FG
693
3,210
3,225
3,218
3,208
3,204
3,204
CG
8,309
16,825
16,555
16,555
16,555
16,555
16,555
WG
231
429
199
199
199
199
199
SG
53
106
114
114
114
114
114
OG
400
976
1,048
1,048
1,048
1,048
1,048
Total Wetlands
43,583
43,338
42,908
42,893
42,874
42,861
42,849
WW
42,824
41,945
41,691
41,677
41,661
41,648
41,636
FW
47
70
72
70
68
68
68
CW
214
378
346
346
346
346
346
GW
452
835
700
700
700
700
700
SW
5
0
1
1
1
1
1
OW
41
110
98
98
98
98
98
Total Settlements
33,361
40,429
43,308
43,291
43,271
43,270
43,270
SS
30,471
31,981
35,849
35,850
35,850
35,851
35,851
FS
330
572
607
589
569
568
568
CS
1,247
3,550
2,982
2,982
2,982
2,982
2,982
GS
1,250
4,102
3,653
3,653
3,653
3,653
3,653
WS
6
25
26
26
26
26
26
OS
58
199
190
190
190
190
190
Total Other Land
21,945
22,585
22,764
22,765
22,775
22,780
22,780
OO
21,026
20,737
20,771
20,776
20,787
20,793
20,793
FO
51
77
90
86
85
84
84
CO
300
613
679
679
679
679
679
GO
481
982
1,109
1,109
1,109
1,109
1,109
WO
82
168
102
102
102
102
102
6-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
so
Grand Total
5
889,923
9
889,913
13 13 13 13 13
889,895 889,895 889,895 889,895 889,895
a The abbreviations are "F" for Forest Land, "C" for Cropland, "G" for Grassland, "W" for Wetlands, "S" for
Settlements, and "O" for Other Lands. Lands remaining in the same land-use category are identified with the
land-use abbreviation given twice (e.g., "FF" is, Forest Land Remaining Forest Land), and land-use change
categories are identified with the previous land use abbreviation followed by the new land-use abbreviation (e.g.,
"CF" is Cropland Converted to Forest Land).
b The land use data for 2013 to 2017 were only partially updated based on new Forest Inventory and Analysis
(FIA) data. In addition, there were no new data incorporated for Alaska. New activity data for the National
Resources Inventory (NRI) and National Land Cover Dataset (NLCD will be incorporated for the next (i.e., 1990
through 2018) Inventory.
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.
Land Use, Land-Use Change, and Forestry 6-11
-------
Figure 6-2: Percent of Total Land Area for Each State in the General Land-Use Categories for
2017
Croplands
Forest Lands
< 10
< 10
Grasslands
Other Lands
31 50
V
\
uTv
k. s m
\
7 ( *\ %
r
y
< 10
] 10- 30
I 30 - SO
I > SO
Wetlands
Settlements
J * 10
10-30
| 30 - SO
| > 50
:c !()
¦ SB
6-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Methodology
IPCC Approaches for Representing Land Areas
IPCC (2006) describes three approaches for representing land areas. Approach 1 provides data on the total area for
each individual land-use category, but does not provide detailed information on changes of area between categories
and is not spatially explicit other than at the national or regional level. With Approach 1, total net conversions
between categories can be detected, but not the individual changes (i.e., additions and/or losses) between the land-
use categories that led to those net changes. Approach 2 introduces tracking of individual land-use changes between
the categories (e.g., Forest Land to Cropland, Cropland to Forest Land, and Grassland to Cropland), using survey
samples or other forms of data, but does not provide spatially-explicit location data. Approach 3 extends Approach 2
by providing spatially-explicit location data, such as surveys with spatially identified sample locations and maps
derived from remote sensing products. The three approaches are not presented as hierarchical tiers and are not
mutually exclusive.
According to IPCC (2006), the approach or mix of approaches selected by an inventory agency should reflect
calculation needs and national circumstances. For this analysis, the NRI, FIA, and the NLCD have been combined to
provide a complete representation of land use for managed lands. These data sources are described in more detail
later in this section. NRI, FIA and NLCD are Approach 3 data sources that provide spatially-explicit representations
of land use and land-use conversions. Lands are treated as remaining in the same category (e.g., Cropland
Remaining Cropland) if a land-use change has not occurred in the last 20 years. Otherwise, the land is classified in a
land-use change category based on the current use and most recent use before conversion to the current use (e.g.,
Cropland Converted to Forest Land).
Definitions of Land Use in the United States
Managed and Unmanaged Land
The United States definition of managed land is similar to the general definition of managed land provided by the
IPCC (2006), but with some additional elaboration to reflect national circumstances. Based on the following
definitions, most lands in the United States are classified as managed:
• Managed Land: Land is considered managed if direct human intervention has influenced its condition.
Direct intervention occurs mostly in areas accessible to human activity and includes altering or maintaining
the condition of the land to produce commercial or non-commercial products or services; to serve as
transportation corridors or locations for buildings, landfills, or other developed areas for commercial or
non-commercial purposes; to extract resources or facilitate acquisition of resources; or to provide social
functions for personal, community, or societal objectives where these areas are readily accessible to
society.13
• Unmanaged Land: All other land is considered unmanaged. Unmanaged land is largely comprised of areas
inaccessible to society due to the remoteness of the locations. Though these lands may be influenced
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
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, all Wetlands are reported as managed, but emissions 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.
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.
Land Use, Land-Use Change, and Forestry 6-13
-------
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.16 The 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
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 with
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
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.
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.
6-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
criteria for Grassland. Woody plant communities of low forbs and shrubs, such as mesquite, chaparral,
mountain shrub, 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 tracts of less than 10 acres (4.05 ha) that may
meet the definitions for Forest Land, 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-CCh 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 > P^Tlptlon arid Application to U.S.
Land Area Classification
U.S. Land-Use Data Sources
The three main sources for land-use data in the United States are the NRI, FIA, and the NLCD (Table 6-8). 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 the surveys contain additional information on management, site conditions, crop types, biometric
measurements, and other data that is needed to estimate C stock changes, N20, and CH4 emissions on those lands. If
NRI and FIA data are not available for an area, however, then the NLCD product is used to represent the land use.
Land Use, Land-Use Change, and Forestry 6-15
-------
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
Alaska
Non-Federal
Federal
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 (except Forest Land), 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 USD A 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 2012 from the NRI. The land use patterns are assumed to remain
the same from 2012 through 2017 for this Inventory, but the time series will be updated when new data are released.
Forest Inventory and Analysis
The FIA program, conducted by the USFS, is another statistically-based survey for the conterminous United States
in addition to the southeast and south central coastal Alaska, and the official source of data on Forest Land area and
management data for the Inventory. 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-
6-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 westernUnited 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 2012 through
2017; see Table A-219).
National Land Cover Dataset
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 a portion of 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), USD A, 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.
NLCD products provide land-cover for 1992, 2001, 2006, and 2011 in the conterminous United States (Homer et al.
2007), and also for Alaska in 2001 and 2011 and Hawaii in 2001. For the conterminous United States, the NLCD
data have been further processed to derive Land Cover Change Products for 2001, 2006, and 2011 (Fry et al. 2011;
Homer et al. 2007; Homer et al. 2015). 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 contain
21 categories of land-cover information, which 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 and 2011 for
the conterminous United States and Alaska, but the time series will be updated when new data are released.
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, as well as federal and non-federal
lands in Alaska. 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.
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 or Other Lands
may be designated as unmanaged land;23
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.
23 All wetlands are considered managed in this Inventory. 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. Regardless, a planned improvement is underway to subdivide managed and unmanaged
Wetlands.
Land Use, Land-Use Change, and Forestry 6-17
-------
• All Forest Lands with active fire protection 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 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 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. The designation of
grasslands as managed is based on grazing livestock population data at the county scale from the USD A 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 approximately 130 petroleum extraction sites and 223 mines. 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. The remaining land represents the unmanaged land base. The resulting spatial product is used to identify NRI
survey locations that are considered managed and unmanaged for the conterminous United States and Hawaii,24 in
addition to determining which areas in the NLCD for Alaska are included in the managed land base.
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 using definitions developed to meet national circumstances, while adhering to IPCC (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
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.
6-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 do not
provide specific land-use categories that are converted to Forest Land, but rather a sum of all Land Converted to
Forest Land. 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 does not provide information
on the specific land-use changes, 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 coastal 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. However, there is
insufficient data at this time to address land use change so forest land in this region is based on the 2001
and 2011 NLCD rather than the FIA. NRI is used in the current report to provide Forest Land areas on non-
Land Use, Land-Use Change, and Forestry 6-19
-------
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 fluxes 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 lands 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.26
• 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 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
26 This analysis does not distinguish between managed and unmanaged wetlands, which is a planned improvement for the
Inventory.
6-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 from these areas are 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 Topologically 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 lias 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 reporting 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 lias 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 percent
difference when open water in coastal regions is removed from the TIGER data.
Recalculations Discussion
The land representation data in the current Inventory were recalculated from the previous Inventory by using
updated FIA data for 1990 to 2017. These data were used as the basis for the forest areas and harmonized with the
other databases as described in the section "Approach for Combining Data Sources". This process also leads to
changes in the areas of other land uses to ensure the total land base area remains the same. Forest land declined by
an average of 4 percent across the time series from 1990 to 2016 based on the new FIA data. Based on the
harmonization. Grassland, Other Land and Settlements increased by an average of 3.6 percent, 0.1 percent, and 0.1
percent, respectively. Wetlands decreased by an average of 0.1 percent and Croplands did not change. New data for
Alaska were not used this year and will be applied during the next Inventory period along with new NRI and NLCD
data.
Planned Improvements
The next (i.e., 1990 through 2018) Inventory will be substantially improved by using new data sets to update the
time series for land representation with the latest NRI and NLCD data sets and ensure consistency between the total
area of managed land in the land-representation description and the remainder of the Inventory. Coastal wetland
areas will also be harmonized with a NOAA data set on coastal wetland land use and land use transitions, as
described in more detail below.
Another key planned improvement for the Inventory is to fully incorporate area data by land-use type for U.S.
Territories. Fortunately, most of the managed land in the United States is included in the current land-use data, but a
complete reporting of all lands in the United States is a key goal for the near future. Preliminary land-use area data
for U.S. Territories by land-use category are provided in Box 6-2.
Box 6-2: Preliminary Estimates of Land Use in U.S. Territories
Several programs have developed land cover maps for U.S. Territories using remote sensing imagery, including the
Gap Analysis Program, Caribbean Land Cover project. National Land Cover Dataset, USFS Pacific Islands Imagery
Project, and the National Oceanic and Atmospheric Administration (NOAA) Coastal Change Analysis Program (C-
CAP). Land-cover data can be used to inform a land-use classification if there is a time series to evaluate the
dominate practices. For example, land that is principally used for timber production with tree cover over most of the
time series is classified as forest land even if there are a few years of grass dominance following timber harvest.
These products were reviewed and evaluated for use in the national Inventory as a step towards implementing a
planned improvement to include U.S. Territories in the land representation for the Inventory. Recommendations are
Land Use, Land-Use Change, and Forestry 6-21
-------
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 a land-cover product 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
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 not used directly in the land representation analysis, however, so a planned
improvement for the next (i.e., 1990 through 2018) Inventory is to reconcile the coastal wetlands data from the C-
CAP product with the wetlands area data provided in the NRI. In addition, the current Inventory does not include a
classification of managed and unmanaged wetlands. However, implementation of the new guidance will require
classification of managed and unmanaged wetlands in the Inventory, and more detailed wetlands datasets will be
evaluated and integrated into the analysis to meet this objective.
NOAA C-CAP data for Hawaii were recently released for 2011, and will be used to analyze land use change for this
state in the near future. 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)
C^ Potest 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.
6-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• 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 photo synthesize
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, CH4, N20, CO, and NOx when 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 harvested timber 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 USD A 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, next the 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. The focus on C
implies that all C-based greenhouse gases are included, and the focus on stock change suggests that specific
ecosystem fluxes do not need to be separately itemized in this report. 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 fire implicitly includes only the C stocks remaining on the NFI plot.
However, changes in C stocks from natural disturbances are highly variable from year to year. 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 that have been classified as forest lands
for less than 20 years. The methods and data used to delineate forest C stock changes by these two categories
Land Use, Land-Use Change, and Forestry 6-23
-------
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 33 percent of the U.S. land area is estimated to be forested based on the U.S. definition of forest land
as provided in the Section 6.1 Representation of the U.S. Land Base. All annual NFI plots included in the public FIA
database as of May 2018 (which includes data through 2017) were used in this Inventory. Since area estimates for
some land use categories were not updated in the Land Representation in the current Inventory there are differences
in the area estimates reported in this section and those reported in Section 6.1 Representation of the U.S. Land Base.
The NFIs from each of the conterminous 48 states (CONUS; USDA Forest Service 2018a, 2018b) and Alaska
comprise an estimated 272 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 2017 for all states (USDA Forest Service 2018b). Sufficient annual
inventory data are not yet available for Hawaii 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 added to the forest C estimates as sufficient data become
available. 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)27 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). Current trends in the estimated forest land area in the
CONUS and Alaska represented here show an average annual rate of increase of 0.02 percent. In addition to the
increase in forest area, the major influences to the net C flux from forest land across the 1990 to 2017 time series are
management activities, natural disturbance, and the ongoing impacts of 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.28 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 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.
27 The Natural Resources Inventory of the USDA Natural Resources Conservation Service is described in Section 6.1
Representation of the U.S. Land Base.
28 The term "biomass density" refers to the mass of live vegetation per unit area. It is usually measured on a dry-weight basis.
Dry biomass is assumed to be 50 percent C by weight.
6-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 6-3: Changes in Forest Area by Region for Forest Land Remaining Forest Land in the
conterminous United States and Alaska (1990-2017, Million Hectares)
100n
t/3
North
Pacific
Coast
t Rocky
' Mountain
| 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
Year
2010
2015
Rocky
Mountain
Pacific
North
South
Forest Carbon Stocks and Stock Change
In the United States, 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 2017. 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 621.1
MMT CO: Eq (169.4 MMT C) in 2017 (Table 6-10 and Table 6-11). The estimated net uptake of C in the Forest
Ecosystem was 517.8 MMT CO2 Eq. (141.2 MMT C) in 2017 (Table 6-10 and Table 6-11). The majority of this
uptake, 357.1 MMT CO2 Eq. (97.4 MMT C), was from aboveground biomass in 2017. Overall, estimates of average
Land Use, Land-Use Change, and Forestry 6-25
-------
C density in forest ecosystems (including all pools) remained stable at approximately 205 MT C ha-1 from 1990 to
2017. 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 2017. The
stable forest ecosystem C density when combined with increasing forest area results in net C accumulation over
time. 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. Aboveground live biomass is
responsible for the majority of net C uptake among all forest ecosystem pools (Figure 6-4).
The estimated net uptake of C in HWP was 103.3 MMT CO2 Eq. (28.2 MMT C) in 2017 (Table 6-10 and Table
6-11). The majority of this uptake, 67.6 MMT CO2 Eq. (18.4 MMT C), was from wood and paper in SWDS.
Products in use were an estimated 35.7 MMT CO2 Eq. (9.7 MMT C) in 2017.
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
2013
2014
2015
2016
2017
Forest Ecosystem
(547.8)
(531.4)
(541.1)
(492.4)
(549.4)
(529.3)
(517.8)
Aboveground Biomass
(378.7)
(361.2)
(379.5)
(361.8)
(377.5)
(371.3)
(357.1)
Belowground Biomass
(90.7)
(86.1)
(89.2)
(84.4)
(88.6)
(87.1)
(83.9)
Dead Wood
(76.0)
(78.9)
(79.4)
(74.5)
(82.6)
(81.9)
(77.4)
Litter
(4.2)
(5.1)
(0.9)
30.0
(3.3)
(1.2)
(3.8)
Soil (Mineral)
1.2
(0.6)
6.8
(1.6)
0.5
9.2
2.3
Soil (Organic)
(0.1)
(0.1)
0.3
(1.0)
1.3
2.2
1.3
Drained Organic Soil3
0.8
0.8
0.8
0.8
0.8
0.8
0.8
Harvested Wood
(123.8)
(108.0)
(75.6)
(76.4)
(95.9)
(99.6)
(103.3)
Products in Use
(54.8)
(44.6)
(13.0)
(13.7)
(31.4)
(33.5)
(35.7)
SWDS
(69.0)
(63.5)
(62.6)
(62.7)
(64.4)
(66.1)
(67.6)
Total Net Flux
(671.6)
(639.4)
(616.7)
(568.8)
(645.2)
(628.9)
(621.1)
3 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 CO2, CH4, and N2O Emissions from
Drained Organic Soils for the methodology used to estimate the CO2 emissions from drained organic soils.
Also, see Table 6-22 and Table 6-23 for 11011-CO2 emissions from drainage of organic soils from both Forest
Land Remaining Forest Land and Land Converted to Forest Land.
Notes: Forest ecosystem C stocks do not include forest stocks in U.S. Territories, Hawaii, or 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
2013
2014
2015
2016
2017
Forest Ecosystem
(149.4)
(144.9)
(147.6)
(134.3)
(149.8)
(144.4)
(141.2)
Aboveground Biomass
(103.3)
(98.5)
(103.5)
(98.7)
(102.9)
(101.3)
(97.4)
Belowground Biomass
(24.7)
(23.5)
(24.3)
(23.0)
(24.2)
(23.8)
(22.9)
Dead Wood
(20.7)
(21.5)
(21.7)
(20.3)
(22.5)
(22.3)
(21.1)
Litter
(1.1)
(1.4)
(0.2)
8.2
(0.9)
(0.3)
(1.0)
Soil (Mineral)
0.3
(0.2)
1.9
(0.4)
0.1
2.5
0.6
Soil (Organic)
+
(+)
0.1
(0.3)
0.4
0.6
0.4
Drained Organic Soil3
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Harvested Wood
(33.8)
(29.5)
(20.6)
(20.8)
(26.1)
(27.2)
(28.2)
Products in Use
(14.9)
(12.2)
(3.5)
(3.7)
(8.6)
(9.1)
(9.7)
6-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
SWDS
(18.8)
(17.3)
(17.1)
(17.1)
(17.6)
(18.0)
(18.4)
Total Net Flux
(183.2)
(174.4)
(168.2)
(155.1)
(176.0)
(171.5)
(169.4)
+ Absolute value does not exceed 0.05 MMT C
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 CO2, CH4, and N2O 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 11011-CO2 gases changes from drainage of organic
soils from Forest Land Remaining Forest Land and Land Converted to Forest Land.
Notes: Forest C stocks do not include forest stocks in U.S. Territories, Hawaii, or 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.
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 fox 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 2017 and woodland areas previously included as forest land have
been separated and included in the Grassland categories in this Inventory.
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
2013
2014
2015
2016
2017
2018
Forest Area (1,000 ha)
269,959
271,883
273,035
273,170
273,346
273,494
273,623
273,791
Carbon Pools (MMT C)
Forest Ecosystem
53,670
55,806
56,969
57,117
57,251
57,401
57,546
57,687
Aboveground Biomass
11,870
13,357
14,160
14,263
14,362
14,465
14,566
14,664
Belowground Biomass
2,378
2,734
2,924
2,949
2,972
2,996
3,020
3,042
Dead Wood
2,153
2,463
2,636
2,658
2,678
2,700
2,723
2,744
Litter
3,663
3,646
3,645
3,645
3,637
3,638
3,638
3,639
Soil (Mineral)
27,824
27,822
27,821
27,819
27,820
27,820
27,817
27,816
Soil (Organic)
5,783
5,784
5,783
5,782
5,783
5,782
5,782
5,781
Harvested Wood
1,895
2,353
2,517
2,538
2,559
2,585
2,612
2,640
Products in Use
1,249
1,447
1,476
1,479
1,483
1,492
1,501
1,510
SWDS
646
906
1,042
1,059
1,076
1,093
1,111
1,130
Total C Stock
55,565
58,159
59,486
59,655
59,810
59,986
60,158
60,328
Notes: Forest area andC stock estimates include all Forest Land Remaining Forest Land in the conterminous 48 states and Alaska
(million ha). Forest C stocks do not include forest stocks in U.S. Territories, Hawaii, or 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). The forest area estimates in this table do not match those in Section 6.1 Representation of the U.S. Land Base, which
includes all managed forest land in Hawaii. Differences also exist because forest land area estimates are based on the latest NFI
data through 2017 and woodland area previously included as forest land has been separated and included in the Grassland
categories in this Inventory. 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 2017 requires estimates of C stocks for 2017
and 2018.
Land Use, Land-Use Change, and Forestry 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 (1990-2017, MMT C per Year)
20-i
0-
| ro -20 H
-S O -40
*5
-60-
-80
.g5
^ O)
11 "10°
i t -12°
¦2 w ,140_
13 §
b ¦§ -160-
u. S
-180-
-200-
i i | i i i
1995
I | i i
2000
i I | I I I
2005
Year
2010
i i | i i
2015
All forest ecosystem pools
Aboveground biomass
• Belowground biomass
Dead wood
• Litter
Soil (mineral)
1 Soil (organic)
Harvested Wood Products (HWP)
Products in use
Solid waste disposal sites
Total net change
(forest ecosystem + HWP)
Box 6-3: CO2 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 CO: emissions from forest fires occurring in the conterminous states as well as the
portion of managed forest lands in Alaska that are captured in the current Inventory. Because it is of interest to
quantify the magnitude of CO2 emissions from fire disturbance, these separate estimates are highlighted here. Note
that these CO2 estimates are based on the same methodology as applied for the no 11-CO: 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. It is important to note that 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 (2006) default
combustion and emission factors were used in this Inventory. This is an improvement over previous Inventories
6-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
where only the IPCC (2006) defaults have been used to estimate fire emissions and resulted in substantial changes to
the estimates provided in this box in comparison to the previous Inventory. The latest information on area burned is
used to compile fire emissions for the United States. At the time this Inventory was compiled, fire data for 2017
were not available so estimates from 2016 were used. The 2017 estimates will be updated in subsequent reports as
fire data becomes available. Estimated CO2 emissions for wildfires in the conterminous 48 states and in Alaska as
well as prescribed fires in 2017 were estimated to be 64.8 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 emissions is necessary in order to associate a portion of emissions, including
estimates of CH4 and N20, with fire. See the discussion in Annex 3.13 for more details on this methodology. Note
that 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
Year
CO2 emitted from
Wildfires in the
Conterminous 48
States (MMT yr1)
CO2 emitted from
Wildfires in Alaska
(MMTyr1)
CO2 emitted from
Prescribed Fires
(MMTyr1)
Total CO2 emitted
(MMTyr1)
1990
14.5
4.98
0.3
19.58
2005
23.41
44.28
1.6
69.28
2013
58.3
11.9
11.6
81.8
2014
64.8
3.4
12.7
81.0
2015
118.9
41.5
7.2
167.6
2016
51.2
1.7
11.9
64.8
2017b
51.2
1.7
11.9
64.8
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 2017 were unavailable when these estimates were summarized; therefore 2016, the most recent
available estimate, is applied to 2017.
Methodology and Data Sources
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. The estimates in this section of the report are based on this land use information
from the NFI and they may differ with the other land use categories where area estimates reported in the Land
Representation were not updated (see Section 6.1 Representation of the U.S. Land Base). 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
Land Use, Land-Use Change, and Forestry 6-29
-------
projected from ti to 2017. 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. Forest land conditions in Alaska were observed on NFI plots from 2004 to 2017.
Plot-level data from the NFI were harmonized with auxiliary data describing climate, forest structure, disturbance,
and other site-specific conditions to develop non-parametric models to predict carbon stocks by forest ecosystem
carbon pool as well as fluxes over the entire inventory period, 1990 to 2017. 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 2017.
Harvested wood C 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 and harvested wood products is provided here. 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 1990s, 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.
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
6-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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. (201 la), 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. (201 la), 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 one 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. 201 la) 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 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 soil C is reported to
a depth of 100 cm for Forest Land Remaining Forest Land to remain consistent with past reporting in this category,
however for consistency across land-use categories mineral (e.g., cropland, grassland, settlements) soil C is reported
to a depth of 30 cm in Section 6.3 Land Converted to Forest Land. Estimates of C from organic soils in this section
(Table 6-10 and Table 6-11) include emissions from drained organic soils and the methods for all estimates can be
found in the Drained Organic Soils section below.
Land Use, Land-Use Change, and Forestry 6-31
-------
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
here and for information purposes in the Energy sector, but the non-CCh 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 that
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, 2016, In preparation).
Estimates for disposal of products reflected 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:
(1 A) 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 2 A and 2B yielded the estimate for HWP contribution under the production estimation
approach. A key assumption for estimating these variables was that products exported from the United States and
held in pools in other countries have the same half-lives for products in use, the same percentage of discarded
products going to SWDS, and the same decay rates in SWDS as they would in the United States.
Uncertainty and Time-Series Consistency
A quantitative uncertainty analysis placed bounds on current flux 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 2017 net annual change for forest C stocks was estimated to be between -922.1 and -
341.5 MMT CO2 Eq. around a central estimate of -621.1 MMT CO2 Eq. at a 95 percent confidence level. This
includes a range of -796.9 to -238.9 MMT CO2 Eq. around a central estimate of -517.8 MMT CO2 Eq. for forest
ecosystems and -122.1 to -84.5 MMT CO2 Eq. around a central estimate of -103.3 MMT CO2 Eq. for HWP.
6-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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)
Source
Gas
2017 Flux Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Flux Estimate
(MMT CO2 Eq.) (%)
Lower
Upper
Lower Upper
Bound
Bound
Bound Bound
Forest Ecosystem C Pools3
CO2
(517.8)
(796.9)
(238.9)
-53.9% 53.9%
Harvested Wood Products'5
CO2
(103.3)
(122.1)
(84.5)
-18.2% 18.2%
Total Forest
CO2
(621.1)
(922.1)
(341.5)
-45.0% 45.0%
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.
Note: Parentheses indicate negative values or net uptake.
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, 2016, In preparation). Factors to convert wood and paper to units of C are based on estimates by industry and
Forest Service published sources (see Annex 3.13). The WOODCARB II model uses estimation methods suggested
by IPCC (2006). Estimates of annual C change in solidwood and paper products in use were calibrated to meet two
independent criteria. The first criterion is that the WOODCARB II model estimate of C in houses standing in 2001
needs to match an independent estimate of C in housing based on U.S. Census and USDA Forest Service survey
data. Meeting the first criterion resulted in an estimated half-life of about 80 years for single family housing built in
the 1920s, which is confirmed by other U.S. Census data on housing. The second criterion is that the WOODCARB
II model estimate of wood and paper being discarded to SWDS needs to match EPA estimates of discards used in
the Waste sector each year over the period 1990 to 2000 (EPA 2006). These criteria help reduce uncertainty in
estimates of annual change in C in products in use in the United States and, to a lesser degree, reduce uncertainty in
estimates of annual change in C in products made from wood harvested in the United States. In addition,
WOODCARB II landfill decay rates have been validated by ensuring that estimates of CH4 emissions from landfills
based on EPA (2006) data are reasonable in comparison to CH4 estimates based on WOODCARB II landfill decay
rates.
Recalculations Discussion
The methods and data used in the current Inventory to compile estimates for forest ecosystem carbon stocks and
stock changes from 1990 through 2017 are consistent with those used in the 1990 through 2016 Inventory for the
eastern United States. In this Inventory the regional approach for carbon stock and stock change estimation in the
Land Use, Land-Use Change, and Forestry 6-33
-------
western United States was replaced by the state-level method used in the eastern United States so carbon stocks and
stock changes are now estimated consistently for the entire 1990 to 2017 time series in all states with
remeasurements in the NFI in the CONUS. This improvement in consistency also improved separation of Forest
Land Remaining Forest Land, Land Converted to Forest Land, and areas with perennial woody biomass that do not
meet the definition of forest land (i.e., woodlands) that are now included in the Grassland Remaining Grassland and
Land Converted to Grassland sections. This resolved approach resulted in a 9 percent decrease in forest area as a
result of transferring approximately 23.5 million ha of land previously included in the Forest Land Remaining
Forest Land to the Land Converted to Forest Land, Grassland Remaining Grassland, and Land Converted to
Grassland categories in 2017 and 27.5 million ha on average annually over the time series (Table 6-16). This
improvement in consistency also corrected problems with imbalanced area estimates which may have resulted in
over or underestimates in carbon stock changes in past Inventories due to different land areas used to calculate stock
differences between years, a problem which stemmed from the time when only the Forest Land Remaining Forest
Land was included in the Inventory and no transfer of carbon between land use categories was estimated. All
managed forest land in Alaska, specifically forest land from interior Alaska, was also included for the first time in
this Inventory, which added more than 24.5 million ha to the Forest Land Remaining Forest Land category (note
that estimates for land use conversion to and from forest land are not currently included for Alaska, Table 6-15).
As a result of these improvements, the estimates reported from 1990 through 2016 are not directly comparable to the
estimates in this Inventory. To illustrate changes in the current Inventory, Table 6-15 includes forest area and carbon
stock estimates for the year 2017 from the previous (1990 to 2016) Inventory and the estimates for 2017 and 2018
from the current Inventory for CONUS and southeast and southcentral coastal Alaska. The forest land area estimates
for the year 2017 decreased by more than 23 million ha from the previous Inventory (Table 6-15) and the forest land
area decreased by an average of 7.5 million ha over the 1990 to 2016 time series. In most cases this was not a loss of
forest land area but rather a reorganization of land into the Land Converted to Forest Land category and the transfer
of 23.5 million ha of land with perennial woody biomass that does not meet the definition of forest land (i.e.,
woodlands) into the Grassland Remaining Grassland and Land Converted to Grassland categories. Despite the
reorganization of substantial land area historically included in the Forest Land Remaining Forest Land (a 9 percent
reduction), there was only a 4 percent decrease in the carbon stocks for the year 2017 between the previous
Inventory and the current Inventory (Table 6-15). This is due to increases in carbon stocks from the transition to the
state-level method used for the western United States in the current Inventory as well as the relatively minor
contribution of the lands reclassified into other land use categories. However, carbon stock changes decreased by 11
percent for the year 2016 in the previous Inventory and the 2016 estimates for the current Inventory (Table 6-16). In
particular, the mineral soil carbon stock changes in the year 2016 decreased by 107 percent between the previous
Inventory and the current Inventory and this change was consistent over the time series (Table 6-16). This was due,
in large part, to the correction of the imbalanced area estimates in the western United States resulting in a substantial
decrease in the contribution of soil carbon stock changes from that region. There was also a 111 percent increase in
dead wood carbon stock changes and an 83 percent decrease in litter carbon stock changes in the year 2016 between
the previous Inventory and the current Inventory. This can be attributed to the incorporation of remeasurements from
the NFI in the western United States for the first time in the current Inventory which allowed for using the state-
level estimation approach.
While not included in the recalculations described in this section, the inclusion of 24.5 million ha of forest area from
interior Alaska contributed an additional 8,604 MMT C stocks, primarily from soil carbon, to the Forest Land
Remaining Forest Land category in 2018 and this increase was consistent with the additions from interior Alaska
over the time series (Table 6-15). The carbon stock changes in interior Alaska were driven, in large part, by
wildfires over the time series and contribute, on average over the time series, approximately -2.2 MMT C per year to
the sink.
There were no changes in the data or methods used to compile estimates of HWP from 1990 through 2016 so no
recalculations were necessary.
6-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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)
Current Estimate
for Year 2017, Current Estimate
Previous Estimate for
Year 2017, from 2018
Inventory,
CONUS+Coastal AK
from 2019
Inventory,
CONUS+Coastal
AK
Current Estimate for
Year 2018, from 2019
Inventory,
CONUS+Coastal AK
for Year 2018,
from 2019
Inventory, Interior
AK
Forest Area (1,000 ha)
272,260
249,084
249,242
24,539
Carbon Pools
Forest
51,131
48,949
49,084
8,604
Aboveground Biomass
14,182
14,090
14,184
480
Belowground Biomass
2,923
2,905
2,924
118
Dead Wood
2,570
2,558
2,578
165
Litter
2,680
2,558
2,589
1,051
Soil (Mineral)
28,422
26,280
26,280
1,536
Soil (Organic)
352
529
528
5,253
Harvested Wood
2,612
2,612
2,640
NA
Products in Use
1,501
1,501
1,510
NA
SWDS
1,111
1,111
1,130
NA
Total Stock
53,743
51,561
51,724
NA
NA - Not Applicable
Table 6-16: Recalculations of Net C Flux from Forest Ecosystem Pools in Forest Land
Remaining Forest Land and Harvested Wood Pools (MMT C)
Carbon Pool (MMT C)
Previous Estimate for
Year 2016,
from 2018 Inventory,
CONUS+Coastal AK
Current Estimate for
Year 2016,
from 2019 Inventory,
CONUS+Coastal AK
Current Estimate for
Year 2017,
from 2019 Inventory,
CONUS+Coastal AK
Current Estimate for
Year 2017,
from 2019 Inventory,
Interior AK
Forest
(155.7)
(138.9)
(134.3)
(6.9)
Aboveground Biomass
(86.0)
(98.6)
(94.2)
(3.2)
Belowground Biomass
(17.9)
(20.3)
(19.5)
(3.4)
Dead Wood
(10.7)
(22.5)
(20.8)
0.4
Litter
(4.4)
(0.7)
(0.6)
0.5
Soil (Mineral)
(36.9)
2.5
(0.3)
0.9
Soil (Organic)
+
0.6
0.8
(0.5)
Drained Organic Soila
0.2
0.2
0.2
NA
Harvested Wood
(27.2)
(27.2)
(28.2)
NA
Products in Use
(9.1)
(9.1)
(9.7)
NA
SWDS
(18.0)
(18.0)
(18.4)
NA
Total Net Flux
(182.9)
(166.1)
(162.5)
NA
NA - Not Applicable
+ Absolute value does not exceed 0.05 MMT C
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 CO2, CH4, and N2O 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-CC>2 gases changes from drainage of organic soils from Forest Land Remaining Forest Land and Land
Converted to Forest Land.
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
Land Use, Land-Use Change, and Forestry 6-35
-------
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 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, 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 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 this year to include all managed forest land in Alaska will be used in the years
ahead for Hawaii and U.S. Territories as forest C data become available (only a small number of plots from Hawaii
are currently available from the annualized sampling design). To that end, research is underway to incorporate all
NFI information (both annual and periodic data) and the dense time series of remotely sensed data in multiple
inferential frameworks for estimating greenhouse gas emissions and removals as well as change detection and
attribution across the entire reporting period and all managed forest land in the United States. Leveraging this
auxiliary information will aid not only the interior Alaska effort but the entire inventory system. In addition to fully
inventorying all managed forest land in the United States, the more intensive sampling of fine woody debris, litter,
and SOC on a subset of FIA plots continues and will substantially improve resolution of C pools (i.e., greater sample
intensity; Westfall et al. 2013) as this information becomes available (Woodall et al. 2011b). Increased sample
intensity of some C pools and using annualized sampling data as it becomes available for those states currently not
reporting are planned for future submissions. The NFI sampling frame extends beyond the forest land use category
(e.g., woodlands, which fall into the grasslands category and urban areas, which fall into the settlements category)
with inventory-relevant information for these lands which will likely become increasingly available in coming years.
Non-C02 Emissions from Forest Fires
Emissions of non-C02 gases from forest fires were estimated using U.S.-specific data for annual area of forest
burned, potential fuel availability, and fire severity as well as the default IPCC (2006) emission factors and some
combustion factors applied to the IPCC methodology. In 2017, emissions from this source were estimated to be 4.9
MMT CO2 Eq. of CH4 and 3.2 MMT CO2 Eq. of N20 (Table 6-17; kt units provided in Table 6-18). The estimates
6-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
of non-C02 emissions from forest fires include wildfires and prescribed fires in the conterminous 48 states and all
managed forest land in Alaska.
Table 6-17: N011-CO2 Emissions from Forest Fires (MMT CO2 Eq.)a
Gas
1990
2005
2013
2014
2015
2016
2017"
CH4
1.5
5.2
6.1
6.1
12.6
4.9
4.9
N2O
1.0
3.4
4.0
4.0
8.3
3.2
3.2
Total
2.4
8.6
10.2
10.1
20.8
8.0
8.0
a These estimates include N011-CO2 Emissions from ForestFires on Forest Land Remaining
Forest Land and Land Converted to Forest Land.
bThe data for 2017 were unavailable when these estimates were developed, therefore 2016,
the most recent available estimate, is applied to 2017.
Table 6-18: N011-CO2 Emissions from Forest Fires (kt)a
Gas
1990
2005
2013
2014
2015
2016
2017"
CH4
59
208
245
243
502
194
194
N2O
3
11
14
13
28
11
11
CO
1,334
4,723
5,574
5,525
11,425
4,425
4,425
NOx
37
133
157
155
321
124
124
a These estimates include Non-CC>2 Emissions from Forest Fires on Forest Land Remaining
Forest Land and Land Converted to Forest Land.
b The data for 2017 were unavailable when these estimates were summarized, therefore
2016, the most recent available estimate, is applied to 2017.
Methodology and Data Sources
Non-C02 emissions from forest fires—primarily CH4 and N20 emissions—were calculated following IPCC (2006)
methodology, which included a combination of U.S.-specific data on area burned, potential fuel available for
combustion, and estimates of combustion based on fire severity along with IPCC default combustion and emission
factors. The estimates were calculated according to Equation 2.27 of IPCC (2006, Volume 4, Chapter 2), which is:
Emissions = Area burned x Fuel available x Combustion factor x Emission Factor x 10 3
where forest area burned is based on Monitoring Trends in Burn Severity (MTBS, Eidenshink et al. 2007 and
National Land Cover NLCD, Homer et al. 2015) data. Fuel estimates are based on current C density estimates
obtained from FIA plot data, combustion is partly a function of burn severity, and emission factors are from IPCC
(2006, Volume 4, Chapter 2). See Annex 3.13 for further details.
Uncertainty and Time-Series Consistency
In order to quantify the uncertainties for non-CCh 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 2017.
Table 6-19: Quantitative Uncertainty Estimates of Non-C02 Emissions from Forest Fires
(MMT CO2 Eq. and Percent)3
Source
Gas
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimateb
(MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Non-CC>2 Emissions from
ForestFires
CH4
4.9
4.1
5.7
-15% 17%
Land Use, Land-Use Change, and Forestry 6-37
-------
Non-CC>2 Emissions from
Forest Fires
3.2
2.8
3.6 -12% 14%
aThese estimates include Non-CCh Emissions from ForestFires 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 1990 through 2017 Inventory to compile estimates of non-C02 emissions from forest fires
are consistent with those used in the 1990 through 2016 Inventory, but also include some additional steps toward
better definition of forest area in Alaska, fuel, and combustion. Modifications in each of these factors affect
estimates. 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 the previous report. Fuel
estimates are based on the distribution of stand-level carbon pools (USDA Forest Service 2018b, 2018d) classified
according to ecological region rather than the state-wide estimates as in the previous report. Combustion estimates
are partly a function of the MTBS severity classifications and thus can vary within a fire. The effects of these
modifications varied across the time series, but more often lowered the estimates for both CH4 and N20.
Planned Improvements
Continued 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 many 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 forest land area.
N additions to soils result in direct and indirect N20 emissions. Direct emissions occur on-site due to the N
additions. Indirect emissions result from fertilizer N that is transformed and transported to another location in a form
other than N20 (ammonia [NH3] and nitrogen oxide [NOx] volatilization, nitrate [NO3] leaching and runoff), and
later converted into N20 at the off-site location. The indirect emissions are assigned to forest land because the
management activity leading to the emissions occurred in forest land.
Direct soil N20 emissions from Forest Land Remaining Forest Land and Land Converted to Forest Land in 2017
were 0.3 MMT C02 Eq. (1 kt), and the indirect emissions were 0.1 MMT C02 Eq. (0.4 kt). Total emissions for 2017
were 0.5 MMT C02 Eq. (2 kt) and have increased by 455 percent from 1990 to 2017. Total forest soil N20
emissions are summarized in Table 6-20.
6-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
2013 2014 2015 2016 2017
Direct N2O Fluxes from Soils
MMT CO2 Eq. 0.1
kt N2O +
Indirect N2O Fluxes from Soils
MMT CO2 Eq. 0.0
kt N2O +
0.3
1
0.1
+
0.3 0.3 0.3 0.3 0.3
11111
0.1 0.1 0.1 0.1 0.1
+ + + + +
Total
MMT CO2 Eq. 0.1
kt N2O +
0.5
2
0.5 0.5 0.5 0.5 0.5
2 2 2 2 2
+ Does not exceed 0.05 MMT CO2 Eq. or 0.5 kt.
Note: Totals may not sum due to independent rounding.
Methodology and Data Sources
The IPCC Tier 1 approach is used to estimate N2O from soils within Forest Land Remaining 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 accounted for N fertilizer application to commercial Douglas-fir stands in western Oregon and
Washington. For the Southeast, estimates of direct N20 emissions from fertilizer applications to forests are based on
the area of pine plantations receiving fertilizer in the southeastern United States and estimated application rates
(Albaugh et al. 2007; Fox et al. 2007). Not accounting for fertilizer applied to non-pine plantations is justified
because fertilization is routine for pine forests but rare for hardwoods (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
2017, 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 2017, so data from 2004 are used for these years.
The annual area estimates are multiplied by the typical rate used in this region (200 lbs. N per acre) to estimate total
N applied (Briggs 2007), and the total N applied to forests is multiplied by the IPCC (2006) default emission factor
of one percent to estimate direct N20 emissions.
For indirect emissions, the volatilization and leaching/runoff N fractions for forest land are calculated using the
IPCC default factors of 10 percent and 30 percent, respectively. The amount of N volatilized is multiplied by the
IPCC default factor of one percent for the portion of volatilized N that is converted to N20 off-site. The amount of N
leached/runoff is multiplied by the IPCC default factor of 0.075 percent for the portion of leached/runoff N that is
converted to N20 off-site. The resulting estimates are summed to obtain total indirect emissions.
Uncertainty and Time-Series Consistency
The amount of N20 emitted from forests depends not only on N inputs and fertilized area, but also on a large
number of variables, including organic C availability, oxygen gas partial pressure, soil moisture content, pH,
temperature, and tree planting/harvesting cycles. The effect of the combined interaction of these variables on N20
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 synthetic N
fertilizers are captured, so applications of organic N fertilizers are not estimated. However, the total quantity of
organic N inputs to soils is included in Section 5.4 Agricultural Soil Management and Section 6.10 Settlements
Remaining Settlements.
Land Use, Land-Use Change, and Forestry 6-39
-------
Uncertainties exist in the fertilization rates, annual area of forest lands receiving fertilizer, and the emission factors.
Fertilization rates are assigned a default level29 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 2017 emission estimates. IPCC (2006)
provided estimates for the uncertainty associated with direct and indirect N20 emission factor for synthetic N
fertilizer application to soils.
Uncertainty is quantified using simple error propagation methods (IPCC 2006). The results of the quantitative
uncertainty analysis are summarized in Table 6-21. Direct N20 fluxes from soils in 2017 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 2017 emission estimate of 0.3 MMT CO2 Eq. Indirect N2O emissions in 2017 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 2017 emission estimate.
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
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate
(MMT CO2 Eq.) (%)
Forest Land Remaining Forest
Lower
Upper
Lower
Upper
Land
Bound
Bound
Bound
Bound
Direct N2O Fluxes from Soils
N2O
0.3
0.1
1.1
-59%
+211%
Indirect N2O Fluxes from Soils
N2O
0.1
+
0.4
-86%
+238%
+ Does not exceed 0.05 MMT CO2 Eq.
Note: Due to rounding the upper and lower bounds may equal the emission estimate in the above table.
The same methods are applied to the entire time series to ensure time-series consistency from 1990 through 2017,
and no recalculations have been done from the previous Inventory. Details on the emission trends through time are
described in more detail in the Methodology section, above.
QA/QC and Verification
The spreadsheet tab containing fertilizer applied to forests and calculations for N20 and uncertainty ranges are
checked and verified.
Planned Improvements
Additional data will be compiled to update estimates of forest areas receiving N fertilizer using surrogate data in the
next Inventory. Another improvement is to further disaggregate emissions by state for southeastern pine plantations
and northwestern Douglas-fir forests to estimate soil N20 emission. This improvement is contingent on the
availability of state-level N fertilization data for forest land. Estimates of the N20 from mineralization of soil C will
also be included in the next Inventory.
C02, CH4/ and N20 Emissions from Drained Organic Soils30
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 N20 emissions (IPCC 2006). In
addition, the 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands
29 Uncertainty is unknown for the fertilization rates so a conservative value of ±50 percent is used in the analysis.
30 Estimates of C and CO2 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.
6-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
(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 (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-CCh emissions on forest land with
drained organic soils in 2017 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-C02 emissions provided here include CH4 and N20. Methane emissions generally associated with anoxic
conditions do occur from the drained land surface but the majority of these emissions originate from ditches
constructed to facilitate drainage at these sites. Emission of N20 can be significant from these drained organic soils
in contrast to the very low emissions from wet organic soils.
Table 6-22: N011-CO2 Emissions from Drained Organic Forest Soilsa'b (MMT CO2 Eq.)
Source
1990
2005
2013
2014
2015
2016
2017
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
+ Does not exceed 0.05 MMT CO2 Eq.
a This table includes estimates from Forest Land Remaining Forest Land and Land Converted to
Forest Land.
b Estimates of C and CO2 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.
Table 6-23: Non-C02 Emissions from Drained Organic Forest Soilsa'b (kt)
Source
1990
2005
2013
2014
2015
2016
2017
CH4
0.6
0.6
0.6
0.6
0.6
0.6
0.6
N2O
+
: + M
+
+
+
+
+
+ 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 CO2 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.
Land Use, Land-Use Change, and Forestry 6-41
-------
Methodology and Data Sources
The Tier 1 methods for estimating CO2, CH4 and N20 emissions from drained inland organic soils on forest lands
follow IPCC (2006), with extensive updates and additional material presented in the 2013 Supplement to the 2006
IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands (IPCC 2014). With the exception of
quantifying area of forest on drained organic soils, which is user-supplied, all quantities necessary for Tier 1
estimates are provided in Chapter 2, Drained Inland Organic Soils of IPCC (2014).
Estimated area of drained organic soils on forest land is 70,849 ha based on analysis of the permanent NFI of the
USDA Forest Service and did not change over the time series. The most recent plot data per state within the
inventories were used in a spatial overlay with the STATSG02 (2016) soils data, and forest plots coincident with the
soil order histosol were selected as having organic soils. Information specific to identifying "drained organic" are
not in the inventory data so an indirect approach was employed here. Specifically, artificially regenerated forest
stands (inventory field STDORGCD=l) on mesic or xeric sites (inventory field 11
-------
Uncertainty and Time-Series Consistency
Uncertainties are based on the sampling error associated with forest area of drained organic soils and the
uncertainties provided in the Chapter 2 (IPCC 2014) emissions factors (Table 6-25). The estimates and resulting
quantities representing uncertainty are based on the IPCC Approach 1-error propagation. However, probabilistic
sampling of the distributions defined for each emission factor produced a histogram result that contained a mean and
95 percent confidence interval. The primary reason for this approach was to develop a numerical representation of
uncertainty with the potential for combining with other forest components. The total non-CCh emissions in 2017
from drained organic soils on Forest Land Remaining Forest Land and Land Converted to Forest Land were
estimated to be between 0.07 and 0.2 MMT CO2 Eq. around a central estimate of 0.1 MMT CO2 Eq. at a 95 percent
confidence level.
Table 6-25: Quantitative Uncertainty Estimates for Non-C02 Emissions on Drained Organic
Forest Soils (MMT CO2 Eq. and Percent)3
2017 Emission
Source Estimate Uncertainty Range Relative to Emission Estimate
(MMT CO2 Eq.) (MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
ch4
n2o
+
0.1
+
+
+
0.2
-76%
-124%
76%
124%
Total
0.1
0.07
0.2
-108%
108%
+ Does not exceed 0.05 MMT CO2 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-C02 emissions may be included in
either the Wetlands or sections on N20 emissions from managed soils. These paths to double counting emissions are
unlikely here because these issues are taken into consideration when developing the estimates and this chapter is the
only section directly including such emissions on forest land.
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.
6.3 Land Converted to Forest Land (CRF
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.31 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 2006 IPCC Guidelines (IPCC 2006), after which they
31 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.
Land Use, Land-Use Change, and Forestry 6-43
-------
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 StatesS2
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 year1.
Over the 20-year conversion period used in the Land Converted to Forest Land category, the conversion of cropland
to forest land resulted in the largest source of C transfer and uptake, accounting for approximately 40 percent of the
uptake annually. Estimated C uptake has remained relatively stable over the time series across all conversion
categories (see Table 6-26). The net flux of C from all forest pool stock changes in 2017 was -120.6 MMT CO2 Eq.
(-32.9 MMT C) (Table 6-26 and Table 6-27).
Mineral soil C stocks are increasing slightly in the early 1990s w ith Land Converted Forest Land, but this trend
reverses in the early 2000's through the remainder of the time series. The small gains in the early part of the time
series are driven by Cropland Converted to Forest Land during the 1990s. Much of this conversion is from annual
crop production, which lias a lower mineral soils C stock than Forest Land. In contrast. Grassland Converted to
Forest Land dominates the trend starting in the early 2000s. Managed pasture to Forest Land is the most common
conversion. This 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 Forest Land by Land Use
Change Category (MMT CO2 Eq.)
Land Use/Carbon Pool
1990
2005
2013
2014
2015
2016
2017
Cropland Converted to Forest Land
(47.4)
(47.8)
(48.0)
(48.1)
(48.0)
(48.0)
(48.0)
Aboveground Biomass
(26.7)
(27.0)
(27.2)
(27.2)
(27.2)
(27.2)
(27.2)
Belowground Biomass
(5.3)
(5.4)
(5.4)
(5.4)
(5.4)
(5.4)
(5.4)
Dead Wood
(6.1)
(6.2)
(6.2)
(6.2)
(6.2)
(6.2)
(6.2)
Litter
(9.0)
(9.1)
(9.1)
(9.1)
(9.1)
(9.1)
(9.1)
Mineral Soil
(0.4)
(0.2)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Grassland Converted to Forest Land
(11.0)
(11.0)
(10.9)
(11.0)
(11.1)
(11.1)
(11.2)
Aboveground Biomass
(5.6)
(5.7)
(5.7)
(5.7)
(5.7)
(5.7)
(5.7)
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
(4.1)
(4.1)
(4.1)
(4.1)
(4.1)
(4.1)
(4.1)
Mineral Soil
0.3
0.4
0.5
0.5
0.3
0.3
0.3
Other Land Converted to Forest Land
(18.1)
(18.2)
(18.3)
(18.3)
(18.3)
(18.3)
(18.3)
Aboveground Biomass
(9.1)
(9.2)
(9.2)
(9.2)
(9.2)
(9.2)
(9.2)
Belowground Biomass
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
Dead Wood
(2.4)
(2.4)
(2.4)
(2.4)
(2.4)
(2.4)
(2.4)
Litter
(4.9)
(4.9)
(5.0)
(5.0)
(5.0)
(5.0)
(5.0)
Mineral Soil
+
+
+
+
+
+
+
Settlements Converted to Forest Land
(41.1)
(41.4)
(41.7)
(41.7)
(41.7)
(41.7)
(41.7)
Aboveground Biomass
(24.7)
(24.9)
(25.0)
(25.1)
(25.1)
(25.1)
(25.1)
Belowground Biomass
(4.8)
(4.9)
(4.9)
(4.9)
(4.9)
(4.9)
(4.9)
32 Hie estimates reported in this section only include the 48 conterminous states in the US. Land use conversion to forest in
Alaska and Hawaii were not included.
6-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Dead Wood
(4.8)
(4.8)
(4.8)
(4.8)
(4.8)
(4.8)
(4.8)
Litter
(6.7)
(6.7)
(6.7)
(6.7)
(6.7)
(6.7)
(6.7)
Mineral Soil
(0.1)
(0.1)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Wetlands Converted to Forest Land
(1.4)
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
Abovegroimd Biomass
(0.7)
(0.7)
(0.7)
(0.7)
(0.7)
(0.7)
(0.7)
Belowground Biomass
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Dead Wood
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Litter
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
Mineral Soil
+
+
+
+
+
+
+
Total Aboveground Biomass Flux
(66.8)
(67.5)
(67.8)
(67.9)
(67.9)
(67.9)
(67.9)
Total Belowground Biomass Flux
(12.9)
(13.0)
(13.1)
(13.1)
(13.1)
(13.1)
(13.1)
Total Dead Wood Flux
(14.1)
(14.2)
(14.3)
(14.3)
(14.3)
(14.3)
(14.3)
Total Litter Flux
(25.0)
(25.3)
(25.4)
(25.5)
(25.5)
(25.5)
(25.5)
Total Mineral Soil Flux
(0.2)
0.1
0.2
0.2
0.1
+
0.1
Total Flux
(119.1)
(120.0)
(120.5)
(120.5)
(120.6)
(120.6)
(120.6)
+ Absolute value does not exceed 0.05 MMT CO2 Eq.
Notes: Totals may not sum due to independent rounding. Parentheses indicate net uptake. These estimates only include land
conversions in the CONUS-land conversions in Alaska and Hawaii were not included in this Inventory.
Table 6-27: Net C Flux from Forest C Pools in Land Converted to Forest Land by Land Use
Change Category (MMT C)
Land Use/Carbon Pool
1990
2005
2013
2014
2015
2016
2017
Cropland Converted to Forest Land
(12.9)
(13.0)
(13.1)
(13.1)
(13.1)
(13.1)
(13.1)
Abovegroimd Biomass
(7.3)
(7.4)
(7.4)
(7.4)
(7.4)
(7.4)
(7.4)
Belowground Biomass
(1.4)
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
Dead Wood
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
(1.7)
Litter
(2.4)
(2.5)
(2.5)
(2.5)
(2.5)
(2.5)
(2.5)
Mineral Soil
(0.1)
(0.1)
+
+
+
+
+
Grassland Converted to Forest Land
(3.0)
(3.0)
(3.0)
(3.0)
(3.0)
(3.0)
(3.0)
Abovegroimd Biomass
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
(1.5)
Belowground Biomass
(0.2)
(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
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
(1.1)
Mineral Soil
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Other Land Converted to Forest Land
(4.9)
(5.0)
(5.0)
(5.0)
(5.0)
(5.0)
(5.0)
Abovegroimd Biomass
(2.5)
(2.5)
(2.5)
(2.5)
(2.5)
(2.5)
(2.5)
Belowground Biomass
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
(0.5)
Dead Wood
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
(0.6)
Litter
(1.3)
(1.3)
(1.4)
(1.4)
(1.4)
(1.4)
(1.4)
Mineral Soil
+
+
+
+
+
+
+
Settlements Converted to Forest Land
(11.2)
(11.3)
(11.4)
(11.4)
(11.4)
(11.4)
(11.4)
Abovegroimd Biomass
(6.7)
(6.8)
(6.8)
(6.8)
(6.8)
(6.8)
(6.8)
Belowground Biomass
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
Dead Wood
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
(1.3)
Litter
(1.8)
(1.8)
(1.8)
(1.8)
(1.8)
(1.8)
(1.8)
Mineral Soil
+
+
(0.1)
+
+
+
+
Wetlands Converted to Forest Land
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
(0.4)
Abovegroimd Biomass
(0.19)
(0.19)
(0.19)
(0.19)
(0.19)
(0.19)
(0.19)
Belowground Biomass
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
Dead Wood
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
Litter
(0.13)
(0.13)
(0.13)
(0.13)
(0.13)
(0.13)
(0.13)
Mineral Soil
+
+
+
+
+
+
+
Total Aboveground Biomass Flux
(18.2)
(18.4)
(18.5)
(18.5)
(18.5)
(18.5)
(18.5)
Total Belowground Biomass Flux
(3.5)
(3.5)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
Total Dead Wood Flux
(3.8)
(3.9)
(3.9)
(3.9)
(3.9)
(3.9)
(3.9)
Total Litter Flux
(6.8)
(6.9)
(6.9)
(6.9)
(6.9)
(6.9)
(6.9)
Total Mineral Soil Flux
(0.1)
+
+
0.1
+
+
+
Total Flux
(32.5)
(32.7)
(32.9)
(32.9)
(32.9)
(32.9)
(32.9)
Land Use, Land-Use Change, and Forestry 6-45
-------
+ Absolute value does not exceed 0.05 MMT C.
Notes: Totals may not sum due to independent rounding. Parentheses indicate net uptake. These estimates only
include land conversions in the CONUS-land conversions in Alaska and Hawaii were not included in this 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).
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 2018 were used in this Inventory. This
may result in inconsistencies with the other land use categories and the area estimates reported in the Land
Representation since new area activity data were not compiled for the other land use categories in this Inventory (see
Section 6.1 Representation of the U.S. Land Base). 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 2017. 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. (201 la), 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. (201 la), 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
6-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
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 2012 National Resources Inventory (NRI) (USDA-NRCS 2013), and National Land
Cover Dataset (NLCD) (Homer et al. 2007). See Annex 3.12 (Methodology for Estimating N20 Emissions, CH4
Emissions and Soil Organic C Stock Changes from Agricultural Soil Management) for more information about this
method. Note that soil C in this Inventory 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 report soil C to a depth of 20
or 30 cm. To ensure consistency in the Land Converted to Forest Land category "where C stock transfers occur
between land-use categories, all soil C estimates were obtained using methods from Ogle et al. (2003, 2006) and
IPCC (2006), which are also used in the Cropland, Grasslands and Settlements land use categories in this Inventory.
Uncertainty and Time-Series Consistency
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 9 percent below to 9 percent above the 2017 C stock change
estimate of -120.6 MMT CO2 Eq.
Table 6-28: Quantitative Uncertainty Estimates for Forest C Pool Stock Changes (MMT CO2
Eq. per Year) in 2017 from Land Converted to Forest Land by Land Use Change
Land Use/Carbon Pool
2017 Flux
Estimate
Uncertainty Range Relative to Flux Range3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Cropland Converted to Forest Land
Aboveground Biomass
Belowground Biomass
(48.0)
(27.2)
(5.4)
(56.8)
(35.8)
(6.5)
(39.2)
(18.6)
(4.3)
-18% 18%
-32% 32%
-20% 20%
Land Use, Land-Use Change, and Forestry 6-47
-------
Dead Wood
(6.2)
(7.4)
(5.0)
-19%
20%
Litter
(9.1)
(10.2)
(8.1)
-12%
12%
Mineral Soils
(0.1)
(0.2)
+
-164%
159%
Grassland Converted to Forest Land
(11.2)
(13.5)
(8.8)
21%
21%
Aboveground Biomass
(5.7)
(7.1)
(4.3)
-25%
25%
Belowground Biomass
(0.9)
(1.2)
(0.6)
-31%
31%
Dead Wood
(0.7)
(0.8)
(0.5)
-22%
22%
Litter
(4.1)
(4.7)
(3.6)
-13%
13%
Mineral Soils
0.2
0.1
0.5
-73%
78%
Other Lands Converted to Forest
Land
(18.3)
(20.6)
(16.0)
-13%
13%
Aboveground Biomass
(9.2)
(11.3)
(7.1)
-23%
23%
Belowground Biomass
(1.7)
(2.2)
(1.3)
-25%
25%
Dead Wood
(2.4)
(2.9)
(1.8)
-24%
24%
Litter
(5.0)
(5.6)
(4.3)
-13%
13%
Mineral Soils
+
+
+
-85%
98%
Settlements Converted to Forest Land
(41.7)
(48.2)
(35.2)
-16%
16%
Aboveground Biomass
(25.1)
(31.2)
(18.9)
-25%
25%
Belowground Biomass
(4.9)
(6.2)
(3.6)
-27%
27%
Dead Wood
(4.8)
(6.0)
(3.7)
-24%
24%
Litter
(6.7)
(7.6)
(5.8)
-14%
14%
Mineral Soils
(0.2)
(0.2)
(0.1)
-23%
20%
Wetlands Converted to Forest Land
(1.5)
(1.7)
(1.3)
-11%
11%
Aboveground Biomass
(0.7)
(0.9)
(0.6)
-20%
20%
Belowground Biomass
(0.1)
(0.2)
(0.1)
-22%
22%
Dead Wood
(0.2)
(0.2)
(0.1)
-27%
27%
Litter
(0.5)
(0.5)
(0.4)
-13%
13%
Mineral Soils
+
+
+
-84%
95%
Total: Aboveground Biomass
(67.9)
(78.8)
(57.0)
-16%
16%
Total: Belowground Biomass
(13.1)
(14.9)
(11.3)
-14%
14%
Total: Dead Wood
(14.3)
(16.1)
(12.5)
-12%
12%
Total: Litter
(25.5)
(27.1)
(23.8)
-6%
6%
Total: Mineral Soils
0.1
(0.2)
0.3
-358%
372%
Total: Lands Converted to Forest
Lands
(120.6)
(131.9)
(109.3)
-9%
9%
+ Absolute value does not exceed 0.05 MMT CO2 Eq.
a Range of flux estimate for 95 percent confidence interval
Note: Parentheses indicate net uptake.
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 2018 were used in this Inventory. This may result in inconsistencies with other land use categories
reported in the Land Representation since new area activity data were not compiled for the current Inventory (see
Section 6.1 Representation of the U.S. Land Base). This is the first 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
6-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
combination of NRI data and NFI data in the eastern U.S. 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
increased by 38 percent in 2016 between the previous Inventory and the current Inventory (Table 6-29). This
increase is directly attributed to the incorporation of annual NFI data into the compilation system. In the previous
Inventory, Grasslands Converted to Forest Land represented the largest transfer and uptake of C across the land use
conversion categories. In this Inventory, Cropland Converted to Forest Land represented the largest transfer and
uptake of C across the land use change categories followed by Settlements Converted to Forest Land (Table 6-29).
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
2016 Estimate,
2016 Estimate,
2017 Estimate,
and Carbon pool (MMT C)
Previous Inventory
Current Inventory
Current Inventory
Cropland Converted to Forest Land
(3.2)
(13.1)
(13.1)
Aboveground Biomass
(1.3)
(7.4)
(7.4)
Belowground Biomass
(0.1)
(1.5)
(1.5)
Dead Wood
(0.7)
(1.7)
(1.7)
Litter
(1.1)
(2.5)
(2.5)
Mineral soil
+
+
+
Grassland Converted to Forest Land
(13.7)
(3.0)
(3.0)
Aboveground Biomass
(7.0)
(1.5)
(1.5)
Belowground Biomass
1.6
(0.3)
(0.3)
Dead Wood
(3.1)
(0.2)
(0.2)
Litter
(5.2)
(1.1)
(1.1)
Mineral soil
+
0.1
0.1
Other Land Converted to Forest
Land
(2.5)
(5.0)
(5.0)
Aboveground Biomass
(1.1)
(2.5)
(2.5)
Belowground Biomass
(0.2)
(0.5)
(0.5)
Dead Wood
(0.4)
(0.6)
(0.6)
Litter
(0.7)
(1.4)
(1.4)
Mineral soil
+
+
+
Settlements Converted to Forest
Land
(0.5)
(11.4)
(11.4)
Aboveground Biomass
(0.2)
(6.8)
(6.8)
Belowground Biomass
(0.0)
(1.3)
(1.3)
Dead Wood
(0.1)
(1.3)
(1.3)
Litter
(0.1)
(1.8)
(1.8)
Mineral soil
+
+
+
Wetlands Converted to Forest Land
(0.6)
(0.4)
(0.4)
Aboveground Biomass
(0.28)
(0.19)
(0.19)
Belowground Biomass
(0.05)
(0.04)
(0.04)
Dead Wood
(0.09)
(0.04)
(0.04)
Litter
(0.19)
(0.13)
(0.13)
Mineral soil
+
+
+
Total Aboveground Biomass Flux
(9.9)
(18.5)
(18.5)
Total Belowground Biomass Flux
1.2
(3.6)
(3.6)
Total Dead Wood Flux
(4.3)
(3.9)
(3.9)
Total Litter Flux
(7.4)
(6.9)
(6.9)
Total SOC (mineral) Flux
+
+
+
Total Flux
(20.5)
(32.9)
(32.9)
+ Absolute value does not exceed 0.05 MMT C.
Notes: Totals may not sum due to independent rounding. Parentheses indicate net uptake.
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 20 or 30 cm. To ensure greater consistency in the Land
Land Use, Land-Use Change, and Forestry 6-49
-------
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 USD A 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. A section on N20 emissions from forest soils that includes estimates of the N20 from
mineralization of soil C will be provided in the next Inventory.
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, and 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.33
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 during the 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., sewage
sludge) and flooding, can modify both organic matter inputs and decomposition, and thereby result in a net C stock
change (Parton et al. 1987; Paustian et al. 1997a; Conant et al. 2001; Ogle et al. 2005). 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.34 Due to the depth and richness of the organic layers, C
loss from drained organic soils can continue over long periods of time, which varies depending on climate and
composition (i.e., decomposability) of the organic matter (Armentano and Menges 1986). Due to deeper drainage
and more intensive management practices, the use of organic soils for annual crop production (and also settlements)
leads to higher C loss rates than drainage of organic soils in grassland or forests (IPCC 2006).
33 Carbon dioxide emissions associated with liming and urea application are also estimated but are included in the Agriculture
chapter of the report.
34 N2O emissions from soils are included in the Agricultural Soil Management section.
6-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Cropland Remaining Cropland includes all cropland in an Inventory year that has been cropland for a continuous
time period of at least 20 years according to the 2012 United States Department of Agriculture (USD A) National
Resources Inventory (NRI) land-use survey for non-federal lands (USDA-NRCS 2015) or according to the National
Land Cover Dataset for federal lands (Homer et al. 2007; Fry et al. 2011; Homer et al. 2015). Cropland includes all
land used to produce food and fiber, in addition to forage that is harvested and used as feed (e.g., hay and silage),
and cropland that has been enrolled in the Conservation Reserve Program (CRP) (i.e., considered reserve 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 total amount of managed area in Cropland Remaining Cropland (see Section 6.1
Representation of the U.S. Land Base) and the cropland area included in the Inventory analysis.35 Improvements are
underway to include croplands in Alaska as part of future C inventories.
Carbon dioxide emissions and removals36 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 will be recalculated in a future Inventory report using the Tier 2 and 3 methods.
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 2017, mineral soils are estimated to sequester
40.0 MMT CO2 Eq. from the atmosphere (10.9 MMT C). This rate of C storage in mineral soils represents about a
44 percent decrease in the rate since the initial reporting year of 1990. Carbon dioxide emissions from organic soils
are 29.7 MMT CO2 Eq. (8.1 MMT C) in 2017, which is a 2 percent decrease compared to 1990. In total, United
States agricultural soils in Cropland Remaining Cropland sequestered approximately 10.3 MMT CO2 Eq. (2.8 MMT
C) in 2017.
Table 6-30: Net CO2 Flux from Soil C Stock Changes in Cropland Remaining Cropland (MMT
COz Eq.)
Soil Type
1990
2005
2013
2014
2015
2016
2017
Mineral Soils
(71.2)
(56.2)
(41.5)
(41.7)
(36.3)
(39.7)
(40.0)
Organic Soils
30.3
29.7
30.1
29.7
30.0
29.8
29.7
Total Net Flux
(40.9)
(26.5)
(11.4)
(12.0)
(6.3)
(9.9)
(10.3)
Note: Parentheses indicate net sequestration.
Table 6-31: Net CO2 Flux from Soil C Stock Changes in Cropland Remaining Cropland (WAT
C)
Soil Type
1990
2005
2013
2014
2015
2016
2017
Mineral Soils
Organic Soils
(19.4)
8.3
(15.3)
8.1
(11.3)
8.2
(11.4)
8.1
(9.9)
8.2
(10.8)
8.1
(10.9)
8.1
Total Net Flux
(11.2)
* (7.2)
(3.1)
(3-3)
(1.7)
(2.7)
(2.8)
35 For the U.S. land representation, land use data for 2013 to 2017 were only partially updated based on new Forest Inventory
and Analysis (FIA) data. These updates led to changes in the land representation data for croplands through the process of
combining FIA data with land use data from the National Resources Inventory and National Land Cover Dataset (See
"Representation of the U.S. Land Base" section for more information). However, an inventory was not compiled for croplands
with the new land representation data so the area estimates in this section are based on the land representation data from the
previous Inventory. This has created additional discrepancies with the reported cropland areas in the "Representation of the U.S.
Land Base" section.
36 Removals occur through uptake of CO2 into crop and forage biomass that is later incorporated into soil C pools.
Land Use, Land-Use Change, and Forestry 6-51
-------
Note: 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 program), 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. However, there is a decline in the net amount of C sequestration (i.e., 2017 is 44 percent less than
1990), and this decline is largely due to lower sequestration rates and less annual cropland enrolled in the CRP37 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 7 percent (based on 2012 estimates) for Cropland
Remaining Cropland on organic soils since 1990.
The spatial variability in the 2012 annual soil C stock changes38 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 with gains in the northern Great
Plains, which lias high rates of CRP enrollment. High rates of net C accumulation in mineral soils also occurred in
the Com Belt region, w hich is the region with the largest amounts of conservation tillage, along with moderate rates
of CRP enrollment. The regions with the highest rates of emissions from drainage of organic soils occur in the
Southeastern Coastal Region (particularly Florida), upper Midwest and Northeast surrounding the Great Lakes, and
isolated areas along the Pacific Coast (particularly California), which coincides with the largest concentrations of
organic soils in the United States that are used for agricultural production.
Figure 6-5: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural
Management within States, 2012, Cropland Remaining Cropland
MT C02 ha1 yr1
¦ < -4 ~ 1 to 2
¦ -4 to -2 ~ 2 to 4
~ -2 to -1 ¦ > 4
~ -1 to 1
w i
Note: Only national-scale soil C stock changes are estimated for 2013 to 2017 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
37 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.
38 Only national-scale emissions are estimated for 2013 to 2017 in this Inventory using the surrogate data method, and therefore
the fine-scale emission patterns in this map are based on inventory data from 2012.
6-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
2012. Negative values represent a net increase in soil C stocks, and positive values represent a net decrease in soil
C stocks.
Figure 6-6: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural
Management within States, 2012, Cropland Remaining Cropland
MT C02 ha-1 yr1
~ < 10
~ 10 to 20
¦ 20 to 30
¦ 30 to 40
¦ > 40
Note: Only national-scale soil C stock changes are estimated for 2013 to 2017 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
2012.
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.
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 2015).
The NRI is a statistically-based sample of all non-federal land, and includes approximately 609,211 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
expansion factor represents the amount of area with 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
2012 (USDA-NRCS 2015). NRI survey locations are classified as Cropland Remaining Cropland in a given year
between 1990 and 2012 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.
Land Use, Land-Use Change, and Forestry 6-53
-------
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 in the United States. These crops include alfalfa hay,
barley, com cotton dry beans, grass hay, grass-clover hay, lentils, oats, onions, peanuts, peas, potatoes, rice,
sorghum, soybeans, sugar beets, sunflowers, tobacco, tomatoes, 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 and soil
nitrous oxide (N20) emissions from agricultural soil management. Carbon and N dynamics are linked in plant-soil
systems through the biogeochemical processes of microbial decomposition and plant production (McGill and Cole
1981). Coupling the two source categories (i.e., agricultural soil C and N20) in a single inventory analysis ensures
that there is a consistent treatment of the processes and interactions between C and N cycling in soils.
The remaining crops on mineral soils are estimated using an IPCC Tier 2 method (Ogle et al. 2003), including some
vegetables, 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, lias 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. Further elaboration on the methodology and data used to
estimate stock changes from mineral soils are described below and in Annex 3.12.
A surrogate data method is used to estimate soil C stock changes from 2013 to 2017 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 2012 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,39 and weather data from the
PRISM Climate Group (PRISM 2015). See Box 6-4 for more information about the surrogate data method. Stock
change estimates for 2013 to 2017 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 data
every year.
A surrogate data method lias 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 1990 to
2012 emissions data that lias been compiled using the inventory methods described in this section. The model to
extend the time series is given by
Y = XP + e,
where Y is the response variable (e.g., soil organic carbon), X(3 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
emissions data for 1990 to 2012 using standard statistical techniques, and these estimates are used to predict the
missing emissions data for 2013 to 2017.
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. 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 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
39 See .
6-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 2012), estimating emissions from each model and deriving confidence
intervals, which propagates uncertainties through the calculations from the original inventory and the surrogate data
method.
Tier 3 Approach. Mineral SOC stocks and stock changes are estimated using the DAYCENT biogeochemical40
model (Parton et al. 1998; Del Grosso et al. 2001, 2011), which is able to simulate 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 lias been
refined to simulate dynamics at a daily time-step. The modeling approach uses daily weather data as an input, along
with information about soil physical properties. 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, and grazing). The model simulates net primary productivity
(NPP) using the NASA-CASA production algorithm MODIS Enhanced Vegetation Index (EVI) products,
MOD13Q1 and MYD13Q1, for most croplands41 (Potter et al. 1993, 2007). The model also simulates soil
temperature, and water dynamics, in addition to turnover, stabilization, and mineralization of soil organic matter C
and nutrients (N, P, K, S). This method is more accurate than the Tier 1 and 2 approaches provided by the IPCC
(2006) because the simulation model treats changes as continuous over time as opposed to the simplified discrete
changes represented in the default method (see Box 6-5 for additional information).
Box 6-5: Tier 3 Approach for Soil C Stocks Compared to Tier 1 or 2 Approaches
A Tier 3 model-based approach is used to estimate soil 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 of combinations in the United States. In contrast, the Tier 3
model simulates soil C dynamics at more than 300,000 individual NRI survey locations in individual fields.
(3) The IPCC Tier 1 and 2 methods use a simplified approach to 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 enviromnental 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.
40 Biogeochemical cycles are the flow of chemical elements and compounds between living organisms and the physical
environment.
41 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 2012. 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 6-55
-------
Historical land-use patterns and irrigation histories are simulated with DAYCENT based on the 2012 USDA NRI
survey (USDA-NRCS 2015). Additional sources of activity data are used to supplement the land-use information
from the NRI. The Conservation Technology Information Center (CTIC 2004) provided annual data on tillage
activity at the county level for the conterminous United States between 1989 and 2004, and these data are adjusted
for long-term adoption of no-till agriculture (Towery 2001). No-till adoption is assumed to remain constant from
2005 through 2012 due to lack of data, but there is a planned improvement to update the tillage histories with a
dataset that was recently released by the USDA (Conservation Effects Assessment Program Data, See Planned
Improvements section). Information on fertilizer use and rates by crop type for different regions of the United States
are obtained primarily from the USDA Economic Research Service. The data collection program was known as the
Cropping Practices Surveys through 1995 (USDA-ERS 1997), and then became the Agricultural Resource
Management Surveys (ARMS) (USDA-ERS 2015). Additional data are compiled through other sources particularly
the National Agricultural Statistics Service (NASS 1992, 1999, 2004). Frequency and rates of manure application to
cropland for 1997 are estimated from data compiled by the USDA Natural Resources Conservation Service
(Edmonds et al. 2003), and then adjusted using county-level estimates of manure available for application in other
years. Specifically, county-scale ratios of manure available for application to soils in other years relative to 1997 are
used to adjust the area amended with manure (see Annex 3.12 for further details). Greater availability of managed
manure N relative to 1997 is assumed to increase the area amended with manure, while reduced availability of
manure N relative to 1997 is assumed to reduce the amended area. Data on the county-level N available for
application are estimated for managed systems based on the total amount of N excreted in manure minus N losses
during storage and transport, and include the addition of N from bedding materials. Nitrogen losses include direct
N20 emissions, volatilization of ammonia and NOx, N runoff and leaching, and the N in poultry manure used as a
feed supplement. More information on livestock manure production is available in Section 5.2 Manure Management
and Annex 3.11.
Daily weather data are another input to the model simulations. These data are based on a 4 kilometer gridded
product from the PRISM Climate Group (2015). Soil attributes are obtained from the Soil Survey Geographic
Database (SSURGO) (Soil Survey Staff 2016). The C dynamics at each NRI point are simulated 100 times as part of
the uncertainty analysis, yielding a total of over 18 million simulation runs for the analysis. Uncertainty in the C
stock estimates from DAYCENT associated with parameterization and model algorithms are adjusted using a
structural uncertainty estimator accounting for uncertainty in model algorithms and parameter values (Ogle et al.
2007, 2010). Carbon stocks and 95 percent confidence intervals are estimated for each year between 1990 and 2012
using the NRI survey data (which is available through 2012). However, the areas may have changed through the
process in which the NRI survey data are reconciled with 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). This process ensures that
the areas of Forest Land Remaining Forest Land and Land Converted to Forest Land are consistent in all three
datasets, and leads to some modification of other lands use areas to ensure the total land area of the United States
does not change. 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).
Soil C stock changes from 2013 to 2017 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 2013 to 2017 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 (Ogle et al. 2003, 2006). 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 provided 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.
6-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 point locations.
Activity data are primarily based on the historical land-use/management patterns recorded in the 2012 NRI (USDA-
NRCS 2015). Each NRI point is classified by land use, soil type, climate region, and management condition. 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). Land-use patterns at
the NRI survey locations on federal lands are based on the National Land Cover Database (NLCD) (Fry et al. 2011;
Homer et al. 2007; Homer et al. 2015). Classification of cropland area by tillage practice is based on data from the
Conservation Technology Information Center (CTIC 2004; Towery 2001) as described in the Tier 3 approach above.
Activity data on wetland restoration of Conservation Reserve Program land are obtained from Euliss and Gleason
(2002). Manure N amendments over the inventory time period are based on application rates and areas amended
with manure N from Edmonds et al. (2003), in addition to the managed manure production data discussed in the
methodology subsection for the Tier 3 approach. Utilizing information from these data sources, SOC stocks for
mineral soils are estimated 50,000 times for 1990 through 2012, 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. 2002; Ogle et al. 2003; Ogle et al. 2006).
Soil C stock changes from 2013 to 2017 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 included a measure of uncertainty as determined from the Monte Carlo Stochastic Simulation
with 50,000 iterations. Emissions are based on the annual data for drained organic soils from 1990 to 2012 for
Cropland Remaining Cropland areas in the 2012 NRI (USDA-NRCS 2015). A surrogate data method is used to
estimate annual C emissions from organic soils from 2013 to 2017 as described in Box 6-4 of this section. Estimates
for 2013 to 2017 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 2013 to 2017, there is additional uncertainty propagated
through the Monte Carlo Analysis 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 ranged from
423 percent below to 423 percent above the 2017 stock change estimate of -10.3 MMT CO2 Eq. The large relative
uncertainty around the 2017 stock change estimate is partly due to variation in soil C stock changes that are not
explained by the surrogate data method, leading to high prediction error with this splicing method. The estimate is
also near zero for the total emissions and the Tier 3 Inventory with a lower bound below zero and an upper bound
above zero, leading to large relative uncertainty.
Land Use, Land-Use Change, and Forestry 6-57
-------
Table 6-32: Approach 2 Quantitative Uncertainty Estimates for Soil C Stock Changes
occurring within Cropland Remaining Cropland (MMT CO2 Eq. and Percent)
2017 Flux Estimate Uncertainty Range Relative to Flux Estimate3
(MMT CO2 Eq.) (MMT CCh Eq.) (%)
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
(36.5)
(3.5)
29.7
(79.8)
(6.9)
26.5
6.8
(0.1)
32.9
-119%
-96%
-11%
119%
96%
11%
Combined Uncertainty for Flux associated
with Agricultural Soil Carbon Stock
Change in Cropland Remaining Cropland
(10.3)
(53.8)
33.2
-423%
423%
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation with a 95 percent confidence interval.
Note: Parentheses indicate net sequestration.
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 biomass
C stocks 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.
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 a
statistical relationship lias been developed to assess uncertainties in the predictive capability of the model. The
comparisons include 92 long-term experiments, representing about 908 combinations of management treatments
across all of the sites (see Ogle et al. 2007 and Annex 3.12 for more information).
Planned Improvements
New land representation data have not been compiled for the current Inventory, and a surrogate data method has
been 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 a future Inventory will be to recalculate the time series for soil C
stock changes by applying the Tier 2 and 3 methods with the latest land use data from the National Resources
Inventory and related management statistics compiled through the Conservation Effects Assessment Program
(discussed below).
There are several other planned improvements underway. The DAYCENT model will be refined to simulate soil
organic C stock changes to a depth of at least 30 cm (currently at 20 cm). Improvements are also underway to more
accurately simulate plant production. 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.
Experimental study sites will continue to be added for quantifying model structural uncertainty.
6-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
There is an effort underway to update the time series of management data with information from the USDA-NRCS
Conservation Effects Assessment Program (CEAP). This improvement will fill several gaps in the management data
including more specific data on fertilizer rates, updated tillage practices, and more information on planting and
harvesting dates for crops.
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 and managed
grassland, 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, but will be
further refined over time to incorporate more of the management data that drive C stock changes on long-term
cropland.
Many of these improvements are expected to be completed for the next 1990 through 2018 Inventory (i.e., 2020
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.
6.5 Land Converted to Cropland (CRF Category
4B2)
Land Converted to Cropland includes all cropland in an inventory year that had been in another land use(s) during
the previous 20 years (USDA-NRCS 2015), 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) or climate zones (e.g., boreal climates). Consequently, there is a discrepancy between the total
amount of managed area in Land Converted to Cropland (see Section 6.1 Representation of the U.S. Land Base) and
the cropland area included in the Inventory.42 Improvements are underway to include croplands in Alaska and
miscellaneous croplands in future C inventories.
Land use change can lead to large losses of C to the atmosphere, particularly conversions from forest land
(Houghton et al. 1983). Moreover, conversion of forest to another land use (i.e., deforestation) is one of the largest
anthropogenic sources of emissions to the atmosphere globally (Schimel 1995), although this source may be
declining according to a recent assessment (Tubiello et al. 2015).
The 2006 IPCC 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
42 For the U.S. land representation, land use data for 2013 to 2017 were only partially updated based on new Forest Inventory
and Analysis (FIA) data. These updates led to changes in the land representation data for croplands through the process of
combining FIA data with land use data from the National Resources Inventory and National Land Cover Dataset (See
"Representation of the U.S. Land Base" section for more information). However, an inventory was not compiled for croplands
with the new land representation data so the area estimates in this section are based on the land representation data from the
previous Inventory. This has created additional discrepancies with the reported cropland areas in the "Representation of the U.S.
Land Base" section.
Land Use, Land-Use Change, and Forestry 6-59
-------
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,43
Forest Land Converted to Cropland is the largest source of emissions from 1990 to 2017, accounting for
approximately 70 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 with the majority of the loss
occurring in the mineral soil C, accounting for approximately 28 percent of the total emissions (Table 6-33 and
Table 6-34).
The net change in total C stocks for 2017 led to CO2 emissions to the atmosphere of 66.9 MMT CO2 Eq. (18.2 MMT
C), including 27.2 MMT CO2 Eq. (7.4 MMT C) from aboveground biomass C losses, 5.4 MMT CO2 Eq. (1.5 MMT
C) from belowground biomass C losses, 6.0 MMT CO2 Eq. (1.6 MMT C) from dead wood C losses, 8.4 MMT CO2
Eq. (2.3 MMT C) from litter C losses, 16.4 MMT CO2 Eq. (4.5 MMT C) from mineral soils and 3.5 MMT CO2 Eq.
(0.9 MMT C) from drainage and cultivation of organic soils. Emissions in 2017 are 12 percent lower than the
emissions in the initial reporting year of 1990, largely due to a reduction in the losses from Grassland Converted to
Cropland and Forest Land Converted to Cropland.
Table 6-33: 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
2013
2014
2015
2016
2017
Grassland Converted to Cropland
24.5
17.3
18.0
17.9
17.8
18.4
18.0
Mineral Soils
21.9
13.9
15.2
15.1
15.0
15.6
15.1
Organic Soils
2.5
3.3
2.9
2.8
2.8
2.8
2.8
Forest Land Converted to
Cropland
50.0
48.2
47.1
47.1
47.1
47.1
47.1
Aboveground Live Biomass
28.9
27.9
27.3
27.2
27.2
27.2
27.2
Belowground Live Biomass
5.8
5.6
5.5
5.4
5.4
5.4
5.4
Dead Wood
6.3
6.1
6.0
6.0
6.0
6.0
6.0
Litter
8.8
8.5
8.4
8.4
8.4
8.4
8.4
Mineral Soils
0.2
0.1
+
+
+
0.1
0.1
Organic Soils
0.1
+
+
+
+
+
+
Other Lands Converted to
Cropland
0.3
0.3
0.1
0.1
0.1
0.1
0.1
Mineral Soils
0.2
0.2
0.1
0.1
0.1
0.1
0.1
Organic Soils
0.1
0.1
+
+
+
+
+
Settlements Converted to
Cropland
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Mineral Soils
0.1
0.1
+
+
+
+
+
Organic Soils
+
+
0.1
0.1
0.1
0.1
0.1
Wetlands Converted to Cropland
0.7
0.8
1.6
1.6
1.7
1.6
1.6
Mineral Soils
0.1
0.1
1.2
1.2
1.2
1.1
1.1
Organic Soils
0.6
0.7
0.4
0.5
0.5
0.5
0.5
Aboveground Live Biomass
28.9
27.9
27.2
27.2
27.2
27.2
27.2
Belowground Live Biomass
5.8
5.6
5.5
5.4
5.4
5.4
5.4
Dead Wood
6.3
6.1
6.0
6.0
6.0
6.0
6.0
Litter
8.8
8.5
8.4
8.4
8.4
8.4
8.4
Total Mineral Soil Flux
22.5
14.4
16.4
16.3
16.3
16.9
16.4
Total Organic Soil Flux
3.4
4.2
3.4
3.4
3.4
3.4
3.5
Total Net Flux
75.6
66.7
66.9
66.7
66.7
67.3
66.9
+ Does not exceed 0.05 MMT CO2 Eq.
43 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.
6-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 6-34: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes in
Land Converted to Cropland (MMT C)
1990
2005
2013
2014
2015
2016
2017
Grassland Converted to Cropland
6.7
4.7
4.9
4.9
4.9
5.0
4.9
Mineral Soils
6.0
3.8
4.1
4.1
4.1
4.3
4.1
Organic Soils
0.7
0.9
0.8
0.8
0.8
0.8
0.8
Forest Land Converted to
Cropland
13.6
13.1
12.9
12.8
12.8
12.8
12.9
Aboveground Live Biomass
7.9
7.6
7.4
7.4
7.4
7.4
7.4
Belowground Live Biomass
1.6
1.5
1.5
1.5
1.5
1.5
1.5
Dead Wood
1.7
1.7
1.6
1.6
1.6
1.6
1.6
Litter
2.4
2.3
2.3
2.3
2.3
2.3
2.3
Mineral Soils
0.1
+
+
+
+
+
+
Organic Soils
+
+
+
+
+
+
+
Other Lands Converted to
Cropland
0.1
0.1
+
+
+
+
+
Mineral Soils
+
0.1
+
+
+
+
+
Organic Soils
+
+
+
+
+
+
+
Settlements Converted to
Cropland
+
+
+
+
+
+
+
Mineral Soils
+
+
+
+
+
+
+
Organic Soils
+
+
+
+
+
+
+
Wetlands Converted to Cropland
0.2
0.2
0.4
0.4
0.5
0.4
0.4
Mineral Soils
+
+
0.3
0.3
0.3
0.3
0.3
Organic Soils
0.2
0.2
0.1
0.1
0.1
0.1
0.1
Aboveground Live Biomass
7.9
7.6
7.4
7.4
7.4
7.4
7.4
Belowground Live Biomass
1.6
1.5
1.5
1.5
1.5
1.5
1.5
Dead Wood
1.7
1.7
1.6
1.6
1.6
1.6
1.6
Litter
2.4
2.3
2.3
2.3
2.3
2.3
2.3
Total Mineral Soil Flux
6.1
3.9
4.5
4.5
4.4
4.6
4.5
Total Organic Soil Flux
0.9
1.1
0.9
0.9
0.9
0.9
0.9
Total Net Flux
20.6
18.2
18.3
18.2
18.2
18.4
18.2
+ 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 for Forest Land Remaining Forest Land using data
from the USDA Forest Service, Forest Inventory and Analysis (FIA) program (USDA Forest Service 2018).
However, a default estimate is used for amount of biomass C in cropland (IPCC 2006), and litter and dead wood C
stocks were assumed to be zero since no reference C density estimates exist for croplands. The difference between
the stocks is reported as the stock change under the assumption that the change occurred in the year of the
conversion. If FIA plots include data on individual trees, aboveground and belowground C density estimates are
based on Woodall et al. (2011). Aboveground and belowground biomass estimates also include live understory
which is a minor component of biomass defined as all biomass of undergrowth plants in a forest, including woody
shrubs and trees less than 2.54 cm dbh. For this Inventory, it was assumed that 10 percent of total understory C mass
is belowground (Smith et al. 2006). Estimates of C density are based on information in Birdsey (1996) and biomass
estimates from Jenkins et al. (2003). 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
Land Use, Land-Use Change, and Forestry 6-61
-------
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
2012 USDA NRI survey for non-federal lands (USDA-NRCS 2015). 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 2012 (USDA-
NRCS 2015). NRI survey locations are classified as Land Converted to Cropland in a given year between 1990 and
2012 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 (Homer et al. 2007; Fry
etal. 2011; Homer etal. 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 2012
for mineral soils on the majority of land that is used to produce annual crops in the United States. These crops
include alfalfa hay, barley, corn, cotton, dry beans, grass hay, grass-clover hay, lentils, oats, onions, peanuts, peas,
potatoes, rice, sorghum, soybeans, sugar beets, sunflowers, tobacco, tomatoes, 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.44
For the years 2013 to 2017, 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
moving-average (ARMA) errors (Brockwell and Davis 2016) are used to estimate the relationship between surrogate
data and the 1990 to 2012 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,45 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 2013 to 2017 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 (USDA-NRCS 2015). Carbon stocks and
44 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).
45 See .
6-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
95 percent confidence intervals are estimated for each year between 1990 and 2012. See the Cropland Remaining
Cropland section for additional discussion of the Tier 3 methodology for mineral soils.
Soil C stock changes from 2013 to 2017 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. This 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 for Cropland Remaining Cropland. 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 2013 to 2017, 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 6-35 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 60
percent below to 60 percent above the 2017 stock change estimate of 66.9 MMT CO2 Eq. The large relative
uncertainty around the 2017 stock change estimate is partly due to large uncertainties in biomass and dead organic
matter C losses with Forest Land Conversion to Cropland. The large relative uncertainty is also associated with
variation in soil C stock change that is not explained by the surrogate data method, leading to high prediction error
with the splicing method.
Table 6-35: 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)
2017 Flux Estimate Uncertainty Range Relative to Flux Estimate3
(MMT CO2 Eq.) (MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Grassland Converted to Cropland 18.0 4.4 31.6 -76% 76%
Mineral Soil C Stocks: Tier 3 14.1 0.6 27.7 -96% 96%
Mineral Soil C Stocks: Tier 2 1.0 0.3 1.7 -71% 71%
Land Use, Land-Use Change, and Forestry 6-63
-------
Organic Soil C Stocks: Tier 2
2.8
1.9
3.8
-34%
34%
Forest Land Converted to Cropland
47.1
9.5
84.8
-80%
80%
Aboveground Live Biomass
27.2
-7.4
61.8
-127%
127%
Belowground Live Biomass
5.4
-1.5
12.4
-127%
127%
Dead Wood
6.0
-1.6
13.5
-127%
127%
Litter
8.4
-2.3
19.1
-127%
127%
Mineral Soil C Stocks: Tier 2
0.1
-0.4
0.5
-592%
592%
Organic Soil C Stocks: Tier 2
+
0.0
0.1
-100%
197%
Other Lands Converted to Cropland
0.1
+
0.1
-105%
104%
Mineral Soil C Stocks: Tier 2
0.1
+
0.1
-105%
104%
Organic Soil C Stocks: Tier 2
+
+
+
0%
0%
Settlements Converted to Cropland
0.1
+
0.1
-57%
57%
Mineral Soil C Stocks: Tier 2
+
+
+
-214%
210%
Organic Soil C Stocks: Tier 2
0.1
+
0.1
-56%
56%
Wetlands Converted to Croplands
1.6
0.7
2.6
-60%
60%
Mineral Soil C Stocks: Tier 2
1.1
0.2
2.0
-83%
83%
Organic Soil C Stocks: Tier 2
0.5
0.2
0.9
-67%
67%
Total: Land Converted to Cropland
66.9
26.8
106.9
-60%
60%
Aboveground Live Biomass
27.2
(7.4)
61.8
-127%
127%
Belowground Live Biomass
5.4
(1.5)
12.4
-127%
127%
Dead Wood
6.0
(1.6)
13.5
-127%
221%
Litter
8.4
(2.3)
19.1
-127%
127%
Mineral Soil C Stocks: Tier 3
14.1
0.6
27.7
-96%
96%
Mineral Soil C Stocks: Tier 2
2.3
1.0
3.5
-55%
55%
Organic Soil C Stocks: Tier 2
3.5
2.4
4.5
-30%
30%
+ Does not exceed 0.05 MMT CO2 Eq.
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or net sequestration.
Methodological recalculations are applied from 2013 to 2016 using the surrogate data method developed for soil C
stock change and from 1990 to 2016 for biomass and dead organic matter C estimates, ensuring consistency across
the time series. Details on the emission trends through time are described in more detail in the Methodology section.
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 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 analysis.
QA/QC and Verification
See the QA/QC and Verification section in Cropland Remaining Cropland for information on QA/QC steps.
Recalculations Discussion
Methodological recalculations are associated with extending the time series from 2013 through 2016 for mineral and
organic soils using a surrogate data method, and from 1990 to 2016 for biomass and dead organic matter C
associated with Forest Land Converted to Cropland. No other recalculations have been implemented in the current
Inventory. Carbon stock change losses increased by an average of 141 percent from 1990 through 2016 based on the
recalculation. This change is almost entirely attributed to the update of biomass and dead organic matter losses for
Forest Land Converted to Cropland with newly available re-measurement data for the western United States. Stock
changes were re-estimated at the plot-level with the new data consistent with the compilation methods described for
Forest Land Remaining Forest Land. In the previous Inventory, state-level averages from the plot data had been
used to approximate the losses of C with Forest Land Converted to Cropland due to a lack of re-measurement data.
6-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Planned Improvements
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. Additional planned improvements
are discussed in the Cropland Remaining Cropland section.
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, but there is currently no reporting of
C stock changes for aboveground and belowground biomass, dead wood and litter pools.46 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.47
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 2015). 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. 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 Section 6.1 Representation of the U.S.
Land Base) and the grassland area included in the Inventory analysis (CRF Category 4C1—Section 6.6)48.
In Grassland Remaining Grassland, there has been considerable variation in soil C stocks between 1990 and 2017.
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. Land use and management generally increased soil C in mineral soils for Grassland
Remaining Grassland between 1990 and 2017. In contrast, organic soils lose a relatively constant amount of C
annually from 1990 through 2017. In 2017, soil C stocks are a net sink, sequestering 0.1 MMT CO2 Eq. (0.0 MMT
C), with an increase of 5.6 MMT CO2 Eq. (1.5 MMT C) in mineral soils, and a loss of 5.6 MMT CO2 Eq. (1.5 MMT
C) from organic soils (Table 6-36 and Table 6-37). Soil C stock changes are 99 percent lower in 2017 compared to
1990, but stock changes are highly variable from 1990 to 2017, with an average annual sequestration of 5.0 MMT
46 There are planned improvements to address all C pools in the future, with an initial effort focused on biomass C.
47 CO2 emissions associated with liming and urea fertilization are also estimated but included in the Agriculture chapter of the
report.
48 For the U. S. land representation, land use data for 2013 to 2017 were only partially updated based on new Forest Inventory
and Analysis (FIA) data. These updates led to changes in the land representation data for grasslands through the process of
combining FIA data with land use data from the National Resources Inventory and National Land Cover Dataset (See
"Representation of the U.S. Land Base" section for more information). However, an inventory was not compiled for grasslands
with the new land representation data so the area estimates in this section are based on the land representation data from the
previous Inventory. This has created additional discrepancies with the reported grassland areas in the "Representation of the U.S.
Land Base" section.
Land Use, Land-Use Change, and Forestry 6-65
-------
CO: Eq. (1.4 MMT C). However, the large inter-annual variability leads to years in which Grassland Remaining
Grassland is a net sink and others in which it is a net source of CO2 emissions.
Table 6-36: Net CO2 Flux from Soil C Stock Changes in Grassland Remaining Grassland (MMT
COz Eq.)
Soil Type 1990
2005
2013 2014 2015 2016 2017
Mineral Soils (11.4)
Organic Soils 7.2
(0.5)
6.0
(9.3) (13.1) 4.1 (7.2) (5.6)
5.5 5.5 5.5 5.5 5.6
Total Net Flux (4.2)
5.5
(3.7) (7.5) 9.6 (1.6) (0.1)
Note: Parentheses indicate net sequestration.
Table 6-37: Net CO2 Flux from Soil C Stock Changes in Grassland Remaining Grassland (MMT
C)
Soil Type 1990
2005
2013 2014 2015 2016 2017
Mineral Soils (3.1)
Organic Soils 2.0
(0.1)
1.6
(2.5) (3.6) 1.1 (2.0) (1.5)
1.5 1.5 1.5 1.5 1.5
Total Net Flux (1.1)
1.5
(1.0) (2.1) 2.6 (0.4) +
+ Absolute value does not exceed 0.05 MMT C
Note: Parentheses indicate net sequestration.
The spatial variability in the 2012 annual soil C stock changes49 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 throughout the country, with a larger concentration of grasslands sequestering soil C in Iowa. 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.
49 Only national-scale emissions are estimated for 2013 to 2017 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 2012.
6-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 6-7: Total Net Annual Soil C Stock Changes for Mineral Soils under Agricultural
Management within States, 2012, Grassland Remaining Grassland
¦ > 40
Note: Only national-scale soil C stock changes are estimated for 2013 to 2017 in the current Inventory using a
surrogate data method, and therefore the line-scale emission patterns in this map are based on inventory data from
2012. Negative values represent a net increase in soil C stocks, and positive values represent a net decrease in soil
C stocks.
Figure 6-8: Total Net Annual Soil C Stock Changes for Organic Soils under Agricultural
Management within States, 2012, Grassland Remaining Grassland
Land Use, Land-Use Change, and Forestry 6-67
-------
Note: Only national-scale soil carbon stock changes are estimated for 2013 to 2017 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 2012.
Methodology
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 2012 USDANRI survey (USDA-NRCS 2015). 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 2012 (USDA-NRCS 2015). NRI survey locations are classified as Grassland Remaining
Grassland in a given year between 1990 and 2012 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 (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 2012
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., sewage sludge) amendments. SOC
stock changes on the remaining 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.50
A surrogate data method is used to estimate soil C stock changes from 2013 to 2017 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 2012 emissions data from the Tier 2 and 3 methods. Surrogate data for these regression models
includes 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
2013 to 2017 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 biogeochemical51 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
2012 USDA NRI survey (USDA-NRCS 2015). Frequency and rates of manure application to grassland during 1997
are estimated from data compiled by the USDA Natural Resources Conservation Service (NRCS) (Edmonds, et al.
2003), and then adjusted using county-level estimates of manure available for application in other years.
Specifically, county-scale ratios of manure available for application to soils in other years relative to 1997 are used
50 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).
51 Biogeochemical cycles are the flow of chemical elements and compounds between living organisms and the physical
environment.
6-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
to adjust the area amended with manure (see Cropland Remaining Cropland section and Annex 3.12 for further
details). Greater availability of managed manure nitrogen (N) relative to 1997 is assumed to increase the area
amended with manure, while reduced availability of manure N relative to 1997 is assumed to reduce the amended
area.
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
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 2012 using the NRI survey data.
Soil C stock changes from 2013 to 2017 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 2015) 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) (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. As with the non-federal lands, the time series
for federal lands has been extended from 2013 to 2017 using a surrogate data method described in Box 6-4 of the
Methodology Section in Cropland Remaining Cropland. Further elaboration on the Tier 2 methodology and data
used to estimate C stock changes from mineral soils are described in Annex 3.12.
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 2017 to
account for additional C stock changes associated withbiosolid (i.e., sewage sludge) amendments. Estimates of the
amounts of biosolids N applied to agricultural land are derived from national data onbiosolids generation,
disposition, and N content (see Section 7.2, Wastewater Treatment for a detailed discussion of the methodology for
estimating 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. N 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. A surrogate data method is used to estimate annual C emissions from organic soils from 2013 to 2017 as
described in Box 6-4 of the Methodology section in Cropland Remaining Cropland. Estimates for 2013 to 2017 will
be updated in future Inventories when new NRI data are available. For more information, see the Cropland
Remaining Cropland section for organic soils.
Land Use, Land-Use Change, and Forestry 6-69
-------
Uncertainty and Time-Series Consistency
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. 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 2013 to 2017, there is
additional uncertainty propagated through the Monte Carlo Analysis associated with the surrogate data method.
Uncertainty estimates are presented in Table 6-38 for each subsource (i.e., 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 1,000 percent below and
above the 2017 stock change estimate of -0.1 MMT CO2 Eq. The large relative uncertainty is primarily due to the
small estimated change in soil C stocks, which is almost zero for 2017.
Table 6-38: Approach 2 Quantitative Uncertainty Estimates for C Stock Changes Occurring
Within Grassland Remaining Grassland (MMT CO2 Eq. and Percent)
2017 Flux Estimate Uncertainty Range Relative to Flux Estimate3
(MMT CO2 Eq.) (MMT CO2 Eq.) (%)
Lower
Upper
Lower
Upper
Bound
Bound
Bound
Bound
Mineral Soil C Stocks Grassland Remaining
Grassland, Tier 3 Methodology
(4.0)
(44.4)
36.5
-1,016%
1,016%
Mineral Soil C Stocks: Grassland Remaining
Grassland, Tier 2 Methodology
(1.5)
(2.9)
+
-100%
100%
Mineral Soil C Stocks: Grassland Remaining
Grassland, Tier 2 Methodology (Change in
Soil C due to Biosolids [i.e., Sewage Sludge]
(0.2)
(0.2)
(0.1)
-50%
50%
Amendments)
Organic Soil C Stocks: Grassland Remaining
Grassland, Tier 2 Methodology
5.6
5.0
6.1
-10%
10%
Combined Uncertainty for Flux Associated
with Agricultural Soil Carbon Stock (0.1) (40.5) 40.4 -74,245% 74,245%
Change in Grassland Remaining Grassland
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
+ Does not exceed 0.05 MMT CO2 Eq.
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or net sequestration.
Details on the emission trends through time are described in more detail in the Methodology section.
Uncertainty is also associated with a lack of reporting on biomass and litter C stock changes. Biomass C stock
changes may be significant for managed grasslands with woody encroachment despite not having attained enough
tree cover to be considered forest lands. Changes in dead organic matter C stocks are assumed to be negligible in
grasslands on an annual basis, although there are certainly significant changes at sub-annual time scales across
seasons.
QA/QC and Verification
See the QA/QC and Verification section in Cropland Remaining Cropland.
Planned Improvements
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. Another key planned
6-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
improvement is to estimate woody biomass C stock changes for grasslands (See Box 6-6). For information about
other improvements, see the Planned Improvements section in Cropland Remaining Cropland.
Box 6-6: Grassland Woody Biomass Analysis
An initial analysis of woodland biomass lias been conducted for regions in the western United States. Woodlands are
areas with trees in a matrix of grass vegetation that do not reach the thresholds for tree cover, diameter at breast
height, and/or tree height to be considered forest land. For this pilot effort, carbon stock densities and stock changes
are estimated using woodland plots in the Forest Inventory and Analysis (FIA) database. The full set of woodland
plots cover 12 states in the western United States, and include two FIA forest type groups, pinyon-juniper and
woodland hardwoods. The results suggest that woodlands are sequestering approximately 20 MMT CO2 Eq. in
biomass, dead wood, and litter pools. The analysis will be expanded to the entire time series and reported in a future
Inventory.
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 (Daubemnire 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 currently the focus is
primarily on herbaceous biomass in this section.52 Biomass burning emits a variety of trace gases including non-
CO2 greenhouse gases, CH4 and N20, as well as CO and NOx that can become greenhouse gases when they react
with other gases in the atmosphere (Andreae and Merlet 2001). IPCC (2006) recommends reporting no 11-CO:
greenhouse gas emissions from all wildfires and prescribed burning occurring in managed grasslands.
Biomass burning in grassland of the United States is a relatively small source of emissions, but it has increased by
over 300 percent since 1990. In 2017, CH4 and N2O emissions from biomass burning in grasslands were 0.3 MMT
CO2 Eq. (12 kt) and 0.3 MMT CO2 Eq. (1 kt), respectively. Annual emissions from 1990 to 2017 have averaged
approximately 0.3 MMT CO2 Eq. (12 kt) of CH4 and 0.3 MMT CO2 Eq. (1 kt) of N20 (see Table 6-39 and Table
6-40).
Table 6-39: ChU and N2O Emissions from Biomass Burning in Grassland (MMT CO2 Eq.)
1990
2005
2013 2014 2015 2016 2017
ch4 0.1
N2O 0.1
O O
u> u>
0.2 0.4 0.3 0.3 0.3
0.2 0.4 0.3 0.3 0.3
Total Net Flux 0.2
0.7
0.4 0.8 0.7 0.6 0.6
Note: Totals may not sum due to independent rounding.
Table 6-40: ChU, N2O, CO, and NOx Emissions from Biomass Burning in Grassland (kt)
1990
2005
2013
2014
2015
2016
2017
ch4
3
13
8
16
13
11
12
n2o
+
1
1
1
1
1
1
CO
84
358
217
442
356
324
345
NOx
5
21
13
27
21
19
21
+ Does not exceed 0.5 kt
52 A planned improvement is underway to incorporate woodland tree biomass into the Inventory.
Land Use, Land-Use Change, and Forestry 6-71
-------
Methodology
The following section includes a description of the methodology used to estimate non-CO; 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-CO:
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 2017, 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 lias 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
20 1 4.53 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 6-41).
Table 6-41: Thousands of Grassland Hectares Burned Annually
Thousand
Year Hectares
1990 317
2005 1,343
2013
815
2014
1,659
2015
NE
2016
NE
2017
NE
Notes: Burned area are not
estimated (NE) for 2015 to 2017
but will be updated in a future
Inventory.
For 1990 to 2014, the total area of grassland burned is multiplied by the IPCC default factor for grassland biomass
(4.1 tonnes dry matter per ha) (IPCC 2006) to estimate the amount of combusted biomass. A combustion factor of 1
is assumed in this Inventory, and the resulting biomass estimate is multiplied by the IPCC default grassland
emission factors for CH4 (2.3 g CH4 per kg dry matter), N20 (0.21 g CH4 per kg dry matter), CO (65 g CH4 per kg
dry matter) and NOx (3.9 g CH4 per kg dry matter) (IPCC 2006). The Tier 1 analysis is implemented in the
Agriculture and Land Use National Greenhouse Gas Inventory (ALU) software (Ogle et al. 2016).54
A linear extrapolation of the trend in the time series is applied to estimate the emissions for 2015 to 2017 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 2017 emissions. The Tier 1
method described previously will be applied to recalculate the 2015 to 2017 emissions in a future Inventory.
53 See .
54 See .
6-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Uncertainty and Time-Series Consistency
Emissions are estimated using a linear regression model with ARMA errors for 2015 to 2017. The linear regression
ARMA model produced estimates of the upper and lower bounds of the emission estimate and the results are
summarized in Table 6-42. Methane emissions from Biomass Burning in Grassland for 2017 are estimated to be
between 0.0 and 0.7 MMT CO2 Eq. at a 95 percent confidence level. This indicates a range of 100 percent below
and 139 percent above the 2017 emission estimate of 0.3 MMT CO2 Eq. Nitrous oxide emissions are estimated to be
between 0.0 and 0.8 MMT CO2 Eq., or approximately 100 percent below and 140 percent above the 2017 emission
estimate of 0.3 MMT CO2 Eq.
Table 6-42: Uncertainty Estimates for Non-C02 Greenhouse Gas Emissions from Biomass
Burning in Grassland (MMT CO2 Eq. and Percent)
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate3
Source
Gas
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Bound Bound
Lower Upper
Bound Bound
Grassland Burning
Grassland Burning
CH4
N2O
0.3
0.3
0.0 0.7
0.0 0.8
-100% 139%
-100% 140%
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.
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.
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 for 2015 to 2017.
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-CCh greenhouse gas emissions from grassland
burning.
Land Use, Land-Use Change, and Forestry 6-73
-------
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 2015).55 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 Section 6.1 Representation of the U.S. Land Base) and the grassland area included in the inventory analysis
(CRF Category 4C2—Section 6.7)56.
Land use change can lead to large losses of C to the atmosphere, particularly conversions from forest land
(Houghton et al. 1983). Moreover, conversion of forest to another land use (i.e., deforestation) is one of the largest
anthropogenic sources of emissions to the atmosphere globally (Schimel 1995), although this source may be
declining 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
The largest C losses with Land Converted to Grassland are associated with aboveground biomass, belowground
biomass, dead wood and litter C losses from Forest Land Converted to Grassland (see Table 6-43 and Table 6-44).
These four pools led to net emissions in 2017 of 12.6, 2.4, -0.9, and 4.8 MMT CO2 Eq. (3.4,0.7, -0.3, and 1.3MMT
C), respectively. Land use and management of mineral soils in Land Converted to Grassland led to an increase in
soil C stocks, estimated at 12.2 MMT CO2 Eq. (3.3 MMT C) in 2017, while drainage of organic soils for grassland
management led to CO2 emissions to the atmosphere of 1.6 MMT CO2 Eq. (0.4 MMT C). The total net C stock
change in 2017 for Land Converted to Grassland is estimated as a loss of 8.3 MMT CO2 Eq. (2.3 MMT C), which is
a 3 percent decrease in emissions compared to the initial reporting year of 1990.
Table 6-43: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Grassland (MMT CO2 Eq.)
1990 2005 2013 2014 2015 2016 2017
Cropland Converted to Grassland (7.5) (11.5) (8.1) (8.4) (6.2) (7.5) (7.6)
55 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.
56 For the U.S. land representation, land use data for 2013 to 2017 were only partially updated based on new Forest Inventory
and Analysis (FIA) data. These updates led to changes in the land representation data for grasslands through the process of
combining FIA data with land use data from the National Resources Inventory and National Land Cover Dataset (See
"Representation of the U.S. Land Base" section for more information). However, an inventory was not compiled for grasslands
with the new land representation data so the area estimates in this section are based on the land representation data from the
previous Inventory. This has created additional discrepancies with the reported grassland areas in the "Representation of the U.S.
Land Base" section.
57 Changes in biomass C stocks are not currently reported for other conversions to grassland (other than forest land), but this is a
planned improvement for a future Inventory. Note: changes in dead organic matter are assumed to negligible for other land use
conversions (i.e., other than forest land) to grassland based on the Tier 1 method in IPCC (2006).
6-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Mineral Soils
(8.0)
(12.7)
(9.3)
(9.5)
(7.4)
(8.6)
(8.7)
Organic Soils
0.5
1.1
1.1
1.1
1.1
1.1
1.1
Forest Land Converted to
Grassland
17.0
18.0
16.3
16.1
15.9
15.9
15.8
Abovegroimd Live Biomass
12.6
12.6
12.6
12.6
12.6
12.6
12.6
Belowground Live Biomass
2.5
2.5
2.4
2.4
2.4
2.4
2.4
Dead Wood
(1.6)
(1.3)
(1.0)
(0.9)
(0.9)
(0.9)
(0.9)
Litter
4.3
4.6
4.8
4.8
4.8
4.8
4.8
Mineral Soils
(0.8)
(0.5)
(2.7)
(2.9)
(3.1)
(3.1)
(3.2)
Organic Soils
+
0.1
0.1
0.1
0.1
0.1
0.1
Other Lands Converted Grassland
(0.5)
(1.0)
+
(0.1)
(0.1)
(0.1)
(0.1)
Mineral Soils
(0.5)
(1.1)
(0.1)
(0.1)
(0.1)
(0.1)
(0.1)
Organic Soils
+
+
+
+
+
+
+
Settlements Converted Grassland
(0.1)
(0.1)
+
+
+
+
+
Mineral Soils
(0.1)
(0.1)
+
+
+
+
+
Organic Soils
+
+
+
+
+
+
+
Wetlands Converted Grassland
(0.2)
(0.2)
0.2
0.2
0.1
0.1
0.1
Mineral Soils
(0.3)
(0.4)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
Organic Soils
0.1
0.3
0.3
0.3
0.3
0.3
0.3
Aboveground Live Biomass
12.6
12.6
12.6
12.6
12.6
12.6
12.6
Belowground Live Biomass
2.5
2.5
2.4
2.4
2.4
2.4
2.4
Dead Wood
(1.6)
(1.3)
(1.0)
(0.9)
(0.9)
(0.9)
(0.9)
Litter
4.3
4.6
4.8
4.8
4.8
4.8
4.8
Total Mineral Soil Flux
(9.7)
(14.8)
(12.3)
(12.6)
(10.8)
(12.0)
(12.2)
Total Organic Soil Flux
0.7
1.5
1.7
1.6
1.7
1.6
1.6
Total Net Flux
8.7
5.1
8.3
7.9
9.8
8.5
8.3
+ Does not exceed 0.05 MMT CO2 Eq.
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or net sequestration.
Table 6-44: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Grassland (MMT C)
1990
2005
2013
2014
2015
2016
2017
Cropland Converted to Grassland
(2.0)
(3.1)
(2.2)
(2.3)
(1.7)
(2.0)
(2.1)
Mineral Soils
(2.2)
(3.5)
(2.5)
(2.6)
(2.0)
(2.3)
(2.4)
Organic Soils
0.1
0.3
0.3
0.3
0.3
0.3
0.3
Forest Land Converted to Grassland
4.6
4.9
4.4
4.4
4.3
4.3
4.3
Aboveground Live Biomass
3.4
3.4
3.4
3.4
3.4
3.4
3.4
Belowground Live Biomass
0.7
0.7
0.7
0.7
0.7
0.7
0.7
Dead Wood
(0.4)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Litter
1.2
1.3
1.3
1.3
1.3
1.3
1.3
Mineral Soils
(0.2)
(0.1)
(0.7)
(0.8)
(0.8)
(0.9)
(0.9)
Organic Soils
+
+
+
+
+
+
+
Other Lands Converted Grassland
(0.1)
(0.3)
+
+
+
+
+
Mineral Soils
(0.1)
(0.3)
+
+
+
+
+
Organic Soils
+
+
+
+
+
+
+
Settlements Converted Grassland
+
+
+
+
+
+
+
Mineral Soils
+
+
+
+
+
+
+
Organic Soils
+
+
+
+
+
+
+
Wetlands Converted Grassland
(0.1)
+
+
+
+
+
+
Mineral Soils
(0.1)
(0.1)
+
+
(0.1)
(0.1)
(0.1)
Organic Soils
+
0.1
0.1
0.1
0.1
0.1
0.1
Aboveground Live Biomass
3.4
3.4
3.4
3.4
3.4
3.4
3.4
Belowground Live Biomass
0.7
0.7
0.7
0.7
0.7
0.7
0.7
Dead Wood
(0.4)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
(0.3)
Litter
1.2
1.3
1.3
1.3
1.3
1.3
1.3
Total Mineral Soil Flux
(2.6)
(4.0)
(3.3)
(3.4)
(2.9)
(3.3)
(3.3)
Total Organic Soil Flux
0.2
0.4
0.5
0.4
0.5
0.4
0.4
Total Net Flux
2.4
1.4
2.3
2.2
2.7
2.3
2.3
+ Does not exceed 0.05 MMT C. Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values
or net sequestration.
Land Use, Land-Use Change, and Forestry 6-75
-------
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 2 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 USD A Forest Service, Forest Inventory and Analysis (FIA) program (USD A Forest Service
2018). 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 re-measurements 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, which 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. 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 grassland. This estimate was used to develop state-specific carbon density estimates for biomass, 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. In addition, 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). In the Eastern United States, there is limited data on grassland carbon stocks following
conversion to grassland so only 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 grassland in the Eastern United States).
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).
Estimates are also derived for changes in dead organic matter with Forest Land Converted to Grassland. 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
6-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
2012 USDA NRI survey for non-federal lands (USDA-NRCS 2015). 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 2012 (USDA-NRCS 2015). NRI survey locations are classified as Land Converted to Grassland
in a given year between 1990 and 2012 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 (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 2013 to 2017 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 2012 emissions data that are derived using the Tier 2 and 3 methods. Surrogate data for these
regression models include 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 2013 to 2017 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 2012 USDA NRI survey (USDA-NRCS 2015). C stocks and 95 percent confidence
intervals are estimated for each year between 1990 and 2012. 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 2013 to 2017 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. 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.
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 6-77
-------
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 for organic soils. A surrogate data method is used to estimate annual C emissions from
organic soils from 2013 to 2017 as described in Box 6-4 of the Methodology section in Cropland Remaining
Cropland. Estimates for 2013 to 2017 will be recalculated in future Inventories when new NRI data are available.
Uncertainty and Time-Series Consistency
The uncertainty analysis for biomass, dead wood and litter C losses with Forest Land Converted to Grassland is
conducted in the same way as the uncertainty assessment for forest ecosystem C flux in the Forest Land Remaining
Forest Land category. Sample and model-based error are combined using simple error propagation methods
provided by the IPCC (2006), by taking the square root of the sum of the squares of the standard deviations of the
uncertain quantities. For additional details see the Uncertainty Analysis in Annex 3.13. The uncertainty analyses for
mineral soil 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. 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 2013 to 2017, 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 6-45 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
the IPCC (2006), as discussed in the previous paragraph. The combined uncertainty for total C stocks in Land
Converted to Grassland ranges from 214 percent below to 214 percent above the 2017 stock change estimate of 8.3
MMT CO2 Eq. The large relative uncertainty around the 2017 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 associated with variation in soil C stock change that is not explained by the surrogate data
method, leading to high prediction error with the splicing method.
Table 6-45: 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)
2017 Flux Estimate3 Uncertainty Range Relative to Flux Estimate3
(MMT CO2 Eq.) (MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Cropland Converted to Grassland
(7.6)
(16.3)
1.2
-116%
116%
Mineral Soil C Stocks: Tier 3
(8.6)
(17.4)
0.1
-102%
102%
Mineral Soil C Stocks: Tier 2
(0.1)
(0.3)
0.1
-333%
333%
Organic Soil C Stocks: Tier 2
1.1
0.8
1.4
-29%
29%
Forest Land Converted to Grassland
15.8
0.3
31.4
-98%
98%
Aboveground Live Biomass
12.6
(0.6)
25.7
-104%
104%
Belowground Live Biomass
2.4
(0.1)
4.9
-105%
104%
Dead Wood
(0.9)
(6.6)
4.7
-597%
599%
Litter
4.8
(0.2)
9.9
-105%
104%
Mineral Soil C Stocks: Tier 2
(3.2)
(5.4)
(1.0)
-70%
70%
Organic Soil C Stocks: Tier 2
0.1
0.1
0.2
-43%
43%
Other Lands Converted to Grassland
(0.1)
(0.3)
0.1
-250%
251%
Mineral Soil C Stocks: Tier 2
(0.1)
(0.3)
0.1
-160%
160%
Organic Soil C Stocks: Tier 2
+
+
0.1
-37%
37%
Settlements Converted to Grassland
+
+
+
-79%
79%
Mineral Soil C Stocks: Tier 2
+
+
+
-575%
550%
Organic Soil C Stocks: Tier 2
+
+
+
-51%
53%
Wetlands Converted to Grasslands
0.1
(0.1)
0.3
-174%
173%
6-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Mineral Soil C Stocks: Tier 2
(0.2)
(0.4)
+
-83%
83%
Organic Soil C Stocks: Tier 2
0.3
0.2
0.4
-42%
42%
Total: Land Converted to Grassland
8.3
(9.5)
26.2
-214%
214%
Aboveground Live Biomass
12.6
(0.6)
25.7
-104%
104%
Belowground Live Biomass
2.4
(0.1)
4.9
-105%
104%
Dead Wood
(0.9)
(6.6)
4.7
-597%
599%
Litter
4.8
(0.2)
9.9
-105%
104%
Mineral Soil C Stocks: Tier 3
(8.6)
(17.4)
0.1
-102%
102%
Mineral Soil C Stocks: Tier 2
(3.6)
(5.8)
(1.3)
-63%
63%
Organic Soil C Stocks: Tier 2
1.6
1.3
2.0
-22%
22%
+ Absolute value does not exceed 0.05 MMT CO2 Eq.
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Notes: Totals may not sum due to independent rounding. Parentheses indicate negative values or net sequestration.
Methodological recalculations are applied from 2013 to 2017 using the surrogate data method developed using the C
stock change estimates from 1990 to 2012, ensuring consistency across the time series. Details on the emission
trends through time are described in more detail in the introductory section above.
Uncertainty is also associated with a lack of reporting on biomass and dead organic matter C stock changes for Land
Converted to Grassland with the exception of forest land conversion. Biomass C stock changes may be significant
for managed grasslands with woody encroachment despite not having attained enough tree cover to be considered
forest lands. Changes in dead organic matter C stocks are assumed to be negligible with conversion of land to
grasslands with the exception of forest lands, which are included in this analysis. This assumption will be further
explored in a future Inventory.
QA/QC and Verification
See the QA/QC and Verification section in Cropland Remaining Cropland fox information on QA/QC steps.
Recalculations Discussion
Methodological recalculations are associated with extending the time series from 2013 through 2016 for mineral and
organic soils using a surrogate data method, and from 1990 to 2016 for biomass and dead organic matter C
associated with Forest Land Converted to Grassland. No other recalculations have been implemented in the current
Inventory. C stock change losses decreased by an average of 67 percent from 1990 through 2016 based on the
recalculation. This change is almost entirely attributed to the update of biomass and dead organic matter losses for
Forest Land Converted to Grassland with newly available re-measurement data for the western United States. Stock
changes were re-estimated at the plot-level with the new data consistent with the compilation methods described for
Forest Land Remaining Forest Land. In the previous Inventory, state-level averages from the plot data had been
used to approximate the losses of C with Forest Land Converted to Grassland due to a lack of re-measurement data.
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. One additional planned improvement for the Land Converted to Grassland category is
Land Use, Land-Use Change, and Forestry 6-79
-------
to develop an inventory of C stock changes for grasslands in Alaska. For information about other improvements, see
the Planned Improvements section in Cropland Remaining Cropland and Grassland Remaining Grassland.
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 includes Peatlands and Coastal Wetlands.
Peatlands Remaining Peatlands
Emissions from Managed Peatlands
Managed peatlands are peatlands that have been cleared and drained for the production of peat. The production
cycle of a managed peatland has three phases: land conversion in preparation for peat extraction (e.g., clearing
surface biomass, draining), extraction (which results in the emissions reported under Peatlands Remaining
Peatlands), and abandonment, restoration, or conversion of the land to another use.
Carbon dioxide emissions from the removal of biomass and the decay of drained peat constitute the major
greenhouse gas flux from managed peatlands. Managed peatlands may also emit CH4 and N20. The natural
production of CH4 is largely reduced but not entirely shut down when peatlands are drained in preparation for peat
extraction (Strack et al. 2004 as cited in the 2006IPCC Guidelines). Drained land surface and ditch networks
contribute to the CH4 flux in peatlands managed for peat extraction. Methane emissions were considered
insignificant under the IPCC Tier 1 methodology (IPCC 2006), but are included in the emissions estimates for
Peatlands Remaining Peatlands consistent with the 2013 Supplement to the 2006 IPCC Guidelines for National
Greenhouse Gas Inventories: Wetlands (IPCC 2013). Nitrous oxide emissions from managed peatlands depend on
site fertility. In addition, abandoned and restored peatlands continue to release greenhouse gas emissions. Although
methodologies are provided for rewetted organic soils (which includes rewetted/restored peatlands) in IPCC (2013)
guidelines, information on the areal extent of rewetted/restored peatlands in the United States is currently
unavailable. The current Inventory estimates CO2, CH4 and N20 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 (2014) recommends reporting CO2, N20, and CH4 emissions from lands undergoing active peat extraction
(i.e., Peatlands Remaining Peatlands) as part of the estimate for emissions from managed wetlands. Peatlands occur
where plant biomass has sunk to the bottom of water bodies and water-logged areas and exhausted the oxygen
supply below the water surface during the course of decay. Due to these anaerobic conditions, much of the plant
matter does not decompose but instead forms layers of peat over decades and centuries. In the United States, peat is
extracted for horticulture and landscaping growing media, and for a wide variety of industrial, personal care, and
other products. It has not been used for fuel in the United States for many decades. Peat is harvested from two types
of peat deposits in the United States: sphagnum bogs in northern states (e.g., Minnesota) and wetlands in states
further south (e.g., Florida). The peat from sphagnum bogs in northern states, which is nutrient-poor, is generally
corrected for acidity and mixed with fertilizer. Production from more southerly states is relatively coarse (i.e.,
fibrous) but nutrient-rich.
IPCC (2006 and 2014) recommend considering both on-site and off-site emissions when estimating CO2 emissions
from Peatlands Remaining Peatlands using the Tier 1 approach. Current methodologies estimate only on-site N2O
and CH4 emissions, since off-site N20 estimates are complicated by the risk of double-counting emissions from
nitrogen fertilizers added to horticultural peat, and off-site CH4 emissions are not relevant given the non-energy uses
of peat, so methodologies are not provided in IPCC (2014) guidelines.
6-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 N20 emissions from saturated ecosystems tend to be low unless there is an exogenous source of
nitrogen, N20 emissions from drained peatlands are dependent on nitrogen mineralization and therefore on soil
fertility. Peatlands located on highly fertile soils contain significant amounts of organic nitrogen in inactive form.
Draining land in preparation for peat extraction allows bacteria to convert the nitrogen into nitrates which leach to
the surface where they are reduced to N20, and contributes to the activity of methanogens and methanotrophs that
result in CH4 emissions (Blodau 2002; Treat et al. 2007 as cited in IPCC 2014). 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 2014).
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 2014). 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 2017 (see Table
6-46) comprising 0.7 MMT C02 Eq. (734 kt) of C02, 0.004 MMT C02 Eq. (0.15 kt) of CH4 and 0.0005 MMT C02
Eq. (0.002 kt) of N20. Total emissions in 2017 were about 0.11 percent greater than total emissions in2016.
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 2017. 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 N20 emissions remained close to zero across the
time series. Nitrous oxide emissions showed a decreasing trend from 1990 until 1995, followed by an increasing
trend through 2001. Nitrous oxide emissions decreased between 2001 and 2006, followed by a leveling off between
2008 and 2010, and a general decline between 2011 and 2017. 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 2017.
Table 6-46: Emissions from Peatlands Remaining Peatlands (MMT CO2 Eq.)
Gas
1990
2005
2013
2014
2015
2016
2017
CO2
1.1
1.1
0.8
0.8
0.8
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
+
+
+
CH4 (On-site)
+
+
+
+
+
+
+
N2O (On-site)
+
+
+
+
+
+
+
Total
1.1
1.1
0.8
0.8
0.8
0.7
0.7
+ Does not exceed 0.05 MMT CO2 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 N2O
emissions are not estimated to avoid double-counting N2O 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 6-47: Emissions from Peatlands Remaining Peatlands (kt)
Gas
1990
2005
2013
2014
2015
2016
2017
CO2
1,055
1,101
770
775
755
733
734
Off-site
985
1,030
720
725
706
686
687
On-site
70
71
50
50
49
47
47
CH4 (On-site)
+
+
+
+
+
+
+
N2O (On-site)
+
+
+
+
+
+
+
Land Use, Land-Use Change, and Forestry 6-81
-------
+ 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 N2O
emissions are not estimated to avoid double-counting N2O 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.
Methodology
Off-Site CO2 Emissions
Carbon dioxide emissions from domestic peat production were estimated using a Tier 1 methodology consistent with
IPCC (2006). Off-site CO2 emissions from Peatlands Remaining Peat lands were calculated by apportioning the
annual weight of peat produced in the United States (Table 6-48) 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 1991
through 2016; USGS 2018). 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 annual Alaska's Mineral Industry reports (DGGS
1993 through 2015). Similar to the U.S. Geological Survey, the Alaska Department of Natural Resources, Division
of Geological & Geophysical Surveys (DGGS) solicits voluntary reporting of peat production from producers for the
Alaska's Mineral Industry report. However, the report does not estimate production for the non-reporting producers,
resulting in larger inter-annual variation in reported peat production from Alaska depending on the number of
producers who report in a given year (Szumigala 2011). In addition, in both the lower 48 states and Alaska, large
variations in peat production can also result from variations in precipitation and the subsequent changes in moisture
conditions, since unusually wet years can hamper peat production. The methodology estimates Alaska emissions
separately from lower 48 emissions because the state conducts its own mineral survey and reports peat production
by volume, rather than by weight (Table 6-49). 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).59 Peat production was not reported for 2015 i 11.11 a ska's Mineral Industry 2014 report (DGGS
2015); and reliable data are not available beyond 2012, so Alaska's peat production in 2013 through 2017 (reported
in cubic yards) was assumed to be equal to the 2012 value.
Consistent with IPCC (2014) 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. Carbon dioxide emissions from dissolved
organic C were estimated by multiplying the area of peatlands by the default emission factor for dissolved organic C
provided in IPCC (2014).
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
59 Peat produced from Alaska was assumed to be nutrient poor; as is the case in Canada, "where deposits of high-quality [but
nutrient poor] sphagnum moss are extensive" (USGS 2008).
6-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 6-48: Peat Production of Lower 48 States (kt)
Type of Deposit 1990
2005
2013 2014 2015 2016 2017
Nutrient-Rich 595.1
Nutrient-Poor 55.4
657.6
27.4
418.5 416.5 405.0 388.1 374.0
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
Sources: United States Geological Survey (USGS) (1991-2016)Minerals Yearbook: Peat (1994-2016);
United States Geological Survey (USGS) (2018) Mineral Commodity Summaries: Peat (2018).
Table 6-49: 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
Sources: Division of Geological & Geophysical Surveys (DGGS), Alaska Department of Natural Resources
(1997-2015).4faifa? 's Mineral Industry Report (1997-2014).
On-site CO 2 Emissions
IPCC (2006) suggests basing the calculation of on-site emission estimates on the area of peatlands managed for peat
extraction differentiated by the nutrient type of the deposit (rich versus poor). Information on the area of land
managed for peat extraction is currently not available for the United States, but consistent with IPCC (2006), an
average production rate for the industry was applied to derive an area estimate. In a mature industrialized peat
industry, such as exists in the United States and Canada, the vacuum method can extract up to 100 metric tons per
hectare per year (Cleary et al. 2005 as cited in IPCC 2006).60 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. The annual land area
estimates were then multiplied by the IPCC (2014) 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.
The estimated areas of land managed for peat extraction are presented in Table 6-50. The total area of peat
production is used to calculate off-site CO2 emissions from dissolved organic carbon and on-site CO2 emissions. The
total area of peat production is also used to calculate on-site CH4 emissions, as described in the On-Site CI 14
Emissions section. The area of nutrient-rich peat production is used to estimate on-site N20 emissions, as described
in the On-Site N2O Emissions section.
Table 6-50: Peat Production Area (Hectares)
1990
2005
2013 2014 2015 2016 2017
Total Area of Peat Production 7,206
Area of Nutrient-Rich Production 554
6,954
274
4,860 4,884 4,759 4,611 4,601
465 515 501 529 660
The IPCC (2006) on-site emissions equation also includes a term which 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 lias been declining since 1990; therefore, it
60 Hie vacuum method is one type of extraction that annually "mills" or breaks up the surface of the peat into particles, which
then dry during the summer months. The air-dried peat particles are then collected by vacuum harvesters and transported from the
area to stockpiles (IPCC 2006).
Land Use, Land-Use Change, and Forestry 6-83
-------
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 2014).
On-site N2O Emissions
IPCC (2006) suggests basing the calculation of on-site N20 emission estimates on the area of nutrient-rich peatlands
managed for peat extraction. These area data are not available directly for the United States, but the on-site CO2
emissions methodology above details the calculation of area data from production data. In order to estimate N20
emissions, the area of nutrient-rich Peatlands Remaining Peatlands was multiplied by the appropriate default
emission factor taken from IPCC (2014).
On-site CH4 Emissions
IPCC (2014) also suggests basing the calculation of on-site CH4 emission estimates on the total area of peatlands
managed for peat extraction. Area data is derived using the calculation from production data described in the On-site
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 (2014). In
order to estimate CH4 emissions from drainage ditches, the total area of peatland was multiplied by the default
fraction of peatland area that contains drainage ditches, and the appropriate emission factor taken from IPCC (2014).
Uncertainty and Time-Series Consistency
A Monte Carlo (Approach 2) uncertainty analysis was applied to estimate the uncertainty of CO2, CH4, and N20
emissions from Peatlands Remaining Peatlands for 2017, 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 2014) 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 2014).
The results of the Approach 2 quantitative uncertainty analysis are summarized in Table 6-51. Carbon dioxide
emissions from Peatlands Remaining Peatlands in 2017 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 2017 emission
estimate of 0.7 MMT CO2 Eq. Methane emissions from Peatlands Remaining Peatlands in 2017 were estimated to
be between 0.002 and 0.007 MMT CO2 Eq. This indicates a range of 57 percent below to 79 percent above the 2017
emission estimate of 0.004 MMT CO2 Eq. Nitrous oxide emissions from Peatlands Remaining Peatlands in 2017
were estimated to be between 0.0002 and 0.0008 MMT CO2 Eq. at the 95 percent confidence level. This indicates a
range of 53 percent below to 54 percent above the 2017 emission estimate of 0.0005 MMT CO2 Eq.
6-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Table 6-51: Approach 2 Quantitative Uncertainty Estimates for CO2, Cm, and N2O Emissions
from Peat lands Remaining Peat/a nds (MMT CO2 Eq. and Percent)
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Upper
Lower
Upper
Bound
Bound
Bound
Bound
Peatlands Remaining Peatlands
CO2
0.7
0.6
0.8
-15%
15%
Peatlands Remaining Peatlands
CH4
+
+
+
-57%
79%
Peatlands Remaining Peatlands
N2O
+
+
+
-53%
54%
+ Does not exceed 0.05 MMT CO2 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
The emissions estimates for Peatlands Remaining Peatlands were updated for 2017 using the Peat section of the
Mineral Commodity Summaries 2017 and Mineral Commodity Summaries 2018. The 2018 edition provided 2016
data and updated 2015 data for the lower 48 states. The 2017 edition provided peat type production estimates for
2016. Although Alaska peat production data for 2017 were unavailable, 2014 data are available in the Alaska's
Mineral Industry 2014 report. However, the reported values represented an apparent 98 percent decrease in
production since 2012. Due to the uncertainty of the most recent data, 2013, 2014, 2015, 2016, and 2017 values
were assumed to be equal to the 2012 value. If updated data are available for the next inventory cycle, this will result
in a recalculation in the next Inventory report.
Planned Improvements
In order to further improve estimates of CO2, N2O, and CH4 emissions from Peatlands Remaining Peatlands, future
efforts will investigate if improved data sources exist for determining the quantity of peat harvested per hectare and
the total area undergoing peat extraction.
Efforts will also be made to 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.
The implied emission factors will be calculated and included in this chapter for future inventories. The N20
emissions calculation uses different land areas than the CO2 and CH4 emission calculations, so estimating the
implied emission factor per total land area is not appropriate and are not generated in the CRF tables. The inclusion
of implied emission factors in this chapter will provide another method of QA/QC and verification.
The 20061PCC Guidelines do not cover all wetland types; they are restricted to peatlands drained and managed for
peat extraction, conversion to flooded lands, and some guidance for drained organic soils. They also do not cover all
of the significant activities occurring on wetlands (e.g., rewetting of peatlands). Since this inventory only includes
Peatlands Remaining Peatlands, additional wetland types and activities found in the 20131PCC Supplement will be
reviewed to determine if they apply to the United States. For those that do, available data will be investigated to
allow for the estimation of greenhouse gas fluxes in future inventory years.
Land Use, Land-Use Change, and Forestry 6-85
-------
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,61 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 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 recognizes both Vegetated Wetlands and Unvegetated Open Water as Coastal Wetlands. Per
guidance provided by the Wetlands Supplement, sequestration of carbon into biomass and soil carbon pools is
recognized only in Vegetated Coastal Wetlands and not to 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 along the oceanic shores on the
conterminous U.S., but does not include Coastal Wetlands Remaining Coastal Wetlands in Alaska or Hawaii.
Seagrasses are not currently included within the Inventory due to insufficient data on distribution, change through
time and carbon (C) stocks or C stock changes as a result of anthropogenic influence.
Under the Coastal Wetlands Remaining Coastal Wetlands category, the following emissions and removals are
quantified in this chapter:
1) Carbon stock changes and CH4 emissions on Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands,
2) Carbon 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 emissions 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 2013 IPCC Wetlands Supplement
methodologies for CH4 emissions, coastal wetlands in salinity conditions less than half that of sea water are sources
of CH4 as result of slow decomposition of organic matter under lower salinity brackish and freshwater, anaerobic
conditions. Conversion of Vegetated Coastal Wetlands to or from Unvegetated Open Water Coastal Wetlands do not
result in a change in salinity condition and are assumed to have no impact on CH4 emissions. The 2013 IPCC
Wetlands Supplement provides methodologies to estimate N20 emissions on coastal wetlands that occur due to
aquaculture. While N20 emissions can also occur due to anthropogenic N loading from the watershed and
atmospheric deposition, these emissions are not reported here to avoid double-counting of indirect N20 emissions
with the Agricultural Soils Management, Forest Land and Settlements categories. The N20 emissions from
61 See .
6-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
aquaculture result from the N derived from consumption of the applied food stock that is then excreted as N load
available for conversion to N20.
The Wetlands Supplement provides procedures for estimating C stock changes and CH4 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. All 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 N20 from aquaculture are included in this Inventory. At the present stage of inventory
development, Coastal Wetlands are not explicitly shown in the Land Representation analysis while work continues
to harmonize data from NOAA's Coastal Change Analysis Program62 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 (602,652 ha), palustrine
scrub shrub (140,602 ha) and estuarine emergent marsh (1,838,461 ha), estuarine scrub shrub (97,231 ha) and
estuarine forest (192,011 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 (52,405 ha), warm temperate (899,026 ha),
subtropical (1,863,204 ha) and Mediterranean (56,322 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, for the
first time, includes changes in aboveground biomass C stocks along with soils. Currently, insufficient data exist on
C stock changes in belowground biomass, DOM and litter. 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 6-52 through Table 6-54 below summarize nationally aggregated aboveground biomass and soil C stock
changes and CHi emissions on Vegetated Coastal Wetlands. Intact Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands hold a relatively small aboveground biomass C stock (9 MMT); however, wetlands
maintain a large C stock in soil (estimated to be 870 MMT C (3,190 MMT CO2 Eq.)) within the top 1 meter of soil
to which C is accumulated at a yearly rate of 9.9 MMT CO2 Eq. over the past five years. Recent yearly CH4
emissions of 3.6 of MMT CO2 Eq. offset C removals resulting in an annual net C removal rate of 6.5 MMT CO2 Eq.
Due to federal regulatory protection, loss of Vegetated Coastal Wetland area slowed considerably in the 1970s and
the current rates of C stock change and CH4 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 6-52: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands (MMT CO2 Eq.)
Year 1990 2005 2013 2014 2015 2016 2017
Soil Flux (9.9) ' (10.0) (9.9) (9.9) (9.9) (9.9) (9.9)
62 See .
Land Use, Land-Use Change, and Forestry 6-87
-------
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.
Table 6-53: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Remaining
Vegetated Coastal Wetlands (MMT C)
Year 1990 2005 2013 2014 2015 2016 2017
Soil Flux (2.7) (2.7) (2.7) (2.7) (2.7) (2.7) (2.7)
Abovegroimd Biomass Flux (0.01) 0.01 (0.01) (0.01) (0.01) (0.01) (0.01)
Total C Stock Change {2.1) {2.1) {2.1) {2.1) {2.1) {2.1) {2.1)
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
Table 6-54: cm Emissions from Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands {imi CCh Eq. and kt CH4)
Year
1990
2005
2013
2014
2015
2016
2017
Methane Emissions (MMT CO2 Eq.)
3.4
3.5l
3.6
3.6
3.6
3.6
3.6
Methane Emissions (kt CFLi)
137
140
142
143
143
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 CH4 for Vegetated Coastal Wetlands Remaining Vegetated Coastal
Wetlands.
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.63 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.64 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).
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
63 See .
64 See .
6-88 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 2017 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 CH4 emissions for Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands are
derived from the same wetland map used in the analysis of wetland soil C fluxes, produced from C-CAP, LiDAR
and tidal data, in combination with default CH4 emission factors provided in Table 4.14 of the Wetlands Supplement.
The methodology follows 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 CH4 include
uncertainties associated with Tier 2 literature values of soil C stocks, aboveground biomass C stocks and CH4 flux,
assumptions that underlie the methodological approaches applied and uncertainties linked to interpretation of remote
sensing data. Uncertainty specific to Vegetated Coastal Wetlands Remaining Vegetated Coastal Wetlands include
differentiation of palustrine and estuarine community classes, which determines the soil C stock and CH4 flux
applied. Soil C stocks and CH4 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 judgement 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 CH4 flux are the Tier 1 default values reported in the Wetlands Supplement. Overall
uncertainty of the NOAA C-CAP remote sensing product is 15 percent. This is in the range of remote sensing
methods (±10-15 percent; IPCC 2003). However, there is significant uncertainty in salinity ranges for tidal and non-
tidal estuarine wetlands and activity data used to apply CH4 flux emission factors (delineation of an 18 ppt
boundary) will need significant improvement to reduce uncertainties.
Table 6-55: 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)
2017 Estimate Uncertainty Range Relative to Estimate
Gas (MMT CO2 Eq.) (MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Soil C Stock Change
CO2
(9.9)
(11.7)
(8.1)
-29.5%
29.5%
Aboveground Biomass C Stock Change
CO2
(0.02)
(0.03)
(0.02)
-16.5%
16.5%
CH4 emissions
CH4
3.6
2.5
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
Land Use, Land-Use Change, and Forestry 6-89
-------
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 CH4 emission factors derived from the IPCC Wetlands Supplement.
Recalculations Discussion
Methodological recalculations are associated with the extension of C-CAP data extrapolation through 2017. Soil
reference carbon sequestration rates were expanded and reanalyzed based upon geometric means; upper and lower
95 percent confidence intervals were calculated and the larger of the two was used (Lu and Megonigal 2017).
Recalculation of carbon sequestration lowered annual removals from 12.1 MMT CO2 Eq. to 9.9 MMT CO2 Eq. per
year over the past five years. New data on aboveground biomass carbon stocks were added that were derived from a
national assessment combining field plot data and aboveground biomass mapping by remote sensing (Byrd et al.
2017; Byrd, etal. 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 stock and aboveground biomass for
coastal wetlands.65 This dataset will be updated periodically. Refined error analysis combining land cover change
and C stock estimates will be provided as new data are incorporated. Through this work, a model is in development
to represent changes in soil C stocks for estuarine emergent wetlands. The C-CAP dataset for 2015 is currently
under development with planned release 2019. Additional data products for years 2003, 2008 and 2013 are also
planned for release. Once complete, land use change for 1990 through 2017 will be recalculated with this updated
dataset.
Emissions from Vegetated Coastal Wetlands Converted to
Unvegetated Open Water Coastal Wetlands
Conversion of intact Vegetated Coastal Wetlands to Unvegetated Open Water Coastal Wetlands is a source of
emissions from both soil and biomass C stocks. It is estimated that 4,828 ha of Vegetated Coastal Wetlands were
converted to Unvegetated Open Water Coastal Wetlands in 2017. 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 recorded by the C-CAP surveys of 2005 and 2010 coincides with two
such events, hurricanes Katrina and Rita both in 2005.
Shallow nearshore open water within the U.S. Land Representation is recognized as falling under the Wetlands
category within the U.S. 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
65 See .
6-90 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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). A Tier 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 enviromnent.
Table 6-56: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to
Unvegetated Open Water Coastal Wetlands (MMT CO2 Eq.)
Year 1990
2005
2013 2014 2015 2016 2017
Soil Flux 4.8
Aboveground Biomass Flux 0.04
3.1
0.03
4.8 4.8 4.8 4.8 4.8
0.04 0.04 0.04 0.04 0.04
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 6-57: CO2 Flux from C Stock Changes in Vegetated Coastal Wetlands Converted to
Unvegetated Open Water Coastal Wetlands (MMT C)
Year 1990
2005
2013 2014 2015 2016 2017
Soil Flux 1.3
Abovegroimd Biomass Flux 0.01
0.8
0.01
1.3 1.3 1.3 1.3 1.3
0.01 0.01 0.01 0.01 0.01
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.
Methodology
The following section includes a brief description of the methodology used to estimate changes in soil and
aboveground biomass 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 2017 from these datasets. C-CAP
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 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 1990-2017 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
Land Use, Land-Use Change, and Forestry 6-91
-------
year in each climate zone by its mean aboveground biomass. Currently, a nationwide dataset for belowground
biomass has not been assembled.
Soil Methane Emissions
A Tier 1 assumption has been applied that salinity conditions are unchanged and hence methane 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. 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 judgement 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; 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). 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 6-58: 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
2017 Flux Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Flux Estimate
(MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
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%
Total Flux
4.8
3.0
6.7
-24.4%
+24.4%
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 2017. Uncertainty in the
defining the long-term trend will be improved with release of the 2015 survey, expected in 2019.
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.
6-92 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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. 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 Monitoring System Science Team, corroborated the assumption that where salinities are
unchanged CH4 emissions are constant with conversion of Unvegetated Open Water Coastal Wetlands to Vegetated
Coastal Wetlands.
Recalculations Discussion
Methodological recalculations are associated with the extension of C-CAP data extrapolation through 2017.
Reference soil carbon stocks were modified to 2701C ha1 to all vegetated intertidal coastal wetland classes and for
all climatic zones, reflecting analysis by Holmquist et al. (2018). This resulted in an increase in soil carbon
emissions due wetland erosions: 1.3 MMT CO2 Eq. per year over the period of 2011 to 2016. New data on
aboveground biomass carbon stocks were added, broken down by climate zone, that were derived from a national
assessment combining field plot data and aboveground biomass mapping by remote sensing (Byrd et al. 2017; Byrd,
et al. 2018), increasing emissions further by 0.04 MMT CO2 Eq. peryear.
Planned Improvements
A refined uncertainty analysis and efforts to improve times series consistency are planned for the 1990 through 2018
Inventory (i.e., 2020 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 of an Approach 3 land representation assessment in future
years.
The C-CAP dataset for 2015 is currently under development with a planned release in 2019. 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 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 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
Land Use, Land-Use Change, and Forestry 6-93
-------
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 and aboveground biomass C stocks are
reported in the Inventory at this time, but improvements are being evaluated to include changes from other C pools.
Table 6-59: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands
Converted to Vegetated Coastal Wetlands {MMJ CO2 Eq.)
Year 1990
2005
2013 2014 2015 2016 2017
Soil C Flux (0.004)
Aboveground Biomass C Flux (0.01)
(0.002)
(0.004)
(0.004) (0.004) (0.004) (0.004) (0.004)
(0.01) (0.01) (0.01) (0.01) (0.01)
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.
Table 6-60: CO2 Flux from C Stock Changes from Unvegetated Open Water Coastal Wetlands
Converted to Vegetated Coastal Wetlands {MMJ C)
Year 1990
2005
2013 2014 2015 2016 2017
Soil C Flux (0.001)
Aboveground Biomass C Flux (0.003)
(0.001)
(0.001)
(0.001) (0.001) (0.001) (0.001) (0.001)
(0.003) (0.003) (0.003) (0.003) (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.
Methodology
The following section includes a brief description of the methodology used to estimate changes in soil C stocks and
CH4 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 2017 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 land area for
managed coastal wetlands. The methodology follows Eq. 4.7, Chapter 4 of the Wetlands Supplement, and is applied
to the area of managed 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.
6-94 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
AbovegroundBiomass Carbon Stock Changes
Quantification of regional coastal wetland aboveground biomass C stock changes for palustrine and estuarine marsh
vegetation are presented this year 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 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 1990-2017 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.
Soil Methane Emissions
A Tier 1 assumption has been applied that salinity conditions are unchanged and hence methane 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 judgement 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). 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).
Table 6-61: 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
2017 Flux Estimate
(MMT CO2 Eq.)
Uncertainty Range
(MMT CO2 Eq.)
Relative to Flux Estimate
(%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Soil C Stock Flux
Aboveground Biomass C Stock Flux
(0.004)
(0.01)
(0.005)
(0.01)
(0.004)
(0.01)
-29.5%
-16.5%
29.5%
16.5%
Total Flux
(0.02)
(0.02)
(0.01)
-38.6%
38.6%
Note: Parentheses indicate net sequestration. Totals may not sum due to independent rounding.
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
Land Use, Land-Use Change, and Forestry 6-95
-------
analysis by the Blue Carbon Monitoring project and reviewed by US Geological Survey prior to publishing, the
peer-review process during publishing, and the Coastal Wetland Inventory team leads before 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 calculation worksheets. Two biogeochemists at the USGS, also members
of the NASA Carbon Monitoring System Science Team corroborated the simplifying assumption that where
salinities are unchanged CH4 emissions are constant with conversion of Unvegetated Open Water Coastal Wetlands
to Vegetated Coastal Wetlands.
Recalculations Discussion
Methodological recalculations are associated with the extension of C-CAP data extrapolation through 2017. Soil
reference carbon sequestration rates were updated based on recalculation by Lu and Megonigal (2017), which
decreased net removals to soil by 0.01 MMT CO2 Eq. per year. New data on aboveground biomass carbon stocks
were added, broken down by climate zone, that were derived from a national assessment combining field plot data
and aboveground biomass mapping by remote sensing (Byrd et al., 2017; Byrd, et al., 2018). This resulted in an
increase in net removal by 0.01 MMT CO2 Eq. per year.
Planned Improvements
Administered by the Smithsonian Enviromnental Research Center, the Coastal Wetland Carbon Research
Coordination Network have 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 2019. Additional data
products for years 2003, 2008, and 2013 are also planned for release. Once complete, land use change for 1990
through 2017 will be recalculated and extended to 2018 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 N20. Nitrous
oxide is generated and emitted as a byproduct of the conversion of ammonia (contained in fish urea) to nitrate
through nitrification and nitrate to N2 gas through denitrification (Hu et al. 2012). Nitrous oxide emissions can be
readily estimated from data on fish production (IPCC 2013 Wetlands Supplement).
Overall, aquaculture production in the United States lias fluctuated slightly from year to year, 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; however, data for 2016 and 2017 are not yet available and in this analysis are held constant with 2015
emissions of 0.1 MMT CO2 Eq.
Table 6-62: N2O Emissions from Aquaculture in Coastal Wetlands (MMT CO2 Eq.)
Year
1990
2005
2013
2014
2015
2016
2017
Emissions (MMT CO2 Eq.)
0.1
0.21
0.1
0.1
0.1
0.1
0.1
Emissions (kt N2O)
0.4
0.6
0.5
0.5
0.5
0.5
0.5
Methodology
The methodology to estimate N20 emissions from Aquaculture in Coastal Wetlands follows guidance in the 2013
IPCC Wetlands Supplementby applying country-specific fisheries production data and the IPCC Tier 1 default
emission factor.
6-96 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Each year NO AA Fisheries document the status of U.S. marine fisheries in the annual report of Fisheries of the
United States (National Marine Fisheries Service, 2016), from which activity data for this analysis is derived.66 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 of fish-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 N20 from aquaculture. It is
not apparent from the data as to the amount of aquaculture occurring above the extent of high tides on river
floodplains. While some aquaculture 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 N20 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 N20 emissions (e.g., Clams, Mussels and Oysters) and have not been included in the
analysis. The IPCC Tier 1 default emissions factor of 0.00169 kg N20-N per kg of fish produced (95 percent
confidence interval - 0, 0.0038) is applied to the activity data to calculate total N20 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 N20 emissions. Uncertainties in N20 emissions from aquaculture are based on expert judgement for
the NOAA Fisheries of the United States fisheries production data (± 100 percent) multiplied by default uncertainty
level for N20 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 of fish production in non-coastal wetland areas,
this is a reasonable initial first approximation for an uncertainty range.
Table 6-63: Approach 1 Quantitative Uncertainty Estimates for N2O Emissions for
Aquaculture Production in Coastal Wetlands (MMT CO2 Eq. and Percent)
2017 Emissions
Estimate Uncertainty Range Relative to Emissions Estimate3
Source (MMT CO2 Eq.) (MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Combined Uncertainty for N2O Emissions
for Aquaculture Production in Coastal
Wetlands
0.1
0.00
0.31
-116%
116%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
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
N20 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 N20 emissions.
66 See .
Land Use, Land-Use Change, and Forestry 6-97
-------
6.9 Land Converted to Wetlands (CRF 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 2017 the rate of annual transition for Land Converted
to Vegetated Coastal Wetlands ranged from 2,619 ha/year to 5,316 ha/year. Conversion rates were higher during the
period 2010 through 2017 than during the earlier part of the time series.
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 Program67
with NRI 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 CH4 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.
Table 6-64: CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal
Wetlands (MMT COz Eq.)
Year 1990
2005
2013 2014 2015 2016 2017
Soil Flux (0.01)
Aboveground Biomass Flux (0.03)
(0.01)
(0.02)
(0.01) (0.01) (0.01) (0.01) (0.01)
(0.03) (0.03) (0.03) (0.03) (0.03)
Total C Stock Change (0.04)
(0.03)
(0.04) (0.04) (0.04) (0.04) (0.04)
Table 6-65: CO2 Flux from C Stock Changes in Land Converted to Vegetated Coastal
Wetlands (MMT C)
Year 1990
2005
2013 2014 2015 2016 2017
Soil Flux (0.004)
Aboveground Biomass Flux (0.01)
(0.002)
(0.01)
(0.005) (0.004) (0.004) (0.004) (0.004)
(0.01) (0.01) (0.01) (0.01) (0.01)
Total C Stock Change (0.01)
(0.01)
(0.01) (0.01) (0.01) (0.01) (0.01)
Table 6-66: ChU Emissions from Land Converted to Vegetated Coastal Wetlands (MMT CO2
Eq. and kt CH4)
Year
1990
2005
2013
2014
2015
2016
2017
Methane Emissions (MMT CO2 Eq.)
0.01
0.01
0.01
0.01
0.01
0.01
0.01
Methane Emissions (kt CFLi)
0.6
0.5
0.6
0.6
0.6
0.6
0.6
67 See .
6-98 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Methodology
The following section includes a description of the methodology used to estimate changes in soil and aboveground
biomass C stock changes 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.68 As a QC step, 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. Delineating Vegetated Coastal Wetlands from ephemerally flooded
upland Grasslands represents a particular challenge in remote sensing. Moreover, at the boundary between wetlands
and uplands, which may be gradual on low lying coastlines, the presence of wetlands may be ephemeral depending
upon weather and climate cycles and as such impacts on the emissions and removals will vary over these time
frames. 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, 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. The methodology follows Eq. 4.7, Chapter 4
of the LPCC Wetlands Supplement, and 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 soil organic content but 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 1990-2017 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. Currently, a nationwide dataset for belowground biomass has not been assembled.
Soil Methane Emissions
Tier 1 estimates of CH4 emissions for Land Converted to Vegetated Coastal Wetlands are derived from the same
wetland map used in the analysis of wetland soil C fluxes, produced from C-CAP, LiDAR and tidal data, in
combination with default CH4 emission factors provided in Table 4.14 of the LPCC Wetlands Supplement. The
methodology follows Eq. 4.9, Chapter 4 of the LPCC Wetlands Supplement, and is applied to the total area of Land
Converted to Vegetated Coastal Wetlands on an annual basis.
68 See .
Land Use, Land-Use Change, and Forestry 6-99
-------
Uncertainty and Time-Series Consistency
Underlying uncertainties in estimates of soil C removal factors, aboveground biomass change, and CH4 emissions
include error in uncertainties associated with Tier 2 literature values of soil C removal estimates, aboveground
biomass stocks, and IPCC default CH4 emission factors, uncertainties linked to interpretation of remote sensing data,
as well as assumptions that underlie the methodological approaches applied.
Uncertainty specific to coastal wetlands include differentiation of palustrine and estuarine community classes which
determines the soil C removal and CH4 flux applied. Soil C removal and CH4 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. Soil and aboveground biomass C removal data for all subcategories
are not available and thus assumptions were applied using expert judgement about the most appropriate assigmnent
of a soil and aboveground biomass C removal factor 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 CH4 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 CH4 flux (e.g., delineation of an 18 ppt boundary),
which will need significant improvement to reduce uncertainties.
Table 6-67: Approach 1 Quantitative Uncertainty Estimates for C Stock Changes occurring
within Land Converted to Vegetated Coastal Wetlands (MMT CO2 Eq. and Percent)
2017 Estimate Uncertainty Range Relative to Estimate3
(MMT CO2 Eq.) (MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
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%
a Range of flux estimates based on error propagation at 95 percent confidence interval.
Note: 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 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 US
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 where based upon peer-reviewed literature and CH4 emission factors derived
from the IPCC Wetlands Supplement.
Recalculations Discussion
Methodological recalculations are associated with the extension of C-CAP data extrapolation through 2017. Soil
reference carbon sequestration rates were updated based recalculation by Lu and Megonigal (2017). New data on
aboveground biomass carbon stocks were added, broken down by climate zone, that were derived from a national
assessment combining field plot data and aboveground biomass mapping by remote sensing (Byrd et al. 2017; Byrd
6-100 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
et al. 2018). A minor transcription error in calculation of Lands Converted to Wetlands for the Mediterranean
climate zone (years 2011 through 2016) was fixed. An error was corrected in aggregating state level activity data to
national level for Land Converted to Vegetated Coastal Wetlands emergent and scrub shrub wetlands (2011 through
2016), decreasing wetland area by 4,345 ha, which reduced net carbon removals to soil by 0.01 MMT CO2 Eq.
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.69 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
The C-CAP dataset for 2015 is currently under development with a planned release in early 2019. 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 with this updated dataset. Currently, biomass from lands converted to wetlands
are only tracked for one year due to lack of available data. In 2019, 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.
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, which is consistent with the assumption of the Tier 1 method in the 2006 IPCC Guidelines (IPCC
2006). 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.70 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 2012 United States Department of Agriculture (USD A) National Resources Inventory
(NRI) (USDA-NRCS 2015)71 or according to the National Land Cover Dataset (NLCD) for federal lands (Homer et
al. 2007; Fry et al. 2011; Homer et al. 2015). The Inventory includes settlements on privately-owned lands in the
conterminous United States and Hawaii. Alaska and the small amount of settlements on federal lands are not
included in this Inventory even though these areas are part of the U.S. managed land base. This leads to a
discrepancy with the total amount of managed area in Settlements Remaining Settlements (see Section 6.1
69 See .
70 N2O emissions from soils are included in the N2O Emissions from Settlement Soils section.
71 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 6-101
-------
Representation of the U.S. Land Base) and the settlements area included in the Inventory analysis72. There is a
planned improvement to include settlements on drained organic soils in these areas as part of a future Inventory.
CO2 emissions from drained organic soils in settlements are 1.3 MMT CO2 Eq. (0.3 MMT C) in 2017. Although the
flux is relatively small, the amount has increased by over 800 percent since 1990.
Table 6-68: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements
(MMT COz Eq.)
Soil Type 1990 2005 2013 2014 2015 2016 2017
Organic Soils 0.1 ^ 0.5 /¦;& 1.3 1.3 1.3 1.3 1.3
Table 6-69: Net CO2 Flux from Soil C Stock Changes in Settlements Remaining Settlements
(MMT C)
Soil Type 1990 2005 2013 2014 2015 2016 2017
Organic Soils + . 0.1 04 04 04 04 0.3
+ Does not exceed 0.05 MMT C
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 2012 NRI (USDA 2015) with additional
information from the NLCD (Fry et al. 2011; Homer et al. 2007; Homer et al. 2015). It is assumed that all settlement
area on organic soils is drained, and those areas are provided in Table 6-70 (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 6-70). The area of land on organic soils in Settlements Remaining Settlements
has increased from 3 thousand hectares in 1990 to over 28 thousand hectares in 2012. The area of land on organic
soils are not currently available from NRI for Settlements Remaining Settlements after 2012.
Table 6-70: Thousands of Hectares of Drained Organic Soils in Settlements Remaining
Settlements
Area
Year
(Thousand Hectares)
1990
3
2005
10
2012
28
2013
ND
2014
ND
2015
ND
2016
ND
2017
ND
Note: No NRI data are available after 2012. ND
(No data)
72 For the land representation, land use data for 2013 to 2017 were only partially updated based on new Forest Inventory and
Analysis (FIA) data. These updates led to changes in the land representation data for settlements through the process of
combining FIA data with land use data from the National Resources Inventory and National Land Cover Dataset (See
"Representation of the U.S. Land Base" section for more information). However, an inventory was not compiled for settlements
in this Inventory, but rather the emissions and removals are based on a surrogate data method. Therefore, the area estimates in
this section are based on the land representation data from the previous Inventory.
6-102 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
To estimate CO2 emissions from drained organic soils across the time series from 1990 to 2012, 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 2013 to 2017 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 2012, and in turn, the
trend is used to approximate the 2013 to 2017 emissions. The Tier 2 method described previously will be applied in
future inventories to recalculate the estimates beyond 2012 as activity data becomes 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 2013 to 2017 based on the linear time series model. The results of
the Approach 2 Monte Carlo uncertainty analysis are summarized in Table 6-71. Soil C losses from drained organic
soils in Settlements Remaining Settlements for 2017 are estimated to be between 0.8 and 1.8 MMT CO2 Eq. at a 95
percent confidence level. This indicates a range of 40 percent below and 40 percent above the 2017 emission
estimate of 1.3 MMT CO2 Eq.
Table 6-71: Uncertainty Estimates for CO2 Emissions from Drained Organic Soils in
Settlements Remaining Settlements (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emission
Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower Upper Lower Upper
Bound Bound Bound Bound
Organic Soils
CO2
1.3
0.8 1.8 -40% 40%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence
interval.
Methodological recalculations are applied from 2013 to 2017 using the linear time series model described above.
Details on the emission trends through time are described in more detail in the Methodology section, above.
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.
Recalculations Discussion
Methodological recalculations are associated with extending the time series from 2013 through 2017 using a linear
time series model. The recalculation had a minor effect on the time series overall with C losses from drainage of
organic soils increasing by less than 1 percent on average.
Planned Improvements
This source will be extended 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. New land representation data
will also be compiled, and the time series will be recalculated for the latter years that are estimated using the data
splicing method in the current Inventory.
Land Use, Land-Use Change, and Forestry 6-103
-------
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. Previous assessments of carbon stock
changes in settlements trees in the Inventory used urban areas as a proxy for settlement area. The past definition of
urban areas was based on population density as delimited by the U.S. Census Bureau. This assessment changes this
approach and uses the settlement areas from Section 6.1 Representation of the U.S. Land Base and tree cover in U.S.
developed land from the NLCD as a proxy for tree cover in settlements, which results in a close, but not precise
alignment with the settlement areas shown in Section 6.1 of this Inventory.
Trees in settlement areas of the United States are estimated to account for an average annual net sequestration of
113.7 MMT CO2 Eq. (31.0 MMT C) over the period from 1990 through 2017. Net C flux from settlement trees in
2017 is estimated to be -123.9 MMT CO2 Eq. (-33.8 MMT C). 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 increasing emissions. In addition, changes in species composition, tree sizes
and tree densities affect base C flux estimates. Annual estimates of CO2 flux (Table 6-72) 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 seven 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.
Carbon flux estimates per unit tree cover for settlement areas are derived from available data on tree cover and C
sequestration in U. S. cities. Percent tree cover in settlement areas was derived from NLCD tree cover data from
developed land, which were adjusted based on photo-interpretation of tree cover in developed land. Photo-
interpretation also includes changes in tree cover in developed lands based on paired photo-interpretation points
between c. 2011 and 2016. Annual sequestration increased by 29 percent between 1990 and 2017 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 6-72: Net C Flux from Settlement Trees (MMT CO2 Eq. and MMT C)
Year
MMT CO2 Eq.
MMT C
1990
(96.2)
(26.2)
2005
(1 16.8)
(31.9)
2013
(125.6)
(34.2)
2014
(125.0)
(34.1)
2015
(124.5)
(33.9)
2016
(123.9)
(33.8)
2017
(123.9)
(33.8)
Note: Parentheses indicate net sequestration.
Methodology
To estimate net carbon sequestration in settlement areas, three types of data are required by state:
6-104 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
1. Settlement area
2. Percent tree cover in settlement areas
3. Carbon sequestration density per unit of tree cover
Settlement Area
Settlements area is defined in Section 6.1 Representation of the U.S. Land Base as a land-use category representing
developed areas. However, as the data used to estimate settlement area comes from the NRI and there hasn't been
an update to this data since 2012, the decision was made to utilize the settlement area data from the previous 1990
through 2016 Inventory for this analysis, while also holding the 2017 value constant with the 2016 value. As a
result, the settlement areas used in this assessment are slightly different from the time series shown in Section 6.1
Representation of the U.S. Land Base (less than 0.24 percent on average over the years).
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 2017: used 2011 NLCD tree cover adjusted with 2016 photo-interpreted values
Land Use, Land-Use Change, and Forestry 6-105
-------
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 database73 and Forest Service urban forest inventory data (e.g.,
Nowak et al. 2016, 2017) (Table 6-73). 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 6-73: 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, MI
12.17
1.88
0.34
0.04
0.13
0.07
0.36
22.1
2.3
Albuquerque, NM
5.61
0.97
0.24
0.03
0.20
0.03
0.82
13.3
1.5
Arlington, TX
6.37
0.73
0.29
0.03
0.26
0.03
0.91
22.5
0.3
Atlanta, GA
6.63
0.54
0.23
0.02
0.18
0.03
0.76
53.9
1.6
Austin, TX
3.57
0.25
0.17
0.01
0.13
0.01
0.73
30.8
1.1
Baltimore, MD
10.30
1.24
0.33
0.04
0.20
0.04
0.59
28.5
1.0
Boise, ID
7.33
2.16
0.26
0.04
0.16
0.06
0.64
7.8
0.2
Boston, MA
7.02
0.96
0.23
0.03
0.17
0.02
0.73
28.9
1.5
Camden, NJ
11.04
6.78
0.32
0.20
0.03
0.10
0.11
16.3
9.9
Casper, WY
6.97
1.50
0.22
0.04
0.12
0.04
0.54
8.9
1.0
Chester, PA
8.83
1.20
0.39
0.04
0.25
0.05
0.64
20.5
1.7
Chicago (region), IL
9.38
0.59
0.38
0.02
0.26
0.02
0.70
15.5
0.3
Chicago, IL
6.03
0.64
0.21
0.02
0.15
0.02
0.70
18.0
1.2
Corvallis, OR
10.68
1.80
0.22
0.03
0.20
0.03
0.91
32.6
4.1
El Paso, TX
3.93
0.86
0.32
0.05
0.23
0.05
0.72
5.9
1.0
Freehold, NJ
11.50
1.78
0.31
0.05
0.20
0.05
0.64
31.2
3.3
Gainesville, FL
6.33
0.99
0.22
0.03
0.16
0.03
0.73
50.6
3.1
Golden, CO
5.88
1.33
0.23
0.05
0.18
0.04
0.79
11.4
1.5
Grand Rapids, MI
9.36
1.36
0.30
0.04
0.20
0.05
0.65
23.8
2.0
73 See .
6-106 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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
Indianab
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'3
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, WI
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
Nebraskab
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 Dakotab
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, WI
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 Dakotab
3.14
0.66
0.13
0.03
0.11
0.02
0.87
16.5
2.2
Syracuse, NY
9.48
1.08
0.30
0.03
0.22
0.04
0.72
26.9
1.3
Tennesseeb
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 takes into
account C emissions associated with tree death and removals. The third step in the methodology estimates net C
emissions from settlement trees based on estimates of annual mortality, tree condition, and assumptions about
whether dead trees were removed from the site. Estimates of annual mortality rates by diameter class and condition
class were obtained from a study of street-tree mortality (Nowak 1986). Different decomposition rates were applied
to dead trees left standing compared with those removed from the site. For removed trees, different rates were
applied to the removed/aboveground biomass in contrast to the belowground biomass (Nowak et al. 2002). The
estimated annual gross C emission rates for each plot were then scaled up to city estimates using tree population
information.
The full methodology development is described in the underlying literature, and key details and assumptions were
made as follows. The allometric models applied to the field data for the Nowak methodology for each tree were
Land Use, Land-Use Change, and Forestry 6-107
-------
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 6-74)
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 6-74). 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 6-74. 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 6-74: 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 (2017)
Gross Annual Net Annual Net: Gross
Sequestration Sequestration Annual
Gross Annual Net Annual Tree per Area of per Area of Sequestration
State
Sequestration
Sequestration
Cover
Tree Cover
Tree Cover
Ratio
Alabama
1,949,043
1,420,218
53.5
0.376
0.274
0.73
Alaska
116,009
84,533
47.4
0.169
0.123
0.73
Arizona
168,252
122,601
4.6
0.388
0.283
0.73
Arkansas
1,205,718
878,576
48.9
0.362
0.264
0.73
California
1,924,163
1,402,089
16.9
0.426
0.311
0.73
Colorado
136,841
99,713
8.0
0.216
0.157
0.73
Connecticut
601,867
438,565
58.7
0.262
0.191
0.73
Delaware
94,692
69,000
24.4
0.366
0.267
0.73
DC
11,995
8,741
25.1
0.366
0.267
0.73
Florida
4,204,004
3,063,350
40.3
0.520
0.379
0.73
Georgia
3,113,443
2,268,687
56.3
0.387
0.282
0.73
6-108 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Hawaii
301,173
219,457
41.7
0.637
0.464
0.73
Idaho
59,881
43,634
7.4
0.201
0.146
0.73
Illinois
655,998
478,009
15.5
0.310
0.226
0.73
Indiana
465,440
430,373
17.1
0.274
0.254
0.92
Iowa
175,849
128,137
8.6
0.263
0.191
0.73
Kansas
287,496
223,720
10.8
0.310
0.241
0.78
Kentucky
902,579
657,686
36.8
0.313
0.228
0.73
Louisiana
1,445,497
1,053,297
47.0
0.435
0.317
0.73
Maine
369,598
269,316
55.5
0.242
0.176
0.73
Maryland
793,137
577,939
40.1
0.353
0.257
0.73
Massachusetts
940,348
685,208
57.2
0.278
0.203
0.73
Michigan
1,317,348
959,918
34.7
0.241
0.175
0.73
Minnesota
311,422
226,926
13.1
0.251
0.183
0.73
Mississippi
1,406,412
1,024,817
57.3
0.377
0.275
0.73
Missouri
836,547
609,570
23.2
0.313
0.228
0.73
Montana
47,429
34,560
4.9
0.201
0.147
0.73
Nebraska
92,271
77,864
7.3
0.261
0.220
0.84
Nevada
38,516
28,066
4.8
0.226
0.165
0.73
New Hampshire
341,910
249,141
59.3
0.238
0.174
0.73
New Jersey
867,597
632,196
40.7
0.321
0.234
0.73
New Mexico
172,828
125,935
10.2
0.288
0.210
0.73
New York
1,472,194
1,072,751
39.9
0.263
0.192
0.73
North Carolina
2,914,053
2,123,396
54.1
0.341
0.249
0.73
North Dakota
18,021
8,563
1.8
0.244
0.116
0.48
Ohio
1,220,678
889,477
28.2
0.271
0.198
0.73
Oklahoma
687,300
500,818
22.1
0.364
0.265
0.73
Oregon
676,245
492,762
39.9
0.265
0.193
0.73
Pennsylvania
1,708,480
1,244,926
40.2
0.267
0.195
0.73
Rhode Island
120,034
87,466
50.0
0.283
0.206
0.73
South Carolina
1,679,448
1,223,771
53.8
0.370
0.269
0.73
South Dakota
28,803
24,978
2.9
0.258
0.224
0.87
Tennessee
1,520,025
1,359,081
41.1
0.332
0.297
0.89
Texas
3,937,047
2,868,826
28.5
0.403
0.294
0.73
Utah
118,115
86,068
11.7
0.235
0.172
0.73
Vermont
174,444
127,113
50.6
0.234
0.170
0.73
Virginia
1,863,143
1,357,625
52.9
0.321
0.234
0.73
Washington
1,032,079
752,049
37.6
0.282
0.206
0.73
West Virginia
628,574
458,026
64.1
0.264
0.192
0.73
Wisconsin
683,179
497,815
25.9
0.246
0.180
0.73
Wyoming
33,049
24,082
4.7
0.199
0.145
0.73
Total
45,870,216
33,791,433
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 6-75). 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
Land Use, Land-Use Change, and Forestry 6-109
-------
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 2017. The results of this quantitative uncertainty analysis are summarized in Table 6-75.
The net C flux from changes in C stocks in urban trees in 2017 was estimated to be between -182.6 and -64.0 MMT
CO2 Eq. at a 95 percent confidence level. This indicates a range of 47 percent more sequestration to 48 percent less
sequestration than the 2017 flux estimate of -123.9 MMT CO2 Eq.
Table 6-75: Approach 2 Quantitative Uncertainty Estimates for Net C Flux from Changes in C
Stocks in Settlement Trees (MMT CO2 Eq. and Percent)
Source
Gas
2017 Flux Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Flux Estimate
(MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Changes in C Stocks in
Settlement Trees
CO2
(123.9)
(182.6)
(64.0)
-47% 48%
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 2017. Details on the emission trends through time are described in more detail in the Methodology section,
above.
QA/QC and Verification
Tier 1 and Tier 2 QA/QC activities were conducted consistent with the U.S. QA/QC plan. Source-specific quality
control measures for settlement trees included checking input data, documentation, and calculations to ensure data
were properly handled through the inventory process. Errors that were found during this process were corrected as
necessary.
Recalculations Discussion
Past estimates of carbon sequestration in settlement areas used urban land and urban tree cover as proxy for the
settlement area estimates. This new approach uses settlement land area and percent tree cover in developed land as a
proxy for percent tree cover in settlement area. This approach to estimating tree cover is believed to be a better
approach as the land area totals between NLCD developed land and settlements align much closer than do urban
land (Table 6-76). Comparing NLCD developed land, urban land (previous method of assessing settlement carbon)
and settlement area in the conterminous United States, reveals:
• 2011 settlement area = 42.51 million ha
• 2010 urban area = 27.35 million ha (-36 percent compared to settlement area)
• 2011 NLCD developed land = 45.41 million ha (+6.8 percent compared to settlement area)
Table 6-76: Comparison of Settlement, Developed and Urban Land Area for Conterminous
United States
Settlement
Developed
Urban
State
ha (2011)
ha (2011)
ha (2010)
Alabama
962,863
977,171
573,377
Arizona
940,105
695,750
566,051
Arkansas
675,412
826,279
284,638
California
2,659,965
2,772,706
2,130,095
Colorado
789,092
763,913
395,419
Connecticut
388,777
314,105
472,596
Delaware
104,101
96,465
105,296
Florida
1,985,843
2,143,229
1,902,388
Georgia
1,399,213
1,529,610
1,236,321
Idaho
401,565
371,793
129,330
6-110 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Illinois
1,362,424
1,739,240
1,022,445
Indiana
987,906
1,015,945
653,408
Iowa
776,671
1,089,338
246,630
Kansas
860,579
1,107,665
252,178
Kentucky
778,060
781,755
364,934
Louisiana
702,575
846,643
512,518
Maine
272,418
296,070
92,849
Maryland
557,088
482,788
519,219
Massachusetts
580,120
529,429
767,917
Michigan
1,572,260
1,590,477
934,804
Minnesota
946,863
1,249,080
444,906
Mississippi
646,988
794,063
288,525
Missouri
1,150,921
1,262,346
531,858
Montana
481,111
552,027
76,888
Nebraska
479,506
732,393
135,555
Nevada
349,974
288,438
198,212
New Hampshire
238,170
189,572
166,613
New Jersey
660,640
610,737
757,507
New Mexico
585,252
381,817
214,415
New York
1,393,123
1,176,401
1,063,658
North Carolina
1,549,227
1,397,659
1,193,342
North Dakota
415,797
732,998
47,801
Ohio
1,584,543
1,569,694
1,144,527
Oklahoma
853,953
1,114,380
338,576
Oregon
641,273
681,309
286,589
Pennsylvania
1,571,368
1,444,560
1,220,442
Rhode Island
83,714
83,486
103,555
South Carolina
833,338
770,522
615,517
South Dakota
390,275
572,579
58,759
Tennessee
1,102,701
1,058,201
751,912
Texas
3,400,132
4,450,649
2,260,511
Utah
423,971
372,832
235,230
Vermont
146,658
135,858
40,335
Virginia
1,087,778
1,024,120
691,376
Washington
963,767
1,045,135
615,435
West Virginia
366,579
442,929
165,875
Wisconsin
1,064,980
1,083,778
487,222
Wyoming
350,006
223,165
50,347
Conterminous
U.S.
42,519,645
45,411,098
27,347,901
The advantages of this newer approach are that the settlement area is now exact (urban method underestimated the
land area as it used urban land instead of settlement land) and percent tree cover is now estimated using areas that
more closely align in total with settlement areas (previous approach used percent urban tree cover). It is not known
how well percent tree cover from developed land represents tree cover in settlement areas, but given the similarities
in definitions and area, the estimate is assumed to be reasonable.
Given that land area now matches with settlement area, the carbon estimates have increased from previous estimates
that used a smaller urban land area. In 2016, the net sequestration values increased from 92.9 MMT CO2 Eq.
(previous urban based estimate) to 123.9 MMT CO2 Eq. (2016 and 2017 settlement estimates) (+33 percent).
This new approach also added changes in percent tree cover based on paired-point analysis of photo interpretation.
Tree cover in developed land dropped from 31.5 percent in c. 2011 to 30.8 percent in c. 2016. This decline in tree
cover will reduce net carbon sequestration. As settlement land was held constant since 2012, tree cover decline led
to a decrease in net sequestration between 2012 and 2016 (Table 6-76). Once settlement area projections are
updated, settlement areas estimates since 2012 should increase and lead to increasing sequestration during this
period, but at a lesser rate than if tree cover was held constant. Tree cover is intended to be reinterpreted using the
same 1,000 paired points in the coming years to monitor tree cover changes in developed lands.
Land Use, Land-Use Change, and Forestry 6-111
-------
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 USD A Forest Service to reconcile the overlap
between urban forest and non-urban forest greenhouse gas inventories. Estimates for settlements 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 "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) Settlement land area will be updated utilizing new data from the most recent National Resources Inventory
(NRI) that will be incorporated into the Section 6.1 Representation of the U.S. Land Base. This update will
provide new data beyond the current NRI that extends through 2012
b) Photo interpretation of settlement tree cover will be updated bi-annually to update tree cover estimates and
trends
c) Areas for photo interpretation of settlement area tree cover will be updated as new NLCD developed land
information becomes available
d) 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
4E1J
Of the synthetic N fertilizers applied to soils in the United States, approximately 3.1 percent are currently applied to
lawns, golf courses, and other landscaping within settlement areas. Application rates are lower than those occurring
on cropland soils, and, therefore, account for a smaller proportion of total U.S. soil N20 emissions per unit area. In
addition to synthetic N fertilizers, a portion of surface applied biosolids (i.e., sewage sludge) is applied to settlement
areas, and drained organic soils (i.e., soils with high organic matter content, known as Histosols) also contribute to
emissions of soil N20.
N additions to soils result in direct and indirect N20 emissions. Direct emissions occur on-site due to the N additions
in the form of synthetic fertilizers and biosolids as well as enhanced mineralization of N in drained organic soils.
Indirect emissions result from fertilizer and biosolids N that is transformed and transported to another location in a
form other than N20 (ammonia [NH3] and nitrogen oxide [NOx] volatilization, nitrate [NO3 ] leaching and runoff),
and later converted into N20 at the off-site location. The indirect emissions are assigned to settlements because the
management activity leading to the emissions occurred in settlements.
Total N20 emissions from soils in Settlements Remaining Settlements74 are 2.5 MMT C02 Eq. (8 kt of N20) in
2017. There is an overall increase of 73 percent from 1990 to 2017 due to an expanding settlement area leading to
more synthetic N fertilizer applications. 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 6-77.
Table 6-77: N2O Emissions from Soils in Settlements Remaining Settlements (MMT CO2 Eq.
and kt N2O)
1990 2005 2013 2014 2015 2016 2017
MMT CO2 Eq.
74 Estimates of Soil N2O for Settlements Remaining Settlements include emissions from Land Converted to Settlements because it
was not possible to separate the activity data.
6-112 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Direct N2O Emissions from Soils
1.1
1.9
2.0
2.0
2.0
1.9
1.9
Synthetic Fertilizers
0.8
1.6
1.7
1.7
1.6
1.6
1.6
Biosolids
0.2
0.2
0.2
0.2
0.2
0.2
0.2
Drained Organic Soils
0.1
0.1
0.2
0.2
0.2
0.2
0.2
Indirect N2O Emissions from Soils
0.4
0.6
0.6
0.6
0.6
0.6
0.6
Total
1.4
2.5
2.6
2.6
2.5
2.5
2.5
ktN20
Direct N2O Emissions from Soils
4
6
7
7
7
7
6
Synthetic Fertilizers
3
5
6
6
6
5
5
Biosolids
1
1
1
1
1
1
1
Drained Organic Soils
+
1
1
1
1
1
1
Indirect N2O Emissions from Soils
1
2
2
2
2
2
2
Total
5
8
9
9
9
8
8
+ Does not exceed 0.5 kt
Methodology
For settlement soils, the IPCC Tier 1 approach is used to estimate soil N20 emissions from synthetic N fertilizer,
biosolids additions, and drained organic soils. Estimates of direct N20 emissions from soils in settlements are based
on the amount of N in synthetic commercial fertilizers applied to settlement soils, the amount of N in biosolids
applied to non-agricultural land and surface disposal (see Section 7.2, Wastewater Treatment for a detailed
discussion of the methodology for estimating biosolids available for non-agricultural land application), and the area
of drained organic soils within settlements.
Nitrogen applications to settlement soils are estimated using data compiled by the USGS (Ruddy et al. 2006). The
USGS estimated on-farm and non-farm fertilizer use is based on sales records at the county level from 1982 through
2001 (Ruddy et al. 2006). Non-farm N fertilizer is assumed to be applied to settlements and forest lands; values for
2002 through 2012 are based on 2001 values adjusted for annual total N fertilizer sales in the United States because
there is no new activity data on application after 2001. Settlement application is calculated by subtracting forest
application from total non-farm fertilizer use. Biosolids applications are derived from national data on biosolids
generation, disposition, and N content (see Section 7.2, Wastewater Treatment for further detail). The total amount
of N resulting from these sources is multiplied by the IPCC default emission factor for applied N (one percent) to
estimate direct N20 emissions (IPCC 2006) for 1990 to 2012. The IPCC (2006) Tier 1 method is also used to
estimate direct N20 emissions due to drainage of organic soils in settlements at the national scale. Estimates of the
total area of drained organic soils are obtained from the 2012 NRI (USDA-NRCS 2015) using soils data from the
Soil Survey Geographic Database (SSURGO) (Soil Survey Staff 2011). To estimate annual emissions from 1990 to
2012, the total area is multiplied by the IPCC default emission factor for temperate regions (IPCC 2006). This
Inventory does not include soil N20 emissions from drainage of organic soils in Alaska and federal lands, although
this is a planned improvement for a future Inventory.
For indirect emissions, the total N applied from fertilizer and biosolids is multiplied by the IPCC default factors of
10 percent for volatilization and 30 percent for leaching/runoff to calculate the amount of N volatilized and the
amount of N leached/runoff The amount of N volatilized is multiplied by the IPCC default factor of one percent for
the portion of volatilized N that is converted to N20 off-site and the amount of N leached/runoff is multiplied by the
IPCC default factor of 0.075 percent for the portion of leached/runoff N that is converted to N20 off-site. The
resulting estimates are summed to obtain total indirect emissions from 1990 to 2012.
A linear extrapolation of the trend in the time series is applied to estimate the direct and indirect N20 emissions
from 2013 to 2017 from synthetic fertilizers and drained organic soils because new activity data for these two
sources have not been compiled for the latter part of the time series. Specifically, a linear regression model with
autoregressive moving-average (ARMA) errors (Brockwell and Davis 2016) is used to estimate the trend in
emissions over time from 1990 to 2012, and in turn, the trend is used to approximate the 2013 to 2017 emissions.
The time series will be recalculated for the years beyond 2012 in a future inventory with the methods described
above for 1990 to 2012. This Inventory does incorporate updated activity data on biosolids application in settlements
through 2017.
Land Use, Land-Use Change, and Forestry 6-113
-------
Uncertainty and Time-Series Consistency
The amount of N20 emitted from settlement soils depends not only on N inputs and area of drained organic soils,
but also on a large number of variables that can influence rates of nitrification and denitrification, including organic
C availability; rate, application method, and timing of N input; oxygen gas partial pressure; soil moisture content;
pH; temperature; and irrigation/watering practices. The effect of the combined interaction of these variables on N20
emissions is complex and highly uncertain. The IPCC default methodology does not explicitly incorporate any of
these variables, except variations in the total amount of fertilizer N and biosolids applications. All settlement soils
are treated equivalently under this methodology.
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.75 Uncertainty in the area of drained
organic soils is based on the estimated variance from the NRI survey (USDA-NRCS 2015). For 2013 to 2017, there
is also additional uncertainty associated with the surrogate data method. Uncertainty in the amounts of biosolids
applied to non-agricultural lands and used in surface disposal is 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.
Uncertainty in the direct and indirect emission factors is provided by IPCC (2006).
Uncertainty is propagated through the calculations of N20 emissions from fertilizer N and drainage of organic soils
using a Monte Carlo analysis. The results are combined with the uncertainty in N20 emissions from the biosolids
application using simple error propagation methods (IPCC 2006). The results are summarized in Table 6-78. Direct
N20 emissions from soils in Settlements Remaining Settlements in 2017 are estimated to be between 1.3 and 2.7
MMT C02 Eq. at a 95 percent confidence level. This indicates a range of 31 percent below to 41 percent above the
2017 emission estimate of 1.9 MMT C02 Eq. Indirect N20 emissions in 2017 are between 0.4 and 0.7 MMT C02
Eq., ranging from a -26 percent to 26 percent around the estimate of 0.6 MMT C02 Eq.
Table 6-78: Quantitative Uncertainty Estimates of N2O Emissions from Soils in Settlements
Remaining Settlements (MMT CO2 Eq. and Percent)
Source
Gas
2017 Emissions
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Settlements Remaining
Settlements
Lower
Bound
Upper
Bound
Lower Upper
Bound Bound
Direct N2O Emissions from Soils
N2O
1.9
1.3
2.7
-31% 41%
Indirect N2O Emissions from
Soils
N2O
0.6
0.4
0.7
-26% 26%
a Range of emission estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Note: These estimates include direct and indirect N2O emissions from Settlements Remaining Settlements and Land
Converted to Settlements because it was not possible to separate the activity data.
Methodological recalculations are applied from 2013 to 2017 using the linear time series model described above.
Details on the emission trends through time are described in more detail in the Methodology section, above.
QA/QC and Verification
The spreadsheet containing fertilizer, drainage of organic soils, and biosolids applied to settlements and calculations
for N20 and uncertainty ranges have been checked and verified.
75 No uncertainty is provided with the USGS fertilizer consumption data (Ruddy et al. 2006) so a conservative ±50 percent is
used in the analysis. Biosolids data are also assumed to have an uncertainty of ±50 percent.
6-114 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Recalculations Discussion
Methodological recalculations are associated with extending the time series from 2013 through 2017 using a linear
time series model. The recalculation had a minor effect on the time series overall with N20 emissions declining by
less than 1 percent on average.
Planned Improvements
This source will be extended to include soil N20 emissions from drainage of organic soils in settlements of Alaska
and federal lands in order to provide a complete inventory of emissions for this category. Updated data on fertilizer
amount and area of drained organic soils will be compiled to update emissions estimates for estimates beyond 2012
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 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. While the majority of food scraps originate from cropland and grassland, in this Inventory they are reported
with the yard trimmings in the Settlements Remaining Settlements section. Additionally, landfills are considered part
of the managed land base under settlements (see Section 6.1 Representation of the U.S. Land Base) and reporting
these C stock changes that occur entirely within landfills fits most appropriately within the Settlements Remaining
Settlements section.
Both the estimated amount of yard trimmings collected annually and the fraction that is landfilled have been
declining. In 1990, over 53 million metric tons (wet weight) of yard trimmings and food scraps are estimated to have
been generated (i.e., put at the curb for collection to be taken to disposal sites or to composting facilities) (EPA
2018). Since then, programs banning or discouraging yard trimmings disposal have led to an increase in backyard
composting and the use of mulching mowers, and a consequent estimated 0.8 percent decrease between 1990 and
2017 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 The net effect of the
reduction in generation and the increase in composting is a 57 percent decrease in the quantity of yard trimmings
disposed of in landfills since 1990.
Food scrap generation has grown by an estimated 67 percent since 1990, and while the proportion of total food
scraps generated that are eventually discarded in landfills has decreased slightly, from an estimated 82 percent in
1990 to 76 percent in 2017, the tonnage disposed of in landfills has increased considerably (by an estimated 55
percent) due to the increase in food scrap generation. Although the total tonnage of food scraps disposed of in
landfills has increased from 1990 to 2017, the difference in the amount of food scraps added from one year to the
next generally decreased, and consequently the annual carbon stock net changes from food scraps have generally
decreased as well (as shown in Table 6-79 and Table 6-80). As described in the Methodology section, the carbon
stocks are modeled using data on the amount of food scraps landfilled since 1960. These food scraps decompose
over time, producing CH4 and 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 food scraps in a single
more recent year; and (2) is much larger than the carbon introduced from food scraps in a single more recent year,
the total carbon stock in the landfill is primarily driven by the more stable 'older' carbon stock, thus resulting in less
annual change in later years.
Land Use, Land-Use Change, and Forestry 6-115
-------
Overall, the decrease in the landfill disposal rate of yard trimmings has more than compensated for the increase in
food scrap disposal in landfills, and the net result is a decrease in annual net change landfill C storage from 26.0
MMT C02 Eq. (7.1 MMT C) in 1990 to 11.9 MMT C02 Eq. (3.2 MMT C) in 2017 (Table 6-79 and Table 6-80).
Table 6-79: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills
(MMT COz Eq.)
Carbon Pool
1990
2005
2013
2014
2015
2016
2017
Yard Trimmings
(21.0)
(7.4)
(8.4)
(8.3)
(8.3)
(8.4)
(8.4)
Grass
(1.8)
(0.6)
(0.8)
(0.8)
(0.8)
(0.8)
(0.8)
Leaves
(9.0)
(3.3)
(3.9)
(3.8)
(3.8)
(3.9)
(3.9)
Branches
(10.2)
(3.4)
(3.8)
(3.8)
(3.7)
(3.7)
(3.7)
Food Scraps
(5.0)
(4.1)
(3.2)
(3.8)
(3.9)
(3.7)
(3.5)
Total Net Flux
(26.0)
(11.4)
(11.7)
(12.1)
(12.3)
(12.1)
(11.9)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net
sequestration.
Table 6-80: Net Changes in Yard Trimmings and Food Scrap Carbon Stocks in Landfills
(MMT C)
Carbon Pool
1990
2005
2013
2014
2015
2016
2017
Yard Trimmings
(5.7)
(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
(2.5)
(0.9)
(1.1)
(1.0)
(1.0)
(1.1)
(1.1)
Branches
(2.8)
(0.9)
(1.0)
(1.0)
(1.0)
(1.0)
(1.0)
Food Scraps
(1.4)
(1.1)
(0.9)
(1.0)
(1.1)
(1.0)
(1.0)
Total Net Flux
(7.1)
(3.1)
(3.2)
(3.3)
(3.3)
(3.3)
(3.2)
Notes: Totals may not sum due to independent rounding. Parentheses indicate net sequestration.
Methodology
When wastes of biogenic origin (such as yard trimmings and food scraps) are landfilled and do not completely
decompose, the C that remains is effectively removed from the C cycle. Empirical evidence indicates that yard
trimmings and food scraps do not completely decompose in landfills (Barlaz 1998, 2005, 2008; De la Cruz and
Barlaz 2010), and thus the stock of C in landfills can increase, with the net effect being a net atmospheric removal of
C. Estimates of net C flux resulting from landfilled yard trimmings and food scraps were developed by estimating
the change in landfilled C stocks between inventory years and are based on methodologies presented for the Land
Use, Land-Use Change, and Forestry sector in IPCC (2003) and the 2006IPCC Guidelines for National
Greenhouse Gas Lnventories (IPCC 2006). Carbon stock estimates were calculated by determining the mass of
landfilled C resulting from yard trimmings and food scraps discarded in a given year; adding the accumulated
landfilled C from previous years; and subtracting the mass of C that was landfilled in previous years and has since
decomposed and been emitted as CO2 and CH4.
To determine the total landfilled C stocks for a given year, the following data and factors were assembled: (1) The
composition of the yard trimmings; (2) the mass of yard trimmings and food scraps discarded in landfills; (3) the C
storage factor of the landfilled yard trimmings and food scraps; and (4) the rate of decomposition of the degradable
C. The composition of yard trimmings was assumed to be 30 percent grass clippings, 40 percent leaves, and 30
percent branches on a wet weight basis (Oshins and Block 2000). The yard trimmings were subdivided, because
each component has its own unique adjusted C storage factor (i.e., moisture content and C content) and rate of
decomposition. The mass of yard trimmings and food scraps disposed of in landfills was estimated by multiplying
the quantity of yard trimmings and food scraps discarded by the proportion of discards managed in landfills. Data on
discards (i.e., the amount generated minus the amount diverted to centralized composting facilities) for both yard
trimmings and food scraps were taken primarily from Advancing Sustainable Materials Management: Facts and
Figures 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. Since the Advancing Sustainable
6-116 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Materials Management: Facts and Figures reports for 2016 and 2017 were unavailable, landfilled material
generation, recovery, and disposal data for 2016 and 2017 were set equal to 2015 values.
The amount of C disposed of in landfills each year, starting in 1960, was estimated by converting the discarded
landfilled yard trimmings and food scraps from a wet weight to a dry weight basis, and then multiplying by the
initial (i.e., pre-decomposition) C content (as a fraction of dry weight). The dry weight of landfilled material was
calculated using dry weight to wet weight ratios (Tchobanoglous et al. 1993, cited by Barlaz 1998) and the initial C
contents and the C storage factors were determined by Barlaz (1998, 2005, 2008) (Table 6-81).
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 6-81).
The modeling approach applied to simulate U.S. landfill C flows builds on the findings of Barlaz (1998, 2005,
2008). The proportion of C stored is assumed to persist in landfills. The remaining portion is assumed to degrade
over time, resulting in emissions of CH4 and CO2. (The CH4 emissions resulting from decomposition of yard
trimmings and food scraps are reported in the Waste chapter.) The degradable portion of the C is assumed to decay
according to first-order kinetics. The decay rates for each of the materials are shown in Table 6-81.
The first-order decay rates, k. for each refuse type were 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, 1990 waste composition
for the United States fromEPA's Characterization of Municipal Solid Waste in the United States: 1990 Update
(EPA 1991) was used to calculate f. This correction factor is then multiplied 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, k= 0.12). As in the Landfills section of the Inventory (Section 7.1), which
estimates CH4 emissions, the overall MSW decay rate is estimated by partitioning the U.S. landfill population into
three categories based on annual precipitation ranges of: (1) Less than 20 inches of rain per year, (2) 20 to 40 inches
of rain per year, and (3) greater than 40 inches of rain per year. These correspond to overall MSW decay rates of
0.020, 0.038, and 0.057 year-1, respectively.
De la Cruz and Barlaz (2010) calculate component-specific decay rates corresponding to the first value (0.020
year1), but not for the other two overall MSW decay rates. To maintain consistency between landfill methodologies
across the Inventory, the correction factors (J) were developed for decay rates of 0.038 and 0.057 year1 through
linear interpolation. A weighted national average component-specific decay rate was calculated by assuming that
decay rates differ for populations that live in differing annual precipitation categories, and 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 6-81.
Land Use, Land-Use Change, and Forestry 6-117
-------
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:
LFCi,t= Ł Wi,n x (1 - MO) x ICOx {[CSix ICG[ + [(1 - (CS,x ICG)) x e-^-")]}
n
where,
MQ
CSi
ICC,
LFCij
WUn
k
n
t
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 I. for waste i (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
-------
Table 6-82: C Stocks in Yard Trimmings and Food Scraps in Landfills (MMT C)
Carbon Pool
1990
2005
2013
2014
2015
2016
2017
Yard Trimmings
156.0
203.1
221.1
223.4
225.7
228.0
230.3
Branches
14.6
18.1
19.8
20.0
20.2
20.4
20.6
Leaves
66.7
87.4
95.6
96.6
97.7
98.7
99.8
Grass
74.7
97.7
105.8
106.8
107.8
108.9
109.9
Food Scraps
17.9
33.2
41.2
42.2
43.3
44.3
45.3
Total Carbon
Stocks
173.9
236.3
262.3
265.7
269.0
272.3
275.5
Note: Totals may not sum due to
independent rounding.
Uncertainty and Time-Series Consistency
The uncertainty analysis for landfilled yard trimmings and food scraps includes an evaluation of the effects of
uncertainty for the following data and factors: disposal in landfills per year (tons of C), initial C content, moisture
content, decay rate, and proportion of C stored. The C storage landfill estimates are also a function of the
composition of the yard trimmings (i.e., the proportions of grass, leaves and branches in the yard trimmings
mixture). There are respective uncertainties associated with each of these factors.
A Monte Carlo (Approach 2) uncertainty analysis was applied to estimate the overall uncertainty of the
sequestration estimate for 2017. The results of the Approach 2 quantitative uncertainty analysis are summarized in
Table 6-83. Total yard trimmings and food scraps CO2 flux in 2017 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 58 percent below to 59 percent above the 2017 flux estimate of -11.9 MMT CO2 Eq.
Table 6-83: Approach 2 Quantitative Uncertainty Estimates for CO2 Flux from Yard
Trimmings and Food Scraps in Landfills (MMT CO2 Eq. and Percent)
2017 Flux
Source
Gas
Estimate
Uncertainty Range Relative to Flux Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower Upper
Lower Upper
Bound Bound
Bound Bound
Yard Trimmings and
Food Scraps
CO2
(11.9)
(18.9) (4.9)
-58% 59%
a Range of flux estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Note: Parentheses indicate negative values or net C sequestration.
QA/QC and Verification
Tier 1 and Tier 2 QA/QC activities were conducted consistent with the U.S. QA/QC plan. Source-specific quality
control measures for Landfilled Yard Trimmings and Food Scraps included checking that input data were properly
transposed within the spreadsheet, checking calculations were correct, and confirming that all activity data and
calculations documentation was complete and updated to ensure data were properly handled through the inventory
process.
Order of magnitude checks and checks of time-series consistency were performed to ensure data were updated
correctly and any changes in emissions estimates were reasonable and reflected changes in activity data. An annual
change trend analysis was also conducted to ensure the validity of the emissions estimates. Errors that were found
during this process were corrected as necessary.
Recalculations Discussion
EPA made the following recalculations:
• The current Inventory lias been revised to reflect updated data from the most recent Advancing Sustainable
Materials Management: Facts and Figures report.
Land Use, Land-Use Change, and Forestry 6-119
-------
• Decay rates (presented in Table 6-81) were also updated using new population distributions from the 2010
U.S. Census.
Recalculations based on these updates resulted in less than 1.0 percent change in the annual carbon stocks and
sequestration values as compared to the previous inventory values, except for 2014 and 2015. The largest changes
occurred in the most recent years: a 1.4 percent increase in sequestration in 2014, a 4.3 percent increase in
sequestration in 2015, and a 0.88 percent decrease in sequestration in 2016.
Planned Improvements
Future work is planned to evaluate the consistency between the estimates of C storage described in this chapter and
the estimates of landfill CH4 emissions described in the Waste chapter. For example, the Waste chapter does not
distinguish landfill CH4 emissions from yard trimmings and food scraps separately from landfill CH4 emissions from
total bulk (i.e., municipal solid) waste, which includes yard trimmings and food scraps. In future years, as time and
resources allow, EPA will further evaluate both categories to ensure consistency.
In addition, data from recent peer-reviewed literature will be evaluated that may modify the default C storage
factors, initial C contents, and decay rates for yard trimmings and food scraps in landfills. Based upon this
evaluation, changes may be made to the default values.
EPA will also investigate updates to the decay rate estimates for food scraps, leaves, grass, and branches. Currently
the inventory calculations use 2010 U.S. Census data. EPA will evaluate using decay rates that vary over time based
on Census data changes over time.
Yard waste composition will also be investigated to determine if changes need to be made based on changes in
residential practices, a review of available literature will be conducted to determine if there are changes in the
allocation of yard trimmings. For example, leaving grass clippings in place is becoming a more common practice,
thus reducing the percentage of grass clippings in yard trimmings disposed in landfills. In addition, agronomists may
be consulted for determining the mass of grass per acre on residential lawns to provide an estimate of total grass
generation for comparison with Inventory estimates.
Finally, EPA will review available data to ensure all types of landfilled yard trimmings and food scraps are being
included in Inventory estimates, such as debris from road construction and commercial food waste not included in
other chapter estimates.
6.11 Land Converted to Settlements (CRF
Category 4E2)
Land Converted to Settlements includes all settlements in an Inventory year that had been in another land use(s)
during the previous 20 years (USDA-NRCS 20 1 5).76 For example, cropland, grassland or forest land converted to
settlements 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 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 analysis77.
76 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.
77 For the land representation, land use data for 2013 to 2017 were only partially updated based on new Forest Inventory and
Analysis (FIA) data. These updates led to changes in the land representation data for settlements through the process of
6-120 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Land use change can lead to large losses of carbon (C) to the atmosphere, particularly conversions from forest land
(Houghton et al. 1983). Moreover, conversion of forest to another land use (i.e., deforestation) is one of the largest
anthropogenic sources of emissions to the atmosphere globally (Schimel 1995), although this source may be
declining globally according to a recent assessment (Tubiello et al. 2015).
IPCC (2006) recommends reporting changes inbiomass, 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 2017, accounting for
approximately 74 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 2017 are 37.5, 7.4,
6.9, and 10.2 MMT CO2 Eq. (10.2, 2.0, 1.9, and 2.8 MMT C). Mineral and organic soils also lost 22.5 and 1.8 MMT
CO2 Eq. in 2017 (6.1 and 0.5 MMT C). The total net flux is 86.2 MMT C02 Eq. in 2017 (23.5 MMT C), which is a
37 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 6-84: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Settlements (MMT CO2 Eq.)
1990
2005
2013
2014
2015
2016
2017
Cropland Converted to
Settlements
4.1
11.9
10.3
10.2
10.2
10.1
10.1
Mineral Soils
3.5
10.7
9.4
9.4
9.3
9.3
9.2
Organic Soils
0.6
1.2
0.9
0.9
0.8
0.9
0.8
Forest Land Converted to
Settlements
54.7
59.9
62.9
63.2
63.2
63.2
63.2
Aboveground Live Biomass
32.6
35.5
37.2
37.5
37.5
37.5
37.5
Belowground Live Biomass
6.4
7.0
7.3
7.4
7.4
7.4
7.4
Dead Wood
5.9
6.5
6.8
6.9
6.9
6.9
6.9
Litter
8.9
9.7
10.1
10.2
10.2
10.2
10.2
Mineral Soils
0.9
1.3
1.3
1.3
1.3
1.3
1.3
Organic Soils
+
+
0.1
+
+
+
+
Grassland Converted
Settlements
4.0
13.5
12.4
12.4
12.4
12.3
12.2
Mineral Soils
3.5
12.3
11.5
11.5
11.4
11.4
11.3
Organic Soils
0.5
1.2
0.9
0.9
0.9
0.9
0.9
Other Lands Converted to
Settlements
0.2
0.7
0.7
0.7
0.7
0.7
0.7
Mineral Soils
0.2
0.6
0.6
0.6
0.6
0.6
0.6
Organic Soils
+
0.1
0.1
0.1
0.1
0.1
0.1
Wetlands Converted to
Settlements
+
0.1
0.1
0.1
0.1
0.1
0.1
Mineral Soils
+
0.1
0.1
0.1
0.1
0.1
0.1
Organic Soils
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Total Aboveground Biomass
Flux
32.6
35.5
37.2
37.5
37.5
37.5
37.5
Total Belowground Biomass
Flux
6.4
7.0
7.3
7.4
7.4
7.4
7.4
Total Dead Wood Flux
5.9
6.5
6.8
6.9
6.9
6.9
6.9
Total Litter Flux
8.9
9.7
10.1
10.2
10.2
10.2
10.2
Total Mineral Soil Flux
8.0
24.9
22.9
22.8
22.7
22.6
22.5
Total Organic Soil Flux
1.1
2.5
2.0
1.9
1.9
1.9
1.8
combining FIA data with land use data from the National Resources Inventory and National Land Cover Dataset (See
"Representation of the U.S. Land Base" section for more information). However, an inventory was not compiled for settlements
in this Inventory, but rather the emissions and removals are based on a surrogate data method. Therefore, the area estimates in
this section are based on the land representation data from the previous Inventory.
Land Use, Land-Use Change, and Forestry 6-121
-------
Total Net Flux 623 86J) 86.4 86.5 86.5 86.4 86.2
+ Does not exceed 0.05 MMT CO2 Eq.
Table 6-85: Net CO2 Flux from Soil, Dead Organic Matter and Biomass C Stock Changes for
Land Converted to Settlements (MMT C)
1990
2005
2013
2014
2015
2016
2017
Cropland Converted to
Settlements
1.1
3.2
2.8
2.8
2.8
2.8
2.7
Mineral Soils
0.9
2.9
2.6
2.6
2.5
2.5
2.5
Organic Soils
0.2
0.3
0.2
0.2
0.2
0.2
0.2
Forest Land Converted to
Settlements
14.9
16.3
17.2
17.2
17.2
17.2
17.2
Aboveground Live Biomass
8.9
9.7
10.2
10.2
10.2
10.2
10.2
Belowground Live Biomass
1.7
1.9
2.0
2.0
2.0
2.0
2.0
Dead Wood
1.6
1.8
1.9
1.9
1.9
1.9
1.9
Litter
2.4
2.6
2.8
2.8
2.8
2.8
2.8
Mineral Soils
0.3
0.4
0.3
0.3
0.3
0.3
0.3
Organic Soils
Grassland Converted
+
+
+
+
+
+
+
Settlements
1.1
3.7
3.4
3.4
3.4
3.4
3.3
Mineral Soils
0.9
3.4
3.1
3.1
3.1
3.1
3.1
Organic Soils
0.1
0.3
0.2
0.2
0.3
0.2
0.2
Other Lands Converted to
Settlements
+
0.2
0.2
0.2
0.2
0.2
0.2
Mineral Soils
+
0.2
0.2
0.2
0.2
0.2
0.2
Organic Soils
Wetlands Converted to
+
+
+
+
+
+
+
Settlements
+
+
+
+
+
+
+
Mineral Soils
+
+
+
+
+
+
+
Organic Soils
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Total Aboveground Biomass
Flux
8.9
9.7
10.2
10.2
10.2
10.2
10.2
Total Belowground Biomass
Flux
1.7
1.9
2.0
2.0
2.0
2.0
2.0
Total Dead Wood Flux
1.6
1.8
1.9
1.9
1.9
1.9
1.9
Total Litter Flux
2.4
2.6
2.8
2.8
2.8
2.8
2.8
Total Mineral Soil Flux
2.2
6.8
6.2
6.2
6.2
6.2
6.1
Total Organic Soil Flux
0.3
0.7
0.5
0.5
0.5
0.5
0.5
Total Net Flux
17.2
23.5
23.6
23.6
23.6
23.6
23.5
+ Does not exceed 0.05 MMT CO2 Eq.
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 USD A Forest Service, Forest Inventory and Analysis (FIA) program (USD A 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
6-122 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 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
2012 USDA NRI survey for non-federal lands (USDA-NRCS 2015). 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 2012 (USDA-NRCS
2015). However, this Inventory only uses NRI data through 2012 because newer data were not available.
NRI survey locations are classified as Land Converted to Settlements in a given year between 1990 and 2012 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 grassland between 1971 and 1978. For federal
lands, the land use history is derived from land cover changes in the National Land Cover Dataset (Homer et al.
2007; Fry et al. 2011; Homer et al. 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 2012. 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 2012 NRI survey data (USDA-NRCS 2015) 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 the loss of soil C with conversion to settlements, which is similar to the estimated losses with
conversion to cropland. 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 2013 to 2017
because NRI activity data are not available for these years. Specifically, a linear regression model with
autoregressive moving-average (ARMA) errors (Brockwell and Davis 2016) is used to estimate the trend in stock
changes over time from 1990 to 2012, and in turn, the trend is used to approximate stock changes from 2013 to
Land Use, Land-Use Change, and Forestry 6-123
-------
2017. The Tier 2 method described previously will be applied to recalculate the 2013 to 2017 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 2012, the total 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 2013 to
2017 because NRI activity data are not available for these years to determine the area of Land Converted to
Settlements.
Uncertainty and Time-Series Consistency
The uncertainty analysis for C losses with Forest Land Converted to Settlements is conducted in the same way as the
uncertainty assessment for forest ecosystem C flux in the Forest Land Remaining Forest Land category. Sample and
model-based error are combined using simple error propagation methods provided by the IPCC (2006), i.e., by
taking the square root of the sum of the squares of the standard deviations of the uncertain quantities. For additional
details see the Uncertainty Analysis in Annex 3.13. The uncertainty analysis for mineral soil 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 6-86 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
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 2013 to
2017. The combined uncertainty for total C stocks in Land Converted to Settlements ranges from 29 percent below
to 29 percent above the 2017 stock change estimate of 86.2 MMT CO2 Eq.
Table 6-86: 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)
2017 Flux Estimate Uncertainty Range Relative to Flux Estimate3
Source (MMT CO2 Eq.) (MMT CO2 Eq.) (%)
Lower
Upper
Lower
Upper
Bound
Bound
Bound
Bound
Cropland Converted to Settlements
10.1
7.8
12.4
-23%
23%
Mineral Soil C Stocks
9.2
6.9
11.5
-25%
25%
Organic Soil C Stocks
0.8
0.5
1.1
-36%
36%
Forest Land Converted to Settlements
63.2
38.2
88.1
-40%
39%
Aboveground Biomass C Stocks
37.5
14.2
60.8
-62%
62%
Belowground Biomass C Stocks
7.4
2.8
11.9
-62%
62%
Dead Wood
6.9
2.6
11.1
-62%
62%
Litter
10.2
3.9
16.5
-62%
62%
Mineral Soil C Stocks
1.3
1.0
1.5
-20%
20%
Organic Soil C Stocks
+
+
+
-39%
39%
Grassland Converted to Settlements
12.2
9.7
14.8
-21%
21%
Mineral Soil C Stocks
11.3
8.8
13.9
-22%
22%
Organic Soil C Stocks
0.9
0.5
1.2
-41%
41%
Other Lands Converted to Settlements
0.7
0.5
0.9
-24%
24%
Mineral Soil C Stocks
0.6
0.5
0.7
-24%
24%
Organic Soil C Stocks
0.1
+
0.2
-80%
80%
Wetlands Converted to Settlements
0.1
+
0.1
-42%
42%
6-124 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Mineral Soil C Stocks
0.1
+
0.1
-42%
42%
Organic Soil C Stocks
+
+
+
0%
0%
Total: Land Converted to Settlements
86.2
61.0
111.4
-29%
29%
Aboveground Biomass C Stocks
37.5
14.2
60.8
-62%
62%
Belowground Biomass C Stocks
7.4
2.8
11.9
-62%
62%
Dead Wood
6.9
2.6
11.1
-62%
62%
Litter
10.2
3.9
16.5
-62%
62%
Mineral Soil C Stocks
22.5
19.1
25.9
-15%
15%
Organic Soil C Stocks
1.8
1.1
2.5
-38%
38%
+ Does not exceed 0.05 MMT CO2 Eq.
a Range of C stock change estimates predicted by Monte Carlo Stochastic Simulation for a 95 percent confidence interval.
Methodological recalculations are applied to the latter part of the time series (2013 to 2017) using the linear time
series model described above. Details on the emission trends through time are described in more detail in the
Methodology section, above.
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.
Recalculations Discussion
Methodological recalculations are associated with extending the time series from 2013 through 2017 using a linear
time series model, and an update of bio mass and dead organic matter losses with Forest Land Converted to
Settlements. The recalculation led to a 31 percent greater loss of C on average. This change is almost entirely
attributed to the update of biomass and dead organic matter losses for Forest Land Converted to Settlements with
newly available re-measurement data for the western United States. New stock changes were estimated at the plot-
level with the new data consistent with the compilation methods described in the Forest Land Remaining Forest
Land section. In the previous Inventory, state-level averages from the plot data had been used to approximate the
losses of C with Forest Land Converted to Settlements due to a lack of re-measurement data.
Planned Improvements
A planned improvement for the Land Converted to Settlements category is to develop an inventory of C stock
changes in Alaska. This includes C stock changes for biomass, dead organic matter and soils. There are plans to
improve classification of urban trees in settlements and to include transfer of biomass from forest land to those areas
in this category. There are also plans to extend the Inventory to included C losses associated with drained organic
soils in settlements occurring on federal lands. New land representation data will also be compiled, and the time
series recalculated for the latter years in the time series that are estimated using data splicing methods in this
Inventory.
6.12 Other Land Remaining Other Land fCRF
Category 4F1)
Land use is constantly occurring, and areas under a number of differing land-use types remain in their respective
land-use type each year, just as other land can remain as other land. While the magnitude of Other Land Remaining
Other Land is known (see Table 6-7), 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, CH4 or N2O fluxes on Other Land Remaining Other Land at this time.
Land Use, Land-Use Change, and Forestry 6-125
-------
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, CH4 or N20 fluxes on Land Converted to
Other Land from fluxes on Other Land Remaining Other Land at this time.
6-126 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
7. Waste
Waste management and treatment activities are sources of greenhouse gas emissions (see Figure 7-1). Landfills
accounted for approximately 16.4 percent of total U.S. anthropogenic methane (CH4) emissions in 2017, the third
largest contribution of any CH4 source in the United States. Additionally, wastewater treatment and composting of
organic waste accounted for approximately 2.2 percent and 0.3 percent of U.S. CH4 emissions, respectively. Nitrous
oxide (N20) emissions from the discharge of wastewater treatment effluents into aquatic enviromnents were
estimated, as were N20 emissions from the treatment process itself. Nitrous oxide emissions from composting were
also estimated. Together, these waste activities account for 1.9 percent of total U.S. N20 emissions. Nitrogen oxides
(NOx), carbon monoxide (CO), and non-CH4 volatile organic compounds (NMVOCs) are emitted by waste
activities, and are addressed separately at the end of this chapter. A summary of greenhouse gas emissions from the
Waste chapter is presented in Table 7-1 and Table 7-2.
Figure 7-1: 2017 Waste Chapter Greenhouse Gas Sources (MMT CO2 Eq.)
Landfills
Wastewater Treatment
Composting
108
Waste as a Portion of All
Emissions
2.0%
0 10 20 30 40 50 60 70 80 90 100 110
MMT COz Eq.
Overall, in 2017, waste activities generated emissions of 131.0 MMT CO2 Eq., or 2.0 percent of total U.S.
greenhouse gas emissions.1
Table 7-1: Emissions from Waste (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CH4
195.2
148.7
129.3
128.9
127.8
124.3
124.1
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
-------
Landfills
179.6
131.4
112.9
112.5
111.2
108.0
107.7
Wastewater Treatment
15.3
15.4
14.3
14.3
14.5
14.2
14.2
Composting
0.4
1.9
2.0
2.1
2.1
2.1
2.2
N2O
3.7
6.1
6.5
6.6
6.7
6.8
6.9
Wastewater Treatment
3.4
4.4
4.7
4.8
4.8
4.9
5.0
Composting
0.3
1.7
1.8
1.9
1.9
1.9
1.9
Total
198.9
154.7
135.8
135.6
134.5
131.1
131.0
Note: Totals may not sum due to independent rounding.
ible 7-2: Emissions from Waste (kt)
Gas/Source
1990
2005
2013
2014
2015
2016
2017
CH4
7,808
5,947
5,171
5,157
5,112
4,972
4,963
Landfills
7,182
5,256
4,517
4,502
4,448
4,319
4,309
Wastewater Treatment
611
616
572
572
579
568
568
Composting
15
75
81
84
85
85
86
N2O
12
20
22
22
22
23
23
Wastewater Treatment
11
15
16
16
16
16
17
Composting
1
6
6
6
6
6
6
Note: Totals may not sum due to independent rounding.
Carbon dioxide (CO2), CH4, and N20 emissions from the incineration of waste are accounted for in the Energy
sector rather than in the Waste sector because almost all incineration of municipal solid waste (MSW) in the United
States occurs at waste-to-energy facilities where useful energy is recovered. Similarly, the Energy sector also
includes an estimate of emissions from burning waste tires and hazardous industrial waste, because virtually all of
the combustion occurs in industrial and utility boilers that recover energy. The incineration of waste in the United
States in 2017 resulted in 11.1 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.
Box 7-1: 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 and sinks provided in this Inventory do 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.
Box 7-2: Waste Data from EPA's Greenhouse Gas Reporting Program
On October 30, 2009, the U.S. Enviromnental 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). 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. Annual reporting is at the facility level, except for certain
suppliers of fossil fuels and industrial greenhouse gases. Data reporting by affected facilities includes the
7-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
reporting of emissions from fuel combustion at that affected facility. In general, the threshold for reporting is
25,000 metric tons or more of CO2 Eq. per year.
EPA presents the data collected by its GHGRP through a data publication tool that allows data to be viewed in
several formats including maps, tables, charts and graphs for individual facilities or groups of facilities.2
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. Within the Waste Chapter, EPA uses directly reported GHGRP data
for net CH4 emissions from MSW landfills for the years 2010 to 2017 of the Inventory. This data is also used to
back-cast emissions from MSW landfills for the years 2005 to 2009.
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 7-3. Disposing of waste in illegal dumping sites is not considered to have occurred in years later
than 1980 and these sites are not considered to contribute to net emissions in this section for the timeframe of 1990
to the current Inventory year. MSW landfills, or sanitary landfills, are sites where MSW is managed to prevent or
minimize health, safety, and environmental impacts. Waste is deposited in different cells and covered daily with
soil; many have environmental monitoring systems to track performance, collect leachate, and collect landfill gas.
Industrial waste landfills are constructed in a similar way as MSW landfills, but are used to dispose of industrial
solid waste, such as RCRA Subtitle D wastes (e.g., non-hazardous industrial solid waste defined in Title 40 of the
Code of Federal Regulations or CFR in section 257.2), commercial solid wastes, or conditionally exempt small-
quantity generator wastes (EPA 2016a).
After being placed in a landfill, organic waste (such as paper, food scraps, and yard trimmings) is initially
decomposed by aerobic bacteria. After the oxygen has been depleted, the remaining waste is available for
consumption by anaerobic bacteria, which break down organic matter into substances such as cellulose, amino acids,
and sugars. These substances are further broken down through fermentation into gases and short-chain organic
compounds that form the substrates for the growth of methanogenic bacteria. These methane (CH4) producing
anaerobic bacteria convert the fermentation products into stabilized organic materials and biogas consisting of
approximately 50 percent biogenic carbon dioxide (CO2) and 50 percent CH4, by volume. Landfill biogas also
contains trace amounts of non-methane organic compounds (NMOC) and volatile organic compounds (VOC) that
either result from decomposition byproducts or volatilization of biodegradable wastes (EPA 2008).
Methane and CO2 are the primary constituents of landfill gas generation and emissions. However, the 2006 IPCC
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 (N20) emissions from
2 See .
Waste 7-3
-------
the disposal and application of sewage sludge on landfills are also not explicitly modeled as part of greenhouse gas
emissions from landfills. Nitrous oxide emissions from sewage sludge applied to landfills as a daily cover or for
disposal are expected to be relatively small because the microbial environment in an anaerobic landfill is not very
conducive to the nitrification and denitrification processes that result in N20 emissions. Furthermore, the 2006IPCC
Guidelines did not include a methodology for estimating N20 emissions from solid waste disposal sites "because
they are not significant." Therefore, only CH4 generation and emissions are estimated for landfills under the Waste
sector.
Methane generation and emissions from landfills are a function of several factors, including: (1) the total amount
and composition of waste-in-place, which is the total waste landfilled annually over the operational lifetime of a
landfill; (2) the characteristics of the landfill receiving waste (e.g., size, climate, cover material); (3) the amount of
CH4 that is recovered and either flared or used for energy purposes; and (4) the amount of CH4 oxidized as the
landfill gas - that is not collected by a gas collection system - passes through the cover material into the atmosphere.
Each landfill has unique characteristics, but all managed landfills employ similar operating practices, including the
application of a daily and intermediate cover material over the waste being disposed of in the landfill to prevent odor
and reduce risks to public health. Based on recent literature, the specific type of cover material used can affect the
rate of oxidation of landfill gas (RTI2011). 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 2017, landfill CHi emissions were approximately 107.7 MMT CO2 Eq. (4,309 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, while industrial waste landfills
accounted for the remainder. Estimates of operational MSW landfills in the United States have ranged from 1,700 to
2,000 facilities (EPA 2018a; EPA 2018c; 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 2018a; 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
2018b; 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 172 facilities with industrial waste landfills met the reporting threshold under Subpart TT (Industrial Waste
Landfills) of EPA's Greenhouse Gas Reporting Program (GHGRP), indicating that there may be several hundred
industrial waste landfills that are not required to report under EPA's GHGRP.
The annual amount of MSW generated and subsequently disposed in MSW landfills varies annually and depends on
several factors (e.g., the economy, consumer patterns, recycling and composting programs, inclusion in a garbage
collection service). The estimated annual quantity of waste placed in MSW landfills increased 10 percent from
approximately 205 MMT in 1990 to 226 MMT in 2000 and then decreased by 8.8 percent to 206 MMT in 2017 (see
Annex 3.14, Table A-235). The total amount of MSW generated is expected to increase as the U.S. population
continues to grow, but the percentage of waste landfilled may decline due to increased recycling and composting
practices. Net CH4 emissions from MSW landfills have decreased since 1990 (see Table 7-3 and Table 7-4).
The estimated quantity of waste placed in industrial waste landfills (from the pulp and paper, and food processing
sectors) has remained relatively steady since 1990, ranging from 9.7 MMT in 1990 to 10.2 MMT in 2017 (see
Annex 3.14, Table A-235). CH4 emissions from industrial waste landfills have also remained at similar levels
recently, ranging from 14.3 MMT in 2005 to 15.9 MMT in 2017 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 2017, LMOP identified 15 new landfill gas-to-
energy (LFGE) projects (EPA 2018a) that began operation. While the amount of landfill gas collected and
7-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
combusted continues to increase, the rate of increase in collection and combustion no longer exceeds the rate of
additional CH4 generation from the amount of organic MSW landfilled as the U.S. population grows (EPA 2018b).
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
2013
2014
2015
2016
2017
MSW CH4 Generation
205.3
-
-
-
-
-
-
Industrial CH4 Generation
12.1
15.9
16.5
16.6
16.6
16.6
16.6
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)
-
117.1
98.1
97.6
96.3
93.0
92.8
Total
179.6
131.4
112.9
112.5
111.2
108.0
107.7
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 uses the first order decay methodology. A methodological change occurs in year 2005. For
years 2005 to 2017, 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. As such, CH4 generation, CH4 recovery, and CH4 oxidized are not calculated separately for 2005 to
2017. See the Time-Series Consistency section of this chapter for more information.
Table 7-4: ChU Emissions from Landfills (kt)
Activity
1990
2005
2013
2014
2015
2016
2017
MSW CH4 Generation
8,214
-
-
-
-
-
-
Industrial CH4 Generation
484
636
661
662
663
664
665
MSW CH4 Recovered
(718)
-
-
-
-
-
-
MSW CH4 Oxidized
(750)
-
-
-
-
-
-
Industrial CH4 Oxidized
MSW net CH4 Emissions
(48)
(64)
(66)
(66)
(66)
(66)
(67)
(GHGRP)
-
4,684
3,923
3,906
3,851
3,722
3,711
Total
7,182
5,256
4,517
4,502
4,448
4,319
4,309
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 uses the first order decay methodology. A methodological change occurs in year 2005. For
years 2005 to 2017, 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. 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.
Waste 7-5
-------
CH4, solid waste = Net CH4 emissions from solid waste
CH4,msw = CH4 generation from MSW landfills
CH4jnd = 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 CFRPart 98.343. The two parts of the equation consider the portion of CH4inthe landfill gas that is not
collected by the landfill gas collection system, and the portion that is collected. First, the recovered CH4 is adjusted
with the collection efficiency of the gas collection and control system and the fraction of hours the recovery system
operated in the calendar year. This quantity represents the amount of CH4 in the landfill gas that is not captured by
the collection system; this amount is then adjusted for oxidation. The second portion of the equation adjusts the
portion of CH4 in the collected landfill gas with the efficiency of the destruction device(s), and the fraction of hours
the destruction device(s) operated during the year.
CH4,soHd Waste = [(—~f r) x(l - OX) + R x (l - (DE x fDest))\
\CE X f REC J v J
where,
CH4j solid waste = Net CH4 emissions from solid waste
R = Quantity of recovered CH4 from Equation HH-4 of EPA's GHGRP
CE = Collection efficiency estimated at the landfill, considering system coverage, operation,
and cover system materials from Table HH-3 of EPA's GHGRP. If area by soil cover type
information is not available, the default value of 0.75 should be used, (percent)
fREc = fraction of hours the recovery system was operating (percent)
OX = oxidation factor (percent)
DE = destruction efficiency (percent)
Fi )esi = fraction of hours the destruction device was operating (fraction)
The current Inventory uses both methods to estimate CH4 emissions across the time series. Prior to the 1990 through
2015 Inventory, only the FOD method was used. Methodological changes were made to the 1990 through 2015
Inventory to incorporate higher tier data (i.e., directly reported CH4 emissions to EPA's GHGRP), which cannot be
directly applied to earlier years in the time series without significant bias. The technique used to merge the directly
reported GHGRP data with the previous methodology is described as the overlap technique in the Time-Series
Consistency chapter of the 2006IPCC Guidelines. Additional details on the technique used is included in the Time
Series Consistency section of this chapter and a technical memorandum (RTI2017).
A summary of the methodology used to generate the current 1990 through 2017 Inventory estimates for MSW
landfills is as follows and 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 EP A's Anthropogenic
Methane Emissions in the United States, Estimates for 1990: Report to Congress (EPA 1993) and an
extensive landfill survey by the EPA's Office of Solid Waste in 1986 (EPA 1988). Although waste placed
in landfills in the 1940s and 1950s contributes very little to current CH4 generation, estimates for those
years were included in the FOD model for completeness in accounting for CH4 generation rates and are
based on the population in those years and the per capita rate for land disposal for the 1960s. For the
Inventory calculations, wastes landfilled prior to 1980 were broken into two groups: wastes disposed in
managed, anaerobic landfills (Methane Conversion Factor, MCF, of 1) and those disposed in uncategorized
solid waste disposal waste sites (MCF of 0.6) (IPCC 2006). Uncategorized sites represent those sites for
which limited information is known about the management practices. All calculations after 1980 assume
waste is disposed in managed, anaerobic landfills. The FOD method was applied to estimate annual CH4
generation. Methane recovery amounts were then subtracted and the result was then adjusted with a 10
percent oxidation factor to derive the net emissions estimates.
7-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• 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 inBioCycle 2006). In-between years were
interpolated based on population growth. For years 1989 to 2000, directly reported total MSW generation
data were used; for other years, the estimated MSW generation (excluding construction and demolition
waste and inerts) were presented in the reports and used in the Inventory. The FOD method was applied to
estimate annual CH4 generation. Landfill-specific CH4 recovery amounts were then subtracted from CH4
generation and the result was then adjusted with a 10 percent oxidation factor to derive the net emissions
estimates.
• 2005 through 2009: Emissions for these years are estimated using net CH4 emissions that are reported by
landfill facilities under EPA's GHGRP. Because not all landfills in the United States are required to report
to EPA's GHGRP, a 9 percent scale-up factor is applied to the GHGRP emissions for completeness.
Supporting information, including details on the technique used to estimate emissions for 2005 to 2009 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 depending on their facility-specific
calculated CH4 flux rate (i.e., 0, 10, 25, or 35 percent). The average oxidation factor from the GHGRP
facilities is 19.5 percent.
• 2010 through 2017: Directly reported net CH4 emissions 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
by incorporating the directly reported GHGRP emissions is presented in the Time-Series Consistency section of this
chapter and in Annex 3.14.
Description of the Methodology for MSW Landfills as Applied for 1990 to 2004
National MSW Methane Generation and Disposal Estimates
States and local municipalities across the United States do not consistently track and report quantities of MSW
generated or collected for management, nor do they report end-of-life disposal methods to a centralized system.
Therefore, national MSW landfill waste generation and disposal data are obtained from secondary data, specifically
the SOG surveys, published approximately every two years, with the most recent publication date of 2014. The SOG
survey was the only continually updated nationwide survey of waste disposed in landfills in the United States and
was the primary data source with which to estimate nationwide CH4 generation from MSW landfills. Currently,
EPA's GHGRP waste disposal data and MSW management data published by EREF are available.
The SOG surveys collect data from the state agencies and then apply the principles of mass balance where all MSW
generated is equal to the amount of MSW landfilled, combusted in waste-to-energy plants, composted, and/or
recycled (BioCycle 2006; Shin 2014). This approach assumes that all waste management methods are tracked and
reported to state agencies. Survey respondents are asked to provide a breakdown of MSW generated and managed
by landfilling, recycling, composting, and combustion (in waste-to-energy facilities) in actual tonnages as opposed
to reporting a percent generated under each waste disposal option. The data reported through the survey have
typically been adjusted to exclude non-MSW materials (e.g., industrial and agricultural wastes, construction and
demolition debris, automobile scrap, and sludge from wastewater treatment plants) that may be included in survey
responses. While these wastes may be disposed of in MSW landfills, they are not the primary type of waste material
Waste 7-7
-------
disposed and are typically inert. In the most recent survey, state agencies were asked to provide already filtered,
MSW-only data. Where this was not possible, they were asked to provide comments to better understand the data
being reported. All state disposal data are adjusted for imports and exports across state lines where imported waste is
included in a state's total while exported waste is not. Methodological changes have occurred over the time frame
the SOG survey has been published, and this has affected the fluctuating trends observed in the data (RTI2013).
State-specific landfill MSW generation data and a national average disposal factor for 1989 through 2004 were
obtained from the SOG survey every two years (i.e., 2002, 2004 as published inBioCycle 2006). The landfill
inventory calculations start with hard numbers (where available) as presented in the SOG documentation for the
report years 2002 and 2004. In-between year waste generation is interpolated using the prior and next SOG report
data. For example, waste generated in 2003 = (waste generation in 2002 + waste generation in 2004)/2. The
quantities of waste generated across all states are summed and that value is then used as the nationwide quantity of
waste generated in each year of the time series. The SOG survey is voluntary and not all states provide data in each
survey year. To estimate waste generation for states that did not provide data in any given reporting year, one of the
following methods was used (RTI 2013):
• For years when a state-specific waste generation rate was available from the previous SOG reporting year
submission, the state-specific waste generation rate for that particular state was used.
— or —
• For years where a state-specific waste generation rate was not available from the previous SOG reporting
year submission, the waste amount is generated using the national average waste generation rate. In other
words, Waste Generated = Reporting Year Population x the National Average Waste Generation Rate
o The National Average Waste Generation Rate is determined by dividing the total reported waste
generated across the reporting states by the total population for reporting states,
o This waste generation rate may be above or below the waste generation rate for the non-reporting
states and contributes to the overall uncertainty of the annual total waste generation amounts used
in the model.
Use of these methods to estimate solid waste generated by states is a key aspect of how the SOG data was
manipulated and why the results differ for total solid waste generated as estimated by SOG and in the Inventory. In
the early years (2002 data in particular), SOG made no attempt to fill gaps for non-survey responses. For the 2004
data, the SOG team used proxy data (mainly from the WB J) to fill gaps for non-reporting states and survey
responses.
Another key aspect of the SOG survey is that it focuses on MSW-only data. The data states collect for solid waste
typically are representative of total solid waste and not the MSW-only fraction. In the early years of the SOG
survey, most states reported total solid waste rather than MSW-only waste. The SOG team, in response, "filtered"
the state-reported data to reflect the MSW-only portion.
This data source also contains the waste generation data reported by states to the SOG survey, which fluctuates from
year to year. Although some fluctuation is expected, for some states, the year-to-year fluctuations are quite
significant (>20 percent increase or decrease in some case) (RTI 2013). The SOG survey reports for these years do
not provide additional explanation for these fluctuations and the source data are not available for further assessment.
Although exact reasons for the large fluctuations are difficult to obtain without direct communication with states,
staff from the SOG team that were contacted speculate that significant fluctuations are present because the particular
state could not gather complete information for waste generation (i.e., they are missing part of recycled and
composted waste data) during a given reporting year. In addition, SOG team staff speculated that some states may
have included C&D and industrial wastes in their previous MSW generation submissions, but made efforts to
exclude that (and other non-MSW categories) in more recent reports (RTI 2013).
Recently, the EREF published a report, MSW Management in the United States, which includes state-specific
landfill MSW generation and disposal data for 2010 and 2013 using a similar methodology as the SOG surveys
(EREF 2016). Because of this similar methodology, EREF data were used to populate data for years where BioCycle
data was not available when possible. State-specific landfill waste generation data for the years in between the SOG
surveys and EREF report (e.g., 2001, 2003, etc.) were either interpolated or extrapolated based on the SOG or EREF
data and the U.S. Census population data (U.S. Census Bureau 2018).
7-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Estimates of the quantity of waste landfilled from 1989 to 2004 are determined by applying an average national
waste disposal factor to the total amount of waste generated (i.e., the SOG data). A national average waste disposal
factor is determined for each year an SOG survey is published and equals the ratio of the total amount of waste
landfilled in the United States to the total amount of waste generated in the United States. The waste disposal factor
is interpolated or extrapolated for the years in-between the SOG surveys, as is done for the amount of waste
generated for a given survey year.
The 2006IPCC Guidelines recommend at least 50 years of waste disposal data to estimate CH4 emissions. Estimates
of the annual quantity of waste landfilled for 1960 through 1988 were obtained from EPA's Anthropogenic Methane
Emissions in the United States, Estimates for 1990: Report to Congress (EPA 1993) and an extensive landfill survey
by the EPA's Office of Solid Waste in 1986 (EPA 1988). Although waste placed in landfills in the 1940s and 1950s
contributes very little to current CH4 generation, estimates for those years were included in the FOD model for
completeness in accounting for CH4 generation rates and are based on the population in those years and the per
capita rate for land disposal for the 1960s. For calculations in the current Inventory, wastes landfilled prior to 1980
were broken into two groups: wastes disposed in landfills (MCF of 1) and those disposed in uncategorized site as
(MCF of 0.6). All calculations after 1980 assume waste is disposed in managed, modern landfills. See Annex 3.14
for more details.
In the current Inventory methodology, the MSW generation and disposal data are no longer used to estimate CH4
emissions for the years 2005 to 2017 because EPA's GHGRP emissions data are now used for those years.
National Landfill Gas Recovery Estimates for MSW Landfills
The estimated landfill gas recovered per year (R) at MSW landfills for 1990 to 2004 was based on a combination of
four databases and includes recovery from flares and/or landfill gas-to-energy (LFGE) projects:
• EPA's GHGRP dataset for MSW landfills (EPA 2015a);3
• A database developed by the Energy Information Administration (EIA) for the voluntary reporting of
greenhouse gases (EIA 2007);
• A database of LFGE projects that is primarily based on information compiled by the EPA LMOP (EPA
2016b);4 and
• The flare vendor database (contains updated sales data collected from vendors of flaring equipment).
The same landfill may be included one or more times across these four databases. To avoid double- or triple-
counting CH4 recovery, the landfills across each database were compared and duplicates identified. A hierarchy of
recovery data is used based on the certainty of the data in each database. In summary, the GHGRP > EIA > LFGE >
flare vendor database. The rationale for this hierarchy is described below.
EPA's GHGRP MSW landfills database was first introduced as a data source for landfill gas recovery in the 1990 to
2013 Inventory. EPA's GHGRP MSW landfills database contains facility-reported data that undergoes rigorous
verification, thus it is considered to contain the least uncertain data of the four CH4 recovery databases. However, as
mentioned earlier, this database is unique in that it only contains a portion of the landfills in the United States
(although, presumably the highest emitters since only those landfills that meet a certain CH4 generation threshold
must report) and only contains data for 2010 and later. In the current Inventory methodology, CH4 recovery for 1990
to 2004 for facilities reporting to EPA's GHGRP has been estimated using the directly reported emissions for those
facilities from 2010 to 2015, and an Excel forecasting function so that the GHGRP data source can be applied to
earlier years in the time series. Prior to 2005, if a landfill in EPA's GHGRP was also in the LFGE or EIA databases,
the landfill gas project information, specifically the project start year, from either the LFGE or EIA databases was
used as the cutoff year for the estimated CH4 recovery in the GHGRP database. For example, if a landfill reporting
under EPA's GHGRP was also included in the LFGE database under a project that started in 2002 that is still
3 The 2015 GHGRP dataset is used to estimate landfill gas recovery from MSW landfills for the years 1990 to 2004 of the
Inventory. This database is no longer updated because the methodology has changed such that the directly reported net methane
emissions from the GHGRP are used and landfill gas recovery is not separately estimated.
4 The LFGE database was not updated for the 1990 to 2017 Inventory because the methodology does not use this database for
years 2005 and later, thus the LMOP 2016 database is the most recent year reflected in the LFGE database for the Inventory.
Waste 7-9
-------
operational, the CH4 recovery data in the GHGRP database for that facility was back-calculated to the year 2002
only.
If a landfill in the GHGRP MSW landfills database was also in the EIA, LFGE, and/or flare vendor database, the
avoided emissions were only based on EPA's GHGRP MSW landfills database to avoid double or triple counting
the recovery amounts. In other words, the CH4 recovery from the same landfill was not included in the total recovery
from the EIA, LFGE, or flare vendor databases.
If a landfill in the EIA database was also in the LFGE and/or the flare vendor database, the CH4 recovery was based
on the EIA data because landfill owners or operators directly reported the amount of CH4 recovered using gas flow
concentration and measurements, and because the reporting accounted for changes over time.
If both the flare data and LFGE recovery data were available for any of the remaining landfills (i.e., not in the EIA
or GHGRP databases), then the avoided emissions were based on the LFGE data, which provides reported landfill-
specific data on gas flow for direct use projects and project capacity (i.e., megawatts) for electricity projects. The
LFGE database is based on the most recent EPA LMOP database (published annually). The remaining portion of
avoided emissions is calculated by the flare vendor database, which estimates CH4 combusted by flares using the
midpoint of a flare's reported capacity. New flare vendor sales data have not been collected since 2015 because
these data are not used for estimates beyond 2005 in the time series. Given that each LFGE project is likely to also
have a flare, double counting reductions from flares and LFGE projects in the LFGE database was avoided by
subtracting emission reductions associated with LFGE projects for which a flare had not been identified from the
emission reductions associated with flares (referred to as the flare correction factor). A further explanation of the
methodology used to estimate the landfill gas recovered can be found in Annex 3.14.
A destruction efficiency of 99 percent was applied to CH4 recovered to estimate CH4 emissions avoided due to the
combusting of CH4 in destruction devices (i.e., flares) in the EIA, LFGE, and flare vendor databases. The median
value of the reported destruction efficiencies to the GHGRP was 99 percent for all reporting years (2010 through
2017). For the other three recovery databases, the 99 percent destruction efficiency value selected was based on the
range of efficiencies (86 to greater than 99 percent) recommended for flares in EPA's AP-42 Compilation of Air
Pollutant Emission Factors, Draft Section 2.4, Table 2.4-3 (EPA 2008). A typical value of 97.7 percent was
presented for the non-CH4 components (i.e., VOC and NMOC) in test results (EPA 2008). An arithmetic average of
98.3 percent and a median value of 99 percent are derived from the test results presented in EPA (2008). Thus, a
value of 99 percent for the destruction efficiency of flares has been used in the Inventory methodology. Other data
sources supporting a 99 percent destruction efficiency include those used to establish New Source Performance
Standards (NSPS) for landfills and in recommendations for shutdown flares used by the EPA LMOP.
National MSW Landfill Methane Oxidation Estimates
The amount of CH4 oxidized by the landfill cover at MSW landfills was assumed to be 10 percent of the CH4
generated that is not recovered (IPCC 2006; Mancinelli and McKay 1985; Czepiel et al. 1996) for the years 1990 to
2004.
National MSW Net Emissions Estimates
Net CH4 emissions are calculated by subtracting the CH4 recovered and CH4 oxidized from CH4 generated at MSW
landfills.
Description of the Methodology for MSW Landfills as Applied for 2005 to 2009
The Inventory methodology uses directly reported net CH4 emissions for the 2010 to 2017 reporting years from
EPA's GHGRP to back-cast emissions for 2005 to 2009. The emissions for 2005 to 2009 are recalculated each year
the Inventory is published to account for the additional year of reported data and any revisions that facilities make to
past GHGRP reports. When EPA verifies the greenhouse gas reports, comparisons are made with data submitted in
earlier reporting years and errors may be identified in these earlier year reports. Facility representatives may submit
revised reports for any reporting year in order to correct these errors. Facilities reporting to EPA's GHGRP that do
not have landfill gas collection and control systems use the FOD method. Facilities with landfill gas collection and
control must use both the FOD method and a back-calculation approach. The back-calculation approach starts with
7-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
the amount of CH4 recovered and works back through the system to account for gas not collected by the landfill gas
collection and control system (i.e., the collection efficiency).
A scale-up factor to account for emissions from landfills that do not report to EPA's GHGRP is also applied
annually. In theory, national MSW landfill emissions should equal the net CH4 emissions reported to the GHGRP
plus net CH4 emissions from landfills that do not report to the GHGRP. EPA estimated a scale-up factor of 9
percent. The rationale behind the 9 percent scale-up factor is included in Annex 3.14 and in (RTI 2018a).
The GHGRP data allows facilities to apply a range of oxidation factors (0.0, 0.10, 0.25, or 0.35) based on the
calculated CH4 flux at the landfill. Therefore, one oxidation factor is not applied for this time frame, as is done for
1990 to 2004. The average oxidation factor across the GHGRP data is 19.5 percent.
Description of the Methodology for MSW Landfills as Applied for 2010 to 2017
Directly reported CH4 emissions to the GHGRP are used for 2010 to 2017 plus the 9 percent scale-up factor to
account for emissions from landfills that do not report to the GHGRP. The average oxidation factor across the
GHGRP data is 19.5 percent.
Description of the First Order Decay Methodology for Industrial Waste Landfills
Emissions from industrial waste landfills are estimated from industrial production data (ERG 2018), 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 172 facilities that reported to subpart TT of the GHGRP in 2017, 93
(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, the 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).
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 two of the 172 facilities, or 1 percent of facilities, have active
gas collection systems (EPA 2018b). 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.
Uncertainty and Time-Seri insistency
Several types of uncertainty are associated with the estimates of CH4 emissions from MSW and industrial waste
landfills when the FOD method is applied directly for 1990 to 2004 in the Waste Model and, to some extent, in the
GHGRP methodology. The approach used in the MSW emission estimates assumes that the CH4 generation
potential (L0) and the rate of decay that produces CH4 from MSW, as determined from several studies of CH4
recovery at MSW landfills, are representative of conditions at U.S. MSW landfills. When this top-down approach is
Waste 7-11
-------
applied at the nationwide level, the uncertainties are assumed to be less than when applying this approach to
individual landfills and then aggregating the results to the national level. In other words, the FOD method as applied
in this Inventory is not facility-specific modeling and while this approach may over- or under-estimate CH4
generation at some landfills if used at the facility-level, the result is expected to balance out because it is being
applied nationwide.
There is a high degree of uncertainty associated with the FOD model, particularly when a homogeneous waste
composition and hypothetical decomposition rates are applied to heterogeneous landfills (IPCC 2006). There is less
uncertainty inEPA's GHGRP data because this methodology is facility-specific, uses directly measured CH4
recovery data (when applicable), and allows for a variety of landfill gas collection efficiencies, destruction
efficiencies, and/or oxidation factors to be used.
Uncertainty also exists in the scale-up 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 emissions from 2010 to 2017. 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 MS W 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. The number of published field studies measuring the rate of oxidation has increased substantially since the
2006 IPCC Guidelines were published and, as discussed in the Potential Improvements section, efforts will continue
to review the literature and revise this value, as appropriate.
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 recovery is used for facilities reporting to the GHGRP for years 2005 to 2015.
The GHGRP MSW landfills database was added as a fourth recovery database starting with the 1990 through 2013
7-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Inventory report. Relying on multiple databases for a complete picture introduces uncertainty because the coverage
and characteristics of each database differs, which increases the chance of double counting avoided emissions.
Additionally, the methodology and assumptions that go into each database differ. For example, the flare database
assumes the midpoint of each flare capacity at the time it is sold and installed at a landfill; the flare may be
achieving a higher capacity, in which case the flare database would underestimate the amount of CH4 recovered.
The LFGE database was updated annually until 2015. The flare database was populated annually until 2015 by the
voluntary sharing of flare sales data by select vendors, which likely underestimated recovery for landfills not
included in the three other recovery databases used by the Inventory. The EIA database has not been updated since
2006 and has, for the most part, been replaced by the GHGRP MSW landfills database. To avoid double counting
and to use the most relevant estimate of CH4 recovery for a given landfill, a hierarchical approach is used among the
four databases. GHGRP data and the EIA data are given precedence because facility data were directly reported; the
LFGE data are given second priority because CH4 recovery is estimated from facility-reported LFGE system
characteristics; and the flare data are given the lowest priority because this database contains minimal information
about the flare, no site-specific operating characteristics, and includes smaller landfills not included in the other
three databases (Bronstein et al. 2012). The coverage provided across the databases most likely represents the
complete universe of landfill CH4 gas recovery; however, the number of unique landfills between the four databases
does differ.
The 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 CH4 estimated to be recovered). This 10 percent uncertainty factor applies to the LFGE database; 12 percent
to the EIA database; and 1 percent for the GHGRP MSW landfills dataset because of the supporting information
provided and rigorous verification process. For flaring without metered recovery data (the flare database), a much
higher uncertainty value of 50 percent is used. The compounding uncertainties associated with the four databases in
addition to the uncertainties associated with the FOD method and annual waste disposal quantities leads to the large
upper and lower bounds for MSW landfills presented in Table 7-5.
The lack of landfill-specific information regarding the number and type of industrial waste landfills in the United
States is a primary source of uncertainty with respect to the industrial waste generation and emission estimates. The
approach used here assumes that most of the organic waste disposed of in industrial waste landfills that would result
in CH4 emissions consists of waste from the pulp and paper and food processing sectors. However, because waste
generation and disposal data are not available in an existing data source for all U.S. industrial waste landfills, a
straight disposal factor is applied over the entire time series to the amount produced to determine the amounts
disposed. Industrial waste facilities reporting under EPA's GHGRP do report detailed waste stream information, and
these data have been used to improve, for example, the DOC value used in the Inventory methodology for the pulp
and paper sector. A 10 percent oxidation factor is also applied to CH4 generation 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)
2017 Emission
Source
Gas
Estimate
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Total Landfills
CH4
107.7
95.7
151.2
-11%
+40%
MSW
ch4
92.8
69.4
116.5
-25%
+26%
Industrial
ch4
15.0
21.4
41.2
-43%
+175%
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 CH4 recovery estimates were not
double-counted and that all LFGE projects and flares were included in the respective project databases. QA/QC
checks performed in the past for the recovery databases were not performed in this Inventory, because new data
were not added to the recovery databases in this Inventory year. For the GHGRP data, EPA verifies annual facility-
level reports through a multi-step process (e.g., combination of electronic checks and manual reviews by staff) to
identify potential errors and ensure that data submitted to EPA 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.5
Recalculations Discussion
Revisions to the individual facility reports submitted to EPA's GHGRP can be made at any time and a portion of
facilities have revised their reports since 2010 for various reasons, resulting in changes to the total net CH4
emissions for MSW landfills. These recalculations increased net emissions for MSW landfills from 2005 to 2015 by
less than 0.5 percent when compared to the previous Inventory report. Each Inventory year, the back-casted
emissions for 2005 to 2009 will be recalculated using the most recently verified data from the GHGRP. Changes in
these data result in changes to the back-casted emissions.
Planned Improvements
EPA has engaged in stakeholder outreach through a series of webinars between December 2016 and August 2017 to
increase the transparency in the Inventory methodology and to identify ideas and supplemental data sources that can
lead to methodological improvements. The areas where EPA is actively working on improvements include the
oxidation factor for 1990 to 2004, the default DOC value, the decay rate (k value), and the scale-up factor.
EPA investigated options to adjust the oxidation factor from the 10 percent currently used for 1990 to 2004 to
another value or approach such as the binned approach used in the GHGRP (e.g., 10 percent, 25 percent, or 35
5 See .
7-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
percent based on methane flux). The oxidation factor currently applied in the later portion of the time series (2005 to
2016) averages at 19.5 percent due to the use of the GHGRP data while the earlier portion of the time series applies
the default of 10 percent. No changes to the oxidation factor have been made to the Inventory as a result of EPA's
recent investigations. Efforts will continue to review new literature and revise the value, as appropriate.
The Inventory currently uses one value of 0.20 for the DOC for years 1990 to 2004. With respect to improvements
to the DOC value, EPA developed a database with MSW characterization data from individual studies across the
United States. EPA will review this data against the Inventory time series to assess the validity of the current DOC
value and how it is applied in the FOD method. Waste characterization studies vary greatly in terms of the
granularity of waste types included and the spatial boundaries of each study (e.g., one landfill, a metro area,
statewide). EPA also notes longer term recommendation from industry stakeholders regarding the DOC values used
in the GHGRP, in the context of new information on the composition of waste disposed in MSW landfills; these
newer values could then be reflected in the 2005 and later years of the Inventory. EPA is continuing to investigate
publicly available waste characterization studies and calculated DOC values resulting from the study data.
EPA began 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. Like the DOC value, the k
values applied through the Waste Model are for the years 1990 to 2004; the k values for 2005 to 2017 are directly
incorporated into the net methane emissions reported to EPA's GHGRP. EPA will continue investigating the
literature for available k value data to understand if the data warrant revisions to the k values used in the Waste
Model between 1990 to 2004.
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 also conducted a brief investigation of the destruction efficiency applied for landfill gas flares and the
fluctuation in natural gas pricing and other potential factors that are impacting the development of new LFGTE
projects. EPA found that flare destruction efficiencies reported by several vendors ranged from 98 to 99.6 percent.
The EPA applies a 99 percent destruction efficiency for all landfill flares incorporated into the Inventory (from 1990
to 2004 because of the GHGRP data used in later years), which aligns well with the identified range. Therefore, no
revisions have been made to the flare destruction efficiency applied in the Inventory.
Box 7-3: Nationwide Municipal Solid Waste Data Sourc
Municipal solid waste generated in the United States can be managed through landfilling, recycling, composting,
and combustion with energy recovery. There are 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
Waste 7-15
-------
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.
Box 7-4: Overview of the Waste Sector
As shown in Figure 7-2 and Figure 7-3, landfilling of MSW is currently and lias 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 7-2: Management of Municipal Solid Waste in the United States, 2015
Recycled
25.8%
MSW to
WTE
12.8%
Landfilled
52.5%
Composted
8.9%
Source: EPA (2018c)
Note: 2015 is the latest year of available data.
7-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Figure 7-3: MSW Management Trends from 1990 to 2015
160
Landfilling
140
120
100
\A
C
o
c
o
Recycling
Combustion with
Energy Recovery
40
Composting
OTHrslrO'=tiO'X>r^CO
-------
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 except 2011 are from the EPA's Advancing Sustainable Materials Management: Facts and Figures 2015
Tables and Figures report (Table 4) published in July 2018 (EPA 2018c).
c 2011 data are not included in the most recent Advancing Sustainable Materials Management: Facts and Figures report (2014),
thus data from the 2013 report (Table 3) was kept in place for 2011 (EPA 2015b).
d Includes electrolytes in batteries and fluff pulp, feces, and urine in disposable diapers. Details may not add to totals due to
rounding.
Note: 2015 is the latest year of available data.
Figure 7-4: Percent of Degradable Materials Diverted from Landfills from 1990 to 2015
(Percent)
90%
Paper and Paperboard
Food Scraps
Yard Trimmings
80%
70%
60%
50%
40%
30%
20%
10%
0%
T
m<3-Ln<Ł>r--cocr)OTHrNiro<3-Ln<Ł>r--oocr)0
a^a>OOOOOOOOOOT-l
HHHHHHHHHHrvl(N(N(N(N(N(N(N(N(N(N(N(N(M(N(N
Source: (EPA 2018c). Note: 2015 is the latest year of available data.
Box 7-5: 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 enviromnent 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
stonnwater 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
7-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• 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
tribal requirements may exist.6
7.2 Wastewater Treatment (CRF Source
Category 5D)
Wastewater treatment processes can produce anthropogenic methane (CH4) and nitrous oxide (N20) emissions.
Wastewater from domestic and industrial sources is treated to remove soluble organic matter, suspended solids,
pathogenic organisms, and chemical contaminants.7 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 2015). Centralized wastewater treatment systems may include a variety of
processes, ranging from lagooning to advanced tertiary treatment technology for removing 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 tertiary 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 N20 may also result from the
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. Recent research suggests that higher emissions of
N20 may in fact originate from nitrification (Ahn et al. 2010). Other more recent research suggests that N20 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 CH4 than wastewater with lower COD (or
BOD) concentrations. BOD represents the amount of oxygen that would be required to completely consume the
organic matter contained in the wastewater through aerobic decomposition processes, while COD measures the total
material available for chemical oxidation (both biodegradable and non-biodegradable). The BOD value is most
commonly expressed in milligrams of oxygen consumed per liter of sample during 5 days of incubation at 20°C, or
BOD5. Because BOD is an aerobic parameter, it is preferable to use COD to estimate CH4 production, since CH4 is
6 For more information regarding federal MSW landfill regulations, see
.
7 Throughout the Inventory, emissions from domestic wastewater also include any commercial and industrial wastewater
collected and co-treated with domestic wastewater.
Waste 7-19
-------
produced only in anaerobic conditions. The principal factor in determining the N20 generation potential of
wastewater is the amount of N in the wastewater. The variability of N in the influent to the treatment system, as well
as the operating conditions of the treatment system itself, also impact the N20 generation potential.
In 2017, CH4 emissions from domestic wastewater treatment were 8.5 MMT CO2 Eq. (339 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 2015). In 2017, CH4 emissions
from industrial wastewater treatment were estimated to be 5.7 MMT CO2 Eq. (229 kt CH4) and include the newly
added sector of breweries. 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 N20, the United States identifies two distinct sources for N20 emissions from domestic wastewater:
emissions from centralized wastewater treatment processes, and emissions from effluent from centralized treatment
systems that lias been discharged into aquatic enviromnents. The 2017 emissions of N20 from centralized
wastewater treatment processes and from effluent were estimated to be 0.4 MMT CO2 Eq. (1.2 kt N20) and 4.6
MMT CO2 Eq. (15.4 kt N20), respectively. Total N20 emissions from domestic wastewater were estimated to be 5.0
MMT CO2 Eq. (16.6 kt N20). Nitrous oxide emissions from wastewater treatment processes gradually increased
across the time series as a result of increasing U.S. population and protein consumption. 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 N20 emission estimates from domestic
wastewater treatment.
Table 7-7: ChU and N2O Emissions from Domestic and Industrial Wastewater Treatment
(MMT COz Eq.)
Activity
1990
2005
2013
2014
2015
2016
2017
CH4
15.3
15.4
14.3
14.3
14.5
14.2
14.2
Domestic
10.4
10.0
OO
OO
8.9
9.0
8.6
8.5
Industrial3
4.9
5.4
5.5
5.4
5.5
5.6
5.7
N2O
3.4
4.4
4.7
4.8
4.8
4.9
5.0
Centralized WWTP
0.2
0.3
0.3
0.3
0.3
0.4
0.4
Domestic Effluent
3.2
4.1
4.3
4.4
4.4
4.5
4.6
Total
18.7
19.8
19.0
19.1
19.3
19.1
19.2
3 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.
Note: Totals may not sum due to independent rounding.
Table 7-8: ChU and N2O Emissions from Domestic and Industrial Wastewater Treatment (kt)
Activity
1990
2005
2013
2014
2015
2016
2017
CH4
611
616
572
572
579
568
568
Domestic
417
399
352
356
360
344
339
Industrial3
194
217
219
216
219
224
229
N2O
11
15
16
16
16
16
17
Centralized WWTP
1
1
1
1
1
1
1
Domestic Effluent
11
14
15
15
15
15
15
3 Industrial activity includes pulp and paper manufacturing, meat and poultry processing, fruit and
vegetable processing, starch-based ethanol production, petroleum refining, and breweries.
Note: Totals may not sum due to independent rounding.
7-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Methodology
Domestic Wastewater CH4 Emission Estimates
Domestic wastewater CH4 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 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
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
= {[(% 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
Emissions from Septic Systems = A
= USpop X (% onsite) X (EFseptic) X 1/109 X 365.25
where,
Total Domestic CH4 Emissions from Wastewater (kt) = A+ B + C+ D
where,
USpop
% onsite
% collected
% aerobicoTcw
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
% aerobiccw
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
% anaerobic
% aerobic w/out primary
% aerobic w/primary
% BOD removed in prim, treat.
% operations not well managed
% anaerobic w/out primary
% anaerobic w/primary
Waste 7-21
-------
EFSEPTIC
= Methane emission factor - septic systems
Total BOD5 produced
= kg BOD/capita/day x U.S. population x 365.25 days/yr
BODcw.inf
= BOD concentration in wastewater entering the constructed wetland
Bo
= Maximum CH4-producing capacity for domestic wastewater
1/106
= Conversion factor, kg to kt
365.25
= Days in a year
3.79
= Conversion factor, gallons to liters
MCF-aerobicnotwellman.
= CH4 correction factor for aerobic systems that are not well managed
MCF-anaerobic
= CH4 correction factor for anaerobic systems
MCF-constructed wetlands
= CH4 correction factor for surface flow constructed wetlands
DE
= CH4 destruction efficiency from flaring or burning in engine
POTWflowCW
= Wastewater flow to POTWs that use constructed wetlands as tertiary
treatment (MGD)
POTWflowAD
= Wastewater influent flow to POTWs that have anaerobic digesters
(MGD)
digester gas
= Cubic feet of digester gas produced per person per day
100
= Wastewater flow to POTW (gallons/person/day)
0.0283
= Conversion factor, ft3 to m3
FRAC CH4
= Proportion of CH4 in biogas
662
= Density of CH4 (g CH4/m3 CH4)
1/109
= Conversion factor, g to kt
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 CH4/capita/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 2018) and include the populations of the United States, American
Samoa, Guam, Northern Mariana Islands, Puerto Rico, and the U.S. Virgin Islands. Table 7-9 presents U.S.
population for 1990 through 2017.
Emissions from Centrally Treated Aerobic and Anaerobic Systems:
Methane emissions from POTWs were estimated by multiplying the total BOD5 produced in the United States by the
percent of wastewater treated centrally, or % collected (about 82 percent) (U.S. Census Bureau 2015), 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 BOD5 treated after primary treatment (67.5 percent, 32.5
percent removed in primary treatment) (Metcalf & Eddy 2014), the maximum CH4-producing capacity of domestic
wastewater (B0, 0.6 kg CH4/kg BOD) (IPCC 2006), and the relative methane correction factors (MCF) 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 BOD5 produced for 1990 through 2017. 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, and 2015 American Housing Surveys conducted by the U.S. Census Bureau
(U.S. Census Bureau 2015), with data for intervening years obtained by linear interpolation and 2017 forecasted
using 1990 to 2016 data. The BOD5 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 BOD5 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 (EPA 1992, 1996, 2000, and 2004). Data for intervening years were obtained by
linear interpolation and the years 2005 through 2016 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
7-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
constructed wetlands as tertiary treatment were obtained from the 1992, 1996, 2000, 2004, 2008, and 2012 Clean
Watersheds Needs Survey (EPA 1992, 1996, 2000, 2004, 2008b, and 2012). Data for intervening years were
obtained by linear interpolation and the years 2013 through 2017 were forecasted from the rest of the time series.
Table 7-9: U.S. Population (Millions) and Domestic Wastewater BODs Produced (kt)
Year Population BODs
1990 253 8,131
2005 300 9,624
2013 320 9,672
2014 323 9,656
2015 325 9,739
2016 327 9,820
201 7 330 9,938
Sources: U.S. Census Bureau (2018);
ERG (2018a).
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
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 CH4 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 inbiogas (0.65), the density of CH4 (662 g CH4/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 .1P-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 inbiogas (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
Enviromnental Managers, "Recommended Standards for Wastewater Facilities (Ten-State StandardsJ" (2004).
Table 7-10 presents domestic wastewater CH4 emissions for both septic and centralized systems, including
anaerobic digesters, in 2017.
Table 7-10: Domestic Wastewater ChU Emissions from Septic and Centralized Systems
(2017, MMT CO2 Eq. and Percent)
CH4 Emissions (MMT CO2 Eq.)
% of Domestic Wastewater CH4
Septic Systems
5.9
69.7%
Centrally-Treated Aerobic Systems
0.03
0.3%
Centrally-Treated Anaerobic Systems
2.3
27.5%
Anaerobic Digesters
0.2
2.4%
Total
8.5
100%
Note: Totals may not sum due to independent rounding.
Waste 7-23
-------
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 2017
are displayed in Table 7-11 below. Table 7-12 contains production data for these industries.
Table 7-11: Industrial Wastewater ChU Emissions by Sector (2017, MMT CO2 Eq. and
Percent)
CH4 Emissions
% of Industrial
(MMT CO2 Eq.)
Wastewater CH4
Meat & Poultry
4.7
81.5%
Pulp & Paper
0.6
10.0%
Fruit &
0.1
0.1
2.4%
2.6%
Vegetables
Petroleum
Refineries
Ethanol
Refineries
0.1
2.6%
Breweries
0.05
1%
Total
5.7
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)
Year Pulp and Paper3
Meat
(Live Weight
Killed)
Poultry
(Live Weight
Killed)
Vegetables,
Fruits and
Juices
Ethanol
Breweries
Petroleum
Refining
1990 82.5
27.3
14.6
38.7
2.5
23.9
702.4
2005 91.8
31.4
25.1
42.9
11.7
23.2
818.6
2013
79.9
33.6
26.5
45.1
39.7
22.5
878.7
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
79.9
35.4
28.9
42.60
47.2
21.8
934.1
a Pulp and paper production is the sum of market pulp production plus paper and paperboard production.
Sources: FAO (2018a) and FAO (2018b); USDA (2018a); Cooper (2018); Beer Institute (2011) and TTB (2018); EIA
(2018).
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 CH4 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 CH4/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
7-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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]
o/0TAp = [%Plants0 x %WWa,P x %CODP]
o/0TAs = [%Plantsa x %WWa,s x %CODs] + [%Plantst x %WWa,t x %CODs]
where,
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 CH4 producing potential of industrial wastewater (kg CH4/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] + [%Plants,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
,, ... Pulp Fruit/ Ethanol Ethanol Breweries
V&ri&blc
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
Waste 7-25
-------
%TAa,t
11.8
0
0
0
0
0
0
0
0
%Plants0
0
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 onEPA'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 CH4 emissions for 1990 through 2017 was developed based on paper and paperboard production
data from the Food and Agricultural Organization of the United Nations (FAO) database FAOSTAT. (FAO 2018a)
and market pulp production data from FAO Pulp and Paper Capacities Reports (FAO 2018b). Market pulp
production values were available directly for 1998, 2000 through 2004, and 2010 through 2016. 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 2018a). 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 2017 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 intervening years were obtained by linear interpolation, while 2015
through 2017 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 2017 (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).
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 CH4/kg COD and default MCF of
7-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
0.8 for anaerobic lagoons were used to estimate the CH4 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 USD A
Agricultural Statistics Database and the Agricultural Statistics Annual Reports (USD A 2018a). 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 CH4/kg COD and default MCF of 0.8 for anaerobic treatment were used to estimate the
CH4 produced from these on-site treatment systems. The USDA National Agricultural Statistics Service (USDA
2018a, 2018c) 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 EPA (1974) for potato,
citrus fruit, and apple processing, and from 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 (m3/ton)
BOD (g/L)
Vegetables
Potatoes
10.27
1.765
Other Vegetables
8.55
0.776
Fruit
Apples
3.66
1.371
Citrus Fruits
10.11
0.317
Non-citrus Fruits
12.42
1.204
Grapes (for wine)
2.78
1.831
Sources: 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 has become more efficient
in recent years (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).
Waste 7-27
-------
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 onsite 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 CH4 produced from
these on-site treatment systems. The amount of CH4 recovered through the use of biomethanators was estimated, and
a 99 percent destruction efficiency was used. Biomethanators are anaerobic reactors that use microorganisms under
anaerobic conditions to reduce COD and organic acids and recover biogas from wastewater (ERG 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 %CODP] + [%Plantsa x %WWa,s x
%CODs] + [%Plantst x %WWa,t x %CODs]) xB0x MCF x % Not Recovered] + [Production x Flow x 3.785 x
COD x ([%PlantSo x %WWa,P x %CODP] + [%Plantsa x %WWa,s x %CODs] + [%Plantst x %WWa,t x %CODs])
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/1)
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 CH4/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 CH4 emissions for 1990 through 2017 was developed based on production data from the Renewable
Fuels Association (Cooper 2018).
Petroleum Refining. Petroleum refining wastewater treatment operations have the potential to produce CH4
emissions from anaerobic wastewater treatment. EPA's Office of Air and Radiation performed an Information
Collection Request (ICR) for petroleum refineries in 2011.8 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 CH4 generation at petroleum refining wastewater treatment systems is presented
below:
Methane = Flow x COD x %TA xB0x MCF
where,
8 Available online at
7-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Flow = Annual flow treated through anaerobic treatment system (m3/year)
COD = COD loading in wastewater entering anaerobic treatment system (kg/m3)
%TA = Percent of wastewater treated anaerobically on site
B0 = Maximum methane producing potential of industrial wastewater (kg CH i/kg COD)
MCF = Methane correction factor
A time series of CH4 emissions for 1990 through 2017 was developed based on production data from the EIA 2018.
Breweries. Since 2010, the number of breweries has increased from less than 2,000 to greater than 6,000 (Brewers
Association 2018). This increase has primarily been driven by craft breweries, which have increased by over 250
percent during that period. Craft breweries were defined as breweries producing less than six million barrels of beer
per year, and non-craft breweries produce greater than six million barrels. With their large amount of water use and
high strength wastewater, breweries generate considerable CH4 emissions from anaerobic wastewater treatment.
However, because many breweries recover their CH4, their emissions are much lower.
The Alcohol and Tobacco Tax and Trade Bureau (TTB) provides total beer production in barrels per year for
different facility size categories from 2007 to the present (TTB 2018). Foryears 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 2017.
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 CH4/kg 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 systems (IPCC 2006).
The amount of CH4 recovered through anaerobic wastewater treatment was estimated, and a 99 percent destruction
efficiency was used (ERG 2018b; Stier J. 2018). Very limited activity data are available on the number of U.S.
breweries that are performing side streaming or pretreatment of wastewater prior to discharge.
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:
Waste 7-29
-------
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
B0 = Maximum methane producing capacity (g CH4/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:
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. That 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 N20 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 N20 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 N20 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 N20-N/kg N produced for constructed wetlands is from
IPCC (2014).
7-30 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
N20 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 2006IPCC Guidelines.
Nitrous oxide emissions from domestic wastewater were estimated using the following methodology:
NzOtotal = NzOplant + NzOeffluent
NzOplant = N20nit/denit + N20woutnit/denit+ N20cwonly + N20cwtertiary
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
where,
N20TOTAL
= Annual emissions of N20 (kt)
N20PLANT
= N20 emissions from centralized wastewater treatment plants (kt)
N20NIT/DENIT
= N20 emissions from centralized wastewater treatment plants with
nitrification/denitrification (kt)
N2OWOUT NIT/DENIT
= N20 emissions from centralized wastewater treatment plants without
nitrification/denitrification (kt)
N2OCW ONLY
= N20 emissions from centralized wastewater treatment plants with constructed
wetlands only (kt)
N2OCW TERTIARY
= N20 emissions from centralized wastewater treatment plants with constructed
wetlands used as tertiary treatment (kt)
N2OEFFLUENT
= N20 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)
POTWflowCW
= 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 N20 -N/kg sewage-N produced) - from effluent
ef4
= Emission factor (kg N20 -N/kg N produced) - constructed wetlands
3.79
= Conversion factor, gallons to liters
44/28
= Molecular weight ratio of N20 to N2
28/44
= Molecular weight ratio of N2 to N20
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 2018)
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, and 2015 American
Waste 7-31
-------
Housing Sun'ev (U.S. Census Bureau 2015). Data for intervening years were obtained by linear interpolation and
2017 was forecasted using 1990 to 2016 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 N20 -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 2018b) and FAO (2018c). Protein consumption data was used directly from USD A for 1990 to 2010 and
2011 through 2013 was calculated using FAO data and a scaling factor. 2014 through 2017 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 2017 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 2017, 298 kt N was removed with 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 Populations 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
2013
320
19.8
81.4
43.3
33.4
285.6
2014
322
20.8
80.8
44.3
34.1
288.7
2015
325
21.8
80.2
44.3
34.1
291.8
2016
327
22.8
81.4
44.3
34.1
294.8
2017
330
23.8
81.7
44.3
34.1
297.9
Sources: Population: U.S. Census Bureau (2018); PopulationND: EPA (1992), EPA (1996), EPA (2000), EPA (2004), EPA
(2008b), EPA (2012); WWTP Population: U.S. Census Bureau (2015); Available Protein: USDA (2018b); 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 2017 CH4 and N20 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 N20 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
7-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 10.3 and 17.3 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 22 percent above the 2017 emissions estimate of 14.2 MMT CO2 Eq. Nitrous oxide emissions from
wastewater treatment were estimated to be between 1.2 and 10.3 MMT CO2 Eq., which indicates a range of
approximately 75 percent below to 108 percent above the 2017 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
2017 Emission Estimate
(MMT CO2 Eq.)
Uncertainty Range Relative to Emission Estimate3
(MMT CO2 Eq.) (%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Wastewater Treatment
CH4
14.2
10.3
17.3
-28%
+22%
Domestic
ch4
8.5
6.0
10.4
-29%
+22%
Industrial
ch4
5.7
2.9
8.5
-49%
+48%
Wastewater Treatment
n2o
5.0
1.2
10.3
-75%
+108%
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 1991
through 2016 values (U.S. Census Bureau 2018). Forecasted protein data was updated which resulted in changes to
2014 through 2016 available protein and protein consumed values. EPA also corrected the emissions calculation for
centrally treated aerobic systems using constructed wetlands as tertiary treatment. This correction affected the entire
time series and resulted in an average decrease of 0.5% for domestic methane emissions and 0.3% for total methane
emissions across the time series.
EPA evaluated pulp and paper production, average BOD concentrations in raw wastewater, and COD:BOD ratio
based on the National Council of Air and Stream Improvement's (NCASI) recommendation and determined updates
to current Inventory data were appropriate. EPA updated production values from summing woodpulp and paper and
paperboard to summing market pulp and paper and paperboard production which resulted in changes for the entire
Waste 7-33
-------
time series. This change also resulted in an update to the data source for pulp and paper production prior to 2002
from the Lockwood-Post to FAO. EPA updated raw wastewater BOD concentrations and the COD:BOD ratio of
influent wastewater based on industry data provided by NCASI (Malmberg 2018) which resulted in changes for
1999 through 2016 and the entire time series, respectively.
EPA evaluated domestic raw BOD production and determined updates to current Inventory data were appropriate to
reflect differences in waste characteristics from households with and without kitchen garbage disposals. The BOD5
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). In addition to applying the distinction of with and without kitchen scraps between
BOD generation rates per capita, the value for the BOD generation rate changed with an updated source (Metcalf &
Eddy 2014). This update further impacted the amount of domestic BOD produced from 2004 through 2016. EPA
now estimates a dynamic BOD generation rate per capita which resulted in changes for the entire times series (ERG
2018a).
On an ongoing basis, EPA reviews other industries that have the potential to emit CH4 from their wastewater
treatment systems because they treat wastewater with significant organics loads. EPA evaluated emissions estimates
from wastewater treatment processes at breweries for potential inclusion in the Inventory. Based on data from the
Brewers Association (Brewers Association 2018; Brewers Association 2017; Brewers Association 2016a; Brewers
Association 2016b), the Beverage Industry Environmental Roundtable (BIER 2017), the Alcohol and Tobacco Tax
and Trade Bureau (TTB 2018), and conversations with industry experts as described above, EPA determined that
this industry generates significant quantities of CH4 from wastewater treatment operations, though a majority of the
emissions are recovered. As a result, EPA determined that the brewery industry is an appropriate category to include
in the Inventory.
Planned Improvements
EPA will continue to investigate the following improvements to the wastewater emissions estimates in the
Inventory:
• Continue working with the NCASI to further refine the market pulp production values as well as update
wastewater characteristic data as new or improved data become available;
• Investigate updated sources of activity data for wastewater treatment system type to distinguish between
aerobic, anaerobic, and other systems with the potential to generate CH4. This includes re-evaluating a
methodology that was developed so that the 2008 and 2012 CWNS data could be used in estimating
emissions from constructed wetlands to determine if it could be extended to all types of systems; and
• Continue reviewing other industrial wastewater treatment sources for those industries believed to discharge
significant loads of BOD or COD, including dairy products processing.
In addition, EPA will continue to monitor potential sources for updating Inventory data, including:
• Sources of data for updating the factor for industrial and commercial co-discharged protein to determine if
the IPCC factor currently used (1.25) is underestimating the contribution of industrial wastewater to N20
emissions;
• WEF biosolid data as a potential source of digester, sludge, and biogas data from POTWs;
• Reports based on international research and other countries' inventory submissions to inform potential
updates to the Inventory's emission factors, methodologies, or included industries;
• Research by groups such as the Water Research Foundation (formerly Water Environment Research
Federation) on emissions from various types of municipal treatment systems, country-specific N20
emission factors, and flare efficiencies and data that indicate septic soil systems are a source of N20 for the
potential development of appropriate emission factors for septic system N20 emissions;
• Sources of data for development of a country-specific methodology for N20 emissions associated with on-
site industrial wastewater treatment operations, including the appropriateness of using IPCC's default factor
for domestic wastewater (0.005 kg N20-N/kg N);
• Additional data sources for stand-alone centralized waste treaters. These data may inform current treatment
assumptions for industrial categories;
7-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
• Additional data sources for improving the uncertainty of the estimate of N entering municipal treatment
systems; and
• Data to update the value used for N content of sludge, the amount of sludge produced, and sludge disposal
practices, along with increasing the transparency of the fate of sludge produced in wastewater treatment.
A refinement of the 2006IPCC Guidelines is currently underway to incorporate abundant new scientific and
empirical knowledge published since 2006 which the IPCC should take into account, particularly with respect to
data for emission factor development. For wastewater treatment, this refinement includes a review of methane and
nitrous oxide emission factors, and an assessment of adding methodologies to account for nitrous oxide emissions
from both domestic and industrial wastewater. EPA will continue to monitor the status of this refinement for
potential updates to the wastewater inventory methodology.
These planned improvements were described in greater detail in a previous Inventory report; please see Section 7.2
of the Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2015.
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. Advantages of composting include reduced volume of the waste,
stabilization of the waste, and destruction of pathogens in the waste. The end products of composting, depending on
its quality, can be recycled as a fertilizer and soil amendment, or be disposed of in a landfill.
Composting is an aerobic process and a large fraction of the degradable organic carbon in the waste material is
converted into carbon dioxide (CO2). Methane (CH4) is formed in anaerobic sections of the compost, which are
created when there is excessive moisture or inadequate aeration (or mixing) of the compost pile. This CH4 is then
oxidized to a large extent in the aerobic sections of the compost. The estimated CH4 released into the atmosphere
ranges from less than 1 percent to a few percent of the initial C content in the material (IPCC 2006). Depending on
how well the compost pile is managed, nitrous oxide (N20) emissions can be produced. The formation of N20
depends on the initial nitrogen content of the material and is mostly due to nitrogen oxide (NOx) denitrification
during the thermophilic and secondary mesophilic stages of composting (Cornell 2007). Emissions vary and range
from less than 0.5 percent to 5 percent of the initial nitrogen content of the material (IPCC 2006). Animal manures
are typically expected to generate more N20 than, for example, yard waste, however data are limited.
From 1990 to 2017, the amount of waste composted in the United States increased from 3,810 kt to 21,333 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. Since then, the amount of waste composted has gradually increased, and when comparing 2010 to
2017, a 16.6 percent increase in waste composted is observed. Emissions of CH4 and N20 from composting from
2010 to 2017 have increased by the same percentage. In 2017, CH4 emissions from composting (see Table 7-18 and
Table 7-19) were 2.2 MMT C02 Eq. (86 kt), and N20 emissions from composting were 1.9 MMT C02 Eq. (6 kt),
representing a slight increase compared to 2016. 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 backyard composting or agricultural composting.
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 population, had
enacted such legislation (ILSR 2014; BioCycle 2010). There are many more initiatives at the metro and municipal
Waste 7-35
-------
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).
Table 7-18: ChU and N2O Emissions from Composting (MMT CO2 Eq.)
Activity 1990
2005
2013 2014 2015 2016 2017
CH4 0.4
N2O 0.3
1.9
1.7
2.0 2.1 2.1 2.1 2.2
1.8 1.9 1.9 1.9 1.9
Total 0.7
3.5
3.9 4.0 4.0 4.0 4.1
Table 7-19: ChU and N2O Emissions from Composting (kt)
Activity
1990 2005 2013
2014
2015
2016
2017
CH4
15 75 81
84
85
85
86
N2O
1 6 6
6
6
6
6
Methodology
Methane and N20 emissions from composting depend on factors such as the type of waste composted, the amount
and type of supporting material (such as wood chips and peat) used, temperature, moisture content (e.g., wet and
fluid versus dry and crumbly), and aeration during the composting process.
The 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):
Ei - MxEFi
where,
Ei = CH4 or N2O emissions from composting, kt CH4 or N20,
M = mass of organic waste composted in kt,
EFi = emission factor for composting, 41 CH4/kt of waste treated (wet basis) and
0.3 t N:0/kt of waste treated (wet basis) (IPCC 2006), and
i = designates either CH4 or N20.
Per IPCC Tier 1 methodology defaults, the emission factors for CH4 and N20 assume a moisture content of 60% in
the wet waste. (IPCC 2006).
Estimates of the quantity of waste composted (M) 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 estimate of the quantity composted for 2012 to
2013 was taken from EPA's Ad\>ancing Sustainable Materials Management: Facts and Figures 2014 report; the
estimate of the quantity composted for 2011 was taken from EPA's Municipal Solid Waste In The United States:
2012 Facts and Figures (EPA 2014); estimates of the quantity composted for 2016 and 2017 were extrapolated
using the 2015 quantity composted and a ratio of the U.S. population growth between 2015 and 2016, and 2016 to
2017 (U.S. Census Bureau 2016, 2017, and 2018).
Table 7-20: U.S. Waste Composted (kt)
Activity
1990
2005
2013
2014
2015
2016
2017
Waste Composted
3,810
18,643
20,358
20,884
21,219
21,332
21,503
7-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Uncertainty and Time-Series Consistency
The estimated uncertainty from the 2006IPCC Guidelines is ±50 percent for the Tier 1 methodology. Emissions
from composting in 2017 were estimated to be between 2.0 and 6.1 MMT CO2 Eq., which indicates a range of 50
percent below to 50 percent above the actual 2017 emission estimate of 4.1 MMT CO2 Eq. (see Table 7-21).
Table 7-21: Tier 1 Quantitative Uncertainty Estimates for Emissions from Composting (MMT
CO2 Eq. and Percent)
Source
Gas
2017 Emission Estimate
Uncertainty Range Relative to Emission Estimate
(MMT CO2 Eq.)
(MMT CO2 Eq.)
(%)
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Composting
CH4, N2O
4.1
2.0
6.1
-50%
+50%
QA/QC and Verification
General QA/QC procedures were applied to data gathering and input, documentation, and calculations consistent
with the U.S. Inventory QA/QC plan, which is in accordance with Vol. 1 Chapter 6 of 2006IPCC Guidelines (see
Annex 8 for more details).
:alcu!ations Discussion
Emissions recalculations were made in this Inventory year for the years 2015 and 2016 per the release of the EPA
Advancing Sustainable Materials Management: 2015 Facts and Figures report (EPA 2018). The tonnage of waste
composted for the year 2015 was previously extrapolated based on the tonnage composted in EPA's Advancing
Sustainable Materials Management: 2014 Facts and Figures report for the year 2014 and a ratio of U.S. population
growth between 2014 and 2015. Because of this change to the 2015 composting tonnage, the extrapolated tonnage
for the year 2016 was also altered. Table 7-19 has been updated to reflect the changes in composting emissions as a
result of these updated tonnage values.
Planned Improvements
For future Inventories, additional efforts will be made to improve the estimates of CH4 and N20 emissions from
composting. For example, a literature search on emission factors and composting systems and management
techniques has been completed and will be documented in a technical memorandum for the 1990 through 2018
Inventory. The purpose of this literature review was to compile all published emission factors specific to various
composting systems and composted materials. This information will be used to determine whether the emission
factors used in the current methodology should be revised, or expanded to account for geographical differences
and/or differences in composting systems used. For example, outdoor composting processes in arid regions typically
require the addition of moisture compared to similar composting processes in wetter climates. Additionally,
composting systems that primarily compost food waste may generate CH4 at different rates than those that compost
yard trimmings because the food waste may have a higher moisture content and more readily degradable material.
Further investigation into accounting of composting emissions estimates across other applicable sections of the
Inventory, in cooperation with the LULUCF Settlements section, will also be completed.
EPA is looking into the possibility of incorporating more specific waste subcategories and category-specific
moisture contents into the emissions estimates for composting in the United States to improve accuracy. However,
to date the EPA has not been able to locate substantial information on the composition of waste at U.S. composting
facilities in order to do so. As additional data becomes available on the composition of waste at these facilities, EPA
will consider using this information in order to create a more detailed calculation of U.S. composting emissions.
Additional efforts are being made to improve the comprehensiveness of the composting Inventory by incorporating
composted waste from U.S. territories. EPA conducted a desk-based investigation into industrial/commercial
Waste 7-37
-------
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.
7.4 Waste Incineration (CRF Source Category
5C1)
As stated earlier in this chapter, carbon dioxide (CO2), nitrous oxide (N20), and methane (CH4) emissions from the
incineration of waste are accounted for in the Energy sector rather than in the Waste sector because almost all
incineration of municipal solid waste (MSW) in the United States occurs at waste-to-energy facilities where useful
energy is recovered. Similarly, the Energy sector also includes an estimate of emissions from burning waste tires and
hazardous industrial waste, because virtually all of the combustion occurs in industrial and utility boilers that
recover energy. The incineration of waste in the United States in 2017 resulted in 11.1 MMT CO2 Eq. of emissions,
over half of which (6.2 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 medical waste incineration. As described in Annex
5 of this report, data are not readily available for that source and emission estimates are not provided. An analysis of
the likely level of emissions was conducted based on a 2009 study of hospital/ medical/ infectious waste incinerator
(HMIWI) facilities in the United States (RTI 2009). Based on that study's information of waste throughput and an
analysis of the fossil-based composition of the waste, it was determined that annual greenhouse gas emissions for
medical waste incineration would be below 500 kt 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 UNFCCC9 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.
Total emissions of NOx, CO, and NMVOCs from waste sources for the years 1990 through 2017 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
2013
2014
2015
2016
2017
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
7
8
8
8
8
9 See .
7-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Landfills
1
6
7
8
8
8
8
Wastewater Treatment
+
+
+
1
1
1
1
Miscellaneous3
+
0
0
0
0
0
0
NMVOCs
673
114
58
68
68
68
68
Wastewater Treatment
57
49
25
29
29
29
29
Miscellaneous3
557
43
22
26
26
26
26
Landfills
58
22
11
13
13
13
13
+ 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.
Note: Totals may not sum due to independent rounding.
Methodology
Emission estimates for 1990 through 2017 were obtained from data published on the National Emission Inventory
(NEI) Air Pollutant Emission Trends web site (EPA 2018), and disaggregated based on EPA (2003). Emission
estimates of these gases were provided by sector, using a "top down" estimating procedure—emissions were
calculated either for individual sources or for many sources combined, using basic activity data (e.g., the amount of
raw material processed) as an indicator of emissions. National activity data were collected for individual categories
from various agencies. Depending on the category, these basic activity data may include data on production, fuel
deliveries, raw material processed, etc.
Uncertainty and Time-Series Consistency
No quantitative estimates of uncertainty were calculated for this source category. Methodological recalculations
were applied to the entire time series to ensure time-series consistency from 1990 through 2017. Details on the
emission trends through time are described in more detail in the Methodology section, above.
Waste 7-39
-------
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."
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. To understand the details of any specific recalculation or methodological improvement, see
the Recalculations Discussion within each source/sink categories' section found in Chapters 3 through 7 of this
report and a discussion of Inventory improvements in Annex 8. Table 9-1 summarizes the quantitative effect of all
changes on U.S. greenhouse gas emissions in the Energy, IPPU, Agriculture, and Waste sectors, while Table 9-2
summarizes the quantitative effect of changes on annual net fluxes from LULUCF. Both tables present results
relative to the previously published Inventory (i.e., the 1990 to 2016 report) in units of million metric tons of carbon
dioxide equivalent (MMT CO2 Eq.).
In general, when methodological changes have been implemented, the previous Inventory's time series (i.e., 1990 to
2016) 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 following source and sink categories underwent the most significant methodological and historical data
changes. A brief summary of the recalculations and/or improvements undertaken are provided for these categories.
• Forest Land Remaining Forest Land: Changes in Forest Carbon Stocks (CO2). In the current Inventory the
regional approach for carbon stock and stock change estimation in the western United States was replaced
by the state-level method used in the eastern United States so carbon stocks and stock changes are now
estimated consistently for the entire 1990 to 2017 time series in all states with remeasurements in the
national forest inventory (NFI) in the conterminous 48 states. This improvement in consistency also
improved separation of Forest Land Remaining Forest Land, Land Converted to Forest Land, and areas
with perennial woody biomass that do not meet the definition of forest land (i.e., woodlands) that are now
included in the Grassland Remaining Grassland and Land Converted to Grassland sections. Next, all
managed forest land in Alaska, specifically forest land from interior Alaska, was also included for the first
time in this Inventory, which added more than 24.5 million ha to the Forest Land Remaining Forest Land
category. The inclusion of 24.5 million ha of forest area from interior Alaska contributed an additional
8,597 MMT C stocks, primarily from soil carbon, to the Forest Land Remaining Forest Land category in
2018 and this increase was consistent with the additions from interior Alaska over the time series (see
Table 6-15). The carbon stock changes in interior Alaska were driven, in large part, by wildfires over the
time series and contribute, on average over the time series, approximately -2.2 MMT C per year to the
sink. As a result of these improvements, the estimates reported in the previous (i.e., 1990 through 2016)
Inventory are not directly comparable to the estimates in this Inventory. In most cases this was not a loss of
forest land area but rather a reorganization of land into the Land Converted to Forest Land category and the
Recalculations and Improvements 9-1
-------
transfer of 23.5 million hectares of land into the Grassland Remaining Grassland and Land Converted to
Grassland categories. The recalculations resulted in an average annual increase in C stock change losses of
39.1 MMT CO2 Eq. (6 percent), across the 1990 through 2016 time series, relative to the previous
Inventory.
• Land Converted to Cropland: Changes in all Ecosystem Carbon Stocks (CO2). Methodological
recalculations are associated with extending the time series from 2013 through 2016 for mineral and
organic soils using a surrogate data method, and from 1990 to 2016 forbiomass and dead organic matter C
associated with Forest Land Converted to Cropland. The increased C stock losses are almost entirely
attributed to the update of biomass and dead organic matter losses for Forest Land Converted to Cropland
with newly available re-measurement data for the western United States. Stock changes were re-estimated
at the plot-level with the new data consistent with the compilation methods described for Forest Land
Remaining Forest Land. In the previous Inventory, state-level averages from the plot data had been used to
approximate the losses of C with Forest Land Converted to Cropland due to a lack of re-measurement data.
C stock change losses increased by an average of 39.1 MMT CO2 Eq. (141 percent) from 1990 through
2016 as a result of the recalculation, relative to the previous Inventory.
• Settlements Remaining Settlements: Changes in Settlement Tree Carbon Stocks (CO2). Past estimates of C
sequestration in settlement areas used urban land and urban tree cover as proxy for the settlement area
estimates. This new approach uses settlement land area and percent tree cover in developed land as a proxy
for percent tree cover in settlement area. The recalculations resulted in an average annual increase in C
stock gains of 35.7 MMT CO2 Eq. (47 percent), across the 1990 through 2016 time series, relative to the
previous Inventory.
• Land Converted to Forest Land: Changes in Forest Carbon Stocks (CO2). The availability of
remeasurement data from the annual national forest inventory (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 current Inventory were based on state-level carbon density
estimates and a combination of Natural Resources Inventory (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. The recalculations resulted in an average annual increase in C stock gains of 30.4 MMT CO2
Eq. (37 percent), across the 1990 through 2016 time series, relative to the previous Inventory.
• Land Converted to Settlements: Changes in Settlement Soil Carbon Stocks (CO2). Methodological
recalculations are associated with extending the time series from 2013 through 2017 using a linear time
series model, and an update of biomass and dead organic matter losses with Forest Land Converted to
Settlements. The recalculation led to a 31 percent greater loss of C on average. This change is almost
entirely attributed to the update of biomass and dead organic matter losses for Forest Land Converted to
Settlements with newly available re-measurement data for the western United States. New stock changes
were estimated at the plot-level with the new data consistent with the compilation methods described in the
Forest Land Remaining Forest Land section. In the previous Inventory, state-level averages from the plot
data had been used to approximate the losses of C with Forest Land Converted to Settlements due to a lack
of re-measurement data. These changes resulted in an average annual increase in C stock change losses of
18.0 MMT CO2 Eq. (31 percent) relative to the previous Inventory.
• Stationary Combustion (N2O). Nitrous oxide emissions from stationary sources were revised across the
entire time series due to revised data from EIA (2019) and EPA (2018) relative to the previous Inventory.
Most notably, EIA updated wood biomass consumption statistics in the residential sector from 2009 to
2016, and the commercial sector from 2014 to 2016 (EIA 2019). Nitrous oxide emission factors for coal
wall-fired boilers used in the electric power sector were also updated from 0.5 kg/TJ to 5.8 kg/TJ to be
consistent with EPA's Compilation of Air Pollutant Emission Factors, AP-42 (EPA 1997). These changes
resulted in an average annual increase in N20 emissions of 15.3 MMT CO2 Eq. (107 percent) relative to the
previous Inventory.
• Land Converted to Grassland: Changes in all Ecosystem Carbon Stock (CO2). Methodological
recalculations are associated with extending the time series from 2013 through 2016 for mineral and
organic soils using a surrogate data method, and from 1990 to 2016 forbiomass and dead organic matter C
associated with Forest Land Converted to Grassland. This change is almost entirely attributed to the update
of biomass and dead organic matter losses for Forest Land Converted to Grassland with newly available
9-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
re-measurement data for the western United States. Stock changes were re-estimated at the plot-level with
the new data consistent with the compilation methods described for Forest Land Remaining Forest Land.
In the previous Inventory, state-level averages from the plot data had been used to approximate the losses
of C with Forest Land Converted to Grassland due to a lack of remeasurement data. These changes
resulted in an average annual increase in C stock of 14.3 MMT CO2 Eq. (67 percent) relative to the
previous Inventory.
• Forest Land Remaining Forest Land: Non-CO 2 Emissions from Forest Fires (CO 2). The methods used in
the current Inventory to compile estimates of non-CO: emissions from forest fires are consistent with those
used in the previous (i.e., 1990 through 2016) Inventory, but also include some additional steps toward
better definition of forest area in Alaska, fuel, and combustion. Modifications in each of these factors affect
estimates. Forest within the Monitoring Trends in Burn Severity (MTBS) defined fire perimeters (MTBS
Data Summaries 2018) are estimated according to National Land Cover Dataset (NLCD) spatial datasets
(Homer et al. 2015) rather than Ruefenacht et al. (2008) as in the previous report. Fuel estimates are based
on the distribution of stand-level carbon pools (USDA Forest Service 2018b, 2018d) classified according to
ecological region rather than the state-wide estimates as in the previous report. Combustion estimates are
partly a function of the MTBS severity classifications and thus can vary within a fire. The effects of these
modifications varied across the time series, but more often lowered the estimates for both CH4 and N20.
These changes resulted in an average annual decrease in CH4 and N20 emissions of 3.6 MMT CO2 Eq. (28
percent) relative to the previous Inventory.
• Wetlands Remaining Wetlands: Changes in Mineral and Organic Soil Carbon Stocks in Coastal Wetlands
(CO2). Methodological recalculations are associated with the extension of the Coastal Change Analysis
Program (C-CAP) data extrapolation through 2017. Soil reference carbon sequestration rates were updated
based on recalculation by Lu and Megonigal (2017), which decreased net removals to soil by 0.01 MMT
CO2 Eq. per year. New data on aboveground biomass carbon stocks were added, broken down by climate
zone, that were derived from a national assessment combining field plot data and aboveground biomass
mapping by remote sensing (Byrd et al., 2017; Byrd, et al., 2018). These changes resulted in an average
annual increase in C stock change losses of 3.5 MMT CO2 Eq. (46 percent) relative to the previous
Inventory.
• Petroleum Systems (CH4). Updates were made to exploration and production segment methodologies,
specifically: using GHGRP data to calculate emissions and activity factors for oil well completions and
workovers with hydraulic fracturing; using Drillinglnfo data (Drillinglnfo 2018) to calculate well drilling
activity; and revising the basis for calculating the number of active wells represented in GHGRP reporting.
The combined impact of revisions to 2016 petroleum systems CH4 emissions, compared to the previous
Inventory, is a decrease from 38.6 to 38.2 MMT CO2 Eq. (0.4 MMT CO2 Eq., or 1 percent). The
recalculations resulted in an average annual increase in CH4 emission estimates across the 1990 through
2016 time series, compared to the previous Inventory, of 3.3 MMT CO2 Eq. (10 percent) with the largest
increases in the estimates for 2005 to 2013 due to the revised data on hydraulically fractured oil well
completions.
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 inclusion of N2O emissions from Natural Gas Systems and Petroleum Systems, and CO2 emissions from
petroleum transport in Petroleum Systems, and breweries as a source of CH4 emissions from industrial wastewater
within Waste.
Table 9-1: Revisions to U.S. Greenhouse Gas Emissions (MMT CO2 Eq.)
Gas/Source
1990
2005
2013
2014
2015
2016
Average
Annual
Change
CO2
-0.1
-1.5
3.3
3.3
2.2
-4.2
-1.0
Fossil Fuel Combustion
-1.6
-2.2
0.5
-1.0
-2.1
-4.2
-2.1
Transportation
-0.9
-0.9
0.1
-0.9
+
-0.4
-0.6
Electric Power Sector
1.5
1.2
5.1
4.4
-1.5
-3.6
0.7
Industrial
-1.4
-2.3
-3.2
-5.4
-1.6
-1.5
-2.1
Recalculations and Improvements 9-3
-------
Residential
-0.2
0.1
-0.4
1.5
1.0
0.4
0.2
Commercial
-0.7
-0.3
-1.1
-0.7
+
0.9
-0.4
U.S. Territories
NC
+
+
+
+
+
+
Non-Energy Use of Fuels
+
0.7
+
1.0
1.3
1.5
0.4
Natural Gas Systems
0.2
0.1
0.3
0.2
0.2
+
0.2
Cement Production
NC
NC
NC
NC
NC
NC
NC
Lime Production
NC
NC
NC
NC
NC
NC
NC
Other Process Uses of Carbonates
NC
NC
NC
NC
-0.1
+
+
Glass Production
NC
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
NC
NC
+
-0.2
0.1
0.1
+
Titanium Dioxide Production
NC
NC
NC
NC
NC
0.1
+
Aluminum Production
NC
NC
NC
NC
NC
NC
NC
Iron and Steel Production & Metallurgical Coke
Production
NC
NC
NC
NC
+
+
+
Ferroalloy Production
NC
NC
NC
NC
NC
NC
NC
Ammonia Production
NC
NC
-0.5
-0.2
-0.2
-1.4
-0.1
Urea Consumption for Non-Agricultural Purposes
NC
NC
0.5
0.3
0.4
1.2
0.1
Phosphoric Acid Production
NC
NC
NC
NC
NC
+
+
Petrochemical Production
+
+
NC
NC
NC
NC
+
Silicon Carbide Production and Consumption
NC
NC
NC
NC
NC
NC
NC
Lead Production
NC
NC
NC
NC
NC
+
+
Zinc Production
NC
NC
NC
NC
NC
NC
NC
Petroleum Systems
1.3
-0.1
2.5
3.3
2.9
-0.6
0.6
Abandoned Oil and Gas Wells
+
+
+
+
+
+
+
Magnesium Production and Processing
NC
NC
NC
NC
NC
NC
NC
Liming
NC
NC
NC
NC
+
-0.7
+
Urea Fertilization
NC
NC
NC
+
-0.2
-0.2
+
Wood Biomass, Ethanol, and Biodiesel
Consumption"
NC
NC
NC
NC
NC
NC
NC
International Bunker Fuels"
NC
NC
-0.9
-1.1
7.3
7.9
1.2
CH4
-0.1
2.8
0.5
-1.9
-4.0
-2.5
1.1
Stationary Combustion
+
+
+
+
0.6
0.6
0.1
Mobile Combustion
0.2
0.2
-0.2
-0.2
-0.2
-0.3
0.1
Coal Mining
NC
NC
NC
NC
NC
+
+
Abandoned Underground Coal Mines
NC
NC
NC
NC
NC
NC
NC
Natural Gas Systems
-2.1
2.3
1.8
0.8
0.9
2.2
0.6
Petroleum Systems
2.2
4.6
5.1
3.5
1.4
-0.4
3.3
Abandoned Oil and Gas Wells
0.1
0.1
+
+
+
0.1
0.1
Petrochemical Production
NC
NC
NC
NC
NC
NC
NC
Silicon Carbide Production and Consumption
NC
NC
NC
NC
NC
NC
NC
Iron and Steel Production & Metallurgical Coke
Production
NC
NC
NC
NC
NC
NC
NC
Ferroalloy Production
NC
NC
NC
NC
NC
NC
NC
Enteric Fermentation
NC
NC
+
+
+
1.8
0.1
Manure Management
+
-2.6
-5.2
-5.1
-5.4
-6.2
-2.3
Rice Cultivation
NC
NC
NC
NC
NC
NC
NC
Field Burning of Agricultural Residues
-0.1
+
-0.1
-0.1
-0.1
-0.1
-0.1
Landfills
NC
-1.3
-0.3
-0.2
-0.5
0.3
-0.3
Wastewater Treatment
-0.4
-0.4
-0.6
-0.7
-0.6
-0.6
-0.4
Composting
NC
NC
NC
NC
+
+
+
Incineration of Waste
NC
NC
NC
NC
NC
NC
NC
International Bunker Fuels"
NC
NC
NC
NC
NC
NC
NC
N2O
15.6
18.0
2.1
1.6
-5.5
-5.1
13.9
Stationary Combustion
14.0
16.9
14.0
13.9
12.4
11.5
15.3
Mobile Combustion
0.3
0.2
-0.4
-0.4
-0.5
-0.5
+
Adipic Acid Production
NC
NC
NC
NC
NC
+
+
9-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Nitric Acid Production
NC
NC
NC
NC
NC
+
+
Manure Management
NC
-0.1
-0.1
-0.1
-0.1
+
-0.1
Agricultural Soil Management
1.2
1.1
-11.4
-11.7
-17.2
-16.0
-1.3
Field Burning of Agricultural Residues
+
+
+
+
+
+
+
Wastewater Treatment
NC
+
+
-0.1
-0.1
-0.1
+
N2O from Product Uses
NC
NC
NC
NC
NC
NC
NC
Caprolactam, Glyoxal, and Glyoxylic Acid
Production
NC
NC
NC
NC
NC
NC
NC
Incineration of Waste
NC
NC
NC
NC
NC
NC
NC
Composting
NC
NC
NC
NC
+
+
+
Semiconductor Manufacture
NC
NC
+
+
+
+
+
Natural Gas Systems
NC*
NC*
NC*
NC*
NC*
NC*
NC*
Petroleum Systems
NC*
NC*
NC*
NC*
NC*
NC*
NC*
International Bunker Fuels"
NC
NC
NC
NC
NC
NC
NC
HFCs, PFCs, SF« and NF3
+
-0.6
-4.8
-6.2
-7.1
-7.2
-1.6
HFCs
+
-0.6
-5.0
-6.1
-6.9
-7.4
-1.5
Substitution of Ozone Depleting Substances
+
-0.6
-5.1
-6.1
-6.9
-7.4
-1.5
HCFC-22 Production
NC
NC
NC
NC
NC
NC
NC
Semiconductor Manufacture
NC
+
0.1
+
+
+
+
Magnesium Production and Processing
NC
NC
NC
NC
NC
NC
NC
PFCs
NC
-0.1
0.2
+
+
+
+
Aluminum Production
NC
NC
NC
NC
NC
NC
NC
Semiconductor Manufacture
NC
-0.1
0.2
+
+
+
+
Substitution of Ozone Depleting Substances
NC
NC
+
+
+
+
+
SF«
+
+
0.1
-0.1
-0.2
0.1
-0.1
Electrical Transmission and Distribution
+
+
+
-0.1
-0.2
+
+
Semiconductor Manufacture
NC
+
0.3
+
+
+
+
Magnesium Production and Processing
NC
NC
-0.2
-0.1
0.1
0.1
-0.1
NF3
NC
+
-0.1
+
+
+
+
Semiconductor Manufacture
NC
+
-0.1
+
+
+
+
Net Emissions (Sources and Sinks)b
28.0
9.9
23.4
67.2
-30.2
-24.9
28.0
Percent Change
0.5%
0.1%
0.4%
1.1%
-0.5%
-0.4%
0.5%
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 CO2 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 Not included in emissions total.
b Sinks are only included in net emissions total.
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
2013
2014
2015
2016
Average
Annual
Change
Forest Land Remaining Forest Land
23.3
18.2
52.4
98.6
11.1
18.9
35.6
Changes in Forest Carbon Stocks3
26.1
25.2
54.2
100.5
21.0
41.5
39.1
N011-CO2 Emissions from Forest Firesb
-2.8
-7.0
-1.8
-1.9
-9.9
-22.6
-3.6
N2O Emissions from Forest Soilsc
NC
NC
NC
NC
NC
NC
NC
N011-CO2 Emissions from Drained
Organic Soils'1
NC
NC
NC
NC
NC
NC
NC
Land Converted to Forest Land
-27.1
-38.4
-45.5
-45.5
-45.6
-45.6
-30.4
Changes in Forest Carbon Stocks6
-27.1
-38.4
-45.5
-45.5
-45.6
-45.6
-30.4
Cropland Remaining Cropland
NC
NC
NC
NC
NC
NC
NC
Changes in Mineral and Organic Soil
Carbon Stocks
NC
NC
NC
NC
NC
NC
NC
Land Converted to Cropland
32.3
40.8
43.7
43.6
43.6
43.6
39.1
Changes in all Ecosystem Carbon Stocksf
32.3
40.8
43.7
43.6
43.6
43.6
39.1
Recalculations and Improvements 9-5
-------
Grassland Remaining Grassland
NC
NC
NC
NC
NC
NC
NC
Changes in Mineral and Organic Soil
Carbon Stocks
NC
NC
NC
NC
NC
NC
NC
N011-CO2 Emissions from Grassland
Firesg
NC
NC
NC
NC
NC
NC
NC
Land Converted to Grassland
-9.1
-14.0
-13.6
-13.5
-13.5
-13.5
-14.3
Changes in all Ecosystem Carbon Stocksf
-9.1
-14.0
-13.6
-13.5
-13.5
-13.5
-14.3
Wetlands Remaining Wetlands
3.5
3.3
3.5
3.5
3.5
3.5
3.5
Changes in Organic Soil Carbon Stocks in
Peatlands
NC
NC
NC
NC
+
+
+
Changes in Aboveground and Soil Carbon
Stocks in Coastal Wetlands
3.5
3.3
3.5
3.5
3.5
3.4
3.5
CH4 Emissions from Coastal Wetlands
Remaining Coastal Wetlands
+
+
+
+
+
+
+
N2O Emissions from Coastal Wetlands
Remaining Coastal Wetlands
NC
NC
NC
NC
NC
+
+
N011-CO2 Emissions from Peatlands
Remaining Peatlands
NC
NC
NC
NC
+
+
+
Land Converted to Wetlands
+
+
+
+
+
+
+
Changes in Aboveground and Soil Carbon
Stocks
+
+
+
+
+
+
+
CH4 Emissions from Land Converted to
Coastal Wetlands
NC
NC
+
+
+
+
+
Settlements Remaining Settlements
-35.9
-36.3
-36.1
-34.6
-33.2
-31.0
-35.8
Changes in Organic Soil Carbon Stocks
NC
NC
NC
NC
NC
NC
NC
Changes in Settlement Tree Carbon
Stocks
-35.8
-36.3
-36.1
-34.4
-32.7
-31.0
-35.7
Changes in Yard Trimming and Food
Scrap Carbon Stocks in Landfills
+
-0.1
-0.1
-0.2
-0.5
0.1
-0.1
N2O Emissions from Settlement Soils'1
NC
NC
NC
NC
NC
NC
NC
Land Converted to Settlements
25.7
17.7
18.0
18.4
18.4
18.4
18.0
Changes in all Ecosystem Carbon Stocksf
25.7
17.7
18.0
18.4
18.4
18.4
18.0
LULUCF Emissions'
-2.9
-7.0
-1.8
-1.9
-9.8
-22.6
-3.6
LULUCF Carbon Stock Change"
15.5
-1.9
24.1
72.2
-6.0
16.8
19.2
LULUCF Sector Net Total1
12.6
-8.9
22.3
70.4
-15.9
-5.8
15.6
Percent Change
1.5%
-1.2%
3.0%
9.5%
-2.3%
-0.8%
2.1%
Note: Totals may not sum due to independent rounding
NC (No Change)
+ Absolute value does not exceed 0.05 MMT CO2 Eq. or 0.05 percent.
a Includes the net changes to carbon stocks stored in all forest ecosystem pools and harvested wood products.
bEstimates 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 abovegroimd/belowgroimd 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 Grassland.
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.
1LULUCF emissions include the CH4 and N2O emissions reported for Peatlands Remaining Peatlands, Forest Fires,
Drained Organic Soils, Grassland Fires, and Coastal Wetlands Remaining Coastal Wetlands; CH4 emissions fromLawrf
Converted to Coastal Wetlands; andN^O emissions from Forest Soils and Settlement Soils.
J LULUCF Carbon Stock Change includes any C stock gains and losses from all land use and land use conversion
categories.
k Hie LULUCF Sector Net Total is the net sum of all CH4 and N2O emissions to the atmosphere plus net carbon stock
changes.
9-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
10. References
Executive Summary
BEA (2019) 2017 Comprehensive Revision of the National Income and Product Accounts: Current-dollar and
"real" GDP, 1929-2017. 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 (2019) Electricity Generation. Monthly Energy Review, February 2019. Energy Information Administration,
U.S. Department of Energy, Washington, D.C. DOE/EIA-0035(2019/02).
EIA (2018) Electricity in the United States. Electricity Explained. Energy Information Administration, U.S.
Department of Energy, Washington, D.C. Available online at:
.
EIA (2017) International Energy Statistics 1980-2017. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EPA (2018a) Acid Rain Program Dataset 1996-2017. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2018b) Greenhouse Gas Reporting Program (GHGRP). 2018 Envirofacts. Subpart HH: Municipal Solid Waste
Landfills and Subpart TT: Industrial Waste Landfills. Available online at:
.
EPA (2018c) "1970 - 2017 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, March 2018.
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 2017) Highway Statistics. Federal Highway Administration, U.S. Department of
Transportation, Washington, D.C. Report FHWA-PL-96-023-annual. Available online at:
.
IEA (2018) 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. [Stacker, 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, L.G. 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 (2018a) Trends in Atmospheric Carbon Dioxide. Available online at:
. 19 December 2018.
NOAA/ESRL (2018b) Trends in Atmospheric Methane. Available online at:
. 19 December 2018.
NOAA/ESRL (2018c) 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 2018.
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 (2018) U.S. Census Bureau International Database (IDB). Available online at:
.
Introduction
Carbon Dioxide Information Analysis Center (CDIAC) (2016) Recent Greenhouse Gas Concentrations. April 2016.
Available online at: .
EPA (2009) Technical Support Document for the Endangerment and Cause or Contribute Findings for Greenhouse
Gases under Section 202(a) of the Clean Air Act. U.S. Environmental Protection Agency. December 2009.
EPA (2017) Climate Change. Available online at: . 6 April 2017.
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.
10-2 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 [Stacker, 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 (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 (1999) Aviation and the Global Atmosphere. Intergovernmental Panel on Climate Change. [J.E. Penner, et al.
(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:
Jacobson, M.Z. (2001) "Strong Radiative Heating Due to the Mixing State of Black Carbon in Atmospheric
Aerosols." Nature, 409:695-697.
NOAA (2017) Vital Signs of the Planet. Available online at: . Accessed on 9
January 2017.
NOAA/ESRL (2018a) Trends in Atmospheric Carbon Dioxide. Available online at:
. 19 December 2018.
NOAA/ESRL (2018b) Trends in Atmospheric Methane. Available online at:
. 19 December 2018.
NOAA/ESRL (2018c) 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 2018.
NOAA/ESRL (2018d) Sulfur Hexafluoride (SF$) 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:
. 10 January 2018.
NRC (2012) Climate Change: Evidence, Impacts, and Choices. Available online at:
.
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: .
References 10-3
-------
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: .
Trends in Greenhouse Gas Emissions
BEA (2019) 2017 Comprehensive Revision of the National Income and Product Accounts: Current-dollar and
"real" GDP, 1929-2017. 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) Monthly Energy Review, February 2019. Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2019/02).
EIA (2019b) Fuel Oil and Kerosene Sales. Energy Information Administration, U.S. Department of Energy,
Washington, D.C. January 2019.
EIA (2018) "In 2017, U.S. electricity sales fell by the greatest amount since the recession" Available online at:
.
EPA (2019) 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 (2018) 1970 - 2017 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, March 2018.
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 (2018) U.S. Census Bureau International Database (IDB). Available online at:
.
USD A (2018) Personal communication. Claudia Hitaj, USD A Economic Research Service, and Vincent
Camobreco, U.S. EPA. December 2018.
Energy
EIA (2019) Monthly Energy Review, February 2019, Energy Information Administration, U.S. Department of
Energy, Washington, DC. DOE/EIA-0035(2019/02).
IEA (2018) 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.
10-4 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
AISI (2004 through 2018) Annual Statistical Report, American Iron and Steel Institute, Washington, D.C.
APTA (2007 through 2017) Public Transportation Fact Book. American Public Transportation Association,
Washington, D.C. Available online at: .
APTA (2006) Commuter Rail National Totals. American Public Transportation Association, Washington, D.C.
BEA (2018) Table 1.1.6. Real Gross Domestic Product, Chained 2012 Dollars. Bureau of Economic Analysis
(BEA), U.S. Department of Commerce, Washington, D.C. September 2018. Available online at:
.
Benson, D. (2002 through 2004) Unpublished data. Upper Great Plains Transportation Institute, North Dakota State
University and American Short Line & Regional Railroad Association.
Browning, L. (2018a). Updated Methodology for Estimating Electricity Use from Highway Plug-In Electric
Vehicles. Technical Memo, October 2018.
Browning, L. (2018b). Updated Non-Highway CH4 and N2O Emission Factors for U.S. GHG Inventory. Technical
Memo, November 2018.
Browning, L. (2017) Updated Methodology for Estimating CH4 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 (2018) Unpublished data from the Fuels Automated System (FAS). Defense Logistics Agency Energy,
U.S. Department of Defense. Washington, D.C.
DOC (1991 through 2018) 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 Edition 37. 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 2017) 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-fro m-
Coal Complex in Kingsport, TN.
EIA (2019a) Monthly Energy Review, February 2019, Energy Information Administration, U.S. Department of
Energy, Washington, DC. DOE/EIA-0035(2019/02).
EIA (2019b) Quarterly Coal Report: July - September 2018. Energy Information Administration, U.S. Department
of Energy. Washington, D.C. DOE/EIA-0121.
EIA (2019c) Form EIA-923 detailed data with previous form data (EIA-906/920), Energy Information
Administration, U.S. Department of Energy. Washington, DC. DOE/EIA.
EIA (2018a) Natural Gas Annual 2017. Energy Information Administration, U.S. Department of Energy.
Washington, D.C. DOE/EIA-O131(17).
EIA (2018b) Annual Coal Report 2017. Energy Information Administration, U.S. Department of Energy.
Washington, D.C. DOE/EIA-0584.
References 10-5
-------
EIA (2018c) Alternative Fuels Data Tables. Energy Information Administration, U.S. Department of Energy.
Washington, D.C. Available online at: .
EIA (2018d) Electric Power Annual 2017. Energy Information Administration, U.S. Department of Energy.
Washington, D.C. DOE/EIA-O348(17).
EIA (2018e) "In 2017, U.S. electricity sales fell by the greatest amount since the recession." Available online at: <
https://www.eia.gov/todayinenergy/detail.php?id=35612>.
EIA (2017) International Energy Statistics 1980-2016. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EIA (1991 through 2018) 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 (2001) U.S. Coal, Domestic and International Issues. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. March 2001.
EPA (2018a) Acid Rain Program Dataset 1996-2017. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2018b) Motor Vehicle Emissions Simulator (Moves) 2014b. Office of Transportation and Air Quality, U.S.
Environmental Protection Agency, Washington, D.C. Available online at: .
EPA (2018c) The Emissions & Generation Resource Integrated Database (eGRID) 2016 Technical Support
Document. Clean Air Markets Division, Office of Atmospheric Programs, U.S. Environmental Protection Agency,
Washington, 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.
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). January 2019.
FHWA (1996 through 2017) 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: .
10-6 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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) (2018a) 2018Mineral Commodity Summaries: Aluminum. U.S. Geological
Survey, Reston, VA.
USGS (2018b) Minerals Industry Surveys: Manufactured Abrasives in the Second Quarter. U.S. Geological Survey,
Reston, VA. Available online at: .
USGS (1991 through 2017) Minerals Yearbook - Iron and Steel Scrap. U.S. Geological Survey, Reston, VA.
USGS (1991 through 2015a) Miner als Yearbook: Manufactured Abrasives Annual Report. U.S. Geological Survey,
Reston, VA. Available online at: .
USGS (1991 through 2015b) Miner als Yearbook: Titanium. U.S. Geological Survey, Reston, VA.
USGS (1991 through 2015c) Miner als Yearbook: Silicon Annual Report. U.S. Geological Survey, Reston, VA.
Available online at: .
USGS (1996 through 2013) Minerals Yearbook: Silicon. U.S. Geological Survey, Reston, VA.
USGS (1995 through 2013) Minerals Yearbook: Lead Annual Report. U.S. Geological Survey, Reston, VA.
USGS (1995, 1998, 2000, 2001, 2002, 2007) Minerals Yearbook: Aluminum Annual Report. U.S. Geological
Survey, Reston, VA.
Stationary Combustion {excluding C02)
EIA (2019) Monthly Energy Review, February 2019. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0035(2019/02).
EIA (2017) International Energy Statistics 1980-2014. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
References 10-7
-------
EPA (2018a) Acid Rain Program Dataset 1996-2017. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
EPA (2018b). 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 2017) 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 (S AIC) 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 (2006) Argonne National Laboratory (2006) GREET model Version 1.7. June 2006.
ANL (2017) The Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model (GREET2017).
Argonne National Laboratory. October 2017. Available online at: .
APTA (2007 through 2017) 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, L. (2018a). Updated Methodology for Estimating Electricity Use from Highway Plug-In Electric
Vehicles. Technical Memo, October 2018.
Browning, L. (2018b) "Updated Non-Highway CH4 and N20 Emission Factors for U.S. GHG Inventory". Technical
Memo, October 2018.
Browning, L. (2017) "Updated Methodology for Estimating CH4 and N20 Emissions from Highway Vehicle
Alternative Fuel Vehicles". Technical Memo, October 18, 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 (2018) 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.
10-8 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
DOC (1991 through 2018) 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, Centerfor
Transportation Analysis, Energy Division, Oak Ridge National Laboratory. ORNL-6978.
DOT (1991 through 2017) 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, February 2019, Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2019/02).
EIA (2019b) Natural Gas Annual. 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 (2018b,) Motor Vehicle Emissions Simulator (MOVES) 2014b. Office of Transportation and Air Quality, U.S.
Environmental Protection Agency. Available online at: .
EPA (2018c) Annual Certification Test Results Report. Office of Transportation and Air Quality, U.S.
Environmental Protection Agency. Available online at: .
EPA (2018d) Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975 -
2017. Office of Transportation and Air Quality, U.S. Environmental Protection Agency. Available online at:
.
EPA (2017a) Annual Certification Test Results Report. Office of Transportation and Air Quality, U.S.
Environmental Protection Agency. Available online at .
EPA (2017b) Confidential engine family sales data submitted to EPA by manufacturers. Office of Transportation
and Air Quality, U.S. Environmental Protection Agency.
EPA (2017c) Confidential Engine Family Sales Data Submitted to EPA by Manufacturers. Office of Transportation
and Air Quality, U.S. Environmental Protection Agency.
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: .
References 10-9
-------
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 (1997) Mobile Source Emission Factor Model (MOBlLE5a). Office of Mobile Sources, U.S. Environmental
Protection Agency, Ann Arbor, Michigan.
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 2017) Highway Statistics. Federal Highway Administration, U.S. Department of
Transportation, Washington, D.C. Report FHWA-PL-96-023-annual. Available online at:
.
FHWA (2017) Off-Highway and Public-Use Gasoline Consumption Estimation Models Used in the Federal
Highway Administration, Publication Number FHWA-PL-17-012. Available online at:
.
Gaffney, J. (2007) Email Communication. John Gaffney, American Public Transportation Association and Joe
Aamidor, ICF International. December 17, 2007.
HybridCars.com (2017). Monthly Plug-In Electric Vehicle Sales Dashboard, 2010-2017. 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.
ICF (2017a) Updated On-highway CH4 and N20 Emission Factors for GHG Inventory. Memorandum from ICF to
Sarah Roberts and Justine Geidosch, Office of Transportation and Air Quality, U.S. Environmental Protection
Agency. October 2017.
ICF (2017b) Updated Non-Highway CH4 and N20 Emission Factors for U.S. GHG Inventory. Memorandum from
ICF to Sarah Roberts and Justine Geidosch, Office of Transportation and Air Quality, U.S. Environmental
Protection Agency. October 2017.
Lipman, T. and M. Delucchi (2002) "Emissions of Nitrous Oxide and Methane from Conventional and Alternative
Fuel Motor Vehicles." Climate Change, 53:477-516.
10-10 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
SAE (2010) Utility Factor Definitions for Plug-In Hybrid Electric Vehicles Using Travel Survey Data. Society of
Automotive Engineers. Report J2841, Available online at: .
Santoni, G., B. Lee, E. Wood, S. Herndon, R. Miake-Lye, S. Wofsy, J. McManus, D. Nelson, M. Zahniser (2011)
Aircraft emissions of methane and nitrous oxide during the alternative aviation fuel experiment. Environ Sci
Technol. 2011 Aug 15; 45(16):7075-82.
U.S. Census Bureau (2000) Vehicle Inventory and Use Survey. U.S. Census Bureau, Washington, D.C. Database
CD-EC97-VIUS.
Whorton, D. (2006 through 2014) Personal communication, Class II and III Rail energy consumption, American
Short Line and Regional Railroad Association.
Carbon Emitted from Non-Energy Uses of Fossil Fuels
ACC (2017a) "Guide to the Business of Chemistry, 2017," American Chemistry Council.
ACC (2018a) "Guide to the Business of Chemistry, 2018" American Chemistry Council.
ACC (2018b) "U.S. Resin Production & Sales 2017 vs. 2016." Available online at:
.
ACC (2017b) "U.S. Resin Production & Sales 2016 vs. 2015."
ACC (2016a) "Guide to the Business of Chemistry, 2016" American Chemistry Council.
ACC (2016b) "U.S. Resin Production & Sales 2015 vs. 2014."
ACC (2015a) "PIPS Year-End Resin Statistics for 2014 vs. 2013: Production, Sales and Captive Use." Available
online at: .
ACC (2013) "U.S. Resin Production & Sales: 2012 vs. 2011," American Chemistry Council. Available online at:
ACC (2012) "Guide to the Business of Chemistry, 2012," American Chemistry Council.
ACC (2003-2011) "PIPS Year-End Resin Statistics for 2010: Production, Sales and Captive Use." 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/>.
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:
.
References 10-11
-------
C&EN (2017) U.S. textile makers look for a revival, Chemical & Engineering News, American Chemical Society,
19 June. Available online at: .
EIA (2019) Monthly Energy Review, February 2019. Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035 (2019/2).
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 (2017) EPA's Pesticides Industry Sales and Usage, 2008 - 2012 Market Estimates. Available online at:
Accessed September 2017.
EPA (2018a) "1970 - 2017 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, December
2016. Available online at: .
EPA (2018b) Advancing Sustainable Materials Management: Facts and Figures 2015, Assessing Trends in Material
Generation, Recycling and Disposal in the United States. Washington, D.C.
EPA (2018c) Resource Conservation and Recovery Act (RCRA) Info, Biennial Report, GM Form (Section 2- Onsite
Management) and WR Form.
EPA (2016a) Advancing Sustainable Materials Management: 2014 Facts and Figures Fact Sheet. Office of Solid
Waste and Emergency Response, U.S. Environmental Protection Agency, Washington, D.C. Available online at:
.
EPA (2016b) Resource Conservation and Recovery Act (RCRA) Info, Biennial Report, GM Form (Section 2- Onsite
Management) and WR Form.
EPA (2015) Resource Conservation and Recovery Act (RCRA) Info, Biennial Report, GM Form (Section 2- Onsite
Management) and WR Form.
EPA (2014a) Municipal Solid Waste in the United States: 2012 Facts and Figures. Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, Washington, D.C. Available online at:
.
EPA (2014b) Chemical Data Access Tool (CDAT). U.S. Environmental Protection Agency, June 2014. Available
online at: . Accessed January 2015.
EPA (2013a) Municipal Solid Waste in the United States: 2011 Facts and Figures. Office of Solid Waste and
Emergency Response, U.S. Environmental Protection Agency, Washington, D.C. Available online at:
.
EPA (2013b) Resource Conservation and Recovery Act (RCRA) Info, Biennial Report, GM Form (Section 2- Onsite
Management) and WR Form.
10-12 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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: .
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.
References 10-13
-------
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) Production 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.
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. EC02-3II-
325611 (RV). Available online at: .
U.S. Census Bureau (1999) Soap and Other Detergent Manufacturing: 1997. Available online at:
.
U.S. International Trade Commission (1990-2017) "Interactive Tariff and Trade DataWeb: Quick Query." Available
online at: . Accessed September 2018.
10-14 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 (2018a) Advancing Sustainable Materials Management: 2015 Data Tables. Office of Land and Emergency
Managements, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2016) Advancing Sustainable Materials Management: 2014 Fact Sheet. Office of Land and Emergency
Managements, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2015) Advancing Sustainable Materials Management: Facts and Figures 2013 - Assessing Trends in Material
Generation, Recycling and Disposal in the United States. Office of Solid Waste and Emergency Response, U.S.
Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2007, 2008, 2011, 2013, 2014) Municipal Solid Waste in the United States: Facts and Figures. Office of Solid
Waste and Emergency Response, U.S. Environmental Protection Agency. Washington, D.C. Available online at:
.
EPA (2006) Solid Waste Management and Greenhouse Gases: A Life-Cycle Assessment of Emissions and Sinks.
Office of Solid Waste and Emergency Response, U.S. Environmental Protection Agency. Washington, D.C.
EPA (2000) Characterization of Municipal Solid Waste in the United States: Source Data on the 1999 Update.
Office of Solid Waste, U.S. Environmental Protection Agency. Washington, D.C. EPA530-F-00-024.
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. https://www.ustires.org/system/files/USTMA_scraptire_summ_2017_072018.pdf. September 27, 2018
RMA (2016) "2015 U.S. Scrap Tire Management Summary." Rubber Manufacturers Association. August 2016.
Available online at: .
References 10-15
-------
RMA (2014) "2013 U.S. Scrap Tire Management Summary." Rubber Manufacturers Association. November 2014.
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.
RMA (2011) "U.S. Scrap Tire Management Summary 2005-2009." Rubber Manufacturers Association. October
2011. Available online at: .
Schneider, S. (2007) E-mail between Shelly Schneider of Franklin Associates (a division of ERG) and Sarah
Shapiro of ICF International, January 10, 2007.
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.
vanHaaren, Rob, Themelis, N., and Goldstein, N. (2010) "The State of Garbage in America." BioCycle, October
2010. Volume 51, Number 10, pg. 16-23.
Coal Mining
AAPG (1984) Coalbed Methane Resources of the United States. AAPG Studies in Geology Series #17.
Creedy, D.P. (1993) Methane Emissions from Coal Related Sources in Britain: Development of a Methodology.
Chemosphere, 26: 419-439.
DMME (2018) DGO Data Information System. Department of Mines, Minerals and Energy of Virginia. Available
online at .
EIA (2018) Annual Coal Report 2017. Table 1. Energy Information Administration, U.S. Department of Energy.
El Paso (2009) Shoal Creek Mine Plan, El Paso Exploration & Production.
EPA (2018) Greenhouse Gas Reporting Program (GHGRP) 2017 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.
Geological Survey of Alabama State Oil and Gas Board (GSA) (2018) Well Records Database. Available online at
.
IEA (2018) 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: EgglestonH.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 (2018) Marshall County VAM Abatement Project Offset Verification Statement submitted to
California Air Resources Board, August 2018.
10-16 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
MSHA (2018) 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) (2018) 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.
MSHA (2018) U.S. Department of Labor, Mine Health & Safety Administration, 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.
Drillinglnfo (2018) July 2018 Download. DI Desktop® 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 (2018) Greenhouse Gas Reporting Program. Environmental Protection Agency. Data reported as of August 19,
2018.
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.
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.
Natural Gas Systems
Drillinglnfo (2018) July 2018 Download. DI Desktop® Drillinglnfo, Inc.
EPA (2018) Greenhouse Gas Reporting Program- Subpart W - Petroleum and Natural Gas Systems. Environmental
Protection Agency. Data reported as of August 19, 2018.
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.
References 10-17
-------
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.
PHMSA (2018) Gas Distribution Annual Data. Pipeline and Hazardous Materials Safety Administration, U.S.
Department of Transportation, Washington, DC. Available online at: .
Zimmerle, et al. (2015) "Methane Emissions from the Natural Gas Transmission and Storage System in the United
States." Environmental Science and Technology, Vol. 49 9374-9383.
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).
Drillinglnfo (2018) July 2018 Download. DI Desktop® 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: .
10-18 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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:
.
jrces of Precursor Greenhouse Gases
EPA (2018) "1970 - 2017 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, March 2018.
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 (2018) 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 2018) 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, February 2019, Energy Information Administration, U.S. Department of
Energy, Washington, D.C. DOE/EIA-0035(2019/02).
FAA (2019) Personal Communication between FAA and Vince Camobreco for aviation emission estimates from the
Aviation Environmental Design Tool (AEDT). January 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, February 2019. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0035(2019/02).
EIA (2019b) Monthly Biodiesel Production Report. February 2019. Energy Information Administration, U.S.
Department of Energy. Washington, D.C.
EPA (2018) Acid Rain Program Dataset 1996-2017. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
References 10-19
-------
EPA (2010) Carbon Content Coefficients Developed for EPA's Mandatory Reporting Rule. Office of Air and
Radiation, Office of Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
Lindstrom, P. (2006) Personal Communication. Perry Lindstrom, Energy Information Administration and Jean Kim,
ICF International.
industrial Processes and Product Use
EPA (2014) Greenhouse Gas Reporting Program. Developments on Publication of Aggregated Greenhouse Gas
Data, November 25, 2014. See .
EPA (2002) Quality Assurance/Quality Control and Uncertainty Management Plan for the U.S. Greenhouse Gas
Inventory: Procedures Manual for Quality Assurance/Quality Control and Uncertainty Analysis, U.S. Greenhouse
Gas Inventory Program, U.S. Environmental Protection Agency, Office of Atmospheric Programs, EPA 430-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.: EgglestonH.S., Srivastava N., Tanabe K., Baasansuren J., Fukuda M., Pub. IGES, Japan 2011.
Cement Production
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) (2017) Mineral Industry Surveys: Cement in April 2017, June 2017.
USGS (2018a) Mineral Commodity Summaries: Cement 2018. U.S. Geological Survey, Reston, VA. January, 2018.
USGS (2018b) Mineral Industry Surveys: Cement in August 2018.
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.
10-20 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 2018a) 2017Mineral Commodities Summary: Lime. U.S. Geological
Survey, Reston, VA (January 2018).
USGS (2018b) (1992 through 2015) Minerals Yearbook: Lime. U.S. Geological Survey, Reston, VA (March 2018).
USGS (2017) 2016Mineral Commodities Summary: Lime. U.S. Geological Survey, Reston, VA (January 2017).
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.
IPCC (2006) 20061PCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
OIT (2002) Glass Industry of the Future: Energy and Environmental Profile of the U.S. Glass Industry. Office of
Industrial Technologies, U.S. Department of Energy. Washington, D.C.
U.S. Bureau of Mines (1991 and 1993a) Minerals Yearbook: Crushed Stone Annual Report. U.S. Department of the
Interior. Washington, D.C.
United States Geological Survey (USGS) (2017) Minerals Industry Surveys; Soda Ash in January 2017. U.S.
Geological Survey, Reston, VA. March, 2017.
USGS (2018) Mineral Industry Surveys: Soda Ash in February 2018. U.S. Geological Survey, Reston, VA.
Accessed September 2018.
USGS (1995 through 2015a) Minerals Yearbook: Crushed Stone Annual Report. U.S. Geological Survey, Reston,
VA.
USGS (1995 through 2015b) Miner als 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 (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.
Otf cess Uses of Carbonates
AISI (2018) 2017Annual 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.
References 10-21
-------
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 (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) Miner als 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.
Ammonia Production
ACC (2018) 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:
.
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: .
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:
.
10-22 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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: 2007 Summary.
Available online at: .
U.S. Census Bureau (2007) Current Industrial Reports Fertilizer Materials and Related Products: 2006 Summary.
Available online at: .
U.S. Census Bureau (2006) Current Industrial Reports Fertilizer Materials and Related Products: 2005 Summary.
Available online at: .
U.S. Census Bureau (2004, 2005) Current Industrial Reports Fertilizer Materials and Related Products: Fourth
Quarter Report Summary. Available online at: .
U.S. Census Bureau (1998 through 2003) Current Industrial Reports Fertilizer Materials and Related Products:
Annual Reports Summary. Available online at: .
U.S. Census Bureau (1991 through 1994) Current Industrial Reports Fertilizer Materials Annual Report. Report No.
MQ28B. U.S. Census Bureau, Washington, D.C.
United States Geological Survey (USGS) (2018) 2018 Mineral Commodity Summaries: Nitrogen (Fixed) -
Ammonia. January 2018. Available online at: .
USGS (2018a) 2016Minerals Yearbook: Nitrogen [Advance Release]. August 2018. Available online at:
.
USGS (2018b) Minerals Commodity Summaries: Nitrogen (Fixed) - Ammonia. Available online at:
. Accessed September 28, 2018.
USGS (2017) 2015 Minerals Yearbook: Nitrogen [Advance Release], August 2017. Available online at:
.
USGS (2016) 2014Minerals Yearbook: Nitrogen [Advance Release]. October 2016. Available online at:
.
USGS (2015) 2013Minerals Yearbook: Nitrogen [Advance Release]. August 2015. Available online at:
.
USGS (2014) 2012Minerals Yearbook: Nitrogen [Advance Release]. September 2014. Available online at:
.
USGS (1994 through 2009) Minerals Yearbook: Nitrogen. Available online at:
.
References 10-23
-------
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.
United States Geological Survey (USGS) (2014 through 2016) Minerals Yearbook: Nitrogen [Advance Release],
Available online at: .
USGS (1994 through 2018a) Minerals Yearbook: Nitrogen. Available online at:
.
USGS (2018b) Minerals Commodity Summaries: Nitrogen (Fixed) - Ammonia. Available online at:
. Accessed September 28, 2018.
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 (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) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
10-24 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 (2018) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.
C&EN (1995) "Production of Top 50 Chemicals Increased Substantially in 1994." Chemical & Engineering News,
73(15): 17. April 10, 1995.
C&EN (1994) "Top 50 Chemicals Production Rose Modestly Last Year." Chemical & Engineering News,
72(15): 13. April 11, 1994.
C&EN (1993) "Top 50 Chemicals Production Recovered Last Year." Chemical & Engineering News, 71(15): 11.
April 12, 1993.
C&EN (1992) "Production of Top 50 Chemicals Stagnates in 1991." Chemical & Engineering News, 70(15): 17.
April 13, 1992.
CMR (2001) "Chemical Profile: Adipic Acid." Chemical Market Reporter. July 16, 2001.
CMR(1998) "Chemical Profile: Adipic Acid." Chemical Market Reporter. June 15, 1998.
CW (2005) "Product Focus: Adipic Acid." Chemical Week. May 4, 2005.
CW (1999) "Product Focus: Adipic Acid/Adiponitrile." Chemical Week, p. 31. March 10, 1999.
Desai (2010, 2011) Personal communication. Mausami Desai, U.S. Environmental Protection Agency and Adipic
Acid Plant Engineers. 2010 and 2011.
EPA (2017) Greenhouse Gas Reporting Program. Full Subpart E, S-CEMS, BB, CC, LL Data Set (XLS) (Adipic
Acid Tab). Office of Air and Radiation, Office of Atmospheric Programs, U.S. Environmental Protection Agency,
Washington, D.C. Accessed on March 5, 2019, Available online at:
EPA (2014 through 2016) Greenhouse Gas Reporting Program. Full Subpart E, O, S-CEMS, BB, CC, LL, RR, F
Data Set (XLS) (Adipic Acid Tab). Office of Air and Radiation, Office of Atmospheric Programs, U.S.
Environmental Protection Agency, Washington, D.C. Accessed on November 15, 2017, 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.
References 10-25
-------
Reimer, R.A., Slaten, C.S., Seapan, M., Koch, T.A. and Triner, V.G. (1999) "Implementation of Technologies for
Abatement of N20 Emissions Associated with Adipic Acid Manufacture." Proceedings of the 2nd Symposium on
Non-C02 Greenhouse Gases (NCGG-2), Noordwijkerhout, The Netherlands, 8-10 Sept. 1999, Ed. J. van Ham etal.,
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 (2018) 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 October 2, 2018.
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 October 2, 2018.
Fibrant (2018). Fibrant LLC Contact Page. Retrieved from: http://www.fibrant52.com/en/contact on October 2,
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.
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
Silicon Carbide Production and Consumption
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
U.S. Census Bureau (2005 through 2018) 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 (2018b) Mineral Industry Surveys: Manufactured Abrasives in the Second Quarter. U.S. Geological Survey,
Reston, VA. Available online at: .
USGS (2018) Mineral Commodity Summaries: Abrasives (Manufactured). Available online at:
http://minerals.usgs.gov/minerals/pubs/commodity/abrasives/mcs-2018-abras.pdf
USGS (199 la through 2015) Miner als Yearbook: Manufactured Abrasives Annual Report. U.S. Geological Survey,
Reston, VA. Available online at: .
USGS (1991b through 2015) Miner als Yearbook: Silicon Annual Report. U.S. Geological Survey, Reston, VA.
Available online at: .
Titanium Dioxide Production
Gambogi, J. (2002) Telephone communication. Joseph Gambogi, Commodity Specialist, U.S. Geological Survey
and Philip Groth, ICF International. November 2002.
10-26 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
United States Geological Survey (2018) 2018 Mineral Commodity Summary: Titanium and Titanium Dioxide. U.S.
Geological Survey, Reston, Va. January, 2018. Available online at:
.
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) (2017) 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 (2018b) Mineral Commodity Summary: Soda Ash. U.S. Geological Survey, Reston, VA. Accessed September
2018.
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) Minerals Yearbook: Soda Ash Annual Report. U.S. Geological Survey, Reston, VA.
USGS (1995) 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, JohnR.
Petrochemical Production
ACC (2018) 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 (2017) Aggregation of Reported Facility Level Data under Subpart X -
National Petrochemical Production for Calendar Years 2010 through 2016. Office of Air and Radiation, Office of
Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.
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) 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
-------
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.
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.
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.
10-28 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
UNFCCC (2014) Report of the Conference of the Parties on its nineteenth session, held in Warsaw from 11 to 23
November 2013. United Nations Framework Convention on Climate Change, Warsaw. (FCCC/CP/2013/10/Add.3).
January 31, 2014. Available online at: .
Carbon Dioxide Consumption
ARI (1990 through 2010) 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 2017. 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.
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.
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) (2018) Mineral Commodity Summaries: Phosphate Rock 2018. January
2018. U.S. Geological Survey, Reston, VA. Available online at:
.
References 10-29
-------
United States Geological Survey (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:
-------
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.
Unites States Geological Survey (USGS) (2018) 2018 USGSMinerals 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) (2016a) 2014Minerals Yearbook: Ferroalloys. U.S. Geological Survey,
Reston, VA. October 2016.
USGS (2018) 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.
USGS (2015a) 2012Minerals Yearbook: Ferroalloys. U.S. Geological Survey, Reston, VA. April 2015.
USGS (2016b) Mineral Industry Surveys: Silicon in December 2016. U.S. Geological Survey, Reston, VA.
December 2016.
USGS (2015b) Mineral Industry Surveys: Silicon in June 2015. U.S. Geological Survey, Reston, VA. September
2015.
USGS (2014) Mineral Industry Surveys: Silicon in September 2014. U.S. Geological Survey, Reston, VA.
December 2014.
USGS (1996 through 2013) Minerals Yearbook: Silicon. U.S. Geological Survey, Reston, VA.
Aluminum Production
EPA (2018) Greenhouse Gas Reporting Program (GHGRP). Envirofacts, Subpart: F Aluminum Production.
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.
USAA (2018) U.S. Primary Aluminum Production: Report for August 2018. U.S. Aluminum Association,
Washington, D.C. September, 2017.
USAA (2017) U.S. Primary Aluminum Production: Report for September 2017. U.S. Aluminum Association,
Washington, D.C. October, 2017.
USAA (2016b) U.S. Primary Aluminum Production: Report for August 2016. U.S. Aluminum Association,
Washington, D.C. August, 2016.
USAA (2016a) U.S. Primary Aluminum Production: Report for February 2016. U.S. Aluminum Association,
Washington, D.C. March, 2016.
References 10-31
-------
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 (2018) 2018Mineral Commodity Summaries: Aluminum. U.S. Geological Survey, Reston, VA.
USGS (2007) 2006Mineral 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.
EPA (2018) 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, 1996Annual Conference Proceedings, International Magnesium
Association. Ube City, Japan.
IPCC (2006) 20061PCC 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 (2018, 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.
10-32 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Lead Production
Dutrizac, J.E., V. Ramachandran, and J.A. Gonzalez (2000) Lead-Zinc 2000. The Minerals, Metals, and Materials
Society.
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, 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) (2018) 2018 Mineral Commodity Summary, Lead. U.S. Geological
Survey, Reston, VA. January 2018. Available online at: <
https://minerals.usgs.gov/minerals/pubs/commodity/lead/mcs-2018-lead.pdl>.
USGS (2017) 2017Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA. January 2017.
USGS (2016) 2016Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA. January 2016.
USGS (2015) 2015Mineral Commodity Summary, Lead. U.S. Geological Survey, Reston, VA. January 2015.
USGS (201'4) 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.
*Jnc 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: .
References 10-33
-------
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.pdI>. 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
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) (2018) 2018 Mineral Commodity Summary: Zinc. U.S. Geological Survey,
Reston, VA. January 2018.
USGS (2017) 2017Mineral Commodity Summary: Zinc. U.S. Geological Survey, Reston, VA. January 2017.USGS
(2016) 2016Mineral Commodity Summary: Zinc. U.S. Geological Survey, Reston, VA. January 2016.
USGS (2015) 2015Mineral 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 ofLCA for the environmental evaluation of the recycling of galvanized steel. ISIJ
International, Vol. 40. No. 3, pp 292-299.
Semiconductor Manufacture
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.
10-34 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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, Q1-Q4 2009, Q1-Q4 2010. Available online at:
.
United States Census Bureau (USCB) (2011, 2012, 2015, 2016, 2017) 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) 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. 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:
U.S. EPA (2019b) Vintaging Model Foam Sector Updates. Greenhouse Gas Reporting Program. Prepared for U.S.
EPA's Stratospheric Protection Division by ICF under EPA Contract Number EP-BPA-16-H-0021. February 13,
2019.
U.S. EPA (2019c) Proposed First-fill Loss Rates for the Refrigeration and Air Conditioning Sector. Greenhouse
Gas Reporting Program. Prepared for U.S. EPA's Stratospheric Protection Division by ICF under EPA Contract
Number EP-BPA-16-H-0021. March 21, 2019.
References 10-35
-------
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. MeiraFilho, B.A. Callander, N. Harris, A. Kattenberg, andK. Maskell (eds.). Cambridge
University Press. Cambridge, United Kingdom.
Levin et al. (2010) "The Global SF6 Source Inferred from Long-term High Precision Atmospheric Measurements
and its Comparison with Emission Inventories." Atmospheric Chemistry and Physics, 10: 2655-2662.
O'Connell, P., F. Heil, J. Henriot, G. Mauthe, H. Morrison, L. Neimeyer, M. Pittroff, R. Probst, J.P. Tailebois
(2002) SF6 in the Electric Industry, Status 2000, CIGRE. February 2002.
RAND (2004) "Trends in SF6 Sales and End-Use Applications: 1961-2003," Katie D. Smythe. International
Conference on SF6 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.
10-36 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Tupman, M. (2003) Personal communication. Martin Tupman, Airgas Nitrous Oxide and Daniel Lieberman, ICF
International. August 8, 2003.
Industrial Processes and Product Use Sources of Precursor
Greenhouse Gases
EPA (2018) "1970 - 2017 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, March 2018.
Available online at: .
EPA (2003) Email correspondence containing preliminary ambient air pollutant data. Office of Air Pollution and the
Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. December 22, 2003.
EPA (1997) Compilation of Air Pollutant Emission Factors, AP-42. Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency. Research Triangle Park, NC. October 1997.
Agriculture
Enteric Fermentation
Archibeque, S. (2011) Personal Communication. Shawn Archibeque, Colorado State University, Fort Collins,
Colorado and staff at ICF International.
Crutzen, P. J., I. Aselmann, and W. Seiler (1986) Methane Production by Domestic Animals, Wild Ruminants, Other
Herbivores, Fauna, and Humans. Tellus, 38B:271-284.
Donovan, K. (1999) Personal Communication. Kacey Donovan, University of California at Davis and staff at ICF
International.
Doren, P.E., J. F. Baker, C. R. Long and T. C. Cartwright (1989) Estimating Parameters of Growth Curves of Bulls,
J Animal Science 67:1432-1445.
Enns, M. (2008) Personal Communication. Dr. Mark Enns, Colorado State University and staff at ICF International.
EPA (2002) Quality Assurance/Quality Control and Uncertainty Management Plan for the U.S. Greenhouse Gas
Inventory: Procedures Manual for Quality Assurance/Quality Control and Uncertainty Analysis, U.S. Greenhouse
Gas Inventory Program, U.S. Environmental Protection Agency, Office of Atmospheric Programs, EPA 430-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 of2001 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,
References 10-37
-------
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.). Cambridge University Press. Cambridge, United
Kingdom 996 pp.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
Johnson, D. (2002) Personal Communication. Don Johnson, Colorado State University, Fort Collins, and ICF
International.
Johnson, D. (1999) Personal Communication. Don Johnson, Colorado State University, Fort Collins, and David
Conneely, ICF International.
Kebreab E., K. A. Johnson, S. L. Archibeque, D. Pape, and T. Wirth (2008) Model for estimating enteric methane
emissions from United States dairy and feedlot cattle. J. Anim. Sci. 86: 2738-2748.
Lippke, H., T. D. Forbes, and W. C. Ellis. (2000) Effect of supplements on growth and forage intake by stacker
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 stacker 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 stacker 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 (2018) 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 /CalfHerds 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.
10-38 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
USD A: APHIS: VS (2002) Reference of2002 Dairy Management Practices. USDA-APHIS-VS, CEAH. Fort
Collins, CO. Available online at .
USD A: APHIS: VS (1998) Beef'97, Parts I-IV. USDA-APHIS-VS, CEAH. Fort Collins, CO. Available online at
.
USD A: APHIS: VS (1996) Reference of1996 Dairy Management Practices. USDA-APHIS-VS, CEAH. Fort
Collins, CO. Available online at .
USD A: APHIS: VS (1994) Beef Cow/Calf Health and Productivity Audit. USDA-APHIS-VS, CEAH. Fort Collins,
CO. Available online at .
USD A: 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, MI.]
Bryant, M.P., V.H. Varel, R.A. Frobish, andH.R. Isaacson (1976) InH.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 (2018) AgSTAR Anaerobic Digester Database. Available online at: . Accessed July 2018.
EPA (2016) AgSTAR Anaerobic Digester Database. Available online at: . Accessed July 2016.
EPA (2008) Climate leaders Greenhouse Gas Inventory Protocol Offset Project Methodology for Project Type
Managing Manure with Biogas Recovery Systems. Available online at:
.
EPA (2006) AgSTAR Digest. Office of Air and Radiation, U.S. Environmental Protection Agency. Washington, D.C.
Winter 2006. Available online at: . Retrieved July 2006.
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 (2003) AgSTAR Digest. Office of Air and Radiation, U.S. Environmental Protection Agency. Washington, D.C.
Winter 2003. Available online at: . Retrieved July 2006.
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 (2000) AgSTAR Digest. Office of Air and Radiation, U.S. Environmental Protection Agency. Washington, D.C.
Spring 2000. Available online at: .
References 10-39
-------
EPA (1992) Global Methane Emissions from Livestock and Poultry Manure, Office of Air and Radiation, U.S.
Environmental Protection Agency. February 1992.
ERG (2018) "Incorporation of USD A 2009 ARMS Swine Data into the Manure Management Greenhouse Gas
Inventory." Memorandum to USDA OCE and EPA from ERG, November 2018.
ERG (2010a) "Typical Animal Mass Values for Inventory Swine Categories." Memorandum to EPA from ERG.
July 19, 2010.
ERG (2010b) Telecon with William Boyd of USDA NRCS and Cortney Itle of ERG Concerning Updated VS and
Nex Rates. August 8, 2010.
ERG (2010c) "Updating Current Inventory Manure Characteristics new USDA Agricultural Waste Management
Field Handbook Values." Memorandum to EPA from ERG. August 13, 2010.
ERG (2008) "Methodology for Improving Methane Emissions Estimates and Emission Reductions from Anaerobic
Digestion System for the 1990-2007 Greenhouse Gas Inventory for Manure Management." Memorandum to EPA
from ERG. August 18, 2008.
ERG (2003a) "Methodology for Estimating Uncertainty for Manure Management Greenhouse Gas Inventory."
Contract No. GS-10F-0036, Task Order 005. Memorandum to EPA from ERG, Lexington, MA. September 26,
2003.
ERG (2003b) "Changes to Beef Calves and Beef Cows Typical Animal Mass in the Manure Management
Greenhouse Gas Inventory." Memorandum to EPA from ERG, October 7, 2003.
ERG (2001) Summary of development ofMDP 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 B0 Literature
Review). ERG, Lexington, MA. June 2000.
Gleason (2018) Personal communication. Karin Gleason, National Oceanic and Atmospheric Administration and
Cortney Itle, ERG. August 2018.
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 theASAE, 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.
10-40 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Robertson, G. P. and P. M. Groffman. 2015. Nitrogen transformations. Pages 421-446 inE. A. Paul, editor. Soil
Microbiology, Ecology, and Biochemistry. 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 (2018a) Quick Stats: Agricultural Statistics Database. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. Available online at: .
USDA (2018b) Chicken and Eggs 2017 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2018. Available online at:
.
USDA (2018c) 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 (2016c) 1987, 1992, 1997, 2002, 2007, and 2012 Census of Agriculture. National Agriculture Statistics
Service, U.S. Department of Agriculture. Washington, D.C. Available online at:
. July 2016.
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) 1987, 1992, 1997, 2002, 2007, and 2012 Census of Agriculture. National Agriculture Statistics
Service, U.S. Department of Agriculture. Washington, D.C. May 2014. Available online at:
.
USDA (2014b) Chicken and Eggs 2013 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2014. Available online at:
.
USDA (2014c) Poultry - Production and Value 2013 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2014. Available online at:
.
References 10-41
-------
USDA (2013a) Chicken and Eggs 2012 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2013. Available online at:
.
USDA (2013b) Poultry - Production and Value 2012 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2013. Available online at:
.
USDA (2012a) Chicken and Eggs 2011 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2012. Available online at:
.
USDA (2012b) Poultry - Production and Value 2011 Summary. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. April 2012. Available online at:
.
USDA (2011a) Chicken and Eggs 2010 Summary. National Agriculture Statistics Service, U.S. Department of
Agriculture. Washington, D.C. February 2011. Available online at:
.
USDA (201 lb) 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:
.
USD A (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:
.
10-42 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
USDA (1998) Chicken and Eggs—Final Estimates 1994-97. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. December 1998. Available online at:
.
USDA (1996) Agricultural Waste Management Field Handbook, National Engineering Handbook (NEH), Part 651.
Natural Resources Conservation Service, U.S. Department of Agriculture. July 1996.
USDA (1995a) Poultry Production and Value—Final Estimates 1988-1993. National Agriculture Statistics Service,
U.S. Department of Agriculture. Washington, D.C. March 1995. Available online at:
.
USDA (1995b) Chicken and Eggs—Final Estimates 1988-1993. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. December 1995. Available online at:
.
USDA (1994) Sheep and Goats—Final Estimates 1989-1993. National Agriculture Statistics Service, U.S.
Department of Agriculture. Washington, D.C. January 31, 1994. Available online at:
.
USDAAPHIS (2003) Sheep 2001, Part I: Reference ofSheep Management in the United States, 2001 and Part IV:
Baseline Reference of2001 Sheep Feedlot Health and Management. USDA-APHIS-VS. Fort Collins, CO.
#N356.0702. Available online at.
USDA APHIS (2000) Layers '99—Part II: References of1999 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 of1995 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 N20 Emissions from
U.S. Cropland Soils." Global Biogeochemical Cycles, 24, GB1009, doi:10.1029/2009GB003544.
Deren, C. (2002) Personal Communication and Dr. Chris Deren, Everglades Research and Education Centre at the
University of Florida and CarenMintz, 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.
References 10-43
-------
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." FEMSMicrobiology 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, L.R. Ahuja and W.J. Parton (eds.), American Society of Agriculture, Crop Society of
America and Soil Science Society of America, pp. 129-148.
10-44 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Parton, W.J., M.D. Hartman, D.S. Ojima, and D.S. Schimel (1998) "DAYCENT: Its Land Surface Submodel:
Description and Testing". Glob. Planet. Chang. 19: 35-48.
Parton, W.J., D.S. Schimel, C.V. Cole, D.S. Ojima (1987) "Analysis of factors controlling soil organic matter levels
in Great Plains grasslands." Soil Science Society of America Journal 51:1173-1179.
Sass, R. L. (2001) CH4 Emissions from Rice Agriculture. Good Practice Guidance and Uncertainty Management in
National Greenhouse Gas Inventories. 399-417. Available online at: .
Sass, R.L., F.M. Fisher, P. A. Harcombe, and F.T. Turner (1990) "Methane Production and Emissions in a Texas
Rice Field." Global Biogeochemical Cycles, 4:47-68.
Sass, R.L., F.M. Fisher, S.T. Lewis, M.F. Jund, and F.T. Turner. (1994) "Methane emissions from rice fields: effect
of soil texture." Global Biogeochemical Cycles 8:135-140.
Schueneman, T. (1997, 1999 through 2001) Personal Communication. Tom Schueneman, Agricultural Extension
Agent, Palm Beach County, FL and 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.
USD A (2005 through 2014) Crop Production Summary. National Agricultural Statistics Service, Agricultural
Statistics Board, U.S. Department of Agriculture, Washington, D.C. Available online at:
.
USD A (2012) Summary of USDA-ARS Research on the Interrelationship of Genetic and Cultural Management
Factors That Impact Grain Arsenic Accumulation in Rice. News and Events. Agricultural Research Service, U.S.
Department of Agriculture, Washington, D.C. Available online at:
. September 2013.
USDA (2003) Field Crops, Final Estimates 1997-2002. Statistical Bulletin No. 982. National Agricultural Statistics
Service, Agricultural Statistics Board, U.S. Department of Agriculture, Washington, D.C. Available online at:
. September 2005.
USDA (1998) Field Crops Final Estimates 1992-1997. Statistical Bulletin Number 947 a. National Agricultural
Statistics Service, Agricultural Statistics Board, U.S. Department of Agriculture, Washington, D.C. Available online
at: . July 2001.
USDA (1994) Field Crops Final Estimates 1987-1992. Statistical Bulletin Number 896. National Agricultural
Statistics Service, Agricultural Statistics Board, U.S. Department of Agriculture, Washington, D.C. Available online
at: . July 2001.
USDA-NRCS (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: .
References 10-45
-------
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, L.V. 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, L.V. 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:
.
Agricultural Soil Management
AAPFCO (2008 through 2018) 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.
10-46 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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 N20 Emissions from
U.S. Cropland Soils." GlobalBiogeochemical 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, WI. USA.
Del Grosso, S.J., W.J. Parton, A.R. Mosier, M.D. Hartman, J. Brenner, D.S. Ojima, and D.S. Schimel (2001)
"Simulated Interaction of Carbon Dynamics and Nitrogen Trace Gas Fluxes Using the DAYCENT Model." In
Schaffer, M., L. Ma, S. Hansen, (eds.). Modeling Carbon and Nitrogen Dynamics for Soil Management. CRC Press.
Boca Raton, Florida. 303-332.
Del Grosso, S.J., T. Wirth, S.M. Ogle, W.J. Parton (2008) Estimating agricultural nitrous oxide emissions. EOS 89,
529-530.
Delgado, J.A., S.J. Del Grosso, and S.M. Ogle (2009) "15N isotopic crop residue cycling studies and modeling
suggest that IPCC methodologies to assess residue contributions to N20-N emissions should be reevaluated."
Nutrient Cycling in Agroecosystems, DOI 10.1007/sl0705-009-9300-9.
Edmonds, L., N. Gollehon, R.L. Kellogg, B. Kintzer, L. Knight, C. Lander, J. Lemunyon, D. Meyer, D.C. Moffitt,
and J. Schaeffer (2003) "Costs Associated with Development and Implementation of Comprehensive Nutrient
Management Plans." Part 1. Nutrient Management, Land Treatment, Manure and Wastewater Handling and Storage,
and Recordkeeping. Natural Resource Conservation Service, U.S. Department of Agriculture.
EPA (2003) Clean Watersheds Needs Survey 2000—Report to Congress, U.S. Environmental Protection Agency.
Washington, D.C. Available online at: .
EPA (1999) Biosolids Generation, Use and Disposal in the United States. Office of Solid Waste, U.S.
Environmental Protection Agency. Available online at: .
EPA (1993) Federal Register. Part II. Standards for the Use and Disposal of Sewage Sludge; Final Rules. U.S.
Environmental Protection Agency, 40 CFR Parts 257, 403, and 503.
Firestone, M. K., and E.A. Davidson, Ed. (1989) Microbiological basis of NO and N20 production and consumption
in soil. Exchange of trace gases between terrestrial ecosystems and the atmosphere. New York, John Wiley & Sons.
ILENR (1993) Illinois Inventory of Greenhouse Gas Emissions and Sinks: 1990. Office of Research and Planning,
Illinois Department of Energy and Natural Resources. Springfield, IL.
IPCC (2013) 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories: Wetlands.
The Intergovernmental Panel on Climate Change. [T, Hiraishi, T. Krug, K. Tanabe, N. Srivastava, B. Jamsranjav,
M. Fukuda and T. Troxler (eds.)]. Hayama, Kanagawa, Japan.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
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.
References 10-47
-------
NASS (2004) Agricultural Chemical Usage: 2003 Field Crops Summary. Report AgChl(04)a, National Agricultural
Statistics Service, U.S. Department of Agriculture. Available online at:
.
NASS (1999) Agricultural Chemical Usage: 1998 Field Crops Summary. Report AgChl(99). National Agricultural
Statistics Service, U.S. Department of Agriculture. Available online at:
.
NASS (1992) Agricultural Chemical Usage: 1991 Field Crops Summary. Report AgChl(92). National Agricultural
Statistics Service, U.S. Department of Agriculture. Available online at:
.
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, 1CF Incorporated. July 30, 1996.
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
fromMODlS satellite data and ecosystem modeling. Earth Interactions 11, Article No. 13, DOl 10.1175/E1228.1.
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 (2015) PRISM Climate Data. Oregon State University. July 24, 2015. Available online at:
.
Ruddy B.C., D.L. Lorenz, and D.K. Mueller (2006) County-level estimates of nutrient inputs to the land surface of
the conterminous United States, 1982-2001. Scientific Investigations Report 2006-5012. U.S Department of the
Interior.
Scheer, C., S.J. Del Grosso, W.J. Parton, D.W. Rowlings, P.R. Grace (2013) Modeling Nitrous Oxide Emissions
from Irrigated Agriculture: Testing DAYCENT with High Frequency Measurements, Ecological Applications, in
press. Available online at: .
Soil Survey Staff (2011) State Soil Geographic (STATSGO) Database for State. Natural Resources Conservation
Service, United States Department of Agriculture. Available online at:
.
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 (2015) 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 (2018) Quick Stats. National Agricultural Statistics Service, United States Department of Agriculture,
Washington, D.C. .
10-48 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
.
Wisconsin Department of Natural Resources (1993) Wisconsin Greenhouse Gas Emissions: Estimates for 1990.
Bureau of Air Management, Wisconsin Department of Natural Resources, Madison, WI.
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," InMinerals 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," InMinerals 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," InMinerals Yearbook 1993. U.S. Department of the Interior/Bureau of
Mines, Washington, D.C. pp. 1107-1147.
Tepordei, V. V. (1994) "Crushed Stone," InMinerals Yearbook 1992. U.S. Department of the Interior/Bureau of
Mines, Washington, D.C. pp. 1279-1303.
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:
.
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:
.
References 10-49
-------
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 of2009, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2008) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of2008, U.S.
Geological Survey, Reston, VA. Available online at:
.
USGS (2007) Mineral Industry Surveys: Crushed Stone and Sand and Gravel in the First Quarter of2007. 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) "Crashed Stone," I n \ finerals Yearbook. U.S. Department of the Interior/U.S. Geological
Survey. Washington, D.C. Available online at: . Accessed: 30 August 2017.
Willett, J.C. (2018) Personal communication. Jason Willett. Preliminary data tables from "Crashed 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 "Crashed 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) Crashed 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," InMinerals 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) "Crashed Stone," InMinerals 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) "Crashed Stone," InMinerals 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) "Crashed Stone," InMinerals Yearbook 2011. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed May 2013.
Willett, J.C. (2013b) Personal Communication. Jason Willet, U.S. Geological Survey and ICF International.
September 9, 2013.
Willett, J.C. (2011a) "Crashed Stone," InMinerals 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) "Crashed Stone," InMinerals Yearbook 2010. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed September 2012.
10-50 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Willett, J.C. (2010) "Crashed 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) "Crashed Stone," InMinerals 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) "Crashed Stone," InMinerals 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) "Crashed Stone," InMinerals Yearbook 2006. U.S. Department of the Interior/U.S. Geological
Survey, Washington, D.C. Available online at:
. Accessed August 2008.
Urea Fertilization
AAPFCO (2008 through 2018) Commercial Fertilizers. Association of American Plant Food Control Officials.
University of Missouri. Columbia, MO.
AAPFCO (1995 through 2000a, 2002 through 2007) Commercial Fertilizers. Association of American Plant Food
Control Officials. University of Kentucky. Lexington, KY.
AAPFCO (2000b) 1999-2000 Commercial Fertilizers Data, ASCII files. Available from David Terry, Secretary,
AAPFCO.
EPA (2000) Preliminary Data Summary: Airport Deicing Operations (Revised). EPA-821-R-00-016. August 2000.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Itle, C. (2009) Email correspondence. Cortney Itle, ERG and Tom Wirth, U.S. Environmental Protection Agency on
the amount of urea used in aircraft deicing. January 7, 2009.
TVA (1991 through 1994) Commercial Fertilizers. Tennessee Valley Authority, Muscle Shoals, AL.
TVA (1992b) Fertilizer Summary Data 1992. Tennessee Valley Authority, Muscle Shoals, AL.
Field Burning of Agricultural Residues
Akintoye, H.A., Agbeyi, E.O., and Olaniyan, A.B. (2005) "The effects of live mulches on tomato (Lycopersicon
esculentum) yield under tropical conditions." Journal of Sustainable Agriculture 26: 27-37.
Bange, M.P., Milroy, S.P., and Thongbai, P. (2004) "Growth and yield of cotton in response to waterlogging." Field
Crops Research 88: 129-142.
Beyaert, R.P. (1996) The effect of cropping and tillage management on the dynamics of soil organic matter. PhD
Thesis. University of Guelph.
Bouquet, D. J., and Breitenbeck, G. A. (2000) "Nitrogen rate effect on partitioning of nitrogen and dry matter by
cotton." Crop Science 40: 1685-1693.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.Cantens, G.
(2004 through 2005) Personal Communication. Janet Lewis, Assistant to Gaston Cantens, Vice President of
Corporate Relations, Florida Crystals Company and ICF International.
Brouder, S.M., and Cassman, K.G (1990) "Root development of two cotton cultivars in relation to potassium uptake
and plant growth in a vermiculitic soilField 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." A iist. J. Agric. Res. 53: 201-210.
References 10-51
-------
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 of ten 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 1: 78-82.
Huett, D.O., and Dettman, E.B. (1991) Effect of nitrogen on growth, quality and nutrient uptake of cabbages grown
in sand culture. Australian Journal of Experimental Agriculture 29: 875-81.
Huett, D.O., and Dettman, B. (1989) "Nitrogen response surface models of zucchini squash, head lettuce and
potato." Plant and Soil 134: 243-254.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
IPCC/UNEP/OECD/IEA (1997) Revised 1996 IPCC Guidelines for National Greenhouse Gas Inventories.
Intergovernmental Panel on Climate Change, United Nations Environment Programme, Organization for Economic
Co-Operation and Development, International Energy Agency, Paris, France.
Jacobs, J.L., Ward, G.N., and Kearney, G. (2004) "Effects of irrigation strategies and nitrogen fertilizer on turnip
dry matter yield, water use efficiency, nutritive characteristics and mineral content in western Victoria." Australian
Journal of Experimental Agriculture 44: 13-26.
10-52 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 junceaL. Field Crops Research 37: 95-101.
LANDFIRE (2008) Existing Vegetation Type Layer, LANDFIRE 1.1.0, U.S. Department of the Interior, Geological
Survey. Accessed 28 October 2010 at .
MacLeod, L.B., Gupta, U.C., and Cutcliffe, J.A. (1971) "Effect of N, P, and K on root yield and nutrient levels in
the leaves and roots of rutabagas grown in a greenhouse." Plant and Soil 35: 281-288.
Mahrani, A., and Aharonov, B. (1964) "Rate of nitrogen absorption and dry matter production by upland cotton
grown under irrigation." Israel J. Agric. Res. 14: 3-9.
Marcussi, F.F.N., Boas, R.L.V., de Godoy, L.J.G., and Goto, R. (2004) "Macronutrient accumulation and
partitioning infertigated 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 ofthe 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, USD A, 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, CYCOCEL 75)." Expl Agric. 40: 381-387.
Moustakas, N.K., and Ntzanis, H. (2005) "Dry matter accumulation and nutrient uptake in flue-cured tobacco
(Nicotiana tabacum L.)." Field Crops Research 94: 1-13.
Peach, L., Benjamin, L.R., and Mead, A. (2000) "Effects on the growth of carrots (Daucus carota L.), cabbage
(Brassica oleracea var. capitata L.) and onion (Allium cepa L.) of restricting the ability of the plants to intercept
resources." Journal of Experimental Botany 51: 605-615.
Pettigrew, W.T., and Meredith, W.R., Jr. (1997) "Dry matter production, nutrient uptake, and growth of cotton as
affected by potassium fertilization." J.PlantNutr. 20:531-548.
Pettigrew, W.T., Meredith, W.R., Jr., and Young, L.D. (2005) "Potassium fertilization effects on cotton lint yield,
yield components, and reniform nematode populations." Agronomy Journal 97: 1245-1251.
PRISM Climate Group (2015) PRISM Climate Data. Oregon State University. July 24, 2015. Available online at:
.
Reid, J.B., and English, J.M. (2000) "Potential yield in carrots (Daucus carota L.): Theory, test, and an application."
Annals of Botany 85: 593-605.
References 10-53
-------
Sadras, V.O., and Wilson, L.J. (1997) "Growth analysis of cotton crops infested with spider mites: II. Partitioning of
dry matter." Crop Science 37: 492-497.
Scholberg, J., McNeal, B.L., Jones, J.W., Boote, K.J., Stanley, C.D., and Obreza, T.A. (2000a) "Growth and canopy
characteristics of field-grown tomato." Agronomy Journal 92: 152-159.
Scholberg, J., McNeal, B.L., Boote, K.J., Jones, J.W., Locasio, S.J., and Olson, S.M. (2000b) "Nitrogen stress
effects on growth and nitrogen accumulation by field-grown tomato." Agronomy Journal 92:159-167.
Singels, A. and Bezuidenhout, C.N. (2002) "A new method of simulating dry matter partitioning in the Canegro
sugarcane model." Field Crops Research 78: 151 - 164.
Sitompul, S.M., Hairiah, K., Cadisch, G., and VanNoordwuk, 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 (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.
.
USDA (2016) Quick Stats: U.S. & All States Data; Crops; Production and Area Harvested; 1990 - 2015. National
Agricultural Statistics Service, U.S. Department of Agriculture. Washington, D.C. U.S. Department of Agriculture,
National Agricultural Statistics Service. Washington, D.C., Available online at: .
Valantin, M., Gary, C., Vaissiere, B.E., and Frossard, J.S. (1999) "Effect of fruit load on partitioning of dry matter
and energy in cantaloupe (Cucumis melo L.)." Annals of Botany 84: 173-181.
Wallach, D., Marani, A., and Kletter, E. (1978) "The relation of cotton crop growth and development to final yield."
Field Crops Research 1: 283-294.
Wells, R., and Meredith, W.R., Jr. (1984) "Comparative growth of obsolete and modern cultivars. I. Vegetative dry
matter partitioning." Crop Science 24: 858-872.4
Wiedenfels, R.P. (2000) "Effects of irrigation and N fertilizer application on sugarcane yield and quality." Field
Crops Research A3: 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: .
10-54 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 IP CC Guidelines for National Greenhouse Gas Inventories: Wetlands.
Hiraishi, T., Krug, T., Tanabe, K., Srivastava, N., Baasansuren, J., Fukuda, M. and Troxler, T.G. (eds.). Published:
IPCC, Switzerland.
IPCC (2010) Revisiting the use of managed land as a proxy for estimating national anthropogenic emissions and
removals. [Eggleston HS, Srivastava N, Tanabe K, Baasansuren J, (eds.)]. Institute for Global Environmental
Studies, Intergovernmental Panel on Climate Change, Hayama, Kanagawa, Japan.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Jin, S., L. Yang, P. Danielson, C. Homer, J. Fry, and G. Xian. (2013) A comprehensive change detection method for
updating the National Land Cover Database to circa 2011. Remote Sensing of Environment, 132: 159-175.
NOAA Coastal Change Analysis Program (C-CAP) Regional Land Cover Database. Data collected 1995-present
Charleston, SC: National Oceanic and Atmospheric Administration (NOAA) Coastal Services Center. Data accessed
at: .
Nusser, S.M. and J.J. Goebel (1997) "The national resources inventory: a long-term multi-resource monitoring
programme." Environmental and Ecological Statistics 4:181-204.
Ogle, S.M., G. Domke, W.A. Zurz, M.T. Rocha, T. Huffman, A. Swan, J.E. Smith, C. Woodall, T. Krug (2018)
Delineating managed land for reporting greenhouse gas emissions and removals to the United Nations Framework
Convention on Climate Change. Carbon Balance and Management 13:9.
Smith, W.B., P.D. Miles, C.H. Perry, and S.A. Pugh (2009) Forest Resources of the United States, 2007. Gen. Tech.
Rep. WO-78. U.S. Department of Agriculture Forest Service, Washington, D.C.
U.S. Census Bureau (2010) Topologically Integrated Geographic Encoding and Referencing (TIGER) system
shapefiles. U.S. Census Bureau, Washington, D.C. Available online at: .
U.S. Department of Agriculture (2015) County Data - Livestock, 1990-2014. U.S. Department of Agriculture,
National Agriculture Statistics Service, Washington, D.C.
References 10-55
-------
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.
Forest Land Remaining Forest Land: Changes in Forest Carbon
Stocks
AF&PA (2006a and earlier) Statistical roundup. (Monthly). Washington, D.C. American Forest & Paper
Association.
AF&PA (2006b and earlier) Statistics of paper, paperboard and wood pulp. Washington, D.C. American Forest &
Paper Association.
Amichev, B.Y. and J.M. Galbraith (2004) "A Revised Methodology for Estimation of Forest Soil Carbon from
Spatial Soils and Forest Inventory Data Sets." Environmental Management 33(Suppl. 1):S74-S86.
Bechtold, W.A.; Patterson, P.L. (2005) The enhanced forest inventory and analysis program—national sampling
design and estimation procedures. Gen. Tech. Rep. SRS-80. Asheville, NC: U.S. Department of Agriculture Forest
Service, Southern Research Station. 85 p.
Birdsey, R. (1996) "Carbon Storage for Major Forest Types and Regions in the Conterminous United States." In
R.N. Sampson and D. Hair, (eds.). Forest and Global Change, Volume 2: Forest Management Opportunities for
Mitigating Carbon Emissions. American Forests. Washington, D.C., 1-26 and 261-379 (appendices 262 and 263).
Coulston, J.W., Wear, D.N., and Vose, J.M. (2015) Complex forest dynamics indicate potential for slowing carbon
accumulation in the southeastern United States. Scientific Reports. 5: 8002.
Domke, G.M., J.E. Smith, and C.W. Woodall. (2011) Accounting for density reduction and structural loss in
standing dead trees: Implications for forest biomass and carbon stock estimates in the United States. Carbon
Balance and Management. 6:14.
Domke, G.M., Woodall, C.W., Smith, J.E., Westfall, J.A., McRoberts, R.E. (2012) Consequences of alternative tree-
level biomass estimation procedures on U.S. forest carbon stock estimates. Forest Ecology and Management. 270:
108-116.
Domke, G.M., 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 U nitcd 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.
10-56 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 Jones, K.C. In preparation. U.S. timber production, trade, consumption, and price statistics 1965
to 2015. Res. Pap. FPL-RP-XXX. Madison, WI: USD A, 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, WI: USD A, 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, WI: USD A, 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, WI: USD A, 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 (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 (2007) Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth
Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen,
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA, 996 pp.
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(1): 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. andFolliott, 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.
References 10-57
-------
Russell, M.B.; Domke, G.M.; Woodall, C.W.; D'Amato, A.W. (2015) Comparisons of allometric and climate-
derived estimates of tree coarse root carbon in forests of the United States. Carbon Balance and Management. 10:
20.
Skog, K.E. (2008) Sequestration of carbon in harvested wood products for the United States. Forest Products
Journal 58:56-72.
Smith, J.E.; Heath, L.S.; Skog, K.E.; Birdsey, R.A. (2006) Methods for calculating forest ecosystem and harvested
carbon with standard estimates for forest types of the United States. Gen. Tech. Rep. NE-343. Newtown Square,
PA: U.S. Department of Agriculture, Forest Service, Northeastern Research Station. 216 p.
Smith, W. B., P. D. Miles, C. H. Perry, and S. A. Pugh (2009) Forest Resources of the United States, 2007. General
Technical Report WO-78, U.S. Department of Agriculture Forest Service, Washington Office.
Smith, J.E., L.S. Heath, and M.C. Nichols (2010) U.S. Forest Carbon Calculation Tool User's Guide: Forestland
Carbon Stocks and Net Annual Stock Change. General Technical Report NRS-13 revised, U.S. Department of
Agriculture Forest Service, Northern Research Station, 34 p.
Steer, Henry B. (1948) Lumber production in the United States. Misc. Pub. 669, U.S. Department of Agriculture
Forest Service. Washington, D.C.
Ulrich, Alice (1985) U.S. Timber Production, Trade, Consumption, and Price Statistics 1950-1985. Misc. Pub.
1453, U.S. Department of Agriculture Forest Service. Washington, D.C.
Ulrich, A.H. (1989) U.S. Timber Production, Trade, Consumption, and Price Statistics, 1950-1987. USDA
Miscellaneous Publication No. 1471, U.S. Department of Agriculture Forest Service. Washington, D.C., 77.
United Nations Framework Convention on Climate Change (2013) Report on the individual review of the inventory
submission of the United States of America submitted in 2012. FCCC/ARR/2012/USA. 42 p.
USDA Forest Service (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., L.S. Heath, G.M. Domke, and M.C. Nichols (2011a) Methods and equations for estimating
aboveground volume, biomass, and carbon for trees in the U.S. forest inventory, 2010. Gen. Tech. Rep. NRS-88.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 30 p.
10-58 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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. (201 lb) "Status and future of the forest health indicators program of the
United States." Environmental Monitoring and Assessment. 177: 419-436.
Woodall, C.W., and V.J. Monleon (2008) Sampling protocol, estimation, and analysis procedures for the down
woody materials indicator of the FIA program. Gen. Tech. Rep. NRS-22. Newtown Square, PA: U.S. Department of
Agriculture, Forest Service, Northern Research Station. 68 p.
Woodall, C.W., Walters, B.F., Oswalt, S.N., Domke, G.M., Toney, C., Gray, A.N. (2013) Biomass and carbon
attributes of downed woody materials in forests of the United States. Forest Ecology and Management 305: 48-59.
Woodall, C.W., Walters, B.F., Coulston, J.W., D'Amato, A.W., Domke, G.M., Russell, M.B., Sowers, P.A. (2015b)
Monitoring network confirms land use change is a substantial component of the forest carbon sink in the eastern
United States. Scientific Reports. 5: 17028.
Zhu, Zhiliang, and McGuire, A.D., eds., (2016) Baseline and projected future carbon storage and greenhouse-gas
fluxes in ecosystems of Alaska: U.S. Geological Survey Professional Paper 1826, 196 p., Available online at:
.
Forest Land Remaining Forest 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) 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.
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).
References 10-59
-------
Binkley, D. (2004) Email correspondence regarding the 95 percent confidence interval for area estimates of southern
pine plantations receiving N fertilizer (±20%) and the rate applied for areas receiving N fertilizer (100 to 200
pounds/acre). Dan Binkley, Department of Forest, Rangeland, and Watershed Stewardship, Colorado State
University and Stephen Del Grosso, Natural Resource Ecology Laboratory, Colorado State University. September
19, 2004.
Binkley, D., R. Carter, and H.L. Allen (1995) Nitrogen Fertilization Practices in Forestry. In: Nitrogen Fertilization
in the Environment, P.E. Bacon (ed.), Marcel Decker, Inc., New York.
Briggs, D. (2007) Management Practices on Pacific Northwest West-Side Industrial Forest Lands, 1991-2005: With
Projections to 2010. Stand Management Cooperative, SMC Working Paper Number 6, College of Forest Resources,
University of Washington, Seattle.
Fox, T.R., H. L. Allen, T.J. Albaugh, R. Rubilar, and C.A. Carlson (2007) Tree Nutrition and Forest Fertilization of
Pine Plantations in the Southern United States. Southern Journal of Applied Forestry, 31, 5-11.
IPCC (2006) 20061PCC 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.
USD A 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.
10-60 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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(1): 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.
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 (1997) "National Soil Survey Laboratory Characterization Data," Digital Data, Natural Resources
Conservation Service, U.S. Department of Agriculture. Lincoln, NE.
USDA-NRCS (2013) Summary Report: 2010 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: .
USDA-NRCS (2009) Summary Report: 2007 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: .
Woodall, C.W., L.S. Heath, G.M. Domke, and M.C. Nichols (2011a) Methods and equations for estimating
aboveground volume, biomass, and carbon for trees in the U.S. forest inventory, 2010. Gen. Tech. Rep. NRS-88.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 30 p.
Woodall, C. W., and V.J. Monleon (2008) Sampling protocol, estimation, and analysis procedures for the down
woody materials indicator of the FIA program. Gen. Tech. Rep. NRS-22. Newtown Square, PA: U.S. Department of
Agriculture, Forest Service, Northern Research Station. 68 p.
Woodall, C.W., Walters, B.F., Coulston, J.W., D'Amato, A.W., Domke, G.M., Russell, M.B., Sowers, P.A. (2015b)
Monitoring network confirms land use change is a substantial component of the forest carbon sink in the eastern
United States. Scientific Reports. 5: 17028.
Woodall, C.W., Walters, B.F., Oswalt, S.N., Domke, G.M., Toney, C., Gray, A.N. (2013) Biomass and carbon
attributes of downed woody materials in forests of the United States. Forest Ecology and Management 305: 48-59.
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.
References 10-61
-------
Brady, N.C. and R.R. Weil (1999) The Nature and Properties of Soils. Prentice Hall. Upper Saddle River, NJ, 881.
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.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
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, andD.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.
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: .
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, GPSR Tech. Report No. 4, USDA/ARS, Ft.
Collins, CO.
10-62 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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:
.
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:
.
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:
.
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., 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.
Ogle, S., M. Eve, M. Sperrow, F.J. Breidt, and K. Paustian (2002) Agricultural Soil C Emissions, 1990-2001:
Documentation to Accompany EPA Inventory Tables. Natural Resources Ecology Laboratory, Fort Collins, CO.
Provided in an e-mail from Stephen Ogle, NREL to Barbara Braatz, ICF International. September 23, 2002.
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, WI, 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.
References 10-63
-------
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
fromMODIS satellite data and ecosystem modeling. Earth Interactions 11, Article No. 13, DOI 10.1175/EI228.1.
PRISM Climate Group, Oregon State University, , created 24 July 2015.
Soil Survey Staff (2016) State Soil Geographic (STATSGO) Database for State. Natural Resources Conservation
Service, United States Department of Agriculture. Available online at:
.
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 (2015) 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-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: .
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.
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
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. 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.
10-64 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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, andD.S. Schimel (2001)
"Simulated Interaction of Carbon Dynamics and Nitrogen Trace Gas Fluxes Using the DAYCENT Model." In
Schaffer, M., L. Ma, S. Hansen, (eds.); Modeling Carbon and Nitrogen Dynamics for Soil Management. CRC Press.
Boca Raton, Florida. 303-332.
Domke, G.M., J.E. Smith, and C.W. Woodall. (2011) Accounting for density reduction and structural loss in
standing dead trees: Implications for forest biomass and carbon stock estimates in the United States. Carbon Balance
and Management. 6:14.
Domke, G.M., et al. (2013) From models to measurements: comparing down dead wood carbon stock estimates in
the U.S. forest inventory. PLoS ONE 8(3): e59949.
Domke, G.M., Perry, C.H., Walters, B.F., Woodall, C.W., and Smith, J.E. (2016) A framework for estimating litter
carbon stocks in forests of the United States. Science of the Total Environment 557-558: 469-478.
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and Wickham, J. (2011) Completion of
the 2006 National Land Cover Database for the Conterminous United States, PE&RS, Vol. 77(9):858-864.
Harmon, M.E., C.W. Woodall, B. Fasth, J. Sexton, M. Yatkov. (2011) Differences between standing and downed
dead tree wood density reduction factors: A comparison across decay classes and tree species. Res. Paper. NRS-15.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 40 p.
Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., and
Wickham, J. (2007) Completion of the 2001 National Land Cover Database for the Conterminous United States.
Photogrammetric Engineering and Remote Sensing, Vol. 73, No. 4, pp 337-341.
Homer, C.G., Dewitz, J.A., Yang, L., Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D.,
and Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
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.
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(1): 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, GPSR Tech. Report No. 4, USDA/ARS, Ft.
Collins, CO.
Ogle, S.M., F.J. Breidt, M. Easter, S. Williams, K. Killian, and K. Paustian (2010) "Scale and uncertainty in
modeled soil organic carbon stock changes for U.S. croplands using a process-based model." Global Change
Biology 16:810-820.
Ogle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003) "Uncertainty in estimating land use and management
impacts on soil organic carbon storage for U.S. agroecosystems between 1982 and 1997." Global Change Biology
9:1521-1542.
Parton, W.J., M.D. Hartman, D.S. Ojima, andD.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, WI, 147-167.
References 10-65
-------
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, Oregon State University, , created 24 July 2015.
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 (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: .
Woodall, C.W., and V.J. Monleon (2008) Sampling protocol, estimation, and analysis procedures for the down
woody materials indicator of the FIA program. Gen. Tech. Rep. NRS-22. Newtown Square, PA: U.S. Department of
Agriculture, Forest Service, Northern Research Station. 68 p.
Woodall, C.W., L.S. Heath, G.M. Domke, and M.C. Nichols (2011) Methods and equations for estimating
aboveground volume, biomass, and carbon for trees in the U.S. forest inventory, 2010. Gen. Tech. Rep. NRS-88.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 30 p.
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.
10-66 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Homer, C.G., Dewitz, J.A., Yang, L., Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing, v. 81,
no. 5, p. 345-354.
IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.
Kellogg, R.L., C.H. Lander, D.C. Moffitt, and N. Gollehon (2000) Manure Nutrients Relative to the Capacity of
Cropland and Pastureland to Assimilate Nutrients: Spatial and Temporal Trends for the United States. U.S.
Department of Agriculture, Washington, D.C. Publication number nps00-0579.
Metherell, A.K., L.A. Harding, C.V. Cole, and W.J. Parton (1993) "CENTURY Soil Organic Matter Model
Environment." Agroecosystem version 4.0. Technical documentation, GPSR Tech. Report No. 4, USDA/ARS, Ft.
Collins, CO.
NEBRA (2007) A National Biosolids Regulation, Quality, End Use & Disposal Survey. North East Biosolids and
Residuals Association. July 21, 2007.
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, WI, 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.
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 (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: .
Grassland Remaining Grassland: PJon-C02 Emiss
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 bio mass burning. Global
Biogeochemical Cycles 15:955-966.
References 10-67
-------
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, L.R. 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), pp.316-335.
Birdsey, R. (1996) "Carbon Storage for Major Forest Types and Regions in the Conterminous United States." In
R.N. Sampson and D. Hair, (eds.). Forest and Global Change, Volume 2: Forest Management Opportunities for
Mitigating Carbon Emissions. American Forests. Washington, D.C., 1-26 and 261-379 (appendices 262 and 263).
Breshears, D.D., Knapp, A.K., Law, D.J., Smith, M.D., Twidwell, D. and Wonkka, C.L., 2016. Rangeland
Responses to Predicted Increases in Drought Extremity. Rangelands, 38(4), pp. 191-196.
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
Del Grosso, S.J., S.M. Ogle, W.J. Parton. (2011) Soil organic matter cycling and greenhouse gas accounting
methodologies, Chapter 1, pp 3-13 DOI: 10.1021/bk-2011-1072.ch001. In: Understanding Greenhouse Gas
Emissions from Agricultural Management (L. Guo, A. Gunasekara, L. McConnell. Eds.), American Chemical
Society, Washington, D.C.
Del Grosso, S.J., W.J. Parton, A.R. Mosier, M.D. Hartman, J. Brenner, D.S. Ojima, andD.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.
10-68 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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.
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 inpinyon-juniper woodlands. Ecological Applications, 19(3), pp.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(1): 12-35.
Jurena, P.N. and Archer, S., (2003). Woody plant establishment and spatial heterogeneity in grasslands. Ecology,
84(4), pp.907-919.
Lenihan, J.M., Drapek, R., Bachelet, D. and Neilson, R.P., (2003). Climate change effects on vegetation distribution,
carbon, and fire in California. Ecological Applications, 13(6), pp.1667-1681.
Metherell, A.K., L.A. Harding, C.V. Cole, and W.J. Parton (1993) "CENTURY Soil Organic Matter Model
Environment." Agroecosystem version 4.0. Technical documentation, GPSR Tech. Report No. 4, USDA/ARS, Ft.
Collins, CO.
Ogle, S.M., F.J. Breidt, M. Easter, S. Williams, K. Killian, and K. Paustian (2010) "Scale and uncertainty in
modeled soil organic carbon stock changes for U.S. croplands using a process-based model." Global Change
Biology 16:810-820.
Ogle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003) "Uncertainty in estimating land use and management
impacts on soil organic carbon storage for U.S. agroecosystems between 1982 and 1997." Global Change Biology
9:1521-1542.
References 10-69
-------
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, WI, 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.
Schimel, D.S. (1995) "Terrestrial ecosystems and the carbon cycle." Global Change Biology 1: 77-91.
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 of ten 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), pp.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 (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 (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: .
Woodall, C. W., and V.J. Monleon (2008) Sampling protocol, estimation, and analysis procedures for the down
woody materials indicator of the FIA program. Gen. Tech. Rep. NRS-22. Newtown Square, PA: U.S. Department of
Agriculture, Forest Service, Northern Research Station. 68 p.
Woodall, C.W., L.S. Heath, G.M. Domke, and M.C. Nichols. (2011) Methods and equations for estimating
aboveground volume, biomass, and carbon for trees in the U.S. forest inventory, 2010. Gen. Tech. Rep. NRS-88.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 30 p.
Wetlands Remaining Wetlands: C02, CH4, and N20 Emissions
lands Remaining Peatlands
Apodaca, L. (2011) Email correspondence. Lori Apodaca, Peat Commodity Specialist, USGS and Emily Rowan,
ICF International. November.
Apodaca, L. (2008) E-mail correspondance. Lori Apodaca, Peat Commodity Specialist, USGS and Emily Rowan,
ICF International. October and November.
10-70 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 (1993-2015)
Alaska's Mineral Industry Report (1997-2014). Alaska Department of Natural Resources, Fairbanks, AK. Available
online at .
IPCC (2014) 2013 Supplement to the 2006 IP CC 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 1, 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 (1991-2016) Minerals Yearbook: Peat (1994-2016). United States Geological Survey, Reston, VA.
Available online at .
USGS (2018) Mineral Commodity Summaries: Peat (2018). United States Geological Survey, Reston, VA.
Available online at .
Wetlands Remaining Coastal Wetlands: Emissions and
Removals from Coastal Wetlands Remaining Coastal Wetlands
Bianchi, T. S., Allison, M. A., Zhao, J., Li, X., Comeaux, R. S., Feagin, R. A., & Kulawardhana, R. W. (2013)
Historical reconstruction of mangrove expansion in the Gulf of Mexico: linking climate change with carbon
sequestration in coastal wetlands. Estuarine, Coastal and Shelf Science 119: 7-16.
Byrd, K. B., Ballanti, L. R., Thomas, N. M., Nguyen, D. K., Holmquist, J. R., Simard, M., Windham-Myers, L.,
Schile, L. M., Parker, V. T., ... and Castaneda-Moya, E. (2017) Biomass/Remote Sensing dataset: 30m resolution
tidal marsh biomass samples and remote sensing data for six regions in the conterminous United States: U.S.
Geological Survey data release, 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., B arras. J. A., St ever, 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.
References 10-71
-------
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) N;0 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 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.
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 (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., 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.
10-72 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Ph.D. Dissertation, University of Maryland, College Park, MD, 342pp.
National Marine Fisheries Service (2016) Fisheries of the United States, 2015. U.S. Department of Commerce,
NOAA Current Fisheries Statistics No. 2015.
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 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.
References 10-73
-------
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.
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.
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. 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.
10-74 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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) 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.
NRCS (1999) Soil Taxonomy: A basic system of soil classification for making and interpreting soil surveys, 2nd
Edition. Agricultural Handbook Number 436, Natural Resources Conservation Service, U.S. Department of
Agriculture, Washington, D.C.
Nusser, S.M. and J.J. Goebel (1997) The national resources inventory: a long-term multi-resource monitoring
programme. Environmental and Ecological Statistics 4:181-204.
Ogle, S.M., M.D. Eve, F.J. Breidt, and K. Paustian (2003) Uncertainty in estimating land use and management
impacts on soil organic carbon storage for U.S. agroecosystems between 1982 and 1997. Global Change Biology
9:1521-1542.
Soil Survey Staff (2011) State Soil Geographic (STATSGO) Database for State. Natural Resources Conservation
Service, United States Department of Agriculture. Available online at:
.
USDA-NRCS (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.
.
Settlements Remaining Settlements: Changes in Carbon Stocks
in Settlement Trees
deVries, R.E. (1987) A Preliminary Investigation of the Growth and Longevity of Trees in Central Park. M.S. thesis,
Rutgers University, New Brunswick, NJ.
Fleming, L.E. (1988) Growth Estimation of Street Trees in Central New Jersey. M.S. thesis, Rutgers University,
New Brunswick, NJ.
Frelich, L.E. (1992) Predicting Dimensional Relationships for Twin Cities Shade Trees. University of Minnesota,
Department of Forest Resources, St. Paul, MN, p. 33.
IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
MRLC (2013) National Land Cover Database 2001 (NLCD2001). Available online at:
. Accessed August 2013.
Nowak, D.J. (1986) Silvics of an Urban Tree Species: Norway maple (Acer platanoides L.). M.S. thesis, College of
Environmental Science and Forestry, State University of New York, Syracuse, NY.
Nowak, D.J. (1994) Atmospheric carbon dioxide reduction by Chicago's urban forest. In: Chicago's Urban Forest
Ecosystem: Results of the Chicago Urban Forest Climate Project. E.G. McPherson, D.J. Nowak, and R.A. Rowntree
(eds.). General Technical Report NE-186. U.S. Department of Agriculture Forest Service, Radnor, PA. pp. 83-94.
References 10-75
-------
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.
Nowak, D.J., R.E. Hoehn, D.E. Crane, J.C. Stevens, J.T. Walton, and J. Bond (2008) A ground-based method of
assessing urban forest structure and ecosystem services. Arboric. Urb. For. 34(6): 347-358.
Nowak, D.J., E.J. Greenfield, R.E. Hoehn, andE. 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
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer.
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.
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.
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 (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.
.
10-76 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Settlements Remaining Settlements: Changes in Yard
Trimmings and Food Scrap Carbon Stocks in Landfills
Barlaz, M.A. (2008) "Re: Corrections to Previously Published Carbon Storage Factors." Memorandum to Randall
Freed, ICF International. February 28, 2008.
Barlaz, M.A. (2005) "Decomposition of Leaves in Simulated Landfill." Letter report to Randall Freed, ICF
Consulting. June 29, 2005.
Barlaz, M.A. (1998) "Carbon Storage during Biodegradation of Municipal Solid Waste Components in Laboratory-
Scale Landfills." Global Biogeochemical Cycles 12:373-380.
De la Cruz, F.B. and M.A. Barlaz (2010) "Estimation of Waste Component Specific Landfill Decay Rates Using
Laboratory-Scale Decomposition Data" Environmental Science & Technology 44:4722- 4728.
Eleazer, W.E., W.S. Odle, Y. Wang, and M.A. Barlaz (1997) "Biodegradability of Municipal Solid Waste
Components in Laboratory-Scale Landfills." Environmental Science & Technology 31:911-917.
EPA (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) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse Gas
Inventories Programme, The Intergovernmental Panel on Climate Change. H.S. Eggleston, L. Buendia, K. Miwa, T.
Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
IPCC (2003) Good Practice Guidance for Land Use, Land-Use Change, and Forestry. The Intergovernmental Panel
on Climate Change, National Greenhouse Gas Inventories Programme, J. Penman et al. (eds.). Available online at
.
Oshins, C. and D. Block (2000) "Feedstock Composition at Composting Sites." Biocycle 41(9):31—34.
Tchobanoglous, G., H. Theisen, and S.A. Vigil (1993) Integrated Solid Waste Management, 1st edition. McGraw-
Hill, NY. Cited by Barlaz (1998) "Carbon Storage during Biodegradation of Municipal Solid Waste Components in
Laboratory-Scale Landfills." Global Biogeochemical Cycles 12:373-380.
Land Converted to Settlements
Birdsey, R. (1996) "Carbon Storage for Major Forest Types and Regions in the Conterminous United States." In
R.N. Sampson and D. Hair, (eds.). Forest and Global Change, Volume 2: Forest Management Opportunities for
Mitigating Carbon Emissions. American Forests. Washington, D.C., 1-26 and 261-379 (appendices 262 and 263).
Brockwell, Peter J., and Richard A. Davis (2016) Introduction to time series and forecasting. Springer. Domke,
G.M., Perry, C.H., Walters, B.F., Woodall, C.W., and Smith, J.E. (2016) A framework for estimating litter carbon
stocks in forests of the United States. Science of the Total Environment 557-558: 469-478.
Domke, G.M., J.E. Smith, and C.W. Woodall. (2011) Accounting for density reduction and structural loss in
standing dead trees: Implications for forest biomass and carbon stock estimates in the United States. Carbon Balance
and Management. 6:14.
References 10-77
-------
Domke, G.M., Woodall, C.W., Walters, B.F., Smith, J.E. (2013) From models to measurements: comparing down
dead wood carbon stock estimates in the U.S. forest inventory. PLoS ONE 8(3): e59949.
Domke, G.M., Perry, C.H., Walters, B.F., Woodall, C.W., and Smith, J.E. (2016) A framework for estimating litter
carbon stocks in forests of the United States. Science of the Total Environment 557-558: 469-478.
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and Wickham, J. (2011)
Completion of the 2006 National Land Cover Database for the Conterminous United States, PE&RS, Vol.
77(9):858-864.
Harmon, M.E., C.W. Woodall, B. Fasth, J. Sexton, M. Yatkov. (2011) Differences between standing and downed
dead tree wood density reduction factors: A comparison across decay classes and tree species. Res. Paper. NRS-15.
Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 40 p.
Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., and
Wickham, J. (2007) Completion of the 2001 National Land Cover Database for the Conterminous United States.
Photogrammetric Engineering and Remote Sensing, Vol. 73, No. 4, pp 337-341.
Homer, C.G., Dewitz, J.A., Yang, L., Jin, S., Danielson, P., Xian, G., Coulston, J., Herold, N.D., Wickham, J.D., and
Megown, K. (2015) Completion of the 2011 National Land Cover Database for the conterminous United States-
Representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing, v. 81,
no. 5, p. 345-354.
IPCC (2006) 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(1): 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 (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.
.
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.
10-78 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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., andR. 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. Ciill, 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 (2018a) Landfill Methane Outreach Program (LMOP). 2018 Landfill and Project Level Data. September 2018.
Available online at: < https://www.epa.gov/lmop/landfill-gas-energy-project-data>.
EPA (2018b) Greenhouse Gas Reporting Program (GHGRP). 2018 Envirofacts. Subpart HH: Municipal Solid Waste
Landfills and Subpart TT: Industrial Waste Landfills. Available online at:
.
EPA (2018c) Advancing Sustainable Materials Management: Facts and Figures 2015. July 2018. Available online
at: .
EPA (2016a) Industrial and Construction and Demolition Landfills. Available online at:
https://www.epa.gov/landfills/industrial-and-construction-and-demolition-cd-landfills.
EPA (2016b) Landfill Gas-to-Energy Project Database. Landfill Methane and Outreach Program. August 2015.
EPA (2015a) Greenhouse Gas Reporting Program (GHGRP). 2015 Envirofacts. Subpart HH: Municipal Solid Waste
Landfills. Available online at: .
EPA (2015b) Advancing Sustainable Materials Management: Facts and Figures 2013. June 2015. Available online
at: < https://www.epa.gov/sites/production/files/2015-09/documents/2013_advncng_smm_rpt.pdi>.
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.
References 10-79
-------
EREF (The Environmental Research & Education Foundation) (2016). Municipal Solid Waste Management in the
United States: 2010 & 2013.
ERG (2018) Draft Production Data Supplied by ERG for 1990-2017 for Pulp and Paper, Fruits and Vegetables, and
Meat. August.
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). In
progress.
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). In
progress.
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 (2013) Review of State of Garbage data used in the U.S. Non-CCh Greenhouse Gas Inventory for Landfills.
Memorandum prepared by K. Weitz and K. Bronstein (RTI) for R. Schmeltz (EPA). November 25, 2013.
RTI (2011) Updated Research on Methane Oxidation in Landfills. Memorandum prepared by K. Weitz (RTI) for R.
Schmeltz (EPA). January 14, 2011.
Shin, D. (2014) Generation and Disposition of Municipal Solid Waste (MSW) in the United States - A National
Survey. Master of Science thesis submitted to the Department of Earth and Environmental Engineering Fu
Foundation School of Engineering and Applied Science, Columbia University. January 3, 2014. Available online at:
.
U.S. Census Bureau (2018) Annual Estimates of the Resident Population: April 1, 2010 to July 1, 2017. Available
online at
.
Waste Business Journal (WBJ) (2016) Directory of Waste Processing & Disposal Sites 2016.
WBJ (2010) Directory of Waste Processing & Disposal Sites 2010.
Wastewater Treatment
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.
10-80 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
Beer Institute (2011) Brewers Almanac. Available online at: .
Benyahia, F., M. Abdulkarim, A. Embaby, and M. Rao. (2006) Refinery Wastewater Treatment: A true
Technological Challenge. Presented at the Seventh Annual U. A.E. University Research Conference.
BIER (2017) Beverage Industry Environmental Roundtable. 2016 Trends and Observations. Available online at:
. Accessed April 2018.
Brewers Association (2018) Statistics: Number of Breweries. Available online at:
. Accessed July 2018.
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.
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 (2018) Energy Information Administration. U.S. Refinery and Blender Net Production of Crude Oil and
Petroleum Products (Thousand Barrels). Available online at:
. Accessed May 2018.
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:
. Accessed December
2015.
References 10-81
-------
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.
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.
10-82 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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 (2018a) FAOSTAT-Forestry Database. Available online at:
. Accessed May 2018, August 2018.
FAO (2018b) "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, and
2016 - 2021 reports. Available online at:< http://www.fao.org/forestry/statistics/81757/en/> Accessed August 2018.
FAO (2018c) FAOSTAT-Food Balance Sheets. Available online at:
. Accessed May 2018.
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 IP CC 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.
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.
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.
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.
References 10-83
-------
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 (2018) Alcohol and Tobacco Tax and Trade Bureau. Beer Statistics. Available online at:
. Accessed May 2018.
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 (2018) International Database. Available online at:
. Accessed May 2018.
U.S. Census Bureau (2015) "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, and 2015 reports. Available
online at . Accessed May 2018.
USDA (2018a) U.S. Department of Agriculture. National Agricultural Statistics Service. Washington, D.C.
Available online at: and <
https://quickstats.nass.usda.gov/>. Accessed May 2018.
USDA (2018b) U.S. Department of Agriculture. Economic Research Service. Nutrient Availability. Washington
D.C. Available online at:
.
Accessed May 2018.
USDA (2018c) U.S. Department of Agriculture. National Agricultural Statistics Service. Vegetables 2017 Summary.
Available online at: . Accessed May 2018.
U.S. Poultry (2006) Email correspondence. John Starkey, USPOULTRY to D. Bartram, ERG. 30 August 2006.
White and Johnson (2003) White, P.J. and Johnson, L.A. Editors. Corn: Chemistry and Technology. 2nd ed.
AACC Monograph Series. American Association of Cereal Chemists. St. Paul, MN.
World Bank (1999) Pollution Prevention and Abatement Handbook 1998, Toward Cleaner Production. The
International Bank for Reconstruction and Development, The World Bank, Washington, D.C. ISBN 0-8213-3638-
X.
Composting
BioCycle (2010) The State of Garbage in America. Prepared by Rob van Haaren, Nickolas Themelis and Nora
Goldstein. Available online at .
Cornell Waste Management Institute (2007) The Science of Composting. 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
.
10-84 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
-------
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) 20061PCC Guidelines for National Greenhouse Gas Inventories. Volume 5: Waste, Chapter4:
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 < http://ilsr.org/wp-content/uploads/2014/07/state-of-composting-in-us.pdl>.
U.S. Census Bureau (2018) Population Estimates: Vintage 2017 Annual Estimates of the Resident Population for the
United States, Regions, States, and Puerto Rico, April 1, 2010 to July 1, 2017. Available online at <
https://www2.census.gov/programs-surveys/popest/tables/2010-2017/state/totals/nst-est2017-01 ,xlsx>.
U.S. Census Bureau (2017) Population Estimates: Vintage 2016 Annual Estimates of the Resident Population for the
United States, Regions, States, and Puerto Rico, April 1, 2010 to July 1, 2016. Available online at <
https://www2.census.gov/programs-surveys/popest/tables/2010-2016/state/totals/nst-est2016-01 ,xlsx>.
U.S. Census Bureau (2016) Population Estimates: Vintage 2015 Annual Estimates of the Resident Population for the
United States, Regions, States, and Puerto Rico, April 1, 2010 to July 1, 2015. Available online at
.
U.S. Composting Council (2010) Yard Trimmings Bans: Impact and Support. Prepared by Stuart Buckner,
Executive Director, U.S., Composting Council. Available online at
.
Waste Incineration
RTI (2009) Updated Hospital/Medical/Infectious Waste Incinerator (HMIWI) Inventory Database. Memo dated July
6, 2009. Available online at: .
Waste Sources of Precursor Greenhouse Gas Emissions
EPA (2018) "1970 - 2017 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, March 2018.
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
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.
References 10-85
-------
Drillinglnfo (2018) July 2018 Download. DI Desktop® Drillinglnfo, Inc.
EIA (2019) Monthly Energy Review, February 2019. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0035(2019/02).
EIA (2018) Monthly Energy Review, February 2018. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. DOE/EIA-0035(2018/02).
EIA (2017) International Energy Statistics 1980-2014. Energy Information Administration, U.S. Department of
Energy. Washington, D.C. Available online at: .
EPA (2018) Acid Rain Program Dataset 1996-2017. Office of Air and Radiation, Office of Atmospheric Programs,
U.S. Environmental Protection Agency, Washington, D.C.
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.
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) 20061PCC 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.
Lu, M & Megonigal, J.P. (2017) Final Report for RAE Baseline Assessment Project. Memo to Silvestrum Climate
Associates by Smithsonian Environmental Research Center, Maryland.
Ruefenacht, B., M.V. Finco, M.D. Nelson, R. Czaplewski, E.H. Helmer, J.A. Blackard, G.R. Holden, A.J. Lister, D.
Salajanu, D. Weyermann, K. Winterberger (2008) Conterminous U.S. and Alaska Forest Type Mapping Using
Forest Inventory and Analysis. USD A Forest Service - Forest Inventory and Analysis Program & Remote Sensing
Applications Center. Available online at: . Accessed 8
September 2015.
USDA Forest Service (2018) 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.
10-86 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2017
------- |