EPA-430-R-19-012
September 2019
Global N011-CO2 Greenhouse Gas
Emission Projections &
Marginal Abatement Cost Analysis:
Methodology Documentation
Report
Prepared for
U.S. Environmental Protection Agency
Office of Atmospheric Programs,
Climate Change Division
Washington, DC
Prepared by
RTI International
3040 E. Cornwallis Road
Research Triangle Park, NC 27709
Abt Associates, Inc.
10 Fawcett Street
Cambridge, MA 02138
ICF
9300 Lee Highway
Fairfax, VA 22031
-------
-------
Contents
Section Page
1 Background and Introduction 1-1
2 Non-COzGHGs 2-1
2.1 GWPs Assigned to Non-CCh GHGs 2-2
2.2 Methane (CH4) 2-2
2.3 Nitrous Oxide (N2O) 2-2
2.4 Fluorinated Gases (F-GHGs) 2-3
2.4.1 Hydrofluorocarbons (HFCs) 2-3
2.4.2 Perfluorocarbons (PFCs) 2-3
2.4.3 Sulphur Hexafluoride (SFs) 2-3
2.4.4 Nitrogen Trifluoride (NF3) 2-3
2.5 References 2-3
3 Baseline Emissions Calculation—General Methodology 3-1
3.1 Data Sources 3-2
3.1.1 Comparative Analysis 3-2
3.2 Calculating Historical and Projected Non-CCh Emissions 3-2
3.3 Generating the Composite Historical and Projected Non-CCh Emissions 3-3
3.3.1 Emission Factors in the Composite Projections 3-6
3.3.2 Reductions in the Baseline Scenario 3-6
3.3.3 Excluded Data 3-6
3.4 References 3-6
4 MAC Analysis—General Methodology 4-1
4.1 Methodology Overview 4-1
4.1.1 Technical Characteristics of Abatement Options 4-2
4.1.2 Economic Characteristics of Abatement Options 4-4
4.1.3 International Adjustment Factors 4-5
4.1.4 MAC Curves and Their Construction 4-6
4.2 MAC Methodological Enhancements 4-8
4.2.1 Incorporating Technological Change 4-9
4.2.1.1 Background 4-9
4.2.1.2 Operationalizing Technological Change 4-10
4.2.1.3 Defining an Experience Curve 4-10
4.2.1.4 Choosing a Learning Rate 4-11
4.2.1.5 Accounting for Technological Maturity and Variation in Country
Development Status 4-12
RTI International is a registered trademark and a trade name of Research Triangle Institute.
-------
4.2.1.6 Incorporating Dynamic Heterogeneity Across Countries 4-14
4.2.1.7 Integration into the Non-CC>2 MACC Model 4-15
4.2.2 Regionalization of Sectoral MAC Curves for the United States 4-17
4.2.2.1 Specification of Baseline GHG Emissions by Region and Sector 4-18
4.2.2.2 Regional Price Adjustment Factors (KLEM) 4-18
4.2.3 MAC Limitations and Uncertainties 4-20
4.3 References 4-21
5 Sector-Level Methods 5-1
5.1 Energy Sector 5-1
5.1.1 Cm in Coal Mining 5-1
5.1.1.1 Coal Mining Projections Methodology 5-1
5.1.1.2 Mitigation Options Considered 5-4
5.1.1.3 Sector-Level Trends/Considerations 5-9
5.1.1.4 References 5-9
5.1.2 Oil and Natural Gas Systems 5-11
5.1.2.1 Oil and Natural Gas Systems Emission Projections 5-12
5.1.2.2 Oil and Natural Gas Systems Mitigation Options Considered 5-14
5.1.2.3 Technical and Economic Characteristics Summary 5-19
5.1.2.4 Sector-Level Trends/Considerations 5-20
5.1.2.5 References 5-23
5.1.3 Stationary and Mobile Combustion 5-25
5.1.3.1 Stationary and Mobile Combustion Projections Methodology 5-25
5.1.3.2 Stationary and Mobile Combustion Mitigation Methodology 5-26
5.1.3.3 References 5-26
5.1.4 Biomass Combustion 5-27
5.1.4.1 Biomass Combustion Projections Methodology 5-27
5.1.4.2 Biomass Combustion Mitigation Methodology 5-28
5.1.4.3 References 5-28
5.1.5 Other Energy 5-29
5.1.5.1 Other Energy Projections Methodology 5-29
5.1.5.2 Other Energy Mitigation Methodology 5-29
5.2 Industrial Processes 5-30
5.2.1 Nitric and Adipic Acid Production 5-30
5.2.1.1 Nitric and Adipic Acid Production Projections Methodology 5-30
5.2.1.2 Nitric and Adipic Acid Production Mitigation Options Considered 5-32
5.2.1.3 Model Facilities 5-36
5.2.1.4 Sector-Level Trends and Considerations 5-39
5.2.1.5 References 5-39
5.2.2 F-GHG Emissions from Semiconductor Manufacturing 5-42
5.2.2.1 Semiconductor Manufacturing Projections Methodology 5-42
5.2.2.2 Mitigation Options Considered for Semiconductor Manufacturing 5-45
-------
5.2.2.3 Model Facilities 5-47
5.2.2.4 Technical and Economic Characteristics Summary 5-47
5.2.2.5 Sector-Level Trends and Considerations 5-49
5.2.2.6 References 5-49
5.2.3 Photovoltaic Cell Manufacturing 5-52
5.2.3.1 Photovoltaic Cell Manufacturing Projections Methodology 5-52
5.2.3.2 Mitigation Options Considered for Photovaltaic Cell Manufacturing 5-55
5.2.3.3 Model Facilities 5-57
5.2.3.4 Technical and Economic Characteristics of Options 5-57
5.2.3.5 References 5-58
5.2.4 PFC Emissions from Flat Panel Display Manufacturing 5-60
5.2.4.1 Flat Panel Display Manufacturing Projections Methodology 5-60
5.2.4.2 Mitigation Options Considered for Flat Panel Display Manufacturing 5-62
5.2.4.3 Model Facilities 5-64
5.2.4.4 Technical and Economic Characteristics Summary 5-64
5.2.4.5 References 5-66
5.2.5 SFs Emissions from Electric Power Systems 5-67
5.2.5.1 Electric Power Systems Projections Methodology 5-67
5.2.5.2 Mitigation Options Considered for Electric Power Systems 5-70
5.2.5.3 Model Facilities 5-74
5.2.5.4 Technical and Economic Characteristics Summary 5-74
5.2.5.5 Sector-Level Trends and Considerations 5-79
5.2.5.6 References 5-80
5.2.6 Primary Aluminum Production 5-83
5.2.6.1 Primary Aluminum Production Emission Projections Methodology 5-83
5.2.6.2 Mitigation Options Considered for Primary Aluminum Production 5-86
5.2.6.3 Model Facilities 5-87
5.2.6.4 Technical and Economic Characteristics Summary 5-88
5.2.6.5 Sector-Level Trends/Considerations 5-89
5.2.6.6 References 5-90
5.2.7 Magnesium Production 5-92
5.2.7.1 Magnesium Production Emission Projections Methodology 5-92
5.2.7.2 Mitigation Options Considered for Magnesium Production 5-96
5.2.7.3 Model Facilities 5-98
5.2.7.4 Technical and Economic Characteristics Summary 5-99
5.2.7.5 References 5-99
5.2.8 Use of Substitutes for Ozone-Depleting Substances 5-102
5.2.8.1 Use of Substitutes for Ozone-Depleting Substances Projections Methodology...5-102
5.2.8.2 Mitigation Options Considered 5-109
5.2.9 HFC-23 Emissions from HCFC-22 Production 5-156
5.2.9.1 HFC-23 from HCFC-22 Production Emission Projections Methodology 5-156
5.2.9.2 Mitigation Options Considered for HCFC-22 Production 5-160
iii
-------
5.2.9.3 Model Facilities 5-161
5.2.9.4 Technical and Economic Characteristics Summary 5-161
5.2.9.5 Sector-Level Trends/Considerations 5-161
5.2.9.6 References 5-162
5.2.10 Other Industrial Processes 5-163
5.2.10.1 Other Industrial Processes Projections Methodology 5-163
5.2.10.2 Other Industrial Processes Mitigation Methodology 5-163
5.3 Agricultural Sector 5-164
5.3.1 Livestock Management 5-164
5.3.1.1 Enteric Fermentation Projections Methodology 5-164
5.3.1.2 Mitigation Options Considered for Enteric Fermentation 5-166
5.3.1.3 Technical and Economic Characteristics of Options 5-167
5.3.1.4 Manure Projections Methodology 5-168
5.3.1.5 Mitigation Options Considered for Manure 5-171
5.3.1.6 Technical and Economic Characteristics of Options 5-171
5.3.1.7 References 5-175
5.3.2 Croplands 5-177
5.3.2.1 Projections Methodology 5-177
5.3.2.2 Mitigation Options Considered for Croplands 5-185
5.3.2.3 Technical Characteristics of Options 5-190
5.3.2.4 Economic Characteristics of Options 5-192
5.3.2.5 References 5-193
5.3.3 Rice Cultivation 5-196
5.3.3.1 Projections Methodology 5-196
5.3.3.2 Mitigation Options Considered for Rice Cultivation 5-199
5.3.3.3 Technical Characteristics of Options 5-202
5.3.3.4 Economic Characteristics of Options 5-203
5.3.3.5 References 5-204
5.3.4 Other Agriculture 5-206
5.4 Waste Sector 5-207
5.4.1 Landfills 5-207
5.4.1.1 Landfill Projections Methodology 5-207
5.4.1.2 Landfill Mitigation Options Considered 5-212
5.4.1.3 Mitigation Options Considered 5-213
5.4.1.4 Technical and Economic Characteristics Summary 5-219
5.4.1.5 Sector-Level Trends and Considerations 5-220
5.4.1.6 References 5-224
5.4.2 Wastewater Management 5-226
5.4.2.1 Wastewater Projections Methodology 5-226
5.4.2.2 Wastewater Mitigation Options Considered 5-229
5.4.2.3 Sector-Level Trends and Considerations 5-233
5.4.2.4 References 5-236
iv
-------
5.4.3 Other Waste
5.4.3.1 Other Energy Projections Methodology
5.4.3.2 Other Energy Mitigation Methodology .
5-237
5-237
5-237
Appendices
A: Refrigeration and Air Conditioning (RefAC) Disaggregation
B: Availability of Country-Reported Emissions Data
C: Data Sources Used to Develop Non-Country-Reported Emissions Estimates
D: Future Mitigation Measures Included in Developing Non-Country-Reported Estimates
E: U.S. EPAVintaging Model Framework
F: Country List
G: Mapping to IPCC Source Categories
V
-------
Figures
Number Page
3-1: Example of Combining Country-Reported and Tier 1 Calculations 3-5
4-1: Illustrative Non-CCh MAC Curve 4-7
4-2: Different Segments of Adopters Influence the Pace of Adoption 4-11
4-3: Illustrative Experience and Learning Curves 4-12
4-4: Cost Factor by Year Based on Technological Maturity and Economic Development 4-14
4-5: Illustrative Example of the Impact of Modeling Foreign-to-Domestic Labor Shift and Economic
Growth on Relative Cost Factors for Abatement Technologies 4-16
4-6: Illustration of Technology Change's Impact on a Non-CC>2 MAC Curve 4-17
5-1: Segments of Oil and Natural Gas Systems 5-11
5-2: Operational Adipic Acid Production Facilities in 2010 by Share of Global Capacity 5-37
5-3: Global SF6 Emissions in 2020 by Facility Type (% of GWP-Weighted Emissions) 5-98
5-4: Waste Generation vs GDP (Per Capita, 2015) 5-211
5-5: Conceptual Model for Estimating Mitigation Potential in the MSW Landfill Sector 5-221
5-6: Five Existing Scenarios Evaluated for Given Wastewater Discharge Pathways Based on
Technology Level, Treatment Alternative, and Collection Method 5-229
5-7: Domestic Wastewater MAC Analysis Flow Chart 5-234
vi
-------
Tables
Number Page
2-1: Sources Included in This Report 2-1
2-2: GWPs from the IPCC Fourth Assessment Report 2-2
3-1: UNFCCC Source Category Mappings to the EPA Source Categories (Example) 3-4
4-1: Calculation of Emission Reductions for an Abatement Option 4-3
4-2: Financial Assumptions in Break-Even Price Calculations for Abatement Options 4-5
4-3: International Economic Adjustment Factors for Selected Countries 4-6
4-4: Combined Effect of Technological Maturity and Economic Development Stage 4-13
4-5: Cost Factor and Emission Factor Results 4-13
4-6: Data Sources for Projecting Economic Growth at the Country Level 4-15
4-7: Regionalized Price Adjustments Factors, Descriptive Statistics 4-19
5-1: Summary of Abatement Measures for Coal Mines 5-4
5-2: Factors Used to Estimate Abatement Potential in Coal Mines 5-6
5-3: Emission Sources from Oil and Natural Gas Systems 5-11
5-4: IPCC Emission Factors by Disaggregation 5-13
5-5: Aggregate IPCC Emission Factor by Driver 5-13
5-6: Abatement Measures Applied in Oil and Gas Production Segments 5-15
5-7: Example Break-Even Price Calculation based on 2010 MAC for the United States 5-21
5-8: International Statistics on Key Activity Drivers: 2010 5-22
5-9: Allocation of Baseline Emissions to the Five Segments of the ONG System 5-22
5-10: Abatement Measures for Nitric and Adipic Acid Production 5-33
5-11: Adipic Acid-Producing Countries' Share of Baseline Emissions3 5-38
5-12: Model Nitric Acid Facilities Assumptions 5-39
5-13: Emission Factor Used for Semiconductor Manufacturing 5-43
5-14: Semiconductor Manufacturing Abatement Options 5-45
5-15: Technical Effectiveness Summary for New Fabs (Constant Over Time) 5-47
5-16: Technical Effectiveness Summary for Old Fabs (in 2020) 5-48
5-17: Engineering Cost Data on a Facility Basis—Semiconductor Manufacturing 5-48
5-18: PV Cell Manufacturing Abatement Options 5-57
5-19: Technical Effectiveness Summary—PV Cell Manufacturing 5-58
5-20: Engineering Cost Data on a Facility Basis—PV Cell Manufacturing 5-58
5-21: FPD Manufacturing Abatement Options 5-65
5-22: Technical Effectiveness Summary—Flat Panel Display Manufacturing 5-65
5-23: Engineering Cost Data on a Facility Basis—Flat Panel Display Manufacturing 5-66
5-24: Technical Effectiveness Summary— Electric Power Systems 5-75
5-25: Engineering Cost Data on a Facility Basis 5-75
vii
-------
5-26: Engineering Cost Inputs 5-76
5-27: Baseline Emission Factors by Year (MtCChe/Mt Al) 5-85
5-28: Description of Primary Aluminum Production Facilities 5-88
5-29: Technical Effectiveness Summary—Primary Aluminum Production 5-89
5-30: Engineering Cost Data on a Facility Basis—Primary Aluminum Production 5-89
5-31: Annual Growth Rates for Primary Production, Casting, and Recycling Production (annual
percentage increase) 5-94
5-32: Emission Factors for Primary Production, Casting, and Recycling Production (1990-1999) 5-95
5-33: Emission Factors for Primary Production, Casting, and Recycling Production (2000-2050) 5-95
5-34: Magnesium Production Abatement Options 5-99
5-35: Engineering Cost Data on a Facility Basis—Magnesium Production 5-99
5-36: Adjustment Factors Applied in by Sector and Country 5-105
5-37: Timing Factors Used for Developing (Article 5) Countries 5-106
5-38: GDP Growth Factors (relative to U.S.) 5-106
5-39: Recycling Adjustment Factors Applied to Refrigeration Emission Estimates 5-108
5-40: Refrigeration and AC Abatement Options 5-109
5-41: Summary of Technical Characteristics of Each Mitigation Option 5-120
5-42: Summary of Economic Characteristics of Each Mitigation Option 5-123
5-43: Solvent Use Abatement Options 5-130
5-44: Technical Effectiveness Summary—Solvent Use 5-133
5-45: Engineering Cost Data on a Facility Basis—Solvent Use 5-134
5-46: Foams Manufacturing Abatement Options 5-136
5-47: Emission Profiles for Foam End Uses 5-140
5-48: Technical Effectiveness Summary—Foams Manufacturing 5-141
5-49: Engineering Cost Data on a Facility Basis—Foams Manufacturing 5-141
5-50: Aerosol Product Use Abatement Options 5-144
5-51: Technical Effectiveness Summary—Aerosol Products 5-147
5-52: Engineering Cost Data on a Facility Basis—Aerosol Products 5-148
5-53: Fire Protection Abatement Options 5-150
5-54: Technical Effectiveness Summary—Fire Protection 5-153
5-55: Engineering Cost Data on a Facility Basis—Fire Protection 5-154
5-56: Portion of Total HCFC-22 Production That is Feedstock HCFC-22 5-157
5-57: Portion of Total HCFC-22 Production That is Feedstock HCFC-22 Production for Non-Al
Countries 5-157
5-58: Technical Effectiveness Summary—HCFC-22 Production 5-161
5-59: Abatement Measures for Enteric Fermentation CH4 5-167
5-60: Abatement Measures for Manure Management 5-172
5-61: Description of the Input Data Used in DAYCENT Simulations 5-187
5-62: DAYCENT Base Mean Yields and Differences from Mean Yield for Mitigation Strategies, by Year
(Metric tons of Grain per Hectare) 5-190
viii
-------
5-63: Alternative Rice Management Scenarios Simulated Using DNDC 5-200
5-64: Rice Management Techniques 5-202
5-65: Model Facility Assumptions for International LFG Mitigation Options 5-212
5-66: Cm Generation Factors by Country 5-213
5-67: Electricity Generation Technology Costs 5-215
5-68: Summary of the Engineering and Cost Assumptions for Mitigation Options at Landfills 5-219
5-69: Model Facilities Share of BAU Emissions: 2010-2030 5-224
5-70: Mitigation Options for the Wastewater Sector 5-230
ix
-------
-------
1 Background and Introduction
This peer-reviewed document outlines methodologies developed and used by the U.S. Environmental Protection
Agency (EPA) to calculate global non-carbon dioxide (non-CCh) greenhouse gas (GHG) projections and mitigation
potential. The methodological approach for emission projections was used to estimate historical non-CCh GHG
emission estimates for individual countries and sectors from 1990 to 2015, and emission projections from 2020
through 2050, at 5-year intervals. The methodological approach for calculating non-CCh GHG mitigation potential
builds off the emission projections and was used to generate marginal abatement cost (MAC) curves, where each
point on the cost curve reflects the average price and reduction potential for mitigation technologies. This
document updates and combines two prior report series, Non-COs Greenhouse Gases: International Emissions and
Projections [EPA, 2006, 2012] and Global Mitigation of
Non-CC>2 Greenhouse Gases (EPA, 2005, 2013).
Combining the "Projections" and "Mitigation"
reports provides an opportunity to better align these
two documents and their respective uses. The period of
analysis has been extended from 2030 through 2050.
The EPA also intends to make the combined results
available electronically to decrease the time required
between updates and to facilitate use of the latest
estimates.
The "Projections" report included historical data;
the latest country-reported emission estimates; and a methodology for estimating non-CCh emission projections
using a combination of country-reported data and calculations based on the Intergovernmental Panel on Climate
Change (IPCC) inventory. The projections methodology appears in Section 3. This methodology update introduces
greater disaggregation of major sources, based on magnitude or trends in emissions, providing higher resolution
on subsource contributions to total emissions. The emission projections are a composite of country-reported
historical emission estimates and trends based on projected activity data. The results are a "business-as-usual"
(BAU) scenario with fixed emission factors and reflect the impact of GHG emission reduction policies and measures
to the extent those are reflected in country-reported historical emissions. The projections include the impact of
existing GHG reduction policies to the extent they are reflected in historical data but exclude additional GHG
reductions because of current and additional planned activities and economy-wide programs that would reduce
emission rates from historical levels.
The "Mitigation" report estimated technical and economic potential mitigation of non-CCh GHG emissions for
20 sectors and 195 countries. The mitigation methodology is presented in Section 4. The mitigation methodology
includes several new developments, including incorporating the impact of technology change on abatement costs
and emission reductions globally and developing a regionalized set of MAC curves that provide increased
heterogeneity in abatement costs and emission reductions for the United States by state. Additionally, this
methodology incorporates new data on mitigation technologies, costs, and country-reported emission baselines.
Finally, a major focus of mitigation methodology development has been greater harmonization with the
projections; specifically, we attempted to use the projections as the "baseline" values to which mitigation technical
and economic potentials are compared when possible and investigate ways in which the estimates may be made
more compatible.
The results of the analysis will be presented in a forthcoming accompanying report as emission projections
and MAC curves for 195 individual countries as well as overall trends by region, gas, and source category for the
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
1-1
-------
METHODOLOGY DOCUMENTATION
years 2020 through 2050. Emission abatement will be shown in both absolute and percentage reductions from the
baseline.
The remainder of this methodology document provides additional detail on the projections and MAC modeling
approaches. Section 2 provides a brief overview of the non-CC>2 GHGs and their global warming potentials (GWPs).
Section 3 provides a detailed description of the methodology used to prepare emission trajectories by source for
1990 through 2050, including a discussion of the sources of uncertainties associated with developing these
emission estimates. Section 4 describes the general methodology for the MAC analysis. Sections 4.1 describes the
anatomy of the MAC curve and the limitations and uncertainties associated with the MAC calculations. Section 4.2
provides a more detailed description of the major methodological enhancements, accounting for technical change
and developing a more regionalized U.S. MACs. Section 5 includes additional methodological detail for the 27
individual non-CC>2 emitting sources analyzed within the energy, industrial, agricultural, and waste sectors. The
technical and economic characteristics of each mitigation option, specific modeling considerations, and projected
baseline emissions are also described by sector and region.
1-2
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
2 Non-C02 GHGs
The gases included in this report are the direct non-CC>2 GHGs covered by the United Nations Framework
Convention on Climate Change (UNFCCC): methane (CH4), nitrous oxide (N2O), and fluorinated greenhouse gases
(F-GHGs) that include hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride (SFs), and nitrogen
trifluoride (NF3). These non-CC>2 GHGs are more potent than CO2 (per unit weight) at trapping heat within the
atmosphere. Additionally, some non-CC>2 GHGs can remain in the atmosphere for longer periods of time than CO2.
This report does not include compounds covered by the Montreal Protocol. Table 2-1 lists the source categories for
the non-CC>2 GHGs considered for this analysis.
Table 2-1: Sources Included in This Report
Sector/Source
Gas
Energy
Natural Gas and Oil Systems
CH4
Coal Mining Activities
ch4
Stationary and Mobile Combustion
ch4, n2o
Biomass Combustion
ch4, n2o
Other Energy
ch4, n2o
Industrial Processes
AdipicAcid and Nitric Acid Production
n20
Use of Substitutes for Ozone-Depleting Substances3
HFCs
HCFC-22 Production
HFCs
Electric Power Systems
SFs
Primary Aluminum Production
PFCs
Magnesium Manufacturing
SFs
Electronics Manufacturing15
HFCs, PFCs, SFs, NFs
Other Industrial Processes
ch4, n2o
Agriculture
Agricultural Soils
N2O
Enteric Fermentation
CH4
Rice Cultivation
ch4
Manure Management
ch4, n2o
Other Agriculture
ch4, n2o
Waste
Landfilling of Solid Waste
CH4
Wastewater
ch4, n2o
Other Waste
ch4, n2o
a Substitutes for ODSs include uses in refrigeration and air conditioning (AC), solvents, foams, aerosols, and fire extinguishers.
b Electronics manufacturing includes semiconductors, photovoltaic, and flat panel displays.
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
2-1
-------
METHODOLOGY DOCUMENTATION
2.1 GWPs Assigned to Non-CC>2 GHGs
The GWP compares the relative ability of each
GHG to trap heat in the atmosphere during a certain
time frame. Per IPCC (2007) guidelines, CO2 is the
reference gas and thus has a GWP of 1. Based on a
time frame of 100 years, the GWP of CH4 is 25 and
the GWP of l\l20 is 298. Table 2-2 lists all GWPs used
in this report to convert the non-CC>2 emissions into
CC>2-equivalent units. This report uses GWPs from
the 2007 IPCC Fourth Assessment Report.
2.2 Methane (CH4)
Cm is approximately 25 times more powerful at
warming the atmosphere than CO2 over a 100-year
period. Additionally, the chemical lifetime for CH4 in
the atmosphere is approximately 12 years,
compared with approximately 100 years for CO2.
These two factors make CH4 a candidate for
mitigating global warming in the near term (i.e.,
within the next 25 years or so) or in the time frame
during which atmospheric concentrations of CH4
could respond to mitigation actions.
Man-made sources of CH4 include coal mining
natural gas and oil systems, stationary and mobile
combustion, certain industrial processes, agricultural
activities, landfills, and wastewater treatment. CH4 is
also a primary component of natural gas and an
important energy source. As such, both the
prevention and capture of CH4 emissions can
provide significant energy, economic, and
environmental benefits.
2.3 Nitrous Oxide (N2O)
Table 2-2: GWPs from the IPCC Fourth Assessment
Report
Greenhouse Gas
GWPa
CO2
1
ch4
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
NFs
17,200
cf4
7,390
C2F6
12,200
C4F10
8,860
C6F14
9,300
SFs
22,800
Source: Intergovernmental Panel on Climate Change (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, United
Kingdom: Cambridge University Press.
a100-year time horizon.
A clear, colorless gas with a slightly sweet odor,
N2O is an important GHG because of its long atmospheric lifetime (approximately 114 years) and heat-trapping
effects that are about 298 times more powerful than CO2.
N2O comes from natural and man-made sources and is removed from the atmosphere mainly by photolysis
(i.e., breakdown by sunlight) in the stratosphere. In the United States, the main man-made sources of N2O are
agricultural soil management, livestock waste management, mobile and stationary fossil fuel combustion, adipic
acid production, and nitric acid production. N2O is also produced naturally from a variety of biological sources in
soil and water, although this report covers anthropogenic sources only.
2-2
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 2 — NON-CO2 GHGS
2.4 Fluorinated Gases (F-GHGs)
There are three major categories or types of F-GHGs: HFCs, PFCs, SF6 and NF3. These compounds are the most
potent GHGs because of their large heat-trapping capacity and, in the cases of the PFCs and SF6, their extremely
long atmospheric lifetimes. Because some of these gases, once emitted, can remain in the atmosphere for
centuries, their accumulation is essentially irreversible. F-GHGs are emitted from a broad range of industrial
sources; most of these gases have few (if any) natural sources.
2.4.1 Hydrofluorocarbons (HFCs)
HFCs are man-made chemicals, many of which have been developed as alternatives to ozone-depleting
substances (ODSs) for industrial, commercial, and consumer products. The GWPs of HFCs range from 124
(HFC-152a) to 14,800 (HFC-23). The atmospheric lifetime for HFCs varies from just over a year (HFC-152a) to
270 years (HFC-23). Most of the commercially used HFCs have atmospheric lifetimes of less than 15 years (for
example, HFC-134a, which is used in automobile air-conditioning and refrigeration, has an atmospheric lifetime of
14 years).
2.4.2 Perfluorocarbons (PFCs)
Primary aluminum production, semiconductor manufacturing and flat panel display manufacturing are the
largest known man-made sources of the PFCs tetrafluoromethane (CF4) and hexafluoroethane (C2F6). PFCs are also
relatively minor substitutes for ODSs. Over a 100-year period, CF4 and C2F6 are, respectively, 7,400 and 12,200
times more effective than CO2 at trapping heat in the atmosphere.
2.4.3 Sulphur Hexafluoride (SF6)
The GWP of SF6 is 22,800, making it the most potent GHG evaluated by IPCC. SF6 is a colorless, odorless,
nontoxic, nonflammable gas with excellent dielectric properties. It is used (1) for insulation and current
interruption in electric power transmission and distribution equipment; (2) to protect molten magnesium from
oxidation and potentially violent burning in the magnesium industry; (3) to create circuitry patterns and to clean
vapor deposition chambers during manufacture of semiconductors and flat panel displays; and (4) for a variety of
smaller uses, including uses as a tracer gas and as a filler for sound-insulated windows. Like the PFCs, SF6 is very
long-lived, so all man-made sources contribute directly to its accumulation in the atmosphere.
2.4.4 Nitrogen Trifluoride (NF3)
NF3 is used in the manufacturing of semiconductors, photovoltaic cells, and flat panel displays. Over a 100-
year period, NF3 is 17,200 times more effective than CO2 at trapping heat in the atmosphere. NF3 was added as a
GHG under the Kyoto Protocol for its second commitment period.
2.5 References
Intergovernmental Panel on Climate Change. 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,
UK: Cambridge University Press.
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
2-3
-------
-------
4 Pas'jlitie Emissions Calculation—General Methodology
The general approach for developing the estimates of historical and projected emissions used a combination of
publicly available emission estimates from nationally prepared GHG reports with the EPA-estimated emissions
consistent with the Revised 1996IPCC Guidelines for National
Greenhouse Gas Inventories (1996 IPCC Guidelines) (IPCC, 1997), the
IPCC Good Practice Guidance and Uncertainty Management in
National Greenhouse Gas Inventories (IPCC Good Practice Guidance)
(IPCC, 2000), and the 2006 IPCC Guidelines for National Greenhouse
Gas Inventories (2006 IPCC Guidelines) (IPCC, 2006). To project
emissions, the EPA used drivers based on globally available growth
rates or activity data specific to each source. The EPA's calculated emission estimates were prepared in a
consistent manner across all countries using IPCC default methodologies, international statistics for activity data,
and the IPCC Tier 1 default emission factors.1 Higher IPCC tiers (i.e., Tier 2 and Tier 3) would be more accurate but
require more specific data requirements, so data limitations prohibited EPA from using higher tier calculations
across all countries, sectors, and subsources. Mitigation options represented in the MAC curves reduce emissions
from the baseline projections described in this section.
The EPA prepared a complete set of non-CC>2 GHG emission estimates, regardless of available country-
reported estimates, to produce a complete global inventory. Depending on available information, the projected
emission estimates for each country and source are either (1) a composite of historical country-reported emissions
and calculated estimates or (2) calculated estimates based on IPCC default emission factors and globally available
activity data. In most cases, some country-reported data are available, so the composite approach was used. The
second approach was only used where no country-reported data are available for a source category. For estimates
based on the composite approach, the Tier 1-calculated emission estimates were used to determine trends
through the time series, but the emission factors (i.e., emissions per unit of activity data) derive primarily from
country-reported information.
The projections estimates used both Excel workbooks and R programs. Source-specific calculations, based on a
Tier 1 calculation methodology,2 were performed in individual analysis workbooks. These source-specific analyses
were subsequently inputted into a program using R programming language to develop a composite of emissions
for the time series of 1990 through 2050.3The EPA used the R programming language to streamline the projections
updating process, as well as to centralize the calculations necessary to combine Tier 1 estimates from the
individual analysis workbooks with country-reported estimates.
The remainder of this section is organized as follows:
• Section 3.1 presents an overview of the main data sources for reported emission estimates and
calculations
• Section 3.2 presents the general methodology for calculating emission estimates from activity data and
emission factors using methods based on Tier 1 methodologies from IPCC inventory guidelines.
1 IPCC inventory guidelines generally include up to three different "tier" emissions methodologies that provide different levels
of complexity and require different data or country-specific information. Tier 1 is the simplest of the three and often involves
default emissions factors that can be used where country-specific emissions factors are not available.
2 In cases where a Tier 1 methodology was not feasible, estimates were based on another established methodology consistent
across all countries using IPCC default methodologies. However, for the purposes of this report, estimates calculated by the EPA
are referred to as "Tier 1" emissions calculations.
3 R is an open-source programming language and software environment.
Baseline Projections
In this report, the terms "business as
usual," "BAU," and "baseline" all refer to
the non-CC>2 emission projection results
and are used interchangeably.
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
3-1
-------
METHODOLOGY DOCUMENTATION
• Section 3.3 discusses the steps taken to develop the final composite emission estimates from the
combination of country-reported and calculated estimates.
3.1 Data Sources
For Annex 1 (Al) countries, the primary source of data for historical emission estimates was data from
National Inventories Common Reporting Format (CRF) files reported to the UNFCCC (UNFCCC, 2018a). The CRF
data contain reported national inventory data from 1990 through 2016. For Non-Annex 1 (Non-Al) countries, the
primary source of data for historical emission estimates was data from the UNFCCC flexible query system (UNFCCC,
2016).4 The UNFCCC flexible query system contains historical Cm, N2O, and F-GHG emission estimates reported to
the UNFCCC. Data for Non-Al countries obtained through the UNFCCC flexible query system contained data
reported through country national communication and biennial update reports. In some cases, more recent data
reported through country national communication and biennial update reports were not available through the
UNFCCC flexible query system and were obtained directly from the reports (UNFCCC, 2018b) or from the UNFCCC
Greenhouse Gas Data Interface (UNFCCC, 2018c).
Activity data were obtained from publicly available sources with global activity data, such as the U.S. Energy
Information Administration (EIA) and the U.N. Food and Agriculture Organization (FAO). Emission factors were
obtained from the 2006 IPCC Guidelines (IPCC, 2006).
Country-reported emission projections are also available from the UNFCCC from Biennial Reports. However,
these projections are only available aggregated at either the gas group or sector level for the years 2020 and 2030,
for Al countries. For this reason, country-reported projections could not be used for source- and gas-specific
projections.
Additional data to inform emission reduction activities, activity data, and emission factor calculations were
obtained from a literature review of academic papers and other databases, which are cited in the source-specific
methodology discussions.
3.1.1 Comparative Analysis
The EPA also obtained data from the European Commission's Emissions Database for Global Atmospheric
Research (EDGAR) (EC-JRC, 2016), and the International Institute for Applied Systems Analysis- Greenhouse Gas-
Air Pollution Interactions and Synergies (NASA-GAINS) database (NASA, 2016) to perform a series of comparative
analyses. The EPA used NASA-GAINS and EDGAR data to compare to the projections generated to observe and
analyze similarities and differences in the projections.
Additionally, a comparative analysis was performed between the composite results and the previous EPA
published report (EPA, 2012) on the source category- and sector-levels in order to assess the consistency of results
with previous projections. Composite results were also compared to other data sources, such as EDGAR and NASA,
to confirm any observed trends and examine sector-level results.
3.2 Calculating Historical and Projected Non-CQ2 Emissions
EPA first calculated Tier 1 emission estimates using IPCC Tier 1 methodology, IPCC Tier 1 emission factors, and
projected activity data through 2050 for all sources and countries. All sources have calculated emissions that rely
4 As identified by the UNFCCC, Al countries include all Organisation for Economic Co-operation and Development (OECD)
countries in 1992, plus countries with economies in transition and several Central and Eastern European States. Non-Al
countries include the rest of the world, consisting of developing countries in South America, Africa, and Asia. For more
information on the distinction between Al and Non-Al countries, see:
http://unfccc.int/Darties and observers/items/2704.php J.
3-2
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 3 — BASELINE EMISSIONS CALCULATION—GENERAL METHODOLOGY
on emission factors that are fixed throughout the time series. Key parameters and assumptions for the
independent set of calculated estimates include:
• For the purposes of calculating projections, the base year for this analysis is the last historical year of
emissions reported by a country. For details on the last historical year of reported estimates by country,
please see Appendix B.
• If a country reported no emissions in a given source in the base year all future emissions were assumed to
be zero.
• GWPs from the IPCC's Fourth Assessment Report (IPCC, 2007) were used to convert non-CC>2 emissions to
CCh-equivalent emissions (COie).5
• The country list included in the analysis is the same as that of the last publication, using the State
Department as a reference, except for two new countries: Kosovo and South Sudan.
• A subset of source categories was considered priority because of their magnitude or projected trend over
the time series: natural gas and oil systems, coal mining activities, adipic acid and nitric acid production,
landfilling of solid waste, substitutes for ODSs, and wastewater.
• Projection results were disaggregated for the following source categories using Tier 1 estimate
proportions:
- Nitric and adipic acid production were disaggregated into nitric acid and adipic acid production.
- Wastewater was disaggregated into urban and rural sources.
- Coal mining activities were disaggregated into above- and under-ground estimates.
- Natural gas and oil systems were disaggregated into gas production; gas transmission, storage, and
distribution; oil production; and oil refining.
- Substitutes for ODSs were disaggregated into refrigeration/air-conditioning, foams, solvents,
aerosols, and fire extinguishers.
- Biomass combustion was disaggregated into biofuels and wood fuel/charcoal.
Section 3.3 provides a detailed description of the methodology used to prepare emission trajectories that
combine the calculations described in this section with country-reported emission estimates by source, including a
discussion on the sources of uncertainties associated with developing these emission estimates.
3.3 Generating the Composite Historical and Projected Non-CQ2 Emissions
To use information from both country-reported and Tier 1-calculated emission estimates, the EPA used a
composite emission methodology, identical to that used in the previous report (EPA, 2012), to combine these two
datasets in a consistent manner. This methodology, described in detail below, creates a complete time series that
is consistent with country-reported data, when available, and fills in gaps using Tier 1-calculated emission
estimates.
5 The IPCC developed the Global Warming Potential (GWP) concept to compare the ability of each greenhouse gas to trap heat
in the atmosphere relative to a reference gas, C02. The IPCC updated GWPs in the IPCC Fourth Assessment Report (AR4) (IPCC
2007) which resulted in significantly different values for some of the gases compared to previously used values in the IPCC
Second Assessment Report (SAR) (IPCC, 1996). For example, the GWP of methane (CH4) has increased from 21 in the SAR to 25
in the AR4, whereas the GWP of N20 has decreased from 310 in the SAR to 298 in the AR4. As such, some of the historical C02e
values in this report may look significantly different than the values in previous versions of the report due to the changes in
GWPs of individual gasses. For a complete list of GWP values see the IPCC Fourth Assessment Report (AR4) (IPCC, 2007). AR4
GWP values were chosen to remain consistent with the GWP-weighted country reported emissions used to develop
projections.
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
3-3
-------
METHODOLOGY DOCUMENTATION
After completing the individual dataset of Tier 1 IPCC-calculated estimates, as described in Section 3.2, EPA
then reviewed for and pulled in any country-reported historical estimates.
• A1 Countries: A full time series of data,
disaggregated at the source category and
subsource category levels, was available and used
as the estimates for A1 countries from CRF files
from 1990 through 2016.
• Non-Al Countries: Historical emission data from
Non-Al countries were available in the UNFCCC
flexible query system, UNFCC detailed data by
party system, country national communication
reports, and country biennial reports, but
generally these reported data did not constitute a
full time series and are sometimes available at the
source, rather than subsource, level (e.g., A1
countries report emissions from nitric and adipic
acid production separately, while Non-Al
countries report aggregated emissions). To
calculate emission estimates for 1990,1995, 2000,
2005, 2010, and 2015, gaps in the reported
estimates were filled by linear interpolation (when
missing data occurred between 2 years with
reported data) or backcasting/projecting (by
applying growth rates from each source category's
activity data, when missing data occur outside of years with reported data).
UNFCCC country-reported data were mapped to corresponding Tier 1 calculations using UNFCCC source and
gas combinations. Because the UNFCCC estimates were available at different levels of disaggregation for A1 and
Non-Al countries, these data were mapped to the corresponding Tier 1 calculations at the source level (see
Table 3-1).
Table 3-1: UNFCCC Source Category Mappings to the EPA Source Categories (Example)
Composite Approach Key Drivers
• Projected Trends. The composite approach
relies on EPA Tier 1-calculated estimates to
determine increases and/or decreases to
emissions through the time series. This is
because country-reported emission
projections are not generally available
disaggregated by gas and source; additionally,
activity data projections are limited in that
emission source categories often combine
multiple activities with different
characteristics.
• Emission Factors. The composite approach
does not have a specific set of emission
factors that corresponds to any particular
estimates. Because the starting point is
country-reported data, emission factors
associated with the composite projections
derive primarily from the country-reported
estimates. See Section 3.3.1 for additional
discussion.
UNFCCC Source Category
EPA's Source Category
l.A Fuel Combustion
Stationary and Mobile Combustion
1.B.2 Oil and Natural Gas
Oil and Natural Gas Systems
Upon completion of the mapping of the reported dataset to the calculation dataset for a given source, a set of
operations was performed in the R code to (1) linearly interpolate any data gaps for intermediate years; (2) back
cast years in which reporting began after 1990, using the growth rates of the Tier 1 calculations to estimate for
missing historical years; and (3) project emissions beyond the years reported using the growth rates of the Tier 1
calculations.
Countries without historical reported emissions were pulled in "as is" from the individual analysis files. This
prioritization process is summarized in the following steps:
1. If country-reported historical data exist, they were used to populate the historical years, 1990 through
2016, which is the latest possible year of available country-reported estimates at the time of the analysis.
3-4
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 3 — BASELINE EMISSIONS CALCULATION—GENERAL METHODOLOGY
2. For those countries, the EPA projected emissions from the most recent country-reported data through
2050 using growth rates calculated by the Tier 1 methodology. For example, if the Tier 1 calculations
indicate a 2015 through 2020 growth rate of 5%, followed by a 2020 through 2050 growth rate of 10%,
the composite calculation increases that country's reported 2015 emissions by 5% to the year 2020, and
then by 10% through 2050.
3. For countries and sources that do not have country-reported historical data, the EPA used the emission
estimates calculated by the Tier 1 methodology as the full time series from 1990 through 2050.
Figure 3-1 provides a visual example of how country-reported and Tier 1 calculations were combined.
Figure 3-1: Example of Combining Country-Reported and Tier 1 Calculations
U.K. Stationary and Mobile CH4 Emissions
3.6
3.2
2.8
2.4
2.0
1.6
1.2
t" t — t — t— t
0.8
0.4
0.0
1990 1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050
Year
Country Reported Emissions — Tier 1 Calculations * Composite Emissions
This process was performed consistently for all countries, sources, and gases. For sources that included
subsources, the composite results were disaggregated based on the proportion of Tier 1-calculated emissions from
each subsource. The composite results for individual sources were extracted as outputs from the R code, indicating
subsource when available.
The "Other" categories in each sector are included for completeness and solely comprise countries that report
data to the UNFCCC database. The EPA did not perform Tier 1 calculations for "Other" source categories. The EPA
obtained historical country-reported estimates for the "Other" categories for 1990 through 2012 and held 2015
through 2050 emission estimates constant at 2012 levels for each country.
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
3-5
-------
METHODOLOGY DOCUMENTATION
3.3.1 Emission Factors in the Composite Projections
Because most of the emission projections are based on a combination of country-reported and calculated
emission estimates, it is not straightforward to specify the emission factor that corresponds to any particular
projection estimate. The starting point for these projections is country-reported data, so the emission factor, or
emissions per unit of activity data differ between countries based on the differing calculations that countries have
incorporated into their emission inventories. The Tier 1 source-specific calculations are based on a BAU scenario,
which generally keep fixed emission factors over time. These calculations use IPCC default emission factors;
however, these are not necessarily representative of the emission factors for the resulting composite projections.
Additionally, in many cases, a source category aggregates emissions for several smaller subsources or categories.
For example, enteric fermentation estimates are driven by estimates for a variety of different animals. For the
composite projections, the IPCC default factors are more significant for disaggregation of subsources and as
weighting factors between the relative trends of different categories within a source category.
3.3.2 Reductions in the Baseline Scenario
Although the baseline or BAU scenario generally does not explicitly model emission reduction policies
undertaken by individual countries, and the default IPCC factors generally reflect uncontrolled emissions, the
composite emission projections do include historical emission reductions. To the extent that emission reductions
are reflected in country-reported base year data, those rates are used throughout the projection time series. Thus,
the degree to which reductions are included in a particular estimate corresponds to the extent to which reductions
are reflected in country-reported data.
3.3.3 Excluded Data
During the review of composite results, the EPA compared country-reported emission estimates to Tier 1-
calculated estimates to identify cases where data in the country-reported datasets may have been misreported or
transcribed incorrectly. As a result, a small number (<0.5%) of country-reported data points were excluded. Two
main metrics were used to identify potential errors: the ratio and difference between country-reported and Tier 1-
calculated emission estimations for the last year of reported data. The country-reported emission estimates that
were reviewed by EPA included those that met any one of the following criteria:
• a difference greater than 500 Gg with Tier 1-calculated emission estimates,
• the highest 2% of differences,
• a ratio greater than 20 of country-reported emission estimates to Tier 1-calculated emission estimates,
and
• the highest 2% of ratios.
The quantitative screening resulted in a list of several hundred data points that were reviewed by source
experts. Based on the EPA's review of the country-reported emission estimates that met the above criteria, 16
country-reported estimates were excluded from the composite calculations. The EPA believes these discrepancies
may have resulted from differences in units, calculations, or problems with data entry in the UNFCCC system and
may not represent countries' current estimates of emissions. Incorporating these data discrepancies would have
resulted in a significant change in projected emissions for some sources and countries in the composite results.
3.4 References
European Commission, Joint Research Centre/Netherlands Environmental Assessment Agency. 2016. Emission
Database for Global Atmospheric Research (EDGAR), release version 4.2. Available online at
http://edgar.irc.ec.europa.eu J
3-6
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 3 — BASELINE EMISSIONS CALCULATION—GENERAL METHODOLOGY
International Institute for Applied Systems Analysis. 2016. Greenhouse Gas-Air Pollution Interactions and
Synergies (GAINS) Model. Available online at http://www.iiasa.ac.at/web/home/research/research
Programs/air/GAINS.html ef.
International Panel on Climate Change. 1996. IPCC Second Assessment Report (SAR). Available online at
https://www.ipcc.ch/report/ipcc-second-assessment-full-report/c?
Intergovernmental Panel on Climate Change. 1997. Revised 1996 IPCC Guidelines for National Greenhouse Gas
Inventories. Paris: Intergovernmental Panel on Climate Change, United Nations Environment Programme,
Organisation for Economic Co-Operation and Development, International Energy Agency.
Intergovernmental Panel on Climate Change. 2000. Good Practice Guidance and Uncertainty Management in
National Greenhouse Gas Inventories. IPCC-XVI/Doc.10 (1.IV.2000). Paris: Intergovernmental Panel on Climate
Change, National Greenhouse Gas Inventories Programme, Montreal.
Intergovernmental Panel on Climate Change. 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.
Intergovernmental Panel on Climate Change. 2007. Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts,
D.W. Fahey, J. Haywood, J. Lean, D.C. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz, and R. Van
Dorland. 2007. Changes in Atmospheric Constituents and in Radiative Forcing. In: 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, UK: Cambridge University Press.
United Nations Framework Convention on Climate Change. 2016. United Nations Framework Convention on
Climate Change Flexible GHG Data Queries. Online Database. Available online at
http://unfccc.int/di/FlexibleQueries/Setup.do J
United Nations Framework Convention on Climate Change. 2018a. 2018 Annex I Party GHG Inventory Submissions
to the UNFCCC. CRF Data Files. Available online at https://unfccc.int/process-and-meetings/transparencv-and-
reporting/reportine-and-review-under-the-convention/greenhouse-eas-inventories-annex-i-parties/national-
inventorv-submissions-2018 riP
United Nations Framework Convention on Climate Change. 2018b. National Communication Submissions from
Non-Annex I Parties to the UNFCCC. Available online at https://unfccc.int/process-and-
meetings/transparencv-and-reporting/reporting-and-review-under-the-convention/national-communications-
and-biennial-update-reports-non-annex-i-parties/national-communication-submissions-from-non-annex-i-
parties Ef
United Nations Framework Convention on Climate Change. 2018c. UNFCCC Greenhouse Gas Data Interface. Online
Database. Available online at http://di.unfccc.int/detailed data by party E?
U.S. Environmental Protection Agency. 2012. Global Anthropogenic Non-CC>2 Greenhouse Gas Emissions: 1990-
2030. Available online at https://www.epa.gov/global-mitigation-non-C02-greenhouse-gases
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
3-7
-------
-------
4 MAC Analysis—General Methodology
The basic methodology is a bottom-up, engineering cost approach and is the same methodology used for the 2013
report (EPA, 2013a). MAC curves were constructed for each region and sector by estimating the carbon price at
which the present-value benefits and costs for each mitigation option equilibrate. The methodology produces a
stepwise curve, where each point reflects the average price and reduction potential if a mitigation technology
were applied across the sector within a given region. In conjunction with appropriate baseline and projected
emissions for a given sector, the results are expressed in terms of absolute reductions of carbon dioxide
equivalents (MtCChe).
The mitigation option analysis throughout this report was conducted using a common methodology and
framework. This section outlines the basic methodology. The sector-specific sections describe the mitigation
estimation methods in greater detail, including any necessary deviations from the basic methodology.
The MAC model continues to offer improvements made in the 2013 report, which included
• disaggregating mitigation potential and costs to the country level for 195 countries;
• updating reduction efficiencies for individual measures by country;
• updating capital and operation and maintenance (O&M) costs for individual measures;
• segmenting O&M costs into labor, materials, and energy components;
• developing international adjustment factors used to construct country-specific abatement costs and
benefits; and
• updating crop process model simulations of changes in crop yields and emissions associated with rice
cultivation and cropland soil management.
4,1 Methodology Overview
The abatement analysis for all non-CC>2 emissions from sources in the agriculture, energy, waste, and industrial
process sectors is based on EPA (2013a) and improves on Ragnauth et al. (2015), Gallaher et al. (2005), and Rose et
al. (2013). These studies provided estimates of potential Cm and N2O emission reductions from major emitting
sectors and quantified costs and benefits of these reductions. Mitigation analysis accounts for country differences
in industry structure and available infrastructure where data are available on a sector-by-sector basis. For example,
the analysis of the natural gas and oil sector relied on country-specific activity data on production,
processing/refining, and transportation infrastructure to distribute baseline emissions to specific subsectors.
Additionally, we accounted for country/regional differences in price of mitigation through a series of international
cost indices (labor, nonenergy materials, energy) to create a more heterogenous representation of emissions and
mitigation costs and benefits across countries. Thus, the EPA analysis provides significant detail at the sector and
subsector levels and across countries.
Given the detailed data available for U.S. sectors, our analysis of the United States used representative facility
estimates but then applied the estimates to a highly disaggregated and detailed set of emission sources for all the
major sectors and subsectors. For example, the EPA analysis of the natural gas sector was based on more than 100
emission sources in that industry, including gas well equipment, pipeline compressors and equipment, and system
upsets. Thus, the EPA analysis provides significant detail at the sector and subsector levels.
The analysis generally began with developing sector-level model facilities or units to which mitigation options
are applied. In many cases, the model facilities, abatement costs, and mitigation potential were based on detailed
U.S. and European Union (EU) inventory estimates and were then extrapolated to "model" facilities for other
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-1
-------
METHODOLOGY DOCUMENTATION
countries. For some sectors, such as wastewater, landfills, and selected industrial sectors, additional detail on
international abatement options and costs was available and was incorporated into the model.
A scaling factor was used to reconcile inventory data with the GER baseline emission data. For the F-GHGs
abatement analysis and oil and natural gas (ONG) and landfills sectors, the EPA assumed that some mitigation
technologies are adopted to meet future regulations or voluntary industry reduction targets. Therefore, the
baseline emissions account for some mitigation options. If an option was assumed to be adopted in the baseline, it
was not included when generating the MAC. In addition, expert judgment determined the market shares for
mitigation technologies competing for the same set of emissions (when multiple options were available that are
substitutes for each other).
The agricultural sector's emission abatement analysis improves on analyses in previous studies supported by
the EPA (2013a) that generated MAC curves by major world region for cropland N2O CFU and soil carbon, livestock
enteric CH4, manure management N2O and CFU, and rice cultivation N2O and CFU. Updates in this report include
the following:
• regionalized livestock estimates for the United States
• more explicit links to the baseline projection estimates
4.1.1 Technical Characteristics of Abatement Options
The non-CC>2 abatement options evaluated in this report were compiled from the studies mentioned above
and from the sector-specific literature. For each region, either the entire set of sector-specific options or the
subset of options determined to be applicable was applied. Options were omitted from individual regions on a
case-by-case basis, using either expert knowledge of the region or technical and physical factors (e.g., appropriate
climate conditions). In addition, the share or extent of applicability of an option within different regions may vary
based on these conditions.
The selective omission of options represents a static view of the region's socioeconomic conditions. In some
instances, the reduction efficiency of an option improves over time reflecting anticipated technology advances.
However, the applicability of options is held constant over time. Ideally, more detailed information on country-
specific conditions, technologies, and experiences will be available in the future, which will enable more rigorous
analyses of abatement option availability over time in each region. Furthermore, the average technical lifetime of
an option (in years), determined using expert knowledge of the technology or recent literature, is held constant
over time and across regions.
Table 4-1 summarizes how the potential emission reduction was calculated for each of the available
abatement options. First the technical effectiveness of each option was calculated by multiplying the option's
technical applicability by its market share by its reduction efficiency. This calculation yields the percentage of
baseline emissions that can be reduced at the national or regional level by a given option. The technical
effectiveness percentage was then multiplied by the applicable baseline emissions (MtCChe) to yield the emission
reductions for the mitigation option.
Technical Effectiveness (TE) = TA x MS x RE
(4.1)
Emissions Reductions (MtC02e) = TE x BE
(4.2)
4-2
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
Table 4-1: Calculation of Emission Reductions for an Abatement Option
Technical
Market
Reduction
Technical
Applicability
X Share3
X Efficiency
= Effectiveness
(%)
(%)
(%)
(%)
Technical
Baseline
Emissions
Effectiveness
X Emissions
= Reductions
(%)
(MtCChe)
(MtCChe)
Percentage
Percentage
Percentage
Percentage of
Emission
Unit
of total
of technically
of technically
baseline
stream to
emission
baseline
applicable
achievable
emissions that
which the
reductions.
emissions
baseline
emission
can be
option is
from a
emissions to
abatement
reduced at the
applied.
particular
which a given
for an option
national or
emission
option is
after it is
regional level
source to
applied;
applied to a
by a given
which a given
avoids
given
option.
option can be
double
emission
potentially
counting
stream.
applied.
among
competing
options.
a Implied market shares for noncompeting mitigation options (i.e., only one option is applicable for an emission streams) sums
to 100%.
where
TA = technical applicability (%)
MS = market share (%)
RE = reduction efficiency (%)
TE = technical efficiency (%)
BE = baseline emissions (MtCChe)
Technical applicability accounts for the portion of emissions from a facility or region that a mitigation option
could feasibly reduce based on its application. For example, if an option applies only to the underground portion of
emissions from coal mining, then the technical applicability for the option would be the percentage of emissions
from underground mining relative to total emissions from coal mining.
The implied market share of an option is a mathematical adjustment for other qualitative factors that may
influence the effectiveness or adoption of a mitigation option. We used market shares for each mitigation option
within every sector. The market shares, determined by various sector-specific methods, must sum to one for each
sector and were assumed constant over time.6 This assumption avoids cumulative reductions of greater than 100%
across options.
When nonoverlapping options are applied, they affect 100% of baseline emissions from the relevant source.
Examples of two nonoverlapping options in the natural gas system are inspection and maintenance of compressors
and replacement of distribution pipes. These options were applied independently to different parts of the sector
and do not compete for the same emission stream.
6 For certain energy, waste, and agriculture sectors, accounting for adoption feasibility, such as social acceptance and
alternative permutations in the sequencing of adoption, was outside the scope of this analysis.
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-3
-------
METHODOLOGY DOCUMENTATION
The reduction efficiency of a mitigation option is the percentage reduction achieved with adoption. The
reduction efficiency was applied to the relevant baseline emissions as defined by technical applicability and
adoption effectiveness. Most abatement options, when adopted, reduce an emission stream less than 100%. If
multiple options are available for the same component, the total reduction for that component is less than 100%.
Once the technical effectiveness of an option was calculated as described above, this percentage was
multiplied by the baseline emissions for each sector and region to calculate the absolute amount of emissions
reduced by employing the option. The absolute amount of baseline emissions reduced by an option in a given year
is expressed in million metric tons of CO2 equivalent (MtCCheq).7
If the options were assumed to be technically feasible in a given region, they were assumed to be
implemented immediately. Furthermore, once options are adopted, they were assumed to remain in place for the
duration of the analysis, and an option's parameters do not change over its lifetime.
4.1.2 Economic Characteristics of Abatement Options
Each abatement option is characterized in terms of its costs and benefits per an abated unit of gas (tCCheq or
tons of emitted gas [e.g., tCFU]). The benefits include a carbon value/price expressed as $/tCC>2e. The carbon price
at which an option's benefits equal the costs is referred to as the option's break-even price.
For each mitigation option, the carbon price (P) at which that option becomes economically viable was
calculated using the equation below (i.e., where the present value of the benefits of the option equals the present
value of the costs of implementing the option). A present value analysis of each option was used to determine
break-even abatement costs in a given region. Break-even calculations are independent of the year the mitigation
option is implemented but are contingent on the life expectancy of the option. The net present value calculation
solves for break-even price P by equating the present value of the benefits with the present value of the costs of
the mitigation option. More specifically,
Lt=1l (1+DR)t J ~ LL + Zt=1 L J ^
Net Present Benefits Net Present Costs
where
P= the break-even price of the option ($/tCC>2e)
ER = the emission reduction achieved by the technology (MtCChe)
R = the revenue generated from energy production (scaled based on regional energy prices) or sales of by-
products of abatement (e.g., compost) or change in agricultural commodity prices ($)
T = the option lifetime (years)
DR = the selected discount rate (%)
CC= the one-time capital cost of the option ($)
RC = the recurring (O&M) cost of the option (portions of which may be scaled based on regional labor and
materials costs) ($/year)
77? = the tax rate (%)
TB = annual tax benefit of depreciation = (^7) ¦ TR
7 One MtC02eq equals 1 teragram of C02 equivalent (TgC02eq); 1 metric ton = 1,000 kg = 1.102 short tons = 2,205 lbs.
4-4
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
Assuming that the emission reduction ER, the recurring costs RC, and the revenue generated R do not change
on an annual basis, then we can rearrange this equation to solve for the break-even price P of the option for a
given year:
P= — + — —_ ££_ ¦ TR (44)
(1-TR)ERT.J-1 1 ER ER ERT (1-Tfi)
Costs include capital or one-time costs and O&M or recurring costs. Most of the agricultural sector options,
such as changes in management practices, do not have applicable capital costs, with the exception of anaerobic
digesters for manure management.
Benefits or revenues from employing an abatement option can include (1) the intrinsic value of the recovered
gas (e.g., the value of Cm either as natural gas or as electricity/heat, the value of HFC-134a as a refrigerant),
(2) non-GHG benefits of abatement options (e.g., compost or digestate for waste diversion options, increases in
crop yields), and (3) the value of abating the gas given a GHG price in terms of dollars per tCC>2 eq ($/tCC>2eq) or
dollars per metric ton of gas (e.g., $/tCH4, $/tHFC-134a). In most cases, the abatement of Cm has two price signals:
one price based on Cl-U's value as energy (because
natural gas is between 90% and 98% Cm) and one
price based on Cl-U's value as a GHG. All cost and
benefit values are expressed in constant-year
2015 U.S. dollars. This analysis was conducted
using a 5% discount rate and a 0% tax rate. For
quick reference, Table 4-2 lists the basic financial
assumptions used throughout this report.
4.1.3 International Adjustment Factors
Costs and benefits of abatement options were adjusted to reflect regional prices. Wages and prices vary by
country. Hence, recurring O&M costs were segmented into labor, energy, and materials costs. Material cost
components ranged from materials and supplies in the industrial and energy sectors to fertilizer costs in the
agricultural sectors—all of which are likely to vary by region. One-time capital costs were assumed to be relatively
stable across regions and were not adjusted from country to country.
For some options, data were available on the relative cost shares between labor, energy, and materials. For
instance, in coal mining, different technologies have different cost shares, which were developed based on expert
judgment. For options without detailed cost breakouts, the shares were generally assigned evenly as 33% each to
labor, energy, and materials. For the agricultural sector, labor, energy, water, and other input costs were
calculated from their shares of agricultural production costs based on social accounting matrix data from the
Global Trade Analysis Project (GTAP) v8 database and agricultural wage data from the International Food Policy
Research Institute (IFPRI).
In regions lacking detailed revenue (benefits) data, revenues were scaled based on the ratio between average
prices of natural gas or of electricity in a given region and in the United States. We used the price of natural gas to
value benefits when captured CH4 is sold as natural gas, and we used the price of electricity to value benefits when
the captured CH4 is used to generate electricity or process heat. The natural gas and electricity prices used come
from the Energy Information Agency's (ElA's) International Energy Statistics. Similarly, revenues from non-CH4
benefits of abatement options were scaled based on the ratio between the gross domestic products (GDPs) per
capita in a given region and in the United States. In the agricultural sector, changes in revenue occur as a change in
either crop yield or livestock productivity. Data on changes in crop yield or livestock productivity were combined
with data on regional producer prices for the relevant agricultural commodity to calculate revenue changes.
Table 4-2: Financial Assumptions in Break-Even Price
Calculations for Abatement Options
Economic Parameter
Assumption
Discount rate
5%
Tax rate
0%
Constant-year dollars
2017$
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-5
-------
METHODOLOGY DOCUMENTATION
Table 4-3 lists the international economic adjustment factors for selected countries. Using publicly available
data on country-specific wage rates and energy prices, along with input from previous MAC analyses, we created
indices reflecting each country's wage rates and prices relative to the United States. Adjustment factors were
created for labor, natural gas, electricity, coal, and materials costs. When data were not available for a country, the
country was either mapped to a similar country (with data) or previously developed Stanford Energy Modeling
Forum (EMF) factors were used.
Table 4-3: International Economic Adjustment Factors for Selected Countries
Country
Labor3
Natural Gasb
Electricity13
Coalb
Materialsc
Afghanistan
0.02
0.75
1.30
0.89
0.01
Brazil
0.24
1.30
1.60
0.76
0.13
Congo
0.19
1.06
0.34
0.37
0.05
China
0.04
0.62
0.63
0.68
0.07
India
0.03
0.67
1.69
0.69
0.02
Madagascar
0.19
1.06
0.34
0.37
0.01
Mexico
0.12
1.04
1.42
0.94
0.20
Norway
1.80
1.62
0.77
2.57
1.61
Poland
0.26
0.98
1.19
1.25
0.24
Russian Federation
0.12
0.19
0.56
0.67
0.15
Switzerland
1.35
1.62
1.41
2.04
1.30
United States
1.00
1.00
1.00
1.00
1.00
Uzbekistan
0.12
0.19
0.38
0.19
0.02
a Wage data were obtained primarily from U.S. Bureau of Labor Statistics' (BLS's) International Labor Comparisons (2010a) and
augmented with BLS (2010b, c) and Russian Federation Federal State Statistics Service (2010).
b Energy prices were obtained from the ElA's International Energy Statistics (2010b).
c Material factors were based on GDP/capita statistics obtained from the United Nations Conference on Trade and Development
Statistical Database (2012).
Note that break-even price calculations for this analysis do not include transaction or monitoring and
reporting costs, because this report contains no explicit assumptions about policies that would encourage and
facilitate adoption of the mitigation options. Refer to Section 4.2.3 for a more complete discussion of the
limitations of this analysis.
4.1.4 MAC Curves and Their Construction
MAC curves are used to show the amount of emission reduction potential at varying carbon price levels. In
theory, a MAC curve illustrates the cost of abating each additional ton of emissions. Figure 4-1 shows an illustrative
MAC curve. The x-axis shows the amount of emission abatement in MtCCheq, and the y-axis shows the break-even
price in $/tCC>2eq required to achieve the level of abatement. Therefore, moving along the curve from left to right,
the lowest cost abatement options are adopted first.
The curve becomes vertical at the point of maximum total abatement potential, which is the sum of all
technically feasible abatement options in a sector or region. At this point, no additional price signals from GHG
credit markets could motivate emission reductions; any additional emission reductions (shifting the vertical axis to
the right) are due to increased energy efficiencies, conservation of production materials, or both.
4-6
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
The points on the MAC curve that appear at or below the zero-cost line ($0/tCC>2eq) illustrate potentially
profitable mitigation options. These "below-the-line" amounts represent mitigation options that are already cost-
effective given the costs and benefits considered (and are sometimes referred to as "no-regret" options) yet have
not been implemented. However, nonmonetary barriers may prevent their adoption.
Figure 4-1: Illustrative Non-C02 MAC Curve
Value of C02
Equivalent
($/tC02eq)
Market Price
$0/tC02eq
Total Abatement Potential
Energy/Commodity
Prices
Abated GHG Emissions (MtC02eq)
The MAC curves in this report were constructed from bottom-up average break-even price calculations. The
average break-even price was calculated for the estimated abatement potential for each mitigation option (see
Section 3.3). The options were then ordered in ascending order of break-even price (cost) and plotted against
abatement potential. The resulting MAC curve is a stepwise function, rather than a smooth curve, as seen in the
illustrative MAC curve (Figure 4-1), because each point on the curve represents the break-even price point for a
discrete mitigation option (or defined bundle of mitigation strategies).
Conceptually, marginal costs are the incremental costs of an additional unit of abatement. However, the
abatement cost curves developed here reflect the incremental costs of adopting the next cost-effective mitigation
option. We estimated the costs and benefits associated with all or nothing adoption of each well-defined
mitigation practice. We did not estimate the marginal costs of incremental changes within each practice (e.g., the
net cost associated with an incremental change in paddy rice irrigation). Instead, the MAC curves developed in this
report reflect the average net cost of each option for the achieved reduction, hence the noncontinuous, stepwise
nature of the curve.
In the energy and waste sectors, representative facilities facing varied mitigation costs employ mitigation
technologies based on the lowest average break-even option price. In calculating the abatement potential, we
evaluated options according to whether they are complements or substitutes. If a group of options are
complements (or independent of one another), the implied market shares are all equal to one. If options are
substitutes for each other, then market shares that sum to one were used to distribute adoption across the
available options. In some instances, the lowest price option was selected for each representative facility. When
limited information was available, the market share was evenly distributed (1/n) across all viable options. In this
way, the implied adoption rate for each technology was estimated.
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-7
-------
METHODOLOGY DOCUMENTATION
In the industrial processes sector, mitigation options were applied to representative facilities in order of
lowest average break-even price to highest average break-even price. Each option was applied to a portion of the
baseline emissions based on the implied adoption rate (the market share factor, as described in Section 3.2),
which, in the industrial sector, was informed by expert insight into potential adoption rates of various mitigation
technologies.
In the agriculture sector, mitigation options were applied to the portion of emissions where they are
technically applicable (e.g., anaerobic digesters were assumed to be applicable only in intensively managed dairy
and hog production systems). The implied market share for competing options was based purely on the number of
available migration options (n) that are applicable to a given subset of emissions and that reduce emissions8 (1/n),
where each option was applied to an equal portion of the cropland base or livestock population and, thus regional
baseline emissions, for each region over time. Given the existence of nonprice and implementation factors that
influence market share and the lack of accurate and detailed information regarding these qualitative
characteristics, we assumed an even distribution of options across the relevant baseline for the agriculture sector.
This approach allows options to share a portion of market penetration, regardless of their cost-effectiveness,
rather than allowing only the least-cost option to completely dominate the market. Our methodology is more
conservative than if we had assumed only price factors exist, thus allowing the least-cost option to penetrate the
sector by 100%.
The MAC curves represent the average economic potential of mitigation technologies in that sector, because
we assumed that if a mitigation technology is technically feasible in a given region, then it is implemented
according to the relevant economic conditions. Therefore, the MAC curves do not represent the market potential
or the social acceptance of a technology. The models used in the analysis are static (i.e., they do not represent
adoption of mitigation technologies over time). This analysis assumed partial equilibrium conditions that do not
represent economic feedbacks from the input or output markets. This analysis made no assumptions regarding a
policy environment that might encourage the implementation of mitigation options. Section 4.2.3 provides
additional discussion of some key limitations of the methodology.
The end result of this analysis is a tabular dataset, available for download from the EPA website at
https://cfpub.epa.gov/nonco2/ for the MAC curves by sector, gas, and region.9
4.2 MAC Methodological Enhancements
This report builds on a study previously conducted by the EPA for Stanford's EMF-21 and the EPA (2006,
2013a) reports. As in previous MAC model updates, we have updated our economic cost for implementing
mitigation options when new data were available, we incorporated the most recent set of international GHG
emission projections and updated the baseline price forecasts and relative international adjustments factors for
labor, energy, and materials.
Additionally, in this update to the MAC model, we introduced two major methodological enhancements that
includes incorporating the effects of technology change on mitigation costs and their reduction efficiencies and
8 Some agricultural mitigation options may increase emissions under certain conditions depending on baseline regional
management and soil, climate, and other considerations. In addition, many mitigation options increase emissions per head of
livestock or per hectare of land but reduce emissions intensity per unit of output. Thus, agricultural MAC curves were calculated
both assuming constant production and constant area/head of livestock to present a range of potential mitigation options. The
options that provide net emissions reductions may differ between these alternative methods of MAC curve generation.
9 Tables are presented that provide the percentage abatement for a series of break-even prices. The MAC data are presented as
tables so that exact values can be determined for use in modeling activities.
4-8
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
developing more regionalized sectoral MACs for the United States. The remainder of this section describes each of
these enhancements in more detail.
4.2.1 Incorporating Technological Change
In this section, we describe the updated model's integration of technological change using learning curves to
model cost reductions resulting from accumulated experience, which is a well-established approach in the
literature. Additionally, we describe two novel approaches for introducing more dynamic heterogeneity between
countries within the model. Preliminary modeling results show that implementing technological change in the
model results in an increase in the supply of emission reductions at a given price. This result is influenced by both,
cumulative experience or learning overtime, and improved reduction efficiency associated with more nascent
abatement measures.
4.2.1.1 Background
Using learning curves—also known as experience curves—to examine the relationship between cost and
experience has been a tool since Wright (1936) first conceptualized "learning-by-doing" in the economics literature
(Papineau, 2006). A learning curve tracks the reduction in cost of a technology, product, or service as a function of
the accumulation of experience. The experience metric is a quantitative measure, usually cumulative, that is
considered a proxy for learning. While cumulative production is the most common experience metric, others
include installed generation capacity, cumulative production capacity, and average production facility size.
In the 1960s, Boston Consulting Group began to leverage the learning curve with consulting clients, and
through the 1970s, it was primarily used as a strategic planning tool and production management tool (Papineau,
2006). Learning curves have been estimated for a wide variety of technologies, including chemical processing,
airplane manufacturing, semiconductor production, and energy generation (Lieberman, 1984; Wright, 1936; Hatch
and Mowery, 1998). More recently, learning curves have been widely leveraged to model technological change in
climate change policy modeling, including in ElA's National Energy Modeling System (Gillingham et al., 2008).
The learning curve is most commonly estimated using the following equations:
Ct = C0*Xf (4.5)
LR = 1-PR = 1-2P (4.6)
where Ct represents the unit cost of a technology in year t and Co represents the cost of the first unit of a
technology when experience begins accumulating. Xt is the experience metric and p is a parameter measuring the
responsiveness of cost to the experience metric. In Equation 5.6, LR is the learning rate, which can be interpreted
as the percentage reduction in unit cost for every doubling of the experience metric. The progress rate (PR) is
simply the inverse of the learning rate.
In addition to using the linear form of the learning curve equation (Equation 4.6), one can employ a log-log
form of the learning curve or estimate a regression to test the statistical significance of multiple experience metrics
at once. See Papineau (2006) or Gillingham et al. (2008) for more detailed exploration of the different approaches
to estimating learning curves.
Broadly, researchers have also implemented technological change in modeling the MAC curve for non-CC>2
GHGs. Carrara and Marangoni (2013) estimated MAC curves and applied a technical progress factor that stretches
the shape of the MAC curve to the right, simulating a larger share of abatement being captured at the same price
as the carbon price relative to the MAC curve without technical change implemented. For this study, we borrowed
from and expanded on the implementation of technological change in the Greenhouse Gas and Air Pollution
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-9
-------
METHODOLOGY DOCUMENTATION
Interactions and Synergies (GAINS) model, which is an integrated assessment model used primarily in European
climate modeling efforts (Hoglund Isaksson et al., 2016). The GAINS model assumes a doubling rate of 15 years for
the experience metric and a 15% learning rate. Additionally, Hoglund Isaksson et al. (2016) also modeled
technological progress by assuming an annual improvement in removal efficiency (RE) of 1%.
4.2.1.2 Operationalizing Technological Change
To operationalize technological change in the MAC curve model, we used the linear form of the learning curve
equation (Equation 4.6) to project cost reductions over the study period of the model. To accomplish this, we had
to specify the shape of an experience curve and then choose a learning rate. In an ideal scenario, historical data
would inform both the shape of the experience curve and the learning rate; however, robust data were not
available on rate of adoption and the cost trajectory of the abatement technologies in the model.
In the absence of better data, we specified a generic experience curve that we applied to all abatement
technologies and we pulled from the literature to identify a reasonable learning rate based on studies of proxy
technologies or industries. Additionally, as previously implemented in the GAINS model, we accounted for
technological maturity and modeled modest improvements in RE (Hoglund Isaksson et al., 2016).
4.2.1.3 Defining an Experience Curve
The rate of cost reductions when modeling technological change is not a function of time, but of experience.
As described in the previous section, the learning rate is the percentage reduction in cost for every doubling of the
experience metric, which is most often a cumulative production metric. Some applications in the learning curve
literature assume a steady doubling rate. For example, the GAINS model assumes a doubling rate of 15 years in its
implementation of technological change for non-CC>2 MAC curve estimation (Hoglund Isaksson et al., 2016).
However, assuming a steady doubling rate builds in an assumption that the market for a product is limitless
because growth over the long term becomes exponential. Rather than apply a steady doubling rate to the model,
we drew from the innovation adoption literature to model a more realistic adoption path. Rogers (1962) was the
first to propose the innovation adoption curve, which follows a sigmoidal shape, or an S-curve. Using a sigmoidal
shape models a slower rate of adoption in the early years that accelerates as a product reaches a more main-
stream audience. In the out-years, the adoption curve tapers off as the market becomes saturated. Figure 4-2,
adapted from Rogers (1962), illustrates how different segments of adopters influence the pace of adoption and
therefore the shape of the curve.
To operationalize this concept in the model, we used the general equation for a sigmoidal function
(Equation 4.7), which gives us the flexibility to manipulate the shape of the curve by changing the asymptote
(max), the slope of the curve (a), and the midpoint of the curve (m). Given the lack of data available on adoption
trends for individual abatement technologies, we chose a conservatively sloped adoption curve with the primary
focus being to model the slower pace of adoption at the tails of the curve (Figure 4-2).
Sigmoidal curve equation: X(t) = (4.7)
4-10
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
100
75
50
25
Innovators Early
2.5 %
Early
Adopters Majority
13.5 % 34 %
Laggards
16 %
Late
Majority
34 %
Figure 4-2: Different Segments of Adopters Influence the Pace of Adoption
Source: Adapted from Rogers, E. 1962. Diffusion of Innovations. London: Free Press.
4.2.1.4 Choosing a Learning Rate
Once the experience curve is defined, we turn to choosing a learning rate. While many applications of the
learning curve concept are directed to estimating the learning rate with real-world data about cost and adoption,
the prospective nature of our modeling required choosing the learning rate in order to estimate the cost reduction.
In choosing a learning rate, we followed Hoglund Isaksson et al.'s (2016) choice of 15%, which is grounded
primarily in the literature that estimates learning rates for energy generation technology. Jamasb and Kohler
(2007) reviewed the learning curve literature for energy technologies, citing 20% as a common "rule of thumb"
learning rate. However, they argued that choosing 20% in a modeling setting may be optimistic if it is fixed over
time because some evidence suggests that learning rates decline over the long term as further improvements
become costlier and more difficult to achieve as low-hanging fruit is exhausted.
Hoglund Isaksson et al. (2016) also argued that 20% may be unrealistic in the case of modeling more mature
technologies for which many cost reductions have already been achieved. Thus, 15% is a conservative choice to
account for the concern about overestimating cost reductions that are achievable over time.
Combining the experience curve and the learning rate yields a learning curve that drives costs down more
rapidly in the early years and flattens out in later years as potential cost reduction opportunities are exhausted
(Figure 4-3). The result of implementing the learning rate equation is a cost factor time series (Ct from
Equation 4.5) that ranges from 0 to land starts at 1 at the beginning of the learning curve. Thus, the cost factor
can be interpreted as 100% minus the percentage reduction in cost achieved through learning. The cost factor is
applied to the cost of an abatement technology to model cost reductions achieved over time through learning.
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-11
-------
METHODOLOGY DOCUMENTATION
Figure 4-3: Illustrative Experience and Learning Curves
IA
a
U
100%
90%
S0%
70%
60%
50%
40%
30%
ZO%
10%
0%
Learn ing Cup/e
fCrt^r Fartnr'i
1
•
.*
»•
¦ *
.*
• •
J ExperienceCuive
#
*
^L-umuiauvc
Production)
» *
• *
.*
••
8
7
6
5
4
3
2
1
2000
2020 2040
2060
Vear
20 SO
2100
2120
u
ro
LL
QJ
(J
c
QJ
QJ
CL
X
¦ Cost Factor
Experience Metric
4.2.1.5 Accounting for Technological Maturity and Variation in Country Development Status
Limited data availability necessitates generalizing some of the implementation of technological change in the
MAC curve model. For example, we used the same experience curve and the same learning rate across all
abatement technologies. However, generalizing can be problematic—some distinctions can be made by taking into
account the maturity of the abatement technology and the development stage of the country where the
technology is being implemented.
As noted in the previous section, more mature technologies may have already achieved significant cost
reductions prior to our study period. Applying the learning curve starting in the model base year could thus
overestimate potential cost reductions for mature abatement technologies. To account for technological maturity,
we classified each technology as either nascent or mature. The maturity level determines (1) when the technology
starts down the learning curve within the model and (2) the assumed annual improvement in RE. For example,
coalbed degasification technology is considered a mature technology because it has been widely used for worker
safety for many years. Ventilation air methane (VAM) oxidation, while already commercialized, is not widely
adopted and was thus classified as a nascent technology.
The development stage of a country may also affect the cost and pace of technological change. We assumed
that emerging economies will not adopt abatement technologies as quickly and will face higher implementation
costs because they may need to import foreign expertise in the earliest stages of technology adoption. To account
for the development stage of countries, we classified each country as either developed or emerging based on their
World Bank Income Group classification (World Bank, 2016). The development stage of the country also affects the
year in which learning starts within the model. Table 4-4 presents the combined effect of technological maturity
and economic development stage.
4-12
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
Table 4-4: Combined Effect of Technological Maturity and Economic Development Stage
Country
Annual Reduction
Technology Status
Developed
Emerging
Efficiency Improvement
Nascent technology
2015
2020
1.0%
Mature technology
2000
2005
0.5%
Reduction Efficiency (RE) Improvements
As mentioned previously, technological change can also result in RE improvements independent of cost
reductions. The GAINS model (2016a) assumes a 1% annual improvement in RE for non-CC>2 GHG abatement
technology in their MAC curve models. We similarly assumed a 1% annual improvement in RE for nascent
technologies but opted for a more conservative 0.5% annual improvement for mature technologies (Table 4-4).10
For the purposes of the model, we express RE in terms of an emission factor, which is represented by the
percentage of a pollutant that is not captured after the abatement technology is implemented.11
Nature of the Results
Accounting for technological maturity and economic development stage of a country yields a matrix of four
cost factor and emission factor time series, which are presented in Table 4-5 to illustrate the interaction between
the different parts of the learning curve implementation. While the emission factor trend varies depending on the
maturity of the technology, the cost factor trend is uniform across the model, varying only when a technology
starts to progress down the learning curve. Figure 4-4 graphically demonstrates the difference between the four
maturity/development combinations. In 2020, mature technologies in developed nations are farthest along the
learning curve, reflecting the assumption that abatement technologies are likely to be introduced in developed
nations first.
Table 4-5: Cost Factor and Emission Factor Results
Developed Nations
Emerging Nations
Nascent Technology
Mature Technology
Nascent Technology
Mature Technology
Emission
Emission
Emission
Emission
Year
Cost Factor
Factor
Cost Factor
Factor
Cost Factor
Factor
Cost Factor
Factor
2015
1.00
1.00
0.88
0.93
1.00
1.00
0.91
0.95
2020
0.95
0.95
0.85
0.91
1.00
1.00
0.88
0.93
2025
0.91
0.90
0.82
0.88
0.95
0.95
0.85
0.91
2030
0.88
0.86
0.80
0.86
0.91
0.90
0.82
0.88
2035
0.85
0.82
0.78
0.84
0.88
0.86
0.80
0.86
2040
0.82
0.78
0.77
0.82
0.85
0.82
0.78
0.84
2045
0.80
0.74
0.75
0.80
0.82
0.78
0.77
0.82
2050
0.78
0.70
0.74
0.78
0.80
0.74
0.75
0.80
10 Note: A 1% improvement in RE is equivalent to RE * 1.01, not RE + 1%.
11 In other words, EF = 1 — RE.
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-13
-------
METHODOLOGY DOCUMENTATION
Figure 4-4: Cost Factor by Year Based on Technological Maturity and Economic Development
100%
\ V ^ V 2020, 1.00
95%
90%
\ N ^ •
\ • \ v 2020, 0.95
W \\
\\
2020, 0.88 \ •.
\ . \ \
\\ w
2020, 0.85
\s •
v. V;-...
- 85%
o
t>
£ 80%
4-»
l/l
<3 75%
70%
65%
60%
1980 2000 2020 2040 2060 2080 2100 2120 2140
Year
Cost (Developed, Nascent) —Cost (Developed, Mature)
Cost (Emerging, Nascent) — • —Cost (Emerging, Mature)
4.2.1.6 Incorporating Dynamic Heterogeneity Across Countries
To estimate abatement costs outside of the United States, we scaled costs up or down using an adjustment
factor based on the difference in GDP per capita between a given country and the United States. In previous
versions of the model, this adjustment factor was static, which does not account for economic growth over time.
Additionally, previous versions of the model assumed that each country faces its own domestic labor costs when
implementing an abatement technology. In reality, we expect that an abatement technology would be more costly
in countries that do not possess the necessary expertise domestically. Rather than facing their own domestic labor
prices, the adopting country would face a higher foreign labor price until the expertise is internalized domestically.
Not accounting for economic growth over time, particularly in rapidly developing countries such as Brazil,
India, and China, will have the effect of depressing projected abatement costs in the out-years of the model.
Similarly, not accounting for the need to import foreign expertise when a new abatement technology is being
introduced will also keep abatement costs lower than they should be in many countries, but in the early years of
the model. The impact of lower abatement costs when not accounting for economic growth and foreign labor
costs is that more abatement will be projected in the early years of the model than may be realistic. Incorporating
these factors, on the other hand, results in a more conservative projection of abatement.
To account for economic growth and the time required to internalize expertise in an adopting country, we
introduced two new features in the updated model. First, we modeled economic growth using projections of
growth in labor and materials costs. To project labor cost growth, we used wage projections from the International
Labor Organization (ILO), which we converted to 2015 Purchasing Power Parity (PPP) dollars (ILO, 2016; World
Bank, 2017). Because of the limited country coverage, we aggregated countries based on World Bank Income
Group classification and used a 5-year average growth rate. To project growth in materials cost, we used GDP per
capita growth projections from the Shared Socioeconomic Pathways Scenario #2 (SSP2), which was drawn from the
MESSAGE-GLOBIOM climate model (International Institute for Applied Systems Analysis [NASA], 2016). GDP per
4-14
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
capita was also quantified in 2015 PPP dollars. Importantly, we did not use projections of actual wage cost or GDP
per capita; instead, we used projected growth rates and applied those rates to 2015 wages and GDP per capita.
Table 4-6 describes the data and data sources used in this part of the analysis.
Table 4-6: Data Sources for Projecting Economic Growth at the Country Level
Cost
Category
Representative
Metric
Units
Baseline Data Source
Growth Rate
Projection Source
Labor
Projected wages3
2015 PPP dollars
per hourb
International Labor
Organization
International Labor
Organization
Materials
GDP per capita
2015 PPP dollars per
capita
NASA SSP2, MESSAGE-
GLOBIOM
NASA SSP2, MESSAGE-
GLOBIOM
a Wages were projected using 5-year average growth rates based on World Bank Income Group classification.
b PPP dollars were calculated using exchange rate data from the World Bank (2017).
Second, we modeled a foreign-to-domestic labor shift by imposing U.S. labor costs on foreign countries in the
early years of the model and phasing them gradually to model the internalization of expertise within a country.
The shift in labor costs was applied only to countries where labor costs are lower than in the United States. Other
components of the cost of abatement—capital, energy, and materials—were assumed to follow the domestic costs
in-country.
Economic growth projections and the foreign-to-domestic labor shift were operationalized by layering the two
time series on top of the learning curve for each country. The result is a composite factor that can be applied to an
abatement technology's cost to make it dynamic and responsive to country context over the study period of the
model.
Figure 4-5 illustrates the impact of each enhancement of the model on the cost of an abatement technology.
Figure 4-5A presents two learning curves, one for the United States (Baseus) and one for a hypothetical developing
nation (Devstatic). These two curves represent a static relationship between the United States and another nation.
Figure 4-5B illustrates how Devstatic changes when integrating a foreign-to-domestic labor shift as described
previously. DevLaborshift starts higher in early years because the nation is facing foreign labor costs in the early years
of adopting an abatement technology. Over time, however, expertise is internalized in-country and labor costs
merge with the original Devstatic curve.
Figure 4-5C illustrates the impact on Devstatic of modeling economic growth through increasing material and
labor costs over time. In this illustrative example, DevGrowth shows that the hypothetical developing nation is
experiencing a faster rate of economic growth than the United States, which causes the gap in costs between the
two nations to close over time.
Finally, Figure 4-5D combines DevLaborshift and DevGrowth to illustrate the combined effect of a labor shift and
economic growth on Devstatic. It is important to reiterate that when modeling economic growth, we calculated
relative factors for each country using the United States as the base. The Baseus curve, while it does not change
throughout Figure 4-5, incorporates projections of the United States' own economic growth.
4.2.1.7 Integration into the Non-C02 MACCModel
The application of technology change to existing MACC calculations implies two important updates to the
previous estimates. First, we switched from using static capital, labor energy, and materials (KLEM) factors as
explained before, allowing them to change every year and representing the cost savings due to technological
change. Second, by applying the RE improvement factor to the current technical effectiveness in the model, we
introduced a dynamic reductions efficiency factor that improves over time.
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-15
-------
METHODOLOGY DOCUMENTATION
Figure 4-5: Illustrative Example of the Impact of Modeling Foreign-to-Domestic Labor Shift arid
Economic Growth on Relative Cost Factors for Abatement Technologies
(A) Static Relationship Between Countries (B) Modeling Foreign-to-Domestic Labor Shift
«/>
o
u
Baseus
Time (Years)
Baseus
Time (Years)
(C) Modeling Economic Growth Over Time (D) Modeling Labor Shift and Growth
Baseus
Time (Years)
4-*
1/1
o
<_>
Baseus
DeVutxvShffl + OeVOrowth
Time (Years)
As described in Section 4.1, the MAC curve model characterizes each abatement option in terms of its costs
and benefits per abated unit of gas (tCChe). Setting each option's benefits equal to costs, the MAC curve model
solves for the carbon price ($/tCChe), also referred to as the option's break-even price (Equation 4.8).
As shown in Equation 4.8, the adjustments to implementation costs affect the initial investment costs (CC) as
well as the recurring costs (RC), while improvements to RE increase the quantity of GHG emissions captured (ER),
and for energy-producing measures (captured and use of Cm), the adjustment also affects the revenue (R)
component.
P =
CC,
¦ +
RCt
R
ER,
(_CCt
I I
TR
ERA-T 1 — TR
Impacted by learning ! Impacted by removal efficiency improvement
(4.8)
The break-even analysis yields the new adjusted break-even price and corresponding incremental emission
reduction for each mitigation measure evaluated across 195 countries.
4-16
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
Operationalizing technological change in the MAC model has two general effects on the shape of the resulting
MAC curve. Reduction efficiency improvements shift the asymptote of the MAC curve outward, increasing the
maximum abatement potential of the technology. Cost reductions through learning yield the same level of
abatement at a lower cost, causing a downward shift in parts of the MAC curve.
Figure 4-6 shows two illustrative MAC curves with and without technological change to demonstrate the
effects of cost reductions and removal efficiency improvements. All other things being equal, the magnitude of the
shifts in the MAC curve will be greater for emerging economies than for developed economies.
Figure 4-6: Illustration of Technology Change's Impact on a Non-C02 MAC Curve
Cost reductions
at the same level
of abatement .
Abated GHG Emissions (MtCC>2e)
Reduction
Efficiency
Improvements
w/o Tech Change w/ Tech Change
4.2.2 Regionalization of Sectoral MAC Curves for the United States
The second major enhancement we have made to the MAC model since the global mitigation report (EPA,
2013a) was to develop regionalized sectoral MAC curves for the United States to better capture the heterogeneity
in cost of mitigation and provide more explicit geospatial representation of the abatement potentials. These
regionalized MAC curves now replace the single national MAC curve previously produced for the United States.
We have developed regional MAC curves for the coal mining, oil and gas, landfills, and nitric/adipic acid
production; livestock; rice cultivation; and cropland soil management sectors. Requisite regional data for
developing comparable regional MACs for the 12 industrial sources were not available at the time of this
publication, but it is our goal to develop regional MAC curves for these sectors in future updates.
We used state-level data on production input costs (labor, energy, and materials), yields, and emission factors
to derive regional MAC curves for the United States. This detailed modeling framework enables us to calculate the
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-17
-------
METHODOLOGY DOCUMENTATION
quantity of GHG reductions and corresponding mitigation costs at the state and regional levels to introduce more
heterogeneity into the MAC model. The methods and approach used are similar to those developed for a study
regionalizing the domestic MAC curves for four U.S. non-CC>2 emitting sectors that included coal mining, oil and
gas, landfills, and nitric/adipic acid production (Rose, Petrusa, and Davis, 2013).
GHG abatement was quantified by technology for each state and can be aggregated to larger regions for
simplified reporting results. The suite of mitigation options evaluated in the regional models is consistent with the
international model. The state-level supply of GHG mitigation for each source was developed following a common
four-step approach.
First, we derived state-level baseline GHG emissions based on the regional share of activity or production. For
example, in coal mining, we used the underground coal mining production to estimate the share of emissions by
state. For the agricultural sectors, we used the state shares of national production to determine state-level
baselines for livestock and rice cultivation. Second, we estimated the direct costs and benefits of state-level GHG
mitigation. We then used the state-level baselines and mitigation measure economics to calculate break-even
prices associated with adopting each mitigation measure in each state. Finally, we constructed MAC curves for
each state in the contiguous United States (Alaska and Hawaii were omitted for this analysis).
4.2.2.1 Specification of Baseline GHG Emissions by Region and Sector
For each sector, we developed state baseline production levels and GHG emissions for the years 2015 through
2050. This initial step defined the pool of GHG emissions and the quantity of emissions applicable to each
mitigation measure. The data sources used to generate state baseline emissions varied across the three source
models. For livestock, we developed state-level baseline activity based on head counts by animal type reported by
the U.S. Department of Agriculture (USDA) National Agricultural Statistics Service (2014a, b, c). Additionally, we
used data from the EPA 2016 GHG Inventory to develop state-level shares of baseline emissions and relative
emission factors by animal type.
For rice cultivation, we relied on the DNDC model to determine regional production and emissions resulting
from various irrigation and nitrogen use regimes to derive the resulting GHG emissions by state. Emissions are
available at a national level from DNDC and are allocated to rice-producing states (Arkansas, California, Louisiana,
Mississippi, Missouri, and Texas for our analysis) using state-level rice production shares and area data. More
information about the DNDC data can be found in the 2013 mitigation report (EPA, 2013a).
For nonrice croplands, we relied on spatially disaggregated data from the DAYCENT model, including crop
production areas and emission rates. The DAYCENT model results are provided on a 25-km grid cell basis for North
America. We calculated weights based on grid area to allocate shares of harvested area, yields, and GHG emissions
(CH4, N2O, and soil carbon) to each state in the contiguous United States.
4.2.2.2 Regional Price Adjustment Factors (KLEM)
We developed state-specific mitigation costs (initial capital investment, annual recurring fixed and variable
O&M) and benefits (e.g., labor savings, energy savings, yield impacts). Accounting for regional differentiation in
prices for labor, energy, and materials allowed us to create heterogeneity in the break-even prices across states for
the same mitigation technology. We developed state price factors using deviations from national average prices
for capital, labor, electricity, and materials. Here, the state price factor represents the ratio of state price to
national average price. Table 4-7 provides summary statistics on the relative price factors calculated across states.
4-18
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
Table 4-7: Regionalized Price Adjustments Factors, Descriptive Statistics
Price Factor
Minimum
Maximum
Median
Mean
Standard Deviation
Labor3
Non-ag sectors
0.24
4.30
1.00
1.08
0.52
Cattle
0.77
1.07
0.98
0.97
0.07
Dairy cattle
0.62
1.30
1.00
1.04
0.16
Swine
0.62
1.30
1.00
1.04
0.16
Maize
0.81
1.43
1.07
1.10
0.15
Soy
0.41
3.41
0.81
0.96
0.63
Sorghum
0.34
3.36
1.00
1.15
0.63
Wheat
0.65
4.30
1.04
1.37
0.89
Barley
0.24
2.63
1.00
0.93
0.57
Rice
0.66
1.31
1.00
0.99
0.08
Electricity13
All sectors
0.61
4.12
0.96
1.14
0.55
Materialsc
Non-ag sectors
0.72
1.90
1.00
1.02
0.12
Cattle
0.79
1.24
1.01
1.02
0.08
Dairy cattle
0.87
1.14
1.00
1.01
0.05
Swine
0.87
1.14
1.00
1.01
0.05
Maize
0.84
1.30
0.98
1.01
0.11
Soy
0.83
1.35
1.00
1.03
0.13
Sorghum
0.86
1.07
1.00
1.00
0.04
Wheat
0.84
1.72
1.00
1.04
0.21
Barley
0.72
1.50
1.02
1.07
0.15
Rice
0.74
1.90
1.00
1.01
0.14
Sources:a [Non-ag sectors] U.S. Bureau of Labor Statistics (BLS). 2018. Annual Mean Wage: Construction and Extraction
Occupations (SOC code 470000); [Agriculture sectors] USDA Economic Research Service (ERS). 2018. Production Costs and
Returns Data by Commodity.
b U.S. Energy Information Administration. 2017. State Energy Data System (SEDS).
c [Non-ag sectors] U.S. Bureau of Economic Analysis. 2018. Gross Domestic Product per Capital by State; [Agriculture sectors]
U.S> Department of Agriculture, Economic Research Service. 2018. Production Costs and Returns Data by Commodity.
To be consistent with the international MAC curve modeling, we assumed that there is no regional
differentiation in capital costs. Hence, the relative cost factor for capital is 1 in all states.
Labor and materials factors were constructed using the Commodity Costs and Returns reports from USDA's
Economic Research Service (USDA, 2018) for the specific crops and livestock types included in the model. Data
were available at the USDA Farm Resource Region level. These data were mapped to states using a county-to-
Global Non-C02 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-19
-------
METHODOLOGY DOCUMENTATION
Resource Region mapping from USDA.12 Materials cost indices were constructed using components of operating
costs detailed in the commodity costs and returns report that were not already accounted for in the model (e.g.,
custom operations, repairs, irrigation, or other variable expenses) relative to U.S. values. Relative costs of inputs
like seeds, fertilizer, fuel, or chemicals were not included in these materials cost indices because the model already
accounts for them. The term "materials" is intended to capture all the nonlabor and nonenergy recurring O&M
costs or potential savings associated with each mitigation measure.
Labor cost indices were calculated using regional hired labor costs as reported in the USDA Commodity Costs
and Returns report relative to U.S. values. These relative costs were allocated to states using an average weighted
by the area of the state in each Resource Region.
Electricity factors were calculated based on industrial sector electricity prices reported by ElA's State Energy
Data System (2016) from 2010 to 2014. These prices were averaged by state and divided by the U.S. 5-year
average from that period to calculate a state price factor for electricity. We used this price factor to adjust any
recurring energy costs for operating a mitigation technology as well as the potential revenue associated with
electricity offsets resulting from converting captured Cm to electricity.
It is important to note that these price factors are static in our analysis. While we allow real price growth for
the national price over time, the relative cost across states does not change.
4.2.3 MAC Limitations and Uncertainties
The results of the MAC analysis cover the major emitting regions, emission sources, and abatement options;
we discuss a few limitations of this analysis briefly below.
Exclusion of Transaction Costs
Ongoing work in the area of mitigation costs continues to focus on including transactions costs. As discussed in
the 2007 version of this report, Assessing Transaction Costs of Project-based Greenhouse Gas Emissions Trading
(Antinori and Sathaye, 2007), transactions costs range between $0.03 per metric ton of CO2 for large projects and
$4.05 per metric ton of CO2 for smaller projects, with a weighted average of $0.36 per metric ton of CO2 for a suite
of projects considered. More recent MAC work by others (Rose et al., 2013) estimated that the unit cost of an
abatement project increases by an average of 15% when transaction costs are included. Transaction costs vary
significantly, contingent on the size of the project, the applicable mitigation technology, and other factors. Given
the lack of comprehensive data, this analysis continues to exclude transaction costs from the analysis.
Limited Use of Regional Data
The analytic framework used in this study is flexible enough to incorporate regional differences in all the
characteristics of abatement options. However, a lack of country-specific data led to a reliance on expert
judgment, as noted in the sector-specific sections. This expert judgment was obtained from source-level technical
experts in government and industry with knowledge of project-level technologies, costs, and specific regional
conditions. The applicability of abatement options, for example, relies on expert judgment, because the makeup of
the current infrastructure in a given country in a given sector is uncertain. A much greater use of data originating
from local experts and organizations is recommended for the follow-up research on CH4 abatement in countries
outside the United States and EU. Incorporating more regional data could also enhance the range of emission
sources and mitigation options addressed in this analysis.
12 USDA Farm Resource Regions were mapped to counties using this mapping: https://wavback.archive-
it.org/5923/20110913212900/httD://www.ers.usda.gov/Briefing/ARMS/resourceregions/resourceregions.htm I
4-20 Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
Exclusion of Indirect Emission Reductions
This analysis does not account for indirect emission reductions, which can result from either the substitution
of electricity from the grid with electricity produced on site from recovered CH4 or from the substitution of natural
gas in pipelines with recovered CH4. Calculation of such indirect reductions requires additional assumptions about
the carbon intensity of electricity in different regions. In the U.S. landfill sector, indirect reductions generally
augment emission reductions by about 15%. In the agricultural sector, although some mitigation options primarily
target a single gas, implementation of the mitigation options will have multiple GHG effects, most of which are
reflected in the agricultural results.
4.3 References
Antinori, C. and J. Sathaye. 2007. Assessing Transaction Costs of Project-based Greenhouse Gas Emissions Trading.
Report No. LBNL-57315. Berkeley, CA: Lawrence Berkeley National Laboratory.
Carrara, S. and G. Marangoni. 2013. Non-CCh greenhouse gas mitigation modeling with marginal abatement cost
curves. Fondazione Eni Enrico Mattei (FEEM). JSTOR. Available online at
http://www.istor.org/stable/resrep01012 J.
Gallaher, M. P., Delhotal, K. C., & Petrusa, J. E. (2005). International marginal abatement costs of non-CC>2
greenhouse gases. Journal of Environmental Sciences, 2(2-3), 327-338.
Gillingham, K., R.G. Newell, and W.A. Pizer. 2008. Modeling endogenous technological change for climate policy
analysis. Energy Economics, 30(6), 2734-2753.
Hoglund Isaksson, L, Winiwarter, W., Purohit, P., & Gomez-Sanabria, A. 2016. Non-C02 greenhouse gas emissions
in the EU-28 from 2005 to 2050: GAINS model methodology. Retrieved from http://pure.iiasa.ac.at/13398AJ
International Institute for Applied Systems Analysis. 2016. Shared Socioeconomic Pathways Scenario #2—
MESSAGE-GLOBIOM. Available online at https://tntcat.iiasa.ac.at/SspDb/dsd?Action=htmlpage&page=10 ef
International Labor Organization. 2016. Global Wage Report 2016/17: Wage Inequality in the Workplace
[Growth_Wage_Rates], Available online at http://www.ilo.org/global/ research/global-reports/global-wage-
report/2016/lang-en/index.htm c?
Jamasb, T., and J. Kohler. 2007. Learning curves for energy technology: A critical assessment. Cambridge, UK:
University of Cambridge. doi:10.17863/CAM.5144
Lieberman, M.B. 1984. The learning curve and pricing in the chemical processing industries. The RAND Journal of
Economics, 15(2), 213-228.
Papineau, M. 2006. An economic perspective on experience curves and dynamic economies in renewable energy
technologies. Energy Policy, 34(4), 422-432.
Ragnauth, S.A., J. Creason, J. Alsalam, S Ohrel, J. Petrusa, and R. Beach. 2015. Global mitigation of non-CC>2
greenhouse gases: marginal abatement costs curves and abatement potential through 2030. Journal of
Integrative Environmental Sciences, 12(supl), 155-168. doi:10.1080/1943815X.2015.1110182
Rogers, E. 1962. Diffusion of Innovations. London, NY: Free Press.
Rose, S., J, Petrusa, and L. Davis. 2013. PRISM 2.0: Regional Non-CC>2 Greenhouse Gas Abatement Potential in the
United States for 2010-2030. Report 3002001675. Palo Alto, CA: Electric Power Research Institute.
Rose, S.K., R. Beach, K. Calvin, B. McCarl, J. Petrusa, B. Sohngen, R. Youngman, A. Diamant, F. de la Chesnaye, J.
Edmonds, R. Rosenzweig, and M. Wise. 2013. Estimating Global Greenhouse Gas Emissions Offset Supplies:
Accounting for Investment Risks and Other Market Realties. Report 1025510. Palo Alto, CA: Electric Power
Research Institute.
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-21
-------
METHODOLOGY DOCUMENTATION
Russian Federation Federal State Statistics Service. 2010. Main Indicators: Living Standards. Table 7.7: Average
Monthly Nominal Accrued Wages of Employees of Organisations by Kinds of Economic Activities. Available
online at http://www.gks.ru/bgd/regl/blO 12/lssWWW.exe/stg/d01/07-07.htm J
U.S. Bureau of Economic Analysis. 2018. Gross Domestic Product per Capita by State [Non-ag sector]. Available
online at https://www.bea.gov/data/gdp/gdp-state
U.S. Bureau of Labor Statistics. 2010a. International Comparisons of Hourly Compensation Costs in Manufacturing,
1996-2008. Table 8: Production Workers: Hourly Compensation Costs in U.S. Dollars in Manufacturing, 33
Countries or Areas and Selected Economic Groups, 1975-2008; and Table 2: All Employees: Hourly
Compensation Costs in U.S. Dollars in Manufacturing, 33 Countries or Areas and Selected Economic Groups,
1996-2008. Released August 26, 2010. Available online at http://www.bls.gov/fls/
U.S. Bureau of Labor Statistics. 2010b. Labor Costs in India's Organized Manufacturing Sector. Table 2: Hourly
compensation costs in India's organized manufacturing sector, 1999-2005. Available online at
http://www.bls.gov/opub/mlr/2010/Q5/artlfull.pdf
U.S. Bureau of Labor Statistics. 2010c. Manufacturing in China. Table 1: Hourly Compensation Costs of
Manufacturing Workers in China, 2002-2008. Available online at
http://www.bls.gOv/fls/china.htm#data comparability
U.S. Bureau of Labor Statistics. 2018. Annual Mean Wage: Construction and Extraction Occupations [Non-ag
sectors] (SOC code 470000). Available online at https://www.bls.gov/opub/ted/2017/annual-mean-wage-of-
construction-and-extraction-occupations-was-48900-in-2016.htm#tab-2
U.S. Department of Agriculture, Economic Research Service. 2010. Agricultural Resource Management Survey
(ARMS): Resource Regions. ERS U.S. Farm Resource Regions. Available online at
https://wavback.archiveit.Org/5923/20110913212900/http:/www.ers.usda.gov/Briefing/ARMS/resourceregio
ns/resourceregions.htm cf
U.S. Department of Agriculture, Economic Research Service. 2018. Production Costs and Returns Data by
Commodity. Agriculture Sectors. Available online at https://www.ers.usda.gov/data-products/commoditv-
costs-and-returns/
U.S. Energy Information Administration. 2010a. International Energy Statistics. Electricity Prices for Industry.
Available online at http://www.eia.gov/emeu/international/elecprii.html
U.S. Energy Information Administration. 2010b. International Energy Statistics. Natural Gas Prices for Industry.
Available online at http://www.eia.gov/emeu/international/ngasprii.html
U.S. Energy Information Administration. 2016. State Energy Data System (SEDS). Available online at
https://www. eia.gov/state/seds/seds-data-complete. php?sid=US
U.S. Energy Information Administration. 2017. State Energy Data System (SEDS). Available online at
https://www. eia.gov/state/seds/seds-data-fuel. php?sid=US
U.S. Environmental Protection Agency. 2006. Global Mitigation of Non-CC>2 Greenhouse Gases. EPA-430-R-06-005.
Washington, DC: EPA.
U.S. Environmental Protection Agency. 2013a. Global Mitigation ofNon-CC>2 Greenhouse Gases: 2010-2030. EPA-
430-R-13-011. Available online at https://www.epa.gov/global-mitigation-non-co2-greenhouse-gases/global-
mitigation-non-co2-ghgs-report-download-report
U.S. Environmental Protection Agency. 2013b. International Non-CCh Greenhouse Gas. Non-CCh Emissions by
Country. Available online at https://cfpub.epa.gov/nonco2/
U.S. Environmental Protection Agency. 2016. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2016.
Available online at https://www.epa.gov/ghgemissions/inventorv-us-greenhouse-gas-emissions-and-sinks-
1990-2016
4-22
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
-------
SECTION 4 — MAC ANALYSIS—GENERAL METHODOLOGY
United Nations Conference on Trade and Development (UNCTAD). 2012. National Accounts: Nominal and Real
GDP, Total and per Capita, Annual 1970-2011. UNCTAD Statistical Database. Available online at
http://unctadstat.unctad.org ef
Weyant, J. and F. de la Chesnaye (eds.). 2006. Multi-greenhouse gas mitigation and climate policy. Energy Journal.
Available online at https://emf.stanford.edu/publications/emf-21-multi-greenhouse-gas-mitigation-and-
climate-policvef
World Bank. 2016. World Bank List of Economies. Available online at
http://databank.worldbank.org/data/download/site-content/CLASS.xlstf
World Bank. 2017. World Development Indicators [Exchange Rates and Prices], Available online at
http://wdi.worldbank.Org/table/4.16 J
Wright, T. P. 1936. Factors affecting the cost of airplanes. Journal of Aeronautical Sciences, 3(4), 122-128.
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
4-23
-------
-------
5 Sector-Level Methods
This section describes the source-specific methodology, organized by the four sectors—energy, industrial
processes, agriculture, and waste. Within each sector, information on the activity data and emission factors used
for both historical and projected emission estimates and the mitigation options considered is provided for each
source category. Assumptions, data sources, and calculations are described along with a concluding section that
identifies the uncertainties and other considerations associated with developing the emission and mitigation
estimates.
5.1 Energy Sector
This section presents the methodology used to estimate Cm and N2O emissions and mitigation from the
following energy sources:
• coal mining activities (CH4) (projections and mitigation)
• ONG systems (CH4) (projections and mitigation)
• stationary and mobile combustion (Cm, N2O) (projections only)
• biomass combustion (Cm, N2O) (projections only)
• other energy sources (CH4, N2O) (projections only), including
- waste incineration (CH4, N2O) and
- fugitives from solid fuels (N2O)
5.1.1 CH4 in Coal Mining
Cm is stored within the coal seams and the surrounding rock strata and is liberated when the pressure above
or surrounding the coal bed is reduced as a result of natural erosions, faulting, or mining. CH4 is produced during
the process of coalification, where vegetation is converted by geological and biological forces into coal. Because
CH4 is explosive, the gas must be removed from underground mines high in CH4 as a safety precaution.
Coal mining is a significant source of anthropogenic GHG emissions. Coal is an important energy resource in
many of the world's economies, used for energy generation or as a feedstock in industrial production processes.
Extracting this important resource through underground and surface mining releases CH4 stored in the coal bed
and surrounding geologic strata.
The quantity of gas emitted from mining operations is a function of two primary factors: coal rank and coal
depth. Coal rank is a measure of the carbon content of the coal, with higher coal ranks corresponding to higher
carbon content and generally higher CH4 content. Coals such as anthracite and semi anthracite have the highest
coal ranks, while peat and lignite have the lowest. Pressure increases with depth and prevents CH4 from migrating
to the surface; as a result, underground mining operations typically emit more CH4 than surface mining. In addition
to emissions from underground and surface mines, post-mining processing of coal and abandoned mines also
release CH4. Abandoned coal mine emissions are included in the Other Energy source estimates.
5.1.1.1 Coal Mining Projections Methodology
Estimates consist of CH4 emissions from surface and underground mining. UNFCCC-reported, country-specific
estimates were used to give historic emission estimates and to determine emission rates of projections. The
methodologies described in this section were used to determine trends over time, or if country-reported emission
estimates were not available for any historical years, EPA used the 2006 IPCC Tier 1 methodology (IPCC, 2006) to
estimate emissions (see Section 3.3, Generating the Composite Emission Projections, for additional information).
Activity data for coal mining included anthracite, metallurgical, bituminous, subbituminous, and lignite coal
Global N011-CO2 Greenhouse Gas Emission Projections & Marginal Abatement Cost Analysis
5-1
-------
METHODOLOGY DOCUMENTATION
production and consumption from EIA (2018). Projections of consumption data were also provided by EIA (2018).
As China is the top coal-producing country, a country-specific methodology was used to estimate China-specific
emission estimates. Emissions from coal mining were disaggregated into surface and underground mining. EPA
also accounted for CFU abated by coal mine CFU projects as reported in the EPA International Coal Mine Methane
Projects Database. The Tier 1 basic equation to estimate Cm emissions from coal mining is as follows:
Fugitive CFU Emissions = (Annual Lignite Production x EFsurface) + (Annual Other Coal Production x EFunderground)
Unless otherwise noted, the EPA assumed that lignite is produced at surface coal mines, while other coal is
produced in underground mines. Because a default methodology for fugitive emissions from abandoned mines is
not currently available, this source is not considered in this report, unless it is included in country-reported
emissions.
Activity Data
Historical
Because China is the top coal-producing country, a country-specific methodology was used to estimate China-
specific emission estimates. A presentation given by the China Coal Research Institute (Lixin, 2016) provides
detailed estimates of China-specific emissions by production type (surface versus underground) and from different
stages of production (VAM, mining, and post-mining). Based on the presentation, a China-specific net emission
value was obtained for 2015. This 2015 net emission estimate was used in conjunction with UNFCCC country-
reported 1994 and 2005 values to calculate historical emissions. To interpolate the years between each of these
country-specific estimates, a GHG intensity was determined using the country-reported estimates and EIA-
reported production volumes for the 3 years 1994, 2005, and 2015. These intensities for each year were then
interpolated and multiplied by that year's associated EIA production volume to give revised net emission estimates
for China.
For all other countries, when UNFCCC country-reported data were not available for any years, historical
emissions were calculated using the Tier 1 equation above and activity data and emission factors as outlined
below.
• Coal production estimates for total primary production, hard coal production, and lignite production for
1990 through 2014 were obtained from ElA's International Energy Statistics Portal (EIA, 2018).
• Production estimates were disaggregated into surface and underground mines, assuming that hard coal is
produced in underground mines and lignite, or soft coal, is produced in surface mines. While total primary
coal production is available for all historic years, the ElA-reported data provide production volumes by
type of coal for 2013 and 2014 only. Therefore, surface and underground proportions of production from
2013 were chosen to apportion primary coal production into surface and underground volumes for all
historic years.
• If historical data were unavailable for a particular country through the UNFCCC or ElA's International
Energy Statistics Portal, it was assumed that coal mining emissions were zero.
Projected
Activity data were forecasted using projected changes in country-level coal consumption from ElA's
International Energy Outlook (EIA, 2017). If EIA did not report country-specific coal consumption forecasts, ElA's
trend estimates for the country's region were used. In some cases, EIA provided estimates for a few countries
within a region and then an estimate for the "rest of the region. For applicable countries, "rest of" estimates of
forecasted coal consumption were used to project activity data.
5-2
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Emission Factors
Historical and Projected
Where IPCC Tier 1 methodology was used, Cm emissions from coal mining activities were determined by
multiplying activity data (i.e., soft and hard coal production) by default Tier 1 IPCC emission factors from the 2006
IPCC Guidelines for National Greenhouse Gas Inventories (IPCC, 2006). The 2006 IPCC Guidelines provide low,
average, and high Tier 1 emission factors. The average emission factor was used for each country-level estimate.
Emission Reductions in Baseline Scenario
The EPA considered accounting for Cm abated by coal mine Cm projects as reported by those projects in the
EPA International Coal Mine Methane Projects Database. However, emission estimates were not adjusted for any
country that self-reported estimates in UNFCCC. It was assumed that countries that self-reported estimates had
the opportunity to account for coal mine Cm projects in their own estimates and that any country-made
adjustments for coal mine Cm projects were captured when projecting emissions forward; however, whether
countries actually accounted for coal mine Cm projects in their estimates is unknown.
For countries that did not self-report estimates, China, Mexico, and South Africa were the only three countries
that have reported reductions that were considered. Of those countries:
• China: country-specific methodology, as discussed above, inherently includes reductions.
• Mexico and South Africa: coal mine recovery-reported reduction volumes are low in magnitude and have
little impact on results.
Based on these criteria, country-level estimates were not adjusted for reported reductions.
The methodology used for this source category does not explicitly model any emission reductions; however,
emission reductions are included to the extent they are reflected in country-reported data.
Uncertainty
Several methodologies were used to calculate historical emissions, depending on data availability for a given
country. While this approach generated more detailed estimates than under a general, one-size-fits-all approach, it
introduces some uncertainty to the estimates.
Emissions were projected using regional coal consumption growth rates and, for the most part, were not
customized to individual countries. While this approach allows regional trends to be consistent with trends
projected by EIA, it introduces uncertainty into emissions for individual countries.
Furthermore, emission estimates were calculated by projecting aggregated total coal consumption rather than
calculating emissions based on trends of underground and surface mining using the Tier 1 equation. This approach
introduces uncertainty because it would not capture any shifts in surface to underground mining (or vice versa),
which are associated with different emission factors.
Emission calculations for this source were based on coal production and consumption statistics, divided into
hard coal and lignite. Classes of coal are often difficult to determine; therefore, production estimates themselves
can contain uncertainties. Additionally, Cm emissions are not necessarily directly related to production. Cm
emissions occur not only during mining, but also during the pre-mining stage and after mining is completed. In
addition, the actual gas levels of a mine can vary significantly based on geologic factors. More accurate estimation
would include information on the gas levels of mines in particular regions and mine operations in the pre-mining
and post-mining stages.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-3
-------
METHODOLOGY DOCUMENTATION
5.1.1.2 Mitigation Options Considered
Five abatement measures were considered for underground coal mining: pipeline injection, on-site electricity
generation, on-site use for process heat, enclosed flare system, and VAM oxidation. We classified the five
abatement measures into three technology categories: energy end uses, excess gas flaring, and mitigation of VAM.
It should be noted that mitigation of gas from degasification systems and mitigation of gas from ventilation
systems are independent of each other.
Abatement measures in the coal mining sector consist of one or more of the following primary components:
(1) a drainage and recovery system (where applicable) to remove Cl-Ufrom the coal seam pre-mining or from the
gob area post-mining, (2) the end-use application for the gas recovered from the drainage system (where
applicable), and (3) the VAM recovery or mitigation system (where applicable). Table 5-1 summarizes the five
mitigation technologies analyzed.
Table 5-1: Summary of Abatement Measures for Coal Mines
Abatement Measure
Total Installed
Capital Costa
(million USD)
Total Annual
O&M Cost
(million USD)
Technical
Lifetime (years)
Technical
Effectiveness13
(%)
Energy End Uses
Pipeline injection
8.4
2.4
15
21%
On-site electricity generation
23.0
2.6
15
28%
On-site use for process heat
2.8
1.2
15
28%
Excess Gas Flaring
Enclosed flare system
2.3
1.5
15
28%
Mitigation of VAM
VAM oxidation
8.0
1.3
15
19-68%
a Capital costs include costs of both recovery and abatement equipment requirements.
b Abatement potential expresses the maximum potential emission reductions at a facility level.
The components of the capital and annual costs for the drainage wells are outlined as given in the EPA's CMOP
Cash Flow Model documentation (EPA, 2011b). The recovery system includes the equipment required for drainage
wells, gas gathering lines, and delivery systems for coal mine methane (CMM). The recovery system was included
in the costs of all abatement measures with the exception of VAM oxidation.13 These costs are additive to the costs
associated with each abatement measure.
• Capital Cost: The capital costs for a drainage system are a function of the recovered gas flow rate.
Equipment requirements include construction of the drainage well(s), a wellhead blower, a satellite
compressor station, and gathering pipelines that connect the compressors to the Cm end-use technology.
The total installed capital costs vary by location and gas flow rate. For example, assuming a 600 Mcf/day
volume of CMM gas (with a CFU concentration of 90%), we estimate the capital costs would be $850,000.
• Annual Operating and Maintenance (O&M) Costs: The annual costs are required to maintain the
drainage system equated to approximately $2.2/Mcf per year. These costs include the ongoing installation
of gob wells and the gathering system piping that connects the wells to satellite compressors. In keeping
13 A recovery system is not required for VAM oxidation because it relies on the mine's existing ventilation system that would be
installed before mining operations commence.
5-4
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
with the example mine of 600 Mcf/day, the annual O&M costs associated with the recovery system would
be approximately $475,000.
• Recovery Efficiency: Recovery efficiency is assumed to be 75%.
Energy End Uses
High-quality CH4 is recovered from coal seams by drilling vertical wells from the surface up to 10 years in
advance of a mining operation or drilling in-mine horizontal boreholes several months or years before mining.
Most mine operators exercise just-in-time management of gate road development; subsequently, horizontal cross-
panel boreholes are installed and drain gas for 6 months or less (EPA, 2011b).
Pipeline Injection
Natural gas companies may purchase Cm recovered from coal mines. Cm suitable for sale into natural gas
pipelines must have a concentration of at least 96% and contain no more than 4% concentration of
noncombustible gases with a maximum of 4% CO2 or nitrogen and 1 ppm oxygen. Although Cm from coal mines
requires water removal, it is typically free of hydrogen sulfide and other impurities found in natural gas. Hence,
little to no additional treatment and processing are necessary to meet the requirements for pipeline injection. In
some cases, high-quality CH4 also can be obtained from gob wells.
Pre-mining degas wells are the preferred recovery method for producing pipeline-quality CH4 from coal seams
because the recovered CH4 is not contaminated with ventilation air from the working areas of the mine.
Gob wells, in contrast, generally do not produce pipeline-quality gas because the CH4 is frequently mixed with
ventilation air. Gob gas CH4 concentrations can range from 30% to over 90%. It is possible to upgrade gob gas for
pipeline quality, although blending with pre-mine drained gas and/or oxygen removal may be necessary, adding to
the cost of gas processing. However, it is possible to maintain a higher and more consistent gas quality through
careful monitoring and adjustment of the vacuum pressure in gob wells as has been demonstrated in the United
States (EPA, 2008).
The viability of a pipeline project is affected by several key factors. First is a coal mine's proximity to a
commercial pipeline. The cost of constructing a pipeline to connect to a commercial pipeline can vary greatly
depending on distance. Secondly, and more importantly, the terrain affects the viability of a commercial pipeline
project. Many mining areas are located in hilly and mountainous regions, thus increasing the difficulty and cost of
constructing both gathering lines and pipeline to connect to the commercial natural gas pipeline (EPA, 2008).
Finally, disposal of water produced from vertical wells may be a significant factor in determining the economic
viability of a pipeline project.
• Initial Capital Cost: The per-facility installed capital costs for pipeline injection of gob gas, as described in
EPA (2011b), include the installation of a pressure swing adsorption system to remove nitrogen and CO2
down to a 4% inert level. The utilization cost is a function of both the inlet gas flow rate and CH4
concentration and includes the cost of dehydration and compression necessary to process the gas and
boost the sales gas to pressure for injection in a natural gas transmission pipeline. Although pipelines
operate at a range of pressures, this analysis assumed an operating pressure of 900 psig. This option also
includes the installation of a catalytic oxygen removal system and a pipeline to connect the mine's gas
processing system to a natural gas pipeline. Pipeline costs estimated for this analysis were adjusted based
on mine proximity to commercial pipeline but do not attempt to account for variations in terrain across
countries.
• Annual O&M Costs: The annual costs include costs of recovery system and cost of gas treatment and
compression required for injection into commercial natural gas pipelines.
• Annual Benefits: Revenues from this option are the gas sales based on the volume of gas produced and
the market price of natural gas.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-5
-------
METHODOLOGY DOCUMENTATION
• Technical Effectiveness: The analysis assumed a technical effectiveness of 21%. As shown in Table 5-2,
this technical effectiveness considers a recovery efficiency of 75% (reflects the loss of 25% of the gas that
cannot be used in pipeline injection because the Cm concentration is too low to process to pipeline
specifications) and a destruction efficiency of 75% to account for losses during transport to point of sale
and injection into a commercial natural gas pipeline.
• Technical Lifetime: 15 years
Table 5-2: Factors Used to Estimate Abatement Potential in Coal Mines
Abatement Measure
Facility
Applicability
Technical
Feasibility
Reduction
Efficiency
Technical
Effectiveness
Energy End Uses: Drained Gasa
Pipeline injection
100%
100%
75%
28%
On-site electricity generation
100%
100%
75%
28%
On-site direct use
100%
100%
75%
28%
Mitigation Only: Drained Gasa
Enclosed flare system
100%
100%
75%
28%
Oxidation of VAMa
VAM oxidation
100%
77%
25%-90%
19-68%
a Drained gas is assumed to represent 32% of a facility's annual emissions, while VAM represents the 68% of annual emissions.
On-Site Electricity Generation
Drained Cm can be used to fire internal combustion (IC) engines that drive generators to make electricity for
sale to the local power grid (EPA, 2011b). The quality of Cm required for use in power generation can be less than
that required for pipeline injection. IC engine generators can generate electricity using gas that has heat content as
low as 300 Btu/cf or about 30% Cm. Mines can use electricity generated from recovered Cm to meet their own on-
site electricity requirements and can also sell electricity generated in excess of on-site needs to utilities (EPA,
2008).
Coal mining is a very energy-intensive industry that could realize significant cost savings by generating its own
power. Nearly all equipment used in underground mining runs on electricity, including mining machines, conveyor
belts, ventilation fans, and elevators. While most of the equipment used in mining operations is used 250 days a
year for two shifts per day, ventilation systems are required to run continuously year-round. These systems require
large amounts of energy, up to 60% of a mine's total electricity usage. Total electricity demand can exceed 24 kWh
per ton of coal produced (EPA, 2008).
• Capital Cost: The cost for this option includes the cost of gas processing to remove gas contaminants
(primarily water vapor and solid particles), the electricity generation equipment, and power grid
connection equipment. Costs were assumed to be $l,300/kW. Assuming a 2 MW facility and a capacity
factor of 90%, total installed capital costs of electricity generation would be $2.7 million. Total installed
capital costs for this abatement measure would be $4.5 million, which includes the $850,000 for recovery,
assuming 20% owner's costs and 5% contingencies.
• Annual O&M Costs: The annual costs include $0.02/kWh for the engine-gen set in addition to the
$2.2/Mcf cost of the recovery system. Assuming a 600 Mcf/d flow and 90% capacity total O&M costs
would be approximately $0.8 million U.S. dollars.
• Annual Benefits: Offsets or cost savings associated with power that would have otherwise been
purchased at market electricity prices represent the annual benefits for this option. A 2 MW capacity
generation facility with a 90% capacity factor would be expected to generate approximately 16,000 MWh
5-6
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
of electricity. Assuming an energy price of $0.075/kWh14, this project would generate $1.2 million in
revenue from electricity sales.
• Technical Effectiveness: The analysis assumed a technical effectiveness of 28%, assuming a recovery
efficiency of 75% and destruction efficiency of 98% in the energy generation unit.
• Technical Lifetime: 15 years
On-Site Use for Process Heat
This category of abatement measures includes (1) recovery for use in the boiler for supporting in-mine heating
and (2) recovery for coal drying.
Mine Boilers
Drained Cm can be used to fuel on-site boilers that provide space or water heat to mine facilities. This analysis
assumes that existing boilers will be retrofitted to burn Cm and that the drained Cm is of sufficient quality to fuel
the mine's boiler and no additional gas processing is required.
• Capital Cost: The costs for this option are primarily associated with the capital cost to retrofit the mine
boiler to fire drained gas. The analysis assumes a 8.1 million British thermal units per hour (MMBTU/hr)
average boiler heat load and a retrofit cost of $7,500/MMBTU/hr. Assuming the mine boiler fuel demand
was 10 Mcf/hr, total installed capital costs for this abatement measure would be $635,000, which
includes $382,000 for the recovery system, $122,000 for boiler retrofit, and an additional 20% in owner's
costs and 5% for contingencies.
• Annual O&M Costs: The annual costs are the $2.4/Mcf to operate the recovery system. Assuming a 240
Mcf/d flow and 90% capacity, the total O&M costs would be approximately $213,000 USD.
• Annual Benefits: Benefits are the energy costs offset by using the drained Cm gas as a substitute fuel
source (offsetting coal consumption). Revenues associated with this project will be the quantity of coal
replaced at the mine mouth coal market price ($/MMBTU).
• Technical Effectiveness: The analysis assumed a recovery efficiency of 75% and destruction efficiency of
98% when combusted in mine boiler.
• Technical Lifetime: 15 years
Coal Drying
Another on-site direct use application for drained CMM is to use it as a fuel in thermal coal drying operations
at coal preparation plants collocated near an underground mine. The existing coal drying process can be retrofitted
to burn Cm as a supplemental fuel in addition to burning coal. Similar to the mine boiler option, we assumed the
CMM will not require further processing to serve as fuel to the dryer.
• Capital Cost: The cost of converting the dryer to burn CMM was assumed to be the same as the cost of
converting the boiler firing system [$7,500/MMBtu/hr], The analysis assumed an average coal drying rate
of 380 tons/hr (EPA, 1998). Assuming the average coal dryer heating requirement is 228 MMBTU/hr,
CMM gas with a lower heating value of 991 BTU/cf, the total installed capital costs for this abatement
measure would be $635,000, which includes $382,000 for the recovery system, $122,000 for boiler
retrofit, and an additional 20% in owner's costs and 5% for contingencies.
• Annual O&M Costs: The annual costs to operate the recovery system are assumed to be $2.4/Mcf.
Assuming a 240 Mcf/d flow and 90% capacity factor, total O&M costs would be approximately $213,000
USD.
14 The actual price utilities would be willing to pay will vary depending on the market and regulatory environment within the
specific country. In the absence of any additional market incentives, purchasers would likely only pay the price of generation in
the range of $0.025/kWh in the United States.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-7
-------
METHODOLOGY DOCUMENTATION
• Annual Benefits: Benefits are the energy costs offset by using the drained Cm gas as a substitute fuel
source (offsetting coal consumption). Revenues associated with this project will be the quantity of coal
replaced based on assumed energy content (MMBTU/ton) at the mine mouth coal market price
($/MMBTU).
• Technical Effectiveness: The analysis used a technical effectiveness of 28%, which assumes a recovery
efficiency of 75% and destruction efficiency of 98% when combusted in the mine boiler.
• Technical Lifetime: 15 years
Excess Gas Flaring
After recovering Cm using the drainage well technique, mines can choose to flare Cm of greater than 30%
concentration (EPA, 2011a). Flare systems considered include an open flare and an enclosed combustion system,
which consist of a mounted burner where the flame is exposed (open) or the flame is enclosed in a stack.
The costs of the flaring system consist of firing equipment and monitoring and metering equipment to validate
Cm destruction levels.
• Capital Cost: The cost of installing a flare system to burn CMM was assumed to be $280/Mcf/day.
Assuming an average daily flow rate of 600 Mcf gas, the total installed capital costs for this abatement
measure would be $1,239,000, which includes $850,000 for the recovery system, $134,000 for the flare
system, and an additional 20% in owner's costs and 5% for contingencies.
• Annual O&M Costs: The annual costs to operate the recovery system were assumed to be $2.4/Mcf.
Assuming a 600 Mcf/d flow and 90% capacity factor, total O&M costs would be approximately $489,000
USD.
• Annual Benefits: No revenues were associated with this option.
• Technical Effectiveness: The analysis used a technical effectiveness of 28%, which assumes a recovery
efficiency of 75% and destruction efficiency of 98% when combusted in a flaring system.
• Technical Lifetime: 15 years
Mitigation of VAM
Oxidation technologies (both thermal and catalytic) have the potential to use CH4 emitted from coal mine
ventilation air. Extremely low CH4 concentration levels (typically below 1%) make it economically infeasible to sell
this gas to a pipeline. However, thermal oxidizers can combust the VAM, converting it to H2O and CO2. VAM
oxidation is technically feasible at CH4 concentrations between 0.25% and 1.25%. For mines with lower VAM
concentrations, a supplemental gas is required to bring the concentration above the 0.25% concentration limit.
• Capital Cost: Abatement measure costs include the ductwork required to collect VAM exhaust from the
mine's ventilation system at the surface vents, the design and installation of a thermal oxidizer unit, and
any supporting auxiliary equipment. The total installed capital cost of the VAM oxidizer unit is $23 per unit
of recoverable ventilation air flow measured in cubic feet per minute [cfm]. Assuming the recoverable
ventilation air flow rate of 100 Mcfm and a CH4 concentration of 0.2%, capital costs would be $2.3 million
USD (=100,000 cfm X $23/cfm). The total installed capital costs for this abatement measure would be
approximately $2.8 million after assuming allowances of 20% in owner's costs and 5% for contingencies.
• Annual O&M Costs: Annual operating costs include costs to maintain the oxidizer unit, the electricity
required to operate the oxidizer blowers (0.075 kWh/mcf), and the periodic relocation costs of moving
equipment to the new location of a mine ventilation shaft (every 5 years at a cost of $4/cfm). Assuming a
100 Mcfm flow rate, total O&M costs would be approximately $462,000 USD, where VAM blower
electricity costs account for 55% of the annual costs, while oxidizer O&M costs represent 28%, and
annualized relocation costs make up the balance.
5-8
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Annual Benefits: Although low-grade heat can be captured from the VAM oxidizer, little economic benefit
can be obtained from it and only under special site-specific conditions; for this reason, we assumed this
abatement measure has no energy-related benefits.
• Technical Effectiveness: The analysis assumes a technical effectiveness of between 19% and 68%, which
assumes a recovery efficiency of 25% (in 2010) to 90% (in 2030) and destruction efficiency of 98% in a
VAM oxidation unit.
• Technical Lifetime: 15 years
Table 5-2 provides the specific factors used to calculate the technical effectiveness parameter for each
abatement measure in the coal mining sector. As discussed in Section 3, the technical effectiveness is the product
of facility applicability, technical feasibility, and reduction efficiency.
5.1.1.3 Sector-Level Trends/Considerations
Based on our review of existing abatement measures, technology improvements have the potential to reduce
the costs of VAM oxidation technology. Despite its abatement potential, VAM oxidation is the measure with the
highest abatement costs largely due to three key factors: (1) the equipment itself is large and costly; (2) there is no
revenue source; and (3) only a handful of technologies have been demonstrated at a commercial scale and, as
such, economies of scale in production have not been realized. The development of an international carbon
market like the UNFCCC's Clean Development Mechanism (CDM) or the Ell's Emissions Trading System (ETS) would
provide a source of revenue from the sale of carbon reduction credits. In addition, improvements in design and
catalysts have the potential to reduce the cost of oxidation over time. All other abatement measures described in
this section were assumed to be mature technologies.
5.1.1.4 References
Intergovernmental Panel on Climate Chang. 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.
Lixin, W. July 20, 2016. China Coal Industry Methane Control Policy and Planning. Presented at U.S.-China Methane
Emissions MRV and Emissions Control and Mitigation. Beijing, China.
Mine Safety and Health Administration. January 7, 2010. 2008 Coal Mine Data. Personal communication with Chad
Hancher at MSHA. methmine08nma.xls.
U.S. Energy Information Administration. 2017. International Energy Outlook 2017. Coal Overview. DOE/EIA-
0484(2017). Washington, DC: USEIA. Available online at
https://www. eia.gov/outlooks/ieo/pdf/0484f 2017). pdf
U.S. Energy Information Administration. 2018. International Energy Statistics Portal. Online Database. Washington,
DC: Energy Information Administration, U.S. Department of Energy. Available online at
www.eia.gov/cfapps/ipdbproiect/IEDIndex3.cfm
U.S. Environmental Protection Agency. 2008. (Revised January 2009). Identifying Opportunities for Methane
Recovery at U.S. Coal Mines: Profiles of Selected Gassy Underground Coal Mines 2002-2006. EPA 430-K-04-003.
U.S. EPA Coalbed Methane Outreach Program. Obtained January 7, 2010, at
http://www.epa.gov/cmop/docs/profiles 2008 final.pdf.
U.S. Environmental Protection Agency. 2011a. inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2009.
Annex 3 Methodological Descriptions for Additional Source or Sink Categories. EPA #430-R-ll-005.
Washington, DC: EPA. Available online at http://epa.gov/climatechange/emissions/usinventorvreport.html.
U.S. Environmental Protection Agency. 2011b. User's Manual for Coal Mine Methane Project Cash Flow Model
(Version 2). Washington, DC: EPA. Available online at http://www.epa.gov/
cmop/docs/cashflow user guide.pdf
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-9
-------
METHODOLOGY DOCUMENTATION
U.S. Environmental Protection Agency. November 1998. Use of Coal Mine Methane in Coal Dryers. EPA Coalbed
Methane Outreach Program Technical Options Series. EPA #6202. Washington, DC: EPA.
5-10 Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.1.2 Oil and Natural Gas Systems
Methane, or Cm, is the principal component of natural gas (95% of pipeline quality natural gas) and is emitted
from natural gas production, processing, transmission, and distribution. Oil production and processing upstream of
oil refineries can also emit Cm in significant quantities because natural gas is often found in conjunction with
petroleum deposits. In both ONG systems, Cm is a fugitive emission from leaking equipment, system upsets, and
deliberate flaring and venting at production fields, processing facilities, natural gas transmission lines and
compressor stations, natural gas storage facilities, and natural gas distribution lines. Figure 5-1 identifies the
facilities and equipment associated with the ONG system segments. Table 5-3 provides examples of the typical
facilities and equipment that comprise ONG systems.
Figure 5-1: Segments of Oil and Natural Gas Systems
~
Production & Processing
1. Drilling and Well Completion
2. Producing Wells
3. Gathering Lines
4. Gathering and Boosting Stations
5. Gas Processing Plant
Natural Gas
Transmission & Storage
6. Transmission Compressor Stations
7. Transmission Pipeline
8. Underground Storage
Distribution
9. Distribution Mains
10. Regulators and Meters for:
a. City Gate
b. Large Volume Customers
c. Residential Customers
d. Commercial Customer
Source: Adapted from American Gas Association (AGA) and Natural Gas STAR Program.
Table 5-3: Emission Sources from Oil and Natural Gas Systems
Segment
Facility
Equipment at the Facility
Production
Wells, central gathering facilities
Separators, pneumatic devices, chemical
injection pumps, dehydrators, compressors,
heaters, meters, pipelines, liquid storage tanks
Processing
Gas plants
Vessels, dehydrators, compressors, acid gas
removal units, heaters, pneumatic devices
Transmission
and storage
Transmission pipeline networks, compressor
stations, meter and pressure-regulating
stations, underground injection/withdrawal
facilities, liquefied natural gas (LNG) facilities
Vessels, compressors, pipelines,
meters/pressure regulators, pneumatic
devices, dehydrators, heaters
Distribution
Main and service pipeline networks, meter
and pressure-regulating stations
Pipelines, meters, pressure regulators,
pneumatic devices, customer meters
Global Non-CCk Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-11
-------
METHODOLOGY DOCUMENTATION
5.1.2.1 Oil and Natural Gas Systems Emission Projections
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite
Emission Projections, for additional information). Activity data for natural gas and oil systems included ONG
production, consumption, and refining statistics and forecasts from EIA (2018a; 2018b) and the BP Statistical
Review of World Energy (2017). Emissions from natural gas and oil systems were disaggregated into Gas
Production and Processing, Gas Transmission, Storage, and Distribution, Oil Production, and Oil Refining. The Tier 1
basic equation to estimate Cm emissions from ONG systems is as follows:
Total Fugitive CHa Emissions = (Annual Oil Production x Emission Factors + Annual Crude Oil Refinedx Emission
Factor) + (Annual Natural Gas Production x Emission Factors + Annual Natural Gas Consumption x Emission
Factors)
Emissions from ONG systems were disaggregated into the following categories:
• gas production and processing
• gas transmission, storage, and distribution
• oil production
• oil refining
These categories represent a combination of IPCC emission factors for each of the given segments. IPCC
factors determine emission estimates based on specific drivers such as total gas or oil production within a country.
Because emission factors are held constant, the driver for determining fugitive Cm emissions from ONG is the
respective production and consumption of these fuels.
Activity Data
Historical
• The EPA obtained historical ONG production and consumption data from EIA (2018a; 2018b) for 1990
through 2015. For refineries, the EPA obtained refinery throughput data from the BP Statistical Review of
World Energy (BP, 2017).
Projected
• Projections of ONG production and consumption were available from the EIA International Energy
Outlook (EIA, 2017). The EPA used growth rates as provided by EIA Reference Case projections for each
year from 2014 through 2050. These growth rates were available by country or region. When available by
region, the EPA assumed all countries within that region have that region's growth rate.
Emission Factors
Historical and Projected
• The EPA used 2006 IPCC Guidelines default factors for natural gas production (IPCC, 2006) (which includes
both natural gas production and processing), natural gas consumption (which includes natural gas
transmission, distribution, and storage), oil production, oil refining, and venting and flaring for 1990,
1995, 2000, 2005, and 2010 emissions. When the 2006 IPCC Guidelines provided only ranges for emission
factors (as opposed to central estimates), the midpoint of the range was used.
• The EPA multiplied appropriate ONG production, consumption, and refining statistics for 1990,1995,
2000, 2005, and 2010 by IPCC (IPCC, 2006) default factors.
5-12
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
If emissions were not reported and EIA production data were not available for a country, the EPA assumed zero
emissions for this source.
Each emission estimate disaggregation relied on a composite IPCC factor and an applicable activity driver. To
illustrate which emission factors were combined and used to drive emissions, Table 5-4 shows the considered
factors per IPCC's definition for each disaggregation of results (IPCC, 2006). Emissions were determined using this
combination (typically a weighted total) of various segments within the ONG industry. Each segment factor used is
representative of the varying pertinent fugitive and venting emission sources per IPCC's methodology. For
instance, associated gas emissions were considered in the oil production venting estimate.
Table 5-4: IPCC Emission Factors by Disaggregation
Disaggregation
Relevant IPCC Factor
Activity Driver Used
Gas production and processing
Gas production, fugitives; gas processing, default
weighted total fugitives and flaring
Gas production
Gas transmission, storage, and
distribution
Gas transmission, fugitives and venting; gas
storage, all; gas distribution, all
Gas consumption
Oil production3
Oil production, default weighted total fugitives,
venting, and flaring; well drilling, testing, and
servicing, venting and flaring
Oil production
Oil refining
Oil refining, all
Oil throughput
a Includes associated gas emissions.
Table 5-5 shows the aggregate emission factor value used for each of the above emission categories. For
clarity, units are given according to IPCC's definition. The factors shown in the table represent A1 (developed)
countries, whereas IPCC gives comparable factors with higher estimates for Non-Al countries. Both country
designation emissions were developed using the same combination of representative IPCC factors within each
segment as shown in Table 5-4. The highest uncertainties are expected to be related to venting and flaring
estimates because present infrastructure and practices can vary significantly by country (Soltanieh, 2016).
Because the developed emission factors for
gas and oil production are so much higher in
magnitude than the gas-consumed and crude-
refined factors that use similar drivers,
respectively, trends in emissions are implicitly
more focused on those drivers. Further, oil
production is universally the highest estimated
segment using this methodology. For example,
based on the application of this methodology, in 2015, oil production comprised 92% of all emissions in Saudi
Arabia with 6% attributed to gas production. In contrast, the IPCC crude-refined factor is so low that emission
estimates in this segment have virtually no overall impact on results.
Emission Reductions in Baseline Scenario
The methodology used for this source category does not explicitly model any emission reductions; however,
emission reductions are included to the extent they are reflected in country-reported data.
Uncertainties
The greatest uncertainties are due to the use of default emission factors and difficulties in projecting ONG
consumption and production through 2050 for rapidly changing global economies such as those in the former
Soviet Union (FSU) and developing Asia. The emission factors provided in the 2006 IPCC Guidelines have a wide
Table 5-5: Aggregate IPCC Emission Factor by Driver
Driver
Aggregate A1 IPCC Factor
Gas produced
0.001932 Gg/MMm3
Gas consumed
0.001580 Gg/MMm3
Oil produced
0.011115 Gg/km3
Crude refined
0.000022 Gg/km3
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-13
-------
METHODOLOGY DOCUMENTATION
range in values, and the midpoint of the range may not be representative of country-level emission estimates.
There could be some minor overlap in emissions between the given disaggregation of categories. The factors used
were also developed in a year prior to this study and may not exactly represent current conditions in a complex
and quickly changing industry. In addition, Cm emissions from ONG systems are not linearly related to throughput,
so the IPCC Tier 1 methodology and emission factors can lead to overestimates.
While the Tier 1 oil production emission estimates derived from IPCC factors are substantially higher than
other segments, this could be due to the nature of how the factors themselves were developed. Suggesting that all
emissions in these categories represent each segment exclusively, as stated above, may not be accurate. As
defined, oil production contains emission estimates from well testing and drilling because IPCC factors are in terms
of total oil production. However, this source may also contain emissions that could be considered gas production.
Given this, there is a possibility of overlap between the emission categories.
IPCC also provides multiple oil production emission factors based on varying resource types such as
conventional oil, heavy oil/bitumen, and synthetic crude. The factor used for this segment is based on the given
default IPCC weighted factor. However, these varying resource types may not be applicable to most countries
because conventional oil is the most widely processed resource type. Further, within the conventional oil resource
type, the onshore fugitive emission factor for oil production has a large variance. Any choice of IPCC factor to use
can have an evident impact on the disaggregation of results.
In addition to the differences in magnitude between segments, emission estimates through this methodology
are affected by the range that IPCC publishes for certain factors as mentioned for oil production fugitives above.
While certain factors may exhibit one value, others can range, providing a drastically different Tier 1 estimate.
These ranges can be associated with various factors such as offshore versus onshore activity or the use of
reciprocating versus centrifugal compressors. For instance, in the gas production fugitives factor, onshore
emissions represent the higher range value, whereas offshore activities are associated with the lower range.
Given this variation, further disaggregation and improvement could be done given the appropriate data. Data
improvements could include onshore and offshore production values for each country with separate Tier 1
estimates for each country in these sectors using the appropriate factor. Further, typical usage of compressor
types in certain countries (reciprocating/centrifugal) could illustrate which factor should be used with gas
consumption drivers. However, because of the lack of these data, the EPA focused primarily on the average of
these ranges to generate results. Given the vast range of global oil and gas industry conditions and the inherent
uncertainty included in using a single emission factor for overall segments, the average range of these factors may
represent the best available estimate. Further, the EPA also reviewed results using only the high range, low range,
and a geometric average when available to investigate the impact that different factors have on overall emission
estimates and to consider alternatives.
5.1.2.2 Oil and Natural Gas Systems Mitigation Options Considered
Within the four segments of ONG systems, a number of abatement measures can be applied to mitigate CH4
losses from activities associated with or directly from the operation of equipment and components. The
abatement measures, such as inspection and maintenance programs for leaks or equipment retrofits or
modifications, may be applied to ONG processes and equipment, including compressors/engines, dehydrators,
pneumatics/controls, pipelines, storage tanks, and wells.
Abatement measures are available to mitigate CH4 losses from activities associated with or directly from the
operation of equipment components common across the ONG system segments of production, processing,
transmission, and distribution. These abatement options in the ONG system segments generally fall into three
categories: equipment modifications/upgrades, changes in operational and maintenance practices including
directed inspection and maintenance, and installation of new equipment. ONG industry-related voluntary
5-14
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
programs such as the Global Methane Initiative (GMI) and the EPA's Natural Gas STAR Program, which are aimed
at identifying cost-effective Cm emission reduction opportunities, have developed a well-documented catalog of
potential Cm abatement measures that are applicable across the segments of the ONG system. Abatement
measures documented by the EPA's Natural Gas STAR Program serve as the basis for estimating the costs of
abatement measures used in this analysis. It is important to note that although abatement measures identified by
the Natural Gas STAR Program are cited as cost-effective based on industry partner-reported experiences, the
abatement measure's cost-effectiveness is determined by the component's emission rate and the value of energy
recovered. This analysis used average emission factors when estimating the break-even prices for each measure. In
many cases, these average emission rates are lower than the case study examples cited in the Natural Gas STAR
Program's documentation. As a result, abatement measures cited as cost-effective by the Natural Gas STAR
Program's partners may not necessarily be the lowest cost options in the MAC analysis.
This section discusses the abatement measures considered for this analysis and presents the costs, benefits,
technical applicability, reductions efficiency, and the expected technology lifetime of each measure. The
abatement measures presented in Table 5-6 provide an overview of the options considered in each segment of the
oil and gas sector.
Table 5-6: Abatement Measures Applied in Oil and Gas Production Segments
Abatement Measure
Component
Total
Installed
Capital Cost
($2008)
Annual
O&M
($2008)
Time
Horizon
Technical
Effectiveness3
Gas Production Segment
Directed inspection & maintenance
at gas production facilities
Chemical
injection pumps
—
6,675
1
40%
Installing surge vessels for
capturing blowdown vents
Compressor
blowdowns
158,940
28,078
15
50%
Installing electronic starters on
production field compressors
Compressor
starts
2,649
5,849
10
75%
Directed inspection & maintenance
at gas production facilities
Deepwater gas
platforms
—
50,000
1
95%
Install flash tank separators on
dehydrators
Dehydrator
vents
6,540
—
5
30% to 60%
Optimize glycol circulation rates in
dehydrators
Dehydrator
vents
—
15
1
33% to 67%
Installing catalytic converters on
gas fueled engines and turbines
Gas engines-
exhaust vented
7,924
4,374
10
56%
Installing plunger lift systems in gas
wells
Gas well
workovers
5,646
(13,855)
5
80%
Replace gas-assisted glycol pumps
with electric pumps
Kimray pumps
2,788
1,949
10
100%
Directed inspection & maintenance
at gas production facilities
Nonassociated
gas wells
817
1
95%
(continued)
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-15
-------
METHODOLOGY DOCUMENTATION
Table 5-6: Abatement Measures Applied in Oil and Gas Production Segments (continued)
Abatement Measure
Component
Total
Installed
Capital Cost
($2008)
Annual
O&M
($2008)
Time
Horizon
Technical
Effectiveness3
Gas Production Segment (continued)
Installing plunger lift systems in gas
wells
Nonassociated
gas wells
5,646
(13,855)
5
80%
Directed inspection & maintenance
on offshore oil platforms
Offshore
platforms,
deepwater oil,
fugitive, vented
and combusted
50,000
1
43%
Flaring instead of venting on
offshore oil platforms
Offshore
platforms,
shallow water
oil, fugitive,
vented and
combusted
165,888,859
4,976,666
15
98%
Installing vapor recovery units on
storage tanks
Oil tanks
473,783
161,507
15
58%
Using pipeline pump-down
techniques to lower gas line
pressure before maintenance
Pipeline
blowdown (BD)
1,352
1
90%
Directed inspection & maintenance
at gas production facilities
Pipeline leaks
—
82
1
60%
Convert gas pneumatic controls to
instrument Air
Pneumatic
device vents
72,311
24,321
10
50% to 90%
Replacing high-bleed pneumatic
devices in the natural gas industry
Pneumatic
device vents
165
—
10
8% to 17%
Directed inspection & maintenance
at gas production facilities
Shallow water
gas platforms
—
33,333
1
95%
Reduced emission completions for
hydraulically fractured natural gas
wells
Unconventional
gas well
completions
30,038
1
90%
Reduced emission completions for
hydraulically fractured natural gas
wells
Unconventional
gas well
workovers
30,039
1
90%
Installing surge vessels for capturing
blowdown vents
Vessel BD
158,940
28,078
15
50%
Installing plunger lift systems in gas
wells
Well clean-ups
(liquid
petroleum gas
wells)
5,646
(13,855)
5
40%
(continued)
5-16
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-6: Abatement Measures Applied in Oil and Gas Production Segments (continued)
Abatement Measure
Component
Total
Installed
Capital Cost
($2008)
Annual
O&M
($2008)
Time
Horizon
Technical
Effectiveness3
Gas Processing Segment
Installing surge vessels for capturing
blowdown vents
Blowdowns/
Venting
158,940
28,078
15
50%
Directed inspection & maintenance
at processing plants and booster
stations—compressors
Centrifugal
compressors
(dry seals)
15,581
1
12%
Directed inspection & maintenance
at processing plants and booster
stations—compressors
Centrifugal
compressors
(wet seals)
6,131
1
12%
Replacing wet seals with dry seals in
centrifugal compressors
Centrifugal
compressors
(wet seals)
380,804
(102,803)
5
66%
Installing catalytic converters on gas
fueled engines and turbines
Gas engines-
exhaust vented
7,924
4,374
10
56%
Replace gas-assisted glycol pumps
with electric pumps
Kimray pumps
2,788
1,949
10
100%
Directed inspection & maintenance
at processing plants and booster
stations
Plants
10,134
5
95%
Directed inspection & maintenance
at processing plants and booster
stations—compressors
Recip.
compressors
6,131
1
10%
Early replacement of reciprocating
compressor rod packing rings
Recip.
compressors
7,800
0
5
1%
Fuel gas retrofit for BD valve—take
recip. compressors offline
Recip.
compressors
2,365
—
5
21%
Reciprocating compressor rod
packing (static-pac)
Recip.
compressors
5,696
—
5
0%
Transmission Segment
Directed inspection & maintenance
at compressor stations-
compressors
Centrifugal
compressors
(dry seals)
15,581
1
13% to 14%
Replacing wet seals with dry seals in
centrifugal compressors
Centrifugal
compressors
(wet seals)
380,804
(102,803)
5
71% to 77%
Install flash tank separators on
dehydrators
Dehydrator
vents
9,504
—
5
67%
Optimize glycol circulation rates in
dehydrators
Dehydrator
vents
15
1
67%
(continued)
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-17
-------
METHODOLOGY DOCUMENTATION
Table 5-6: Abatement Measures Applied in Oil and Gas Production Segments (continued)
Abatement Measure
Component
Total
Installed
Capital Cost
($2008)
Annual
O&M
($2008)
Time
Horizon
Technical
Effectiveness3
Transmission Segment (continued)
Installing catalytic converters on gas
fueled engines and turbines
Engine/turbine
exhaust vented
7,924
4,374
10
56%
Directed inspection & maintenance
at gate stations and surface facilities
M&R (Trans.
Co.
Interconnect)
1,741
1
72%
Directed inspection & maintenance
on transmission pipelines
Pipeline leaks
—
41
1
60%
Using pipeline pump-down
techniques to lower gas line
pressure before maintenance
Pipeline venting
1,352
1
90%
Convert gas pneumatic controls to
instrument air
Pneumatic
devices
72,311
24,321
10
50% to 90%
Replacing high-bleed pneumatic
devices in the natural gas industry
Pneumatic
devices
165
—
10
8% to 17%
Directed inspection & maintenance
at compressor stations-
compressors
Recip
compressor
15,581
1
10% to 12%
Early replacement of reciprocating
compressor rod packing rings
Recip
compressor
7,800
—
5
1%
Early replacement of reciprocating
compressor rod packing rings and
rods
Recip
compressor
41,068
5
1% to 74%
Fuel gas retrofit for BD valve—take
recip. compressors offline
Recip
compressor
2,365
—
5
36% to 39%
Reciprocating compressor rod
packing (static-pac)
Recip
compressor
5,696
—
5
6% to 9%
Installing surge vessels for capturing
blowdown vents
Station venting
158,940
28,078
15
50%
Directed inspection & maintenance
at compressor stations
Stations
—
1,398
1
85%
Directed inspection & maintenance
at gas storage wells
Wells (storage)
—
651
1
95%
Distribution Segment
Directed inspection & maintenance
at gate stations and surface facilities
M&R <100
—
1,604
1
30% to 80%
Replace cast iron pipeline
Mains—cast
iron
373,633
182
5
95%
(continued)
5-18
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-6: Abatement Measures Applied in Oil and Gas Production Segments (continued)
Abatement Measure
Component
Total
Installed
Capital Cost
($2008)
Annual
O&M
($2008)
Time
Horizon
Technical
Effectiveness3
Distribution Segment (continued)
Replace unprotected steel pipeline
Mains-
unprotected
steel
373,633
182
5
95%
Replace unprotected steel service
lines
Services-
unprotected
steel
418,023
311
5
95%
a Technical effectiveness reflects the percentage reduction achievable from implementing the abatement measure considering
the presence of complementary options. Technical effectiveness is the product of three separate factors—the reduction
efficiency, technical applicability, and market penetration.
b Lower technical effectiveness is due to limited applicability at liquid petroleum gas wells.
5.1.2.3 Technical and Economic Characteristics Summary
The MAC analysis approach consisted of four sequential steps. Step 1 was to assess the sectoral trends, which
entailed reviewing recent international energy statistics for oil and gas. The second step was to develop source-
level emission estimates that could be used to build different model ONG systems. These model systems reflect
country-specific variations in production process techniques, level of maintenance, and vintage of the existing
infrastructure. Step 3 was to estimate country-specific abatement costs and benefits based on the relative cost
factors for labor, energy, and nonenergy inputs. Step 4 was to compute the break-even prices for each country-
specific abatement measure.
Defining International Model Facilities for the Analysis
For this analysis, we developed model ONG systems for each segment based initially on the EPA ONG system
emission inventory. Scaling factors were developed based on country-specific activity factors developed from the
international statistics illustrated in Error! Reference source not found.. Where reliable data were available, we
made international adjustments to reflect specific country systems. For countries for which data were not
available, this analysis assumed the oil and gas system was similar to that in the United States in terms of the
distribution of emissions (total BAU emissions for each country are exogenous to the MAC model obtained from
EPA [2012a]). The relative international factor was multiplied by the percentage share of U.S. oil and gas Cm
emission inventory at the segment/component source level (e.g., compressors, valves, connections, pneumatic
devices). The resulting Technical Applicability factor was used to allocate a fraction of the national baseline
emissions to each component source in the ONG inventory (e.g., wells, tanks, compressors, valves).
Multiplying the technical applicability factor by the baseline emissions yields the subset of emissions available
for reductions from each component source and abatement measure. The technical applicability factor comprises
two parts. The U.S. 2010 GHG emission inventory serves as the basis for the distribution of emissions across the
constituent components (see EPA, 2012b, Annex 3). The second component of the technical applicability factor is
the country- and segment-specific relative activity factor (e.g., total oil production, gross natural gas production).
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-19
-------
METHODOLOGY DOCUMENTATION
Estimating Abatement Project Costs and Benefits
The analysis begins with technology costs for the United States as reported in the EPA Lessons Learned
documentation. We applied the Nelson-Farrar15 Oil Field and Refinery Operation cost indices to convert from
reported-year costs to 2008 dollars (USD) for capital and O&M costs, respectively. Next, we applied the country-
specific relative price factors for labor, energy, and nonenergy components of annual costs and benefits. This final
step yielded country-specific costs and benefits used to compute the break-even price for each abatement
measure. Abatement measure costs and technical efficiencies were applied to estimate the break-even prices.
Table 5-7 presents the break-even prices for selected ONG system abatement measures for the United States in
2010. Break-even prices for other countries differ based on differences in ONG operational spatial distribution and
production technologies, reservoir types, and differences in input prices for mitigation technologies. For this
analysis, we used the abatement measure costs, revenue, and reduction efficiency as described in the previous
section to estimate the break-even price for each abatement measure.
The first step is to estimate the reduced emissions on a per-unit basis for each technology. This value is
calculated by multiplying the abatement measure's technical efficiency by the annual emissions per unit of the
component or process to which the abatement measure is being applied. The resulting annual reduced emissions
served as the denominator in the break-even price calculation.
In Table 5-7, we present abatement cost and revenues per metric ton of CO2 equivalent (tC02e) reduced for
the abatement measures with the largest national emission reductions. Costs include the annualized total installed
capital cost and annual O&M costs. Offsetting these costs are the annual revenue in terms of gas savings and the
tax benefit of depreciation. The break-even prices reported in Table 5-7 were calculated by subtracting the annual
revenues from the annualized costs.
5.1.2.4 Sector-Level Trends/Considerations
The objective in assessing the sectoral trends is to understand how emissions differ across countries and how
they vary over time. Assessing trends not only considers aggregate growth or decline in emissions but also any
potential shift in sector emissions across the oil and gas segments. To this end, we reviewed the current
international oil and gas industry activity data for 2010. Statistics reviewed included gross natural gas production,
oil production, LNG imports, and gas processing throughput (EIA, 2011; Oil & Gas Journal, 2011). In the absence of
real infrastructure data, these statistics provide insights on the relative importance of segments internationally.
Error! Reference source not found, presents these key statistics for the 10 largest emitting countries in 2010.
15 Nelson-Farrar Annual Cost Indices are available in the first issue of each quarter of the Oil and Gas Journal.
5-20
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-7: Example Break-Even Price Calculation based on 2010 MAC for the United States
Abatement Measure
System
Component/
Process
Reduced
Emissions
per Unit
(tC02e)
Annualized
Capital
Costs
($/tC02e)
Annual
Cost
($/tC02e)
Annual
Revenue
($/tC02e)
Tax Benefit
of
Depreciation
($/tC02e)
Break-
even
Price
($/tC02e)
National
Incremental
Reductions
(MtC02e)
Production
Convert gas pneumatic Pneumatic
controls to instrument air device vents
71.0
$335.68
$441.41
$10.01
$82.50
$684.58
15.29
Reduced emission
completions for
hydraulically fractured
natural gas wells
Unconventional
gas well
completions
2,703.96
$0.00
$11.11
$10.01
$0.00
$1.10
8.82
Replacing high-bleed
pneumatic devices in the
natural gas industry
Pneumatic
device vents
9.7
$7.38
$0.00
$10.01
$1.81
-$4.44
2.30
Processing
Directed inspection &
maintenance at
processing plants and
booster stations
Plants
1,109.0
$0.00
$9.14
$10.01
$0.00
-$0.87
0.50
Fuel gas retrofit for bd
valve—take recip.
compressors offline
Recip.
compressors
351.9
$2.96
$0.00
$10.01
$0.90
-$7.95
1.34
Replacing wet seals with
dry seals in centrifugal
compressors
Centrifugal
compressors
(wet seals)
5,000.8
$33.48
-$20.56
$10.01
$10.15
-$7.24
2.53
Transmission
Convert gas pneumatic Pneumatic
controls to instrument air devices
89.9
$2,898.32
$3,811.28
$10.01
$712.36
$5,987.24
2.88
Directed inspection &
maintenance at
compressor stations
Stations
3,655.9
$0.00
$0.41
$10.01
$0.00
-$9.60
6.61
Fuel gas retrofit for bd
valve—take recip.
compressors offline
Reciprocating
compressor
1,014.8
$1.07
$0.00
$10.01
$0.32
-$9.26
5.65
Distribution
Directed inspection &
maintenance at gate
stations and surface
facilities
M&R >300
511.6
$0.00
$3.40
$10.01
$0.00
-$6.60
1.58
Directed inspection &
maintenance at gate
stations and surface
facilities
M&R 100-300
220.2
$0.00
$7.90
$10.01
$0.00
-$2.10
2.48
Replace cast iron pipeline Mains—cast
iron
91.7
$1,790.73
$1.99
$10.01
$543.06
$1,239.65
2.54
Note: Break-even price assumed a 10% discount rate and a 40% tax rate. Annual energy benefits were based on a natural gas
price of $4/Mcf.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-21
-------
METHODOLOGY DOCUMENTATION
Table 5-8: International Statistics on Key Activity Drivers: 2010
Country
2010 Emissions
(MtCOze)
Dry Natural Gas
Production3
(Bcf/year)
Crude Oil
Production13
(Mbbl/day)
Gas Processing
Plant
Th rough putc
(MMcfd)
Gas
Transmission
Pipelines'* (km)
Russia
332.0
22,965
10,146
926
160,952
United States
247.8
26,858
9,688
45,808
548,665
Kuwait
106.0
422
2,450
1,034
269
Iraq
94.1
596
2,408
1,550
3,365
Angola
84.9
364
1,988
137
—
Uzbekistan
84.7
2,123
105
NA
10,253
Libya
77.4
1,069
1,789
2,567
—
Canada
53.3
6,695
3,483
29,154
75,835
Iran
47.2
7,774
4,252
10,509
20,725
Venezuela
30.2
2,510
2,375
3,555
5,347
a EIA. 2018a. International Energy Statistics: Gross Natural Gas Production.
b EIA. 2018b. International Energy Statistics: Total Oil Supply.
c Oil & Gas Journal [OGJ]. June 6, 2011. Worldwide Processing Survey.
d CIA. 2011. The World Factbook.
e EIA. 2012. Country analysis brief-Uzbekistan. http://www.eia.gov/countries/cab.cfm?fips=UZ
Although differences in annual production and throughput provide some indication of the size of a country's
ONG system, considerations of age and the condition of the infrastructure are major factors in determining the
rate of source-level emissions and in turn the abatement potential associated with each abatement measure. In
general, countries with aging infrastructure will have "leakier" components and in turn have a greater abatement
potential. Conversely, countries with newly developed infrastructure will have less abatement potential.
Another important trend to consider is the expansion of unconventional gas (shale gas) production. The
growth in unconventional gas production (e.g., United States, Canada, and China) is likely to result in an increased
frequency of hydraulically fractured gas well completions and related workovers. In the absence of any regulatory
or voluntary actions to reduce emissions from these sources, this trend suggests that the gas production segment
will represent an even greater proportion of these nations' baseline emissions over time.
Table 5-9 shows the allocation of baseline emissions to the five segments of the ONG system. These
percentages determine the distribution of emissions over the production supply chain.
Table 5-9: Allocation of Baseline Emissions to the Five Segments of the ONG System
Regions
OIL
GAS_PRODUCTION
GAS_PROCESSING
GAS_TRANSMISSION
GAS_DIST
Africa
0%
69%
10%
17%
3%
Asia
1%
42%
35%
14%
8%
Central and
South America
1%
56%
10%
25%
8%
Eurasia
0%
40%
0%
47%
13%
Europe
2%
43%
0%
33%
21%
Middle East
1%
63%
8%
21%
7%
North America
5%
60%
9%
17%
8%
5-22
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.1.2.5 References
BP. 2017. Statistical Review of World Energy 2017. Available online at
http://www.bp.com/en/global/corporate/energv-economics/statistical-review-of-world-energv.html of
Intergovernmental Panel on Climate Change. 2006. 2006IPCC Guidelines for National Greenhouse Gas Inventories:
Volume 2 Energy. Chapter 4 Fugitive Emissions. Available online at http://www.ipcc-
nggip.iges.or.ip/public/2006gl/vol2.html J
Oil & Gas Journal [OGJ]. June 6, 2011. Worldwide Processing Survey.
Soltanieh. 2016. A Review of Global Gas Flaring and Venting and Impact on the Environment. International Journal
of Greenhouse Gas Control. Available online at https://www.researchgate.net/publication/299545498 ef
U.S. Energy Information Administration. 2011. World Shale Gas Resources: An Initial Assessment of 14 Regions
Outside the United States. Washington, DC: EIA. Available online at
http://www.eia.gov/analvsis/studies/worldshalegas/
U.S. Energy Information Administration. 2017. International Energy Outlook 2017. DOE/EIA—0484(2017).
Washington, DC: EIA. Available online at https://www.eia.gov/outlooks/ieo/pdf/0484f2017).pdf
U.S. Energy Information Administration. 2018a. International Energy Statistics: Gross Natural Gas Production.
Washington, DC: EIA. Available online at
http://tonto.eia.doe.gov/cfapps/ipdbproiect/IEDIndex3.cfm?tid=3&pid=3&aid=l
U.S. Energy Information Administration. 2018b. International Energy Statistics: Total Oil Supply. IEA: Washington,
DC: EIA. Available online at http://www.eia.gov/cfapps/ipdbproiect/IEDIndex3.cfm?tid=5&pid=53&aid=l
U.S. Environmental Protection Agency. 2012a. Global Anthropogenic Non-C02 Greenhouse Gas Emissions: 1990-
2030. EPA #430-R-12-006. Washington, DC: EPA. Available online at
http://www.epa.gov/climatechange/economics/international.html
U.S. Environmental Protection Agency. 2012b. U.S. Greenhouse Gas Inventory Report: Inventory of U.S. Greenhouse
Gas Emissions and Sinks: 1990-2010. EPA#430-R-12-001. Washington, DC: EPA. Available online at
http://epa.gov/climatechange/emissions/usinventorvreport.html
Natural Gas Star Program Lessons Learned Documents Referenced in Model
U.S. Environmental Protection Agency. October 2006. Reducing Methane Emissions from Compressor Rod Packing
Systems: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Replacing Wet Seals with Dry Seals in Centrifugal
Compressors: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Reducing Emissions When Taking Compressors Offline:
Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Replacing Glycol Dehydrators with Desiccant Dehydrators:
Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Optimize Glycol Circulation and Install Flash Tank Separators
in Dehydrators: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2003. Directed Inspection and Maintenance at Compressor
Stations: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-23
-------
METHODOLOGY DOCUMENTATION
U.S. Environmental Protection Agency. October 2003. Directed Inspection and Maintenance at Gate Stations and
Surface Facilities: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. O Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2003. Directed Inspection and Maintenance at Gas Processing
Plants and Booster Stations: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. Available
online at http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Replacing Gas-Assisted Glycol Pumps with Electric Pumps:
Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Using Pipeline Pump-Down Techniques to Lower Gas Line
Pressure Before Maintenance: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA.
Available online at http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Composite Wrap for Non-Leaking Pipeline Defects: Lessons
Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Using Hot Taps for In Service Pipeline Connections: Lessons
Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Convert Gas Pneumatic Controls to Instrument Air: Lessons
Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Options for Reducing Methane Emissions from Pneumatic
Devices in the Natural Gas Industry: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA.
Available online at http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. (2006a, October). Installing Vapor Recovery Units on Storage Tanks. Lessons
Learned from Natural Gas STAR Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. 2011. Reduced Emission Completions for Hydraulically Fractured Natural
Gas Wells: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. October 2006. Installing Plunger Lift Systems in Gas Wells: Lessons Learned
from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. 2011. Options for Removing Accumulated Fluid and Improving Flow in Gas
Wells: Lessons Learned from Natural Gas Star Partners. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
U.S. Environmental Protection Agency. 2011. Install Electric Motor Starters: Partner Reported Opportunities (PROs)
for Reducing Methane Emissions. PRO Fact Sheet No. 105. Washington, DC: EPA. Available online at
http://www.epa.gov/gasstar/tools/recommended.html
5-24
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.1.3 Stationary and Mobile Combustion
Stationary and mobile combustion consists of Cm and N2O emissions from the combustion of fossil fuels in
vehicles; power plants; and residential, commercial, and industrial stationary sources.
5.1.3.1 Stationary and Mobile Combustion Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite
Emission Projections, for additional information). Activity data for stationary and mobile combustion included coal,
oil and gas consumption from EIA (2018) and projected fuel consumption data from EIA (2017). The Tier 1 basic
equation to estimate Cm and N2O emissions from stationary and mobile combustion is as follows:
CH4 Emissions = Annual Fuel Consumption (by fuel type) x Emission Factor (by fuel type) (5.1)
N20 Emissions = Annual Fuel Consumption (by fuel type) x Emission Factor (by fuel type) (5.2)
The driving factors for determining emissions from stationary and mobile combustion are activity data (fuel
consumption) and activity data growth rates.
Activity Data
Historical
• Fuel consumption data by fuel type (coal, oil, and gas) were obtained from ElA's International Energy
Statistics Database (EIA, 2018). While the database contained full time-series data in mass units for 1990
through 2014, data in energy units were accurate only for 2010 through 2014. Thus, the trend from the
mass data was used to back cast the energy data for 1990 through 2010.
Projected
• Projected fuel consumption data by fuel type (coal, oil, and gas) were obtained from ElA's International
Energy Outlook database (EIA, 2017). The data were then converted to annual growth rates, which were
used to project the historical data. The data were provided through 2040, so the 2040 growth rate was
assumed to be constant from 2041 through 2050. Projected fuel consumption data were available at the
country level for some developed countries; for all other countries, a regional growth rate was applied.
Emission Factors
Historical and Projected
• Tier 1 emission factors were obtained from IPCC's 2006IPCC Guidelines. Because the activity data are
broken out by fuel type and the IPCC emission factors are broken out by IPCC sector, the emission factor
for each fuel type was mapped to its closest matching IPCC sector (e.g., the emission factor for jet fuel
was mapped to the IPCC's Aviation sector). The same emission factors were applied for each country.
Emission Reductions in Baseline Scenario
The methodology used for this source category does not explicitly model any emission reductions; however,
emission reductions are included to the extent they are reflected in country-reported data.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-25
-------
METHODOLOGY DOCUMENTATION
Uncertainty
Uncertainties in the emission estimates include differences between regional- and country-level growth rates,
assumptions that activity data growth rates from 2041 through 2050 will equal that of 2040, and the IPCC Tier 1
default emission factors. For stationary combustion sources, this high degree of uncertainty is a result of the lack
of relevant measurements, uncertainties in measurements, or an insufficient understanding of the emission
generating process (IPCC, 2006). The IPCC Good Practice Guidance and Uncertainty Management in National
Greenhouse Gas Inventories (IPCC, 2000) estimates uncertainty for the stationary Cm combustion emission factors
at ±50 to 150%. Stationary combustion N2O combustion emission factors are highly uncertain due to limited
testing data on which the factors are based. In addition, the use of uncontrolled stationary IPCC default emission
factors may overestimate emissions in those developing countries that have adopted some level of emission
control strategies for combustion sources.
Uncertainty in N2O and Cm emission factors for mobile combustion depends on a number of factors, including
uncertainties in fuel composition, fleet age distribution and other vehicle characteristics, and maintenance
patterns of the vehicle stock, to name a few (IPCC, 2000).
5.1.3.2 Stationary and Mobile Combustion Mitigation Methodology
The EPA has not estimated mitigation potential from stationary and mobile combustion because of the lack of
available data on mitigation options.
5.1.3.3 References
Intergovernmental Panel on Climate Change. 2000. Good Practice Guidance and Uncertainty Management in
National Greenhouse Gas Inventories. IPCC-XVI/Doc.10 (1.IV.2000). Montreal, Canada: Intergovernmental
Panel on Climate Change, National Greenhouse Gas Inventories Programme.
Intergovernmental Panel on Climate Change. 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. Energy Information Administration. 2017. International Energy Outlook 2017. Report# DOE/EIA-0484(2017).
Washington, DC: Energy Information Administration, U.S. Department of Energy. Available online at
https://www. eia.gov/outlooks/ieo/pdf/0484f 2017). pdf
U.S. Energy Information Administration. 2018. International Energy Statistics Portal. Washington, DC: Energy
Information Administration, U.S. Department of Energy. Online Database. Available online at
www.eia.gov/cfapps/ipdbproiect/IEDIndex3.cfm
5-26
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.1.4 Biomass Combustion
Biomass combustion consists of Cm and N2O emissions from the incomplete combustion of biofuels, wood,
and charcoal.
5.1.4.1 Biomass Combustion Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3 Generating the Composite
Emission Projections, for additional information). Activity data for biomass combustion included biofuel
consumption from EIA (2018) with growth rates from EIA (2017) and charcoal and wood fuel consumption from
FAO (2016) with growth rates from FAO (2010). Emissions from the biomass combustion source category were
disaggregated according to emissions from biofuel consumption and emissions from wood fuel and charcoal
consumption.
The Tier 1 basic equation to estimate Cm and N2O emissions from biomass combustion is as follows:
CH4 Emissions = Fuel Consumption (by fuel type) x CH4 Emission Factor (by fuel type) (5.3)
N20 Emissions = Fuel Consumption (by fuel type) xN20 Emission Factor (by fuel type) (5.4)
Biofuel (in barrels), charcoal (in metric tons), and wood fuel (in cubic meters) were converted to energy units
using energy conversion values before multiplying each by its fuel-specific emission factor. The driving factors for
determining emissions were both historical activity data and projected growth rates.
Activity Data
Historical
• Charcoal and wood fuel activity data were obtained from the FAO's FAOSTAT database (FAO, 2016).
• Biofuel activity data were obtained from ElA's International Energy Statistics database (EIA, 2018).
Projected
• Growth rates for wood fuel consumption, broken out by region, were obtained from FAO (2010). Growth
rates of charcoal consumption were assumed to equal that of wood fuel. Because growth rates are not
provided after 2030, the growth rate for 2031 through 2050 was assumed to equal that of 2021 through
2030.
• Growth rates for biofuels consumption, broken out by region, were obtained from ElA's International
Energy Outlook (EIA, 2017).
• Because of the absence of country-specific growth rate information, individual countries were mapped to
each region to obtain country-level growth rates.
Emission Factors
Historical and Projected
• Tier 1 emission factors were obtained from IPCC (2006). For charcoal and biofuels, the exact IPCC values
were used for all countries. For wood fuel, an energy-weighted emission factor was calculated. Because
the IPCC emission factors are different in each sector and the wood fuel data from FAOSTAT were not
broken out by sector, an energy-weighted factor was calculated to reflect the previous EPA-published
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-27
-------
METHODOLOGY DOCUMENTATION
report (EPA, 2012) in the Energy, Industrial, Transport, and Other sectors. Given that most wood fuel
burning occurs in the Other sector (i.e., the residential and commercial sectors) based on the previous
EPA-published report (EPA, 2012), the emission factor is heavily weighted (80%) to the IPCC's Other
emission factor.
Emission Reductions in Baseline Scenario
The methodology used for this source category does not explicitly model any emission reductions; however,
emission reductions are included to the extent they are reflected in country-reported data.
Uncertainties
Uncertainties in the emission estimates occur because emission factors for biomass combustion are not as
well developed as those for fossil fuels because of limited test data for the variety of types and conditions under
which these fuels are burned. Uncertainties are at least as great as those for fossil fuel CFU and N2O factors (± 50 to
150%). Activity data for biomass fuel combustion also tend to be much more uncertain than fossil fuels because of
the smaller, dispersed, and localized collection and use of these fuels, which makes tracking consumption more
difficult. Detailed data related to the location and magnitude of biomass combustion are limited.
Furthermore, energy values used to convert mass and volume of biomass sources into energy units contain
significant uncertainty.
Uncertainty also occurs due to the lack of sectoral granularity in the wood fuel activity data. The emission
factor for energy and industry for CFU is an order of magnitude smaller than that of the residential and commercial
sectors.
5.1.4.2 Biomass Combustion Mitigation Methodology
The EPA has not estimated the mitigation potential from biomass combustion because of the lack of available
data on mitigation options.
5.1.4.3 References
Food and Agriculture Organization of the United Nations. 2010. Future Trends in Energy, Climate and Woodfuel
Use. Rome, Italy: Food and Agriculture Organization of the United Nations. Available online at
http://www.fao.org/docrep/013/il756e/il756e05.pdfGP
Food and Agriculture Organization of the United Nations. 2016. Forestry Production and Trade. Rome, Italy: Food
and Agriculture Organization of the United Nations. Online Databased Accessed: August 2016. Available online
at http://www.fao.0rg/faostat/en/#data/FO J
Intergovernmental Panel on Climate Change. 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. Energy Information Administration. 2017. International Energy Outlook 2017. Washington, DC: U.S. Energy
Information Administration. Available online at https://www.eia.gov/outlooks/ieo/pdf/0484f2017).pdf
U.S. Energy Information Administration. 2018. International Energy Statistics. Washington, DC: U.S. Energy
Information Administration. Online Database. Available online at
www.eia.gov/cfapps/ipdbproiect/IEDIndex3.cfm
U.S. Environmental Protection Agency. 2012. Global Anthropogenic Non-CC>2 Greenhouse Gas Emissions: 1990-
2030. Available online at https://www.epa.gov/sites/production/files/2016-
08/documents/epa global nonco2 projections dec2Q12.pdf
5-28
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.1.5 Other Energy
This source category includes solid fuel transformation and waste incineration.
5.1.5.1 Other Energy Projections Methodology
This source category solely comprises countries that report data to the UNFCCC database. The EPA did not
perform Tier 1 calculations for other energy sources. The EPA obtained historical values for 1990 through 2012 and
held 2015 through 2050 values constant at 2012 levels for each country.
5.1.5.2 Other Energy Mitigation Methodology
The EPA has not estimated mitigation potential from other energy because of the lack of available data on
mitigation options.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-29
-------
METHODOLOGY DOCUMENTATION
5.2 Industrial Processes
This section presents the methodologies used to develop global non-CC>2 emission projections and mitigation
estimates from the industrial processes sector. The industrial processes sector includes industrial sources of N2O
and Cm, along with several sources of high-GWP gases. The high-GWP sources include the use of substitutes for
ODSs and industrial sources of HFCs, PFCs, and SF6. The categories and their GHG emission and mitigation
estimates presented in this section are as follows:
• adipic acid and nitric acid production (N2O)
• use of substitutes for ozone-depleting substances (HFCs, PFCs)
• hydrochlorofluorocarbons (HCFCs)-22 production (HFCs)
• operation of electric power systems (SFs)
• primary aluminum production (PFCs)
• semiconductor manufacturing (HFCs, PFCs, SFs)
• magnesium manufacturing (SFs)
• flat panel display manufacturing (PFCs, SFs)
• photovoltaic manufacturing (PFCs)
• other industrial processes sources (CH4, N2O) (projections only), including:
- chemical production (CH4)
- iron and steel production (CH4)
- metal production (CH4, N2O)
- mineral products (CH4)
- petrochemical Production (CH4)
- silicon carbide production (CH4)
- solvent and other product use (N2O)
5.2.1 Nitric and Adipic Acid Production
The nitric and adipic acid production source category consists of N2O emissions from the production of adipic
acid and nitric acid. Nitric acid (HNO3) is an inorganic compound used primarily to make synthetic commercial
fertilizer. The production of nitric and adipic acid results in significant N2O emissions as a by-product. Adipic acid
(hexane-1, 6-dioxic acid) is a white crystalline solid used as a feedstock in the manufacture of synthetic fibers,
coatings, plastics, urethane foams, elastomers, and synthetic lubricants.
5.2.1.1 Nitric and Adipic Acid Production Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite
Emission Projections, for additional information). Activity data for adipic acid and nitric acid production included
adipic acid production capacity from Chemical Week (2007) and ICIS Europe (2011, 2013, and 2015) and
production utilization from SRI (2010) and Chemical Week (1999 and 2007); adipic acid consumption growth rate
from SRI (2010); nitrogenous fertilizer production from International Fertilizer Industry Association(IFA) (2016);
and projections of fertilizer consumption from Tenkorang and Lowenberg-DeBoer (2008). Emissions from the nitric
and adipic acid production emission source category were disaggregated to nitric acid production and adipic acid
production source categories. Emission reductions from currently-installed bio-based adipic acid production
5-30
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
capacity were included in the emission estimates, while no abatement technology was assumed to be already
installed for nitric acid production capacity.
The Tier 1 base case equations to estimate N2O emissions from adipic acid and nitric acid production are as
follows:
N20 emissions = Adipic Acid Production * Partially Abated Emission Factor (5.5)
N20 emissions = Nitric Acid Production * Unabated Emission Factor (5.6)
For adipic acid production, the driving factor for determining projected emissions is the projected 3.5% annual
growth rate in global production capacity, although the assumed increasing adoption of bio-based adipic acid
production technologies that do not emit N2O mitigates the projected emissions increase to some extent. For nitric
acid production, the driving factor for determining projected emissions is projected increases in nitrogenous
fertilizer production, which is used to approximate production of the compound.
Activity Data
Historical
Adipic Acid
• When country-reported emission data were unavailable, production data were estimated based on adipic
acid plant production capacity figures and estimated capacity utilization. Capacity utilization was assumed
to be 75% in 1990, 80% in 1995, 90% in 2000 and 2005, and 88% in 2010 through 2050 (SRI, 2010;
Chemical Week, 2007,1999). Adipic acid production data at the country level were obtained for 1990,
1995,1998, 2007, 2009, 2011, 2013, and 2015 (Chemical Week, 2007; ICIS Europe, 2011, 2013, 2015)
Nitric Acid
• When country-reported emission data were unavailable, country-specific fertilizer production data were
used (International Fertilizer Industry Association [IFA], 2016).
Projected
Adipic Acid
• Global adipic acid consumption was forecasted to increase by 3.5% annually for the period 2008 through
2013 (SRI. 2010). In this analysis, projections of global adipic acid consumption were used as a surrogate
for production projections, and the 3.5% growth rate was applied from 2015 through 2050. Known bio-
based adipic acid capacity was assumed through 2007, after which bio-based capacity was maintained
constant.
Nitric Acid
• The growth rates of fertilizer consumption from 2015 through 2030 were estimated by using the regional
N fertilizer consumption projections available from Tenkorang and Lowenberg-DeBoer et al. (2008).
Tenkorang and Lowenberg-DeBoer et al. (2008) provided regional fertilizer use for 2015 and 2030.
Fertilizer use for 2020 and 2025 were interpolated. These consumption projections were then used to
calculate average annual growth rates for the 5-year increments between 2015 and 2030, which in turn
were used to project fertilizer use by country.
• The average annual percentage change in fertilizer use by region for the remainder of the projected time
series (i.e., 2030 through 2050),was available from FAO (2012). The average annual regional growth rates
for the 5-year increments between 2030 and 2050 were used to project fertilizer use by country.
Countries were assigned to regions based on Annex I of Tenkorang and Lowenberg-DeBoer (2008) and
Appendix 1 of FAO (2012).
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-31
-------
METHODOLOGY DOCUMENTATION
Emission Factors
Historical and Projected
Adipic Acid
• The IPCC uncontrolled default emission factor for N2O emissions is 300 kilograms N2O per metric ton
adipic acid (IPCC, 2006). The N2O emission factor is partially abated because the emission factor for bio-
based production was assumed to equal zero based on review of the bio-based technologies. Using
known installations through 2007, bio-based adipic acid capacity increases from 0% of installed capacity in
2002 to a peak of 28% in 2011 and declines to 7% in 2050 because of an assumed constant amount of bio-
based adipic acid production capacity and increasing unabated adipic acid production. The emission factor
for conventional adipic acid production was assumed to be the same across all countries.
Nitric Acid
• The unabated emission factor used for Tier 1 calculations is 9 kilograms N2O per metric ton nitric acid
(IPCC, 2006). The unabated emission factor was assumed to be the same across all countries.
Emission Reductions in Baseline Scenario
Adipic Acid
• Emission reductions from currently installed bio-based adipic acid production capacity were included in
the base case Tier 1 estimates, based on a market assessment from Shen et al. (2009).
Nitric Acid
• While no emission reductions were included in the base case Tier 1 estimates for nitric acid production,
historical emission reductions from N2O abatement technology are included in this analysis to the extent
that abatement is reflected in country-reported emission estimates.
Uncertainties
In general, the IPCC default adipic acid emission factor is more certain than the IPCC default nitric acid
emission factor because the adipic acid emission factor is derived from stoichiometry of the process chemical
reaction. The 2006 IPCC Guidelines (IPCC, 2006) estimate an uncertainty range for the unabated adipic acid
emission factor of ±10%. The uncertainty range given for the unabated nitric acid emission factor is ±40%. A more
thorough understanding of country-specific production processes and control technologies would reduce
uncertainty in these estimates by allowing the use of more specific emission factors. Proxying projected nitric acid
production to projected nitrogen fertilizer demand growth is a source of uncertainty because many countries that
use nitrogenous fertilizer do not produce it. Other sources of uncertainty include assumptions in the projected
growth rate of global adipic acid production and assumptions of currently installed and projected bio-based adipic
acid production.
5.2.1.2 Nitric and Adipic Acid Production Mitigation Options Considered
This analysis considered four abatement measures applied to the chemical processes used to produce nitric
and adipic acid to reduce the quantity of N20 emissions released during the production process. Thermal
destruction is the abatement measure applied to the adipic acid production process. The three remaining
measures are applicable to the nitric acid production process.
Nitric acid facilities have the option of using specially designed catalysts to control N20 emissions. The location
of catalyst placement within the nitric acid production process determines the catalyst design, composition, and
terminology. Abatement measures applicable to nitric acid are characterized by where in the production process
they are implemented. These options include primary abatement, secondary abatement, and tertiary abatement.
Primary abatement measures occur within the ammonia burner, preventing the formation of N20. Secondary
abatement measures such as homogeneous thermal decomposition and catalytic decomposition are installed at an
intermediate point in the production process, removing the N20 formed through ammonia oxidation. Tertiary
5-32
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
abatement measures, such as catalytic decomposition and nonselective catalytic reduction (NSCR) units are
applied to the tail gas streams at the end of the nitric acid production process. The implementation of one
technology over another is driven largely by facility design constraints and/or cost considerations. The high
operating costs of NSCR units and improvement in modern facility design will drive most future abatement
projects to adopt secondary or tertiary catalysts over NSCR units.
This section briefly characterizes each abatement measure and the supporting technical assumptions that
were used to compute the break-even prices. Table 5-10 summarizes the costs and technical assumptions for the
four abatement measures. Abatement measure costs were derived from a variety of sources reporting in euros
and dollars over a number of base years. For consistency, we assumed a fixed exchange rate of 1.32 (USD/EUR),
and the Chemical Engineering Plant Cost Index (CEPI) was used to adjust costs for inflation. Consistent with other
sectors evaluated in this study, the costs of abatement developed for this analysis exclude capital and O&M costs
attributable to monitoring, reporting, and verification activities.
Table 5-10: Abatement Measures for Nitric and Adipic Acid Production
1
2010 USD
Total
Annual Benefits
Abatement Option
Installed
Capital
Cost
Annual
O&M
Cost
Time
Horizon
(years)
Technical Non-
Efficiency^ Energy energy
Average
Reductions
(tC02e/yr)
Adipic Acid Production3
Thermal/catalytic decomposition
11.4
2.2
20
96% - 0.3
4,206,218
Nitric Acid Production13
Secondary Abatement
Catalytic decomposition in the
burner
1.3
0.4
15
85% - -
779,571
Tertia ry Abatement
Direct catalytic decomposition
2.3
0.2
15
95% - -
871,286
Tertia ry Abatement
NSCR unit
4.0
2.1
20
95% -
871,286
a Based on adipic acid plant capacity of 200 metric tons of adipic acid per day.
b Based on nitric acid plant capacity of 1,000 tHNOs/day.
Adipic Acid—N20 Abatement Methods
Adipic acid facilities typically direct the flue gas to a reductive furnace in a thermal destruction process to
reduce nitric oxide (NOx) emissions. Thermal destruction is the combustion of off-gases (including N20) in the
presence of Cm. The combustion process converts N20 to nitrogen, resulting primarily in emissions of NO and
some residual N20 (Ecofys, Fraunhofer ISIR, and Oko-lnstitute, 2009). Facilities may also employ a catalytic
decomposition method to abate the N20 generated. The EU Emissions Trading System [ETS] and CDM
methodologies for this abatement measure suggest that heat generated from the decomposition of N20 can be
used to produce process steam for use in local processes, substituting for more expensive steam generated using
fossil fuel alone.
• Applicability: This option applies to adipic acid production facilities that do not currently control N20
emissions. Based on a recent analysis (Schneider et al., 2010), only 9 of the 23 operational facilities in
2010 had unabated N20 emissions.
• Technical Effectiveness: This analysis assumed a 95% efficiency converting N20 into nitrogen and water.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-33
-------
METHODOLOGY DOCUMENTATION
• Technical Lifetime: 20 years
• Capital Cost: The initial capital cost is $156 per metric ton of production capacity in 2010 dollars. This cost
includes the costs of engineering design and process modifications in addition to equipment and
installation costs. Assuming a plant with capacity of 200 tonnes adipic acid production per day, the initial
capital cost would be approximately $11.4 million (2010 USD).
• Annual O&M Costs: Annual costs total $38 per metric ton of production in 2010 dollars, which includes
the costs of annual energy requirements and system maintenance. Assuming a plant with capacity of 200
tonnes and a utilization factor of 80%, the annual operating cost would be $2.2 million (2010 USD).
Catalyst consumption represents 60% of the annual costs.
• Annual Benefits: Steam produced through the decomposition of N20 under this abatement measure can
offset steam generated using more expensive energy sources providing a fuel cost savings. These annual
benefits can equal up to 60% of operating costs (Ecofys et al., 2009). This analysis assumes a more
conservative estimate of 16% of operating costs or $5.6 per metric ton of adipic acid production based on
CDM project documentation.
Nitric Acid—Primary Abatement Measures
This group of abatement measures can be applied at the ammonia oxidation stage of the nitric acid production
process. Perez-Ramirez et al. (2003) identified three alternative approaches categorized as primary abatement
options: optimized oxidation, modification of the Pt-Rh gauzes, and oxide-based combustion catalysts. All three
technologies prevent the formation of N20 in the ammonia burner and would require making adjustments to the
ammonia oxidation process and/or catalyst (Perez-Ramirez et al., 2003). Although the primary abatement
technology options are technically feasible, they are not modeled in this analysis because of a lack of technology
cost data and the fact that the alternative options discussed below achieve greater abatement and are better
defined.
Nitric Acid—Secondary Abatement Measures
Secondary abatement measures remove N20 immediately following the ammonia oxidation stage between
the ammonia converter and the absorption column (Perez-Ramirez et al., 2003). Abatement measures include
thermal decomposition and catalytic decomposition inside or immediately following the ammonia burner. Thermal
decomposition, developed by Norsk Hydro in the 1990s, is better suited for inclusion in new plants, because it
requires redesigning the reaction chamber immediately following the ammonia burner. This design change can
increase the capital cost of a new plant by 5 to 6% but has no impact on operating costs (Perez-Ramirez et al.,
2003). The catalytic decomposition option is better suited for retrofitting and can be incorporated as an add-on
technology at minimal cost. For this analysis, the catalytic decomposition costs were used as the representative
costs of the secondary abatement option.
• Applicability: This option is applicable to all existing nitric acid plants.
• Technical Effectiveness: This analysis assumed an 80% efficiency converting N20 into nitrogen and water.
• Technical Lifetime: 20 years
• Capital Cost: Capital costs include the purchase and installation of the catalyst and any technical
modifications made to the production process. This analysis assumed a capital cost of $3.5/tonne of HNO3
production capacity16 and a plant capacity of 1,000 tHNOs/day. Using these assumptions, the initial capital
costs would equal $1.3 million (2010 USD).
16 Based on costs of € 0.25/tHN03 reported in 2008 euros (EC, 2008) scaled to 2010 USD using the CEPI and an exchange rate of
1.32 (USD/EUR).
5-34
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Annual O&M Costs: Annual O&M costs include catalyst replacement and recycling of spent catalyst,
replacement of spare catalyst, and loss of production due to catalyst disruptions. This analysis assumed an
annual cost of $1.3/tHNC>3 produced and a plant utilization rate of 90% (Perez-Ramirez et al., 2003).
Following the plant example of a 1,000 tHNOs/day, the annual cost would be $0.4 million (2010 USD).
• Annual Benefits: No benefits are associated with this option.
Nitric Acid—Tertiary Abatement Measure: Direct Catalytic Decomposition
Tertiary abatement measures are located after the absorption tower where tail gas leaving the absorption
column is treated to destroy N20 (Perez-Ramirez et al., 2003). Similar to earlier abatement measures, this measure
reduces the N20 into nitrogen and oxygen through thermal or catalytic decomposition.
• Applicability: This option is applicable to most existing nitric acid plants but is highly dependent on site-
specific factors, such as age of the facility and the footprint of the facility. Tertiary abatement measures
may require additional space and additional equipment.
• Technical Effectiveness: The analysis assumed a 82% efficiency converting N20 into nitrogen and water.
• Technical Lifetime: 20 years
• Capital Cost: Capital costs include the purchase and installation of the catalyst and any technical
modifications made to the production process. This analysis assumed a capital cost of $6.3/tonne of HNO3
production capacity17 and a plant capacity of 1,000 tHNOs/day. Using these assumptions, the initial capital
costs would equal $2.3 million (2010 USD).
• Annual O&M Costs: Annual costs include catalyst replacement and recycling of spent catalyst,
replacement of spare catalyst, and loss of production due to catalyst disruptions or the lowering of the
process pressure. This analysis assumed an annual cost of $0.6/tHNC>3 produced and a plant utilization
rate of 90% (Perez-Ramirez et al., 2003). Following the plant example of a 1,000 tHNOs/day, the annual
cost would be $0.2 million (2010 USD).
• Annual Benefits: Minor benefits are associated with this option. Decomposition is an exothermic process,
so a small amount of heat could be recovered from the process and converted to steam. However, the
costs of the equipment needed to recover the heat and convert it to steam could outweigh the benefit.
The ability to accrue benefits would also be limited by the amount of space available to add the
equipment.
Nitric Acid—Tertiary Abatement Measure: NSCR
One specialized type of tertiary catalyst is an NSCR system. The NSCR typically costs more than other types of
tertiary catalysts because it requires a reagent fuel, such as natural gas, propane, butane, or hydrogen, to reduce
NOx and N20 over a catalyst. If an NSCR system is used at a nitric acid plant that is collocated with other chemical
processes, the costs of these reagent fuels may be lessened. For example, if ammonia is produced near the nitric
acid production plant, the waste gas stream from ammonia production is a hydrogen-rich gas stream that could be
used as the reagent fuel for an NSCR.
• Applicability: This option is applicable to all nitric acid production facilities without existing tertiary
abatement measures. Although it is theoretically possible to employ multiple abatement measures, the
likelihood of multiple retrofitted abatement measures operating together in an efficient manner is very
low.
• Technical Effectiveness: The analysis assumed 90% efficiency converting N20 into nitrogen and water.
17 Based on costs of € 0.5/ tHNOs reported in 2008 euros (EC, 2008) scaled to 2010 USD using the Chemical Engineering Plant
Cost Index (CEPI) and an exchange rate of 1.32 (USD/EUR).
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-35
-------
METHODOLOGY DOCUMENTATION
• Technical Lifetime: 20 years
• Capital Cost: Capital costs include the purchase and installation of the NSCR unit and catalyst. This
analysis assumed a capital cost of $12.6/tonne of HNO3 production capacity based on $8.2/tHNC>3
reported in 1991 USD (EPA, 1991) scaled to 2010 USD using the CEPI. Assuming a plant capacity of 1,000
tHNOs/day, the initial capital cost would equal $4.6 million (2010 USD).
• Annual O&M Costs: Annual costs total $8.8/tHNC>3 produced. Annual costs include the cost of reagent
fuel, labor, maintenance, and other fixed costs for capital recovery and insurance. The total annual cost
for the example plant would be $2.9 million per year (2010 USD).
• Annual Benefits: Energy benefits are associated with this option. The NSCR reaction is exothermic, which
means that the reaction generates heat. This heat can be recovered and converted into steam for use as
an energy source.
5.2.1.3 Model Facilities
The MAC analysis is based on project costs developed for a set of model facilities based on the abatement
measure costs discussed earlier in this section. Similar to the steps taken in other sectors, we developed an
inventory of facilities that are representative of existing facilities. Next, we applied the abatement costs to
calculate the break-even prices for each option and applicable facility pair. Finally, the model estimates the
mitigation potential based on the country-specific share of emissions attributed to nitric versus adipic acid
production. This analysis takes the N20 emission projections (given) and allocates emissions based on the
production process to derive the model facility inventories.
Adipic acid facilities are defined through a detailed inventory of the 23 production facilities worldwide
operating in 2010. While no comprehensive inventory was available for nitric acid plants, it is believed that there
are roughly 500 to 600 nitric acid plants globally (Kollmuss and Lazarus, 2010). Instead, we developed a series of 4
model nitric acid production units based on plant characteristics obtained from a detailed inventory of 67 nitric
acid plants that varied in age and production processes.18
Adipic Acid—Facility Inventory
The first step in the analysis was to determine the allocation of projected emissions to nitric and adipic acid
production by country. For example, in the United States, adipic acid production accounted for approximately 15%
of total baseline emissions in 2010, while the majority of emissions were attributed to nitric acid production. Once
the share of baseline emissions is determined, the MAC model can assess the abatement potential on the
technically applicable pool of emissions available for abatement.
To estimate the technically applicable share of emissions, we developed a detailed inventory of operational
adipic acid plants in 2010. Adipic acid plants were used as the starting point because the number of international
adipic acid plants is small (<30 globally), supported by recent literature providing detail on existing plants in 2010
obtained from Schneider, Lazarus, and Kollmuss (2010). The detailed inventory includes 23 adipic acid production
facilities operating in 11 countries totaling approximately 3,000 kt of production capacity.19 Schneider and co-
authors also identified the N20 abatement technologies and plant utilization factors. Figure 5-2 summarizes the
18 Although a number of different processes are employed at nitric acid production facilities, single-pressure plants are much
more common in the United States. Based on information gathered, most nitric acid production plants were constructed to
maximize the yield from Stages 2 and 3 of the production process and, therefore, operate at high pressures.
19 Major changes to previously reported adipic acid inventories (Mainhardt and Kruger, 2000; OECD, 2004) includes the opening
of five new plants in China between 2008 and 2009 and the closure of two plants located in Canada and the U.K. In addition, a
fourth plant located in the United States was idle between 2008 and 2009 and assumed to continue to idle in 2010 (EPA, 2012).
5-36 Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Figure 5-2: Operational Adipic Acid Production Facilities in 2010 by Share of Global Capacity
c
3
O
u
aa
c
3
"D
O
United States (3)
China (7)
Germany (3)
France (1)
South Korea (1)
Singapore (1)
Japan (2)
Italy (1)
Brazil (1)
Ukraine (2) ~ 2%
India (1) ] 0%
0%
5%
¦ With N20 Abatement
~ Unabated
1
10% 15% 20% 25% 30%
Share of Global Production Capacity
35%
Source: Adapted from Schneider, L., M. Lazarus, and A. Kollrnuss. 2010. Industrial N20 Projects Under the CDM: Adipic Acid—A
Case of Carbon Leakage? Working Paper No. WP-US-1006. Sormervilie, MA: Stockholm Environment Institute (SEI j.
http://ec.europa.eu/clima/consultations/0004/unregistered/cdm watch 2 en.pdfJ.
Note: Facility counts are listed in parentheses beside country names.
global adipic acid production capacity breakdown by country, and facility counts are reported in parenthesis after
the country labels. The bottom-up inventory was used to estimate NzO emission from adipic acid production by
country.
Although 11 countries currently produce adipic acid, only 4 countries (China, Ukraine, Japan, and India) have
operational facilities that are known to have no N20 emission controls in place. As the figure shows, all but 15% of
the adipic acid capacity has N20 abatement controls in place. The 15% of capacity with no N20 abatement controls
is represented by the nine smallest facilities in the industry located in China (5), Ukraine (2), Japan (1), and
India (1).
In the 1990s, most of the adipic acid producers in developed countries voluntarily adopted N20 abatement
measures (Schneider et al., 2010; Ecofys et al., 2009; EPA, 2012). In 2005, with the establishment of the CDM
methodology for crediting N,0 abatement projects at adipic acid plants, producers in developing countries began
to adopt N20 abatement measures. Schneider and coauthors point out that although the CDM methodology was
effective in achieving N20 reductions in developing countries, it was limited to facilities that were in operation
before 2005,
Since 2005, much of the growth in adipic acid production capacity has been in China, with five plants coming
online between 2008 and 2009 (Schneider et al., 2010). Future growth is also projected to be highly concentrated
in Asia (Global Industry Analysts Inc., 2010). China alone was expected to see its capacity more than double in the
near term with the opening of five new adipic acid plants between 2011 and 2013 (Zhao, 2011). At the end of
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-37
-------
METHODOLOGY DOCUMENTATION
2015, China had seen significant growth in adipic acid capacity, expanding from 320 kt in 2010 to approximately
1,800 kt (CCFGroup, 2016).
Only 15% of global capacity continues to operate with no known N20 abatement. China and Ukraine account
for over 95% of the capacity with unabated N20 emissions. In China, the five plants operating without abatement
controls account for two-thirds of the country's total adipic acid capacity. For this analysis, we assumed that future
abatement potential is limited to the nine plants identified as having no known N20 abatement measure in place.
Although no information was available on specific plant utilization rates, we assumed utilization rates of 60%
for all non-CDM facilities, 85% for CDM facilities,20 and 45% for non-CDM facilities in other parts of Asia (Schneider
et al., 2011). Combining plant capacities and corresponding utilization rates yields a total adipic acid production in
2010 of 1.84 million metric tons.
Next, we estimated net emissions for each country by applying the IPCC emission factor of 300 kg N20 per
metric ton of adipic acid produced to the plant production estimated above. Net emissions estimated account for
existing abatement activity assuming a control efficiency of 96%. This analysis yields net emissions by country
totaling 103,800 tonnes of N20 (32.2 MtC02e) in 2010.
We assumed the net emissions calculated for each country represent adipic acid's representative share of
total projected baseline emissions. Table 5-11 provides the percentages used to break out the N20 emission
baseline to adipic acid.
The analysis assumed that N20 emissions from adipic acid production account for the percentage of total
sectoral baseline listed in Table 5-11. We attribute the balance of baseline emissions to nitric acid production.
Table 5-11: Adipic Acid-Producing Countries' Share of Baseline Emissions3
Share of N2O Baseline, %
Country
Adipic Acid
Nitric Acid
Brazil
5
95
China"
36
64
France
30
70
Germany
21
79
India
1
99
Italy
27
73
Japanb
36
64
Singapore
25
75
South Korea
5
95
Ukraine15
36
64
United States
15
85
Other Countries
0
100
a For China, Japan, and Ukraine, the more detailed inventory-based estimate of emissions developed for this analysis yielded
emission values greater than the total baseline projections for 2010. Hence, we defaulted back to percentages assumed in EPA
(2013) (36%).
b China, Japan, and Ukraine percentages used are from the EMF 21 MAC model (EPA, 2006).
20 Facilities located in Brazil, China, and South Korea.
5-38
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Nitric Acid Model Facility Description
While it is believed that there are roughly 500 to 600 nitric acid plants globally (Kollmuss and Lazarus, 2010),
no comprehensive inventory was available for nitric acid plants. Instead, we developed a series of 4 model nitric
acid production units based on plant characteristics obtained from a detailed inventory of 67 nitric acid plants that
varied in age and production processes. We organized the model facilities based on production capacity. All four
facility types were assumed to have an uncontrolled emission factor of 8.5 kg N20 per tHNC>3 produced21 (IPCC,
2006). Table 5-12 summarizes the model facilities for nitric acid production by capacity and resulting annual N20
emissions.
Table 5-12: Model Nitric Acid Facilities Assumptions
Annual N2O Emissions
Production
(uncontrolled)
Model Plants
(tHNOs/yr)
(tN20)
Small
30,600
261
Medium
113,333
968
Large
226,667
1,936
Modern plant
340,000
2,904
5.2.1.4 Sector-Level Trends and Considerations
The following additional data and detail would improve our abatement potential estimates:
• Abatement technology utilization rates: Active CDM and Joint Implementation abatement projects in this
sector have reported N20 reduction efficiencies and utilization rates significantly higher than the default
assumptions provided by the IPCC.
• Technology applicability: Across various nitric acid production processes having a better understanding of
how costs for abatement measures would vary with each process.
5.2.1.5 References
CCFGroup. May 2016. China adipic acid export surges to new high in Mar 2016. Available online at
http://www.ccfgroup.com/newscenter/newsview.php7Class ID=D00000&lnfo ID=20160512110cf
Chemical Week. March 10, 1999. Product focus: Adipic acid/adiponitrile. Chemical Week, pg. 31.
Chemical Week. August 1-8, 2007. Product focus: Adipic acid. Chemical Week.
European Commission. February 2008. Support for the Development and Adoption of Monitoring and Reporting
Guidelines and Harmonised Benchmarks for N2O Activities for Unilateral Inclusion for the EU ETS for 2008-12.
Available online at:
https://ec.europa.eu/clima/sites/clima/files/ets/allowances/docs/entec study 2008 en.pdfc?
Ecofys, Fraunhofer ISIR (Institute for Systems and Innovation Research), and the Oko-lnstitute. 2009. Methodology
for the Free Allocation of Emission Allowances in the EU ETS Post 2012 (pp. 6-20, and 42-48). European
Commission. Available online at http://ec.europa.eu/clima/policies/ets/benchmarking/docs/ bm study-
chemicals en.pdfc?
21 The default emissions factor for the high-pressure process is 9 kg N20 per ton of nitric acid; the default emissions factor for
the medium-pressure processes is 7 kg N20 per ton of nitric acid produced.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-39
-------
METHODOLOGY DOCUMENTATION
Food and Agriculture Organization of the United Nations. 2012. World Agriculture towards 2030/2050 Report.
Available online at http://www.fao.org/docrep/016/apl06e/apl06e.pdfc?
Global Industry Analysts Inc. 2011. Global Adipic Acid Market to Reach 5.5 Billion Pounds by 2015, According to a
New Report by Global Industry Analysts, Inc. Obtained on March 11, 2011 at:
http://www.prweb.eom/releases/2011/l/prweb8043623.htm
ICIS Europe. 2011. European chemical profile: Adipic acid. ICIS. Available online at
https://www.icis.com/resources/news/2011/02/28/9439026/european-chemical-profile-adipic-acid/bP
ICIS Europe. 2013. Asia Chemical Profile: Adipic Acid. ICIS. Available online at
https://www.icis.com/resources/news/2013/04/20/966Q653/asia-chemical-profile-adipic-acid/cf
ICIS Europe. 2015. US INVISTA to close adipic acid plant in Texas. ICIS. Available online at
https://www.icis.com/resources/news/2015/10/06/9930535/us-invista-to-close-adipic-acid-plant-in-texas/cP
Intergovernmental Panel on Climate Change. 2006. 2006IPCC Guidelines for National Greenhouse Gas Inventories:
Volume 3 Industrial Processes and Product Use. Available online at http://www.ipcc-
nggip.iges.or.ip/public/2006gl/pdf/3 Volume3/V3 3 Ch3 Chemical lndustrv.pdf cP
International Fertilizer Industry Association, 2016. IFADATA Statistical Database. International Fertilizer Industry
Association. Available online at http://www.fertilizer.org/ifa/ifadata/search eP
Kollmuss, A., and M. Lazarus. 2010. Industrial N20 Projects Under the CDM: The Case of Nitric Acid Production.
Working Paper No. WP-US-1007. Somerville, MA: Stockholm Environment Institute. Available online at
http://www.sei-international.org/mediamanager/documents/ Publications/Climate/sei-nitricacid-
9nov2010.pdf cf
Mainhardt, H. and D. Kruger. 2000. N20 emissions from adipic acid and nitric acid production. Good Practice and
Uncertainty Management in National Greenhouse Gas Inventories. Montreal: Intergovernmental Panel on
Climate Change, National Greenhouse Gas Inventories Programme. Available online at http://www.ipcc-
nggip.iges.or.ip/public/gp/bgp/3 2 Adipic Acid Nitric Acid Production.pdf cP
Organisation of Economic Cooperation and Development. 2004. SIDS Initial Assessment Report for SIAM18 Adipic
Acid. Paris, France: OECD.
Perez-Ramirez, J., F. Kapteijn, K. Schoffel, and J. A. Moulijn. 2003. Formation and control of N20 in nitric acid
production: Where do we stand today? Applied Catalysis B: Environmental, 44(2), 117-151.
doi:10.1016/S0926-3373(03)00026-2.
Schneider, L., M. Lazarus, and A. Kollmuss. 2010. Industrial N20 projects under the CDM: Adipic acid—A case of
carbon leakage? Working Paper No. WP-US-1006. Somerville, MA: Stockholm Environment Institute. Available
online at http://ec.europa.eu/clima/consultations/0004/ unregistered/cdm watch 2 en.pdf cP
Shen, L., J. Haufe, and M.K. Patel. 2009. Product overview and market projection of emerging bio-based plastics
PRO-BIP 2009. Report for European polysaccharide network of excellence (EPNOE) and European bioplastics
243. Available online at
https://www.researchgate.net/profile/Li Shenl5/publication/216092211 Product overviw and market proi
ection of emerging bio-based plastics PRO-BIP 2009/links/0c9605279efb4e96a8000000.pdfoP
SRI. January 2010. World Petrochemical Report: Adipic Acid. SRI Consulting. Access Intelligence LLC Inc. Abstract
available online at http://www.sriconsulting.com/CEH/Public/Reports/608.500Q/ cP
Tenkorang, F. and J. Lowenberg-DeBoer. 2008. Forecasting Long-term Global Fertilizer Demand. Rome: Food and
Agriculture Organization of the United Nations.
U.S. Environmental Protection Agency. 1991. Alternative Control Techniques Document—Nitric and Adipic Acid
Manufacturing Plants. EPA-450/3-91-026. Research Triangle Park, NC: EPA. Available online at
http://www.epa.gov/ttn/catc/dirl/nitric.pdf
5-40
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
U.S. Environmental Protection Agency. 2012. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2010.
Washington, DC: EPA. Available online at http://epa.gov/climatechange/emissions/usinventorvreport.html
U.S. Environmental Protection Agency. 2013. Global Mitigation ofNon-CC>2 Greenhouse Gases: 2010-2030. EPA-
430-R-13-011. Washington DC: EPA. Available online at https://www.epa.gov/global-mitigation-non-co2-
greenhouse-gases/global-mitigation-non-co2-ghgs-report-download-report
Zhao, D. November 7, 2011. China adipic acid tumbles 36% on weak demand from PU sector. Market news article
from ICIS.com. Available online at http://www.icis.com/Articles/2011/ll/07/ 9505810/china-adipic-acid-
tumbles-36-on-weak-demand-from-pu-sector.html ef
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-41
-------
METHODOLOGY DOCUMENTATION
5.2.2 F-GHG Emissions from Semiconductor Manufacturing
Semiconductor manufacturing consists of HFC, PFC, SF6, and NF3 emissions from two repeated activities: (1)
cleaning tool chambers used to deposit thin films on substrate surfaces, a process referred to as chemical vapor
deposition (CVD) chamber cleaning, and (2) etching intricate patterns into successive layers of films and metals, a
process referred to as plasma etching. Film deposition and etching processes begin with the semiconductive
crystalline silicon (Si) wafer (or another substrate) and are completed once successive films (layers) are deposited
and etched to form a device. EPA GHGRP data indicate that approximately 27% to 46% of emissions result from
chamber cleaning processes and 54% to 73% from etching processes (EPA, 2014, 2015, 2016).
5.2.2.1 Semiconductor Manufacturing Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3 Generating the Composite
Emission Projections, for additional information). Activity data for semiconductor manufacturing included annual
semiconductor manufacturing capacity by country from World Fab Watch (SEMI 1996, 2001, 2002, 2003, 2006 and
2007) and World Fab Forecast (SEMI 2008, 2009, 2010, 2011, 2012, and 2016), and growth of manufacturing
capacities based on the growth in each country's gross domestic product. Emission factors were based on EPA's
GHGRP data and activity from the aforementioned sources (2016). Emission reductions were incorporated based
on the EPA and U.S. semiconductor industry's voluntary partnership and the global industry commitment through
the World Semiconductor Council to reduce F-GHG emissions.
The Tier 1 equation to estimate HFC, PFC, SF6, and NF3 emissions from semiconductor manufacturing is as
follows:
Ei = EFi * cu * Cd * [1 - (ai * di)] (5.7)
where:
Ei = Emissions of gas /
EFi = IPCC Tier 1 emission factor for gas /
Cu = Annual plant production capacity utilization
Cd = Manufacturing design capacity
o, = Abatement fraction of gas /
di = Destruction or removal efficiency of gas /
Emissions vary over time as a function of all the variables above: the emission factor is driven by
advancements in manufacturing technology, the capacity utilization and capacity design are driven by the demand
of the semiconductor products, the abatement fraction is facility specific and depends on whether the
manufacturing facility uses abatement devices, and the destruction or removal efficiency depends on the type of
abatement technology used.
5-42
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Activity Data
Historical
Historical activity data consisted of the annual semiconductor manufacturing capacity by country, broken out
by 200-mm and 300-mm wafer size.22 These data are provided by the World Fab Watch (SEMI, July 1996, 2001,
2002, April 2003, 2006 and 2007 editions) and World Fab Forecast (SEMI, 2008, 2009, 2010, 2011, 2012, and 2016
editions).23 These sources were used to develop country-specific capacity shares for 1995, 2000, 2005, 2010, and
2015.24 Beginning in 2006, World Fab Forecast provided activity data separated out by 200-mm and 300-mm wafer
sizes. However, for the years 1990 through 2005 when World Fab Watch was used, manufacturing capacities were
not distinguished by wafer size in the data source. Therefore, it was assumed that all global capacity before 2006
was for the 200-mm wafer size.
Projected
For all countries, their manufacturing capacities from 2020 through 2050 were estimated by growing the
manufacturing capacity at a rate equivalent to the growth in each country's GDP over the same time period. These
total capacities were then disaggregated by 200-mm and 300-mm wafer sizes based on projected shares of each
wafer size. The projected shares of capacity for 200-mm wafers were developed through linear extrapolation of
historical data from 2006 through 2017. The shares of 300-mm wafer capacity were calculated as (1-200-mm
share). This projection by 200-mm and 300-mm shares was done to ensure that the recent increase in shares of
300-mm fabs compared against 200-mm fabs (growing from only 42% of global production capacity in 2006 to 51%
in 2017) was represented for the projected years. In addition, if a particular country had no 300-mm wafer sizes in
the reported historical data, then that country was assumed to continue using only 200-mm fabs in the projections
through 2050.
Emission Factors
Historical and Projected
Emissions were estimated using EPA-derived emission factors for F-GHG emissions in the units of kgCChe/cm2.
Emission factors were developed using EPA GHGRP-reported data (EPA, 2017) and activity data as described
above. Emission factors were obtained for each 200-mm and 300-mm wafer size using a regression-through-the-
origin (RTO) model: facility-reported aggregate emissions of F-GHGs were regressed on the corresponding
manufacturing capacity (m2) to estimate an aggregate F-GHG emission factor (CO2 Eq/m2). The slope of the RTO
model is the emission factor of the subpopulation.
The EPA developed 2011 F-GHG emission factors (subject to the availability of both the GHGRP data and
World Fab Forecast data) and applied them to both the historical and projected data. The emission factors used
are presented in Table 5-13.
Table 5-13: Emission Factor Used for Semiconductor Manufacturing
Emission Factor
kg COie/cm2
Wafer Size
Historical
Projected
200 mm
1.06
0.89
300 mm
0.43
0.36
22 All wafer sizes less than 200 mm were included in the 200-mm category, and all wafer sizes greater than 300 mm were
included in the 300-mm category.
23 The World Fab Watch was the predecessor of the World Fab Forecast.
24 Country-specific capacity shares in 1990 were assumed to be equivalent to those in 1995.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-43
-------
METHODOLOGY DOCUMENTATION
When comparing country emissions as calculated using the approach described above (Tier 1) with emissions
reported to the UNFCCC, cases of over- or underestimations were identified. To address these instances, when
feasible, country emission projections were estimated using the most recent country-reported data (2015) and the
growth rates and country shares described above.
Finally, shares of emissions by gas were calculated from the historical reported emissions to give percentages
of HFC, PFC, SFs, and NF3 emissions of total semiconductor emissions. These shares were applied to country
emissions for all years, holding 2015 shares constant for all future years, to determine country emissions by type of
gas. The assumption to hold 2015 shares constant in the projections was made because it cannot be known at this
time what new processes and technologies will be used in semiconductor manufacturing, which is a high-tech and
rapidly evolving industry. As a result, if new technologies come online in future years, the shares of emissions
could change.
Emission Reductions in Baseline Scenario
In the 1990s, the absence of emission control measures, rapid growth of the semiconductor industry (11 to
12% per year through the late 1990s), and increasing complexity of microchips could have potentially resulted in
significantly increased future emissions from semiconductor manufacturing. In response, the EPA and the U.S.
semiconductor industry launched a voluntary partnership to reduce F-GHG emissions in 1996. In 1999, the U.S.
partnership catalyzed a global industry commitment through the World Semiconductor Council (WSC). Most WSC
member countries—the United States, EU, Japan, South Korea, and Taiwan25—voluntarily committed to reduce
HFC, PFC, SFs, and NF3 emissions to 90% of 1995 levels by 2010.26 At the end of the 2010 WSC goal, the member
countries reduced absolute emissions by 32% below the baseline to a level of 2.7 MMTCE, surpassing the 10%
reduction target (WSC, 2016). WSC set a subsequent goal for 2020 to implement best practices for new
semiconductor fabs to result in a normalized emission rate (NER) of 0.22 kgCChe/cm2, which is equivalent to a 30%
NER reduction from a 2010-aggregated baseline.
Uncertainty
This analysis projected emissions based on the assumption that current semiconductor manufacturing
processes continue, and that currently available abatement technologies are used to reduce the resulting F-GHG
emissions. It did not model a possible future in which new technologies are applied or F-GHGs use in
semiconductor manufacturing is significantly reduced or eliminated. Thus, this analysis may overestimate
emissions. In addition, the emission factors used could be revisited in the future, because there are multiple
options for emission factor analysis that could affect the results. Finally, the percentages of semiconductor
manufacturing emissions that result from chamber cleaning and etching processes were drawn from industry
reports; however, these breakdowns could be updated based on the EPA GHGRP data to show a range for 200-mm
and 300-mm wafer sizes.
25 For purposes of this report, emissions presented for China include emissions from manufacture in China and Taiwan;
however, emissions for these countries were estimated separately because they are treated separately under the WSC and
have different industry associations.
26 For the U.S. Semiconductor Industry Association, Japan Electronic and Information Technology Industries Association, and
European Semiconductor Industry Association, the baseline year is 1995; for the Korean Semiconductor Industry Association,
the baseline year is 1997; and for the Taiwan Semiconductor Industry Association, the baseline is the average of the emission
values in 1997 and 1999.
5-44
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.2.2.2 Mitigation Options Considered for Semiconductor Manufacturing
Six mitigation technology options were considered for the semiconductor manufacturing sector: thermal
abatement, catalytic abatement, plasma abatement, NF3 remote chamber clean, gas replacement, and process
optimization.
• Thermal Abatement: These point-of-use abatement systems that use heat to destroy or remove F-GHGs
from effluent process streams are connected directly to a manufacturing tool.
• Catalytic Abatement: Tool effluent process streams are run through abatement systems with catalysts
(e.g., CuO, ZnO, Al203) that destroy or remove F-GHGs.
• Plasma Abatement: Plasma, in a point-of-use abatement system, is used to react with (thereby destroying
or removing) F-GHGs from the process effluent stream.
• NF3 Remote Chamber Clean: Highly ionized NF3 is used to clean CVD chambers. This process is very
efficient (using ~98% of the gas in a process), resulting in lower emissions on a mass and C02 basis than
traditional in situ chamber clean processes that use approximately 20% to 50% of the gas in a process and
have lower efficiencies (EPA, 2010).
• Gas Replacement: Higher GWP gases are replaced with lower GWP gases and in some cases more
efficient gases (e.g., C4F8 may replace C2F6 in a traditional chamber-cleaning process).
• Process Optimization: Processes are adjusted to become more efficient, using more gas within the
process or inputting less gas into the process, and thus resulting in lower emissions.
These technologies reduce emissions from either etch or chamber-cleaning processes or in some cases both.
Table 5-14 demonstrates the applicability of each mitigation technology to each process type. While some of these
technologies can be stacked or used together (e.g., a process can be optimized and then abatement can be applied
to that process), the cost and mitigation analysis did not model this situation.
Table 5-14: Semiconductor Manufacturing Abatement Options
Thermal
Catalytic
Plasma
NF3 Remote
Gas
Process
Fab/Emissions Type
Abatement
Abatement
Abatement
Clean
Replacement
Optimization
Reduction efficiency
95%
99%
97%
95%
77%
54%
New fab
Etch emissions
X
X
X
Clean emissions
X
X
Old fab
Etch emissions
X
X
X
Clean emissions
X
X
X
X
Thermal Abatement
Thermal abatement systems can be used to abate emissions from both etching and CVD chamber-cleaning
processes by heating process effluent streams to high temperatures to remove or destroy F-GHGs. The use of
thermal abatement offers the benefit of not affecting the manufacturing process (Applied Materials, 1999);
however, the systems do require space that may not be available in sub-fabs, particularly in older facilities. In
addition, these systems require large amounts of cooling water, and the system's use results in regulated NOx
emissions. Thermal abatement systems are currently the most widely used abatement system in the
semiconductor industry.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-45
-------
METHODOLOGY DOCUMENTATION
Catalytic Abatement
A catalytic abatement system uses a catalyst to destroy or remove F-GHG emissions from the effluents of both
plasma etching and CVD chamber-cleaning processes. This type of abatement is applicable at most facilities, but
there may be some space constraints as mentioned above for thermal abatement systems. Additionally, because
these systems are based on destruction via catalyst, they must be process/stream specific to achieve the 99%
emission reductions quoted in the literature and used in this analysis (Fthenakis, 2001; Burton, 2003).
Because catalytic destruction systems operate at relatively low temperatures, their use results in little or no
emissions, and the required amounts of water are also low. Because of the high cost of catalyst replacement, these
systems are the least widely used type of abatement (expert judgment).
Plasma Abatement
These systems, which use plasma to remove or destroy F-GHGs, are applicable to etch processes in most
facilities, with some physical space limitations. (These systems, though, are smaller in size compared with thermal
and catalytic systems.) Plasma abatement systems use a small plasma source that effectively dissociates the F-GHG
molecules that react with fragments of the additive gas (hydrogen, oxygen, water, or CH4) to produce low
molecular weight by-products such as HF with little or no GWP. After disassociation, wet scrubbers can remove the
molecules. The presence of additive gas is necessary to prevent later downstream reformation of F-GHG molecules
(Motorola, 1998).
NF3 Remote Chamber Clean
NF3 remote chamber clean is an alternative cleaning technology that offers the benefit of having a particularly
high (~98%) utilization rate of NF3 (IPCC, 2006 and EPA Subpart I of the Greenhouse Gas Reporting Program),
resulting in relatively low emissions compared with traditional chamber cleans. NF3 remote clean systems
dissociate NF3 using argon gas and converting the source gas to active F-atoms in the plasma upstream of the
process chamber. These electrically neutral atoms can selectively remove material in the chamber. The by-
products of remote clean include HF, F2, and other gases, most of which are removed by facility acid scrubber
systems. The use of NF3 remote clean systems is much more prevalent in newer fabs because the technology was
not available when many older fabs were constructed.
Capital costs for NF3 remote clean systems will differ for new and old fabs because of the "readiness" for NF3
remote clean installation. "Readiness" consists of having the current infrastructure (e.g., duct work, hookups) for
system installation. It was assumed that old fabs do not have the current infrastructure to use NF3 remote clean,
whereas new fabs do. Therefore, the capital costs for old fabs reflect the needed infrastructure changes for the
fab.
Gas Replacement
Gas replacement can be used to mitigate emissions from the traditional CVD chamber-cleaning process. This
method can be applied in most facilities and has already been used throughout the industry in many instances. For
this strategy, a lower GWP gas replaces a higher GWP gas. The most common replacement seen is using C4F8to
replace C3F8 or C2F6. In addition, the replacement gas (C4F8) is often used/consumed more efficiently during CVD
chamber cleaning than the original gas C2F6 or C3F8, which, combined with the differences in GWP, yields lower
emissions.
Process Optimization
Process optimization is the reduction in GHG emissions from a process by modifying or adding to the process
recipe. Process optimization is considered to be only applicable for chamber cleans because these processes offer
the opportunity for more flexibility than etch processes. Etch processes are typically developed to optimize
5-46
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
production yield, and they are only adjusted to increase this yield; a company would not risk negatively affecting it
(Beu, 2005; Fthenakis, 2001). Process gas optimizations for CVD clean processes can be implemented because
adjustments to these processes are much less precise than etch processes. There is room to reduce emissions
without affecting yield. Optimization of clean processes to reduce emissions usually results in small production
gains but sometimes can result in large increases in efficiency.
Facilities optimizing processes incur labor costs of an estimated $121,370; it was assumed that old fabs incur
this cost, while new fabs do not implement this technology because of their assumed use of NF3 remote clean for
most of clean processes.
5.2.2.3 Model Facilities
The analysis considers two types of facilities, distinguished generally by the size of wafer manufactured:
• New Facilities: Facilities that manufacture wafers that are 300 mm or larger. These types of facilities tend
to have more complex processes, particularly for plasma etching, which can lead to higher emissions.
These facilities have a high likelihood of using remote chamber clean processes and are more likely to
have abatement installed. Using data from Subpart I of EPA's GHGRP and information from the World Fab
Forecast, average new facility emissions, size of facility in terms of manufactured wafer area, and the
percentage of etch versus clean emissions were determined. For purposes of this analysis, these
characteristics of the facility did not change over time.
• Old Facilities: Facilities that manufacture wafers that are 200 mm or smaller. These types of facilities tend
to have fewer complex processes and rely on more traditional in situ chamber clean processes as well as
remote clean processes. They also have more physical space limitations in facility subflooring, which can
limit the use of abatement systems. Using data from Subpart I of EPA's GHGRP and information from the
World Fab Forecast, average new facility emissions, size of facility in terms of manufactured wafer area,
and the percentage of etch versus clean emissions were determined. For purposes of this analysis, these
characteristics of the facility did not change over time.
5.2.2.4 Technical and Economic Characteristics Summary
Table 5-15 and Table 5-16 report the technical applicability, market penetration, and reduction efficiency
assumptions used to develop the abatement measures' technical effectiveness at new and old fabs. The technical
effectiveness is the weighted average of the abatement measures using the process emissions presented in
Table 5-15 for each process as the weight multiplied by the product of the technical applicability, market
penetration, and reduction efficiency.
Table 5-15: Technical Effectiveness Summary for New Fabs (Constant Over Time)
Etch (54%) Clean (46%)
Abatement Technical Market Technical Market Reduction Technical
Measure Applicability Penetration Applicability Penetration Efficiency Effectiveness
Thermal abatement
0%
0%
90%
50%
95%
20%
Catalytic abatement
0%
0%
0%
0%
99%
0%
Plasma abatement
0%
0%
0%
0%
97%
0%
NF3 remote clean
0%
0%
10%
50%
95%
2%
Gas replacement
0%
0%
0%
0%
77%
0%
Process
optimization
0%
0%
0%
0%
54%
0%
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis 5-47
-------
METHODOLOGY DOCUMENTATION
Table 5-16: Technical Effectiveness Summary for Old Fabs (in 2020)
Etch (34%)
Clean (66%)
Abatement
Measure
Technical
Applicability
Market
Penetration
Technical
Applicability
Market
Penetration
Reduction
Efficiency
Technical
Effectiveness
Thermal abatement
30%
90%
30%
15%
95%
15%
Catalytic abatement
30%
5%
0%
5%
99%
1%
Plasma abatement
30%
5%
0%
0%
97%
1%
NF3 remote clean
0%
0%
100%
5%
95%
4%
Gas replacement
0%
0%
10%
40%
77%
2%
Process
optimization
0%
0%
10%
40%
54%
2%
Technical applicability assumptions presented in Table 5-15 and Table 5-16 are intended to reflect the space
limitations or preexisting process performance issues that are likely to be found at a fraction of all facilities,
particularly old facilities, preventing the total implementation of the abatement measures. Assumed market
penetration rates are based on cost (lower cost options will penetrate the market more) and expert knowledge of
industry trends. For example, fabs tend to use thermal abatement more than the other technologies in their
etching processes. In addition, it was assumed that because most new fabs already have NF3 remote systems in
place less market share would go to gas replacement and process optimization.
The technical effectiveness estimates were then multiplied by the share of total emissions for each facility
type to estimate the abatement potential achievable under each abatement measure.
Table 5-17 presents a summary of the engineering cost data for each of the mitigation technologies.
Table 5-17: Engineering Cost Data on a Facility Basis—Semiconductor Manufacturing
Abatement
Project Lifetime
(years)
Capital Costs
(2010 USD)
Annual Costs
(2010 USD)
Abatement Amount
(tCOze)
Option
New Old
New
Old
New
Old
New
Old
Thermal
abatement
7 7
$12,551,949
$6,275,974
$723,935
$361,967
33,724
8,143
Catalytic
abatement
7 7
$15,203,729
$7,601,864
$957,481
$501,109
502
358
Plasma
abatement
7 7
$3,994,685
$1,997,342
$114,134
$57,067
492
350
NF3 remote
clean
22 11
$3,307,599
$10,127,096
$1,337,192
$3,714,600
2,784
2,181
Gas
replacement
22 11
n/a
$1,539,176
n/a
$83,783
n/a
1,414
Process
optimization
22 11
n/a
$121,370
n/a
($168,359)
n/a
992
5-48
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.2.2.5 Sector-Level Trends and Considerations
Several important industry trends drive changes in emissions and mitigation potential from semiconductor
manufacturing: (1) rapid production growth, (2) evolving manufacturing processes and increasing complexity in
devices produced, and (3) impacts of mitigation efforts resulting from voluntary emission reduction goals. These
trends are described below.
Etch and chamber-cleaning processes have evolved as semiconductor technologies have advanced and
understanding of the emission pathways associated with manufacturing has improved. As technologies advanced,
the semiconductor industry used larger wafer sizes to increase chip production (e.g., 150 mm to 200 mm to 300
mm to 450 mm). Fabs that produce semiconductors on smaller wafers, on average, tend to be older and use
manufacturing processes that result in a different breakdown of F-GHG emissions from etch and clean processes as
compared with newer fabs. Older fabs may emit approximately 80% of F-GHG emissions total from chamber-
cleaning processes and about 20% of emissions from etch processes. These percentages change to about 45%/55%
clean/etch for newer fabs. This shift in the source of emissions over time is a result of the following: (1) newer fabs
generally are trending to NF3 remote-clean technologies that result in lower emissions on a C02e basis than
traditional older C2F6- or C4F8-based clean systems; (2) more technologically advanced etch processes have a
significantly greater number of steps, resulting in more F-GHG emissions; and (3) newer fabs can have less physical
limitations on using abatement. As a new generation of fabs comes online using 450-mm wafers, it is expected
they will continue to use NF3 remote clean technologies, abatement, and more advanced etch processes.
In the 1990s, the absence of emission control measures, rapid growth of the semiconductor industry (11% to
12% per year through the late 1990s), and increasing complexity of microchips could have potentially resulted in
significantly increased future emissions from semiconductor manufacturing. In response, the EPA and the U.S.
semiconductor industry launched a voluntary partnership to reduce F-GHG emissions in 1996. In 1999, the U.S.
partnership catalyzed a global industry commitment through the World Semiconductor Council (WSC). Most WSC
member countries—the United States, European Union, Japan, South Korea, and Taiwan27—voluntarily committed
to reduce HFC, PFC, NF3, and SF6 emissions to 90% of 1995 levels by 2010.28 This emission reduction goal was met
in 2010. Achievement of the 2010 WSC emission reduction goal has occurred in the context of significantly
increasing underlying manufacturing activity (WSC, 2011). At the end of the 2010 WSC goal, the member countries,
including China, reduced absolute emissions by 32% below the baseline to a level of 2.7 MMTCE, surpassing the
10% reduction target (WSC, 2016). WSC set a subsequent goal for 2020 to implement best practices for new
semiconductor fabs to result in a Normalized Emission Rate (NER) of 0.22 kgCChe/cm2, which is equivalent to a
30% NER reduction from a 2010-aggregated baseline. For the with measures scenario, it was assumed that all of
WSC member countries met and maintained the new WSC goal.
5.2.2.6 References
Applied Materials. October 18,1999. Catalytic abatement of PFC emissions. Presented at Semicon Southwest 99: A
Partnership for PFC Emissions Reductions, Austin, TX.
27 For purposes of this report, emissions presented for China include emissions from manufacture in China and Taiwan;
however, emissions for these countries were estimated separately as they are treated separately under the WSC and have
different industry associations.
28 For the U.S. Semiconductor Industry Association (SIA), Japan Electronic and Information Technology Industries Association
(JEITA), and European Semiconductor Industry Association (ESIA), the baseline year is 1995; for the Korean Semiconductor
Industry Association (KSIA), the baseline year is 1997; and for the Taiwan Semiconductor Industry Association (TSIA), the
baseline is the average of the emission values in 1997 and 1999.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-49
-------
METHODOLOGY DOCUMENTATION
Beu, L December 2005. Reduction of Perfluorocarbon (PFC) Emissions: 2005 State-of-the-Technology Report.
TT#0510469A-ENG. International SEMATECH Manufacturing Initiative, Albany, NY. Available online at
http://www.epa.gov/highgwp/semiconductor-pfc/documents/final tt report.pdf
Burton, S. 2003. Personal communication with Brown of Motorola (2002) supplemented by personal
communication with Von Gompel of BOC Edwards (2003), research of DuPont's Zyron Web site (2003), and
personal communication with Air Liquide regarding thermal destruction, NF3 remote clean, and capture
membrane unit costs.
Fthenakis, V. December 2001. Options for Abating Greenhouse Gases from Exhaust Streams. Brookhaven National
Laboratory. Available online at http://www.bnl.gov/isd/documents/23784.pdf
Intergovernmental Panel on Climate Change. 2006. 2006 IPCC Guidelines for National Greenhouse Gas Inventories.
The National Greenhouse Gas Inventories Programme (Volume 3, Chapter 6: Electronics Industry Emissions),
The Intergovernmental Panel on Climate Change, H.S. Eggleston, L. Buendia, K. Miwa, T. Ngara, and K. Tanabe
(eds.). Hayama, Kanagawa, Japan.
Motorola. October 18,1998. Long-Term Evaluation of Litmus "Blue" Inductively-Coupled Plasma Device for Point-
of-Use PFC and HFC Abatement. Presented at Semicon Southwest 99: A Partnership for PFC Emissions
Reductions, Austin, TX.
Semiconductor Equipment and
Semiconductor Equipment and
Semiconductor Equipment and
http://dom.semi.org/ ef
Semiconductor Equipment and
http://dom.semi.org/ J
Semiconductor Equipment and
http://dom.semi.org/ E?
Semiconductor Equipment and
http://dom.semi.org/ tiP
Semiconductor Equipment and
http://dom.semi.org/ ef
Semiconductor Equipment and
http://dom.semi.org/ ef
U.S. Environmental Protection Agency. 2010. Draft Emission Factors for Refined Semiconductor Manufacturing
Process Categories. Office of Air and Radiation Office of Atmospheric Programs, Climate Change Division, U.S.
Environmental Protection Agency, Washington, DC. Available in docket EPA-HQ-OAR-2009-0927.
U.S. EPA. 2014. U.S EPA Greenhouse Gas Reporting Program (GHGRP) Envirofacts. Subpart I: Electronics
Manufacture.
U.S. EPA. 2015. U.S EPA Greenhouse Gas Reporting Program (GHGRP) Envirofacts. Subpart I: Electronics
Manufacture.
U.S. EPA. 2016. U.S EPA Greenhouse Gas Reporting Program (GHGRP) Envirofacts. Subpart I: Electronics
Manufacture.
U.S. EPA. 2017. U.S. EPA Greenhouse Gas Reporting Program (GHGRP) Envirofacts. Subpart I: Electronics
Manufacture. Available online at: http://www.epa.gov/enviro/facts/ghg/search.html
Materials Industry (SEMI). 2008, 2009, 2010, 2011, 2012. World Fab Forecast.
Materials Industry (SEMI). 2016. World Fab Forecast.
Materials Industry (SEMI). 1996. World Fab Watch: 1996 Edition. Available online at
Materials Industry (SEMI). 2001. World Fab Watch: 2001 Edition. Available online at
Materials Industry (SEMI). 2002. World Fab Watch: 2002 Edition. Available online at
Materials Industry (SEMI). 2003. World Fab Watch: 2003 Edition. Available online at
Materials Industry (SEMI). 2006. World Fab Watch: 2006 Edition. Available online at
Materials Industry (SEMI). 2007. World Fab Watch: 2007 Edition. Available online at
5-50
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Other Sources Reviewed for this Analysis:
• Semiconductor Equipment and Materials Industry. 2011 World Fab Forecast, May 2011 Edition.
• Semiconductor Equipment and Materials Industry. 2012 World Fab Forecast, August 2012 Edition.
• Semiconductor Equipment and Materials Industry. 2018 World Fab Forecast, May 2018 Edition.
Clean Development Mechanism Proposed Methodologies Reviewed:
• NM0289: PFC gas emission reduction by gas replacement for CVD cleaning at 200 mm (8 inches) process
by Hynix Semiconductor Inc. (submitted September 2008)
• NM0303: PFC gas emissions reduction by gas replacement for CVD cleaning processes in semiconductor
processing operations (submitted April 2009)
• NM0317: Substitution of fluorinated compound (FC) gases for cleaning CVD reactors in the semiconductor
industry (submitted June 2009)
• NM0330: Substitution of fluorinated compound (FC) gases for cleaning CVD reactors in the semiconductor
industry (submitted December 2009)
• NM0332: PFCs emission reduction from installation of an abatement device in a semiconductor
manufacturing facility (submitted January 2010)
• NM0335: PFC emission reduction by gas replacement in the process of CVD cleaning in semiconductor
production (submitted February 2010)
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-51
-------
METHODOLOGY DOCUMENTATION
5.2.3 Photovoltaic Cell Manufacturing
The photovoltaic (PV) cell manufacturing process can use multiple F-GHGs during production, including NF3
and the PFCs CF4 and C2F6. PV manufacturing emissions occur during the etching and chamber-cleaning processes.
Etching is done on various substrates, including crystalline silicon, amorphous silicon, and other thin films. CF4 and
C2F6 are used during the manufacture of crystalline silicon (c-Si) PV cells, and NF3 is used during the manufacture of
amorphous silicon (a-Si) and tandem a-Si/nanocrystaline (nc) silicon PV cells. Processes for PV cells manufactured
on other thin films do not require the use of GHGs; therefore, these processes were not considered in this
analysis.29
5.2.3.1 Photovoltaic Cell Manufacturing Projections Methodology
UNFCCC-reported, country-specific estimates were not available for any historical years for this emission
source. Therefore, the 2006 IPCCTier 1 methodology (IPCC, 2006) was used to estimate emissions. Activity data for
Photovoltaic Manufacturing include maximum design capacities from DisplaySearch database (2009) as well as
global installed photovoltaic capacities from EIA (2017). Manufacturing processes of PV cells with c-Si PV cells and
a-Si and tandem a-Si/nanocrystalline silicon PV cells use F-GHGs, while other thin film technologies do not.
Therefore, this analysis was limited to the c-Si and a-Si PV cell markets. Emission factors were obtained from IPCC
(2006) and a EPA-developed NF3 emission factor.
The Tier 1 basic equation to estimate PFC and NF3 emissions from photovoltaic manufacturing is as follows:
FO = EFi *Cu*Cd (5.8)
where:
FCi = Emissions of gas / (mass)
EFi = Emission factor for gas / (mass/m2)
Cu = Fraction of annual plant production capacity utilization (%)30
Cd = Annual maximum design capacity of substrate processed (m2)
The main data sources for this source category were the DisplaySearch Q4'09 Quarterly PV Cell Capacity
Database & Trends Report ("DisplaySearch database") (DisplaySearch, 2009) and the International Energy Outlook
2016 from EIA (2017). The most recent DisplaySearch database available to the EPA, published in 2009, supplied
historical (2000 through 2009) and projected (2010 through 2013) annual data through 2013 about all PV
manufacturing facilities in the world. This included location (country), type of technology manufactured at a facility
(c-Si, a-Si, or other thin film), maximum design capacity (megawatts) of a facility, and in some cases conversion
efficiency of the PV technology manufactured at a facility. The International Energy Outlook provided global
installed PV capacities; these data were used to estimate the global PV cell production data for 2015 through 2040.
As evident by data in the DisplaySearch database, a variety of substrates is used in the production of PV cells,
including c-Si, a-Si, and other thin films. Manufacturing processes of PV cells with c-Si PV cells and a-Si and tandem
29 For this analysis, technology market shares were calculated based on a PV market that was assumed to only include c-Si and
a-Si technologies, since other thin film technologies do not use fluorinates gases. No correction was made in the global solar
manufacturing capacity to account for thin film technologies that do not use fluorinated gases. Projected maximum design
capacities also included capacities for c-Si and a-Si only, the two technologies that use PFCs in their manufacturing processes.
30 The fraction of annual plant production capacity utilization (Cu) was assumed to be equivalent to 100%; that is, EPA assumed
the maximum design capacity was used.
5-52
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
a-Si/nanocrystalline (nc) silicon PV cells use F-GHGs, while other thin film technologies do not. Therefore, this
analysis was limited to the c-Si and a-Si PV cell markets.
Activity Data
Historical
The activity data for emission estimates from PV manufacturing were area (m2) of PV panels produced, which
was derived from maximum design capacities expressed in total peak power production, in megawatts (MW), for
each country and the world.
Historical maximum design capacities,31 in units of MW, were determined by the following various methods:
• 1990 and 1995: Maximum design capacities in these years were assumed to be 0 MW because the sector
was so small in this time period that any associated manufacturing emissions were assumed to be
negligible.
• 2000, 2005, and 2010: Maximum design capacities by country and for the world in 2000, 2005, and 2010
were extracted directly from the DisplaySearch database (DisplaySearch, 2009).
• 2015: Maximum design capacities by country and the world in 2015 were estimated from the
International Energy Outlook (EIA, 2017). It was assumed that the number of total PV panels installed in
2015, calculated based on the difference between the total PV capacity in 2015 and 2014, was the total
PV panels produced after taking into account assumed yield, fraction sold, and fraction inventoried. The
relationship is provided in the equation below:
Maximum Design Capacity (MW) = Panels Yield *
(1+Fraction of Panels Inventoried/Fraction of Panels Sold) * Total PV Installed (MW) (5.9)
Maximum design capacity was converted to area of produced PV panels (m2) using technology-specific and time-
varying market shares (of substrates types) and average electrical conversion efficiencies for c-Si and a-Si and the
expected power produced per unit of solar power absorbed at the Earth's equator at noon (0.001 W/m2). The
equation used for this conversion is as follows:
Area of PV Panel Produced (m2) = Maximum Design Capacity (MW) / [£t(Market Share of
Technology t (%) * Average Electrical Conversion Efficiency of Technology t (%)) * Expected Power
Produced (.001 MW/m2)] (5.10)
where t = technology type
Technology market shares32 and average conversion efficiencies33 were determined using data from the
DisplaySearch database (DisplaySearch, 2009). In instances where data were not available to calculate these values
(i.e., DisplaySearch information was incomplete or for future years), technology conversion efficiencies and market
shares were assumed based on historical data and expert judgment.
31 Historical maximum design capacities included maximum design capacities for c-Si and a-Si, the two technologies that use
PFCs in their manufacturing processes.
32 For this report, technology market shares were calculated based on a PV market that was assumed to only include c-Si and a-
Si technologies.
33 Technology conversion efficiencies were supplied for some years for both c-Si and a-Si technologies in the DisplaySearch
database. For each year this information was supplied, a simple average of the available conversion efficiencies was taken for
each technology.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-53
-------
METHODOLOGY DOCUMENTATION
Projected
Projected maximum design capacities,34 in units of MW, were determined by the following methods:
• 2020 through 2040: World maximum design capacities in 2020 through 2040 were based on the global
installed solar generating capacity from ElA's International Energy Outlook reference case (EIA, 2017).
Projected global installed solar manufacturing capacity in any given year was assumed to be the
difference between the solar generating capacity between that year and the previous year.
• 2045 through 2050: World maximum design capacities in 2045 through 2050 were determined by
applying 5-year CAGRs for each time period. These CAGRs of design capacity for the 2040 through 2045
and 2045 through 2050 periods are assumed to be the same as the CAGR of new design capacity from
2035 through 2040.
Country-specific maximum design capacities were determined by applying capacity shares to world maximum
design capacity estimates. Country-specific capacity shares were held constant at 2015 levels through 2050.
Maximum design capacity was converted to area of produced PV panels (m2), using the conversion equation
described in the previous section. Technology market shares35 for the a-Si type were assumed to be 20% in 2050
and for the c-Si type were assumed to be 80%. For the in-between years, the technology market shares were
interpolated between the years 2015 and 2050. The average conversion efficiencies36 for a-Si type were assumed
to increase by 1% every 5 years and were assumed to hold constant for the c-Si type at 2015 levels through 2050.
Emission Factors
Historical and Projected
Area of PV panels (m2) for each country and the world were converted to emissions (MMTCChe) using the
emission factors (MMTCChe/m2) for c-Si and a-Si and the respective market shares of each technology in a given
year. CF4 and C2F6 are used during manufacture of c-Si PV cells. Tier 1 emission factors for both of these PFCs for
PV manufacturing were obtained from the 2006 IPCC Guidelines (IPCC, 2006).
NF3 is also used during manufacture of a-Si PV cells; however, no published emission factor for NF3 used
during PV manufacturing was identified. NF3 is used routinely for cleaning during the manufacture of a-Si PV cells,
and the emissions are not negligible, depending on emission abatement practices. Therefore, the EPA developed
an emission factor for NF3 using measured, unpublished NF3 usage and NF3 emission data for currently operating a-
Si PV manufacturing facilities.37
Emission Reductions in Baseline Scenario
No emission reduction assumptions were incorporated into the base case scenario.
34 Projected maximum design capacities included capacities for c-Si and a-Si, the two technologies that use PFCs in their
manufacturing processes.
35 For this report, technology market shares were calculated based on a PV market that was assumed to only include c-Si and a-
Si technologies.
36 Technology conversion efficiencies were supplied for some years for both c-Si and a-Si technologies in the DisplaySearch
database. For each year this information was supplied, a simple average of the available conversion efficiencies was taken for
each technology.
37 In developing an emission factor for NF3, EPA also considered using NF3 emissions from the manufacturing stage of solar PV
cells, from "Life-Cycle Nitrogen Trifluoride Emissions from Photovoltaics" by Fthenakis et al. (2010). However, given that this
emission factor considers abatement, EPA did not use it in this report because this report does not consider abatement in the
BAU scenario.
5-54
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Uncertainty
In developing global projections of PFC emissions from the PV sector, a broad perspective was adapted to
determine future capacity for manufacturing PV cells. This forecast was framed by the assumption that the
reference case from the International Energy Outlook took into account the advances in the use of natural gas and
coal along with the use of alternative renewable energy technologies—wind, hydro, geothermal, and solar thermal
technologies—that serve as alternatives to both conventional fossil fuels and PV solar. Pressure to develop sources
of clean, renewable energy is growing because of the increasing costs and risks of securing traditional energy
supplies, the increasing focus on ensuring energy reliability and resilience, the increasing need for more energy as
countries like China and India industrialize further, and a growing understanding of the environmental effects of
traditional sources of energy.
Although this perspective was useful in framing these projections, many uncertainties surround it. First and
foremost are uncertainties in future clean energy and GHG policy, which is one of the main drivers in the use of
renewable energy. Demand for renewable energy is highly dependent on the design of such policies, and what
these policies will look like and how they will be implemented in some developed nations, as well as developing
nations, are still unknown. Additionally, large-scale use of renewable energy to meet base-load power needs will
also rely on the use of battery storage, which is still an emerging technology.
Another uncertainty is a longer term shift away from centralized sources of electricity generation to more
distributed sources of electricity. It is this distributive benefit that gives solar, over the long term, an advantage
relative to other renewable energy sources. This advantage, however, might not become evident in trends until
2025 or sometime thereafter.
Lastly, emissions estimated in these projections do not explicitly consider PFC abatement. Abatement may
occur when point-of-use abatement systems are used at a manufacturing facility for PFCs. Additionally, all NF3
used during chamber cleaning passes through required silane abatement systems for safety purposes, which are
capable without modification of abating NF3 and more capable with some modification. Emission estimates are
sensitive to the use of abatement. This sensitivity may be considered in future versions of this report, when more
information about this newly emerging sector is available.
5.2.3.2 Mitigation Options Considered for Photovoltaic Cell Manufacturing
Four mitigation technologies were considered in this analysis as options for reducing F-GHG emissions from PV
manufacturing: thermal abatement systems, catalytic abatement systems, plasmas abatement systems, and the
NF3 remote chamber clean process. Because of the lack of mitigation cost information specific to PV production,
data were drawn from experience reducing emissions from similar processes in semiconductor manufacturing.
• Thermal Abatement: These point-of-use abatement systems that use heat to destroy or remove F-GHGs
from effluent process streams are connected directly to a manufacturing tool.
• Catalytic Abatement: Tool effluent process streams are run through abatement systems with catalysts
(e.g., CuO, ZnO, Al203) that destroy or remove F-GHGs.
• Plasma Abatement: Plasma in a point-of-use abatement system is used to react (destroy or remove) F-
GHGs from the process effluent stream.
• NF3 Remote Chamber Clean: Highly ionized NF3 is used to clean CVD chambers. This process is highly
efficient (~98%), resulting in lower emissions on a mass and C02 basis than traditional in situ chamber-
cleaning processes with utilization efficiencies around 20% to 50% (IPCC, 2006).
Thermal Abatement
Thermal abatement systems can be used to abate F-gas emissions from both etching and chamber-cleaning
processes. The use of thermal abatement offers the benefit of not affecting the manufacturing process (Applied
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-55
-------
METHODOLOGY DOCUMENTATION
Materials, 1999); however, the systems do require space that may not be available in some facilities. In addition,
these systems require large amounts of cooling water, and the use of the systems results in regulated NOx
emissions.
The total facility capital cost for installing thermal abatement systems is estimated to be $6.3 million. This
estimate includes costs for the systems, the necessary ducting, water recirculation and hookup, and natural gas
costs (Fthenakis, 2001; Burton, 2003). The annual operating cost is estimated to be $361,967 at the facility level.
No annual cost savings are associated with using this technology.
Catalytic Abatement
A catalytic abatement system uses a catalyst to destroy or remove F-gas emissions from the effluents of both
plasma etching and CVD chamber-cleaning processes. This type of abatement is applicable at most facilities, but
again there may be some space constraints as mentioned for thermal abatement systems. Additionally, because
these systems are based on destruction via catalyst, they must be process/stream specific to achieve the 99%
emission reductions quoted in the literature and used in this analysis (Fthenakis, 2001; Burton, 2003).
Because catalytic destruction systems operate at relatively low temperatures, their use results in little or nox
emissions, and the required amounts of water are low as well. Because of the high cost of catalyst replacement,
these systems are the least widely used type of abatement (expert judgment).
The capital cost associated with purchasing and installing the abatement systems is estimated to be $7.6
million per fab (i.e., facility). To use catalytic abatement systems, facilities must factor in the annual cost of
resources such as water, waste chemicals, electricity, and catalyst replacements. To cover these operating
expenses, a facility manufacturing PV cells is estimated to incur an annual cost of $501,109. As with other
abatement technologies considered in this sector, the use of catalytic abatement systems will not result in annual
cost savings.
Plasma Abatement
These systems are applicable to etch processes in most facilities, with some physical space limitations. (These
systems though are relatively smaller in size compared with thermal and catalytic systems.) Plasma abatement
systems use a small plasma source that effectively dissociates the F-gas molecules that react with fragments of the
additive gas (hydrogen, oxygen, water, or CH4) to produce low molecular weight by-products such as HF with little
or no GWP. After disassociation, wet scrubbers can remove the molecules. The presence of additive gas is
necessary to prevent later downstream reformation of PFC molecules (Burton, 2003).
The capital cost for plasma abatement systems is estimated to be $2.0 million per facility (Burton, 2003),
which covers the purchase and installation of plasma systems. Plasma abatement systems require an annual
operation cost of $1,304 per chamber, which includes general maintenance and use of the systems. Total annual
facility costs are $57,067 based on an assumed four chambers per tool and 25 tools per facility. The use of plasma
abatement systems will not result in annual cost savings.
NF3 Remote Chamber Clean
NF3 remote chamber clean is an alternative cleaning technology that has the benefit of having a particularly
high (~98%) utilization rate of NF3 (IPCC, 2006), resulting in relatively low emissions compared with traditional
chamber cleans. (The stated utilization is based on uses for semiconductor manufacturing; as a result of large gas
flows in PV manufacturing, the actual utilization may be lower.) NF3 remote clean systems dissociate NF3 using
argon gas and converting the source gas to active F-atoms in the plasma upstream of the process chamber. These
electrically neutral atoms can selectively remove material in the chamber. The by-products of remote clean include
HF, F2, and other gases, most of which are removed by facility acid scrubber systems.
5-56
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
We assumed that PV facilities are not "NF3 ready"; in other words, these facilities do not have the current
infrastructure to handle the direct installation of NF3 remote systems because this technology is relatively new.
Therefore, facilities incur capital costs, in addition to system costs, associated with items such as gas hookups and
necessary hardware such as manifolds and values. The facility cost is estimated to be $10.1 million. The annual
facility cost for NF3 remote clean is estimated to be $3.7 million (Burton, 2003). These costs are associated with the
purchase of larger volumes of gas (NF3 versus traditional chamber clean gases), general maintenance, and F2
scrubs to remove the highly explosive gas from the effluent. No annual cost savings are associated with using this
technology.
5.2.3.3 Model Facilities
The manufacture of PV uses F-GHGs depending on the substrate and process used in the production.
Substrates used in the industry include crystalline silicon, amorphous silicon, and other thin films. F-GHGs are used
during the manufacture of crystalline silicon (c-Si) PV cells, amorphous silicon (a-Si), and tandem a-
Si/nanocrystaline (nc) silicon PV cells. Other thin film PV technologies do not require the use of F-GHGs. As with
the other electronics manufacturing sectors, emissions in this sector result from two main types of manufacturing
processes: etching substrates and cleaning CVD chambers. Manufacturing processes and uses of GHGs across the
industry are generally similar; therefore, only one type of model facility was considered for this analysis.
• The model facility represents a PV manufacturing facility of average manufacturing capacity
(DisplaySearch, 2009) of 80 MW with an estimated 25 tools with 3.5 chambers. The facility uses only three
F-GHGs: CF4, C2F6, and NF3.38 The emission breakdown for a PV manufacturing facility is estimated to be
25% etch emissions and 75% clean emissions.
The model facility emission breakdown is important because some mitigation technologies are applicable to either
both or just one type of manufacturing process.
5.2.3.4 Technical and Economic Characteristics of Options
The mitigation technologies reduce emissions from either etch or chamber clean processes or in some cases
both. Table 5-18 presents the applicability and the reduction efficiency of each abatement measure.
Table 5-18: PV Cell Manufacturing Abatement Options
Abatement Option
Applicable
Reduction
Efficiency
Information Source
Thermal abatement
Etch and clean
95%
Fthenakis (2001), Beu (2005), and EPA (2009)
Catalytic abatement
Etch and clean
99%
Fthenakis (2001), Brown et al. (2012)
Plasma abatement
Etch
97%
Fthenakis (2001), Hattori et al. (2006)
NF3 remote chamber clean
Clean
95%
Beu (2005)
Similar to the methods employed for analyzing abatement in the semiconductor manufacturing sector, this
analysis developed a technical effectiveness parameter, defined as the percentage reductions achievable by each
technology/process combination. Estimating this parameter required making a number of assumptions regarding
the distribution of emissions by manufacturing process (etch and clean), in addition to making process-specific
estimates of technical applicability and market penetration. We held these assumptions constant for all model
38 Although these gases are used for different PV technologies, for simplicity in this analysis, one general facility producing an
unidentifiable PV technology was considered.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-57
-------
METHODOLOGY DOCUMENTATION
years. Table 5-19 presents the technical applicability, market penetration, and reduction efficiency assumptions
used to develop the abatement measures' technical effectiveness parameters.
Table 5-19: Technical Effectiveness Summary—PV Cell Manufacturing
Etch (25%)
Clean (75%)
Abatement
Measure
Technical
Applicability
Market
Penetration
Technical
Applicability
Market
Penetration
Reduction
Efficiency
Technical
Effectiveness
Thermal
abatement
85%
65%
85%
20%
95%
25%
Catalytic
abatement
85%
10%
85%
10%
99%
8%
Plasma abatement
85%
25%
0%
0%
97%
5%
NF3 remote clean
0%
0%
100%
70%
95%
50%
The technical effectiveness is the weighted average of the abatement measures using the emissions attributed
to each process (i.e., 25% etching, and 75% cleaning) as the weight multiplied by the product of the technical
applicability, market penetration, and reduction efficiency for each abatement measure. We then multiplied the
technical effectiveness estimates by the share of total emissions to estimate the abatement potential achievable
under each abatement measure. Table 5-20 summarizes the information used to estimate the break-even prices in
the MAC analysis.
Table 5-20: Engineering Cost Data on a Facility Basis—PV Cell Manufacturing
Abatement Option
Project
Lifetime
(years)
Capital Costs
(2016 USD)
Annual
Revenue
(2016 USD)
Annual 0&M
Costs
(2016 USD)
Abatement
Amount
(tC02e)
Thermal abatement
7
$6,275,974
$0
$361,967.33
4,802
Catalytic abatement
7
$7,601,864
$0
$501,108.98
1,601
Plasma abatement
7
$1,997,342
$0
$ 57,066.92
985
NF3 remote clean
25
$10,127,095
$0
$3,714,600.05
9,492
5.2.3.5 References
Applied Materials. October 18,1999. Catalytic Abatement ofPFC Emissions. Presented at Semicon Southwest 99:
Partnership for PFC Emissions Reductions, Austin, TX.
Beu, L December 2005. Reduction of Perfluorocarbon (PFC) Emissions: 2005 State-of-the-Technology Report.
TT#0510469A-ENG. Albany, NY: International SEMATECH Manufacturing Initiative. Available online at
http://www.epa.gov/highgwp/semiconductor-pfc/documents/final tt report.pdf
Brown et al., 2012. Catalytic technology for PFC emissions control. Solid State Technology. Available online at
http://www.electroiq.com/content/eiq-2/en/articles/sst/print/volume-44/issue-7/features/emissions-
control/catalvtic-technology-for-pfc-emissions-control.html cf
Burton, S. 2003. Personal communication with Brown, Motorola (2002) supplemented by personal communication
with Von Gompel, BOC Edwards (2003), research on DuPont's Zyron Web site (2003), and personal
communication with Air Liquide regarding thermal destruction, NF3 remote clean, and capture membrane unit
costs.
5-58
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
DisplaySearch. 2009. DisplaySearch Q4'09 Quarterly PV Cell Capacity Database & Trends Report. DisplaySearch,
LLC.
Fthenakis, V. December 2001. Options for Abating Greenhouse Gases from Exhaust Streams. Upton, NY.
Brookhaven National Laboratory. Available online at http://www.bnl.gov/isd/documents/23784.pdf
Fthenakis, V., D. Clark, M. Moalem, P. Chandler, R. Ridgeway, F. Hulbert, D. Cooper, P. Maroulis. 2010. "Life-Cycle
Nitrogen Trifluoride Emissions from Photovoltaics". Environ. Sci. Technology. 2010 44 22 8750-8757.
doi:10.1021/esl00401y.
Hattori, K., K. Sakurai, N. Watanabe, H. Mangyou, S. Hasaka, and K. Shibuya. September 25-27, 2006. Application
of atmospheric plasma abatement system for exhausted gas from MEMS etching process. Presented at the
Institute of Electrical and Electronics Engineers International Symposium on Semiconductor Manufacturing,
Tokyo, Japan.
Intergovernmental Panel on Climate Change. 2006. 2006IPCC Guidelines for National Greenhouse Gas Inventories.
The National Greenhouse Gas Inventories Programme (Volume 3, Chapter 6: Electronics Industry Emissions),
H.S. Eggleston, L. Buendia, K. Miwa, T. Ngara, and K. Tanabe (eds.). Hayama, Kanagawa, Japan.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-59
-------
METHODOLOGY DOCUMENTATION
5.2.4 PFC Emissions from Flat Panel Display Manufacturing
Flat panel display (FPD), namely liquid crystal display (LCD) panels, are components of screens used in
consumer electronics, computers, and mobile devices. FPD manufacturing is a relatively new industry that
continues to grow as demand for these screens increases. The FPD manufacturing process uses SF6, PFCs, including
CF4, and NF3 in the etching and chamber cleaning processes. These F-GHGs are used for CVD cleaning processes
and plasma dry etching during manufacture of arrays of thin-film transistors on glass substrates, which switch
pixels of LCDs and organic light-emitting diode displays. Various F-GHGs play different roles in each of these
processes (e.g., CF4 is mainly used for etching, while SF6 and NF3 are used for both etching and cleaning) and are
selected by manufacturers based on their chemical and physical properties and ability to efficiently perform in the
process (EPA, 2017). In general, almost all manufacturers used SF6, NF3, and CF4, with a smaller mix of other HFCs
and PFCs in some instances.
5.2.4.1 Flat Panel Display Manufacturing Projections Methodology
UNFCCC-reported, country-specific estimates were not available for any historical years for this emission
source. Therefore, the 2006 IPCCTier 1 methodology (IPCC, 2006) was used to estimate emissions. Activity data for
flat panel display (FPD) manufacturing include maximum design capacities and input area of manufacturing FPDs
from the DisplaySearch database (2009 and 2016). Emission factors were obtained from IPCC (2006) and calculated
from CDP-reported data. The Tier 1 basic equation to estimate HFC, PFC, SF6, and NF3 emissions from FPD
manufacturing is as follows:
FQ = EFi * Cu * Cd (5.11)
where:
FO
EFi
Cu
Cd
Activity Data
Historical
The main sources of data for this source category are the DisplaySearch Q4'09 Quarterly FPD Capacity
Database & Trends report ("DisplaySearch database") (DisplaySearch, 2009), and DisplaySearch Q4 2016 Large
Area Display Production Strategy Tracker (DisplaySearch, 2016). These databases supply historical and projected
annual data through 2017 about all FPD facilities in the world, including location (country), maximum design
capacity for substrate processing of a facility (in 1,000 m2), and in some cases the utilized capacity of a facility
(percentage). DisplaySearch 2016 provides total input area of large-area displays from each manufacturer for 2012
through 2017.
For 2000 through 2010, "utilized capacity (m2) of FPD area produced" was the activity data used to estimate
emissions. The activity data were derived from maximum design capacity expressed in area (1,000 m2) for each
country and the world. Maximum design capacity was converted to utilized capacity (m2) by applying a utilized
capacity factor (%). For simplicity, a single, global average utilized capacity factor of 88% was applied to all
countries and to the world for all years. This factor was derived by taking a simple average of the world utilized
39 EPA assumed that Cu is equivalent to 88% for 1990 through 2010.
Emissions of gas / (mass)
Emission factor for gas / (mass/m2)
Fraction of annual plant production capacity utilization (%)39
Annual maximum design capacity of substrate processed (m2)
5-60
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
capacity factors (percentage) for all years provided in the DisplaySearch FPD database (DisplaySearch, 2009).40 For
2012 through 2017, the "input area of manufactured FPDs (m2)" was the activity data used to estimate emissions.
Total input area for 2012 through 2017 was estimated by adjusting the large-area panel input area from
DisplaySearch (2016) based on the share of small and medium panels from each manufacturer in 2012 from
DisplaySearch (2009).
Total maximum design capacities were determined by the following various methods:
• 1990,1995: Total world maximum design capacities in 1990 and 1995 were determined by applying 5-
year global CAGRs for each period. These 5-year global CAGRs were assumed based on expert judgment
about past demand in the FPD market.
where:
Using the world maximum design capacity estimate for each year and country-specific shares of world
capacity ("capacity shares"), the EPA determined country-specific CAGRs for each 5-year interval using the
following equation:
1/5
-1 (5.12)
Country5YearCAGRj =
WorldManufacturedCapacity(Yf) * CountryShare! (Yf)
WorldManufacturedCapacity(Yo) * CountryShare! (Yo)
Y/ = Future year
Yo = Initial year
/ = Country index
Country-specific capacity shares for 1990 and 1995 were assumed to be equivalent to the 2000 country-
specific capacity shares, determined by using country and world capacity data extracted from the
DisplaySearch database (DisplaySearch, 2009).
Maximum design capacity was back casted for each country for the years 1990 and 1995 by applying a
country-specific 5-year CAGR to maximum design capacity in the appropriate adjacent time frame.
• 2000, 2005, and 2010: Total maximum design capacities were determined by country and for the world in
2000, 2005, and 2010 directly from the DisplaySearch database (DisplaySearch, 2009).
As noted above, once total maximum design capacities were determined for each country and the world
for the 1990 through 2010 time series, these values were converted to utilized capacity (m2) using a world
average utilized capacity factor.
• 2015: Total large-area display input area was obtained from DisplaySearch2016 and adjusted to account
for small and medium area display area. The input area used in this time series was the same as the
"utilized capacity" in historical time series.
Projected
• 2020 through 2050: Total display input area was determined starting from the last available
DisplaySearch2016 data and grown using the historical annual growth rate of 8.5%, the average of the
growth rates between 2012 and 2017, until 2025. For the time frame of 2025 through 2050, a more
conservative growth rate of 4.3% was applied, representing half of the growth rate used until 2025.
40 In the DisplaySearch FPD database, capacity utilizations (%) were only available for the years 2005 through 2010. The
capacity utilization provided for the world in each of these years was simply averaged together to get the capacity utilization
factor used in this analysis (88%). While the DisplaySearch databases provided some country-specific capacity utilizations for
specific fabs in a country, there were many gaps in these data. Therefore, using the database may have led to an
underestimation of actual emissions.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-61
-------
METHODOLOGY DOCUMENTATION
Emission Factors
Historical and Projected
To determine emissions for each country, the EPA converted the total utilized capacity and total input area
(m2) to PFC, HFCs, NF3, and SF6 emissions (MMTCChe) using:
• IPCC Tier 1 emission factors for PFCs, NF3, and SF6 (MMTCChe/m2) (IPCC, 2006) and
• the calculated emission factor for HFC using emission data reported to the CDP (formerly the Carbon
Disclosure Project). The EPA used reported emission data for one representative facility to develop an HFC
emission factor (MMTCChe/m2), which was applied for each country.
Emission Reductions in Baseline Scenario
Without incentives and or emission targets, it was assumed that the FPD sector does not employ additional
abatement technologies beyond what had been installed as of 2012 (TTLA, 2012). To reduce emissions, this sector
may employ abatement technologies, including fueled combustion, plasma and catalytic technologies explicitly
intended for F-GHG abatement.
Uncertainty
These global emission projections are highly sensitive to the assumption that China's domestic demand for
FPDs will substantially increase in the future (DisplaySearch, 2016), thereby increasing Chinese domestic capacity
and production of FPDs and hence increasing emissions. If actual domestic demand in China varies in the future,
China's large contribution to global emissions may change.
5.2.4.2 Mitigation Options Considered for Flat Panel Display Manufacturing
We identified six mitigation technology options for the FPD manufacturing sector: central abatement, thermal
abatement, catalytic abatement, plasma abatement, NF3 remote chamber clean, and gas replacement.
• Central Abatement: These large-scale abatement systems are generally located on the roof of facilities
and are applicable to etch emissions (SFs).
• Thermal Abatement: These point-of-use (POU) abatement systems, which use heat to destroy or remove
F-GHGs from effluent process streams, are connected directly to a manufacturing tool.
• Catalytic Abatement: Tool effluent process streams are run through POU abatement systems with
catalysts (e.g., CuO, ZnO, Al203) that destroy or remove F-GHGs.
• Plasma Abatement: Plasma in a POU abatement system is used to react (destroy or remove) F-GHGs from
the process effluent stream.
• NF3 Remote Chamber Clean: Highly ionized NF3 is used to clean CVD chambers. This process is highly
efficient (using ~98% of the gas in a process), resulting in lower emissions on a mass and C02 basis than
traditional in situ chamber clean processes that use approximately 20% to 50% of the gas in a process
(IPCC, 2006).
• Gas Replacement: Higher GWP gases are replaced with lower GWP gases.
Central Abatement
Central abatement systems (CASs) have begun to be designed and used to handle the generally high gas (SFs)
flows used in FPD manufacturing because of the large size of the substrate being etched. A CAS is a large-scale
thermal oxidation abatement system that is located on the roof of FPD facilities, so there are few expected space
limitations for this technology. This technology has recently started to come online and is only applicable to etch
emissions. Two CDM projects (one from LG and one from Samsung) in Korea have used this technology (CDM
5-62
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Project #3440 and CDM project #3333). Its use is, however, limited throughout the rest of the industry because it is
expensive and relatively new.
The capital cost for a CAS is $4.9 million. The annual O&M cost, which includes items such as utilities and
parts, is estimated to be $2.8 million for a facility. No revenues are generated from using a CAS.
Thermal Abatement
Thermal abatement systems can be used to abate emissions from both etching and CVD chamber-cleaning
processes. The use of thermal abatement offers the benefit of not affecting the manufacturing process (Applied
Materials, 1999); however, the systems do require space that may not be available in some facilities. In addition,
these systems require large amounts of cooling water, and the use of the systems results in regulated NOx
emissions.
The total facility capital cost for installing thermal abatement systems is estimated to be $6.3 million. This
estimate includes costs for the systems, the necessary ducting, water recirculation and hook up, and natural gas
costs (Fthenakis, 2001; Burton, 2003). The annual operating cost is estimated to be $361,967 at the facility level.
No annual savings are associated with using this technology.
Catalytic Abatement
A catalytic abatement system is used to abate emissions from both etching and CVD chamber-cleaning
processes. This type of abatement is applicable at most facilities, but again there may be some space constraints,
as noted for thermal abatement systems, thus limiting the use of these systems in the market. Another limitation
to their use is high catalyst replacement costs.
The capital cost associated with purchasing and installing the abatement systems is estimated to be $7.6
million per facility. To use catalytic abatement systems, facilities must factor in the annual cost of resources such
as water, waste chemicals, electricity, and catalyst replacements. To cover these operating expenses, a facility
manufacturing FPDs is estimated to incur an annual cost of $501,107. As with other abatement technologies
considered in this sector, the use of thermal abatement systems will not result in annual savings.
Plasma Abatement
Plasma abatement systems are assumed to be applicable to etch processes in most facilities, with some
physical space limitations. (These systems, however, are relatively smaller in size compared with thermal and
catalytic systems.) Plasma abatement systems use a small plasma source that effectively dissociates the F-GHG
molecules that react with fragments of the additive gas (hydrogen, oxygen water, or CH4) to produce low
molecular weight by-products such as HF with little or no GWP. After disassociation, wet scrubbers can remove the
molecules. The presence of additive gas is necessary to prevent later downstream reformation of F-gas molecules
(Motorola, 1998).
The capital cost for plasma abatement systems is estimated to be $1.9 million per facility (Fthenakis, 2001;
Burton, 2003), which covers the purchase and installation of plasma systems. Plasma abatement systems require
an annual operation cost of $1,304 per chamber, which includes general maintenance and use of the systems. The
total annual facility cost is $114,114 based on an assumed number of tools per facility and chambers per tool. The
use of plasma abatement systems will not result in annual cost savings.
NF3 Remote Chamber Clean
NF3 remote chamber clean is an alternative cleaning technology that has the benefit of having a particularly
high utilization rate of NF3 (~98%; IPCC, 2006), resulting in relatively low emissions compared with traditional
chamber cleans. (Note: The stated utilization is based on uses in semiconductor manufacturing; as a result of large
gas flows in FPD manufacturing, the actual utilization may be lower.) NF3 remote clean systems dissociate NF3 using
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-63
-------
METHODOLOGY DOCUMENTATION
argon gas, converting the source gas to active F-atoms in the plasma upstream of the process chamber. These
electrically neutral atoms can selectively remove material in the chamber. The by-products of remote clean include
HF, F2, and other gases, most of which are removed by facility acid scrubber systems. The use of NF3 remote clean
systems is much more prevalent in new facilities because the technology was not available when many old
facilities were constructed.
We assumed FPD facilities are not "NF3 ready"; in other words, these facilities do not have the current
infrastructure to handle the direct installation of NF3 remote systems because this technology is relatively new.
Therefore, facilities incur capital costs, in addition to system costs, associated with items such as gas hookups and
necessary hardware such as manifolds and values. The facility cost is estimated to be $10.1 million. The annual
facility cost for NF3 remote clean is estimated to be $3.7 million (Burton, 2003). This cost is associated with the
purchase of larger volumes of gas (NF3 vs. traditional chamber clean gases), general maintenance, and F2 scrubs to
remove the highly explosive gas from the effluent. No annual cost savings are associated with using this
technology.
Gas Replacement
Gas replacement can be used to mitigate emissions from the traditional CVD chamber-cleaning process. Gas
replacement can be applied in most facilities and has already been used throughout the industry in many
instances. For this strategy, a lower GWP gas replaces a higher GWP gas. The most common replacement is using
NF3 to replace SFs.
Facilities replacing SFs with NF3 incur an estimated capital cost of $1.3 million for items such as gas hookups
and implementation. Annual savings for this option result from the lower cost of the replacement gas and are
estimated to be $47,611, based on the incremental cost of the gases and the average amount of gas consumed per
facility. Gas replacement has no operational costs.
5.2.4.3 Model Facilities
The manufacture of flat panels uses F-gases. Emissions in this sector result from two main types of
manufacturing processes: etching substrates and cleaning CVD chambers. Manufacturing processes and uses of
GHGs across the industry are generally similar; therefore, only one type of model facility was considered for this
analysis.
• The model facility represents an average flat panel display manufacturing facility. The average
manufacturing capacity (DisplaySearch, 2009) of a fabrication lab is 760,839.9 m2 and is estimated to have
25 tools with four chambers. The facility only uses CFU, SF6, and NF3. Model facilities are NF3 remote clean
ready. The emission breakdown for a FPD manufacturing facility is estimated to be 23% etch emissions
and 77% clean emissions.
The model facility emission breakdown is important because some mitigation technologies are applicable to either
both or just one type of manufacturing process.
5.2.4.4 Technical and Economic Characteristics Summary
This section describes the mitigation options in detail and includes technical and economic data that were
used to calculate the break-even price and reduction potential for each option.
Table 5-21 presents the reduction efficiency and the applicability of each mitigation technology to emissions
from a type of process (etch and/or clean).
5-64
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-21: FPD Manufacturing Abatement Options
Abatement Option
Applicable Process
Emission Type(s)
Reduction
Efficiency
Information Source
Central abatement
Etch
77%
CDM project #3333
Thermal abatement
Etch and clean
95%
Fthenakis (2001), Beu (2005), and EPA (2009)
Catalytic abatement
Etch and clean
99%
Fthenakis (2001), Brown et al. (2012)
Plasma abatement
Etch
97%
Fthenakis (2001), Hattori et al. (2006)
NF3 remote chamber clean
Clean
95%
Beu (2005)
Gas replacement
Clean
77%
CDM methods NM0289, NM303, NM0317,
NM0335
Similar to the methods employed for analyzing abatement in the semiconductor and photovoltaics
manufacturing sectors, this analysis developed a technical effectiveness parameter, defined as the percentage
reductions achievable by each technology/process combination. Estimating this parameter requires assumptions
regarding the distribution of emissions by manufacturing process (etch and clean), in addition to process-specific
assumptions on technical applicability and market penetration. These assumptions are held constant for all model
years in the MAC analysis. Table 5-22 presents the technical applicability, market penetration, and reduction
efficiency assumptions used to develop the abatement measures' technical effectiveness parameters.
Table 5-22: Technical Effectiveness Summary—Flat Panel Display Manufacturing
Etch (23%)
Clean (77%)
Abatement
Measure
Technical
Applicability
Market
Penetration
Technical
Applicability
Market
Penetration
Reduction
Efficiency
Technical
Effectiveness
Central abatement
100%
40%
0%
n/a
77%
7%
Thermal abatement
85%
30%
85%
55%
95%
40%
Catalytic abatement
85%
10%
85%
15%
99%
12%
Plasma abatement
85%
20%
0%
n/a
97%
4%
NF3 remote clean
0%
n/a
100%
20%
95%
15%
Gas replacement
0%
n/a
50%
10%
77%
3%
The technical effectiveness is a weighted average of the abatement measure's emission reductions when
applied to each applicable process(es). The share of total emissions attributed to each process (i.e., 23% etching
and 77% cleaning) is the weight that is multiplied by the product of the technical applicability, market penetration,
and reduction efficiency for each abatement measure. The technical effectiveness estimates are then multiplied by
the facility annual emissions to estimate the abatement potential achievable through each of the six abatement
measures. Summing the technical effectiveness across the six abatement measures yields the maximum level of
emission reductions that is technically achievable.
Table 5-23 summarizes the engineering cost data and abatement potential for each abatement option
considered in this analysis.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-65
-------
METHODOLOGY DOCUMENTATION
Table 5-23: Engineering Cost Data on a Facility Basis—Flat Panel Display Manufacturing
Abatement Option
Project
Lifetime
(years)
Capital Costs
(2010 USD)
Annual
Revenue
(2010 USD)
Annual 0&M
Costs
(2010 USD)
Abatement
Amount
(tCOe)
Central abatement
15
$4,939,147
$0
$2,803,407
5,889
Thermal abatement
7
$6,275,974
$0
$361,967
32,749
Catalytic abatement
7
$7,601,864
$0
$501,109
9,600
Plasma abatement
7
$1,997,342
$0
$114,134
3,167
NF3 remote clean
21
$10,127,096
$0
$3,714,600
12,029
Gas replacement
21
$1,298,791
$47,611
$0
2,427
5.2.4.5 References
Applied Materials. October 18,1999. Catalytic Abatement ofPFC Emissions. Presented at Semicon Southwest 99:
Partnership for PFC Emissions Reductions, Austin, TX.
Beu, L December 2005. Reduction of Perfluorocarbon (PFC) Emissions: 2005 State-of-the-Technology Report.
TT#0510469A-ENG. Austin, TX: International SEMATECH Manufacturing Initiative.
Brown S., et al. 2012. Catalytic technology for PFC emissions control. Solid State Technology. Available online at
http://www.electroiq.com/content/eiq-2/en/articles/sst/print/volume-44/issue-7/features/emissions-
control/catalvtic-technology-for-pfc-emissions-control.html cf
Burton, S., 2003. Personal communication with Brown, Motorola (2002) supplemented by personal communication
with Von Gompel, BOC Edwards (2003), research of DuPont's Zyron Web site (2003), and personal
communication with Air Liquide regarding thermal destruction, NF3 Remote Clean, and Capture Membrane
unit costs.
DisplaySearch. 2009. DisplaySearch Q4'09 Quarterly FPD Supply/Demand and Capital Spending Report Database.
DisplaySearch, LLC.
DisplaySearch. 2016. DisplaySearch Q4'16 Large Area Display Production Strategy Tracker. DisplaySearch, LLC.
Fthenakis, V. December 2001. Options for Abating Greenhouse Gases from Exhaust Streams. Upton, NY:
Brookhaven National Laboratory. Available online at http://www.bnl.gov/isd/documents/23784.pdf
Hattori, K., K. Sakurai, N. Watanabe, H. Mangyou, S. Hasaka, and K. Shibuya. September 25-27, 2006. Application
of atmospheric plasma abatement system for exhausted gas from MEMS etching process. Presented at the
Institute of Electrical and Electronics Engineers International Symposium on Semiconductor Manufacturing
2006, Tokyo, Japan.
Intergovernmental Panel on Climate Change. 2006. 2006IPCC Guidelines for National Greenhouse Gas Inventories.
The National Greenhouse Gas Inventories Programme (Volume 3, Chapter 6: Electronics Industry Emissions),
The Intergovernmental Panel on Climate Change, H.S. Eggleston, L. Buendia, K. Miwa, T. Ngara, and K. Tanabe
(eds.). Hayama, Kanagawa, Japan.
Motorola. October 18,1998. Long-term evaluation of litmus "blue" inductively-coupled plasma device for point-of-
use PFC and HFC abatement. Presented at Semicon Southwest 99: A Partnership for PFC Emissions Reductions,
Austin, TX.
Taiwan TFT LCD Association. 2012. PFCs Reduction Experiences in TTLA.
U.S. Environmental Protection Agency. 2017. Center for Corporate Climate Leadership Sector Spotlight: Electronics.
Available online at https://www.epa.gov/climateleadership/center-corporate-climate-leadership-sector-
spotlight-electronics.
5-66
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.2.5 SFe Emissions from Electric Power Systems
SFs is used as both an arc-quenching and insulating medium in electrical transmission and distribution
equipment. SF6 emissions from electrical equipment used in transmission and distribution systems occur through
leakage and handling losses. Leakage losses can occur at gasket seals, flanges, and threaded fittings and are
generally larger in older equipment. Handling emissions occur when equipment is opened for servicing, SF6 gas
analysis, or disposal. The manufacture of equipment for electrical transmission and distribution can also result in
SFs emissions, but this source is not included in this report.41
Several factors affect SF6 emissions from electrical equipment, including the type and age of SF6-containing
equipment, and the handling and maintenance protocols used by electric utilities. Historically, approximately 20%
of total global SF6 sales have been attributed to electric power systems, where the SF6 is believed to have been
used primarily to replace emitted SF6. Approximately 60% of global sales have gone to manufacturers of electrical
equipment, where the SF6 is believed to have been mostly banked in new equipment (Smythe, 2004).
5.2.5.1 Electric Power Systems Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this category,
when available. The 2006 IPCC Tier 1 methodology (IPCC, 2006) was not well suited to estimate emissions from
electric power systems given the type of data available on global SF6 use. Instead, if country-reported emission
estimates were not available for any historical years, estimates were calculated based on a methodology derived
from an equation in the IPCC Good Practice Guidance (IPCC, 2000). Activity data included a RAND survey of global
SFs sales to electric utilities and equipment manufacturers from Knopman and Smythe (2007), estimates of net
electricity generation and consumption from ElA's International Energy Outlook (2017), and a report of emissions
from equipment use and decommissioning in the European Union from Ecofys (2005). The successful attainment of
developed country SF6 reduction goals are accounted in the emission projections. Methods contained in the 2006
IPCC Guidelines for National Greenhouse Gas Inventories (IPCC, 2006) were not well suited to estimate emissions
from electric power systems given the type of data available on global SF6 use. Instead, estimates were calculated
using the following equation, which was derived from the equation for emissions in the IPCC Good Practice
Guidance (IPCC, 2000):
Emissions = SF6 purchased to refill existing equipment + nameplate capacity of retiring equipment (5.13)
This equation holds true regardless of whether the gas is released or recovered from retiring equipment.
Recovered gas is used to refill existing equipment, lowering the amount of SF6 purchased by utilities for this
purpose.
The primary driving factors for determining historical emissions in the electric power systems methodology
are the source data from UNFCCC, RAND, and EIA. For determining projected emissions, the primary factor in
determining these estimates is the methodology for the projection calculation, described in-depth below.
Activity Data
Historical
RAND, EIA, and Ecofys data were used to estimate historical emissions in the absence of country-level
estimates from UNFCCC. First, total world SF6 emissions were estimated using data from the 2007 RAND survey
(Knopman and Smythe, 2007). The data included total gas purchases by utilities and equipment manufacturers
41 Although these emissions were not explicitly estimated in this study, some A1 countries report emissions from the
manufacture of equipment for electrical transmission and distribution equipment manufacture within this source category. In
these cases, this source category includes these emissions.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-67
-------
METHODOLOGY DOCUMENTATION
from 1961 through 2006 (believed to include all SF6-consuming countries except Russia and China). From these
data, the two components of the above emission equation were calculated using the following assumptions:
1. SFs emitted by utilities in a given year is approximately equal to SF6 purchased to refill existing equipment
in a given year.42,43
2. The nameplate capacity of retiring equipment in a given year was assumed to be equal to 77.5% of the
amount of gas purchased by electrical equipment manufacturers 30 years previous (e.g., in 1990, the
nameplate capacity of retiring equipment is equal to 77.5% of the gas purchased by original equipment
manufacturers in 1960).44The remaining 22.5%45 was assumed to have been emitted at the time of
manufacture.
The RAND survey data were adjusted to include sales for China and Russia, which were not included in the
survey. To make this adjustment, it was assumed Russian and Chinese SF6 sales were proportional to the net
electricity consumption of these countries available from EIA (2017). The total sales for the countries represented
in the RAND survey were multiplied by the proportion of China's and Russia's net electricity consumption relative
to the world net electricity consumption. A 3-year smoothing was applied to the sum of the results to reduce the
potential impact of inventory fluctuations on the estimates.
Next, emission estimates were developed for all countries from 1990 through 2006 using the RAND world SF6
emission totals. Known emission values for the United States, the EU,46 and Japan47 were subtracted from the
world emission total. The remaining emissions were allocated to countries according to each country's share of net
electricity consumption to the world net electricity consumption minus consumption from the United States, the
EU, and Japan (EIA, 2017).
For 2010 historical emissions, each country's 2006 emission estimates were extrapolated based on the change
in world net electricity consumption from 2006 through 2010, as provided by EIA (2017). It was necessary to use
extrapolation for 2010 emissions because RAND ceased publication of its survey in 2007, so 2006 was the last year
for which RAND survey data were available.
Projected
Global emission projections were developed using the following approaches:
• Projections were extrapolated from a country's last reported estimate using a growth rate based on
estimates from the last 5 years of reported estimates. The extrapolation method was used for the United
States based on the 2015 estimate (EPA, 2016) and for EU countries based on the 2015-reported
estimates from Ecofys (2005). The growth rates reflect a declining growth rate in emissions and hence a
decline in forecasted emissions for these countries.
42 Communications with electrical equipment manufacturers indicated that beginning in the late 1990s, a small but increasing
fraction of new equipment was being filled with gas purchased by utilities rather than by equipment manufacturers. In this
analysis, EPA assumed that in 1999 1% of new equipment was filled using gas purchased by utilities and by 2003 this fraction
had grown to 5%. This assumption has the effect of decreasing estimated global refills and emissions by 11% in 2003.
43 See the country-by-country emissions section for information on how emissions were estimated for Russia and China.
44 The volume of SF6 sold for use in new equipment before 1961 was assumed to have increased linearly from 0 tons in 1950 to
91 tons in 1961, the first year for which the RAND survey has data.
45 The 22.5% emission rate is an average of IPCC SF6 emission rates for Europe and Japan before 1996 (IPCC, 2000).
46 EU emissions were based on those provided for equipment use and decommissioning in Reductions ofSF6 Emissions from
High and Medium Voltage Electrical Equipment: Final Report to CAPIEL (Ecofys, 2005). The Ecofys study relied on bottom-up
estimates of emission rates and of the SF6 bank in equipment, both of which varied by region and over time. A newer report
was published by Ecofys (2010); however, specific country-level estimates were not presented in a table format.
47 Historical emissions data for Japan used in this analysis were available through the UNFCCC flexible query system (UNFCCC,
2016).
5-68
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• For Japan, Canada, Australia, the rest of Europe, Hong Kong, Singapore, and Eurasia, projections were
based on holding the 2006 emission estimate constant for the remainder of the time series. The
underlying assumption for this approach was that any emissions associated with system growth are
expected to be offset by decreases in the equipment's average SF6 capacity and emission rate as new,
small, leak-tight equipment gradually replaces old, large, leaky equipment. With regard to Japan, while
the SFs bank in Japan is expected to grow substantially in the future, it was assumed that, in addition to
adoption of newer equipment with smaller SF6 capacity, continued reduction measures would maintain
historical emission levels through the time series (Yokota et al., 2005).
• For all other countries, primarily developing countries, projections were extrapolated from the 2010
values using the country's electricity generation growth rate (EIA, 2017). Country-specific growth rates
were used when available; when not available, region-level growth rates were used. This approach was
used for most developing countries based on the assumption that these countries began to install SF6
equipment relatively recently. Consequently, as infrastructure expands, emissions from developing
countries are anticipated to grow at the same rate as country- or region-specific net electricity generation
projections.
Emission Reductions in Baseline Scenario
Since the mid to late 1990s, various developed countries have implemented voluntary (and in some cases
mandatory) programs aimed at reducing SF6 emissions from electric power systems. These countries include the
United States, Japan, and countries in the EU. The successful attainment of developed-country SF6 reduction goals
is accounted for in the emission projections; however, the EPA did not consider any enhanced future mitigation
from these programs beyond existing levels.
Uncertainties
In developing emission estimates for this source, the EPA used multiple international datasets and IPCC
guidance. The robustness of the bottom-up estimates is believed to have improved since the previous EPA-
published report (EPA, 2012) because of the use of updated RAND and EIA data in this version of the report.
Nevertheless, this analysis is subject to a number of uncertainties that affect both global- and country-specific
emission estimates, particularly estimates for countries other than the United States, Japan, and the EU.
First, the SF6 producers represented in the RAND survey do not represent 100% of global SF6 production and
consumption. The EPA accounted for unreported Russian and Chinese SF6 production, consumption, and emissions
by assuming a relationship between net electricity consumption and SF6 emissions. However, this assumption is
subject to uncertainty. One source of this uncertainty is the fact that net exports from or imports into Russia and
China affect the relationship between SF6 consumption and net electricity consumption in the rest of the world.
Net exports from Russia and China would make the "consumption factor" (SF6 consumption/net electricity
consumption) in the rest of the world appear to be smaller than it actually is, while net imports would have the
opposite impact. Information from manufacturers of electrical equipment indicates that exports from Russia and
China have fluctuated over time, peaking around the year 2000 and declining more recently. Thus, the apparent
dip in global emissions between 1995 and 2000, and the subsequent rise between 2000 and 2005, may be partly
an artifact of these export trends rather than purely a result of changes in emissions from electric power
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-69
-------
METHODOLOGY DOCUMENTATION
systems.48 Another source of uncertainty is that the relationship between SF6 emissions and net electricity
consumption varies from country to country, even when imports and exports are properly accounted for.49
Second, the RAND survey's attribution of SF6 sales to particular end uses is also uncertain, because SF6
producers frequently sell to distributors rather than directly to end users. Although producers would be expected
to have a reasonably good understanding of their markets, this understanding is not always accurate. Thus, some
of the SFs sales that the survey attributed to utilities could have actually been attributed to other uses or vice
versa.
Third, the typical lifetime of electrical equipment, and therefore the amount of equipment that is now being
retired, is uncertain. This analysis used a lifetime of 30 years; however, other publications have estimated the
lifetime at 40 years. The typical lifetime assumption is important because the amount of equipment manufactured
30 years ago is considerably larger than the amount of equipment manufactured 40 years ago. If the average
lifetime of equipment was assumed to be less than 30 years, then the estimate of 2006 global emissions would
decrease.
Fourth, for countries other than the EU, Japan, and countries that have reported to the UNFCCC, the EPA
assumed that each country's share of past and current global emissions is directly proportional to that country's
share of past and current global net electricity consumption. In fact, as noted above, the relationship between
emissions and electricity consumption varies between regions and over time, particularly as regions make efforts
to reduce their emission rates. Thus, there is an associated uncertainty in the allocation of global emissions to
individual regions within this analysis.
Fifth, emission estimates based on RAND sales data do not include SF6 emissions from electrical equipment
manufacturing. However, some of the UNFCCC-reported data that were used do include emissions from the
manufacture of electrical equipment.
Finally, emission projections were based on the assumption that emissions in developing countries will
increase with increasing net electricity generation. However, the application, design, and maintenance of
equipment all affect equipment banks and emission rates. These factors may change over time, which may alter
the trends observed to date. For example, switchgear dimensions have changed since the 1970s, resulting in a
reduction in the amount of SF6 required in switchgear (Ecofys, 2010).
5.2.5.2 Mitigation Options Considered for Electric Power Systems
This analysis considers six abatement options for this sector: SF6 recycling, leak detection and repair (LDAR)
using a thermal imaging camera, LDAR using a handheld gas detector, equipment replacement, the use of SF6 free
gas insulated equipment, and improved SF6 handling. Refurbishing existing equipment has been considered
previously in this analysis; however, industry experts have stated that this is no longer a common practice in the
field.
For the purposes of this analysis, four distinct emission streams were analyzed for the sector—improper
handling of SFs; venting during equipment maintenance and disposal; periodic leakage from equipment; and
chronic leakage from equipment. Each abatement option can only target one emission stream. Two emission
48 The bottom-up studies cited above indicate that emissions from this sector declined between 1995 and 2000, and
atmospheric studies confirm that emissions declined globally (Maiss and Brenninkmeijer, 2000). Other atmospheric studies
indicate that emissions increased after 2000 (Peters et al., 2005). However, the post-2000 increase may be from other sectors
(e.g., magnesium or electronics).
49 S. Reiman and M. Vollmer of EMPA have performed a preliminary analysis of this relationship, comparing the SF6 emissions
reported through national inventories with the net electricity consumption reported by EIA. They find that the ratios between
these two values vary by more than a factor of 10.
5-70
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
streams—chronic leakage from equipment and periodic leakage from equipment can be mitigated by two
abatement options. Specifically, chronic leakage from equipment can be mitigated by either equipment
replacement or the use of SF6 free gas insulated equipment and periodic leakage from equipment can be mitigated
by either of the two leak detection and leak repair options (thermal imaging camera and hand-held gas detectors).
SF6 Recycling
This option involves transferring SFs from electrical equipment into storage containers during equipment
servicing or decommissioning so that the SFs can be reused. Recycling is conducted using an SFs reclamation cart
(commonly referred to as a gas cart). The gas cart recovers the SFs from the equipment and purifies it for future
use; the recovered and purified SFs gas can then be stored within the cart, in a separate storage container, or
transferred back to the equipment for reuse. Proper recycling techniques are documented in the technical
literature (International Council on Large Electric Systems, 2005; International Electrotechnical Commission, 2008;
Institute of Electrical and Electronics Engineers, 2012). The alternative to using a gas cart is venting the used SFs
into the atmosphere and then replacing it with fresh SFs. Venting is typically performed in areas where
environmental consideration is low because the cost of purchasing new gas is often cheaper than purchasing gas
carts and paying technicians to reclaim gas from equipment.
• Capital Cost: The average total capital costs associated with purchasing gas carts are estimated to be
about $520,000 for the uncontrolled system and $78,000 for the partially controlled system. The cost per
gas cart unit is the same for both systems at approximately $104,000. Gas carts can range in cost from as
low as $21,000 to as high as $187,000 depending on their size (Rothlisberger, 2011a), and a mid-range gas
cart size was assumed for both system types in this analysis. The average capital costs for the partially
controlled system are less than the cost per unit, because U.S. systems have implemented SF6 recycling to
a greater extent and fewer gas carts are needed across U.S. systems: less than one per system.
• Annual O&M Costs: O&M costs are estimated to be $8,000 for the uncontrolled system and $22,000 for
the partially controlled system. The lower O&M costs for the uncontrolled system are driven by the
significantly lower labor cost in developing countries relative to the United States.
• Annual Benefits: Annual revenue, which was estimated based on the reduction of SF6 emissions
multiplied by the cost per pound of SF6 gas, is close to $102,000 for the uncontrolled system and $2,000
for the partially controlled system. Annual revenues are significantly higher for the uncontrolled system
because the uncontrolled system has not implemented the option at all, while the partially controlled
system has implemented the option to 91% of its potential; therefore, the potential for reductions is
greater. The cost of SF6 per pound varies regionally and is relatively low in the United States
(Rothlisberger, 2018). In addition, partially controlled systems have a significantly lower emission rate,
which also results in a comparatively lower annual revenue.
LDAR
LDAR is a two-step process. First, a leak detection technique is used to identify gas leaks from SF6-insulated
equipment. Leak detection methods vary and can involve simple techniques such as using soap and water
solutions or more sophisticated techniques such as those requiring detection devices, such as cameras to visualize
the source of the SF6 leaks by exploiting the strong infrared adsorption of SF6 for detection. Identified leaks are
typically repaired by applying a sealing material to the component that is leaking, although in some cases the
component needs to be replaced completely. Once the leak is repaired, the equipment tends to last months to
years without another major leak. SF6 emissions from periodic equipment leakage account for 10% of emissions
from both uncontrolled systems (in developing countries) and partially controlled systems (in the United States)
(Rothlisberger, 2011a; 2011b) for this abatement option. Two LDAR options were analyzed, as described below.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-71
-------
METHODOLOGY DOCUMENTATION
Thermal Imaging Camera
This first LDAR abatement option analyzes the use of a thermal imaging camera. Considered the best method
for finding leaks on energized equipment, the camera is able to detect even very minor leaks (as small as 0.5
pounds annually). However, some drawbacks of this method are that it requires light from the sun to properly view
the leak and an experienced technician (Wolf, 2017).
• Capital Cost: Costs associated with purchasing thermal imaging cameras are estimated to be $21,000 for
an uncontrolled abatement system and $50,000 for a partially controlled system. The cost for a single
thermal imaging camera is approximately $107,000 (Czerepuszko, 2011a).
• Annual O&M Costs: O&M costs are estimated to be $150 for the uncontrolled system and $3,500 for the
partially controlled system. The lower O&M costs for the uncontrolled system are driven by the
significantly lower labor cost in developing countries relative to the United States.
• Annual Revenue: Annual revenue is estimated based on the potential reduction of SF6 emissions
multiplied by the cost per pound of SF6 gas and is $2,000 for the uncontrolled system and $250 for the
partially controlled system. Annual revenues are significantly higher for the uncontrolled system primarily
because it was assumed that the uncontrolled system has implemented the option to a lesser extent than
the partially controlled system; therefore, the potential for reductions is greater. In addition, because the
cost of SF6 per pound varies regionally and costs significantly less in the United States (Rothlisberger,
2018), the cost of SF6 per pound is significantly less for the partially controlled system relative to systems
in other regions, so less money is saved through reduced emissions.
Handheld Leak Detectors
A handheld detector is a simple, low-cost device that may be used as an alternative to the thermal imaging
camera that can be cost prohibitive in developing countries. The detector is used by tracing the entire substation in
a continuous path. If a leak is identified, it is verified by blowing shop air into the area of suspected leak and
repeating the check of the area using the detector (lonScience, n.d.). This must be done while equipment is in
service and cannot be used for decommissioned parts (Wolf, 2017).
• Capital Cost: Handheld detectors are estimated to cost $1,400 in uncontrolled and $700 in the partially
controlled systems, with costs varying based on the assumed potential application of additional detectors
that could be utilized. (Ladzinski, 2018).
• Annual O&M Costs: For uncontrolled systems, the annual O&M costs are expected to be $1,600 while in
partially controlled systems, they are expected to be $7,500. This is due to the lower cost of labor in
developing countries.
• Annual Revenue: The annual revenue is calculated the same way as the LDAR thermal imaging camera
option. However, because the handheld detector is less cost prohibitive than the thermal imaging camera,
it is assumed to have a higher market penetration, and more than double the percent technical
effectiveness than the camera option for the uncontrolled system, resulting in annual revenues of $7,600.
For the partially controlled, the annual revenue is estimated at $250, the same as the LDAR thermal
imaging camera method.
Equipment Replacement
Unlike Equipment replacement is an option by which chronic leaking equipment is identified and replaced with
new and less leak prone equipment. Engineering design changes have reduced the amount of SF6 necessary for the
operation of switchgear and increased the tightness of equipment, resulting in smaller leakage amounts and less
frequent leakage over time. Replacing gas insulated equipment is a costly and timely option but eliminates the
worst performing breakers and helps to increase system reliability.
5-72
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Capital Cost: Assuming the replacement of a total of eight 145 kV breakers annually, costs associated with
equipment replacement are estimated to be $2,9 million for both an uncontrolled system and partially
controlled system. The estimated cost to replace a single 145 kV circuit breaker is estimated to be
$350,000 (based on a range of $200,000 for a 120 KV breaker up to $500,000 for a 345-kV breaker, as
provided by McNulty and Jasinski [2012]).
• Annual O&M Costs: No incremental O&M costs are assumed to be associated with this option.
• Annual Revenue: Annual revenue, which was estimated based on the reduction of SF6 emissions
multiplied by the cost per pound of SF6 gas, is $18,000 for the uncontrolled system and $2,000 for the
partially controlled system. Annual revenues are significantly higher for the uncontrolled system primarily
because it was assumed that the uncontrolled system has a higher SF6 rate compared to developing
countries, and therefore more emissions to abate relative to the partially controlled systems; therefore,
the potential for reductions is greater. In addition, the cost of SF6 per pound is significantly less for the
partially controlled system, so less money is saved through reduced emissions.
SFe-free Gas Insulated Equipment
New circuit breaker technology has been developed that uses alternative gases to SF6 that have lower or no
global warming potential. Traditional gases such as dry air, nitrogen, CO2 and related mixtures are low global
warming but are limited in their dielectric strength as compared to SF6 (Kieffel et al., 2015). A class of g3 mixtures
based on 3M™ Novec™ fluids have a higher GWP of less than 500 compared to the traditional gases (yet
significantly lower GWP as compared to SFs), their dielectric properties are stronger (Kieffel et al., 2015). While not
commercialized yet in the United States, early adopters are beginning to transition to g3 alternatives in Europe (GE,
2017). Vacuum is a technology that does not use any SF6 and is well established in the medium voltage class of
equipment; in 2015, the technology entered pilot application for high voltage equipment at 145 kV (Kieffel et al.,
2015). Vacuum technology for high voltage equipment is beginning to enter the market (Rak, 2017b; Glaubitz,
2017; Ecofys, 2018). This option is based on similar assumptions to the equipment replacement option. For the
purpose of this analysis, vacuum technology was assumed to be the replacement technology. Vacuum-based
technology is also considered an option in Europe and Japan for partially controlled systems and was included in
our analysis for this abatement option.
• Capital Costs: The newer SF6-free gas insulated equipment is currently more expensive than the
traditional breakers that contain SF6. Assuming the replacement of a total of eight 145 kV breakers
annually, the capital cost for this equipment in each of the three model facility systems is $4.3 million. The
estimated cost to replace a single 145 kV circuit breaker is estimated to be $550,000. It is assumed that
the technology is 50% more expensive than a traditional SF6 breaker, and the McNulty and Jasinski (2012)
capital cost estimate is used as the basis. It is assumed that a team of four field personnel require on week
(40 hours) of labor time to install one breaker.50 Information on costs is limited, and in lieu of a source on
high voltage replacement costs, medium voltage capital costs were referenced. Ecofys (2018) reports that
SFs free medium voltage (up to 36 kV) GIE can be 30 to 50% more expensive than regular SF6 breakers.
• Annual O&M Costs: Equipment maintenance is not necessary until at least 25 years after installation (Rak
2017b. Glaubitz, 2017). No incremental O&M costs are assumed to be associated with this option.
• Annual Revenue: Annual revenue for the uncontrolled systems will be $18,000, $2,000 for the partially
controlled-United States system, and $1,000 for the partially controlled-Europe/Japan system. These
50 Planning hours associated with the replacement, which may include time needed to receive necessary approvals for the
installation were not quantitatively considered. Any prior effort needed by a utility to accept this option, such as a smaller pilot-
scaled study of the technology, also were not quantified.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-73
-------
METHODOLOGY DOCUMENTATION
estimates are based on the technical applicability of this technology and the potential for it to reduce the
emissions assumed for each model facility.
Improved SF6 Handling
This option involves improving the procedures and techniques for handling SFs, especially when maintenance
is being performed on gas-insulated circuit breakers. Handling-related leaks can occur when (1) inappropriate
fittings are used to connect transfer hoses to cylinders or equipment, (2) SFs is not cleared from transfer hoses
before the hoses are disconnected from cylinders/equipment, (3) gas cylinders are not monitored/maintained
because they have been misplaced or stored improperly, and (4) a technician accidently vents SFs. Improving SFs
handling involves both training technicians to properly handle gas and purchasing adapter kits that ensure proper
fittings are available for connecting hoses to all gas-insulated equipment throughout the system. SFs emissions
from handling-related leaks account for 40% of emissions from both uncontrolled systems (in developing
countries) and partially controlled systems (in the United States) (Rothlisberger, 2011a; 2011b).
• Capital Cost: The capital cost associated with improved SFs handling is estimated to be $15,00 for both
the uncontrolled system and the partially controlled system. This capital cost consists entirely of
purchasing adapter kits, which are estimated to cost $1,500 each (middle of cost range provided by
Rothlisberger [2011a]).
• Annual O&M Costs: O&M costs are estimated to be $400 for the uncontrolled system and $3,00 for the
partially controlled system. The lower O&M cost for the uncontrolled system is driven by the significantly
lower labor cost in developing countries relative to the United States.
• Annual Revenue: Annual revenue, which was estimated based on the reduction of SFs emissions
multiplied by the cost per pound of SFs gas, is $136,000 for the uncontrolled system and $7,000 for the
partially controlled system.
5.2.5.3 Model Facilities
The analysis considers several possibilities for the maintenance and SF6 handling procedures used at the
typical electric transmission and distribution system, reflecting different levels of emissions. For the purpose of this
analysis, the three types of systems include the following:
• Partially controlled system—United States: Abatement options have been partially to fully applied in the
United States. The partially controlled system represents an EPS containing SF6-insulated equipment
located in the United States.
• Partially controlled system—Europe/Japan: In Europe and Japan, abatement options are close to fully
implemented, except for the use of SF6 free gas insulated equipment, which only recently began to gain
market momentum, and represents an opportunity for further abatement for these systems.
• Uncontrolled system: In contrast, abatement options have only been minimally applied or not applied at
all in most developing countries (Czerepuszko, 2011a; North China Grid Company [NCGC], 2010).
Therefore, the uncontrolled containing SF6-insulated equipment located in a developing country, which
for this analysis would mean any country outside of Europe, Japan, and the United States. EPS containing
SF6-insulated equipment located in a developing country, which for this analysis means any country
outside of Europe, Japan, and the United States.
5.2.5.4 Technical and Economic Characteristics Summary
Table 5-24 shows the technical applicability, market penetration, and reduction efficiency assumptions used to
develop the abatement measures' technical effectiveness for each model facility type. The technical effectiveness
parameter represents the percentage reductions achievable by each technology/facility type combination.
5-74
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-24: Technical Effectiveness Summary—Electric Power Systems
Abatement
Technical
Market
Reduction
Technical
Option
Model Facility Type
Applicability
Penetration
Efficiency
Effectiveness
SFs recycling
Uncontrolled
30%
100%
90%
27%
Partially controlled (US)
10%
100%
90%
9%
LDAR (Thermal
Uncontrolled
5%
20%
50%
1%
Imaging Camera)
Partially controlled (US)
5%
50%
50%
1%
LDAR (Handheld
Uncontrolled
5%
80%
50%
2%
Leak Detector)
Partially controlled (US)
5%
50%
50%
1%
Equipment
Uncontrolled
10%
50%
95%
5%
Replacement
Partially controlled (US)
20%
50%
95%
10%
SF6-free Gas
Insulated
Uncontrolled
10%
50%
95%
5%
Partially controlled (US)
20%
50%
95%
10%
Equipment
Partially controlled
(Europe/Japan)
100%
25%
95%
24%
Improved SF6
Uncontrolled
40%
100%
90%
36%
Handling
Partially controlled (US)
40%
100%
90%
100%
Table 5-25 summarizes the economic characteristics of each option. More details on the data sources and
assumptions used to develop these costs are described below.
Table 5-25: Engineering Cost Data on a Facility Basis
Annual
Annual O&M
Project Lifetime
Capital Cost
Revenue(2010
Costs
Abatement
Abatement Option
(years)
(2010 USD)
USD)
(2010 USD)
Amount (tC02e)
SFs recycling
1 Q
$520,156
$101,919
$7,592
44,678
ID
$78,023
$1,716
$20,097
2,048
LDAR (Thermal
c
$20,700
$1,887
$151
827
Imaging Camera)
Z>
$49,615
$238
$3,468
284
LDAR (Handheld Leak
c
$1,387
$7,550
$1,557
3,309
Detector)
Z>
$693
$238
$7,458
284
Equipment
/in
$2,886,599
$17,930
-
7,860
Replacement
4U
$2,844,205
$1,811
-
2,161
SF6-free Gas
$4,328,671
$17,930
-
7,860
Insulated Equipment
jU
$4,373,246
$1,811
-
2,161
Improved SF6
1
$4,369,910
$1,029
-
1,228
handling
1
$14,671
$135,893
$358
59,570
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-75
-------
METHODOLOGY DOCUMENTATION
The cross-cutting engineering cost inputs used for the assessment of all abatement options are supplied in
Table 5-26 (all costs in 2010 USD):
Table 5-26: Engineering Cost Inputs
Option
Uncontrolled
System
(developing
country)
Partially
Controlled
System (United
States)
Partially
Controlled System
(Europe and Japan)
Source
Size of system (SF6
nameplate capacity)
100,000 pounds
100,000 pounds
100,000 pounds
N/A
Emission rate
16%
2.2%
0.5%
Expert judgment
(uncontrolled
systems); 2016
average rate from U.S.
inventory (partially
controlled system-US);
Conservative estimate
assumed perGlaubitz
(2017) (partially
controlled system-
Europe/Japan)
Cost of bulk SFs (per
pound)
$23.59
$8.67
$8.67
Rothlisberger (2018a)a
Labor cost of
technician (per hour)
$1.95
$37.29
$34.64
BLS, 2012
SFs Recycling
Capital cost per gas
cart
$104,000
$104,000
N/A
Expert judgment
(middle of range
provided by
Rothlisberger [2011a])
Number of gas carts
that could be used at
100,000-pound
system
5
5
N/A
Expert judgment
(middle of range
provided by
Rothlisberger [2011b])
Existing penetration
of gas carts at facility
0%
85%
N/A
NCGC (2010) and
National Electric
Power Authority
(NEPA) (2005)
(uncontrolled
systems); expert
judgment (partially
controlled systems)
Annual O&M labor
per gas cart (hours)
780
780
N/A
Expert judgment
(middle of range
provided by
Rothlisberger [2011b])
(continued)
5-76
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-26: Engineering Cost Inputs (continued)
Option
Uncontrolled
System
(developing
country)
Partially
Controlled
System (United
States)
Partially
Controlled System
(Europe and Japan)
Source
Capital cost per gas
cart
$104,000
$104,000
N/A
Expert judgment
(middle of range
provided by
Rothlisberger [2011a])
LDAR (Thermal Imaging Camera)
Capital cost per unit
$107,000
$107,000
N/A
Czerepuszko (2011a)
Number of cameras
that could be used at
100,000-pound
system
1
1
N/A
Czerepuszko (2011a)
Annual O&M labor
per camera and
associated repairs
(hours)
400
200
N/A
Expert judgment
Existing penetration
in region
3%
7%
N/A
Czerepuszko (2011a)
SFs reduced through
application (pounds)
827
284
N/A
N/A
LDAR (Handheld Leak Detector)
Capital cost per unit
$700
$700
N/A
Ladzinski (2018)
Number of handheld
detectors that could
be used at 100,000-
pound system
2
2
N/A
Expert judgment
Annual O&M labor
per handheld
detectors and
associated repairs
(hours)
400
200
N/A
Expert judgment
Existing penetration
in region
50%
50%
N/A
Expert judgment
Technical
applicability to
baseline emissions
5%
5%
N/A
Rothlisberger (2011b)
(uncontrolled
systems);
Rothlisberger (2011a)
(partially controlled
systems)
SFs reduced through
application (pounds)
3,309
284
N/A
N/A
(continued)
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-77
-------
METHODOLOGY DOCUMENTATION
Table 5-26: Engineering Cost Inputs (continued)
Option
Uncontrolled
System
(developing
country)
Partially
Controlled
System (United
States)
Partially
Controlled System
(Europe and Japan)
Source
Equipment Replacement
Capital cost per
breaker
$366,000
$366,000
N/A
McNulty and Jasinski
(2012)
Number of 145kV
breakers (that would
be subject to
replacement) in a
given year
8
8
N/A
Expert judgment
(Using PG&E as an
example facility and
230kV breakers as a
proxy [Rak 2017b])
Percentage of leak-
prone equipment
already replaced
0%
10%
N/A
Expert judgment
Capital labor hours
per replaced breaker
160
160
N/A
ICF Expert Judgement
SFs reduced through
application (pounds)
7,860
2,161
N/A
N/A
SF6-Free Gas Insulated Equipment
Capital cost per
breaker
$549,000
$549,000
$549,000
McNulty and Jasinski
(2012) show the cost
of breaker
replacement ranges
from $200,000 for a
120kV breaker up to
from $200,000 for a
120kV breaker up to
$500,000 for a 345kV
breaker. Ecofys (2018)
reports that SF6 free
MV (up to 36kv) GIE
can be 30 to 50% more
expensive than regular
SFs breakers (see Table
24 of Ecofys 2018)
Number of 145kV
breakers (that would
be subject to
replacement) in a
given year
8
8
8
Expert judgment
(Using PG&E as an
example facility and
230kV breakers as a
proxy [Rak 2017b])
Percentage of leak-
prone equipment
already replaced
0%
0%
0%
Expert judgment
(continued)
5-78
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-26: Engineering Cost Inputs (continued)
Option
Uncontrolled
System
(developing
country)
Partially
Controlled
System (United
States)
Partially
Controlled System
(Europe and Japan)
Source
Annual O&M labor
hours per
replacement
160
160
160
ICF Expert Judgement
SFs reduced through
application (pounds)
7,860
2,161
1,228
N/A
Improved SF6 Handling
Cost per adapter kit
$1,500
$1,500
N/A
Expert judgment
(middle of range
provided by
Rothlisberger [2011a])
Number of adapter
kits that could be
used at 100,000-
pound system
20
20
N/A
Expert judgment
Percentage of kits
already purchased
50%
50%
N/A
Rothlisberger (2011a)
Number of
technicians per
system
23
23
N/A
Expert judgment
(middle of range
provided by
Rothlisberger [2011b])
Number of annual
training per
technician (hours)
16
16
N/A
Rothlisberger (2011a)
Percentage of
technicians already
trained
50%
80%
N/A
Rothlisberger (2011a)
SFs reduced through
application (pounds)
59,570
8,191
N/A
N/A
a Rothlisberger (2011a) provided a range of $12 to $20 USD for the estimated cost of bulk SFs in developing countries.
5.2.5.5 Sector-Level Trends and Considerations
Research into a gas or gas mixture to replace SFsfor use in medium- and high-voltage equipment has been
underway through pilot applications, primarily in Europe. SF6-free alternative technology availability depends on
the application in question. In medium-voltage equipment, established alternatives, such as vacuum technology,
are commercially available; however, further research and development is needed to establish commercially
viable, widespread adoption of SF6-free alternatives for high-voltage equipment. That said, promising
developments are underway. For example, National Grid, a utility in the United Kingdom, became the first utility to
energize a SF6-free 420 kV gas-insulated line in their Sellindge substation in April 2017. Additionally, in late 2017,
the Etzel substation in Switzerland successfully tested the alternative Green Gas for Grid or g3, a fluoroketone-
based mixture, on a 14- kV gas-insulated substation (T&D World, 2018). In the United States, utilities are beginning
to take similar action to learn and plan for adoption of SF6-free technology. In particular, in California, a state that
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-79
-------
METHODOLOGY DOCUMENTATION
is currently proposing to phase out SF6 gas, an insulating gas in transmission and distribution electrical equipment,
several utilities are working together toward viable alternatives and a transition plan that can reduce costs,
standardize replacement gases, and avoid duplication in their research and development (Rak, 2017a).
While additional efforts are necessary for a widespread transition to SF6 alternatives in high-voltage
switchgear, technology enhancements have optimized SF6-insulated equipment over the past 30 to 40 years. As
such, the amount of SF6 necessary for the operation of switchgear has declined significantly, and the tightness of
equipment has improved. Such engineering design changes have resulted in smaller leakage amounts and less
frequent leakage over time (Rhiemeier et al., 2010; Ecofys, 2018). The average age of SF6-insulated equipment in
developed countries (including Europe, Japan, and the United States) is considerably older than in developing
countries. Until recently, developing countries were slow to adopt SF6-insulated equipment because of its
relatively high cost compared with other traditional technologies, such as oil-insulated circuit breakers
(Rothlisberger, 2011b). Also, the electrical grid in developing countries has grown rapidly over the last decade with
economic growth, so the average age of all types of electrical equipment tends to be newer in developing
countries. The average SF6-insulated circuit breaker in developing countries, therefore, was assumed to leak less
than the average SF6-insulated circuit breaker in developed countries.
Employee training and investments in SF6 handling technologies (such as SF6 recovery carts) vary widely
among countries and regions. The use of equipment and accessories to properly handle SF6 is high in developed
countries (Rothlisberger, 2011a) yet low to nonexistent in at least some developing countries (NCGC, 2010; NEPA,
2005). Employee training is perhaps strongest in Europe, where the European Commission requires that personnel
who handle SF6 receive formal training and certification (EC Reg. No. 842/2006). In the United States, employee
training has improved significantly since 1999 by companies participating in the EPA's voluntary SF6 Emission
Reduction Partnership for Electric Power Systems. Employee training is low to nonexistent in at least some
developing countries (NEPA, 2005).
5.2.5.6 References
Czerepuszko, P. May 13, 2011a. Documentation of meeting between Paul Czerepuszko of FLIR Systems and Paul
Stewart of ICF International.
Czerepuszko, P. May 23, 2011b. Documentation of meeting between Paul Czerepuszko of FLIR Systems and Paul
Stewart of ICF International.
Ecofys. 2005. Reductions ofSFe Emissions from High and Medium Voltage Electrical Equipment in Europe, Final
Report to Capiel.
Ecofys. 2010. Update on Global SF6 Emissions Trends from Electrical Equipment—Edition 1.1. Ecofys Emission
Scenario Initiative on Sulphur Hexafluoride for Electrical Industry (ESI-SFs).
Ecofys. 2018. Concept for SF6-Free Transmission and Distribution of Electrical Energy. Available online at
https://www.umweltbundesamt.de/sites/default/files/medien/2503/dokumente/final-report-sf6 en.pdf ef
GE. 2017. National Grid Begins Journey to SF6-Free HV Substations. Case Study. Available online at
GEGridSolutions.com J
Glaubitz. January 2017. Sustainable Reduction of SF6 Emission—OEMs, Users, New EU-F-gas Regulation. Presented
at EPA's Workshop on SF6 Emission Reduction Strategies. Available online at
https://www.epa.gov/sites/production/files/2017-02/documents/glaubitz presentation 2017 workshop.pdf
Institute of Electrical and Electronics Engineers. 2012. IEEE Guide for Sulphur Hexafluoride (SFe) Gas Handling for
High-Voltage (over 1000 Vac) Equipment. IEEE Std C37.122.3.
International Council on Large Electric Systems. 2005. Guide for the Preparation of Customised "Practical SF6
Handling Instructions." Brochure No. 276.
5-80
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
International Electrotechnical Commission. 2008. High-voltage Switchgear and Controlgear— Part 303: Use and
Handling of Sulphur Hexafluoride (SFe). IEC Technical Report 62271-303. Geneva, Switzerland: International
Electrotechnical Commission.
lonScience. No date. Owners' Manual. SF6 GasCheck 6000 handheld leak detector UK V.1.1. Available online at
https://www.ionscience.eom/products/sf6-gascheck-6000-leak-detector/#downloads ef
Intergovernmental Panel on Climate Change. May 2000. Good Practice Guidance and Uncertainty Management in
National Greenhouse Gas Inventories. IPCC-XVI/Doc.10 (1.IV.2000). Montreal: Intergovernmental Panel on
Climate Change, National Greenhouse Gas Inventories Programme.
Intergovernmental Panel on Climate Change. 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.
Kieffel, Y., F. Biquez, and P. Ponchon. 2015. Alternative gases to SF6 for use in high-voltage switchgear: g3. Paper
0230. 23rd International Conference on Electricity Distribution, Lyon, France.
Knopman, D. and K. Smythe. 2007. 2004-2006 SF6 data summary. Project memorandum prepared by the RAND
Corporation for the National Electrical Manufacturers Association.
Ladzinski, R. June 6, 2018. Email correspondence between Robert Ladzinski of lonScience and Neha Vaingankar of
ICF.
Maiss, M. and C. Brenninkmeijer. 2000. A reversed trend in emissions of SF6 into the atmosphere? Proceedings of
the 2nd International Symposium Non-CCh Greenhouse Gases: Scientific Understanding, Control and
Implementation, Noorwijkerhout, The Netherland 1999,199-204. doi:10.1007/978-94-015-9343-4_30
McNulty, M. and J. Jasinski. April 2012. SF6 Equipment maintenance, repair, and replacement and emissions
program. Presented at EPA's 2012 Workshop on SF6 Emission Reduction Strategies. Available online at
https://www.epa.gov/sites/production/files/2016-02/documents/confl2 mcnultv.pdf
National Electric Power Authority. 2005. Reducing SFs emissions in high-voltage transmission/distribution systems
in Nigeria—Project design document. NM0135. Available online at
http://cdm.unfccc.int/methodologies/PAmethodologies/pnm/bvref/NM0135 ef
North China Grid Company. 2010. SFs recycling project of North China Grid—Project design document. CDM
Project 3707. Available online at http://cdm.unfccc.int/Proiects/Validation/DB/
MPHV4MMDLVU2i0E4ZP69TWKSRAT3HB/view.htmltf
Peters, W., E. Dlugokencky, J. Olivier, G. Dutton, and K. Smythe. 2005. Surface measurements show a 17 percent
increase in the release of sulfur-hexafluoride (SFs) to the atmosphere in 2003. Proceedings of the Fourth
International Symposium NCGG-4. Milpress, Rotterdam, 2005.
Rak, T. January 2017a. SF6 Free HV GIS and breakers. Presented at EPA's 2017 Workshop on SF6 Emission Reduction
Strategies Available online at https://www.epa.gov/sites/production/files/2017-
02/documents/rak presentation 2017 workshop.pdf
Rak, T. 2017b. PG&E's initiative to use SF6 gas alternatives in gas insulated equipment. Technical paper
presentation, COMET 2017. Available online at
https://utwlQ356.utweb.utexas.edu/sites/default/files/COMET%20NON-
SF6%20GIS%20and%20Breakers 2018-01-22%20%28Tom%20Rak%29.pdf ef
Rhiemeier, J., S. Wartmann, M. Pagnotta, N. Makowska, and X. Li. 2010. Update on Global SF6 Emissions Trends
from Electrical Equipment—Edition 1.1. Ecofys Emission Scenario Initiative on Sulphur Hexafluoride for Electric
Industry. Available online at http://www.ecofvs.com/fiIes/fiIes/esi-
sf6 finalreport editionll 100701 vOl.pdfcf
Rothlisberger, L. May 5, 2011a. Documentation of meeting between Lukas Rothlisberger of DILO and Paul Stewart
of ICF International.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-81
-------
METHODOLOGY DOCUMENTATION
Rothlisberger, L. June 2, 2011b. Documentation of meeting between Lukas Rothlisberger of DILO and Paul Stewart
of ICF International.
Rothlisberger, L. May 23, 2018. Documentation of meeting between Lukas Rothlisberger of DILO and Mollie Averyt
of ICF.
Smythe, K. December 1-3, 2004. Trends in SFs sales and end-use applications: 1961-2003. Presented at the
International Conference on SFs and the Environment: Emission Reduction Technologies, Scottsdale, AZ.
T&D World. 2018. 'Transforming the Transmission Industry". April 27,2018. Available online at:
https://www.tdworld.com/substations/transforming-transmission-industrvc?
UNFCCC. 2016. United Nations Framework Convention on Climate Change Flexible GHG Data Queries. Online
Database Accessed: Fall 2016. Available online at: http://unfccc.int/di/FlexibleQueries/Setup.do ef
U.S. Energy Information Administration. 2017. International Energy Outlook 2017. Report# DOE/EIA-0484(2017).
Washington, DC: Energy Information Administration, U.S. Department of Energy. Available online at
https://www. eia.gov/outlooks/ieo/pdf/0484f 2017). pdf
U.S. Environmental Protection Agency. 2011a. SFs Emission Reduction Partnership for Electric Power Systems—
2010 Annual Report. Washington, DC: EPA. Available online at http://www.epa.gov/electricpower-
sf6/documents/sf6 2010 ann report.pdf
U.S. Environmental Protection Agency. 2011b. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2009.
EPA #430-R-11-005. Washington, DC: EPA. Available online at
http://epa.gov/climatechange/emissions/usinventorvreport.html
U.S. Environmental Protection Agency. 2012. Global Anthropogenic Non-CC>2 Greenhouse Gas Emissions: 1990-
2030. EPA #430-R-12-006. Washington, DC: EPA. Available online at
http://www.epa.gov/climatechange/economics/intemational.html
U.S. Environmental Protection Agency. 2016. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2015.
USEPA #430-P-17-001. Washington, DC: USEPA.
Wolf, M. November 2017. SF6 leak management, repairs, and considerations. Presented at DILO 2nd Annual SF6
Gas Management Seminar. Tampa, Florida. Available online at http://www.dilo.com/wp-
content/uploads/2Q17/ll/DILO-Version-SF6-Leak-Management-and-Repair-Technioues-Wolf.pdf cf
Yokota, T., K. Yokotsu, K., Kawakita, H. Yonezawa, T. Sakai, and T. Yamagiwa. 2005. Recent practice for huge
reduction of SFs gas emissions from GIS&GCB in Japan. Presented at the CIGRE SC A3 & B3 Joint Colloquium,
Tokyo.
5-82
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.2.6 Primary A lumin um Production
Emissions of the PFCs CF4 and C2F6 are generated during brief process upset conditions in the aluminum smelting
process. During the aluminum smelting process, when the alumina (AI2O3) in the electrolytic bath falls below
critical levels required for electrolysis, rapid voltage increases occur. These voltage excursions are termed "anode
effects" (AEs). AEs produce CF4 and C2F6 emissions when carbon from the anode, instead of reacting with alumina
as it does during normal operating conditions, combines with fluorine (F2) from the dissociated molten cryolite
bath. In general, the magnitude of emissions for a given level of production depends on the frequency and
duration of these AEs; the more frequent and long-lasting the AEs, the greater the emissions.51
The most significant process emissions are:
• CO2 emissions from the consumption of carbon anodes in the reaction to convert aluminum oxide to
aluminum metal and
• PFC emissions of CF4 and C2F6 during AEs (IPCC, 2006).
5.2.6.1 Primary Aluminum Production Emission Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full-time series from 1990 through 2050 (see Section 3.3, Generating the Composite
Emission Projections, for additional information). Activity data for primary aluminum production included primary
aluminum production data from United States Geological Survey (1995 through 2016) and historical global
percentages of cell types derived from reports from the International Aluminium Institute (2016). Emission factors
were obtained from IAI (2011), IPCC (2006), and Marks and Nunez (2018).
The Tier 1 basic equation (IPCC, 2006) used in this analysis to estimate PFC emissions from primary aluminum
production is as follows:
EcF4 = EFcF4,i • MP
and
EC2F6= EFc2F6,i • MP (5.14)
where:
EcF4 =
Ec2F6 =
EFcF4 =
E Fc2F6 =
MP
Activity Data
Historical
The EPA estimated primary aluminum production for all aluminum-producing countries based on data from
U.S. Geological Survey (USGS) Mineral Yearbooks for Aluminum (USGS, 1995 through 2016a; USGS, 2016b).
51 It should be noted that over the last several years there has been the discovery and documentation of nonanode effect
(NAE)-related emissions. EPA has supported some of the most significant work on NAE emissions. These emissions can be a
significant, perhaps the major, source of PFC emissions in some smelter cells. It should also be noted that NAE emissions and
NAE abatement measures are not addressed in this report.
Emissions of CF4 from aluminum production, kg CF4
Emissions of C2F6 from aluminum production, kg C2F6
Default emission factor by cell technology type i for kg CF4/tonne aluminum
Default emission factor by cell technology type i for kg C2Fs/tonne aluminum
Metal production, tonnes aluminum
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-83
-------
METHODOLOGY DOCUMENTATION
Country-specific aluminum production was disaggregated to cell type using historical global percentages for 1990
through 2015 derived from the International Aluminium Institute's (lAI's) Results of the 2010 Anode Effects Survey
(IAI, 2011) and Results of the 2015 Anode Effects Survey (IAI, 2016) reports.
Projected
Country-specific production projections from 2020 to 2030 were estimated based on aluminum production
compounded annual growth rates (CAGRs) estimated from historical production data from 2005 through 2015
global production estimates (USGS, 2016a). To estimate production from 2030 to 2050, the CAGR was reduced by
50% based on expert judgment to account for slowing growth in global demand and slowing global population
growth. Growth in aluminum production is dependent on fluctuations in the global economic situation (Dudin et
al., 2017). The CAGRs used in this analysis reflect the expectation that aluminum production will continue to grow
in the coming years as urbanization, industrialization, and economic development in BRIC and other emerging
countries increase global demand (IAI, 2013b).
Country-specific aluminum production for 1990 through 2020 was disaggregated to cell type using historical
global percentages (for 1990 through 2015) derived from lAI's Results of the 2010 Anode Effects Survey (IAI, 2011)
and Results of the 2015 Anode Effects Survey (IAI, 2016) reports. Country-specific production projections from
2025 through 2050 were first disaggregated into "existing" or "new-build" production by comparing a country's
production projection against that country's total production capacity in 2018 (Light Metal Age, 2018). Production
less than or equal to a country's capacity in 2018 was considered existing production, and production greater than
that amount was considered new-build production. Existing SWPB, HSS, and VSS facilities were assumed to be fully
retired from 2015 to 2050, except in Russia where VSS facilities are expected to remain online (Marks, 2018).
SWPB, HSS, and VSS production is linearly interpolated between 2015 and 2050. Existing production was
disaggregated to cell type based on IAI production data (IAI, 2011 through 2016), and new-build production was
assigned to the point-fed center work prebake (PFPB) (i.e., newer) cell type. Given the efficiency of this cell type, it
was assumed that aluminum manufacturers will opt for this cell type for all new-build production (Marks, 2017).
Emission Factors
Historical and Projected
The EPA estimated PFC emission factors using the IPCC Tier 1 methodology (i.e., default emission factors
multiplied by quantity of metal produced) (IPCC, 2006). Average cell type-specific emission factors for 1990
through 2010 are taken from lAI's annual Anode Effect Surveys (IAI, 2011). Emission factors for 2010 through 2015
were provided by Marks and Nunez (2018). Projected emission factors were assumed to be constant at 2015 levels
through 2050 (Table 5-27).
Emission Reductions in Baseline Scenario
The base case analysis is intended to model the hypothetical scenario in which no further action is taken by
the aluminum industry to reduce its emission rates below the 2015 levels. Although this scenario represents a
break from the historical trend, future action by the aluminum sector is not guaranteed, and the rate of decline in
emission intensities (metric ton CChe/metric ton aluminum) has decreased in recent years (i.e., since 2005).
Uncertainty
In developing these emission estimates, the EPA used multiple international datasets and the most recent
IPCC guidance on estimating emissions from this source. Nevertheless, uncertainties exist in both the activity data
and the emission rates used to generate these emission estimates.
5-84
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-27: Baseline Emission Factors by Year (MtC02e/Mt Al)
Technology Type
1990
2000
2010
2020
2030
2040
2050
China
PFPB
0.8
0.8
0.8
1.3
1.3
1.3
1.3
CWPB
2.9
1.8
0.6
1.3
1.3
1.3
1.3
SWPB
15.4
13.9
4.5
4.0
4.0
4.0
4.0
HSS
3.1
2.4
0.9
4.1
4.1
4.1
4.1
VSS
5.0
3.2
1.0
1.7
1.7
1.7
1.7
Rest of World
PFPB
1.7
0.6
0.3
0.2
0.2
0.2
0.2
CWPB
3.5
1.8
0.6
0.2
0.2
0.2
0.2
SWPB
15.4
13.9
4.5
4.0
4.0
4.0
4.0
HSS
3.1
2.4
0.9
4.1
4.1
4.1
4.1
VSS
5.7
3.2
1.0
1.7
1.7
1.7
1.7
First, while this study incorporated recent data on total aluminum production by country from USGS
Commodity Surveys and USGS Mineral Yearbooks, to disaggregate historical aluminum production by cell type, the
EPA used information derived from lAI's Results of the 2010 Anode Effects Survey (IAI, 2011) and Results of the
2015 Anode Effects Survey (IAI, 2016) reports. This information provided the percentage breakout of total global
production (which adjusts for nonreporters) by cell type for 1990 through 2015 used for the disaggregation.
Therefore, these data may not be representative of the percentage breakout by cell type (and hence the
emissions) for an individual country (or region). Cell type is important because emissions per ton of aluminum (i.e.,
effective emission factors) can vary by a factor of five or more across different cell types (IPCC, 2000). To
disaggregate current (i.e., through 2025) and projected (i.e., post-2025) aluminum production by cell type, the EPA
first disaggregated the data into existing or new-build production by comparing it with projected production in
2025; then for existing production, we adopted the percentage breakout of total global production estimated for
2025 for 2030 through 2050, with new-build production assigned as aluminum. Therefore, the resultant total
production projection percentage breakout may not truly represent the future breakouts that would be derived
from reported production data for the technology in place through 2050.
Second, the EPA used a single aluminum production compounded annual growth rate to project country-
specific production through 2030, and a single growth rate to project country-specific production from 2030 to
2050. Future production in individual countries is likely to follow actual trends not reflected by an annual growth
rate, and the value of an individual country's annual growth rate might be significantly different from that of the
rest-of-world rate.
In addition, as previously discussed, the EPA assumed the emission factors estimated for 2015 when
estimating emissions for 2020 through 2050. Therefore, these emission factors may not truly represent values that
would be derived from reported data for the technology in place through 2050. Further, the methodology does not
explicitly account for reductions from CDM projects, but the impacts of these projects are captured to some extent
in the emission factors and cell type allocations taken from IAI.
It was assumed in this analysis that the vast majority of PFC emissions are generated during high-voltage AE
events. However, recent research using more sensitive measurement equipment has shown that PFC emissions
also occur during low-voltage AE events (Marks and Nunez, 2018), which are not accounted for in this analysis.
Emission rates from low-voltage AEs are lower than rates from high-voltage AEs, but because emissions from low-
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-85
-------
METHODOLOGY DOCUMENTATION
voltage effects can occur almost continuously, the associated emissions can be significant (Marks and Nunez,
2018). This analysis also does not account for additional PFC emissions associated with starting operations at a
new facility (Marks, 2017).
The base case analysis assumes that no further action is taken by the aluminum industry to reduce their
emission rates below the 2015 levels. However, the IAI notes that there has been an 88% reduction in global PFC
emissions per metric ton since 1990 and a 35% reduction since 2006 (IAI, 2016). Although the rate of reduction in
emission intensities has slowed in recent years, it is unlikely that actual emissions will be as high as those
presented in the analysis.
5.2.6.2 Mitigation Options Considered for Primary Aluminum Production
Five different electrolytic cell types are used to produce aluminum: Vertical Stud Soderberg (VSS), Horizontal
Stud Soderberg (HSS), Side-Worked Prebake (SWPB), Center-Worked Prebake (CWPB), and PFPB.52 PFPB is
considered the most technologically advanced process to produce aluminum, and all new greenfield smelters built
in the world today use this technology. Existing, older, and higher PFC-emitting PFPB systems can further improve
their AE performance by implementing management and work practices, as well as improved control software.
Retrofits to CWPB cells were not analyzed in this analysis because they already operate at close to the same
efficiency as PFPB cells in terms of emissions. Retrofits to CWPB cells are unlikely to occur given the economics and
reductions would not be material (Marks 2018). Facilities using VSS, HSS, and SWPB cells can reduce emissions by
retrofitting smelters with emission-reducing technologies such as computer control systems and point-feeding
systems, by shifting production to PFPB technology, and by adopting management and work practices aimed at
reducing PFC emissions. However, in practice, the greatest potential for reduction in PFC emissions is through
addition of new greenfield PFPB capacity in concert with the shutdown of existing high-emitting facilities.53
PFC emission reductions can primarily be achieved by installing/upgrading process computer control systems54
and installing alumina point-feed systems.55 The two abatement options considered for this analysis are (1) a
minor retrofit involving the upgrade of process computer control systems only and (2) a major retrofit involving
both the installation/upgrade of process computer control systems and the installation of alumina point-feed
systems.56 The installation of alumina point-feed systems was not analyzed on its own because it would be very
unlikely that an aluminum production facility would install alumina point-feed systems without also installing or
upgrading process computer control systems.57
52 It should be noted that PFPB and CWPB are essentially the same cell design but with different alumina feed processes.
53 More information on how global primary aluminum production according to cell type changed from 1990-2012 is available in
"Figure 2: Primary aluminium smelting technology mix, 1990-2012" in Results of the 2012 Anode Effect Survey: Report on the
Aluminium Industry's Global Perfluorocarbon Gases Emissions Reduction Programme, International Aluminium Institute,
London, U.K. http://www.world-aluminium.org/media/filer public/2013/08/20/2012 anode effect survey report.pdf J.
54 Process computer control systems control the repositioning of carbon anodes as they are consumed and provide greater
control over raw material (alumina) feeding. All smelters operate with process control computers. The upgrade would involve
changes in the algorithms controlling feed and anode effect detection.
55 Point-feed systems allow more precise alumina feeding.
56 A major retrofit results in PFPB technology, which is the state-of-the-art technology in aluminum production. Conversion to
PFPB technology results in the most reliable increases in operational and production efficiency, although the capital outlay for
this option is significant. In addition, retrofit options are usually implemented after extensive computer modeling and large-
scale development work is conducted on test cells.
57 It should be noted that, as previously mentioned, existing, older, and higher PFC-emitting systems can further improve their
anode effect performance by implementing management and work practices, as well as improved control software.
5-86
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Minor Retrofit
A minor retrofit involves the installation/upgrade of process computer control systems. Minor retrofits can be
performed at any facility type other than the state-of-the-art PFPB facilities. For the cost analysis, a minor retrofit
has a lifetime of 10 years for VSS, HSS, and SWPB facility types and 30 years at the PFPB facilities, based on expert
judgment. The lifetime of the minor retrofit at older facilities is shorter because the estimated remaining lifetime
of the facilities themselves is shorter.
Capital costs represent the costs associated with purchasing and installing the process computer control
systems at the aluminum production facilities. The capital costs for minor retrofit, obtained from International
Energy Agency (IEA) (2000) and confirmed by Marks (2018), range from $6.5 million to $8.8 million (2016 USD),
depending on the facility type. The annual O&M costs associated with minor retrofits are strictly the additional
operating costs for the increased aluminum production. The additional operating costs were assumed to equal the
percentage increase in current efficiency multiplied by the capital costs of the retrofit, which is the method used to
estimate O&M costs by IEA (2000). For minor retrofits, these costs range from approximately $65,000 to $130,000
(2016 USD), depending on the facility type.
It was assumed that model facilities would use the increased current efficiency (aluminum production/unit of
electricity) resulting from the retrofits to produce more aluminum with the same amount of electricity
consumption as before (rather than producing the previous levels of aluminum production and realizing the
electricity savings).58 Increased current efficiencies for each facility and retrofit are available from IEA (2000). The
additional revenues associated with the minor retrofit option, depending on the facility type range between $0.5
million to $1 million (2016 USD).
Major Retrofit
A major retrofit involves both the installation/upgrade of process computer control systems and the
installation of alumina point-feed systems. Major retrofits result in AE performance approaching that of PFPB
technology, which is the state-of-the-art technology in aluminum production. A major retrofit also results in
increases in operational and production efficiency, although the capital outlay for this option is significant. In
addition, retrofit options are usually implemented after extensive computer modeling and large-scale
development work are conducted on test cells. Major retrofits can be performed for the older facility types (VSS,
HSS, and SWPB). According to Marks (2011b), there is no opportunity for conventional CWPBs to install point
feeders because they already have "bar break" feed systems, which have roughly the same AE performance as
point feeders. By definition, a PFPB model facility has point-feeding systems, so there is no opportunity for
additional application.
The capital costs for major retrofits represent the costs associated with purchasing and installing the process
computer control systems and alumina point-feeding technologies at the aluminum production facilities. The
capital costs for major retrofits, obtained from IEA (2000) and confirmed by Marks (2018), range from $13 million
to $98 million (2016 USD), depending on facility type. Additionally, annual O&M costs range from $390,000 to $3.7
million (2016 USD), depending on facility type. Annual revenues for major retrofits range from $1 million to $2
million (2016 USD), depending on facility type. The expected lifetimes for major retrofits was assumed to be 10
years.
5.2.6.3 Model Facilities
A facility's performance may be represented by the mean or median (depending on the size of the cohort of
facilities and the range of performance) PFC emission factor—PFC emissions per unit production (e.g., metric tons
58 Note that this is a simplifying assumption for the purpose of this analysis; any increase in production would be driven by
demand for aluminum, not specifically driven by a set level of electricity consumption.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-87
-------
METHODOLOGY DOCUMENTATION
C02e/metric ton aluminum)—for a particular cell technology type. However, in the case of PFPB technology, the
universe of facilities using this technology is further subdivided into state-of-the-art (i.e., newer) PFPB facilities for
which no abatement measures are applicable and other (i.e., older) PFPB facilities for which certain abatement
measures are an option. The performance for state-of-the-art (as opposed to other) PFPB technology is therefore
better represented by a PFC emission factor less than the average (i.e., the median). Table 5-28 presents a
description of the model facilities considered for this analysis.
Table 5-28: Description of Primary Aluminum Production Facilities
Facility Type
Description
VSS
This facility uses VSS technology with an average PFC emission factor of 1.21 metric tons
CC>2e/metric ton aluminum. The production capacity of the facility is 200,000 metric tons
per year.
HSS
This facility uses HSS technology with an average PFC emission factor of 1.39 metric tons
CC>2e/metric ton aluminum. The production capacity of the facility is 200,000 metric tons
per year.
SWPB
This facility uses SWPB technology with an average PFC emission factor of 3.49 metric tons
CC>2e/metric ton aluminum. The production capacity of the facility is 200,000 metric tons
per year.
PFPB (state of
the art)
This facility uses state-of-the-art PFPB technology with a median PFC emission factor of 0.18
metric tons CC>2e/metric ton aluminum. The production capacity of the facility is 200,000
metric tons per year.
PFPB (other)
This facility uses other PFPB technology, with an average PFC emission factor of 0.20 metric
tons C02e/metric ton aluminum. The production capacity of the facility is 200,000 metric
tons per year.3
a It should be noted that the "state of the art" has been improving rapidly with respect to AE performance, and the best PFPB
facilities (top 10%) are performing at better than 0.06 metric tons C02e/metric ton aluminum. Median performance for all IAI
non-Chinese producers is about 0.23 metric tons C02e/metric ton aluminum, while median Chinese PFPB performance is about
0.8 metric tons C02e/metric ton aluminum (IAI 2017).
5.2.6.4 Technical and Economic Characteristics Summary
Technical effectiveness is the parameter used to assess the abatement potential from each technology option.
The technical effectiveness parameter determines the share of emission reductions attributed to each abatement
measure. Similar to other industrial process sectors covered in this report, the technical effectiveness parameter is
defined as the percentage of emission reductions achievable by each technology/facility combination. Estimating
this parameter required assumptions regarding the distribution of emissions by manufacturing process (i.e., VSS,
HSS, SWPB, and PFPB) in addition to process-specific estimates of technical applicability and market penetration.
The technical applicability, market penetration, and reduction efficiency assumptions are held constant for all
model years. Table 5-29 presents the market penetration, technical applicability, and reduction efficiency
assumptions used to develop the abatement measures' technical effectiveness parameter. Technical effectiveness
is equal to the product of the technical applicability, market penetration, and reduction efficiency.
5-88
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-29: Technical Effectiveness Summary—Primary Aluminum Production
Abatement Option
Model Facility
Type
Technical
Applicability
Market
Penetration
Reduction
Efficiency
Technical
Effectiveness
Minor retrofit
VSS
100%
50%
39%
11%
HSS
100%
50%
39%
20%
SWPB
100%
50%
24%
12%
PFPB
100%
100%
55%
55%
Major retrofit
VSS
73%
50%
11%
56%
HSS
100%
50%
78%
39%
SWPB
100%
50%
96%
48%
The technical applicability factor for VSS assumes that roughly 27% of VSS capacity already has point feeding
(Marks, 2011). Table 5-30 documents the facility-level engineering cost data used in the MAC analysis for the
primary aluminum production sector.
Table 5-30: Engineering Cost Data on a Facility Basis—Primary Aluminum Production
Abatement
Option
Facility Type
Project
Lifetime
(years)
Capital Cost
(2016 USD)
Annual
Revenue
(2016 USD)
Annual O&M
Costs
(2016 USD)
Abatement
Amount
(tCOze)
Minor retrofit
VSS
10
$6,582,874
$1,052,475
$131,657
103,000
HSS
10
$6,582,874
$526,237
$65,829
121,000
SWPB
10
$6,866,347
$789,356
$102,995
165,500
PFPB
30
$8,834,909
$526,237
$88,349
4,000
Major retrofit
VSS
10
$93,057,897
$2,104,949
$3,722,316
206,000
HSS
10
$98,002,926
$1,052,475
$1,960,059
242,000
SWPB
10
$12,992,514
$1,578,712
$389,775
662,000
5.2.6.5 Sector-Level Trends/Considerations
The emission projections (i.e., baseline emissions) account for the historical reduction in the effective emission
factor (i.e., metric ton C02e/metric ton aluminum) realized by facilities but do not assume that aluminum
producers have conducted retrofits or will continue to introduce technologies and practices aimed at reducing PFC
emissions. That said, the global primary aluminum industry through the IAI has a voluntary PFC emission reduction
goal of reduce emissions of PFCs per metric ton of aluminum by at least 50% by 2020 as compared with 2006 (IAI,
2013a). In addition, commissioning of new, less emissive facilities to meet global demand will also have the result
of reducing the effective emission factor.
This analysis does not consider "breakthrough" technologies, which, if developed and widely implemented,
could significantly reduce GHG emissions from aluminum production. One potential breakthrough technology is
the use of inert anodes. An inert anode is chemically nonreactive and so does not react to the electrolysis process
and is not consumed during production. Due to the high corrosiveness of the cryolitic melts, it is difficult to
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-89
-------
METHODOLOGY DOCUMENTATION
develop an anode that is truly inert or even slowly consumable. Alcoa announced in 2000 that it was working with
inert anodes but had not proved the commercial feasibility (Kvande and Drabl0s 2014). In May of 2018, Apple
announced that it has partnered with Alcoa and Rio Tinto to commercialize the technology, and along with the
governments of Canada and Quebec, invest a combined $144 million in research and development. Sufficient
economic data was not available to include this technology as part of this analysis (Apple 2018).
5.2.6.6 References
Apple. 2018. Apple paves the way for breakthrough carbon-free aluminum smelting method. May 10, 2018.
Available online at https://www.apple.com/newsroom/2018/Q5/apple-paves-the-wav-for-breakthrough-
carbon-free-aluminum-smelting-method/eP
Dudin, M.N., N.A. Voykova, E.E. Frolova, J.A. Artemieva, E.P. Rusakova, and A.H. Abashidze. 2017. Modern Trends
and Challenges of Development of Global Aluminum Industry. Metalurgija, 56(1-2), 255-258. Available online
at https://hrcak.srce.hr/168955 E?
International Aluminium Institute. 2011. Results of the 2010 Anode Effects Survey: Report on the Aluminum
Industry's Global Perfluorocarbon Gases Emissions Reduction Program. London, UK: IAI.
International Aluminium Institute. 2013a. Results of the 2012 Anode Effects Survey: Report on the Aluminum
Industry's Global Perfluorocarbon Gases Emissions Reduction Program. London, UK: IAI.
International Aluminium Institute. 2013b. The Global Aluminium Industry: 40 years from 1972. London, UK: IAI.
International Aluminium Institute. 2014. Results of the 2013 Anode Effects Survey: Report on the Aluminum
Industry's Global Perfluorocarbon Gases Emissions Reduction Program. London, UK: IAI.
International Aluminium Institute. 2015. Results of the 2014 Anode Effects Survey: Report on the Aluminum
Industry's Global Perfluorocarbon Gases Emissions Reduction Program. London, UK: IAI.
International Aluminium Institute. 2016. Results of the 2015 Anode Effects Survey: Report on the Aluminum
Industry's Global Perfluorocarbon Gases Emissions Reduction Program. London, UK: IAI
International Aluminium Institute. 2017. Results of the 2016 Anode Effects Survey: Report on the Aluminum
Industry's Global Perfluorocarbon Gases Emissions Reduction Program. London, UK: IAI.
International Energy Agency. 2000. Greenhouse Gas Emissions from the Aluminium Industry. Cheltenham, UK: The
IEA Greenhouse Gas R&D Program.
Intergovernmental Panel on Climate Change. May 2000. Good Practice Guidance and Uncertainty Management in
National Greenhouse Gas Inventories. IPCC-XVI/Doc.10 (1.IV.2000). Montreal: Intergovernmental Panel on
Climate Change, National Greenhouse Gas Inventories Programme.
Intergovernmental Panel on Climate Change. 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.
Kvande, H. and P.A. Drabl0s. 2014. The aluminum smelting process and innovative alternative technologies. JOEM,
56(55), S23-S32. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4131935/pdf/ioem-56-s23.pdf
Light Metal Age. 2018. Primary Aluminum Producers, https://www.lightmetalage.com/resources-section/primarv-
producers/c?
London Metals Exchange. 2011. LME official prices for aluminum. Available online at
http://www.lme.com/aluminium.asp J
Marks, J., IAI. July 14, 2011. Personal communication with Jerry Marks, J Marks & Associates, LLC.
Marks, J. IAI. May 2017. Personal communication with Jerry Marks, J Marks & Associates, LLC. 2017.
Marks, J. April 27, 2018. Personal communication with Jerry Marks, J Marks & Associates, LLC.
5-90
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Marks, J. and P. Nunez. 2018. Updated Factors for Calculating PFC Emissions from Primary Aluminum Production.
The Minerals, Metals, and Materials Society. Light Metals.
U.S. Environmental Protection Agency. 2012. Global Anthropogenic Non-CCh Greenhouse Gas Emissions: 1990-
2030. EPA #430-R-12-006. Washington, DC: EPA. Available online at
http://www.epa.gov/climatechange/economics/international.html
U.S. Geological Survey. 1995 through 2016a. Mineral Yearbook: Aluminum. Reston, VA: U.S. Geological Survey.
Available online at http://minerals.usgs.gOv/minerals/pubs/commoditv/aluminum/index.html#mvb
U.S. Geological Survey. 2016b. Mineral Commodity Survey: Aluminum. Reston, VA: U.S. Geological Survey. Available
online at http://minerals.usgs.gov/minerals/pubs/commoditv/aluminum/index.html
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-91
-------
METHODOLOGY DOCUMENTATION
5.2.7 Magnesium Production
The magnesium metal production and casting industry uses SF6 as a cover gas to prevent the spontaneous
combustion of molten magnesium in the presence of air. Fugitive SF6 emissions occur primarily during three
magnesium manufacturing processes: primary production, die casting, and recycling-based production. Additional
processes that may use SF6 include sand and gravity casting, as well as wrought, anode, and permanent mold
casting; however, these are not included in this analysis.
5.2.7.1 Magnesium Production Emission Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite
Emission Projections, for additional information). Activity data for magnesium manufacturing included primary
magnesium production data from USGS (2007, 2009, and 2017), die-casting production data from Harnisch and
Schwarz (2003) and Edgar (2004), automobile production from Ward's Motor Vehicle Data (2001) and the
International Organization of Motor Vehicle Manufacturers (2017), and production growth rates from Webb
(2005). Emission factors were obtained from Gjestland and Magers (1996) and EPA (2010). Emission reductions
were incorporated based on the EPA's SF6 Emission Reduction Partnership for the Magnesium Industry, facility-
reporting to the EPA Greenhouse Gas Reporting Program (GHGRP), regulations phasing-out and banning SF6 in the
European Union, and several magnesium CDM projects.
The 2006 IPCC Tier 1 methodology (IPCC, 2006) was used to estimate emissions from magnesium
manufacturing as follows:
Esf6 = MGc * EFsf6 * 10"3 (5.15)
where:
Esf6 = SFs emissions from magnesium casting, tonnes
MGC = Total amount of magnesium casting or handling in the country, tonnes
EFsf6 = Default emission factor for SF6 emissions from magnesium casting, kg SFs/tonnes Mg casting
In the absence of emission control measures, the rapid growth of the magnesium manufacturing industry is
expected to significantly increase future SF6 emissions from magnesium production and processing. However,
efforts in recent years to eliminate the use of SF6 in this application around the world have reduced potential
emission growth. In 2003, the EPA's SF6 Emission Reduction Partnership for the Magnesium Industry formed a
global industry commitment through the International Magnesium Association (IMA), which represents
approximately 80% of magnesium production and processing outside of China, to eliminate SF6 emissions from
magnesium operations by the end of 2010. The U.S. partnership has ended, but facilities in the United States that
contain magnesium production processes are required to annually report emissions under subpart T of EPA's
Greenhouse Gas Reporting Program (GHGRP) (40 CFR Part 98). In addition, regulatory efforts in Europe and Japan
and CDM projects in Brazil and South Korea have resulted in reduced emissions. So far, two magnesium projects
have been registered with CDM: 1 project from Israel and 1 from Brazil (CDM Project Activities, 2018).
Activity Data
Historical
Activity data for the three different magnesium manufacturing types are discussed below.
Primary Production
Primary magnesium production data for 1990 through 2015 were obtained from the USGS (2007, 2009, 2017).
5-92
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Die-Casting Production
Die-casting production activity data were obtained from various sources as outlined below.
• EU. For Portugal, Spain, and the U.K., historical SF6 emissions were estimated using country-specific data
on SF6-based magnesium casting production from Harnisch and Schwarz (2003). For 1990, emissions were
estimated using the 1995 estimates and two trends between 1990 and 1995: (1) EU auto production and
(2) magnesium used per car in the United States between 1990 and 1995. Based on these trends, SF6
emissions in the EU from die-casting production that were based on car production were assumed to have
increased by 30% between 1990 and 2015, and emission factors were assumed to remain constant over
the same period.
• China (2000, 2005, and 2010 only). The activity data used for Chinese historical SF6 emissions from
magnesium manufacturing was Chinese casting volume for the years 2000, 2005, and 2010 from Edgar
(2004). For the intervening years, die-casting production was estimated using the ratio of country-specific
automobile production to U.S. automobile production. This ratio was multiplied by U.S. die-casting
production to obtain an estimate of die-casting production in each country. Automobile production for
1990 to 2000 was obtained from Ward's Motor Vehicle Data (Ward's, 2001), and 2001 through 2015
production data were obtained from the International Organization of Motor Vehicle Manufacturers
(2017).
• Other countries. For other countries, die-casting production for the years 1990 through 2012 was
estimated as a function of automobile production. For example, for Brazil, Russia, and Ukraine, die-casting
production was estimated using the same methodology for estimating die-casting production in China for
years other than 2000, 2005, and 2010. For countries that did not produce automobiles but had growing
die-casting industries, such as Kazakhstan and Israel (IMA, 2002), production was estimated from the ratio
of primary production to die-casting production for a similar country. Russia was used as a proxy for
estimating production in Kazakhstan, while the United States was used as a proxy for Israel. Taiwan was
assumed to acquire 12.5% of Japan's die-casting activity starting in 2002, increasing linearly to 50% by
2005.
Recycling-Based Production
Recycling-based production for Brazil, China, Russia, the U.K., and countries in the EU was estimated using die-
casting activity and a "remelt factor" of 30%. The secondary production to die-casting ratio ranges from 30% to
55% across countries that actively recycle scrap magnesium (Edgar, 2006), and 30% was chosen as a conservative
default for those countries where emissions were calculated for this source. The Czech Republic was reported to
have a new recycling plant come online in 2002 and was assumed to have an annual growth rate of 3.4% through
2010 and then 1.7% from 2011 through 2015 (Webb, 2005). Table 5-31 presents the growth rates used in this
analysis.
Projected
Annual growth rates for primary casting and recycling production are summarized in Table 5-31. In general,
annual growth rates used in this analysis were assumed to account for new facility construction and facility
capacity expansion, driven by growing global demand for magnesium in applications such as automotive
lightweighting to improve fuel economy. Primary production and die-casting production growth rates were based
on information supplied by Webb (2005) for the rest of the countries' estimates. Recycling production was
estimated based on the growth rates associated with die-casting production.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-93
-------
METHODOLOGY DOCUMENTATION
Table 5-31: Annual Growth Rates for Primary Production, Casting, and Recycling Production (annual
percentage increase)
Primary Production
Annual Growth Rates3
(%)
Casting Annual Growth Rates
(%)
Recycling
Annual Growth
Rates3
(%)
Years
China
Rest of World
Rest of World Asia
China
Europe
Russia
World
2016-2035
5.5b
1.7
00
r*>
t-H
5.0
1.7
4.8
Same as casting0
2036-2050
0.0°
0.8
0.8 2.4
2.5
0.8
2.4
Same as casting0
a Source: Primary and casting growth rates were based on Webb (2005). For recycling, it was assumed that growth rates will be
driven by increasing use in automotive applications; consequently, growth rates will be the same as casting estimates.
b Annual growth for China was estimated to be 5.5% through 2020 and then held at zero for 2020 through 2050.
c For EU countries, this growth rate was applied until 2015 when the EU regulation to phase out the use of SF6 in recycling came
into effect
Primary Production
Primary production was assumed to grow 1.7% per year between 2016 and 2035. Between 2036 and 2050,
growth was assumed to decrease to an annual rate of 0.8%. Primary production in China is projected to grow at a
rate of 5.4% from 2016 through 2030, to be consistent with 2010 through 2015 historical growth (USGS, 2017) and
then projected to hold steady at 2030 levels through 2050. For South Korea, primary production was projected to
grow at 3.4% per year from 2016 through 2035 and 1.7% per year from 2036 through 2050. This is consistent with
Roskill (2016), which estimates that by 2020 the global production growth rate of magnesium metal will average
3.4% per year and the die-casting growth rate will be 4.0%.
Die-Casting Production
In Asia (except China) and Russia, die-casting production is expected to grow at 4.8% per year from 2016
through 2035 (Webb, 2005) and at 2.4% annually between 2036 and 2050. For Europe and other countries such as
Brazil, Israel, Kazakhstan, and Ukraine, die-casting production is estimated to grow at 1.7% annually from 2016
through 2035, and 0.8% annually from 2036 through 2050. From 2016 through 2035, die-casting production in
China is estimated to grow at 5.5% annually. This growth is spurred by increasing investments by western,
Japanese, and Taiwanese companies in China to meet domestic demand for camera, computers, and automobile
parts. Die-casting production between 2036 and 2050 for China was assumed to hold constant.
Recycling-Based Production
For all countries with projected emission estimates, recycling growth rates were set equal to die-casting
growth rates.
Emission Factors
Historical and Projected
SFs emissions were conservatively assumed to be equivalent to SF6 consumption (i.e., it was assumed that no
SFs is destroyed during the metal processes). Table 5-32 and Table 5-33 summarize the emission factors used to
estimate historical emissions for each of the production processes for all countries in this analysis. The emission
factor for primary production was based on measurements made in 1994 and 1995 by U.S. producers. Because of
the similarity between the primary and recycling production processes, the emission factor for recycling
production was assumed to be the same as the emission factor for primary production. The emission factor for die
casting was obtained from a 1996 international survey of die casters performed by Gjestland and Magers (1996). It
was assumed that the 2010 emission factors obtained from the EPA's SF6 Emission Reduction Partnership for the
5-94
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-32: Emission Factors for Primary Production, Casting, and Recycling Production (1990-1999)
Emission Factors
Process
(kg SFe/metric ton Mg produced)3
Source
Primary production
1.10
EPA, 2010
Casting
4.10
Gjestland and Magers, 1996
Recycling
1.10
EPA, 2010
a Emission factors used to estimate emissions from Brazil, China, the Czech Republic, Israel, Kazakhstan, Portugal, Russia, Spain,
Ukraine, and the U.K., as appropriate.
Table 5-33: Emission Factors for Primary Production, Casting, and Recycling Production (2000-2050)
Emission Factors
Process
(kg SFe/metric ton Mg produced)3
Source
Primary production
0.75
EPA, 2010
Casting
1.00
EPA, 2010
Recycling
0.75
EPA, 2010
a Emission factors used to estimate emissions from Brazil, China, the Czech Republic, Israel, Kazakhstan, Portugal, Russia, Spain,
Ukraine, and the U.K., as appropriate.
Magnesium Industry remain constant from 2011 through 2050. The 2000 emission factor for primary production
was based on measurements made by four producers (i.e., producers with U.S. and international operations) (EPA,
2010) and was held constant from 2010 through 2050. For all countries including China, the emission factors for
die casting were estimated based on reports from U.S. die casters and a report on emissions from European die
casters (Harnisch and Schwarz, 2003).
Emission Reductions in Baseline Scenario
Emission reductions that are applicable for the base case scenario include the following:
• In the United States, reductions that occurred as a result of the voluntary partnership are reflected in the
estimates. Additionally, some facilities have reported to EPA's GHGRP program that they use HFC-134a for
magnesium production and processing. Reductions in SF6 emissions due to HFC-134a emissions in the
United States were accounted for in the UNFCCC-reported emission data.
• In the EU, a 2006 regulation, as defined in Article 13 of the EU F-GHG Regulation No. 517/2014 (Office
Journal of the European Union, 2014), banned the use of SF6 in magnesium die casting for plants using
more than 850 kg per year. Starting in 2008, all EU die-casting facilities that use more than 850 kg of SF6
per year were required to stop using SF6. SF6 emissions from smaller facilities (those using less than 850 kg
in die casting), which was assumed to be the average proportion of die-casting emissions reported to the
EPA GHGRP relative to total die-casting emissions (i.e., 41%), were assumed to linearly phase out between
2008 and 2018.
• In 2014, the EU regulation phasing out SF6 emissions from die casting was extended to SF6 emissions from
recycling. Starting in 2015, all EU recycling facilities that use more than 850 kg of SF6 per year were
required to stop using SF6. For recycling facilities smaller than this threshold, emissions were assumed
negligible.
• In 2018, SFs is banned from use in the EU for any magnesium die-casting and recycling process, regardless
of facility size. Therefore, emissions were zeroed out for EU countries from 2018 through 2050.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-95
-------
METHODOLOGY DOCUMENTATION
• Magnesium recyclers in the U.K. have switched to sulfur dioxide (SO2) since 2000; therefore, the U.K.'s SF6
emissions from magnesium recycling from 2000 through 2050 were assumed to be zero.
• Emission reductions from several magnesium CDM projects, including those in Brazil and Israel, are
reflected in the base case. RIMA, a large-scale magnesium production and processing facility in Brazil,
implemented a full conversion to SO2 for its primary, die-casting, and recycling activities and is monitoring
with CDM through 2023 (UNFCCC, 2018). Dead Sea Magnesium and Ortal Diecasting 1993 Ltd. in Israel
implemented a conversion from primary production to HFC-134a and is implementing the SF6 abatement
action plan, with financial support from CDM, through 2019 (UNFCCC, 2010; UNFCCC, 2017). The Brazil
and Israel CDM project reports provided SF6 emission reductions for primary and die-casting production
processes. From the last year of CDM available data (i.e., 2016 for Brazil and 2012 for Israel) through
2050, emission reductions were assumed to remain constant and continue from CDM project
implementation. Emission reductions were assumed to be held constant in absolute terms from the latest
CDM monitoring reports (i.e., from 2016 for Brazil and from 2012 for Israel). The EPA assumed that the
technology implemented to achieve SF6 emission reductions remains in use by the facilities after the end
of the CDM project.
Uncertainties
Historical and projected emissions from this source are affected by both activity levels and emission rates.
Although country-specific activity levels are fairly well known for primary production, they are less well known for
recycling-based production, particularly the share consisting of magnesium-base alloys and the share for die
casting. In addition, emission rates vary widely across different processes and over time. The EPA accounted for
these variations (e.g., the decline in emission rates that occurred between 1995 and 2000), but some regional- and
process-based variability may exist.
In addition, projected emissions from magnesium production and processing are sensitive to estimated
activity growth rates and to assumptions regarding the adoption and/or retention of alternative melt protection
technologies. The EPA used relatively high activity growth rates to project emissions; therefore, slight changes in
these rates would lead to large changes in projected emissions. Furthermore, population growth and economic
development resulting from purchasing power growth will likely affect magnesium consumption on a per capita
basis, and therefore total emissions from magnesium production processes. This analysis does not include
projections based on the effects of such growth.
This analysis also assumed that some, but not all, of Chinese magnesium producers have adopted SF6 in place
of solid sulfur powder as these producers seek to increase metal quality. Because China is currently the world's
largest producer of magnesium, greater penetration of the Chinese market by SF6 would significantly increase both
Chinese and global emissions. On the other hand, penetration of the Chinese die-casting market by alternative
cover gases would lower Chinese emissions below those projected in this analysis.
Finally, the assumption that SF6 emissions are equivalent to SF6 consumption may overstate emissions,
because recent EPA studies have shown that 5% to 20% of the SF6 degrades during its use as a cover gas during at
least one type of casting process (Bartos et al., 2003).
5.2.7.2 Mitigation Options Considered for Magnesium Production
Use of SFs as a cover gas is the only source of GHG emissions from magnesium production. Although studies
indicate some destruction of SFs in its use as a cover gas (Bartos et al., 2003), the analysis described here follows
current IPCC guidelines (2006), which assumes that all SFs used is emitted to the atmosphere.
Replacement with Alternative Cover Gas—Sulfur Dioxide (S02)
Historically, S02 has been used as a cover gas in magnesium production and processing activities. However,
because of toxicity, odor, and corrosivity concerns, S02 use is not common or was discontinued in most countries.
5-96
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Historic and recent S02 technology research aims to improve process feed systems and control technology and to
address the toxicity and odor issues with improved containment and pollution control systems (Environment
Canada, 1998), nonetheless this option is still not very common in most countries. The use of S02 has the potential
to reduce SFs emissions by 100% because a complete replacement of the cover gas system is involved. Currently,
S02 is being used as a cover gas; for example, it is used as a cover gas at one die-casting facility in Brazil (UNFCCC,
2010b). This option was assumed to be technically applicable to all three model facilities. The maximum market
penetration for this option was assumed to be 80% of the emissions of SFs for recycle/remelt facilities and 10% for
both die-casting and primary production facilities. The lifetime of this option was assumed to be 15 years.
Facilities implementing S02 as an alternative cover gas incur capital costs related to the cost for new piping,
pollution control equipment, and safety equipment for workers. The total capital cost was $522,070 for all three
facility types. Facilities also incur annual costs (or generate annual cost savings) based on the purchase price of the
alternative cover gas. This option results in annual gas purchase costs of $5,953 each for die-casting and primary
production facilities and an annual gas purchase cost of $26,591 for recycle/remelt facilities. S02 is significantly less
expensive than SFs, and the required gas replacement ratio is 1:1, resulting in a significant net savings in material
costs. Replacing SFs with S02 also results in avoided costs of $36,910 each for both die-casting and primary
production model facilities and $588,018 for the recycle/remelt model facility associated with the purchase of SFs.
Replacement with Alternative Cover Gas—HFC 134a
Research has shown that candidate fluorinated compounds such as HFC-134a can be a cover gas substitute for
SFs (Milbrath, 2002; Ricketts, 2002; Hillis, 2002). In addition, currently, HFC-134a is used as a cover gas at two die-
casting facilities in Israel (UNFCCC, 2008a, 2008b) and at least four facilities in the United States (EPA, 2018). While
F-GHGs have an advantage over S02 because they have potentially fewer associated health, safety, odor, and
corrosive impacts, some current F-GHG alternatives (including HFC-134a) still have climate impacts, albeit
relatively minimal. The GWP of HFC-134a is significantly less than that of SFs: thus, the GWP-weighted cover gas
emissions could be reduced by 94%. HFC-134a was assumed to be technically applicable to all model facilities. The
maximum market penetration for this option was assumed to be 45% of the emissions of SFs for die-casting and
primary production facilities and 10% for recycle/remelt facilities. The lifetime of this option was assumed to be 15
years.
Facilities implementing HFC-134a as an alternative cover gas do not incur upfront capital costs, as use of HFC-
134a is a simple drop-in option and does not require additional/new systems or training. They incur annual costs
(or generate annual cost savings) based on the purchase price of the alternate cover gas. Use of HFC-134a results
in annual gas purchase costs of $49,055 each for die-casting and primary production facilities and $219,135 for the
recycle/remelt facility. HFC-134a is not only less expensive than SFs, but additionally HFC-134a has a gas
replacement ratio of 0.71:1, resulting in significant net savings in material costs. Replacing SFs with HFC-134a
results in avoided costs of $36,910 each for both die-casting and primary production model facilities and $164,880
for the recycle/remelt facility associated with the purchase of SFs.
Replacement with Alternative Cover Gas—Novec™ 612
Research has shown that candidate fluorinated compounds such as Novec 612 can be a cover gas substitute
for SFs (Milbrath, 2002; Ricketts, 2002; Hillis, 2002). Additionally, currently, Novec612 is being used at one die-
casting facility and one remelt facility in the United States (EPA, 2018). The use of Novec 612 as an alternative
cover gas represents an advantage over S02 because, like other F-GHGs, Novec 612 has potentially fewer
associated health, safety, odor, and corrosive impacts. Novec 612 is a zero-GWP gas and, therefore, has a
reduction efficiency of 100% compared with SFs. Novec 612 was assumed to be technically applicable to all model
facilities.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-97
-------
METHODOLOGY DOCUMENTATION
Facilities implementing Novec 612 as an alternative cover gas incur capital costs related to purchasing
computerized mass flow control cabinets and piping material to direct the gas. The total capital cost was
$270,092.85 for the die-casting facility, $36,462,59 for the recycle/remelt facility, and $528,595.94 for the primary
production facility. Use of Novec 612 results in annual gas purchase costs of $114,799.27 for die-casting and
primary production facilities and $512,819.56 for the recycle/remelt facility. However, because the replacement
ratio of Novec 612 to SF6 is 0.3:1, significantly less Novec 612 is required to process the same quantity of
magnesium. These costs are offset by the avoided costs of purchasing SF6, an annual cost savings of $36,910 for
both die-casting and primary production model facilities and $164,880 for the recycle/remelt model facility.
5.2.7.3 Model Facilities
To evaluate the cost of reducing SF6 emissions from magnesium production, this analysis considered reduction
costs for three typical magnesium production facilities, which were generally characterized based on facility-
specific case studies measuring average SF6 consumption, production capacity, and type. Global SFs emissions from
magnesium production by facility type are shown in Figure 5-3. Model facilities are based on industry data from
the United States but apply to magnesium facilities globally.
We characterize these typical facilities as follows:
• Die-Casting Facility: This model facility represents a medium-sized die-casting facility. The facility was
characterized based on real data from a case study (EPA, 2011) where a given abatement option was
implemented in 2008. The facility produces 26,014 metric tons of magnesium per year and emits 0.17 kg
of SF6 per metric ton of magnesium
produced, representing a total
annual facility emission of 4,483 kg of
SF6. Production and emission data
from 2007 were used to define the
pre-abatement emission baseline
(EPA, 2011).59
• Recycle/Remelt Facility: This model
facility represents a medium-sized
recycle facility. The facility was
characterized based on real data
from a case study where a given
abatement option was implemented
in 2008. The facility produces 18,453
metric tons of magnesium per year
and emits 1.09 kg of SF6 per metric
ton of magnesium produced,
representing a total annual facility
emission of 20,026 kg of SFs.
Production and emission data from
the prior year (i.e., 2007) were used to define the pre-abatement emission baseline (EPA, 2011).
• Primary Production Facility: This facility assumes the same characteristics as the die-casting facility.
Figure 5-3: Global SFe Emissions in 2020 by Facility Type
(% of GWP-Weighted Emissions)
Primary
casting,
51%
production,
Recycle,
59 We relied on data from the EPA case study as it is reliable arid complete. Data from EPA's GHGRP contains confidential
business activity data that was not accessible for this study.
5-98 Global Ncn-CO; Greenhouse Gas Emissions Projections & Klarginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.2.7.4 Technical and Economic Characteristics Summary
Three potential options are available for reducing SFs emissions from magnesium production and processing
operations. These emission abatement measures all include replacing SFs with an alternative cover gas: S02, HFC-
134a, or Novec 612. Table 5-34 presents the reduction efficiency and applicability for the three alternative cover
gas options.
Table 5-34: Magnesium Production Abatement Options
Abatement Option
Reduction Efficiency, %
Applicability
Alternative cover gas—Novec 612
100%
• Die casting
• Recycle/remelt
• Primary production
Alternative cover gas—HFC-134a
95%
• Die casting
• Recycle/remelt
• Primary production
Alternative cover gas—SO2
100%
• Die casting
• Recycle/remelt
• Primary production
Table 5-35 presents the costs and other assumptions associated with the options analyzed. All options have an
assumed lifetime of 15 years.
Table 5-35: Engineering Cost Data on a Facility Basis—Magnesium Production
Project
2016 USD Costs, $
Abatement
Abatement
Lifetime
Annual
Annual
Amount
Option
Facility Type
(years)
Capital
Savings3
O&M
(tCOze)
SO2
Die casting
15
522,070
36,910
5,953
102,212
Recycle/remelt
15
522,070
164,880
26,591
456,593
Primary production
15
522,070
36,910
5,953
102,212
HFC-134a
Die casting
15
—
36,910
49,055
97,661
Recycle/remelt
15
—
164,880
219,135
436,260
Primary production
15
—
36,910
49,055
97,661
Novec 612
Die casting
15
270,093
36,910
114,799
102,210
Recycle/remelt
15
36,463
164,880
512,820
456,582
Primary production
15
5428,596
36,910
114,799
102,210
a These numbers are not net annual savings.
5.2.7.5 References
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.
Bartos, S., J. Marks, R. Kantamaneni, and C. Laush. March 2003. Measured SF6 emissions from magnesium die
casting operations. Magnesium Technology 2003, Proceedings of The Minerals, Metals & Materials Society
(TMS) Conference.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-99
-------
METHODOLOGY DOCUMENTATION
CDM Project Activities. 2018. Project Search. UNFCCC, last viewed November 7, 2018. Available online at
https://cdm.unfccc.int/Proiects/proisearch.html J
Edgar, B. March 2004. SFs Usage in the Chinese Magnesium Industry: 2000-2010. Report prepared for U.S.
Environmental Protection Agency.
Edgar, B. 2006. Personal communication with Bob Edgar, Norsk Hydro Magnesium.
Environment Canada. 1998. Powering GHG Reductions through Technology Advancement. Clean Technology
Advancement Division, Environment Canada.
Gjestland H., and D. Magers. 1996. Practical usage of sulfur hexafluoride for melt protection in the magnesium die
casting industry. Annual Conference Proceedings, International Magnesium Association, Ube City, Japan.
Harnisch, J. and W. Schwarz. 2003. Costs of the Impact on Emissions of Potential Regulatory Framework for
Reducing Emissions of Hydrofluorocarbons, Perfluorocarbons, and Sulphur Hexafluoride. (B4-
3040/2002/336380/MAR/E1). Final Report prepared on behalf of the European Commission (DG ENV).
Hillis, J.E. 2002. The International program to identify alternatives to SF6 in magnesium melt protection. Presented
at the International Conference on SF6 and the Environment: Emissions Reduction Technologies, November
21-22, San Diego, CA.
IMA. 2002. Personal communication with Rick Opatick, International Magnesium Association.
Intergovernmental Panel on Climate Change. 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.
Milbrath, D. 2002. Development of 3M Novec 612 Magnesium Protection Fluid as a Substitute for SF6 in the
Magnesium Industry. Presented at the International Conference on SF6 and the Environment: Emissions
Reduction Technologies, November 21-22, San Diego, CA.
Motor Vehicle Manufacturers. 2017. 2017 Production Statistics. Available online at
http://www.oica.net/categorv/production-statistics/2017-statistics/cf
Office Journal of the European Union. 2014. Regulation (EU) No. 517/2014 of the European Parliament and of the
Council of 16 April 2014 on fluorinated greenhouse gases and repealing Regulation (EC) No. 842/2006.
Available online at http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32014R0517&from=EN J
Organisation Internationale des Constructeurs d'Automobiles. 2017. 2016 Production Statistics. Available online at
http://oica.net/categorv/production-statistics/iJ
Ricketts, N. 2002. Environmental implications of using HFC-134a as a replacement for SF6 in the magnesium
industry. Presented at the International Conference on SF6 and the Environment: Emissions Reduction
Technologies, November 21-22, San Diego, CA.
Roskill. 2016. Magnesium Metal Global Industry, Markets & Outlook. Available online at:
https://roskill.com/market-report/magnesium-metal/cf
United Nations Framework Convention on Climate Change. 2008a. SF6 Switch at Ortal Diecasting 1993 Ltd.
Production. New York: UNFCCC. Available online at http://cdm.unfccc.int/Proiects/DB/TUEV-
SUED1233931497.2/view of"
United Nations Framework Convention on Climate Change. 2008b. SF6 Switch at Dead Sea Magnesium. New York:
UNFCCC. Available online at http://cdm.unfccc.int/Proiects/DB/TUEV-SUED1235638608.46/viewcf
United Nations Framework Convention on Climate Change. 2018. Conversion of SF6 to the alternative cover gas
S02 at RIMA magnesium production. CDM Project Design Document. Available online at
https://cdm.unfccc.int/Proiects/DB/TUEV-SUED1239262577.48/viewJ
5-100
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
United Nations Framework Convention on Climate Change. 2010. Conversion of SF6 to the Alternative SChat RIMA
Magnesium Production. New York: UNFCCC. Available online at http://cdm.unfccc.int/Proiects/DB/TUEV-
SUED1239262577.48/view cf
U.S. Environmental Protection Agency. 2011. Case Study: U.S. Magnesium Industry Adopts New Technology for
Climate Protection. Washington, DC: EPA.
U.S. Environmental Protection Agency. 2012. Global Anthropogenic Non-CCh Greenhouse Gas Emissions: 1990-
2030. EPA 430-R-12-006. Washington, DC: EPA. Available online at
http://www.epa.gov/climatecliange/economics/intemational.litinl
U.S. Environmental Protection Agency. 2018. Greenhouse Gas Reporting Program Data. Washington, DC: EPA.
Available online at https://www.epa.gov/ghgreporting/ghg-reporting-program-data-sets
U.S. Geological Survey. 2007. Minerals Yearbook 2007: Magnesium. Reston, VA: U.S. Geological Survey. GPO Stock
#024-004-02538-7.
U.S. Geological Survey. 2009. Minerals Yearbook 2009: Magnesium. Reston, VA: U.S. Geological Survey. GPO Stock
#024-004-02538-7.
U.S. Geological Survey. 2017. Minerals Yearbook 2015: Magnesium. Reston, VA: U.S. Geological Survey.
Ward's. 2001. Ward's World Motor Vehicle Data. Southfield, MO.
Webb, D. 2005. Magnesium supply and demand 2004. International Magnesium Association Conference, May 22-
24, Berlin, Germany.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-101
-------
METHODOLOGY DOCUMENTATION
5.2.8 Use of Substitutes for Ozone-Depleting Substances
This section covers projections and mitigation options for the following source categories:
• HFC emissions from refrigeration and ACs
• HFC emissions from solvent use
• HFC emissions from foams manufacturing
• HFC emissions from aerosol product use
• HFC and PFC emissions from fire protection
5.2.8.1 Use of Substitutes for Ozone-Depleting Substances Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. If country-reported emission estimates were not available for any historical years, the
EPA used a modeling approach to determine emissions from the various ODS substitute end-use sectors
(refrigeration/AC, foams, aerosols, fire extinguishing, and solvents). EPA modeled emissions based on country-
reported ODS consumption data to United Nations Environment Programme (UNEP) (2018). Allocation of
emissions to each end-use sector in the United States was modeled using the Vintaging Model from EPA (2017).
Emissions by end-use sector from non-U.S. countries were estimated by building on the U.S. assessment; then
country-specific adjustments were made using factors for economic growth, timing of the ODS phase-out, types of
alternatives employed, distribution of ODSs across end-uses in each region or country, and country-specific HFC
consumption from the Multilateral Fund (MLF) from UNEP (2017a). The model incorporated measures to reduce or
eliminate future emissions of these gases based on regulations by the United States, European Union, Australia,
Canada, and Japan. Specific deviations from this basic methodology were made for several sectors and are
discussed in the sections that follow. This methodology is described in more detail in the following sections.
Activity Data
The EPA used the Vintaging Model of ODS- and ODS-substitute-containing equipment and products to
estimate the use and emissions of ODS substitutes in the United States. The model tracks the use and emissions of
each of the substances separately for each of the ages or "vintages" of equipment. The Vintaging Model is used to
produce the ODS substitute emission estimates in the official U.S. GHG Inventory and is updated and enhanced
annually. For this analysis, the Vintaging Model was adapted slightly to include data sources common to each
source category (e.g., GDP).60
After U.S. emissions were calculated using the Vintaging Model, the EPA developed emission estimates for
non-U.S. countries by building on the detailed U.S. assessment. In developing these estimates, the EPA initially
assumed that the transition from ODSs to HFCs follows the same substitution patterns as the United States. The
U.S.-based substitution scenarios were then customized to each region or country using adjustment factors that
take into consideration differences in historical and projected economic growth, the timing of the ODS phase-out,
the type of alternatives employed, and the distribution of ODSs across end uses in each region or country. The
general methodology and assumptions used by the EPA are discussed below, although the methodology was
modified for several sectors when necessary. Specific deviations from this basic methodology are discussed
following the general methodology description.
The consumption of ODS and ODS substitutes was modeled by estimating the quantity of equipment or
products sold, serviced, and retired each year and the amount of the chemical required to manufacture and/or
maintain the equipment over time. The model estimates emissions by applying an emission profile (e.g., annual
leak rates, service emission rates, and disposal emission rates for AC and refrigeration end uses) to each
60 A discussion of the Vintaging Model can be found in the U.S. Inventory of Greenhouse Gas Emissions and Sinks (EPA, 2017).
5-102
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
population of equipment. The model estimates and projects annual use and emissions of each compound over
time by aggregating the consumption and emission output from approximately 65 different end uses.
The major end-use sectors defined in the Vintaging Model for characterizing ODS use in the United States are
refrigeration and AC, aerosols (including metered-dose inhalers [MDIs]), solvent cleaning, fire extinguishing
equipment, foam production, and sterilization. The Vintaging Model estimates the use and emissions of ODS
substitutes by taking the following steps:
1. Gather historical emission data. The Vintaging Model is populated with information on each end use,
taken from published and confidential sources and industry experts.
2. Simulate the implementation of new, non-ODS technologies. The Vintaging Model uses detailed
characterizations of the historical and current uses of the ODSs, as well as data on how the substitutes are
replacing the ODSs, to simulate the implementation of new technologies that ensure compliance with
ODS phase-out policies. As part of this simulation, the ODS substitutes are introduced in each of the end
uses over time as seen historically and as projected for the future considering the need to comply with
the ODS phase-out.
3. Estimate emissions of the ODS substitutes. The chemical use is estimated from the amount of substitutes
that are required each year for the manufacture, installation, use, or servicing of products. The emissions
were estimated from the emission profile for each vintage of equipment or product in each end use. By
aggregating the emissions from each vintage, a time profile of emissions from each end use is developed.
Historical
The following general steps were applied to estimate country-specific emissions. Steps 1 through 7 result in
preliminary emission estimates calculated by Equation 5.16 below. The preliminary consumption estimates were
adjusted to emissions and based on a regional or economic factor as discussed in Steps 8 through 11.
1. Gather base ODS consumption data for each country. UNEP (2010) provided reported ODS consumption in
terms of ozone depletion potential (ODP)-weighted totals for the major types of ODSs: CFCs, HCFCs,
halons, carbon tetrachloride, and methyl chloroform. The base year for estimates was 1989 because, in
general, ODS substitution had not yet taken place; when data for 1989 were unavailable, the earliest
available data were used as a proxy. Because data were only available in ODP-weighted totals by ODS
"group," groups were divided into component chemicals (e.g., CFC-11, CFC-12) according to 1990 U.S.
percentages as modeled in the Vintaging Model. After disaggregating the ODP-weighted consumption by
chemical, ODPs were used to determine the total consumption in metric tons.
2. Calculate the percentage of base ODS consumption of each chemical group used in each end-use sector.
The amount of ODSs used in various industrial sectors differs by country. Data on the end-use
distributions of ODS in 1990 were available for the following countries:
- United States from the Vintaging Model
- UKfrom U.K. Use and Emissions of Selected Halocarbons, prepared for the Department of the
Environment (March 1996)
- Russia from Phaseout of Ozone Depleting Substances in Russia, prepared for the Ministry for
Protection of the Environment and Natural Resources of The Russian Federation and the Danish
Environmental Protection Agency (Russian Federation, 1994)
The 1990 end-use sector distribution for the United States was applied to Canada and Japan. The UK's
distribution was applied to the EU-15,61 non-EU Western Europe,62 Australia, and New Zealand. Russia's
distribution was applied to the FSU and Eastern European countries. For developing countries, data on the 1990
61 For the purposes of this report, the U.K. is considered part of the EU. Hence, the EU-15 is defined as these EU members:
Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain,
Sweden, and the UK.
62 Iceland, Liechtenstein, Monaco, Montenegro, Norway, and Switzerland.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-103
-------
METHODOLOGY DOCUMENTATION
consumption of ODS were available for many nations63 by sector and substance from the Multilateral Secretariat.
For developing countries that did not have data available, the EPA used a representative average of the published
data (Rest of World).
3. Calculate the base consumption ofODSsfor each end-use sector. This step involves multiplying the
amount of consumption of each chemical group from Step 1 by the end-use sector distribution
percentages from Step 2.
4. Obtain conversion ratios. Ratios of HFC consumption to base ODS consumption and HFC emissions to base
HFC consumption were obtained from the Vintaging Model for each given year, chemical, and end use.
These ratios are used to convert ODS consumption to HFC emissions.
6. Estimate HFC consumption in metric tons. This step involves multiplying the country-specific base level
consumption of ODSs (Step 3) by the ratio of HFC consumption to base-level ODS consumption (Step 4).
6. Estimate HFC emissions in metric tons. This step requires multiplying the HFC consumption (Step 5) by the
ratio of HFC emissions to HFC consumption (Step 4).
7. Estimate GWP-weighted ODS substitute emissions in metric tons ofC02 equivalent. This step involves
multiplying HFC emissions (Step 6) by an average GWP to derive GWP-weighted HFC emissions. The
average GWP, which varies by sector, is determined by examining the estimated ODS substitute emissions
in 2015 in the United States, as obtained from the Vintaging Model. The year 2015 was used as a
representative average; the U.S. HFC market was assumed to be mature by this date and, under a BAU
scenario, the mix of HFCs and other ODS substitutes (and hence the average GWP) is not expected to
change significantly thereafter. For instance, this year is beyond the recent (January 1, 2010) U.S. and
Montreal Protocol HCFC phaseout step. Regional policies that are expected to change the mix of HFCs
(and hence average GWP), such as the EU F-GHG Regulation, are incorporated in a later step.
Equation 5.16: Preliminary Estimate Emissions Calculations
HFC HFC
Consumption Emissions
ODS (MT) (MT) Average GWP of
HFC Emissions Consumption [U.S., year] [U.S., year] HFC Emissions
(MtCOze) = (MT) X ODS X HFC X (MtC02e/MT)
[country, year] [country, 1989 or Consumption Consumption [U.S., 2015]
as available] (MT) (MT)
[U.S., 1989] [U.S., year]
Step 3 Step 5 Step 6 Step 7
63 Algeria, Antigua and Barbuda, Argentina, Bahrain, Bangladesh, Barbados, Belize, Benin, Bolivia, Brazil, Burkina Faso, Burma,
Cameroon, Chile, China, Columbia, Costa Rica, Croatia, Cuba, Dominica, Dominican Republic, Ecuador, Egypt, El Salvador,
Ethiopia, Georgia, Ghana, Grenada, Guatemala, Guyana, Honduras, India, Indonesia, Iran, Jamaica, Jordan, Kenya, Lebanon,
Lesotho, Macedonia, Madagascar, Malawi, Malaysia, Maldives, Mali, Malta, Mauritius, Mexico, Moldova, Mongolia, Morocco,
Mozambique, Namibia, Nepal, Nicaragua, Niger, Nigeria, Pakistan, Panama, Paraguay, Peru, Philippines, Saint Lucia, South
Korea, Sri Lanka, Sudan, Swaziland, Syria, Thailand, Togo, Trinidad and Tobago, Tunisia, Turkey, Uganda, Uruguay, Venezuela,
Vietnam, Yemen, Zambia, and Zimbabwe.
5-104
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
This methodology is followed for each country, given year, and end-use category (e.g., refrigeration). This
equation thus produces preliminary estimates based on the general assumption that all countries will transition
away from ODSs in a similar manner as the United States. (For example, CFC-12 mobile air-conditioners
transitioned to HFC-134a beginning in 1994 in the United States. Thus, as a first estimation, it was assumed that
CFC-12 mobile air-conditioners transition to HFC-134a in other countries.) In many cases, options for ODS
substitutes in each end use are technically limited to the same set of alternatives, regardless of geographic region.
Furthermore, alternative technologies used in the United States are available and in many cases are used
worldwide. These assumptions may be adjusted in later steps to account for differences between the United States
and other countries, as explained below.
8. Develop and apply adjustment factors. In this analysis, the EPA applied adjustment factors to modify the
emission estimates for countries based on what is known qualitatively about how their transition to
alternatives and technology preferences will likely differ from that of the United States. For example, the
EPA multiplied the estimates produced in Step 7 by adjustment factors of less than 1 to refrigeration and
AC end uses because some nations have been more likely to use hydrocarbon (HC) refrigerants than HFCs
and/or because some nations may choose less emissive designs or practices. Table 5-36 shows the
adjustment factors used for each sector and country grouping.
Table 5-36: Adjustment Factors Applied in by Sector and Country
Sector
Country
Ref/AC
Aerosols
Foams
Solvents
Fire-Ext.
Australia/New Zealand
0.90
1.00
1.00
1.00
1.00
China/economies in transition
0.80
1.00
1.00
1.00
1.00
EU
0.70
1.00
1.00
1.00
1.00
Non-EU Europe
0.80
1.00
0.00
1.00
1.00
Japan
0.70
1.00
1.00
1.00
1.00
Rest of World
0.80
1.00
0.00
1.00
1.00
Projected
The following general steps were applied to project country-specific emissions following the adjustments in
the steps above.
9. Develop timing factors. Because most developing countries will transition to substitutes more slowly, the
EPA reduced the adjusted emission estimates by multiplying the results in each year by a timing factor to
reflect the assumed delay in their transition. In the Montreal Protocol, developing countries are listed
under Article 5.64 Timing factors for CFCs start at 25% in 1995 and increase by 25% at each 5-year interval,
until they reach 100% in 2010, consistent with the Montreal Protocol data for their CFC phaseout and
when they are assumed to have caught up to the developed countries. Article 5 countries also have a
delayed phase-out of HCFCs to account for the fact that these countries can continue consuming new
HCFCs through 2040 with specific step-downs based on the 2007 Adjustment to the Montreal Protocol.
These factors are outlined in Table 5-37.
64 A complete list of Article 5 countries is available at
http://ozone.unep.ore/Ratification status/list of article 5 parties.shtm cf.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-105
-------
METHODOLOGY DOCUMENTATION
Table 5-37: Timing Factors Used for Developing (Article 5) Countries
Factor
1995
2000
2005
2010
2015
2020
2025
2030
2035
2040
2045
2050
CFCs
0.25
0.50
0.75
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
HCFCs
0.00
0.00
0.00
0.00
0.11
0.35
0.68
0.98
0.98
1.00
1.00
1.00
10. Develop economic growth factors. Because other countries' economies are growing at different rates than
the United States' economy, the EPA altered emissions based on comparisons between U.S. and regional
historical and projected GDP. These GDP growth factors are shown in Table 5-38.
Table 5-38: GDP Growth Factors (relative to U.S.)
Region
1995
2000
2005
2010
2015
2020
2025
2030
2035
2040
2045
2050
Africa
0.92
0.89
1.02
1.26
1.31
1.39
1.50
1.61
1.69
1.76
1.82
1.87
Asia
1.23
1.14
1.28
1.60
1.83
2.09
2.37
2.69
2.86
3.04
3.22
3.41
Australia/New
Zealand
0.99
0.98
1.02
1.12
1.15
1.19
1.21
1.23
1.27
1.31
1.35
1.40
Brazil
1.02
0.92
0.94
1.12
1.06
1.00
1.04
1.06
1.09
1.12
1.15
1.18
Canada
0.96
0.95
0.95
0.97
0.96
0.96
0.96
0.95
0.95
0.95
0.95
0.97
Central/South
America
1.10
1.00
1.03
1.24
1.30
1.29
1.35
1.43
1.48
1.53
1.58
1.63
China
1.57
1.92
2.70
4.44
5.81
6.99
8.06
9.27
9.87
10.53
11.00
11.21
Eastern Europe
0.46
0.39
0.48
0.54
0.47
0.48
0.51
0.54
0.56
0.58
0.60
0.62
Economies in
Transition
0.55
0.48
0.57
0.65
0.62
0.58
0.57
0.57
0.57
0.55
0.51
0.49
EU
0.95
0.89
0.86
0.87
0.82
0.80
0.78
0.76
0.75
0.75
0.74
0.74
Europe (non-
EU)
0.95
0.89
0.86
0.90
0.87
0.85
0.84
0.83
0.82
0.81
0.81
0.81
India
1.13
1.23
1.50
2.15
2.67
3.49
4.53
5.68
6.64
7.66
8.74
9.84
Japan
0.94
0.80
0.75
0.73
0.68
0.63
0.59
0.56
0.53
0.51
0.50
0.48
Mexico
0.98
1.01
0.97
1.03
1.06
1.09
1.14
1.19
1.27
1.38
1.51
1.63
Middle East
0.97
0.97
1.09
1.33
1.42
1.51
1.61
1.72
1.80
1.87
1.93
1.98
South Korea
1.29
1.34
1.49
1.76
1.83
1.89
1.90
1.89
1.88
1.87
1.86
1.86
Source: USDA(2017).
11. Estimate adjusted HFC emissions in metric tons ofCC>2 equivalent in a given year by country. The EPA
estimated emissions and projections for each year by multiplying the estimates in Step 7 by the
adjustment factors (Step 8), the timing factors (Step 9), and the growth factors (Step 10).
Country-Specific Adjustments
Nations that have ratified the Montreal Protocol are required to report ODS consumption by chemical "group"
(e.g., CFCs) to the UNEP Ozone Secretariat; as of this report, all 197 nations and the European Union had ratified
the Montreal Protocol. Where available, EPA made country-specific adjustments using reported HFC consumption
data published by the MLF for 77 low volume consuming (LVC) and 42 non-LVC countries (UNEP, 2017b).
5-106
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
For countries included in the MLF study, 2015 HFC consumption estimates replace the country-specific HFC
consumption estimates developed under Steps 1 through 4 using ODS consumption data. Total HFC consumption
for the 77 LVC and 42 non-LVC countries in 2015 was disaggregated by country using each country's relative 2012
HCFC consumption (UNEP, 2017b). Consumption estimates for each country were further disaggregated by sector
based on reported survey data. Consumption estimates from 1990 through 2050 were then determined for each
year by back- and forecasting consumption from the 2015 estimate (by applying the economic growth factors
developed in Step 10) and then emissions were estimated using the approach outlined in Step 11.
Emission Factors
In addition to the adjustments discussed above, the EPA adjusted the methodology for some sectors to
account for information that was available on a country or regional scale. These adjustments are discussed by
sector in more detail below.
Historical and Projected
Fire Extinguishing
The EPA adjusted global emissions in the fire extinguishing sector by region by developing Vintaging Model
scenarios that were representative of country- and region-specific substitution data. In addition, the EPA adjusted
emissions in the EU to account for the rapid halon phase-out due to regulation. Details of these adjustments
include the following:
• To estimate baseline emissions, information collected on current and projected market characterizations
of international total flooding sectors was used to create country-specific versions of the Vintaging Model
(i.e., country-specific ODS substitution patterns). For this report, current and projected market
information was obtained on new total flooding systems in which halons have been previously used.
Information for Australia, Brazil, China, India, Japan, Russia, and the UK was obtained from Halon
Technical Options Committee members from those countries.65 Information for the United States was
taken from the Vintaging Model. General information was also collected on northern, southern, and
eastern Europe. Baseline emission information from some of these countries was used to adjust the
substitution patterns for all other countries not listed above, as described below:
- Australia: proxy for New Zealand
- Brazil: proxy for countries in Latin America and the Caribbean
- India: proxy for all other developing countries
- Eastern, northern, and southern Europe: proxies for European countries (based on geography)
- Russia: proxy for economies in transition
An adjustment factor was applied to EU countries to account for European Regulation 2037/2000 on
Substances that Deplete the Ozone Layer, which mandates the decommissioning of all halon systems and
extinguishers in the EU-15 by the end of 2003 (with the exception of those applications that are defined as critical
uses). To reflect this, the methodology assumed that all halon systems in the EU-15 will be decommissioned by
2004. No adjustments were made to the 13 countries that joined the EU since May 2004, because the regulation
makes exceptions for these countries.
Refrigeration and Air-Conditioning
The EPA adjusted estimates for the refrigeration and AC sector to account for less refrigerant recovery (i.e.,
more venting) in developing countries. These estimates assumed that recovery does not occur in these countries in
65 Fire protection experts in these countries provided confidential information on the status of national halon transition
markets and average costs to install the substitute extinguishing systems in use (on a per volume of protected space basis) for
2001 through 2020.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-107
-------
METHODOLOGY DOCUMENTATION
any small refrigeration and AC units but does occur in larger units, such as chillers. The resulting adjustment factors
are shown in Table 5-39.
Table 5-39: Recycling Adjustment Factors Applied to Refrigeration Emission Estimates
Year
1995
2000
2005
2010
2015
2020
2025
2030
2035
2040
2045
2050
Adjustment
factor
1.00
1.01
1.03
1.09
1.12
1.28
1.51
1.77
1.79
1.74
1.69
1.68
Aerosols
Because the ban on CFC use in MDI aerosols caused the United States to transition out of CFCs earlier than
other countries, the U.S. consumption of ODSs in 1990 for non-MDI aerosols was assumed to be zero. To
determine a nonzero denominator for the ratio calculated in Step 4, it was assumed that 15% of the non-MDI
aerosols ODS consumption transitioned to HFCs, while the remainder was assumed to transition to not-in-kind
(NIK) or HC alternatives.
Foams
Most global emissions were estimated in the foam-blowing sector by developing Vintaging Model scenarios
that were representative of country- or region-specific substitution and consumption patterns. To estimate
baseline emissions, current and projected characterizations of international total foams markets were used to
create country- or region-specific versions of the Vintaging Model. The market information was obtained from
Ashford (2004), based on research conducted on global foam markets. Scenarios were developed for Japan,
Europe (both EU and non-EU countries combined), other developed countries (excluding Canada), countries with
economies in transition, and China. It was assumed that other Non-Al countries would not transition to HFCs
during the scope of this analysis, as reflected by the foams adjustment factor (Step 8 above). Once the Vintaging
Model scenarios had been run, the emissions were disaggregated to a country-specific level based on estimated
1989 CFC consumption for foams developed for this analysis. Emission estimates were adjusted slightly to account
for relative differences in countries' economic growth as compared with the United States (Step 10 above).
Emission Reductions in the Baseline Scenario
For this analysis, the model calculated a BAU case that does not incorporate measures to reduce or eliminate
the future emissions of these gases, other than those regulated by U.S., EU, Australian, Canadian, and Japanese
law. For the United States, the Vintaging Model includes transitions to low-GWP alternatives to reflect compliance
with rules issued under EPA's Significant New Alternatives Policy Program.66 For the EU, the HFC consumption
reduction schedule from the F-GHG regulations was applied evenly, starting in 2015, to the consumption estimates
across all countries in the EU and across all sectors. For Australia and Canada, the reduction schedule from their F-
GHG regulations was applied evenly, starting in 2020, to the consumption estimates. For Japan, consumption
estimates have been adjusted according to mandated transitions to alternatives with GWPs below a target
threshold in select refrigerants/AC, foam, and aerosol propellants end uses. No other F-GHG policies have been
applied.
Furthermore, the model does not project future market transitions, including those anticipated by industry.
There is significant uncertainty as to what compounds will replace HFCs in ODS substitutes' applications,
particularly in developing countries.
Uncertainties
ODSs and their substitutes are first consumed during manufacture (e.g., to charge a refrigerator). These gases
are then mostly emitted to the atmosphere over time from equipment leaks, services, and disposal. Some
66 At the time of publication, the Court of Appeals for the District of Columbia Circuit vacated these rulemakings.
5-108
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
consumption may be recovered or recycled, depending on the end use and country. The relationship between
initial consumption and eventual emission is complex and uncertain.
5.2.8.2 Mitigation Options Considered
Refrigeration and AC Mitigation Options Considered
A number of HFCs are used in refrigeration and AC systems and are emitted to the atmosphere during
equipment operation, repair, and disposal, unless recovered, recycled, and ultimately destroyed. The most
common HFCs are HFC-134a, R-404A, R-410A, R-407C, and R-507A.67 In response to the ODS phaseout, equipment
is being retrofitted or replaced to use HFC-based substitutes or intermediate substitutes (e.g., HCFCs) that will
eventually need to be replaced by HFCs or other non-ozone-depleting alternatives.
This analysis considered reduction costs for applying 19 new technologies and using three types of improved
technician practices. Table 5-40 summarizes the technology and practice options reviewed, the types of equipment
that were assumed to adopt such options, and the associated system type definitions (i.e., the equipment
characteristics assumed to develop the option costs).
Table 5-40: Refrigeration and AC Abatement Options
Abatement Option
Reduction
Efficiency
Applicability
HFO-1234yf in MVACs
100%
New MVACs in light-duty vehicles
R-513A in buses and trains
57%
New AC systems in buses and trains
R-407A/R-407F in new large retail food
50%
New large retail food refrigeration systems
HFC secondary loop and/or cascade systems in
new large retail food
50%
New large retail food refrigeration systems
C02 transcritical systems in new large retail
food
100%
New large retail food refrigeration systems
CO2 in medium retail food refrigeration
systems
100%
New medium retail food refrigeration systems
HC in ice makers
100%
New ice makers
R-452A in refrigerated trucks/trailers and
intermodal containers
45%
New refrigerated trucks/trailers and intermodal
containers
NH3 or CO2 in large refrigeration systems
100%
New industrial process refrigeration (IPR) and
cold storage systems
HCs in new domestic refrigeration
100%
New domestic refrigerators
MCHX in new commercial unitary AC systems
37.5%
New commercial unitary AC equipment
R-32 in new commercial unitary AC Equipment
67%
New commercial unitary AC equipment
R-32 with MCHX in new commercial unitary AC
87%
New commercial unitary AC equipment
R-452B with MCHX in new residential and
commercial unitary AC
87%
New residential unitary AC equipment
(continued)
67 R-404A, R-410A, R-407C, and R-507A refrigerant blends are composed of HFCs. Specifically, R-404A is 44% by weight HFC-
125, 52% HFC-143a, and 4% HFC-134a. R-410A is 50% HFC-32 and 50% HFC-134a. R-407C is 23% HFC-32, 25% HFC-125, and 52%
HFC-134a. R-507A (also called R-507) is 50% HFC-125 and 50% HFC-143a.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-109
-------
METHODOLOGY DOCUMENTATION
Table 5-40: Refrigeration and AC Abatement Options (continued)
Abatement Option
Reduction
Efficiency
Applicability
R-290 in new residential unitary ac equipment
100%
New residential unitary AC equipment
R-452B in new heat pumps
67%
New heat pumps
HCs in new window ac and dehumidifiers
100%
New window AC and dehumidifiers
R-32 in new window ac and dehumidifiers
68%
New window AC and dehumidifiers
R-452B in new positive displacement chillers
64%
New positive displacement chillers
R-450A/R-513A in new centrifugal chillers
57%
New centrifugal chillers
HCFO-1233zd(E) in new centrifugal chillers
99%
New centrifugal chillers
Leak repair
40%
All existing large equipment (i.e., large retail
food, IPR, cold storage, and chillers)
Refrigerant recovery at servicing
95%
All small equipment (i.e., MVACs, unitary AC)
Refrigerant recovery at disposal
85%
All existing refrigeration/AC reaching disposal
HFO-1234yf in New MVACs
Hydrofluoroolefin (HFO)-1234yf has a GWP of only 4 and performs similarly to HFC-134a, making the use of
current MVAC system designs with minimal changes feasible. HFO-1234yf is, however, slightly flammable
(designated 2L flammability in Addendum H to ANSI-ASHRAE Standard 34-2010), which may necessitate certain
safety mitigation strategies. This option has already begun penetrating the EU and U.S. markets in a couple of
models (Refrigeration and Air Conditioning Magazine, 2013). This option was assumed to be as efficient as
conventional HFC-134a MVAC systems (Oko-Recherche et al., 2011; Koban, 2009). This option is applicable to a
newly manufactured MVAC system in a light-duty vehicle in developing countries because it was assumed to
penetrate the baseline market in the EU and other developed countries.
• Capital Cost: The one-time capital cost was estimated at approximately $110 per MVAC system, resulting
from incremental refrigerant costs and hardware changes (EPA and NHTSA, 2011; Centro Ricerche Fiat,
2008). The hardware costs were assumed to be 10% greater in developing countries.
• Annual O&M Costs: Annual costs were assumed to be approximately $3 per system associated with
incremental refrigerant replacement costs.
• Annual Revenue: No annual savings were assumed.
R-513A in New Buses and Trains
This abatement option applies to the use of R-513A, a nonflammable blend of HFC-134a and HFO-1234yf in AC
systems for buses and trains. The refrigerant charge size for bus and train AC systems is much larger than for light-
duty vehicle and car MVACs; therefore, the use of flammable or mildly flammable alternatives is more of a concern
(UNEP 2018). Although the use of HFO-1234yf in buses and trains has been modeled in the past, HFO-1234yf is not
a likely option because it is flammable and egress options are limited. This option is applicable to newly
manufactured AC systems in buses and trains in developed and developing countries.
• Capital Cost: R-513A is very similar to HFC-134a, such that existing equipment requires little to no change
to achieve the desired operational characteristics. Therefore, no one-time costs were assumed.
• Annual O&M Costs: Annual costs were assumed to be approximately $9 per system associated with
incremental refrigerant replacement costs.
• Annual Revenue: No annual savings were assumed.
5-110
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
R-407A/R-407F in New Large Retail Food
This abatement option was applied to a newly manufactured large retail food system in a large supermarket.
This option replaces HFC refrigerant blends of R-404A and R507A with lower GWP HFC refrigerant blends R-407A
and R-407F, which requires little to no change to achieve the desired operational characteristics. When designing
new systems, additional modifications may be needed, such as changing the orifice or thermostatic expansion
valve size to achieve the same efficiency (ACHR News, 2012). In addition, HFC refrigerant blends R-407A and R-
407F offer a 10% to 15% increase in energy efficiency compared with R-404A (Linde Gas, 2014; Honeywell, 2014).
This option is applicable to newly manufactured large retail food systems in developing countries because this
transition is assumed to have largely occurred in developed countries.
• Capital Cost: R-404A and R-507A are very similar to R-407A and R-407F in their key chemical properties,
such that existing equipment requires little to no change to achieve the desired operational
characteristics. Therefore, no one-time costs were assumed.
• Annual O&M Costs: R-404A and R-507A have prices roughly equal to R-407A and R-407F. Therefore, this
analysis assumed no annual costs are incurred when transitioning to these alternative refrigerants.
• Annual Revenue: Annual savings are estimated to equal almost $17,200 per supermarket, associated with
reduced energy consumption of 12.5% compared with a baseline energy usage of 1.2 million kWh/year
and assuming electricity cost savings are 66% greater for developing countries.
HFC Secondary Loop and/or Cascade Systems in New Large Retail Food
Secondary loop systems use two fluids: a primary refrigerant and a secondary fluid. The secondary fluid is
cooled by the primary refrigerant in the machine room and then pumped throughout the store to remove heat
from the display equipment. In supermarkets, secondary loop systems are also sometimes used in combination
with a cascade system. Cascade designs consist of two independent refrigeration systems that share a common
cascade heat exchanger. The heat exchanger acts as the low-temperature refrigerant condenser and serves as the
high-temperature refrigerant evaporator. Each component of a cascade design uses a different refrigerant that is
most suitable for the given temperature range, with C02 commonly used in the low-temperature circuit and an
HFC used as the refrigerant at the medium-temperature phase (Refrigeration, Air Conditioning and Heat Pumps
Technical Options Committee, 2011). Because the HFC refrigerant is contained in the machine room in a secondary
loop system and is not required for use in the low-temperature circuit of a cascade design, these systems require a
significantly lower refrigerant charge and have lower leakage rates, resulting in approximately 90% less annual
leakage. While historically these systems were less efficient than conventional DX systems, today these systems
are just as efficient as conventional DX systems, if not more so, because of simplified piping, newly designed
circulating pumps, and fewer components (Wang et al., 2010; DelVentura et al., 2007; Heath and Armer, 2012;
WalMart, 2006; Hinde, Zha, and Lan, 2009). In addition, HFC refrigerant blends R-407A and R-407F offer up to a
15% increase in energy efficiency compared with R-404A (Linde Gas, 2014; Honeywell, 2014).
This abatement option was applied to a newly manufactured large retail food system in a large supermarket.
• Capital Cost: The one-time cost in developed countries was estimated to be up to 25% more expensive
than conventional DX systems (IPCC, 2005); it was assumed that conventional DX systems cost roughly
$182,000 for a large (60,000 sq. ft.) supermarket. The incremental increase is, therefore, approximately
$45,500 per supermarket; this capital cost was estimated to be 10% greater in developing countries.
• Annual O&M Costs: This analysis did not assume O&M costs.
• Annual Revenue: Secondary loop systems were assumed to reduce annual direct emissions by reducing
charge size by 70% and reducing the annual leak rate from 15% to 5%. Annual savings are estimated to
equal almost $4,600 per supermarket associated with reduced refrigerant charge and leakage and
reduced energy consumption of 5% compared with a baseline energy usage of 1.2 million kWh/year;
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-111
-------
METHODOLOGY DOCUMENTATION
annual electricity cost savings were assumed to be 66% greater for developing countries, resulting in
annual savings of approximately $7,300.
CO2 Transcritical Systems in New Large Retail Food
This option eliminates the use of HFCs in large retail food refrigeration systems through the use of C02 as the
primary refrigerant in a transcritical cycle. C02 transcritical systems are similar to traditional centralized DX designs
but must operate at high pressures to accommodate the low critical temperature of C02 (GTZ Proklima, 2008). As a
result, special controls and component specifications must be incorporated into the system design, which often
results in higher upfront costs (Environmental Investigation Agency, 2012). Additionally, C02 transcritical systems
operate most efficiently in cooler climates, performing an estimated 5% to 10% more efficiently than DX systems
using an HFC refrigerant in regions with an average annual temperature below 50°F (Supermarket News, 2012). At
the same time, because of a possible energy penalty, the use of C02 transcritical systems in warmer climates is
currently considered less viable. Specifically, use of these systems is most widely accepted in areas where the
maximum ambient temperature is frequently below 88°F (ACHR News, 2010) or where the average annual
temperature is lower than 59°F (Hill PHOENIX, 2012). In 2012, it was estimated that over 1,300 C02 transcritical
systems were in operation in Europe with installations as far south as Italy and Spain, in addition to a handful of
systems that have been installed in Canada (Shecco, 2012). In 2018, it was reported that 196 U.S. supermarkets
within the GreenChill partnership used CO2 transcritical systems (R744, 2019a).
This abatement option was assumed to be applied to a newly manufactured large retail food system in large
supermarkets in cooler climates in developed and developing countries.
• Capital Cost: One-time costs for CO2 transcritical systems in developed countries are estimated to be
17.5% more expensive than conventional HFC centralized DX systems (Australian Green Cooling Council,
2008; R744, 2012); it was assumed that conventional DX systems cost roughly $182,000 for a large
(60,000 sq. ft.) supermarket, equivalent to an incremental cost of nearly $32,000 per supermarket; these
capital costs are estimated to be 10% greater in developing countries.
• Annual O&M Costs: This analysis did not assume annual O&M costs.
• Annual Revenue: Annual savings are estimated at about $13,400 per supermarket, which results from
both refrigerant savings (due to avoided HFC refrigerant leaks) that total approximately $1,800 per
supermarket and energy savings (due to increased efficiency) that total approximately $11,600 per
supermarket. In developing countries, where electricity rates were assumed to be 66% higher, annual
savings were assumed to total more than $21,100.
C02 in Medium Retail Food Refrigeration Systems
This option eliminates the use of HFCs in medium retail food refrigeration systems and condensing units
through the use of CO2 as the primary refrigerant in a transcritical cycle. Although CO2 transcritical systems are
typically used in large retail food systems, CO2 provides an opportunity for reductions at smaller stores, such as
convenience stores or smaller grocery stores in city-center locations. CO2 systems are similar to traditional HFC
designs but must operate at high pressures to accommodate the low critical temperature of CO2 (GTZ Proklima,
2008). These systems are often referred to as mini-boosters. As a result, special controls and component
specifications must be incorporated into the system design, which often results in higher upfront costs
(Environmental Investigation Agency, 2012). These systems have already been launched in Japan and Europe, with
capacities of 2 to 10 kW, suitable for small supermarkets and convenience stores (R744, 2019b).
This abatement option applies the use of CO2 in refrigeration systems for medium retail food equipment as a
replacement for R-404A or R-507A. This option is applicable to newly manufactured refrigeration systems in all
countries.
5-112
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Capital Cost: One-time savings are estimated at approximately $110 per system associated with
incremental refrigerant costs. Additional costs may be realized associated with designing a system to
operate at a higher pressure, but they were not quantified in this analysis.
• Annual O&M Costs: No annual O&M costs were assumed.
• Annual Revenue: Annual savings are estimated at about $13 per system associated with incremental
refrigerant replacement costs. Annual energy savings are also likely to be associated with this option, but
they were not quantified in this analysis.
R-290 in New Ice Makers
This abatement option was applied to the use of R-290 as a replacement for HFC-134a or R-404A in ice
makers. Ice makers generate ice by freezing water as a stand-alone appliance or an industrial machine for making
ice on a large scale. Ice makers with R-290 are designed with separate specifications from HFC equipment because
of its flammability. Some ice maker manufacturers, for example, Hoshizaki in Japan, have pledged to transition to
R-290 refrigerant by 2018 (Hydrocarbons21, 2016).
This option is applicable to newly manufactured ice makers in all countries, except the EU, where it was
already assumed to penetrate the baseline market. For cost modeling purposes, this option was applied to a
manufacturing facility that produces 10,000 ice makers per year.
• Capital Cost: The one-time capital cost was estimated at approximately $325,000 per facility, resulting
from the additional structural and safety precautions applied to the design of equipment using flammable
refrigerants, including purchasing new charging equipment, testing and certifying equipment to the
appropriate safety standards, and converting manufacturing facilities to accommodate a flammable
refrigerant, if applicable (e.g., installing sensors and storage capacity).
• Annual O&M Costs: No annual O&M costs were assumed.
• Annual Revenue: Annual savings are estimated at about $9,600 per facility associated with incremental
refrigerant replacement costs.
R-452A in New Refrigerated Trucks and Trailers
This abatement option was applied to the use of R-452A as a replacement for refrigerated trucks/trailers and
intermodal containers. Refrigerated trucks/trailers and intermodal containers are designed to carry perishable
freight at specific cold temperatures. R-404A is the most used refrigerant in transport refrigeration today. R-452A,
an HFO blend with significantly lower environmental impacts, is considered a drop-in replacement for R-404A. R-
452A has been tested in refrigerated transport applications and showed no loss of performance and reliability in
terms of refrigeration capacity, pull-down, and fuel efficiency (Thermo King, 2014). This option is applicable to
newly manufactured systems in all countries.
• Capital Cost: R-452A is very similar to R-404A, such that existing equipment requires little to no change to
achieve the desired operational characteristics. Therefore, no one-time costs associated with equipment
modifications were assumed. A one-time cost equal to roughly $83 per system was assumed because of
the incremental refrigerant cost.
• Annual O&M Costs: Annual costs were assumed to be approximately $28 per system associated with
incremental refrigerant replacement costs.
• Annual Revenue: No savings were assumed.
NH3 or C02 in New IPR and Cold Storage Systems
This abatement option was assumed to be applicable to cold storage and industrial process refrigeration
systems. Although NH3 refrigeration systems are already common in refrigerated spaces over 200,000 sq. ft.,
additional penetration of NH3 systems is possible in facilities that are less than 200,000 sq. ft. but greater than
50,000 sq. ft. In addition, modern NH3 absorption refrigeration units are compact, lightweight, efficient,
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-113
-------
METHODOLOGY DOCUMENTATION
economical, and safe, which has made more applications possible. Improved technologies have also expanded the
technical feasibility of using C02 systems. C02 systems are being used in low-temperature refrigeration (-30°C to
-56°C), while ammonia/C02 systems can be used for higher temperature refrigeration (-35°C to -54°C). The lower
temperature for both systems is limited primarily by the -56°C triple point of C02 being used on the low side. The
choice between these systems is primarily due to outdoor temperatures; in colder climates, a C02 system is both
energy efficient and simpler, while in hotter climates, a cascade system may be needed to maintain energy
efficiency. In Europe and the United States, storage and production facilities have been built with ammonia/C02
cascade systems. These systems are estimated to be 2% to 20% more energy efficient than their HFC counterparts
(Gooseff and Horton, 2008).
• Capital Cost: For cost modeling purposes, this option was applied to a newly constructed I PR/cold storage
refrigeration system/facility. The incremental one-time cost is estimated at approximately $212,000 per
system in developed countries (Gooseff and Horton, 2008) and assumed to be 10% more in developing
countries. In addition, a one-time savings of $18,000 was assumed due to incremental refrigerant cost.
• Annual O&M Costs: This analysis did not assume annual O&M costs.
• Annual Revenue: The annual savings of approximately $50,200 per system was associated with lower
refrigerant replacement costs and reduced energy consumption of 11%; annual electricity cost savings
were assumed to be 66% greater for developing countries, resulting in annual savings of approximately
$83,000.
HCs in New Domestic Refrigeration Systems
HFC-134a may be replaced with HCs in household refrigerators. HCs, such as butane and propane, have very
low GWPs of 4.0 and 3.3, respectively. The main disadvantage of HCs is that they are flammable, but engineering
design changes and safety features in manufacturing plants have been successfully implemented to overcome
these challenges. In 2017, roughly one-third of the 100 million household refrigerators/freezers manufactured
annually around the world used hydrocarbon refrigerants (Hydrocarbon21, 2017). This option is applicable to
newly manufactured domestic refrigerators in developing countries because it was assumed to penetrate the
baseline market in the EU and other developed countries. For cost modeling purposes, this option was applied to a
manufacturing facility that produces 10,000 domestic refrigerators per year.
• Capital Cost: The one-time capital cost is estimated at approximately $325,000 per facility, resulting from
the additional structural and safety precautions applied to the design of equipment using flammable
refrigerants, including purchasing new charging equipment, testing and certifying equipment to the
appropriate safety standards, and converting manufacturing facilities to accommodate a flammable
refrigerant, if applicable (e.g., installing sensors and storage capacity). In addition, a one-time savings of
about $530,000 per facility is assumed associated with incremental refrigerant replacement costs and a
lower refrigerant charge size.
• Annual O&M Costs: No annual O&M costs were assumed.
• Annual Revenue: Annual savings are estimated at about $3,200 per facility associated with incremental
refrigerant replacement costs and a lower refrigerant charge size.
MCHX in New Unitary AC Equipment
This option explores the use of microchannel heat exchangers (MCHX) in unitary AC equipment using R-410A.
MCHXs are a modification of conventional heat exchangers, which transfer heat in AC and refrigeration systems
(e.g., for the rejection of heat from indoor cooled spaces to the outside ambient space). Because MCHXs transfer
heat through a series of small tubes instead of a single or multiple large-diameter tubes, systems using them
require between 35% and 40% less refrigerant to operate than those using conventional heat exchangers.
5-114
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Likewise, if average leak rates remain the same,68 the actual amount of refrigerant emitted would be less. In
addition, MCHX systems perform better and are more energy efficient than conventional systems. They also
require smaller components, which results in reduced quantities of metals and other materials required per unit,
although potential savings in material costs are not analyzed here. MCHXs are already used widely by multiple
manufacturers in the automotive industry and in certain models of screw and scroll chillers.
For cost modeling purposes, this option was applied to a newly manufactured small and large commercial
unitary AC system.
• Capital Cost: A one-time savings equal to roughly $27 was assumed because of the smaller refrigerant
charge.
• Annual O&M Costs: This analysis did not assume annual O&M costs.
• Annual Revenue: The annual savings associated with avoided refrigerant losses due to a smaller
refrigerant charge was estimated at approximately $2.30 per system.
R-32 in New Commercial Unitary AC Equipment
In this option, R-32, a mildly flammable refrigerant with a GWP of 650, and R-452B, a blend of HFC-125, HFC-
32, and HFO-1234yf with a GWP of 698, are used in new unitary AC equipment to replace R-410A, which has a
GWP of 1,725. R-32 is category A2L in ANSI/ASHRAE Standard 34.
R-32 performs with a reduced charge volume ratio of 66% compared with R-410A (Xu et al., 2012). It is also
reportedly 2% to 3% more energy efficient than R-410A (Pham and Sachs, 2010). In addition, the equipment used
has the potential to be downsized by up to 15%, which can decrease one-time costs by reducing the amount of
materials used. R-32 AC products are already available in Japan and India (Daikin, 2012; Stanga, 2012).
Manufacturers in Algeria, China, Thailand, and Indonesia also plan to transition to R-32 AC systems (Stanga, 2012).
For cost modeling purposes, this option was applied to a newly manufactured commercial unitary AC system
(e.g., small commercial and large commercial unitary AC).
• Capital Cost: The option was conservatively assumed to result in a one-time savings of approximately $30
per system because of the reduced quantity of refrigerant required and lower cost of the alternative
refrigerant. Additional savings may be realized through reduced material costs; however, costs also may
be associated with designing a system to use a mildly flammable refrigerant.
• Annual O&M Costs: This analysis did not assume annual O&M costs.
• Annual Revenue: The annual savings associated with avoided refrigerant losses due to a smaller
refrigerant charge and incremental cost of replacement refrigerant was estimated at approximately $2.30
per system. Annual energy savings are also likely to be associated with this option but were not quantified
in this analysis.
R-32 with MCHX in New Commercial Unitary AC Equipment
Similar to the options described above, this option explores the use of MCHX in commercial unitary AC
equipment but with R-32 refrigerant in place of R-410A. The use of MCHX results in a refrigerant charge reduction
of between 35% and 40% compared with conventional heat exchangers, while the use of R-32 refrigerant allows a
further charge size reduction of 66% compared with R-410A. For cost modeling purposes, this option was applied
to all newly manufactured commercial unitary AC systems.
68 For example, if average leak rates are dominated by failures or service errors that lead to a catastrophic (100%) loss and the
MCHX system has the same reliability, then average leak rates would be the same.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-115
-------
METHODOLOGY DOCUMENTATION
• Capital Cost: The option was conservatively assumed to result in a one-time savings of approximately $46
per system because of the reduced quantity of refrigerant required and lower cost of the alternative
refrigerant.
• Annual O&M Costs: No annual O&M costs were assumed.
• Annual Revenue: Annual savings are associated with reduced charge size and incremental refrigerant
replacement costs, estimated at approximately $4 per system. Annual energy savings are also likely to be
associated with this option but were not quantified in this analysis.
R-452B with MCHX in New Residential Unitary AC Equipment
Similar to the options described above, this option explores the use of MCHX in residential unitary AC
equipment with R-452B refrigerant in place of R-410A. R-452B, a blend of HFC-125, HFC-32, and HFO-1234yf, has a
GWP of 698 and is also category A2L in ANSI/ASHRAE Standard 34. R-452B operates at the same pressure as R-
410A, even in hot climates, so no redesign is necessary; however, safety design changes may be needed (Alliance
for Responsible Atmospheric Policy, 2018). R-452B performs with a reduced charge volume ratio of 90% compared
with R-410A and offers slightly increased efficiency (Shen et al., 2016). The use of MCHX results in a refrigerant
charge reduction of between 35% and 40% compared with conventional heat exchangers, while the use of R-452B
refrigerant allows a further charge size reduction of 10% compared with R-410A.
For cost modeling purposes, this option was applied to all newly manufactured residential unitary AC systems
in all countries except the EU.
• Capital Cost: The option was conservatively assumed to result in a one-time cost of approximately $17
per system despite the reduced quantity of refrigerant required because of the higher cost of the
alternative refrigerants. Additional costs may be realized associated with designing a system to use a
mildly flammable refrigerant.
• Annual O&M Costs: Annual costs are associated with incremental refrigerant replacement costs,
estimated at approximately $1.40 per system.
• Annual Revenue: Annual energy savings are also likely to be associated with this option but were not
quantified in this analysis.
R-290in New Residential Unitary AC Equipment
This abatement option applies the use of R-290 in unitary AC equipment systems as a replacement for R-410A.
Although its use is not yet widespread, Petra, a major manufacturer of HVAC equipment in the Middle East, has
begun testing R-290 in large commercial HVAC systems. Petra expects that the testing will conclude in 2020
(Hydrocarbons21, 2019). In addition, both Godrej in India and Gree in China are producing units with HC
refrigerants (Godrej, 2012; Gree, 2012).
This option is applicable to newly manufactured domestic refrigerators in developed and developing countries
except the EU. For cost modeling purposes, this option was applied to a manufacturing facility that produces
10,000 residential unitary AC systems per year.
• Capital Cost: The one-time capital cost was estimated at approximately $325,000 per facility, resulting
from the additional structural and safety precautions applied to the design of equipment using flammable
refrigerants, including purchasing new charging equipment, testing and certifying equipment to the
appropriate safety standards, and converting manufacturing facilities to accommodate a flammable
refrigerant, if applicable (e.g., installing sensors and storage capacity). In addition, a one-time savings of
about $234,000 per facility was assumed associated with incremental refrigerant replacement costs and a
lower refrigerant charge size.
• Annual O&M Costs: No annual O&M costs were assumed.
5-116
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Annual Revenue: Annual savings are estimated at about $21,000 per facility associated with incremental
refrigerant replacement costs.
R-32/R-452B in New Heat Pumps
Similar to the options described above, this option explores the use of R-32 and R-452B in heat pumps in place
of R-410A. For cost modeling purposes, this option was applied to all newly manufactured heat pumps in all
countries.
• Capital Cost: The option was conservatively assumed to result in a one-time cost of approximately $1.80
per system despite the reduced quantity of refrigerant required because of the higher cost of the
alternative refrigerants. Additional costs may be realized associated with designing a system to use a
mildly flammable refrigerant.
• Annual O&M Costs: Annual costs were associated with incremental refrigerant replacement costs,
estimated at approximately $0.44 per system.
• Annual Revenue: Annual energy savings are also likely to be associated with this option but were not
quantified in this analysis.
R-290 in New Window AC and Dehumidifiers
R-410A is widely used in window AC units and dehumidifiers, brought about by regulations phasing out HCFC-
22, which was previously used. By replacing R-410A with HCs such as propane (R-290), which was assumed to have
a negligible GWP, significant emissions can be avoided. Two Chinese AC manufacturers have already
commercialized room AC units using R-290. The R-290 AC designs achieve lower refrigerant charge sizes than is
currently required by international standard (IEC 60335-2-240) and include additional safety features, such as a
special compressor design and refrigerant leak alarm system. Mass production of the R-290 units started in 2009
for initial sale in Europe and China (GTZ Proklima, 2009). Additionally, the first R-290-based room AC units were
launched in India in 2013. Sales from launch until 2014 total over 100,000 units (NRDC, 2014).
This option is applicable to newly manufactured domestic refrigerators in developed and developing countries
other than the EU. For cost modeling purposes, this option was applied to a manufacturing facility that produces
10,000 window units and/or dehumidifiers per year.
• Capital Cost: This option was conservatively assumed to have one-time capital costs of approximately
$325,000 per system, resulting from the additional structural and safety precautions applied to the design
of equipment using flammable refrigerants, including purchasing new charging equipment, testing and
certifying equipment to the appropriate safety standards, and converting manufacturing facilities to
accommodate a flammable refrigerant, if applicable (e.g., installing sensors and storage capacity). There is
indication that R-290 AC units can be produced more cheaply than R-410A units as a result of the better
heat transfer properties and lower pressure drop of R-290, which allows for the use of narrower tubes in
the condenser and evaporator (GTZ Proklima, 2009). In addition, a one-time savings of about $27,000 per
facility was assumed associated with incremental refrigerant replacement costs and a lower refrigerant
charge size.
• Annual O&M Costs: No annual O&M costs were assumed.
• Annual Revenue: Annual savings are estimated at about $160 per facility associated with incremental
refrigerant replacement costs.
R-32 in New Window AC and Dehumidifiers
Similar to the options described above, this option explores the replacement R-410A in new window AC and
dehumidifiers with R-32 refrigerant. R-32 has a lower refrigerant flow rate and a higher discharge temperature
than R-410A, so some redesign is necessary. R-32 is mildly flammable and requires some safety design changes
(Alliance for Responsible Atmospheric Policy, 2018). Advantages of R-32 include higher operating efficiencies,
meaning units that use R-32 refrigerant consume less power than similar units with R-410A. Also, units using R-32
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-117
-------
METHODOLOGY DOCUMENTATION
generally require a lower charge size, reducing the total amount of refrigerant needed. In 2016, it was estimated
that more than 10 million R-32 ACs had been sold since their introduction (Proud Green Building, 2016).
This abatement option applies the use of R-32 in new window AC and dehumidifier systems. This option is
applicable to newly manufactured systems in all countries except the EU.
• Capital Cost: The option was conservatively assumed to result in a one-time savings of approximately
$1.80 per system because of the reduced quantity of refrigerant required and lower cost of the
alternative refrigerant. Additional savings may be realized through reduced material costs; however, costs
also may be associated with designing a system to use a mildly flammable refrigerant.
• Annual O&M Costs: This analysis did not assume annual O&M costs.
• Annual Revenue: The annual savings associated with avoided refrigerant losses due to a smaller
refrigerant charge and incremental cost of replacement refrigerant was estimated at approximately $0.01
per system. Annual energy savings are also likely to be associated with this option but were not quantified
in this analysis.
R-452B in New Positive Displacement Chillers
This abatement option applies the use of R-452B as a replacement for R-410A or R-407C in positive
displacement chiller systems. R-452B is a bend of HFC-32, HFC-125, and HFO-1234yf and offers similar efficiency to
R-410A units. Several equipment manufacturers offer R-452B in positive displacement chillers (Cooling Post, 2019).
• Capital Cost: R-452B is considered to be a drop-in replacement for R-410A; therefore, this analysis
assumed no one-time equipment modification costs for transitioning to R-452B. A one-time cost of
approximately $4,200 per system associated with incremental refrigerant cost was assumed.
• Annual O&M Costs: This analysis assumed an annual cost of $250 per system associated with incremental
refrigerant replacement cost.
• Annual Revenue: This analysis did not assume a cost savings.
R-450A/R-513A in New Centrifugal Chillers
This abatement option applies the use of R-450A and R-513A in centrifugal chillers. R-450A and R-513A are
blends of HFC-134a and HFO-1234yf. Equipment manufacturers in Asia offer both R-450A and R-513A centrifugal
chillers (GlobeNewswire, 2018).
• Capital Cost: R-450A and R-513A are considered drop-in replacements for HFC-134a (Chemours, 2015);
therefore, this analysis assumed no one-time equipment modification costs for transitioning to R-450A
and R-513A. A one-time cost of approximately $13,100 per system associated with incremental
refrigerant costs was assumed.
• Annual O&M Costs: This analysis assumed an annual cost of $780 per system associated with incremental
refrigerant replacement costs.
• Annual Revenue: This analysis did not assume a cost savings.
HCFO-1233zd(E) in New Centrifugal Chillers
This abatement option applies the use of hydrochlorofluoroolefin (HCFO)-1233zd(E) in low-pressure
centrifugal chillers in place of HFC-245fa (or centrifugal chillers that historically used HCFC-123). Chiller
manufacturers have already begun producing chillers with HCFO-1233zd(E). The abatement option that replaces
HFC-245fa with HCFO-1233zd(E) is applicable to newly manufactured centrifugal chillers in developing countries
and the EU.
• Capital Cost: Capital costs of $51,000 were assumed for production line conversion to HCFO-1233zd(E),
including the refrigerant system, charging machines, and safety system (UNEP, 2012). A one-time cost of
approximately $2,900 per system associated with incremental refrigerant cost was assumed.
5-118
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Annual O&M Costs: This analysis assumed an annual cost of $170 per system associated with incremental
refrigerant replacement costs.
• Annual Revenue: This analysis did not assume a cost savings.
Leak Repair for Existing Large Equipment
This abatement option was assumed to be applicable to large retail food, cold storage and industrial
refrigeration, and positive displacement chiller systems. Some level of leak repair activity is already practiced in the
baseline, but this option explores additional efforts to repair leaks.
• Capital Cost: For cost modeling purposes, this option was applied to large supermarkets requiring
significant small repairs (e.g., maintenance of the purge system or replacement of a gasket or O-ring). A
one-time cost of approximately $1,870 was estimated per supermarket for parts and labor needed to
perform the repair in developed countries (EPA, 1998); in developing countries, this cost was estimated to
be 10% greater.
• Annual O&M Costs: This analysis did not assume annual O&M costs.
• Annual Revenue: The annual savings associated with avoided refrigerant replacement was estimated at
$1,220 per supermarket.
Refrigerant Recovery at Servicing for Existing Small Equipment
Similar to disposal recovery, this option assumed more widespread and thorough refrigerant recovery
practices while servicing HFC refrigeration/AC systems. Because it was assumed that significant refrigerant is
already recovered during servicing of large equipment, this abatement option was only applied to MVAC and small
unitary AC systems.
• Capital Cost: For cost modeling purposes, this option was applied to a U.S. auto servicing facility assumed
to perform MVAC servicing jobs using a recovery/recycling (recharge) device designed to meet the SAE
J2788 standard. The incremental one-time cost was estimated at approximately $4,050 per servicing
facility for the purchase of a refrigerant recovery device in developed countries (ICF, 2008); this cost was
estimated to be 10% greater in developing countries.
• Annual O&M Costs: The annual cost was estimated at roughly $870 per auto servicing facility in
developed countries for technician labor time and the purchase of new filters for the recovery device (ICF,
2008); in developing countries, technician labor costs were assumed to be one-fifth the cost of that in
developed countries; therefore, the annual cost was assumed to be nearly $194.
• Annual Revenue: The annual savings was estimated at roughly $350 per auto servicing facility based on
the value of the recovered refrigerant for reclamation/reuse.
Refrigerant Recovery at Disposal for All Existing Equipment Types
Some level of refrigerant recovery at equipment disposal already occurs in the baseline of developed and
developing countries, because it is illegal to vent HFCs when equipment is discarded in the United States and
elsewhere. However, this option explores more widespread, thorough efforts to recover refrigerant at disposal
across all equipment types. The approach involves using a refrigerant recovery device that transfers refrigerant
into an external storage container before disposal of the equipment. Once the recovery process is complete, the
refrigerant contained in the storage container may be cleaned by using recycling devices, sent to a reclamation
facility to be purified,69 or destroyed using approved technologies (e.g., incineration).
• Capital Cost: For cost modeling purposes, this option was applied to an auto dismantling facility assumed
to use a single refrigerant recovery device that meets SAE J2788 standards to perform MVAC recovery
69 Recycling cleans and reclamation purifies recovered refrigerant; reclamation is more thorough and involves repeated
precision distillation, filtering, and contaminant removal. Recycling is used for on-site servicing of MVACs and other equipment,
whereas reclamation requires sending the refrigerant off-site to a reclaimer.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-119
-------
METHODOLOGY DOCUMENTATION
jobs. The incremental one-time cost was estimated at approximately $2,025 per facility for the purchase
of a refrigerant recovery device in developed countries (ICF, 2008); this cost was estimated to be 10%
greater in developing countries.
• Annual O&M Costs: Annual costs are estimated at roughly $1,100 per auto dismantling facility for
technician labor time and the purchase of new filters for the recovery device (ICF, 2008). In developing
countries, technician labor costs were assumed to be one-fifth the cost of that in developed countries;
therefore, annual costs were assumed to be about $220.
• Annual Revenue: The annual savings was estimated at about $440 per auto dismantling facility based on
the value of the recovered refrigerant for reclamation/reuse.
Technical and Economic Characteristics Summary
Table 5-41 presents the technical characteristics of each mitigation option considered in the analysis.
Table 5-41: Summary of Technical Characteristics of Each Mitigation Option
Technical
Market
Technical
Applicability
Penetration
Reduction
Effectiveness
Facility/Abatement Option
(2030)
Rate (2030)a
Efficiency
(2030)b
MVACs—Developing
HFO-1234yf
46%
100%
100%
46%
Buses and Trains—U.S./Other Developed
R-513A
21%
55%
57%
7%
Buses and Trains—EU
R-513A
21%
55%
57%
7%
Buses and Trains—Developing
R-513A
6%
9%
57%
0%
Large Retail Food—U.S./Other Developed
CO2 transcritical systems
59%
33%
100%
19%
Large Retail Food—EU
CO2 transcritical systems
59%
33%
100%
19%
Large Retail Food—Developing
DX R-407A/R-407F
7%
34%
50%
1%
HFC secondary loop and/or cascade systems
11%
33%
50%
2%
CO2 transcritical systems
4%
33%
100%
1%
Medium Retail Food—U.S./Other Developed
CO2 transcritical systems
44%
33%
100%
14%
Medium Retail Food—EU
CO2 transcritical systems
28%
33%
100%
9%
Medium Retail Food—Developing
CO2 transcritical systems
24%
33%
100%
8%
Ice Makers—U.S./Other Developed
HCs
20%
50%
100%
10%
Ice Makers—Developing
HCs
9%
19%
100%
2%
(continued)
5-120
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-41: Summary of Technical Characteristics of Each Mitigation Option (continued)
Technical
Market
Technical
Applicability
Penetration
Reduction
Effectiveness
Facility/Abatement Option
(2030)
Rate (2030)a
Efficiency
(2030)b
Refrigerated Trucks/Trailers and Intermodal
Containers—U.S./Other Developed
R-452A
23%
50%
45%
5%
Refrigerated Trucks/Trailers and Intermodal
Containers—Developing
R-452A
0%
0%
45%
0%
IPR & Cold Storage—U.S./Other Developed
NHsorCOz
44%
100%
100%
44%
IPR & Cold Storage—EU
NHsorCOz
44%
100%
100%
44%
IPR & Cold Storage—Developing
NHsorCOz
19%
33%
100%
1%
Refrigerated Appliances—Developing
HCs
7%
100%
100%
7%
Commercial Unitary AC—U.S./Other
Developed
MCHX
0%
0%
38%
11%C
R-32
25%
50%
68%
9%
R-32 with MCHX
5%
50%
87%
2%
Commercial Unitary AC—EU
MCHX
0%
0%
38%
11%C
R-32
25%
50%
68%
9%
R-32 with MCHX
5%
50%
87%
2%
Commercial Unitary AC—Developing
MCHX
18%
86%
38%
6%
R-32
0%
0%
68%
0%
R-32 with MCHX
0%
0%
87%
0%
Residential Unitary AC—U.S./Other Developed
R-32/R-454B with MCHX
44%
100%
81%
13%
Residential Unitary AC—Developing
R-32/R-454B with MCHX
0%
0%
81%
0%
R-290
0%
0%
100%
0%
Heat Pumps—U.S./Other Developed
R-32/R-454B
10%
100%
67%
7%
Heat Pumps—EU
R-32/R-454B
15%
100%
67%
4%
Heat Pumps—Developing
R-32/R-454B
0%
0%
67%
0%
(continued)
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis 5-121
-------
METHODOLOGY DOCUMENTATION
Table 5-41: Summary of Technical Characteristics of Each Mitigation Option (continued)
Technical
Market
Technical
Applicability
Penetration
Reduction
Effectiveness
Facility/Abatement Option
(2030)
Rate (2030)a
Efficiency
(2030)b
Window AC Units and Dehumidifiers—
U.S./Other Developed
R-290
1%
50%
100%
1%
R-32
7%
50%
68%
2%
Window AC Units and Dehumidifiers—
Developing
R-290
4%
27%
100%
1%
R-32
4%
27%
68%
1%
PD Chillers—EU
R-454B
28%
100%
64%
18%
PD Chillers—Developing
R-454B
9%
100%
64%
6%
Centrifugal Chillers—EU
R-513A
27%
100%
57%
15%
Centrifugal Chillers—Developing
R-513A
12%
60%
57%
4%
HCFO-1233zd(E)
2%
60%
99%
1%
Cross-Cutting Practice Options—U.S./Other
Developed
Leak repair (large equipment)
51%
100%
40%
20%
Refrigerant recovery at servicing (small
equipment)
10%
100%
95%
10%
Refrigerant recovery at disposal
17%
100%
85%
14%
Cross-Cutting Practice Options—EU
Leak repair (large equipment)
61%
100%
40%
25%
Refrigerant recovery at servicing (small
equipment)
12%
100%
95%
11%
Refrigerant recovery at disposal
17%
100%
85%
14%
Cross-Cutting Practice Options—Developing
Leak repair (large equipment)
64%
100%
40%
26%
Refrigerant recovery at servicing (small
equipment)
11%
100%
95%
10%
Refrigerant recovery at disposal
23%
100%
85%
20%
a Market penetration assumptions for this analysis vary over time, and the technical effectiveness values are based on the
cumulative market penetration rates assumed until that point.
b Technical effectiveness figures represent the percentage of baseline emissions from the relevant facility type that can be
abated in 2030; figures do not account for indirect GHG impacts (i.e., increases or decreases in electricity or fuel consumption),
which were accounted for in the cost analysis.
c This option is no longer assumed to penetrate the market in new equipment as of 2030; however, emission reductions are still
experienced from existing equipment.
5-122
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-42 presents the economic characteristics of each mitigation option considered in the analysis.
Table 5-42: Summary of Economic Characteristics of Each Mitigation Option
Project
Annual
Annual
Abatement
Lifetime
Capital Cost
Revenue
O&M Costs
Amount
Abatement Option/Facility Type
(years)
(2015 USD)
(2015 USD)
(2015 USD)
(tC02e)a
HF0-1234yf
MVAC—Developing, new
16
$111
—
$3
0.2
R-513A
Buses and trains—U.S./other developed,
new
12
$47
—
$9
0.7
Buses and trains—EU, new
12
$47
—
$9
0.7
Buses and trains—Developing, new
12
$47
—
$9
0.7
R-407A/R-407F
Large retail food —Developing, new
18
$0
$17,206
—
695.4
HFC Secondary Loop and/or Cascade
Systems
Large retail food —Developing, new
18
$41,482
$7,311
—
429.4
CO2 Transcritical Systems
Large retail food—U.S./other developed,
new
18
$19,610
$13,445
—
1,096
Large retail food —EU, new
18
$19,610
$13,445
—
1,096
Large retail food —Developing, new
18
$22,795
$21,107
—
1,096
CO2
Medium retail food—U.S./other developed,
existing
20
-$108
$13
—
8.1
Medium retail food—EU, existing
20
-$108
$13
—
8.1
Medium retail food—Developing, existing
20
-$108
$13
—
8.1
HCs
Ice makers—U.S./other developed, new
8
$107,125
$9,587
—
14,213
Ice makers—Developing, new
8
$107,125
$9,587
—
14,213
R-452A
Refrigerated trucks/trailers and intermodal
containers—U.S./other developed, new
12
m
00
-00-
—
$28
4.6
Refrigerated trucks/trailers and intermodal
containers—Developing, new
12
m
00
-00-
—
$28
4.6
NH3 or C02
IPR/cold storage—U.S./other developed,
new
25
$193,000
$50,180
—
711.6
IPR/cold storage—EU, new
25
$193,000
$50,180
711.6
IPR/cold storage—Developing, new
25
$214,100
$82,705
711.6
(continued)
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-123
-------
METHODOLOGY DOCUMENTATION
Table 5-42: Summary of Economic Characteristics of Each Mitigation Option (continued)
Project
Annual
Annual
Abatement
Lifetime
Capital Cost
Revenue
O&M Costs
Amount
Abatement Option/Facility Type
(years)
(2015 USD)
(2015 USD)
(2015 USD)
(tC02e)a
HCs
Domestic refrigerators—Developing, new
14
-$201,075
$3,156
—
8,798
MCHX
Commercial Unitary AC—U.S./other
developed, new
15
-$27
$2
—
1.7
Commercial Unitary AC—EU, new
15
-$27
$2
—
1.7
Commercial Unitary AC—Developing, new
15
-$27
$2
—
1.7
R-32
Commercial Unitary AC—U.S./other
developed, new
15
-$30
$3
—
2.1
Commercial Unitary AC—EU, new
15
-$30
$3
—
2.1
Commercial Unitary AC—Developing, new
15
-$30
$3
—
2.1
R-32 with MCHX
Commercial unitary AC—U.S./other
developed, new
15
-$46
$4
—
2.4
Commercial unitary AC—EU, new
15
-$46
$4
2.4
Commercial unitary AC—Developing, new
15
-$46
$4
2.4
R-452B with MCHX
Residential unitary AC—U.S./other
developed, new
15
$16
—
51
1.2
Residential unitary AC—Developing, new
15
$16
—
$1
1.2
R-290
Residential unitary AC—Developing, new
15
$91,250
$21,038
—
5,915
R-32/R-452B
Heat pumps—U.S./other developed, new
15
$2
—
—
0.3
Heat pumps—EU, new
15
$2
—
0.3
Heat pumps—Developing, new
15
$2
—
0.3
HCs
Window units/dehumidifiers—U.S./other
developed, new
12
$298,325
51(>()
—
1,073
Window units/dehumidifiers—Developing,
new
12
$298,325
$160
—
1,073
R-32
Window units/dehumidifiers—U.S./other
developed, new
12
-$2
0.1
Window units/dehumidifiers—Developing,
new
12
-$2
0.1
(continued)
5-124
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-42: Summary of Economic Characteristics of Each Mitigation Option (continued)
Abatement Option/Facility Type
Project
Lifetime
(years)
Capital Cost
(2015 USD)
Annual
Revenue
(2015 USD)
Annual
O&M Costs
(2015 USD)
Abatement
Amount
(tC02e)a
R-452B
Positive displacement chiller—EU, new
20
$4,167
—
$250
51.1
Positive displacement chiller—Developing,
new
20
$4,167
—
$250
51.1
R-450A/R-513A
Centrifugal chiller—EU, new
27
$13,057
—
$783
73.2
Centrifugal chiller—Developing, new
27
$13,057
—
$783
73.2
HCF0-1233zd(E)
Centrifugal chiller—Developing, new
27
$53,880
—
$173
71.7
Leak Repair
Large retail food—U.S./Other developed,
existing
5
$1,870
$1,224
—
533.4
Large retail food —EU, existing
5
$1,870
$1,224
533.4
Large retail Food—Developing, existing
5
$2,057
$1,224
533.4
Recovery at Servicing
Auto servicing station—U.S./other
developed
7
$4,050
$351
58/0
62.8
Auto servicing station—EU
7
$4,050
$351
$870
62.8
Auto servicing station —Developing
7
$4,455
$351
$174
62.8
Recovery at Disposal
Auto disposal yard—U.S./other developed
7
$2,026
$445
$1,084
79.6
Auto disposal yard—EU
7
$2,026
$445
$1,084
79.6
Auto disposal yard—Developing
7
$2,229
$445
$217
79.6
a Emission reductions shown include only reductions associated with HFCs; they do not include indirect (C02) emissions
associated with differences in energy consumption.
Sector-Level Trends/Considerations
The development of alternative refrigerants and technologies is quickly evolving in this sector, with
efficiencies increasing and costs decreasing as research and market share expand. Thus, the costs and reduction
efficiencies of the alternatives reviewed in this analysis are subject to change and likely conservative. Moreover,
new options not quantified in this analysis are entering the market and will continue to do so; additional options,
such as C02 in transport refrigeration and low-GWP refrigerants for comfort cooling chillers, could be
quantitatively considered in future analyses.
The costs for the options explored in this analysis are highly variable, depending on the types of systems
reviewed. Estimates of the amount of refrigerant recoverable from equipment at service and disposal are highly
uncertain and highly variable based on the type of equipment. Recovery from large equipment is generally more
cost-effective than for small equipment, because the amount of refrigerant recoverable is greater and the relative
amount of technician time needed to perform the recovery is smaller. Similarly, because leak repair can be
performed on many different equipment types and can involve many different activities/tools, determining an
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-125
-------
METHODOLOGY DOCUMENTATION
average cost of such repairs or the average emission reduction associated with them is difficult. This analysis,
therefore, relies on broad assumptions available in the published literature, which may not reflect specific or even
average values for the leak repair activities modeled.
Finally, it was assumed that numerous abatement options result in increased or decreased energy
consumption (e.g., R-407A/R-407F in large retail food refrigeration systems, C02 transcritical large retail food
refrigeration systems, NH3 or C02 in new I PR and cold storage systems). While the cost associated with the
increase or decrease in energy consumption, which would vary widely based on region as well as particular
application, was quantified as part of this analysis, the increase or decrease in C02 emissions associated with this
energy use was not quantified. To accurately capture net emission reductions of these abatement options,
emissions associated with the increase or decrease in energy use should also be calculated.
References
ACHR News. December 3, 2012. Retrofitting with HCFs. Air Conditioning, Heating, Refrigeration News. Available
online at http://digital.bnpmedia.com/publication/?i=136503 eP
ACHR News. September 6, 2010. C02 in supermarket refrigeration. Air Conditioning, Heating, Refrigeration News.
Available online at http://www.achrnews.com/articles/co2-in-supermarket-refrigeration J
Alliance for Responsible Atmospheric Policy. 2018. Residential Air Conditioning. Available online at
http://www.alliancepolicv.org/resources/technologv-informationcf
Ashford, P. 2004. Peer review comments on U.S. EPA Draft Report, Draft Analysis of International Costs of Abating
HFC Emissions from Foams. Caleb Management Services Ltd. March 3, 2004.
Australian Green Cooling Council. 2008. Transcritical C02 (R744) Supermarket Refrigeration in Australia. Available
online at http://ec.europa.eu/clima/events/0007/australian green cooling en.pdf cf
Automotive Engineering International. September 2008. Consensus Building on Refrigerant Type. Automotive
Engineering International.
Chemours. 2015. Opteon XP10 Refrigerant. Available online at
https://www.chemours.com/Refrigerants/en US/products/Opteon/Stationarv Refrigeration/assets/downloa
ds/Opteon XP10 prodinfo.pdf cf
Cooling Post. 2019. Water-cooled chillers with R452B option. Available online at
https://www.coolingpost.com/products/water-cooled-chillers-with-r452b-option/cf
Daikin. September 27, 2012. News release: World's first commercialization of air conditioning equipment using
next-generation refrigerant HFC32. Available online at
http://www.daikin.com/press/2012/120927/index.html cf
DelVentura, R. et al., September 2007. BOHN the cold standard: Secondary loop systems for the supermarket
industry. White Paper.
Environmental Investigation Agency. 2012. HFC-Free Technologies Are Available in the US Market for the
Supermarket-Retail Refrigeration Sector. Available online at http://eia-global.org/PDF/USSUPERMARKET
REPORT.pdf cf
Environmental Leader: Environmental and Energy Management News. December 1, 2010. Carrier unveils natural
refrigerated transport solution. Available online at http://www.environmentalleader.com/2010/12/Ql/carrier-
unveils-natural-refrigerated-transport-solution/cf
GlobeNewswire. 2018. Carrier AquaEdge® 19XR, 23XRV and AquaForce® 30KA, 30XA, 30XB, 30XW and 30HXC
Chillers are Available with Lower Global Warming Refrigerant Solutions. Available online at
https://www.globenewswire.com/news-release/2018/02/13/1339517/0/en/Carrier-AouaEdge-19XR-23XRV-
and-AquaForce-30KA-30XA-30XB-30XW-and-30HXC-Chillers-are-Available-With-Lower-Global-Warming-
Refrigerant-Solutions.html cf
5-126
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Godrej. 2012. "HC Technology for Room Air Conditioners." Presented by Dilip Rajadhyaksha at Advancing Ozone
and Climate Protection Technologies: Next Steps, Bangkok, Thailand. July 21-22, 2012.
Gooseff, A., and J. Horton. March 16-19, 2008. Ammonia as the sustainable refrigerant: an ammonia-halocarbon
comparison. Presented at 30th Annual Meeting International Institute of Ammonia Refrigeration, Colorado
Springs, CO.
Gree Electronic Appliances. 2012. "Application of HC Refrigerant in Room Air Conditioners." Presented at
Advancing Ozone and Climate Protection Technologies: Next Steps, Bangkok, Thailand. July 21-22, 2012.
GTZ Proklima. July 2008. Natural Refrigerants: Sustainable Ozone- and Climate-Friendly Alternatives to HCFCs.
Deutsche Gesellschaft fur Technische Zusammenarbeit (GTZ) GmbH. Eschborn, Germany. Available online at
http://epa.gov/greenchill/downloads/en-gtz-proklima-natural-refrigerants.pdf
GTZ Proklima. May 10, 2009. Gree Electric R290 air conditioner. Presented at the Joint West Asia and South Asia
Network Meeting. Available online at http://www.hvdrocarbons21.com/files/news/gree-ac-unit-
hvdrocarbons.pdf ef
Heath, R. and J. Armer. June 12, 2012. SuperValu's development of next generation refrigeration systems:
application of ammonia, C02 and propane. Presented at ATMOsphere America the Business Case: Natural
Refrigerants.
Hill PHOENIX. 2012. Advances in C02 supermarket refrigeration in North America. Presented by S. Martin, Hill
PHOENIX, at the ATMOsphere 2012 conference, Washington, DC. Available online at
http://www.atmo. org/media. presentation. php?id=139 ef
Hinde, D., S. Zha, and L. Lan. February 2009. Carbon dioxide in North American supermarkets. ASHRAE Journal, 51.
Honeywell. 2014. Genetron Performax® LT (R407F): first choice solution for commercial refrigeration. Available
online at http://www.honevwell-refrigerants.com/europe/wp-content/uploads/2014/Q2/Performax-LT-4pp-
130204-low-res.pdf J
Hydrocarbons21. December 3, 2017. Are home fridges in U.S. turning toward hydrocarbons? Available online at
http://hvdrocarbons21.com/articles/8018/are home fridges in u s turning toward hydrocarbons cf
Hydrocarbons21. February 2019. Petra converting HVAC production lines to R290. Available online at
http://hvdrocarbons21.com/articles/8839/petra converting hvac production lines to r290ef
Hydrocarbons21. March 18, 2016. Hoshizaki ice makers to go all natural by 2018. Available online at
http://hvdrocarbons21.com/articles/6925/hoshizaki ice makers to go all natural by 2018 eP
ICF International. 2008. Regulatory Impact Analysis (RIA) of Current Section 609 Requirements and Potential
Updates: Draft Report. Washington, DC: ICF International.
Intergovernmental Panel on Climate Change. 2005.1PCC/TEAP Special Report on Safeguarding the Ozone Layer and
the Global Climate System: Issues Related to Hydrofluorocarbons and Perfluorocarbons. In B. Metz, L. Kuijpers,
S. Solomon, S.O. Andersen, O. Davidson, J. Pons, et al. (Eds.). Prepared by Working Groups I and III of the IPCC,
and the Technical and Economic Assessment Panel. New York, NY: Cambridge University Press.
Koban, M. April 2009. HFO-1234yf low GWP refrigerant LCCP analysis. Presented at SAE 2009 World Congress.
Linde Gas. 2014. R407A: lower global warming potential replacement for R404A. Available online at
https://www.linde-gas.com/en/images/R407A-Refrigerant-Brochure tcml7-114449.pdf cP
Malvicino, C. Centro Ricerche Fiat 2008. The 4 Fiat pandas experiment and other considerations on refrigerants.
Presented at the SAE Alternative Refrigerant Systems Symposium, Phoenix, AZ.
National Resource Defense Council. November 2014. Hydrocarbon room air conditioners: Energy-efficient,
commercially viable and safe. Available online at https://www.nrdc.org/experts/avinash-kar/hvdrocarbon-
room-air-conditioners-energv-efficient-commerciallv-viable-and-safeeP
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-127
-------
METHODOLOGY DOCUMENTATION
Oko-Recherche, Oko-lnstitute, HEAT International, Danish Technological Institute, Re-phridge, Karlsruhe University
of Applied Sciences, Estonia Environmental Research Centre, and Ammonia Partnership. September 2011.
Preparatory Study for a Review of Regulation (EC) No 842/2006 on Certain Fluorinated Greenhouse Gases.
Prepared for the European Commission.
Pham, H. and H. Sachs. 2010. Next Generation Refrigerants: Standards and Climate Policy Implications of
Engineering Constraints. ACEEE Summer Study on Energy Efficiency in Buildings. Available online at
http://www.aceee.org/files/proceedings/2010/data/papers/1933.pdfc?
Proud Green Building. May 2016. New window air conditioning units using more sustainable refrigerant. Available
online at https://www.proudgreenbuilding.com/news/new-window-air-conditioning-units-using-more-
sustainable-refrigerant/ J
R744. December 18, 2012. Cascade CO2 could be winning option in emerging economies—CCAC Forum, Part 118.
Available online at http://www.r744.com/news/view/3755 C?
R744. January 24, 2019b. Lennox to launch CO2 condensing unit. Available online at
http://www.r744.com/articles/8810/lennox to launch co2 condensing unit J
R744. March 2019a. GreenChill reports growth in CO2 systems at U.S. supermarkets. Available online at
http://www.r744.com/articles/8926/greenchill reports growth in co2 systems at u s supermarkets t?
Refrigeration and Air Conditioning (RAC) Magazine. January 7, 2013. US car giant GM publicly backs HFO-1234yf.
Available online at http://www.racplus.com/news/us-car-giant-gm-publiclv-backs-
1234vf/8640815.article?blocktitle=Latest-news&contentlD=2332 c?
Refrigeration, Air Conditioning and Heat Pumps Technical Options Committee. 2011. 2010 Assessment. Nairobi:
UNEP. Available online at http://ozone.unep.org/teap/Reports/RTOC/RTOC-Assessment-report-201Q.pdfcf
Russian Federation. 1994. Achievable ODS Phase-out Schedule in the Russian Federation. Available online at:
http://documents.worldbank.org/curated/pt/326351468776405093/pdf/multi0page.pdfc?
Shecco. 2012. 2012: Natural refrigerants market growth for Europe.
Shen, B., O. Abdelaziz, and L. Liudahl. 2016. Assessment of DR-55 as a drop-in replacement for R410A.
International Refrigeration and Air Conditioning Conference. Paper 1708. Available online at
https://pdfs.semanticscholar.org/5cla/db29505534d9686e9560c7e0410118deal29.pdfc?
Stanga, M. December 13, 2012. Correspondence: cost of R-32 for unitary AC equipment.
Supermarket News. 2012. Refrigeration systems chillin' with carbon dioxide. Available online at
http://supermarketnews.com/technology/refrigeration-svstems-chillin-carbon-
dioxide?page=lshowtext.cfm?t=ptb0810 c?
Technology and Economic Assessment Panel. 2012. Decision XXIII/9 Task Force Report: Additional Information on
Alternatives to Ozone-Depleting Substances.
Thermo King. 2014. Refrigerant Conversion Guide. Available online at
http://thermoking.ipublishpro.com/119135919RefrGuideEN/ib/pdf.pdfc?
U.S. Department of Agriculture. 2017. International Macroeconomic Data Set. Accessed November 2017. Available
online at https://www.ers.usda.gov/data-products/international-macroeconomic-data-set/international-
macroeconomic-data-set/. #Baseline Data Files.%20
U.S. Environmental Protection Agency and National Highway Traffic Safety Administration. 2011. Draft Joint
Technical Support Document: Proposed Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards.
U.S. Environmental Protection Agency. 2017. Vintaging Model. Version VM 10 file_v4.4_02.22.17.xls.
U.S. Environmental Protection Agency. May 1998. Draft Regulatory Impact Analysis: The Substitutes Recycling Rule.
Prepared by ICF Incorporated for U.S. Environmental Protection Agency. Washington, DC: ICF International.
5-128
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
United Nations Environment Programme. 2012. Report of the Technology and Economic Assessment Panel,
Additional Information on Alternatives to Ozone Depleting Substances. Decision XXII1/9 Task Force Report.
Available online at https://ozone.unep.org/science/assessment/teap?page=l J
United Nations Environmental Programme. 2017a. Overall Analysis of the Results of the Surveys ofODS
Alternatives (Decision 79/42(C)). Executive Committee of the Multilateral Fund for the Implementation of the
Montreal Protocol. Eightieth Meeting. Montreal, 13-17 November 2017. Available online at
http://www.multilateralfund.Org/80/Document%20Librarvl/l/8054.pdfc?
United Nations Environment Programme. 2017b. Production and Consumption of Ozone Depleting Substances,
1986-2016. United Nations Environment Programme (UNEP) Ozone Secretariat. Available online at
http://ozone.unep.org/reporting/cf
United Nations Environment Programme. 2018. OzonAction Kigali Fact Sheet No. 17—Refrigerant Trends in Mobile
Air-Conditioning. Available online at
https://wedocs.unep.org/bitstream/handle/20.500.11822/26651/7946FS17MACSector.pdf?sequence=l&isAII
owed=vc?
WalMart. September 2006. DOW Case Study: Secondary Loop Refrigeration and DOWFROST™ Fluid Contribute to
Wal-Mart "Green Initiative."
Wang, K., M. Eisele, Y. Hwang, and R. Radermacher. October 2010. Review of secondary loop refrigeration systems.
International Journal of Refrigeration, 33(2), 212-234.
Xu, X., Y. Hwang, and R. Radermacher. July 16-19, 2012. Performance measurement ofR32 in vapor injection heat
pump system. Presented at the International Refrigeration and Air Conditioning Conference at Purdue.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-129
-------
METHODOLOGY DOCUMENTATION
Solvent Use Mitigation Options Considered
Historically, CFCs (in particular CFC-113), methyl chloroform, and, to a lesser extent, carbon tetrachloride were
used as the predominant solvent cleaning agents. HFCs, hydrofluoroethers (HFEs), PFCs, and aqueous and semi-
aqueous NIK solvents have since replaced these legacy solvents, with HFC emissions currently dominating the
GWP-weighted emissions from the solvents sector.
A total of four abatement options were identified and analyzed for the solvent sector: (1) replacement of HFCs
with HFEs, (2) retrofitting of equipment, (3) transition to NIK aqueous, and (4) transition to NIK semi-aqueous. This
section briefly describes each option. Table 5-43 provides a technology overview of each abatement option in
terms of its assumed reduction efficiency and applicability by facility type.
Table 5-43: Solvent Use Abatement Options
Abatement Option
Reduction Efficiency
Applicability
HFC to HFE
76.4%
All facilities
Retrofit
50.0%
Nonretrofitted facilities
NIK aqueous
100.0%
Electronics cleaning
NIK semi-aqueous
100.0%
Electronics cleaning
For the purposes of evaluating the cost of reducing HFC emissions, this analysis characterized four model
facilities of emission sources:
Precision cleaning applications with retrofitted equipment: Precision cleaning may apply to electronics
components; medical devices; or metal, plastic, or glass surfaces and is characterized by applications that require a
high level of cleanliness to ensure the satisfactory performance of the product being cleaned. This facility is
defined as a vapor degreaser that is 10 sq. ft. in size, uses HFC-4310mee as a solvent, and emits approximately 250
pounds of solvent annually (Owens, 2008). The facility was assumed to have already retrofitted its equipment
through engineering control changes and improved containment to minimize emissions to comply with stringent
environmental and safety regulations (e.g., the National Emissions Standards for Hazardous Air Pollutants
[NESHAPs] in the United States) that limit emissions from solvent cleaning equipment in the United States and
other developed countries.
Precision cleaning applications with nonretrofitted equipment: This facility is characterized to generally
distinguish between precision cleaning facilities in developed and developing countries. The degreaser size and
type of solvent used are identical to the precision cleaning facility mentioned above; however, this facility is
assumed to not have retrofitted equipment to better control emissions because of the lack of regulations requiring
such controls. Thus, the amount of HFC solvent lost annually is higher; this analysis assumed a loss of
approximately 500 pounds annually for this facility based on the assumption that retrofitted equipment emits 50%
less than nonretrofitted equipment (Durkee, 1997).
Electronics cleaning applications with retrofitted equipment: Electronics cleaning, including defluxing and
other cleaning operations, is defined as a process that removes contaminants, primarily solder flux residues, from
electronics and circuit boards. This facility is defined as a vapor degreaser 10 sq. ft. in size, which uses HFC-
4310mee as a solvent and emits approximately 250 pounds of solvent annually (Owens, 2008). Similar to the
precision cleaning applications with equipment retrofits, this facility was assumed to have already retrofitted its
equipment through engineering control changes and improved containment to minimize emissions because of
regulations in place to control VOC emissions.
5-130
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Electronics cleaning applications with nonretrofitted equipment: This facility is characterized to generally
distinguish between electronics cleaning facilities in developed and developing countries. The degreaser size and
type of solvent used by this facility are identical to the electronics cleaning facility mentioned above; however, this
facility was assumed to not have retrofitted equipment to better control emissions because of the lack of
regulations requiring such controls. This analysis assumed emissions of approximately 500 pounds annually from
this facility based on the assumption that retrofitted equipment emits 50% less than nonretrofitted equipment
(Durkee, 1997).
HFCs to HFEs
This option, which is applicable to all facilities in the baseline, examines the replacement of HFC-4310mee with
lower GWP HFE solvents. Although other low-GWP chemicals may be feasible, HFE-7100 and HFE-7200 were used
as proxies for this abatement option because they display material compatibility properties similar to HFCs, a
prime factor that has led to their success in the market. To model emission reductions, this option assumes that
the degreaser transitions to the use of 75% HFE-7100 and 25% HFE-7200.70 For the purpose of this analysis, the
100-year GWP of alternative solvents was calculated as the weighted average of 75% HFE-7100 with a GWP of 297
and 25% HFE-7200 with a GWP of 59 for a GWP of 238.
• Capital Cost: HFE solvents are very similar to HFC-4310mee in their key chemical properties, such that
existing equipment designed with low emission features can still be used with HFE solvents, although the
equipment might need minor adjustments, such as resetting of the heat balance. These modifications are
not likely to amount to a substantial one-time cost (ICF Consulting, 2003; Owens, 2003); therefore, this
analysis assumed no one-time costs for converting to an HFE solvent.
• Annual O&M Costs: HFE solvents have pricing structures roughly equal to the pricing structure of HFCs
(Owens, 2003). Therefore, this analysis assumed no annual costs are incurred when transitioning to an
HFE solvent.
• Annual Revenue: This analysis did not assume a cost savings. A net cost savings may occasionally be
experienced by end users that choose HFE solvents that are lower in density than HFC-4310mee (Owens,
2003). For example, because the same volume of solvent is used and solvents are sold on a mass basis,
formulations blended with HFE-7200 may be lower in cost relative to formulations containing HFC-
4310mee.
Retrofit
This abatement option is applicable to nonretrofitted facilities using solvents for precision cleaning and
electronics cleaning. Retrofits, including engineering control changes (e.g., increased freeboard height, installation
of freeboard chillers, and use of automatic hoists), improved containment, and implementation of other
abatement technologies can reduce emissions of HFCs used in solvent cleaning. Retrofitting a vapor degreaser,
combined with proper O&M, can reduce solvent emissions from 46% to as much as 70%, depending on the specific
retrofit methods chosen (Durkee, 1997). For example, installing a freeboard refrigeration device, sometimes
referred to as a chiller (i.e., a set of secondary coils mounted in the freeboard), and maintaining a freeboard ratio
of 1.0 to minimize diffusional solvent losses, can reduce emissions by 46%, while installing heating coils to produce
superheated vapor along with installing a chiller can reduce emissions by 70%. For this analysis, the reduction
efficiency of the retrofit option was assumed to equal 50%.
In the United States, many enterprises have bought new equipment or retrofitted aging equipment into
compliance with the NESHAP, which limits emissions from degreasers using traditional chlorinated solvents such as
trichloroethylene. Fluorinated solvents such as HFCs are not covered by this regulation; nonetheless, a number of
companies using HFCs and other nonchlorinated solvents have adopted NESHAP-compliant solvent cleaning
70 In actuality, a facility would choose one of the two HFEs for adoption; however, for modeling purposes this assumption was
used to reflect the market presence of the two HFEs.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-131
-------
METHODOLOGY DOCUMENTATION
machines because of the associated economic, occupational, and environmental benefits (i.e., reduced emissions)
(Durkee, 1997). Likewise, many European countries have imposed stringent environmental and safety regulations
that require the lowest level of emissions attainable by solvent degreasing equipment. Retrofit techniques were
either already implemented or simply not required if the user had purchased new emission-tight vapor degreasers.
Consequently, this analysis assumed that end users in the United States and developed countries have already
adopted these controls and that any emissions from these facilities cannot be further abated by this option. In
contrast, solvent users in Non-Al (i.e., developing) countries were assumed to not have retrofitted their
equipment but may consider the equipment retrofit option, because updating their equipment may be preferred
over investing in entirely new units.
• Capital Cost: To retrofit equipment, significant upgrades must be made. One-time costs were based on
the assumption that a user chooses to retrofit their equipment through increasing freeboard height,
installing a cover, and installing a freeboard refrigeration device. Based on these upgrades, one-time costs
were assumed to be $24,500 (Durkee, 1997).
• Annual O&M Costs: No annual costs are associated with this abatement option.
• Annual Revenue: Annual savings are associated with the avoided consumption of HFC that results from a
reduction in emissions. An annual cost savings of almost $4,500 was assumed based on the assumed
reduction in emissions of 250 pounds per year of HFCs that would otherwise need to be replaced.
NIK Aqueous
This abatement option is applicable only to facilities that use solvents for the purposes of electronics cleaning.
This option replaces HFC-containing systems used for electronics cleaning end uses with an aqueous cleaning
process. In the aqueous process, a water-based cleaning solution is used as the primary solvent and is usually
combined with a detergent to remove contaminants. Because all HFCs are replaced with a solvent that does not
have a GWP, the reduction efficiency of this option is 100%.
• Capital Cost: Vapor degreasers are not suitable for retrofit to aqueous cleaning processes (Crest
Ultrasonic, 2008). Therefore, the cost of replacing an HFC-containing cleaning system with an aqueous
system was based on the initial investment in tanks, equipment, and space (Beeks, 2008). This analysis
assumed a one-time cost of $50,000 for the investment in the equipment and the additional space
needed for that equipment (Owens, 2008).
• Annual O&M Costs: The major operating costs for an aqueous system are associated with the cost of
energy and the cost of the continuous flow of de-ionized water (Crest Ultrasonic, 2008). An annual cost of
$7,400 was used to represent energy and water consumption costs; this cost was based on consumption
of 9 kilowatt (kW) per day and $10 worth of de-ionized water per day (Owens, 2008).
• Annual Revenue: Annual savings were based on the savings associated with not using an HFC-based
cleaning system. An annual savings of $6,700 was used to represent energy and HFC solvent cleaner costs
associated with using a retrofitted HFC-based cleaning system, while an annual savings of $11,200 was
used to represent energy and HFC solvent cleaner costs associated with using a nonretrofitted HFC-based
cleaning system; this savings was based on consumption of 4 kW per day and 250 to 500 pounds of HFC
lost per year (Owens, 2008; Durkee, 1997).
NIK Semi-aqueous
This abatement option is applicable only to facilities that use solvents for the purposes of electronics cleaning.
This option replaces HFC-containing systems used for electronics cleaning end uses with a semi-aqueous cleaning
process. In the semi-aqueous process, the cleaning solution is an organic solvent that is blended with a surfactant,
making it water soluble. An example of a solvent/surfactant blend is a terpene/water combination blended with
glycol ethers. Because all HFCs are replaced with solvents that have no GWP, the reduction efficiency is 100%.
5-132
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Capital Cost: Compared with aqueous systems, semi-aqueous systems often require an extra tank or two
as well as the need for ventilation. Therefore, semi-aqueous systems were assumed to be slightly higher
in cost than aqueous systems (Crest Ultrasonic, 2008). Additionally, vapor degreasers are not suitable for
retrofit to semi-aqueous cleaning processes (Crest Ultrasonic, 2008). Therefore, the cost of replacing an
HFC-containing cleaning system with a semi-aqueous system was based on the total initial investment in
tanks and equipment (Beeks, 2008). This analysis assumed a one-time cost of $55,000 for the investment
in the equipment and the additional space needed for that equipment (Owens, 2008).71
• Annual O&M Costs: The major operating costs for a semi-aqueous system are associated with the cost of
energy and the cost of the continuous flow of de-ionized water. Compared with aqueous systems, semi-
aqueous systems add a level of complication and were, therefore, assumed to require more energy. As a
result, an annual cost of $9,100 was used to represent energy and water consumption costs; this cost was
based on consumption of 12 kW per day and $10 worth of de-ionized water per day (Owens, 2008).
• Annual Revenue: Annual savings are based on the savings associated with not using an HFC-based
cleaning system. An annual savings of $6,700 was used to represent energy and HFC solvent cleaner costs
associated with using a retrofitted HFC-based cleaning system, while an annual savings of $11,200 was
used to represent energy and HFC solvent cleaner costs associated with using a nonretrofitted HFC-based
cleaning system; this savings was based on consumption of 4 kW per day and 250 to 500 pounds of HFC
lost per year (Owens, 2008; Durkee, 1997).
Technical and Economic Characteristics Summary
The analysis also developed a technical effectiveness parameter, defined as the percentage reductions
achievable by each technology/facility type combination. Estimating this parameter required making several
assumptions regarding the distribution of emissions from model facilities in addition to process-specific estimates
of technical applicability and market penetration. Market penetration rates vary over time as systems are
upgraded and the options are applied in the future. Table 5-44 summarizes these assumptions and presents
technical effectiveness parameters used in the MAC model.
Table 5-44: Technical Effectiveness Summary—Solvent Use
Technical
Applicability
Facility/Abatement Option (2030)
Market
Penetration Rate
(2030)a
Reduction
Efficiency
Technical
Effectiveness
(2030)b
Precision Retrofitted—U.S. and Other Developed and EU
HFCtoHFE 60%
87%
85%
44%
Precision Nonretrofitted—Developing
HFCtoHFE 60%
65%
85%
26%
Retrofit 100%
18%
50%
14%
Electronics Retrofitted—U.S. and Other Developed and EU
HFCtoHFE 100%
67%
85%
57%
Aqueous 100%
7%
100%
7%
Semi-aqueous 100%
7%
100%
7%
(continued)
71 Although these costs are higher than the NIK aqueous abatement option, it was assumed that the semi-aqueous option will
nonetheless be adopted in some facilities, for example, where the NIK aqueous option might not be effective for the particular
cleaning required.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-133
-------
METHODOLOGY DOCUMENTATION
Table 5-44: Technical Effectiveness Summary—Solvent Use (continued)
Facility/Abatement Option
Technical
Applicability
(2030)
Market
Penetration Rate
(2030)a
Reduction
Efficiency
Technical
Effectiveness
(2030)b
Electronics Nonretrofitted—
Developing
HFC to HFE
100%
20%
85%
17%
Retrofit
100%
16%
50%
8%
Aqueous
100%
2%
100%
2%
Semi-aqueous
100%
2%
100%
2%
a Market penetration assumptions for this analysis vary over time, and the technical effectiveness values are based on the
cumulative market penetration rates assumed until that point.
b Technical effectiveness figures represent the percentage of baseline emissions from the relevant facility type that can be
abated in 2030; figures do not account for indirect GHG impacts associated with decreased electricity consumption (e.g.,
aqueous and semi-aqueous cleaning), which are accounted for in the cost analysis.
Table 5-45 presents the engineering cost data for each mitigation option outlined above, including all cost
parameters necessary to calculate the break-even price. This section provides additional detail on the base cost
estimates.
Table 5-45: Engineering Cost Data on a Facility Basis—Solvent Use
Abatement Option
Project
Lifetime
(years)
Capital Cost
(2015 USD)
Annual
Revenue
(2015 USD)
Annual
O&M Costs
(2015 USD)
Abatement
Amount
(tCOze)
HFC to HFE
Retrofitted—Developed
15
—
—
—
159
Nonretrofitted—Developing
15
—
—
—
191
Inert gas
Nonretrofitted—Developing
15
$24,500
$4,500
—
186
NIK aqueous
Retrofitted—Developed
15
$50,000
$6,700
$7,400
186
Nonretrofitted—Developing
15
$50,000
$11,200
$7,400
224
NIK semi-aqueous
Retrofitted—Developed
15
$55,000
$6,700
$9,100
186
Nonretrofitted—Developing
15
$55,000
$11,200
$9,100
224
Sector-Level Trends/Considerations
In developed countries, retrofits were assumed to have already been fully adopted, and in developing
countries all equipment was assumed to remain nonretrofitted. In addition, although NIK replacement alternatives
and HFE solvent applications currently exist worldwide, the baseline emissions considered here only cover that
portion of the market still using HFCs and PFCs. Hence, for the purposes of transitioning away from the high-GWP
solvents in this analysis, we modeled no technology adoption of the NIK and HFE solvents in the baseline.
5-134
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
References
Beeks, M., Brulin & Company, Inc. October 2008. Personal communication with Emily Herzog and Mollie Averyt,
ICF International.
Crest Ultrasonic. September/October 2008. Personal communication between Eric Larson of Crest Ultrasonics and
Emily Herzog, ICF International.
Durkee, J.B. November 12-13,1997. Chlorinated solvents NESHAP—results to date, recommendations and
conclusions. Presented at the International Conference on Ozone Layer Protection Technologies, Baltimore,
MD.
ICF Consulting, October 7, 2003. Personal communication between solvent industry experts and William Kenyon,
ICF Consulting.
Owens, J., 3M. August 2008. Personal communication with Emily Herzog, ICF International.
Owens, J.G., P.E., 3M Performance Materials. October 27, 2003. Written correspondence with Mollie Averyt and
Marian Martin Van Pelt, ICF Consulting.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-135
-------
METHODOLOGY DOCUMENTATION
Foam Manufacturing Mitigation Options Considered
Foam is used as insulation in a range of equipment and products, including refrigerated appliances, reefers,
and other refrigerated transport systems; in buildings (e.g., walls, roofs, floors) and pipes; and in the production of
other products, such as steering wheels, furniture, and shoes, for example. A wide variety of foam types are used
for these applications, which have historically been produced with blowing agents that are ODSs (i.e., CFCs and
HCFCs), but since the ODS phaseout under the Montreal Protocol, HFCs have commonly been adopted in their
place-primarily HFC-134a, HFC-152a, HFC-245fa, and HFC-365mfc.
This analysis considered the costs of reducing foam emissions by replacing HFCs with low-GWP blowing agents
in various types of foam manufacturing operations. Specifically, four abatement options were identified and
analyzed for reducing emissions at product/equipment production for various polyurethane (PU) and extruded
polystyrene (XPS) foam products by using HC, HFO, and/or HCFO blowing agents (or blends thereof) in place of
HFCs. By replacing HFCs with the identified abatement options, this analysis assumed that the total amount of
high-GWP blowing agent that would have been emitted during the lifetime of the foam products produced by a
facility in a given year is eliminated.
Table 5-46 summarizes the reduction efficiency assumed for each mitigation option used in the MAC analysis.
Table 5-46: Foams Manufacturing Abatement Options
Abatement Option
Reduction
Efficiency
Applicability
PU commercial refrigeration foam: HFC-245fa
to HCFO-1233zd(E)-EU
99%
New PU commercial refrigeration units
Low-pressure (LP), two-component PU spray
foam: HFC-245fa/C02 to HFO-1234ze(E)/HCFO-
1233zd(E)—Developed
99%
New LP two-component spray foam products
XPS: HFC-134a/C02to HFO-1234ze(E)-EU
53%
New XPS boardstock foam products
PU OCF: HFC-134a to HCs-Developing
100%
New PU OCF applications products
HCFO-1233zd(E) in Commercial Refrigeration
This option replaces HFC-245fa used in commercial refrigeration foam with unsaturated HCFCs or HCFOs,
namely HCFO-1233zd(E). HCFO-1233zd(E) has a GWP of 4.7 to 7 (EPA, 2012). Unsaturated HFCs and HCFCs with
low GWPs have emerged in the market on a commercial scale as alternative blowing agents in various foam
applications over the last several years.
This abatement option was applied to a facility that manufactures 50,000 units annually, with each unit
containing 1.4 kilograms of blowing agent. This option is applicable to HFC-245fa in newly manufactured
commercial refrigeration equipment in the EU.
• Capital Costs: HCFO-1233zd(E) is very similar to HFC-245fa, such that existing manufacturing equipment
requires little to no change to achieve the desired operational characteristics. Therefore, no one-time
costs were assumed.
• Annual O&M Costs: Annual costs were assumed to be a result of the incremental cost of the blowing
agent replacement, which is approximately $4/kg for HCFO-1233zd(E) relative to HFC-245fa (UNEP, 2012).
• Annual Revenue: No annual savings were assumed.
5-136
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
HFO-1234ze(E)/HCFO-1233zd(E) in Low-Pressure, Two-Component Spray Foam (Frothing Insulation and
Sealant)
This abatement option applies the use of a blend of HFO-1234ze(E) and HCFO-1233zd(E) as a replacement for
HFC-134a and HFC-245fa in low-pressure (LP), two-component spray foam (i.e., frothing insulation and sealant).
The abatement option was assumed to be a blowing agent blend of HFO-1234ze(E) and HCFO-1233zd(E), with the
same composition as the respective gaseous and liquid HFC blowing agent components assuming a 1:1
replacement ratio (i.e., approximately 18% and 82%, respectively).
This abatement option was applied to a facility that manufactures 2.1 million pounds of LP two-component PU
spray foam using HFC-134a/HFC-245fa blowing agent annually (Dow, 20 18).72 It was assumed that the foam
formulation is 17% by weight blowing agent (approximately 3% of that is HFC-245fa and the remainder is HFC-
134a) (Dow, 2018), so the facility uses 350,000 pounds of blowing agent annually. This option is applicable to HFC-
134a and HFC-245fa in newly manufactured LP two-component spray foam products in developed countries,
except the EU.
• Capital Costs: Capital costs for transitioning to HFO-1234ze(E)/HCFO-1233zd(E) blowing agent are
estimated to be up to $2 million per facility (i.e., $1 million for equipment modifications and $1 million for
manufacturing plant upgrades to handle flammable material [BASF, 2014; Dow, 2018]). HFO-1234ze(E) is
mildly flammable, so the manufacturing plant will need to be upgraded to accommodate the use of
flammable material. These upgrades include safety modifications such as new leak detection monitors,
new holding tanks for raw material with containment barriers, and ventilation (BASF, 2014; Dow, 2018).
Furthermore, the electrical classification must be upgraded for all equipment, such as mixing tanks,
motors, light switches, and light fixtures (Dow, 2018). Costs depend on the size, layout, and existing
equipment in the plant (BASF, 2014). Standard spraying equipment at end use was also assumed to be
used for this abatement option (UNEP, 2008).
• Conversion Costs: Conversion costs include foam reformulation, product development, certifications,
shelf-life studies, field and external testing, and commercialization (American Chemistry Council Center
for Polyurethanes Industry [ACC CPI], 2014; BASF, 2014; Clayton Corporation, 2014; Dow, 2014; Fomo
Products, 2014; RHH, 2014). Reformulation includes studying, optimizing, and testing the new foam
product and requires third-party testing approvals, internal manufacturer or customer standards, the use
of third-party testing, and iterative research and development (ACC CPI, 2014). New products must be
tested to meet internal quality controls, dimensional stability, voluntary performance standards, and/or
mandatory standards and regulations. The tests verify performance metrics such as thermal performance,
fire safety, and other life and safety requirements (ACC CPI, 2014; BASF, 2014). Conversion costs are
estimated to be several million dollars (BASF, 2014). For the purposes of this analysis, conversion costs
were assumed to be $3 million.
• Annual O&M Costs: Annual O&M costs include labor, energy, and training. Annual O&M costs for HFO-
1234ze(E)/HCFO-1233zd(E) in LP two-component PU spray foam are expected to be the same as with
HFC-134a/HFC-245fa. Additional training required to handle flammable material is not expected to be
significantly different (i.e., may add one to two hours per employee [Dow, 2018]). Thus, the only annual
incremental cost increase is due to the price differential between the blowing agents, which is
approximately $7.5/kg for HFO-1234ze(E) relative to HFC-134a and $4/kg for HCFO-1233zd(E) relative to
HFC-245fa (UNEP, 2012).
72 This estimate was calculated using blowing agent consumption for the spray foam market in the United States modeled in
EPA's Vintaging Model and assuming six LP two-component spray foam manufacturers in the United States, each with one
facility, that comprise 10% of the total spray foam market. These six manufacturers are Dow Chemical Company, Fomo
Products Inc., RHH Foam Systems Inc., Clayton Corporation, BASF Corporation, and Commercial Thermal Solutions, Inc.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-137
-------
METHODOLOGY DOCUMENTATION
• Annual Revenue: Because the alternative blowing agents are more expensive on a mass basis than HFC-
134a/HFC-245fa, no annual cost savings are associated with this option. Cost savings associated with the
increase in yield (product coverage) and R-value (insulation value) were not quantitatively accounted for
in this analysis.
HFC-134a/HFO-1234ze(E) in XPS Boardstock
This abatement option applies the use of blends of HFO-1234ze(E) and HFC-134a as a replacement for HFC-
134a and CCh-based blends in extruded XPS boardstock foam. The abatement option is assumed to be a blowing
agent blend of HFC-134a (42%) and HFO-1234ze(E) (58%), which is equivalent to the formulation for R-450A (i.e., a
refrigerant blend currently available in the market).
This abatement option was applied to an XPS boardstock manufacturing facility that produces 100 million
board feet of HFC-134a/CC>2 XPS boardstock across 1.5 lines per year (i.e., 66.7 million board feet per line [Owens
Corning, 2014; Extruded Polystyrene Foam Association [XPSA], 2018]).73 This option applies to all HFC-134a/C02
blends in newly manufactured XPS boardstock foam in the EU.
• Capital Costs: The capital costs for switching to HFC-134a/HFO-1234ze(E) are estimated to be $1.8 million
across 1.5 lines with up to an additional $1 million per manufacturing plant for plant upgrades for
handling flammable material (BASF, 2014; Dow, 2018). HFO-1234ze(E) is mildly flammable, so the
manufacturing plant will need to be upgraded to accommodate the use of flammable material. These
upgrades include safety modifications such as new leak detection monitors, new holding tanks for raw
material with containment barriers, and ventilation (Dow, 2018). Furthermore, the electrical classification
must be upgraded for all equipment, such as mixing tanks, motors, light switches, and light fixtures (Dow
2018). Costs depend on the size, layout, and existing equipment in the plant (BASF, 2014).74
• Conversion Costs: According to XPSA, conversion costs for transitioning to a new blowing agent are
required for conducting research and development to ensure the finished product can meet R-5
performance requirements per ASTM C-578 standards, pilot-scale trials, full-scale trials, certification
testing, building code and fire code listing updates, emission permits, and negotiating of contracts with
new suppliers (XPSA, 2014). Owens Corning spent $132 million for the conversion from CFCs to HCFC-
142b to its current HFC-134a/HFC-152a blend across six lines in four facilities in North America (Owens
Corning, 2014). Assuming the conversion costs were equivalent for each blowing agent transition and
incurred equally across the facilities (in the absence of more definitive information about each transition
and facility), conversion costs are estimated to be $16.5 million per facility and $11 million per line.
• Annual O&M Costs: Annual O&M costs include labor, energy, and training. Total O&M costs are
estimated to be $240,000, based on an estimated cost of $0.03 per board foot (Russell, 2005). It was
assumed that blowing agent conversions would result in a 10% decrease in capacity because of slower
production throughputs caused by the need for rapid expansion control for the alternative blend (Russell,
2005; XPSA, 2018). The annual incremental cost increase due to the price differential between the
blowing agents is approximately $5/kg. The goal is to have the density and thickness of the foam not
change with this abatement option. Density is critical to performance properties of XPS boardstock foam,
including R-value and compressive strength, which are required to meet building codes. Furthermore, XPS
boardstock foam blown with an alternative blowing agent must maintain typical thickness because of the
limited space requirements for building construction (Dow, 2018).
73 Owens Corning manufactures XPS boardstock across six lines in four facilities in North America (Owens Corning, 2014).
According to XPSA, 100 million board feet per year is a reasonable estimate for XPS boardstock production per line and, on
average, manufacturing facilities have four to six lines (XPSA, 2018).
74 Comments submitted by BASF are specific to facilities manufacturing PU foam but are applicable to XPS foam manufacturing
given the similarities between plant upgrades needed for handling flammable material.
5-138
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Annual Revenue: Because the alternative blowing agent is more expensive on a mass basis than HFC-
134a/CC>2, no annual cost savings are associated with this option.
HCs in PU One-Component Foam (Sealant Cylinders)
This abatement option replaces HFC-134a used in PU one-component foam (OCF) (sealant cylinders) with an
HC blowing agent. HCs are inexpensive and have very low direct GWPs. Technical issues exist with using HCs, such
as flammability and lower insulation performance, but these can be overcome through proper safety controls and
engineering design (EPA, 2009). A significant advantage of HCs is that they can be easily blended to affect a range
of properties, such as thermal performance, cell gas pressure, and foam density, as well as cost (UNEP, 2012).
Flammability, however, may cause a high incremental capital cost for facilities, which may be uneconomic for
small- or medium-sized enterprises; otherwise, HCs have low operating costs (UNEP, 2012). Use of butane and
propane in OCFs is well proven and is associated with low operating costs (UNEP, 2012).
This abatement option was applied to a facility that manufactures 160,000 pounds of LP PU OCF sealant
cylinders across one line using HFC-134a blowing agent annually.75 It was assumed that the foam formulation is
8.7% by weight blowing agent, so the facility uses 140,000 pounds of blowing agent annually. This option applies to
all HFC-134a in newly manufactured PU OCF in developing countries.
• Capital Costs: Capital costs for switching to HCs are estimated to be up to $2 million (i.e., $1 million per
facility for equipment modifications for one line and $1 million for facility upgrades for handling
flammable material).76 HCs are flammable, so the manufacturing plant will need to be upgraded to
accommodate the use of flammable material. These upgrades include safety modifications such as new
leak detection monitors, new holding tanks for raw material with containment barriers, and ventilation
(BASF, 2014; Dow, 2018). Furthermore, the electrical classification must be upgraded for all equipment,
such as mixing tanks, motors, light switches, and light fixtures (Dow, 2018). Costs depend on the size,
layout, and existing equipment in the plant (BASF, 2014). Standard spraying equipment at end use was
also assumed to be used for this abatement option (UNEP, 2008).
• Conversion Costs: Conversion costs include foam reformulation, product development, certifications,
shelf-life studies, field and external testing, and commercialization (ACC CPI, 2014; BASF, 2014; Clayton
Corporation, 2014; Dow, 2014; Fomo Products, 2014; RHH, 2014). Reformulation includes studying,
optimizing, and testing the new foam product and requires third-party testing approvals, internal
manufacturer or customer standards, the use of third-party testing, and iterative research and
development (ACC CPI, 2014). New products must be tested to meet internal quality controls,
dimensional stability, voluntary performance standards, and/or mandatory standards and regulations. The
tests verify performance metrics such as thermal performance, fire safety, and other life and safety
requirements (ACC CPI, 2014; BASF, 2014). Conversion costs are estimated to be several million dollars
(BASF, 2014). For the purposes of this analysis, conversion costs were assumed to be $3 million.
• Annual O&M Costs: Annual O&M costs include labor, energy, and training. Annual O&M costs for
switching production to HCs from HFC-134a are estimated to be equal. Additional training required to
handle flammable material is not expected to be significant (Dow, 2018). Thus, no annual O&M costs
were assumed.
75 This estimate was calculated using blowing agent consumption for the OCF market in the United States modeled in EPA's
Vintaging Model and assuming six OCF manufacturers in the United States, each with one facility. These six manufacturers are
Dow Chemical Company, Fomo Products Inc., RHH Foam Systems Inc., Clayton Corporation, BASF Corporation, and Commercial
Thermal Solutions, Inc.
76 Capital costs were assumed to be equivalent for any abatement option using flammable blowing agents in OCFs.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-139
-------
METHODOLOGY DOCUMENTATION
• Annual Revenue: Annual savings are estimated at about $587,000 per facility associated with incremental
blowing agent replacement costs. Cost savings associated with the increase in yield (product coverage)
and R-value (insulation value) were not quantitatively accounted for in this analysis.
Technical and Economic Characteristics Summary
The analysis also developed a technical effectiveness parameter, defined as the percentage reductions
achievable by each technology/facility type combination. Estimating this parameter required developing
assumptions regarding the distribution of emissions by abatement option in addition to process-specific estimates
of technical applicability and market penetration. This analysis assumed that the targeted abatement amount for
each abatement option in the foam sector is the total amount of high-GWP blowing agent that would have been
emitted during the lifetime of the foam products produced by a facility in a given year.
Closed-cell foams are assumed to emit a portion of their total blowing agent content upon manufacture, a
portion at a constant rate over the lifetime of the foam, and a portion at disposal; these portions vary by end use.
The emission profiles of the foams included in this analysis are shown in Table 5-47. According to the emission
profiles, 100% of the blowing agent used in each foam product is eventually emitted over its lifetime. Thus, the
targeted abatement amount (i.e., emissions) for each option is equivalent to the amount of blowing agent
consumed in a given year by the manufacturing facility (i.e., the total amount of blowing agent that would have
been emitted from all products manufactured in that year throughout their lifetime).
Table 5-47: Emission Profiles for Foam End Uses
Foam End Use
Loss at
Manufacturing
(%)
Annual
Leakage
Rate (%)
Leakage
Lifetime
(years)
Loss at
Disposal
(%)
Total
Emissions
(%)
PU commercial refrigeration units
4
0.25
15
92.25
100%
LP two-component PU spray foam
15
1.50
50
10.00
100%
XPS boardstock foam
25
0.75
25
56.25
100%
PU one-component foam
95
2.50
2
0
100%
Market penetration is a modeled value that considers the market's willingness to adopt the option, the rate of
uptake of the alternative into new foams, and the lifetime of the existing foam base. Because foam lifetimes can
be decades, replacing the stock of foams with non-HFC blowing agents will take many years. The market
penetration rate is modeled to capture such time lapses. Technical effectiveness figures do not account for indirect
GHG impacts associated with changes in electricity consumption (e.g., for foam-blowing processes), which are
accounted for in the cost analysis.77 Table 5-48 summarizes these assumptions and presents technical
effectiveness parameters used in the MAC model.
77 Indirect GHG emissions were not accounted for in the technical effectiveness calculations so that the analysis can show
purely ODS substitute (i.e., HFC) emission reductions achievable. It is recognized that indirect GHG emissions can be significant,
and such differences, to the extent data are available on them, are accounted for in the cost analyses.
5-140
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-48: Technical Effectiveness Summary—Foams Manufacturing
Facility/Abatement Option
Technical
Applicability
(2030)
Market
Penetration
Rate (2030)a
Reduction
Efficiency
Technical
Effectiveness
(2030)b
PU Commercial Refrigeration Units—EU
HFC-245fa to HCFO-1233zd(E)
13%
100%
99%
13%
LP Two-component PU Spray Foam—
U.S./Other Developed
HFC-245fa/C02 to HFO-1234ze(E)/HCFO-
1233zd(E)
52%
30%
99%
15%
XPS Boardstock Foam—EU
HFC-134a/C02to HFO-1234ze(E)
73%
80%
53%
31%
PU OCF—Developing
HFC-134a to HCs
95%
100%
100%
95%
a Market penetration assumptions for this analysis vary over time, and the technical effectiveness values are based on the
cumulative market penetration rates assumed until that point.
b Technical effectiveness figures represent the percentage of baseline emissions from the relevant facility type that can be
abated in 2030; figures do not account for indirect GHG impacts associated with increased electricity consumption (e.g., for
foam blowing processes or for EOL appliance processing), which are accounted for in the cost analysis.
Table 5-49 presents the engineering cost data for each mitigation option outlined above, including all cost
parameters necessary to calculate the break-even price.
Table 5-49: Engineering Cost Data on a Facility Basis—Foams Manufacturing
Abatement Option
Project
Lifetime
(years)
Capital Cost
(2015 USD)
Annual
Revenue
(2015 USD)
Annual O&M
Costs (2015
USD)
Abatement
Amount
(tC02e)
Commercial refrigeration: HFC-245fa to HCFO-1233zd(E)
PU commercial refrigeration foam
manufacturing facility—EU
25
$0
$0
$280,000
71,610
LP two-component spray: HFC-245fa/C02 to HFO-1234ze(E)/HCFO-1233zd(E)
PU spray foam manufacturing facility—
U.S./other developed
25
$5,000,000
$0
$230,124
58,912
XPS: HFC-134a/C02 to HFO-1234ze(E)
XPS boardstock production facility—EU
25
$19,300,000
$0
$4,039,260
516,331
OCF: HFC-134a to HCs
OCF manufacturing facility—Developing
25
$5,000,000
$587,088
$0
185,716
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-141
-------
METHODOLOGY DOCUMENTATION
Sector-Level Trends/Considerations
Available data on costs for abatement technologies were not scaled to reflect potential differences in the costs
outside of the United States. Additional research may be required to determine actual variability in costs across
regions. Moreover, the development of alternative blowing agents is quickly evolving; thus, new options may enter
the market, which should be considered quantitatively in future analyses (e.g., unsaturated fluorinated compounds
and methyl formate).
References
American Chemistry Council Center for Polyurethanes Industry. October 17, 2014. Comment submitted by Steve
Russell, Vice President, Plastics Division, CPI, ACC, for Docket ID No. EPA-HQ-OAR-2014-0198, Protection of
stratospheric ozone: change of listing status for certain substitutes under the significant new alternatives
policy program. Federal Register. Vol. 79. 46,126 (Aug. 6, 2014). Available online at
https://www. regulations.gov/document?D=EPA-HQ-QAR-2014-0198-0175
BASF Corporation. October 17, 2014. Comments of BASF Corporation on US EPA's Proposal to Change the Listings
from Acceptable to Unacceptable for HFC-134a, HFC-245fa, HFC-365mfc, and Any Blends Containing These
Blowing Agents for All Foam End-Uses and Applications Except for Spray Foam as of January 1, 2017. Available
online at https://www.regulations.gov/document?D=EPA-HQ-OAR-2014-0198-0093
Clayton Corporation. 2014. Comment submitted by Clayton Corporation for Docket ID No. EPA-HQ-OAR-2014-
0198, Protection of stratospheric ozone: change of listing status for certain substitutes under the significant
new alternatives policy program. Federal Register. Vol. 79. 46,126 (Aug. 6, 2014). October 20, 2014. Available
online at https://www.regulations.gov/document?D=EPA-HQ-QAR-2014-0198-0133
Dow Chemical Company. October 17, 2014. Comment submitted by the Dow Chemical Company for Docket ID No.
EPA-HQ-OAR-2014-0198, Protection of stratospheric ozone: change of listing status for certain substitutes
under the significant new alternatives policy program. Federal Register. Vol. 79.46,126 (Aug. 6, 2014).
Available online at https://www.regulations.gov/document?D=EPA-HQ-QAR-2014-0198-0204
Dow Chemical Company. August 7, August 29, September 28, and October 10, 2018. Personal communication
regarding foam abatement option assumptions between Dow (Lisa Massaro and Kristie Murray), EPA (Dave
Godwin), and ICF (Kasey Knoell).
Extruded Polystyrene Foam Association. October 20, 2014. Comment submitted by John Ferraro, Executive
Director, XPSA, for Docket ID No. EPA-HQ-OAR-2014-0198, Protection of stratospheric ozone: change of listing
status for certain substitutes under the significant new alternatives policy program, Federal Register. Vol. 79.
46,126 (Aug. 6, 2014). Available online at https://www.regulations.gov/document?D=EPA-HQ-OAR-2014-
0198-0189
Extruded Polystyrene Foam Association. September 17, 2018. Personal communication regarding foam abatement
option assumptions between EPA (Dave Godwin) and XPSA (John Heinze).
Fomo Products, Inc. October 17, 2014. Comment submitted by Dr. Thomas Fishback, Vice President, Research and
Development, Fomo Products, Inc., for Docket ID No. EPA-HQ-OAR-2014-0198, Protection of stratospheric
ozone: change of listing status for certain substitutes under the significant new alternatives policy program.
Federal Register. Vol. 79. 46,126 (Aug. 6, 2014). Available online at
https://www. regulations.gov/document?D=EPA-HQ-QAR-2014-0198-0139
Owens Corning. October 20, 2014. Comment submitted by Greg Mather, Vice President and General Manager
Foam Insulation Building Materials Group, Owens Corning, for Docket ID: No. EPA-HQ-OAR-2014-0198,
Protection of stratospheric ozone: change of listing status for certain substitutes under the significant new
alternatives policy program. Federal Register. Vol. 79. 46,126 (Aug. 6, 2014). Available online at
https://www. regulations.gov/document?D=EPA-HQ-OAR-2014-0198-0Q85
5-142
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
RHH Foam Systems, Inc. October 20, 2014. Comment submitted by Peter J. Bartell, Vice President/Co-Owner, RHH
Foam Systems, Inc., for Docket ID: No. EPA-HQ-OAR-2014-0198, Protection of stratospheric ozone: change of
listing status for certain substitutes under the significant new alternatives policy program. Federal Register.
Vol. 79. 46,126 (Aug. 6, 2014). Available online at https://www. regulations.gov/document?D=E PA-HQ-OAR-
2014-0198-0112
Russell, R. April 2005. Personal communication regarding foam production costs between ICF and Robert Russell.
United Nations Environment Programme. June 2008. Revised analysis of relevant cost considerations surrounding
the financing of HCFC phase-out (Decisions 53/37(1) and 54/40). 55th Meeting of the Executive Committee of
the Multilateral Fund for the Implementation of the Montreal Protocol. Available online at
http://www.multilateralfund.Org/sites/55th/Document%20Librarv2/l/5547.pdfcf
United Nations Environment Programme. 2012. Report of the Technology and Economic Assessment Panel (TEAP)
Decision XXIII/9 Task Force Report Additional Information on Alternatives to Ozone Depleting Substances.
Available online at http://ozone.unep.org/Assessment Panels/TEAP/Reports/TEAP Reports/teap-task-force-
XXIII-9-report-mav2012.pdf ef
U.S. Environmental Protection Agency. October 30, 2009. 2009 Marginal Abatement Cost Curve Analysis for
Reduction ofHFCs in Traditional Ozone Depleting Substance (ODS) End-Use Applications: Draft Report.
Prepared by ICF International.
U.S. Environmental Protection Agency. August 10, 2012. Protection of stratospheric ozone: determination 27 for
significant new alternatives policy program. Federal Register. Vol. 77, No. 155. Available online at
http://www.gpo.gov/fdsvs/pkg/FR-2012-Q8- 10/pdf/2012-19688.pdf
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-143
-------
METHODOLOGY DOCUMENTATION
Aerosol Product Use Mitigation Options Considered
Aerosol propellant formulations containing HFCs are used in a variety of consumer products and are
completely emitted during product use. This analysis estimates emissions from consumer aerosol products, such as
spray deodorants and hair sprays, and specialty aerosol uses, such as freeze spray and dust removal products, and,
separately, emissions from metered dose inhalers (MDIs). HFC-134a has been introduced as an alternative
propellant to CFCs in both MDIs and consumer aerosol products; in addition, HFC-227ea is used in MDIs and HFC-
152a is used in consumer aerosol applications.
A total of five abatement options were identified for the aerosols sector. For consumer aerosol products, the
options are transitioning to a replacement propellant—including HCs, HFO-1234ze, and HFC-152a (for those
products containing HFC-134a)—and converting to an NIK alternative, such as a stick, roller, or finger/trigger
pump. Costs were analyzed by looking at a model facility that uses HFC to fill 10 million aerosol cans a year. For
MDIs, the abatement measure examined by this analysis is further use of dry powder inhaler (DPI) technology
where suitable for the patient. Costs were analyzed based on a single DPI compared with a single MDI, with
estimated cost data that incorporate the cost associated with avoided use of HFC-134a propellant, the increase in
the cost of DPI treatment, the cost to market the new treatment, and the cost to retrain patients in using the DPI
(Ecofys, 2000; Enviros, 2000). Table 5-50 summarizes the applicability of each abatement option to the aerosol
emission categories. The subsequent subsections describe each abatement option in more detail.
Table 5-50: Aerosol Product Use Abatement Options
Reduction Efficiency
Consumer
Aerosol
Facility/
Abatement Option HFC-134a
Consumer
Aerosol
Facility/
HFC-152a
MDI
Applicability
Consumer Aerosol Products
HC 99.7%
97.5%
NA
Consumer aerosol facility/HFC-134a/HFC-152a
NIK 100%
100%
NA
Consumer aerosol facility/HFC-134a/HFC-152a
HFO-1234ze 99.5%
95.7%
NA
Consumer aerosol facility/HFC-134a/HFC-152a
HFC-134a to HFC-152a 89.2%
Consumer aerosol facility/HFC-134a
Pharmaceutical Aerosol Products (MDIs)
DPIs NA
NA
100%
Metered dose inhaler
For the purposes of evaluating the cost of reducing HFC emissions, this analysis characterized three categories
of emission sources:
• a facility that produces 10 million consumer aerosol cans per year, with each can containing an HFC-134a
aerosol propellant charge of 2 ounces;
• a facility that produces 10 million consumer aerosol cans per year, with each can containing an HFC-152a
aerosol propellant charge of 2 ounces; and
• a single 200-dose MDI aerosol unit with a charge size of 15 grams that uses HFC-134a propellant.
5-144
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
HCs
This option replaces HFC-134a or HFC-152a in non-MDI aerosols with an HC-based propellant. HC aerosol
propellants are usually mixtures of propane, butane, and isobutane.78 Their primary advantage lies in their
affordability; the price of HC propellants ranges from one-third to one-half that of HFCs. The main disadvantages of
HC aerosol propellants are flammability concerns and, because they are VOCs, their contribution to ground-level
ozone and smog. Despite these concerns, HC aerosol propellants already hold a sizable share of the market and
may be acceptable for additional applications.
• Capital Cost: Costs of converting filling facilities to accept HC propellants can range from $10,000 to
potentially as high as $1.2 million ($325,000 was assumed); the one-time cost varies based on the need
for investments in new equipment and the need to relocate to regions where the use of HCs is considered
safe (Nardini, 2002). To accommodate any flammable propellant, a company is required to build a storage
tank to house the product. This tank will need to be connected to the main facility through a plumbing
system (Cook, 2008; Tourigny, 2008). According to discussions with industry, the majority of companies
would already have fire insurance and other fire safety precautions intact; therefore, no significant
additional costs would be associated with housing a flammable chemical, and the increase in annual costs
would be zero (Cook, 2008; Tourigny, 2008).
• Annual O&M Costs: This analysis does not assume O&M costs.
• Annual Revenue: Given that HCs (estimated at $3.50/kg) are lower cost than HFC-134a (estimated at
$8/kg) and equivalent in cost to HFC-152a (estimated at $3.50/kg), the adoption of this abatement
measure is expected to result in an annual savings associated with gas purchases, ranging from no savings
to nearly $3 million.
NIK
NIK aerosol devices include finger/trigger pumps, powder formulations, sticks, rollers, brushes, nebulizers, and
bag-in-can/piston-can systems. These systems often prove to be a better and more cost-effective option than HFC-
propelled aerosols, particularly in areas where a unique HFC property is not specifically needed. Because all of the
HFC (either HFC-134a or HFC-152a) contained in the aerosol can is replaced with a device that does not use any
GHGs, the reduction efficiency of this option is 100%.
• Capital Cost: Significant variability exists in financial components of projects targeting NIK replacements
for HFC-containing aerosol products. This variability is attributable to the wide range of potential aerosol
and NIK product types. A one-time cost to make the conversion is estimated at $250,000.
• Annual O&M Costs: Annual costs of $500,000 are estimated to address higher material costs of the
particular sticks, rollers, and pumps being used (UNEP, 1999).
• Annual Revenue: An annual savings is expected, ranging from $2.0 million to $4.5 million, as a result of
eliminating the need for an HFC propellant.
HFO-1234ze
HFO-1234ze has potential application both as a propellant and as the active ingredient in aerosol dusters.
HFO-1234ze is nonflammable (at room temperature) and has physical properties that are very similar to both HFC-
134a and HFC-152a. Hence, it may be used as a "drop-in" replacement for HFC propellants (Tourigny, 2011). The
manufacturer of this chemical indicates that Europe and Japan have already begun to adopt HFO-1234ze, while
interest is also rising in the United States because of awareness of environmental sustainability (Malerba, 2011). A
number of dusters using HFO-1234ze are available today (Amazon, 2013; ITW Chemtronics, 2013; Miller
Stephenson, 2013; Stanley Supply and Services, 2013). In the absence of regulations, adoption in Europe and Japan
78 For calculation purposes, a GWP of 3.48 was used based on an average of the GWP of propane (GWP=3.3) and isobutane
(GWP=3.65).
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-145
-------
METHODOLOGY DOCUMENTATION
is expected to grow continuously at a moderate rate (reaching a maximum of 15% to 20% of today's HFC volume);
therefore, this option is expected to penetrate up to 15% of the non-MDI HFC-134a market and up to 20% of the
non-MDI HFC-152a market. In the United States, adoption of HFO-1234ze is expected to follow a similar path but
with a later start. In developing countries, no interest in HFO-1234ze is expected in the foreseeable future because
of inexpensive options that are the preferred solutions today.
• Capital Cost: For this analysis, a one-time cost of roughly $500,000 was assumed because of the need for
bulk storage. According to Tourigny (2011), although it is possible to use 1-ton cylinders and avoid the
costs of adding bulk storage, using ton cylinders is inefficient and adds to the unit cost of the HFO
material. Therefore, any facility using this material would almost certainly need to use bulk storage. This is
likely a conservative (high) one-time cost estimate, considering it is about the same capital cost
considered in the next section for a flammable propellant, whereas HFO-1234ze(E) is not flammable at
room temperatures.
• Annual O&M Costs: Because HFO-1234ze has a higher cost than the other HFCs (i.e., HFC-134a and HFC-
152a), a facility making the transition would incur a higher annual cost when adopting this propellant,
ranging from $4.3 million to $6.8 million.
• Annual Revenue: No annual savings are assumed.
HFC-134a to HFC-152a
This abatement measure examines replacing HFC-134a (with a GWP of 1,300) with HFC-152a (with a GWP of
140). HFC-134a was assumed to represent 58% of non-MDI aerosols; therefore, this abatement option is only
applicable to 58% of the non-MDI aerosol model facilities. HFC-134a is the primary nonflammable propellant in
certain industrial products. HFC-152a possesses only moderate flammability hazards and might, therefore, be
acceptable for some applications that use HFC-134a, but it may present problems for other applications.
• Capital Cost: Costs of converting filling facilities to accept HFC-152a may range from $500,000 to
$600,000, (Cook, 2008; Tourigny, 2008). To accommodate HFC-152a (or any flammable propellant), a
company is required to build a storage tank to house the product. This tank will need to be connected to
the main facility through a plumbing system (Cook, 2008).
• Annual O&M Costs: Aside from the costs associated with building a storage house, no other significant
expenses would be incurred. According to discussions with industry, the majority of companies would
already have fire insurance and other fire safety precautions intact; therefore, no significant additional
costs would be associated with housing a flammable chemical, and the increase in annual costs would be
zero (Cook, 2008; Tourigny, 2008).
• Annual Revenue: The lower cost of HFC-152a (compared with HFC-134a) results in an annual savings
associated with gas purchases, estimated at $2.6 million for a typical aerosol filling facility.
DPIs
DPIs are a viable abatement measure for most anti-asthma drugs, although they are not successful with all
patients or all drugs. Micronised dry powder, which contains the drug agent, is contained in the DPI, a
nonpressurized delivery system, and is inhaled and deposited in the lungs. DPIs are suitable only in patients who
are able to inhale robustly enough to transport the powder to the lungs. DPIs are not suitable for persons with
severe asthma or for young children. Unlike MDIs, powdered drug particles contained in DPIs tend to aggregate
and may cause problems in areas with hot and humid climates. Other issues that doctors and patients consider
when choosing a treatment device include the patient's manual dexterity, ability to adapt to a new device,
perception of the effectiveness of the medicine, and taste of any added ingredients. Ultimately, these and other
critical patient care issues must be assessed by the doctor and patient in choosing whether a DPI, MDI, or other
type of therapy is most appropriate (Price et al., 2004; UNEP, 2010). Where feasible, DPIs—which do not contain
GHGs—could be used in lieu of HFC-containing MDIs; hence, the reduction efficiency of this option is 100%.
5-146
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Capital Cost: No one-time costs are assumed.
• Annual O&M Costs: The annual cost associated with using DPIs was estimated to be approximately
$700,000 per metric ton of substance. This cost was based on €533,000 (in 1999 Euros) per metric ton of
substance (Enviros, 2000), which translates to an annual cost of $552,544 using the 1999 exchange rate of
$0.964629 Euros to 1 U.S. dollar. According to the source cited by Ecofys (2000), this annual cost incurred
by the industry takes into account the increase in the cost of DPI treatment, the cost to market the new
treatment, and the cost to retrain the patients in using the DPI (Enviros, 2000). It is unknown to what
extent this value includes capital and annual costs and savings.
• Annual Revenue: No cost savings are assumed. A DPI treatment of 200 doses costs, on average, around
$10 more than an MDI (Enviros, 2000).
Technical and Economic Characteristics Summary
The analysis developed a technical effectiveness parameter, defined as the percentage reductions achievable
by each technology/facility type combination. Market penetration rates vary over time as systems are upgraded in
the future. Table 5-51 summarizes the assumptions regarding technical applicability, market penetration, and
technical effectiveness of each option.
Table 5-51: Technical Effectiveness Summary—Aerosol Products
Technical
Market
Technical
Applicability
Penetration
Reduction
Effectiveness
Facility/Abatement Option
(2030)
Rate (2030)a
Efficiency
(2030)b
Consumer Aerosol Products—U.S. and Other Developed and EU
HFC-134a to HC
58%
20%
100%
10%
HFC-134a to NIK
58%
20%
100%
10%
HFC-134a to HFO-1234ze
58%
20%
100%
9%
HFC-134a to HFC-152a
58%
10%
91%
5%
HFC-152a to HC
42%
20%
95%
9%
HFC-152a to NIK
42%
40%
100%
19%
HFC-152a to HFO-1234ze
42%
20%
95%
8%
Consumer Aerosol Products—Developing
HFC-134a to HC
58%
20%
100%
10%
HFC-134a to NIK
58%
20%
100%
10%
HFC-134a to HFO-1234ze
58%
20%
100%
9%
HFC-134a to HFC-152a
58%
10%
91%
5%
HFC-152a to HC
42%
20%
95%
9%
HFC-152a to NIK
42%
40%
100%
19%
HFC-152a to HFO-1234ze
42%
20%
95%
8%
Pharmaceutical Aerosol Products (MDIs)
—U.S. and Other Developed and EU
DPIs
100%
20%
100%
20%
Pharmaceutical Aerosol Products (MDIs)
—Developing
DPIs
100%
20%
100%
20%
a Market penetration assumptions for this analysis vary over time, and the technical effectiveness values are based on the
cumulative market penetration rates assumed until that point.
b Technical effectiveness figures represent the percentage of baseline emissions from the relevant facility type that can be
abated in 2030.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-147
-------
METHODOLOGY DOCUMENTATION
Table 5-52 presents the engineering cost data for each mitigation option outlined above, including all cost
parameters necessary to calculate the break-even price. For more detailed costs information.
Table 5-52: Engineering Cost Data on a Facility Basis—Aerosol Products
Abatement Option
Project
Lifetime
(years)
Capital Cost
(2015 USD)
Annual
Revenue
(2015 USD)
Annual O&M
Costs (2015
USD)
Abatement
Amount
(tC02e)
HCs
Consumer aerosol/HFC-134a
10
$325,000
$2,800,000
—
807,125
Consumer aerosol/HFC-152a
10
$325,000
$1,000,000
—
66,623
NIK
Consumer aerosol/HFC-134a
10
$250,000
$4,100,000
$500,000
810,810
Consumer aerosol/HFC-152a
10
$250,000
$2,300,000
$500,000
70,308
HF0-1234ze
Consumer aerosol/HFC-134a
10
$500,000
$1,400,000
807,408
Consumer aerosol/HFC-152a
10
$500,000
$3,200,000
66,906
HFC-134a to HFC-152a
Consumer aerosol/HFC-134a
19
$500,000
51.800.000
740,502
DPIs
Pharmaceutical Aerosols (MDI)
10
5/00.000
1,430
Sector-Level Trends/Considerations
The treatment of international variability of mitigation technology costs is a significant area of uncertainty in
this analysis. The analysis is currently limited by the lack of detail on cost assumptions, which may not accurately
represent the transition costs regionally. Additionally, the cost assumptions for the transition to DPIs are based on
a study released in 2000, which may not reflect the latest technical and economic parameters. Finally, the general
methodology used here projects increasing use of HFC aerosols based on historical growth and expanding GDPs.
Some market sectors may not expand that quickly.
References
Amazon. 2013. Falcon DPSGRN Dust-Off ECO duster, 5 oz Canister (Case of 12).
http://www.amazon.com/industrial-scientific/dp/B009L9A2X6 cf
Cook, S., TechSpray. August 2008. Personal communication with Emily Herzog, ICF International.
Ecofys. 2000. Abatement of Emissions of Other Greenhouse Gases: Engineered Chemicals. Prepared for the
International Energy Agency Greenhouse Gas Research and Design Programme.
Enviros. March 2000. Study on the Use ofHFCsfor Metered Dose Inhalers in the European Union. Commissioned by
the International Pharmaceutical Aerosol Consortium (IPAC).
ITW Chemtronics. 2013. Typhoon Blast™ duster. http://www.chemtronics.com/products/product.asp?id=598 cf
Malerba, A., Honeywell. November 2011a. E-mail communication with Pamela Mathis, ICF International.
Miller-Stephenson. 2013. Aero-Duster® product information. http://www.miller-
stephenson.eom/assets/l/Store%20ltem/MS-222L.pdfE?
Nardini, G., May 2002. Personal communication with lliriana Mushkolaj, ICF Consulting.
5-148
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Price, D., E. Valovirta, and J. Fischer. 2004. The importance of preserving choice in inhalation therapy: The CFC
transition and beyond. Journal of Drug Assessment, 7, 45-61.
Stanley Supply & Services, 2013. Techspray Renew 1580-10S Eco-Duster, lOoz.
http://www.stanlevsupplvservices.eom/techsprav-renew-1580-10s-eco-dyster-10oz/p/477-260lj1
Tourigny, J., MicroCare Corporation. August 2008. Personal communication with Emily Herzog, ICF International.
Tourigny, J., MicroCare Corporation. November 2011. E-mail communication with Emily Herzog, ICF International.
United Nations Environment Programme. 1999.1998 Report of the Solvents, Coatings, and Adhesives Technical
Options Committee (STOC): 1998 Assessment. Nairobi, Kenya: UNEP Ozone Secretariat.
United Nations Environment Programme. 2010. Report of the UNEP Medical Technical Options Committee: 2010
Assessment. Nairobi, Kenya: UNEP Ozone Secretariat.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-149
-------
METHODOLOGY DOCUMENTATION
Fire Protection Mitigation Options Considered
The fire protection sector encompasses total flooding fire protection systems and portable (hand-held) fire
extinguishers and includes emissions of HFCs and PFCs.
The alternatives to HFCs/PFCs in total flooding applications can be categorized as in-kind gaseous agent
alternatives (i.e., CO2, inert gases, fluorinated ketones) and NIK alternatives (i.e., dispersed and condensed aerosol
extinguishing systems, water sprinklers, water mist, foam, or inert gas generators). Already, climate-friendly clean
agents and new NIK alternative technologies have been introduced to the market.
This analysis reviewed options to reduce emissions from the fire protection sector by using zero-GWP or low-
GWP extinguishing agents in lieu of HFCs/PFCs in new total flooding equipment. Specifically, this analysis assessed
alternative agents used in newly built total flooding systems to protect against Class A surface fire hazards and
newly built total flooding systems to protect against Class B fuel hazards in large (>3,000 m3) marine applications.
All costs are presented in 2010 dollars based on the Consumer Price Index (U.S. Department of Labor, 2011).
Facilities/emissions for which no abatement options were considered in this analysis include existing total
flooding systems (used to protect against all fire hazards) and all new and existing portable extinguishers. Existing
flooding systems were not assessed because alternative fire protection agents require larger space requirements,
rendering system retrofit costs highly dependent on the facility and possibly cost-prohibitive. Portable
extinguishers were not assessed because emissions from this source are small, and climate-friendly alternatives
were already assumed to be used widely in the baseline.
The analyzed facilities were assessed on a per-cubic meter of protected space basis, assuming an average
emission rate of 2% per year. Specifically, for Class A surface fire hazards, an average of 0.633 kg of HFC-227ea is
needed to protect each cubic meter of protected space; while 0.630 kg is required for large Class B fire hazards
(Wickham, 2003).
Three abatement options were considered for this analysis: in new Class A total flooding systems, replacement
with either FK-5-1-12 or inert gas, and in new Class B total flooding systems, replacement with water mist. Each of
these options is described in Table 5-53.
Table 5-53: Fire Protection Abatement Options
Abatement Option
Applicable System Types
Reduction Efficiency
FK-5-1-12
New Class A total flooding
100%
Inert gas
New Class A total flooding
100%
Water mist
New Class B total flooding
100%
FK-5-1-12 in New Class A Total Flooding Applications
FK-5-l-12-mmy2 (also known as l,l,l,2,2,4,5,5,5-nonafluoro-4-[trifluoromethyl]-3-pentanone, and commonly
referred to as FK-5-1-12) is a fluorinated ketone with an atmospheric lifetime of 5 days and a 100-year GWP of
approximately 1 (Kidde Fire Protection, 2011). This option examines use of FK-5-1-12 in total flooding systems. The
option is applicable in new Class A total flooding applications, replacing HFCs (primarily HFC-227ea). Class A total
flooding application end uses represent an estimated 95% of the total flooding sector; the additional adoption of
FK-5-1-12 was assumed to only occur when new systems are installed because replacing installed systems may be
cost prohibitive.
• Capital Cost: Capital costs of FK-5-1-12 systems in developed countries associated with installation and
equipment are estimated to be $9.40 more than conventional HFC systems per cubic meter of protected
space. Also, although the floor space requirements for this option are very similar to those of HFC
5-150
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
systems, there is a slight increase in the floor space needed to protect each cubic meter of space
(approximately 0.0005 sq. ft.) (Wickham, 2003). Assuming an average construction cost of approximately
$176 per square foot (R.S. Means, 2007), this translates into an incremental one-time construction cost of
$0.09 per cubic meter of protected space. Therefore, the total incremental one-time cost of this option is
$9.49 per cubic meter of protected space in developed countries. Capital costs were assumed to be 10%
greater in developing countries to account for higher tariffs.
• Annual O&M Costs: Because the additional space requirement associated with this option relative to
conventional HFC systems is so small (an average of 0.0005 square feet per cubic meter of protected
space [Wickham, 2003]), the additional annual costs associated with heating and cooling are also very
small—less than $0.01 annually per cubic meter of protected space. This cost was derived by multiplying
the additional space requirement (0.0005 sq. ft./cubic meter of protected space) by the average
electricity cost to heat/cool space, which was assumed to be roughly $7.60 per square foot in developed
countries (EIA, 2011; ICF, 2009). In developing countries, annual costs associated with electricity
consumption were assumed to be 66% greater. In addition, an annual cost of $0.09 per cubic meter of
protected space was assumed to be associated with annual emissions/agent replacement costs. This cost
is based on the assumption that approximately 0.74 kilograms of FK-5-1-12 agent is required to protect
every cubic meter of protected space, that 2% of this amount is leaked each year, and that FK-5-1-12 has
an incremental cost (relative to HFC-227ea) of approximately $2/kg (Werner, 2011).
• Annual Revenue: Because the agent cost of FK-5-1-12 is greater than that of HFC-227ea, no annual cost
savings were assumed for this option.
Inert Gas Systems in New Class A Total Flooding Applications
Inert gas systems extinguish fires using argon, nitrogen, or a blend of the two, sometimes incorporating CO2 as
a third component (UNEP, 2001). Inert gas systems provide an equivalent level of both fire protection and life
safety/health protection in most Class A (ordinary combustible) fire hazards, including electronics and
telecommunications applications. Limitations of the inert gas systems include a slower discharge time than that of
HFC systems—60 seconds or more compared with 10 to 15 seconds (Kucnerowicz-Polak, 2002)—and a larger
volume of agent needed than in HFC systems to extinguish fires. The weight-support structures and space needed
for additional steel cylinders of gas may prohibit the retrofit of many existing HFC systems, such as those on small
ships and in other applications where the system infrastructure is fixed.
This technology option was assumed to be applicable in new Class A application end uses, replacing HFCs
(primarily HFC-227ea). Class A total flooding application end uses represent an estimated 95% of the total flooding
sector; the additional adoption of inert gas systems was assumed to only occur when new systems are installed
because replacing installed systems may be cost prohibitive.
• Capital Cost: Inert gas systems were assumed to cost $7.13 more than conventional HFC-227ea systems in
developed countries, which were estimated to cost roughly $33 per cubic meter of protected space
(average across all space sizes) (Wickham, 2003). In addition, because inert gas systems require more
space to house gas cylinders than conventional HFC systems (an additional 0.023 sq. ft. per cubic meter of
protected space [Wickham, 2003]), in some cases there will be additional one-time costs to construct the
additional space for storage. Assuming a construction cost of about $176 per square foot (R.S. Means,
2007), this additional space requirement translates into an incremental one-time cost of $4.03 per cubic
meter of protected space. Therefore, the total incremental capital cost of this option is $11.16 per cubic
meter of protected space. Capital costs were assumed to be 10% greater in developing countries to
account for higher tariffs.
• Annual O&M Costs: Depending on the application, the space required to house additional gas cylinders
(an additional 0.023 sq. ft. per cubic meter of protected space) will need to be heated and cooled. Based
on average U.S. electricity costs of about $7.60 per square foot (ICF, 2009; EIA, 2011), the heating and
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-151
-------
METHODOLOGY DOCUMENTATION
cooling costs associated with this option result in an assumed annual cost of $0.17 per cubic meter of
protected space for developed countries. In developing countries, annual costs were assumed to be 66%
greater because of higher electricity costs.
• Annual Revenue: Because, on average, 0.633 kilogram of HFC-227ea is needed to protect 1 cubic meter of
space (Wickham, 2003) and assuming a release rate of 2% of the installed base, the emission of
approximately 13 grams of HFC-227ea is avoided each year per cubic meter of protected space. Based on
an average HFC-227ea cost of about $24 per kilogram (Werner, 2011), this translates into an annual
savings of $0.30 per cubic meter of protected space. These annual savings are assumed to be the same in
all regions.
Water Mist Systems in New Class B Total Flooding Applications
Water mist systems use relatively small droplet sprays under low, medium, or high pressure to extinguish
fires. These systems use specially designed nozzles to produce much smaller droplets than are produced by
traditional water-spray systems or conventional sprinklers, so they use less water to extinguish a fire (UNEP, 2001;
Wickham, 2002). However, some barriers have impeded broad use of water mist systems. First, these systems may
be cost prohibitive in small spaces and have not proven effective in extinguishing small fires in large-volume spaces
(>3,000 m3) (IMO, 2001; Wickham, 2002). Additionally, because there is a nonlinear relationship between the
volume of space and the amount of water mist needed to extinguish a given fire and because this relationship
(referred to as the "mechanism of extinguishment") is not well understood, applications of water mist systems
have been limited to those where fire test protocols have been developed, based on empirically tested system
performance. Other market barriers for this option include additional space requirements for system storage
compared with conventional HFC-227ea systems. Water mist systems can provide equivalent fire protection and
life safety/health protection for Class B fuel hazards, where low-temperature freezing is not a concern (EPA, 2004).
This technology option was assumed to be applicable in large (>3,000 m3), new Class B total flooding
application end uses, replacing HFCs (primarily HFC-227ea). This analysis assumed that systems designed to protect
against Class B fire hazards represent an estimated 5% of the total flooding sector; the additional adoption of
water mist systems was assumed to only occur when new systems are installed because replacing installed
systems may be cost prohibitive.
• Capital Cost: The capital cost of water mist systems used in marine systems to protect spaces of 3,000 m3
and larger in developed countries is estimated to be $4.82 more per cubic meter of protected space than
conventional HFC-227ea systems in large spaces (which are estimated to cost an average of about $30 per
cubic meter of protected space in spaces of these sizes) (Wickham, 2003).79 In addition, because water
mist systems require more space than conventional HFC systems (an additional 0.0472 sq. ft. per cubic
meter of protected space [Wickham, 2003]), one-time costs associated with constructing additional space
are also considered. Assuming a construction cost of roughly $176 per square foot (R.S. Means, 2007), this
additional space requirement translates into an incremental one-time cost of $8.32 per cubic meter of
protected space. Therefore, the total incremental capital cost of this option was assumed to be $13.14
per cubic meter of protected space in developed countries. Capital costs were assumed to be 10% greater
in developing countries to account for higher tariffs.
• Annual O&M Costs: Depending on the application, the space required to house additional gas cylinders
(an additional 0.0472 sq. ft. per cubic meter of protected space) will need to be heated and cooled. Based
on average U.S. electricity costs of roughly $7.60 per square foot (ICF, 2009; EIA, 2011), the heating and
cooling costs associated with this option result in an annual cost of $0.36 per cubic meter of protected
space in developed countries. In developing countries, annual costs were assumed to be 66% greater
because of higher electricity costs.
79 The cost of conventional HFC-227ea systems is less per cubic meter of protected space in large spaces than in small ones.
5-152
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Annual Revenue: Because an average of 0.63 kilogram of HFC-227ea is needed to protect 1 cubic meter of
space (for 3,000 m3 to 5,000 m3 spaces) (Wickham, 2003) and assuming a release rate of 2% of the
installed base, we assumed that the emission of approximately 13 grams of HFC-227ea is avoided each
year (i.e., 0.63 kilogram x 2%). Based on an average HFC-227ea cost of roughly $24 per kilogram, this
translates into an annual savings of $0.30 per cubic meter of protected space (Werner, 2011). These
annual savings were assumed to be the same in all regions.
Technical and Economic Characteristics Summary
The analysis developed a technical effectiveness parameter, defined as the percentage reductions achievable
by each technology/class/facility type combination. Estimating this parameter required making assumptions
regarding the distribution of emissions by class (A or B), in addition to process-specific estimates of technical
applicability and market penetration. Market penetration rates vary over time; the market penetration used in this
calculation is a modeled value that represents the assumed rate of penetration of the abatement option into fire
protection systems over time, market willingness to adopt the option, and the turnover rate of existing fire
protection systems. Table 5-54 summarizes these assumptions and presents technical effectiveness parameters
used in the MAC model.
Table 5-54: Technical Effectiveness Summary—Fire Protection
Technical
Market
Technical
Applicability
Penetration
Reduction
Effectiveness
Facility/ Abatement Option
(2030)
Rate (2030)a
Efficiency
(2030)b
New Class A total flooding—U.S. and
Other Developed
FK-5-1-12
30%
35%
100%
19%
Inert gas systems
30%
19%
100%
6%
New Class A total flooding—EU
FK-5-1-12
31%
35%
100%
20%
Inert gas systems
31%
19%
100%
6%
New Class A total flooding—Developing
FK-5-1-12
17%
35%
100%
7%
Inert gas systems
17%
10%
100%
2%
New Class B total flooding—U.S. and
Other Developed
Water mist systems
55%
3%
100%
1%
New Class B total flooding—EU
Water mist systems
57%
3%
100%
1%
New Class B total flooding—Developing
Water mist systems
19%
1%
100%
0%
a Market penetration assumptions for this analysis vary over time, and the technical effectiveness values were based on the
cumulative market penetration rates assumed until that point.
b Technical effectiveness figures represent the percentage of baseline emissions from the relevant facility type that can be
abated in 2030; figures do not account for indirect GHG impacts associated with increased electricity consumption for
heating/cooling of additional space, which is accounted for in the cost analysis.
The analysis was based on representative project costs for model facilities in the developing and developed
world, summarized in Table 5-55. We applied the costs to calculate the break-even prices for each appropriate
option for each country. The model estimates the mitigation potential based on the percentage of the total ODS
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-153
-------
METHODOLOGY DOCUMENTATION
substitutes baseline attributable to each representative facility and the technical effectiveness for each technology
in each facility.
Table 5-55: Engineering Cost Data on a Facility Basis—Fire Protection
Project
Annual
Annual
Abatement
Lifetime
Capital Cost
Revenue
0&M Costs
Amount
Abatement Option
(years)
(2015 USD)
(2015 USD)
(2015 USD)
(tC02e)
FK-5-1-12
New Class A total flooding—Developed
20
$9.49
—
$0.03
0.04
New Class A total flooding—Developing
20
$10.44
—
$0.03
0.04
Inert gas
New Class A total flooding—Developed
20
$11.16
$0.30
$0.17
0.04
New Class A total flooding—Developing
20
$12.28
$0.30
$0.28
0.04
Water mist
Large, new Class B total flooding—
Developed
20
$13.14
$0.30
$0.36
0.04
Large, new Class B total flooding—
Developing
20
$14.45
$0.30
$0.60
0.04
References
ICF International. March 2009. Opportunities for Combined Heat and Power in Data Center. Prepared for Oak Ridge
National Laboratory. Available online at
https://wwwl.eere.energy.gov/manufacturing/datacenters/pdfs/chp data centers.pdf
International Maritime Organization. November 30, 2001. Performance Testing and Approval Standards for Fire
Safety Systems: Fire Test Protocols for Fire-Extinguishing Systems. Submitted by Germany to the International
Maritime Organization Subcommittee on Fire-Protection, 46th session, Agenda item 12.
Kidde Fire Protection. 2011. Fire Suppression System Engineered for Use with 3M™ Novec™ 1230 Fire Protection
Fluid. Ref. 6351-5-05/11. Available online at http://www.kfp.co.uk/utcfs/ws-438/Assets/6351-
5 Novec System.pdfc?
Kucnerowicz-Polak, B. March 28, 2002. Halon sector update. Presented at the 19th Meeting of the Ozone
Operations Resource Group, The World Bank, Washington, DC.
R.S. Means Company, Inc. 2007. Means Square Foot Costs, 29th Annual Edition 2008. Kingston, MA: R.S. Means
Company, Inc.
United Nations Environment Programme. 2001. Standards and Codes of Practice to Eliminate Dependency on
Halons: Handbook of Good Practices in the Halon Sector. United Nations Publication ISBN 92-807-1988-1.
UNEP Division of Technology, Industry and Economics (DTE) under the OzonAction Programme under the
Multilateral Fund for the Implementation of the Montreal Protocol, in cooperation with The Fire Protection
Research Foundation.
U.S. Department of Labor, Bureau of Labor Statistics. 2011. Consumer price index. Available online at
http://www.bls.gov/cpi/
U.S. Environmental Protection Agency. 2004. Analysis of Costs to Abate International Ozone-Depleting Substance
Substitute Emissions. EPA #40-R-04-006. Washington, DC: EPA. Available online at
http://www.epa.gov/ozone/snap/emissions/downloads/ODSsubstituteemissions.pdf
5-154
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Werner, K., 3M. October 14, 2011. Personal communication with Pamela Mathis, ICF International.
Wickham, R. 2002. Status of Industry Efforts to Replace Halon Fire Extinguishing Agents. Stratham, NH: Wickham
Associates. Available online at http://www.epa.gov/ozone/snap/fire/status.pdf
Wickham, R. 2003. Review of the Use of Carbon Dioxide Total Flooding Fire Extinguishing Systems. Stratham, NH:
Wickham Associates. Available online at http://www.epa.gov/ozone/snap/fire/co2/ co2report2.pdf
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-155
-------
METHODOLOGY DOCUMENTATION
5.2.9 HFC-23 Emissions from HCFC-22 Production
Trifluoromethane (HFC-23) is generated and emitted as a by-product during the production of
chlorodifluoromethane (HCFC-22). HCFC-22 is used in emissive applications (primarily AC and refrigeration) and as
a feedstock for production of synthetic polymers. Because HCFC-22 depletes stratospheric ozone, its production
for dispersive uses is scheduled to be phased out under the Montreal Protocol. However, feedstock production, a
nondispersive use, is permitted to continue indefinitely.
5.2.9.1 HFC-23 from HCFC-22 Production Emission Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite
Emission Projections, for additional information). Activity data for HCFC-22 production included HCFC-22
production data from UNEP (2017a); production capacity information from IHS Markit (2001), Will et al. (2004 and
2008) and Jebens et al. (2014); field data on HFC-23 emissions from Montzka et al. (2010); HFC-23 emission
projections from the United States National Communications (2017); and the growth rate of feedstock HCFC-22
production from Montzka et al. (2010). In some cases, the emission estimates were reduced due to assumed
market penetrations of thermal abatement technologies.
The Tier 1 basic equation to estimate HFC-23 emissions from HCFC-22 production is as follows:
HFC — 23 emissions = HCFC — 22 Production * Emission Factor (5.17)
For HCFC-22 production, the driving factors in determining emissions are the use of process optimization
and/or thermal reduction to reduce HFC-23 emissions and the projected increase in HCFC-22 production for
feedstock use, which is not regulated under the Montreal Protocol.
Activity Data
Historical
Historical HCFC-22 production was estimated using an emission rate to estimate the HFC-23 emissions and
subtracting any emissions that were abated through technology. The data used to estimate historical HFC-23
emissions from HCFC-22 production included:
• Country-specific HCFC production data as reported to the UNEP Ozone Secretariat (UNEP, 2017a, 2017b);
• 2001, 2004, 2007, and 2013 country-specific production capacity information from the Chemical and
Economics Handbook (IHS Markit, 2001; Will et al., 2004; Will et al., 2008; Jebens et al., 2014); and
• Field data on HFC-23 emissions from HCFC-22 production (Montzka et al., 2010).
Estimating Production in Europe
Information on historical HCFC-22 production was used to estimate HFC-23 emissions. According to Jebens et
al. (2014), HCFC production from Greece, the Netherlands, and Spain is only HCFC-22 (based on plant capacities).
UNEP (2017a) reports total nonfeedstock HCFC production by country in ODP-weighted tons. As a result,
nonfeedstock HCFC-22 production for these countries was assumed to be the total reported for each country in
UNEP (2017a) after "unweighting" the production estimates by HCFC-22's ODP of 0.055. The ratio of nonfeedstock
production to feedstock production was then used to estimate nonfeedstock HCFC-22 production to total HCFC-22
production, without exceeding the production capacities reported in the Chemical and Economics Handbook (IHS
Markit, 2001; Will et al., 2004; Jebens et al., 2014). The ratio of nonfeedstock production to feedstock production,
5-156
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
as shown in Table 5-56, was estimated over the time series based on data for 1990 from EPA (2006) and data for
1996 and 2007 from Montzka et al. (2010) and by linearly interpolating the intervening years. The ratio of
nonfeedstock production to feedstock production from 2008 through 2015 was assumed to be equal to the 2007
estimate from Montzka et al. (2010).
Table 5-56: Portion of Total HCFC-22 Production That is Feedstock HCFC-22
Production for Al Countries
1990
1995
2001
2002
2003
2004
2005
2006
2007-2015
20%
26%
41%
44%
47%
50%
52%
55%
58%
This total was subtracted from western Europe production from Jebens et al. (2014) across the time series,
and the remaining HCFC-22 production for western Europe is allocated to France, Germany, Italy, and the U.K.
based on total HCFC-22 production capacity for each country as reported in Chemical and Economics Handbook
(IHS Markit, 2001; Will et al., 2008; Will et al., 2004; Jebens et al., 2014). For all European countries, it was
assumed that production from 1990 through 2003 could not exceed 2001-reported capacity, production in 2004
through 2006 could not exceed 2004-reported capacity, and production in 2007 through 2015 could not exceed
2013-reported capacity.
Estimating Production in the Rest of the World
Based on plant capacities from Will et al. (2004), HCFC production in Mexico, Argentina, Venezuela, and India
was also only HCFC-22. Again, UNEP (2017a)-reported HCFC production was assumed to be the total nonfeedstock
HCFC-22 production reported for each country by "unweighting" the production estimates by dividing the total
production by HCFC-22's ODP of 0.055.
For South Korea, 33% of total HCFC production capacity was HCFC-22 (Will et al., 2004, 2008; Jebens et al.,
2014). This percentage was applied across the UNEP-reported nonfeedstock HCFC production time series to
estimate nonfeedstock HCFC-22 production totals. The ratio of nonfeedstock production to feedstock production
was then used to estimate nonfeedstock HCFC-22 production to total HCFC-22 production. For North Korea, UNEP
(2017a) provided HCFC-22 production estimates for 2007 through 2015.
Jebens et al. (2014) reported China's HCFC-22 production for 2003 through 2013. HCFC-22 production was
back casted using the ratio of total HCFC-22 production reported in Jebens et al. (2014) to UNEP-reported
nonfeedstock HCFC production for 2003. This ratio was applied across the UNEP-reported time series for 1990 to
2002 and 2014 to estimate China's HCFC-22 production for those years. UNEP (2017b) reported HCFC-22
production in China for 2015. The ratio of nonfeedstock production to feedstock production across the time series
for China and other Non-Al countries and Russia is shown in Table 5-57.
Table 5-57: Portion of Total HCFC-22 Production That is Feedstock HCFC-22 Production for Non-A1
Countries
Countries
1990
1995
2000
2001
2002
2003
2004
2005
2006
2007-2015
Non-Al
20%
31%
29%
28%
27%
26%
26%
25%
24%
23%
China
44%
55%
53%
52%
51%
50%
49%
48%
47%
46%
Projected
HFC-23 emission projections were developed for Al countries including Germany, Japan, the Netherlands,
Russia, Spain, and the United States. For the United States, national communications projections of emissions were
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-157
-------
METHODOLOGY DOCUMENTATION
used for 2015 through 2020 (UNFCCC, 2017a). Emission trends were used to project HFC-23 emissions for the
remainder of the time series (2025 through 2050).
For all other A1 countries, the dispersive production and feedstock production portion of emissions were
projected separately to account for the decline in the production for dispersive purposes because of the phase-out
requirements of the Montreal Protocol. The following assumptions for these countries were applied to estimate
dispersive production:
• UNFCCC reported zero emissions of HFC-23 for Australia and Canada, beginning in 2000 and 1995,
respectively (UNFCCC, 2017b). No further data were available on Australia, so the EPA assumed Australia
will not produce HCFC-22 in the future. Jebens et al. (2014) reports that Canada only produces one HCFC,
HCFC-123, so it was assumed that Canada will not produce HCFC-22 in the future.
• For Greece and the U.K., HCFC-22 production was assumed to end due to plant closures in 2006 and 2008,
respectively, and therefore emissions were set equal to zero. For Spain, HCFC production has not been
reported to UNEP since 2011, so it was assumed that nonfeedstock production of HCFC-22 ended.
• For developed countries other than Australia, Canada, the U.K., and Spain, emissions from nonfeedstock
production were assumed to decrease linearly from 2015 so that no emissions resulted from HCFC-22
nonfeedstock production by the 2020 phase-out date under the Montreal Protocol.
To project the feedstock production portion of HFC-23 emissions for developed countries, a 5% global growth
rate of feedstock HCFC-22 production as reported in Montzka et al. (2010) was applied for all countries.
HFC-23 emission projections were developed for Non-Al countries including Argentina, China, India, Mexico,
North Korea, South Korea, and Venezuela. To do so, Non-Al aggregate HCFC-22 production was projected for both
dispersive and feedstock production.
• HCFC-22 dispersive production for developing countries was projected using a 2010 HCFC-22 production
estimate of 395,000 metric tons, as provided by Miller and Kuijpers (2011), and a baseline estimate of
383,000 metric tons and the percentage reductions from that baseline as prescribed by the accelerated
phase-out schedule of the Montreal Protocol.
• HCFC-22 feedstock production was projected for developing countries based on the 5% global growth
rate of feedstock HCFC-22 production, as reported in Montzka et al. (2010).
Production was then disaggregated by country using the percentage of each country's contribution to 2015 Non-
Al total HCFC-22 production.
Each country's HCFC-22 projected production was then apportioned into four different model facilities for
each developing country. The model facilities for which HCFC-22 production projections were apportioned are as
follows:
• Residual emissions: These facilities already have abatement controls in place. Facilities that have CDM
projects (mitigation projects funded by developed countries under the Kyoto Protocol) in developing
countries are considered "residual emission facilities." In addition, China implemented 15 destruction
facilities not covered by CDM that are considered "residual emission facilities" (UNEP, 2017b).
• Non-CDM and uncontrolled facility: Non-CDM facilities are existing facilities that are uncontrolled. These
facilities exist in China, Venezuela, Mexico, and North Korea (UNEP, 2017b).
• New uncontrolled facility: New facilities were assumed to be uncontrolled when built. It was assumed
that a new facility enters the market once projected production exceeds current capacity. In other words,
the percentage of emissions from new facilities is zero until projected production exceeds capacity.
• Post-CDM facility: These are facilities that were previously controlled under the CDM ("residual emission
model facility"). It was assumed that the incineration device installed (via a CDM project) is not kept in
5-158
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
operation. In 2013, the EU Emissions Trading System, New Zealand, and Australia imposed a ban on the
use of Certified Emission Reduction credits from HFC-23 destruction, which significantly lowered the value
of credits obtained from HFC-23 abatement projects. Signed in 2016, the Paris Agreement universally
rejects HFC-23 projects for offset credits. Several facilities with CDM projects are no longer destroying
HFC-23 emissions without the generation of these credits. Post-CDM facilities exist in Argentina, Mexico,
India, and South Korea.
Emission Factors
Historical and Projected
To estimate emissions of HFC-23, the estimated HCFC-22 production levels were multiplied by emission rates
(i.e., tons of HFC-23 emitted per ton of HCFC-22 produced). The emission rate for A1 countries was assumed to be
2% across the entire time series (Montzka et al., 2010).
To address the varying use of abatement technologies by facilities, HFC-23 emissions for developing countries
were then projected using two HFC-23/HCFC-22 co-production ratios to develop estimates. The HFC-23/HCFC-22
co-production ratio of 2.9%, representative of the CDM's annual mean ratio for 2009, was used to estimate
emissions (Miller et al., 2010). For emissions associated with the "residual" model facility, the HFC-23/HCFC-22 co-
production ratio was modified by 95% to account for a reduction in efficiency associated with the incinerator.80
Emission Reductions in Baseline Scenario
In some cases, the emission estimates were reduced due to assumed market penetrations of thermal
abatement technologies. The emission rate for Non-Al countries and Russia was assumed to be 3% from 1990
through 2005 (EPA, 2006) and 2.4% from 2006 through 2015 (Miller et al., 2010). The decreased emission rate
takes into account any HFC-23 emission offsets from CDM projects in these countries and the Joint
Implementation project at Russia's HCFC-22 plant in Perm.
To reflect the adoption of thermal oxidation technology between 1995 and 2015, we reduced current
emission rates relative to historical emission rates in some regions. The following market penetrations were
incorporated into the analysis:
• In 2000, the baseline market penetration of thermal oxidation was estimated to be 100% in Germany,
France, and Italy and 75% in the U.K. (Harnisch and Hendriks, 2000). Except for the U.K., these levels were
assumed to be maintained through 2050.
• In 2005, the baseline market penetration of thermal oxidation in the U.K. was estimated to be 87.5%. This
was intended to reflect the 2005 commissioning of a thermal oxidizer at the one U.K. plant that had not
had one previously (Campbell, 2006). For 2006 through 2008, the level of baseline market penetration in
the U.K. was estimated to be 100%. No emissions were estimated for the U.K. after 2008 as a result of two
HCFC-22 plants closings in 2008 (MacCarthy et al., 2010).
• In 2005, the baseline market penetration of thermal oxidation in the Netherlands was assumed to be
100%.
Uncertainty
In developing these emission estimates, the EPA made use of multiple international datasets, country-specific
information on abatement levels (where available), and the 2006IPCC Guidelines on estimating emissions from
this source. Nevertheless, uncertainties exist in both the activity data and the emission rates used to generate
these emission estimates. Although the EPA used four separate sources to estimate country-by-country production
of HCFC-22 (UNEP-reported, country-specific HCFC production, country-by-country production capacities from the
80 The assumption of 95% destruction efficiency is conservative. Although reduction efficiency is closer to 99.99% for
incineration, a lower reduction efficiency takes into account startups, shutdowns (e.g., for cleaning), and malfunctions.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-159
-------
METHODOLOGY DOCUMENTATION
Chemical and Economics Handbook; and field data on HFC-23 emissions from HCFC-22 production), none of these
sources is comprehensive. Specifically, none provides country-by-country production of HCFC-22 for all countries.
As a result, the EPA used different ratios to estimate total HCFC-22 production over time for several countries (e.g.,
percentage of total HCFC production capacity that is HCFC-22 for South Korea). These ratios may add uncertainty
to the extent that the ratios fluctuate over time.
Future emission and abatement levels are particularly uncertain. Future policies (e.g., under the Montreal
Protocol) could affect total production of HCFC-22 and therefore emissions of HFC-23. For example, the Kigali
Amendment to the Montreal Protocol mandates all HCFC-22-producing facilities to collect and destroy HFC-23 by-
products from 2020 to the extent practicable. Because most, if not all, HCFC-22 production plants have access to
existing destruction facilities, they could restart the equipment that was used to destroy HFC-23 previously if the
equipment is not currently in use. Changing emission rates may also have a significant impact on emissions. There
is a significant probability that many of these emissions will be averted. In this case, HFC-23 emissions would be
lower than projected in this analysis.
5.2.9.2 Mitigation Options Considered for HCFC-22 Production
One abatement option, thermal oxidation, was examined in this analysis of the HCFC-22 production sector.
Thermal oxidation, the process of oxidizing HFC-23 to CO2, hydrogen fluoride (HF), and water, is a demonstrated
technology for the destruction of halogenated organic compounds. For example, destruction of more than 99% of
HFC-23 can be achieved under optimal conditions (i.e., a relatively concentrated HFC-23 vent stream with a low
flow rate) (Rand et al., 1999). In practice, actual reductions will be determined by the fraction of production time
that the destruction device is operating. Units may experience some downtime because of the extreme corrosivity
of HF and the high temperatures required for complete destruction. This analysis assumed a reduction efficiency of
95%.81 The destruction of HFC-23 by thermal oxidation was assumed to be 100% applicable to all facilities, and the
analysis assumed a project lifetime of 20 years. Cost estimates for installing and operating a thermal oxidizer are
summarized below:82
• Capital Cost: The capital cost for a thermal oxidation system is estimated to be approximately $5.2 million
to install at an existing plant and $4.0 million to install during construction of a new plant (Irrgang, 2018).
The capital cost for restarting an existing incinerator is estimated to be approximately $400,000 (Irrgang,
2018 and UNEP, 2017b).
• Annual O&M Costs: O&M costs are estimated at $200,000 (Irrgang, 2018).
• Annual Revenue: No annual savings or revenues are associated with the thermal oxidation abatement
option.83
• Technical Lifetime: 20 years
• Reduction Efficiency: Thermal oxidation technology was assumed to be 95% efficient in abating HFC-23
emissions.
81 A representative of a company that manufactures thermal oxidation systems stated that new systems are built using
materials that better resist corrosion than the materials used in older systems. The representative indicated that such new
systems were likely to experience very limited downtime, considerably less than 5% (Rost, 2006).
82 Estimates developed for this analysis are based on communication with industry and best available industry assessments;
actual costs of some systems could differ from these estimates.
83 It should be noted that annual revenue is generated for participants of CDM projects; however, CDM projects were not
assumed to cover further abatement of emissions in this analysis.
5-160
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.2.9.3 Model Facilities
This analysis developed four potential model facilities to model the mitigation potential in this sector. These
facilities included the following:
• Residual emissions: These facilities have abatement controls in place already. All facilities in the A1
countries and facilities that have CDM projects (mitigation projects funded by developed countries under
the Kyoto Protocol) in the developing countries are considered "residual emission facilities."
• Non-CDM and uncontrolled facility: Non-CDM facilities are existing facilities that are uncontrolled. These
facilities exist in China, South Korea, and Venezuela.
• New uncontrolled facility: New facilities were assumed to be uncontrolled when built. It was assumed
that a new facility enters the market once projected production exceeds current capacity. In other words,
the percentage of emissions from new facilities is 0% until projected production exceeds capacity. It was
assumed that new facilities will only be built in Non-Al countries.
• Post-CDM facility: Similar to the "less mitigation scenario" of Miller and Kuijpers (2011), this analysis
assumed that the 12 CDM projects that opted for a 7-year crediting period (in China, South Korea, Mexico,
and Argentina) are not renewed after their first terms (note the remaining seven facilities opted for a one-
time fixed crediting period that cannot exceed 10 years). Under this assumption, by 2020, all facilities
previously controlled via CDM ("residual emission model facility") are considered a "post-CDM" facility. It
was assumed that the incineration device installed (via a CDM project) will not be kept in operation once
the CDM crediting period is over. This analysis costs out mitigation from these facilities differently than a
new uncontrolled facility by considering capital costs associated with restarting the incinerator.
5.2.9.4 Technical and Economic Characteristics Summary
The analysis developed a technical effectiveness parameter, defined as the percentage reductions achievable
by each technology/facility type combination. Market penetration rates vary over time as systems are upgraded in
the future. Table 5-58 summarizes the assumptions regarding technical applicability, market penetration, and
technical effectiveness of thermal oxidation for each facility type.
Table 5-58: Technical Effectiveness Summary—HCFC-22 Production
Model Facility Type
Technical
Applicability
Market
Penetration
Rate
Reduction
Efficiency
Technical
Effectiveness
(2030)
Non-CDM and uncontrolled facility
100%
100%
95%
95%
New uncontrolled facility
100%
100%
95%
95%
Post-CDM facility
100%
100%
95%
95%
5.2.9.5 Sector-Level Trends/Considerations
This analysis evaluates how thermal oxidation can be applied to facilities that are current CDM participants
after the crediting period is over and the CDM project is completed. Because an incineration device is already
installed due to the CDM project, the costs to adopt the abatement measure relate only to its annual operation.
Facilities participating in CDM were assumed to have completed their crediting periods by 2020.
This analysis also assumed that new facilities will enter the market to meet future global demand of HCFC-22.
New facilities were assumed to enter the market once projected production for a Non-Al country exceeds current
plant capacities. According to industry, the costs of installing thermal oxidation systems in new plants are generally
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-161
-------
METHODOLOGY DOCUMENTATION
less expensive than the cost of installation at existing plants. This analysis used a capital cost for new facilities that
is approximately 23% less than the cost of installation at existing facilities (Irrgang, 2018).
5.2.9.6 References
Campbell, N., Arkema. 2006. Personal communication with Deborah Ottinger Schaefer of EPA, April 24, 2006.
Harnisch, J. and C. Hendricks. April 2000. Economic Evaluation of Emission Reductions ofHFCs, PFCs and SF6 in
Europe. Contribution to the study Economic Evaluation of Sectoral Emission Reduction Objectives for Climate
Change. Commission of the European Union, Directorate General Environment.
IHS Markit. 2001. CEH Marketing Research Report: Fluorocarbons. Chemical and Economics Handbook.
Irrgang, G., Selas Linde. June 11, 2018. Personal communication with Rebecca Ferenchiak, ICF. Information
provided by Mr. Rost to Ms. Ottinger in 2006 confirmed by Mr. Irrgang and adjusted for inflation.
Jebens, A., T. Kalin, Y. Yamaguchi, and Y. Zhang. 2014. CEH Marketing Research Report: Fluorocarbons. Chemical
Economics Handbook—IHS Chemical.
MacCarthy, J., J. Thomas, S. Choudrie, N. Passant, G. Thistlethwaite, T. Murrells, J. Watterson, L. Cardenas, and A.
Thomson. 2010. UK Greenhouse Gas Inventory, 1990 to 2008: Annual Report for Submission under the
Framework Convention on Climate Change. AEA: Oxfordshire, UK.
Miller, B.R., M. Rigby, L.J.M. Kuijpers, P.B. Krummel, LP. Steele, M. Leiste, P.J. Fraser, A. McCulloch, C. Harth, P.
Salameh, J. Miihle, R.F. Weiss, R.G. Prinn, R.H.J. Wang, S. O'Doherty, B.R. Greally, and P.G. Simmonds. 2010.
HFC-23 (CHF3) emission trend response to HCFC-22 (CHCIF2) production and recent HFC-23 emission
abatement measures. Atmospheric Chemistry and Physics, 10, 7875-7890.
Miller, B.R. and L.J.M. Kuijpers. 2011. Projecting Future HFC-23 emissions. Atmospheric Chemistry and Physics,
Discussion, 11, 23081-23102.
Montzka, S.A., L. Kuijpers, M.O. Battle, M.A.K. Verhulst, E.S. Saltzman, and D.W. Fahey. 2010. Recent increases in
global HFC-23 emissions. Geophysical Research Letters, 37, L02808.
Rand, S., D. Ottinger, and M. Branscome. May 26-28,1999. Opportunities for the Reduction of HFC-23 Emissions
from the Production of HCFC-22. IPCC/TEAP Joint Expert Meeting. Petten, Netherlands.
Rost, M., T-thermal. April 24, 2006. Personal communication with Debora Ottinger, EPA.
United Nations Environment Programme. 2017a. Data Access Centre. HCFC Production. Available online at
http://ozone.unep.org/Data Reporting/Data Access/J
United Nations Environment Programme. 2017b. Executive Committee of the Multilateral Fund for the
Implementation of the Montreal Protocol: Seventy-eight Meeting-Key Aspects Related to HFC-23 By-Product
Control Technologies. Available online at www.multilateralfund.org/78/English/l/7809.docx C?
U.S. Environmental Protection Agency. 2006. Global Anthropogenic Non-C02 Greenhouse Gas Emissions: 1990-
2020. Office of Atmospheric Programs: Climate Change Division. Available online at http://www.epa.gov/
climatechange/economics/international.html
National Communications. 2010. Fifth National Communications - Annex I. Available online at:
https://unfccc.int/process-and-meetings/transparencv-and-reporting/reporting-and-review-under-the-
convention/national-communications-and-biennial-reports-annex-i-parties/national-communication-
submissions/fifth-national-communications-annex-i ef
Will, R., A. Kishi, and S. Schlag. 2004. CEH Marketing Research Report: Fluorocarbons. Chemical Economics
Handbook. SRI Consulting.
Will, R. K. and H. Mori. 2008. CEH Marketing Research Report: Fluorocarbons. Chemical Economics Handbook. SRI
Consulting.
5-162
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.2.10 Other Industrial Processes
• Other industrial processes sources (Cm, N2O), include
- chemical production (CH4)
- iron and steel production (CH4)
- metal production (Cm, N2O)
- mineral products (CH4)
- petrochemical production (CH4)
- silicon carbide production (CH4)
- solvent and other product use (N2O)
5.2.10.1 Other Industrial Processes Projections Methodology
The source category solely comprises countries that report data to the UNFCCC database. The EPA did not
perform Tier 1 calculations for other industrial process sources, which include the production of mineral products,
ammonia, silicon carbide, calcium carbide, iron and steel, ferroalloys, and solvent and other product use. The EPA
obtained historical values for 1990 through 2012 and held 2015 through 2050 values constant at 2012 levels for
each country.
5.2.10.2 Other Industrial Processes Mitigation Methodology
The EPA has not estimated mitigation potential from other industrial processes because of a lack of available
data on mitigation options.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-163
-------
METHODOLOGY DOCUMENTATION
5.3 Agricultural Sector
This section presents the methodology for estimating global CFU and N2O emissions and mitigation for the
following agricultural sources:
• enteric fermentation (CH4)
• manure management (Cm, N2O)
• cropland soils (N2O)
• rice cultivation (CH4)
The baseline projections also include other agricultural sources, including
• field burning of agricultural residues (Cm, N2O) and
• prescribed burning of savannas (Cm, N2O).
Mitigation options are only developed and applied to the first four categories, which account for most agricultural
sector non-CC>2 GHG emissions. Additional details are described below.
5.3.1 Livestock Management
Livestock operations generate CH4 and N2O emissions. The GHG emissions mainly come from two sources:
enteric fermentation and manure management. Enteric fermentation refers to a process whereby microbes in an
animal's digestive system break down cellulose, and CFU is produced as a by-product and can be exhaled by the
animal. Domesticated ruminants such as cattle, buffalo, sheep, goats, and camels account for the majority of
enteric fermentation CFU emissions. Other domesticated nonruminants such as swine and horses also produce CH4
as a by-product of enteric fermentation, but emissions per animal species vary significantly.
Global livestock inventories typically include cattle, buffalo, sheep, goats, camels, horses, mules, asses, deer,
alpaca, poultry, and swine. Our mitigation analysis, however, is restricted to the animal types for which mitigation
(and mitigation estimates) is most feasible, including enteric fermentation and manure management for cattle and
manure management for swine. Enteric fermentation calculations are described first. In the next section, we
describe the methodology for calculating baseline emissions of Cl-Ufrom all animal types. We then describe the
methodology for calculating mitigation potential for cattle.
5.3.1.1 Enteric Fermentation Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category when available. For those countries with country-reported emission estimates, emission projections were
estimated from the most recent country-reported data through 2050 using growth rates calculated by the Tier 1
methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission estimates
were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite Emission
Projections, for additional information). Activity data for enteric fermentation included animal population data
from FAO (2016) and livestock product growth rates from IFPRI's IMPACT model (2016).
The Tier 1 basic equation to estimate CH4 emissions from enteric fermentation from each individual type of
animal is as follows:
Emission Factor (kg/head/yr) x Animal Population (head) /(106 kg/Gg) = Emissions (Gg/yr) (5.18)
5-164
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
The primary driver for determining Cm emissions from enteric fermentation was animal population. It was
assumed that the animal characteristics on which the default emission factors are based do not change
significantly over time.84 Emissions from enteric fermentation were broken out by livestock type.
Activity Data
Historical
• Animal population data for years 1990,1995, 2000, and 2005 through 2014 were obtained from FAO
(2016). Populations of nondairy cattle were calculated by subtracting FAO dairy cattle populations from
FAO total cattle populations.85
• If reported emissions were available only for a portion of the time series, emissions were interpolated
using the available data in conjunction with the growth rate associated with the estimated Tier 1
emissions calculated for the country.
• The FAO population data were further modified in instances where country data were aggregated for part
of the time series. For example, in 1990, animal population data were not available for certain countries
that were formed after the dissolution of the FSU. Therefore, for each region, the percentage contribution
of each country to its regional total was determined using animal population data for subsequent years
with available data. These percentages were then applied to fill in gaps in animal populations for these
countries.
Projected
• Emissions from 2015 through 2050 were projected based on livestock product growth rates developed by
the IFPRI's IMPACT model (Robinson et al., 2015; Sulser et al., 2015).86 The IMPACT model projects growth
rates by country for the demand of beef, pork, lamb, and milk for the years 2015 through 2050 in 5-year
increments. These estimates were used to proxy average annual growth rates for the livestock species,
nondairy cattle, swine, sheep, and dairy cattle, respectively. For the remaining livestock types, the
average population growth rates from 2005 through 2014 in the FAO data were applied.87
• The growth rates described above were applied to the 2014 FAO animal populations to calculate
projected populations for 2015, 2020, 2025, 2030, 2035, 2040, 2045, and 2050 for each livestock species.
Emission Factors
Historical and Projected
• Tier 1 default emission factors from the 2006IPCC Guidelines were used in the calculated emissions (IPCC,
2006).
84These projections do not take into account changes in CH4 emissions due to variations in feed composition, feed intake, or
improved animal genetics, all of which could increase or decrease CH4 emissions compared to the Tier 1 emission factors (Rojas-
Downing et al., 2017).
85 FAO animal population data are based on total meat production (from both commercial and farm slaughter) and milk
production for a given year and do not take into account fluctuations in animal populations due to seasonal births or slaughter.
FAO data on livestock numbers are intended to cover all domestic animals irrespective of their age and the place or purpose of
their breeding and, as such, do not distinguish between slaughtered dairy or beef cattle.
86 The IMPACT model incorporates supply and demand parameters to determine the estimated growth rates. These parameters
include the feed mix applied according to relative price movements, international trade, national income, population, and
urban growth rates, as well as anticipated changes in these rates over time.
87 Basing livestock population growth on the 2005 through 2014 historical trend led to unrealistically high growth rates in some
countries that have experienced large livestock increases in recent years. In countries where the growth between 2014 and
2050 was greater than 200%, the trend was adjusted to draw on a longer historical period. When possible, the period used was
1990 through 2014; however, in some cases, a shorter period was necessary to keep growth as close as possible to the range
considered reasonable (i.e., 200% or less).
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-165
-------
METHODOLOGY DOCUMENTATION
• For buffalo, sheep, goats, camels, horses, mules and asses, deer, alpacas, and swine, the appropriate
enteric fermentation emission factors for either "developed" or "developing" countries were used. For
dairy and nondairy cattle, enteric fermentation emission factors for world regions were used with factors
assigned to countries based on the region in which they are located.
Emission Reductions in Baseline Scenario
The methodology used for this source category does not explicitly model any emission reductions; however,
emission reductions are included to the extent they are reflected in country-reported data.
Uncertainty
The greatest uncertainty of emissions from enteric fermentation is associated with the use of default IPCC
emission factors because of the lack of information on country-specific animal diets. Emission estimates for
countries with a variety of animal diets could be inaccurate, particularly when projecting emissions, because there
is a lack of information on potential changes in the quality, quantity, and type of feed that could affect projected
emissions in future years. Additionally, the emission projections do not take into account improved animal genetics
that could reduce emissions in enteric fermentation (Rojas-Downing et al., 2017). There are also uncertainties
associated with the animal population data. FAO data are derived from country, self-reported amounts of animals
slaughtered within national boundaries or milk produced (or estimated values if countries do not self-report).
Animal population data are based on total annual meat (from both commercial and farm slaughter) and milk
production for a given country and do not take into account fluctuations in animal populations due to seasonal
births or slaughter (FAO, 2016). Additionally, FAO data cover all domestic animals irrespective of their age and the
place or purpose of their breeding and, as such, do not distinguish between slaughtered dairy or beef cattle, which
have different emission factors. Finally, the impacts of either historical population trends or world markets and
consumption patterns on national livestock production patterns are often difficult to predict, further increasing the
uncertainty of projected emissions from this source.
5.3.1.2 Mitigation Options Considered for Enteric Fermentation
This section characterizes the mitigation technologies that can be applied to reduce enteric Cm emissions. A
significant number of livestock GHG mitigation measures can be identified in the literature (e.g., Hristov et al.,
2013; Archibeque et al., 2012; UNFCCC, 2008a; Whittle et al., 2013). However, developing consistent and region-
specific cost estimates for emerging mitigation measures or options that are not widely adopted is a challenging
task. The measure performance and cost data are scarce and often reflect anecdotal experience reported in a
specific country, region, or livestock production system.88
Based on the availability and quality of mitigation measure cost and emission reduction efficiency information,
this analysis evaluated six mitigation options for enteric fermentation Cm emissions (Table 5-59). Each technology
is briefly characterized followed by a discussion of the abatement measures' implementation costs, potential
benefits, and system design assumptions used in the MAC analysis. Many of the currently available enteric
fermentation mitigation options work indirectly by increasing animal growth rates and reducing time-to-finish (or
increasing milk production for dairy cows). The potential GHG mitigation estimated here depends on the
assumption that total production of meat or milk remains the same as in the baseline. Simply put, these strategies
work because increased productivity means fewer animals are required to produce the same amount of meat or
In addition, although there are potential opportunities for mitigation from dietary shifts relative to the baseline (Popp et al.,
2010), our assessment focuses on technical mitigation options within the agricultural sector rather than structural or demand
adjustments (see Frank et al. [2018] for a study that incorporates all three within a market modeling framework).
5-166 Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
milk, and fewer animals mean reduced GHGs (e.g., there is a reduction in GHG intensity for production of livestock
products).89
• Improved Feed Conversion: Improved feed conversion encompasses a number of management practices
that would improve the proportion of feed energy converted to final products. The practices include
increased amount of grain fed to livestock and inclusion of dietary additives. This option is more effective
in reducing emissions in regions where baseline feed is of relatively low quality.
• Antibiotics: Antibiotics (e.g., monensin) may be fed to cattle to promote increased weight gain and
reduce feed intake per metric ton of meat produced.
• Bovine Somatotropin (bST): bST may be administered to dairy cattle to increase milk production. Because
of opposition to the use of growth hormones like bST in many countries, this option was only applied in a
subset of countries where the use of bST is currently expected to be legal and feasible.
• Propionate Precursors: Propionate precursors (malate, fumarate) may be administered to animals daily.
Hydrogen produced in the rumen through fermentation can react to produce either Cm or propionate. By
adding propionate precursors to animal feed, more hydrogen is used to produce propionate and less Cm
is produced.
• Antimethanogen: Antimethanogen is a vaccine that can be administered to animals to suppress Cm
production in the rumen. The vaccine is currently in infancy of development with limited information on
emission reduction efficiency, long-term mitigation effects, and animal health impacts.
• Intensive Grazing: Intensive grazing means improving nutrition through more intensive pasture
management and cattle rotations to allow for regrowth while decreasing reliance on prepared rations.
5.3.1.3 Technical and Economic Characteristics of Options
Table 5-59 provides descriptions of the technical and economic characteristics of the abatement measures for
enteric fermentation included in this analysis. For all enteric mitigation options considered, we assumed no initial
capital costs. Recurring annual costs range from -$180 to $300 per head of livestock. Negative costs indicate cost
savings in this table such as reductions in feeding operations costs.
Table 5-59: Abatement Measures for Enteric Fermentation CH4
Abatement
Option
Total
Installed
Capital
Cost (2010
USD)
Annual O&M
Cost (2010
USD)
Capital
Lifetime
(years)
Reduction
Efficiency
(change in
emissions per
head)
Benefits
(changes in
livestock or
energy
revenue)
Technical Applicability
Improved feed
conversion
0
$25-$295 per
head
NA
CH4: -39.4% to
+39.6%
0-79%
increase in
animal yield
Beef and dairy cattle in areas
with low baseline livestock
growth and milk production rates
Antibiotics
0
$4-$9 per head
NA
CH4: -0.4% to
-6%
5% increase in
animal yield
Beef cattle in urban and
intensively managed livestock
production systems (LPS)
bST
0
$123-$300 per
head
NA
CH4: -0.2% to
+10.3%
12.5%
increase in
animal yield
Dairy cattle in urban and
intensively managed LPS within
countries where bST is currently
expected to be legal and feasible
(continued)
89 There are a variety of additional mitigation options under development that may be incorporated in future analyses as
sufficient data become available (e.g., wide array of different changes to livestock feed, additional supplements, breeding
programs).
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-167
-------
METHODOLOGY DOCUMENTATION
Table 5-59: Abatement Measures for Enteric Fermentation CH4 (continued)
Total
Reduction
Benefits
Installed
Efficiency
(changes in
Capital
Annual O&M
Capital
(change in
livestock or
Abatement
Cost (2010
Cost (2010
Lifetime
emissions per
energy
Option
USD)
USD)
(years)
head)
revenue)
Technical Applicability
Propionate
0
$40-$120 per
NA
CH4: -10% beef
5% increase in Beef cattle, sheep, and dairy
precursors
head
cattle and sheep;
animal yield
animals in urban and intensively
-25% dairy
managed LPS
animals
Anti-
0
$9-$33 per
NA
CH4: -10%
5% increase in
All ruminants in urban and
methanogen
head
animal yield
intensively managed LPS
Intensive
0
-$180 to +$1
NA
CH4: -13.3% beef
-11.2%
Beef and dairy cattle in
grazing
per head
cattle; -15.5%
reduction in
developed regions and Latin
dairy cattle
dairy cattle
America in urban and intensively
yield
managed LPS
Note: Annual costs and reduction efficiencies were calculated for each region based on Gerbens (1998), Bates (2001), and
professional judgment based on consultations with experts.
5.3.1.4 Manure Projections Methodology
When available, UNFCCC-reported, country-specific estimates were used for historical emission estimates in
this source category. For those countries with country-reported emission estimates, emission projections were
estimated from the most recent country-reported data through 2050 using growth rates calculated by the Tier 1
methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission estimates
were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite Emission
Projections, for additional information). Activity data for manure management included animal population data
from FAO (2016) and livestock product growth rates form IFPRI's IMPACT model (2016).
The Tier 1 basic equation to estimate Cm emissions from manure management is as follows:
Emission Factor (kg/head/yr) x Animal Population (head)/(106 kg/Gg) = Emissions (Gg/yr) (5.19)
The primary driver for determining Cm emissions from manure management is animal population, assuming
that waste management and animal characteristics do not change significantly over time.90 Within manure
management, emissions were broken out by gas and livestock type.
Activity Data
Historical
• Animal population data for 1990,1995, 2000, and 2005 through 2014 were obtained from FAO (2016).
Populations of nondairy cattle were calculated by subtracting FAO dairy cattle populations from FAO total
cattle populations.91
90 These projections do not take into account changes in manure production or composition due to variations in feed
composition, feed intake, or improved animal genetics, nor do they include potential changes to future waste management
practices, all of which could increase or decrease methane emissions compared to the Tier 1 emission factors (Rojas-Downing
etal., 2017).
91 FAO animal population data are based on total meat production (from both commercial and farm slaughter) and milk
production for a given year and do not take into account fluctuations in animal populations due to seasonal births or slaughter.
FAO data on livestock numbers are intended to cover all domestic animals irrespective of their age and the place or purpose of
their breeding and, as such, do not distinguish between slaughtered dairy or beef cattle.
5-168
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• If country-reported emissions were available only for a portion of the time series, emissions were
interpolated using the available data in conjunction with the growth rate associated with the estimated
Tier 1 emissions calculated for the country.
• The FAO population data were further modified in instances where country-reported data were
aggregated for part of the time series. For example, in 1990, animal population data were not available
for certain countries that were formed after the breakup of the FSU (Armenia, Azerbaijan, Belarus,
Estonia, Georgia, Kazakhstan, Kyrgyzstan, Latvia, Lithuania, Moldova, Russian Federation, Tajikistan,
Turkmenistan, Ukraine, and Uzbekistan), Yugoslavia (Bosnia, Croatia, Macedonia, Slovenia, and Serbia and
Montenegro), Czechoslovakia (Czech Republic and Slovakia), and Ethiopia (Ethiopia and Eritrea). In
addition, animal population data from Belgium and Luxembourg were reported jointly until 2000, and
animal population data from Serbia and Montenegro continued to be reported together until 2006.
Therefore, for each region, the EPA determined the percentage contribution of each country to its
regional total using 1995 (1993 for Czechoslovakia), 2000, or 2006 animal population data. The EPA then
applied these percentages to estimate 1990,1995, 2000, and/or 2005 animal populations for these
countries.
• Default manure management system usage values were taken from the IPCC (2006) Guidelines, Tables
10A-4 to 10A-9, except from poultry emission factors, which were taken from the IPCC (1996)
methodology employed in the 2006 EPA-published report (EPA, 2006).
Projected
• Country-reported emission estimates for 2010, 2015, 2020, 2025, and 2030, if available, were used. If
country-reported emission projections were not available, the EPA projected emissions from 2005
through 2050 based on livestock product growth rates developed by the International Food Policy
Research Institute's IMPACT model (IFPRI, 2016; Robinson et al., 2015; Sulser et al., 2015).92 The IMPACT
model projects growth rates by country for the demand of beef, pork, lamb, poultry, and milk for the
years 2005 through 2050 in 5-year increments. These estimates were used to proxy average annual
growth rates for the livestock species, nondairy cattle, swine, sheep, all poultry (including turkeys, ducks,
geese, and chickens) and dairy cattle, respectively. For the remaining livestock types, the average
population growth rates from 2005 through 2014 in the FAO data were applied.93
• The growth rates described above were applied to the 2014 FAO animal populations to calculate
projected populations for 2015, 2020, 2025, 2030, 2035, 2040, 2045, and 2050 for each livestock species.
Emission Factors
Historical and Projected
• For sheep, goats, camels and other camelids, horses, mules and asses, and poultry, Cm emission factors
for both developed and developing countries were obtained from the 2006 IPCC Guidelines (IPCC, 2006)
by climate type (i.e., cool, temperate, or warm).
92 The IFPRI IMPACT model incorporates supply and demand parameters to determine the estimated growth rates. These
parameters include the feed mix applied according to relative price movements, international trade, national income,
population, and urban growth rates, as well as anticipated changes in these rates over time.
93 Basing livestock population growth on the 2005 through 2014 historical trend led to unrealistically high growth rates in some
countries that have experienced large livestock increases in recent years. In countries where the growth between 2014 and
2050 was greater than 200%, the trend was adjusted to draw on a longer historical period. Where possible, the period used was
1990 through 2014; however, in some cases, a shorter period was necessary to keep growth as close as possible to the range
considered reasonable (i.e., 200% or less).
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-169
-------
METHODOLOGY DOCUMENTATION
• For cattle, swine, and buffalo, Cm emission factors from the 2006 IPCC Guidelines were used and were
selected based on region and average annual temperature (provided in increments of 1° Celsius) for the
country.
• According to IPCC (2006) Tier 1 default assumptions, N2O from manure for animal categories other than
cattle, buffalo, swine, and poultry was assumed to be managed in pasture and grazing operations and was
therefore not included in the manure management estimates. Therefore, manure management emissions
from these animal types were assumed to be zero and were estimated under N2O from agriculturally
managed soils.
• For cattle, buffalo, swine, and poultry, all default data for determining N2O emission factors were
obtained from the 2006 IPCC Guidelines (IPCC, 2006). Nitrogen excretion rates (kilograms nitrogen per
1,000 kg animal mass) were obtained by animal type and region and were used in conjunction with typical
animal mass estimates (in kilograms, available by animal type and region for cattle, swine, and buffalo and
by developed or developing country designation for poultry) to calculate a nitrogen excretion rate per
head per year for each animal type and region and also by developed or developing country designation
for poultry. The nitrogen excretion rate was used with default manure management system usage
estimates and the associated emission factors for each management system to calculate default emission
factors per head per year by animal type and region for cattle, buffalo, and swine and by region and
developed or developing country designation for poultry. Climate type for most countries was identified
using data from the Global Historical Climatology Network, which are published by the National Climatic
Data Center and contain annual average temperatures for most countries' capitals or major cities. These
annual averages are for a range of years, which vary by country. Given the lack of animal population data
by areas within a country, it was assumed that 100% of the animal populations are located in a climate
defined by the average temperature of the country capital. When climate data were not available for a
specific country, either a nearby country was used as a proxy or the general climate type (warm,
temperate, cool) was taken from a previous manure CH4 model from 2002 and the median temperature
for that climate type was used.
Emission Reductions in Baseline Scenario
The methodology used for this source category does not explicitly model any emission reductions; however,
emission reductions are included to the extent they are reflected in country-reported data.
Uncertainty
The default IPCC emission factors represent the greatest source of uncertainty because of the lack of
information on country-specific manure management systems and the geographic concentration of animal
populations, which affects the climate zone assignment. Considerable uncertainty in projected emissions is due to
the lack of information on potential changes to management system types and animal feeding characteristics that
could affect emissions in the projected years. Additionally, the emission projections do not take into account
improved animal genetics that can reduce emissions in enteric fermentation, nor do they take in into consideration
future potential changes to manure management practices (Rojas-Downing et al., 2017). There are also
uncertainties associated with the animal population data. FAO data are derived from total annual meat (from both
commercial and farm slaughter) and milk production data for a given country and do not take into account
fluctuations in animal populations due to seasonal births or slaughter. (FAO, 2016). Additionally, the impacts of
either historical population trends or world markets and livestock product consumption patterns on national
livestock production patterns are often difficult to predict, further increasing the uncertainty of projected
emissions from this source.
5-170
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.3.1.5 Mitigation Options Considered for Manure
Mitigation options for reducing Cm from livestock manure focus on changes in manure management practices
that capture the Cm to flare or use for energy production (see Table 5-60). This analysis included 10 options for
manure management of Cm emissions including both large capital-intensive digesters applied in developed
regions and small-scale digesters for developing regions. Revenues were generated from using captured Cm for
either heat or electricity on the farm; these revenues were scaled to other regions based on an electricity price
index. Capital costs and O&M costs for digester systems were mainly based on the EPA AgSTAR program data and
experience in the United States and developing countries (EPA, 2010; Roos, personal communication 2012; Costa,
personal communication 2012), supplemented by information from USDA (2007, 2011). For the EU, technology
cost and performance parameters were based on Bates et al. (2009). For developing countries, the U.S. technology
cost data were assumed for large digester systems with adjustments made to represent O&M costs in the
developing countries. Capital costs for small-scale systems were based on EPA (2006), which estimates the capital
cost per 1,000 pounds live weight.94
• Complete-Mix Digesters: Complete-mix digesters are more common in warmer climates, where manure is
flushed out of barns or pens with water, lowering the solids' concentration to a level generally between
3% and 10%. Often the manure accumulates in a mixing tank before entering the digester. These digesters
make use of gravity and pumps to move the manure through the system. These digesters are typically
heated to maintain a constant temperature and gas flow.
• Plug-Flow Digesters: Plug-flow digesters consist of long and relatively narrow heated tanks, often built
below ground level, with gas-tight covers. Plug-flow digesters are only used for dairy manure because
they require higher manure solids' content, around 11% to 13%.
• Fixed-Film Digesters: Fixed-film digesters may be appropriate when concentrations of solids are very low,
such as in swine manure management situations where manure is very diluted with water. Fixed-film
digesters consist of a tank packed with inert media on which bacteria grow as a biofilm.
• Covered Lagoons: Covered earthen lagoons are the simplest of the systems used in developed countries
and generally the least expensive, although there is quite a bit of variation in the systems that have been
built. This system is used with low manure solids' concentration (less than 3%) and can be used for swine
or dairy cattle. Cm is captured by covering the lagoon where manure is stored with a floating cover and
piping the gas out to a flare or used on-farm. Because these digesters are not generally heated, the
available gas flow varies significantly over the course of the year.
• Dome Digesters: These are small, unheated digesters used in some developing countries, including China
and India. A typical dome digester is a brick-lined cylinder sunk in the ground with a wall dividing the
cylinder in two with inlet and outlet ports connected to the bottom of the tank. Biogas generated is
typically used by the household for cooking and other household energy needs.
• Centralized Digesters: Centralized digesters are large digesters to which individual farmers transport their
waste for large-scale digestion and dispersion of capital costs.
5.3.1.6 Technical and Economic Characteristics of Options
Table 5-60 provides a description of the technical and economic characteristics of the abatement measures for
manure management included in this analysis.
94 Additional strategies that may be considered in future assessments include the addition of manure separation prior to
anaerobic digestion and other changes in the way that manure is captured and stored.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-171
-------
Table 5-60: Abatement Measures for Manure Management
Abatement
Option
Total Installed
Capital Cost
(2010 USD)
Annual O&M
Cost
(2010 USD)
Capital
Lifetime
(years)
Reduction
Efficiency
(change in
Emissions
per Head)
Benefits (changes in
livestock or energy
revenue)
Technical Applicability
Adjustments
Across Regions
Complete-Mix Digester, Hogs
With engine
100 per head
(US)
0.11 per head
(US)
20
CH4: -85%
$8 energy revenue/
savings per head (US)
Hogs in selected LPS and
management intensities
Labor costs,
labor share,
energy prices
Without
engine
61 per head
(US)
0.07 per head
(US)
20
CH4: -85%
None
Hogs in selected LPS and
management intensities
Labor costs,
labor share
Complete-Mix Digester, Dairy Cattle
With engine
958 per head
(US)
3.35 per head
(US)
20
CH4: -85%
$65 energy revenue/
savings per head (US)
Dairy cattle in selected LPS
and management intensities
Labor costs,
labor share,
energy prices
Without
engine
588 per head
(US)
2.06 per head
(US)
20
CH4: -85%
none
Dairy cattle in selected LPS
and management intensities
Labor costs,
labor share
Plug-Flow Digester, Dairy Cattle
With engine
1288 per head
(US)
2.3
20
CH4: -85%
$65 energy revenue/
savings per head (US)
Dairy cattle in selected LPS
and management intensities
Labor costs,
labor share,
energy prices
Without
engine
790 per head
(US)
8.9
20
CH4: -85%
None
Dairy cattle in selected LPS
and management intensities
Labor costs,
labor share
Fixed-Film Digester, Hogs
With engine
128 per head
(US)
0.15 per head
(US)
20
CH4: -85%
$8 energy revenue/
savings per head (US)
Hogs in selected LPS and
management intensities
Labor costs,
labor share,
energy prices
Without
engine
102 per head
(US)
0.12 per head
(US)
20
CH4: -85%
None
Hogs in selected LPS and
management intensities
Labor costs,
labor share
(continued)
-------
Table 5-60: Abatement Measures for Manure Management (continued)
Abatement
Option
Total Installed
Capital Cost
(2010 USD)
Annual O&M
Cost
(2010 USD)
Capital
Lifetime
(years)
Reduction
Efficiency
(change in
emissions
per Head)
Benefits (changes in
livestock or energy
revenue)
Technical Applicability
Adjustments
Across Regions
Covered Lagoon, Large-Scale, Hogs
With engine
43 per head (US)
0.13 per head
(US)
20
CH4: -85%
$8 energy revenue/
savings per head (US)
Hogs in selected LPS and
management intensities
Labor costs,
labor share,
energy prices
Without
engine
25 per head (US)
0.06 per head
(US)
20
CH4: -85%
None
Hogs in selected LPS and
management intensities
Labor costs,
labor share
Covered Lagoon, Large-Scale, Dairy Cattle
With engine
1182 per head (US)
3.43 per head
(US)
20
CH4: -85%
$65 energy revenue/
savings per head (US)
Dairy cattle in selected LPS
and management intensities
Labor costs,
labor share,
energy prices
Without
engine
773 per head (US)
2.01 per head
(US)
20
CH4: -85%
None
Dairy cattle in selected LPS
and management intensities
Labor costs,
labor share
Small-Scale Digesters
Dome
digester,
cooking fuel
and light
50 per 1,000 lbs live
weight
1.25 per
1,000 lbs live
weight
10
CH4: -50%
$7 energy
revenue/savings per
head hogs (Vietnam),
$48 energy
revenue/savings per
head dairy cattle
(Tanzania)
Hogs and dairy cattle in
selected LPS and management
intensities in developing
countries
Labor costs,
labor share,
energy prices
Polyethylen
e bag
digester,
cooking fuel
and light
20 per 1,000 lbs live
weight
0.5 per 1,000
lbs live
weight
10
CH4: -50%
$7 energy
revenue/savings per
head hogs (Vietnam),
$48 energy
revenue/savings per
head dairy cattle
(Tanzania)
Hogs and dairy cattle in
selected LPS and management
intensities in developing
countries
Labor costs,
labor share,
energy prices
(continued)
-------
Table 5-60: Abatement Measures for Manure Management (continued)
Abatement
Option
Total Installed
Capital Cost
(2010 USD)
Annual O&M
Cost
(2010 USD)
Capital
Lifetime
(years)
Reduction
Efficiency
(change in
emissions
per Head)
Benefits (changes in
livestock or energy
revenue)
Technical Applicability
Adjustments
Across Regions
Centralized Digester
Centralized
digester
163 per head
average for hogs
across the EU, 1,007
per head average
for dairy cattle
across the EU
0.07 per head
for hogs, 2.06
dairy cattle
20
CH4: -85%
$8 energy
revenue/savings per
head for hogs (U.S.)
and $65 energy
revenue/savings per
head for dairy cattle
(U.S.)
Hogs and dairy cattle in
selected LPS and management
intensities in the EU-27 region
Labor costs,
labor share,
energy prices
Note: Cm reduction efficiencies were assumed to be 85% from baseline for the complete-mix, plug-flow, fixed-film, and large-scale covered lagoon digesters
based on the difference between IPCC default emission factors for anaerobic manure management, where Cl-Uis released into the atmosphere, and digesters.
For the smaller-scale digesters applied in developing countries, the reduction efficiency was assumed to be 50% from baseline, where baseline emissions are
much lower because of a different distribution of manure management practices and the likelihood of less efficient CH4 capture. Because all systems are
assumed to capture the same quantity of Cl-U and it was assumed that none of these options affects livestock yields, the benefits per head are constant across
system types, providing a given level of Cl-U reduction and energy generation/substitution. Values shown in the table are illustrative based on a single identified
country for which the values were estimated but vary widely across countries based on relative labor, energy, and other input costs.
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.3.1.7 References
Archibeque, S., K. Haugen-Kozyra, K. Johnson, E. Kebreab, and W. Powers-Schilling. 2012. Near-Term Options for
Reducing Greenhouse Gas Emissions from Livestock Systems in the United States: Beef Dairy, and Swine
Production Systems. Report Nl R12-04. Durham, NC: Nicholas Institute for Environmental Policy Solutions.
Bates, J., N. Brophy, M. Harfoot, and J. Webb. 2009. Agriculture: Methane and nitrous oxide. Sectoral Emission
Reduction Potentials and Economic Costs for Climate Change.
Bates, J. 2001. Economic Evaluation of Emissions Reductions of Nitrous Oxides and Methane in Agriculture in the
EU: Bottom-Up Analysis. Contribution to a Study for DG Environment, European Commission by Ecofys Energy
and Environment, AEA Technology Environment and National Technical University of Athens.
Costa, A. 2012. Personal communication.
Food and Agriculture Organization of the United Nations. 2016. FAOSTAT Statistical Database. Food and
Agriculture Organization of the United Nations. Available online at http://www.fao.Org/faostat/en/#home J
Frank, S., R. Beach, P. Havlik, H. Valin, M. Herrero, A. Mosnier, T. Hasegawa, J. Creason, S. Ragnauth, and M.
Obersteiner. 2018. Structural change as a key component for agricultural non-C02 mitigation efforts. Nature
Communications, 9,1060. doi:10.1038/s41467-018-03489-l
Gerbens, S. 1998. Cost-Effectiveness of Methane Emissions Reduction from Enteric Fermentation of Cattle and
Buffalo. Draft Report.
Hristov, A.N., J. Oh, C. Lee, R. Meinen, F. Montes, T. Ott, J. Firkins, A. Rotz, C. Dell, A. Adesogan, W. Yang, J.
Tricarico, E. Kebreab, G. Waghorn, J. Dijksta, and S. Oosting. 2013. Mitigation of Greenhouse Gas Emissions in
Livestock Production—A Review of Technical Options for Non-C02 Emissions. P.J. Gerber, B. Henderson and
H.P.S. Makkar (eds.). FAO Animal Production and Health Paper No. 177. Rome: FAO.
Intergovernmental Panel on Climate Change. 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.
International Food Policy Research Institute. 2016. Impact Model Growth Rate Spreadsheet for Livestock
Populations e-mailed from Timothy Sulser of IFPRI to Katrin Moffroid of ICF, September 30, 2016.
Popp, A., H. Lotze-Campen, B. Bodirsky. 2010. Food consumption, diet shifts and associated non-C02 greenhouse
gases from agricultural production. Global Environmental Change-Human and Policy Dimensions, 20(3), 451-
462.
Robinson, S., D. Mason-D'Croz, S. Islam, T.B. Sulser, R.D. Robertson, T. Zhu, A. Gueneau, G. Pitois, and M.W.
Rosegrant. 2015. The International Model for Policy Analysis of Agricultural Commodities and Trade (IMPACT):
Model description for version 3. IFPRI Discussion Paper 1483. Washington, DC: International Food Policy
Research Institute. Available online at http://ebrarv.ifpri.org/cdm/ref/collection/pl5738coll2/id/129825 bF
Rojas-Downing, M.M., A.P. Nejadhashemi, T. Harrigan, and S.A. Woznicki. 2017. Climate change and livestock:
Impacts, adaptation, and mitigation. Climate Risk Management, 16,145-163.
Roos, K., EPA. 2012. Personal communication with Robert Beach, RTI.
Sulser, T.B., D. Mason-D'Croz, S. Islam, S. Robinson, K. Wiebe, and M.W. Rosegrant. 2015. Africa in the global
agricultural economy in 2030 and 2050. In Beyond a Middle Income Africa: Transforming African Economies for
Sustained Growth with Rising Employment and Incomes. Chapter 2. O. Badiane and T. Makombe (eds.)
ReSAKSS Annual trends and outlook report 2014. Washington, DC: International Food Policy Research
Institute. Available online at http://ebrarv.ifpri.org/cdm/ref/collection/pl5738coll2/id/130003 cP
U.S. Department of Agriculture, Economic Research Service. February 2011. Climate Change Policy and the
Adoption of Methane Digesters on Livestock Operations. ERR-111. Washington, DC: USDA, ERS.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-175
-------
METHODOLOGY DOCUMENTATION
U.S. Department of Agriculture, Natural Resources Conservation Service. October 2007. Analysis of Energy
Production Costs of Anaerobic Digestion Systems of U.S. Livestock Production Facilities. Washington, DC: USDA,
NRCS. Available online at http://www.agmrc.org/media/cms/manuredigesters FC5C31F0F7B78.pdf cf
U.S. Environmental Protection Agency. 2006. Global Mitigation ofNon-CC>2 Greenhouse Gases. EPA #430-R-06-005.
Washington, DC: EPA.
U.S. Environmental Protection Agency. 2010. Anaerobic Digestion Capital Costs for Dairy Farms. Washington, DC:
EPA. Available online at http://www.epa.gov/AgSTAR/pdf/digester cost fs.pdf
Whittle, L., B. Hug, S. White, E. Heyhoe, K. Harle, E. Mamun, and H. Ahammad. March 2013. Costs and Potential of
Agricultural Emissions Abatement in Australia. Technical Report 13.2. Australian Bureau of Agricultural and
Resource Economics and Sciences.
5-176
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.3.2 Croplands
Land management in croplands influences soil N2O emissions, Cm fluxes, and soil organic carbon stocks (and
associated CO2 fluxes to the atmosphere). Soil N2O emissions are influenced by human activity, including synthetic
nitrogen fertilization practices, application of organic fertilizers such as manure, drainage of organic soils,
cultivation of nitrogen-fixing crops, and enhancement of nitrogen mineralization in soils through practices such as
cultivation/management of native grasslands and forests (Mosier et al., 1998; Smith et al., 2007).
N2O is produced naturally in soils through the microbial process of denitrification and nitrification. A number
of anthropogenic activities add nitrogen to the soils, thereby increasing the amount of nitrogen available for
nitrification and denitrification, and ultimately the amount of N2O emitted. Anthropogenic activities may add
nitrogen to the soils either directly or indirectly.
Direct additions of nitrogen occur from the following activities:
• Various cropping practices, including (1) application of fertilizers; (2) incorporation of crop residues into
the soil, including those from nitrogen-fixing crops (e.g., beans, pulses, and alfalfa); and (3) cultivation of
high organic content soils (histosols); and
• Livestock waste management, including (1) spreading of livestock wastes on cropland and pasture and (2)
direct deposition of wastes by grazing livestock.
Indirect additions occur through volatilization and subsequent atmospheric deposition of ammonia and oxides
of nitrogen that originate from (1) the application of fertilizers and livestock wastes onto cropland and pastureland
and (2) subsequent surface runoff and leaching of nitrogen from these same sources.
Cultivation of histosols, application of sewage sludge, asymbiotic fixation of soil nitrogen, and mineralization
of soil organic matter are additional sources of direct and indirect N2O emissions on croplands. Emissions from
these sources are not calculated or included in these estimates because of a lack of available activity data at the
country level. This may result in an underestimate of emissions.
N2O emissions from agricultural soils calculated herein consist of the following six components:
1. Direct emissions from commercial synthetic fertilizer application (Equation 5.20)
2. Indirect emissions from commercial synthetic fertilizer application (Equation 5.21)
3. Direct emissions from the incorporation of crop residues (Equation 5.22)
4. Indirect emissions from the incorporation of crop residues (Equation 5.23)
5. Direct emissions from manure (pasture, range and paddock and all applied manure) (Equation 5.25)
6. Indirect emissions from manure (Equation 5.26)
5.3.2.1 Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category when available. For those countries with country-reported emission estimates, emission projections were
estimated from the most recent country-reported data through 2050 using growth rates calculated by the Tier 1
methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission estimates
were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite Emission
Projections, for additional information). Activity data for agricultural soils included nitrogen fertilizer consumption
from IFA (2016) or FAO (2016) with growth rates of fertilizer consumption from Tenkerong and Lowenber-DeBoer
(2008) and FAO (2012), crop production and area from FAO (2016) with growth rates from IFPRI's IMPACT Model
(2017), and animal population data from FAO (2016) with livestock product growth rates from IFPRI (2016).
The Tier 1 basic equations to estimate N2O emissions from agricultural soils are as follows:
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-177
-------
METHODOLOGY DOCUMENTATION
Equation 5.20: Direct N2O Emissions from Synthetic Fertilizer
44
Direct Emissions from Synthetic Fertilizer (Gg N2 0) = Fsn * EF1*— (5.20)
where:
Fsn = The annual amount of synthetic fertilizer nitrogen applied to soils (Gg nitrogen)
EFi = Emission factor (equal to 0.01 Gg INhO-N/Gg nitrogen input)
44/28 = Conversion of N2O-N to N2O
Equation 5.21: Indirect N2O Emissions from Synthetic Fertilizer
Indirect Emissions from Synthetic Fertilizer (Gg N20) = [(FSN * FracLEACH * EF3) +
(.Fsn * FracCASF * EF4)] * ^ (5.21)
where:
Fsn = Annual amount of synthetic fertilizer nitrogen applied to soils (Gg nitrogen)
FradEACH = Nitrogen lost from leaching and runoff (equal to 0.30 Gg N/Gg N applied in humid countries and
equal to 0 in arid countries where leaching and runoff are not likely to occur)
EF3 = Emission factor for N2O emissions from nitrogen leaching and runoff (equal to 0.0075 Gg N2O-
N/Gg N leached or runoff)
FracGASF = Fraction of synthetic fertilizer nitrogen that volatilizes as NH3 and NOx (equal to 0.10 Gg
nitrogen volatilized/Gg nitrogen applied)
EF4 = Emission factor for N2O emissions from nitrogen volatilization (equal to 0.01 Gg N20-N/(Gg NH3-
N + NOx-N volatilized))
44/28 = Conversion of N2O-N to N2O
Equation 5.22: Direct N2O Emissions from Crop Residues
Direct Emissions from Crop Residues (Gg N20) = FCR * EF1 *^* 106 (5.22)
where:
Fcr = The annual amount of nitrogen in crop residues and forage/pasture renewal (kg nitrogen)
EFi = Emission factor (equal to 0.01 kg INhO-N/kg nitrogen input)
44/28 = Conversion of N2O -N to N2O
10s = Conversion from kg to Gg
Equation 5.23: Indirect N2O Emissions from Crop Residues
Indirect Emissions from Crop Residues (Gg N20) — Fcr * PracLEACH * EF3 * — * 106 (5.23)
where:
Fcr = The annual amount of nitrogen in crop residues and forage/pasture renewal (kg nitrogen)
FradEACH = Nitrogen lost from leaching and runoff (equal to 0.30 kg N/kg nitrogen applied in humid
countries and equal to zero in arid countries where leaching and runoff are not likely to occur)
EF3 = Emission factor for N2O emissions from nitrogen leaching and runoff (equal to 0.0075 kg N2O-
N/kg nitrogen leached or runoff)
44/28 = Conversion of N2O -N to N2O
5-178
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
10s = Conversion from kg to Gg
Nitrogen additions to soils from crop residues depend on the crop type and yield because different crop types have
different nitrogen contents and different amounts of residue typically left in the soil. The equation for Fcr is as
follows:
Equation 5.24: Annual amount of nitrogen in crop residues and forage/pasture renewal (Fcr)
Fcr (Gg N20) = £r(Yield FreshT * DRYT * ST + IT) * AreaT * (Nag(j^ + Rbg-Bio(r) * Nbgtj)) (5-24)
where:
T
Crop or forage type
Yield Fresh =
Fresh weight yield of crop (kg fresh weight/ha)
DRY
Dry matter fraction of harvested crop (kg dry matter/kg fresh weight)
S
Slope for above-ground residue dry matter
1
Intercept for above-ground residue dry matter
Area =
Total annual area harvested (ha)
Nag =
Nitrogen content of above-ground residues (kg N/kg dry matter)
Rbg-BIO =
Ratio of below-ground residues to above-ground biomass
Nbg =
Nitrogen content of below-ground residues (kg N/kg dry matter)
Direct N2O emissions result from livestock manure that is applied to soils through daily spread operations;
through application to soils of the residues of already managed manure; or through direct deposition on pasture,
range, and paddock by grazing livestock.
Equation 5.25: Direct N2O Emissions from Manure Applied to Soils
Direct Emissions from Manure Applied to Soils (Gg N20) = [(FAM * EF-J
(¦Fprp,so * EF3pRp S0)\ * —
where:
Fam = Annual amount of manure applied to soils (Gg N/year)
EFi = Emission factor (equal to 0.01 Gg INhO-N/Gg nitrogen input)
Fprp = Annual amount of manure deposited by grazing animals on pasture, range, and paddock (PRP)
(Gg N/year)
EF3,prp = Emission factor for N2O emissions from manure deposited by grazing animals on PRP (Gg N2O-
N/Gg N)
CPP = Cattle, poultry, and pigs
SO = Sheep and other animals
44/28 = Conversion of N2O-N emissions to N2O emissions
Equation 5.26: Indirect N2O Emissions from Manure Applied to Soils
Indirect Emissions from Manure Applied to Soils (Gg N20)
{{Fam + Fprp) * FracGASM * EF4J\ * —
+
(FpRP.cpp * FF3pRpcppj +
(5.25)
— [((^4m + Fprp) * FracLEACH * EF-^) +
(5.26)
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-179
-------
METHODOLOGY DOCUMENTATION
where:
Fam = Annual amount of manure applied to soils (Gg N/year)
Fprp = Annual amount of manure deposited by grazing animals on pasture, range, and paddock (PRP)
(Gg N/year)
FradEACH = Nitrogen lost from leaching and runoff (equal to 0.30 Gg N/Gg N)
EF3 = Emission factor for N2O emissions from nitrogen leaching and runoff (equal to 0.0075 Gg N2O-
N/Gg N leached or runoff)
FracGASM = Fraction of animal manure nitrogen that volatilizes as NH3 and NOx (equal to 0.10 Gg nitrogen
volatilized/Gg nitrogen applied)
EF4 = Emission factor for N2O emissions from nitrogen volatilization (equal to 0.01 Gg N20-N/(Gg NH3-
N + NOx-N volatilized))
44/28 = Conversion of N2O-N to N2O
Equation 5.27: Annual amount of managed manure applied to soils (Fam)
Fam = NMMS Avb x [1 — (FracFEED + FracFUEL + FracCNST)] x 106
(5.27)
where:
Fam = Annual amount of manure applied to soils (Gg N/year)
NMMs_Avb = Amount of managed manure nitrogen available for application to managed soils or for feed,
fuel, or construction purposes (kg nitrogen yr
FracFEED = Fraction of managed manure used for feed (%)
FracFUEL = Fraction of managed manure used for fuel (%)
FraceNST = Fraction of managed manure used for construction (%)
Using IPCC (2006) Equation 10.34, the EPA estimated managed manure nitrogen available for application to
managed soils as follows:
Equation 5.28: Annual amount of managed manure available (Nr>
N,
MMS_Avb
= MM[C
JV(T) x NeXfj) x MSfj
x x NbeddingMSj
)]} (5.28)
where:
NMMs_Avb = Amount of managed manure nitrogen available for application to managed soils or for feed,
fuel, or construction purposes (kg nitrogen yr-1)
N(t; = Number of head of livestock species/category T in the country
Nex(r) = Annual average nitrogen excretion per animal of species/category T in the country (kg nitrogen
animal1 yr
MS(t,s> = Fraction of total annual nitrogen excretion for each livestock species/category T that is managed
in manure management system S in the country (dimensionless)
FradossMs = Amount of managed manure nitrogen for livestock category T that is lost in the manure
management system S (%)
NbeddingMs = Amount of nitrogen from bedding (to be applied for solid storage and deep bedding manure
management system if known organic bedding usage) (kg nitrogen animal1 yr
S = Manure management system (MMS)
T = Species/category of livestock
In all six components of emissions from agricultural soils, activity data (i.e., fertilizer consumption, crop
production/area harvested, and livestock populations) were the driving factor for determining emissions.
5-180
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Activity Data
Direct and Indirect Emissions from Commercial Synthetic Fertilizer Application
Historical
• Commercial synthetic fertilizer consumption data were obtained from the IFA database of fertilizer
statistics, known as IFADATA (IFA, 2016), and from the FAO database of agricultural statistics, known as
FAOSTAT (FAO, 2016).
• IFA was the preferred source of activity data, and when IFA data were unavailable, FAO data were used.
Activity data from at least one of these sources were available for most countries from 1990 through
2014. Specifically, data on the consumption of nitrogenous fertilizer, reported in metric tons of nitrogen95
(FAO) or thousand metric tons of nitrogen (IFA), were used.
Projected
• The growth rates of fertilizer consumption from 2015 through 2030 were estimated by using the regional
nitrogen fertilizer consumption projections available from Tenkorang and Lowenberg-DeBoer et al. (2008)
who provided regional fertilizer use for 2015 and 2030. Countries in a specific region assigned by
Tenkorang and Lowenberg-DeBoer et al. (2008) were assigned the same fertilizer consumption growth
rates. Fertilizer use for 2020 and 2025 was interpolated. These consumption projections were then used
to calculate average annual growth rates for the 5-year increments between 2015 and 2030, which in turn
were used to project fertilizer use by country.
• The average annual percentage change in fertilizer use by region for the remainder of the projected time
series (i.e., 2030 through 2050) was available from FAO (2012). The average annual regional growth rates
for the 5-year increments between 2030 and 2050 were used to project fertilizer use by country.
Countries were assigned to regions based on Annex 1 of Tenkorang and Lowenberg-DeBoer (2008) and
Appendix 1 of FAO (2012).
Direct and Indirect Emissions from the Incorporation of Crop Residues
Historical
• Historical production and area statistics for the following major crops (residues of which are typically
incorporated into soils)—barley, maize, pulses,96 rice, sorghum, soybeans, and wheat—were obtained
from FAO. Historical production and area data for these crops were available for most countries for 1990
through 2014 (FAO, 2016). For countries for which data were not available, the EPA assumed zero
production.
Projected
• The growth rates of crop production and area by crop type for 2015 through 2050 were estimated based
on country and regional crop production and area projections developed by IFPRI's IMPACT model (IFPRI,
2017). Projected crop production and area data for all crops through 2050 were obtained from IFPRI by
country and subcontinent. When country data were not available, data by subcontinent (e.g., Middle East)
were used.
• These production and area projections were used to calculate average annual growth rates for the 5-year
increments between 2015 and 2050. For countries for which specific data were unavailable, but the
country is known to produce the crop according to FAO (2016), the EPA used the 5-year growth rates for
the relevant region and then used the growth rates to project crop production and area by country.
95 In the FAO online database, fertilizer data appear to be reported in metric tons, but data are actually reported in metric tons
of nitrogen. This was corroborated by paper copies of the FAO statistics.
96 Pulses include lentils, dry beans, dry broad beans, dry horse beans, chickpeas, and pulses not elsewhere specified.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-181
-------
METHODOLOGY DOCUMENTATION
Direct and Indirect Emissions from Manure (Pasture, Range, and Paddock and All Applied Manure)
Historical
• Animal population data for 1990,1995, 2000, and 2005 through 2014 were obtained from FAO (2016).
Populations of nondairy cattle were obtained by subtracting FAO dairy cattle populations from FAO total
cattle populations.
Projected
• Emissions from 2015 through 2050 were projected based on livestock product growth rates developed by
the IFPRI IMPACT model (IFPRI, 2016). The IMPACT model projects growth rates by country for the
demand of beef, pork, lamb, poultry, and milk for the years 2005 through 2050 in 5-year increments.
These estimates were used to proxy average annual growth rates for the livestock species, nondairy
cattle, swine, sheep, all poultry (including turkeys, ducks, geese, and chickens), and dairy cattle,
respectively. For the remaining livestock types, the average population growth rates from 2005 through
2014 in the FAO data were applied.97
• Starting with the historical year 2014, FAO animal population statistics, growth rates were applied to
calculate projected populations for 2015, 2020, 2025, 2030, 2035, 2040, 2045, and 2050 for each livestock
species.
Emission Factors
Direct and Indirect Emissions from Commercial Synthetic Fertilizer Application
Historical and Projected
• The default Tier 1 emission factors from IPCC (2006) were used to calculate indirect emissions from
synthetic fertilizer consumption.
• As recommended in the 2006 IPCC Guidelines (IPCC, 2006), the EPA assumed that 1% of all nitrogen from
fertilizer consumption is directly emitted as N2O.
Direct and Indirect Emissions from the Incorporation of Crop Residues
Historical and Projected
• Crop residue factors by crop type as shown in Table 11.2 in the 2006 IPCC Guidelines (IPCC, 2006) were
used; a proxy was used if a default factor was not available for a particular crop.
• Nbg for rice and Rbg-Bio for sorghum were based on the general "grains" category in the 2006 IPCC
Guidelines (IPCC, 2006).
Direct and Indirect Emissions from Manure (Pasture, Range, and Paddock and All Applied Manure)
Historical and Projected
• IPCC default nitrogen excretion rates by region and development category were used to estimate nitrogen
excretion per head by country for each animal type, based on the country's region and development
category (IPCC, 2006). To use these defaults, we assigned countries to regions (i.e., Africa, Asia, Eastern
Europe, Indian Subcontinent, Latin America, Middle East, North America, Oceania, and Western Europe)
and development categories (i.e., developed and developing).
• Next, the IPCC guidance methodology on "Coordination with reporting for N2O emissions from managed
soils," found in Section 10.5.4 of the 2006 IPCC Guidelines was applied to determine the amount of
97 Basing livestock population growth on the 2005 through 2014 historical trend led to unrealistically high growth rates in some
countries that have experienced large livestock increases in recent years. In countries where the growth between 2014 and
2050 was greater than 200%, the trend was adjusted to draw on a longer historical period. When possible, the period used was
1990 through 2014; however, in some cases, a shorter period was necessary to keep growth as close as possible to the range
considered reasonable (i.e., 200% or less).
5-182
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
nitrogen that remains in manure following management in MMSs. The amount of nitrogen remaining
corresponds to the amount available for application to agricultural soils.
Emission Reductions in Baseline Scenario
The methodology used for this source category does not explicitly model any emission reductions; however,
emission reductions are included to the extent they are reflected in country-reported data.
Uncertainty
The greatest uncertainties are associated with the completeness of the activity data used to derive the
emission estimates. Emissions from fertilizers are only estimated from synthetic fertilizer use. However, organic
fertilizers (other than the estimated manure and crop residues) also contribute to N2O emissions from soils, but
this activity is not captured in these estimates. Crop residues from crops other than those covered (including from
nitrogen-fixing crops other than soybeans and pulses) may be left on the field, thus resulting in N2O emissions. The
identity and quantity of these crops vary among the different countries.
The livestock nitrogen excretion values, while based on detailed population statistics and regional nitrogen
excretion factors, do not accurately reflect country-to-country variations in animal weight or feeding regimes. Any
contribution of animal bedding materials to manure nitrogen was not considered. The "other" category for manure
management is a large unknown—the EPA assumed no emissions from this category, except for from poultry,
where the "other" category was assumed to represent an average of the "poultry with litter" and "poultry without
litter" management systems. Finally, emissions from histosols, sewage sludge, asymbiotic fixation of soil nitrogen,
and mineralization of soil organic matter are not calculated or included in these estimates because of a lack of
available activity data at the country level. The last two sources, in particular, can be a significant component of
agricultural soil emissions.
Uncertainty also exists in the projected emissions. For some subcategories, projections are not available to
2050, so projections from earlier periods were used. Additionally, in some cases projections are on a regional level,
not a country-specific level, and using regional projections increases uncertainty. Economic and environmental
agricultural policies and improved farming practices are also factors that affect emissions from agricultural soil
management. Because of the complexities of agricultural product markets and the influences of disruptions in the
industry (such as food safety issues), many of these factors are hard to predict, thereby contributing to the
uncertainty of emission projections for this source.
Historical Data Assumptions
The EPA used the following assumptions for countries with incomplete data.
Direct and Indirect Emissions from Commercial Synthetic Fertilizer Application
Eritrea before 1993. In 1993, the former People's Democratic Republic of Ethiopia (Ethiopia PDR) divided into
Ethiopia and Eritrea. Data for Ethiopia for 1990 through 2014 were available from IFA, but data for Eritrea were
not. To estimate the fertilizer consumption of Eritrea in 1990, the EPA determined the relative ratio of the fertilizer
consumption of the current Eritrea and Ethiopia in 1993. This ratio (2% for fertilizer consumption) was then applied
to the fertilizer consumption of Ethiopia PDR to estimate the fertilizer consumption of Eritrea for 1990. This
method assumed that the IFA data for Ethiopia in 1990 included only the portion of Ethiopia PDR that would
become Ethiopia and not the portion that would become Eritrea.
Belgium-Luxembourg before 2000. In 2000, Belgium and Luxembourg began reporting separately to FAO,
rather than together, as had been the case previously. The distribution of fertilizer consumption between these
two countries in 2000 was assumed to be the same for 1990 and 1995. Consequently, Belgium-Luxembourg
consumption in 1990 and 1995 was allocated between Belgium and Luxembourg by their relative percentages in
2000.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-183
-------
METHODOLOGY DOCUMENTATION
The former Yugoslavia before 1995. In 1995, Yugoslavia divided into separate countries. The distribution of
fertilizer consumption among the former Yugoslav countries in 1995 was assumed to be the same for 1990.
Consequently, Yugoslavia consumption in 1990 was allocated among the former Yugoslav countries according to
their relative percentages in 1995. Montenegro was not reported separately from Serbia at any point, and it was
assumed that this country had zero synthetic fertilizer consumption (i.e., all consumption was allocated to Serbia).
The former Czechoslovakia before 1993. In 1993, Czechoslovakia divided into the Czech and Slovak Republics.
The distribution of fertilizer consumption between these two countries in 1993 was assumed to be the same for
1990. Consequently, Czechoslovakia consumption in 1990 was allocated between the Czech and Slovak Republics
by their relative percentages in 1993.
IFA reports data for FSU states dating back to 1990 (before the breakup of the Soviet Union), so there was no
need to separate out Soviet Union data for 1990, as would have to be done with FAO data, which were not
reported separately in 1990.
Portions of the FAO time series for particular countries were determined to be outliers because they differed
significantly from other parts of the time series and did not lineup with trends in other parts of the time series. In
such cases, the rest of the time series was extrapolated to replace the outlier data point. This was the case for
Algeria, Benin, Nepal, and United Arab Emirates for 2005. In addition, the entire FAO time series for Bahrain was
not used because of significant and extreme variations in reported fertilizer use. In these two cases, no other data
were available, and fertilizer use was assumed to be zero.
Direct and Indirect Emissions from the Incorporation of Crop Residues
The FSU before 1993. In 1993, the Soviet Union divided into separate countries (in the context of FAO
reporting—the political dissolution occurred in 1991). The distribution of crop production and area among the FSU
countries in 1993 was assumed to be the same for 1990. Consequently, Soviet crop production and area in 1990
were allocated among the FSU countries by their relative percentages in 1993.
The former Yugoslavia before 1995. In 1995, Yugoslavia divided into separate countries. The distribution of
crop production and area among the former Yugoslav countries in 1995 was assumed to be the same for 1990.
Consequently, Yugoslavia crop production and area in 1990 were allocated among the former Yugoslav countries
according to their relative percentages in 1995. Montenegro was not reported separately from Serbia at any point,
and it was assumed that this country had zero crop production and area (i.e., all production and harvested area
were allocated to Serbia).
The former Czechoslovakia before 1993. As noted above, in 1993, Czechoslovakia divided into the Czech and
Slovak Republics. The distribution of crop production and area between these two countries in 1993 was assumed
to be the same for 1990. Consequently, Czechoslovakia crop production and area in 1990 were allocated between
the Czech and Slovak Republics by their relative percentages in 1993.
Ethiopia and Eritrea before 1993. As noted above, in 1993, Ethiopia PDR divided into Ethiopia and Eritrea. The
distribution of crop production and area between these two countries in 1993 was assumed to be the same for
1990. Consequently, Ethiopia PDR crop production and area in 1990 were allocated between Ethiopia and Eritrea
by their relative percentages in 1993.
Belgium-Luxembourg before 2000. As noted above, in 2000, Belgium and Luxembourg began reporting
separately to FAO, rather than together, as had previously been the case. The distribution of crop production and
area between these two countries in 2000 was assumed to be the same for 1990 and 1995. Consequently,
Belgium-Luxembourg crop production and area in 1990 and 1995 were allocated between Belgium and
Luxembourg by their relative percentages in 2000.
5-184
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Direct and Indirect Emissions from Manure (Pasture, Range, and Paddock and All Applied Manure)
In 1990, animal population data were not available for certain countries that were formed after the breakup
of the FSU (Armenia, Azerbaijan, Belarus, Estonia, Georgia, Kazakhstan, Kyrgyzstan, Latvia, Lithuania, Moldova,
Russian Federation, Tajikistan, Turkmenistan, Ukraine, and Uzbekistan), Yugoslavia (Bosnia, Croatia, Macedonia,
Slovenia, and Serbia and Montenegro), Czechoslovakia (Czech Republic and Slovakia), and Ethiopia (Ethiopia and
Eritrea).
In addition, Belgium and Luxembourg were reported jointly until 2000. Therefore, for each region, the EPA
determined the percentage contribution of each country to its regional total using 1995 (1993 for Czechoslovakia)
or 2000 animal population data. The EPA then applied these percentages to estimate 1990 and/or 1995 animal
population for these countries. The animal types included were dairy cows, other cattle, buffalo, sheep, goats, pigs,
chickens, turkeys, ducks, geese, horses, mules, asses, camels, and other camelids (assumed to be llamas and
alpacas).
5.3.2.2 Mitigation Options Considered for Croplands
This section estimates global non-CC>2 mitigation potential from croplands using DAYCENT, a gridded global
biophysical model. Given that a key goal of this study is to generate global mitigation estimates for all regions of
the globe, we focused on applying models capable of characterizing emissions and yields across the globe under
alternative scenarios rather than limiting the study to crop/region combinations where long-term empirical data
are available for case studies. Using a spatially differentiated biophysical model improves the analysis by
incorporating differences in soil characteristics and dynamic soil processes, climate, and management practices, as
well as crop-specific nutrient uptake and plant physiology. It allows us to model mitigation consistently across the
entire world, capturing differentiation in underlying characteristics that affect simulation results for both baseline
and mitigation practices between countries. All gridded mitigation estimates are defined relative to the baseline
conditions for that point. This enables us to capture differences in yield response to altering nutrient management
practices, for instance, based on the initial level of nutrient availability.
The mitigation options considered in this study include:98
• No Till: All cultivation and field preparation events were removed except for seeding, which occurred
directly into the residue.
• Reduced Fertilization: Baseline fertilizer application levels were reduced by 20%.
• Increased Fertilization: Baseline fertilizer application levels were increased by 20%.
• Split Nitrogen Fertilization: Under this option, the baseline nitrogen application amount was applied
in three separate and equal amounts (planting day, 16 days after planting day, and 47 days after
planting day) instead of once on planting day (following Del Grosso et al. [2009]).
• Nitrification Inhibitors: The baseline nitrogen application amount was applied once annually on the
date of planting. Nitrification inhibitors were applied at the time of fertilization and reduced
nitrification by 50% for 8 weeks (following Del Grosso et al. [2009] and Branson et al. [1992]).
• 100% Residue Incorporation: In this option, all crop residue was assumed to remain after harvest.
This option serves to evaluate how reducing removals would affect soil organic carbon stocks.
While using a biophysical model is important for globally consistent spatially differentiated estimation of
mitigation potential and cost, it comes with a number of restrictions. First, to assess cropland mitigation potential
and costs, DAYCENT-simulated yields and emissions under each mitigation option are compared to DAYCENT-
98 Additional mitigation options (e.g., reduced tillage, other N process inhibitors, slow release fertilizers) may be considered in
future studies, depending on the level of potential mitigation and data availability for inclusion within DAYCENT.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-185
-------
METHODOLOGY DOCUMENTATION
simulated baseline conditions" to calculate proportionate changes in a consistent manner. However, baseline
emissions simulated by DAYCENT differ from the baseline estimates of emissions as calculated above.100 Because
dynamic soil processes are represented, treatments were held constant over time to focus on the dynamics of
changes in yields and emissions relative to baseline conditions. Specifically, fertilizer applications are held constant
in the DAYCENT baseline. Instead, alternative fertilizer rates appear in the GHG mitigation strategies evaluated
below.
Second, it is important to note that this analysis currently models the effects of applying mitigation strategies
to only a fraction of the total emissions from agriculture, focusing on emissions from major crops (with mitigation
scenario estimates applied to similar minor crops in some cases) where there are sufficient data available for
model calibration. Crops accounting for about two-thirds of global cropland are simulated within DAYCENT.101 See
Table 5-61 for a description of the input data used in the DAYCENT simulations. Specifically, the following types of
emissions and crops are included:
• Direct and indirect emissions from mineral-based cropland soils processes
- Synthetic and organic fertilization
- Residue N
- Mineralization and asymbiotic fixation, based on temperature and moisture, etc.
• Major crops supplemented by selected similar minor crops
- Barley (plus rye)
- Maize (plus green corn)
- Sorghum
- Soybeans (plus lentils, other beans)
- Wheat (plus oats)
99 Baseline conditions for each modeled crop in each region were simulated using DAYCENT based on data characterizing
climate, soils, and management practices.
100 There are substantial differences in the baseline emissions from agricultural soils presented in the GER analyses and those
utilized in calculating mitigation potential. The GER emissions are based on country-reported values or Tier 1 calculations,
whereas DAYCENT calculates emissions using global gridded data and crop-specific nonlinear functions. The DAYCENT
simulations are expected to provide estimates of mitigation potential that better reflect differences across crops as well as
spatial differences in soil, climate, cultivar, management, and other key factors, but they are only available for the subset of
crops for which sufficient data were available to adequately calibrate the model. In addition, the DAYCENT simulations held
fertilizer use constant during the simulation period to focus on the dynamic response of emissions to changes in management.
However, fertilizer use is projected to rise substantially over time in the GER, leading to increasing divergence in the baseline
emission estimates over time for the subset of crops that are consistent between the two baselines. Given differences in
mitigation options between cropland and grassland as well as the nonlinear relationships between soil, climate, and
management and the resulting yields and GHG emissions, there is no readily available method for applying the available
estimates of the effects of cropland mitigation options from DAYCENT to additional crops or to grasslands. Thus, the baseline
used for comparison to the mitigation scenarios reflects only a subset of total emissions from agricultural soils. Similarly, the
GHG emissions potential reported reflect only a subset of the total potential reductions available from agricultural soils.
Expanding the scope of GHG emissions from agricultural soils for which mitigation potential and costs have been estimated is
an important area of future research.
101 Thus, mitigation potential presented here is an underestimate of mitigation that could be achieved if these mitigation
options were applied to nonmodeled crops. There is additional mitigation potential associated with managed grasslands that is
not captured here but is an important topic for future research.
5-186
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-61: Description of the Input Data Used in DAYCENT Simulations
Data Type Description Source
Daily weather Daily weather for 1901-2010 at 0.5° The original data source was the MsTMIP
resolution in latitude by longitude. This project's 6-hour CRU + NCEP combined
includes daily minimum temperature, daily data. This was aggregated to daily, and all
maximum temperature, and daily nonland cells were removed,
precipitation. http://nacp.ornl.gov/MsTMIP.shtml
Soils These data are the same as data used for the Food and Agriculture Organization. 1996.
previous DAYCENT global simulations. The The Digitized Soil Map of the World
data are at 0.5° resolution in latitude by Including Derived Soil Properties. CDROM.
longitude and include sand, silt, clay, bulk Rome: FAO.
density, pH, and number of soil layers.
Agricultural cells to This mask was computed from the fraction of
simulate agricultural area. The fraction of agricultural
area is provided at a 5-minute resolution in
latitude by longitude. These data were
aggregated to 0.5° resolution by latitude and
longitude then we selected cells where
fraction of cropland area > 5% of the grid cell
area.
Crop-specific masks indicating where to These files were produced by Mirella
simulate each crop. Each crop mask is a Salvatore at the FAO and Aaron Berdanier.
subset of the agricultural cells to simulate, Personal communication,
described above. This data was provided at
0.5° resolution in latitude by longitude.
Note: Although separate crop masks were
provided for winter and spring wheat, there
was almost no difference between these
masks. Likewise for winter and spring barley.
The main difference between winter and
spring varieties was the planting and harvest
dates (see below).
Irrigated areas by Crop-specific data with the fraction of
crop type cropland area that is irrigated. These data
were provided at 0.5° resolution in latitude by
longitude for all years between 1985 and
2008. Irrigation was simulated for modern
agriculture (year 1951 or later) for cells where
the irrigated fraction > 0.0 for any year
between 1985 and 2008. The fraction of
cropland irrigated in 2008 was used in the
post-processing step to aggregate model
results.
(continued)
Agricultural Lands in the Year 2000.
Described in the publication, Ramankutty et
al. 2008. Farming the planet: 1. Geographic
distribution of global agricultural lands in
the year 2000. Global Biogeochemical
Cycles 22, GB1003.
doi:10.1029/2007GB002952.
Crop masks for
maize, winter
wheat, spring
wheat, winter
barley, spring
barley, sorghum,
and soybean
These files were produced by Mirella
Salvatore at the FAO and Aaron Berdanier.
Personal communication.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-187
-------
METHODOLOGY DOCUMENTATION
Table 5-61: Description of the Input Data Used in DAYCENT Simulations (continued)
Data Type Description
Initial year of Fraction of area in agriculture for years 1700-
cultivation 2007 at 0.5° resolution in latitude by longitude.
We computed the first year when the fraction
of agricultural area was 50% of the fraction of
cropland area in 2000; this determined the
year of plow-out for the cell.
Source
Global Cropland and Pasture Data from
1700-2007. This is a beta release of an
updated version of the DAYCENT original
historical cropland dataset that spanned
the 1700-1992 period. The original
dataset was described in the publication
by Ramankutty and Foley (1999) in Global
Biogeochemical Cycles. This release
updates the data to the 1700-2007 time
period.
Crop-specific Planting date (day of year) and harvest date
planting and (day of year) for each crop at 0.5° resolution in
harvest dates latitude by longitude: barley (winter), barley
(spring), maize (main season), maize(second
season), sorghum (main season), sorghum
second season), soybeans, wheat (winter),
wheat (spring)
Sacks, W.J., D. Deryng, J.A. Foley, and N.
Ramankutty. 2010. Crop planting dates: An
analysis of global patterns. Global Ecology
and Biogeography 19, 607-620.
doi: 10.1111/j. 1466-8238.2010.00551.x.
Harvest type and
residue removal
rate by crop.
Harvest type and residue removal rate by crop
at 0.5° resolution in latitude by longitude by
crop. The harvest type designates a grain or
nongrain harvest (for this exercise, all crops
had grain harvests). The residue removal rate
determines the percentage of residue removed
from the field at time of harvest. Residue
includes all above-ground plant material after
grain is removed.
These files were produced by Mirella
Salvatore at the FAO and Aaron Berdanier.
Personal communication.
Tillage, planting, Tillage, planting, and weeding practices by crop
and weeding for developed countries (conventional),
practices by country developing countries (conservation), and less
and by crop developed countries. Crops are categorized as
small grain (barley, wheat) or large grain
(maize, sorghum, soybean). These practices
determine the intensity of soil disturbance
simulated for each event.
These files were produced by Mirella
Salvatore at the FAO and Aaron Berdanier.
Personal communication.
N application rates Annual nitrogen application rates including These files were produced by Mirella
include fertilizer nitrogen fertilizer plus manure nitrogen (gN m"
nitrogen and yrat 0.5° resolution in latitude by longitude
manure nitrogen by crop for years 1985-2008. Nitrogen
application rates from 1950-1984 were linearly
interpolated between 0.0 in 1950 and the 1985
rate. Nitrogen application rates for 2009-2035
were set to the 2008 rate.
Note: There were no data about the relative
amount of fertilizer nitrogen and manure
nitrogen.
Salvatore at the FAO and Aaron Berdanier,
Personal communication.
(continued)
5-188
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-61: Description of the Input Data Used in DAYCENT Simulations (continued)
Data Type
Harvested areas
and yields by crop
type in year 2000.
Description
Harvested area (proportion of grid cell area)
and yield (tons/ha). The data are provided at
5-minute resolution in latitude by longitude.
We aggregated the data to a 0.5° resolution.
The measured yields were compared with
simulated yields from the baseline simulation.
The harvested area fraction was used in the
post-processing step for aggregating model
results.
Source
Harvested area and yields of 175 crops
(M3-Crops Data). Monfreda et al. 2008.
Farming the planet: 2. Geographic
distribution of crop areas, yields,
physiological types, and net primary
production in the year 2000. Global
Biogeochemical Cycles 22, GB1022.
doi:10.1029/2007GB002947.
Compared with the projections reported above, this discussion excludes rice (soils and cultivation) which
treated separately below. Furthermore, compared to the estimates typically developed for GHG inventories, the
emissions presented in this section are lower because the following types of emissions were excluded because of
data limitations and a lack of mitigation options:
• Drainage of organic soils.
• Grassland soils, including pasture
• Other crops not mentioned above (e.g., vegetables)
• Restoration of degraded lands
• Burning of residues or biofuel
The focus is on emissions from major crops, which is consistent with our evaluation of options that can be
applied to mitigate emissions from these major crops in this section.
The mitigation options evaluated in this analysis were based on a review of the literature to identify the most
promising options, while also considering data availability and the potential for modeling within DAYCENT. The
mitigation options represent alternative management practices that would alter crop yields and the associated
GHG emissions, including adoption of no-till management, split nitrogen fertilization applications, application of
nitrification inhibitors, increased nitrogen fertilization (20% increase over BAU),102 decreased nitrogen fertilization
(20% reduction from BAU), and 100% crop residue incorporation.
The nitrogen management practices (split nitrogen fertilization, nitrification inhibitors, increased and
decreased nitrogen fertilization) influence N2O emissions in addition to soil organic carbon stocks due to reduced
or enhanced carbon inputs associated with the level of crop production. Smith et al. (2007) estimated that 89% of
the overall technical potential for mitigation of agricultural GHG emissions is associated with carbon sequestration
in soils. Although soil organic carbon stock fluxes are negligible in the baseline, there is considerable opportunity to
modify stocks in the future. Levels of soil organic matter and, in particular, soil C both influence and are influenced
102 Baseline fertilizer levels vary across grid cells. As noted in Table 6-61, global gridded fertilizer rates were provided by M i re 11 a
Salvatore at the FAO and Aaron Berdanier, personal communication. For the purposes of this analysis, we explored the
potential effects of increasing and decreasing N fertilizer application rates by 20%. Those values were selected to represent a
reasonable range of changes in fertilizer application. Depending on baseline fertilizer rates, some regions see relatively little
reduction in yields from a 20% reduction, while others see a much more substantial yield decrease. Increasing fertilizer levels
generally increases N20 emissions (though also tends to increase soil carbon), but there are some countries, particularly in Sub-
Saharan Africa, where increasing fertilizer rates results in a reduction in GHG intensity (i.e., reduction in emissions per unit of
output). The primary rationale for including an increase in fertilizer rate was to reflect the potential for reduced emission
intensity and lower emissions in a constant crop production scenario.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-189
-------
METHODOLOGY DOCUMENTATION
by cropland productivity. Other things being equal, higher crop yields may increase soil carbon wherever more
crop residue can be incorporated into the soil. Similarly, reducing crop residue removal would affect soil organic
carbon stocks by changing the amount of carbon input to the soil. Practices such as adoption of conservation
tillage, restoration of degraded lands, improved water and nutrient management, and cropping intensification can
increase soil C by enhancing carbon inputs to soils from greater crop production or decrease the losses of carbon
from soils with lower decomposition rates (Paustian et al., 1997; Six et al., 2000).
5.3.2.3 Technical Characteristics of Options
For the purposes of this analysis of the technical mitigation potential from croplands soils, all options are
available in all regions and all time periods, though only options that result in emission reductions for a given
crop/region combination are retained in the MAC curves. Options related to fertilization are applicable where
baseline fertilizer application levels are nonzero. The changes in emissions and yield for each crop/region/year
combination modeled were simulated in the DAYCENT model. Table 5-62 presents the base yields for each
modeled crop type and the difference from base yield impacts associated with the mitigation option over time.
Table 5-62: DAYCENT Base Mean Yields and Differences from Mean Yield for Mitigation Strategies, by Year
(Metric Tons of Grain per Hectare)
Mitigation Strategy
2015
2020
2025
2030
Maize
Base yield3
3.64
3.64
3.59
3.60
No till
-0.25
-0.17
-0.12
-0.07
Optimal nitrogen fertilization13
2.9
3.05
3.1
3.08
Split nitrogen fertilization
0.16
0.17
0.19
0.18
100% residue incorporation
0.23
0.24
0.24
0.24
Nitrification inhibitors
-0.01
-0.01
-0.01
-0.01
Reduced fertilization
-0.36
-0.39
-0.4
-0.4
Increased fertilization
0.28
0.29
0.31
0.31
Sorghum
Base yield3
2.34
2.35
2.33
2.32
No till
-0.18
-0.13
-0.1
-0.06
Optimal nitrogen fertilization13
3.07
3.27
3.19
3.25
Split nitrogen fertilization
0.14
0.14
0.13
0.14
100% residue incorporation
0.15
0.17
0.16
0.17
Nitrification inhibitors
-0.02
-0.03
-0.02
-0.02
Reduced fertilization
-0.22
-0.25
-0.26
-0.27
Increased fertilization
0.19
0.22
0.22
0.23
(continued)
5-190
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-62: DAYCENT Base Mean Yields, and Differences from Mean Yield for Mitigation Strategies, by Year
(Metric Tons of Grain per Hectare) (continued)
Mitigation Strategy
2015
2020
2025
2030
Winter Wheat
Base yield3
2.92
2.89
2.8
2.87
No-till
-0.13
-0.11
-0.07
-0.05
Optimal nitrogen fertilization13
1.55
1.82
1.87
1.78
Split nitrogen fertilization
0.09
0.1
0.11
0.11
100% residue incorporation
0.11
0.12
0.13
0.12
Nitrification inhibitors
0.03
0.04
0.04
0.05
Reduced fertilization
-0.22
-0.26
-0.25
-0.27
Increased fertilization
0.19
0.2
0.2
0.21
Spring Wheat
Base yield3
2.94
2.92
2.85
2.83
No till
-0.16
-0.13
-0.1
-0.08
Optimal nitrogen fertilization13
1.49
1.46
1.4
1.36
Split nitrogen fertilization
0.07
0.08
0.08
0.08
100% residue incorporation
0.11
0.11
0.11
0.11
Nitrification inhibitors
0.02
0.03
0.03
0.03
Reduced fertilization
-0.2
-0.22
-0.21
-0.21
Increased fertilization
0.14
0.15
0.14
0.14
Winter Barley
Base yield3
3.59
3.58
3.5
3.57
No till
-0.2
-0.21
-0.15
-0.1
Optimal nitrogen fertilization13
2.64
3.11
3.07
3
Split nitrogen fertilization
0.04
0.06
0.06
0.05
100% residue incorporation
0.37
0.39
0.39
0.39
Nitrification inhibitors
0.01
0.03
0.03
0.03
Reduced fertilization
-0.34
-0.39
-0.41
-0.43
Increased fertilization
0.31
0.35
0.36
0.38
Spring Barley
Base yield3
2.83
2.79
2.77
2.77
No till
-0.29
-0.24
-0.2
-0.17
Optimal nitrogen fertilization13
1.8
1.8
1.67
1.63
Split nitrogen fertilization
0.08
0.09
0.09
0.08
100% residue incorporation
0.21
0.22
0.21
0.21
Nitrification inhibitors
0.01
0.02
0.02
0.02
Reduced fertilization
-0.28
-0.31
-0.31
-0.32
Increased fertilization
0.24
0.26
0.25
0.25
(continued)
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-191
-------
METHODOLOGY DOCUMENTATION
Table 5-62: DAYCENT Base Mean Yields, and Differences from Mean Yield for Mitigation Strategies, by Year
(Metric Tons of Grain per Hectare) (continued)
Mitigation Strategy
2015
2020
2025
2030
Soybeans
Base yield3
2.95
2.94
2.92
2.92
No till
-0.02
-0.02
-0.01
-0.01
Optimal nitrogen fertilization13
0.06
0.07
0.07
0.07
Split nitrogen fertilization
0
0
0
0
100% residue incorporation
0.02
0.02
0.02
0.02
Nitrification inhibitors
0
0.01
0.01
0.01
Reduced fertilization
-0.01
-0.01
-0.01
-0.01
Increased fertilization
0.01
0.01
0.01
0.01
a Base yield values represent production with no mitigation options, subsequent rows under each commodity presents relative
change from base yield values. bOptimal nitrogen fertilization simulates providing sufficient N to meet crop needs in each daily
timestep of the DAYCENT model. This option was excluded from the main MAC analysis; values are presented for information
only.
5.3.2.4 Economic Characteristics of Options
In cases where yields decrease as a result of the mitigation option, we valued the reduction in production at
the market price and included no tax or other benefits. Details of the costs associated with each mitigation option
are listed below.
• No Till: Reductions in labor costs are associated with the reduction in field preparation that are based on
data from USDA's Agricultural Resource Management Survey data, which provides labor estimates for
conventional and conservation tillage on both irrigated and rain-fed land by major crop. Conversion to no
till would require purchasing equipment for direct planting. However, if this equipment is purchased in
place of equipment used for traditional tillage, there may be little incremental capital cost associated with
no till. Some crop budgets actually indicate lower capital costs for no till because fewer passes over the
field are needed, which leads to reduced equipment depreciation. Thus, no incremental capital costs were
assumed for no-till adoption.
• Reduced Fertilization: This option reduces operation costs by the value of the fertilizer withheld, which
varies across crop/region/year combinations based on baseline fertilizer application rates and prices. We
assumed that other costs of fertilizer application are unchanged.
• Increased Fertilization: This option increases operation costs by the value of the additional fertilizer used,
which varies across crop/region/year combinations based on baseline fertilizer application rates and
prices. We assumed that other costs of fertilizer application are unchanged.
• Split Nitrogen Fertilization: This option was assumed to require 14% more labor to account for additional
passes over the fields to apply fertilizer multiple times rather than only once.
• Nitrification Inhibitors: The costs of this option include the cost of the nitrification inhibitor, assumed to
be $20 per hectare for the United States (Scharf et al., 2005) and scaled to other regions. We assumed
that this option does not affect any other costs of fertilizer application.
• 100% Residue Incorporation: No cost is associated with this option.
5-192
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Description of the DAYCENT Model
This analysis used the DAYCENT ecosystem model to estimate crop yields, N2O and Cm emissions, and soil
carbon stocks. DAYCENT is a process-based model (Parton et al., 1998; Del Grosso et al., 2001) that simulates
biogeochemical carbon and nitrogen fluxes between the atmosphere, vegetation, and soil by representing the
influence of environmental conditions on these fluxes, including soil characteristics and weather patterns, crop
and forage qualities, and management practices. DAYCENT uses the soil carbon modeling framework developed
in the Century model (Parton et al., 1987,1988,1994; Metherell et al., 1993), with refinement to simulate
carbon dynamics at a daily time-step. Key processes simulated by DAYCENT include crop production, organic
matter formation and decomposition, and soil water and temperature regimes by layer, in addition to
nitrification and denitrification processes. DAYCENT has been evaluated in several studies (e.g., Del Grosso et
al., 2002, 2005, 2009) and has also been recently adopted by the EPA to develop the soil carbon and soil
estimates for the annual Inventory of U.S. Greenhouse Gas Emissions and Sinks (EPA, 2013) submitted to the
UNFCCC.
DAYCENT simulated crop yields, direct N2O and Cm emissions, and soil organic carbon stock changes at a
0.5°grid resolution. Indirect N2O emissions were estimated base on simulated amounts of nitrate leaching,
nitrogen runoff in overland water flow, and NOx emissions from a site according to the DAYCENT model
combined with the IPCC default factors for indirect N2O emissions (De Klein et al., 2006). To represent the
longer term effect of cultivation on soil C, simulations started in 1700 after a simulation of 3,000 years of native
vegetation, which is a similar procedure to the methods applied in the U.S. Greenhouse Gas Inventory for
agricultural soil carbon and N2O (EPA, 2013). Weather data were based on a dataset generated by the North
American Carbon Program at a 0.5° resolution with daily minimum and maximum temperatures and daily
precipitation. The soils data were based on the FAO Digitized Soil Map of the World (FAO, 1996). Major cropland
areas of the world were simulated according to a global cropland map developed by Ramankutty et al. (2008),
with grid cells with less than 5% cropland area excluded in the analysis.
We established projected baseline emissions and crop production for both irrigated and rainfed production
systems using simulated yields and GHG emission rates from the DAYCENT model and adjusting with projected
growth rates of these production systems from IFPRI's IMPACT model. In DAYCENT, crop production areas were
held constant at the 2010 level to obtain the biophysical effects of management practice changes on crop yields
and GHG fluxes. Projected acreage changes from the IMPACT model reflect socioeconomic drivers such as
population growth and technological changes to meet global food demand (Nelson et al., 2010).
The croplands analysis, through its use of a dynamic biophysical crop model, captures the effect of fertilizer
applications on plant growth and, hence, on CO2 sequestration and soil carbon storage. These effects, where
applicable, are important. However, soil carbon levels tend to stabilize in a period of a decade or so, so the
resulting abatement potential in later years is reduced.
Several limitations are worth noting in the croplands analysis. Coverage was limited to major crops, and
pasture was excluded. As a result, the mitigation potential, compared with the sector baseline as a whole, is
limited.
5.3.2.5 References
Bronson, K.F., A.R. Mosier, and S.R. Bishnoi. 1992. Nitrous oxide emissions in irrigated corn as affected by
nitrification inhibitors. Soil Science Society of America Journal, 56,161-165.
De Klein, C., R.S.A. Novoa, S. Ogle, K.A. Smith, et al. 2006. Chapter 11: N2O emissions from managed soil, and CO2
emissions from lime and urea application. In 2006 IPCC guidelines for national greenhouse gas inventories, Vol.
4: Agriculture, forestry and other land use, edited by S. Eggleston, L Buendia, K. Miwa, T. Ngara and K. Tanabe.
Kanagawa, Japan: IGES.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-193
-------
METHODOLOGY DOCUMENTATION
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: M. Schaffer et al.
(eds.), Modeling Carbon and Nitrogen Dynamics for Soil Management. Boca Raton, FL: CRC Press.
Del Grosso, S.J., D.S. Ojima, W.J. Parton, A.R. Mosier, G.A. Peterson, and D.S. Schimel. 2002. Simulated effects of
dryland cropping intensification on soil organic matter and greenhouse gas exchanges using the DAYCENT
ecosystem model. Environmental Pollution, 116, S75-S83.
Del Grosso, S.J., A.R. Mosier, W.J. Parton, and D.S. Ojima. 2005. DAYCENT model analysis of past and contemporary
soil N20 and net greenhouse gas flux for major crops in the USA. Soil Tillage and Research, 83, 9-24.
Del Grosso, S.J., D.S. Ojima, W.J. Parton, E. Stehfest, M. Heistemann, B. DeAngelo, and S. Rose. 2009. Global Scale
DAYCENT Model Analysis of Greenhouse Gas Mitigation Strategies for Cropped Soils. Global and Planetary
Change, 67,44-50.
Del Grosso, S., S.M. Ogle, W.J. Parton, and F.J. Breidt. 2010. Estimating uncertainty in N2O emissions from US
cropland soils. Global Biogeochemical Cycles, 24, GB1009, doi:10.1029/2009GB003544
Food and Agriculture Organization of the United Nations. 1996. The Digitized Soil Map of the World Including
Derived Soil Properties. CD ROM. Rome: Food and Agriculture Organization.
Food and Agriculture Organization of the United Nations. 2012. World Agriculture towards 2030/2050 Report.
Food and Agriculture Organization of the United Nations. Available online at
http://www.fao.ore/docrep/016/apl06e/apl06e.pdfJ
Food and Agriculture Organization of the United Nations. 2016. FAOSTAT Statistical Database. Food and
Agriculture Organization of the United Nations. Available online at http://faostat.fao.org J
International Fertilizer Industry Association. 2016. IFADATA Statistical Database. International Fertilizer Industry
Association. Available online at http://www.fertilizer.org/ifa/ifadata/search J
International Food Policy Research Institute. 2016. International Food Policy Research Institute, Impact Model
Growth Rate Spreadsheet for Livestock Populations e-mailed from Timothy Sulser of IFPRI to Katrin Moffroid
of ICF, September 30, 2016.
International Food Policy Research Institute. 2017. Impact Model Growth Rate Spreadsheet for Crop Area and Yield
e-mailed from Timothy Sulser of IFPRI to Katrin Moffroid of ICF, February 27, 2017.
Intergovernmental Panel on Climate Change. 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.
Metherell, A.K., L.A. Harding, C.V. Cole, and W.J. Parton. 1993. CENTURY Soil Organic Matter Model Environment.
Agroecosystem version 4.0. Technical documentation, GPSRTech. Report No. 4, Ft. Collins, CO: USDA/ARS.
Mosier, A.R., J.M. Duxbury, J.R. Freney, O. Heinemeyer, and K. Minami. 1998. Assessing and mitigating N2O
emissions from agricultural soils. Climatic Change, 40, 7-38.
Parton, W.J., D.S. Schimel, C.V. Cole, and 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, and C.V. Cole. 1988. Dynamics of C, N, P, and S in grassland soils: a model.
Biogeochemistry, 5,109-131.
Parton, W.J., D.S. Ojima, C.V. Cole, and D.S. Schimel. 1994. A General Model for Soil Organic Matter Dynamics:
Sensitivity to litter chemistry, texture and management, in Quantitative Modeling of Soil Forming Processes.
Special Publication 39, Soil Science Society of America, Madison, Wl, 147-167.
Parton, W.J., M.D. Hartman, D.S. Ojima, and D.S. Schimel. 1998. DAYCENT: Its Land Surface Submodel: Description
and Testing. Global and Planetary Change, 19, 35-48.
5-194
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Paustian, K., H.P. Collins, and E.A. Paul. 1997. Management controls on soil carbon. In E.A. Paul, K. Paustian, and
C.V. Cole (eds.), pp. 15-49. Soil Organic Matter in Temperate Agroecosystems: Long-Term Experiments in
North America. Boca Raton, FL: CRC Press.
Ramankutty, N. and J.A. Foley. 1998. Characterizing patterns of global land use: an analysis of global croplands
data. Global Biogeochemical Cycles, 12(4), 667-685.
Ramankutty, N., A.T. Evan, C. Monfreda, and J.A. Foley. 2008. Farming the planet: 1. Geographic distribution of
global agricultural lands in the year 2000. Global Biogeochemical Cycles, 22, GB1003,
doi: 10.1029/2007GB002952
Scharf, P., L. Mueller, and J. Medeiros. 2005. Making Urea Work in No-Till. Grant progress report. Columbia, MO:
University of Missouri, Missouri Agricultural Experiment Station. Available online at
http://aes.missouri.edu/pfcs/researcli/prop604a.pdfJ
Six, J., E.T. Elliott, and K. Paustian. 2000. Soil macroaggregate turnover and microaggregate formation: A
mechanism for C sequestration under no-tillage agriculture. Soil Biology and Biochemistry, 32, 2099-2103.
Smith, P., D. Martino, Z. Cai, D. Gwary, H. Janzen, P. Kumar, B. McCarl, S. Ogle, F. O'Mara, C. Rice, B. Scholes, and O.
Sirotenko. 2007. Agriculture. In Climate Change 2007: Mitigation. Contribution of Working Group III to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change. B. Metz, O.R. Davidson, P.R.
Bosch, R. Dave, and L.A. Meyer (eds.), Cambridge, UK: Cambridge University Press.
Tenkorang, F. and J. Lowenberg-DeBoer. 2008. Forecasting Long-term Global Fertilizer Demand. Rome: Food and
Agriculture Organization of the United Nations.
U.S. Environmental Protection Agency. 2012. Global Anthropogenic Non-CC>2 Greenhouse Gas Emissions: 1990-
2030. EPA #430-R-12-006. Washington, DC: EPA.
U.S. Environmental Protection Agency. 2013. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2011.
Washington, DC: EPA.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-195
-------
METHODOLOGY DOCUMENTATION
5.3.3 Rice Cultivation
Rice cultivation consists of Cm emissions from rice production. The anaerobic decomposition of organic
matter in flooded rice fields produces Cm. When fields are flooded, aerobic decomposition of organic material
gradually depletes the oxygen present in the soil and flood water, causing anaerobic conditions in the soil to
develop. Once the environment becomes anaerobic, CFU is produced through anaerobic decomposition of soil
organic matter by methanogenic bacteria. Several factors influence the amount of Cm produced, including water
management practices and the quantity of organic material available to decompose.
N2O emissions from rice cultivation associated with fertilizer, crop residues, and other N additions to soils are
captured in Section 5.3.3.1, Projections Methodology. This section evaluates CH4 emissions from rice cultivation as
described above. Conversely, the mitigation analysis in Section 5.3.3.2, Mitigation Options Considered for Rice
Cultivation, presents both N2O and CFU mitigation from rice cultivation together.
5.3.3.1 Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite
Emission Projections, for additional information). Activity data for rice cultivation included rice area harvested
from FAO (2016), type of water management regime and rice-growing season length from GRiSP (2013), and
growth rate of rice area harvested from IFPRI's IMPACT model (2017).
The Tier 1 basic equation to estimate CH4 is as follows:
CH4 Emissions from Rice Cultivation (Gg CH4) = Y.i,j,k{EFi,j,k * h,],k * ^i,j,k * 10~6) (5.29)
where:
EFij,k = A daily emission factor for i,j, and k conditions (kg CFU ha 1 day
tij,k = Cultivation period of rice for i, j, and k conditions (days)
Aij,k = Annual harvested area of rice for i,j, and k conditions (ha yr"1)
i, j, and k = Represent different ecosystems, water regimes, type and amount of organic amendments, and
other conditions under which CFU emissions from rice may vary
Rice emissions vary according to the conditions under which rice is grown. Using the approach outlined above,
the harvested area can be subdivided by different growing conditions (e.g., water management regime) and
multiplied by an emission factor appropriate to the conditions. The sum of these individual products represents
the total national estimate.
In practice, it is difficult to obtain specific emission factors for each commonly occurring set of rice production
conditions in a country, so the 2006IPCC Guidelines instruct countries to first obtain a baseline emission factor, the
seasonally integrated emission factor for continuously flooded fields without organic amendments (EFC). Different
scaling factors were then applied to this seasonally integrated emission factor to obtain an adjusted seasonally
integrated emission factor for the harvested area as follows:
EFi = EFC * SFW * SF0 * SFs (5.30)
5-196
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
where:
EFi = Adjusted seasonally integrated emission factor for a particular harvested area
EFC = Seasonally integrated emission factor for continuously flooded fields without organic amendments
SFW = Scaling factor to account for the differences in ecosystem and water management regime
SF0 = Scaling factors for organic amendments (should vary for both type and amount of amendment
applied)
SFs = Scaling factor for soil type, if available.
For emissions from rice cultivation, activity data (i.e., rice area harvested), water management regime, and the
length of the rice-growing season were the driving factors for determining emissions.
Activity Data
Historical
• Area harvested for rice cultivation from 1990 through 2014 was obtained from FAO's FAOSTAT Statistical
Database (FAO, 2016). If the harvested area was not available from FAO statistics, it was assumed that the
country did not grow rice.
• Information on the type of water management regime (irrigated, rainfed lowland, upland, or deepwater)
was obtained from the Global Rice Science Partnership (GRiSP) Rice Almanac, 4th edition (GRisP, 2013). If a
country's water management regime was not available in the Rice Almanac, the EPA used the regime data
from the IRRI World Rice Statistics, which is consistent with previous updates to this report (IRRI, 2009).
• Information on the length of the rice-growing season in each country was also obtained from the Rice
Almanac (GRiSP, 2013). If a country's rice-growing season lengths were not available in the Rice Almanac,
the EPA used the rice-growing season length from the IRRI World Rice Statistics, which is consistent with
previous updates to this report (IRRI, 2009).
Projected
• A 5-year growth rate of rice area harvested data was estimated for 2015 through 2050 by using the
country rice area harvested projections developed by IFPRI's IMPACT model (IFPRI, 2017).103
• For countries where projected area data were not available, regional (i.e., Africa, Americas, Asia, Europe,
or Oceania) area 5-year growth rates from the same IFPRI IMPACT model (IFPRI, 2017) were used.
• A 1-year area growth rate based on IFPRI data was applied to historical 2014 emissions to develop 2015
projections. Five-year area growth rates were then applied to the 2015 emissions attributed to rice
cultivation to develop projections at 5-year intervals through 2050.
Emission Factors
Historical and Projected
• Country-applicable daily emission factors were developed for each of the five main water management
types: continuously flooded, irrigated, rainfed lowland, upland, or deepwater.
• The starting point (baseline) emission factor (1.3 kg CFU/ha-day) obtained from the 2006IPCC Guidelines
(IPCC, 2006) assumes fields with no flooding for less than 180 days prior to rice cultivation and
continuously flooded fields during rice cultivation without organic amendments. Scaling factors from the
2006 IPCC Guidelines (IPCC, 2006) were then applied to adjust the starting-point emission factor for each
of the other water regimes based on the factors for the pre-cultivation and cultivation periods. A scaling
factor of 1.22 was used for the pre-cultivation period for all water regimes except upland cultivation. The
103 The IFPRI IMPACT model incorporates supply and demand parameters to determine the estimated growth rates. These
parameters include the feed mix applied according to relative price movements, international trade, national income,
population, and urban growth rates, as well as anticipated changes in these rates over time.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-197
-------
METHODOLOGY DOCUMENTATION
scaling factors 0.78, 0.28, 0.31, and 0 were used for the rice cultivation period for irrigated, regular rainfed
(lowland), deepwater, and upland, respectively.
- The combination of all the above adjustment factors provided the adjusted country-specific emission
factors used in the emission equation above.
- A weighted average of the water-regime-based emission factors for each country was calculated
based on the percentage of each regime in that country. This weighting gives the combined final daily
emission factor for each country.
Season Lengths
Country-applicable season lengths were based on the Rice Almanac (GRiSP, 2013). Season lengths were given
as month ranges for planting and harvest (e.g., planting: February through March, harvest: mid-June through mid-
July). To estimate the number of days corresponding to the given range, the EPA made the following assumptions:
• A single month given (e.g., March, rather than a range, March-April) was assumed to refer to the 15th of
that month; "mid" refers to the 15th of the month; "early" refers to the 1st of the month; and "late"
refers to the last day of the month.
• A range of months was assumed to refer to the 1st or 15th day of the month, falling in the approximate
middle of the range, as applicable. For example, April-May would return May 1st; April-June would
return May 15th; late November-January would return Jan 1st.
• For countries with more than one season per year (i.e., "main," "second"), season lengths were added.
For countries with early and late seasons, the longer of the two seasons was used. For countries where
GRiSP identified different rice-growing regions, the regional season lengths were averaged.
Area harvested data for 1990,1995, 2000, 2005, 2010, and 2014 were multiplied by the combined final daily
emission factor and by the season length.
Emission Reductions in Baseline Scenario
The methodology used for this source category does not explicitly model any emission reductions; however,
emission reductions are included to the extent they are reflected in country-reported data.
Uncertainties
Significant uncertainties exist in the CFU emission estimates from rice cultivation. The greatest uncertainties
are associated with the use of default emission factors. The IPCC emission factors are not country specific and are
adjusted for some parameters (e.g., water management) but not adjusted for other parameters (e.g., rationing).
Water regime information was not available for several other countries, and using the default emission factor for
these countries may lead to an overestimate of emissions. In addition, country-specific information is not readily
available on the amount of flooding and aeration in irrigated areas, so assumptions had to be developed based on
country conditions.
If a country-specific emission factor was not available and a country was used as a proxy for season length, the
same country proxy was used. Otherwise the baseline emission factor (1.3) was used. The following country
proxies were applied:
• Madagascar's emission factor was applied to Comoros.
• Malaysia's emission factor was applied to Brunei Darussalam.
Because of limited information, all irrigated land was assumed to be continuously flooded with no aeration.
This assumption is conservative and could lead to overestimates in emissions.
Also, no scaling adjustment was made to account for organic amendments, because of a lack of data on the
use of such amendments. This may result in an underestimate of emissions.
5-198
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
The rice season length is also an area of uncertainty, because many assumptions were made to turn a rough
estimate of month ranges into a specific number of days. In addition, a number of countries' rice season lengths
were proxied because of a lack of data, and these proxies for season length might not be accurate. For some
countries where FAO indicated that rice is grown, no season length data were available, and for some countries
the available data were problematic (e.g., planting dates overlapped with harvest dates). In both of these cases,
countries in the same region deemed to have similar climates or rice-growing schemes were used as proxies.
5.3.3.2 Mitigation Options Considered for Rice Cultivation
This section presents the methodology for estimating mitigation from the rice sector. Both N2O and Cm
mitigation are analyzed together, in contrast to the baseline projection where they are estimated separately. This
section estimates global non-CC>2 mitigation potential from rice production using DNDC, a gridded global
biophysical model. Using a biophysical model improves the analysis by incorporating soil characteristics and
dynamic soil processes, as well as crop-specific nutrient uptake and plant physiology.
The mitigation options included in this analysis were based on a review of the literature to identify the most
promising options, while also considering data availability and the potential for modeling within DNDC. We then
analyzed 26 mitigation scenarios using DNDC 9.5.104 The scenarios addressed management techniques in various
combinations hypothesized to reduce GHG emissions from rice systems: water management regime (continuous
flooding, mid-season drainage, dry seeding, alternate wetting and drying, and switching to dryland rice production
system), residue management (partial or total residue incorporation), tillage, and various fertilizer management
alternatives (ammonium sulfate in place of urea, urea with nitrification inhibitor, slow release urea, 10% reduced
fertilizer, 20% reduced fertilizer, and 30% reduced fertilizer).
The water management system under which rice is produced is one of the most important factors influencing
Cm emissions. Specifically, switching from continuous flooding of rice paddy fields to draining flooded fields
periodically during the growing season—a water conservation practice that is increasingly adopted in the baseline
to reduce water use—would significantly reduce CH4 emissions. Other practices (e.g., fertilizer applications, tillage
practices, and residue management) also alter the soil conditions and hence affect crop yields and the soil carbon-
and nitrogen-driving processes such as decomposition, nitrification, and denitrification (Neue and Sass, 1994; Li et
al., 2006).
Table 5-63 reports the 26 alternative rice management scenarios simulated using DNDC.105
104 Note that 38 different scenario names are reported in the outputs. Because water management practices were assumed not
to affect nonirrigated rice emissions, the simulation results for options combined with continuous flooding or midseason
drainage are the same for nonirrigated rice. The analogous options that alter fertilizer and other management practices but do
not affect water management were identified as beginning with "base" rather than "cf" or "md."
105 Another potential mitigation option is varietal selection toward cultivars that have lower emissions per unit of output.
However, data were insufficient to represent such changes within DNDC for this study.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-199
-------
METHODOLOGY DOCUMENTATION
Table 5-63: Alternative Rice Management Scenarios Simulated Using DNDC
Abbreviation
Scenario
Flooding
Residue
Incorporation, %
Alternative
Management
Fertilization
cf_r50
Continuous flooding
CF
50
Conventional
cf_rl00
Continuous flooding,
100% residue
incorporation
CF
100
Conventional
cf_r50_amsu
Continuous flooding,
ammonium sulphate
fertilizer
CF
50
Ammonium
sulfate
cf_r50_ninhib
Continuous flooding,
nitrification inhibitor
fertilizer
CF
50
Nitrification
inhibitor
cf_r50_slowrel
Continuous flooding,
slow release fertilizer
CF
50
Slow release
cf_r50_notill
Continuous flooding, no
till
CF
50
N o till
Conventional
cf_r50_f70
Continuous flooding,
30% reduced fertilizer
CF
50
—
30% reduced
cf_r50_f90
Continuous flooding,
10% reduced fertilizer
CF
50
—
10% reduced
cf_r50_auto
Continuous flooding,
auto-fertilization to
maximize yields
CF
50
Automatically
adjusted by DNDC
to maximize yields
md_r50
Mid-season drainage
MD
50
Conventional
md_rl00
Mid-season drainage
w/100% residue
incorporation
MD
100
Conventional
md_r50_amsu
Mid-season drainage,
ammonium sulphate
fertilizer
MD
50
Ammonium
sulfate
md_r50_ninhi
b
Mid-season drainage,
nitrification inhibitor
fertilizer
MD
50
Nitrification
inhibitor
md_r50_slowr
el
Mid-season drainage,
slow release fertilizer
MD
50
Slow release
md_r50_notill
Mid-season drainage, no
till
MD
50
N o till
Conventional
md_r50_f70
Mid-season drainage,
30% reduced fertilizer
MD
50
30% reduced
md_r50_f90
Mid-season drainage,
10% reduced fertilizer
MD
50
10% reduced
(continued)
5-200
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-63: Alternative Rice Management Scenarios Simulated Using DNDC (continued)
Abbreviation
Scenario
Flooding
Residue
Incorporation, %
Alternative
Management
Fertilization
md_r50_ds
Mid-season drainage,
dry seeding
MD
w/DS
50
—
Conventional
md_r50_auto
Mid-season drainage,
auto-fertilization to
maximize yields
MD
50
Automatically
adjusted by DNDC
to maximize yields
awd_r50
Alternate wetting &
drying (AWD)
AWD
50
—
Conventional
awd_r50_ninh
ib
AWD w/nitrification
inhibitor
AWD
50
—
Nitrification
inhibitor
awd_r50_slow
re I
AWD w/slow release
AWD
50
—
Slow release
base_r50_ds
Dry seeding
DS
50
—
Conventional
base_r50_f80_
ds
Dry seeding, 20%
reduced fertilizer
DS
50
—
20% reduced
dry_r50
Dryland rice
dryland
rice
50
Conventional
dry_r50_f80
Dryland rice, 20%
reduced fertilizer
dryland
rice
50
20% reduced
For nonirrigated rice, there is no difference between scenarios with alternative water management. Thus, we
refer to those scenarios for the nonirrigated rice with "base_" in front rather than "cf" or "md."
Most of the major rice-producing countries have some mix of flood regimes in DNDC (see Table 5-64). To
determine baseline emissions for each country, simulation results were combined based on the fraction of rice
area in each rice category (deepwater, upland, rainfed, and irrigated) and flood regime for irrigated rice. For
instance, baseline emissions for Bangladesh were determined by averaging the results of the CF and MD scenarios
with 50% residue removal (cf_r50 * 0.2 + md_r50 * 0.8).
However, for the purposes of calculating emission reductions, we compared mitigation options with the
portions of the baseline to which they could potentially be applied rather than to the national weighted average.
For instance, application of the mitigation option of switching to ammonium sulphate fertilizer (cf_r50_amsu) was
compared with baseline emissions from continuously flooded rice with conventional fertilizer (cf_r50) and with
baseline emissions from rice managed using mid-season drainage with conventional fertilizer (md_r50) rather than
being compared with the baseline weighted average emissions per hectare. We made this comparison to better
represent the mitigation potential from adopting each mitigation option on each baseline subcategory. As an
example, an option such as cf_r50_amsu may result in emission reductions relative to cf_r50 but increases in
emissions relative to md_r50 (and possibly the weighted baseline emissions as well) in many countries. This results
from the change in water management regime in moving from mid-season drainage to continuous flooding,
whereas we are trying to isolate the effects of changing fertilizer for a given baseline water management strategy
in that example.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-201
-------
METHODOLOGY DOCUMENTATION
Table 5-64: Rice Management Techniques
Management
Technique
Description
Rice flooding
Continuous
flooding (CF)
Rice paddy is flooded on planting date and drained 10 days before harvest date; applies to
both irrigated and rainfed rice
Mid-season
drainage (MD)
Rice paddy is drained twice during growing season for 8 days; final drainage is 10 days before
harvest date—applies only to irrigated rice
AWD
Rice paddy is initially flooded to 10 cm; water level is reduced at rate of -0.5 cm/day till to -5
cm and then reflooded at rate of 0.5 cm/day till to 10 cm—applies only to irrigated rice
Dryland rice
All irrigated and rainfed rice are swapped for dryland rice; no flooding occurs
Rice seeding
Direct seeding
Rice paddy is flooded 40 days after planting date and drained 10 days before harvest date-
applies to both irrigated and rainfed rice
Residue incorporation
50%
50% of above-ground crop residue is removed; remaining residue is incorporated at next
tillage
100%
All residue remains in place and is incorporated at next tillage
Tillage
Conventional
Before first crop in rotation tillage to 20 cm depth; subsequent tillages (following each crop in
rotation) to 10 cm depth
No till
Tillage only mulches residue
Fertilizer
Conventional
Fertilizer nitrogen applied as urea on plant date using a crop-specific rate
Ammonium
sulfate
Fertilizer nitrogen applied as ammonium sulfate on plant date using a crop-specific rate
Nitrification
inhibitor
Nitrification inhibitor is used with urea; reduced conversion of NH4 to N03 is simulated with
60% efficiency over 120 days
Slow release
Slow-release urea applied on planting date; nitrogen is released over 90 days at a linear rate
10% reduced
Crop-specified baseline fertilizer nitrogen rate is reduced by 10% (applied as urea)
20% reduced
Crop-specified baseline fertilizer nitrogen rate is reduced by 20% (applied as urea)
30% reduced
Crop-specified baseline fertilizer nitrogen rate is reduced by 30% (applied as urea)
Auto
fertilization
Fertilizer nitrogen is applied at the rate that maximizes crop yield
5.3.3.3 Technical Characteristics of Options
Rice management techniques are defined in further detail in Table 5-64.
• Applicability: All options applicable for a given cropping pattern were assumed available to all acres in all
countries. However, water management options (e.g., shifting from continuous flooding to midseason
drainage) are only applicable to irrigated systems. No water management options are available for
rainfed, deepwater, or upland rice cultivation areas.
5-202
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Technical Efficiency: Technical efficiency was determined by the DNDC model for each country,
production type, and water management combination for each mitigation option.
• Technical Lifetime: Technical lifetime is indefinite; no capital costs are included for which a lifetime must
be defined.
5.3.3.4 Economic Characteristics of Options
• Capital Cost: None of the options considered for this analysis were assumed to have any capital cost.
• Annual O&M Cost: these recurring costs reflect the changes in labor, fertilizer, and other inputs
associated with each option.
• Annual Benefits: Calculated based on changes in production associated with changes in yield, valued at
market prices.
Description of the DNDC Model
For this analysis, we used a modified version of the DNDC 9.5 global database to simulate crop yields and GHG
fluxes from global paddy rice cultivation systems. Details of management (e.g., crop rotation, tillage,
fertilization, manure amendment, irrigation, weeding, and grazing) have been parameterized and linked to the
various biogeochemical processes (e.g., crop growth, litter production, soil water infiltration, decomposition,
nitrification, denitrification, fermentation) embedded in DNDC (e.g., Li et al., 2004; Li et al., 2006; Li, 2011;
Abdalla et al., 2011; Giltrap et al., 2011; Dai et al., 2012).106 The DNDC 9.5 global database contains information
on soil characteristics, crop planted area, and management conditions (fertilization, irrigation, season, and
tillage) on a 0.5- by 0.5-degree grid cell of the world. The database is used to establish the initial conditions in
the model in 2000. The model considers all paddy rice production systems, including irrigated and rainfed rice,
and single, double, and mixed rice, as well as deep water and upland cropping systems. For this study, baseline
and mitigation scenario modeling was carried out for all rice-producing countries in the world that produce a
substantial quantity of rice.
The FAO country-level statistics (FAOSTAT 2010) were used to establish the harvested area for rice. The total
area was calculated for each country in the global database for each type and evenly distributed across all grid
cells within a country in the absence of subnational information. The global meteorological data from the
National Oceanic and Atmospheric Administration's National Centers for Environmental Prediction climate
reanalysis product were used to establish climate data for 2010 in the model. The 2010 climate data were used
for all model years. Planting and harvest dates were matched approximately to local growing season. Tillage,
flooding, and drainage dates for irrigated rice were established based on the planting dates.
A baseline scenario was established for each country using DNDC 9.5. Rice yields and GHG fluxes (Cm, direct
and indirect N2O, and changes in soil organic carbon) were simulated in the DNDC model for each grid cell, and
results were aggregated at the country level for irrigated, rainfed, deep water, and upland production systems
for each scenario, in both mean annual rates per hectare and mean annual national totals. Results were
reported for 2010 and by 5-year increments through 2030.
We adjusted results from DNDC with projected acreage of these production systems by IFPRI's IMPACT model.
In DNDC, rice production areas were held constant at the 2010 level to obtain the biophysical effects of
management practice changes on crop yields and GHG fluxes. Projected acreage changes from the IMPACT
model reflect socioeconomic drivers (such as population growth) and technological changes to meet the global
food demand (Nelson et al., 2010). The IMPACT modeling projects that while global rice production would
increase by 11% between 2010 and 2030, the total area dedicated to rice cultivation would decrease by 5%
during the same period because of productivity improvements.
106 The paddy-rice version of DNDC has been validated in many countries and world regions and is used for national trace gas
inventory studies in North America, Europe, and Asia (e.g., Smith et al., 2002; Follador et al., 2011; Leip et al., 2011; Li et al.,
2002; Cai et al., 2003; Li et al., 2005).
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-203
-------
METHODOLOGY DOCUMENTATION
5.3.3.5 References
Abdalla M., S. Kumar, M. Jones, J. Burke, and M. Williams. 2011. Testing DNDC model for simulating soil respiration
and assessing the effects of climate change on the CO2 gas flux from Irish agriculture. Global and
Planetary Change, 201178,106-115. doi:10.1016/j.gloplacha.2011.05.011
Cai, Z., T. Sawamoto, C. Li, G. Kang, J. Boonjawat, A. Mosier, R. Wassmann, and H. Tsuruta, 2003. Field validation of
the DNDC model for greenhouse gas emissions in East Asian cropping systems. Global Biogeochemical Cycles,
17(4), 1107, doi: 10.1029/2003GB002046
Dai, Z.H., C.C. Trettin, C.S. Li, H. Li, G. Sun, and D.M. Amatya. 2012. Effect of assessment scale on spatial and
temporal variations in CH(4), CO(2), and N(2)0 fluxes in a forested wetland. Water Air and Soil Pollution,
223(1),253-265 doi: 10.1007/sll270-011-0855-0
Follador, M., A. Leip, and L. Orlandini. 2011. Assessing the impact of cross-compliance effects on nitrogen fluxes
from European farmlands with DNDC-EUROPE. Environmental Pollution, 159(11), 3233-324.
doi:10.1016/j.envpol.2011.01.025
Food and Agriculture Organization of the United Nations. 2016. FAOSTAT Statistical Database. Available online at
http://faostat.fao.org J
Giltrap, D.L., C. Li, and S. Saggar. 2010. DNDC: A process-based model of greenhouse gas fluxes from agricultural
soils. Agriculture, Ecosystems & Environment, 136, 292-230.
Global Rice Science Partnership. 2013. Global Rice Science Partnership (GRiSP) rice almanac, 4th edition.
Philippines: International Rice Research Institute. Available online at
http://irri.org/resources/publications/books/rice-almanac-4th-edition cP
International Food Policy Research Institute. 2017. Impact model growth rate spreadsheet for crop area and yield.
E-mail from Timothy Sulser, IFPRI, to Katrin Moffroid, ICF.
Intergovernmental Panel on Climate Change. 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.
International Rice Research Institute. 2009. World Rice Statistics. Available online at
http://beta.irri.org/solutions/index.php?option=com content&task=view&id=250 cf
Leip, A., M. Follador, S. Tarantola, M. Busto, and N. Villa-Vialaneix. 2011. Sensitivity of the process-based model
DNDC on microbiological parameters. Nitrogen and Global Change - Key Findings and Future Challenge.
Edinburgh, United Kingdom.
Li, C. 2011. Mitigating greenhouse gas emissions from agroecosystems: Scientific basis and modeling approach. In
Understanding Greenhouse Gas Emissions from Agricultural Management (pp. 299-330), Guo, L., et al. (eds.).
ACS Symposium Series American Chemical Society: Washington, DC. doi:10.1021/bk-2011-1072.ch016
Li, C., A. Mosier, R. Wassman, Z. Cai, X. Zheng, Y. Huang, H. Tsuruta, J. Boonjawat, and R. Lantin. 2004. Modelling
greenhouse gas emissions from rice-based production systems sensitivity and upscaling. Global
Biogeochemical Cycles, 18,1-19.
Li, C., J Qiu, S. Frolking, X. Xiao, W. Salas, B. Moore III, S. Boles, Y. Huang, and R. Sass. 2002. Reduced methane
emissions from large-scale changes in water management in China's rice paddies during 1980-2000.
Geophysical Research Letters, 29(20). doi:10.1029/2002GL015370
Li, C., S. Frolking, X. Xiao, B. Moore III, S. Boles, J. Qiu, Y. Huang, W. Salas, and R. Sass. 2005. Modeling impacts of
farming management alternatives on CO2, CH4, and N2O emissions: A case study for water management of rice
agriculture of China. Global Biogeochemical Cycles, 19. doi:10.1029/2004GB002341
Li, C., W. Salas, B. DeAngelo, and S. Rose. 2006. Assessing alternatives for mitigating net greenhouse gas emissions
and increasing yields from rice production in China over the next twenty years. Journal of Environmental
Quality, 35, 1554-1565. doi:10.2134/jeq2005.0208
5-204
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Nelson, G.C., M.W. Rosegrant, A. Palazzo, I. Gray, C. Ingersoll, R. Robertson, S. Tokgoz, T. Zhu, T.B. Sulser, C.
Ringler, S. Msangi, and L. You. 2010. Food Security, Farming, and Climate Change to 2050: Scenarios, Results,
Policy Options. Washington, DC: International Food Policy Research Institute.
Smith, W.N., R.L. Desjardins, B. Grant, C. Li, R. Lemke, P. Rochette, M.D. Corre, and D. Pennock. 2002. Testing the
DNDC model using N2O emissions at two experimental sites in Canada. Canada Journal of Soil Science, 82, 365-
374.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-205
-------
METHODOLOGY DOCUMENTATION
5.3.4 Other Agriculture
The source category solely comprises countries that report data to the UNFCCC database. The EPA did not
perform Tier 1 calculations for other agriculture sources, which include prescribed burning of savannas and field
burning of agricultural residues. The EPA obtained historical values for 1990 through 2012 and held 2015 through
2050 values constant at 2012 levels for each country.
5-206
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
This section presents the methodology used to estimate GHG emissions and mitigation potential from the
following waste sector sources:
• landfilling of solid waste (Cm)
• wastewater management (Cm, N2O)
• other waste (Cm, N2O) (projections only)
5.4.1 Landfills
Landfills produce CH4 in combination with other landfill gases (LFGs) through the natural process of bacterial
decomposition of organic waste under anaerobic conditions. LFGs are generated over a period of several decades,
with gas flows usually beginning 1 to 2 years after the waste is put in place. Cm makes up approximately 50% of
LFG. The remaining 50% is CO2 mixed with small quantities of other gases, including volatile organic compounds
(VOCs). The amount of CH4 generated by landfills per country is determined by a number of factors that include
population size, the quantity of waste disposed of per capita, composition of the waste disposed of, and the waste
management practices applied at the landfill. Changes in these key factors drive projected trends in CH4 emissions.
5.4.1.1 Landfill Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite
Emission Projections, for additional information). Activity data for landfilling of solid waste included population
data from the UN (2015); gross domestic product (GDP) data from USDA (2016) or the World Bank (2017); GDP
growth rates by country from OECD (2014); municipal solid waste (MSW) generation per capita and MSW landfill
rates from OECD.stat (2016), Eurostat (2016) or UN Statistics Division (2016); and industrial waste generation rates
from OECD (2011) or the China Statistical Yearbook (NBSC, 2015).
Emission estimates for this source are based on the IPCC First Order Decay (FOD) method using a mix of IPCC
default and country-specific input data. This method assumes that the degradable organic carbon (DOC) in landfills
decays under anaerobic conditions and releases Cm over time. Part of the Cm generated is oxidized or can be
recovered for energy or flaring; as a result, the Cm actually emitted will be less than the amount generated. The
2006 IPCC Tier 1 equation used is below.
CHAEmissions =
^CH_generated. ; -ft
'(1 -OX,) (5.31)
where:
CH4 Emissions = Cm emitted in year T, Gg
T = Inventory year
x = Waste category or type/material
Rt = Recovered Cm in year T, Gg
OXt = Oxidation factor in year T(fraction)
For further explanation regarding the methodology, please refer to Chapter 3, Solid Waste Disposal, of the 2006
IPCC Guidelines (IPCC, 2006).
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-207
-------
METHODOLOGY DOCUMENTATION
Solid waste generation, disposal, and emissions were analyzed separately for municipal solid waste (MSW) and
industrial solid waste. MSW is generally defined as waste collected by municipalities or other local authorities and
typically includes household waste, garden (yard) and park waste, and commercial/institutional waste. Industrial
waste includes organic, process waste generated by industry that is not included in the MSW stream. Industrial
waste generation and composition vary depending on the type of industry and processes/technologies in the
concerned country. Only those industrial wastes that are expected to contain DOC and fossil carbon should be
considered for the purpose of estimating emissions from waste. These industries primarily include food/beverage
manufacturing, wood product manufacturing, textile manufacturing, paper manufacturing, and construction and
demolition waste.
Activity Data
Historical
The key activity data driving emissions from landfilling of MSW are population, the rate of solid waste
generation (usually expressed in kilograms per person per day), and the rate of landfill disposal. Similarly, for
industrial solid waste, the key activity data are GDP, the rate of industrial process waste generation, and the rate of
landfill disposal for industrial waste. To develop the Tier 1 estimates, the analysis required historical activity data
going back at least 50 years, because of the nature of landfill emissions and the FOD methodology. These data
were developed using a combination of country-reported data and IPCC default assumptions.
• Population data were obtained from the United Nations Department of Economic and Social Affairs,
Population Division, World Population Prospects; custom data acquired via the website (UN, 2015).
Population data for each country were obtained in 5-year increments from 1950 through 2050.
Population data for the years between the reported values were linearly interpolated.
• Annual real GDP data by country from 1960 through 2050 were pulled from either the U.S. Department of
Agriculture (for years 1980 through 2030) (USDA, 2016) or the World Bank (for years 1960 through 2030)
(World Bank, 2017). When GDP data were not available for a country before 1980, it was back casted
using either the USDA world growth rate (USDA, 2016) or the World Bank world growth rate (World Bank,
2017).
• MSW generation per capita data were available from the OECD Questionnaire on the state of the
environment: municipal waste generation and treatment (OECD.stat, 2016); Eurostat, municipal waste
generation and treatment (Eurostat, 2016); and the United Nations Statistics Division (UNSD),
environmental statistics database: municipal waste collected (UNSD, 2016). These data provide country-
reported information for 75 countries from 1990 through 2015. Most countries do not have data reported
for every year, so missing values were linearly interpolated. To complete the time series back to 1960, we
assumed a constant per capita value based on the last reported year available in the country-reported
historical data. For countries with no reported data, IPCC default values for the year 2000, were used to
specify waste generation per capita, and the rest of the historical time series similarly assumed constant
waste generation historically. IPCC default values were specified for regions and mapped to individual
countries (IPCC Table 2.1).
• Total annual waste generation was calculated as the product of population and waste per capita.
• MSW composition was based on IPCC default values specified by region, which were mapped to individual
countries (IPCC Table 2.3). The composition does not vary by year.
• MSW landfill rate data were taken from the OECD Questionnaire on the state of the environment:
municipal waste generation and treatment (OECD.stat, 2016) and Eurostat, municipal waste generation
and treatment (Eurostat, 2016). These data provide country-reported information for 43 countries over
the period 1990 through 2015. Most countries do not have data reported for every year, so missing values
were linearly interpolated. To complete the time series back to 1960, the EPA assumed the same disposal
5-208
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
rate as the last year of reported data. For countries with no reported data, the year 2000 IPCC default
value was selected as the landfill disposal rate for all years. IPCC default values were specified for regions
and mapped to individual countries (IPCC Table 2.1).
• Industrial waste generation data from key industrial sources of organic process waste were available from
OECD, Environment Statistics Database, waste generation by sector (OECD, 2011). Data for China were
obtained from the China Statistical Yearbook (NBSC, 2015). These data provide tons of industrial waste for
key contributing industries, including food, textiles, wood, paper, and construction and demolition, for 32
countries from 1990 through 2012. Most countries have data for a subset of years, so missing values were
linearly interpolated.
• Per IPCC guidance, overall average industrial waste generation rates by country are based on reported
data. These rates do not vary over time because of limitations in the country-reported data. For countries
with no reported industrial waste data, input values were assigned based on similar or proximate
countries for which data are reported.
• Total annual industrial waste generation was calculated as the product of the generation rate and annual
GDP.
• Industrial waste composition was estimated as the proportion coming from each key industrial waste
source (i.e., food, textiles, wood, paper, and construction and demolition). The quantity of waste by
source category is required because the DOC values in the emission calculations vary by material. This
composition does not vary over time because of limitations in the country-reported data. For countries
with no reported industrial waste composition data, input values were assigned based on similar or
proximate countries for which data are reported.
• Industrial waste landfill rates were from country-reported data. When country-reported data were not
available, industrial landfill rates were assumed to be half of the country's reported MSW landfill rate.
Projected
• Consistent with historical population data, the population data projection relies on the United Nations
Department of Economic and Social Affairs, Population Division, World Population Prospects, custom data
acquired via the website (UN, 2015). Population data for each country were obtained in 5-year increments
from 1950 through 2050. Populations for years between the reported values were linearly interpolated.
• GDP data were projected beyond 2030 using the OECD country growth rates for available countries.
When country-specific growth rates were not available, the OECD world growth rate was used (OECD,
2014).
• MSW generation per capita is assumed to be constant and equal to the most recently available country-
reported or IPCC default data from either the last country-reported value or the IPCC default 2000 value
based on the annual percentage change in GDP per capita. Total future MSW was based on the product of
future population and the future waste generation rate.
• MSW landfill rates were assumed to be the same as the most recent year of country-reported data. When
no country-reported data were available, the IPCC default value was used.
• Because the average industrial waste generation rate by country does not vary over time, it was assumed
constant for projected years. Total future industrial waste generation is the product of future GDP and the
industrial waste generation rate.
• Projected industrial waste landfill rates were assumed to be half of the projected MSW landfill rates.
Emission Factors
Historical and Projected
The analysis used IPCC Tier 1 emission factors, including the following assumptions:
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-209
-------
METHODOLOGY DOCUMENTATION
• DOC by waste material was obtained from 2006IPCC Guidelines Table 2.4 for MSW and Table 2.5 for
industrial waste.
• The fraction of DOC dissimilated (DOCf) is a constant from 2006 IPCC Guidelines (0.5) for all materials and
countries.
• Cm generation rates by waste material and climatic zone were obtained from 2006 IPCC Guidelines Table
3.3.
• Oxidation (OX) and recovery (R) were assumed to equal zero, per IPCC guidance.
• Default values from 2006 IPCC Guidelines Table 3.1 and the IPCC Waste Model were used for estimated
distribution of landfill types used to manage landfilled waste (i.e., managed or unmanaged, deep or
shallow, and uncategorized).
Emission Reductions in Baseline Scenario
No emission reductions assumptions were applied in the methodology to estimate emissions from this source.
However, emission reductions are included to the extent they are reflected in country-reported data.
Uncertainty
Uncertainties in the emission estimates are directly related to the quality and availability of the input data
used to derive the emissions. As noted in 2006 IPCC Guidelines (2006), there are several key areas of uncertainty to
consider in the emission estimates, including the following:
• MSW generation subject to landfilling is based on the country's total population. To the extent that total
population does not reflect the population for which waste is collected, the analysis may overstate waste
generation.
• The analysis relies on 2006 IPCC Guidelines default assumptions to describe the Cm correction factor
(MCF) for each type of landfill in the model and the associated mix of different types of landfills used to
manage MSW. These characteristics are critical in determining emissions from a landfill; however, the
default values do not vary by country or over time. This introduces significant uncertainty, particularly at
the level of the individual country (versus global), because the default will not reflect the characteristics of
each country's landfills. It is not clear in which direction this uncertainty biases the results.
• Similarly, the composition of materials in MSW is based on broad IPCC default assumptions. In reality,
waste composition varies widely even within countries (e.g., between urban and rural populations,
between households with different incomes) as well as between countries. It is not clear in which
direction this uncertainty biases the results.
• Per IPCC Tier 1 guidance, the Tier 1 estimates do not include CH4 oxidation or recovery for flaring/energy,
which may result in overstated emissions.
• Industrial solid waste generation data for key industries that produce waste with DOC are based on limited
country-reported information. In addition, there are no country-reported data describing the management
of industrial solid waste. These limited data require assumptions (e.g., mapping data from one country to
multiple countries and using MSW landfilling rates as proxy) that introduce significant uncertainty into the
industrial waste estimates. An additional consideration is that it is not clear what is included in the
available industrial waste generation data. The analysis would ideally be limited to the process waste from
these industries (e.g., food waste and paper waste), rather than also including material such as office
waste, which is captured by the MSW stream. To the extent that the OECD industrial waste data are not
limited to process waste, the analysis likely overstates industrial waste generation.
5-210
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
A Note on Per Capita Waste Generation
As described above, the analysis assumed constant per capita waste generation based on the most recently
available country-reported or IPCC default data to complete the annual forecast and backcast that underlie the
Tier 1 approach. To the degree that per capita waste generation changes over time based on economic
development, generally, this approach may overstate historical generation and understate future estimates of
per capita waste generation at the country level. EPA also considered approaches to forecast per capita waste
generation based on a quantified relationship to GDP per capita. However, the empirical relationship between
per capita generation and economic indicators such as GDP per capita is not clear.
The literature generally agrees that GDP or income and waste generation are linked, but factors such as the
degree of urbanization and changes in waste management policies and infrastructure, which are by-products of
growth, complicate the relationship. For instance, "after 20 years of rapid economic growth through 1997, the
government of Taiwan enforced aggressive MSW management practices, which contributed to a large reduction
in the per capita MSW generation from 1.14 kg day-i in 1997 to 0.81 kg day-i in 2002, even though the economy
continued to grow" (Kawai and Tasaki, 2016).
Figure 5-4 below presents the relationship between GDP per capita and waste generation per capita for 2015
from EPA's Tier 1 methodology for 191 countries.
Figure 5-4: Waste Generation vs GDP (Per Capita, 2015)
TO
'a. TT
rc ru
O *
CU s_
Q. CU
c °-
o c
¦¦p Q
-------
METHODOLOGY DOCUMENTATION
5.4.1.2 Landfill Mitigation Options Considered
Model Facilities
This analysis considers abatement measures' impacts on three model facilities representing the solid waste
management alternatives with different levels of CH4-generating capacity:
• Open dump sites: Defined as solid waste disposal facilities where the waste is left uncompacted and
without cover.
• Basic landfills (also referred to as managed dump sites): Defined as solid waste disposal facilities where
the waste is compacted and covered but there are no additional engineered systems.
• Engineered sanitary landfills; Defined as facilities that include waste compaction and cover and are
designed and constructed with gas and leachate collection systems.
Various data sources were consulted to define the characteristics of the model facilities in the different
countries and regions, and a proxy country approach was used when data were not found for a given country.
Under this approach, countries for which no data were available were paired with a representative proxy country
based on similarities in socioeconomic and technology development trends that are closely correlated with a
country's waste composition. Furthermore, waste composition is the only parameter that affects both Lo (CFU
generation rate) and k constant (decay rate), two key factors used to estimate gas generation from the model
facilities.
To ensure project costs and benefits were comparable, we assumed annual waste acceptance rates were fixed
at 100,000 tonnes/yr, and the average depth of waste was assumed to be between 25 and 50 feet. Open dumps
have shallower waste depths sprawling over large areas. In contrast, basic and engineered landfills concentrate the
disposed of waste over a smaller area and at increased depths of between 40 and 50 feet. Facility CFU recovery
(also referred to as capture efficiency) also varies by landfill type and range from 10% for open dumps to 85% for
engineered landfills. Table 5-65 summarizes the standardized model facility assumptions.
Table 5-65: Model Facility Assumptions for International LFG Mitigation Options
Facility Type
No. Years
Open
Annual Waste
Acceptance Rate
(tonnes/yr)
Project Design
Area (Acres)
Waste Depth
(ft)
Facility CH4
Recovery
Engineered landfill
15
100,000
40
50
85%
Basic landfill
15
100,000
50
40
75%
Open dump
15
100,000
80
25
10%
To improve the heterogeneity in the break-even options across countries, we developed a dataset of country-
specific data of Lo (CFU generation potential) and k constant (decay rate) values, the two key parameters in the first
order decay model, which was used to estimate LFG generation. Both parameters were calculated based on the
composition of the waste being landfilled, which was determined by the country-specific socioeconomic
conditions, consumption patterns, and waste management practices. Therefore, the CFU generation results and,
consequently, the amount of CFU potentially mitigated by each LFG control measure are driven by the waste
composition, which is related to consumption patterns and socioeconomic conditions. We grouped the countries
according to the following logic:
First, we identified the decay constant (k) and Cm generation potential of waste (Lo) for 16 countries that
included at least one country within each major region (Africa, Asia, Caribbean/Central & South America, Eurasia,
Europe, Middle East, and North America). This information was obtained from a number of sources, including
5-212
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
international studies conducted by the World Bank, the EPA's voluntary program, the MSW Decision Support Tool
(DST), and other peer-reviewed literature.
Second, we used expert judgment, taking into consideration trends of socioeconomic and technological
development to associate countries with other countries for which we have Cm generation data (e.g., we have Cm
generation data for Jordan and considered that Algeria, Egypt, and South Africa have similar socioeconomic and
technological conditions). Alternatively, we have Cm generation data for Guinea, but we think that the
socioeconomic and technological conditions in Egypt, Algeria, and South Africa are closer to those in Jordan than to
those in Guinea. Table 5-66 presents the data used to characterize the model facilities for specific countries
identified for this analysis.
Table 5-66: CH4 Generation Factors by Country
Country
Region3
k Constant
(i/yr)
Lo
(ft3/short ton)
Data Source
Guinea
Africa
0.18
4,690
WB
China
Asia
0.11
1,532
LMOP
India
Asia
0.11
3,988
Zhu et al. (2007)
Japan
Asia
0.11
4,620
WB
Nepal
Asia
0.04
6,890
WB
Pakistan
Asia
0.11
3,193
WB
Philippines
Asia
0.18
1,922
MSW DST
Argentina
CCSA
0.11
4,122
WB
Belize
CCSA
0.12
2,499
MSW DST
Colombia
CCSA
0.11
2,948
LMOP
Nicaragua
CCSA
0.11
2,627
MSW DST
Panama
CCSA
0.11
3,236
MSW DST
Bosnia and Herzegovina
Eurasia
0.06
4,295
WB
Ukraine
Eurasia
0.06
4,886
LMOP
Jordan
Middle East
0.02
5,984
WB
United States
North America
0.04
3,055
LMOP
a CCSA = Central & South America
Sources: WB—World Bank Studies by Country; LMOP—EPA's LMOP country-specific LFG models; MSW DST—decision support
model; and Zhu et al. (2007).
The international assessment of other OECD countries assumed waste management practices and landfill
designs similar to those in the United States. For this reason, we leveraged the existing United States-based landfill
population, scaling the landfill size and emissions to meet projected baselines. For all non-OECD countries for
which we had no data, we developed three model facilities to represent the allocation of waste to each type of
waste management facility (i.e., engineered landfill, sanitary landfill, and open dump). Each facility type was
assumed to have similar characteristics in terms of capacity, average depth of waste in place, and annual waste
acceptance rates.
5.4.1.3 Mitigation Options Considered
This analysis considers two types of abatement measures: mitigation technologies and diversion alternatives.
Mitigation technologies represent add-on technologies that can be applied to one or more landfill types described
above (i.e., open dump, basic landfill, engineered landfill) intended to capture and destroy the CH4. Diverting
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-213
-------
METHODOLOGY DOCUMENTATION
organic waste from the landfill for alternative uses is the second approach to reduce the quantity of LFG generated
at existing landfills.
Mitigation Technologies
LFG Collection and Flaring
Most basic landfills and engineered landfills have (or are applicable for) LFG collection systems for both public
health and facility safety concerns. These systems prevent the migration of Cm to on-site structures and adjacent
property and prevent the release of non-CFU organic compounds to the atmosphere. Flares ignite and burn LFG.
Flare designs include open and enclosed flares. Enclosed flares are more expensive but provide greater control of
combustion conditions, allow for stack testing, reduce light and noise nuisances, and might have higher
combustion efficiencies (EPA, 2010).
• Applicability: This option applies to all basic landfills and engineered landfills.
• Technical Efficiency: This analysis assumed a collection efficiency of 75% for basic landfills and of 85% for
engineered landfills and a flaring efficiency of 98%.
• Technical Lifetime: 15 years
• Capital Cost: Capital cost includes the construction of wells, wellheads, and laying of gathering lines that
make up the collection system, as well as the flare system with monitoring and control systems. Costs
were derived from the EPA Landfill Methane Outreach Program (LMOP) Project Cost Estimation Model.
• Annual O&M Cost: Typical annual O&M costs for collection systems are $2,250 per well and $4,500 per
flare. Electricity costs to operate the blower for a 600-cfm active gas collection system average $44,500
per year107 (EPA, 2010), assuming an electricity price of 7 cents/kWh and consumption rate of 0.002 kWh
per ft3.
• Annual Benefits: No economic benefits (energy production) are associated with this option.
LFG Collection for Electricity Generation
Converting LFG to electricity offers a potentially cost-effective way to use the gas being generated by the
landfill. This option requires an LFG collection and flare system as described earlier in this section, as well as the
electricity generation system. Components of the electricity generation system include the equipment for
generating energy (e.g., internal combustion engine, gas turbine, or microturbine) and the interconnections for
transmitting electricity produced to the energy grid.
• Applicability: This option applies to all basic landfills and engineered landfills.
• Technical Efficiency: This analysis assumed a collection efficiency of 75% for basic landfills and 85% for
engineered landfills and combustion efficiency of 98%.
• Technical Lifetime: 15 years
• Capital Cost: Capital cost includes the costs of the collection and flare system discussed and the treatment
system, energy generation equipment, and interconnection equipment for selling electricity to the power
grid. Costs were derived from the EPA LMOP Project Cost Estimation Model, which is available at EPA's
LMOP web page (see Table 5-67).
• Annual O&M Cost: Annual O&M costs are between $130 and $380 per kilowatt of capacity.
• Annual Benefits: Annual revenues are derived from the sale of electricity.
107 For this analysis, we assumed an electricity price of 7.5 cents/kWh and an energy consumption rate of 0.002 kWh/ft3.
5-214
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Table 5-67: Electricity Generation Technology Costs
Technology
Capital Cost
(2010 $/kW)
Annual O&M Costs
(2010 $/kW)
Internal combustion engine (>0.8 MW)
$1,700
$180
Small IC engine (<1 MW)
$2,300
$210
Gas turbine (>3 MW)
$1,400
$130
Microturbine (<1 MW)
$5,500
$380
CHP with IC engine (<1 MW)
$2,300
$210
LFG Collection for Direct Use
Direct use provides an alternative use of LFG with minimal treatment. Under this option, LFG collected at the
landfill is pumped to a nearby (<5 miles) end user. The gas delivered can serve as a medium-BTU fuel for boiler or
drying operations, kiln operations, and cement and asphalt production.108 Although little condensate removal and
filtration is needed, combustion equipment might need slight modifications to run with LFG (EPA, 2010). However,
these modification costs are not considered part of the technology costs.
• Applicability: This option is available to all basic landfills and engineered landfills.
• Technical Efficiency: This analysis assumed a collection efficiency of 75% for basic landfills and 85% for
engineered landfills and an end-use combustion efficiency of 98%.
• Technical Lifetime: 15 years
• Capital Cost:109 Costs include the equipment and installation cost of a skid-mounted filter, compressor,
and dehydrator, and the cost to construct a gas pipeline to carry the gas to a nearby (<5 miles) end
user(s).
• Annual Cost: Annual O&M costs include the cost of electricity and maintenance of the filters,
compressors, and dehydrators. The electricity costs were calculated by multiplying electricity price times
the energy required to power the equipment and transmit gas to end users, assuming a system power
demand of 0.002 kWh/ft3. Non-energy-related O&M costs were scaled to LFG project volumes assuming a
cost of $0.0014/ft3.
• Annual Benefits: Annual revenue accrues to the project through the sale of LFG to an end user at an
assumed price that is 80% of the current natural gas price; the discounted price reflects the lower BTU
content of the gas.
Enhanced Oxidation Systems
Enhanced oxidation systems are considered mitigation technologies that exploit the propensity of some
naturally occurring bacteria to oxidize Cm.110 By providing optimum conditions for microbial habitation and
efficiently routing LFGs to where they are cultivated, a number of bio-based systems, passively or actively vented
biofilters, and biowindows have been developed that can alone, or with gas collection, mitigate landfill Cm
emissions.
108 Other direct use applications include use in infrared heaters, greenhouses, artisan studios, leachate evaporation, and biofuel
production.
109 It is important to note that direct use of LFG may require equipment modifications at the end-user site to handle the lower
BTU content of LFG or additional treatment systems to improve the energy content; these costs are not considered part of this
abatement measure's project costs. Including these costs would increase project costs by more than $200,000 (EPA, 2010).
110 Oxidation of methane entails mixing the gas (CH4) with oxygen and converting the CH4 to C02 and water.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-215
-------
METHODOLOGY DOCUMENTATION
• Applicability: This option applies to basic landfills and engineered landfills.
• Technical Efficiency: This option analysis assumed a reduction efficiency of 44% of the remaining 15% of
Cm not collected by LFG collection system (Weitz, 2011).
• Technical Lifetime: 50 years
• Capital Cost: Capital costs are the incremental costs of enhanced oxidation systems above the traditional
clay/soil cover. These costs assume an incremental cost of $6 million for 100 acres of cover. The cost of
designing and constructing the biocover assumes $3/yd3 for earth moving, a compost price of
$5/tonne,m and an average cover depth of 3 feet.
• Annual O&M Cost: The O&M cost is assumed to be less than 0.1% of installed capital costs.
• Annual Benefits: No revenues are associated with this option.
Diversion Alternatives
Diversion alternatives redirect biodegradable components of the waste stream from the landfill for reuse
through recycling or conversion to a value-add product (e.g., energy or compost). The following diversion
alternatives were considered for this analysis:
Composting
Composting consists of the aerobic digestion of the fermentable organic fraction of MSW to produce a
reusable product. In the presence of oxygen, microorganisms decompose the biodegradable organic matter to
form compost, which contains nutrients used in agriculture as soil conditioner.
• Applicability: This option applies to yard and food components of the waste stream.
• Technical Efficiency: This analysis assumes reduction efficiency of 95%, which represents the avoided Cm
potential.
• Technical Lifetime: 15 years
• Capital Cost: Capital cost includes the purchase of land and equipment, site preparation, and facility
construction equal to $1.8 million (2010 USD). Capital costs were obtained from the composting process
model documentation of the MSW DST (MSW DST Documentation), which presents this cost for 100
tons/day facilities producing marketable high-quality compost products.
• Annual Cost: The O&M cost of the windrow composting facility includes the labor, overhead, fuel,
electricity, and equipment maintenance costs.112 This analysis assumes an O&M cost of $19/tonne-yr
(MSW DST Documentation).
• Annual Benefits: Revenue from compost is from sales and cost savings from avoided landfilling. The
composting process is not perfectly efficient, and this analysis assumed that 80% of the incoming organic
waste is converted to marketable compost product.
Anaerobic Digestion (AD)
AD is a complex biological process that uses anaerobic microorganisms to hydrolyze complex organics to
simple monomers and hence to volatile fatty acids; the volatile fatty acids are converted to CH4 and CO2 in the
biogasification step. The biogas can be recovered and used to generate energy. This analysis considers AD that
produces electricity using a gas engine, which is the most common product. A small amount of CH4 may be
111 The compost price assumes a weight by volume of 0.32 tonnes/yd3 (DST Model Documentation).
112 This analysis assumes that no precomposting screening will take place. Therefore, there will not be organics rejects from the
process needing disposal at a landfill facility, which is consistent with the data provided for high-quality compost production in
the composting process model documentation of the MSW DST (MSW DST Documentation).
5-216
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
released as fugitive emissions during the digestion process. This analysis assumes Cm emissions of 1 to 2 g/kg of
waste (dry weight) as reported in IPCC (2006).
• Applicability: This option assumes removal of wood, paper, and food waste.
• Technical Efficiency: This analysis assumed a capture efficiency of 75% and a reduction efficiency of 95%.
• Technical Lifetime: 20 years
• Capital Cost: The plant's capital cost includes the cost of land, the digestors, the gas engine, and air
pollution control and monitoring devices. The capital cost for this analysis is $472/design tonne was
obtained from Eunomia (2008), which describes this cost for facilities of 20,000 to 30,000 tonnes/yr in the
United Kingdom (U.K.).
• Annual Cost: The O&M cost includes the labor, overhead, fuel, electricity, and maintenance cost. An O&M
cost of $55/tonne yr1 (reported as £35 GBP/tonne) was considered in this analysis and obtained from
Eunomia (2008), which presents costs typical of U.K. facilities.
• Annual Benefits: Revenue from the sale of electricity generated with the biogas is sold to an end user.
The biogas recovery from the digestion process is not perfectly efficient and was assumed to be 75% of
total value, and the biogas composition was assumed to be 60/40% CH4/CO2 according to Eunomia (2008).
Similarly, the efficiency of the biogas conversion to electricity in the gas engine is assumed to be 37% as
reported by Eunomia (2008).
Mechanical Biological Treatment (MBT)
MBT can be defined as the processing or conversion of solid waste with biologically degradable components
via a combination of mechanical and other physical processes (e.g., cutting or crushing, sorting) with biological
processes (aerobic composting, anaerobic digestion). The primary objective is to reduce the mass and the volume
of the waste. A secondary objective is a lower environmental impact of the waste after its deposition (i.e., low
emissions of LFG, small amounts of leachate, and a reduced settlement of the landfill body).
• Applicability: This option applies to all landfill types,
• Technical Efficiency: This analysis assumed a reduction efficiency of 95%.
• Technical Lifetime: 20 years
• Capital Cost: The plant's capital cost includes the cost of land, facility, equipment, and air pollution
control and monitoring devices. The analysis assumed a capital cost of $15 million based on a reported
facility cost of $244/design tonne (reported as £150 British pounds/tonne) obtained from Eunomia
(2008). Costs are reported for a 60,000 tonne/yr facility in the U.K.
• Annual O&M Costs: The O&M cost of the MBT facility was $2 million in 2010. This cost includes the labor,
overhead, taxes, administration, insurance, indirect costs, energy, and maintenance costs. It does not
include residues disposal. A 2007 annual O&M cost of $22/tonne (reported as £13 British pounds/tonne)
was considered in this analysis and obtained from Eunomia (2008).
• Annual Benefits: Annual revenues from the sale of refuse-derived fuel (RDF) and recyclables that are
produced from the MBT process are sold to an end user (i.e., cement kilns or coal-fired utility). According
to Eunomia (2008), RDF is produced at a typical rate of 0.48 tonne/tonne of waste. Eunomia (2008) also
reports that 1 tonne of RDF can be assumed to replace 0.90 tonne of coal used to fuel a cement kiln and
0.38 tonne of coal for power generation. The market coal price of $40/tonne was used to estimate the
revenues. Similarly, Eunomia (2008) reports an 80% recovery rate for ferrous metals, 70% recovery rate
for nonferrous metals, and 70% recovery rate for glass. Sale prices of $352/tonne for ferrous metals
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-217
-------
METHODOLOGY DOCUMENTATION
(USGS, 2012), $l,881/tonne113 for nonferrous metals, and $25/tonne for glass were used to estimate the
revenues from recyclables sale.
Paper Recycling
Recycling typically consists of two major processes: the separation process at a material recovery facility
(MRF) and the remanufacturing process where recyclables are used to produce new products. For consistency with
other mitigation option included in this report, the costing component of this analysis only considered the
separation process.
• Applicability: This option applies to the entire waste stream.
• Technical Efficiency: This analysis assumed a reduction efficiency of 95% of potential CFU.
• Technical Lifetime: 20 years
• Capital Costs. The capital cost for this option is $35 million in (2010 USD). It consists of construction,
engineering, and equipment costs. It assumes a handling capacity of 100,000 tonnes of waste per year.
This analysis relied on a $297/tonne of annual capacity (2006 prices), which is an average of reported
capital costs from CalRecycle (2009) for similar sized facilities.
• O&M Cost. The O&M cost of the MRF facility includes wages, overhead, equipment and building
maintenance, and utilities. An O&M cost of $66/tonne of annual waste capacity before residue disposal
was based on reported operating costs used in the CalRecycle (2009) report. The cost of disposal of the
MRF rejects can be estimated assuming an MRF separation efficiency of 55% of the incoming organic
waste and that the rejects are sent to a regular landfill with a tipping fee of $29/tonne, which represents a
U.S. national average tipping fee obtained from Municipal Solid Waste Facility Directory (Chartwell, 2004).
• Annual Benefits: Annual benefits come from the sale of recyclables and decreased waste. The recyclables
that are separated at the MRF are sold to an end user (e.g., a remanufacturing facility) sometimes through
brokers. The 55% separation efficiency and recyclables sale prices were used to estimate the revenues
from recyclables sale. The following prices were used in the analysis: mixed paper114—$140/tonne; scrap
metals115—$l,307/tonne; and scrap glass—$25/tonne. Tonnage sold for reuse avoids landfilling costs.
Annual cost savings are equal to tonnage sold for reuse times the tipping fee of $29/tonne.
Waste to Energy (WTE)
WTE is a combustion process; thus, its main emissions include CO2, CO, NOx, and non-CFU VOCs. Municipal
waste is incinerated to reduce its volume to save landfill costs and recover energy from its combustion either for
heating and/or electricity generation. This analysis considers WTE using mass-burn incineration and electricity
recovery, which is the most common WTE design. Representative CH4 emissions of 0.2 to 60 kg/Gg of waste (wet
weight) and N2O emissions of 41 to 56 g/ton of waste (wet weight) were obtained from IPCC (2006).
• Applicability: This option applies to the entire waste stream.
• Technical Efficiency: This analysis assumed a reduction efficiency of 100%.
• Technical Lifetime: 20 years
• Capital Costs. The plant's capital cost of $165 million includes the facility design engineering and
construction. Capital equipment includes the cost of land, incinerators, ash handling system, turbine, and
113 Price obtained from MetalPrices.com at http://www.metalDrices.com/FreeSite/metals/al scrap/al scrap.asottTables cf.
114 Prices were obtained from: http://www.recvcle.cc/freepapr.htm cP.
115 Assumes a weighted average price of aluminum can scrap and ferrous metal scrap prices. The aluminum can scrap price was
obtained from http://www.metalprices.com/ J. The ferrous metal price was obtained from 2012 USGS Mineral Commodities
Summary: Iron & Steel Scrap at: http://minerals.usgs.gov/minerals/pubs/commoditv/iron & steel scrap/ CP.
5-218
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
air pollution control and monitoring devices. Costs assume $829/tonne of design capacity. This cost was
derived from Eunomia (2008), which describes this cost for a 200,000 tonne/yr facility in the U.K.
• O&M Cost. The annual O&M cost of the WTE facility is $8 million, approximately 4% of installed capital
costs. Annual costs include labor, overhead, taxes, administration, insurance, indirect costs, auxiliary fuel
cost, electricity cost, and maintenance cost. They do not include the cost for disposing of the combustion
residue and spray dryer residue. The cost is based on an annual O&M cost of $41/tonne/year. Annual
avoided landfilling is also included as a cost savings. The cost of disposal of the fly and bottom ash from
the incineration process assumes an estimated 15% of the incoming organic waste will be converted to
ash (MSW DST Documentation). No reuse of the bottom ash (e.g., in construction projects) was assumed,
and the bottom and fly ash will be mixed and sent to a landfill. Both the avoided landfilling costs and
residual waste landfilling costs assume a tipping fee of $29/tonne.
• Annual Benefits: Revenues from electricity sales represent the annual benefits for this option. Electricity
is generated by recovering heat from combusting waste. The recovery of the heat is inefficient, which is
represented by the heat rate of the plant, reported as 18,000 (BTU/kWh) in the WTE process model
documentation of the MSW DST (MSW DST Documentation). The electricity produced per tonne of waste
can then be estimated according to the heat value of the waste incinerated (4,750 BTU/tonne of waste).
The market price of electricity was used to estimate the revenues.
5.4.1.4 Technical and Economic Characteristics Summary
Table 5-68 summarizes the engineering and costs assumptions for mitigation options considered for MSW
landfills.
Table 5-68: Summary of the Engineering and Cost Assumptions for Mitigation Options at Landfills
Abatement Option
Total Installed
Capital Cost
(millions 2010
USD)
Annual
O&M Cost
(millions 2010
USD)
Time
Horizon
(years)
Reduction
Efficiency
(%)a
LFG Mitigation Options
LFG collect and flaring system
1.7
0.3
15
85%
LFG for electricity generation
85%
Internal combustion engine
6.3
0.8
15
85%
Gas turbine (>3 MW)
5.6
0.6
15
85%
Micro-turbine (<1 MW)
4.1
0.1
15
85%
Combined heat and power production
7.9
0.8
15
85%
Direct gas use
2.6
0.5
15
85%
Enhanced oxidation systems
5.4
0.0
50
44%
Waste Diversion Options
Composting
1.8
0.7
15
95%
Anaerobic digestion
16.9
1.7
20
95%
Mechanical biological treatment
15.4
1.8
20
95%
Paper recycling
34.9
8.9
20
95%
Waste to energy
165.7
8.0
20
100%
a Reduction efficiency reflects the abatement measure's ability to mitigate/avoid CH4 generation. However, this does not reflect
the total mitigation potential.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-219
-------
METHODOLOGY DOCUMENTATION
5.4.1.5 Sector-Level Trends and Considerations
It is important to note that depending on the scenario considered in the model, diversion options may or may
not be included. If diversion options are considered, BAU emissions (indexed by facility type) are uniformly
distributed by the total number of technologies (N = 12). If diversion options are omitted, BAU emissions are
distributed by the number of landfill-based mitigation technologies (N = 7) (see Figure 5-5).
Underlying the general modeling approach, the MAC analysis also incorporated additional international
considerations to capture shifts in the share of BAU emissions allocated to the three model landfill types defined
earlier in this section (i.e., open dump, basic landfill, and engineered landfill). Table 5-69 presents the facility share
of BAU emissions over time. In the United States and the EU, we assumed advanced waste management practices
were already in place. Reflecting this assumption, we assumed zero emissions coming from open dumps in these
countries and assumed all emissions come from basic and engineered landfills. Given the existing level of
infrastructure in place, there is very little change in the assumed distribution over the 20-year modeling horizon.
For emerging economies and developing countries, the analysis assumed a greater share of emissions is
represented by open dumps in 2010. Over the next 20 years, this distribution is projected to shift away from open
dumps as countries begin to adopt advanced waste management practices with greater shares of total waste going
to basic sanitary and engineered landfills. These shares were developed using expert judgment after reviewing
existing literature on waste disposal trends and abatement opportunities provided through various studies by the
World Bank, EPA's LMOP program, and the GMI.
5-220
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
Distribute 50% of BAU
uniformly to
Diversion Technologies
(N = 5)
Composting
Anaerobic Digester
MBT
Paper Recycling
WTE
Distribute 50% of BAU
uniformly to
Mitigation Technologies
(N = 7)
Enhanced Oxidation
Yes
Collection & Flare
LFGTE-IC Engine
LFGTE-Turbine
BAU
Emissions,
diversion
LFGTE - Mircoturbine
(yes/no)
LFGTE-CHP Engine
Direct Use
No
¦100%-
' Allocation of BAU
emissions (mitigation
v vs. diversion) .
Technical
Effectiveness
Country Weight
Country Weight
Technical
Effectiveness (%)
Open Dumps
(% of BAU)
Engineered LFs
(% of BAU)
Basic LFs
(% of BAU)
% of BAU to
Diversion
Technology
BAU share by
landfill type
% of BAU to
Landfill Mitigation
Technologies
Abatement Potential
(MtC02e)
Abatement Potential
(MtC02e)
LFGTE - Mircoturbine
Distribute BAU to
Mitigation Technologies
Enhanced Oxidation
LFGTE-CHP Engine
Collection & Flare
LFGTE - IC Engine
LFGTE-Turbine
Direct Use
Figure 5-5: Conceptual Model for Estimating Mitigation Potential in the MSW Landfill Sector
(J)
m
O
—I
O
7J
r~
m
<
I
O
D
Cf)
-------
cn
ro
ls>
KJ
o
9
Ci
Distribute 50% of BAU
uniformly to
Diversion Technologies
(N = 5)
% of BAU to
Diversion
Technology
Composting
Anaerobic Digester
MBT
—~ Paper Recycling
—~m WTE
Allocation of BAU
emissions (mitigation
vs. diversion)
Distribute 50% of BAU
uniformly to
Mitigation Technologies
(N = 7)
Technical
Effectiveness (%)
Country Weight
Abatement Potential
(MtC02e)
Open Dumps
(% of BAU)
Yes
—~ Collection & Flare
LFGTE - II
% of BAU to
Landfill Mitigation
Technologies
LFGTE-Turbine
BAU
Emissions,
diversion
measures
(yes/no)
Basic LFs
(% of BAU}
LFGTE - Mircoturbine
LFGTE-CHP Engine
Direct Use
Engineered LFs
(% of BAU)
No
Distribute BAU to
Mitigation Technologies
Enhanced Oxidation
Collection & Flare
LFGTE -IC Engine
Technical
Effectiveness
(%)
BAU share by
landfill type
Country Weight
Abatement Potential
(MtC02e)
¦100°/
LFGTE-Turbine
LFGTE - Mircoturbine
LFGTE-CHP Engine
Direct Use
I
O
O
o
r~
O
o
-<
o
o
o
c
>
H
o
-------
Allocation of BAU
emissions (mitigation
vs. diversion)
Yes
BAU
Emissions,
diversion
(yes/no)
No
100%-
Basic LFs
(% of BAU)
Open Dumps
(% of BAU)
Engineered LFs
(% of BAU)
% of BAU to
Diversion
Technology
% of BAU to
Landfill Mitigation
Technologies
BAU share by
landfill type
Distribute 50% of BAU
uniformly to
Diversion Technologies
(N = 5)
Composting
Anaerobic Digester
Paper Recycling
WTE
Distribute 50% of BAU
uniformly to
Mitigation Technologies
(N = 7)
Collection & Flare
LFGTE - IC Engine
LFGTE- Turbine
LFGTE - Mircoturbine
LFGTE-CHP Engine
Abatement Potential
(MtC02e)
Country Weight
Technical
Effectiveness (%)
Distribute BAU to
Mitigation Technologies
Enhanced Oxidation |
Collection & Flare
LFGTE-IC Engine
LFGTE-Turbine
LFGTE - Mircoturbine
LFGTE - CHP^Engine
Direct Use
Technical
Effectiveness
Country Weight
Abatement Potential
(MtC02e)
(D
m
O
O
7)
r~
m
<
I
O
O
(D
-------
METHODOLOGY DOCUMENTATION
Table 5-69: Model Facilities Share of BAU Emissions: 2010-2030
2010
2020
2030
Country/
Region
Dump
Sites
Basic
LF
Engineered
LF
Dump
Sites
Basic
LF
Engineered
LF
Dump
Sites
Basic
LF
Engineered
LF
China
20%
60%
20%
10%
60%
30%
10%
50%
40%
Brazil
10%
60%
30%
10%
50%
40%
0%
50%
50%
Mexico
10%
60%
30%
10%
50%
40%
0%
50%
50%
Russia
20%
40%
40%
20%
40%
40%
10%
40%
50%
Ukraine
20%
40%
40%
20%
40%
40%
10%
40%
50%
Australia
10%
30%
60%
10%
30%
60%
0%
30%
70%
Canada
10%
30%
60%
10%
30%
60%
0%
30%
70%
Japan
10%
30%
60%
0%
30%
70%
0%
20%
80%
Turkey
20%
40%
40%
20%
40%
40%
10%
40%
50%
United States
0%
20%
80%
0%
20%
80%
0%
10%
90%
India
20%
60%
20%
10%
60%
30%
10%
50%
40%
South Korea
10%
30%
60%
0%
30%
70%
0%
20%
80%
EU-27
0%
20%
80%
0%
20%
80%
0%
10%
90%
Africa
40%
40%
20%
30%
40%
30%
20%
40%
40%
Central & South
America
10%
60%
30%
10%
50%
40%
0%
70%
30%
Middle East
20%
60%
20%
10%
60%
30%
10%
60%
30%
Eurasia
20%
60%
20%
10%
60%
30%
10%
60%
30%
Asia
20%
60%
20%
10%
60%
30%
10%
60%
30%
Source: Based on expert judgment in consultation with World Bank (2010) and EPA (2009, 2011).
5.4.1.6 References
CalRecycle. 2009. Life Cycle Assessment and Economic Analysis of Organic Waste Management and Greenhouse
Gas Reduction Options. Available online at
http://www.calrecvcle.ca.gov/climate/Events/LifeCvcle/2009/default.htm
Chartwell Information Publisher Services. 2004. Chartwell Municipal Solid Waste Facility Directory. Solid Waste
Market Intelligence, Analysis, and Strategy Services.
Eunomia. 2008. Development of Marginal Abatement Cost Curves for the Waste Sector. UK: Committee on Climate
Change. Defra and Environment Agency. Available online at
http://www.theccc.org.uk/pdfs/Eunomia%20Waste%20MACCs%20Report%20Final.pdfEf
Eurostat. 2016. Eurostat. Municipal waste generation and treatment. Available online at
https://ec.europa.eu/eurostat/web/waste/municipal-waste-generation-and-treatment-bv-treatment-
method of
Intergovernmental Panel on Climate Change. 2006. 2006 IPCC Guidelines for National Greenhouse Gas Inventories:
Reference Manual (Volume 5-Waste). Available online at http://www.ipcc-
nggip.iges.or.jp/public/2006gl/index. html cf
Kawai, K. and T. Tasaki. 2016. Revisiting estimates of municipal solid waste generation per capita and their
reliability. Journal of Material Cycles Waste Management, 18,1-13.
5-224
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
National Bureau of Statistics of China. 2015. China Statistical Yearbook. Available online at
http://www.stats.gov.cn/tisi/ndsi/2015/indexeh.htm
Organisation for Economic Cooperation and Development. 2011. Environment Statistics Database, Waste
generation by sector. doi:10.1787/data-00674-en
Organisation for Economic Cooperation and Development. 2014. Long-term baseline projections, No. 95 (Edition
2014). Organisation for Economic Cooperation and Development. Available online at
https://data.oecd.org/gdp/gdp-long-term-forecast.htm bP
OECD.stat. 2016. Questionnaire on the state of the environment: municipal waste generation and treatment.
Available online at http://stats.oecd.org/lndex.aspx?DataSetCode=MUNW J
RTI International. 2012. Municipal Solid Waste Decision Support Tool. MSW DST Documentation. Research Triangle
Park, NC: RTI International. Available online at https://mswdst.rti.org/ cP
United Nations (UN). 2015. World Population Prospects data. Available at https://population.un.org/wpp/cf
U.S. Department of Agriculture. 2016. Real GDP (2010 Dollars) Projections. Available online at
https://www.ers.usda.gov/data-products/international-macroeconomic-data-set.aspx
U.S. Environmental Protection Agency. 2009. Municipal Solid Waste in the United States: 2009 Facts and Figures.
Washington, DC: EPA. Available online at http://www.epa.gov/osw/nonhaz/municipal/msw99.htm
U.S. Environmental Protection Agency. 2010. Landfill Gas Energy Cost Model. Washington, DC: Landfill Methane
Outreach Program. Available online at http://www.epa.gov/lmop/publications-tools/index.html
U.S. Environmental Protection Agency. 2011. Draft Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-
2011. Washington, DC: EPA. Available online at
http://www.epa.gov/climatechange/emissions/usinventorvreport.html
U.S. Geologic Survey. 2012. Mineral Commodities Summary: Iron & Steel Scrap. Available online at
http://minerals.usgs.gov/minerals/pubs/commoditv/iron & steel scrap/
United Nations Statistics Division (UNSD). 2016. United Nations Statistics Division (UNSD), Environmental statistics
database: municipal waste collected. Available online at
h ttp://data. un. org/Data. aspx?d=ENV&f=variablelD%3a 1814 eP
World Bank Group. 2010. 2009—World Development Indicators: Table 2.1 Population Dynamics. Available online at
http://data.worldbank.org/data-catalog/world-development-indicatorscP
World Bank. 2012. What a waste—A global review of solid waste management. Urban Development Series
Knowledge Papers, No. 15.
World Bank. 2017. GDP (Constant 2010 US$). Available online at
http://data.worldbank.org/indicator/NY.GDP.MKTP.KD c?
Weitz, K. January 2011. Updated Research on Methane Oxidation in Landfills. Technical memorandum providing an
update to previous (March 2006) research on methane oxidation rates in landfills. Research Triangle Park, NC:
RTI International.
Zhu, D., P.U. Asnani, C. Zurbriigg, S. Anapolsky, and S. Mani. 2007. Improving Municipal Solid Waste Management
in India. Washington DC: The World Bank.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-225
-------
METHODOLOGY DOCUMENTATION
5.4.2 Wastewater Management
Wastewater originates from a variety of residential, commercial, and industrial sources and may be treated on
site (uncollected), sewered to a centralized plant (collected), or disposed of untreated nearby or via an outfall.
Domestic and industrial wastewater treatment activities can result in deliberate venting and fugitive emissions of
Cm. In addition, domestic wastewater is a source of N2O emissions. CH4 is produced when the organic material
present in the wastewater flows decomposes under anaerobic conditions. Although most developed countries rely
on centralized aerobic wastewater treatment systems, which limit the level of CH4 generated, less developed
countries often rely on a broader suite of wastewater treatment technologies with a significant proportion of
wastewater flows handled by anaerobic systems such as septic tanks, latrines, open sewers, and lagoons. Industrial
wastewater can also be treated anaerobically, with significant CH4 being emitted from those industries with high
organic loadings in their wastewater stream, such as food processing and pulp and paper facilities (EPA, 2012).
5.4.2.1 Wastewater Projections Methodology
UNFCCC-reported, country-specific estimates were used for historical emission estimates in this source
category, when available. For those countries with country-reported emission estimates, emission projections
were estimated from the most recent country-reported data through 2050 using growth rates calculated by the
Tier 1 methodology. For countries that do not have country-reported historical data, Tier 1-calculated emission
estimates were used for the full time series from 1990 through 2050 (see Section 3.3, Generating the Composite
Emission Projections, for additional information). Activity data for wastewater management included population
data from the UN (2015), urban population data from the UN (2014), percentage of the population with
wastewater collection and treatment from the UN Statistics Division (2016) and per capita protein generation from
FAO (2016). While country-reported estimates include both domestic and industrial wastewater, the emission
estimates calculated using Tier 1 methodology only include domestic wastewater.116
Methane from Wastewater Treatment/Discharge
The general equation to estimate CH4 emissions from domestic wastewater is as follows:
CH4 Emissions = ¦ TLJ ¦ EFj)] ¦ (TOW - S) - R (5.32)
where:
CH4 Emissions = CH4 emitted in year T, kg CFU/yr
TOW = Total organics in wastewater in year, kg biochemical oxygen demand (BOD)/yr
S = Organic component removed as sludge, kg BOD/yr (S=0)
Ui = Fraction of population in income group, /
Tij = Degree of utilization of each treatment pathway, j, for each income group, /
/ = Income group: rural, urban high-income, urban low-income
j = Wastewater treatment/discharge pathways
EFj = Methane emission factor, kg CH4/ kg BOD
R = Quantity of methane recovered, kg CFU/yr
Key drivers of wastewater emissions include the quantity of degradable organic material in the wastewater and
the type of treatment system used. Treatment systems or discharge pathways that provide anaerobic
environments will generally produce CH4, whereas systems that provide aerobic environments will normally
produce little or no CH4. BOD is used to measure the organic component of domestic wastewater. The total
116 While industrial wastewater emissions were not explicitly estimated in this report, some countries report industrial
wastewater emissions within this source category. In these cases, this source category includes these emissions.
5-226
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
quantity of domestic BOD in the base year and subsequent years is driven by changes in population and per capita
BOD generation.
Nitrous Oxide Emissions from Wastewater
The general equation to estimate N2O emissions from domestic wastewater is as follows:
N2O = Neffluent * EFcffluent * (44/28) (5.33)
where:
N2O Emissions = N2O emission, kg INhO/yr
N = Nitrogen (N) in the effluent discharged into aquatic environments, kg N/yr
EF = Emission factor for N2O emission, kg N2O-N / kg N
The factor 44/28 is the conversion of kg N2O-N into kg N2O.
Activity Data
Methane from Wastewater Treatment/Discharge
Historical
• Annual total population data by country in 5-year increments from 1950 through 2015 were obtained
from the United Nations Department of Economic and Social Affairs, Population Division, World
Population Prospects (UN, 2015). Population data for years between the reported values were linearly
interpolated.
• The annual percentage of urban population was obtained from The 2014 Revision (UN, 2014). Urban
population is split into high- and low-income urban populations based on IPCC default proportions,
provided for a sample of countries and regions (see Table 6.5 of the 2006 IPCC Guidelines) (IPCC, 2006).
• The percentage of the population with wastewater collection and treatment was obtained from the
United Nations Statistics Division, environmental statistics database: percentage of population with
wastewater collection and treatment (UNSD, 2016). These data provide country-reported information for
58 countries over the period 1990 through 2009, although most countries do not have data reported for
every year; missing values were linearly interpolated. These data provide a basis for determining the
percentage of wastewater that is (1) collected and treated, (2) collected but untreated, and (3)
uncollected. For countries that do not have reported data, the EPA used the IPCC default assumptions for
the percentage utilization of these pathways, based on Table 6.5 of the 2006 IPCC Guidelines (IPCC, 2006).
Values do not vary by year for countries using IPCC default values.
• The total quantity of organics in wastewater is calculated as the product of population and BOD per
person. IPCC default BOD values were used. Countries without IPCC default values were assigned values
based on a nearby country or the associated region (see Table 6.4 of the 2006 IPCC Guidelines) (IPCC,
2006).
• The total quantity of BOD in wastewater, by collection/treatment path and urban/rural designation, was
estimated based on total BOD, the fraction of population in each income class, and proportion of
wastewater in each discharge pathway, per above.
Projected
• Annual total population data by country in 5-year increments from 2016 through 2050 were obtained
from the United Nations Department of Economic and Social Affairs, Population Division, World
Population Prospects (UN, 2015). Population data for years between the reported values were linearly
interpolated.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-227
-------
METHODOLOGY DOCUMENTATION
• For all projected activity data, the EPA assumed the last year of country-reported data, if available, or
otherwise relied on the IPCC defaults, which do not change over time.
Nitrous Oxide Emissions from Wastewater
The activity data needed to estimate N2O emissions include the nitrogen content in the wastewater effluent,
country population, and average annual per capita protein generation.
Historical
• Annual total population data by country in 5-year increments from 1950 through 2015 were obtained
from the United Nations Department of Economic and Social Affairs, Population Division, World
Population Prospects (UN, 2015). Population data for years between the reported values were linearly
interpolated.
• Per capita protein generation, which consists of intake (consumption), was obtained for 1990 through
2011 for 162 countries from FAO (2016). Missing values over this period were linearly interpolated.
Countries for which no data are available were mapped to proximate country data.
Projected
• Annual total population data by country in 5-year increments from 2016 through 2050 were obtained
from the United Nations Department of Economic and Social Affairs, Population Division, World
Population Prospects (UN, 2015). Population data for years between the reported values were linearly
interpolated.
• Projected per capita protein generation was assumed to be equal to the last year of country-reported
data for all future years, 2011 through 2050.
Emission Factors
Methane from Wastewater Treatment/Discharge
Historical and Projected
The analysis used IPCC Tier 1 emission factors, including the following assumptions:
• Maximum CFU production capacity (kg CFU/kg BOD) is the IPCC default emission factor (see Section 6.2 of
the 2006 IPCC Guidelines), which was further adjusted based on the specific discharge pathway in use
based on IPCC default MCF (see Table 6.3 of the 2006 IPCC Guidelines) (IPCC, 2006).
• IPCC default MCFs are provided for specific wastewater treatment approaches, which are more detailed
than the categories of collected-treated, collected-untreated, and uncollected (e.g., septic, sewer, latrine).
The EPA developed weighted average MCFs to represent the three categories in the analysis. Using these
three categories enabled the analysis to leverage country-reported wastewater management data.
Nitrous Oxide Emissions from Wastewater
Historical and Projected
The analysis used IPCC (2006) Tier 1 emission factors, including the following assumptions:
• fraction N in protein (kg N/kg protein) based on IPCC 6.3.3
• fraction of nonconsumption protein for developed countries based on IPCC 6.3.1.3 and for developing
countries based on IPCC 6.3.1.3
• fraction of industrial and commercial co-discharged protein based on IPCC 6.3.1.3
• N removed with sludge based on IPCC Ch. 6
• emission factor (kg N20/kg N) based on IPCC 6.3.1.2
• conversion factor to convert kg N2O-N to kg N2O based on IPCC Ch. 6
5-228
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Emission Reductions in Baseline Scenario
Cm emissions from wastewater can be reduced through improved wastewater treatment practices, including
reducing the amount of organic waste anaerobically digested and flaring or using CH4 from anaerobic digesters for
cogeneration or other beneficial reuse. Such emission reduction activities are not widespread, so they are not
explicitly included in these estimates. The estimates do not account for possible future modernization of domestic
wastewater handling that may see a shift to aerobic treatments and the implementation of CH4 capture from
anaerobic digesters that would result in a reduction of emissions.
Uncertainty
For domestic wastewater, the primary uncertainties lie in specifying the annual utilization of different
wastewater management practices and the associated IPCC default emission factors for those practices. In
addition, large uncertainties are associated with the IPCC default emission factors for N2O from effluent.
5.4.2.2 Wastewater Mitigation Options Considered
This analysis focused on domestic wastewater treatment and implementation of abatement measures aimed
at reducing CH4 emissions, which can be mitigated through investment in infrastructure and/or equipment.
Conversely, there are no proven and reliable technologies for mitigating N2O emissions. Mitigation steps to limit
N2O emissions from wastewater treatment require technical expertise and experience rather than an engineered
solution; thus, they fall outside the scope of an engineered cost analysis. This analysis considers abatement
measures that may be applied to one of five existing wastewater treatment systems currently being used in a given
country. Scenarios 1 and 2 correspond to the upper half of the sanitation ladder, while Scenarios 3 through 5
correspond to the lower half of the sanitation ladder. Figure 5-6 presents the five baseline scenarios for the
existing status quo.
Figure 5-6: Five Existing Scenarios Evaluated for Given Wastewater Discharge Pathways Based on
Technology Level, Treatment Alternative, and Collection Method
Domestic
wastewater
Centralized
collection
No centralized
collection
Treated
Untreated
Treated
(1) Centralized
WWTP
(2) (Open)
Sewer
(5) Other
system
(4) Latrine
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-229
-------
METHODOLOGY DOCUMENTATION
Table 5-70 compares the three abatement alternatives for an example population of 400,000 people,
population density of 3,000/km2, and wastewater generation rate of 340 L/person/day.
Table 5-70: Mitigation Options for the Wastewater Sector
Abatement Option
Total Installed
Capital Cost
(2010 10s USD)
Annual O&M
Cost
(2010 10s USD)
Time
Horizon
(years)
Technical
Efficiency
Anaerobic biomass digester with
Cm collection and cogen.
$21.1
$5.0
20
60-80%
Aerobic wastewater treatment plant
(WWTP)
$97.2
$4.7
20
60-80%
Centralized wastewater collection
(+ aerobic WWTP)
$55.9 ($153.1)
$1.6 ($6.3)
50
60-80%
CH4 Mitigation Technology for Existing Decentralized Treatment
This section characterizes the reduction in CH4emissions by adding a collection system and centralized
treatment facility in developing countries where the current practice is decentralized wastewater treatment.
Wastewater Collection System—New Construction
For areas of the developing world without centralized wastewater treatment, latrines and/or septic tanks are
typically used to dispose of domestic wastewater. In both of these cases, the organic matter in the wastewater will
undergo anaerobic degradation to produce Cm. The construction and implementation of a collection system and
subsequent treatment at a centralized facility would significantly reduce Cm formation because transporting
wastewater through sewers promotes aerobic conditions and reduces the fraction of organic content that
undergoes anaerobic digestion.
The design and size of a wastewater collection system depend on the population served, the service area size,
and water use characteristics of the population. Wastewater collection systems link all household and commercial
discharges through underground piping, conveying the water to either a centralized treatment facility or directly to
an outfall point where it is released into the environment. Pipelines can vary from 6 inches in diameter to
concrete-lined tunnels up to 30 feet in diameter. Collection systems are built with a gradient so gravity can
facilitate the water flow; where large distances must be covered, periodic pump stations (also called lift stations)
are sometimes used to pump the sewage to a higher elevation and again allow gravity to transport the sewage.
Sewage pumps are typically centrifugal pumps with open impellers, designed to have a wide opening to prevent
the raw sewage from clogging the pump. This scenario evaluates the impact of installing a sewer collection system
without a centralized treatment facility.
• Applicability: This option applies to all scenarios having no existing centralized collection system.
• Technical Efficiency: This analysis assumed an initial collection efficiency of 60%, which increases by 10%
each year, because of an assumed improvement in technical efficiency.
• Technical Lifetime: 50 years
• Capital Cost: We used the cost estimation model Water and Wastewater Treatment Technologies
Appropriate for Reuse (WAWTTAR) (Finney and Gearheart, 2004) to determine the capital cost of the
sewer construction. The model is used by engineers, planners, decision-makers, and financiers to estimate
the costs of making improvements to wastewater treatment systems while minimizing impacts to water
resources. The capital cost curve for wastewater collection systems is based on the population density:
capital cost ($MM/km2) = 360.54 x Dp"0 844, where DP is population density in (persons/km2).
5-230
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Annual O&M Cost: Annual O&M costs for collection systems were scaled from the capital cost and
assumed to be a factor of 0.028 x initial capital cost, which for this case gives the following cost curve,
based on population density: O&M cost ($MM/km2) = 10.095 x DP"0844.
• Annual Benefits: No benefits are associated with this option.
Aerobic Wastewater Treatment Plant (WWTP)-New Construction
Contaminants in wastewater are removed using a variety of physical, chemical, and biological methods. A
WWTP typically comprises many unit operations from each of these broad categories. Wastewater treatment
technologies are also divided into stages of treatment, each of which comprises one or more individual treatment
processes. We provide a brief summary of each of these classifications:
• Pretreatment: This stage involves the removal of wastewater constituents. These constituents can include
rags, sticks, floatables, grit, and grease that may cause maintenance or operational problems with the
treatment operations, processes, and ancillary systems.
• Primary Treatment: This stage focuses on the removal of a portion of the total suspended solids (TSS) and
organic matter from the wastewater. Primary treatment is a physical unit process in which the sewage
flows into large tanks, known as primary clarifiers or primary settling tanks.
• Secondary Treatment: This stage focuses on the removal of biodegradable organic matter (in solution or
suspension) and TSS by aerobic or anaerobic biological treatment. Disinfection is also typically included in
the definition of conventional secondary treatment. Secondary treatment is a biological process that
cultivates and uses a consortium of microorganisms to degrade the organic wastes and reduce nutrient
levels in wastewater. Secondary treatment can either be aerobic (with oxygen) or anaerobic (without
oxygen). By far, the most common approach used in WWTPs is the activated sludge process. This process
is an aerobic suspended-growth system containing a biomass that is maintained with oxygen and is
capable of stabilizing organic matter found in wastewater.
• Tertiary Treatment: This stage involves the removal of residual suspended solids (after secondary
treatment), usually by granular medium filtration or microscreens. Disinfection is also typically a part of
tertiary treatment. Nutrient removal is often included in this stage.
• Applicability: This option applies to all conditions when new WWTPs are constructed.
• Technical Efficiency: This analysis assumed an initial collection efficiency of 60%, which increases by 10%
each year.
• Technical Lifetime: 20 years.
• Capital Cost: Capital costs were estimated using EPA's cost curves detailing the construction costs of
publicly owned wastewater treatment facilities (EPA, 1980). The cost curves in EPA (1980) are based on
actual winning bids for treatment plants, which include detailed equipment and materials requirements,
including labor, amortization, land, concrete, pumps, pipes, power, haulage, chemicals, and design fees.
All cost curves were updated to year 2010 dollars. The cost curve is based on the flow rate of the WWTP:
capital cost ($MM) = 0.0174 x Q0 73, where Q is the flow rate in m3/day.
• Annual O&M Cost: Typical annual O&M costs of an aerobic WWTP are due to electricity used to provide
aeration and operation equipment, labor to operate the plant, chemicals, and equipment replacement.
EPA's cost curves (updated to 2010 dollars) provide the following cost curve for an aerobic WWTP, based
on the flow rate: 0.0002 x Q°8517.
• Annual Benefits: None.
CH4 Mitigation Technology for Existing Collection System without Treatment
This section characterizes the reduction in Cm emissions for the existing condition of a centralized collection
system without a treatment facility. As noted above, contaminants in wastewater are removed via a variety of
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-231
-------
METHODOLOGY DOCUMENTATION
physical, chemical, and biological methods. An anaerobic WWTP typically comprises many unit operations divided
into stages of treatment: pretreatment, primary treatment, secondary treatment, and tertiary treatment.
• Applicability: This option applies to all conditions when new WWTPs are constructed.
• Technical Efficiency: This analysis assumed an initial collection efficiency of 60%, which increases by 10%
each year.
• Technical Lifetime: 20 years.
• Capital Cost: Capital costs were estimated using EPA's cost curves detailing the construction costs of
publicly owned wastewater treatment facilities. The cost curve is based on the flow rate of the WWTP:
capital cost ($MM) = 0.0174 x Q0 73, where Q is the flow rate in m3/day.
• Annual O&M Cost: Typical annual O&M costs of an aerobic WWTP are due to electricity used to provide
aeration and operation equipment, labor to operate the plant, chemicals, and equipment replacement.
CapdetWorks v2.5 was used to estimate O&M costs. The costs were based on a detailed equipment and
materials database that uses published cost indices, including labor, amortization, and energy
requirements. CapdetWorks provides the following cost curve for an aerobic WWTP, based on the flow
rate: O&M cost ($MM) = 0.0002 x Q0-8517.
• Annual Benefits: None.
CH4 Mitigation Technology for Existing Centralized Aerobic WWTPs
Anaerobic Biomass Digester with CH4 Collection
The top of the technology ladder evaluated assumes that an existing centralized WWTP is used to treat all
wastewater generated in the region. The significant quantity of biomass generated during the decomposition of
the sewage is a major operational component of WWTP operation. Typical approaches to sludge handling include
dewatering to reduce the overall volume and further water reduction in open-air drying beds. The sludge is rich in
organic matter and has the potential to produce high amounts of Cm during degradation. Anaerobic digestion is
an additional sludge-handling step that can be employed to further reduce the sludge volume; it is a process that
involves the decomposition of this organic material in an oxygen-free environment to produce and collect Cm.
Anaerobic digesters are large covered tanks that are heated to optimize the CH4-generating process. The tanks
typically employ a mixing mechanism to ensure uniform conditions throughout the tank and are designed with
headspace to collect the gas generated, which is typically a mix of 60 to 70% Cm and the 30 to 40% CO2, along with
trace gases. The remaining solid material is nutrient rich and is a suitable fertilizer for land application. The heat
from the flared gas can be used to heat the digester, lowering the overall energy requirements of the system.
Alternatively, the gas can be used to produce electricity with a turbine.
• Applicability: This option applies to all existing WWTP types.
• Technical Efficiency: This analysis assumed an initial collection efficiency of 60%, which increases by 10%
each year.
• Technical Lifetime: 20 years
• Capital Cost: Costs were derived from EPA's process cost curves for new construction of an anaerobic
digester. The capital cost covers the construction of the tank with heater and cover and includes concrete,
all equipment, process piping, and steel required for digester construction. Costs were derived from
CapdetWorks. The cost curve is based on the flow rate of the WWTP: capital cost ($MM) = 0.0004 x Q0 92,
where Q is the flow rate in m3/day.
• Annual O&M Cost: Typical annual O&M costs for collection systems are based on CapdetWorks.
CapdetWorks provides the following cost curve for aerobic WWTP, based on the flow rate: O&M cost
($MM) = 0.00042 x Q0-7939.
5-232
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
• Annual Benefits: Stabilized sludge can be land applied as fertilizer. The cogeneration option provides
electricity. Flared gas can be used elsewhere at the plant to reduce overall energy requirements.
5.4.2.3 Sector-Level Trends and Considerations
The reader should bear in mind throughout the analysis that the wastewater sanitation technology is likely to
be fixed by external factors, and improvements in technology will be driven by the population's desire/capacity for
improved sanitation and hygiene, with any improvements to GHG emissions a secondary result of the change.
Thus, although abatement measures are available, they should not be considered to be a viable control measure
that could be implemented for the sole purpose of reducing a country's GHG emissions, but rather a by-product of
a country's position on the sanitation ladder.
The MAC analysis is based on project costs developed for a set of model facilities based on the technical and
economic parameters discussed above. Similar to the steps taken in other sectors, we developed an inventory of
facilities that are representative of existing facilities. Next, we applied the abatement costs reported above to
calculate the break-even prices for each option and wastewater treatment scenario. Finally, the model estimates
the mitigation potential based on the country-specific share of emissions attributed to each wastewater treatment
scenario. Figure 5-7 shows the organization of the domestic wastewater MAC model. The country-specific
distributions are based on an analysis conducted by the EPA (2012).
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-233
-------
METHODOLOGY DOCUMENTATION
Figure 5-7: Domestic Wastewater MAC Analysis Flow Chart
National Population,
Rural
Urban
> r
Aerobic WWTP
(Individual or
centralized)
Anaerobic
Sludge Digester
with Co-Gen
Collection
System and
Aerobic WWTP
Anaerobic
Sludge Digester
with Co-Gen
Aerobic WWTP
Aerobic WWTP
National Population.
Rural
Urban
Aerobic WWTP
(Individual or
centralized)
Anaerobic
Sludge Digester
with Co-Gen
Collection
System and
Aerobic WWTP
Anaerobic
Sludge Digester
with Co-Gen
Aerobic WWTP
Aerobic WWTP
5-234
Global Non-CCk Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
Rural
Urban
Anaerobic
Sludge Digester
with Co-Gen
Aerobic WWTP
Collection
System and
Aerobic WWTP
Aerobic WWTP
(Individual or
centralized)
Anaerobic
Sludge Digester
with Co-Gen
Aerobic WWTP
National Population,
The analysis allocates, when information is available, a percentage of annual emissions to domestic
wastewater treatment. For each country, the remaining share of emissions is allocated to industrial wastewater
treatment.
Shares allocated to each source (domestic/industrial) were based on historical emission data obtained from
the UNFCCC's GHG emission reporting database. Data were limited to 24 A1 countries accounting for 15% of
emissions in 2010. For these 24 countries, we calculated a 5-year average share of ChU emissions attributable to
domestic sources based on emissions reported between 2002 and 2007. For all other countries, because of a lack
of data, we assumed emission projections are wholly attributable to domestic wastewater treatment systems to be
consistent with EPA's (2012) projections methodology.
The analysis also leverages estimated changes in wastewater disposal activity along each wastewater
treatment pathway discussed earlier in this section. These data were obtained from previous EPA analyses used to
develop international wastewater projections. Trends in wastewater disposal activity are determined by
population projections, distribution of population between rural and urban settings, population density, and
wastewater flow rates per person. These parameters are used to estimate country- and technology-specific
abatement project costs.
Other trends applied for this analysis include increasing the technical applicability factor and technical
effectiveness factor. The technical applicability factor is assumed to increase at 1% per year between 2010 and
2030. The technical effectiveness factor increases at a similar rate, growing from 60% to 80% over the 20-year time
period. These assumptions were based on expert judgment and intended to reflect increases in both the adoption
of improved sanitation systems and improvements through learning best management practices for the alternative
treatment systems that reduce CFUemissions.
Global Non-C02 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-235
-------
METHODOLOGY DOCUMENTATION
5.4.2.4 References
Food and Agriculture Organization of the United Nations. 2016. FAO Statistical Yearbook 2016. Available online at
http://www.fao.Org/faostat/en/#data/cf
Finney, B., and R. Gearheart. 2004. WAWTTAR: Water and Wastewater Treatment Technologies Appropriate for
Reuse, http://firehole.humboldt.edu/wawttar/wawttar.html cP
Intergovernmental Panel on Climate Change. 2006. 2006IPCC Guidelines for National Greenhouse Gas Inventories:
Volume 5: Waste; Chapter 6: Wastewater Treatment and Discharge, http://www.ipcc-
nggip.iges.or.ip/public/2006gl/pdf/5 Volume5/V5 6 Ch6 Wastewater.pdfcf
United Nations. 2014. Department of Economic and Social Affairs, Population Division (2014). World Population
Prospects: The 2014 Revision.
United Nations. 2015. Department of Economic and Social Affairs, Population Division (2015). World Population
Prospects: The 2015 Revision.
United Nations Statistics Division. 2016. Environmental statistics database: percentage of population with
wastewater collection and treatment. Available online at
https://unstats.un.org/unsd/environment/wastewater.htm J
U.S. Environmental Protection Agency. 1980. Construction Costs for Municipal Wastewater Plants: 1973-1978.
EPA/430/9-80-003. Washington, DC: EPA.
U.S. Environmental Protection Agency. 2012. Global Anthropogenic Non-CC>2 Greenhouse Gas Emissions: 1990-
2030. EPA #430-R-12-006. Washington, DC: EPA.
5-236
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
-------
SECTION 5 — SECTOR-LEVEL METHODS
5.4.3 Other Waste
This source category includes solid fuel transformation and waste incineration.
5.4.3.1 Other Energy Projections Methodology
This source category solely comprises countries that report data to the UNFCCC database. The EPA did not
perform Tier 1 calculations for other energy sources. The EPA obtained historical values for 1990 through 2012 and
held 2015 through 2050 values constant at 2012 levels for each country.
5.4.3.2 Other Energy Mitigation Methodology
The EPA has not estimated mitigation potential from other energy because of the lack of available data on
mitigation options.
Global N011-CO2 Greenhouse Gas Emissions Projections & Marginal Abatement Cost Analysis
5-237
------- |