Renewable Fuel Standard (RFS) Program:
Standards for 2023-2025 and
Other Changes
Regulatory Impact Analysis
SEPA
United States
Environmental Protection
Agency
-------
Renewable Fuel Standard (RFS) Program:
Standards for 2023-2025 and
Other Changes
This technical report does not necessarily represent final EPA decisions
or positions. It is intended to present technical analysis of issues using
data that are currently available. The purpose in the release of such
reports is to facilitate the exchange of technical information and to
inform the public of technical developments.
Regulatory Impact Analysis
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
NOTICE
4>EPA
United States
Environmental Protection
Agency
EPA-420-R-23-015
June 2023
-------
Table of Contents
Executive Summary iii
Overview vii
List of Acronyms and Abbreviations ix
Chapter 1: Review of the Implementation of the Program 1
1.1 Progression of the Fuels Market 1
1.2 In-Use Consumption of Renewable Fuels 3
1.3 2010 Biofuel Projections Versus Reality 9
1.4 Gasoline, Diesel, and Crude Oil 15
1.5 Cellulosic Biofuel 23
1.6 Biodiesel and Renewable Diesel 24
1.7 Ethanol 28
1.8 Other Biofuels 34
1.9 RIN System and Prices 37
1.10 Carryover RIN Proj ections 44
1.11 Gasoline and Diesel Projections 53
Chapter 2: Baselines 62
2.1 No RFS Baseline 64
2.2 2022 Baseline 98
Chapter 3: Candidate Volumes and Volume Changes 101
3.1 Mix of Renewable Fuel Types for Candidate Volumes 101
3.2 Volume Changes Analyzed With Respect to the No RFS Baseline 103
3.3 Volume Changes Analyzed with Respect to the 2022 Baseline 107
3.4 2023 Supplemental Volume Requirement 109
Chapter 4: Environmental Impacts Ill
4.1 Air Quality Ill
4.2 Climate Change 120
4.3 Conversion of Wetlands, Ecosystems, and Wildlife Habitats 210
4.4 Soil and Water Quality 223
4.5 Water Quantity and Availability 238
4.6 Ecosystem Services 253
Chapter 5: Energy Security Impacts 256
5.1 Review of Historical Energy Security Literature 257
5.2 Review of Recent Energy Security Literature 260
5.3 Cost of Existing U.S. Energy Security Policies 267
5.4 Energy Security Impacts 270
Chapter 6: Rate of Production and Consumption of Renewable Fuel 277
6.1 Cellulosic Biofuel 277
6.2 Biomass-Based Diesel 296
6.3 Imported Sugarcane Ethanol 319
i
-------
6.4 Other Advanced Biofuel 321
6.5 Total Ethanol Consumption 322
6.6 Corn Ethanol 331
6.7 Conventional Biodiesel and Renewable Diesel 332
Chapter 7: Infrastructure 334
7.1 Biogas 334
7.2 Biodiesel 335
7.3 Renewable Diesel 338
7.4 Ethanol 339
7.5 Deliverability of Materials, Goods, and Products Other Than Renewable Fuel 349
Chapter 8: Other Factors 351
8.1 Job Creation 351
8.2 Rural Economic Development 356
8.3 Supply of Agricultural Commodities 357
8.4 Price of Agricultural Commodities 360
8.5 Food Prices 367
Chapter 9: Environmental Justice 371
9.1 Proximity Analysis of Facilities Participating in the RFS Program 372
9.2 Non-GHG Air Quality Impacts 381
9.3 Water & Soil Quality Impacts 382
9.4 Impacts on Fuel and Food Prices 384
9.5 Greenhouse Gas Impacts 386
9.6 Effects on Specific Populations of Concern 388
Chapter 10: Estimated Costs and Fuel Price Impacts 391
10.1 Renewable Fuel Costs 391
10.2 Gasoline, Diesel Fuel and Natural Gas Costs 422
10.3 Fuel Energy Density and Fuel Economy Cost 426
10.4 Costs 427
10.5 Estimated Fuel Price Impacts 439
10.6 Analysis of Alternative Scenarios 449
Chapter 11: Screening Analysis 462
11.1 Background 462
11.2 Screening Analysis Approach 464
11.3 Cost-to-Sales Ratio Result 464
11.4 Conclusion 465
ii
-------
Executive Summary
The Renewable Fuel Standard (RFS) program began in 2006 pursuant to the requirements
in Clean Air Act (CAA) section 21 l(o) that were added through the Energy Policy Act of 2005
(EPAct). The statutory requirements for the RFS program were subsequently amended and
extended through the Energy Independence and Security Act of 2007 (EISA). In addition to
increasing the number of renewable fuel categories from one to four, increasing the volume
targets, and extending those volume targets from 2012 to 2022, EISA also expanded the waiver
provisions in CAA section 21 l(o)(7) that authorize EPA to waive the statutory volume targets
under certain conditions.
The statute includes annual, nationally applicable volume targets through 2022 for
cellulosic biofuel, advanced biofuel, and total renewable fuel, and through 2012 for biomass-
based diesel (BBD). For years after those for which the statute specifies volume targets, the
statute directs EPA to establish volume requirements based on a review of implementation of the
program in prior years and an analysis of a set of specified factors. In order to effectuate those
volume requirements, through 2022 EPA must also translate them into percentage standards that
obligated parties then use to determine the compliance obligations that they must meet every
year. As discussed in Preamble Section VII, we are continuing to use percentage standards as the
implementing mechanism for 2023-2025.
In this action we are establishing the applicable volume targets for all four categories of
renewable fuel for the years 2023, 2024, and 2025, as well as establishing a supplemental
standard for 2023 to address the remand of the 2016 annual rule by the D.C. Circuit Court of
Appeals, in Americans for Clean Energy v. EPA, 864 F.3d 691 (2017) (hereinafter "ACE'). We
are also establishing the annual percentage standards for all four categories that will apply to
gasoline and diesel fuel produced or imported by obligated parties in 2023-2025, as well as the
percentage standard for the 2023 supplemental standard.
This Regulatory Impact Analysis (RIA) supports our rulemaking in several ways. First,
this RIA addresses our statutory obligations under CAA section 21 l(o)(2)(B)(ii) for determining
the applicable volume requirements for cellulosic biofuel, BBD, advanced biofuel, and total
renewable fuel. Specifically, this section of the statute directs us to establish the applicable
volumes based upon a review of the implementation of the program and an analysis of various
environmental, economic, and other factors. We provide this analysis here, in conjunction with
the analysis in the preamble and several technical support memoranda to the docket. Second, this
RIA supports the 2023 supplemental standard in response to the ACE remand. Among other
things, Chapters 3 and 6 describe the availability of renewable fuel to meet the supplemental
standard.
Table ES-1 summarizes certain potential impacts associated with the final volumes in this
rule, including both quantified and unquantified impacts. Tables ES-2 and 3 contain more detail
on the annual costs and energy security benefits respectively. The table is not a comprehensive
listing of all the potential impacts that EPA considered in this rulemaking. The inclusion of an
impact in this table also does not indicate that EPA gave it greater weight than impacts not listed
in this table. A full discussion of each impact, including the uncertainties associated with
in
-------
estimating the impact, is contained in the RIA Chapter identified under the "More Information"
column. EPA compiled this table to provide additional information to the public regarding this
rulemaking and to comply with Circular A-4.
iv
-------
Table ES-1: Potential Quantified and Unquantified Impacts Associated with the Final Volumes in this Rule"
Potential impacts associated
Effect
More
with the volumes in this rule
Effect
Effect Quantified
Monetized
Information
Increases in CO, NH3, NOx, PM10, PM2.5, SO2 and VOC emissions associated with biorefinery
production and product transport.
Emission inventory impacts
-
4.1
Impacts on air quality from biofuel
production and use
Higher ambient concentrations of NOx, HCHO and SO2 downwind of production facilities
Emission inventory impacts
-
4.1
Varying emission impacts from vehicles running on ethanol blends
Emission inventory impacts
-
4.1
Decrease for THC, CO, and PM2.5, but increase slightly for NOx emissions from pre-2007 diesels
running on biodiesel
Emission inventory impacts
-
4.1
Impacts on climate change from
biofuel feedstocks production and
Reduced GHG Emissions
Illustrative
Illustrative
4.2
displacement of petroleum fuels
Increased conversion of pasture, grasslands and other habitats to cropland
Qualitative
-
4.3
Impacts on wetlands, ecosystems,
and wildlife habitat from land use
Decreased plant diversity and decreased natural forage for wildlife, particularly for birds and insects
Qualitative
-
4.3
change
Increased use of pesticides and reduced access to natural forage leading to reduced insect
biodiversity, especially in pollinators
Qualitative
-
4.3
Increased erosion, fertilizer and pesticide runoff and/or leachate
Qualitative
-
4.4
Depletion of natural soil organic matter, thereby depleting the soils nutrients
Qualitative
-
4.4
Impacts on soil and water quality
from biofuel feedstock production
Chemical contamination from releases and spills
Qualitative
4.4
Increased erosion from tilling and other land management practices
Qualitative
-
4.4
Increased chance of a cyanobacterium bloom occuring
Qualitative
4.4
Increased turbidity and sedimentation in aquatic ecosystems; Nutrient loading in waterways;
Increased stress on aquatic organisms
Qualitative
-
4.4
Impacts on water quantity and
Aquifer depletion
Qualitative
-
4.5
availability from biofuel and
feedstock production
Use of limited water resources for irrigation instead of meeting human needs
Qualitative
-
4.5
Energy security
Increased energy security
Energy security benefits
$513 million
5
Production and use of renewable
fuels
Increased production and use of renewable fuels
Increased production and use of
renewable fuels
-
6
Increased development of infrastructure of deliver and use renewable fuels
Qualitative
7
Infrastructure
No adverse impact on deliverability of materials, goods, and products other than renewable fuel
Qualitative
7
Jobs
Increased employment
Qualitative
-
8.1
Rural economic development
Support for rural economic development associated with biofuel and feedstock production
Qualitative
8.2
Commodity supply and price
impacts
Increased supply of certain agricultural commodities
Qualitative
8.3
Higher corn, soybean, and soybean oil prices
Commodity price increases
-
8.4
Higher food prices
Food price increases
-
8.5
Increased societal cost
Fuel cost increases
$23.8 billion
10.4
Costs
Changes to costs to consumers of transportation fuel
Cost changes
-
10.5
Increased costs to transport goods
Cost increases
-
10.5
a This table includes both societal costs and benefits (fuel costs, energy security, GHG emissions) as well as distributional effects or transfers (jobs, rural
economic development, etc.).
v
-------
Table ES-2: Fuel Costs of the 2023-2025 Volumes (2022 dollars, millions)"
Year
Discount Rate
0%
3%
7%
2023
Excluding Supplemental Standard
Including Supplemental Standard
$8,110
$8,738
$8,110
$8,738
$8,110
$8,738
2024
$7,352
$7,138
$6,871
2025
$8,455
$7,970
$7,385
Cumulative Discounted Costs
Excluding Supplemental Standard
Including Supplemental Standard
$23,917
$24,545
$23,218
$23,846
$22,366
$22,994
a These costs represent the costs of producing and using biofuels relative to the petroleum fuels they displace. They
do not include other factors, such as the potential impacts on soil and water quality or potential GHG reduction
benefits.
Table ES-3: Energy Security Benefits
of the 2023-2025 Volumes (2022 dollars, millions)
Discount Rate
Year
0%
3%
7%
2023
Excluding Supplemental Standard
Including Supplemental Standard
$180
$192
$180
$192
$180
$192
2024
$161
$156
$150
2025
$175
$165
$153
Cumulative Discounted Benefits
Excluding Supplemental Standard
Including Supplemental Standard
$515
$528
$501
$513
$483
$495
vi
-------
Overview
Chapter 1: Review of the Implementation of the Program
This chapter reviews the implementation of the RFS program, focusing on renewable fuel
production and use in the transportation sector since the RFS program began.
Chapter 2: Baselines
This chapter identifies the appropriate baselines for comparison.
Chapter 3: Candidate Volumes and Volume Changes
This chapter identifies the specific biofuel types and associated feedstocks that are projected to
be used to meet the final volume requirements.
Chapter 4: Environmental Impacts
This chapter discusses the environmental factors EPA analyzed in developing the final volume
requirements.
Chapter 5. Energy Security Impacts
This chapter reviews the literature on energy security impacts associated with petroleum
consumption and imports and summarizes EPA's estimates of the benefits that would result from
the final volume requirements.
Chapter 6: Rate of Production and Consumption of Renewable Fuel
This chapter discusses the expected annual rate of future commercial production of renewable
fuels, including advanced biofuels in each category (cellulosic biofuel and BBD).
Chapter 7: Infrastructure
This chapter analyzes the impact of renewable fuels on the distribution infrastructure of the U.S.
Chapter 8: Other Factors
This chapter provides greater detail on our evaluation of impacts of renewable fuels on job
creation, rural economic development, supply and price of agricultural commodities, and food
prices.
Chapter 9: Environmental Justice
This chapter describes potential environmental justice impacts associated with the production
and use of renewable fuels.
Chapter 10: Estimated Costs and Fuel Price Impacts
This chapter assess the impact of the use of renewable fuels on the social cost, the cost to
consumers of transportation fuel, and on the cost to transport goods.
Chapter 11: Screening Analysis
This chapter discusses EPA's screening analysis evaluating the potential impacts of the final
RFS standards on small entities.
vii
-------
Note: Unless otherwise stated, all documents cited in this document are available in the docket
for this action (EPA-HQ-OAR-2021-0427). We have generally not included in the docket
Federal Register notices, court cases, statutes, or regulations. These materials are easily
accessible to the public via the Internet and other means.
-------
List of Acronyms and Abbreviations
Numerous acronyms and abbreviations are included in this document. While this may not be an
exhaustive list, to ease the reading of this document and for reference purposes, the following
acronyms and abbreviations are defined here:
AAA
American Automobile Association
ACE
Americans for Clean Energy v. EPA, 864 F.3d 691 (2017)
AEO
Annual Energy Outlook
ASTM
American Society for Testing and Materials
BBD
Biomass-Based Diesel
bbl
Barrel
BOB
Gasoline Before Oxygenate Blending
bpd
Barrels Per Day
CAA
Clean Air Act
CAFE
Corporate Average Fuel Economy
CBI
Confidential Business Information
CBOB
Conventional Gasoline Before Oxygenate Blending
CG
Conventional Gasoline
CI
Carbon Intensity
CNG
Compressed Natural Gas
CO
Carbon Monoxide
CWC
Cellulosic Waiver Credit
DCO
Distillers Corn Oil
DDGS
Dried Distillers Grains with Solubles
DGS
Distillers Grains with Solubles
DOE
U.S. Department of Energy
DRIA
Draft Regulatory Impact Analysis
EIA
U.S. Energy Information Administration
EISA
Energy Independence and Security Act of 2007
EJ
Environmental Justice
EMTS
EPA-Moderated Transaction System
EO
Executive Order
EPA
U.S. Environmental Protection Agency
EPAct
Energy Policy Act of 2005
EV
Electric Vehicle
FFV
Flex-Fuel Vehicle
FOG
Fats, Oils, and Greases
gal
Gallon
GDP
Gross Domestic Product
GHG
Greenhouse Gas
GREET
Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model
LCA
Lifecycle Analysis
IEA
International Energy Agency
IEO
International Energy Outlook
IPCC
Intergovernmental Panel on Climate Change
IX
-------
LCFS
Low Carbon Fuel Standard
LNG
Liquified Natural Gas
MMBD
Million Barrels per Day
MSW
Municipal Solid Waste
MTBE
Methyl Tertiary Butyl Ether
MY
Model Year
NAICS
North American Industry Classification System
NASS
National Agricultural Statistics Service
NEMS
National Energy Modeling System
NGLs
Natural Gas Liquids
NHTSA
National Highway Transportation Administration
NOx
Nitrogen Oxides
NREL
National Renewable Energy Laboratory
OPEC
Organization of Petroleum Exporting Countries
OPIS
Oil Price Information Service
ORNL
Oak Ridge National Laboratory
PADD
Petroleum Administration for Defense District
PHEV
Plug-in Hybrid Electric Vehicle
PM
Particulate Matter
PTD
Product Transfer Document
RBOB
Reformulated Gasoline Before Oxygenate Blending
RFA
Regulatory Flexibility Act
RFF
Resources for the Future
RFG
Reformulated Gasoline
RFRA
Renewable Fuels Reinvestment Act
RFS
Renewable Fuel Standard
RIA
Regulatory Impact Analysis
RIN
Renewable Identification Number
RNG
Renewable Natural Gas
RVO
Renewable Volume Obligation
RVP
Reid Vapor Pressure
SBA
Small Business Administration
SBREFA
Small Business Regulatory Enforcement Fairness Act of 1996
SES
Socioeconomic Status
SOx
Sulfur Oxides
SPR
Strategic Petroleum Reserve
SRE
Small Refinery Exemption
STEO
Short Term Energy Outlook
UCO
Used Cooking Oil
ULSD
Ultra-Low-Sulfur Diesel
USD A
U.S. Department of Agriculture
USGCRP
U.S. Global Change Research Program
VEETC
Volumetric Ethanol Excise Tax Credit
VOC
Volatile Organic Compounds
WTI
West Texas Intermediate
-------
Chapter 1: Review of the Implementation of the Program
The statute directs EPA to establish volumes based on several factors, including "a
review of the implementation of the program during calendar years specified in the tables
CAA section 21 l(o)(2)(B)(ii). This chapter reviews the implementation of the RFS program in
previous years, focusing on renewable fuel production and use in the transportation sector since
the beginning of the RFS program. Of particular interest is a comparison of what the
expectations were when the RFS program was initially designed and implemented to what
actually occurred, and an investigation into the reasons that the renewable fuels market
developed as it did. To this end, the focus of this chapter is on factors related to the production
and use of renewable fuels:
• Feedstock availability, production, and collection
• Renewable fuel production technology and capacity
• Distribution, storage, blending, and dispensing of renewable fuels
• The consumption of renewable fuels in vehicles and engines
1.1 Progression of the Fuels Market
At the time that the RFS program was initially created by the Energy Policy Act of 2005
(EPAct), the transportation fuels market was already undergoing changes. Multiple state bans on
the use of methyl tertiary butyl ether (MTBE) in gasoline—due to concerns about leaking
underground storage tanks and groundwater contamination—had caused refiners to look for
replacement sources of high octane gasoline blendstocks. Crude oil prices had also begun to rise
over the lower levels seen in the previous decade, improving the relative economic value of
alternative fuels. Both of these factors provided an incentive for the increased use of ethanol in
gasoline even before the RFS program went into effect.
Congressional activity related to MTBE also had an impact on ethanol use in the years
leading up to the EPAct. For instance, Congress had considered providing liability protection to
refiners using MTBE under the premise that they had no choice but to use an oxygenate in the
reformulated gasoline (RFG) and oxyfuels programs.1 Congressional consideration of some sort
of liability protection for refiners, as well as the lack of sufficient infrastructure between 2000-
2005 for distributing and blending ethanol, likely contributed to the continued use of MTBE
despite state bans and concerns expressed by EPA and the public about MTBE in the years prior
to and including 2005.2
Ultimately, however, Congress rejected any form of liability protection for MTBE in the
EPAct. While the EPAct did not include a nationwide ban on the use of MTBE, it did remove the
RFG oxygen mandate, eliminating any argument that MTBE use was necessary to comply with
the statute. In combination with the removal of the RFG oxygen mandate, the creation of the
RFS program, and the increased economic value of ethanol in light of increasing crude oil prices,
1 "Timeline - A Very Short History of MTBE in the US," available at
https://www.icis.com/explore/resonrces/news/2006/07/05/1070674/timeiine-a-veiy-short-historo-of-mtbe-in-the-ns.
2 "Clinton-Gore Administration Acts To Eliminate MTBE, Boost Ethanol," available at
https://www.epa.gov/archive/epapages/newsroom arc hive/newsreleases/2054b28bf.l. 55afaa852568a80066c805.html
1
-------
refiners now had increased disincentives to continue using MTBE after 2005. In addition,
although the oxygen requirement for RFG was removed in the EPAct, the emission standards for
RFG were neither eliminated nor modified.3 Without MTBE, something was quickly needed to
replace the lost volume and octane that had been provided by MTBE while also ensuring that the
RFG emission standards would continue to be met. The net result of these factors is that the
market made a dramatic shift away from MTBE to ethanol in a very short period of time. By the
end of 2006, MTBE use in gasoline had fallen by about 80% in comparison to 2005 levels and
by 2007 was essentially zero, while ethanol replaced MTBE on an almost one-for-one energy-
equivalent basis over those same two years, growing by 56%.4 The sudden demand for ethanol
use in RFG areas, representing about one-third of all gasoline, was so great that its use was
temporarily reduced in much of the rest of the country (conventional gasoline (CG) areas) where
ethanol was not needed to meet state fuel program requirements until additional ethanol supply
could be brought online. This occurred despite the fact that E10 in CG areas benefitted from a 1
psi Reid Vapor Pressure (RVP) waiver, while RFG's emission standards precluded that waiver.
After the RFS program first went into effect in 2006, other factors continued to affect the
biofuels market. Crude oil prices continued to rise, state mandates for ethanol and biodiesel use
expanded, California's Low Carbon Fuel Standard (LCFS) program was implemented, and
foreign demand for biofuels increased. At the same time, the federal ethanol tax subsidy expired
at the end of 2011,5 and the federal oxygenated fuels (oxyfuels) program was largely phased out
as areas came into attainment with ambient wintertime carbon monoxide (CO) standards.
Furthermore, the statutory requirements for the RFS program were amended by the Energy
Independence and Security Act of 2007 (EISA), replacing the single total renewable fuel
standard with four nested standards (cellulosic biofuel, BBD, advanced biofuel, and total
renewable fuel). EPA implemented these changes through what became known as the RFS2
program, which began in the midst of these other changes, first with a single but considerably
higher total renewable fuel standard in 2009 compared to previous years, and then with the
addition of separate standards for cellulosic biofuel, BBD, and advanced biofuel beginning in
2010. In the following years, cellulosic ethanol production struggled to develop despite
Congressional aspirations, and increases in ethanol use slowed as the nationwide average ethanol
concentration approached 10%.6 BBD volume, in contrast, expanded beyond the Congressional
targets, outcompeting other advanced biofuels with the help of an ongoing tax incentive, and
EPA reflected this by setting higher BBD volume requirements for years after 2012.
The history of the progression of the fuels market indicates that consumption of
renewable fuels has been a function of many factors, of which the RFS program was only one.
Many of these factors can be expected to contribute to renewable fuel production and
consumption in the future. These factors include other federal and state fuels programs and
incentives, the octane value of ethanol, and foreign demand for renewable fuel.
3 See 40 CFR 80.41(e) and (f).
4 Based on EPA batch data: https://www.epa.gov/fuels-registration-reporting-and-coiiipIiance-help/gasoIine-
properties-over-time (excludes California).
5 The Volumetric Ethanol Excise Tax Credit (VEETC) was instituted through the American Jobs Creation Act of
2004 and the deadline was extended to December 31, 2011, through the Renewable Fuels Reinvestment Act
(RFRA).
6 Here and elsewhere in this document, "ethanol concentration" refers to the concentration of denatured ethanol in
gasoline.
2
-------
1.2 In-Use Consumption of Renewable Fuels
There are several reasons why actual renewable use may differ from the renewable fuel
volume targets specified in the statute or even the volumes required through the RFS regulations.
First, the statutory provisions of the RFS program provide EPA with several waiver authorities to
reduce the statutory volumes under particular circumstances.7 The statutory volumes minus
waived volumes equal applicable volumes. In turn, the applicable percentage standards, which
are the mechanism through which the obligations of an individual "obligated party" are
determined under the RFS program, are based on the applicable volumes.8
The "general waiver" authority at CAA section 21 l(o)(7)(A) was enacted by EPAct and
maintained in EISA. It permits EPA to reduce any of the four applicable volume targets in the
statute if EPA makes one of the following findings:
(i) based on a determination by the Administrator, after public notice and opportunity for
comment, that implementation of the requirement would severely harm the economy or
environment of a State, a region, or the United States; or
(ii) based on a determination by the Administrator, after public notice and opportunity for
comment, that there is an inadequate domestic supply.
The "cellulosic waiver" authority at CAA section 21 l(o)(7)(D) was introduced by EISA.
It requires (not merely permits) EPA to reduce the statutory cellulosic volume target to the
projected volume available in years that the projected volume of cellulosic biofuel production is
less than the statutory target. When making such a reduction, EPA may also reduce the statutory
volume targets for total renewable fuel and advanced biofuels by the same or a lesser volume.
The "biomass-based diesel waiver" authority at CAA section 21 l(o)(7)(E) was also
introduced by EISA. It requires a reduction from the statutory BBD volume for up to 60 days if
EPA determines that there is a significant renewable feedstock disruption or other market
circumstances that would make the price of BBD increase significantly. When making such a
reduction in BBD volume, EPA may also reduce the statutory volume targets for total renewable
fuel and advanced biofuels by the same or a lesser volume, similar to the cellulosic waiver
authority.
The "reset" authority at CAA section 21 l(o)(7)(F) was also introduced by EISA. It
requires EPA, after 2015, to modify the volumes in the tables in CAA section 21 l(o)(2)(B) in
compliance with CAA section 21 l(o)(2)(B)(ii) if EPA either: waives 20 percent of the volumes
in any table in CAA section 21 l(o)(2)(B) for two consecutive years or waives 50 percent of the
volumes in one year.
The statute only specifies volume targets for BBD for 2009 through 2012, and EPA did
not reduce the statutory target for any of those years under either the general or BBD waiver
7 CAA section 21 l(o)(7).
8 Obligated parties are producers and importers of gasoline and diesel. See 40 CFR 80.1406.
3
-------
authorities. Under the cellulosic waiver authority, however, EPA has reduced the statutory target
for cellulosic biofuel in every year since 2010 and the statutory targets for advanced biofuel and
total renewable fuel in every year since 2014.
EPA has used the general waiver authority only once, for the 2016 compliance year,
based on a finding of inadequate domestic supply.9 However, the D.C. Circuit vacated EPA's use
of this waiver authority in Americans for Clean Energy v. EPA, 864 F.3d 691 (D.C. Cir. 2017)
("ACE'). Specifically, the court found that EPA had impermissibly considered demand-side
factors in its assessment of inadequate domestic supply, rather than limiting that assessment to
supply-side factors. The court remanded the rule back to EPA for further consideration in light of
its ruling. EPA took the first step to respond to that remand when it established the applicable
volume requirements for 2022,10 and is completing its response to that remand in this
rulemaking.
In addition to the waiver authorities mentioned above, there are at least five other reasons
why actual renewable fuel use may differ from either the statutory or applicable volume
requirements in any given year. The first is that the percentage standards are based on projected
volumes of non-renewable gasoline and diesel consumption provided by the U.S. Energy
Information Administration (EIA), which typically deviate to some degree from what actually
occurs. For compliance years 2007-2022, the EIA source for such projections was the Short
Term Energy Outlook (STEO);11 for compliance years 2023 and later, the EIA source for such
projections is the Annual Energy Outlook (AEO).
Since the first percentage standard was applied in 2007, this forecast has both over- and
under-predicted actual consumption. In the event that the actual consumption of non-renewable
gasoline and diesel is lower than the projection that EPA used to set the applicable percentage
standards, the obligations applicable to individual obligated parties are likewise lower and, all
other things being equal, the actual volumes of renewable fuel used as transportation fuel will
fall short of the volumes EPA used in setting the percentage standards. Likewise, if the actual
consumption of non-renewable gasoline and diesel is higher than the projection that EPA used to
set the applicable percentage standards, the actual volumes of renewable fuel used as
transportation fuel will exceed the volumes EPA used in setting the percentage standards.
Despite the fact that the statute directs EPA to set standards that ensure that transportation fuel
sold or introduced into commerce contains the applicable volumes of renewable fuel, the statute
also directs EPA to use projections of gasoline and diesel for this purpose, and does not mandate
that EPA correct the volume requirements based on deviations in those projections from the
volumes actually consumed.
Another reason that the volume requirements may not be reached by the market in a
particular year is related to the credit system that is used to demonstrate compliance with the
RFS program.12 These credits are called Renewable Identification Numbers, or "RINs."
9 80 FR 77420 (December 14, 2015).
10 87 FR 36900 (July 1, 2022).
11 CAA section 21 l(o)(3)(A).
12 CAA section 21 l(o)(5) establishes the provisions for credits under the RFS program. This system is discussed in
more detail in Chapter 1.9.
4
-------
Obligated parties have the flexibility to use RINs representing renewable fuel produced in the
previous year, often called "carryover RINs," to demonstrate compliance rather than by using
RINs representing current year renewable fuel production.13'14 The nationwide total number of
carryover RINs grew dramatically in the early years of the RFS program, and obligated parties
have at times drawn down the total number of carryover RINs to help fulfill their obligations.
For instance, consumption of renewable fuels fell more than 500 million ethanol-equivalent
gallons short of the applicable volume requirement in 2013, and obligated parties used carryover
RINs to make up the shortfall. See Chapter 1.10 for further discussion of carryover RINs.
The third reason that the applicable volume requirements may vary from actual
renewable fuel use is the difficulty in projecting the future market's ability to make available and
consume renewable fuels. For instance, in several cases, producers of cellulosic biofuel made
plans that did not come to fruition, such as Cello Energy, Range Fuels, and KiOR.15 In the past,
there was also considerable uncertainty associated with estimating the ability of the RFS
standards to incentivize increases in the consumption of ethanol above the E10 blendwall.16
Other unforeseen circumstances, such as the drought in 2012 that adversely affected crops yields
and the impacts of the COVID-19 pandemic in 2020, have also contributed to shortfalls in
renewable fuel production in comparison to the intended volume requirements. By contrast, in
some other years, the market used more renewable fuel than what EPA projected, typically when
the economics of doing so were favorable or as a result of other incentives such as state Low
Carbon Fuel Standard (LCFS) programs.
A fourth reason that the applicable volume requirements may vary from actual renewable
fuel use is that there are other drivers for renewable fuel use besides the RFS program. For
instance, as discussed in Chapter 1.1, in the early years of the RFS program, renewable fuel use
significantly outpaced the RFS requirements, spurred by the transition from MTBE to ethanol as
an oxygenate. We discuss numerous other, non-RFS economic drivers for renewable fuel use
throughout this section.
Finally, exemptions given to small refineries in past years due to disproportionate
economic hardship have effectively reduced the required volume of renewable fuel for those
years in comparison to the volumes on which the percentage standards were based. Small
refineries may request these exemptions under CAA section 21 l(o)(9)(B) and are evaluated on a
refinery-by-refinery basis. In cases where a small refinery exemption (SRE) was granted after the
applicable percentage standards were set, the percentage standards remained unchanged but were
then applicable to a smaller number of parties, resulting in smaller effective aggregate renewable
fuel requirements.
Historically, once the percentage standards were established for a given year, EPA has
not adjusted them to account for SREs that were subsequently granted. Rather, from the start of
the RFS program through the 2019 compliance year, EPA's standard-setting process only
13 This flexibility is a function of the two-year life of RINs as discussed more fully in Chapter 1.9.
14 The use of previous-year RINs for compliance with the applicable standards is limited to 20% of an obligated
party's Renewable Volume Obligation (RVO). See 40 CFR 80.1427(a)(5).
15 80 FR 77506 (December 14, 2015).
16 80 FR 77457 (December 14, 2015).
5
-------
accounted for SREs that had been granted at the time of the final annual rule. In essence, this
meant that non-exempt obligated parties did not have to make up for volumes that would not be
attained by the exempt small refineries.17 This approach is consistent with that taken for the
projected non-renewable gasoline and diesel volumes used to calculate the percentage standards,
where errors in projected volumes could likewise result in actual consumption of renewable fuel
falling short of the intended volume requirements.
Figure 1.2-1: Volume of SREs Granted After the Applicable Percentage Standards Were
Set, By Compliance Year3
2000
1800
1600
1400
2 1200
EC
§ 1000
I 800
600
400
200
0
Compliance year
a No SREs have been granted for years after RFS compliance year 2018. This chart shows the impact of certain
SREs previously granted for compliance years 2016, 2017, and 2018 that have since been remanded, reconsidered,
and denied. However, as a result of subsequent EPA action, these small refineries were required to resubmit their
RFS annual compliance reports with zero deficit carryforward and no additional RIN retirements. See "April 2022
Alternative RFS Compliance Demonstration Approach for Certain Small Refineries," EPA-420-R-22-006, April
2022; see also "June 2022 Alternative RFS Compliance Demonstration Approach for Certain Small Refineries,"
EPA-420-R-22-012, June 2022.
As shown in Figure 1.2-1, SREs granted after the standards were set varied significantly
by compliance year. However, these SREs did not necessarily translate into an equivalent
reduction in actual consumption of renewable fuel. Other factors also played a role in
determining whether and when actual consumption was affected by SREs.18 For instance, the
combination of the economic attractiveness of marketing ethanol to consumers as E10 and the
infrastructure to blend, distribute, and dispense E10, along with longer-term contracts for ethanol
blending, meant that the nationwide average ethanol concentration remained very near 10.00%
ethanol even when large numbers of SREs were granted.
¦ Advanced biofuel
¦ Conventional renewable fuel
I....ill
2011 2012 2013 2014 2015 2016 2017 2018
17 75 FR 76805 (December 9, 2010).
18 Another action that decoupled actual consumption from a specific year's RFS standards is the Alternative RIN
Retirement Schedule, which provides additional time and opens a broader range of RIN vintages for small refineries
to comply with their 2020 RFS obligations through a series of quarterly retirement deadlines through February 1,
2024. 87 FR 54158 (September 2, 2022).
6
-------
With regard to the timing of the impacts, SREs generally affected the demand for RINs in
the calendar year in which they were granted and the following years, rather than in the RFS
compliance year to which they applied, as shown in Figure 1.2-2. This was often due to EPA
granting the SREs after the compliance year had passed.
Figure 1.2-2: Volume of SREs Granted After the Applicable Percentage Standards Were
Set, By Calendar Year When Exemptions Were Granted3
2000
2011 2012 2013 2014 2015 2016 2017 2018 2019
Calendar year
a No SREs have been granted after calendar year 2019. This chart shows the impact of certain SREs previously
granted for compliance years 2016, 2017, and 2018 in calendaryear 2019 that have since been remanded,
reconsidered, and denied. However, as a result of subsequent EPA action, these small refineries were required to
resubmit their RFS annual compliance reports with zero deficit carryforward and no additional RIN retirements. See
"April 2022 Alternative RFS Compliance Demonstration Approach for Certain Small Refineries," EPA-420-R-22-
006, April 2022; see also "June 2022 Alternative RFS Compliance Demonstration Approach for Certain Small
Refineries," EPA-420-R-22-012, June 2022.
However, it was not always the case that SREs affected the demand for RINs only in the
calendar year in which they were granted and the following years. For instance, some small
refineries adjusted their RIN acquisition efforts to reflect anticipated grants of their SRE
petitions, effectively resulting in SREs having a market impact before they were actually
granted. In all or almost all cases, a small refinery that was granted an exemption continued to
blend renewable fuel into its own gasoline and diesel due to the economic attractiveness of doing
so. In such cases, the total number of RINs generated may not have been reduced by the SRE,
but the number of carryover RINs may have increased. Finally, as discussed above, higher-than-
projected gasoline and diesel demand could offset the effect of SREs to some degree.
In the final rule that established the original 2020 standards, EPA revised the RFS
regulations to account for a projection of exempt small refinery volumes, increasing the 2020
percentage standards applicable to non-exempt refineries.19 Given that EPA subsequently made a
decision not to exempt any volumes of gasoline and diesel for 2020 (i.e., no SREs were granted),
19 85 FR 7016 (February 6, 2020).
7
-------
the original 2020 percentage standards were applicable to a larger volume of gasoline and diesel,
effectively increasing the renewable fuel requirements.20
In sum, due to the many factors that affect renewable fuel consumption, actual
consumption has been both higher and lower than the volumes that EPA originally intended to be
achieved in setting the percentage standards.
Figure 1.2-3: Intended3 Versus Actual Consumption of Total Renewable Fuel
Source for actual consumption: EPA-Moderated Transaction System (EMTS)
a Intended volumes represent the volumes used to calculate the applicable percentage standards. As such, the
intended volumes do not account for the effects of SREs granted after the percentage standards were established,
errors in projected demand for gasoline and diesel, or the use of carryover RINs for compliance.
b The "intended consumption" for 2020 represents the 2020 rulemaking that established the original 2020 standards
on February 6, 2020 (85 FR 7016), not the rulemaking that revised those standards on July 1, 2022 (87 FR 39600).
The total volume of renewable fuel that was intended to be used between 2007-2022
(i.e., the volume that was used to calculate the applicable percentage standards) was about 229
billion ethanol-equivalent gallons. In comparison, actual consumption was about 251 billion
ethanol-equivalent gallons over the same time period. Thus, actual consumption has exceeded
what was intended over the life of the RFS program through 2022. In 2007 and 2008, the
significant oversupply in comparison to the intended volumes was due primarily to the expansion
of E10 when the market as a whole had not yet reached the El0 blendwall and blending ethanol
as E10 was economically attractive relative to gasoline. In years after 2016, the significant
undersupply in comparison to the intended volumes affected all types of renewable fuel more
equitably rather than just ethanol, and was precipitated by a combination of the approval of SREs
after the applicable percentage standards had been set, lower than projected gasoline and diesel
consumption, and other economic factors.
Economic factors impact conventional renewable fuel and non-cellulosic advanced
biofuel differently. These factors include crude oil prices, renewable fuel production costs
(which are in turn a function of feedstock, process heat, and power costs), tax subsidies, and the
market pressures created by the RFS standards to increase ethanol use above the E10 blendwall.
211 We note, however, that on July 1, 2022, EPA revised the 2020 standards to account for the fact that no SREs were
granted for 2020, as well as to address impacts of the COVID-19 pandemic. See 87 FR 39600 (July 1, 2022).
8
-------
Economic factors are coupled with the use of carryover RINs for compliance, the number of
carryover RINs available, and deficit carry-forwards. In 2013, for instance, the implied
conventional renewable fuel standard was 13.8 billion gallons, which was considerably higher
than the E10 blendwall. The market responded by producing less conventional renewable fuel
but more non-cellulosic advanced biofuel than required. The net effect of these two outcomes
nevertheless still fell short of the applicable volume requirements, and the market thus relied on
some carryover RINs for compliance.
1.3 2010 Biofuel Projections Versus Reality
In the 2010 rule that established the RFS2 program, EPA projected volumes of each type
of renewable fuel that in the aggregate would meet the applicable volume targets in the statute
for cellulosic biofuel, BBD, advanced biofuel, and total renewable fuel.21 These projections did
not include any consideration of potential future waivers or any other factor that might cause the
statutory volumes not to be met. In reality, actual consumption of renewable fuel typically fell
short of the statutory targets for all renewable fuel categories except for BBD. Moreover, the
specific types of renewable fuel that were projected in 2010 to be used to fulfill the mandates
differed from what was actually used, most notably in regard to the relative amounts of ethanol
and non-ethanol renewable fuels.
This section highlights the aspirational nature of the statutory volume targets, especially
for cellulosic biofuel and its carry-through impact on advanced biofuel and total renewable fuel.
This section also highlights the difficulty in projecting the ability of the market to meet
applicable standards as well as the specific mix of biofuels that will be produced, imported, and
consumed.
1.3.1 Shortfalls in Comparison to Statutory Targets
As explained in Chapter 1.2, there are many reasons why actual use of renewable fuels
fell short of the statutory targets. Figure 1.3.1-1 compares the statutory targets to actual
consumption for the four categories of renewable fuel.
21 75 FR 14670 (March 26, 2010).
9
-------
Figure 1.3.1-1: Comparison of Statutory Volume Targets to Actual Consumption
Cellulosic Biofuel
Biomass-Based Diesel3
16,000 Statutory target ,
— — Statutory target
14,000 Actual consumption
3,000
Actual consumption
w 2,500
i 2,000 /
12,000 -f
2 /
cc 10,000
c ~
— 8 000 ^
i x
6,000 — '
4,000 ^ —
= 1,500 y
5 1,000 t ^ ' — — — — — — — — — — — — — -
500 /
2,000 — *
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
0
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Advanced Biofuel
Total Renewable Fuel
40,000
35 000 ~ — Statutory target
— — Statutory target
20,000 ; ^
^^^—Actual consumption y
S
Z 15,000 ^
~
1
10,000 ^
J*
Actual consumption y
w 30,000 *
1
§ 25,000 «*"
20,000 - X
10,000
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
^ ^
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Source for actual consumption: EMTS
a The statute specifies BBD volume targets only through 2012. Thereafter, the required BBD volume can be no less
than 1.0 billion gallons, but can be more based on an analysis of specified factors.
The significant shortfalls in advanced biofuel and total renewable fuel for more recent
years are primarily the result of shortfalls in cellulosic biofuel. This fact is more evident in
Figure 1.3.1-2, which shows that consumption is considerably closer to the implied statutory
volume targets for non-cellulosic advanced biofuel and conventional renewable fuel.
Figure 1.3.1-2: Comparison of Implied Statutory Volume Targets to Actual Consumption
Non-Cellulosic Advanced Biofuel3 Conventional Renewable Fuelb
c r\nr\ 1 c nnn
c rtno ~ — Implied statutory target S
b,UUO
Actual consumption
w 4'000 ^
5 / '
o 3'000 y *
^ 2,000 ^ '
1,000 *
w 14,000 ^ ^
J 13,000 _ |./ / \ /
1 .' v
12,000 '
— — Implied statutory target
11,000
Actual consumption
10,000
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Source for actual consumption: EMTS
a Non-cellulosic advanced biofuel represents D4 and D5 RINs.
b Conventional renewable fuel represents D6 RINs.
10
-------
The oversupply in non-cellulosic advanced biofuel between 2011 and 2017 partially
offset some of the shortfall in conventional renewable fuel in the same years, and also
contributed to increases in the number of carryover RINs in some years.
A direct comparison of shortfalls in consumption of cellulosic biofuel to shortfalls in the
other categories of renewable fuel makes it clear that consumption of advanced biofuel and total
renewable fuel was directly affected by the shortfall in cellulosic biofuel, while the consumption
of non-cellulosic advanced biofuel and conventional renewable fuel was not. This is to be
expected since the cellulosic biofuel category is nested within advanced biofuel and total
renewable fuel categories, but cellulosic biofuel is independent of non-cellulosic advanced
biofuel and conventional renewable fuel.
Figure 1.3.1-3: Comparative Shortfalls (Statutory Target Minus Actual Consumption)
18,000
16,000 ^^»Shortfall in cellulose biofuel *—
16,000 —Shortfall in cellulosic biofuel
14,000 Shortfall in advanced biofuel
12,000 —— - S
14,000 ^^—shortfall in total renewable fuel ^S
Shortfall in non-cellulosic advanced biofuel j
w 10,000
^ 8,000
12,000 Shortfall in conventional renewable fuel w f
z 10,000 /y
0 8'000
1 6,000
4,000
2,O0°o
-2,000
-4,000
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
-2,000
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Source for actual consumption: EMTS
1.3.2 Relative Proportions of Ethanol and Non-Ethanol Renewable Fuel
In the RFS2 rule, non-cellulosic advanced biofuel through 2022 was projected to be
composed of biodiesel, renewable diesel, and imported sugarcane ethanol. This has proved
largely true as volumes of renewable jet fuel, biogas, heating oil, domestic advanced ethanol, and
naphtha—the only other eligible advanced biofuels—have represented only a very small fraction
of non-cellulosic advanced biofuel consumption. However, the relative proportions of biodiesel,
renewable diesel, and imported sugarcane ethanol have been far different in actual consumption
than in the projections from the RFS2 rule.
11
-------
Figure 1.3.2-1: Volumetric Proportions of Each Fuel Type in Non-Cellulosic Advanced
Biofuel
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Actual Consumption
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
2010 Projection
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Source: "2010 projection" is from Table 1.2-3 of the RIA for the RFS2 rule. "Actual consumption" is from EMTS.
Actual consumption of imported sugarcane ethanol has been considerably lower than in
the 2010 projection, and consumption of advanced biodiesel and renewable diesel has been
higher. This outcome for imported sugarcane ethanol is mirrored in the outcome for total
ethanol: actual consumption of ethanol has been lower than the 2010 projection and actual
biodiesel and renewable diesel has been higher.
Figure 1.3.2-2: Actual Versus 2010 Projection of Ethanol Consumption in Non-Cellulosic
Renewable Fuel3
18,000
17,000
w 16,000
c
_o
c 15,000
o
S 14,000
13,000
12,000
Source: "2010 projection" is from Table 1.2-3 of the RIA for the RFS2 rule. "Actual consumption" is from EMTS.
a The 2010 projection of ethanol shown here represents the "primary control case" from the RFS2 rule. EPA also
analyzed a "low ethanol control case" and a "high ethanol control case".
2010 projection
Actual—
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
12
-------
Figure 1.3.2-3: Actual Versus 2010 Projection of Biodiesel + Renewable Diesel
Consumption in Non-Cellulosic Renewable Fuel
5,500
1,000
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Source: "2010 projection" is from Table 1.2-3 of the RIA for the RFS2 rule. "Actual consumption" is from EMTS.
This pattern of lower ethanol and higher non-ethanol volumes in comparison to
expectations appears to be linked to the E10 blendwall and the difficulty that the market has had
in increasing sales of higher-level ethanol blends (e.g., E15 and E85). The 2010 projections
included a significant volume of E85 that did not materialize. The result is that, rather than being
met entirely with corn ethanol as projected in 2010, the implied conventional renewable fuel
volume requirement has included volumes of ethanol up to and just slightly greater than the E10
blendwall, while biodiesel and renewable diesel have made up the difference.
Figure 1.3.2-4: Actual Versus 2010 Projection of Ethanol Consumption in Conventional
Renewable Fuel
18,000
17,000
2010 projection
„ 16,000 — Actual
C
o
Source: "2010 projection" is from Table 1.2-3 of the RIA for the RFS2 rule. "Actual consumption" is from EMTS.
The expectation at the time that EISA was enacted in 2007 was that the implied
conventional renewable fuel volume requirement could be met entirely with ethanol as E10
without the nation as a whole exceeding an average ethanol content of 10.00%, and without the
need for E15 or E85. This expectation was based on the assumption that gasoline demand would
continue to increase in the future, as had been projected by EIA since 2000. By the time RFS2
13
-------
regulations were finalized in 2010, however, EIA's Annual Energy Outlook (AEO) projected
that future gasoline demand was likely to decrease rather than increase.
Figure 1.3.2-5: EIA Projections of Future Gasoline Demand
24.0000
23.0000
22.0000
21.0000
S 20.0000
¦o
= 19.0000
a
18.0000
17.0000
16.0000
15.0000
While EPA's projections in the 2010 rule for how the statutory targets through 2022
might be met included significant volumes of drop-in renewable diesel, it also included total
ethanol volumes in excess of the implied statutory conventional renewable fuel volume targets;
EPA's projections assumed that substantial volumes of ethanol would also be used to meet the
cellulosic biofuel and implied non-cellulosic advanced biofuel volume targets. These projections
were based on what EPA believed at that time was reasonable to expect for production and
consumption of all renewable fuel types under the influence of the RFS standards, as well as the
growth in the flex-fuel vehicle (FFV) fleet and the availability of E85 at retail service stations
that would be needed in order for the projected ethanol volumes to be consumed (El5 had not
been approved at that time). Based on EPA's projections of total ethanol volume in the RFS2
rule and EIA's projection of gasoline demand in AEO 2010, the nationwide average ethanol
content would have first exceeded 10.00% in 2014 in the primary case and would have continued
upwards to 15.5% by 2022. In reality, the actual increase in the nationwide average ethanol
concentration over time has been considerably slower; the same is true even when ignoring
cellulosic ethanol (i.e., when comparing actual ethanol use to the projected volume of
conventional ethanol and non-cellulosic advanced ethanol such as imported sugarcane ethanol).
•AE02000
•AE02002
•AE02005
¦AE02007
•AEO2010
OnirMrn<3,LniDr-.ooa^o*HrNim'3,u->^Dr-.coa^OrHrNj
OOOOOOOOOOrHrHrHrHr-It—I*—It—It—lrH
-------
Figure 1.3.2-6: 2010 Projected Versus Actual Ethanol Concentration
16%
1%
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Source for actual ethanol concentration: Gasoline and ethanol consumption from EIA's Monthly Energy Review
The considerably slower-than-projected approach to and exceedance of the El 0
blendwall suggests that increasing sales of E85 were more difficult to achieve than either EPA or
ethanol proponents had projected it would be when the RFS program was established.
1.4 Gasoline, Diesel, and Crude Oil
This section compares crude oil prices with crude oil price projections, and discusses
observed changes in petroleum imports, refinery margins, and transportation fuel demand prior
to and during the years of the implementation of the RFS program.
1.4.1 Crude Oil Prices vs. Crude Oil Price Projections
Crude oil prices have a significant impact on the economics of increased use of
renewable fuels. When crude oil prices increase, both renewable fuel feedstock prices and
gasoline and diesel prices tend to increase as well, although gasoline and diesel prices generally
increase more relative to renewable fuel feedstock prices. Thus, higher crude oil prices generally
improve the economics of renewable fuels relative to gasoline and diesel. Conversely, lower
crude oil prices tend to hurt the economics of renewable fuels.
When EPA was projecting the cost of future renewable fuels for the RFS2 rule, crude oil
prices were very high compared to historical crude oil prices. For estimating the cost of
rulemakings, EPA uses projections for the future prices of petroleum products. The cost analysis
for the RFS2 rule was based on crude oil, gasoline, and diesel prices projected by EIA in AEO
2008, which projected crude oil prices for decades into the future. Figure 1.4.1-1 shows AEO
2008 projected crude oil prices, as well as actual crude oil prices, for both West Texas
Intermediate (WTI, a light, sweet crude produced in the U.S.) and Brent (a light, sweet European
15
-------
crude oil).22-23-24-25 When there was separation in Brent and WTI crude oil prices and Brent
prices were higher than WTI, Brent crude oil prices likely represented the marginal price of
crude oils purchased by U.S. refiners and set the marginal price of U.S. refined products, while
WTI tended to reflect crude purchase price for many U.S. refiners.
Figure 1.4.1-1: AEO 2008 Projected and Actual Crude Oil Prices (2007 dollars)3
120
AEO 2008 Reference Case AEO 2008 High Price Case
0 I
2000 2005 2010 2015 2020 2025 2030
Year
a Actual crude oil prices have been adjusted to 2007 dollars to be consistent with the value of money used in AEO
2008; 2022 data represents the first 6 months only.
Figure 1.4.1-1 shows actual crude oil prices beginning to increase in 2004 and reaching
an average price of nearly $100 per barrel in 2008. Furthermore, some reports at that time
projected even higher crude oil prices due to crude oil production not keeping up with demand.26
Nevertheless, EIA crude oil price projections during this time were much lower, and it was
during this time that the RFS2 rule was written. The AEO 2008 reference case projected crude
oil prices decreasing to under $60 per barrel and remaining that low all the way out to 2030.
Because the AEO 2008 reference case projected much lower crude oil prices than actual prices
and many other independent predictions at that time, EPA also analyzed the cost of the RFS2
program based on AEO 2008 high crude oil prices. The AEO 2008 high price case estimated
crude oil prices rising from $70 per barrel to mid-$90s per barrel out to 2030. Actual crude oil
prices decreased in 2014 back down to the $40 to $60 per barrel price range (after adjusting the
prices back to 2007 dollars—the dollar value used in AEO 2008), which were much lower than
the peak prices, but higher than the typical historical crude oil prices prior to 2004. In retrospect,
22 Light crude oils are comprised of more lower temperature boiling, shorter chain hydrocarbons, while heavy crude
oils are comprised of more higher temperature boiling, longer chain hydrocarbons. Sweet crude oils have less sulfur,
while sour crude oils have more sulfur. Increased sulfur in crude oils make them more expensive to refine to meet
gasoline and diesel sulfur specifications; thus, sour crude oils are typically priced lower than sweet crude oils.
23 AEO 2008 - Petroleum Product Prices; Reference Case; EIA; June 2008.
24 AEO 2008 - Petroleum Product Prices; High Price Case; EIA: June 2008.
25 Petroleum and Other Liquids - Spot Prices WTI - Cushing and Brent - Europe; EIA;
https://www.eia.gov/dnav/pet/pet pri spt si a.htm.
26 Hirsch, Robert L.; Peaking of World Oil Production: Impacts, Mitigation & Risk Management; Report to the
Department of Energy; February 2005.
16
-------
the reference case and high crude oil price projections of AEO 2008 essentially represented the
range of crude oil prices since the RFS2 program was promulgated.
1.4.2 Petroleum Imports
As discussed further in Chapter 5, energy security is an important goal of the RFS
program. Importing a significant amount of crude oil and finished petroleum products from
abroad creates an energy security concern if the foreign petroleum supply is disrupted. A good
example is the oil embargo by the Organization of Petroleum Exporting Countries (OPEC)
against the U.S. in 1973 and 1974, which drove up prices, reduced supply, and is attributed to
causing the U.S. economy to slide into a recession.27 It also led to Congress banning the export
of U.S. crude oil from 1975 to 2015.28
At the time that Congress passed EPAct and EISA and EPA promulgated the RFS1 and
RFS2 rules, the U.S. was importing a large portion of its crude oil. That trend was expected to
continue because the eventual increase in U.S. tight oil crude oil production was not known at
that time. Below we consider the petroleum trade imbalance during that time and what has
transpired since.
EIA collects data on imports of crude oil and petroleum products, receives data on crude
oil and petroleum product exports from the U.S. Bureau of the Census, and calculates net imports
of petroleum into the U.S.29 The EIA-reported net imports of petroleum values account for
imports and exports of crude oil, petroleum products, and biofuels.30 For the net imports figures
shown in Figure 1.4.2-1, the renewable fuel volumes were removed to only show the U.S. net
imports of petroleum for the years from 2000-2021. Because the production volume of U.S. tight
oil impacted the net petroleum imports in such a significant way, those volumes are also shown
in the figure, along with the individual net imports of gasoline and diesel.
27 Verrastro, Frank A., The Arab Oil Embargo-40 Years Later; Center for Strategic & International Studies; October
16, 2013.
28 1975 Energy Policy and Conservation Act; Consolidated Appropriations Act of 2016.
29 U.S. Net Imports by Country; Petroleum and Other Liquids; EIA;
https://www.eia.gov/dnav/pet/pet move neti a EP00 IMN mbbtpd ni.fatm.
30 To calculate net petroleum imports, EPA subtracted net biofuel imports from the U.S. Net Imports reported by
EIA.
17
-------
Figure 1.4.2-1: U.S. Net Petroleum Imports and U.S. Tight Oil Production
Year
Net Petroleum Imports Net Gasoline Imports
Net Distil late Imports Tight Oil Production
Figure 1.4.2-1 shows that net petroleum imports (crude oil and refined product net
imports) increased from just over 10 million barrels per day (bpd) in 2000 to a maximum of
about 12.5 million bpd in 2005. After peaking in 2005 when EPAct was passed, net petroleum
imports started to decrease, very slowly at first, in 2006 and 2007. Starting in 2008, net
petroleum imports declined each year by roughly 1 million bpd.
Increased tight oil production and changes in gasoline and distillate (comprised largely of
diesel) net imports were responsible for reducing net petroleum imports. Figure 1.4.2-1 clearly
shows that tight oil production—which increased from about zero in 2009 to 8 million bpd in
2018—had a very large impact on net petroleum imports. Distillate exports began to increase
starting in 2006 and they continued to increase through 2017. As a result, net distillate imports—
which were somewhat positive at 0.2 million bpd initially—trended downward starting in 2006
to negative 1.1 million bpd in 2017. Gasoline net imports reached a maximum of over 1 million
bpd in 2007. Like distillate, gasoline exports also began to increase, which likewise
corresponded with a reduction in net gasoline imports. By 2017, gasoline net imports were 1.1
million bpd lower than in 2007.
Renewable fuels likely contributed to reducing net petroleum imports by a relatively
modest amount. The volume of corn ethanol consumption increased from about 2 billion gallons
(0.13 million bpd) in 2000 to over 14 billion gallons (0.91 million bpd) in 20 1 5.31 32 Biodiesel
consumption increased from 10 million gallons in 2001 to over 1 billion gallons (0.07 million
bpd) in 2013, and biodiesel and renewable diesel consumption totaled over 2 billion gallons
31 EIA, Monthly Energy Review, Table 10.3, Fuel Ethanol Review;
https://www.eia. gov/totalenergy/data/monthlv/pdf/sec 10 7 .pdf.
32 Note that "corn ethanol" also includes small amounts of ethanol produced from other sources of starch such as
wheat and grain sorghum.
18
-------
(0.13 million bpd) in 2019.33'34 Assuming that this total renewable fuel volume displaced an
energy-equivalent volume of petroleum imports, corn ethanol and biodiesel/renewable diesel
combined would have displaced about 0.75 million bpd of petroleum equivalent volume in
2019—which is equivalent to 6% of the highest petroleum import volume of 12.5 million bpd.
Petroleum imports also contributed to a monetary trade imbalance that was of particular
concern prior to the passage of EISA in 2007. What made the continued increase of net
petroleum imports until 2005 of particular concern was that crude oil prices were increasing at
the same time.35 Crude oil spot prices (both WTI and Brent) had doubled in 2005 to over $50/bbl
compared to average crude oil spot prices prior to 2004. Crude oil prices continued to increase,
nearly doubling again in 2008 compared to 2005. Thus, the U.S. imported petroleum trade
imbalance quadrupled in monetary terms. However, petroleum imports have decreased in recent
years.
The total U.S. trade imbalance increased to just under $800 billion in 2005 and increased
further to over $800 billion per year in 2006 through 2008.36 The increasing crude oil prices on
top of the increasing petroleum imports contributed to this increasing trade imbalance. The 12
million bpd net petroleum import volume combined with the approximately $70/bbl crude oil
price in 2006 contributed to about $300 billion of the total U.S. trade imbalance. While
petroleum imports directly comprised a large portion of the increasing trade imbalance, higher
crude oil prices also increased the prices of many other goods that were imported into the U.S.,
which likely indirectly contributed to the trade imbalance.37 In 2009, the U.S. trade imbalance
dropped to $500 billion. Since then, the U.S. trade imbalance increased back into the $600-700
billion per year range until 2018 and 2019, when it increased again back above $800 billion per
year. Then, in 2020, the U.S. trade imbalance further increased above $900 billion. As shown in
Figure 1.4.2-1, the decreasing net imports of petroleum means that petroleum is not a factor for
this increasing trade imbalance.
We recognize that because the U.S. is a participant in the world market for petroleum
products, its economy cannot be shielded from world-wide price shocks.38 However, the
potential for petroleum supply disruptions due to supply shocks has been significantly
diminished due to the increase in tight oil production and, to a lesser extent, renewable fuels,
which has shifted the U.S. to being a modest net petroleum importer in the world petroleum
market in 2023-2025. Nevertheless, the potential for supply disruptions (discussed further in
Chapter 5) has not been eliminated.39
33 Biodiesel consumption data from EIA, Monthly Energy Review, Table 10.4, Biodiesel and Other Renewable
Fuels Overview; https://www.eia.gov/totalenergy/data/monthlv/pdf/sec.1.0 8.pdf.
34 Renewable consumption data from Public Data for the Renewable Fuel Standard; EPA Moderated Transaction
System; https://www.epa.gov/fiiels-registration-reporting-and-compliance-help/pnblic-data-renewable-fiiel-standard.
35 Spot Prices - Petroleum and Other Liquids; EIA; https://www.eia.gov/dnav/pet/pet pri spt s i a.htm.
36 U.S. Trade in Goods with World, Seasonally Adjusted; United States Census Bureau;
https://www.census.gov/foreign-trade/balance/cQ004.htmi.
37 U.S. Trade Deficit and the Impact of Changing Oil Prices; Congressional Research Service; February 24, 2020;
https://fas.org/sgp/crs/niisc/RS22204.pdf.
38 Bordoff, Jason; The Myth of US Energy Independence has Gone Up in Smoke; Foreign Policy; September 18,
2019; https://foreignpolicv.com/2019/09/18/the-mvth-of-n-s-energv-independence-has-gone-np-in-smoke.
39 Foreman, Dean; Why the US must Import and Export Oil; American Petroleum Institute; June 14, 2018;
https://www.api.org/news-policv-and-issiies/blog/2018/06/14/wltv-the-us-must-import-and-export-oil.
19
-------
1.4.3 Refinery Margins
Refinery margins reveal the economic health of refineries. The higher the margins for a
refinery, the greater its profitability and economic viability. Over time, refinery margins vary
considerably, but must average at least a certain level in order to remain viable long term.
Publicly available refinery margin data from BP is shown in Figure 1.4.3-1 for three
different types of refineries: (1) A U.S. Gulf Coast coking refinery; (2) A Northwest European
sweet cmde oil cracking refinery; and (3) A medium crude oil hydrocracking refinery in
Singapore.40 The refinery margin data is for three refineries owned by BP; thus, it may not
represent the margins of other refineries in the same regions. The margin data is on a semi-
variable basis, accounting for all variable costs and fixed energy costs.
Figure 1.4.3-1: Refinery Margins in Three Different Regions (S/bbl)
4 i
I SINGAPORE MEDIUM SOUR HYDROCRACKING
-10
1995 2000 2005 2010 2015 2020
Figure 1.4.3-1 shows that from 1993-2004, refinery margins were modest, and the
Singapore refinery in particular experienced zero or near-zero margins over much of 1998-2004.
From 2004-2009, crude oil prices were rising and it was a much better period for these
refineries' margins, particularly for the Gulf Coast refinery. The Gulf Coast refinery's margins
were likely much higher due to the heavy sour crude oil processed there being much less
expensive than the crude oils processed at the other two refineries. The blending of corn ethanol
into E10 gasoline played an important role in improving refinery margins in the 2005 to 2007
timeframe.41 All three refineries' margins decreased dramatically after 2008, likely due to the
large decrease in refined product demand associated with the Great Recession. As the world
emerged from the Great Recession, the three refineries' margins started improving in 2010, and
in particular, the refinery margins improved more dramatically for the heavy sour coking refinery
40 Oil Refinery Margins - Regional; NASDAQ Data Link; ;tps://data.nasdaa.com/data/BP/OIL REF MARG-oil-
refining-margins-regional.
41 Economics of Blending 10 Percent Corn Ethanol into Gasoline; Office of Transportation and Air Quality,
Environmental Protection Agency; (EPA-420-R-22-034. December 2022.
20
-------
in the Gulf Coast. However, refinery margins for U.S. refineries that refine light, sweet crude oil
are not represented in Figure 1.4.3-1. As shown in Figure 1.4.1-1, but not reflected in Figure
1.4.3-1, light sweet crude oil prices were depressed in the U.S. during 2011-2014. Lower prices
of sweet crude oil provided high margins for U.S. refineries that processed sweet crude oil
during this time period. The Gulf Coast refinery margins stayed elevated all the way up to 2020,
at which point refinery margins declined steeply for all three refinery types due to the COVID-19
pandemic. Refinery margins returned to their pre-pandemic levels in 2022, but then increased
significantly starting in March 2022 due to geopolitical factors.42
1.4.4 Transportation Fuel Demand
At the time the RFS2 program was being enacted through EISA in 2007, there had been a
consistent increase in U.S. petroleum demand. However, transportation fuel demand fell short of
historical demand increases starting in 2008 and has remained relatively stable since that time.
Figure 1.4.4-1 shows the actual volume of gasoline, distillate, and jet fuel consumed in the U.S.
from 2000-2021, as well as the projected demand of gasoline and distillate if transportation fuel
demand growth had continued at the historic rate based on AEO 2008.43 44
Figure 1.4.4-1: Actual and Projected Transportation Fuel Demand
250.0
_ 200.0
'i/T
c
_o
"ro
g 150.0
CQ
"O
OJ
'~E- 100.0
Q.
D
LO
"H3
U_
50.0
0.0
2000 2005 2010 2015 2020 2025
Year
Gasoline Supplied
Projected Gasoline Demand
Distillate Demand
~
Projected Distillate Demand
Jet Fuel Demand
y
Figure 1.4.4-1 shows that both gasoline and distillate demand increased up to 2007.
During previous years, gasoline and distillate demand was increasing 1.3% and 1.7% per year on
average, respectively. The dashed lines in Figure 1.4.4-1 show projected gasoline and distillate
42 McGurty, Janet; Refinery Margin Tracker: Russian crude cargoes taper off as margins rise; S&P Global; April 4,
2022.
43 Product Supplied; Petroleum and Other Liquids, Energy Information Administration,
https://www.eia.gov/dnav/pet/pet cons psup dc nus mbbl a.htm.
44 Annual Energy Outlook 2008; Energy Information Administration; June 2008;
https://www.eia.gov/outlooks/arcliive/aeo08/index.html.
21
-------
demand if they had continued to increase at the same rate as that prior to 2008. The figure clearly
shows that actual gasoline and distillate demand fell far short of projected demand after 2007.
Conversely, jet fuel demand was essentially flat during the entire period.
Several factors led to the decrease of transportation fuel demand after 2007 relative to
projected values:
• The Great Recession. The Great Recession began in 2008 and officially lasted for 18
months, although employment did not return to pre-recession levels until over 6 years
after the onset of the recession. The Great Recession caused a large impact on economic
activity, which reduced transportation fuel demand during these years.
• Increased crude oil prices. Sustained, higher crude oil prices resulted in increased
transportation fuel prices over this time period, which affected consumer behavior by
impacting the number of miles traveled and vehicle purchase decisions. After 2014, crude
oil prices decreased to the $40-50 price range, which brought gasoline prices back down
and likely reversed some of the consumer behavior changes.
• Increasing fuel economy of the motor vehicle fleet. EPA and the National Highway
Transportation Administration (NHTSA) finalized standards which reduced light-duty
motor vehicle greenhouse gas (GHG) emissions and increased the Corporate Average
Fuel Economy (CAFE) of motor vehicles. The GHG/CAFE standards applied to light-
duty vehicles sold in 2012-2025 and thereafter.45 EPA and NHTSA also established
GHG/CAFE standards for new heavy-duty vehicles and their trailers.46 The phase 1 and
phase 2 heavy-duty GHG standards began to phase-in in 2014 and will continue to do so
through 2027.47 In addition, EPA proposed additional light-duty, medium-duty and
heavy-duty standards which will affect the greenhouse gas emissions, and thus the fuel
economy of, future motor vehicles.48 49 If finalized, these proposed rules would further
reduce the petroleum consumption of the motor vehicle fleet. The GHG standards only
affect new internal combustion vehicles; thus, as consumers purchase new motor
vehicles, these new vehicles consume less gasoline and diesel compared to the vehicles
sold in previous years, reducing overall petroleum demand.
• Electric vehicle penetration andfuel displacement. Electric vehicles (EVs) and plug-in
hybrid electric vehicles (PHEVs) reduce consumption of petroleum fuel by either
partially displacing petroleum fuels (in the case of PHEVs) or completely displacing
petroleum demand (in the case of EVs). Data on annual electrified vehicle sales indicates
45 75 FR 25324 (May 7, 2010) and 86 FR 74434 (December 30, 2021).
46 76 FR 57106 (September 15, 2011).
47 81 FR 73478 (October 25, 2016).
48 Proposed Rulemaking: Multi-Pollutant Emissions Standards for Model Years 2027 and Later Light-Duty and
Medium-Duty Vehicles, 88 FR 29184 (May 5, 2023).
49 Proposed Rulemaking: Greenhouse Gas Emissions Standards for Heavy-Duty Vehicles - Phase 3, 88 FR 25926
(April 27, 2023).
22
-------
that EVs and PHEVs displaced an estimated 5 million gallons of fuel in 2011 and that
this increased to over 400 million gallons in 2019.50
1.5 Cellulosic Biofuel
Actual production of cellulosic biofuel through 2022 has been significantly less than the
statutory volumes, which reached 16 billion gallons in 2022. Minimal volumes of cellulosic
biofuel were produced through 2013. Since 2013, volumes of the types of liquid cellulosic
biofuels projected in the RFS2 rule have remained limited. There are numerous reasons that
liquid cellulosic biofuel production has not developed as anticipated. In some years, the lower
than anticipated crude oil prices discussed in Chapter 1.4.1 certainly impacted the market's
ability to produce liquid cellulosic biofuels at competitive prices. The relatively low production
costs estimated in the RFS2 rule (generally $1.00-2.50 per gallon of liquid cellulosic biofuel
based on NREL modeling, depending on the production technology and technology year) have
not been realized.51 While the issues associated with each individual company and facility are
unique, and the reasons facilities fail to consistently produce cellulosic biofuel at the expected
volumes are not always publicly disclosed, there appear to be several common challenges across
the liquid cellulosic biofuel industry. These challenges include: (1) Feedstock quality and
handling issues; (2) Higher than anticipated feedstock and capital costs; and (3) Difficulty
scaling up technology to commercial scale. The inability of several first-of-a-kind cellulosic
biofuel production facilities to continue operating has also likely impacted investment in the
commercialization of similar technologies. As we discuss further in Chapter 6.1, the availability
of liquid cellulosic biofuel has historically been very low and has typically fallen short of EPA's
projections.
Although production of liquid cellulosic biofuel from commercial scale production
facilities has been far lower than projected in the RFS2 rule, smaller volumes of qualifying
cellulosic biofuel have been produced using technologies not discussed in that rule. The
production of compressed natural gas and liquified natural gas (CNG/LNG) derived from biogas,
which was not one of the cellulosic biofuel production technologies discussed in the RFS2 rule,
has accounted for the vast majority of the cellulosic biofuel produced since 2010. The RFS2 rule
contained a pathway52 for the production of biogas from landfills, sewage and waste treatment
plants, and manure digesters to generate advanced biofuel (D5) RINs.53 In response to questions
from multiple companies, EPA subsequently evaluated whether biogas from several different
sources could be considered not just an advanced biofuel, but also a cellulosic biofuel. In the
Pathways II rule, EPA added a pathway for CNG/LNG derived from biogas from landfills,
municipal wastewater treatment facility digesters, agricultural digesters, and separated MSW
digesters, as well as biogas from the cellulosic components of biomass processed in other waste
digesters, to generate cellulosic biofuel (D3) RINs when used as a transportation fuel.54
50 Transportation Research Center at Argonne National Laboratory, https://www.anl.gov/es/light~diitv-electric~drive~
vehicles-monthly-sales-updates.
51 Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis. EPA-420-R-10-006. February 2010.
52 A pathway is a combination of feedstock, production process, and fuel type. EPA has evaluated a number of
different pathways to determine the category of renewable fuel that fuel produced using the various pathway
qualifies for. The list of generally applicable pathways can be found in 40 CFR 80.1426(f).
53 75 FR 14872 (March 26, 2010).
54 79 FR 42128 (July 18, 2014).
23
-------
Following this decision, production of CNG/LNG derived from biogas increased rapidly, from
approximately 33 million RINs in 2014 to over 660 million RINs in 2022.55 Through 2022, over
98% of all of the cellulosic RINs generated in the RFS program have been for CNG/LNG
derived from biogas. We anticipate that CNG/LNG derived from biogas will continue to be the
source of the vast majority of cellulosic biofuel in the RFS program through 2025. Actual
cellulosic biofuel production for each year from 2014-2022 is shown in Figure 1.5-1.
Figure 1.5-1: Cellulosic RINs Generated (2013-2022)
700
600
500
1 400
c
Q
= 300
illl
2013 2014 2015 2016 2017 201S 2019 2020 2021 2022
¦ CNG/LNG Derived from Biogas ¦ Liquid Cellulosic Biofuels
1.6 Biodiesel and Renewable Diesel
The actual supply of biodiesel and renewable diesel has significantly exceeded the supply
projected by EPA in the RFS2 rule. In that rule, EPA projected that 1.62 billion gallons of
biodiesel and 0.15 billion gallons of renewable diesel would be supplied in 2021, all of which
was projected to be produced in the U.S.56 The actual supply of biodiesel and renewable diesel in
2022 was 1.74 billion gallons and 1.36 billion gallons, respectively. While the majority of these
volumes were produced domestically, significant volumes were imported. Further, while the vast
majority of biodiesel and renewable diesel supplied since 2010 has met the requirements for
BBD or advanced biofuel, smaller volumes were produced from grandfathered facilities using
renewable biomass that does not qualify for BBD or advanced RINs and therefore only qualify to
generate conventional renewable fuel (D6) RINs. The most likely feedstock used to produce
grandfathered biodiesel and renewable diesel is palm oil; however, other types of renewable
biomass that have not been approved to generate advanced or BBD RINs could also have been
used.
55 One RIN can be generated for each ethanol-equivalent gallon of renewable fuel. One gallon of ethanol is eligible
to generate one RIN; other types of fuel generate RINs based on their energy content per gallon relative to ethanol.
For CNG/LNG derived from biogas, every 77,000 BTU of qualifying biogas generates one RIN.
56 2022 is the most recent year for which data are available for comparison.
24
-------
Figure 1.6-1: 2010 Projected vs. Actual Biodiesel and Renewable Diesel Supply (2014-2022)
3.50
3.00
2-50 —
on
£=
O
2014 2015 2016 2017 2018 2019 2020 2021 2022
¦ Projected Renewable Diesel ¦ Actual Renewable Diesel
¦ Projected Biodiesel ¦ Actual Biodiesel
Projected volumes are from the RFS2 rule. Actual volumes are from EMTS data.
Figure 1.6-2: Source of Biodiesel and Renewable Diesel Consumed in the U.S. (2014-2022)
3.00
2.50
trt
C
2.00
Q
(3
1.50
c
_o
£
1.00
0.50
0.00
I
I
I
I II
2014 2015 201S 2017 2018 2019 2020 2021 2022
I
I
I
I Biodiesel Net Imports
I Domestic Biodiesel
i Renewable Diesel Net Imports
Domestic Renewable Diesel
The reason that the supply of biodiesel and renewable diesel has been much higher than
projected in the RFS2 rule is primarily related to challenges associated with consuming ethanol
as higher-level blends with gasoline (i.e., greater that 10% ethanol), which we discuss further in
Chapter 1.7. The limited use of higher-level ethanol blends, together with lower than projected
gasoline demand, resulted in total ethanol consumption in 2020 - 2022 (12.68, 13.94, and 13.98
billion gallons, respectively) that was lower than the projected ethanol consumption volume in
2022 even under the low ethanol case from the RFS2 rule (17.04 billion gallons).5' Since the
"s1 Ethanol consumption volume are from EIA's Monthly Energy Review, while the ethanol projections are from the
RFS2 rale. Ethanol consumption in 2020 was significantly impacted by the COVID-19 pandemic. Ethanol
25
-------
primary fuels available to meet the advanced biofuel requirements are biodiesel, renewable
diesel, and sugarcane ethanol, the challenges associated with increasing ethanol consumption is a
significant factor in a much smaller-than-projected supply of sugarcane ethanol. Instead, greater
volumes of biodiesel and renewable diesel have been used to meet the advanced biofuel
requirement and at times even the total renewable fuel requirement, as further discussed in
Chapter 6.
The feedstocks used to produce biodiesel and renewable diesel each year from 2014-
2021 for domestically produced and imported biodiesel and renewable diesel are shown in
Figures 1.6-2 and 1.6-3.
Figure 1.6-2: Domestic Biodiesel and Renewable Diesel Feedstocks (2014-2022)
3.00
2014 2015 2016 2017 201S 2019 2020 2021 2022
¦ FOG ¦Corn Oil ¦Soybean Oil BCanolaOil ¦ Grandfathered (Unknown)
Source: EMTS
consumption in the U.S. reached a peak of 14.49 billion gallons in 2017, still far short of the volumes projected in
the RFS2 rule.
26
-------
Figure 1.6-3: Imported Biodiesel and Renewable Diesel Feedstocks (2014-2022)
1.00
0.90
0.80
0.70
d
O
0.60
"ro
u
0.50
c
G
0.40
IE
0.30
0.20
0.10
0.00
I
I
2014 2015 2016 2017 2018 2019 2020 2021 2022
I FOG ¦ Corn Oil ¦ Soybean Oil BCanolaOil BAIgalOil ¦ Grandfathered (Unknown)
Source: EMTS
There are several notable differences between the quantities of feedstock projected to be
used to produce biodiesel and renewable diesel in the RFS2 rule and the actual feedstocks used
to produce these fuels in 2022. Domestic biodiesel production in 2022 was fairly similar to the
volume of biodiesel projected in the RFS2 rule for that year (all of which was projected to be
produced domestically), but there were significant differences in the feedstocks used to produce
this biodiesel. Relative to the quantities projected in the RFS2 rule, the use of soybean oil, fats,
oils, and greases (FOG), and other sources were all higher than projected, while the use of corn
oil from ethanol plants was lower than projected. These differences largely reflect the greater
than anticipated demand for biodiesel as a result of the limitations on ethanol consumption (see
Chapter 1.7). The lower than expected use of corn oil is likely the result of production of non-
food grade corn oil being a relatively new feedstock at the time of the RFS2 rule, EPA's ethanol
projections being over-ambitious (since corn oil is a co-product of ethanol production), and
demand for this feedstock in animal feed and other sectors.
Domestic renewable diesel production in 2022 was significantly higher than projected in
the RFS2 rule, in which EPA projected that all renewable diesel would be produced domestically
from FOG. While the majority of domestic renewable diesel was produced from FOG in 2022,
significant volumes were also produced from soybean oil and corn oil from ethanol plants. The
U.S. also imported significant volumes of biodiesel and renewable diesel in 2022, as well as in
previous years. By 2022, the majority of the imported biodiesel and renewable diesel was
produced from FOG; however, in earlier years the U.S. also imported large volumes of biodiesel
produced from soybean oil.58 These data, particularly the significant decrease in imported
biodiesel and renewable diesel from 2016-2018 after the U.S. announced tariffs on imported
biodiesel from Argentina and Indonesia, also highlight the importance of U.S. trade policy on the
supply of these fuels.
58 Source: EMTS.
27
-------
1.7 Ethanol
The predominant form of biofuel used to meet the standards under the RFS program—
and in particular the total renewable fuel standard—has been ethanol. In 2005, just prior to
implementation of the RFS1 program, ethanol accounted for 97% of all biofuel consumed in the
U.S. transportation sector.59 Since then, the total volume of ethanol used in the U.S. has more
than tripled from 5.5 million gallons in 2006 to 14.5 million gallons in 2019.60 61 By 2010,
ethanol use in the U.S. was approaching the "E10 blendwall" (as represented by the nationwide
average ethanol concentration) and actually exceeded 10.00% in 2016. By 2022, ethanol
accounted for 78% of the 18.1 billion gallons of biofuel consumed in the U.S.62
In all years since ethanol was approved for use in gasoline in 1979, the vast majority of
ethanol consumed in the U.S. has been produced domestically from corn starch with small
amounts from other starches. Cellulosic ethanol has represented at most 0.07% (2019) of all
ethanol consumed in the U.S., while the proportion of imported sugarcane ethanol has been small
but highly variable.
Figure 1.7-1: Proportions of Ethanol Fuel Sources
i£l^-OOOlO«H(NrO^J"U")lDf^COOOHrM tDI^-OOCTlOt-H4
-------
Figure 1.7-2: Domestic Production and Consumption of Ethanol
18,000
16,000
14,000
12,000
to
c
= 10,000
BD
c
¦2 8,000
i
6,000
4,000
2,000
0
—
**
m
¦ Domestic production
¦ Domestic consumptior
1
E10 blendwall
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
Source: Domestic consumption from EIA's Monthly Energy Review. The E10 blendwall was derived from total
gasoline energy demand. Domestic production = domestic consumption - imports + exports.
The E10 blendwall appears to have been a decisive factor in limiting growth in domestic
consumption of ethanol. As illustrated in Figure 1.7-3, the nationwide average ethanol
concentration did not increase at the same pace after 2010 as it did in previous years, but instead
slowed significantly, approaching and then slightly exceeding 10.00% at a comparative crawl
after 2010.63 Ethanol production exceeded domestic consumption through exports.
63 As discussed further in Chapter 1.10, a comparison of historical gasoline + diesel volumes reported by obligated
parties with gasoline + diesel estimated by EIA revealed a significant difference. Insofar as EIA estimates of
gasoline consumption are lower than actual consumption, the nationwide average ethanol concentration calculated
from those gasoline consumption estimates are then higher than actuals. However, even if this is the case, we would
expect the general trend shown in Figure 1.7-3 to be unchanged.
29
-------
Figure 1.7-3: Nationwide Average Concentration of Ethanol in Gasoline Consumed in the
U.S.
12%
Q)
C
s 10%
0%
27%
, 1
.535
—3
'4)
.54%
3.86%
1.
32%
2
.069
h
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
Source: Derived from EIA's Monthly Energy Review - ethanol consumption divided by motor gasoline
consumption
After E10 was approved for use in all vehicles in 1979, consumers had a choice between
EO (gasoline without ethanol) and E10. Consumers likely made their choice based on knowledge
of what fuels were available based on pump labeling, relative price, perceptions (or lack thereof)
of impacts on vehicle fuel economy, vehicle operability or longevity, comfort with an unfamiliar
fuel, and perceived benefits to the environment or economy. Since approaching and exceeding
the El0 blendwall between 2010 and 2016, virtually all gasoline nationwide contains 10%
ethanol. As a result, most consumers today have little choice but to use E10. However, with the
expansion of retail service stations offering El 5 and E85, the choice for consumers has now
shifted to between E10 and these higher-level ethanol blends. For higher-level ethanol blends,
consumers likely consider all of the factors they considered when the choice was between EO and
E10, plus whether the fuel is legally permitted to be used in their vehicle and whether the
manufacturer has warranted their vehicle for its use.
1.7.1 E85
The earliest form of a higher-level ethanol blend was E85. In 1996, the first FFV was
produced that could operate on fuel containing up to 85% denatured ethanol (83% ethanol).64
Starting in 2007, ASTM International limited the maximum ethanol content of E85 to 83% in
specification D5798, with a minimum ethanol content of 51%. EIA assumes that the annual,
nationwide average ethanol concentration of E85 is 74%.65
E85 is not considered gasoline under EPA's regulations, and as such is permitted to be
used only in FFVs. However, FFVs can operate on either gasoline or E85. Under basic economic
64 "Alternative Fuel Ford Taurus," available in the docket for this action.
65 "AEO2022 Table 2," available in the docket fortius action. See footnote 11.
30
-------
theory, and assuming all other factors are equal, FFV owners are more likely to purchase E85 if
they believe that doing so reduces their fuel costs. E85 reduces fuel economy in comparison to
E10, and E85 must sell at a discount to E10 if it is to represent an equivalent value in terms of
energy content. For an average E85 containing 74% ethanol, its volumetric energy content is
approximately 21% less than E10 (or 24% lower than that of EO, though EO is rarely the point of
comparison as sales volumes of EO are considerably lower than sales volumes of E10).66-67 In
order for E85 to be priced equivalently to gasoline on an energy-equivalent basis, then, its price
must be on average 21% lower than that of E10. As shown in Figure 1.7.1-1, the nationwide
average price of E85 compared to E10 has only rarely achieved the requisite energy equivalent
pricing needed for FFV owners who are aware of and concerned about the fuel economy impacts
of E85. Furthermore, E85 purchasers generally have no way of knowing whether their fuel
contains 83% ethanol, 51% ethanol, or something in-between.
a The 21% energy equivalence level of E85 compared to E10 assumes that E85 contains 74% ethanol.
EPA has estimated the nationwide volume of E85 consumed in past years using two
different methods.68 The results of those analyses are shown in Figure 1.7.1-2.
66 Assumes ethanol energy content is 3.555 mill Btu per barrel and gasoline energy content is 5.222 mill Btu per
barrel. EIA Monthly Energy Review for April 2021, Tables A1 and A3.
67 A comparison to EO would be more relevant prior to 2010 when there remained significant volumes of EO for sale
at retail stations.
68 "Estimate of E85 consumption in 2020," available in the docket for this action.
31
-------
a The ethanol concentration of E85 is assumed to be 74% on average.
As discussed in Chapter 6.5, we do not need estimates of E85 (or El5) for the purposes
of estimating total ethanol consumption for the years covered by this rule. However, for cost
purposes only, we have estimated El5 and E85 volumes using a methodology and dataset
different from that used to estimate the values in Figure 1.7.1-2. The resulting estimates of E15
and E85 volumes for 2023, 2024, and 2025 that are used in the cost analysis are provided in
Chapter 6.5.2.
1.7.2 E15
In 2011, gasoline containing up to 15% ethanol was permitted to be used in model year
(MY) 2001 and newer vehicles.69 El 5 has since been offered at an increasing number of retail
service stations.70 However, there is currently no publicly available data on actual nationwide
El5 sales volumes.
Sales of El 5 prior to 2019 were seasonal due to the fact that El 5 did not qualify for the
1-psi RVP waiver for summer gasoline in CG areas that has been permitted for El0 since the
summer volatility standards were implemented in 1989.71 As shown in Figure 1.7.2-1, monthly
El 5 sales in Minnesota from 2015-2018 demonstrate that sales volumes of El 5 in summer
months were notably lower than in non-summer months in this time period.72
69 76 FR 4662 (January 26, 2011).
711 See Chapter 6.4.3.
71 54 FR 11883(March 22, 1989).
72 Minnesota is the only source of data on E15 sales by month of which we are aware.
32
-------
Figure 1.7.2-1: Normalized Monthly E15 Sales per Station in Minnesota3
2.00
0.20
0.00
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Source: Minnesota Commerce Department
a Normalized values derived by dividing the monthly E15 sales volume per station by the annual average E15 sales
volume per station.
In 2019, EPA extended the 1-psi waiver to El5 by regulation.73 EPA estimated that the
annual average El5 sales per station in Minnesota would have been 16% higher had the 1-psi
waiver been in place from 2015-2018.74 On July 2, 2021, the U.S. Court of Appeals for the D.C.
Circuit ruled that EPA's extension of the 1-psi waiver to El 5 was based on an impermissible reading
of the statute and vacated it. EPA subsequently issued emergency fuel waivers for the summer of
2022, and the summer for 2023 that allowed El 5 to take advantage of the 1-psi waiver to address
issues related to fuel price and supply.75 Insofar as the 1-psi waiver for El5 had an impact on
summer sales of E15, therefore, it did so only for 2019-2022. For these years, data from
Minnesota on per-station sales of El 5 indicates that those sales were no longer seasonal as they
were prior to 2019.
73 84 FR 26980 (June 10, 2019).
74 "Estimating the impacts of the lpsi waiver for E15," memorandum from David Korotney to EPA docket EPA-
HQ-0 AR-2019-0136.
75 On April 28, 2023 EPA also issued an emergency fuel waiver for E15 RVP that took effect May 1, 2023
33
-------
Figure 1.7.2-1: Normalized Monthly E15 Sales per Station in Minnesota; Before and After
1-psi Waiver3
0.40
q 2q ^^Average pre-waiver
^^Average post-waiver
0.00
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Source: Minnesota Commerce Department
a Normalized values derived by dividing the monthly E15 sales volume per station by the annual average E15 sales
volume per station.
On March 6, 2023, EPA proposed to remove the 1-psi waiver for E10 in eight states.76 77
If this proposal is finalized, the net result would be that E10 and E15 are treated the same in
these states with regard to RVP beginning with the summer of 2024, and there may be no
reduction in summer sales of El 5 compared to other months in these states.
1.8 Other Biofuels
Although domestic corn ethanol and BBD have dominated the biofuels landscape since
implementation of the RFS program began in 2006, other biofuels have also contributed to the
total renewable fuel pool, sometimes providing the marginal volumes needed to meet the other
applicable standards. As shown in Figures 1.8-1 and 2, biofuels other than corn ethanol and BBD
represented between 2-5% of total renewable fuel from 2012-2022.78
76 "Request from States for Removal of Gasoline Volatility Waiver" 88 FR 13758 (March 6, 2023).
77 Illinois, Iowa, Minnesota, Missouri, Nebraska, Ohio, South Dakota, and Wisconsin.
78 Detailed data prior to 2012 on RIN generation, adjustments to account for invalid RINs, and exports is less robust
and is therefore not presented here.
34
-------
Figure 1.8-1: Contribution of Biofuels to Total Renewable Fuel Consumption"-1'
20,000
¦ Imported conventional ethanol
19,000
¦ Imported conventional renewable diesel
¦ Imported conventional biodiesel
18,000
¦ Domestic conventional biodiesel
¦ Domestic advanced renewable diesel
¦ Domestic advanced gasoline/naphtha
to
17,000
¦ Domestic advanced heating oil
2
cd
c
o
¦ Imported advanced ethanol
i
16,000
15,000
14,000
¦ Domestic advanced ethanol
¦ Domestic advanced biogas
¦ Domestic advanced jet fuel
¦ Imported cellulosic ethanol
¦ Domestic cellulosic ethanol
¦ Imported cellulosic biogas
¦ Domestic cellulosic biogas
13,000
¦ Domestic corn ethanol + BBD
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Source: EMTS
a Ignores any biofuels that contributed less than 1 million RINs in aggregate over all years shown. This affects
domestic cellulosic gasoline/naphtha, domestic cellulosic diesel, and domestic conventional butanol.
b Fuel type and D code of exports is known, but whether the exported fuel was originally produced domestically or
was imported is not known. For purposes of this chart, exports were assumed to be distributed to domestic
production and imports in proportion to the relative production volumes of each.
Figure 1.8-2: Biofuels Other than Corn Ethanol and BBD
1,000
Imported conventional ethanol
¦ Imported conventional renewable diesel
¦ Imported conventional biodiesel
¦ Domestic conventional biodiesel
¦ Domestic advanced renewable diesel
¦ Domestic advanced gasoline/naphtha
¦ Domestic advanced heating oil
¦ Imported advanced ethanol
¦ Domestic advanced ethanol
¦ Domestic advanced biogas
¦ Domestic advanced jet fuel
¦ Imported cellulosic ethanol
¦ Domestic cellulosic ethanol
¦ Imported cellulosic biogas
¦ Domestic cellulosic biogas
0
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
35
-------
As illustrated in Figure 1.8-3, advanced biofuel exclusive of cellulosic biofuel or BBD
(i.e., renewable fuel having a D code of 5) has been met with the greatest variety of fuel types
compared to the other statutory categories.
Figure 1.8-3: Advanced Biofuel Types Excluding Cellulosic Biofuel and BBD
700
600
500
¦ Domestic advanced renewable diesel
^ 400
¦ Domestic advanced gasoline/naphtha
ct:
c
o
¦ Domestic advanced heating oil
^ 300
¦ Imported advanced ethanol
¦ Domestic advanced ethanol
¦ Domestic advanced biogas
200
100
0
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Source: EMTS
These sources of advanced biofuel varied widely in both their overall contributions to the
advanced biofuel pool from 2012-2022, as well as in each individual year. As the largest overall
contributor, imported advanced ethanol produced from sugarcane in Brazil is discussed
separately in Chapter 6.3. Production of domestic advanced renewable diesel,79
gasoline/naphtha, and ethanol were of approximately similar magnitude and demonstrated no
consistent increasing or decreasing trends between 2012-2022. Domestic advanced biogas fell to
near zero in 2015 after biogas from landfills was recategorized as cellulosic biofuel in 2014.80
Domestic advanced heating oil has grown steadily since 2012 but has never generated more than
3 million RINs in a single year.
As described in Chapter 1.5, cellulosic biofuel has been composed predominately of
biogas-based CNG/LNG. Smaller volumes of cellulosic ethanol and heating oil and very small
volumes of gasoline/naphtha and renewable diesel have also been used.
79 Small quantities of renewable diesel are not BBD but are nonetheless advanced biofuel.
80 79 FR 42128 (July 18, 2014).
36
-------
1.9 RIN System and Prices
1.9.1 RIN System
RINs were created by EPA under CAA section 21 l(o)(5) as a flexible credit and
compliance mechanism to enable obligated parties across the country to meet their renewable
fuel blending obligations under the RFS program without having to blend the renewable fuel
themselves.81 RINs allow: (1) Obligated parties (i.e., the refining industry) to comply with the
RFS program without producing, purchasing, or blending the renewable fuel themselves; (2)
Non-obligated blenders of renewable fuel to maintain their preexisting blending operations; and
(3) The ethanol and other biofuel industries to continue to produce biofuels, now with the
support of the RIN value. Obligated parties, of course, can and do produce, purchase, and blend
their own renewable fuel, but the RIN system allows them the option of not doing so and instead
relying on the business practices of other market participants that are already set up to do so.
RINs are generated by renewable fuel producers (or in some cases renewable fuel importers) and
are assigned to the renewable fuel they produce. These RINs are generally sold together with the
renewable fuel to refiners or blenders. RINs can be separated from renewable fuel by obligated
parties or when renewable fuel is blended into transportation fuel. Once separated, RINs can be
used by obligated parties to demonstrate compliance with their RFS obligations or can be traded
to other parties.
Under the RFS program, EPA created five different types of RINs: cellulosic biofuel
(D3) RINs, BBD (D4) RINs, advanced biofuel (D5) RINs, conventional renewable fuel (D6)
RINs, and cellulosic diesel RINs (D7).82 The type of RIN that can be generated for each
renewable fuel depends on a variety of factors, including the feedstock used to produce the fuel,
the type of fuel produced, and the lifecycle GHG reductions relative to petroleum fuel. As shown
in Figure 1.9-1, the obligations under the RFS regulations are nested, such that some RIN types
can be used to satisfy obligations in multiple categories.
81 The RIN system was created in the RFS1 rule (72 FR 23900, May 1, 2007) and modified in the RFS2 rule (75 FR
14670, March 26, 2010).
82 40 CFR 80.1425(g).
37
-------
Figure 1.9-1: Nested Structure of the RFS Program
Total renewable fuel
Advanced biofuel
r
03/7
Non-cellulosic advanced
D4 D5 06
t t t t
CelJutosic BBD "Other" advanced Conventional
biofuel (sugarcane ethanol, (mostlv corn-ethanol)
y
1.9.2 RIN Prices
RIN prices have varied significantly since 2010. There have also been significant and
notable differences between the prices of each of the four major RIN types. A chart of RIN
prices, as reported to EPA through EMTS, is shown in Figure 1.9.2-1.83 While there are a wide
variety of factors that impact REN prices, including both market-based and regulatory factors, a
review of RTN prices reveals several notable aspects of the RFS program.
83 RIN prices are reported publicly on EPA's website (https://www.epa.gov/fiiels-registration-reporting-and-
compliance-help/rin-trades-and-price-information). These prices are reported to EPA by the parties that trade RINs
and are inclusive of all RIN trades (with the exception of RIN prices that appear to be outliers or data entiy errors).
Several other sendees also report daily RIN prices; however, these reports are generally not publicly available.
Further, the prices reported by these services generally represent only spot trades and do not include RINs traded
through long-term contracts.
38
-------
Figure 1.9.2-1: RIN Prices (June 2010 - December 2022)
$4.00
$3.50
$3.00
Source: EMTS Price Data
Prior to 2013, D6 RIN prices were low (less than $0.05 per RIN). These low prices were
likely due to the fact that from 2010-2012 it was cost-effective to blend ethanol into gasoline as
E10 even without the incentives provided by the RFS program. The low RIN prices during this
period also indicate that the RFS requirements were not the driving force behind increased use of
E10.
Beginning in 2013, D6 RIN prices rose sharply. 2013 marked the first time the implied
conventional renewable fuel requirement exceeded the volume of ethanol that could be
consumed as E10.84 While it has generally been cost-effective to blend ethanol as E10, higher-
level ethanol blends (e.g., E15 and E85) have generally not been cost effective, even with the
incentives provided by the RFS program. This is largely because: (1) Fuel blends that contain
greater than 10% ethanol are currently not optimized to take advantage of the high octane value
of ethanol; (2) The lower energy content of ethanol is more noticeable as the amount of ethanol
increases; and (3) Infrastructure limitations have restricted the availability of higher-level ethanol
blends (see Chapter 6.4).
In subsequent years, D6 RIN prices have varied significantly, but they have never
returned to the low prices observed prior to 2013. It is also notable that, from 2013-2016, D6
RIN prices remained close to, but slightly less than, D4 and D5 RIN prices. During this time,
obligated parties were purchasing D4 and D5 RINs in excess of their BBD and advanced biofuel
obligations to make up for the shortfall in conventional biofuel volume and used those RINs to
meet their total renewable fuel obligations. Essentially, given the inability to successfully
introduce higher-level ethanol blends into the market in sufficiently large quantities, the market
relied upon biodiesel and renewable diesel (primarily advanced biofuel and BBD, but also some
volume of conventional biodiesel and renewable diesel) as the marginal RFS compliance option
84 The conventional renewable fuel requirement is the difference between the total renewable fuel requirement and
the advanced biofuel requirement.
39
-------
when other sources of conventional biofuel were not available at competitive prices. After 2018,
D6 RIN prices were, for some time, significantly lower than D4 and D5 RIN prices, but still
higher than the D6 RIN prices observed prior to 2013. These lower D6 RIN prices are largely the
result of: (1) SREs granted in 2018, which reduced the total number of D6 RINs needed for
compliance with the RFS obligations to a number that was below the E10 blendwall; and (2) The
large number of carryover RINs available, as discussed in Chapter 1.9.1. More recently, D4, D5,
and D6 RIN prices have risen dramatically, reaching nearly $2 per RIN in the summer of 2021
before decreasing slightly to between $1.50-2.00 by the end of 2022. These prices reflect the
cost of biodiesel and renewable diesel production (the marginal supply) at a time of unusually
high commodity prices for soybean and other oil feedstocks, less the value of other subsidies and
credits (e.g., the $1.00 per gallon federal tax subsidy and state LCFS credits).
While D6 RIN prices have remained relatively high in recent years, these price levels
have not translated into higher ethanol prices for ethanol producers. After examining market
data, EPA found no correlation between D6 RIN prices and ethanol prices from 2010-2022.
Instead, higher D6 RIN prices have resulted in lower effective prices for ethanol after the RINs
have been separated and sold.8' Higher D6 RIN prices have thus served to subsidize fuel blends
that contain higher proportions of conventional biofuel (e.g., E85 and B20 biodiesel/renewable
diesel blends) and increased the cost of fuel blends that contain little or no conventional biofuel
(e.g., E0 and BO).86
Figure 1.9.2-2: Ethanol Prices and D6 RIN Prices (June 2010 - December 2022)
$4.00
$3.50
$3.00
"E
0
=B $2-50
DO
¦or
S $2.00
01
~o
£ $1.50
JZ
LU
$1.00
$0.50
i
I
1
v\
in i
u,
UJ
k
Vi
| 11
'n
/ '
z
- *
rr
$0.00
1/1/10 1/1/11 1/1/12 1/1/13 1/1/14 1/1/15 1/1/16 1/1/17 1/1/18 1/1/19 1/1/20 1/1/21 1/1/22
Ethanol Price RiN Price
Sources: Ethanol Price from USDA Weekly Ag Roundup, D6 RIN Price from EMTS data
D5 RINs were priced at a level between D4 and D6 RINs from 2010-2013. However,
since 2013, D5 RIN prices have been nearly identical to D4 RIN prices. This shift in the relative
pricing of D5 and D4 RINs also corresponds with the market reaching the E10 blendwall. This is
85 The effective price is the price of the ethanol after subtracting the RIN value from the price of the ethanol with the
attached RIN.
86 Burkholder, Dallas. "A preliminary Assessment of RIN Market Dynamics, RIN Prices, and Their Effects." U.S.
EPA Office of Transportation and Air Quality. May 2015.
40
-------
because there are two primary fuel types that have been used to satisfy the advanced biofuel
requirements: sugarcane ethanol and BBD. From 2010-2012, obligated parties generally met
their implied requirements for "other advanced biofuel" with sugarcane ethanol.87 This is
apparent in the volumes of sugarcane ethanol (which supplied the vast majority of volume
requirement for "other advanced" biofuels) and BBD (which did not exceed the volume
requirement for BBD by an appreciable volume) used in the U.S. in these years.88 It is also
indicated by the prices for D5 RINs, which were significantly lower than the price of D4 RINs
during this time, suggesting that it was more cost effective for obligated parties to meet their
compliance obligations with D5 RINs (generated for sugarcane ethanol) than D4 RINs
(generated for biodiesel and renewable diesel). When the E10 blendwall was reached in 2013,
however, it became much more expensive to increase the volume of ethanol blended into the
gasoline pool. While obligated parties could still import sugarcane ethanol to satisfy their
advanced biofuel obligations, doing so would reduce the volume of corn ethanol that could be
used as E10. Available non-ethanol renewable fuels were almost entirely advanced biodiesel and
renewable diesel, so obligated parties generally used these fuels (rather than sugarcane ethanol)
to meet the advanced biofuel requirements so that they could use corn ethanol to satisfy the
remaining total renewable fuel requirements. RIN prices responded, and since 2013 the prices of
D4 and D5 RINs have been nearly identical.
D4 RIN prices, much like all RIN prices, have varied significantly since 2010. The
pricing of these RINs, however, has been fairly straightforward. D4 RINs are generally priced to
account for the price difference between biodiesel and petroleum diesel, which in turn are largely
a function of the pricing of their respective oil supplies. Other factors can also impact this
relationship; most significantly are the presence or absence of the biodiesel tax credit and the
impact of other subsidies and credits (e.g., the $1.00 per gallon federal tax subsidy and state
LCFS credits).89 Most recently, in 2021 and 2022, D4 RIN prices have increased quite
significantly, tracking with an increase in feedstock commodity prices (e.g., soybean oil), which
comprise greater than 80% of the cost of production of BBD. Generally, D4 RIN prices have
increased to a level that allows BBD to be cost-effective with petroleum-based fuels, increasing
BBD production and use. A 2018 paper exploring the relationship between the price of D4 RINs
and economic fundamentals concluded that "movements in D4 biodiesel RIN price at
frequencies of a month or longer are well explained by two economic fundamentals: (a) the
spread between the biodiesel and ULSD prices and (b) whether the $1 per gallon biodiesel tax
87 "Other advanced biofuel" is not an RFS standard category, but is the difference between the advanced biofuel
requirement and the sum of the cellulosic biofuel and BBD requirements, both of which are nested within the
advanced biofuel category.
88 See Chapters 6.3 and 6.2 for volumes of sugarcane ethanol and BBD used in the U.S., respectively.
89 A $1 per gallon biodiesel blenders tax credit has been available to biodiesel blended every year from 2010-2022.
However, at various times this credit has expired and been reinstated retroactively. The biodiesel tax credit expired
at the end of 2009 and was not reinstated until December 2010, applying to all biodiesel blended in 2010 and 2011.
The biodiesel tax credit has since been again reauthorized semi-regularly, including in January 2013 (applying to
biodiesel produced in 2012 and 2013), December 2014 (applying to biodiesel produced in 2014), December 2015
(applying to biodiesel produced in 2015 and 2016), and February 2018 (applying to biodiesel produced in 2017). In
December 2019 the tax credit was retroactively reinstated for 2018 and 2019 and put in place prospectively through
2022. In August 2022, the tax credit was extended through 2024. Beginning in 2025 biodiesel and renewable diesel
could qualify for the clean fuel production credit.
41
-------
credit is in effect."90 This same paper discusses in greater detail the strong correlation between
weekly D4 RIN prices and predicted D4 RIN price values using a model based on economic
fundamentals. As state LCFS programs have come online and increased in stringency, the value
of these credits is now another increasingly important factor.
Data on cellulosic RIN (D3 and D7) prices were not generally available until 2015. This
is likely due to the fact that prior to 2015, the market for cellulosic RINs was too small to support
commercial reporting services; very few cellulosic RINs were generated and traded in years prior
to 2016. From 2015—when D3 RIN prices were first regularly available—through 2018, the
price of these RINs was very closely related to the sum of the D5 RIN price plus the price of the
cellulosic waiver credit (CWC).91 This is as expected, since obligated parties can satisfy their
cellulosic biofuel obligations through the use of either cellulosic RINs or CWCs plus D5 RINs.
The slight discount for D3 RINs (as opposed to the combination of a CWC and a D5 RIN) is also
as expected, as CWCs can be purchased directly from EPA when obligated parties demonstrate
compliance and carry no risk of RIN invalidity.92 This discount tends to be larger at the
beginning of the year, before narrowing near the end of the year as the RFS compliance deadline
nears for obligated parties. Starting in 2019, the D3 RIN price was significantly lower than the
CWC plus D5 RIN price. This is likely due to an over-supply of D3 RINs caused by EPA
granting a relatively large number of SREs for the 2017 and 2018 compliance years, lowering the
effective RFS standards (see Chapter 1.2). The average D3 RIN price fell to near the D5 RIN
price, before slowly increasing relative to the D5 RIN price starting in the second half of 2019
and remaining between the D5 RIN price and the D5 plus CWC price through the end of 2022.
90 Irwin, S.H, K. McCormack, and J. H. Stock (2018). "The price of biodiesel RINs and economic fundamentals."
NBER working paper series, working paper 25341.
91 CAA section 21 l(o)(7)(D)(ii) established a price cap mechanism for cellulosic biofuel RINs. In implementing this
provision, EPA makes CWCs available for sale to obligated parties at a price determined by a statutory formula in
any year in which EPA reduces the required volume of cellulosic biofuel using the cellulosic waiver authority. A
CWC satisfies an obligated party's cellulosic biofuel obligation. However, unlike a cellulosic RIN, which also helps
satisfy an obligated party's advanced biofuel and total renewable fuel obligations, a CWC does not help satisfy an
obligated party's advanced biofuel and total renewable fuel obligations. A cellulosic RIN (which can be used to
meet all 3 obligations) has similar compliance value as a CWC (which can only be used to satisfy the cellulosic
biofuel obligation) and an advanced RIN (which can be used to satisfy the advanced biofuel and total renewable fuel
obligations).
92 During a few time periods (such as late 2016), the price for D3 RINs was higher than the price for a CWC + D5
RIN. This was likely due to the fact that up to 20% of a previous year's RINs can be used towards compliance in
any given year, while CWCs can only be used towards compliance obligations in that year. Obligated parties likely
purchased 2016 D3 RINs at a premium anticipating the sharp increase in the CWC price in 2017.
42
-------
Figure 1.9.2-3: D3 RIN Prices and D5 RIN Price Plus CWC Price93
$5.00
$4 JO
$4.00 -
$350
J
/
tT
$3.00
$2.50
$2.00
$150 *4,
$1.00 "|ii
$0.50 If |
i
i 38
| II.
1/1/15
1/1/16 1/1/17 1/1/18 1/1/19 1/1/20 1/1/21
D3 RIN Price D5 RIN Price D5 + CWC Price
1/1/22
Source: RIN price data from EMTS
The fact that the price of D3 RINs, with very few exceptions, has not exceeded the CWC
plus D5 RIN price has potentially significant consequences for both the cellulosic biofuel and
petroleum fuel markets. For obligated parties, the CWC price effectively sets a maximum price
for cellulosic RINs (CWC plus the D5 RIN price) and protects these parties from excessively
high cellulosic RIN prices. The CWC price is also informational to potential cellulosic biofuel
producers. Potential cellulosic biofuel producers can use the CWC price, along with the price of
the petroleum fuel displaced by the cellulosic biofuel they produce and any tax credits or other
incentives available for the fuel, as an approximation of the maximum price they can reasonably
expect to receive for the cellulosic biofuel they produce. Knowing this price can help potential
cellulosic biofuel producers determine whether their cellulosic biofuel production processes are
economically viable under both current and likely future market conditions.
At the same time, the relatively high value of the CWC plus D5 RIN price, in conjunction
with EPA's statutory obligation from 2010 to 2022 to set the required volume of cellulosic
biofuel at the volume expected to be produced each year,94 has resulted in generally high D3
RIN prices. These RIN prices are realized for all cellulosic RINs, even those generated for
biofuels such as CNG/LNG derived from biogas that can often be produced at a cost that is
competitive with the petroleum fuels they displace even without the RIN value. While some of
this excess RIN value may be passed on to consumers who use CNG/LNG derived from biogas
as transportation fuel in the form of lower cost fuel and/or longer term fixed-price fuel contracts,
a significant portion of the RIN value may remain with the biofuel producer, the parties that
dispense CNG/LNG derived from biogas, and any other parties involved in the production of this
93 EPA offers cellulosic waiver credits for years in which we reduce the cellulosic biofuel volume from the statutory
target. Cellulosic waiver credit prices are available at: Utps://www.epa.gov/renewable-fuel-standard-
program/cellulosic-waiver-credits-under-renewable-fuel-standard-program.
54 CAA section 21 l(o)(7)(D).
43
-------
type of cellulosic biofuel.95 Unlike other RIN costs that are generally transferee! within the liquid
fuel pool (e.g., from consumers of fuels with relatively low renewable fuel content such as EO or
BO to consumers of fuels with relatively high renewable fuel content such as E85 or B20), much
of the RIN value for CNG/LNG derived from biogas may be transferred from consumers who
purchase gasoline and diesel to parties outside of the liquid fuel pool (e.g., landfill owners). For
example, the average cellulosic RIN price was $3.02 in 2022.96 Thus, the total cost associated
with the 630 million cellulosic RINs required for compliance in 2022 was approximately $1.9
billion. Therefore, the cellulosic biofuel requirement likely increased the price of gasoline and
diesel sold in the U.S. in 2022 by approximately $0.01 per gallon.97 These transfers are expected
to increase significantly through 2025 as a result of the cellulosic biofuel volumes we are
finalizing in this rule. For example, using the average cellulosic RIN price for May 2022 - April
2023 (the last six months for which data are available) of $2.79 and the final cellulosic biofuel
volume for 2025 of 1.38 billion gallons, we estimate that the cost associated with cellulosic RIN
purchases would be $3.85 billion, and would be expected to increase the price of gasoline and
diesel in 2025 by approximately $0.02 per gallon.98
1.10 Carryover RIN Proj ections
This section details the calculations performed by EPA to project the number of available
carryover RINs in the context of developing the final 2023-2025 RFS standards. While the actual
number of carryover RINs available for use by obligated parties to use towards these standards
will not be known until after compliance with the preceding year's standards is complete, we are
able to project these values by using 2021 compliance data and assumptions about RIN
generation relative to RIN obligations in 2022. Chapter 1.10.1 calculates the number of carryover
RINs available after compliance with the 2021 standards. Chapter 1.10.2 projects the number of
carryover RINs that will be available for compliance with the 2023-2025 standards. Chapter
1.10.3 provides historical data on the number of available carryover RINs. Chapter 1.10.4
summarizes EMTS data on RIN retirements and errors.
1.10.1 Number of Available Carryover RINs After Compliance With the 2021
Standards
In order to calculate the number of 2021 carryover RINs available for compliance with
the 2022 standards, we began with the 2021 RFS compliance year data in Table 1.10.1-1. From
this data, we calculated that approximately 20.39 billion total RINs were retired for compliance
95 EPA currently does not have sufficient data to determine the proportion of the RIN value that is used to discount
the retail price of CNG/LNG derived from biogas when used as transportation fuel.
96 Average D3 RIN price in 2022 according to EMTS RIN price data.
97 In the February 2023 STEO, EIA forecasted gasoline and diesel consumption in 2022 at 8.78 million bpd (134.6
billion gallons per year) and 3.68 million bpd (56.4 billion gallons per year) respectively. Dividing the total cost of
cellulosic RINs in 2021 ($1.9 billion) by the total consumption of gasoline and diesel (191.0 billion gallons) results
in an estimated cost of $0.01 per gallon of gasoline and diesel as a result of the cellulosic biofuel requirement.
98 In the 2023 AEO, EIA forecasted gasoline and diesel consumption in 2025 at 138.4 billion gallons and 52.4
billion gallons respectively. Dividing the total cost of cellulosic RINs in 2025 ($3.85 billion) by the total
consumption of gasoline and diesel (190.8 billion gallons) results in an estimated cost of $0,020 per gallon of
gasoline and diesel as a result of the cellulosic biofuel requirement.
44
-------
in the 2021 compliance year." Of this total, approximately 18.58 billion 2021 RINs and 1.81
billion 2020 carryover RINs were used.
Table 1.10.1-1: RINs Retired by Obligated Parties and Exporters in the 2021 Compliance
Year3
RIN Type
RIN Year
Total
2020
2021
D3
40,866,327
538,249,180
579,115,507
D4
79,946,630
4,757,209,816
4,837,156,446
D5
34,020,833
229,546,265
263,567,098
D6
1,650,618,764
13,058,319,692
14,708,938,456
D7
0
0
0
Total
1,805,452,554
18,583,324,953
20,388,777,507
a RINs include those retired by companies with an RVO as a gasoline/diesel fuel importer or refiner, as well as RINs
retired by companies with an RVO as renewable fuel exporters. Renewable fuel exporters include exporters of neat
renewable fuel, as well as exporters of renewable fuel blended with other fuels (including, but not limited to,
gasoline, diesel fuel, heating oil, and jet fuel). See Table 1.10.4-1 for more detailed data.
Next, we calculated the net number of RINs that were generated in 2021. To do this, we
took the total number of RINs generated in 2021 and then removed any RINs that were generated
in error, as well as any RINs that were retired for purposes other than satisfying an obligated
party or exporter RVO (e.g., for spills, remedial actions, enforcement obligations, etc.). Using
the data in Table 1.10.1-2, we calculated that a net of approximately 19.72 billion RINs were
generated in 2021.
Table 1.10.1-2: 2021 Net RINs Generated3
RIN Type
Total RINs
Generatedb
RIN
Errors0
Other RIN
Retirements'1
Net RINs
Generated®
D3
568,414,154
5,205,881
188,456
563,019,817
D4
4,873,631,040
1,952,566
56,904,276
4,814,774,198
D5
233,872,099
830,150
142,240
232,899,709
D6
14,258,274,476
13,588,686
139,641,775
14,105,044,015
D7
247,518
0
0
247,518
Total
19,934,439,287
21,577,283
196,876,747
19,715,985,257
a Data from April 2023 and compiled from https J/www .epa. gov/sv s te m/fi tes/o the r-fi les/2023 -
05/avaitabterins Apr2023.csv and https://www.epa.gov/svstem/files/other-files/2023-
05/retiretransaction_Apr2023.csv.
b The total number of RINs generated includes those RINs generated for exported fuel.
0 See Table 1.10.4-2 for more detailed data.
d See Table 1.10.4-3 for more detailed data.
e Net RINs Generated = Total RINs Generated - (RIN Errors + Other RIN Retirements).
To determine the total number of 2021 carryover RINs available for compliance with the
2022 standards, we then subtracted the number of 2021 RINs retired in the 2021 compliance year
99 Includes RINs retired in the 2021 compliance year to satisfy 2020 compliance deficits.
45
-------
from the net number of 2021 RINs generated. We calculate that there are approximately 1.13
billion 2021 carryover RINs available, as shown in Table 1.10.1-3.
Table 1.10.1-3: 2021 Carryover RINs
RIN Type
Net 2021 RINs
Generated
2021 RINs Retired
for Compliance
2021 Carryover
RINs
D3
563,019,817
538,249,180
24,770,637
D4
4,814,774,198
4,757,209,816
57,564,382
D5
232,899,709
229,546,265
3,353,444
D6
14,105,044,015
13,058,319,692
1,046,724,323
D7
247,518
0
247,518
Total
19,715,985,257
18,583,324,953
1,132,660,304
Obligated parties are also able to carryforward a compliance deficit from one year to the
next year,100 increasing their RVO for 2022 and effectively decreasing the number of 2021
carryover RINs available for compliance with the 2022 standards. In order to account for this, we
calculate the effective number of 2021 carryover RINs available for compliance with the 2022
standards by subtracting out the 2021 compliance deficits, which have to be satisfied at the time
of compliance with the 2022 standards.101 We note, however, that 2021 compliance deficits
exceeded the number of available 2021 carryover RINs for several standards, which means that
there was a shortfall in the number of RINs available to comply with these standards in 2021 and
that some obligated parties had to carry forward a deficit into 2022. After accounting for this
adjustment, the effective number of 2021 carryover RINs available for compliance with the 2022
standards are shown in Table 1.10.1-4.102
100 See 40 CFR 80.1427(b).
101 The compliance deadline for the 2022 standards will be the first quarterly reporting deadline after the effective
date of this action.
102 In other words, the number of available carryover RINs is effectively reduced in light of the volume of 2021
deficits carried forward to 2022. We note, moreover, that these numbers could change based on, for instance,
enforcement actions or obligated parties truing up their RVOs pursuant to the attest engagement required by 40 CFR
80.1464.
46
-------
Table 1.10.1-4: Efl
ective 2021 (
Carryover RINsa
RFS Standard
RIN Type
2021
Carryover
RINs
2021
Compliance
Deficits
Net Surplus/
Deficitb
Effective 2021
Carryover
RINs
Cellulosic
Biofuel
D3+D7
25,018,155
32,558,093
-7,539,938
0
Non-Cellulosic
Advanced
Biofuelc
D4+D5
60,917,826
457,984,877
-397,067,051
0
Conventional
Renewable Fueld
D6
1,046,724,323
553,171,784
493,552,539
493,552,539
a Data current as of May 17, 2023, and compiled from Table 5 at https://www.epa.gov/fuels-registration-reporting-
and-compliance-heip/annnal-compliance-data-obligated-parties-and.
b Net Surplus/Deficit = Carryover RINs - Compliance Deficits. Negative values represent a shortfall in the number
of RINs available to comply with the applicable standard and are counted as zero for purposes of determining the
effective number of available carryover RINs.
0 Non-cellulosic advanced biofuel is not an RFS standard category but is calculated by subtracting the number of
cellulosic RINs from the number of advanced RINs.
d Conventional renewable fuel is not an RFS standard category but is calculated by subtracting the number of
advanced RINs from the number of total renewable fuel RINs.
1.10.2 Number of Available Carryover RINs for 2023-2025
Given the uncertainty of the impact of compliance with the 2022 standards on the number
of available carryover RINs, we are unable to provide a quantitative analysis of the number of
carryover RINs that may be available for compliance with the 2023-2025 standards.103 However,
if we assume that the uncertainties result in neither a net gain nor net loss of excess RINs for
2022, then the carryover RINs that we projected to be available in Chapter 1.10.1 would
represent the number of carryover RINs available for compliance with the 2023-2025 standards,
as shown in Table 1.10.2-1.104
103 Sources of uncertainty that could potentially increase the number of carryover RINs include lower actual gasoline
and diesel fuel use than the projection used to derive the standards. Sources of uncertainty that could potentially
decrease the number of carryover RINs include enforcement actions and higher actual gasoline and diesel fuel use
than the projection used to derive the standards.
104 The actual number of RINs that will be available for use by obligated parties to use towards the 2023-2025
standards will not be known until the compliance deadline for the preceding compliance year. Even after this date,
however, this number could change based on, for instance, obligated parties truing up their RVOs pursuant to the
attest engagement required by 40 CFR 80.1464 or enforcement actions.
47
-------
Table 1.10.2-1: Projected Carryover RINs for 2023-2025
RFS Standard
RIN Type
Projected Effective
Carryover RINs
Cellulosic Biofuel
D3+D7
0
Non-Cellulosic Advanced BiofueP
D4+D5
0
Conventional Renewable Fuelb
D6
493,552,539
a Non-cellulosic advanced biofuel is not an RFS standard category but is calculated by subtracting the number of
cellulosic RINs from the number of advanced RINs.
b Conventional renewable fuel is not an RFS standard category but is calculated by subtracting the number of
advanced RINs from the number of total renewable fuel RINs.
We note that while we project that there will effectively be no cellulosic biofuel or non-
cellulosic advanced biofuel carryover RINs available for compliance with the 2023-2025
standards, this does not mean that actual carryover RINs will not be available in these years. As
discussed in Chapter 1.10.1, the actual number of carryover RINs available relative to the
"effective" number is a function of the volume of RIN deficits that obligated parties carry
forward from one year into the next. For example, if obligated parties carry forward a significant
volume of RIN deficits, then the absolute number of carryover RINs available for compliance
with the following year's standards will be larger than were obligated parties to carry forward a
smaller volume of RIN deficits.
1.10.3 Carryover RIN History
In order to provide a historical perspective on the number of available carryover RINs,
we calculated the absolute and effective number of carryover RINs for each year since 2013
using the same methodology described in Chapter 1.10.1. The results are provided in Table
1.10.3-1 and Figures 1.10.3-1 through 3 and represent the number of RINs of a given vintage
available for compliance with the subsequent year's standard (e.g., the number of available
carryover RINs in 2021 are those 2021 RINs that can be used to comply with the 2022
standards).
48
-------
Table 1.10.3-1: Number of Available Carryover RINs History (million RINs)
Non-Cellulosic
Conventional
Compliance
Year
Cellulosic Biofuel
Advanced Biofuel
Renewa
)le Fuel
Absolute3
Effectiveb
Absolute3
Effectiveb
Absolute3
Effectiveb
2013
0
0
565
538
1,088
1,046
2014
12
12
467
445
1,373
1,254
2015
39
39
372
367
1,248
1,242
2016
39
34
888
826
1,947
1,623
2017
29
9
844
727
3,061
2,517
2018
54
51
645
619
2,896
2,800
2019
53
41
221
23
2,186
1,743
2020
42
12
131
0
1,764
1,178
2021
25
0
61
0
1,047
494
a Represents the absolute number of carryover RINs that are available for compliance with the subsequent year's
standards and does not account for carryforward deficits.
b Represents the effective number of carryover RINs that are available for compliance with the subsequent year's
standards after accounting for carryforward deficits. Standards for which deficits exceed the number of available
carryover RINs are represented as zero.
Figure 1.10.3-1: Number of Available Conventional Renewable Fuel Carryover RINs
3,500
£ 3,000
C
o
= 2,500
15 2,000
'ro
< 1,500
V)
2
t 1,000
a;
>
o
500
2013 2014 2015 2016 2017 2018 2019
RVO Year
2020 2021
-Absolute
¦ Effective
49
-------
Figure 1.10.3-2: Number of Available Non-Cellulosic Advanced Biofuel Carryover RINs
1,000
CC
£
O
900
800
700
600
500
400
300
200
100
2013 2014 2015 2016 2017 2018 2019 2020 2021
RVO Year
Absolute
¦ Effective
Figure 1.10.3-3: Number of Available Cellulosic Biofuel Carryover RINs
100
"t/T
90
cd
c.
80
o
1
70
60
-Q
50
>
<
40
V)
2
C£L
30
O
20
£¦
ro
10
u
0
2013
2020
2021
¦Absolute
Effective
50
-------
1.10.4 EMTS RIN Data
Table 1.10.4-1: RINs Retired by Importers, Refiners, and Exporters in the 2021
Compliance Year"
RIN Type
Year
Importers
Refiners
Exporters
Total
D3
2020
1,278,365
39,587,962
0
40,866,327
2021
20,458,104
517,791,076
0
538,249,180
D4
2020
3,113,161
72,344,765
4,488,704
79,946,630
2021
166,008,756
4,046,472,627
544,728,433
4,757,209,816
D5
2020
259,135
33,761,698
0
34,020,833
2021
7,816,733
221,242,368
487,164
229,546,265
D6
2020
45,309,368
1,582,509,068
22,800,328
1,650,618,764
2021
506,006,642
12,255,894,529
296,418,521
13,058,319,692
D7
2020
0
0
0
0
2021
0
0
0
0
Total
750,250,264
18,769,604,093
868,923,150
20,388,777,507
a Data current as of May 17, 2023, and compiled from Table 3 at https://www.epa.gov/fuels-registration-reporting-
and-compliance-heip/annnal-compliance-data-obligated-parties-and.
Table 1.10.4-2: 2021 RIN Errorsa
RIN Type
Import Volume
Correction
Invalid RIN
Volume error
correction
Total
Retirement Code
30
SO
60
—
D3
0
5,205,881
0
5,205,881
D4
0
1,622,929
329,637
1,952,566
D5
0
824,414
5,736
830,150
D6
0
9,859,895
3,728,791
13,588,686
D7
0
0
0
0
Total
0
17,513,119
4,064,164
21,577,283
a Data from April 2023 and compiled from https J/www .epa. gov/sv s te in/fi tes/o the r-fi les/2023 -
05/retiretransaction Apr2023.csv.
51
-------
Table 1.10.4-3: Of
her 2021 RIN Retirements3
RIN Type
Reported
spill
Contaminated
or spoiled fuel
Renewable fuel used in
an ocean-going vessel
Enforcement
Obligation
Retirement Code
10
20
40
70
D3
0
0
0
0
D4
6,277
170,062
46,179
37,458
D5
0
36,820
0
0
D6
15,723
80,910
0
621,738
D7
0
0
0
0
Total
22,000
287,792
46,179
659,196
RIN Type
Renewable fuel used or designated
to be used in any application that is
not transportation fuel heating oil or
jet fuel
Delayed RIN
Retire per
80.1426(g)(3)
only
Remedial action
- Retirement
pursuant to
80.1431(c)
Retirement Code
90
100
110
D3
0
0
188,456
D4
47,826,467
0
638,669
D5
0
0
105,420
D6
135,074,542
0
3,650,317
D7
0
0
0
Total
182,901,009
0
4,582,862
Remediation of
Voluntary
Feedstock using
Invalid RIN Use
RIN
renewable fuel
RIN Type
for Compliance
Retirement
with RINs
Total
Retirement Code
130
160
170
—
D3
0
0
0
188,456
D4
0
0
8,179,164
56,904,276
D5
0
0
0
142,240
D6
198,545
0
0
139,641,775
D7
0
0
0
0
Total
198,545
0
8,179,164
196,876,747
a Data from April 2023 and compiled from https J/www .epa. gov/sv s te m/fi tes/o the r-fi les/2023 -
05/retiretransaction_Apr2023.csv.
52
-------
1.11 Gasoline and Diesel Proj ections
This section reviews the accuracy of the projected volumes of gasoline and diesel used to
calculate the percentage standards. In Chapter 1.11.1, we discuss the differences between EIA's
projections of gasoline and diesel volumes and those volumes reported by obligated parties under
the RFS program. In Chapter 1.11.2, we provide possible explanations for these differences. In
Chapter 1.11.3, we detail the calculations used to create an additional adjustment factor to EIA's
gasoline and diesel volumes in order to improve their agreement with those volumes reported by
obligated parties.
1.11.1 Difference Between EIA and Obligated Party Reported Volumes
In the 2020-2022 final rule, we revised the 2020 percentage standards such that the
required volumes of renewable fuel for each RFS category were set to the actual volumes of
those renewable fuels used in that year, as provided by EMTS RIN generation data.105 In tandem,
we used the actual volumes of gasoline and diesel used in 2020, taken from EIA's STEO Data
Browser, to calculate the revised percentage standards for 2020. Based on these calculations, we
expected that the obligated volume of gasoline and diesel reported by obligated parties for 2020
would be virtually identical to the volume provided by EIA such that the total number of RINs
required to be retired for 2020 compliance would equal the volume of total renewable fuel used
to establish the revised 2020 total renewable fuel percentage standard. This would result in
neither a net gain nor net loss of RINs in the number of available carryover RINs as described in
the 2020-2022 final rule. We used the same approach to establish the 2021 standards in that
same action—namely, using the actual volumes of renewable fuel and gasoline and diesel to
calculate the 2021 percentage standards, with the expectation that the number of available
carryover RINs would remain static going into 2022, and allow us to establish market-forcing
standards for 2022.106
Upon review of the 2020 compliance report data submitted on December 1, 2022,
however, we noticed an unexpected and significant difference between the volume of obligated
gasoline and diesel reported by obligated parties and the volume of gasoline and diesel provided
by EIA that we used to calculate the revised 2020 percentage standards. Specifically, obligated
parties reported a total obligated gasoline and diesel volume of 167.4 billion gallons, whereas the
2020 percentage standards were based on a total gasoline and diesel volume of 158.2 billion
gallons. As a result, rather than the intended 17.13 billion RINs being required for 2020
compliance, obligated parties reported a total RVO of 18.11 billion RINs for 2020. In other
words, this 9-billion-gallon discrepancy in the gasoline and diesel volume resulted in a total 2020
RVO 980 million RINs more than the actual volume of renewable fuels used in 2020 that was
intended.
A similar difference was observed upon review of the 2021 compliance report data
submitted on March 31, 2023. Specifically, obligated parties reported a total obligated gasoline
and diesel volume of 177.1 billion gallons, whereas the 2021 percentage standards were based on
a total gasoline and diesel volume of 168.4 billion gallons. As a result, rather than the intended
105 87 FR 39600, 39602 (July 1, 2022).
106 87 FR 39600, 39602-03 (July 1, 2022).
53
-------
18.84 billion RINs being required for 2021 compliance, obligated parties reported a total RVO of
19.82 billion RINs for 2021. In other words, this 9-billion-gallon discrepancy in the gasoline and
diesel volume resulted in a total 2021 RVO 980 million RINs more than the actual volume of
renewable fuels used in 2021 that was intended.
In order to determine whether this difference between obligated party-reported and EIA
gasoline and diesel volumes was an issue limited to the 2020 and 2021 compliance years—
possibly attributable to the onset of the COVID-19 pandemic—we made the same comparison
for all compliance years dating back to 2013. To do so, we made an adjustment to the gasoline
and diesel volumes reported by obligated parties to better align it with EIA's estimate of non-
renewable gasoline and diesel volumes. Small refineries that receive an exemption from their
RFS obligations under 40 CFR 80.1441 are not required to submit a compliance report for the
year in which they receive an exemption and, as a result, the volumes of gasoline and diesel they
produce in that year are not reported to EPA.107 To adjust for this, we added the unreported
volume of gasoline and diesel from exempt small refineries—using the volumes projected by the
exempted small refineries in their small refinery exemption (SRE) petitions—to the total
obligated volume of gasoline and diesel reported by obligated parties.108 The results are provided
in Table 1.11.1-1.
107 EIA's estimate is based on the total volume of gasoline and diesel consumed in the United States in a given year,
irrespective of where or by whom the fuel was produced. This means that fuel produced by exempt small refineries
is included in EIA's estimates, but is not captured in the volumes reported to EPA.
108 While for most years we were able to simply add the estimated volume of exempt gasoline and diesel from
EPA's SRE website (Table 1, https://www.epa.gov/fuels-registration-reporting-and-compliance-help/rfs-smaH-
refi nerv -exe nipt ions') to the total reported volume of gasoline and diesel for that year (Table 1,
https://www.epa.gov/fiieis-registration-reporting-and-compliance-heip/aniinal-compliance-data-obligated-parties-
and), a different approach was used for 2018. For this year, EPA had initially exempted over 13.4 billion gallons of
gasoline and diesel, but in the "April 2022 Denial of Petitions for RFS Small Refinery Exemptions," EPA-420-R-
22-005 ("April 2022 SRE Denial Action"), EPA subsequently denied all of those previously-granted SRE petitions
upon remand and the total reported exempt volume of gasoline and diesel is now listed as 0 gallons on the SRE
website. In a companion action issued at the same time, the "April 2022 Alternative RFS Compliance
Demonstration Compliance Approach for Certain Small Refineries," EPA-420-R-22-006 ("April 2022 Compliance
Action"), EPA allowed the 31 small refineries covered by the April 2022 SRE Denial Action to resubmit their 2018
RFS annual compliance reports with zero deficit carryforward and no additional RIN retirements. In analyzing the
2018 data, however, only 2.8 billion gallons of additional gasoline and diesel has been reported since the April 2022
SRE Denial Action, and so much of that initially exempt 13.4 billion gallons seemingly remains unreported. As
such, we added 10.6 billion gallons (13.4 - 2.8 = 10.6) to the reported volume of gasoline and diesel for 2018.
54
-------
Table 1.11.1-1: Adjusted Volumes of Gasoline and Diesel (billion gallons)
Year
Reported
Volume of
Gasoline
and Diesel3
Unreported
SRE
Volumesb
Total Adjusted
Volume of
Gasoline and
Diesel
EIA Actual
Volume of
Gasoline
and Diesel0
Absolute
Difference
Relative
Difference
2013
173.0
2.0
175.0
171.1
-3.9
-2.2%
2014
175.9
2.3
178.2
174.6
-3.6
-2.0%
2015
178.4
3.1
181.5
178.0
-3.5
-1.9%
2016
173.6
7.0
180.6
178.5
-2.1
-1.2%
2017
167.4
16.1
183.4
179.5
-3.9
-2.1%
2018
176.0
10.6
186.6
182.1
-4.5
-2.4%
2019
187.4
0.0
187.4
180.7
-6.7
-3.6%
2020
167.4
0.0
167.4
157.7
-9.7
-5.8%
2021
177.1
0.0
177.1
171.5
-5.6
-3.2%
a Source: Table 1, https://www.epa.gov/fuels-registration-reporting-aiKl-compliance-help/annual-compliance-data-
ob t i ga ted-part ies-a nd.
b Source: Table 1, https://www.epa.gov/fuels-registration-reporting-and-compliance-help/rfs-small-refinery-
exemptions (for all years except 2018). For 2018, see footnote 108.
0 Source: EIA STEO Data Browser (adjusted to exclude fuel used in Alaska and ocean-going vessels).
As the figures in Table 1.11.1-1 demonstrate, the volume of gasoline and diesel reported
by obligated parties and adjusted to reflect unreported SRE volumes has been consistently
greater than the volume reported by EIA.109 More importantly for purposes of the present
discussion, this difference has grown in recent years, leading to the retirement of more RINs by
obligated parties in 2020 and 2021 than EPA intended.
It is significant that the difference between the adjusted reported volume of gasoline and
diesel by obligated parties and the volume reported by EIA is consistently in one direction (i.e.,
EIA's volume of gasoline and diesel has always been below the adjusted reported volume of
gasoline and diesel by obligated parties) and appears to be growing in magnitude. When the
obligated volume of gasoline and diesel (used by obligated parties for compliance) is higher than
volume of gasoline and diesel provided by EIA (used by EPA to set the standards), the
percentage standards are consequentially more stringent than necessary to achieve the intended
renewable fuel volumes. This is due to the fact that, in simple terms, percentage standards are
calculated by dividing renewable fuel volume by gasoline and diesel volume. For 2020 and 2021,
the ultimate result of this difference in gasoline and diesel volumes is that approximately 980
million RINs and 880 million RINs, respectively, will be retired by obligated parties in excess of
109 This difference went largely unnoticed until now because, until recently, in many years we had granted
significant numbers of SREs such that the volume of gasoline and diesel reported to EPA was below that of EIA's
projected volume of gasoline and diesel used to calculate the percentage standards. See, e.g., 2016-2018 in Table 1,
https://www.epa.gov/fiiels-registration-reporting-aiid-compliance-help/aniinal-compliaiice-data-obligated-parties-
and. In other words, SREs were masking the discrepancy and it was not until 2020—when the percentage standards
were revised retroactively to the actual volumes and no SREs were granted—that it became apparent that there was
an issue with the gasoline and diesel volumes used to set the standards in comparison to the volumes used for
compliance.
55
-------
the level that we intended, leading to an unanticipated drawdown of the number of available
carryover RINs.110
1.11.2 Possible Reasons for Difference Between EIA and Obligated Party
Reported Volumes
At the time the RFS program was created, exports of gasoline and diesel averaged around
4-5 billion gallons per year. Since that time, however, at the same time renewable fuel use has
expanded, exports of gasoline and diesel have grown significantly and now average around 30
billion gallons per year, as shown in Figure 1.11.2-1. For the reasons discussed in the following
sections, we believe that exported gasoline and diesel is the main source of the difference
discussed in Chapter 1.11.1.
Figure 1.11.2-1: Gasoline and Distillate Exports
35
30
25
1,0
fij
U
c
~ 15
S
10
5
0
Data Source: EIA Finished Motor Gasoline and Distillate Fuel Oil Export Data
1.11.2.1 Treatment of Exports Under the RF S Program
While EPA's implementation of the RFS program is intended to impose an obligation
exclusively on gasoline and diesel that is consumed in the U.S.,111 only gasoline and diesel that a
refiner designates for export at the refinery112 is actually able to be excluded from incurring an
1111 In the 2020-2022 final rule, EPA expected that the number of available carryover RINs would remain static after
compliance with the 2020 and 2021 standards. However, as discussed in Chapter 1.10 and Preamble Section III.C.4,
the effective number of available carryover RINs has been significantly and unexpectedly drawn down as a result of
2020 and 2021 compliance.
111 Gasoline and diesel used outside the RFS covered area is specifically excluded from incurring an RVO under 40
CFR 1407(f)(5).
112 We have always intended that exported gasoline and diesel obligated under the RFS program be treated
consistently with how exported gasoline and diesel is treated under EPA's fuel quality regulations. In order for a fuel
to be exempt from EPA's fuel quality standards, the fuel must be segregated from non-exempted fuel from the point
o
fN
ID
00
O
fN
¦^r
ID
00
o
fN
UD
00
O
fN
cn
CD
CFi
cn
CTI
o
O
o
O
O
T—1
1—1
T—1
T—1
fN
fN
<71
CTi
CTi
CTi
CTi
o
o
o
O
O
o
O
o
o
o
O
O
t-H
*—1
*—1
t-H
*—1
fN
fN
fN
fN
fN
fN
fN
fN
fN
fN
fN
fN
56
-------
RVO.113 Conversely, a refiner remains responsible for satisfying the RVO associated with the
fuel it produces even if that fuel is sold in the U.S. but is later redesignated for export
downstream. We believe that the majority of the exported gasoline and diesel volumes over the
years have likely been excluded by obligated parties in their RVO calculations in order to avoid
the added cost of the RIN obligation.
However, a considerable portion of these gasoline and diesel exports are also likely
exported by downstream parties opportunistically (i.e., downstream parties redesignate fuel that
was intended to be used domestically as instead fuel for export) in situations where market
pricing supports it despite the RIN cost differential. These exports show up as exported volume
collected by the U.S. Census Bureau in EIA data, but are not accounted for by obligated parties
in the volumes reported to EPA (i.e., obligated parties report an obligated volume of gasoline and
diesel that likely includes some volume of fuel that was in fact exported downstream).
In past years, this difference was consistently around 2-3% based on the data shown in
Table 1.11.1-1. However, in 2020 and 2021 the difference increased significantly. The timing of
this corresponded to significant changes domestically and internationally in gasoline and diesel
prices, which may have increased the opportunity for exports to occur by downstream parties
despite the RIN cost associated with the exported fuel. While we sought comment in the 2020
proposed rule on creating a mechanism by which refiners could exclude gasoline and diesel that
is redesignated for export by a downstream party,114 we did not finalize such provisions in the
2020 final rule.115
1.11.2.2 Treatment of Exports by EIA
While the treatment of exports under the RFS program is likely a significant contributor
to the difference in obligated party and EIA gasoline and diesel volumes, it is likely that the
treatment of exports by EIA is also a contributor. Using EIA's reported volumes of gasoline and
ethanol consumption, the average concentration of ethanol in gasoline is calculated to have been
10.39% in 2022. However, this is an impossibly high number given what we know about
consumption of E15 and E85 in the United States. Specifically, the number of retail service
stations offering these two fuel blends is too small to support the necessary sales volumes for the
of designation to the point that the fuel is ultimately exported (40 CFR 1090.645) and the fuel must be designated
for export prior to leaving the facility where it was produced (40 CFR 1090.1005(b)). As noted in the Fuels
Regulatory Streamlining Rule, these provisions were transferred from 40 CFR part 80 (see 85 FR 78430-31,
December 4, 2020).
113 When the RFS regulatory provisions were designed, EPA allowed obligated parties to exclude exported gasoline
and diesel volumes that they themselves exported or controlled to the point of export, but for implementation
streamlining reasons did not account for downstream opportunistic exports. For example, if an entity took delivery
of gasoline or diesel off a pipeline, but chose to export the fuel rather than sell it domestically, the upstream refinery
that produced the fuel would still incur an RVO for the volumes in question. This was also at a time when RIN
prices were low and renewable fuels were being blended into the gasoline and diesel fuel pools in excess of the RFS
program mandates, so any excess renewable fuel obligation that this may have created was negligible.
114 84 FR 36801 (July 29, 2019).
115 As part of the 2020 final rule, we stated that "While we are not at this time expanding the NTDF redesignation
provisions to allow refiners to exclude exporter gasoline, we may consider doing so in the future." "Renewable Fuel
Standard Program - Standards for 2020 and Biomass-Based Diesel Volume for 2021 and Other Changes: Response
to Comments," Chapter 9.1, pg 194.
57
-------
average ethanol concentration to be that high. Since EIA's estimate of overall ethanol
consumption mirrors very closely that reported to EPA in EMTS, the implication is that the
consumption volume of gasoline reported by EIA is likely an underestimate of the actual
volumes consumed in the U.S.
Based on recent discussions with EIA, we believe that some portion of gasoline is being
exported after being blended with renewable fuel (e.g., E10), but is being reported to EIA as EO.
This means that the ethanol portion of this exported gasoline is being incorrectly subtracted from
the total volume of gasoline consumed in the U.S., resulting in EIA's gasoline volumes being
lower than they should be. A similar issue with exports of diesel fuel may exist, but because of
how biodiesel and renewable diesel is blended, we believe this is less likely to be a significant
volume.
1.11.3 Gasoline and Diesel Projection Adjustment
As discussed in the previous section, at the outset of RFS program implementation,
exports of gasoline and diesel were relatively low and there was comparatively little
opportunistic exporting of obligated fuel by downstream entities. Today, however, exports of
gasoline and diesel have increased substantially. We now have reason to believe that a non-
trivial volume of this exported fuel is obligated gasoline and diesel that downstream parties have
redesignated for export because they are able to make a profit doing so even with the RIN
obligation attached to the fuel.
Without a mechanism by which obligated parties can account for this fuel when
calculating their production of obligated fuel, the nationwide volume of gasoline and diesel
reported by obligated parties is expected to remain higher than EIA's reported volume of
gasoline and diesel consumed in the U.S. As a result, compliance with the RFS standards is
likely to require higher volumes of renewable fuels than were intended by EPA in establishing
the percentage standards using EIA's projections. This bias is clearly not providing an accurate
projection of the volume of gasoline and diesel reported by obligated parties for RFS
compliance, and therefore requires that we make a correction to maintain program integrity.
In this final rule we are not in a position to finalize and implement the mechanism on
which we sought comment in the 2020 proposed rule given the time available. Doing so would
also not address the portion of the difference attributable to exports in EIA's projections.
However, in light of the unintended consequences of the difference between the volumes of
gasoline and diesel reported by obligated parties and those reported by EIA, we are compelled to
take steps to more accurately project the gasoline and diesel volumes for 2023-2025 that will be
reported by obligated parties such that the renewable fuel volumes ultimately required by the
RFS percentage standards would be expected to match the volumes used in Preamble Section
VII.C to calculate those percentage standards. As discussed above, the total adjusted volume of
gasoline and diesel reported by obligated parties has been consistently higher than the volume of
gasoline and diesel reported by EIA, upon whose forecast we are basing our percentage
standards. In using EIA's forecasts, we have, in the past, required more renewable fuel than our
rulemakings intended, which has resulted in a corresponding drawdown of the number of
available carryover RINs.
58
-------
In order to better match the percentage standards with the volumes we are promulgating,
we are applying an additional adjustment factor to EIA's projected volume of gasoline and diesel
used to calculate the percentage standards in order to make it more representative of the gasoline
and diesel volumes obligated parties are expected to report when they ultimately comply with the
2023-2025 standards.116 This factor will increase the projected gasoline and diesel volumes
beyond those projected by EIA for each of these three years, thereby slightly lowering the
percentage standards in order to compensate for the difference between EIA's projected volume
of gasoline and diesel and the volume reported by obligated parties. This adjustment does not
affect EPA's determination of the renewable fuel volumes used in setting the percentage
standards. Rather, it is a necessary step to ensure that the actual renewable fuel volumes required
to be consumed in the market match the targeted volumes.
To determine this adjustment factor, we compared EIA's forecasts of gasoline and diesel
in previous AEOs to the actual volume of gasoline and diesel reported by obligated parties.117
While we currently have compliance data through 2021, we believe that the sudden and dramatic
reduction of gasoline and diesel production in 2020 as a result of the COVID-19 pandemic
warrants the exclusion of 2020 from any comparison of forecast versus actual data. As such, we
limited our comparison to the 2013-2019 and 2021 compliance years.
Under the presumption that AEO forecasts of gasoline and diesel are likely to be less
accurate for years further into the future, we conducted our analysis separately for one-year
forecasts, two-year forecasts, and three-year forecasts. For instance, the volumes for 2016 that
were forecast in AEO2016 represented one-year forecasts, the volumes for 2017 that were
forecast in AEO2016 represented two-year forecasts, and so on. We used future gasoline and
diesel forecasts for AEO2011 through AEO2021 for forecast years 2013 through 2021, adjusted
downward to remove all renewable fuel, consumption in Alaska, and consumption by ocean-
going vessels to be consistent with the volumes used to calculated percentage standards and the
production volumes reported by obligated parties. We then compared the resulting adjusted AEO
forecasts to the volumes reported by obligated parties for the same years. The results are shown
in Tables 1.11.3-1 through 3.
116 We note that we already make several other adjustments to EIA's projected gasoline and diesel volumes,
including subtracting volumes of gasoline and diesel used in Alaska and volumes of diesel used in ocean-going
vessels. These adjustments—including the new AEO projection adjustment factor—are simply intended to more
accurately reflect the actual volume of gasoline and diesel reported by obligated parties.
117 As discussed in Preamble Section VILA, while we previously relied on EIA's STEO as our source of projected
gasoline and diesel volumes, for this action covering years through 2025 we must use EIA's AEO for gasoline and
diesel projections and therefore need to evaluate its projections—rather than STEO's—against the gasoline and
diesel volumes reported by obligated parties in order to calculate the adjustment factors.
59
-------
Table 1.11.3-1: Comparison of One-Year AEO Forecasts and Corresponding Obligated
Party Reported Volumes of Gasoline and Diesel
AEO
Edition
Target
Year
Gasoline + Diesel Volume
billion gallons)
Percent
Difference
AEO
Forecast3
Adjusted AEO
Forecastb'c
Obligated Party
Reported Volume
2013
2013
186.9
170.8
175.0
2.4%
2014
2014
188.1
172.5
178.2
3.3%
2015
2015
196.2
178.7
181.5
1.5%
2016
2016
200.0
180.7
180.6
0.0%
2017
2017
202.2
182.8
183.4
0.3%
2018
2018
202.5
181.6
186.6
2.8%
2019
2019
204.1
182.9
187.4
2.5%
2021
2021
193.6
173.2
177.1
2.2%
a Gasoline and diesel consumption derived from AEO Table 11.
b Adjustments include subtraction of renewable fuels, consumption in Alaska, and consumption by ocean-going
vessels.
0 Ethanol consumed in gasoline and E85 derived from AEO Table 2. Biodiesel, renewable diesel, and other biomass-
derived liquids consumption derived from AEO Table 11. Consumption in Alaska and ocean-going vessels
consistent with the volumes used to establish the applicable percentage standards.
Table 1.11.3-2: Comparison of Two-Year AEO Forecasts and Corresponding Obligated
Party Reported Volumes of Gasoline and Diesel
AEO
Edition
Target
Year
Gasoline + Diesel Volume
billion gallons)
Percent
Difference
AEO
Forecast3
Adjusted AEO
Forecastb'c
Obligated Party
Reported Volume
2012
2013
191.8
175.5
175.0
-0.3%
2013
2014
188.3
171.8
178.2
3.8%
2014
2015
190.8
174.9
181.5
3.8%
2015
2016
196.5
179.0
180.6
0.9%
2016
2017
201.5
181.9
183.4
0.8%
2017
2018
199.0
179.2
186.6
4.1%
2018
2019
198.7
177.8
187.4
5.4%
a Gasoline and diesel consumption derived from AEO Table 11.
b Adjustments include subtraction of renewable fuels, consumption in Alaska, and consumption by ocean-going
vessels.
0 Ethanol consumed in gasoline and E85 derived from AEO Table 2. Biodiesel, renewable diesel, and other biomass-
derived liquids consumption derived from AEO Table 11. Consumption in Alaska and ocean-going vessels
consistent with the volumes used to establish the applicable percentage standards.
60
-------
Table 1.11.3-3: Comparison of Three-Year AEO Forecasts and Corresponding Obligated
Party Reported Volumes of Gasoline and Diesel
AEO
Edition
Target
Year
Gasoline + Diesel Volume
billion gallons)
Percent
Difference
AEO
Forecast3
Adjusted AEO
Forecastb'c
Obligated Party
Reported Volume
2011
2013
199.6
182.1
175.0
-3.9%
2012
2014
193.0
176.3
178.2
1.1%
2013
2015
189.6
172.7
181.5
5.1%
2014
2016
191.8
175.9
180.6
2.7%
2015
2017
192.6
174.2
183.4
5.3%
2016
2018
200.6
180.7
186.6
3.3%
2017
2019
198.9
178.9
187.4
4.8%
a Gasoline and diesel consumption derived from AEO Table 11.
b Adjustments include subtraction of renewable fuels, consumption in Alaska, and consumption by ocean-going
vessels.
0 Ethanol consumed in gasoline and E85 derived from AEO Table 2. Biodiesel, renewable diesel, and other biomass-
derived liquids consumption derived from AEO Table 11. Consumption in Alaska and ocean-going vessels
consistent with the volumes used to establish the applicable percentage standards.
While the percent difference values appeared to exhibit a slightly increasing trend over
time, there was also considerable variability. Under the premise that data from more recent years
is likely to provide a better basis for making future projections than data for earlier years, we
used a weighted average of percent differences for all years where the weighting was higher for
more recent years and lower for earlier years. Specifically, the weighting factor for any given
year was twice as large as the weighting factor for the previous year. This approach is consistent
with that taken to project volumes of imported sugarcane ethanol and other advanced biofuel as
discussed in Chapters 6.3 and 6.4, respectively. The resulting weighted averages of the percent
differences are shown in Table 1.11.3-4.
Table 1.11.3-4: Weighted Average of Percent Differences Between Adjusted AEO Forecasts
One-year forecasts
2.2%
Two-year forecasts
4.1%
Three-year forecasts
4.2%
Given the variability in the percent difference values shown in Tables 1.11.3-1 through 3,
we believe that taking the average of these three values in Table 1.11.3-4 to produce a single
adjustment factor for simplicity's sake makes the most sense. Therefore, we have chosen to use
the average of the values from Table 1.11.3-4 as the adjustment factor when calculating the
applicable percentage standards for all three years 2023-2025. This value is 3.5%. In Preamble
Section VII.C, we have used this adjustment factor to increase the projected volumes for 2023-
2025.
61
-------
Chapter 2: Baselines
This document contains a collection of analyses examining factors identified in the CAA,
as well as other analyses EPA conducted to evaluate the impacts of this rule. The choice of
baseline has a first-order impact on the outcome of those analyses. In Preamble Section III.D, we
discuss the fact that a "No RFS" baseline is the most appropriate among available options for
purposes of evaluating the impacts of the final volumes for 2023-2025. This chapter describes
our derivation of the No RFS baseline, as well as an alternative baseline representing actual
renewable fuel consumption in 2022.
The No RFS baseline represents our projection of the world as it would exist if EPA did
not establish volume requirements for 2023-2025.118 Conceptually, the No RFS baseline allows
EPA to directly project the impacts of the candidate volumes for 2023-2025 relative to a
scenario without volume requirements. For the No RFS baseline, we assumed that the RFS
program existed as administered by EPA from its inception through 2022, and that renewable
fuel production developed with the support of the RFS program in these years. We also assumed
that non-RFS federal and state programs that support renewable fuel production and use (e.g., the
BBD tax credit and state LCFS programs), would continue to exist in 2023-2025.
While the No RFS baseline represents the renewable fuel volumes we expect would be
used in the U.S. if EPA did not establish RFS volume requirements for 2023-2025, we note that
this baseline is a hypothetical scenario because we have a statutory requirement to establish
volume requirements for each year.119 Moreover, the statute places a few key conditions on the
volume requirements for years after those established in the statute,120 and these conditions
would not permit the volumes to be set equivalent to the No RFS baseline. The No RFS baseline
volumes projected in this chapter would not meet the statutory requirement that the advanced
biofuel volume requirement be at least the same percentage of the total renewable fuel volume
requirement as in calendar year 2022.121 Nevertheless, the No RFS baseline is an appropriate
point of reference since it allows us to estimate the impacts of this action alone.
To project the No RFS baseline, we began by projecting renewable fuel use in the U.S. in
2023-2025 in the absence of RFS volume requirements for these years.122 We assumed that all
state mandates for renewable fuel use would continue, and that additional volumes of renewable
fuel would be used if these fuels could be provided at a lower price than petroleum-based fuels,
after taking into account available federal and state incentives. The differences between the
candidate volumes and the No RFS baseline represent the volume changes that we analyzed for
this rule. These volume changes, as detailed in Chapter 3, are the starting point for the analyses
presented in this document, except where noted.
118 Or, alternatively, if EPA established volume requirements at levels lower than what the market would have
supplied anyway.
119 CAA section 21 l(o)(2)(B)(ii).
120 See Preamble Section II.C.3.
121 CAA section 21 l(o)(2)(B)(iii). The ratio of advanced to total for the 2022 volume requirements is 0.273, while
the ratio of advanced to total for the No RFS baseline is 0.124 in 2023.
122 The analyses conducted to make this projection are described in Chapter 2.1.
62
-------
In some cases, the volume changes between the No RFS baseline and the candidate
volumes was sufficient to assess the impacts of the various factors enumerated in the statute. For
example, the GHG impacts and the costs are directly dependent on the volume of renewable fuel
used in the U.S. In other cases, however, these volume changes alone were insufficient and
potentially misleading. For example, the candidate volume for total domestic ethanol
consumption is 660-787 million gallons per year higher than under the No RFS baseline. This
projected volume increase could imply that additional ethanol production capacity and
distribution infrastructure would be needed to supply the candidate volumes. But total domestic
ethanol consumption in the candidate volumes for 2023-2025 is lower than total domestic
ethanol consumption achieved in previous years. Thus, no additional ethanol production capacity
or distribution infrastructure are projected to be needed to meet the candidate ethanol volumes
for 2023-2025. Where appropriate, such as in our assessment of infrastructure, we have therefore
considered not only the change in domestic renewable fuel consumption from the No RFS
baseline to the candidate volumes, but also other relevant factors as they exist in 2022.
There are some cases where we have insufficient information to project a No RFS
baseline for 2023-2025, such as U.S. crop production. U.S. crop production has an impact on a
number of the statutory factors, such as the projected conversion of wetlands, ecosystems, and
wildlife habitat, water quality, and water availability. At this time, we have insufficient
information to determine what U.S. crop acreage and production would be under a No RFS
baseline. One potential scenario is that total U.S. crop acreage and production would decrease in
2023-2025 if there was lower demand for crops for biofuel production. But other scenarios are
also possible and may be more likely. If demand for biofuel in the U.S. were lower in 2023-2025
in the absence of the RFS program, it is possible that biofuel exports would increase, and the
market would see little to no change in domestic biofuel production or biofuel feedstock crop
production. For instance, there have been significant exports of ethanol in recent years,123 and
both imports and exports of biodiesel and renewable diesel.124 Foreign markets may be able to
absorb additional renewable fuel exports from the U.S. Alternatively, domestic biofuel
production could decrease with little change in U.S. crop acreage and production if there is
sufficient demand for these crops in other markets, or production of crops used for biofuel
production could decrease and farmers could plant other crops on land previously used for
production of biofuel feedstocks. In cases where we have insufficient information to determine
what would happen under the No RFS baseline, we have used the most recent data available
(generally from 2021 or 2022) as a proxy for the No RFS baseline.
Finally, for our assessment of costs and fuel price impacts we have considered the
impacts of the candidate volumes relative to both the No RFS baseline and a 2022 baseline. We
recognize that the 2022 baseline may be of interest to the public as it gives an indication of
changes in volume requirements over time and how costs and fuel prices may change from
current levels as a result of this action. Nevertheless, we believe that the No RFS baseline better
represents the overall impacts of taking an action to establish volume requirements for 2023-
2025 versus not taking that action.
123 See Chapter 6.6.
124 See Chapter 6.2.4.
63
-------
2.1 No RFS Baseline
The No RFS baseline was derived from the relative economics of biofuels and the
petroleum fuels that those biofuels are blended into. If the blending cost of a biofuel is less than
the petroleum fuel that it is blended into, we assume that the biofuel would be used and displace
the respective petroleum fuel. The blending cost of a biofuel includes the value that the biofuel
has when blending it into the petroleum fuel. There are several components that must be
considered for each fuel:
• Production cost
• Distribution cost
• Blending value to the fuel blender (i.e., octane value and RVP cost of ethanol)
• Federal and state subsidies
• Relative energy value of the fuel, which may or not be a factor
• Cost to upgrade retail stations to enable them to offer the renewable fuel
These various cost components of each renewable fuel are added together to determine
the value of each fuel at the point that it is to be blended into petroleum fuel. For each renewable
fuel, the combination of these various cost components is represented using an equation that will
be described in each case.
There are many similarities between this No RFS baseline analysis and that of the cost
analysis described in Chapter 10, but there are differences as well. Table 2.1-1 summarizes the
various cost components considered for this analysis and provides comments how this analysis
differs from the cost analysis.
64
-------
Table 2.1-1: Comparison of No RFS Baseline Analysis to Cost Analysis
Included in No RFS
and Cost Analysis
Notes
No RFS
Cost
Production Cost
Yes
Yes
For the No RFS baseline, capital costs are
amortized using higher return on investment with
taxes, while cost analysis uses lower pre-tax return
on investment used for social analyses
Distribution Cost
Yes
Yes
Same
Blending Cost
Yes
Yes
Same
Fuel Economy
Cost
Yes
Yes
The cost analysis always accounts for fuel economy
cost, while the No RFS baseline only does so if it
impacts the value of the renewable fuel to fuel
blenders
Federal and State
Subsidies
Yes
No
The social cost analysis never takes subsidies into
account
Conducted on a
State-by-State,
Fuel Type-by-
Fuel Type Basis
Yes
No
While a national-average cost is sufficient for the
cost analysis, it was necessary to estimate the
economics of blending renewable fuel in individual
states that offer subsidies, and by fuel type, to
assess whether the renewable fuel would be
blended into each fuel in that state
For the No RFS baseline analysis, we use the latest projected feedstock prices (e.g., corn,
vegetable oil) for estimating the production costs for their associated fuels. For some renewable
fuels, the estimated volume under a No RFS scenario is projected to be significantly smaller than
under the RFS program. This result could in turn result in lower market prices for the agricultural
feedstocks, making the renewable fuels made from them more attractive. We did not evaluate
such a feedback mechanism. The various economic factors shown in Table 2.1-1 are further
discussed below for each renewable fuel.125
For the gasoline and diesel fuel prices, we use the most recent wholesale price projections
by the Annual Energy Outlook 2023. Since the Energy Information Administration models much
of the RFS program in its AEO modeling, some price impacts of the RFS program are likely
represented in these wholesale gasoline and diesel fuel prices. The RFS impact on the AEO
gasoline and diesel fuel prices will slightly bias the analysis conducted for the No RFS Baseline,
however, the impact is minimal and within the accuracy of the No RFS Baseline analysis.
2.1.1 Ethanol
By far the largest volume of ethanol blended into U.S. gasoline is produced from corn
and is mostly blended into gasoline at 10% (i.e., E10). However, some volume of ethanol is also
125 The spreadsheets used to estimate the No RFS baseline for corn ethanol "Corn Ethanol No RFS Baseline for SET
Final Rule" and biodiesel and renewable diesel "Biodiesel Renewable Diesel No RFS Baseline for SET final rule"
are available in the docket for this action.
65
-------
blended at higher blend percentages of 15% and 51-83% (i.e., E15 and E85, respectively).126
This section discusses the blending economics of ethanol and estimates the No RFS baseline for
all three of these ethanol fuel blends.
2.1.1.1 E10
The cost of blending ethanol into gasoline at 10% was analyzed by EPA in a peer
reviewed technical report.127 That report and its appendix provides both an historical review and
prospective analysis for the economics of blending ethanol into gasoline. The methodology used
in that analysis and its conclusion are summarized here.
A number of key factors were considered when evaluating the relative economics of
blending ethanol into gasoline. These factors depend on the type of gasoline the ethanol is
blended into, the season or year, and tax policies. Since ethanol is blended into gasoline at the
gasoline distribution terminal, it is most straightforward to consider those economic factors that
impact the decision to blend ethanol at that point. From that vantage point, the relative
economics of blending ethanol into gasoline—or the value of replacing ethanol in gasoline with
other components—can be summarized by the following equation:
EBCeio = (ESP + EDC - ERV-FETS - SETS) - GTP
Where:
• EBCeio is ethanol blending cost for E10
• ESP is ethanol plant gate spot price
• EDC is ethanol distribution cost
• ERV is ethanol replacement value
• FETS is federal ethanol tax subsidy
• SETS is state ethanol tax subsidy
• GTP is gasoline terminal price; all are in dollars per gallon
This equation allows us to break down these factors by year, by state, and by gasoline
type, enabling a detailed assessment of the relative blending economics of ethanol to gasoline
over time and by location. If the resulting ethanol blending cost is negative, it is assumed to be
cost-effective to blend ethanol. Since gasoline is marketed based on volume, not energy content,
the lower energy density of ethanol is not part of the ethanol blending cost equation. E10
contains about 3% less energy content than E0, and the cost of the lower energy content of the
gasoline is paid by consumers through lower fuel economy and more frequent refueling. Since
this small change in energy content is largely imperceptible to consumers and because gasoline
without ethanol is not widely available, refiners are able to price ethanol based on its volume
(unlike E85, for example, which must be priced lower at retail due to its lower energy density).
Thus, energy density is not a factor in this blending cost equation for E10. It is an important part
126 E85 (Flex Fuel), Alternative Fuels Data Center, https://afctc.energv.gov/fneis/ethanol e85.html
127 "Economics of Blending 10 Percent Corn Ethanol into Gasoline," EPA-420-R-22-034, November 2022.
66
-------
of assessing the overall social costs of ethanol use, but does not factor into the decision to blend
ethanol as E10.
Ethanol Plant Gate Spot Price (ESP)
We estimated future ethanol plant gate prices by gathering projected ethanol plant input
information (e.g., future corn prices projected by USDA and utility prices projected by EIA) to
estimate ethanol production costs that we presume represents plant gate prices. This is essentially
the same information used for estimating ethanol production costs for the cost analysis, except
that the capital costs are handled differently. Instead of amortizing the capital costs using a 7%
before tax rate of return on investment, capital costs are amortized using a 10% after tax return
on investment. As shown in Table 2.1.1.1-1, the capital amortization factor increases to 0.16
from 0.11 used for the cost analysis.
Table 2.1.1.1-1: Capital Amortization Factor Used for Estimating Plant Gate Spot Prices
Based on Production Costs
Depreciation
Life
Economic and
Project Life
Federal and
State Tax Rate
Return on
Investment
Resulting Capital
Amortization Factor
10 Years
15 Years
39%
10%
0.16
The year-by-year ethanol plant gate price projections are based on production costs and
are summarized in Table 2.1.1.1-2. There are two sets of ethanol price projections, one made by
the Energy Information Administration (EIA) and the second by the Food and Agricultural
Policy Research Institute (FAPRI), which are also summarized in Table 2.1.1.1-2.
Table 2.1.1.1-2: Projected Ethanol Plant Gate Prices
Year
Price ($/gal)
2023
2.45
2024
2.13
2025
1.89
Ethanol Distribution Cost (EDC)
This factor represents the added cost of moving ethanol from production plants to
gasoline distribution terminals, reflecting its different modes of transport (the gasoline terminal
prices in the equation already includes distribution costs). Because ethanol is primarily produced
in the Midwest and distributed longer distances to the rest of the country, the terminal price of
ethanol is usually lower in the Midwest than in other parts of the U.S. Ethanol distribution costs
were estimated for EPA on a regional basis, but to conduct the analysis on a state-by-state basis,
these costs were interpolated or extrapolated to estimate state-specific costs based on ethanol
spot prices.128 The estimated distribution costs for ethanol ranged from 110/gal in the Midwest to
290/gal when moved to the furthest distances along the U.S. coasts, and over 500/gal when
shipped to Alaska and Hawaii. The distribution cost to each state is summarized in Table 2.1.1.1-
3.
128 Modeling a "No-RFS" Case; refinery modeling conducted by Mathpro for EPA under ICF Contract EP-C-16-
020, July 17, 2018.
67
-------
Table 2.1.1
.1-3: Ethanol Distribution Cost by State
Region
States
Average Ethanol
Distribution
Cost (0/gal)
PADD 1
New York, Pennsylvania, West Virginia
28.7
District of Columbia, Connecticut, Delaware, Maryland,
Massachusetts, New Jersey, Rhode Island, Virginia
20.7
Georgia, South Carolina Vermont, New Hampshire,
North Carolina
22.7
Florida, Maine
28.8
PADD2
Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota,
Missouri, Nebraska, Ohio, South Dakota, Wisconsin
11.0
Kentucky, North Dakota, Oklahoma, Tennessee
20.7
PADD 3
Arkansas, Louisiana, Mississippi, Texas
15.5
Alabama, New Mexico
20.7
PADD 4
Colorado, Idaho, Montana, Utah, Wyoming
17.2
PADD 5
Oregon, Washington
21.4
Arizona, California, Nevada
25.4
Alaska, Hawaii
51.0
Ethanol Replacement Value (ERV)
Ethanol has properties that provide value (primarily octane) or cost (vapor pressure
impacts) when it is blended into gasoline. We use the term "ethanol replacement value" to refer
to the sum of the costs due to these properties, including properties that increase and decrease
ethanol's blending value. Depending on where and when the ethanol is used, the ethanol
blending value is an important consideration when gasoline production is modified to take into
account the subsequent addition, or potential removal, of ethanol.
Essentially all E10 blending in the U.S. now occurs by "match-blending," where the base
gasoline ("gasoline before oxygenate blending" or BOB) is modified to account for the
subsequent addition of ethanol, in which the blending value of ethanol is important. In RFG
areas, refiners produce a reformulated gasoline before oxygenate blending (RBOB) that has both
a lower octane value and lower RVP tailored to still meet the RFG standards after the addition of
ethanol. This has been typical for ethanol-blended RFG since the mid-1990s. As the use of
ethanol expanded into CG areas, a similar match-blending process began to be used there as
well, replacing splash-blending. In these areas, a conventional gasoline before oxygenate
blending (CBOB) is produced by refiners for match-blending with ethanol. CG is also adjusted
to account for the octane value of ethanol, but unlike RFG, most CG is not adjusted for RVP due
to a 1-psi RVP waiver provided for E10 in most locations. When RBOB and CBOB are
produced, the refiner makes the decision that ethanol will be blended into their gasoline since the
BOBs cannot be sold as finished gasoline without adding 10% ethanol, but the ethanol is still
blended into the gasoline at the terminal.129 It is likely that refiners make their decision on
129 The exception to this is a small amount of premium grade BOB that is sold as regular or midgrade E0.
68
-------
producing BOBs based on the economics of producing finished gasoline at terminals. In the case
of such match blends, the economic value of ethanol relative to gasoline includes a consideration
of not only its value on a volumetric basis as a substitute for gasoline, but also the blending value
of ethanol resulting from its higher octane, and in some cases, its impact on volatility.
The full value of ethanol is best reflected by the cost associated with meeting all of the
gasoline standards and requirements through some means other than blending ethanol, including
any capital costs to produce ethanol replacements. To assess this, ICF conducted refinery
modeling for EPA for removing ethanol from the gasoline pool.130 After aggregating the refinery
cost modeling results—which account for the octane value and volatility of ethanol, as well as
replacing its volume—the replacement costs of ethanol in regular grade CG and RFG are
summarized in Table 2.1.1.1-4. The ethanol replacement costs were estimated based on a certain
set of modeling conditions—projected prices for the year 2020 with crude oil priced at $72/bbl.
The economics for replacing ethanol, however, would be expected to vary over time based on
changing market factors, such as the market value of RVP control costs, crude oil prices, and
particularly the market value for octane. The ethanol replacement costs were adjusted for the
years analyzed under the No RFS baseline based on crude oil prices, which likely provides a
reasonable estimate of how refiners would value the octane, RVP, and other replacement costs of
ethanol over time.
Table 2.1.1.1-4 Ethanol Replacement Value (J
>/gal)
Gasoline Type
Gasoline Grade
Year
2023
2024
2025
Conventional
Gasoline
Summertime Regular
2.10
2.12
1.98
Summertime Premium
1.58
1.60
1.50
Reformulated
Gasoline
Summer Regular
1.81
1.83
1.71
Summer Premium
1.30
1.31
1.23
Conventional and
Reformulated
Winter Regular
0.85
0.86
0.80
Winter Premium
0.65
0.65
0.61
Federal and State Ethanol Tax Subsidies (FETS and SETS)
The federal ethanol blending tax subsidy expired in 2011, so it did not figure into the No
RFS baseline analysis. Various state tax subsidies, however, have been provided for the use of
ethanol. These tax subsidies incentivize the blending of ethanol into the gasoline pool and
directly impact the decision of whether to use ethanol. Iowa and Illinois offer an ethanol
blending subsidy of 250/gal and 290/gal, respectively.131 The California LCFS program is
estimated to provide ethanol a blending credit of 330/gal in 2019.132>133 Several states also have
130 The results of this refinery modeling are summarized in Chapter 10.1.3.1.1; Analysis of the Effects of Low-
Biofuel Use on Gasoline Properties; An Addemdum to the "No-RFS" Study; refinery modeling study conducted by
Mathpro for EPA under ICF Contract EP-C-16-020; June 7, 2019.
131 States' Biofuels Statutory Citations; The National Agricultural Law Center; University of Arkansas,
https://nationalaglawcenter.org/state-compilations/biofneis.
132 California Air Resources Board (CARB), Fuel Pathway Table; LCFS Pathway Certified Carbon Intensities;
https://ww2.arb.ca.gOv/resonrces/docnments/lcfs-pathwav-certi:fied-carbon-intensities.
133 Weekly LCFS Credit Transfer Activity Reports; California Air Resources Board;
https://ww3.arb.ca.gov/fiiels/lcfs/credit/lrtweeklvcreditreports.htm.
69
-------
ethanol use mandates that require the use of ethanol regardless of the economics for doing so.134
These mandates cannot be factored into the ethanol blending cost equation, but are accounted for
in EPA's overall analysis by including the ethanol volume in gasoline in these states regardless
of the blending economics. Other federal and state subsidies—such as ethanol production
subsidies, loan guarantees, grants, and any other subsidies—were not considered by this analysis.
Gasoline Terminal Price (GTP)
Refinery rack price data from 2018—which already included the distribution costs for
moving gasoline to downstream terminals—were used to represent the price of gasoline to
blenders on a state-by-state basis.135 However, these prices were not projected for future years.
Instead, we used projected refinery wholesale price data from AEO 2023 to adjust the 2018
refinery rack price data to represent gasoline rack prices in future years. We used 2018 data
instead of the most recent data to avoid abnormal pricing effects caused by the COVID-19
pandemic or the subsequent supply issues that emerged when the pandemic was subsiding. This
gasoline price data, summarized in Table 2.1.1.1-5, was collected for each states and is assumed
to represent the average gasoline price for all the terminals in each state.136
134 States' Biofuels Statutory Citations; The National Agricultural Law Center; https://nationalagtawcenter.org/state~
co mpilat ions/b iofuels.
135 EIA; Spot Prices; https://www.eia.gov/dnav/pet/pet pri spt s i a.htm.
136 EIA; Prime Supplier Sales Volume; https://www.eia.gov/dnav/pet/pet cons prim den tins m.htm.
70
-------
Table 2.1.1.1-5: Gasoline Terminal Prices in 2019 ($/gal)a
Gasoline Grade
Gasoline Grade
State
Regular
Premium
State
Regular
Premium
Alaska
2.37
2.44
Montana
1.84
2.30
Alabama
1.68
2.11
North Carolina
1.69
2.07
Arkansas
1.70
2.03
North Dakota
1.77
2.18
Arizona
2.00
2.29
Nebraska
1.74
2.55
California
2.37
2.61
New Hampshire
1 Si)
2 oo
Colorado
1.85
2.26
New Jersey
1.72
2.91
Connecticut
1.77
: 00
New Mexico
1.82
2.18
DC.
1.70
:.
-------
Table 2.1.1.1-6: National Average Gasoline Prices
Year
Price
Actual National
2018
$1.98
Average Gasoline Price
AEO 2023 Projected
2023
$2.81
National Average
2024
$2.48
Gasoline Prices
2025
$2.26
The No RFS Baseline analysis revealed that it is economic to blend ethanol into the entire
gasoline pool up to 10%. As shown in Figure 2.1.1.1-1, ethanol is over 400/gal less expensive
than gasoline in the most expensive market for blending ethanol, and about $2/gal less expensive
than gasoline in the least expensive market for blending ethanol (in which a state subsidy
applies).
Figure 2.1.1.1-1: Economics of Blending Ethanol up to the E10 Blendwall
O)
en
0.00
-40.00
(T3
CtQ
U
(3 -80.00
&D
c
-120.00
c
-------
which affects ethanol's value to fuel blenders. This E85 fuel pricing effect is captured in a
breakeven price for ethanol.
The economics for using ethanol in E85 is estimated in two steps. First, we estimated the
breakeven price for ethanol blended in E85 based on the price of gasoline price in each state.
This calculation is made for regular and premium grades of both CG and RFG in each state. In
the second step, the estimated ethanol plant gate price, ethanol distribution cost, retail cost, and
E85 subsidies are combined together in the following equation to estimate whether ethanol
blended into E85 is economical:
EBCess = (ESP + EDC - FETS - SETS + RC) - EBBV
Where:
• EBCess is ethanol breakeven price for ethanol blended as E85
• ESP is ethanol plant gate spot price
• EDC is ethanol distribution cost
• FETS is federal ethanol tax subsidy
• SETS is state ethanol tax subsidy
• RC is retail cost (service station revamp to sell E85)
• EBBV is ethanol breakeven blending value; all are in dollars per gallon
Ethanol Replacement Value (ERV)
Blending ethanol into gasoline for E85 is different than blending for E10 because refiners
do not make a separate E85 BOB; thus, the E10 RBOBs and CBOBs are blended with ethanol to
produce E85 and there is significant octane giveaway. Conversely, there is no risk that the E85
blend will exceed any RVP limits because E85 has a very low RVP. In fact, the resulting E85
blend is so low in vapor pressure that it causes most E85 blends to not meet the RVP minimum
standards. In those cases, E85 is blended with less ethanol—usually 70% in the winter and up to
79% in the summer—and the year-round average is 74%, which allows ethanol to comply with
the ASTM RVP minimum standards.137
Although refiners do not create a lower octane BOB for blending into E85, ethanol
producers nonetheless saw the opportunity to blend natural gas liquids (NGLs) with ethanol to
produce E85. NGLs are a low cost, low octane, higher RVP petroleum blending material that
ethanol producers use to denature their ethanol. Since ethanol plants already have this blendstock
material on hand, they blend E85 on-site using NGLs and then distribute the finished E85 from
there. When blending up E85 with NGLs, the higher RVP of the NGLs allows blending a higher
ethanol content of 83% in the summer. However, the RVP of NGLs is about the same or slightly
higher than winter gasoline, so the winter blend percentage is the same. Because the more
volatile NGLs are smaller hydrocarbons, they contain lower volumetric energy content, which is
137 ASTM D5798-21, Standard Specification for Ethanol Fuel Blends for Flexible-Fuel Automotive Spark-Ignition
Engines.
73
-------
a factor in considering their value as well. Because NGLs are used as an E85 blendstock, we also
evaluated the economics of blending E85 blended with NGLs.
Federal and State Ethanol Tax Subsidies (FETS and SETS)
There is no federal ethanol blending tax subsidy for E85. Various state tax subsidies,
however, have been provided for the use of ethanol. These tax subsidies incentivize the blending
of ethanol into the gasoline pool and directly impact the decision of whether to use ethanol.
Table 2.1.1.2-1 provides the E85 subsidies offered by different states.
Table 2.1.1.2-1: State E85 Subsidies
State
E85 Subsidy (0/gal)
New York
53
Pennsylvania
25
Iowa
16
South Dakota
14
Kansas
12.5
Michigan
11
The California and Oregon LCFS blending credits for ethanol apply when ethanol is
blended into E85 as well (Oregon's blending credit is assumed to be the same as California's).
The blending credit applies to E85, so its credit is amortized over the ethanol portion of E85 to
assess the blending value of ethanol. Aside from the retail cost credit offered by USD A described
below, other federal and state subsidies—such as ethanol production subsidies, loan guarantees,
grants, and any other subsidies—were not considered by this analysis.
Retail Cost (RC)
The retail costs for E85 are estimated based on the investments needed to offer E85 at
retail stations and the estimated throughput at E85 stations.138 We estimated the total cost for a
typical retail station revamp to enable selling E85 to be $50,300 and that these stations sell on
average 39,000 gallons of E85 per year. When amortizing this capital cost over the gallons of
E85 sold, the total cost of the revamp adds 21.90/gal to the cost of blending ethanol into E85
(accounting only for the 64% of ethanol in E85 above the ethanol in E10).
Ethanol Breakeven Blending Value (EBBV)
There are downstream pricing effects for E85 that require the economics of E85 be
assessed differently when blending ethanol into E85 compared to blending ethanol into E10.
These downstream pricing effects exist because E85 contains less energy content compared to
E10—22% and 30% less when blended with gasoline and NGLs, respectively. This lower energy
density of E85 is noticeable to consumers in their fuel economy, so they demand a lower price at
retail stations, which therefore requires that the economics of E85 be assessed at retail. Price
138 The methodology used and the estimated costs for these revamps are discussed in Chapter 10.1.4.1.2.
74
-------
information collected for E85 shows that it is typically priced 16% lower than E10 at retail.139140
For the No RFS analysis, we assumed that gasoline-blended E85 is priced 16% lower than E10
and that NGL-blended E85—which has much lower volumetric energy content—is priced 21%
lower than E10.
Figures 2.1.1.2-1 and 2 show how the breakeven price for ethanol is estimated for E85
when blended with gasoline and NGLs, respectively, using the example of regular grade CG sold
in Pennsylvania and Missouri. At the top of each figure, the pricing of gasoline is shown from
terminal to retail, depicting the price impacts when distribution costs and taxes are added on. At
the bottom of each figure, the pricing of E85 is shown when blended with gasoline and NGLs,
respectively. The E85 prices are then estimated at the terminal after the retail, tax, and
distribution costs are subtracted from the retail prices. Finally, the ethanol breakeven price is
estimated for the ethanol blended into E85 based on the price of gasoline at the terminal and the
fraction of gasoline and ethanol in E85.
Figure 2.1.1.2-1: Example Calculations for Ethanol Breakeven Price for Gasoline-Blended
E85
2023 Gasoline and E85 Pricing, and Ethanol Breakeven Price
Gasoline Blendstock, Two different State Tax Rates - Crude oil Priced $63/bbl
Gasoline Pricing
RETAIL PRICE
PA 279 c/gal
MO 239 c/ea|
RETAIL PRICE
PA 234 c/gal for ESS
MO 200 c/gal for E85
279 c/gal
X 0.84 FE effect
= 228 c/gal
12 c/gal
12 c/gal
TAX
76 c/gal
36 c/gal
10 c/gal
76 c/gal
36 c/gal
10 c/gal
181 c/gal
181 c/gal
Ethanol
Breakeven
Pricing
136 c/gal E85 120 c/gal Ethanol
142 c/gal ESS 128 c/gal Ethanol
136 c/gal = 0.26*181 + 0,74*120
Ethanol would have to be priced at 120/128 c/gal or
less to be attractive to refiners
139 Retailing E85: An Analysis of Market Performance, July 2014 - August 2015; Fuels Institute;
https://www.fuelsinstitute.org/Researcli/Reports: March 23, 2017.
1411 AAA Gas Prices; https://gasprices.aaa.com: downloaded June 15, 2022.
75
-------
Figure 2.1.1.2-2: Example Calculations for Ethanol Breakeven Price for NGL-Blended E85
2023 Gasoline and E85 Pricing, and Ethanol Breakeven Price
NGL Blendstock ($1.26/gal), Two different State Tax Rates - Crude oil Priced $63/bbl
Gasoline Pricing
RETAIL PRICE
PA 279 c/gal
MO 239 c/gal
RETAIL PRICE
PA 220 c/gal for E85
MO 188 c/gal for E85
279 c/gal
X 0.79 FE effect
= 220 c/gal
12 c/gal
12 c/gal
TAX
76 c/gal
36 c/gal
ES5 Pricing
TAX
76 c/gal
36 c/gal
10 c/gal
10 c/gal
181 c/gal
181 c/gal
Ethanol
Breakeven
Pricing
122 c/gal E85 121 c/gal Ethanol
130 c/gal E85 131 c/gal Ethanol
122 c/gal = 0.24*126 + 0,76*121
Ethanol would have to be priced at 121/131 c/gal or
less to be attractive to refiners
Figure 2.1.1.2-1 shows that when the E85 is blended with gasoline, the breakeven price
of ethanol in E85 is 1200/gal and 1290/gal, which is 510/gal and 600/gal lower than the gasoline
price, depending on whether the state gasoline tax is high (Pennsylvania) or low (Missouri),
respectively. Similarly, Figure 2.1.1.2-2 shows that when the E85 is blended with NGLs, the
breakeven price of ethanol in E85 is 1210/gal and 1310/gal, which is 500/gal and 600/gal lower
than the gasoline price, depending on whether the state gasoline tax is high (Pennsylvania) or
low (Missouri), respectively. A list of gasoline tax rates by state (including all federal and state
taxes) is provided in Table 2.1.1.2-2.
76
-------
Table 2.1.1.2-2: Gasoline Tax Rates by State (Includes Federal and State Taxes; ^/gal)
State
Tax Rate
State
Tax Rate
Alaska
27
Montana
51
Alabama
45
North Carolina
55
Arkansas
43
North Dakota
41
Arizona
37
Nebraska
49
California
79
New Hampshire
42
Colorado
40
New Jersey
60
Connecticut
61
New Mexico
37
DC
42
Nevada
52
Delaware
41
New York
63
Florida
61
Ohio
57
Georgia
49
Oklahoma
38
Hawaii
67
Oregon
54
Iowa
49
Pennsylvania
76
Idaho
51
Rhode Island
54
Illinois
58
South Carolina
47
Indiana
65
South Dakota
48
Kansas
49
Tennessee
67
Kentucky
44
Texas
38
Louisiana
38
Utah
50
Massachusetts
45
Virginia
39
Maryland
55
Vermont
49
Maine
48
Washington
68
Michigan
46
Wisconsin
51
Minnesota
47
West Virginia
54
Missouri
36
Wyoming
42
Mississippi
37
As for E10, if the ethanol blending cost is negative, ethanol is considered economical to
blend E85 in comparison to gasoline; if it is positive, it is not economical. Figure 2.1.1.2-3
provides some key results of the No RFS baseline analysis for E85, showing a range in blending
values for ethanol in E85, which vary from economic to blend to not economic to blend. For the
highest cost market for E85, ethanol is priced 80-900/gal higher than its breakeven price. But for
lowest cost market for E85, ethanol is around 500/gal lower than its breakeven price. It is
important to understand which gasoline in which states are economically attractive to E85 since
this determines the potential market size.
77
-------
Figure 2.1.1.2-3: Economics of Blending Ethanol in E85
Range in Ethanol Blending Cost in E85
E85 priced 16% lower than E10
200
ru
do
u 150
-------
sold in New York is likely quite small. New York has about 80 retail stations that sell E85. If we
assume that the average volume of E85 sold by these retail stations is the same as that sold by
E85 stations nationwide, then New York retail stations would sell 3.1 million gallons of E85 per
year under the No RFS baseline.
2.1.1.3 E15
The analysis for estimating the E15 baseline has similarities with how both E10 and E85
were estimated. Of the variables in the ethanol blending cost equation in Chapter 2.1.1.1, Ethanol
Plant Gate Spot Price (ESP), Ethanol Distribution Cost (EDC), and Gasoline Terminal Price
(GTP) are again the same. Like for E85, an additional cost applies to El5 to account for the cost
to modify retail stations to carry El 5 and we believe that Ethanol Replacement Value (ERV)
does not apply as well, although we keep as a term and explain the possibility below for how it
could apply.
The economics to determine whether ethanol blended into El 5 is economical is estimated
by combining the ethanol plant gate price, ethanol distribution cost, ethanol replacement cost,
and retail cost in the following equation:
EBCeis = (ESP + EDC - ERV-FETS - SETS + RC) - GTP
Where:
• EBCeis is ethanol blending cost for El5
• ESP is ethanol plant gate spot price
• EDC is ethanol distribution cost
• ERV is ethanol replacement value
• FETS is federal ethanol tax subsidy
• SETS is state ethanol tax subsidy
• RC is retail cost (service station revamp to sell El 5)
• GTP is gasoline terminal price; all are in dollars per gallon
Ethanol Replacement Value (ERV)
Blending ethanol into gasoline for E15 is different than blending for E10 because we
believe that refiners do not make a separate E15 BOB; thus, E10 BOBs are blended with ethanol
to produce El 5, in which case there is octane giveaway and no blending value to refiners for
ethanol. It is possible, though, that some refineries with extra gasoline storage tanks could blend
an El 5 BOB to sell off their refinery racks; however, we have no knowledge of this currently
happening, Similarly, there should be no RVP cost for blending ethanol above that of E10
because ethanol-gasoline blends reach a maximum RVP at 10%.
A larger issue for E15 is that it does not receive a 1-psi waiver like E10 does in the
summer, which means that ethanol cannot be blended into E10 to produce El5 without either
exceeding summer RVP limits or incurring an additional cost. However, as discussed in Chapter
1.7.2, E15 did receive a regulatory 1-psi waiver for 2019-2021 and EPA-issued emergency fuel
79
-------
waivers throughout the summer of 2022 and now again into 2023 which further allowed El 5 to
take advantage of the 1-psi waiver. A number of Midwestern states petitioned EPA to remove the
1-psi waiver for E10. If the E10 1-psi waiver were to be removed in those states, a new lower
RVP, higher-cost BOB would be required for both E10 and E15, which would remove the hurdle
for selling El 5 in the summer months in those states. EPA proposed in a rulemaking to grant
those states' request to remove the 1 psi waiver for E10 starting in 2024;142 however, that
rulemaking would need to be finalized for that proposed action to take place.143
Federal and State Ethanol Tax Subsidies (FETS and SETS)
There is no federal nor state ethanol blending tax subsidy for El5. It is important to know
that California does not allow the sale of El 5. Other federal and state subsidies—such as ethanol
production subsidies, loan guarantees, grants, and any other subsidies—were not considered by
this analysis.
Retail Cost (RC)
The retail costs for El 5 are estimated based on the investments needed to offer El 5 at
retail stations and the estimated throughput at El 5 stations.144 We estimated the total cost for a
typical retail station revamp to enable selling E15 to be $133,000, and that these stations sell on
average 180,000 gallons of E15 per year. When amortizing this capital cost over the gallons of
E15 sold, the total cost of the revamp adds 2030/gal to the cost of blending ethanol into E15
(accounting only for the 5% of ethanol in E15 above the ethanol in E10).
E15 has different properties than E10 that allow it to be priced differently than E10. E15
has higher octane than E10, so the fuels industry could set E15 prices higher on that basis.
Conversely, El 5 has lower energy density than E10, which means that consumers are not able to
drive the same distance on a tankful of E15. The website e85prices.com, which collects
information on gasoline and ethanol-gasoline blend prices, reported that El5 is priced 8.50/gal
cheaper than E10. A conversation with a gasoline retail marketer explained that when beginning
to offer E15 for sale, marketers will typically price it lower than E10 as a means to promote E15
to consumers and increase its sales. If E15 is priced 80/gal lower than E10, it adds 1600/gal
(8/0.05) to the blending cost for blending ethanol into E15. However, if this is a marketing
strategy, this practice would likely diminish over time. We do not know what the ultimate price
of E15 will be relative to E10 since many retail station owners only began to offer E15 in recent
years. To maximize their profit, retail station owners will seek the optimal El5 price that
balances sales volume and pricing. For this analysis, we assumed that El 5 is priced lower than
E10 consistent with how E85 is priced.145 Since E15 contains 1.8% less energy than E10, we
assumed that E15 is priced 1.2%, or about 30/gal, less than E10.
142 88 FR 13758, March 6, 2023.
143 Providing El5 with a 1-psi waiver or removing the E10 1-psi waiver—either of which would allow El5 to use
the same BOB as E10—would simply remove a logistical barrier to the use of E15 during summer months.
However, E15 use under a No RFS Baseline would still be governed by the relative economics of blending
additional ethanol into E10 gasoline relative to continuing to use petroleum gasoline.
144 The methodology used and the estimated costs for these revamps are discussed in Chapter 10.1.4.1.2.
145 E85, which contains 74% ethanol and 21% less energy than E10, is typically priced 16% lower than E10.
80
-------
Similar to E10, if the ethanol blending cost is negative, then ethanol is considered
economic to blend into gasoline to produce El 5, while it would not be economic if the value is
positive. Figure 2.1.1.3-1 provides some key results of the No RFS baseline analysis for E15,
showing a range in blending values for ethanol in El5, which vary from economic to blend to not
economic to blend.
Figure 2.1.1.3-1: Economics of Blending Ethanol in E15
Range in Ethanol Blending Costs in E15
E15 priced 1% lower than E10
o
cj
2025
Year
- High Cost E15 - No Retail Cost
¦ Low Cost E15 - No Retail Cost
- Avg Cost E15 w/1/2 Retail Cost
• Avg Cost E15 - No Retail Cost
Low Cost E15 exc Prem & Summer CG
It is important to recognize the cost impact due to revamping the retail station to enable it
to sell E15. Assuming a typical retail station revamp cost of $132,000, and that the HBIIP
program subsidized half the cost, the retail station is estimated to need to cover a cost of about
$ 1/gal for that 5% increment of ethanol in El5. This is shown in Figure 2.1.1.3-1 as the
difference between the dashed blue line and the solid blue line, which represents the average El5
cost without any retail cost included. None of the solid lines in the figure include this retail
revamp cost; adding in this retail cost component immediately makes every gasoline market
uneconomic for blending additional ethanol into E10 to produce El5.
Assuming a best-case scenario in which a retail station was able to secure an additional
local subsidy that covered the balance of the El 5 revamp cost, then the lowest cost market for
the additional 5% of ethanol in El 5 would have about a -400/gal blending cost for ethanol.
However, similar to E85, this gasoline market is comprised primarily of premium gasoline, and
in a few cases summertime regular grade gasoline. Since the premium gasoline market is very
small, the uncertainty about summertime blending would likely dissuade retailers from wanting
to sell El5. The next most economic gasoline market includes regular grade gasoline, but its
ethanol blending cost is more than 200/gal in 2023 to 2025. If refiners and terminal operators
could overcome the steep logistical hurdles of producing and moving a separate El 5 BOB to
terminals and eventually to retail stations, the gained ethanol replacement value described above
in the discussion about the economics of E10 ethanol for the E15 BOB would more than offset
the retail cost of making El 5 available, and El 5 could be economical in some summertime
regular gasoline markets. However, we don't believe that refiners and terminal operators would
create a separate El5 BOB due to the significant hurdles until the point that El5 becomes a
81
-------
significant portion of the gasoline market. Thus, the ethanol blending cost analysis finds the
gasoline market uneconomical for El 5 in the absence of the RFS program.
After reviewing the El5 blending economics, we project that without the RFS program in
place, the fuels market would not offer El 5 for sale.
2.1.2 Cellulosic Biofuel
The primary type of cellulosic biofuel that we project will generate appreciable quantities
of cellulosic RINs in 2023-2025 is CNG/LNG derived from biogas. We also project that some
volumes of liquid cellulosic ethanol from corn kernel fiber (CKF) will be produced in these
years. Cellulosic biofuels generally cost more to produce than the fossil fuels they displace, and
therefore generally would not be used absent the incentives provided by the RFS program. There
are, however, state incentive programs (e.g., the California and Oregon LCFS programs) that we
project would be sufficient to incentivize the use of some types of cellulosic biofuels without the
additional incentives provided by the RFS program. Furthermore, it is our expectation that the
majority of ethanol produced from corn kernel fiber will be produced concurrent with and in the
same process as ethanol produced from the corn starch. Thus, the economics for blending ethanol
from corn kernel fiber are the same as those of corn starch ethanol. This section describes our
projections of cellulosic biofuel use for the No RFS baseline.
2.1.2.1 CNG/LNG Derived from Biogas
As described in greater detail in Chapter 10, CNG/LNG derived from biogas is generally
more expensive to produce than natural gas. Because of this higher cost, and because of the
demand for renewable natural gas (RNG) in sectors other than the transportation sector, we
project that without incentives for the use of renewable CNG/LNG in the transportation sector,
very little or none of this fuel would be used in the transportation sector.
There are, however, two state LCFS programs (California and Oregon) that currently
offer incentives for the use of CNG/LNG in the transportation sector. We have assumed that the
incentives provided by these states would be sufficient for some quantity of CNG/LNG to be
used in the transportation sector in the absence of the RFS program. To project the quantity of
CNG/LNG used as transportation fuel in these states (including both fossil natural gas and
RNG), we have used data provided by California and Oregon and extrapolated the use of these
fuels through 2025. Specifically, we calculated a year-over-year growth rate for each year for
California and Oregon separately. We then calculated the average observed annual growth rate
from 2015-2021 for California and 2017-2019 for Oregon146 to determine an average annual
rate of growth and used this growth rate to project CNG/LNG volumes in California and Oregon
through 2025. This growth rate was applied to the reported use of renewable CNG/LNG in the
transportation sector in 2021, the latest year for which data were available at the time this
analysis for the No RFS baseline was completed. We assumed that all CNG/LNG used as
transportation fuel in these states in 2023-2025 was from renewable sources and did not include
the growth rate in 2020 due to the impacts of the COIVD-19 pandemic. The projected volume of
146 The Oregon LCFS program began in 2016, and therefore we cannot calculate an annual rate of growth prior to
2017.
82
-------
renewable CNG/LNG used as transportation fuel in California and Oregon under the No RFS
baseline is summarized in Table 2.1.2.1-1.
Table 2.1.2.1-1: CNG/LNG Derived from Biogas in the No RFS Baseline (million ethanol-
equivalent gallons)
State
Annual Growth Rate
2021
2022
2023
2024
2025
California
4.3%
303
316
330
344
359
Oregon
20.3%
4
5
6
7
8
Total
N/A
307
321
336
351
367
2.1.2.2 Liquid Cellulosic Biofuels
In recent years there have been very small quantities of liquid cellulosic biofuels
produced. This is despite the fact that the combination of the RFS program, federal tax credit,
and state incentives (e.g., the California LCFS program) have provided very large financial
incentives for liquid cellulosic biofuels. While the incentives provided by state programs and the
federal tax credit are expected to continue in future years, we do not expect that these incentives
alone will be sufficient to support most types of liquid cellulosic biofuel production in 2023-
2025.
The one likely exception is ethanol produced from CKF at existing ethanol production
facilities. Many corn ethanol producers have claimed that their existing facilities are capable of
producing ethanol from CKF, in some cases with the addition of cellulose enzymes and in other
cases using only the natural occuring enzymes from the corn kernel. In either case, we project
that the cost of producing ethanol from CKF would be the same as, or only slightly more
expensive than, producing ethanol from corn starch. Because ethanol produced from CKF has
the potential to receive greater incentives through programs such as California's LCFS we
project that the production of ethanol from CKF in the No RFS baseline would be equal to the
production of this fuel with the volumes we are finalizing in this rule.
2.1.3 Biomass-Based Diesel
2.1.3.1 Biodiesel
Estimating the economics of blending biodiesel is different than ethanol because, unlike
corn ethanol plants that are almost exclusively located in the Midwest, biodiesel plants are more
scattered around the country. The more diffuse location of biodiesel plants affects how we
estimate distribution costs for using biodiesel. Also, refiners do not change the properties of the
diesel they produce to accommodate the downstream blending of biodiesel, and as such there is
no additional blending value associated with its use like there is for E10. However, blending
biodiesel does often require the addition of additives to accommodate some of its properties. The
blending cost of biodiesel is estimated using the following equation:
83
-------
BBC = (BSP + BDC - FBTS - SBTS) - DTP
Where:
• BBC is biodiesel blending cost
• BSP is biodiesel plant gate spot price
• BDC is biodiesel distribution cost
• FBTS is federal biodiesel tax subsidy
• SBTS is state biodiesel tax subsidy
• DTP is diesel terminal price; all are in dollars per gallon
Biodiesel Plant Gate Spot Price (BSP)
USDA collects biodiesel plant gate pricing data, which is the price paid to biodiesel
producers when they sell their biodiesel; however, USDA does not project future biodiesel
prices.147 Instead, we assumed that biodiesel production costs reflected plant gate prices and then
estimated biodiesel production costs based on future vegetable oil and utility prices. This is
essentially the same information used for estimating biodiesel production costs for the cost
analysis in Chapter 10, except that the capital costs are amortized using the capital amortization
factor in Table 2.1.1.1-1. The resulting projected biodiesel plant gate prices are summarized in
Table 2.1.3.1-1 and we also list the Food and Agricultural Policy Research Institute (FAPRI)
biodiesel price projections for comparison, although the FAPRI price projections are not used in
our No RFS baseline analysis.
Table 2.1.3.1-1: Projected Biot
iesel Plant Gate Prices ($/ga
Projected Production Cost
2023
2024
2025
Soybean Oil
5.76
4.88
4.39
Corn Oil
4.85
4.13
3.72
Waste Oil
4.46
3.81
3.44
FAPRI (for comparison only)
5.49
5.12
4.99
Biodiesel Distribution Cost (BDC)
This factor represents the added cost of moving biodiesel from production plants to
terminals where it is blended into diesel. Unlike ethanol, which is almost exclusively produced in
the Midwest and distributed elsewhere from there, biodiesel is predominantly produced in the
Midwest, but there are also biodiesel plants dispersed around the country. For this reason, we
took a very different approach for this analysis. Using 2019 EIA data, we estimated the quantity
of biodiesel produced within each PADD, the movement of biodiesel between PADDs, and the
147 USDA Economic Research Service; US Bioenergy Statistics. 2021. Table 17 Biodiesel and Diesel Prices by
month.
84
-------
imports and exports of biodiesel into and out of each PADD, as summarized in Table 2.1.3.1-
2 148,149,150,151,152
Table 2.1.3.1-2: Biodiesel Production, Imports, Export, and Movement Between PADDs
and Consumption in 2019 (million gallons)
From
From
Other
PADD
Production
Imports
Exports
PADD 2
PADD 3
Movement
Consumption
PADD 1
88
82
7
103
4
0
271
PADD 2
1,166
42
60
-
0
-363
785
PADD 3
337
12
14
140
-
115
450
PADD 4
0
9
0
21
0
5
35
PADD 5
134
22
32
99
22
-6
239
Total
1,725
168
114
363
26
-249
1,779
ICF estimated the distribution costs for distributing biodiesel both within and between
PADDs, as summarized in Table 2.1.3.1-3.153
Table 2.1.3.1-3: Biodiesel Distribution Costs (Wgal)
PADD
Within PADD
From Outside the PADD
PADD 1
15
35
PADD 2
15
15
PADD 3
15
18
PADD 4
15
25
PADD 5
15
32
As expected, distribution costs for distributing biodiesel within a PADD are less than
when the biodiesel is distributed further away from outside the PADD. Since imports come from
outside the PADD, we used outside the PADD values for imports. Comparing these biodiesel
distribution costs to ethanol, distributing biodiesel is expected to be more expensive, which
recognizes that the larger volume of ethanol provides the opportunity to optimize the distribution
system more so than biodiesel. For example, the greater volume of ethanol allows for greater use
of unit trains and more streamlined logistics overall. Like for ethanol, distribution costs of
148 Petroleum Administration for Defense District (PADD): The 50 U.S. states and the District of Columbia are
divided into five districts. Each PADD comprise a subset of U.S. states; PADD 1: Eastern states; PADD 2: Midwest
states; PADD 3: Gulf Coast; PADD 4: Rocky Mountain States; PADD 5: Pacific Coast states.
149 Table 5; Biodiesel (B100) production by petroleum administration for defense district (PADD); US Monthly
biodiesel production Report; Energy Information Administration; https://www.eia.gov/biofuels/biodiesel/production
150 Exports; Petroleum and Other Liquids; Energy Information Administration;
https://www.eia.gov/dnav/pet/pet move exp dc NUS-ZOO rnbbt m.fatm (some values were updated since this data
was downloaded which would have a negligible impact on the analysis)
151 Imports by Area of Entry; Petroleum and Other Liquids; Energy Information Administration;
https://www.eia.gov/dnav/pet/pet move imp dc NUS-ZOO mbbl m.fatm (some values were updated since this data
was downloaded which would have a negligible impact on the analysis)
152 Movements by Pipeline, Tanker, Barge, and Rail between PAD Districts; Petroleum and Other Liquids Energy
Information Administration; https://www.eia.gov/dnav/pet/pet_move ptb dc R20-R10_mbbl_m.htm
153 Modeling a "No-RFS" Case; refinery modeling conducted by Mathpro for EPA under ICF Contract EP-C-16-
020, July 17, 2018.
85
-------
biodiesel to the East and West Coasts are higher compared to distribution in the Midwest where
most of the biofuels are produced. Although the Rocky Mountain states are located much closer
to the Midwest, it is expensive to distribute biodiesel to the rural areas there.
Federal and State Biodiesel Tax Subsidies (FBTS and SBTS)
In 2004, the federal government established a $1.00 tax subsidy for blending biodiesel
and renewable diesel into diesel as part of the American Jobs Creation Act of 2004, which has
been extended multiple times over the past 18 years. Since 2005, the federal biodiesel tax
subsidy has expired multiple times, but it was always reestablished sometimes retroactively,
effectively, maintaining this subsidy for each and every year. In August 2022, the biodiesel tax
credit was again extended to December 31, 2024 in the Inflation Reduction Act. Given this trend
for offering this blending subsidy, we assume that this $1.00 per gallon biodiesel blending
subsidy will continue to be in place through 2025.
States also provide subsidies to blend biodiesel into diesel. These state subsidies were
enacted in previous years and are presumed to continue through 2025. Table 2.1.3.1-4
summarizes the states that offer such subsidies and their amounts.
Table 2.1.3.1-4: State Biodiesel Subsidies (^/gal)
State
Biodiesel Subsidy
Hawaii
36.5
Iowa
3.5
Illinois
14
North Dakota
100
Rhode Island
30
Texas
20
The California and Oregon LCFS programs do not offer specific subsides per se, but
through the cap-and-trade nature of their programs, they can be equated to subsidies. Oregon also
has a biodiesel blending mandate, which requires that their diesel contain 5% biodiesel. We
assumed that, on average, each state would only blend up to 5% biodiesel, which means that
Oregon's mandate would satisfy its biodiesel volume regardless of its LCFS program. In the case
for California, which does not have a biodiesel mandate, we estimated the equivalent per-gallon
subsidy amount from the incentives offered by its LCFS program. From 2023-2025, biodiesel
produced from soybean oil is estimated to receive an LCFS blending incentive of $1.01/gal in
2023 decreasing to $0.96/gal in 2025. Biodiesel produced from non-soybean oil vegetable oils is
expected to receive a blending incentive of $1.84 each year from 2023-2025.
Although different than subsidies, several states have mandates that require that the diesel
within their state contain a minimum quantity of biodiesel. Table 2.1.3.1-5 lists the states that
have such a mandate and the percentage of biodiesel required to be blended into diesel.
86
-------
Table 2.1.3.1-5: State Biodiesel Mandates
State
Minimum % of Biodiesel
Minnesota
12.5
New Mexico
5
Oregon
5
Pennsylvania
2
Washington
2
Diesel Terminal Price (DTP)
Refinery rack price data from 2019—which already included the distribution costs for
moving diesel to downstream terminals—were used to represent the price of diesel to blenders
on a state-by-state basis. However, these prices were not projected for future years.154 Instead,
we used projected refinery wholesale price data from AEO 2023 to adjust the 2019 refinery rack
price data to represent diesel rack prices in future years. We used 2019 data instead of more
recent data to avoid abnormal pricing effects caused by the COVID-19 pandemic or the
subsequent supply issues that emerged when the pandemic was subsiding. This diesel price data,
summarized in Table 2.1.3.1-6, was collected by states and is assumed to represent the average
diesel price for all the terminals in each state.
154 EIA; Spot Prices; https://www.eia.gov/dnav/pet/pet pri spt s i a.fatm.
87
-------
Table 2.1.3.1-6: Diesel Terminal Prices ($/gal)
2019
2023
2024
2025
2019
2023
2024
2025
Alaska
2.43
4.46
4.27
3.97
Montana
2.00
3.67
3.52
3.27
Alabama
1.93
3.53
3.38
3.15
North Carolina
1.95
3.57
3.42
3.18
Arkansas
1.94
3.56
3.41
3.17
North Dakota
1.98
3.63
3.47
3.23
Arizona
2.07
3.80
3.64
3.39
Nebraska
1.98
3.64
3.48
3.24
California
2.20
4.04
3.87
3.60
New Hampshire
1.98
3.63
3.48
3.24
Colorado
2.02
3.70
3.54
3.29
New Jersey
1.93
3.54
3.39
3.15
Connecticut
1.96
3.60
3.45
3.21
New Mexico
2.05
3.75
3.59
3.34
District of Columbia
1.95
3.58
3.42
3.19
Nevada
2.08
3.82
3.66
3.41
Delaware
1.95
3.58
3.43
3.19
New York
2.00
3.66
3.51
3.26
Florida
1.98
3.63
3.47
3.23
Ohio
1.91
3.51
3.36
3.12
Georgia
1.94
3.56
3.41
3.17
Oklahoma
1.91
3.50
3.36
3.12
Hawaii
2.17
3.97
3.80
3.54
Oregon
2.04
3.75
3.59
3.34
Iowa
1.98
3.62
3.47
3.23
Pennsylvania
1.94
3.55
3.40
3.16
Idaho
2.01
3.68
3.53
3.28
Rhode Island
1.95
3.58
3.43
3.19
Illinois
1.88
3.44
3.29
3.06
South Carolina
1.94
3.57
3.41
3.18
Indiana
1.90
3.48
3.33
3.10
South Dakota
2.00
3.67
3.51
3.27
Kansas
1.94
3.56
3.40
3.17
Tennessee
1.94
3.56
3.41
3.17
Kentucky
1.97
3.62
3.46
3.22
Texas
1.91
3.50
3.35
3.12
Louisiana
1.88
3.45
3.30
3.07
Utah
2.05
3.77
3.61
3.36
Massachusetts
1.98
3.63
3.47
3.23
Virginia
1.95
3.58
3.42
3.19
Maryland
1.95
3.58
3.42
3.19
Vermont
1.99
3.65
3.50
3.25
Maine
1.99
3.64
3.49
3.24
Washington
1.98
3.64
3.48
3.24
Michigan
1.91
3.51
3.36
3.13
Wisconsin
1.93
3.55
3.40
3.16
Minnesota
1.99
3.66
3.50
3.26
West Virginia
1.97
3.61
3.45
3.21
Missouri
1.95
3.59
3.43
3.19
Wyoming
2.12
3.89
3.73
3.47
Mississippi
1.91
3.50
3.36
3.12
U.S. Average
1.96
3.59
3.44
3.20
Because there are state mandates and blending subsidies for biodiesel, each state is
represented in EPA's analysis. Since biodiesel distribution volumes and costs are estimated on a
PADD basis, the states are grouped together within their respective PADDs. We then established
a hierarchy for how biodiesel is consumed. First, state mandates are satisfied by biodiesel
volume that is available to each state within its PADD—the biodiesel volume determined by the
percent mandate requirement and the diesel fuel volume sold in that state in 2019, adjusted to
2023 to 2025 by the national diesel fuel demand in those years from AEO 2023 relative to the
national diesel fuel demand in 2019.155 Next, biodiesel is allocated to states based on its blending
cost—the state with the lowest biodiesel blending cost in each PADD (e.g., states with biodiesel
blending subsidies) would receive biodiesel, with any one state assumed to blend biodiesel only
up to 5%.156 Therefore, once a state reaches 5% biodiesel content in its diesel and more biodiesel
is available in the PADD, biodiesel is blended to the next lowest blending cost state, and so on
until the biodiesel available in the PADD is exhausted.
Similar to the other biofuels analyzed for the No RFS baseline, mandates are satisfied
regardless of the blending economics. If the biodiesel blending cost is negative, biodiesel is
155 EIA; Prime Supplier Sales Volume; https://www.eia.gov/dnav/pet/pet cons prim den tins m.fatm.
156 Minnesota is an exception because it mandates a higher volume. Limiting biodiesel blends to 5% in the
remaining states is appropriate because at least one engine manufacturer does not warranty their truck engines if
operated on diesel containing more than 5% biodiesel. Furthermore, California does not allow biodiesel blends
above 5% in order to avoid increases in NOx emissions.
88
-------
considered economical to blend into diesel and additional nonmandated volumes are assumed to
be blended. Conversely, biodiesel is assumed to not be blended into diesel if the biodiesel
blending value is positive. Because of its relative cost, biodiesel consumption without the RFS
program would be driven mostly by the state mandates, but would also occur absent the RFS
program due to state subsidies, mainly the California LCFS program. The volume of biodiesel
estimated to be blended into diesel in each state is determined by the following methodology:
Using the estimated year-by-year biodiesel volumes estimated or projected by the no RFS
baseline analysis would result in large volumetric swings in some years based on the changing
economics of biodiesel in those years. In reality, the marketplace is unlikely to make such
swings. To avoid this problem, the following steps were taken to normalize the growth and use
of biodiesel:
• Biodiesel economics were assessed historically starting in 2009 through 2022, and
projected in 2023 to 2025, to determine the states where biodiesel would be economical
to blend, and in what volumes.
• Biodiesel demand in any one historical year was not allowed to exceed the demand that
occurred under the RFS program. Since biodiesel demand has been declining in recent
years, this volume was 2376 million gallons in the year 2020.
• When combined with renewable diesel, the volumetric demand for the lowest cost
biogenic oils (i.e., waste oils (FOG) and corn oil) was not allowed to exceed the total
projected demand for those oils that occurred in each year under the RFS program.157
• The maximum demand for biodiesel in any one year for a given state was then calculated
to be the average of biodiesel demand for that year and the previous three years. This step
attempted to reflect how potential biodiesel investors or banks would seek to assess the
economics for investing in expanding biodiesel plant capacity.
157 Since biodiesel is a less expensive process, biodiesel use of the lower cost vegetable oils took precedence over
renewable diesel. However, the mandated biodiesel volume was assumed to use a mix of vegetable oils consistent
with the current fractional use of vegetable oils.
89
-------
Table 2.1.3.1-7 Year-by-Year Analysis of Biodiesel Volumes for the No RFS Baseline
Mandated
Economic
4-Year Average
Best Fit of
Total of Mandated
Biodiesel
Biodiesel
Volume of Economic
4-Year
Biodiesel Volume
Year
Volume
Volume
Biodiesel Volume
Volume
and 4-Year Volume
2009
102
245
2010
135
56
2011
125
360
2012
126
50
178
283
409
2013
126
734
300
293
419
2014
160
571
429
304
464
2015
156
78
358
314
470
2016
170
226
402
324
494
2017
181
623
374
334
515
2018
244
309
309
345
588
2019
239
240
349
355
594
2020
242
257
357
365
607
2021
252
208
254
375
627
2022
254
222
232
386
640
2023
237
294
245
396
633
2024
225
1005
432
406
631
2025
223
1186
677
416
640
Table 2.1.3.1-8 lists the states expected to consume biodiesel under the No RFS baseline
in the years 2023 to 2025 and summarizes their volume of biodiesel by the biogenic oil feedstock
types estimated to be used to produce the biodiesel. For the states that mandate the percentage of
biodiesel to be blended into diesel, we apportioned the biogenic oil feedstock types based on the
current mix of these vegetable oils currently being used to produce biodiesel. For the states that
would use biodiesel based on economics, the use is a function of the biodiesel economics when
using the various feedstocks - the volume of biodiesel would only be produced from a particular
oil feedstock if it is economic to use in that year.158 The table lists the mandated volume by each
state at the top and the volume for states where it is economical to use biodiesel just below the
states with mandates, although only California and North Dakota are listed separately since these
states have the largest subsidies without a mandate, and tend to be economical for multiple
vegetable oil feedstocks, while the projected volumes for the other states are aggregated together.
In the next row in the table, the biodiesel volume by vegetable oil type is totaled.
There are several calculations which follow to convert the volume and vegetable oil
feedstock types from this yearly analysis to the estimated volume and vegetable oil feedstock
types under the No RFS baseline. In the next row of the table, the "Maximum Volumes" of
biodiesel volume by feedstock type is listed. This is important to reflect limits on the maximum
amount of biodiesel by feedstock type. Because of the favorable economics in 2024 and 2025 for
158 Historical diesel sales volumes from EIA and projected diesel volumes in AEO 2023 were used to project the
volume of diesel sold in each state. EIA; Prime Supplier Sales Volume;
https://www.eia.gov/dnav/pet/pet cons prim dcu tins m.fatm.
90
-------
using FOG and corn oil vegetable feedstocks, the maximum limit of 979 and 228 million gallons
of FOG and corn oil feedstocks serves to prevent the over-consumption of these vegetable oil
feedstock types.159 This establishes the Vegetable Oil Fraction which is shown in the next row of
the table. The No RFS Baseline biodiesel volumes for years 2023 to 2025 from Table 2.1.3.1-7
are then reproduced in the table, multiplied by the Vegetable Oil Fraction and summarized in the
last row of the table as Biodiesel Volume by Feedstock Type.
Table 2.1.3.1-8: Biodiesel in No RFS Baseline (million gal/yr)
2023
2024
2025
Soy
Corn
Soy
Corn
Soy
Corn
State
Oil
Oil
FOG
Oil
Oil
FOG
Oil
Oil
FOG
Oregon
18
3
15
17
3
14
17
3
14
Volume in
States
with
New Mexico
17
2
15
16
2
14
16
2
14
Minnesota
57
11
49
54
11
46
53
11
46
Mandates
Washington
11
2
9
10
2
8
10
2
8
Pennsylvania
16
2
10
15
2
9
15
2
9
Economic
Volume
California
North Dakota
31
5
134
23
76
13
16
3
68
12
75
13
16
3
67
11
Other States
2031
317
1820
Total of Mandated and
Economic Volumes
119
58
326
200
39
2202
200
356
1990
Maximum Volumes
1093
228
979
1093
228
979
1093
228
979
Biodiesel Volume by
Vegetable Oil Type
119
58
326
200
39
979
200
228
979
Vegetable Oil Fraction
0.22
0.11
0.67
0.16
0.04
0.80
0.14
0.16
0.70
Volume from Table
2.1.3.1-7
633
631
640
Biodiesel Volume by
Feedstock Type
142
69
422
103
25
502
92
103
445
For the most part, this mix of vegetable oil types is used for biodiesel for estimating costs
for the No RFS Baseline, however, a few minor adjustments were made to the vegetable oil
feedstock types after the No RFS Baseline analysis was conducted for renewable diesel (see
Chapter 2.1.3.3 below).
159 Although FOG is the lowest priced feedstock type and economics would normally dictate using as much of that
feedstock type as the lowest cost option, many biodiesel plants cannot use FOG because of the fatty acid content
which causes operational problems in their plants. Also, biodiesel plants tend to be located more in the Midwest
which is the agricultural centers for the production of corn and soy oil, and the plants may actually have a lower cost
and more reliable option to purchase these vegetable oil types that are produced close to their plants.
91
-------
2.1.3.2 Renewable Diesel
While renewable diesel is produced in a much different process than biodiesel, it uses the
same feedstocks and so much of the blending cost analysis is similar. The blending cost of
renewable diesel is estimated using the following equation:
RDBC = (RDSP + RDDC - FRDTS - SRDTS) - DTP
Where:
• RDBC is renewable diesel blending cost
• RDSP is renewable diesel plant gate spot price
• RDDC is renewable diesel distribution cost
• FRDTS is federal renewable diesel tax subsidy
• SRDTS is state renewable diesel tax subsidy
• DTP is diesel terminal price; all are in dollars per gallon
Some of the equation inputs, including the distribution costs (RDDC), federal tax subsidy
(FRDTS), state tax subsidies (SRDTS), and diesel terminal price (DTP) are the same as that
described in Chapter 2.1.3.1 for biodiesel, so they are not discussed further here. However, the
state mandates described in Chapter 2.1.3.1 are assumed to not apply to renewable diesel.
Renewable Diesel Plant Gate Spot Price (RDSP)
Similar to biodiesel, we estimated future renewable diesel plant gate prices by gathering
projected renewable diesel plant input information (e.g., future biogenic oil and utility prices) to
estimate renewable diesel production costs, which we assumed represent plant gate prices. This
is essentially the same information used for estimating renewable diesel production costs for the
cost analysis in Chapter 10, except that the capital costs are amortized using the capital
amortization factor in Table 2.1.1.1-1. Imports are assumed to be half produced from soybean oil
and half from palm oil, and have the same production costs as that produced domestically. The
resulting projected renewable diesel plant gate prices are summarized in Table 2.1.3.2-1.
Table 2.1.3.2-1: Projected Renewable Diesel Plant Gate Prices ($/gal)
Feedstock
2023
2024
2025
Soybean Oil
5.76
5.62
5.10
Corn Oil
4.85
4.81
4.38
Waste Oil
4.46
4.47
4.08
The methodology for analyzing renewable diesel volumes is structured the same as that
for biodiesel described in Chapter 2.1.3.1. States are grouped together within their respective
PADDs and a hierarchy is established for how renewable diesel is consumed, except that we did
not include any state mandates. The state with the lowest renewable diesel blending cost (e.g.,
states with blending subsidies) would receive renewable diesel first. An important difference
92
-------
from the analysis for biodiesel, however, is that states are able to displace up to 95% of their
diesel volume with renewable diesel.160
Similar to the other biofuels analyzed for the No RFS baseline, if the renewable diesel
blending cost is negative, renewable diesel is considered economical to blend into diesel.
Conversely, renewable diesel is assumed to not be blended into diesel if the blending value is
positive. Because of its relative cost, renewable diesel consumption without the RFS program
would only be blended into diesel if a state offers a significant subsidy, mainly the California and
Oregon LCFS programs. The volume of renewable diesel estimated to be blended into diesel in
each state is determined by the volume of diesel sold in that state.161
Allowing up to 95% of the diesel in a state to be supplanted with renewable diesel would
allow the results of the analysis to swing wildly from year to year based on even small changes
in the economics of renewable diesel in any given year. In reality, the marketplace is unlikely to
make such swings. To avoid this problem, the following steps were taken to rationalize the
growth and use of renewable diesel which is very similar to those conducted for biodiesel:
• Renewable diesel economics were assessed from 2009-2025 to determine the states
where renewable diesel would be economic to blend, and what the maximum volume
could be.
• Renewable diesel demand in any one historical year was not allowed to exceed the
demand that occurred in that year under the RFS program. This data was extrapolated to
determine the maximum renewable diesel demand for future years (see Table 2.1.3.2-2).
• When combined with biodiesel, the demand for the lowest cost biogenic oils (i.e., waste
oils (FOG) and corn oil) was not allowed to exceed the total demand for those oils that
occurred in that year under the RFS program.
• The maximum demand for renewable diesel in any one year for a given state was then
calculated to be the average of renewable diesel demand for that year and the previous
three years. This step attempted to reflect how potential renewable diesel investors or
banks would seek to assess the economics for investing in expanding renewable diesel
plant capacity.
These steps are implemented in Table 2.1.3.2-2. The first column of the table shows the
estimated historical and projected future renewable diesel volumes for each individual year. The
next column shows the 4-year average renewable diesel volume from 2012 to 2025. The last
column shows a best fit to the 4-year average volumes, and the volumes shown for 2023 to 2025
are used for the No RFS Baseline volume.
160 Renewable diesel has properties similar to petroleum diesel, so it can displace petroleum diesel without causing
vehicle compatibility or drivability issues.
161 Historical diesel sales volumes from EIA and projected diesel volumes in AEO 2023 were used to project the
volume of diesel sold in each state. EIA; Prime Supplier Sales Volume;
https://www.eia.gov/dnav/pet/pet cons prim dcu tins m.fatm.
93
-------
Table 2.1.3.2-2: Year-by-Year Analysis of Renewable Diesel (million gallons)
Economic Renewable
4-Year Average Volume of
Best Fit of 4-
Year
Diesel Volume
Economic Renewable Diesel Volume
Year Volume
2009
2
2010
34
2011
0
2012
139
44
32
2013
127
75
90
2014
234
125
148
2015
386
222
206
2016
353
275
265
2017
367
335
323
2018
402
377
381
2019
621
436
439
2020
580
493
497
2021
960
641
681
2022
1439
900
883
2023
1524
1126
1085
2024
1039
1241
1287
2025
2005
1502
1489
Table 2.1.3.2-3 lists the volume of renewable diesel which is economically favorable for
blending into diesel fuel for the years 2023, 2024 and 2025. Although many states are
economical for renewable diesel, particularly in 2024 and 2025 the renewable diesel is
essentially only being consumed in California. For this reason, we show the potential maximum
volume of renewable diesel which can be consumed in California, and we aggregated the
potential consumption volume in other states. The volume of economical renewable diesel is
shown by vegetable oil type, assuming that the mix of vegetable oils is consistent with the
percentage of vegetable oils consumed in a recent year under the RFS program. These vegetable
oil quantities are just a starting point and are adjusted when estimating the mix of vegetable oils
consumed by both biodiesel and renewable diesel plants under a No RFS Baseline to ensure that
the vegetable oil volume is below the established maximum volumes. The final vegetable oil
volumes are shown in Table 2.1.3.3-3.
Table 2.1.3.2-3: Potential Volume of Renewable Diesel by Feedstock Type (million gallons)
Year
State
Feedstock
Total
Soybean Oil
Corn Oil
FOG
2023
California
0
593
2,552
3,146
Other States
0
0
1,245
1,245
2024
California
0
564
2,427
2,991
Other States
0
133
1,164
1,297
2025
California
1,423
297
1,275
2,995
Other States
0
239
12,654
12,893
94
-------
2.1.3.3 Final No RFS Baseline Volumes for Biodiesel and Renewable Diesel
While the volume of biodiesel and renewable diesel by feedstock type were initially
estimated in Tables 2.1.3.1-8 and 2.1.3.2-3, using these volumes, particularly the renewable
diesel volumes, would exceed a total volume by feedstock type that reflects a reasonable growth
increase from current trends, and exceed the maximum expected volume of renewable diesel
estimated in Table 2.1.3.2-2. To estimate the maximum vegetable oil volumes which could be
available for producing biodiesel and renewable diesel in 2023 - 2025, we reviewed the trend in
vegetable oil consumption for previous years and projected their future volumes, which is
summarized in Table 2.1.3.3-1.
Table 2.1.3.3-1: Maximum Vegetable Oil Volumes
Year
Soy
Corn Oil
FOG
2023
1,948
320
1,431
2024
2,128
328
1,379
2025
2,450
336
1,442
The final No RFS Baseline volumes for biodiesel and renewable diesel that result from
the calculations described in Chapters 2.1.3.1 and 2.1.3.2 above are shown in Table 2.1.3.3-2.
Since FOG vegetable oil is the lowest in cost, the renewable diesel plants are assumed to use this
vegetable oil feedstock up to the maximum projected amount after removing the amount
consumed by biodiesel plants. A similar calculation is conducted for corn oil which is the next
most cost-effective vegetable oil. Both FOG and corn oil are consumed up to the maximum
projected amount. In 2025, when soy oil becomes economical to use for producing renewable
diesel, some soy oil renewable diesel is projected to be produced up to the maximum amount of
renewable diesel estimated to be produced in that year. Some small adjustments are made to the
quantities of vegetable oils projected to be consumed by biodiesel plants to make the vegetable
volumes more consistent over the three years.
Table 2.1.3.3-2 Final No RFS Baseline Volumes for Biodiesel and Renewable Diesel (million
Biodiesel
Renewable Diesel
Year
Soy
Corn Oil
FOG
Soy
Corn Oil
FOG
2023
142
69
422
0
75
1,009
2024
210
26
395
0
303
984
2025
198
43
398
153
292
1,044
The amount of renewable diesel in the No RFS Baseline is estimated to be much higher
for this final rule RIA than it was for the DRIA. This primarily reflects the improved economics
for renewable diesel fuel due to the higher projected crude oil prices.
2.1.4 Other Advanced Biofuel
In addition to ethanol, cellulosic biofuel, and BBD, we also estimated volumes of other
advanced biofuel for the No RFS baseline. These biofuels include imported sugarcane ethanol,
domestically produced advanced ethanol, non-cellulosic RNG used in CNG/LNG vehicles,
95
-------
heating oil, naphtha, and advanced renewable diesel that does not qualify as BBD (coded as D5
rather than as D4). In Chapters 6.3 and 6.4, we present a derivation of the projected volumes of
these other advanced biofuels for 2023-2025 in the context of the candidate volumes that we
analyzed. Here we discuss the deviations from those projections that we believe would apply
under a No RFS baseline.
According to data from EIA, all ethanol imports entered the U.S. through the West Coast
in 2019-2021, and the majority did so in 2022. We believe that these imports were likely used to
help refiners meet the requirements of the California LCFS program, which provides significant
additional incentives for the use of advanced ethanol beyond that of the RFS program. In the
absence of the RFS program, we believe that these incentives would remain. Thus, we have
assumed that the volume of imported sugarcane ethanol would be the same regardless of whether
the RFS program were in place in 2023-2025. For similar reasons, we believe that domestically
produced advanced ethanol would also continue to find a market in California in the absence of
the RFS program.
As discussed in Chapter 6.2.4, a similar situation exists for advanced renewable diesel.
The vast majority of the renewable diesel consumed in the U.S. has been consumed in California
to fulfill the mandates of its LCFS program, and possibly Oregon for the same reason. Some
renewable diesel would continue to be consumed in these states in the absence of the RFS
program, particularly that produced from FOG due to the lower Carbon Intensity (CI) value
assigned to it under the LCFS program. We believe that this would also be the case for advanced
renewable diesel that does not qualify as BBD since the statutory threshold of 50% GHG
reduction is the same for advanced biofuel and for BBD, and because such renewable diesel is
generally produced from FOG. Thus, we have assumed that the volume of advanced renewable
diesel that does not qualify as BBD would be the same regardless of whether the RFS program
were in place in 2023-2025.
Remaining forms of other advanced biofuel (i.e., non-cellulosic RNGused in CNG/LNG
vehicles, heating oil, and naphtha) are much less likely to find their way to markets such as the
California LCFS program, where the incentive would be insufficient to continue supporting their
use in the absence of the RFS program. Therefore, we have assumed that consumption of these
biofuels would be zero under the No RFS baseline.
2.1.5 Summary of No RFS Baseline
Following our analysis of individual biofuel types as described above, we estimated the
constituent mix of both renewable fuel types and feedstocks that could be used under a No RFS
baseline, as shown in Table 2.1.5-1 (in million RINs) and Table 2.1.5-2 (in million gallons).
96
-------
Table 2.1.5-1: No RFS Baseline for 2023-2
025 (million RINs]
2023
2024
2025
Cellulosic Biofuel
343
402
444
CNG/LNG from biogas
336
351
367
Ethanol from CKF
7
51
77
Diesel/jet fuel from wood waste/MSW
0
0
0
Total Biomass-Based Diesel
2,796
3,139
3,496
Biodiesel
948
947
959
Soybean oil
212
315
297
FOG
633
593
597
Corn oil
104
39
65
Canola oil
0
0
0
Renewable Diesel
1,843
2,188
2,532
Soybean oil
0
0
261
FOG
1,715
1,673
1,775
Corn oil
128
515
496
Canola oil
0
0
0
Jet fuel from FOG
5
5
5
Other Advanced Biofuels
226
226
226
Renewable diesel from FOG
104
104
104
Imported sugarcane ethanol
95
95
95
Domestic ethanol from waste ethanol
27
27
27
Othera
0
0
0
Conventional Renewable Fuel
13,185
13,224
12,992
Ethanol from corn
13,185
13,224
12,992
Renewable diesel from palm oil
0
0
0
a Composed of non-cellulosic biogas, heating oil, and naphtha.
-------
Table 2.1.5-2: No RFS Baseline for 2023-2
025 (million gallons)
2023
2024
2025
Cellulosic Biofuel
343
402
444
CNG/LNG from biogas
336
351
367
Ethanol from CKF
7
51
77
Diesel/jet fuel from wood waste/MSW
0
0
0
Total Biomass-Based Diesel
1,719
1,921
2,132
Biodiesel
632
621
639
Soybean oil
141
210
198
FOG
422
395
398
Corn oil
69
26
43
Canola oil
0
0
0
Renewable Diesel
1,084
1,287
1,490
Soybean oil
0
0
0
FOG
1,009
984
1,044
Corn oil
75
303
292
Canola oil
0
0
0
Jet fuel from FOG
3
3
3
Other Advanced Biofuels
183
183
183
Renewable diesel from FOG
61
61
61
Imported sugarcane ethanol
95
95
95
Domestic ethanol from waste ethanol
27
27
27
Othera
0
0
0
Conventional Renewable Fuel
13,185
13,224
12,992
Ethanol from corn
13,185
13,224
12,992
Renewable diesel from palm oil
0
0
0
a Composed of non-cellulosic biogas, heating oil, and naphtha.
2.2 2022 Baseline
As discussed in Preamble Section III.D.3, while we believe that the No RFS baseline is
preferable as a point of reference for analyzing the impacts of the candidate volumes, we have
also estimated the costs of this rule relative to renewable fuel consumption in 2022 as an
additional informational case. These alternative estimated costs allow a comparison to those
presented in recent RFS annual rules and provide an appreciation for what the impacts of the rule
may be relative to the recent past.
For the proposal, we needed to estimate the mix of biofuels that could be used to meet the
2022 volume requirements in order to be able to use those volume requirements as a point of
reference. In the 2020-2022 annual rule, we made just such an estimate of the mix of biofuels,
but we adjusted those estimates for the proposal to be more precise.162 For this final rule, we
have instead used the actual consumption volume of each type of renewable fuel as determined
162 See Table 2.1-1, Renewable Fuel Standard (RFS) Program: RFS Annual Rules - Regulatory Impact Analysis,
EPA-420-R-22-008, June 2022.
98
-------
from EMTS data. The results are shown in Table 2.2-1 (in million RINs) and Table 2.2-2 (in
million gallons).
Table 2.2-1: Actual Mix of Biofuels in 2022 (million RINs)
Cellulosic Biofuel
666
CNG/LNG from biogas
665
Ethanol from CKF
1
Diesel/jet fuel from wood waste/MSW
0
Total Biomass-Based Diesela
4,944
Biodiesel
2,606
Soybean oil
1,492
FOG
519
Corn oil
195
Canola oil
400
Renewable Diesela
2,314
Soybean oila
498
FOG
1,460
Corn oil
356
Canola oil
0
Jet fuel from FOG
24
Other Advanced Biofuels
318
Renewable diesel from FOG
124
Imported sugarcane ethanol
81
Domestic ethanol from waste ethanol
29
Otherb
84
Conventional Renewable Fuel
14,034
Ethanol from corn
14,034
Renewable diesel from palm oil
0
a Includes 250 million RINs assumed to be used to meet the supplemental volume requirement for 2022.
b Composed of non-cellulosic biogas, heating oil, and naphtha.
-------
Table 2.2-2: Actual Mix of Biofuels in 2022 (million gallons)
Cellulosic Biofuel
666
CNG/LNG from biogas
665
Ethanol from CKF
1
Diesel/jet fuel from wood waste/MSW
0
Total Biomass-Based Diesela
3,113
Biodiesel
1,737
Soybean oil
995
FOG
346
Corn oil
130
Canola oil
267
Renewable Diesela
1,361
Soybean oila
293
FOG
859
Corn oil
209
Canola oil
0
Jet fuel from FOG
14
Other Advanced Biofuels
247
Renewable diesel from FOG
73
Imported sugarcane ethanol
81
Domestic ethanol from waste ethanol
29
Otherb
65
Conventional Renewable Fuel
14,034
Ethanol from corn
14,034
Renewable diesel from palm oil
0
a Includes 147 million gallons assumed to be used to meet the supplemental volume requirement for 2022.
b Composed of non-cellulosic biogas, heating oil, and naphtha.
-------
Chapter 3: Candidate Volumes and Volume Changes
For analyses in which we have quantified the impacts of the candidate volumes for 2023-
2025 and the 2023 supplemental standard, we have identified the specific biofuel types and
associated feedstocks that are projected to be used to meet those volumes. While we
acknowledge that there is significant uncertainty about the types of renewable fuels that would
be used to meet the candidate volumes, we believe that the mix of biofuel types described in this
chapter are reasonable projections of what could be supplied for the purpose of assessing the
potential impacts. As described in Chapter 2, we also acknowledge that the choice of baseline
affects the estimated impacts of the candidate volumes and the 2023 supplemental standard. This
chapter describes both the methodology for identifying the mix of biofuels that could result from
the candidate volumes and the 2023 supplemental standard and the change in volumes in
comparison to the No RFS and 2022 baselines.
3.1 Mix of Renewable Fuel Types for Candidate Volumes
The candidate volumes that we developed for 2023-2025 (excluding the 2023
supplemental standard) are presented in Preamble Section III.C.5 and are repeated in Tables 3.1-
1 and 2.
Table 3.1-1: Candidate Volume Components (million RINs)a
D Codeb
2023
2024
2025
Cellulosic biofuel
D3 +D7
838
1,090
1,376
Biomass-based diesel
D4
5,965
6,205
6,881
Other advanced biofuel
D5
290
290
290
Conventional renewable fuel
D6
13,845
13,955
13,779
a Does not include RINs used to meet the 2023 supplemental standard.
b The D codes given for each component category are defined in 40 CFR 80.1425(g). D codes are used to identify
the statutory categories that can be fulfilled with each component category according to 40 CFR 80.1427(a)(2).
Table 3.1-2: Candidate Volumes in Statutory and Implied
Categories i
million RIP
D Code
2023
2024
2025
Cellulosic biofuel
D3 +D7
838
1,090
1,376
Non-cellulosic advanced biofuelb
D4 + D5
6,255
6,495
7,171
Advanced biofuel
D3 + D4 +
D5 +D7
7,093
7,585
8,547
Conventional renewable fuelb
D6
13,845
13,955
13,779
Total renewable fuel
All
20,938
21,540
22,326
a Does not include RINs used to meet the 2023 supplemental standard.
b These are implied volume requirements, not regulatory volume requirements.
101
-------
We estimated the constituent mix of renewable fuel types and feedstocks that could be
used to meet the candidate volumes (absent the 2023 supplemental standard) as shown in Tables
3.1-3 (in million RINs) and 3.1-4 (in million gallons).163
Table 3.1-3: Candidate Volumes Assessed for 2023-2025 (million I
UNs)
2023
2024
2025
Cellulosic Biofuel
838
1,090
1,376
CNG/LNG from biogas
831
1,039
1,299
Ethanol from CKF
7
51
77
Total Biomass-Based Diesela
5,965
6,205
6,881
Biodiesel
2,565
2,500
2,436
Soybean oil
1,473
1,451
1,430
FOG
481
454
427
Corn oil
173
134
95
Canola oil
438
461
484
Renewable Diesel
3,376
3,681
4,421
Soybean oil
777
1,141
1,501
FOG
1,883
1,825
1,962
Corn oil
348
406
463
Canola oil
368
309
495
Jet fuel from FOG
24
24
24
Other Advanced Biofuels
290
290
290
Renewable diesel from FOG
104
104
104
Imported sugarcane ethanol
95
95
95
Domestic ethanol from waste ethanol
27
27
27
Otherb
64
64
64
Conventional Renewable Fuel
13,845
13,955
13,779
Ethanol from corn
13,845
13,955
13,779
Renewable diesel from palm oil
0
0
0
a Includes BBD in excess of the candidate volume for advanced biofuel. The excess would be used to help meet the
candidate volume for conventional renewable fuel.
b Composed of non-cellulosic biogas, heating oil, and naphtha.
163 The analyses leading to the mix of renewable fuel types and feedstocks are presented in Chapter 6. We have also
analyzed the impacts of the 2023 supplemental standard under the assumption that it will be met with soybean oil-
based renewable diesel in Chapter 3.4.
102
-------
Table 3.1-4: Candidate Volumes Assessed for 2023-2025 (million gallons)
2023
2024
2025
Cellulosic Biofuel
838
1,090
1,376
CNG/LNG from biogas
831
1,039
1,299
Ethanol from CKF
7
51
77
Total Biomass-Based Diesela
3,710
3,846
4,239
Biodiesel
1,710
1,667
1,624
Soybean oil
982
967
953
FOG
321
303
285
Corn oil
115
89
63
Canola oil
292
307
323
Renewable Diesel
1,986
2,165
2,601
Soybean oil
457
641
883
FOG
1,108
1,074
1,154
Corn oil
205
239
272
Canola oil
216
182
291
Jet fuel from FOG
14
14
14
Other Advanced Biofuels
232
232
232
Renewable diesel from FOG
61
61
61
Imported sugarcane ethanol
95
95
95
Domestic ethanol from waste ethanol
27
27
27
Otherb
49
49
49
Conventional Renewable Fuel
13,845
13,955
13,779
Ethanol from corn
13,845
13,955
13,779
Renewable diesel from palm oil
0
0
0
a Includes BBD in excess of the candidate volume for advanced biofuel. The excess would be used to help meet the
candidate volume for conventional renewable fuel.
b Composed of non-cellulosic biogas, heating oil, and naphtha.
Unlike for 2022, wherein we projected that some palm-based, imported conventional
renewable diesel would be needed in order to meet the applicable standards and the 2022
supplemental standard,164 we do not believe that any palm-based, imported conventional
renewable diesel would be needed in 2023-2025. Our assessment of BBD, described more fully
in Chapter 6.2, leads us to the conclusion that there will be sufficient volumes available to meet
the candidate volumes for non-cellulosic advanced biofuel and conventional renewable fuel and,
in the case of 2023, the supplemental standard.
3.2 Volume Changes Analyzed With Respect to the No RFS Baseline
For those factors for which we quantified the impacts of the candidate volumes for 2023-
2025, the impacts were based on the difference in the volumes of specific renewable fuel types
between the candidate volumes and the No RFS baseline. These differences are shown in Tables
3.2-1 and 2 in terms of RINs and physical volumes, respectively. The values in these tables
reflect the difference between values in Tables 3.1-3 and 2.1.5-1.
164 87 FR 39600 (July 1, 2022).
103
-------
Table 3.2-1: Volume Changes for Candidate Volumes Relative to the No RFS Baseline
(million RINs)
2023
2024
2025
Cellulosic Biofuel
495
688
932
CNG/LNG from biogas
495
688
932
Ethanol from CKF
0
0
0
Total Biomass-Based Diesel
3,169
3,066
3,385
Biodiesel
1,617
1,554
1,478
Soybean oil
1,262
1,136
1,133
FOG
-152
-139
-170
Corn oil
70
95
31
Canola oil
438
461
484
Renewable Diesel
1,533
1,493
1,889
Soybean oil
777
1,141
1,240
FOG
168
152
187
Corn oil
221
-109
-33
Canola oil
368
309
495
Jet fuel from FOG
19
19
19
Other Advanced Biofuels
64
64
64
Renewable diesel from FOG
0
0
0
Imported sugarcane ethanol
0
0
0
Domestic ethanol from waste ethanol
0
0
0
Othera
64
64
64
Conventional Renewable Fuel
660
731
787
Ethanol from corn
660
731
787
Renewable diesel from palm oil
0
0
0
a Composed of non-cellulosic biogas, heating oil, and naphtha.
-------
Table 3.2-2: Volume Changes for Candidate Volumes Relative to the No RFS Baseline
(million gallons)"
2023
2024
2025
Cellulosic Biofuel
495
688
932
CNG/LNG from biogasa
495
688
932
Ethanol from CKF
0
0
0
Total Biomass-Based Diesel
1,991
1,925
2,107
Biodiesel
1,078
1,036
985
Soybean oil
841
757
755
FOG
-101
-92
-113
Corn oil
46
63
20
Canola oil
292
307
323
Renewable Diesel
902
878
1,111
Soybean oil
457
671
729
FOG
99
90
110
Corn oil
130
-64
-20
Canola oil
216
182
291
Jet fuel from FOG
11
11
11
Other Advanced Biofuels
49
49
49
Renewable diesel from FOG
0
0
0
Imported sugarcane ethanol
0
0
0
Domestic ethanol from waste ethanol
0
0
0
Otherb
49
49
49
Conventional Renewable Fuel
660
731
787
Ethanol from corn
660
731
787
Renewable diesel from palm oil
0
0
0
a CNG/LNG remain in ethanol-equivalent gallons in this table.
b Composed of non-cellulosic biogas, heating oil, and naphtha.
Note that the changes in ethanol from corn shown in Tables 3.2-1 and 2 can be entirely
attributed to ethanol used as E15 and E85, since under the No RFS baseline we project that E10
would be used regardless of the RFS program but there would not be any El 5 or E85 use.165 For
the analyses conducted in support of this rule we generally projected that any increase in ethanol
consumption would result in a gallon-for-gallon increase in ethanol production.
Tables 3.2-1 and 2 represent the change in biofuel use in the transportation sector that
could occur if the candidate volumes were to become the basis for the applicable percentage
standards.
In the 2020-2022 annual rule, we made some simplifications to the projected volume
changes for the purposes of our analyses. Namely, we grouped fuels with very small changes in
volumes with similar fuels having much larger volume changes. We did this because: (1) We had
more limited data on the impacts of those renewable fuel types with smaller volume changes; (2)
The impacts on many of the factors evaluated in Chapter 4 are expected to be similar; and (3) We
165 See Chapter 2.1.1 for more discussion on El5 andE85.
105
-------
expect small volume changes to have little material impact on the overall conclusions of the
analyses. For this rule, we have taken a similar approach. This simplification fell into three areas:
1. We have treated all volume changes in canola oil as if they were changes in soybean
oil.
2. We have treated all volume changes in renewable jet fuel as if they were changes in
renewable diesel.
3. We have treated all volume changes in "Other" advanced biofuel, which is dominated
by naphtha, as if they were changes in renewable diesel.166
As a result of these adjustments and simplifications, the volume changes that we used in
our analyses were as follows:
Table 3.2-3: Volume Changes Analyzed for the Candidate Volumes With Respect to the No
2023
2024
2025
CNG/LNG from biogasa
495
688
932
Biodiesel from soybean oil
1,133
1,065
1,078
Biodiesel from FOG
-101
-92
-113
Biodiesel from corn oil
46
63
20
Renewable diesel from soybean oil
710
901
1,066
Renewable diesel from FOG
115
106
126
Renewable diesel from corn oil
137
-68
-21
Ethanol from corn
660
731
787
a CNG/LNG remain in ethanol-equivalent gallons in this table.
For the climate change analyses, we determined that a more robust analysis could be
performed if BBD produced from FOG could be disaggregated into specific types. Therefore,
using data from EIA's Monthly Biofuels Capacity and Feedstocks Update, we determined that
FOG on average consists of 71% used cooking oil (UCO) and about 29% tallow.167 These
fractions were applied to the volume changes shown in Table 3.2-3 for both biodiesel and
renewable diesel produced from FOG in the context of the climate change analyses.
Table 3.2-4: Disaggregated Biofuels Made From FO<
2023
2024
2025
Biodiesel from FOG
-101
-92
-113
UCO
-72
-66
-80
Tallow
-29
-27
-33
Renewable diesel from FOG
113
106
126
UCO
82
75
90
Tallow
33
31
37
(million gallons)
166 We assumed that the feedstocks used to produce these "other" advanced biofuels were proportional to the
feedstocks used to produce renewable diesel.
167 EIA Monthly Biofuels Capacity and Feedstocks Update - Table 2b, https://www.eia.gov/biofuels/update.
Average over both 2021 and 2022.
106
-------
3.3 Volume Changes Analyzed with Respect to the 2022 Baseline
As described in Chapter 2.2, for cost purposes only, we also analyzed the impacts of
volume changes with respect to the 2022 baseline. These differences are shown in Tables 3.3-1
and 2 in terms of RINs and physical volumes, respectively. The values in these tables reflect the
difference between values in Tables 3.1-3 and 2.2-1.
Table 3.3-1: Volume Changes for Candidate Volumes Relative to 2022 Baseline (million
RINs)
2023
2024
2025
Cellulosic Biofuel
172
424
710
CNG/LNG from biogas
166
374
634
Ethanol from CKF
6
50
76
Total Biomass-Based Diesel
1,271
1,511
2,187
Biodiesel
-41
-106
-170
Soybean oil
-19
-41
-62
FOG
-38
-65
-92
Corn oil
-22
-61
-100
Canola oil
38
61
84
Renewable Diesel
1,312
1,617
2,357
Soybean oil
529
893
1,253
FOG
423
365
502
Corn oil
-8
50
107
Canola oil
368
309
495
Jet fuel from FOG
0
0
0
Other Advanced Biofuels
-28
-28
-28
Renewable diesel from FOG
-20
-20
-20
Imported sugarcane ethanol
14
14
14
Domestic ethanol from waste ethanol
-2
-2
-2
Other
-20
-20
-20
Conventional Renewable Fuel
-189
-79
-255
Ethanol from corn
-189
-79
-255
Renewable diesel from palm oil
0
0
0
107
-------
Table 3.3-2: Volume Changes for Candidate Volumes Relative to 2022 Baseline (million
gallons)"
2023
2024
2025
Cellulosic Biofuel
172
424
710
CNG/LNG from biogasa
166
374
634
Ethanol from CKF
6
50
76
Total Biomass-Based Diesel
744
881
1,273
Biodiesel
-27
-71
-113
Soybean oil
-13
-27
-41
FOG
-25
-43
-61
Corn oil
-15
-41
-67
Canola oil
25
41
56
Renewable Diesel
772
951
1,386
Soybean oil
311
525
737
FOG
249
215
295
Corn oil
-5
29
63
Canola oil
216
182
291
Jet fuel from FOG
0
0
0
Other Advanced Biofuels
-15
-15
-15
Renewable diesel from FOG
-12
-12
-12
Imported sugarcane ethanol
14
14
14
Domestic ethanol from waste ethanol
-2
-2
-2
Other
-15
-15
-15
Conventional Renewable Fuel
-189
-79
-255
Ethanol from corn
-189
-79
-255
a CNG/LNG remains in ethanol-equivalent gallons in this table
Unlike for the comparison to the No RFS baseline, the changes in ethanol from corn
shown in Tables 3.3-1 and 2 are a function of both changes in total gasoline demand as well as
changes in the consumption of E15 and E85. Table 3.3-3 shows the amount of ethanol that can
be attributed to each.
Table 3.3-3: Source of Ethanol Changes in Comparison to the 2022 Baseline (million
gallons)
2023
2024
2025
Changes in ethanol consumption attributable to
changes in gasoline demand
309
87
203
Changes in ethanol consumption attributable to
changes in El 5 and E85 consumption
27
95
129
Total
336
182
332
We made the same adjustments and simplifications to the volume changes in comparison
to the 2022 baseline as we made to the volume changes in comparison to the No RFS baseline.
The results are shown in Table 3.3-4.
108
-------
Table 3.3-4: Volume Changes Analyzed for Candidate Volumes With Respect to the 2022
2023
2024
2025
CNG/LNG from biogasa
166
374
634
Ethanol from CKF
6
50
76
Biodiesel from soybean oil
13
13
15
Biodiesel from FOG
-25
-43
-61
Biodiesel from corn oil
-15
-41
-67
Renewable diesel from soybean oil
517
696
1,017
Renewable diesel from FOG
244
211
292
Renewable diesel from corn oil
-5
29
62
Ethanol from corn
-189
-79
-255
a CNG/LNG remains in ethanol-equivalent gallons in this table
3.4 2023 Supplemental Volume Requirement
As discussed in Preamble Section V, we are establishing a supplemental volume
requirement of 250 million gallons of renewable fuel that will apply in 2023, which completes
our response to the ACE remand. Although we are requiring this supplemental volume
requirement in concert with the final volumes for 2023-2025 established under CAA section
21 l(o)(2)(B)(ii), the 2023 supplemental volume requirement is not established under our "set"
authority, but rather our outstanding obligation from 2016 to promulgate standards under CAA
section 21 l(o)(3)(B)(i). Additionally, as we have in the past, we rely on our authority in CAA
section 21 l(o)(2)(A)(i) to promulgate late standards.168 It is in fact an independent requirement
that is separately justified. For this reason, our analysis of the statutory factors listed in CAA
section 21 l(o)(2)(B)(ii)(I) through (VI) has been focused on the candidate volumes exclusive of
the supplemental volume requirement.
The requirements of CAA section 21 l(o)(2)(B)(ii) do not apply to the 250-million-gallon
supplemental volume requirement for 2023; we have not conducted an analysis of all of the
factors listed in CAA section 21 l(o)(2)(B)(ii)(I) through (VI) as part of our assessment of the
appropriateness of imposing the supplemental volume requirement on obligated parties.
Nevertheless, it is both prudent and consistent with the requirements of Executive Order 12866
and Circular A-4 that we assess the costs, GHG, and energy security impacts of the 250-million-
gallon supplemental volume requirement for 2023.
In our assessment for 2023, we have projected that biodiesel and renewable diesel would
be the fuels most likely to be supplied to satisfy the 250-million-gallon supplemental volume
requirement. We also determined that there would be sufficient quantities of biodiesel and
renewable diesel available to satisfy the supplemental volume requirement beyond the quantity
of these fuels needed to satisfy the BBD, advanced biofuel, and total renewable fuel
requirements for 2023. However, it is difficult to identify the precise mix of biofuel types and
feedstocks that would make up this 250 million gallons since it is not a segregated and uniquely
categorized pool of renewable fuel. For the purposes of analyzing its impacts, we have made the
1® [n promulgating the 2009 and 2010 combined BBD standard, upheld by the D.C. Circuit in NPRA v. EPA, 630
F.3d 145 (2010), we utilized express authority under CAA section 21 l(o)(2). 75 FR 14670, 14718.
109
-------
simplifying assumption that it would be composed entirely of soybean oil renewable diesel, as
we project that this is the highest cost type of biodiesel or renewable diesel available, and
therefore the fuel type that is likely to make up the marginal gallons used to satisfy the
supplemental volume requirement.
Under the No RFS baseline, there would be no supplemental volume requirement
because there would be no RFS obligations of any kind. However, under the 2022 baseline there
is in fact a supplemental volume requirement.169 As described in the 2020-2022 annual rule, we
projected that the 250-million-gallon supplemental volume requirement for 2022 would be met
with imported palm-based renewable diesel. The net result is that the 250-million-gallon
supplemental volume requirement for 2023 would result in the following changes in fuel types in
comparison to the No RFS and 2022 baselines:
Table 3.3-1: Volume Changes for 2023 Supplemental Volume Requirement (million
In comparison to No RFS baseline
Soybean oil renewable diesel
+147
Palm oil renewable diesel
0
In comparison to 2022 baseline
Soybean oil renewable diesel
0
Palm oil renewable diesel
0
a The 250-million-allon supplemental volume requirement represents ethanol-equivalent gallons. Values are
presented in physical gallons of renewable diesel, where 1 gallon of renewable diesel has the same amount of energy
as 1.7 gallons of ethanol.
169 87 FR 36900 (July 1, 2022).
110
-------
Chapter 4: Environmental Impacts
The statute requires EPA to analyze a number of environmental factors in its determination
of the appropriate volumes to establish under the set authority. This chapter discusses those
environmental factors required by the statute. Due to its close association with water quality,
which is a factor listed in the statute, we also investigated soil quality even though it is not listed
in the statute. In addition to the analysis presented here, we also considered the Second Triennial
and draft Third Triennial Report to Congress on Biofuels, which provides additional information
on environmental impacts.170'171
4.1 Air Quality
Air quality, as measured by the concentration of air pollutants in the ambient atmosphere,
can be affected by increased production and use of biofuels. Some air pollutants are emitted
directly (e.g., nitrogen oxides (NOx)), other air pollutants are formed secondarily in the
atmosphere (e.g., ozone), and some air pollutants have directly emitted and secondarily formed
components (e.g., particulate matter (PM) and aldehydes). Health and environmental effects of
criteria pollutants and air toxics which can be impacted by biofuel use are discussed in a
memorandum to the docket.172 Air quality can be affected by emissions from combustion of
biofuels in vehicles, as well as emissions from production and transport of feedstocks,
conversion of feedstocks to biofuels, and transport of the finished biofuels. Recent dispersion
modeling has shown elevated pollutant concentrations near corn, soybean, and wood
biorefineries, which were associated with adverse respiratory outcomes.173
In addition to the type of biofuel, other factors may affect air quality, including but not
limited to the blend level, the vehicle technology, emissions control technology, and operating
conditions. However, overall, the impacts on air quality resulting from the biofuel volume
changes due to this rule are expected to be relatively minor and thus provide little basis to favor
higher or lower volumes. First, the largest volume changes are for renewable diesel, primarily
produced from soybean oil, particularly in comparison to 2022 volumes, with smaller volumes of
biodiesel and renewable diesel from fats, oils, and greases (FOG), ethanol, and biogas. Much of
the increase in renewable diesel is produced at traditional petroleum refineries that have been
converted to renewable fuel production; at such facilities, the emission impact is not likely to be
significant because the processes used to produce renewable diesel are similar to processes used
in the production of petroleum-based diesel. In addition, while data on end- use impacts of
renewable diesel are limited, the impacts are expected to be minor. It should also be noted that
170 EPA. Biofuels and the Environment: Second Triennial Report to Congress (Final Report, 2018). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-18/195, 2018.
https://cfpnb.epa.gov/si/si public record report.cfm?Lab=IO&dirEntrvId=34.1.491.
171 EPA. Biofuels and the Environment: Third Triennial Report to Congress (Draft Report, 2022). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-18/195, 2018.
172 EPA (2023). "Health and environmental effects of pollutants discussed in Chapter 4 of Regulatory Impact
Analysis (RIA) supporting final RFS standards for 2023-2025." Memorandum from Margaret Zawacki to Docket
No. EPA-HQ-OAR-2021-0427, June 2023.
173 Lee, E. K., Romeiko, X. X., Zhang, W. Feingold, B., Khwaja, H., Zhang, X., and Lin, S. (2021). Residential
proximity to biorefinery sources of air pollution and respiratory diseases in New York State. Environ. Sci., Technol.
55, 10035-10045.
Ill
-------
EPA's "anti-backsliding study" (ABS), required under CAA Section 21 l(v)(l), examined the
impacts on air quality that might result from changes in vehicle and engine emissions associated
with renewable fuel volumes of ethanol under the RFS, relative to approximately 2005 levels.174
Hoekman et al. (2018) also reviewed available literature on potential air quality impacts for E10
versus E0 across the entire lifecycle.175 Both studies found potential increases and decreases in
ambient concentration levels of pollutants, but none of them were large, even when they
considered much greater changes in ethanol volumes than are being established in this rule. The
Second Triennial Report to Congress on Biofuels176 summarized information on air quality
associated with biofuels and emphasized that emissions of NOx, SOx, CO, VOCs, NH3, PM2.5,
and PM10, can be impacted at each stage of biofuel production, distribution, and usage. The
report also noted that impacts associated with feedstock and fuel production and distribution are
important to consider when evaluating the air quality impacts of biofuel production and use,
along with those associated with fuel usage.177
Table 3.2-3 summarizes the changes in renewable production volume assessed for this
rule. The discussion below focuses on potential impacts for these fuel/feedstock combinations.
4.1.1 Production and Transport Emissions of Liquid Biofuels
Corn Ethanol
Air quality impacts of corn ethanol are associated with each step in the supply chain: (1)
agricultural feedstock production and storage, (2) feedstock transport to the biorefinery, (3)
ethanol production at the biorefinery, (4) ethanol distribution, blending and storage, and (5) end
use.
There is little recent literature that addresses air quality impacts of processes upstream of
the end-use vehicle emissions from corn ethanol. A 2009 analysis using the GREET model
concluded that criteria pollutant emissions from corn ethanol production are substantially higher
than for gasoline on a mass per gasoline equivalent gallon basis.178 A significant source of
174 EPA (2020). Clean Air Act Section 21 l(v)(l) Anti-Backsliding Study.
https://nepis.epa. gov/Exe/ZvPDF.cgi?Dockev=P100ZBYl.pdf.
175 Hoekman, S. K., Broch, A., & Liu, X. (2018). Environmental implications of higher ethanol production and use
in the U.S. Renewable and Sustainable Energy Reviews, 81, 3140-3158.
176 U.S. EPA (2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Environmental
Protection Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June 2018.
177 The Third Triennial Report to Congress on Biofuels is in progress and the AQ impacts summarized in that draft
version are consistent with what was in the Second Triennial Report to Congress on Biofuels, see
https://cfpub.epa.gov/ncea/biofuels/recordisplav.cfm?deid=353055.
178 Hess P, Johnston M, Brown-Steiner B, Holloway T, de Andrade JB, Artaxo P. Chapter 10: air quality issues
associated with biofuel production and use. In: Howarth RW, Bringezu S. editors. Biofuels: environmental
consequences and interactions with changing land use. Gummersbach, Germany; 2009. p. 169-94.
https://ecom.mons.co rneil.edu/bitstream/handle/ .1.8.1.3/46218/scope..1.245782010.pdf?sequence=2.
112
-------
upstream emissions from corn ethanol is production facilities.179'180 Table 4.1.1-1 summarizes
corn ethanol plant emissions, using data from the 2017 National Emissions Inventory (NEI)
where available.181 For facilities not found in the 2017 NEI, we used data from the 2016
emissions modeling platform version 1182 or inventory estimates using facility-level volume data
from the EPA moderated transaction system.183 Only a few plants used coal or coal in
combination with other energy sources, although the small number of wet mill plants contributed
disproportionately to emissions, especially sulfur dioxide.
Table 4.1.1-1: Pollutant Emissions (short tons) From Biodiesel and Corn Ethanol
Biorefineries in U.S. in 2017
Finished
Fuel
Number
of
Facilities
CO
NH3
NOx
PMio
PM25
S02
VOCs
Corn
Ethanol
(total)
176
7362.8
278.7
9045.5
5218.7
4088.5
1854.4
8908.7
Coal; Dry
Mill
2
75.3
0
55.8
20.7
20.0
n.a.
39.7
Coal; Wet
Mill
2
455.9
23.2
603.2
376.5
260.0
547.1
827.9
Natural
Gas; Dry
Mill
160
6389.6
246.4
7880.6
4533.5
3647.2
904.4
7560.3
Natural
Gas; Wet
Mill
3
251.8
9.0
142.2
184.5
102.7
74.7
270.5
Unknown;
Unknown
9
190.1
0.0
363.7
103.4
58.5
327.6
210.3
Biodiesel6
175
960.5
39.7
1277.0
815.7
556.2
3384.1
3987.2
Total
351
8323.2
318.4
10,322.5
6034.4
4644.6
5238.5
12,895.9
Sources: EPA 2017 NEI (https://www.epa.gov/air-eniissions-inventories/2017-tiatlonal-eniissions-inventore-nei-
data) and EPA 2016 version 1 modeling platform (https://www.epa.gov/air-em.issions-modeiing/2016vl-platform')
Once the ethanol is produced at biorefineries, it is transported to terminals for blending
and storage. At the blending terminal, ethanol is blended with gasoline for various fuel
combinations such as E10, El 5, or E85. The blended fuel is then sent to retail gasoline outlets
where it is sold to the customer. Primary modes of distributing ethanol to the blending terminal
and the blended fuel to the retail outlets are rail, road, or barges. Emissions come from
combustion and evaporation during transport by mobile sources, as well as evaporative losses
179179 £pa Biofuels and the Environment: Second Triennial Report to Congress (Final Report, 2018). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-18/195, 2018.
https://cfpnb.epa.gov/si/si public record report.cfm?Lab=IO&dirEntrvId=341491.
180 de Gouw, J. A., McKeen, S. A., Aikin, K. C., Brock, C. A., ABrown, S. S., Gilman, J. B., Graus, M., AHanisco,
T., Holloway, J. S., Kaiser, J., Keutsch, F. N., Lerner, B. M., Liao, J., Markovic, M. Z., Middlebrook, A. M., Min,
K.-E., Neuman, J. A., Nowak, J. B., Peischl, J., Pollack, I. B., Roberts, J. M., et al. (2015). Airborne measurements
of the atmospheric emissions from a fuel ethanol refinery. Journal of Geophysical Research: Atmospheres, 120(9),
4385-4397. https://doi.org/10.1002/2015JDQ23138.
181 https://www.epa.gov/air-emissions-inyentories/2017-national~emissions-inyentorv-nei~data.
182 https://www.epa.gov/air-emissions-modeling/2016vl-platform.
183 https://www.epa.gov/fiiels-registration-reporting-and-compliance-help/reporting-rfs-rin-transactions-epa-
mode rated.
113
-------
during storage and transport. The largest emission contribution is for VOCs due to evaporation.
Table 4.1.1-2 presents emissions associated with transport. Air quality impacts associated with
changes in ethanol production and transport are expected to be primarily in the local area where
the emissions occur.184 Ambient measurements also indicate concentrations of several pollutants
such as NOx, formaldehyde, and SO2 are greater directly downwind of production facilities, up
to a distance of 30 kilometers.185
Table 4.1.1-2. Emissions From Transportation of Et
lanol (short t
ons) in 2016
CO
nh3
NOx
PM10
PM25
SO2
voc
4,225
26
19,270
630
533
340
660,674
Source: EPA 2016 version 1 modeling platform (https://www.epa.gov/air-emissions-modeling/2016vl-platform')
Using the production and transport emissions data, along with total production in 2017,
we calculated emission rates in grams per gallon for production of and transport of corn ethanol.
We then multiplied the grams per gallon emission rates by the volume impacts for this rule,
relative to the No RFS baseline (from Table 3.2-3), to estimate the impacts associated with
ethanol production and transport (see Table 4.1.1-3). In doing so, we assumed that additional
ethanol use in the U.S. is associated with ethanol production in the U.S. We note that ethanol
volumes used domestically could be sourced from imports. Thus, it is unclear what impact
imports would have on domestic production and associated emissions. We note, moreover, that
significant quantities of domestically produced ethanol are exported, and thus not used for RFS
compliance; the below table does not capture emissions related to such exports.
184 Cook, R., Phillips, S., Houyoux, M., Dolwick, P., Mason, R., Yanca, C., Zawacki, M., Davidson, K., Michaels,
H., Harvey, C., Somers, J., Luecken, D.. 2011. Air quality impacts of increased use of ethanol under the United
States' Energy Independence and Security Act. Atmospheric Environment, 45: 7714-7724.
https://www.sciencedirect.com/science/article/pii/S1352231010007375.
185 See, e.g., de Gouw, J. A., McKeen, S. A., Aikin, K. C., Brock, C. A., ABrown, S. S., Gilman, J. B., Graus, M.,
AHanisco, T., Holloway, J. S., Kaiser, J., Keutsch, F. N., Lerner, B. M., Liao, J., Markovic, M. Z., Middlebrook, A.
M., Min, K.-E., Neuman, J. A., Nowak, J. B., Peischl, J., Pollack, I. B., Roberts, J. M., et al. (2015). Airborne
measurements of the atmospheric emissions from a fuel ethanol refinery. Journal of Geophysical Research:
Atmospheres, 120(9), 4385-97. https://doi.org/10.1002/2015JDQ23138.
114
-------
Table 4.1.1-3: Pollutant Emission Impact Estimates for Production and Transport of Corn
Ethanol of the 2023-2025 Final Volumes Relative to ]>
o RFS Baseline (shorl
tons)
CO
nh3
NOx
PMio
pm25
so2
voc
Bioreftnery
Emissions
7,363
279
9,046
5,219
4,089
1,854
8,909
Transport
Emissions
4,225
26
19,270
630
533
340
660,674
Total
Emissions
11,558
305
28,316
5,849
4,622
2,194
669,583
Impacts
per Million
Gallons
Ethanol11
0.75
0.02
1.83
0.38
0.30
0.14
43.45
2023
Volume
Changes
495
13
1,208
251
198
92
28,677
2024
Volume
Changes
548
15
1,338
278
219
102
31,762
2025
Volume
Changes
590
16
1,440
299
236
110
34,195
a Emissions per million gallons ethanol is calculated using total domestic ethanol production in 2017 as reported in
the EI A Monthly Energy Review (15.41 billion gallons)
We also compared emission rates per energy unit produced for production of ethanol
versus gasoline, using emissions data from the 2017 NEI and production for 2017 from the EIA.
The portion of refinery emissions attributable to gasoline production was estimated using data
from GREET.186 As seen in Table 4.1.1-4, emissions per BTU produced are much higher for
ethanol than gasoline.
Table 4.1.1-4: Emissions Per Energy Unit Produced for Ethanol Versus Gasoline
(g/mmBTU) in 2017
Pollutant
g/mmBTU
EtOH
g/mmBTU
Gasoline
VOC
6.64
0.64
CO
5.49
0.55
NOx
6.75
0.76
PMio
3.89
0.21
PM2.5
3.05
0.19
SO2
1.38
0.27
NH3
0.21
0.03
186 Sun, P., Zhu, L. Emissions Updates for Petroleum Products in GREET 2019,
https://greet.es.anl.gov/files/petro 20.1.9.
115
-------
Biodiesel/Renewable Diesel
Although biodiesel is sourced from a variety of feedstocks, domestic soybean and
domestic FOGs made up nearly 80% of the biodiesel in 2022, with most of that being domestic
soybean. Data are lacking on emission and air quality impacts of either soybean oil biodiesel or
FOGs that address the feedstock production (soybean) or collection (FOGs), storage, and
transport stages. In the soybean diesel production phase, emission impacts depend on the oil
extraction method used. Mechanical expelling is the least efficient with the highest emissions of
NOx, VOCs, CO, and PM2.5, followed by hexane extraction and then enzyme assisted aqueous
extraction process (EAEP).187 Hum et al. (2016) compared life cycle emissions for low sulfur
diesel (LSD), soybean-based biodiesel, and grease trap waste (GTW)-based biodiesel.188 This
study relied on GREET-2014 for soybean-based biodiesel impacts.189 The study found decreases
in PM and CO (5% and 66%), but increases in NOx and SOx (10% and 39%, respectively).
However, the comparison's end use emission estimates included only pre-2007 engines.
A smaller amount of biodiesel is derived from FOG. FOGs are waste products of
processes like animal rendering. Overall, since FOG is a generally a byproduct, farming
emissions are not attributed to it, and the effects from FOGs may be expected to be much lower
than for soybean oil biodiesel.
Table 4.1.1-1 provides estimated emissions from biodiesel refineries in the U.S. Given
the limited impact of this rule on biodiesel production, national-scale impacts are small.
However, there could be localized impacts.
We also compared emission rates per energy unit produced for production of biodiesel
versus distillate, using emissions data from the 2017 NEI and production for 2017 from the EIA.
As seen in Table 4.1.1-5, emissions per BTU produced are much higher for biodiesel than
distillate.
187 Cheng, M., Sekhon, J. J. K., Rosentrater, K. A., Wang, T., Jung, S., Johnson, L. A. "Environmental Impact
Assessment of Soybean Oil Production: Extruding-Expelling Process, Hexane Extraction and Aqueous Extraction."
Food and Bioproducts Processing 108 (2018): 58-68.
https://www.sciencedirect.com/science/article/abs/pii/S0960308518300014.
188 Hums, M., Cairncross, R., & Spatan, S. (2016). Life-cycle assessment of biodiesel produced from grease trap
waste. Environmental Science & Technology, 50(5), 2718-2726. https://doi.Org/.l.0.1021/acs.est.5b02667.
189 Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model; Argonne National
Laboratory: Argonne, IL, 2014. https://greet.es.ani.gov.
116
-------
Table 4.1.1-5: Emissions Per Energy Unit Produced for Biodiesel Versus Distillate
(g/mmBTU) in 2017
Pollutant
g/mmBTU
Biodiesel
g/mmBTU
Distillate
voc
19.21
1.37
CO
4.63
1.19
NOx
5.91
1.63
PMio
3.93
0.45
PM2.5
2.68
0.40
SO2
16.30
0.59
NH3
0.19
0.07
While biodiesel is the predominant advanced biofuel used in diesel engines, renewable
diesel is projected to account for roughly similar increases in biomass-based diesel during the
2023 through 2025 timeframe. Much renewable diesel is produced at traditional petroleum
refineries; at such facilities the emission impact is not likely to be significant because the
processes used to produce renewable diesel are similar to processes used in the production of
petroleum-based diesel. However, there will be emission impacts from new facilities constructed
to produce renewable diesel. Reported emissions data for such facilities are extremely limited
and inadequate to draw any conclusions about potential level of impacts. Furthermore, these
emission increases may be offset by emission decreases resulting from decreased petroleum
distillate refining at other locations. Thus, given the limited research available on renewable
diesel production and end use emissions, we have not been able to quantify the air quality
impacts of the additional renewable diesel use associated with this rule.
4.1.2 End Use Emissions of Liquid Biofuels
Ethanol
After distribution to the retail outlet stations, end use at the vehicle occurs. Emissions at
this step include both evaporative losses during dispensing the fuel, diurnal tank venting
processes, and exhaust emissions from combustion during vehicle operation. Impacts of ethanol
blends on vehicle exhaust emissions are the result of complex interactions between fuel
properties, vehicle technologies, and emission control systems. Depending on the pollutant and
blend concentration, the impacts vary both in direction and magnitude.
Several test programs in recent years have evaluated the impacts of fuel properties,
including those of certain ethanol blends on emissions from vehicles meeting Tier 2 and Tier 3
117
-------
standards.190'191'192'193 However, because the projected changes in volume of ethanol resulting
from this action are much smaller than the total amount of fuel consumed across the country, and
given the magnitude of the changes in emission rates when burning E10 vs EO, the overall end
use impacts are expected to be small. The volume changes we are projecting are largely due to
increased use of E10. We expect only very small increases in E15 and E85 use, as we discuss in
Chapter 6.5, and thus emission changes due to increased use of these fuels are also anticipated to
be very minor.
Biodiesel
Biodiesel consists of straight-chain molecules that boil in the diesel range and typically
contain at least one double bond as well as oxygen incorporated into a methyl ester group. These
chemical features can cause differences in emissions relative to petroleum diesel, primarily when
used in older engines. EPA's MOVES3 model assumes no emission impacts of biodiesel fuel for
engines meeting 2007 and later standards due to their highly efficient emission controls.
However, the model does estimate criteria pollutant emission impacts for pre-2007 engines based
on data generated for B20 (20 vol%) blends of soybean-based biodiesel in petroleum diesel
(Table 4.1.2-1; EPA, 2020, Table 8-1).194 The biodiesel effects implemented in MOVES are
obtained from an analysis conducted as part of the 2010 Renewable Fuel Standard Program.195
Table 4.1.2-1: Emission Impacts on Pre-2007 Heavy-Duty Diesel Engines for All Cycles
Tested on 20 vol% Soybean-Based Biodiesel Fuel Relative to an Average Base Petroleum
Diesel Fuel
Pollutant
Percent Change in Emissions
THC (Total Hydrocarbons)
-14.1
CO
-13.8
NOx
+2.2
PM2.5
-15.6
Renewable Diesel
Renewable diesel (RD) is made by hydrotreating vegetable oils or other fats or greases, a
process that removes oxygen and double bonds and produces paraffins in the diesel boiling
190 EPA (2013a). Assessing the Effect of Five Gasoline Properties on Exhaust Emissions from Light-Duty Vehicles
Certified to Tier 2 Standards: Analysis of Data from Epact Phase 3 (Epact/V2/E-89).
191 EPA (2013b). Epact/V2/E-89: Assessing the Effect of Five Gasoline Properties on Exhaust Emissions from
Light-Duty Vehicles Certified to Tier 2 Standards - Final Report on Program Design and Data Collection.
192 Morgan, P., Lobato, P., Premnath, V., Kroll, S., Brunner, K.. Impacts of Splash-Blending on Particulate
Emissions for SIDI Engines. Coordinating Research Council (2018). http://crcsite.wpengine.com/wp-
eontent/nptoads/20.1.9/05/CRC~E~94~3 Final-Report 2018~06~26.pdf.
193 Morgan, P., Smith, I., Premnath, V., Kroll, S., Crawford, R.. Evaluation and Investigation of Fuel Effects on
Gaseous and Particulate Emissions on SIDI in-Use Vehicles. Coordinating Research Council (2017).
http://crcsite.wpengine.com/wp-content/nploads/2019/05/CRC_2017-3-21_03-20955_E94-2FinalReport-Revlb.pdf.
194 EPA. Fuel Effects on Exhaust Emissions from Onroad Vehicles in MOVES3. U. S. Environmental Protection
Agency, Ann Arbor, MI, EPA-420-R-20-016. https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P.1.0.1.0M6C.pdf.
195 USEPA Office of Transportation and Air Quality. Regulatory Impact Analysis: Renewable Fuel Standard
Program (RFS2). EPA-420-R-10-006. Assessment and Standards Division, Ann Arbor, MI. February, 2010.
(Appendix A).
118
-------
range. As a result, it has a very high cetane index and essentially zero aromatics or sulfur
content.196 Given the paucity of data at the time MOVES3 was released, and the fact that RD is
chemically identical to material that comprises a significant portion of petroleum diesel, we did
not include any emission impacts for RD blends in the model. Since we are now forecasting a
significant increase in the use of RD, it seems appropriate to provide a brief review of recent
studies assessing its emission impacts.
In 2020, McCaffery, el al., compared ULSD (ultra-low sulfur petroleum diesel) to a
98.5% RD / 1.5% ULSD blend in a 2012 Chevrolet Silverado with Duramax engine.197 The
study used the LA92 test cycle to represent real-world driving as well as eight steady-state
speed-load combinations for additional data. Emissions were sampled upstream of aftertreatment
to focus on the effects of the fuel itself. Results for the LA92 cycle showed reductions in
particulate mass and number, hydrocarbons, and NOx for RD compared to ULSD, while the
steady-state tests showed lower hydrocarbons at all points, lower PM for six of the eight points
(higher PM at two), and no change in NOx at seven of the eight points (higher NOx at one).
In a 2015 study, Na, et al., compared RD to California ULSD and two intermediate
blends at 20% and 50% RD using a model year 2000 Freightliner truck with Caterpillar CI5
engine.198 Test conditions included the EPA Urban Dynamometer Driving Schedule (UDDS) and
the California Heavy Heavy-Duty Diesel Truck (HHDDT) cruise procedure. Results showed
reductions or no statistically significant differences in PM, hydrocarbon, and NOx across both
test conditions for RD and its blends.
Singh, et al., in a 2018 literature review, concluded that RD blends consistently reduced
particulate mass and number emissions relative to petroleum diesel.199 They observed that NOx
emission impacts were less consistent across test cycles and engine and injection technologies,
but that the majority of studies that measured NOx found a trend of reductions with RD.
A 2015 multimedia evaluation of renewable diesel prepared by the California Air
Resources Board concluded that RD reduced emissions of PM, NOx, hydrocarbons, and CO in
diesel engine exhaust compared to petroleum diesel.200 They also observed that RD is likely to
reduce exhaust PAHs, a conclusion supported by Singer, et al., in a 2015 study.201
196 Coordinating Research Council, "Combustion and Engine-Out Emissions Characteristics of a Light Duty Vehicle
Operating on a Hydrogenated Vegetable Oil Renewable Diesel", Project CRC E-l 17, July 2022.
197 McCaffery, C., Karavalakis, G., Durbin, T. Johnson, K. (2020) Engine-Out Emission Characteristics of a Light
Duty Vehicle Operating on a Hydrogenated Vegetable Oil Renewable Diesel. SAE Paper 2020-01-0337.
198 Na, K., Biswas, S., Robertson, W., Sahay, K., Okamoto, R., Mitchell, A., & S., L. (2015). Impact of biodiesel
and renewable diesel on emissions of regulated pollutants and greenhouse gases on a 2000 heavy duty diesel truck.
Atmospheric Environment, 107, 307-14.
199 Singh, D., Subramanian, K.A., Garg, M.O. (2018). Comprehensive review of combustion, performance and
emissions characteristics of a compression ignition engine fueled with hydroprocessed renewable diesel. Renewable
and Sustainable Energy Reviews, 81, 2947-2954.
200 California EPA (2015). Staff report: Multimedia evaluation of renewable diesel.
https://ww2.arb.ca.gov/sites/defanit/files/2018-08/Renewable Diesel Multimedia Evaluation 5-21-15.pdf.
201 Singer, A., Schroder, O., Pabst, C., Munack, A., Biinger, J., Ruck, W., Krahl, J. (2015). Aging studies of
biodiesel and HVO and their testing as neat fuel and blends for exhaust emissions in heavy-duty engines and
passenger cars. Fuel, 153, 595-603.
119
-------
Overall, these studies suggest that emission increases of NOx are not expected with
additional RD use, while emission reductions of PM and hydrocarbons are likely. We will
continue to evaluate the need to include emissions impacts of RD in future MOVES updates as
more data becomes available on RD volumes and blend-levels in the fuel supply.
4.1.4 Air Quality Modeling
As mentioned at the beginning of Chapter 4.1, air quality impacts resulting from this
action are expected to be relatively minor. Any significant impacts are likely to be highly
localized, and thus, would likely not be captured at the geographic scale used in photochemical
air quality modeling. Thus, no air quality modeling was done. Geographic distribution of
emissions also varies, and a comprehensive evaluation of offsetting impacts is very complex.
Furthermore, to the extent that this rule is associated with reductions in imported refined
petroleum products, those upstream emissions and the adverse impacts they cause would occur in
foreign countries. Such upstream international impacts are typically considered outside the scope
of an RIA or other analysis used to support a rulemaking.
4.2 Climate Change
CAA section 21 l(o)(2)(B)(ii) states that the basis for setting applicable renewable fuel
volumes after 2022 must include, among other things, "an analysis of.. .the impact of the
production and use of renewable fuels on the environment, including on.. .climate change."
While the statute requires that EPA base its determinations, in part, on an analysis of the climate
change impact of renewable fuels, it does not require a specific type of analysis. While the
impacts of climate change include rising temperatures and sea levels, ocean acidification,
increased occurrence and intensity of wildfires and extreme weather events, and other impacts,202
these impacts are driven by changes in greenhouse gas emissions. Since the CAA requires
evaluation of lifecycle greenhouse gas (GHG) emissions as part of the RFS program, we believe
the CAA gives us the discretion to use lifecycle GHG emissions estimates as a reasonable proxy
for climate change impacts.
Our assessment of the climate change impacts of the candidate volumes relies on an
extrapolation of lifecycle analysis (LCA) GHG emissions estimates.203 As we did in the 2020-
2022 rule, this approach involves multiplying LCA emissions of individual fuels by the change
in the consumption of each fuel in the candidate volumes scenario relative to the No RFS
baseline to quantify the GHG impacts. We repeat this process for each fuel (e.g., corn ethanol,
soybean oil biodiesel, landfill biogas CNG) to estimate the overall GHG impacts of the candidate
202 Wuebbles, D.J., D.W. Fahey, K.A. Hibbard, B. DeAngelo, S. Doherty, K. Hayhoe, R. Horton, J.P. Kossin, P.C.
Taylor, A.M. Waple, and C.P. Weaver, 2017: Executive summary. In: Climate Science Special Report: Fourth
National Climate Assessment, Volume I [Wuebbles, D.J., D.W. Fahey, K.A. Hibbard, D.J. Dokken, B.C. Stewart,
and T.K. Maycock (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, pp. 12-34, doi:
10.7930/J0DJ5CTG.
203 In this section, we use a range of terminology, consistent with the scientific literature, to describe the concept of
lifecycle GHG emissions. We sometimes call lifecycle GHG emissions "LCA emissions," "LCA ranges," "LCA
values," "LCA estimates," "carbon intensity (CI)," or some combination of these terms. For purposes of this
discussion, the meaning of these terms is the same, namely the GHG emissions associated with all stages of fuel
production and use, including significant indirect GHG emissions.
120
-------
volumes. In the 2020-2022 rule, we applied the LCA estimates that we developed in the March
2010 RFS2 rule and in subsequent agency actions. While our existing LCA estimates for the RFS
program remain within the range of more recent estimates, we acknowledge that the biofuel
GHG modeling framework EPA has previously relied upon is old, and that an updated
framework is needed. Thus, for this rulemaking, we are updating the approach from the 2020-
2022 rule to use a range of more recent LCA estimates from the scientific literature. Given the
uncertainty and variation in LCA estimates, instead of providing one estimate of the GHG
impacts of each candidate volume, we provide a high and low estimate of the potential GHG
impacts. We then use this range of values for considering the GHG impacts of the candidate
renewable fuel volumes that change relative to the No RFS baseline.
This section discusses our evaluation of the potential effects of the candidate volumes on
GHG emissions. We start with background on our LCA of the GHG emissions associated with
biofuels since the beginning of the RFS2 program in 2010. Following this, we present a range of
LCA estimates from the literature. We use these ranges along with the volume scenarios
discussed in Chapter 3 to produce a range of potential GHG emissions impacts. Finally, we
monetize this range of GHG emissions to produce an estimate of the monetized GHG benefits
associated with the candidate volumes.
The science associated with lifecycle assessment of biofuels remains an active area of
research and discussion. Recent examples include the October 2022 report from the National
Academies of Science, Engineering and Medicine (NASEM) titled "Current Methods for Life
Cycle Analyses of Low-Carbon Transportation Fuels in the United States (2022)."204 While the
NASEM report does not endorse any particular numerical result or model, it provides
recommendations on the application of LCA methods and areas for additional research. Another
example of ongoing research and discussions comes from the Administration's sustainable
aviation fuel (SAF) Grand Challenge. A workgroup that includes DOE, EPA, FAA and USDA is
currently examining LCA methodologies and data needs specifically related to SAF. Our model
comparison exercise, discussed in a Model Comparison Exercise Technical Document included
in the docket for this rulemaking, contributes to this continuing scientific research and
discussion. As EPA uses LCA modeling of transportation fuels not just for RFS analysis of
program performance and feedstock assessment but also broader policy analysis, the Agency
would benefit from updating its existing set of transportation fuel LCA modeling capabilities.
Data and findings from recent science and the modeling comparison exercise will help inform
EPA's next steps on updating its lifecycle GHG estimation methodology as part of a separate
action.
4.2.1 Background on Renewable Fuel GHG Analysis for the RFS Program
To support the GHG emission reduction goals of EISA, Congress required that biofuels
used to meet the RFS obligations achieve certain lifecycle GHG reductions. To qualify as a
renewable fuel under the RFS program, a fuel must be produced from qualifying feedstocks and
have lifecycle GHG emission that are at least 20% less than the baseline petroleum-based
204 National Academies of Sciences, Engineering, and Medicine 2022. Current Methods for Life Cycle Analyses of
Low-Carbon Transportation Fuels in the United States. Washington, DC: The National Academies Press.
https://doi.org/10.17226/26402.
121
-------
gasoline and diesel fuels.205 The CAA specifically defines the term "lifecycle greenhouse gas
emissions" to mean "the aggregate quantity of greenhouse gas emissions (including direct
emissions and significant indirect emissions such as significant emissions from land use
changes), as determined by the Administrator, related to the full fuel lifecycle, including all
stages of fuel and feedstock production and distribution, from feedstock generation or extraction
through the distribution and delivery and use of the finished fuel to the ultimate consumer, where
the mass values for all greenhouse gases are adjusted to account for their relative global warming
potential."206 In the March 2010 RFS2 rule (75 FR 14670), EPA interpreted the provision
"including direct emissions and significant indirect emissions" as requiring our LCA to consider
the consequential, or market-mediated, impacts of increased demand for renewable fuels.
Indirect emissions, by definition, cannot be directly measured in the way that direct emissions
can be calculated. Indirect emissions result from changes in prices (e.g., agricultural commodity
or petroleum prices) that ripple through the economy. For example, if increased consumption of
renewable fuel in the U.S. diverts U.S. exports of corn from the global markets, the market-
mediated impact could be for other countries to produce more corn to supply the global demand
for cereal grains. While all of the corn used to produce ethanol in the scenario may have been
grown in the U.S., the land use change emissions and other crop production emissions (e.g., from
increased fertilizer use) in these other countries would be considered "indirect" or "induced"
land use change emissions. Other examples of market-mediated impacts include changes in
livestock production that result from increased production of renewable fuel co-products such as
soybean meal. If the increased production of soybean meal leads to a decrease in feed prices for
cattle, an indirect impact could include the increased production of beef, with associated GHG
emissions.
While the term "significant indirect emissions" requires analytical judgement, prior
modeling work has indicated that the indirect impacts from land use change, livestock, and crop
production can result in emissions that have a large impact on the lifecycle GHG estimates.
Therefore, to be consistent with the CAA requirements, our lifecycle analyses of crop-based
biofuels have taken into account global agricultural and livestock markets, since many biofuel
feedstocks use globally traded commodities. In addition, the increasing interdependence of the
energy and agricultural markets suggests that capturing indirect energy sector impacts could have
important implications for lifecycle analysis.
As part of the March 2010 RFS2 rule, EPA estimated lifecycle GHG emissions
attributable to different various production pathways; that is, the emissions associated with the
production and use of a biofuel, including indirect emissions, on a per-unit energy basis. At the
time of the analysis for the 2010 RFS2 rule, there were no models available off the shelf that
could perform the type of lifecycle analysis required under our interpretation of the CAA
definition of lifecycle GHG emissions. Thus, EPA developed a new modeling framework to
perform the analysis. The framework we developed used multiple models and data sources.207
We used the Forest and Agricultural Sector Optimization Model with Greenhouse Gases model
(hereinafter referred to as "FASOM") and the FAPRI-CARD model (Food and Agricultural
205 See 42 USC 7545(o)(l), (2)(A)(i).
206 See 42 USC 7545(o)(l)(H).
207 EPA (2010). Renewable fuel standard program (RFS2) regulatory impact analysis. Washington, DC, US
Environmental Protection Agency Office of Transportation Air Quality. EPA-420-R-10-006. Chapter 2.4.
122
-------
Policy Research Institute international model; hereinafter referred to as "FAPRI") developed at
the Center for Agriculture and Rural Development at Iowa State University. We ran aligned
scenarios in both models and used FASOM to estimate domestic agricultural and forestry sector
impacts, and the FAPRI model to estimate international agricultural sector impacts. Our
framework included data from many other sources, including emissions factors and other data
from the GREET model. We sought public comment on this new framework, organized four peer
reviews of different aspects of it and held a public workshop. Based on all of this input we
refined the modeling for the final March 2010 RFS2 rule. We also estimated the uncertainty
associated with the land use change satellite data and emissions factors used in our analysis. The
framework we developed in 2010 used the best science, data, and models available at the time.
Since the 2010 RFS2 rule, we have used the RFS2 modeling framework to conduct
numerous (over 140) analyses of new pathways and their lifecycle GHG emissions. Based on
these analyses, we have approved additional pathways for participation in the RFS program.
These pathways rely on novel feedstocks (e.g., canola oil, grain sorghum, camelina oil,208
distillers sorghum oil209) and novel production processes involving existing feedstocks (e.g.,
catalytic pyrolysis and upgrading of cellulosic biomass,210 gasification and upgrading of crop
residues211). EPA maintains a summary of lifecycle greenhouse gas intensities estimated for the
Renewable Fuel Standard program, which are available in spreadsheet form in a document titled
"Summary Lifecycle Analysis Greenhouse Gas Results for the U.S. Renewable Fuels Standard
Program."212 Our lifecycle analyses of various pathways are also published online.213 A list of
pathways that have been approved by regulation can also be found at 40 CFR 80.1426(f)(1).
Depending on the renewable fuel, the feedstocks used to produce it, the amount of fossil
energy used in growing the feedstocks and producing the fuel, land use change and associated
agricultural emissions, and other factors, the GHG emission reductions vary considerably. In
general, we have found that renewable fuels that are not expected to have significant impacts on
land use—such as fuels produced from wastes, residues, or by-products—have greater GHG
emission reductions than renewable fuels produced from crops intended to be used as feedstock
for renewable fuel production. For instance, with respect to biodiesel and renewable diesel
production, the use of waste fats, oils, and greases (FOG) as feedstocks typically results in lower
lifecycle GHG emissions compared to use of vegetable oils, such as soybean or canola oil.214 In
addition, most cellulosic biofuels—which are required to meet the highest statutory lifecycle
208 Pathways I rule. 78 FR 14190 (March 5, 2013).
209 Sorghum oil rule. 83 FR 37735 (August 2, 2018).
210 Pathways I rule. 78 FR 14190 (March 5, 2013).
211 "San Joaquin Renewables Fuel Pathway Determination under the RFS Program." May 11, 2020.
https://www.epa.gov/renewable-fiiel-standard-program/san-ioaanin-renewables-approval.
212 See https://www.epa.gOv/fiieis-registration-reportlng-and-compliance-heip/li:fecvcle-greenhonse-gas-resnits.
213 See faMs:/feww£Biy^M^ and
214 According to EPA's assessment biodiesel produced from yellow grease has lifecycle GHG emissions of 13.8 kg
CChe/mmBTU while biodiesel produced from soybean oil and canola oil have lifecycle GHG emissions of 42.2 kg
CChe/mmBTU and 48.1 kg CChe/mmBTU respectively. See https://www.epa.gov/fiiels-iegistration-reporting-and-
compliance-help/lifecvcle-greenhouse-gas-resiilts.
123
-------
GHG reduction threshold of 60%—are currently produced from wastes, residues, or by-products,
including landfill biogas.215
Since the time when EPA developed that 2010 LCA methodology, multiple researchers
and analytical teams have further studied and assessed the lifecycle GHG emissions associated
with transportation fuels in general and crop-based biofuels in particular. While our existing
LCA estimates for the RFS program remain within the range of more recent estimates, we
acknowledge that the biofuel GHG modeling framework we have relied upon to date is
comparatively old, and that a better understanding of these newer models and data is needed.
Accordingly, EPA has initiated work to develop updated modeling of the GHG impacts
associated with biofuels. In consultation with our interagency partners at USDA and DOE, we
hosted a virtual public workshop on biofuel GHG modeling on February 28 and March 1,
2022.216 In the proposed rule we invited public input on various aspects of biofuel GHG
modeling and LCA, and in the DRIA we included a discussion and review of available models
and LCA estimates. The workshop proceedings, including the workshop presentations and the
comments submitted to the workshop docket, touch on a broad and complex set of topics.217 A
general theme that emerged from this process is that, in support of a better understanding of the
lifecycle GHG impacts of biofuels, it would be helpful to compare available models, identify
how and why the model estimates differ, and evaluate which models and estimates align best
with available science and data. Our model comparison exercise, discussed in the Model
Comparison Exercise Technical Document, contributes to this continuing scientific research and
discussion. However, as explained in Preamble Section IV.A.2, we did not ultimately rely on the
model comparison exercise to evaluate the candidate volumes or to inform the volumes in this
final rule.
4.2.2 Range of LCA Estimates by Fuel Pathway for Illustrative Scenario
As discussed at the beginning of Chapter 4.2, our assessment of the climate change
impacts of this action involves multiplying LCA emissions of individual fuels by the change in
the candidate volumes of that fuel to quantify the GHG impacts. We repeat this process for each
fuel (e.g., corn ethanol, soybean oil biodiesel, landfill biogas CNG) to estimate the overall GHG
impacts of the candidate volumes. In the 2020-2022 RVO rulemaking, we applied the LCA
estimates that we developed in the March 2010 RFS2 rule (75 FR 14670) and in subsequent
agency actions. In this rulemaking, we use a range of LCA emissions estimates that are in the
literature. Instead of providing one estimate of the GHG impacts of each candidate volume, we
provide a high and low estimate of the potential GHG impacts, which is inclusive of the values
we estimated in the 2010 RFS final rule and subsequent agency actions. We then use this range
215 According to data from EMTS in 2020 over 92% of all cellulosic biofuel RINs were produced from biogas from
landfills or biogas from municipal wastewater treatment facilities. An additional 7% of cellulosic biofuel was
produced from agricultural residues or biogas from agricultural digesters.
216 For more information see the Federal Register Notice, "Announcing Upcoming Virtual Meeting on Biofuel
Greenhouse Gas Modeling." 86 FR 73756. December 28, 2021. More information is also available on the workshop
webpage: https://www.epa.gov/renewable-fiiel-standard-program/workshop-biofiiel-green.honse-gas-modeling.
217 Docket ID No. EPA-HQ-OAR-2021-0921.
124
-------
of values for considering the GHG impacts of the candidate renewable fuel volumes that change
relative to the No RFS baseline.218
To develop the range of LCA values, we conducted a high-level review of relevant
literature for the biofuel pathways that would be most likely to satisfy the candidate renewable
fuel volumes.219 Based on our review, we compiled the LCA estimates in the literature for each
pathway. Given that all LCA studies and models have particular strengths and weaknesses, as
well as uncertainties and limitations, our goal for this compilation of literatures estimates is to
consider the ranges of published estimates, not to adjudicate which particular studies, estimates
or assumptions are most appropriate. We include estimates from peer-reviewed journal articles,
authoritative governmental reports, and other credible publications, such as studies by non-
governmental organizations. Our review is intentionally broad and inclusive, and is informed by
our experience conducting LCA evaluations of transportation fuels for the RFS program. Our
review includes studies that estimate the lifecycle GHG emissions associated with the relevant
biofuel pathways and the petroleum-based fuels they replace. We focused on estimates of the
average type of each fuel produced in the United States.220 For example, for corn ethanol, we
focused on estimates for average corn ethanol production from natural gas-fired dry mill
facilities, as that is the predominant mode of corn ethanol production in the United States.221
We made minor changes to the LCA ranges used in the proposed rule.222 We reviewed
the public comments and searched the literature to identify new or additional studies to add to
our review. However, public commenters did not identify any additional LCA estimates that we
had not already considered. Likewise, our updated search of the literature did not identify any
additional estimates. The one update we made was replacing estimates from the 2021 version of
the Greenhouse gases, Regulated Emissions, and Energy use in Technologies (GREET) Model
with estimates from the 2022 version. Some of the public comments recommended removing
some of the studies considered in the proposed rule. We considered these comments carefully but
decided not to remove any of the studies considered in the proposed rule as they meet the broad
criteria for our compilation of published estimates. We discuss these comments and our
218 As explained in Chapter 4.2.3, for the illustrative scenario we require an annual stream of GHG emissions for
crop-based fuels, and the only available results that report an annual stream are from EPA's prior LCA modeling for
the RFS program. Thus, for crop-based fuels, the range of values comes only from EPA's LCA results.
219 Details on the sourcing of each estimate from our literature review are available in the memo, "Notes on
Literature Review of Transportation Fuel Greenhouse Gas (GHG) Lifecycle Analysis (LCA)," available in the
docket for this action.
220 We note that lifecycle GHG emissions are also influenced by the use of advanced technologies and improved
production practices. For example, corn ethanol produced with the adoption of advanced technologies or climate
smart agricultural practices can lower LCA emissions. Corn ethanol facilities produce a highly concentrated stream
of CO2 that lends itself to carbon capture and sequestration (CCS). CCS is being deployed at ethanol plants and has
the potential to reduce emissions for corn-starch ethanol, especially if mills with CCS use renewable sources of
electricity and other advanced technologies to lower their need for thermal energy. Climate smart farming practices
are being widely adopted at the feedstock production stage and can lower the GHG intensity of biofuels. For
example, reducing tillage, planting cover crops between rotations, and improving nutrient use efficiency can build
soil organic carbon stocks and reduce nitrous oxide emissions.
221 Lee, U., et al. (2021). "Retrospective analysis of the US corn ethanol industry for 2005-2019: implications for
greenhouse gas emission reductions." Biofuels, Bioproducts and Biorefining.
222 We do not include estimates from the model comparison exercise in the range of LCA estimates from the
literature, as explained in Preamble Section IV.A.2, we did not ultimately rely on the model comparison exercise to
evaluate the candidate volumes or to inform the volumes in this final rule.
125
-------
reasoning in the summary and analysis of comments document that is part of this rulemaking
package.
We reviewed relevant literature and identified a range of lifecycle GHG estimates for the
biofuel pathways with increased consumption in the candidate volumes scenario relative to the
No RFS baseline scenario (see Chapter 3 for description of these scenarios). We also identified a
range of lifecycle GHG estimates for the conventional fossil-based fuels that the biofuels are
likely to replace. We include estimates from the March 2010 RFS2 rule and studies published
after the March 2010 RFS2 rule, as that rule considered the available science at the time. We do
not claim that our compilation is fully comprehensive, but we attempt to include relevant studies
published before March 2023. In cases where there are multiple studies that include updates to
the same general model and approach, we include only the most recent study. However, we
include a subset of older estimates that are still used for major regulatory programs or that
continue to be widely cited for other reasons. In this section we focus on studies that estimate full
lifecycle (or "well-to-wheel") GHG emissions.223 We focus our compilation on estimates of the
average type of each fuel produced in the United States (e.g., natural gas-fired corn ethanol
plants), though we include discussion in relevant sections about how advanced technologies
could lead to more significant emissions reductions in the future.
Many of the studies we compiled include sensitivity analysis, where parameters or other
assumptions are varied to produce a number of estimates. In these cases, we include
representative high and low estimates. For example, when studies report a 95% interval of
estimates, we include only the central estimate (usually the default, mean or median estimate)
and the estimates at the top and bottom of the interval. This approach simplifies the presentation
of results relative to including every estimate in between. We believe this approach is
appropriate given that the primary purpose of our literature review is to produce a range (high
and low estimate) for each pathway. We intentionally do not calculate or present any statistics
(e.g., mean, median) derived from the estimates included in our literature review, as we do not
believe such statistics would be meaningful or appropriate based on the design of our literature
compilation.
As discussed below, in a very small number of cases we remove outlier estimates that are
representative of localized or special circumstances. We do this in order to form a range that we
believe is representative of each biofuel pathway on a national average basis. We believe this is
appropriate as the purpose of our review is to consider national average fuel production, not
regional variation or unique conditions that are unlikely to represent the impacts of the candidate
volumes relative to the No RFS baseline. While we do remove a small number of estimates that
represent special circumstances, our general approach to the literature compilation is to be
inclusive of a wide range of estimates based on a variety of study types and assumptions.
Figure 4.2.2-1 provides an overview of the lifecycle GHG estimates in our literature
compilation. All of the pathways in our compilation are included with the exception of
compressed natural gas (CNG) produced from manure digester biogas, as some of the estimates
for this pathway (e.g., -533 gCChe/MJ) are so low that they skew the rest of the chart. All of the
223 For crop-based biofuels, there are many studies that only estimate land use change GHG emissions; these studies
are discussed in DRIA Chapter 4.2.2.8.
126
-------
estimates in this chart report lifecycle GHG estimates as carbon dioxide-equivalent (CChe)
emissions per megajoule (MJ) of fuel consumed. All CChe estimates are based on 100-year
global warming potential (GWP) from the IPCC.224 This allows us to compare all of the
estimates on a gCChe/MJ of fuel basis. However, we stress that many of the studies in this chart
do not align in terms of their scope, system boundaries, time horizon, year of analysis, or other
factors. Therefore, the estimates reported in this figure give us a sense for the range of estimates
for each pathway, but we must exercise caution when comparing estimates and drawing
conclusions.
224 The reviewed estimates use GWP values from the IPCC Second Assessment Report (SAR), Fourth Assessment
Report (AR4) or Fifth Assessment Report (AR5). We did not attempt to harmonize GWP assumptions across studies
as many studies only reported CO2Q results and not emissions by gas.
127
-------
Figure 4.2.2-1 Lifecycle GHG Emissions Estimates by Pathway
Petroleum Gasoline -
Petroleum Diesel -
• ••
Natural Gas CNG -
Corn Starch Ethanol -
Soybean Oil Biodiesel ¦
Soybean Oil Renewable Diesel -
• • •• •
• <"••>•••• • • •
% •• •
• A A •
• • • • •• ••• • •• •• •
UCO Biodiesel -
• •
• • • •
UCO Renewable Diesel -
Tallow Biodiesel -
Tallow Renewable Diesel - •
•••# •
•MM* • ••
• • •
• •„•••
•• • *
• •••• •
• •
DCO Biodiesel -
•• •
DCO Renewable Diesel -
• •
• •
Landfill Gas CNG -
• •
• •
0 40 80
Lifecycle GHG Emissions (gC02e/MJ)
120
Notes: CNG = compressed natural gas, DCO = distillers corn oil. UCO = used cooking oil. Other than reporting all
estimates in gCOie/MJ no effort has been made to harmonize estimates. Estimates for CNG produced from manure
digester biogas are excluded, as some of the estimates for this pathway (e.g., -533 gC02e/MJ) are so low they would
skew the rest of the chart.
Figure 4.2.2-1 shows that, in general, the LCA estimates for biofuel pathways tend to be
lower than those for petroleum gasoline, diesel, and natural gas. GHG emissions for biofuels
produced from corn and soybeans tend to be higher than those produced from used cooking oil,
tallow or landfill gas. However, there are some high estimates for the tallow-based pathways
when animal production emissions are allocated to the tallow.225 Below, we summarize the
225 Seber, G., et al. (2014). "Environmental and economic assessment of producing hydroprocessed jet and diesel
fuel from waste oils and tallow." Biomass and Bioenergy 67: 108-18.
128
-------
results of our literature compilation for each of the relevant biofuel pathways. For better
legibility we provide a list of references at the end of this section rather than using footnotes.
In Chapter 4.2.3, we describe how the LCA estimates for each pathway are used to
estimate a potential range of GHG impacts associated with the candidate volumes relative to the
No RFS baseline for an illustrative 30-year scenario. Chapter 4.2.4 describes how the GHG
estimates are used to estimate the potential monetized benefits of the GHG impacts associated
with the candidate volumes relative to the No RFS baseline for an illustrative 30-year scenario.
4.2.2.1 Length of Time Period for Analysis
The time period over which land use change emissions are quantified influences the GHG
estimates for crop-based renewable fuels. If increased demand for biofuels leads to land
conversion, an initial pulse of emissions would likely be released when the land is converted, and
there would also be foregone sequestration over time based on the carbon that would have been
sequestered through plant growth and soil carbon accumulation absent the biofuel induced land
use change.226 Over time, if the biofuel production continues, the GHG benefits of displacing
fossil fuels may eventually "pay back" the initial increase in GHG emissions from the first year.
Thus, when increased biofuel production is expected to result in land conversion, longer
analytical time horizons should result in greater GHG reduction estimates than shorter time
horizons, provided other assumptions are met. The question of the appropriate time horizon over
which to evaluate the net emissions can depend on many factors (e.g., the lifetime of the project,
the goals of the program, future projections of renewable fuels use). After considering public
comments and the input of an expert peer review panel, in the March 2010 RFS2 rule (75 FR
14670), EPA determined that our lifecycle greenhouse gas emissions analysis for renewable
fuels would quantify the GHG impacts over a 30-year period. One of the reasons for using 30
years as a reasonable time horizon for analysis is that biofuel production facilities last multiple
decades after they are constructed.
EPA continues to believe that 30 years is an appropriate timeframe for evaluating the
lifecycle GHG emissions of renewable fuels for purposes of determining which fuel pathways
satisfy the statutory GHG reduction thresholds for qualification under each of the four categories
of renewable fuel. With respect to estimating the GHG impacts of this rulemaking specifically,
the CAA gives us discretion to choose the appropriate analytical time period. On one hand, this
Set rule is part of the broader RFS program that has been in existence since 2005, so there have
been long-term market impacts of standards that were set in past individual years. Furthermore,
once the cost of clearing and converting land is incurred, and given that global cropland areas are
expected to continue expanding, it seems likely that land will continue to be used for agricultural
226 The initial pulse of emissions may take longer than one year depending on the fate of the biomass cleared from
the land. For example, if the biomass is burned, the emissions will indeed occur in the first year. If it is left on the
ground or landfilled the emissions associated with biomass decay may occur over several years. The lifecycle GHG
analyses for the March 2010 RFS2 rule allocated international biomass clearing emissions to the first year. We said
at the time that this was a simplification that was appropriate for the purposes of the analysis. EPA (2010).
Renewable fuel standard program (RFS2) regulatory impact analysis. Washington, DC, US Environmental
Protection Agency Office of Transportation Air Quality. EPA-420-R-10-006. Section 2.4.4.2.6.8.
129
-------
purposes in the future for a period of time.227 On the other hand, the volumes in this rule do not
extend beyond 2025 and making projections about future policies, volume requirements, and
renewable fuel use are inherently uncertain. Since we have chosen to use LCA GHG estimates as
the proxy for evaluating the climate change impacts of this rule, it is consistent to use the same
30-year analytical time period for the purposes of estimating the GHG impacts of this rule.
Therefore, in the illustrative GHG scenario presented in Chapter 4.2.3, the analyses assume that,
in each of the 29 years following the introduction of the final standards, aggregate renewable fuel
consumption (and consequent increased demand for agricultural goods) for each category
exceeded baseline levels by the same volume as required by this rule.
4.2.2.2 Petroleum Gasoline and Diesel
The net GHG impacts of the production and use of biofuels depends on the GHG
emissions associated with the conventional fuels they displace. For the purposes of conducting
the lifecycle GHG emissions analysis and determining which biofuels meet the GHG
requirements, CAA Section 21 l(o)(l)(C) defines baseline lifecycle greenhouse gas emissions as
"the average lifecycle greenhouse gas emissions, as determined by the Administrator, after notice
and opportunity for comment, for gasoline or diesel (whichever is being replaced by the
renewable fuel) sold or distributed as transportation fuel in 2005." As the baseline lifecycle GHG
emissions are used for a specific purpose under the RFS program, we are not required to use it
for evaluating the GHG impacts of this action. Given that this rule involves biofuel production
and use in 2022 and beyond, we believe it is appropriate to consider LCA estimates for gasoline
and diesel production that occurred more recently than 2005. Furthermore, given that we are
developing a range LCA estimates from literature for biofuels, we believe a similar approach is
appropriate for the conventional fuels they replace. Thus, our literature review for this action
includes studies that estimate the lifecycle GHG emissions associated with petroleum-based
gasoline and diesel.
For the March 2010 RFS2 rule, EPA estimated the lifecycle GHG emissions associated
with average 2005 gasoline and diesel. Our review includes the 2010 RFS2 rule and studies that
estimated lifecycle GHG emissions for average U.S. gasoline and diesel that were published
following the 2010 RFS2 rule. Studies that estimate only the GHG emissions associated with
crude oil extraction228 or refining229 are excluded from our review, as we require estimates of the
227 Globally cropland areas have been expanding, suggesting that once land is put into cultivation it is likely to stay
under production. Potapov et al. (2022) report that cropland area increased by 9% globally from 2003 to 2019.
Furthermore, integrated assessment modeling of future scenarios suggests that global cropland areas, including
bioenergy cropland, are expected to increase through 2100. See for example the figure on page 32 of IPCC (2019).
Potapov, P., Turubanova, S., Hansen, M.C. et al. Global maps of cropland extent and change show accelerated
cropland expansion in the twenty-first century. Nat Food 3, 19-28 (2022). https://doi.org/10.1038/s43016-Q21-
00429-z: IPCC, 2019: Summary for Policymakers. In: Climate Change and Land: an IPCC special report on climate
change, desertification, land degradation, sustainable land management, food security, and greenhouse gas fluxes in
terrestrial ecosystems [P.R. Shukla, J. Skea, E. Calvo Buendia, V. Masson-Delmotte, H.- O. Portner, D. C. Roberts,
P. Zhai, R. Slade, S. Connors, R. van Diemen, M. Ferrat, E. Haughey, S. Luz, S. Neogi, M. Pathak, J. Petzold, J.
Portugal Pereira, P. Vyas, E. Huntley, K. Kissick, M. Belkacemi, J. Malley, (eds.)]. In press.
228 See for example, Masnadi, M. S., et al. (2018). "Global carbon intensity of crude oil production." Science
361(6405): 851-53.
229 See for example, Jing, L., et al. (2020). "Carbon intensity of global crude oil refining and mitigation potential."
Nature Climate Change 10(6): 526-32.
130
-------
full lifecycle GHG emissions. It is not appropriate to simply sum crude oil extraction estimates
with refining estimates as the properties of the crude oil from different can affect the refining
emissions.
We recognize that the LCA of the average gallon of gasoline and diesel replaced by
biofuels may be different than the LCA of the marginal gallons replaced. While there is at least
one study that suggests the lifecycle emissions intensity of the marginal oil supplies may be
higher than average oil supplies,230 we did not identify any studies that estimate the full lifecycle
GHG emissions of the marginal volumes that would most likely be displaced, including oil
extraction, oil transport, refining, fuel distribution and use.
The following figures show the range of published LCA estimates for gasoline and diesel
compiled from our review of relevant literature. There are two changes to these charts relative to
the DRIA versions: 1) we updated estimates from GREET-2021 to GREET-2022, and 2) we
updated from AR5 to AR6 GWP values, except the BEIOM estimate is still based on AR5 as we
do not have results by gas for BEIOM. Updating the estimates from GREET-2021 to GREET-
2022 increases the LCA estimates by 0.4 gCChe/MJ for gasoline and by 0.9 gCChe/MJ for diesel.
Updating from AR5 to AR6 increased the other LCA estimates by less than 0.1 gC02e/MJ.
230 Masnadi, M. S., et al. (2021). "Carbon implications of marginal oils from market-derived demand shocks."
Nature 599(7883): 80-84.
131
-------
Figure 4.2.2.2-1: Petroleum Gasoline Lifecycle GHG Estimates
0 25 50 75 1 00
Petroleum Gasoline GHG Emissions (gC02e/MJ)
Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by In order, these
descriptors are the author or other name (e.g., RFS2 rule) and a brief descriptor of the scenario modeled. The
Upstream stage includes all of the emissions associated with extracting, handling and delivering crude oil to the
refinery gate. The Conversion stage includes emissions associated with refining. The Downstream stage includes
emissions associated with gasoline distribution and tailpipe combustion emissions. All values in this chart use 100-
year AR6 GWP values, except for BEIOM which uses AR5.
132
-------
Figure 4.2.2.2-2: Petroleum Diesel Lifecycle GHG Estimates
0 25 50 75
Petroleum Diesel GHG Emissions (gC02e/MJ)
Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by In order, these
descriptors are the author or other name (e.g., RFS2 rule) and a brief descriptor of the scenario modeled. The
Upstream stage includes all of the emissions associated with extracting, handling and delivering crude oil to the
refinery gate. The Conversion stage includes emissions associated with refining. The Downstream stage includes
emissions associated with diesel distribution and tailpipe combustion emissions. All the values in this chart use 100-
year AR6 GWP values, except BEIOM which uses AR5.
The 2010 RFS2 estimates were largely based on a study by the National Energy
Technology Laboratory (NETL).231 A team of NETL researchers published new estimates of the
lifecycle GHG emissions associated with 2005 and 2014 average U.S. gasoline and diesel
(Cooney et al. 2017). For 2005 average diesel, the Cooney et al. (2017) estimates are very similar
to our estimates for the 2010 RFS2 rule. For 2005 average gasoline the Cooney et al. (2017)
estimates are higher by approximately 4 gCChe/MJ. The Cooney et al. (2017) estimates for 2005
average gasoline and diesel values represent the high end of the range used in our illustrative
GHG impacts assessment.
231 U.S. EPA, (2010). 2005 Petroleum Baseline Lifecycle GHG (Greenhouse Gas) Calculations. U.S. Enviromnental
Protection Agency, Docket Item No. EPA-HQ-OAR-2005-0161-3151. Washington DC, January.
133
-------
The GREET-2022 model estimates lifecycle GHG emissions for average U.S. gasoline
and diesel. The GREET estimates for average gasoline and diesel are lower than the estimates
from RFS2 rule (2010) and Cooney et al. (2014) but all of these estimates are within 5 gCChe/MJ
of each other. The figures above also report results from a study with the BEIOM model
(Avelino et al. 2021), an "environmentally extended input-output model." The BEIOM estimates
are significantly lower than those from the other studies. BEIOM's methodology differs
significantly from the other studies in our review and is limited in geographic scope to the United
States, which may explain its lower estimates for the carbon intensity of gasoline and diesel.
Based on our review of published estimates, we use a range of 84 to 98 gC02e/MJ for gasoline
and for diesel we use a range of 84 to 94 gC02e/MJ.
4.2.2.3 Corn Starch Ethanol
More studies have been published on the GHG emissions associated with corn starch
ethanol than any of the other biofuel pathways considered for this rule. Our literature review
includes 9 studies that estimate the lifecycle GHG emissions associated with corn ethanol. Many
of these studies include multiple emissions estimates based on different assumptions about the
energy efficiency of dry mill ethanol production, co-products and other factors. Some of these
studies report a large number of estimates. Figure 4.2.2.3-1 below includes 19 estimates from
these studies that are representative of the range of results that each of them reports.
134
-------
Figure 4.2.2.3-1: Corn Starch Ethanol Lifecycle Greenhouse Gas Estimates
Lark et al. (2022)/Other/RFS2 RIA -
Brandao (2022) -
Lark et al. (2022)/Other/CA-LCFS -
CARB (2018)/GTAP-BIO+AEZ-EF/Dry Mill/High LUC -
RFS2 rule (2010)/FASOM-FAPRI/NG Dry DDGS/High LUC -
Lark et al. (2022)/Other/GREET -
CARB (2018)/GTAP-BIO+AEZ-EF/Dry Mill/Mean LUC -
RFS2 rule (2010)/FASOM-FAPRI/2022 Avg NG Dry Mill/Mean LUC -
BEIOM (2021)/Avg. Dry Mill -
CARB (2018)/GTAP-BIO+AEZ-EF/Dry Mill/Low LUC -
Scully et al. (2021)/GTAP-BIO+CCLUB/High LUC -
GREET (2022)/GTAP-BIO+CCLUB/NG Dry Mill Dry DGS/Default-
GREET (2022)/GTAP-BIO+CCLUB/Avg. Plant/Default-
Lewandrowski et al. (2019)/FASOM+GTAP-BIO/2022 BAU -
Scully et al. (2021 )/GTAP-BIO+CCLUB/Central LUC -
RFS2 rule (2010)/FASOM-FAPRI/Adv. NG Dry Mill/Low LUC -
GREET (2022)/GTAP-BIO+CCLUB/NG Dry Mill Wet DGS/Default-
Lee etal. (2021)/GTAP-BIO+CCLUB/2019-
Scully et al. (2021)/GTAP-BIO+CCLUB/Low LUC -
Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by In order, these
descriptors are the author or other name (e.g., RFS2 rule) and year of the study; the model used to estimate the LUC
emissions; the type of natural gas-fired dry mill used for ethanol production (e.g., 2022 Avg. NG Dry Mill); the
LUC estimate case (e.g.. Low LUC): and the non-LUC estimate case (e.g.. Low CI). The Upstream stage includes
all of the emissions associated with corn production and transport upstream of the ethanol production facility. The
Conversion stage includes emissions associated with fuel production at the ethanol production facility. The
Downstream stage includes emissions associated with ethanol transport and non-CO: combustion emissions. The
LUC stage includes emissions from induced land use changes. For studies that do not report disaggregated results,
results are reported as LUC and Non-LUC emissions.
We include estimates for different natural gas-fired dry mill configurations from RFS2
(2010) and GREET-2022. We include the high, low and mean land use change GHG estimates
from RFS2 (2010). The CARB (2018) estimates are based on the default assumptions in the most
recent version of the CA-GREET model (version 3.0), a version of GREET developed by CARB
for the CA-LCFS. We include a range of land use change emissions for CARB (2018) based on
the CARB (2014) report describing the indirect land use change modeling that continues to be
used for CA-LCFS implementation. Lewandrowski et al. (2019) is a study that attempts to
update the RFS2 (2010) estimates based on more recent data and swaps in land use change
135
-------
estimates from GTAP-BIO. Lee et al. (2021) uses GREET to estimate U.S. corn ethanol carbon
intensity from 2005 to 2019. We include the estimate for 2019 and add the default land use
change estimate from GREET. The lowest estimate is from Scully et al. (2021), a review paper
that developed a range of LCA estimates by combining elements of prior studies. The highest
estimates are from Lark et al. (2022), a study that modeled historical U.S. land use change GHG
emissions attributable to corn ethanol and added these estimates to the LCA estimates from RFS
(2010), CARB (2018) and GREET.232 We also include the estimate from BEIOM (2021) which
uses economic input-output methodology (see Chapter 4.2.2.6 for more information). Finally, we
include the estimate from Brandao (2022), a retrospective consequential lifecycle analysis of the
GHG emissions associated with ramping up ethanol production to 15 billion gallons from 1999-
2018.
Among the estimates in the above figure, upstream emissions range from 9 to 51
gC02e/MJ. These include emissions associated with feedstock production and transport,
including non-LUC market-mediated impacts in the agricultural sector. BEIOM reports the
highest upstream emissions, with the next highest estimate being 25 gCChe/MJ from GREET
(2021). The lowest estimate comes from Scully et al. (2021), which includes a relatively large
credit (-13.5 gC02e/MJ) for DGS displacing other sources of livestock feed.
We include estimates for ethanol produced at a U.S. dry mill facility using natural gas
and electricity for energy,233 as dry mills produce over 90% of U.S. fuel ethanol and natural gas
and electricity account for almost all of the energy use at these facilities.234 Among the studies in
Figure 4.2.2.2-1, conversion emissions range from 13 to 33 gC02e/MJ.235 The highest estimates
are from CARB (2018) based on the default assumptions used in the CA-GREET3.0 model. The
GREET-2022 estimate for "industrial average" dry mill corn ethanol production is 29
gC02e/MJ, and the RFS2 (2010) estimate for a projected 2022 natural gas dry mill facility is 26
gC02e/MJ. The lowest GHG estimate of 13 gC02e/MJ for the fuel production stage comes from
the most advanced natural gas-fired dry mill facility evaluated in the 2010 RFS2 rule. This
advanced facility includes wet DGS, corn oil fractionation, combined heat and power (CHP) and
membrane separation technologies.236
The largest source of variation between estimates are the land use change emissions,
ranging from -1 to 65 gC02e/MJ in Figure 4.2.2.3-1. The highest land use change estimates are
from Lark et al. (2021), which produced estimates of U.S. land use change emissions attributable
232 Lark et al. caveat that incorporating their U.S. land use change emissions into other fuel program estimates is
only a partial analysis, and that to accurately assess the carbon intensity of corn ethanol, a full reanalysis is needed
to ensure consistent treatment and systems boundaries.
233 In accordance with CAA 211(o)(2)(A)(i) renewable fuel production from facilities that commenced construction
prior to December 19. 2007, are exempt from the 20% GHG reduction requirement to qualify as renewable fuel
under the RFS program. Our review in this section focuses on average dry mill corn ethanol production in the U.S.
regardless of facility status pursuant to this "grandfathering" exemption.
234 Lee, U., et al. (2021). "Retrospective analysis of the US corn ethanol industry for 2005-2019: implications for
greenhouse gas emission reductions." Biofuels, Bioproducts and Biorefining
235 This excludes the studies that do not report disaggregated non-LUC emissions.
236 Including this facility in our review also allows us to include a wider range of RFS2 (2010) estimates which is
beneficial for the 30-year illustrative scenario as the RFS2 (2010) estimates are the only ones that report a
differentiated 30-year stream of annual emissions. Without this estimate the range used for the illustrative scenario
would be further from the full range identified in the literature.
136
-------
to corn ethanol and added them to non-U.S. land use change emissions estimates from RFS2
(2010), GREET and CARB. Scully et al. (2021) reports negative land use change emissions in
part because they assumed that planting annual crops on land categorized as cropland pasture
would result in net sequestration of soil carbon.237 For more discussion of corn ethanol land use
change estimates see DRIA Chapter 4.2.2.238
Downstream emissions range from 1 to 7 gC02e/MJ. Downstream emissions are
associated with fuel distribution from ethanol production facilities to retail gasoline stations and
tailpipe emissions. Some studies also include emissions from production and use of denaturant
which is added to ethanol in small volume percentages to render it undrinkable. The highest
downstream estimate is from Scully et al. (2021) including emissions associated with denaturant.
All of the other estimates are 4 gC02e/MJ or less.
Overall, our literature review of estimates representative of average corn ethanol
production at natural gas-fired U.S. dry mills produces a range from 38 to 116 gC02e/MJ. The
largest source of variation across studies continues to be estimated emissions associated with
direct and indirect land use change. Although this is already a wide range, corn ethanol can have
higher or lower GHG emissions depending on farm specific or facility specific factors.
There are plausible scenarios whereby corn ethanol currently being produced at particular
facilities or under certain conditions may be associated with greater lifecycle GHG emissions
than the upper end of the range formed by our compilation of literature LCA estimates. For
example, the GHG emissions could be higher for ethanol currently being produced from facilities
fired with coal (or that use electricity produced from coal-fired power plants), ethanol produced
from corn grown on marginal lands with lower yields, or in a scenario where prolonged drought
or other factors substantially lowers crop yields. More generally, there are also different
analytical approaches that produce higher estimates of the GHG impacts of corn ethanol
production, such as estimates that consider the "carbon opportunity cost" of the land used to
grow the corn feedstock, as proposed by Searchinger et al. (20 1 8).239 This study estimates these
opportunity costs two different ways: 1) As the average direct land use change emissions
associated with producing corn across the globe, or 2) As the foregone sequestration associated
with not devoting the same land to regenerating forest. Based on the estimates in Searchinger et
al. (2018), the carbon opportunity cost of corn ethanol using either approach is approximately
160 gCChe/MJ.240 We do not include this estimate in our compilation of LCA values, as it is
237 Spawn-Lee, S. A., et al. (2021). "Comment on 'Carbon Intensity of corn ethanol in the United States: state of the
science'." Environmental Research Letters 16(11): 118001.
238 The DRIA for the proposed rule includes a discussion of available models and land use change estimates that is
not part of this RIA for the final rule. The review of studies and land use change estimates in the DRIA remains
relevant, but we determined it did not bear repeating in this document as it does not factor directly into our analysis
of the climate impacts of the candidate volumes.
239 Searchinger, T. D., Wirsenius, S., Beringer, T., & Dumas, P. (2018). Assessing the efficiency of changes in land
use for mitigating climate change. Nature, 564(7735), 249-253.
240 Ibid., Extended Data Table 3 using either the COC "loss" method or the COC "gain" method.
137
-------
presented as a "carbon opportunity cost", "carbon cost" or "consumption cost," not an LCA
941
estimate.
On the other hand, corn ethanol produced with the adoption of advanced technologies or
climate smart agricultural practices can have lower LCA emissions. Corn ethanol facilities
produce a highly concentrated stream of C02 that lends itself to carbon capture and sequestration
(CCS). CCS is being deployed at ethanol plants and has the potential to result in negative
emissions at the ethanol production facility, especially if mills with CCS use renewable sources
of electricity and other advanced technologies to lower their needs for thermal energy.242 Climate
smart practices are being adopted at the feedstock production stage. For example, planting cover
crops between corn rotations can build soil organic carbon stocks. Collecting data on and
evaluating these trends in corn and ethanol production are areas for additional effort that will
inform future LCA estimates for corn ethanol.
4.2.2.4 Soybean Oil Biodiesel
Relative to corn ethanol, there have been fewer studies published on the GHG emissions
associated with soybean oil biodiesel. Our literature review includes 7 studies that estimate the
lifecycle GHG emissions associated with soybean oil biodiesel. The figure below includes 19
estimates of the lifecycle GHG emissions associated with soybean oil biodiesel production and
use.
241 Although Searchinger et al. (2018) does not present the carbon opportunity cost (COC) as an LCA estimate, the
paper states that under a particular set of assumptions the COC could be treated as an ILUC estimate (p. 251, "Our
biofuel COC estimates are equivalent to ILUC estimates if crops diverted to biofuels (after deducting by-products)
are fully replaced at the average global carbon loss per kilogram of crop.") Furthermore, Figure 2 in this paper
combines the COC estimates with production emissions to compare the total "carbon costs" of different fuel sources,
including corn ethanol and gasoline. These estimates could potentially be compared with LCA values, but for this
RIA we believe it is more appropriate to consider them separately from the rest of the LCA literature. Treating the
COC estimates as equivalent to ILUC requires a special set of assumptions that no other study in the literature
includes. The other studies in the literature use economic models to estimate the ILUC associated with a change in
biofuel production and use relative to a business as usual baseline scenario. In contrast, Searchinger et al. (2018)
either assumes, 1) that land dedicated to biofuel production would otherwise be used for forest regeneration, or 2)
that biofuel feedstocks are "fully replaced at the average global carbon loss per kilogram of crop." In our view, these
assumptions make the Searchinger et al. (2018) estimates categorically different from other LUC estimates in the
literature, supporting our choice to consider them separately.
242 Yoo, E., Lee, U., & Wang, M. (2022). Life-Cycle Greenhouse Gas Emissions of Sustainable Aviation Fuel
through a Net-Zero Carbon Biofuel Plant Design. ACS Sustainable Chemistry & Engineering, 10(27), 8725-8732.
138
-------
Figure 4.2.2.4-1: Soybean Oil Biodiesel Lifecycle Greenhouse Gas Estimates
CARB (2018)/GTAP-BIO+AEZ-EF/High LUC -
RFS2 rule (2010)/FASOM-FAPRI/High LUC -
CARB (2018)/GTAP-BIO+AEZ-EF/Mean LUC-
BEIOM (2021)/USA Avg. -
Xu etal. (2021 )/GTAP-BIO+AEZ-EF/USA avg./Mass/CARB (2014)-
Knoope et al. (2018)/High -
GREET (2022)/GTAP-BIO+CCLUB/Energy/Avg. Proxy -
Chen et at. (2018)/GTAP-BIO+AEZ-EF/GTAP 2011/High Non-LUC -
CARB (2018)/GTAP-BIO+AEZ-EF/Low LUC -
Chen et al. (2018)/GTAP-BIO+AEZ-EF/GTAP 2011/Central Non-LUC -
RFS2 rule (2010)/FASOM-FAPRI/Mean LUC -
Knoope et al. (2018)/Average -
GREET (2022)/GTAP-BIO+CCLUB/Mass/Case 8-
Xu et al. (2021 )/GTAP-BIO+CCLUB/USA avg./Mass/CCLUB -
GREET (2022)/GTAP-BIO+CCLUB/Mass/GTAP 2011 -
Chen et al. (2018)/GTAP-BIO+CCLUB/GTAP 2011/Central Non-LUC -
Knoope et al. (2018)/Low-
Chen et al. (2018)/GTAP-BIO+CCLUB/GTAP 2011/Low Non-LUC -
RFS2 rule (2010)/FASOM-FAPRI/Low LUC -
Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by In order, these
descriptors are the author or other name (e.g., RFS2 rule) and year of the study; the model used to estimate the LUC
emissions: the allocation approach used for soybean meal (e.g., mass, energy); the LUC estimate case (e.g.. Low
LUC); and the non-LUC estimate case (e.g.. Low CI). The Upstream stage includes all of the emissions associated
with soybean oil production and transport upstream of the biodiesel production facility. The Conversion stage
includes emission associated with fuel production at the biodiesel production facility. The Downstream stage
includes emissions associated with biodiesel transport and non-CO: combustion emissions. The LUC stage includes
emissions from induced land use changes. For studies that do not report disaggregated results, results are reported as
LUC and Non-LUC emissions.
RFS2 (2010) estimated uncertainty in land use change GHG emissions and reported a
relatively wide range of estimates. The only estimate in our review that is outside the range of
estimates from RFS2 (2010) is from CARB (2018), using CARB's high estimate for soy
biodiesel land use change emissions from CARB (2014). GREET-2022 allows users to choose
from land use change results from three different GTAP-BIO runs, and four different allocation
approaches to account for soybean meal coproduct. We include three estimates from GREET-
2022 based on different combinations of these factors to provide a representative range of
estimates from GREET-2022. Chen et al. (2018) used the GREET and the GTAP-BIO models to
139
-------
estimate soy biodiesel carbon intensity. Chen et al. (2018) reports a range of estimates based on
sensitivity analysis of the GREET input parameters and prior land use change GHG estimates
based on GTAP-BIO two different sets of land conversion emissions factors. Knoope et al.
(2018) is a GREET-style LCA study that excludes land use change GHG emissions and reports a
range of estimates based on sensitivity analysis of input parameters. We also include the estimate
from BEIOM (2021) which uses economic input-output methodology (see Chapter 4.2.2.6 for
more information).
Among the estimates in the above figure, upstream emissions range from 7 to 37
gC02e/MJ. These include emissions associated with feedstock production and transport,
including non-LUC market-mediated impacts in the agricultural sector. Upstream emissions
estimates depend on the methodology used to estimate them and the co-product accounting
methods applied to the soybean meal co-product. By default, GREET uses mass allocation for
the meal co-product, but GREET allows users to select market, energy or displacement
approaches. We include the mass- and energy-based allocation approaches in the figure above.243
The energy-based allocation approach produces a higher estimate (18 gCChe/MJ) than the mass-
based approach (9 gCChe/MJ). The lowest overall estimate for upstream emissions is from RFS2
(2010). RFS2 (2010) is the only study in our review that uses a consequential modeling approach
for non-land use change emissions, whereby the GHG impacts were modeled using agricultural-
economic models. The highest estimate is from BEIOM which is also unique in its modeling
approach. All of the other studies use an attributional approach to estimate upstream GHG
emissions, and most of them apply a mass-based allocation approach to account for the soybean
meal co-product.
We include estimates for biodiesel produced at U.S. facilities that use a transesterification
process. The range of conversion emissions in Figure 4.2.2.4-1 range from -1 to 15 gC02e/MJ.
Most of the fuel production estimates are from 8-12 gC02e/MJ. The RFS2 (2010) estimate is -1
gC02e/MJ based on the assumption that the glycerin co-product from biodiesel production is
burned for thermal process energy displacing the use of petroleum residual oil. Most of the other
studies use energy, market, or mass-based allocation to account for the glycerin co-product
which results in a larger estimate for fuel production GHG emissions.
Similar to corn ethanol, the largest source of variation between soybean oil biodiesel
LCA estimates are the land use change emissions, ranging from 5 to 64 gC02e/MJ in Figure
4.2.2.4-1.244 The highest and lowest land use change GHG estimates are from RFS2 (2010)
based on the upper and lower bounds of the reported 95% interval. The land use change
uncertainty analysis for RFS2 (2010) considered uncertainty in land conversion types and
emissions factors but did not consider uncertainty in economic model parameters. As discussed,
in DRIA Chapter 4.2.2.8,245 the range of soybean oil biodiesel land use change GHG estimates in
243 Using displacement results in upstream emissions of -17 gC02e/MJ and total LCA emissions of 38 gC02e/MJ.
Although the inclusion of the displacement method provides interesting variation in the estimates for each stage the
overall estimate is near the middle of the range, thus we exclude it from Figure 4.2.3.3 to improve legibility.
244 This excludes Knoope et al. (2021) and BEIOM (2021) which exclude land use change emissions.
245 The DRIA for the proposed rule includes a discussion of available models and land use change estimates that is
not part of this RIA for the final rule. The review of studies and land use change estimates in the DRIA remains
relevant, but we determined it did not bear repeating in this document as it does not factor directly into our analysis
of the climate impacts of the candidate volumes.
140
-------
the literature is wider when we consider studies that only estimate land use change emissions,
ranging from 5 to 80 gCChe/MJ.
Particular characteristics of soybean oil biodiesel production introduce greater potential
for uncertainty relative to corn ethanol. For example, the quantity of biofuel that can be produced
from an acre of U.S. soybeans is substantially smaller than that produced from an acre of U.S.
corn. Based on data from USD A and GREET, an average acre of U.S. farmland yields about four
times as much corn ethanol as soybean oil biodiesel on an energy basis.246 The difference in per-
acre fuel yield means that soybean oil biodiesel modeling results are far more sensitive to
tradeoffs between cropland extensification and other means of obtaining additional soybean oil.
For example, every acre of cropland extensification projected in a given soybean oil biodiesel
scenario represents four times as much new cropland per megajoule of biodiesel relative to a
corn ethanol scenario.
Another key sensitivity for soybean oil is the impact on livestock markets. While soybean
oil represents about 20 percent, by mass, of the crush product of soybeans, meal represents about
80 percent of that crush. Soybean meal is an important source of protein in livestock feed diets.
Corn ethanol also has a livestock feed co-product in the form of distillers grains. On a weight
basis, one megajoule of soybean oil biodiesel is associated with approximately 4 times as much
feed coproduct as one megajoule of corn ethanol.247 Soybean meal and distillers grains are used
differently in feed rations and their nutritional contents are also not identical. However, the
general comparison of the mass of each co-product demonstrates that the quantity of feed
product associated with a given quantity of soybean oil biodiesel is substantially greater than that
associated with the same quantity of corn ethanol. The impact of biofuel feed coproducts on
GHG emissions is highly complex. As a brief example, to the extent greater production of feed
coproducts allows cattle producers to intensify production, reducing the use of grazing lands,
these feed coproducts may mitigate LUC emissions. Conversely, to the extent that greater
availability of feed products reduces costs for livestock producers, this may lead to greater
livestock production and increased livestock-related emissions. It is unclear which of these
market dynamics may prove dominant in the future, making the net signal of livestock emissions
highly uncertain. The larger quantities of feed coproducts associated with the production of
soybean oil biodiesel relative to corn ethanol amplify this uncertainty at both ends of the
emissions range, contributing to the wider overall range of GHG impacts we observe in the
literature.
The markets for soybean oil and meal have changed significantly over time, which may
be a source of variation in soybean oil biodiesel GHG estimates in the literature. Soybean oil
246 Assumes soybean yield of 3,084 lbs/acre and corn yield of 9,912 lbs/acre, based on 2021 average yields from
USDA NASS. Assumes 93.6 lbs of soybean oil per MMBTU of biodiesel output and 163.7 lbs of corn per MMBTU
of ethanol based on GREET-2021.Thus, one acre yields 6.26 MMBTU (128.6 gallons ethanol-equivalent) of
soybean oil biodiesel, or 60.6 MMBTU (506.2 gallons of ethanol) of corn ethanol. USDA NASS data from
QuickStats database, https://anickstats.nass.nsda.gov.
247 For every lb of soybean oil produced, approximately 5.26 lbs of soybean meal are produced. At GREET-2021
average biodiesel yields, production of the one MMBTU of soybean oil biodiesel is associated with the production
of approximately 251.4 lbs of soybean meal. For comparison, according to GREET-2021, production of one
MMBTU of corn ethanol using the aforementioned dry mill ethanol process coproduces about 60.4 lbs of dried
distillers grains (assuming 100% drying).
141
-------
demand and prices have grown dramatically in recent years. Surging demand for soybean oil,
including for biofuel, drove soybean oil prices to record levels in 2021-2022. In past decades,
soybean oil supply was primarily dictated by demand for soybean meal. Now the USD A reports
that "U.S. supplies of soybean oil will largely depend on expanding the capacity of the U.S.
soybean crushing industry."248 Whereas meal used to be the primary driver of soybean crushing,
USD A now reports that, "Higher soybean oil prices may improve the profitability of soybean
processing in the near term; however, an excess supply of meal may cut into profits."249
Another key sensitivity present in the literature is uncertainty about the source of soybean
oil and biodiesel supplies. The U.S. has a dominant position in the global corn and corn ethanol
markets. The U.S. is also a major producer of soybeans, but other countries such as Brazil and
Argentina are competitive with the U.S. in the global market for growing soybeans. Furthermore,
the vegetable oil markets is highly integrated internationally. For example, soybean oil is just one
part of the larger global market for vegetable oils and is often substituting with other vegetable
oils, including palm oil in many use cases. Given the potential for marginal palm oil
extensification into carbon-rich peat lands in Southeast Asia,250 the extent to which palm oil
backfills for soybean oil diverted to biofuel production from other uses also substantially impacts
LUC emissions. In addition, increased demand for soybean oil in the U.S. could lead to land
conversion in other large soybean-producing countries such as Argentina and Brazil that are
home to carbon-dense tropical ecosystems. Therefore, the potential for these types of impacts on
sensitive high-carbon lands creates additional uncertainty in soybean oil biodiesel GHG
modeling.
Downstream emissions range from 1 to 5 gCChe/MJ. Downstream emissions are
associated with fuel distribution from biodiesel production facilities to retail gasoline stations
and tailpipe emissions. The highest downstream estimates are from GREET and the lowest are
from RFS2 (2010).
Overall, our compilation of estimates representative of average U.S. soybean oil biodiesel
production provides a range from 14 to 73 gCChe/MJ. The largest source of variation across
studies continues to be estimated emissions associated with induced land use change. Although
this is a wide range, biodiesel produced under particular conditions may produce emissions that
are higher or lower than this range on a per MJ basis based on particular factors.
There are plausible scenarios whereby soy biodiesel currently being produced at
particular facilities or under certain conditions may be associated with greater lifecycle GHG
emissions than the upper end of the range formed by our compilation of literature LCA
estimates. For example, LCA emissions may be higher if economic conditions result in soybean
oil used for biodiesel to be backfilled with palm oil or soybeans grown in tropical regions with
high rates of deforestation. More generally, there are also different analytical approaches that
248 Ates, A.M., and Bukowski, M., (2022). "Examining Record Soybean Oil Prices in 2021-22." Amber Waves.
USDA Economic Research Service. December 21, 2022. https://www.ers.usda.gov/amber-
waves/2022/december/examining-record-sovbean-oil-prices-in-202.1.-22.
249 Ibid.
250 See for example: Austin, K. G., et al. (2017). "Shifting patterns of oil palm driven deforestation in Indonesia and
implications for zero-deforestation commitments." Land Use Policy 69: 41-48.
142
-------
produce higher estimates of the GHG impacts of soy biodiesel production, such as estimates that
consider the "carbon opportunity cost" of the land used to grow the soybean oil feedstock, as
proposed by Searchinger et al. (2018).251 This study estimates these opportunity costs two
different ways: 1) as the average direct land use change emissions associated with producing
soybeans across the globe, or 2) as the foregone sequestration associated with not devoting the
same land to regenerating forest. Based on the estimates in Searchinger et al. (2018), the carbon
opportunity cost of soy biodiesel is 300 gCChe/MJ with the first approach, or 270 gCChe/MJ
using the second approach.252 We do not include this estimate in our compilation of LCA values,
as it is presented as a "carbon opportunity cost", "carbon cost" or "consumption cost," not an
LCA estimate.253
It may also be possible to produce soybean oil biodiesel with lower LCA emissions with
the adoption of climate smart agricultural practices. For example, planting cover crops between
soybean rotations has the potential to build soil organic carbon stocks. Collecting data on and
evaluating these trends in soybean production and vegetable oil markets are areas for additional
research that will inform future LCA estimates for soybean oil biodiesel.
4.2.2.5 Soybean Oil Renewable Diesel
Relative to soybean oil biodiesel, there have been fewer studies published on the GHG
emissions associated with soybean oil renewable diesel. Our literature review includes estimates
from 5 sources that estimate the GHG emissions associated with soybean oil renewable diesel.
These studies include numerous estimates based on different scenarios for land use change and
assumptions related to co-product accounting. The figure below summarizes these estimates.
Relative to the DRIA version, the only changes are updating the GREET-2021 estimates to
GREET-2022.
251 Searchinger, T. D., Wirsenius, S., Beringer, T., & Dumas, P. (2018). Assessing the efficiency of changes in land
use for mitigating climate change. Nature, 564(7735), 249-253.
252 Ibid., Extended Data Table 3 using either the COC "loss" method or the COC "gain" method.
253 Although Searchinger et al. (2018) does not present the carbon opportunity cost (COC) as an LCA estimate, the
paper states that under a particular set of assumptions the COC could be treated as an ILUC estimate (p. 251, "Our
biofuel COC estimates are equivalent to ILUC estimates if crops diverted to biofuels (after deducting by-products)
are fully replaced at the average global carbon loss per kilogram of crop.") Furthermore, Figure 2 in this paper
combines the COC estimates with production emissions to compare the total "carbon costs" of different fuel sources,
including corn ethanol and gasoline. These estimates could potentially be compared with LCA values, but for this
RIA we believe it is more appropriate to consider them separately from the rest of the LCA literature. Treating the
COC estimates as equivalent to ILUC requires a special set of assumptions that no other study in the literature
includes. The other studies in the literature use economic models to estimate the ILUC associated with a change in
biofuel production and use relative to a business as usual baseline scenario. In contrast, Searchinger et al. (2018)
either assumes, 1) that land dedicated to biofuel production would otherwise be used for forest regeneration, or 2)
that biofuel feedstocks are "fully replaced at the average global carbon loss per kilogram of crop." In our view, these
assumptions make the Searchinger et al. (2018) estimates categorically different from other LUC estimates in the
literature, supporting our choice to consider them separately.
143
-------
Figure 4.2.2.5-1: Soybean Oil Renewable Diesel Lifecycle Greenhouse Gas Estimates
RFS2 rule (2010)/FASOM-FAPRI/CA-LCFS Avg./High LUC -
CARB (2022)/GTAP-BIO+AEZ-EF/Mass/High LUC/High CI -
CARB (2018)/GTAP-BIO+AEZ-EF/Mass/High LUC-
RFS2 rule (2010)/FASOM-FAPRI/CA-LCFS Avg./Mean LUC-
Xu etal. (2021)/GTAP-BIO+AEZ-EF/USAavg./Mass/CARB (2014)-
CARB (2018)/GTAP-BIO+AEZ-EF/Mass/Mean LUC-
Xu etal. (2021)/GTAP-BIO+AEZ-EF/USAavg./Mass/ICAO (2019)-
GREET (2022)/GTAP-BIO+CCLUB/Market/Avg. Proxy -
CARB (2022)/GTAP-BIO+AEZ-EF/Mass/Low LUC/Low CI -
CARB (2018)/GTAP-BIO+AEZ-EF/Mass/Low LUC-
GREET (2022)/GTAP-BIO+CCLUB/Mass/Case 8 -
Xu etal. (2021)/GTAP-BIO+CCLUB/USAavg./Mass/CCLUB-
GREET (2022)/GTAP-BIO+CCLUB/Mass/GTAP 2011 -
RFS2 rule (2010)/FASOM-FAPRI/CA-LCFS Avg./Low LUC-
Stage
LUC
Non-LUC
Downstream
Conversion
Upstream
0 25 50 75
Soybean Oil Renewable Diesel GHG Emissions (gC02e/MJ)
Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by In order, these
descriptors are the author or other name (e.g., RFS2 rule) and year of the study; the model used to estimate the LUC
emissions; the allocation approach used for soybean meal (e.g., mass, energy); the LUC estimate case (e.g.. Low
LUC); and the non-LUC estimate case (e.g.. Low CI). The Upstream stage includes all of the emissions associated
with soybean oil production and transport upstream of the renewable diesel production facility. The Conversion
stage includes emission associated with fuel production at the renewable diesel production facility. The Downstream
stage includes emissions associated with renewable diesel transport and non-CO; combustion emissions. The LUC
stage includes emissions from induced land use changes. For studies that do not report disaggregated results, results
are reported as LUC and Non-LUC emissions.
The estimates from RFS2 (2010) in the figure above are based on the "upstream" GHG
modeling for soybean oil from the 2010 RFS2 rule combined with estimates that EPA published
more recently for renewable diesel production and downstream fuel distribution and use.254
Similar to the review for soybean oil biodiesel, CARB (2018) provides a range of land use
change GHG estimates and GREET (2022) includes multiple land use change scenarios and co-
product allocation approaches for soybean meal. Given the relative scarcity of LCA estimates for
254 April 2022 Canola Oil Pathways NPRM (87 FR 22823, April 18, 2022).
144
-------
soybean oil renewable diesel, we also include the highest and lowest carbon intensities for
individual U.S. facilities as certified by CARB for the CA-LCFS (CARB 2022) using their
central land use change GHG estimates.
Among the estimates in the above figure, upstream emissions range from 8 to 14
gC02e/MJ. Upstream emissions vary for the same reasons discussed for soybean oil biodiesel,
including the methodology used to estimate them and the co-product accounting methods applied
to the soybean meal co-product. The lowest upstream emissions estimates are from RFS2
(20 1 0),255 and the highest upstream emissions are from CARB (2018). As discussed above for
soybean oil biodiesel, the RFS2 (2010) analysis uses an entirely different methodology than
GREET or CARB for estimating GHG emissions associated with feedstock production.
Our review includes estimates representative of U.S. renewable diesel production via a
hydrotreating process. The range of conversion emissions in Figure 4.2.2.5-1 range from 8 to 16
gC02e/MJ. This range does not include the facility-specific carbon intensities from CARB
(2022) as this source does not report carbon intensity disaggregated into lifecycle stages. The
lowest estimate is from CARB (2018) and the highest estimate is from GREET-2022 using
market-based allocation. The RFS2 (2010) estimates in the figure above use hydrotreating
processing data provided by CARB representing the average of renewable diesel production
facilities registered for the CA-LCFS as of June 2021.256 The estimate uses an energy allocation
approach to account for co-products of renewable diesel production. The lowest conservation
stage estimates come from CARB (2018) based on the default assumptions in CA-GREET
version 3.0.
Similar to soybean oil biodiesel, the largest source of variation between soybean oil
renewable diesel LCA estimates are the land use change emissions. The same factors, discussed
above, that introduce additional complexity into LUC modeling for soybean oil biodiesel also
apply to soybean oil renewable diesel. For renewable diesel the land use change GHG estimates
range from 6 to 67 gCChe/MJ in Figure 4.2.2.4-1.257 The highest and lowest land use change
GHG estimates are from RFS2 (2010) based on the upper and lower bounds of the reported 95%
confidence interval. As discussed, in DRIA Chapter 4.2.2.8,258 the range of soybean oil biodiesel
land use change GHG estimates in the literature is wider when we consider studies that only
estimate land use change emissions, ranging from 5 to 80 gC02e/MJ.
255 The renewable diesel upstream emissions from RFS2(2010) are lower than those for soybean oil biodiesel,
because we have updated the soybean oil upstream estimates for renewable diesel using more recent emissions
factors from GREET and AR5 GWP values. More details are provided in a technical memo to the docket titled
"Notes on Literature Review of Transportation Fuel Greenhouse Gas (GHG) Lifecycle Analysis (LCA)."
256 For more information on hydrotreating process data evaluated by EPA, see April 2022 Canola Oil Pathways
NPRM (87 FR 22823), Section II.C.9.
257 This excludes Knoope et al. (2021) and BEIOM (2021) which exclude land use change emissions.
258 The DRIA for the proposed rule includes a discussion of available models and land use change estimates that is
not part of this RIA for the final rule. The review of studies and land use change estimates in the DRIA remains
relevant, but we determined it did not bear repeating in this document as it does not factor directly into our analysis
of the climate impacts of the candidate volumes.
145
-------
Downstream emissions are associated with fuel distribution from renewable diesel
production facilities to retail gasoline stations and tailpipe emissions. For renewable diesel, there
is little variation in the review estimates, as they are all approximately 1 gCChe/MJ.
Overall, our literature review of estimates representative of average U.S. soybean oil
renewable production provides a range from 26 to 87 gCChe/MJ. This is a relatively wide range,
and the largest source of variation between studies continues to be estimated emissions
associated with induced land use change. Although this is a wide range, renewable diesel
produced under particular conditions may produce emissions that are outside of this range on a
per MJ basis. Some of the main factors that could result in emissions higher or lower than the
literature range are the same as those discussed above in the Chapter 4.2.2.7 for soybean oil
biodiesel. Renewable diesel carbon intensities could be reduced by consuming less hydrogen
during the conversion stage, or sourcing the hydrogen from low carbon sources.
It is worth noting that the International Civil Aviation Organization (ICAO) has been
conducting similar lifecycle GHG analysis in support of the Carbon Offsetting and Reduction
Scheme for International Aviation (CORSIA). Given ICAO's focus on jet fuel, they have not
specifically published an LCA value for soybean oil renewable diesel. However, the lifecycle
analysis for soybean oil jet fuel is similar to an LCA for soybean oil renewable diesel since both
are produced through the same hydrotreating process. While most current hydrotreating
processes yield renewable diesel with small amounts of naptha and LPG co-products, these
facilities can be configured to produce a separate jet fuel stream from the rest of the products
produced. Producing jet fuel requires additional refining, therefore jet fuel LCA estimates tend to
be slightly more GHG intensive than producing renewable diesel alone. The soybean oil jet fuel
results from ICAO (2021) are summarized in the table below.
Table 4.2.2.5-1: U.S. Soybean Oil Jet Fuel Estimates from ICAO (2021) (gCOie/MJ)
Estimate
Core259
Land Use Change
LCA Value
GLOBIOM LUC
40
14
54
(Low end of 95% CI)
GTAP-BIO LUC
40
20
60
ICAO Default
40
25
65
GLOBIOM LUC
40
50
91
GLOBIOM LUC
40
92
132
(High end of 95% CI)
Notes: For their default land use change estimate, ICAO uses the GTAP-BIO estimate plus 4.45 gC02e/MJ, see
ICAO (2021) p. 149 for explanation. The GTAP-BIO and central GLOBIOM land use change estimates are from
ICAO (2021) Table 67. The low and high GLOBIOM estimates are from ICAO (2021) Table 72. The low estimate
is the 2.5% quantile and the high estimate is the 97.5% quantile from sensitivity analysis (300 runs). LCA values in
table might not be the sum or core and LUC values due to rounding.
259 ICAO (2021) includes estimates from GREET of the direct, or "core," GHG emissions associated with jet fuel
produced from soybean oil through a hydrotreating process.
146
-------
Given the similarities in the hydrotreating process, it is a relatively straightforward
adjustment to modify a soybean oil jet fuel LCA to a soybean oil renewable diesel LCA.260 If the
ICAO soybean oil jet fuel LCA was adjusted for the lower energy needs for renewable diesel, the
ICAO estimates for soybean oil renewable diesel would be 50 to 128 gCChe/MJ. If the ICAO
LCA values were included in Figure 4.2.2.5-1 and Table 4.2.2.12, the overall range of values for
soybean oil renewable diesel would be wider (26 gCChe/MJ to 128 gCChe/MJ).
4.2.2.6 FOGBiodiesel
We reviewed literature on the GHG emissions associated with biodiesel produced from
fats, oils, and greases (FOG). Specifically, we reviewed estimates for biodiesel produced from
used cooking oil (UCO) and animal tallow. Figure 4.2.2.6-1 includes the LCA estimates for
UCO biodiesel, and Figure 4.2.2.6-2 includes the LCA estimates for animal tallow biodiesel.
Argonne National Laboratory added UCO biodiesel and renewable diesel as new pathways in
GREET-2022. Relative to the DRIA, we updated these figures here to include the GREET-2022
estimates.
260 ICAO's core GHG estimates are based on analysis with GREET using an energy allocation approach for co-
products. The GREET-2021 core GHG estimate for soybean oil renewable diesel using energy allocation is 36
gC02e/MJ. This GREET-2021 estimate can be substituted for the 40 gC02e/MJ core jet fuel value to produce an
LCA range for soybean oil renewable diesel.
147
-------
Figure 4.2.2.6-1: UCO Biodiesel Lifecycle Greenhouse Gas Estimates
CARB (2022)/Highest CI -
GD
RFS2 (2010) + O'Malley (2021) -
CARB (2018)-
RFS2 rule (2010) -
GREET (2022)-
Xu etal. (2021)/USA avg.-
i
1
4
d|
Stage
¦
LUC
Non-LUC
Downstream
Conversion
Upstream
CARB (2022)/Lowest CI -
GD
0 10 20 30
UCO Biodiesel GHG Emissions (gC02e/MJ)
Notes: The Upstream stage includes all of the emissions associated with UCO pre-treatment/rendering and transport
upstream of the biodiesel production facility. O'Malley et al. (2021) also includes indirect GHG emissions in the
Upstream stage. The Conversion stage includes emission associated with fuel production at the biodiesel production
facility. The Downstream stage includes emissions associated with biodiesel transport and non-CO; combustion
emissions. Estimates that only report total lifecycle GHG emissions are depicted only with a label and no bars.
Estimates for UCO biodiesel range from 12 to 32 gC02e/MJ. Given the relative scarcity
of LCA studies on UCO biodiesel, we include the highest and lowest certified carbon intensities
for individual biodiesel facilities under the CA-LCFS in our review. The highest and lowest
estimates come from the CA-LCFS range (CARB 2022). This is not surprising given that CARB
(2022) evaluates individual facilities whereas the other estimates represent a U.S. average. The
CARB and RFS2 estimates assume that the only upstream emissions for supplying UCO are
associated with rendering/cooking the raw UCO and transporting it to biodiesel production
facilities. O'Malley et al. (2021) looked at the current uses for UCO apart from biofuel
production and evaluated a case study where UCO diverted from livestock feed and
oleochemical uses is backfilled with corn, soybean oil and palm oil. Based on this case study,
they estimated potential indirect emissions of 12.2 gC02e/MJ associated with UCO use for
148
-------
biodiesel. In the figure above, we include the potential indirect emissions estimate from
O'Malley et al. (2021) added to the RFS2 (2010) estimates.
Figure 4.2.2.6-2: Animal Tallow Biodiesel Lifecycle Greenhouse Gas Estimates
GREET (2022) + O'Malley et al. (2021)/Energy-
CARB (2018)-
CARB (2022)/Highest CI -
Chen et al. (2018)/High-
GREET (2022)/Energy-
CARB (2022)/Lowest CI -
GREET (2022)/Market-
Chen et al. (2018)/NBB (2016)/Central-
GREET (2022)/Mass -
gd
Stage
¦
LUC
Non-LUC
Downstream
Conversion
Upstream
GREET (2022)/Displacement-
0 20 40 60
Tallow Biodiesel GHG Emissions (gC02e/MJ)
Notes: The Upstream stage includes all of the emissions associated with animal tallow pre-treatment/rendering and
transport upstream of the biodiesel production facility. O'Malley et al. (2021) also includes indirect GHG emissions
in the Upstream stage. The Conversion stage includes emission associated with fuel production at the biodiesel
production facility. The Downstream stage includes emissions associated with biodiesel transport and non-CO;
combustion emissions. For studies that do not report disaggregated results, results are reported as LUC and Non-
LUC emissions. Estimates that only report total lifecycle GHG emissions are depicted only with a label and no bars.
Estimates for tallow biodiesel range from 16 to 58 gCChe/MJ. Most of the estimates
assume that tallow is a byproduct of meat production and assume zero upstream emissions from
livestock production allocated to the tallow. For these estimates the ranges are primarily based
on different assumptions about the energy requirements for rendering, as well as different
assumptions about the co-products from rendering and the accounting methods for these co-
products. The exception is the case study by O'Malley et al. (2021) which estimates emissions of
34.8 gC02e/MJ associated with backfilling tallow used in livestock feed and oleochemical
149
-------
production with corn, soybean oil and palm oil. To inform our range of estimates we add this
backfill emissions estimate to the estimates from GREET-2022.
O'Malley et al. (2021) estimates the backfilling emissions associated with using tallow as
a biofuel feedstock are almost three-times greater than such emissions for UCO. The reason for
the difference appears to be that, according to O'Malley et al. (2021), approximately 83% of
UCO that is not currently used for biofuel is fed to swine and poultry, whereas a much greater
share of tallow is used in the oleochemical industry. They assume that vegetable oils (soybean oil
and palm oil) will substitute for oleochemical uses and cattle feed, and corn will substitute for
cattle feed. Thus, their analysis assumes tallow will primarily be replaced with soybean and palm
oil, while UCO will primarily be replaced with corn. In their GHG analysis, they assign greater
LCA emissions to soybean oil and palm oil than corn, which explains that higher GHG estimates
for tallow relative to UCO.
4.2.2.7 FOG Renewable Diesel
We reviewed literature on the GHG emissions associated with renewable diesel produced
from FOG. Specifically, we reviewed estimates for renewable diesel produced from used
cooking oil (UCO) and animal tallow. Figure 4.2.2.7-1 includes the LCA estimates for UCO
renewable diesel, and Figure 4.2.2.7-2 includes the LCA estimates for animal tallow renewable
diesel. Argonne National Laboratory added UCO biodiesel and renewable diesel as new
pathways in GREET-2022. Relative to the DRIA, we update these figures to include the
GREET-2022 estimates.
150
-------
Figure 4.2.2.7-1: UCO Renewable Diesel Lifecycle Greenhouse Gas Estimates
CARB (2022)/Highest CI -
RFS2 rule (2010)/CA-LCFS Avg. •
RFS2 rule (2010)/Pearlson et al. Energy-
@
Stage
¦
LUC
Non-LUC
Downstream
Conversion
Upstream
CARB (2018) -
CARB (2022)/Lowest CI -
GREET (2022)-
Xu et al. (2021 )/USA avg.-
Seber et al. (2014)/Pearlson et al. (2013)/High -
Seber et al. (2014)/Pearlson et al. (2013)/Base -
Seber et al. (2014)/Pearlson et al. (2013)/Low-
0 10 20 30
UCO Renewable Diesel GHG Emissions (gC02e/MJ)
Notes: The Upstream stage includes all of the emissions associated with UCO pre-treatment/rendering and transport
upstream of the biodiesel production facility. O'Malley et al. (2021) also includes indirect GHG emissions in the
Upstream stage. The Conversion stage includes emission associated with fuel production at the biodiesel production
facility. The Downstream stage includes emissions associated with biodiesel transport and non-CO; combustion
emissions. Estimates that only report total lifecycle GHG emissions are depicted only with a label and no bars.
Estimates for UCO renewable diesel range from 12 to 37 gCChe/MJ. The CARB and
RFS2 estimates assume that the only upstream emissions for supplying UCO are associated with
rendering/cooking the raw UCO and transporting it to biodiesel production facilities. O'Malley
et al. (2021) looked at the current uses for UCO apart from biofuel production and evaluated a
case study where UCO diverted from livestock feed and oleochemical uses is backfilled with
corn, soybean oil and palm oil. Based on this case study, they estimated potential indirect
emissions of 12.2 gC02e/MJ associated with UCO use for biodiesel. In the figure above, the
highest estimate is based on the sum of the potential indirect emissions estimate from O'Malley
et al. (2021) added to the RFS2 (2010) estimate. The lowest estimates are from Seber et al.
(2014), which do not include any backfilling emissions.
151
-------
Figure 4.2.2.7-2: Animal Tallow Renewable Diesel Lifecycle Greenhouse Gas Estimates
Seber et al. (2014)/Pearlson et al. (2013)/Coproduct High -
Seber et al. (2014)/Pearlson et al. (2013)/Coproduct Base -
Seber et al. (2014)/Pearlson et al. (2013)/Coproduct Low -
GREET (2022) + O'Malley et al. (2021 )/Energy -
Riazi (2020)/Aspen Plus/Poultry High/Mass -
CARB (2022)/Highest CI -
CARB (2018)-
Riazi (2020)/Aspen Plus/Beef High/Mass -
Seber etal. (2014)/Pearlson etal. (2013)/Byproduct High-
Riazi (2020)/Aspen Plus/Poultry Low/Mass -
Seber et al. (2014)/Pearlson et al. (2013)/Byproduct Base -
Riazi (2020)/Aspen Plus/Beef Low/Mass -
GREET (2022)/Energy -
Seber et al. (2014)/Pearlson et al. (2013)/Byproduct Low -
GREET (2022)/Market -
CARB (2022)/Lowest CI -
GREET (2022)/Mass -
Xu etal. (2021)/USAavg. -
GREET (2022)/Displacement-
Tallow Renewable Diesel GHG Emissions (gC02e/MJ)
Notes: The Upstream stage includes all of the emissions associated with tallow pre-treatment/rendering and
transport upstream of the biodiesel production facility. Seber et al. (2021) also includes livestock production GHG
emissions in the Upstream stage. The Conversion stage includes emission associated with fuel production at the
biodiesel production facility. The Downstream stage includes emissions associated with biodiesel transport and non-
CO2 combustion emissions. Estimates that only report total lifecycle GHG emissions are depicted only with a label
and no bars.
Estimates for tallow renewable diesel range from 14 to 80 gC02e/MJ. The highest
estimates are from Seber et al. (2014), which is the only study that includes scenarios where
GHG emissions associated with livestock raising and meat production are allocated the tallow. In
other words, in these scenarios the tallow is considered a co-product of meat production rather
than a byproduct. Our review also includes a case study by O'Malley et al. (2021) that evaluates
emission associated with backfilling tallow used in livestock feed and oleochemical production
with corn, soybean oil and palm oil. The lowest estimate is from GREET-2022 using a
displacement approach for the co-products from tallow rendering.
152
-------
4.2.2.8 Distillers Corn Oil Biodiesel
We reviewed published estimates of the GHG emissions associated with biodiesel
produced from distillers corn oil (DCO). DCO is a co-product from ethanol production whereby
oil is removed the DGS before it is sold as livestock feed. The DCO can then be used as a biofuel
feedstock or added back into livestock feed at desired levels. Figure 4.2.2.8-1 includes the LCA
estimates for DCO biodiesel.
Figure 4.2.2.8-1: DCO Biodiesel Lifecycle Greenhouse Gas Estimates
CARB (2022)/Highest CI -
0
RFS2 rule (2020)-
0 10 20 30
DCO Biodiesel GHG Emissions (gC02e/MJ)
Notes: The Upstream stage includes all of the emissions associated with DCO extraction and in some cases
backfilling with corn in livestock feed. The Conversion stage includes emission associated with fuel production at
the biodiesel production facility. The Downstream stage includes emissions associated with biodiesel transport and
non-COi combustion emissions. Estimates that only report total lifecycle GHG emissions are depicted only with a
label and no bars.
DCO biodiesel estimates range from 14 to 37 gC02e/MJ. Most of the estimates assume
that DCO is a byproduct and assign none of the emissions associated with corn or ethanol
production to it. For the 2020 RFS2 rule, we estimated the emissions associated with corn
153
-------
backfilling for DCO in animal feed. As discussed in that rule and a prior rule on distillers
sorghum oil, we determined that DCO is used as a source of energy/calories in feed diets and that
corn is a likely product to backfill when DCO is used as a biofuel feedstock. The lowest
estimates do not include emissions associated with backfilling corn or other products for the
DCO.
4.2.2.9 Distillers Corn Oil Renewable Diesel
We reviewed published estimates of the GHG emissions associated with renewable diesel
produced from DCO. Figure 4.2.2.9-1 includes the LCA estimates for DCO renewable diesel.
Figure 4.2.2.9-1: DCO Renewable Diesel Lifecycle Greenhouse Gas Estimates
CARB (2022)/Highest CI -
RFS2 rule (2020)-
CARB (2018) -
Stage
CARB (2022)/Lowest CI -
GE)
LUC
Non-LUC
Downstream
Conversion
Upstream
GREET (2022)-
Xu et al. (2021)/USA avg. -
0 10 20 30 40
DCO Renewable Diesel GHG Emissions (gC02e/MJ)
Notes: The Upstream stage includes all of the emissions associated with DCO extraction and in some cases
backfilling with corn in livestock feed. The Conversion stage includes emissions associated with fuel production at
the renewable diesel production facility. The Downstream stage includes emissions associated with biodiesel
transport and non-CO; combustion emissions. Estimates that only report total lifecycle GHG emissions are depicted
only with a label and no bars.
154
-------
DCO renewable diesel estimates range from 12 to 46 gCChe/MJ. Most of the estimates
assume that DCO is a byproduct and assign none of the emissions associated with corn or
ethanol production to it. For the 2020 RFS2 rule, we estimated the emissions associated with
corn backfilling for DCO in animal feed. The lowest estimate is from Xu et al. (2021), which
does not include backfilling emissions.
4.2.2.10 Natural Gas CNG
As discussed above for petroleum gasoline and diesel, for the purposes of conducting the
lifecycle GHG emissions analysis and determining which biofuels meet the GHG requirements,
CAA Section 21 l(o)(l)(C) defines baseline lifecycle greenhouse gas emissions as "the average
lifecycle greenhouse gas emissions, as determined by the Administrator, after notice and
opportunity for comment, for gasoline or diesel (whichever is being replaced by the renewable
fuel) sold or distributed as transportation fuel in 2005." While the baseline lifecycle GHG
emissions are used for a different specific purpose under the RFS program, we are not required
to use it here in this analysis for evaluating the GHG impacts of the candidate volumes.
To inform a range of potential GHG impacts associated with renewable CNG we
consider two scenarios for the conventional fuels it displaces. In the first scenario we assume the
candidate volumes of renewable CNG, relative to the No RFS baseline, cause some miles
traveled with diesel vehicles to be replaced with miles traveled with vehicles that run on
renewable CNG. This scenario assumes that the candidate volumes make CNG vehicles more
economically attractive than diesel vehicles in some cases, leading to a marginal increase in
CNG vehicle miles traveled relative to diesel vehicle miles traveled. In the second scenario, we
assume the candidate volumes of renewable CNG do not shift the relative miles traveled for
diesel vehicles relative to CNG vehicles, but instead cause CNG vehicles to be fueled with
renewable CNG instead of conventional CNG.
Thus, our literature review for this action includes studies that estimate the lifecycle GHG
emissions associated with natural gas CNG. Figure 4.2.2.9-1 includes the range of LCA estimate
for natural gas CNG identified in our review of the literature. Based on our review, LCA
estimates for diesel are higher than those for natural gas CNG on a per MJ of fuel basis. For the
illustrative 30-year GHG scenario discussed in Chapter 4.2.3, the scenario where renewable
CNG replaces diesel fuel produces the high estimate of the GHG benefits of renewable CNG.
The low estimate of renewable CNG GHG benefits is based on the scenario that assumes
renewable CNG displaces conventional CNG.
155
-------
Figure 4.2.2.10-1: Natural Gas CNG Well-to-Wheel Greenhouse Gas Estimates
CARB (2022)/Highest CI -
0
CARB (2022)/Lowest CI -
0
EPA and NHTSA (2016)-
GREET (2022)/Default CH4 Leak -
GREET (2022)/"EPA" CH4 Leak -
Stage
Non-LUC
Downstream
Conversion
Upstream
0 20 40 60 80
Natural Gas CNG GHG Emissions (gC02e/MJ)
Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by In order, these
descriptors are the author or other name (e.g., RFS2 rule) and a brief descriptor of the scenario modeled. The
Upstream stage includes all of the emissions associated with extracting, processing and delivering natural gas to a
compression facility. The Conversion stage includes emissions associated compressing natural gas to CNG. The
Downstream stage includes emissions associated with fueling a CNG vehicle and tailpipe combustion emissions.
Estimates that only report total lifecycle GHG emissions are depicted only with a label and no bars.
The natural gas CNG estimates in our review range from 73 to 81 gCChe/MJ. Our review
did not identify many applicable studies, as there are many more studies on the lifecycle
emissions associated with natural gas production than natural gas for CNG vehicles. The EPA
and NHTSA (2016) estimate is from the RIA for the Greenhouse Gas Emissions and Fuel
Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles - Phase 2 rule. It
represents lifecycle emissions for CNG used in a 2014 or later dedicated CNG vehicle. The
lowest estimates are from GREET-2022. By default, GREET-2022 assumes a set of assumptions
for methane leakage during natural gas production. The model also gives users the option of
choosing methane leakage assumptions derived by Argonne from the EPA GHG Inventory. The
highest estimates are the natural gas CNG pathways certified under the CA-LCFS program. Our
156
-------
review includes studies of the emissions associated with national average natural gas CNG. CNG
produced from natural gas that is in turn produced from wells or systems with high leakage rates
may have much greater carbon intensity than the estimates in our review. This is an area where
additional LCA research would be helpful.
4.2.2.11 Landfill Biogas CNG
Our literature review did not identify many studies on the lifecycle GHG emissions
associated with CNG produced from landfill gas. Our review is limited to estimates derived from
the GREET model and estimates by CARB as part of their implementation of the CA-LCFS
program. Figure 4.2.2.11-1 includes the LCA estimates for CNG produced from landfill biogas.
157
-------
Figure 4.2.2.11-1: Landfill Biogas CNG Lifecycle Greenhouse Gas Estimates
CARB (2018)/1% Leak/NG/Offsite-
CARB (2022)/Highest Cl/Offsite-
GREET (2022)/3% Leakage/Electricity/Offsite -
CARB (2022)/Lowest Cl/Offsite -
©
Stage
Non-LUC
Downstream
Conversion
Upstream
GREET (2022)12% Leakage/RNG/Onsite -
GREET (2022)/1% Leakage/RNG/Onsite -
-50 0 50
Landfill Gas CNG GHG Emissions (gC02e/MJ)
Notes: The Upstream stage includes all of the emissions associated with capturing landfill gas, processing it to
pipeline quality, and transporting it to the fueling location. Downstream emission include tailpipe emissions
including CO2 emissions. In most studies upstream emissions are negative because they assume landfill gas will be
flared in the counterfactual baseline scenario. Because reductions in CO2 emissions are included in the Upstream
emissions, CO2 tailpipe emissions are included in the Downstream emissions. CARB (2018) reports more
disaggregated results and excludes tailpipe CO2 emissions. Estimates that only report total lifecycle GHG emissions
are depicted only with a label and no bars.
The range of estimates in Figure 4.2.2.11-1 range from 6 to 70 gCChe/MJ. The higher
estimates are from CARB and the lower estimates are from GREET-2022. We varied two
parameters in GREET-2022 to provide a range of estimates. By default, GREET-2022 assumes a
2% methane leakage rate associated with processing landfill gas to pipeline quality. We include
estimates assuming 1% and 3% methane leakage and show that each 1% of additional methane
leakage increases the LCA estimates by about 6 gC02e/MJ. By default, GREET-2022 assumes
CNG fueling occurs offsite from where the landfill gas is produced, and the majority of CA-
LCFS certified LFG CNG pathways are for offsite CNG fueling. Based on GREET-2022, onsite
fueling reduces the LCA estimate by approximately 3 gCChe/MJ. GREET-2022 assumes that
158
-------
landfills will use renewable natural gas (RNG) as process fuel for gas cleanup and compression.
GREET allows users to assume that landfills will instead use grid electricity to power these
processes. Putting these factors together, the highest emissions scenarios that we include for
GREET-2022 assumes 3% methane leakage, electricity for process energy and offsite CNG
refueling.
The highest estimates are from CARB, with the very highest estimate coming from the
default CA-GREET version 3.0 model. GREET-2022 includes negative upstream emissions
based on reduced GHG emissions in a baseline scenario absent the use of the landfill gas as fuel.
The CA-GREET model has much larger upstream GHG emissions than GREET-2022 as it does
not include the relatively large emissions reductions relative to the baseline. Landfills with high
leakage rates associated with capturing and cleaning up the biogas may have much greater
carbon intensity than the estimates in our review. This is an area where additional LCA research
would be helpful.
4.2.2.12 Manure Digester CNG
Our literature review did not identify many studies on the lifecycle GHG emissions
associated with CNG produced from manure digester biogas. Our review is limited to estimates
derived from the GREET model and estimates by CARB as part of their implementation of the
CA-LCFS program. Figure 4.2.2.12-1 includes the LCA estimates for CNG produced from
landfill biogas.
159
-------
Figure 4.2.2.12-1: Manure Digester CNG Lifecycle Greenhouse Gas Estimates
GREET (2022)/Broiler/Offsite -
GREET (2022)/Layer/Offsite -
GREET (2022)/Dairy/Offsite -
0
GD
0
Stage
CARB (2022)/Highest Cl/Dairy-
EEI
LUC
Non-LUC
Downstream
Conversion
Upstream
GREET (2022)/Swine/Offsite -
CARB (2022)/Lowest Cl/Dairy- | -533 ]
400 -200 0
Manure Biogas CNG GHG Emissions (gC02e/MJ)
Notes: The Upstream stage includes emissions associated with operating the digester. The Conversion stage includes
emissions associated with processing the biogas to pipeline quality, and transporting it to the fueling location.
Downstream emissions include tailpipe emissions including CO2 emissions. In many of the studies, the upstream
emissions are negative as they include emissions avoided relative to a counterfactual scenario where the manure is
not treated in a digester. Because reductions in CO2 emissions are included in the Upstream emissions, tailpipe C02
emissions are included in the Downstream stage. Estimates that only report total lifecycle GHG emissions are
depicted only with a label and no bars.
There are relatively few studies but a very large range of LCA estimates (-533 to 52
gCChe/MJ) for biogas CNG produced from manure digesters. CARB has certified pathways for
CNG produced from over 50 different sources of manure biogas. All of these pathways have
negative carbon intensities meaning they reduce GHG emissions even before displacing any
conventional transportation fuels. The negative emissions are due to the assumed high methane
and nitrous oxide emissions in the baseline scenario absent the collection and treatment of animal
manure in anaerobic digesters. Consequently, the biggest area of uncertainty in the LCA for
manure digester GHG emissions is how the manure will be treated in the baseline scenario and
the associated emissions. Based on estimates from GREET-2022, CNG produced from broiler
manure biogas has positive carbon intensities, as high as 52 gC02e/MJ. The GREET manure
160
-------
biogas LCA estimates are extremely sensitive to the type of animal manure processed. Once
again, this is due to large variations in emissions in the counterfactual baseline scenario
assuming manure is treated without a digester. If the baseline emissions estimates were
harmonized, the LCA estimates would not be sensitive to animal type, as the emissions
associated with digester operations are similar for different types of manure.
4.2.2.13 Summary of LCA Ranges
Based on the literature review for each pathway discussed above, the range of LCA
estimates are summarized in Table 4.2.2.12-1.
Table 4.2.2.13-1: Lifecycle GHG Ranges Based on Literature Review (gCChe/MJ)
Pathway
LCA Range
Petroleum Gasoline
84 to 98
Petroleum Diesel
84 to 94
Natural Gas CNG
73 to 81
Corn Starch Ethanol
38 to 116
Soybean Oil Biodiesel
14 to 73
Soybean Oil Renewable Diesel
26 to 87
Used Cooking Oil Biodiesel
12 to 32
Used Cooking Oil Renewable Diesel
12 to 37
Tallow Biodiesel
16 to 58
Tallow Renewable Diesel
14 to 81
Distillers Corn Oil Biodiesel
14 to 37
Distillers Corn Oil Renewable Diesel
12 to 46
Landfill Gas CNG
6 to 70
Manure Biogas CNG
-533 to 52
In the sections that follow we present a range of monetized climate benefits associated
with the candidate volumes for an illustrative 30-year scenario. In order to appropriately
monetize GHG impacts over this period an annual stream of net GHG emissions is required. For
the non-crop based fuel pathways we assume a constant stream of GHG emissions per MJ over
the 30-year period. The land use change emissions associated with crop-based biofuels are highly
dynamic, as the majority of emission increases associated with land use changes occur relatively
quickly (e.g., in the first few years) with the reduced emissions associated with the biofuel use
occuring over time. Thus, for the 30-year illustrative scenario, we use estimates for crop-based
biofuels that report an annual stream of land use change emissions. The majority of the land use
change GHG estimates in the literature do not report an annual stream. In many cases, these LUC
estimates are derived by estimating land conversions induced by the crop-based biofuel
production and then multiplying these conversions by emissions factors that estimate the
resulting total emissions over a 20-30 year period. The only study identified in our review that
does report an annual stream of land use change emissions is the analysis for the 2010 RFS2 rule.
The reasons that no other studies report annual emissions are not entirely clear, but many studies
use static models to estimate land use change that are not conducive to reporting annual
emissions. Other studies use models that have the capability to estimate an annual stream but did
161
-------
not report them for reasons that were not discussed in the publication. Thus, for the illustrative
GHG scenario we use the highest and lowest LCA estimates from the 2010 RFS2 rule for the
crop-based biofuel pathways. The LCA ranges used for the illustrative 30-year scenario are
summarized in the following table. The results for the 30-year scenario are described in Chapter
4.2.3.
Table 4.2.2.13-2: Lifecycle GHG Ranges for Illustrative 30-Year Scenario (gCChe/MJ)
Pathway
LCA Range
Petroleum Gasoline
84 to 98
Petroleum Diesel
84 to 94
Corn Starch Ethanol
49 to 91
Soybean Oil Biodiesel
14 to 72
Soybean Oil Renewable Diesel
26 to 87
Used Cooking Oil Biodiesel
12 to 32
Used Cooking Oil Renewable Diesel
12 to 37
Tallow Biodiesel
16 to 58
Tallow Renewable Diesel
14 to 81
Distillers Corn Oil Biodiesel
14 to 37
Distillers Corn Oil Renewable Diesel
12 to 46
Natural Gas CNG
73 to 81
Landfill Gas CNG
6 to 70
Manure Biogas CNG
-533 to 52
4.2.2.14 References for LCA Literature Review
For ease of reference, the following is the list of references cited in Chapter 4.2.2, unless
otherwise cited with the footnote:
• BEIOM. (2021). Avelino, A. F. T., et al. "Creating a harmonized time series of
environmentally-extended input-output tables to assess the evolution of the US
bioeconomy - A retrospective analysis of corn ethanol and soybean biodiesel." Journal of
Cleaner Production 321: 128890.
• Brandao, M. (2022). "Indirect Effects Negate Global Climate Change Mitigation
Potential of Substituting Gasoline With Corn Ethanol as a Transportation Fuel in the
USA." Frontiers in Climate 4.
• CARB (2014). Detailed Analysis for Indirect Land Use Change. California Air Resources
Board. Sacramento, CA. 113 pages
• CARB (2018). CA-GREET3.0 Model. Sacramento, CA, California Air Resources Board.
• CARB (2022). CA-LCFS Current Pathways Certified Carbon Intensities. California Air
Resources Board. Sacramento, CA. =
• Carriquiry, M., et al. (2019). "Incorporating Sub-National Brazilian Agricultural
Production and Land-Use into U.S. Biofuel Policy Evaluation." Applied Economic
Perspectives and Policy.
162
-------
• Chen, R., et al. (2018). "Life cycle energy and greenhouse gas emission effects of
biodiesel in the United States with induced land use change impacts." Bioresource
Technology 251: 249-258.
• Cooney, G., et al. (2017). "Updating the U.S. Life Cycle GHG Petroleum Baseline to
2014 with Projections to 2040 Using Open-Source Engineering-Based Models."
Environmental Science & Technology 51(2): 977-987.
• EPA (2010). Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis.
U.S. Environmental Protection Agency, Office of Transportation Air Quality.
Washington, DC. EPA-420-R-10-006.
• EPA (2020). Renewable Fuel Standard Program: Standards for 2020 and Biomass-Based
Diesel Volume for 2021 and Other Changes. 85 FR 7016.
• EPA and NHSTA (2016). Greenhouse Gas Emissions and Fuel Efficiency Standards for
Medium- and Heavy-Duty Engines and Vehicles - Phase 2: Regulatory Impact Analysis.
• EPA (2014). July 2014 Pathways II final rule. 79 FR 42128.
• GREET (2022). Wang, Michael, Elgowainy, Amgad, Lee, Uisung, Baek, Kwang H.,
Bafana, Adarsh, Benavides, Pahola T., Burnham, Andrew, Cai, Hao, Cappello, Vincenzo,
Chen, Peter, Gan, Yu, Gracida-Alvarez, Ulises R., Hawkins, Troy R., Iyer, Rakesh K.,
Kelly, Jarod C., Kim, Taemin, Kumar, Shi shir, Kwon, Hoyoung, Lee, Kyuha, Liu, Xinyu,
Lu, Zifeng, Masum, Farhad, Ng, Clarence, Ou, Longwen, Reddi, Krishna, Siddique,
Nazib, Sun, Pingping, Vyawahare, Pradeep, Xu, Hui, and Zaimes, George. Greenhouse
gases, Regulated Emissions, and Energy use in Technologies Model ® (2022 Excel).
Computer Software. USDOE Office of Energy Efficiency and Renewable Energy
(EERE). 10 Oct. 2022. Web. doi:10.11578/GREET-Excel-2022/dc.20220908.1.
• ICAO (2021). CORSIA Eligible Fuels — Lifecycle Assessment Methodology. CORSIA
Supporting Document. Version 3: 155.
• Knoope, M. M. J., et al. (2019). "Analysing the water and greenhouse gas effects of soya
bean-based biodiesel in five different regions." Global Change Biology Bioenergy 11(2):
381-399.
• Laborde, D., et al. (2014). Progress in estimates of ILUC with MIRAGE model. JRC
Scientific and Policy Reports. Italy, European Commission Joint Research Centre
Institute for Energy and Transport: 46.
• Lark, T. J., et al. (2022). "Environmental outcomes of the US Renewable Fuel Standard."
Proceedings of the National Academy of Sciences 119(9)
• Lee, U., et al. (2021). "Retrospective analysis of the US corn ethanol industry for 2005-
2019: implications for greenhouse gas emission reductions." Biofuels, Bioproducts and
Biorefining.
• Lewandrowski, J., et al. (2019). "The greenhouse gas benefits of corn ethanol - assessing
recent evidence." Biofuels: 1-15.
• O'Malley, J., et al. (2021). Indirect Emissions from Waste and Residue Feedstocks: 10
Case Studies from the United States, The International Council of Clean Transportation:
49.
• Plevin, R. J., et al. (2015). "Carbon Accounting and Economic Model Uncertainty of
Emissions from Biofuels-Induced Land Use Change." Environmental Science &
Technology 49(5): 2656-2664.
163
-------
• Plevin, R. J., et al. (2022). "Choices in land representation materially affect modeled
biofuel carbon intensity estimates." Journal of Cleaner Production: 131477.
• Riazi, B., et al. (2020). "Renewable diesel from oils and animal fat waste: implications of
feedstock, technology, co-products and ILUC on life cycle GWP." Resources,
Conservation and Recycling 161: 104944.
• Scully, M. J., et al. (2021). "Carbon intensity of corn ethanol in the United States: state of
the science." Environmental Research Letters.
• Seber, G., et al. (2014). "Environmental and economic assessment of producing
hydroprocessed jet and diesel fuel from waste oils and tallow." Biomass and Bioenergy
67: 108-118.
• Taheripour, F., et al. (2017). "The impact of considering land intensification and updated
data on biofuels land use change and emissions estimates." Biotechnology for Biofuels
10(1): 191.
• Xu, H., et al. (2022). "Life Cycle Greenhouse Gas Emissions of Biodiesel and Renewable
Diesel Production in the United States." Environmental Science & Technology 56(12):
7512-7521.
4.2.3 GHG Results for Illustrative Scenario
In order to estimate the monetized social cost or benefit of the candidate biofuel volumes
in Chapter 4.2.4, annual streams of emissions are required. As discussed in Chapter 4.2.2.13, to
develop ranges for purposes of estimating the monetized GHG impacts we rely on the high and
low LCA estimates, from the ranges discussed in Chapter 4.2.2, that report annual streams of
emissions.
For each of the 2023, 2024, and 2025 standards, we estimate a 30-year stream of changes
in GHG emissions for renewable fuel volumes above the No RFS baseline for each analyzed fuel
using the carbon intensity analyses discussed above. While the standards established in this
action only apply in individual years, this analysis portrays what might be expected if, in each of
the ensuing 29 years, aggregate renewable fuel consumption for each category exceeded baseline
levels by the same volume as required by the rule.
Table 4.2.3-2 summarizes the annual low biofuel emission estimates and high petroleum
baseline emission estimates in grams CChe per megajoule (Table 4.2.3-4 does the same for high
biofuel and low petroleum baseline estimates). Table 4.2.3-3 presents the high petroleum
subtracted from the low biofuel emission estimates to show net emissions from displacing
petroleum fuels with biofuels on a per fuel-equivalent megajoule basis (Table 4.2.3-5 does the
same for high biofuel and low petroleum baseline estimates). GHG benefits from biofuels
displacing fossil fuel use include the GHG emissions associated with biofuel production and use,
including land use change emissions relative to the baseline scenario.
Emissions streams based on the 2023 through 2025 standards are presented in Tables
4.2.3-6 through 4.2.3-8 for low biofuel/high petroleum lifecycle analysis estimates, and Tables
4.2.3-10 through 4.2.3-12 for high biofuel/low petroleum lifecycle analysis estimates
respectively, both compared to the No RFS baseline. These are derived by first converting the
net emission streams presented in Tables 4.2.3-3 and 4.2.3-5 from grams CChe per megajoule to
164
-------
million metric tons CChe per megajoule, then multiplying these streams of emissions factors by
changes in renewable fuel volumes. The volume changes in 2023 reflect the difference between
the target volumes and the No RFS baseline as presented in Table 4.2.3-1. As discussed in
Chapter 4.2.2.1, our GHG analysis of the 2023 standard assumes that the target volumes will
produce GHG benefits for the subsequent 29 years due to ongoing use of renewable fuels (and
their consequent displacement of fossil fuels). In analyzing the GHG impacts of the 2024
standard we only consider the difference in volumes between the 2024 standard and 2023
standard because the emissions benefits of the increase in use of renewable fuels to meet the
2023 standards are already accounted for in the 29 years following 2023 (i.e., 2023-2053). Thus,
we only attribute emissions to the 2024 standard for the volumes that have changed compared to
the previous year (2023). Similarly, for 2024, we only include the emission impact for volumes
that have changed from the 2024 levels. These resulting annual sequences of emissions for the
2023 through 2025 standards are then summed, resulting in a combined stream of estimated
annual emissions from 2023 through 2054. These are presented in Tables 4.2.3-9 and 4.2.3-13
respectively below.
Table 4.2.3-
: Volume Changes Used for Illustrative GHG Scenario
Landfill
Biogas
CNG/
LNG261
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/Jet
Fuel
Soybean/
Canola
Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
Volume
Changes
Relative to No
RFS Baseline
[Table 3.2-3]
(million
gallons)
2023
248
248
-
1,133
(101)
46
710
115
137
660
2024
344
344
-
1,065
(92)
63
901
106
(68)
731
2025
466
466
-
1,078
(113)
20
1,066
126
(21)
787
Volume
Changes
Relative to
Previous Year
(million
gallons)
2023
248
248
-
1,133
(101)
46
710
115
137
660
2024
97
97
-
(68)
9
17
190
(9)
(205)
71
2025
122
122
-
13
(21)
(43)
165
20
47
56
Table 4.2.3-5 shows positive net GHG emissions for the corn ethanol and soybean oil
renewable diesel and biodiesel volumes in the first year of a volume increase due to the initial
pulse of land use change emissions in the estimates used for this illustrative scenario. For corn
ethanol, volumes increase year over year from 2023 through 2025, which results in positive
emissions in 2023, 2024, and 2025 in Tables 4.2.3-6, 7, and 8 respectively. Conversely for
example, as shown in Table 4.2.3-6, soybean oil biodiesel volumes are negative in 2024 because
those volumes decrease relative to the previous year's (2023) volume increases from the No RFS
baseline. As noted above, this scenario assumes that the biofuel production continues for 30
years, irrespective of volume mandates in future years.
261 Table 3.2-3 presents total volume changes for CNG/LNG from biogas. We assume for purposes of this
illustrative GHG scenario that half of that biogas is sourced from landfills and half from agricultural digesters.
165
-------
We separately estimate a 30-year stream of changes in GHG emissions for renewable fuel
volumes from the 2023 supplemental volume requirement as described in Chapter 3.3. As shown
in Table 3.3-1, the supplemental volume requirement of 250 million ethanol-equivalent gallons is
represented by an energy-equivalent 147 million soybean oil renewable diesel gallons in 2023.262
Table 4.2.3-14 shows the illustrative GHG scenario using the same process described above for
both low biofuel/high petroleum and high biofuel/low petroleum lifecycle analysis estimates for
the supplemental volumes.
262 See Chapter 3.3 for more details.
166
-------
Table 4.2.3-2: Gross low biofuel/high petroleum annual lifecycle analysis estimates for
Landfill
Biogas
CNG/
LNG
Ag.
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuelb
Soybean/
Canola
Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean
Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
Gasoline
(High
Estimate)
Diesel
(High
Estimate)
Year 0
5.7
(532.7)
37.6
710.3
13.1
13.7
755.9
12.8
12.4
395.0
98.1
94.1
Year 1
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 2
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 3
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 4
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 5
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 6
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 7
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 8
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 9
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 10
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 11
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 12
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 13
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 14
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 15
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 16
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 17
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 18
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 19
5.7
(532.7)
37.6
(20.0)
13.1
13.7
(8.7)
12.8
12.4
38.9
98.1
94.1
Year 20
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
Year 21
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
Year 22
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
Year 23
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
Year 24
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
Year 25
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
Year 26
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
Year 27
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
Year 28
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
Year 29
5.7
(532.7)
37.6
7.5
13.1
13.7
19.9
12.8
12.4
34.5
98.1
94.1
a Parentheses indicate a reduction in GHG emissions.
b Wood waste/MSW diesel and jet fuels are comprised of a wide variety of feedstocks and represent a small volume
of fuels in this rule. We have made a simplifying assumption that these fuels meet a 60% GHG reduction (equal to
the cellulosic threshold) compared to the diesel GHG estimate shown in this table.
167
-------
Table 4.2.3-3: Net low biofuel/high petroleum (low biofuel minus high petroleum baseline)
annual lifecycle analysis estimates for individual biofuels, presented in grams CChe per
megajoule of fuel."
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
Year 0
(88.4)
(626.8)
(56.5)
616.2
(81.0)
(80.4)
661.8
(81.3)
(81.7)
296.9
Year 1
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 2
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 3
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 4
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 5
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 6
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 7
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 8
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 9
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 10
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 11
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 12
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 13
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 14
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 15
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 16
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 17
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 18
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 19
(88.4)
(626.8)
(56.5)
(114.1)
(81.0)
(80.4)
(102.8)
(81.3)
(81.7)
(59.2)
Year 20
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
Year 21
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
Year 22
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
Year 23
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
Year 24
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
Year 25
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
Year 26
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
Year 27
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
Year 28
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
Year 29
(88.4)
(626.8)
(56.5)
(86.6)
(81.0)
(80.4)
(74.2)
(81.3)
(81.7)
(63.6)
a Parentheses indicate a net reduction in GHG emissions
168
-------
Table 4.2.3-4: Gross high biofuel/low petroleum annual lifecycle analysis estimates for
Landfill
Biogas
CNG/
LNG
Ag.
Digester
Biogas
CNG/
LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel"
Soybean/
Canola
Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean
Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
Gasoline
(Low
Estimate)
Diesel
(Low
Estimate)
Natural
Gas (Low
Estimate)
Year 0
69.8
51.9
33.4
1,044.0
39.4
36.6
1,102.7
49.3
46.3
665.1
83.6
83.5
72.7
Year 1
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 2
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 3
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 4
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 5
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 6
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 7
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 8
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 9
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 10
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 11
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 12
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 13
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 14
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 15
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 16
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 17
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 18
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 19
69.8
51.9
33.4
54.2
39.4
36.6
68.4
49.3
46.3
79.8
83.6
83.5
72.7
Year 20
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
Year 21
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
Year 22
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
Year 23
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
Year 24
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
Year 25
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
Year 26
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
Year 27
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
Year 28
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
Year 29
69.8
51.9
33.4
8.4
39.4
36.6
20.9
49.3
46.3
53.6
83.6
83.5
72.7
a Wood waste/MSW diesel and jet fuels are comprised of a wide variety of feedstocks and represent a small volume
of fuels in this rule. We have made a simplifying assumption that these fuels meet a 60% GHG reduction (equal to
the cellulosic threshold) compared to the diesel GHG estimate shown in this table.
169
-------
Table 4.2.3-5: Net high biofuel/low petroleum (high biofuel minus low petroleum baseline)
annual lifecycle analysis estimates for individual biofuels, presented in grams CChe per
megajoule of fuel."
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
Year 0
(3.0)
(20.8)
(50.1)
960.5
(44.1)
(46.9)
1,019.2
(34.2)
(37.2)
581.5
Year 1
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 2
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 3
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 4
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 5
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 6
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 7
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 8
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 9
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 10
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 11
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 12
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 13
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 14
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 15
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 16
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 17
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 18
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 19
(3.0)
(20.8)
(50.1)
(29.3)
(44.1)
(46.9)
(15.1)
(34.2)
(37.2)
(3.8)
Year 20
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
Year 21
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
Year 22
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
Year 23
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
Year 24
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
Year 25
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
Year 26
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
Year 27
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
Year 28
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
Year 29
(3.0)
(20.8)
(50.1)
(75.1)
(44.1)
(46.9)
(62.6)
(34.2)
(37.2)
(30.0)
a Parentheses indicate a net reduction in GHG emissions.
170
-------
Table 4.2.3-6: 30-year stream of emissions for 2023 standards using low biofuel/high
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS
baseline, presented in millions of metric tons CQ2e.a
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
(12.5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12-5)
(12.5)
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
88.1
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
16.3)
12.4)
12.4)
12.4)
12.4)
12.4)
12.4)
12.4)
12.4)
12.4)
12.4)
Fats/Oils/
Greases
Biodiesel
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
Corn Oil
Biodiesel
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0-5)
(0.5)
Soybean Oil
Renewable
Diesel
60.9
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(9-5)
(6-8)
(6-8)
(6-8)
(6-8)
(6-8)
(6-8)
(6-8)
(6-8)
(6-8)
(6-8)
Fats/Oils/
Greases
Renewable
Diesel
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1-2)
(1.2)
Corn Oil
Renewable
Diesel
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1-4)
(1.4)
Corn
Starch
Ethanol
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. Parentheses
indicate a net reduction in GHG emissions.
171
-------
Table 4.2.3-7: 30-year stream of emissions for 2024 standards using low biofuel/high
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS
baseline, presented in millions o
' metric
tons CC>2e.a
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
2023
-
-
-
-
-
-
-
-
-
-
2024
(0.7)
(4.9)
-
(5.3)
(0.1)
(0.2)
16.3
0.1
2.2
1.7
2025
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2026
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2027
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2028
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2029
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2030
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2031
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2032
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2033
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2034
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2035
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2036
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2037
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2038
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2039
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2040
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2041
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2042
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2043
(0.7)
(4.9)
-
1.0
(0.1)
(0.2)
(2.5)
0.1
2.2
(0.3)
2044
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2045
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2046
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2047
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2048
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2049
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2050
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2051
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2052
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2053
(0.7)
(4.9)
-
0.7
(0.1)
(0.2)
(1.8)
0.1
2.2
(0.4)
2054
-
-
-
-
-
-
-
-
-
-
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. Parentheses
indicate a net reduction in GHG emissions.
172
-------
Table 4.2.3-8: 30-year stream of emissions for 2025 standards using low biofuel/high
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
2023
-
-
-
-
-
-
-
-
-
-
2024
-
-
-
-
-
-
-
-
-
-
2025
(0.9)
(6.2)
-
1.0
0.2
0.4
14.2
(0.2)
(0.5)
1.3
2026
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2027
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2028
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2029
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2030
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2031
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2032
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2033
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2034
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2035
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2036
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2037
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2038
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2039
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2040
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2041
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2042
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2043
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2044
(0.9)
(6.2)
-
(0.2)
0.2
0.4
(2.2)
(0.2)
(0.5)
(0.3)
2045
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
2046
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
2047
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
2048
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
2049
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
2050
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
2051
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
2052
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
2053
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
2054
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. Parentheses
indicate a net reduction in GHG emissions.
173
-------
Table 4.2.3-9: 30-year stream of emissions for combined 2023-2025 standards using low
biofuel/high petroleum lifecycle analysis estimates for individual biofuels, relative to the No
RFS baseline, presented in millions of metric tons CQ2e.a
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
2023
(1.8)
(12.5)
-
88.1
1.0
(0.5)
60.9
(1.2)
(1.4)
15.8
2024
(2.5)
(17.4)
-
(21.6)
0.9
(0.6)
6.9
(1.1)
0.7
(1.5)
2025
(3.3)
(23.6)
-
(14.3)
1.2
(0.2)
2.2
(1.3)
0.2
(2.1)
2026
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2027
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2028
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2029
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2030
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2031
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2032
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2033
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2034
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2035
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2036
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2037
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2038
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2039
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2040
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2041
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2042
(3.3)
(23.6)
-
(15.5)
1.2
(0.2)
(14.2)
(1.3)
0.2
(3.8)
2043
(3.3)
(23.6)
-
(11.6)
1.2
(0.2)
(11.6)
(1.3)
0.2
(4.0)
2044
(3.3)
(23.6)
-
(11.8)
1.2
(0.2)
(10.9)
(1.3)
0.2
(4.0)
2045
(3.3)
(23.6)
-
(11.8)
1.2
(0.2)
(10.3)
(1.3)
0.2
(4.0)
2046
(3.3)
(23.6)
-
(11.8)
1.2
(0.2)
(10.3)
(1.3)
0.2
(4.0)
2047
(3.3)
(23.6)
-
(11.8)
1.2
(0.2)
(10.3)
(1.3)
0.2
(4.0)
2048
(3.3)
(23.6)
-
(11.8)
1.2
(0.2)
(10.3)
(1.3)
0.2
(4.0)
2049
(3.3)
(23.6)
-
(11.8)
1.2
(0.2)
(10.3)
(1.3)
0.2
(4.0)
2050
(3.3)
(23.6)
-
(11.8)
1.2
(0.2)
(10.3)
(1.3)
0.2
(4.0)
2051
(3.3)
(23.6)
-
(11.8)
1.2
(0.2)
(10.3)
(1.3)
0.2
(4.0)
2052
(3.3)
(23.6)
-
(11.8)
1.2
(0.2)
(10.3)
(1.3)
0.2
(4.0)
2053
(1.6)
(11.1)
-
0.6
0.1
0.3
(3.4)
(0.1)
1.7
(0.7)
2054
(0.9)
(6.2)
-
(0.1)
0.2
0.4
(1.6)
(0.2)
(0.5)
(0.3)
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. Parentheses
indicate a net reduction in GHG emissions.
174
-------
Table 4.2.3-10: 30-year stream of emissions for 2023 standards using high biofuel/low
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS
baseline, presented in millions o
' metric
tons CC>2e.a
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
2023
(0.1)
(0.4)
-
137.3
0.6
(0.3)
93.9
(0.5)
(0.7)
31.0
2024
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2025
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2026
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2027
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2028
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2029
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2030
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2031
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2032
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2033
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2034
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2035
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2036
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2037
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2038
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2039
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2040
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2041
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2042
(0.1)
(0.4)
-
(4.2)
0.6
(0.3)
(1.4)
(0.5)
(0.7)
(0.2)
2043
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2044
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2045
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2046
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2047
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2048
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2049
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2050
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2051
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2052
(0.1)
(0.4)
-
(10.7)
0.6
(0.3)
(5.8)
(0.5)
(0.7)
(1.6)
2053
-
-
-
-
-
-
-
-
-
-
2054
-
-
-
-
-
-
-
-
-
-
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. Parentheses
indicate a net reduction in GHG emissions.
175
-------
Table 4.2.3-11: 30-year stream of emissions for 2024 standards using high biofuel/low
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS
baseline, presented in millions o
' metric
tons CC>2e.a
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
2023
-
-
-
-
-
-
-
-
-
-
2024
(0.0)
(0.2)
-
(8.3)
(0.1)
(0.1)
25.2
0.0
1.0
3.3
2025
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2026
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2027
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2028
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2029
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2030
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2031
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2032
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2033
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2034
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2035
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2036
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2037
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2038
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2039
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2040
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2041
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2042
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2043
(0.0)
(0.2)
-
0.3
(0.1)
(0.1)
(0.4)
0.0
1.0
(0.0)
2044
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2045
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2046
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2047
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2048
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2049
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2050
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2051
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2052
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2053
(0.0)
(0.2)
-
0.6
(0.1)
(0.1)
(1.5)
0.0
1.0
(0.2)
2054
-
-
-
-
-
-
-
-
-
-
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. Parentheses
indicate a net reduction in GHG emissions.
176
-------
Table 4.2.3-12: 30-year stream of emissions for 2025 standards using high biofuel/low
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
2023
-
-
-
-
-
-
-
-
-
-
2024
-
-
-
-
-
-
-
-
-
-
2025
(0.0)
(0.2)
-
1.6
0.1
0.3
21.8
(0.1)
(0.2)
2.6
2026
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2027
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2028
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2029
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2030
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2031
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2032
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2033
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2034
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2035
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2036
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2037
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2038
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2039
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2040
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2041
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2042
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2043
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2044
(0.0)
(0.2)
-
(0.0)
0.1
0.3
(0.3)
(0.1)
(0.2)
(0.0)
2045
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
2046
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
2047
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
2048
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
2049
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
2050
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
2051
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
2052
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
2053
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
2054
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggrega
consumption for each category exceeded baseline levels by the same volume as required by
indicate a net reduction in GHG emissions.
e renewable fuel
the rule. Parentheses
177
-------
Table 4.2.3-13: 30-year stream of emissions for combined 2023-2025 standards using high
biofuel/low petroleum lifecycle analysis estimates for individual biofuels, relative to the No
RFS baseline, presented in millions of metric tons CQ2e.a
Landfill
Biogas
CNG/LNG
Agricultural
Digester
Biogas
CNG/LNG
Wood
Waste/
MSW
Diesel/
Jet Fuel
Soybean/
Canola Oil
Biodiesel
Fats/Oils/
Greases
Biodiesel
Corn Oil
Biodiesel
Soybean Oil
Renewable
Diesel
Fats/Oils/
Greases
Renewable
Diesel
Corn Oil
Renewable
Diesel
Corn
Starch
Ethanol
2023
(0.1)
(0.4)
-
137.3
0.6
(0.3)
93.9
(0.5)
(0.7)
31.0
2024
(0.1)
(0.6)
-
(12.5)
0.5
(0.4)
23.8
(0.5)
0.3
3.1
2025
(0.1)
(0.8)
-
(2.3)
0.6
(0.1)
20.0
(0.6)
0.1
2.4
2026
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2027
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2028
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2029
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2030
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2031
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2032
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2033
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2034
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2035
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2036
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2037
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2038
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2039
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2040
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2041
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2042
(0.1)
(0.8)
-
(4.0)
0.6
(0.1)
(2.1)
(0.6)
0.1
(0.2)
2043
(0.1)
(0.8)
-
(10.5)
0.6
(0.1)
(6.5)
(0.6)
0.1
(1.6)
2044
(0.1)
(0.8)
-
(10.1)
0.6
(0.1)
(7.6)
(0.6)
0.1
(1.8)
2045
(0.1)
(0.8)
-
(10.2)
0.6
(0.1)
(8.7)
(0.6)
0.1
(1.9)
2046
(0.1)
(0.8)
-
(10.2)
0.6
(0.1)
(8.7)
(0.6)
0.1
(1.9)
2047
(0.1)
(0.8)
-
(10.2)
0.6
(0.1)
(8.7)
(0.6)
0.1
(1.9)
2048
(0.1)
(0.8)
-
(10.2)
0.6
(0.1)
(8.7)
(0.6)
0.1
(1.9)
2049
(0.1)
(0.8)
-
(10.2)
0.6
(0.1)
(8.7)
(0.6)
0.1
(1.9)
2050
(0.1)
(0.8)
-
(10.2)
0.6
(0.1)
(8.7)
(0.6)
0.1
(1.9)
2051
(0.1)
(0.8)
-
(10.2)
0.6
(0.1)
(8.7)
(0.6)
0.1
(1.9)
2052
(0.1)
(0.8)
-
(10.2)
0.6
(0.1)
(8.7)
(0.6)
0.1
(1.9)
2053
(0.1)
(0.4)
-
0.5
0.1
0.2
(2.9)
(0.0)
0.8
(0.3)
2054
(0.0)
(0.2)
-
(0.1)
0.1
0.3
(1.3)
(0.1)
(0.2)
(0.1)
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. Parentheses
indicate a net reduction in GHG emissions.
178
-------
Table 4.2.3-14: 30-year stream of lifecycle analysis estimates for volume changes from 2023
supplemental volume requirement, represented by soybean renewable diesel compared to
petroleum diesel, presented in millions of me
Low Biofuel/
High Petroleum
High Biofuel/
Low Petroleum
2023
12.6
19.4
2024
(2.0)
(0.3)
2025
(2.0)
(0.3)
2026
(2.0)
(0.3)
2027
(2.0)
(0.3)
2028
(2.0)
(0.3)
2029
(2.0)
(0.3)
2030
(2.0)
(0.3)
2031
(2.0)
(0.3)
2032
(2.0)
(0.3)
2033
(2.0)
(0.3)
2034
(2.0)
(0.3)
2035
(2.0)
(0.3)
2036
(2.0)
(0.3)
2037
(2.0)
(0.3)
2038
(2.0)
(0.3)
2039
(2.0)
(0.3)
2040
(2.0)
(0.3)
2041
(2.0)
(0.3)
2042
(2.0)
(0.3)
2043
(1.4)
(1.2)
2044
(1.4)
(1.2)
2045
(1.4)
(1.2)
2046
(1.4)
(1.2)
2047
(1.4)
(1.2)
2048
(1.4)
(1.2)
2049
(1.4)
(1.2)
2050
(1.4)
(1.2)
2051
(1.4)
(1.2)
2052
(1.4)
(1.2)
2053
2054
-
-
ric tons CC>2e.a
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. Parentheses
indicate a net reduction in GHG emissions.
179
-------
4.2.4 Monetized GHG Impacts
4.2.4.1 Social Cost of Greenhouse Gases
For assessing GHG impacts in this illustrative scenario, we rely upon past biofuel
emissions reductions estimates that are available in CChe—carbon equivalent emissions using
the global warming potentials utilized in those analyses.263 We estimate the social benefits of
GHG reductions in this illustrative scenario using estimates of the social cost of greenhouse
gases (SC-GHG), specifically using the social cost of carbon (SC-CO2). The SC-GHG is the
monetary value of the net harm to society associated with a marginal increase in GHG emissions
in a given year, or the benefit of avoiding that increase. In principle, SC-GHG includes the value
of all climate change impacts (both negative and positive), including (but not limited to) changes
in net agricultural productivity, human health effects, property damage from increased flood risk
and natural disasters, disruption of energy systems, risk of conflict, environmental migration, and
the value of ecosystem services. The SC-GHG therefore, reflects the societal value of reducing
emissions of the gas in question by one metric ton and is the theoretically appropriate value to
use in conducting benefit-cost analyses of policies that affect GHG emissions. In practice, data
and modeling limitations naturally restrain the ability of SC-GHG estimates to include all the
important physical, ecological, and economic impacts of climate change, such that the estimates
are a partial accounting of climate change impacts and will therefore, tend to be underestimates
of the marginal benefits of abatement. EPA and other Federal agencies began regularly
incorporating SC-GHG estimates in their benefit-cost analyses conducted under Executive Order
(E.O.) 12866264 since 2008, following a Ninth Circuit Court of Appeals remand of a rule for
failing to monetize the benefits of reducing CO2 emissions in that rulemaking process.
In 2017, the National Academies of Sciences, Engineering, and Medicine published a
report that provides a roadmap for how to update SC-GHG estimates used in Federal analyses
going forward to ensure that they reflect advances in the scientific literature.265 The National
Academies' report recommended specific criteria for future SC-GHG updates, a modeling
framework to satisfy the specified criteria, and both near-term updates and longer-term research
needs pertaining to various components of the estimation process. The research community has
made considerable progress in developing new data and methods that help to advance various
components of the SC-GHG estimation process in response to the National Academies'
recommendations.
263 It would be preferable to use estimates for each gas (e.g., CO2, CH4, N20), but we use CChe estimates for this
illustrative scenario as they are the most readily available biofuel carbon intensity estimates.
264 Presidents since the 1970s have issued executive orders requiring agencies to conduct analysis of the economic
consequences of regulations as part of the rulemaking development process. E.O. 12866, released in 1993 and still in
effect today, requires that for all economically significant regulatory actions, an agency provide an assessment of the
potential costs and benefits of the regulatory action, and that this assessment include a quantification of benefits and
costs to the extent feasible. Many statutes also require agencies to conduct at least some of the same analyses
required under E.O. 12866, such as the Energy Policy and Conservation Act, which mandates the setting of fuel
economy regulations. For purposes of this action, monetized climate benefits are presented for purposes of providing
a complete benefit-cost analysis under E.O. 12866 and other relevant executive orders. The estimates of change in
GHG emissions and the monetized benefits associated with those changes play no part in the record basis for this
action.
265 National Academies. (2017). Valuing Climate Damages: Updating Estimation of the Social Cost of Carbon
Dioxide. Washington, DC: The National Academies Press.
180
-------
In a first-day executive order (E.O. 13990), Protecting Public Health and the
Environment and Restoring Science to Tackle the Climate Crisis, President Biden called for a
renewed focus on updating estimates of the social cost of greenhouse gases (SC-GHG) to reflect
the latest science, noting that "it is essential that agencies capture the full benefits of reducing
greenhouse gas emissions as accurately as possible." Important steps have been taken to begin to
fulfill this directive of E.O. 13990. In February 2021, the Interagency Working Group on the SC-
GHG (IWG) released a technical support document (hereinafter the "February 2021 TSD") that
provided a set of IWG recommended SC-GHG estimates while work on a more comprehensive
update is underway to reflect recent scientific advances relevant to SC-GHG estimation.266 In
addition, as discussed further below, EPA has developed a draft updated SC-GHG methodology
within a sensitivity analysis in the regulatory impact analysis of EPA's November 2022
supplemental proposal for oil and gas standards that is currently undergoing external peer review
and a public comment process.267
EPA has applied the IWG's recommended interim SC-GHG estimates in the Agency's
regulatory benefit-cost analyses published since the release of the February 2021 TSD and is
likewise using them in this document. We have evaluated the SC-GHG estimates in the February
2021 TSD and have determined that these estimates are appropriate for use in estimating the
social benefits of GHG reductions in this illustrative scenario. These SC-GHG estimates are
interim values developed for use in benefit-cost analyses until updated estimates of the impacts
of climate change can be developed based on the best available science and economics. After
considering the TSD, and the issues and studies discussed therein, EPA finds that these
estimates, while likely an underestimate, are the best currently available SC-GHG estimates until
revised estimates have been developed reflecting the latest, peer-reviewed science. The SC-GHG
estimates presented in the February 2021 SC-GHG TSD and used in this document were
developed over many years, using a transparent process, peer-reviewed methodologies, the best
science available at the time of that process, and with input from the public. Specifically, in
2009, an interagency working group (IWG) that included EPA and other executive branch
agencies and offices was established to develop estimates relying on the best available science
for agencies to use. The IWG published SC- CO2 estimates in 2010 that were developed from an
ensemble of three widely cited integrated assessment models (IAMs) that estimate global climate
damages using highly aggregated representations of climate processes and the global economy
combined into a single modeling framework. The three IAMs were run using a common set of
input assumptions in each model for future population, economic, and CO2 emissions growth, as
well as equilibrium climate sensitivity (ECS)—a measure of the globally averaged temperature
response to increased atmospheric CO2 concentrations. These estimates were updated in 2013
based on new versions of each jam.268'269'270 In August 2016 the IWG published estimates of the
social cost of methane (SC-CH4) and nitrous oxide (SC-N2O) using methodologies that are
266 IWG. (2021). Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim
Estimates under Executive Order 13990. Retrieved from Washington, DC: https://www.whitehouse.gov/wp-
eontent/nptoads/2021/02/TeefaiiiealSupportDoeument SocialCostofCarbonMethaneNitronsOxide.pdf
267 See https://www.epa.gov/eiwironmental-economies/scghg
268 Dynamic Integrated Climate and Economy (DICE) 2010 (Nordhaus 2010).
269 Climate Framework for Uncertainty, Negotiation, and Distribution (FUND) 3.8 (Anthoff and Tol 2013a, 2013b)
270 Dynamic Integrated Climate and Economy (DICE), Climate Framework for Uncertainty, Negotiation, and
Distribution (FUND), and Policy Analysis of the Greenhouse Gas Effect (PAGE) 2009 (Hope 2013).
181
-------
consistent with the methodology underlying the SC-CO2 estimates. The modeling approach that
extends the IWG SC-CO2 methodology to non-CCh GHGs has undergone multiple stages of peer
review. The SC-CH4 and SC-N2O estimates were developed by Marten, Kopits, Griffiths,
Newbold, and Wolverton (2015) and underwent a standard double-blind peer review process
prior to journal publication. These estimates were applied in regulatory impact analyses of EPA
proposed rulemakings with CH4 and N2O emissions impacts. EPA also sought additional
external peer review of technical issues associated with its application to regulatory analysis.
Following the completion of the independent external peer review of the application of the
Marten et al. (2015) estimates271, EPA began using the estimates in the primary benefit-cost
analysis calculations and tables for a number of proposed rulemakings in 2015.272 EPA
considered and responded to public comments received for the proposed rulemakings before
using the estimates in final regulatory analyses in 2016.273 In 2015, as part of the response to
public comments received to a 2013 solicitation for comments on the SC-CO2 estimates, the
IWG announced a National Academies of Sciences, Engineering, and Medicine review of the
SC-CO2 estimates to offer advice on how to approach future updates to ensure that the estimates
continue to reflect the best available science and methodologies. In January 2017, the National
Academies released their final report, Valuing Climate Damages: Updating Estimation of the
Social Cost of Carbon Dioxide, and recommended specific criteria for future updates to the SC-
GHG estimates, a modeling framework to satisfy the specified criteria, and both near-term
updates and longer-term research needs pertaining to various components of the estimation
process.274 Shortly thereafter, in March 2017, President Trump issued Executive Order 13783,
which disbanded the IWG, withdrew the previous TSDs, and directed agencies to ensure SC-
GHG estimates used in regulatory analyses are consistent with the guidance contained in OMB's
Circular A-4, "including with respect to the consideration of domestic versus international
impacts and the consideration of appropriate discount rates" (E.O. 13783, Section 5(c)). Benefit-
cost analyses following E.O. 13783 used SC-GHG estimates that attempted to focus on the
specific share of climate change damages in the U.S. as captured by the models (which did not
reflect many pathways by which climate impacts affect the welfare of U.S. citizens and
residents) and were calculated using two discount rates recommended by Circular A-4, 3 percent
and 7 percent.275 All other methodological decisions and model versions used in SC-GHG
calculations remained the same as those used by the IWG in 2010 and 2013, respectively.
271 Alex L. Marten, Elizabeth A. Kopits, Charles W. Griffiths, Stephen C. Newbold & Ann Wolverton (2015)
Incremental CH4 andN20 mitigation benefits consistent with the US Government's SC-C02 estimates, Climate
Policy, 15:2, 272-298, DOI: 10.1080/14693062.2014.912981
272 U.S. Environmental Protection Agency (EPA), 2015b. Regulatory Impact Analysis for the Proposed Revisions to
the Emission Guidelines for Existing Sources and Supplemental Proposed New Source Performance Standards in
the Municipal Solid Waste Landfills Sector. fattPs://www.regiilations.gov/doeument?D=EPA~HO~OAR~2Q14~
0086.
273 The SC-CH4 and SC-N20 estimates were first used in sensitivity analysis for the Proposed Rulemaking for
Greenhouse Gas Emissions and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles-
Phase 2 (U.S. EPA, 2015a).
274 National Academies of Sciences, Engineering, and Medicine (National Academies). 2017. Valuing Climate
Damages: Updating Estimation of the Social Cost of Carbon Dioxide. Washington, D.C.: National Academies Press.
275 EPA regulatory analyses under E.O. 13783 included sensitivity analyses based on global SC-GHG values and
using a lower discount rate of 2.5%. OMB Circular A-4 (OMB, 2003) recognizes that special considerations arise
when applying discount rates if intergenerational effects are important. In the IWG's 2015 Response to Comments,
OMB—as a co-chair of the IWG—made clear that "Circular A-4 is a living document," that "the use of 7 percent is
182
-------
On January 20, 2021, President Biden issued Executive Order 13990, which re-
established an IWG and directed it to develop an update of the social cost of carbon and other
greenhouse gas estimates that reflect the best available science and the recommendations of the
National Academies. In February 2021, the IWG recommended the interim use of the most
recent SC-GHG estimates developed by the IWG prior to the group being disbanded in 2017,
adjusted for inflation.276 As discussed in the February 2021 TSD, the IWG's selection of these
interim estimates reflected the immediate need to have SC-GHG estimates available for agencies
to use in regulatory benefit-cost analyses and other applications that were developed using a
transparent process, peer reviewed methodologies, and the science available at the time of that
process. As noted above, EPA participated in the IWG but has also independently evaluated the
interim SC-GHG estimates published in the February 2021 TSD and determined they are
appropriate to use here to estimate climate benefits EPA and other agencies intend to undertake a
fuller update of the SC-GHG estimates that takes into consideration the advice of the National
Academies (2017) and other recent scientific literature. EPA has also evaluated the supporting
rationale of the February 2021 TSD, including the studies and methodological issues discussed
therein, and concludes that it agrees with the rationale for these estimates presented in the TSD
and summarized below.
In particular, the IWG found that the SC-GHG estimates used under E.O. 13783 fail to
reflect the full impact of GHG emissions in multiple ways. First, the IWG concluded that those
estimates fail to capture many climate impacts that can affect the welfare of U.S. citizens and
residents. Examples of affected interests include direct effects on U.S. citizens and assets located
abroad, international trade, and tourism, and spillover pathways such as economic and political
destabilization and global migration that can lead to adverse impacts on U.S. national security,
public health, and humanitarian concerns. Those impacts are better captured within global
measures of the social cost of greenhouse gases.
In addition, assessing the benefits of U.S. GHG mitigation activities requires
consideration of how those actions may affect mitigation activities by other countries, as those
international mitigation actions will provide a benefit to U.S. citizens and residents by mitigating
climate impacts that affect U.S. citizens and residents. A wide range of scientific and economic
experts have emphasized the issue of reciprocity as support for considering global damages of
GHG emissions. Using a global estimate of damages in U.S. analyses of regulatory actions
allows the U.S. to continue to actively encourage other nations, including emerging major
economies, to take significant steps to reduce emissions. The only way to achieve an efficient
allocation of resources for emissions reduction on a global basis—and so benefit the U.S. and its
citizens—is for all countries to base their policies on global estimates of damages.
not considered appropriate for intergenerational discounting," and that "[t]here is wide support for this view in the
academic literature, and it is recognized in Circular A-4 itself." OMB, as part of the IWG, similarly repeatedly
confirmed that "a focus on global SCC estimates in [regulatory impact analyses] is appropriate" (IWG, 2015).
276 IWG. (2021). Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim
Estimates under Executive Order 13990. Retrieved from Washington, DC: fattps://www.wfaitehonse.gov/wp-
eontent/nptoads/2021/02/TeefaiiiealSnpportDoeument SocialCostofCarbonMethaneNitronsOxide.pdf
183
-------
As a member of the IWG involved in the development of the February 2021 SC-GHG
TSD, EPA agrees with this assessment and, therefore, in this document EPA centers attention on
a global measure of SC-GHG. This approach is the same as that taken in EPA regulatory
analyses over 2009 through 2016. A robust estimate of climate damages to U.S. citizens and
residents that accounts for the myriad of ways that global climate change reduces the net welfare
of U.S. populations does not currently exist in the literature. As explained in the February 2021
TSD, existing estimates are both incomplete and an underestimate of total damages that accrue to
the citizens and residents of the U.S. because they do not fully capture the regional interactions
and spillovers discussed above, nor do they include all of the important physical, ecological, and
economic impacts of climate change recognized in the climate change literature, as discussed
further below. EPA, as a member of the IWG, will continue to review developments in the
literature, including more robust methodologies for estimating the magnitude of the various
damages to U.S. populations from climate impacts and reciprocal international mitigation
activities, and explore ways to better inform the public of the full range of carbon impacts.
Second, the IWG concluded that the use of the social rate of return on capital (7 percent
under current OMB Circular A-4 guidance) to discount the future benefits of reducing GHG
emissions inappropriately underestimates the impacts of climate change for the purposes of
estimating the SC-GHG. Consistent with the findings of the National Academies (2017) and the
economic literature, the IWG continued to conclude that the consumption rate of interest is the
theoretically appropriate discount rate in an intergenerational context, and recommended that
discount rate uncertainty and relevant aspects of intergenerational ethical considerations be
accounted for in selecting future discount rates.277'278'279'280 Furthermore, the damage estimates
developed for use in the SC-GHG are estimated in consumption-equivalent terms, and so an
application of OMB Circular A-4's guidance for regulatory analysis would then use the
277 IWG. (2010). Technical Support Document: Social Cost of Carbon for Regulatory Impact Analysis under
Executive Order 12866. Retrieved from Washington, DC: https://www.epa.gov/sites/default/files/2016-
12/doeuments/see tsd 2010.pdf Interagency Working Group on Social Cost of Carbon (IWG). 2010. Technical
Support Document: Social Cost of Carbon for Regulatory Impact Analysis under Executive Order 12866. February.
United States Government.
278 IWG. (2013). Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory
Impact Analysis Under Executive Order 12866. Retrieved from Washington, DC: https://
obamawhitehouse.archives.gov/sites/default/files/omb/inforeg/social cost of carbon for ria 2013 update.pdf
Interagency Working Group on Social Cost of Carbon (IWG). 2013. Technical Support Document: Technical
Update of the Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866. May.
279 IWG. (2016a). Addendum to Technical Support Document on Social Cost of Carbon for Regulatory Impact
Analysis under Executive Order 12866: Application of the Methodology to Estimate the SociafCost of Methane
and the Social Cost of Nitrous Oxide. Retrieved from Washington, DC: https://www.epa.gov/sites/defanlt/
files/20]/)~.1.2/doeuments/addendiim to sc-ghg tsd august 2016.pdf Greenhouse Gases (IWG). 2016a.August.
280 IWG. (2016b). Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory
Impact Analysis Under Executive Order 12866. Retrieved from Washington, DC:
https://www.epa.gov/sites/default/files/2016-12/documents/sc co2 tsd august 2016.pdf. GHG emissions are stock
pollutants, where damages are associated with what has accumulated in the atmosphere over time, and they are long
lived such that subsequent damages resulting from emissions today occur over many decades or centuries
depending on the specific greenhouse gas under consideration. In calculating the SC-GHG, the stream of future
damages to agriculture, human health, and other market and non-market sectors from an additional unit of emissions
are estimated in terms of reduced consumption (or consumption equivalents). Then that stream of future damages is
discounted to its present value in the year when the additional unit of emissions was released. Given the long time
horizon over which the damages are expected to occur, the discount rate has a large influence on the present value of
future damages.
184
-------
consumption discount rate to calculate the SC-GHG. EPA agrees with this assessment and will
continue to follow developments in the literature pertaining to this issue. EPA also notes that
while OMB Circular A-4, as published in 2003, recommends using 3 percent and 7 percent
discount rates as "default" values, Circular A-4 also reminds agencies that "different regulations
may call for different emphases in the analysis, depending on the nature and complexity of the
regulatory issues and the sensitivity of the benefit and cost estimates to the key assumptions." On
discounting, Circular A-4 recognizes that "special ethical considerations arise when comparing
benefits and costs across generations," and Circular A-4 acknowledges that analyses may
appropriately "discount future costs and consumption benefits.. .at a lower rate than for
intragenerational analysis." In the 2015 Response to Comments on the Social Cost of Carbon for
Regulatory Impact Analysis, OMB, EPA, and the other IWG members recognized that "Circular
A-4 is a living document" and "the use of 7 percent is not considered appropriate for
intergenerational discounting. There is wide support for this view in the academic literature, and
it is recognized in Circular A-4 itself." Thus, EPA concludes that a 7 percent discount rate is not
appropriate to apply to value the social cost of greenhouse gases in the analysis presented in this
final rule. In this analysis, to calculate the present and annualized values of climate benefits, EPA
uses the same discount rate as the rate used to discount the value of damages from future GHG
emissions, for internal consistency. That approach to discounting follows the same approach that
the February 2021 TSD recommends "to ensure internal consistency—i.e., future damages from
climate change using the SC-GHG at 2.5 percent should be discounted to the base year of the
analysis using the same 2.5 percent rate." EPA has also consulted the National Academies' 2017
recommendations on how SC-GHG estimates can "be combined in RIAs with other cost and
benefits estimates that may use different discount rates." The National Academies reviewed
"several options," including "presenting all discount rate combinations of other costs and
benefits with [SC-GHG] estimates".
While the IWG works to assess how best to incorporate the latest, peer reviewed science
to develop an updated set of SC-GHG estimates, it recommended the interim estimates to be the
most recent estimates developed by the IWG prior to the group being disbanded in 2017. The
estimates rely on the same models and harmonized inputs and are calculated using a range of
discount rates. As explained in the February 2021 TSD, the IWG has concluded that it is
appropriate for agencies to revert to the same set of four values drawn from the SC-GHG
distributions based on three discount rates as were used in regulatory analyses between 2010 and
2016 and subject to public comment. For each discount rate, the IWG combined the distributions
across models and socioeconomic emissions scenarios (applying equal weight to each) and then
selected a set of four values for use in agency analyses: an average value resulting from the
model runs for each of three discount rates (2.5 percent, 3 percent, and 5 percent), plus a fourth
value, selected as the 95th percentile of estimates based on a 3 percent discount rate. The fourth
value was included to represent the extensive evidence in the scientific and economic literature
of the potential for lower-probability, higher-impact outcomes from climate change, which
would be particularly harmful to society and thus relevant to the public and policymakers.
Absent formal inclusion of risk aversion in the modeling, considering values above the mean in a
right skewed distribution with long tails acknowledges society's preference for avoiding risk
when high consequence outcomes are possible. As explained in the February 2021 TSD, this
update reflects the immediate need to have an operational SC-GHG that was developed using a
transparent process, peer-reviewed methodologies, and the science available at the time of that
185
-------
process. Those estimates were subject to public comment in the context of dozens of proposed
rulemakings as well as in a dedicated public comment period in 2013.
Table 4.2.4.1-1 summarizes the interim SC-CO2 estimates for the years 2023-2054.
These estimates are reported in 2020 dollars in the IWG's 2021 TSD but are otherwise identical
to those presented in the IWG's 2016 TSD.281 For purposes of capturing uncertainty around the
SC-CO2 estimates in analyses, the February 2021 TSD emphasizes the importance of considering
all four of the SC-CO2 values. The SC-CO2 increases over time within the models (i.e., the
societal harm from one metric ton emitted in 2030 is higher than the harm caused by one metric
ton emitted in 2025) because future emissions produce larger incremental damages as physical
and economic systems become more stressed in response to greater climatic change, and because
GDP is growing over time and many damage categories are modeled as proportional to GDP.
281IWG. (2021). Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim
Estimates under Executive Order 13990. Retrieved from Washington, DC: fattps://www.wfaitehonse.gov/wp-
eontent/nptoads/2021/02/TeefaiiiealSnpportDoeument SocialCostofCarbonMethaneNitronsOxide.pdf.
186
-------
Table 4.2.4.1-1: Interim Social Cost of Carbon Values, 2023-2054 (2022$/Metric Ton
CP2)282
Emissions
Year
Discount Rate and Statistic
5% Average
3% Average
2.5% Average
3% 95th Percentile
2023
$18
$61
$90
$182
2024
$18
$62
$91
$185
2025
$19
$63
$93
$189
2026
$19
$64
$94
$193
2027
$20
$66
$96
$197
2028
$21
$67
$97
$201
2029
$21
$68
$99
$205
2030
$22
$69
$100
$209
2031
$22
$70
$102
$213
2032
$23
$72
$103
$218
2033
$24
$73
$105
$222
2034
$24
$74
$106
$226
2035
$25
$76
$108
$230
2036
$26
$77
$109
$235
2037
$26
$78
$111
$239
2038
$27
$79
$112
$243
2039
$28
$81
$114
$248
2040
$28
$82
$115
$252
2041
$29
$83
$117
$256
2042
$30
$85
$118
$260
2043
$30
$86
$120
$264
2044
$31
$87
$121
$267
2045
$32
$88
$123
$271
2046
$33
$90
$124
$275
2047
$33
$91
$126
$279
2048
$34
$92
$127
$283
2049
$35
$93
$129
$287
2050
$35
$95
$130
$291
2051
$36
$95
$132
$292
2052
$37
$96
$133
$293
2053
$38
$97
$135
$294
2054
$38
$99
$136
$295
282 The 2023-2050 SC-C02 values are identical to those reported in the February 2021 TSD (IWG 2021) adjusted to
2022 dollars using the annual GDP Implicit Price Deflator values in the U.S. Bureau of Economic Analysis' (BEA)
NIPA Table 1.1.9 (U.S. BEA 2022). This table displays the values rounded to the nearest dollar; the annual
unrounded values used in the calculations in this analysis are available on OMB's website:
https://www.whitehonse.gOv/omb/information-regiiiatorv-affairs/regniatore-matters/#scghgs.
187
-------
The February 2021 TSD provides SC-GHG estimates through emissions year 2050.
Estimates were extended for the period 2051 to 2054 using the IWG methods, assumptions, and
parameters identical to the 2020-2050 estimates. Specifically, 2051-2054 SC-GHG estimates
were calculated in Mimi.jl, an open-source modular computing platform used for creating,
running, and performing analyses on IAMs (www.mimiframework.org). For CO2, the 2051-2054
SC-GHG values were calculated by linearly interpolating between the 2050 TSD values and the
2055 Mimi-based values. The annual unrounded 2051-2054 values used in the calculations in
this document are available in the docket for this action, and the replication code is available
upon request.
There are a number of limitations and uncertainties associated with the SC-CO2 estimates
presented in Table 4.2.4.1-1. Some uncertainties are captured within the analysis, while other
areas of uncertainty have not yet been quantified in a way that can be modeled. Figure 4.2.4.1-1
presents the quantified sources of uncertainty in the form of frequency distributions for the SC-
CO2 estimates for emissions in 2030 (in 2022$). The distribution of the SC-CO2 estimate reflects
uncertainty in key model parameters such as the equilibrium climate sensitivity, as well as
uncertainty in other parameters set by the original model developers. To highlight the difference
between the impact of the discount rate and other quantified sources of uncertainty, the bars
below the frequency distributions provide a symmetric representation of quantified variability in
the SC-CO2 estimates for each discount rate. As illustrated by the figure, the assumed discount
rate plays a critical role in the ultimate estimate of the SC-CO2. This is because CO2 emissions
today continue to impact society far out into the future, so with a higher discount rate, costs that
accrue to future generations are weighted less, resulting in a lower estimate. As discussed in the
February 2021 TSD, there are other sources of uncertainty that have not yet been quantified and
are thus not reflected in these estimates.
188
-------
Figure 4.2.4.1-1: Frequency Distribution of SC-CO2 Estimates for 2030
283
if)
^ -1
o
m
o -J
5% Average = $22
"TT
3% Average = $69
2.5% Average = $100
3%
95th Pet.
$209
I
Discount Rate
~ 5.0%
3.0%
2.5%
~
~
5th - 95"' Percentile
~ i of Simulations
TTT
nth I
T
0 20
"TTTI
40
TTTI
60
TTTI
80
I I I I I
100
I III
140
TTT
"TT
180
TTTT
220
I I I I I I I I I
260
300
I I I I
340
TTT
380
Social Cost of Carbon in 2030 [2022$ / metric ton C02]
The interim SC-CO2 estimates presented in Table 4.2.4.1-1 have a number of other
limitations. First, the current scientific and economic understanding of discounting approaches
suggests discount rates appropriate for intergenerational analysis in the context of climate change
are likely to be less than 3 percent, near 2 percent or lower.284 Second, the IAMs used to produce
these interim estimates do not include all of the important physical, ecological, and economic
impacts of climate change recognized in the climate change literature and the science underlying
their "damage functions" - i.e., the core parts of the IAMs that map global mean temperature
changes and other physical impacts of climate change into economic (both market and
nonmarket) damages - lags behind the most recent research. For example, limitations include the
incomplete treatment of catastrophic and non-catastrophic impacts in the integrated assessment
models, their incomplete treatment of adaptation and technological change, the incomplete way
in which inter-regional and intersectoral linkages are modeled, uncertainty in the extrapolation of
damages to high temperatures, and inadequate representation of the relationship between the
discount rate and uncertainty in economic growth over long time horizons. Likewise, the
socioeconomic and emissions scenarios used as inputs to the models do not reflect new
information from the last decade of scenario generation or the full range of projections.
283 Although the distributions and numbers are based on the full set of model results (150,000 estimates for each
discount rate and gas), for display purposes the horizontal axis is truncated with 0.584 to 0.96 percent of the
estimates falling below the lowest bin displayed and 0.38 to 3.9 percent of the estimates falling above the highest bin
displayed, depending on the discount rate and GHG.
284 IWG. (2021). Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim
Estimates under Executive Order 13990. Retrieved from Washington, DC: https://www.whitehouse.gov/wp-
content/uploads/2021/02/TechnicalSupportDocument SocialCostofCarbonMethaneNitrousOxide.pdf
189
-------
The modeling limitations do not all work in the same direction in terms of their influence
on the SC-GHG estimates. However, as discussed in the February 2021 TSD, the IWG has
recommended that, taken together, the limitations suggest that the SC-GHG estimates used in
this final rule likely underestimate the damages from GHG emissions. EPA concurs that the
values used in this rulemaking conservatively underestimate the climate benefits associated with
GHG emission reductions. In particular, the Intergovernmental Panel on Climate Change (IPCC)
Fourth Assessment Report, which was the most current IPCC assessment available at the time
when the IWG decision over the ECS input was made, concluded that SC-CO2 estimates "very
likely.. .underestimate the damage costs" due to omitted impacts.285 Since then, the peer-
reviewed literature has continued to support this conclusion, as noted in the IPCC's Fifth
Assessment report and other recent scientific assessments.286'287'288'289'290'291'292'293 These
assessments confirm and strengthen the science, updating projections of future climate change
and documenting and attributing ongoing changes. For example, sea level rise projections from
the IPCC's Fourth Assessment report ranged from 18 to 59 centimeters by the 2090s relative to
1980-1999, while excluding any dynamic changes in ice sheets due to the limited understanding
of those processes at the time. A decade later, the Fourth National Climate Assessment projected
a substantially larger sea level rise of 30 to 130 centimeters by the end of the century relative to
285 Intergovernmental Panel on Climate Change (IPCC). 2007. Core Writing Team; Pachauri, R.K; and Reisinger, A.
(ed.), Climate Change 2007: Synthesis Report, Contribution of Working Groups I, II and III to the Fourth
Assessment Report of the Intergovernmental Panel on Climate Change, IPCC, ISBN 92-9169-122-4.
286 Intergovernmental Panel on Climate Change (IPCC). 2014. Climate Change 2014: Synthesis Report.
Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on
Climate Change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 pp.
287 Intergovernmental Panel on Climate Change (IPCC). 2018. Global Warming of 1.5°C. An IPCC Special Report
on the impacts of global warming of 1.5°C above pre-industrial levels and related global greenhouse gas emission
pathways, in the context of strengthening the global response to the threat of climate change, sustainable
development, and efforts to eradicate poverty [Masson-Delmotte, V., P. Zhai, H.-O. Portner, D. Roberts, J. Skea,
P.R. Shukla, A. Pirani, W. Moufouma-Okia, C. Pean, R. Pidcock, S. Connors, J.B.R. Matthews, Y. Chen, X. Zhou,
M.I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T. Waterfield (eds.)].
288 Intergovernmental Panel on Climate Change (IPCC). 2019a. Climate Change and Land: an IPCC special report
on climate change, desertification, land degradation, sustainable land management, food security, and greenhouse
gas fluxes in terrestrial ecosystems [P.R. Shukla, J. Skea, E. Calvo Buendia, V. Masson-Delmotte, H.-O. Portner, D.
C. Roberts, P. Zhai, R. Slade, S. Connors, R. van Diemen, M. Ferrat, E. Haughey, S. Luz, S. Neogi, M. Pathak, J.
Petzold, J. Portugal Pereira, P. Vyas, E. Huntley, K. Kissick, M. Belkacemi, J. Malley, (eds.)].
289 Intergovernmental Panel on Climate Change (IPCC). 2019b. IPCC Special Report on the Ocean and Cryosphere
in a Changing Climate [H.-O. Portner, D.C. Roberts, V. Masson-Delmotte, P. Zhai, M. Tignor, E. Poloczanska, K.
Mintenbeck, A. Alegria, M. Nicolai, A. Okem, J. Petzold, B. Rama, N.M. Weyer (eds.)].
290 U.S. Global Change Research Program (USGCRP). 2016. The Impacts of Climate Change on Human Health in
the United States: A Scientific Assessment. Crimmins, A., J. Balbus, J.L. Gamble, C.B. Beard, J.E. Bell, D. Dodgen,
R.J. Eisen, N. Fann, M.D. Hawkins, S.C. Herring, L. Jantarasami, D.M. Mills, S. Saha, M.C. Sarofim, J. Trtanj, and
L. Ziska, Eds. U.S. Global Change Research Program, Washington, DC, 312 pp.
https://dx.doi.org/10.7930/J0R49NOX.
291 U.S. Global Change Research Program (USGCRP). 2018. Impacts, Risks, and Adaptation in the United States:
Fourth National Climate Assessment, Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel,
K.L.M. Lewis, T.K. Maycock, and B.C. Stewart (eds.)]. U.S. Global Change Research Program, Washington, DC,
USA, 1515 pp. doi: 10.7930/NCA4.2018.
292 National Academies of Sciences, Engineering, and Medicine (National Academies). 2016b. Attribution of
Extreme Weather Events in the Context of Climate Change. Washington, DC: The National Academies Press.
https://doi.org/10.17226/21852.
293 National Academies of Sciences, Engineering, and Medicine (National Academies). 2019. Climate Change and
Ecosystems. Washington, DC: The National Academies Press, https://doi.org/10.17226/25504.
190
-------
2000, while not ruling out even more extreme outcomes.294 EPA has reviewed and considered
the limitations of the models used to estimate the interim SC-GHG estimates and concurs with
the February 2021 SC-GHG TSD's assessment that, taken together, the limitations suggest that
the interim SC-GHG estimates likely underestimate the damages from GHG emissions.
The February 2021 TSD briefly previews some of the recent advances in the scientific
and economic literature that the IWG is actively following and that could provide guidance on,
or methodologies for, addressing some of the limitations with the interim SC-GHG estimates.
The IWG is currently working on a comprehensive update of the SC-GHG estimates taking into
consideration recommendations from the National Academies of Sciences, Engineering and
Medicine, recent scientific literature, public comments received on the February 2021 TSD, and
other input from experts and diverse stakeholder groups (National Academies, 2017). While that
process continues EPA is continuously reviewing developments in the scientific literature on the
SC-GHG, including more robust methodologies for estimating damages from emissions, and
looking for opportunities to further improve SC-GHG estimation going forward. Most recently,
EPA presented a draft set of updated SC-GHG estimates within a sensitivity analysis in the
regulatory impact analysis of EPA's November 2022 supplemental proposal for oil and gas
standards that that aims to incorporate recent advances in the climate science and economics
literature. Specifically, the draft updated methodology incorporates new literature and research
consistent with the National Academies near-term recommendations on socioeconomic and
emissions inputs, climate modeling components, discounting approaches, and treatment of
uncertainty, and an enhanced representation of how physical impacts of climate change translate
to economic damages in the modeling framework based on the best and readily adaptable
damage functions available in the peer reviewed literature. EPA solicited public comment on the
sensitivity analysis and the accompanying draft technical report, which explains the methodology
underlying the new set of estimates, in the docket for the proposed oil and gas rule. EPA is also
conducting an external peer review of this technical report. More information about this process
and public comment opportunities is available on EPA's website.295 EPA's draft technical report
will be among the many technical inputs available to the IWG as it continues its work. Tables
4.2.3-6 through Tables 4.2.3-13 show the estimated changes in CChe for the volume changes
analyzed in each year, 2023-2025. This analysis portrays what might be expected if, in each of
the ensuing 29 years, aggregate renewable fuel consumption for each category exceeded baseline
levels by the same volume associated with the rule. EPA estimated the dollar value of these
GHG-related effects for each analysis year between 2023 through 2054 by applying the SC-CO2
estimates, shown in Table 4.2.4.1-1, to the estimated changes in GHG emissions inventories
resulting from the candidate volumes. EPA then calculated the present value and annualized
benefits from the perspective of each year by discounting each year-specific value to that year
using the same discount rate used to calculate the SC-CO2.296
294 U.S. Global Change Research Program (USGCRP). 2018. Impacts, Risks, and Adaptation in the United States:
Fourth National Climate Assessment, Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel,
K.L.M. Lewis, T.K. Maycock, and B.C. Stewart (eds.)]. U.S. Global Change Research Program, Washington, DC,
USA, 1515 pp. doi: 10.7930/NCA4.2018.
295 See https://www.epa.gov/eiivironmental-economies/scghg.
296 According to OMB's Circular A-4 (OMB, 2003), an "analysis should focus on benefits and costs that accrue to
citizens and residents of the United States", and international effects should be reported, but separately. Circular A-4
also reminds analysts that "[different regulations may call for different emphases in the analysis, depending on the
191
-------
4.2.4.2 Results
For this illustrative scenario, the interim estimates for carbon dioxide from the February
2021 TSD, presented in Table 4.2.4.1-1, were used to estimate the social benefits of the
estimated 30-year stream of GHG impacts presented in Chapter 4.2.3. For each year, the total of
emissions changes presented in Tables 4.2.3-6 through 4.2.3-13 are multiplied by each of the
four SC-CO2 values for the same year. Values for each year and discount rate statistic are then
converted to present value using the corresponding discount rates. The resulting streams of
estimated social benefits of the biofuel volume changes assumed in this illustrative scenario for
the 2023-2025 standards are presented in Tables 4.2.4.2-1 through 4.2.4.2-8. Note that in these
tables, volume changes for the 2023-2025 standards are relative to the No RFS baseline. We
separately estimate the social benefits of the biofuel volume changes assumed in this illustrative
scenario from the 2023 supplemental volume requirement as described in Chapter 3.3. We
present the results for these supplemental volumes in Tables 4.2.4.2-9 and 4.2.4.2-10. All
calculations are available in a spreadsheet in the docket for this rule.297
nature and complexity of the regulatory issues." To correctly assess the total climate damages to U.S. citizens and
residents, an analysis should account for all the ways climate impacts affect the welfare of U.S. citizens and
residents, including how U.S. GHG mitigation activities affect mitigation activities by other countries, and spillover
effects from climate action elsewhere. The SC-GHG estimates used in regulatory analysis under revoked EO 13783
were a limited approximation of some of the U.S. specific climate damages from GHG emissions. These estimates
range from $8 per metric ton CO2 for emissions occurring in 2023 to $13 per metric ton CO2 for emissions occurring
in 2054. However, as discussed at length in the IWG's February 2021 SC-GHG TSD, these estimates are an
underestimate of the benefits of GHG mitigation accruing to U.S. citizens and residents, as well as being subject to a
considerable degree of uncertainty due to the manner in which they are derived. In particular, as discussed in this
analysis, EPA concurs with the assessment in the February 2021 SC-GHG TSD that the estimates developed under
revoked E.O. 13783 did not capture significant regional interactions, spillovers, and other effects and so are
incomplete underestimates. As the U.S. Government Accountability Office (GAO) concluded in a June 2020 report
examining the SC-GHG estimates developed under E.O. 13783, the models "were not premised or calibrated to
provide estimates of the social cost of carbon based on domestic damages" p.29 (U.S. GAO, 2020). Further, the
report noted that the National Academies found that country-specific social costs of carbon estimates were "limited
by existing methodologies, which focus primarily on global estimates and do not model all relevant interactions
among regions" p.26 (U.S. GAO, 2020). It is also important to note that the SC-GHG estimates developed under
E.O. 13783 were never peer reviewed, and when their use in a specific regulatory action was challenged, the U.S.
District Court for the Northern District of California determined that use of those values had been "soundly rejected
by economists as improper and unsupported by science," and that the values themselves omitted key damages to
U.S. citizens and residents including to supply chains, U.S. assets and companies, and geopolitical security. The
Court found that by omitting such impacts, those estimates "fail[ed] to consider.. .important aspect[s] of the
problem" and departed from the "best science available" as reflected in the global estimates. California v. Bernhardt,
472 F. Supp. 3d 573, 613-14 (N.D. Cal. 2020). EPA continues to center attention in this analysis on the global
measures of the SC-GHG as the appropriate estimates given the flaws in the U.S. specific estimates, and as
necessary for all countries to use to achieve an efficient allocation of resources for emissions reduction on a global
basis, and so benefit the U.S. and its citizens.
297 See "GHG Scenario for 2023-25 Set Rule (FRM).xlsx," available in the docket for this rule.
192
-------
Table 4.2.4.2-1: Present value of 30-year stream of climate benefits for 2023 standards,
using low biofuel/high petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2022$)a
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$(2,649)
$(9,021)
$(13,350)
$(26,943)
2024
$794
$2,726
$4,040
$8,156
2025
$778
$2,697
$4,004
$8,086
2026
$763
$2,668
$3,968
$8,014
2027
$747
$2,639
$3,931
$7,939
2028
$731
$2,609
$3,894
$7,861
2029
$714
$2,578
$3,856
$7,781
2030
$698
$2,547
$3,818
$7,699
2031
$685
$2,518
$3,781
$7,628
2032
$671
$2,490
$3,744
$7,554
2033
$657
$2,460
$3,707
$7,479
2034
$643
$2,431
$3,669
$7,401
2035
$629
$2,400
$3,631
$7,321
2036
$615
$2,370
$3,593
$7,240
2037
$601
$2,339
$3,554
$7,158
2038
$586
$2,308
$3,515
$7,074
2039
$572
$2,277
$3,476
$6,988
2040
$558
$2,246
$3,437
$6,902
2041
$545
$2,215
$3,396
$6,805
2042
$532
$2,183
$3,356
$6,708
2043
$446
$1,852
$2,853
$5,689
2044
$435
$1,825
$2,818
$5,605
2045
$424
$1,798
$2,783
$5,522
2046
$413
$1,771
$2,748
$5,438
2047
$402
$1,744
$2,714
$5,355
2048
$391
$1,717
$2,679
$5,272
2049
$381
$1,690
$2,644
$5,190
2050
$370
$1,663
$2,610
$5,107
2051
$362
$1,625
$2,582
$4,975
2052
$351
$1,595
$2,542
$4,846
2053
$-
$-
$-
$-
2054
$-
$-
$-
$-
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
193
-------
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
Table 4.2.4.2-2: Present value of 30-year stream of climate benefits for 2024 standards,
using low biofuel/high petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2022$)a
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$-
$-
$-
$-
2024
$(161)
$(551)
$(817)
$(1,650)
2025
$94
$325
$483
$975
2026
$92
$322
$478
$966
2027
$90
$318
$474
$957
2028
$88
$314
$469
$947
2029
$86
$311
$465
$938
2030
$84
$307
$460
$928
2031
$83
$304
$456
$919
2032
$81
$300
$451
$910
2033
$79
$296
$447
$901
2034
$78
$293
$442
$892
2035
$76
$289
$438
$882
2036
$74
$286
$433
$873
2037
$72
$282
$428
$863
2038
$71
$278
$424
$852
2039
$69
$274
$419
$842
2040
$67
$271
$414
$832
2041
$66
$267
$409
$820
2042
$64
$263
$404
$808
2043
$63
$259
$400
$797
2044
$56
$235
$363
$721
2045
$55
$231
$358
$711
2046
$53
$228
$354
$700
2047
$52
$224
$349
$689
2048
$50
$221
$345
$678
2049
$49
$217
$340
$668
2050
$48
$214
$336
$657
2051
$47
$209
$332
$640
2052
$45
$205
$327
$624
2053
$44
$201
$322
$608
2054
$-
$-
$-
$-
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
194
-------
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
195
-------
Table 4.2.4.2-3: Present value of 30-year stream of climate benefits for 2025 standards,
using low biofuel/high petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2022$)a
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$-
$-
$-
$-
2024
$-
$-
$-
$-
2025
$(162)
$(561)
$(833)
$(1,683)
2026
$164
$575
$855
$1,727
2027
$161
$569
$847
$1,711
2028
$157
$562
$839
$1,694
2029
$154
$556
$831
$1,677
2030
$150
$549
$823
$1,659
2031
$148
$543
$815
$1,644
2032
$145
$537
$807
$1,628
2033
$142
$530
$799
$1,612
2034
$139
$524
$791
$1,595
2035
$136
$517
$783
$1,578
2036
$133
$511
$774
$1,561
2037
$129
$504
$766
$1,543
2038
$126
$498
$758
$1,525
2039
$123
$491
$749
$1,506
2040
$120
$484
$741
$1,488
2041
$117
$477
$732
$1,467
2042
$115
$471
$723
$1,446
2043
$112
$464
$715
$1,425
2044
$109
$457
$706
$1,404
2045
$99
$421
$652
$1,293
2046
$97
$415
$643
$1,273
2047
$94
$408
$635
$1,254
2048
$92
$402
$627
$1,234
2049
$89
$396
$619
$1,215
2050
$87
$389
$611
$1,196
2051
$85
$380
$604
$1,165
2052
$82
$373
$595
$1,135
2053
$80
$367
$586
$1,105
2054
$77
$360
$577
$1,077
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
196
-------
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
197
-------
Table 4.2.4.2-4: Present value of 30-year stream of climate benefits for the combined 2023-
2025 standards, using low biofuel/high petroleum lifecycle analysis estimates, relative to the
No RFS baseline, presented with four values for the social cost of carbon (SC-CO2)
(millions of 2022$)"
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$(2,649)
$(9,021)
$(13,350)
$(26,943)
2024
$633
$2,174
$3,222
$6,506
2025
$710
$2,461
$3,653
$7,378
2026
$1,019
$3,565
$5,301
$10,707
2027
$998
$3,525
$5,252
$10,606
2028
$976
$3,485
$5,202
$10,503
2029
$954
$3,444
$5,152
$10,396
2030
$932
$3,403
$5,101
$10,286
2031
$915
$3,365
$5,052
$10,191
2032
$897
$3,326
$5,003
$10,093
2033
$878
$3,287
$4,953
$9,992
2034
$860
$3,247
$4,902
$9,888
2035
$841
$3,207
$4,852
$9,782
2036
$822
$3,166
$4,800
$9,673
2037
$802
$3,126
$4,749
$9,563
2038
$783
$3,084
$4,697
$9,451
2039
$764
$3,043
$4,644
$9,337
2040
$745
$3,001
$4,592
$9,222
2041
$728
$2,959
$4,538
$9,092
2042
$710
$2,917
$4,484
$8,963
2043
$621
$2,575
$3,967
$7,911
2044
$600
$2,517
$3,887
$7,731
2045
$578
$2,450
$3,793
$7,525
2046
$563
$2,413
$3,745
$7,412
2047
$548
$2,376
$3,698
$7,298
2048
$533
$2,340
$3,651
$7,185
2049
$519
$2,303
$3,603
$7,072
2050
$504
$2,267
$3,556
$6,960
2051
$494
$2,214
$3,518
$6,780
2052
$479
$2,174
$3,464
$6,605
2053
$124
$568
$908
$1,713
2054 b
$77
$360
$577
$1,077
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
198
-------
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
b Combined impacts presented in Table 4.2.4.2-4 are the sum of the three thirty-year streams of impacts for the 2023
through 2025 standards presented in Tables 4.2.4.2-1, 4.2.4.2-2, and 4.2.4.2-3. Because we assess thirty years of
impacts for each year standards, the period of analysis for the 2023 standards extends to 2052.
199
-------
Table 4.2.4.2-5: Present value of 30-year stream of climate benefits for 2023 standards,
using high biofuel/low petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2022$)a
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$(4,653)
$(15,845)
$(23,449)
$(47,324)
2024
$125
$430
$637
$1,286
2025
$123
$425
$631
$1,275
2026
$120
$421
$625
$1,263
2027
$118
$416
$620
$1,251
2028
$115
$411
$614
$1,239
2029
$113
$406
$608
$1,226
2030
$110
$401
$602
$1,213
2031
$108
$397
$596
$1,202
2032
$106
$392
$590
$1,191
2033
$104
$388
$584
$1,179
2034
$101
$383
$578
$1,167
2035
$99
$378
$572
$1,154
2036
$97
$374
$566
$1,141
2037
$95
$369
$560
$1,128
2038
$92
$364
$554
$1,115
2039
$90
$359
$548
$1,101
2040
$88
$354
$542
$1,088
2041
$86
$349
$535
$1,073
2042
$84
$344
$529
$1,057
2043
$223
$924
$1,423
$2,838
2044
$217
$910
$1,406
$2,797
2045
$212
$897
$1,389
$2,755
2046
$206
$883
$1,371
$2,713
2047
$201
$870
$1,354
$2,672
2048
$195
$857
$1,336
$2,630
2049
$190
$843
$1,319
$2,589
2050
$185
$830
$1,302
$2,548
2051
$181
$811
$1,288
$2,482
2052
$175
$796
$1,268
$2,418
2053
$-
$-
$-
$-
2054
$-
$-
$-
$-
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
200
-------
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
201
-------
Table 4.2.4.2-6: Present value of 30-year stream of climate benefits for 2024 standards,
using high biofuel/low petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2022$)a
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$-
$-
$-
$-
2024
$(366)
$(1,258)
$(1,865)
$(3,765)
2025
$(9)
$(33)
$(49)
$(98)
2026
$(9)
$(32)
$(48)
$(97)
2027
$(9)
$(32)
$(48)
$(97)
2028
$(9)
$(32)
$(47)
$(96)
2029
$(9)
$(31)
$(47)
$(95)
2030
$(8)
$(31)
$(46)
$(94)
2031
$(8)
$(31)
$(46)
$(93)
2032
$(8)
$(30)
$(46)
$(92)
2033
$(8)
$(30)
$(45)
$(91)
2034
$(8)
$(30)
$(45)
$(90)
2035
$(8)
$(29)
$(44)
$(89)
2036
$(7)
$(29)
$(44)
$(88)
2037
$(7)
$(28)
$(43)
$(87)
2038
$(7)
$(28)
$(43)
$(86)
2039
$(7)
$(28)
$(42)
$(85)
2040
$(7)
$(27)
$(42)
$(84)
2041
$(7)
$(27)
$(41)
$(83)
2042
$(6)
$(27)
$(41)
$(82)
2043
$(6)
$(26)
$(40)
$(80)
2044
$4
$18
$27
$54
2045
$4
$17
$27
$54
2046
$4
$17
$27
$53
2047
$4
$17
$26
$52
2048
$4
$17
$26
$51
2049
$4
$16
$26
$50
2050
$4
$16
$25
$50
2051
$4
$16
$25
$48
2052
$3
$15
$25
$47
2053
$3
$15
$24
$46
2054
$-
$-
$-
$-
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
202
-------
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
203
-------
Table 4.2.4.2-7: Present value of 30-year stream of climate benefits for 2025 standards,
using high biofuel/low petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2022$)a
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$-
$-
$-
$-
2024
$-
$-
$-
$-
2025
$(444)
$(1,540)
$(2,286)
$(4,617)
2026
$10
$34
$50
$101
2027
$9
$33
$50
$100
2028
$9
$33
$49
$99
2029
$9
$32
$49
$98
2030
$9
$32
$48
$97
2031
$9
$32
$48
$96
2032
$8
$31
$47
$95
2033
$8
$31
$47
$94
2034
$8
$31
$46
$93
2035
$8
$30
$46
$92
2036
$8
$30
$45
$91
2037
$8
$29
$45
$90
2038
$7
$29
$44
$89
2039
$7
$29
$44
$88
2040
$7
$28
$43
$87
2041
$7
$28
$43
$86
2042
$7
$28
$42
$85
2043
$7
$27
$42
$83
2044
$6
$27
$41
$82
2045
$19
$82
$127
$253
2046
$19
$81
$126
$249
2047
$18
$80
$124
$245
2048
$18
$79
$123
$241
2049
$17
$77
$121
$237
2050
$17
$76
$119
$234
2051
$17
$74
$118
$228
2052
$16
$73
$116
$222
2053
$16
$72
$114
$216
2054
$15
$70
$113
$210
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
204
-------
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
205
-------
Table 4.2.4.2-8: Present value of 30-year stream of climate benefits for the combined 2023-
2025 standards, using high biofuel/low petroleum lifecycle analysis estimates, relative to the
No RFS baseline, presented with four values for the social cost of carbon (SC-CO2)
(millions of 2022$)"
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$(4,653)
$(15,845)
$(23,449)
$(47,324)
2024
$(241)
$(829)
$(1,228)
$(2,479)
2025
$(331)
$(1,148)
$(1,704)
$(3,440)
2026
$121
$422
$627
$1,267
2027
$118
$417
$621
$1,255
2028
$115
$412
$615
$1,242
2029
$113
$407
$609
$1,230
2030
$110
$402
$603
$1,217
2031
$108
$398
$598
$1,205
2032
$106
$393
$592
$1,194
2033
$104
$389
$586
$1,182
2034
$102
$384
$580
$1,170
2035
$99
$379
$574
$1,157
2036
$97
$375
$568
$1,144
2037
$95
$370
$562
$1,131
2038
$93
$365
$556
$1,118
2039
$90
$360
$549
$1,105
2040
$88
$355
$543
$1,091
2041
$86
$350
$537
$1,076
2042
$84
$345
$530
$1,060
2043
$223
$925
$1,425
$2,841
2044
$228
$955
$1,475
$2,933
2045
$235
$997
$1,543
$3,061
2046
$229
$982
$1,524
$3,015
2047
$223
$967
$1,504
$2,969
2048
$217
$952
$1,485
$2,923
2049
$211
$937
$1,466
$2,877
2050
$205
$922
$1,447
$2,831
2051
$201
$901
$1,431
$2,758
2052
$195
$884
$1,409
$2,687
2053
$19
$87
$139
$262
2054 b
$15
$70
$113
$210
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
206
-------
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
b Combined impacts presented in Table 4.2.4.2-8 are the sum of the three thirty-year streams of impacts for the 2023
through 2025 standards presented in Tables 4.2.4.2-5, 4.2.4.2-6, and 4.2.4.2-7. Because we assess thirty years of
impacts for each year standards, the period of analysis for the 2023 standards extends to 2052.
207
-------
Table 4.2.4.2-9: Present value of 30-year stream of climate benefits for 2023 supplemental
volume requirement, using low biofuel/high petroleum lifecycle analysis estimates,
presented with
'our values for the social cost of carbon (SC-CO2) (millions of 2022$)a
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$(225)
$(767)
$(1,135)
$(2,290)
2024
$34
$118
$175
$353
2025
$34
$117
$173
$350
2026
$33
$115
$172
$347
2027
$32
$114
$170
$343
2028
$32
$113
$168
$340
2029
$31
$112
$167
$337
2030
$30
$110
$165
$333
2031
$30
$109
$164
$330
2032
$29
$108
$162
$327
2033
$28
$106
$160
$324
2034
$28
$105
$159
$320
2035
$27
$104
$157
$317
2036
$27
$103
$155
$313
2037
$26
$101
$154
$310
2038
$25
$100
$152
$306
2039
$25
$99
$150
$302
2040
$24
$97
$149
$299
2041
$24
$96
$147
$294
2042
$23
$94
$145
$290
2043
$16
$67
$104
$207
2044
$16
$66
$102
$204
2045
$15
$65
$101
$200
2046
$15
$64
$100
$197
2047
$15
$63
$99
$194
2048
$14
$62
$97
$191
2049
$14
$61
$96
$188
2050
$13
$60
$95
$185
2051
$13
$59
$94
$181
2052
$13
$58
$92
$176
2053
$-
$-
$-
$-
2054
$-
$-
$-
$-
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
208
-------
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
Table 4.2.4.2-10: Present value of 30-year stream of climate benefits for 2023 supplemental
volume requirement, using high biofuel/low petroleum lifecycle analysis estimates,
presented with
'our values for the social cost of carbon (SC-CO2) (millions of 2022$)a
Year
Rate: 5%
Rate: 3%
Rate: 2.5%
Rate: 3%
95th percentile
2023
$(347)
$(1,181)
$(1,748)
$(3,527)
2024
$5
$17
$26
$52
2025
$5
$17
$25
$51
2026
$5
$17
$25
$51
2027
$5
$17
$25
$50
2028
$5
$17
$25
$50
2029
$5
$16
$24
$49
2030
$4
$16
$24
$49
2031
$4
$16
$24
$48
2032
$4
$16
$24
$48
2033
$4
$16
$24
$48
2034
$4
$15
$23
$47
2035
$4
$15
$23
$47
2036
$4
$15
$23
$46
2037
$4
$15
$23
$45
2038
$4
$15
$22
$45
2039
$4
$14
$22
$44
2040
$4
$14
$22
$44
2041
$3
$14
$22
$43
2042
$3
$14
$21
$43
2043
$14
$57
$87
$174
2044
$13
$56
$86
$172
2045
$13
$55
$85
$169
2046
$13
$54
$84
$167
2047
$12
$53
$83
$164
2048
$12
$53
$82
$162
2049
$12
$52
$81
$159
2050
$11
$51
$80
$156
2051
$11
$50
$79
$152
2052
$11
$49
$78
$148
2053
$-
$-
$-
$-
2054
$-
$-
$-
$-
a This analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
209
-------
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
We note that the methodology underlying the SC-CO2 estimates used in this analysis has
been subject to public comment in the context of dozens of rulemakings as well as in a dedicated
public comment period in 2013. We note that there is an ongoing interagency process to update
the SC-GHG estimates, and there will be further opportunity to provide public input on the SC-
GHG methodology through that process.298 As part of that separate process, the EPA welcomes
the opportunity to continually improve its understanding through public input on the analytical
issues associated with the presentation of anticipated costs, benefits, and other impacts of its
actions, as done through RIAs.
4.3 Conversion of Wetlands, Ecosystems, and Wildlife Habitats
The Second Triennial Report to Congress on Biofuels299 summarized the numerous
studies that have examined changes in wetlands, ecosystems, and wildlife habitats. The Report
noted, for example, there has been an observed increase in acreage planted with soybeans and
corn between the decade leading up to enactment of EISA and the decade following enactment.
Evidence from observations of land use change suggests that some of this increase in acreage
and crop use is a consequence of increased biofuel production. It is likely that the environmental
and natural resource impacts associated with land use change are, at least in part, due to
increased biofuel production and use. A more in-depth evaluation of cropland conversion and its
impacts to the environment can be found in the May 19, 2023 biological evaluation that EPA
submitted to the U.S. Fish and Wildlife Service and the National Marine Fisheries Service
(together, "the Services") on May 20, 2023, and the May 31, 2023 addendum that EPA provided
to NMFS in response to a follow-up request (together, "the May 19 BE").300 Although the
discussion which follows is not as detailed as in the May 19 BE, some aspects of it are discussed.
As can be seen in DRIA Chapter 4.2.2, there is a wide range of estimates for the area and types
of land use change depending on feedstock, model choice, scenario design and input
assumptions.301 This section focuses on impacts related to the domestic production of renewable
298 For example, EPA, on behalf of the IWG, published a Federal Register notice on January 25, 2022, to solicit
public nominations of scientific experts for the upcoming peer review the forthcoming update (87 FR 3801). EPA
has a webpage where additional information regarding the peer review process will be posted as it becomes
available: https://www.epa.gov/enviromnental-economics/scghg-tsd-peer-review. There will be a separate Federal
Register notice for the public comment period on the forthcoming SC-GHG technical support document once it is
released.
299 U.S. EPA (2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Environmental
Protection Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June 2018.
300 "Biological Evaluation of the Renewable Fuel Standard (RFS) Set Rule," May 19, 2023 & Email from
T. Phillips, EPA, to D. Baldwin, NOAA (May 31, 2023) are both available in the docket for this action.
301 The DRIA for the proposed rule includes a discussion of available models and land use change estimates that is
not part of this RIA for the final rule. The review of studies and land use change estimates in the DRIA remains
relevant, but we determined it did not bear repeating in this document as it does not factor directly into our analysis
of the climate impacts of the candidate volumes.
210
-------
fuels and their underlying feedstocks. Effects from the end use of renewable fuel (i.e., retail
station storage and dispensing, and combustion of renewable fuel in vehicles and engines) are
mostly from air quality effects (Chapter 4.1), climate effects (Chapter 4.2), and possible leakage
from underground storage tanks (Chapter 4.4.4). Insofar as there are impacts of renewable fuel
on the conversion of wetlands, ecosystems, and wildlife habitats, they are associated with crop-
based feedstocks rather than waste fats, oils, and greases, or biogas. The impacts of the candidate
volumes in many of these sections is compared to the 2022 baseline as a substitute to the No RFS
baseline as land use change for crop-based feedstocks would not be drastically affected by the
removal of the RFS program as referenced in Chapter 2. We note that to the extent the RFS
standards in this action are associated with increased palm oil production, either as a biofuel
feedstock or for other purposes (e.g., backfilling of soybean oil that has been diverted to biofuel
production), there is strong evidence that palm oil production is linked with degradation of
wetlands, ecosystems and wildlife habitats outside of the U.S., and other adverse environmental
impacts on air quality, soil quality and water quality outside of the U.S. Tropical forests are
carbon sinks, and their conversion for oil palm production results in both sequestered carbon
emissions and foregone future carbon sequestration. These impacts mitigate the potential GHG
benefit otherwise provided by biofuel displacement of fossil fuels.302
302 Austin, K. G., A. Schwantes, Y. Gu and P. S. Kasibhatla (2019). "What causes deforestation in Indonesia?"
Environmental Research Letters 14(2): 024007; Austin, K. G., M. Gonzalez-Roglich, D. Schaffer-Smith, A. M.
Schwantes and J. J. Swenson (2017). "Trends in size of tropical deforestation events signal increasing dominance of
industrial-scale drivers." Environmental Research Letters 12(5); Austin, K. G., A. Mosnier, J. Pirker, I. McCallum,
S. Fritz and P. S. Kasibhatla (2017). "Shifting patterns of oil palm driven deforestation in Indonesia and implications
for zero-deforestation commitments." Land Use Policy 69: 41-48; Babel, M. S., B. Shrestha and S. R. Perret (2011).
"Hydrological impact of biofuel production: A case study of the Khlong Phlo Watershed in Thailand." Agricultural
Water Management 101(1): 8-26.; Carlson, K. M., L. M. Curran, G. P. Asner, A. M. Pittman, S. N. Trigg and J. M.
Adeney (2013). "Carbon emissions from forest conversion by Kalimantan oil palm plantations." Nature Climate
Change 3(3): 283-287; Gatto, M., M. Wollni and M. Qaim (2015). "Oil palm boom and land-use dynamics in
Indonesia: The role of policies and socioeconomic factors." Land Use Policy 46: 292-303; Gaveau, D. L. A., D.
Sheil, M. A. Salim, S. Aijasakusuma, M. Ancrenaz, P. Pacheco and E. Meijaard (2016). "Rapid conversions and
avoided deforestation: Examining four decades of industrial plantation expansion in Borneo." Scientific reports 6(1):
1-13; Gunarso, P., M. E. Hartoyo, F. Agus and T. J. Killeen (2013). Oil palm and land use change in Indonesia,
Malaysia and Papua New Guinea. Reports from the Technical Panels of the 2nd greenhouse gas working Group of
the Roundtable on Sustainable Palm Oil (RSPO). the Netherlands, Tropenbos International: 29-63; Hooijer, A., S.
Page, J. Jauhiainen, W. A. Lee, X. X. Lu, A. Idris and G. Anshari (2012). "Subsidence and carbon loss in drained
tropical peatlands." Biogeosciences 9(3): 1053; Koh, L. P., J. Miettinen, S. C. Liew and J. Ghazoul (2011).
"Remotely sensed evidence of tropical peatland conversion to oil palm." Proc Natl Acad Sci U S A 108(12): 5127-
5132; Koh, L. P. and D. S. Wilcove (2008). "Is oil palm agriculture really destroying tropical biodiversity?"
Conservation letters 1(2): 60-64; Luskin, M. S., J. S. Brashares, K. Ickes, I.-F. Sun, C. Fletcher, S. Wright and M. D.
Potts (2017). "Cross-boundary subsidy cascades from oil palm degrade distant tropical forests." Nature
communications 8(1): 1-7; Miettinen, J., C. Shi and S. C. Liew (2016). "Land cover distribution in the peatlands of
Peninsular Malaysia, Sumatra and Borneo in 2015 with changes since 1990." Global Ecology and Conservation 6:
67-78; Miettinen, J., A. Hooijer, D. Tollenaar, S. Page, C. Malins, R. Vernimmen, C. Shi and S. C. Liew (2012).
"Historical analysis and projection of oil palm plantation expansion on peatland in Southeast Asia." ICCT White
Paper 17; Mukheijee, I. and B. K. Sovacool (2014). "Palm oil-based biofuels and sustainability in southeast Asia: A
review of Indonesia, Malaysia, and Thailand." Renewable and sustainable energy reviews 37: 1-12; Omar, W., N.
Aziz, A. T. Mohammed, M. H. Harun and A. K. Din (2010). "Mapping of oil palm cultivation on peatland in
Malaysia." MPOB Information Series; Vijay, V., S. L. Pimm, C. N. Jenkins and S. J. Smith (2016). "The impacts of
oil palm on recent deforestation and biodiversity loss." PLoS One 11(7): e0159668.
211
-------
4.3.1 Wetlands
There are several federal reports that describe the status and trends of U.S. wetlands,303
including the U.S. Fish and Wildlife Service (USFWS) Status and Trends of Wetlands in the
Conterminous United States,304 the USFWS and NOAA Status and Trends of Wetlands in the
Coastal Watersheds of the Coterminous United States,305 the USFWS Status and Trends of
Prairie Wetlands in the United States,306 EPA's National Wetland Condition Assessment307
(NWCA), and USDA's Natural Resources Inventory (NRI).308 The USGS NWALT (National
Water-Quality Assessment (NAWQA) Program's Wall-to-Wall Anthropogenic Land Use
Trends) series does not model changes in wetlands.309 Although these federal wetland reports are
a wealth of information on wetland status and trends in the U.S., many of them are unfortunately
not particularly useful in evaluating the impact of biofuels or the RFS program. The most recent
versions of the three USFWS reports only cover up to 2009, and, therefore, are of limited utility
given that EISA was enacted in 2007 and the RFS2 program was promulgated in 2010. The 2011
NCWA was the first in the series, thus, trends cannot be inferred from that report alone. The
second field sampling for NWCA was conducted in 2016 and may be used to infer trends once
the report is available.
The most pertinent federal program that monitors and reports the status and trends of U.S.
wetlands in the context of biofuels is the USDA NRI.310 Wetlands are not an independent land
cover class in the NRI, but are overlaid on other land cover types (e.g., wetlands on forested
lands made up 66,053,800 acres in 2007). The changes in wetland acres between 2007 and 2017
are shown in Table 4.3.1-1. There was an overall reduction by roughly 52,800 acres between
2007 and 2012, and a further reduction of 64,300 acres between 2012 and 2017. Over the full
2007 to 2017 timeframe, these changes represent a reduction of 0.11%. These reductions were
mostly from losses of wetlands on cropland and rangeland, which were partly offset by gains in
303 Summarized and listed here: https://www.epa.gov/wetlands/how-does-epa-keep-track-status-and-trends-
wetlands-iis.
304 Dahl, T.E. 2011. Status and trends of wetlands in the conterminous United States 2004 to 2009. U.S. Department
of the Interior; Fish and Wildlife Service, Washington, D.C. 108 pp.
305 T.E. Dahl and S.M. Stedman. 2013. Status and trends of wetlands in the coastal watersheds of the Conterminous
United States 2004 to 2009. U.S. Department of the Interior, Fish and Wildlife Service and National Oceanic and
Atmospheric Administration, National Marine Fisheries Service. (46 p.)
306 Dahl, T.E. 2014. Status and trends of prairie wetlands in the United States 1997 to 2009. U.S. Department of the
Interior; Fish and Wildlife Service, Ecological Services, Washington, D.C. (67 pages).
307 NATIONAL WETLAND CONDITION ASSESSMENT 2011: A Collaborative Survey of the Nation's
Wetlands. U.S. Environmental Protection Agency Office of Wetlands, Oceans and Watersheds Office of Research
and Development Washington, DC 20460. EPA-843-R-15-005. May 2016
308 U.S. Department of Agriculture. 2020. Summary Report: 2017 National Resources Inventory, Natural Resources
Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State University,
Ames, Iowa. https://www.nrcs.nsda.gov/sites/defanit/files/2022-10/2017NRISnmmare Finai.pdf.
309 Falcone JA (2015). U.S. conterminous wall-to-wall anthropogenic land use trends (NWALT), 1974-2012. U.S.
Geological Survey: 33 pp. Washington, DC.
310 See Table 7 - Changes in land use/cover between 2012 and 2017, U.S. Department of Agriculture. 2020.
Summary Report: 2017 National Resources Inventory, Natural Resources Conservation Service, Washington, DC,
and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.
https://www.nrcs.nsda.gOv/sites/defanit/files/2022-.l.0/2017NRISnm.mare Finai.pdf.
212
-------
developed and water areas.311 The report does not provide the information needed to determine
the portion of wetland acres lost in order to grow feedstocks for biofuels, nor does it attempt to
identify the portion of lost wetland acres attributable to the RFS program.
Table 4.3.1-1: Changes in palustrine312 and estuarine313 wetlands on different land
use/cover types between 2007, 2012, and 2017314
Acres (in thousands)
Wetlands on
2007
2012
2017
Change
(2017-2007)
Change
(%)
Cropland, pastureland,
& CRP land
17,623.5
17,552.5
17,426.4
-197.1
-1.12
Rangeland
7,969.2
7,913.0
7,876.8
-92.4
-1.16
Forest land
66,053.8
66,035.9
65,983.6
-70.2
-0.11
Other rural land
14,731.1
14,736.6
14,801.5
70.4
0.48
Developed land
1,411.0
1,450.9
1,486.5
75.5
5.35
Water areas
3,556.0
3,602.9
3,652.7
96.7
2.72
Total
111,344.6
111,291.8
111,227.5
-117.1
-0.11
There are several other regional studies examining changes in wetland area, including
several from the Prairie Pothole Region.315 In the only other national assessment to date, Wright
et al. (2017) found that within 50 miles of an ethanol biorefinery there was a 14,000-acre loss of
wetland between 2008 and 2012. While one might infer a causal connection between proximity
to an ethanol biorefinery and loss of wetlands (a question that was not investigated directly), this
study nevertheless does not demonstrate a connection to the RFS program specifically. As
discussed in Chapter 1, there are and have been numerous other drivers for ethanol use in the
U.S., most significantly the economic benefits of using ethanol in E10 blends. Additionally, a
significant portion of U.S. ethanol production is exported and therefore cannot be attributed to
the RFS program.
311 "Water areas" are defined in the USD A NRI as "[a] broad land cover/use category comprising water bodies and
streams that are permanent open water."
312 The NRI defines "palustrine wetlands" as "[w]etlands occurring in the Palustrine System, one of five systems in
the classification of wetlands and deepwater habitats (Cowardin et al. 1979). Palustrine wetlands include all nontidal
wetlands dominated by trees, shrubs, persistent emergent plants, or emergent mosses or lichens, as well as small,
shallow open water ponds or potholes. Palustrine wetlands are often called swamps, marshes, potholes, bogs, or
fens." NRI Glossary, https://www.nrcs.usda.gov/sites/default/files/2022-10/NRI gtossarv.pdf.
313 The NRI defines "estuarine wetlands" as "[w]etlands occurring in the Estuarine System, one of five systems in
the classification of wetlands and deepwater habitats (Cowardin et al. 1979). Estuarine wetlands are tidal wetlands
that are usually semienclosed by land but have open, partly obstructed or sporadic access to the open ocean, and in
which ocean water is at least occasionally diluted by freshwater runoff from the land. The most common example is
where a river flows into the ocean." NRI Glossary, https://www.nrcs.usda.gov/sites/default/files/2022-
10/NRI giossarv.pdf.
314 U.S. Department of Agriculture. 2020. Summary Report: 2017 National Resources Inventory, Natural Resources
Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State University,
Ames, Iowa. https://www.nrcs.usda.gov/sites/default/fiies/2022-10/2017NRISummarv Final.pdf.
315 Johnston, C. A. (2013). "Wetland losses due to row crop expansion in the Dakota Prairie Pothole Region."
Wetlands 33(1): 175-182. Johnston, C. A. (2014). "Agricultural expansion: land use shell game in the U.S. Northern
Plains." Landscape Ecology 29(1): 81-95: 10.1007/sl0980-013-9947-0.
213
-------
There are also many differences between Wright et al. (2017) and the NRI that make
direct comparison of these two studies not relevant. These differences stem from numerous
sources, including the geographic extent (the entire contiguous U.S. for the NRI versus only
areas in the contiguous U.S. within 100 miles of a biorefinery in Wright et al. (2017)), and
source data (fixed random points in the NRI versus satellite-derived data from the USDA's
Cropland Data Layer in Wright et al. (2017)). Reconciling these estimates is beyond the scope of
this rulemaking. Nonetheless, when we consider these two national assessments and the other
studies cited above overall, they demonstrate that agricultural extensification may affect
wetlands,316 but any losses are relatively small compared with the total amount of wetland.
Moreover, as stated above, where these studies were directed at the potential impacts of biofuels,
they considered the impact of increased biofuel production generally, not the incremental impact
in biofuel production attributable to the RFS program or the volumes established in this action.
As discussed in further detail in Chapter 2, much of the biofuel projected to be used in 2023-
2025 is expected to be used even in the absence of the RFS volume requirements. Any land use
change associated with biofuels that would be used in the absence of the RFS volume
requirements is therefore not attributable to this action.
In the most recent NRI, the USD A reported that there was a decrease of 24,300 acres in
the total wetland and deepwater habitat area, including palustrine and estuarine wetlands and
other aquatic habitats, between 2012 and 2017.317 The bulk of the wetland losses were in the
Prairie Pothole region, as reported elsewhere,318 with some very high rates (i.e., >15%, Wright et
al. 2017). The conversion reported by Wright, Larson et al. (2017) explicitly included only lands
that had not been in cropland for at least 20 years; although these areas may not represent
pristine habitats, they are expected to represent habitats that are in a relatively natural state.
The studies discussed above show that total wetland acres in the contiguous U.S. have
been decreasing since 2007. The volume increases for 2023-2025 compared to the No RFS
baseline that are described in Chapter 3 and the Biological Evaluation, due to biofuels produced
from agricultural feedstocks (especially corn and soybeans) would suggest the potential for an
associated increase in crop production. As such, they may be associated with increased pressure
to convert wetlands into cropland or otherwise impact wetlands. However, if we consider the
potential impacts relative to the current situation in 2022 (i.e., the 2022 baseline discussed in
Chapter 2.2) there would be much less potential impact. Additional information on land use
change from corn and soybean can be found in the May 19 BE. (In addition to corn ethanol, and
especially soy BBD, the main volume changes in this rule are from soy biodiesel and biogas
derived CNG/LNG which is not expected to have associated impacts on land-use change and
316 Agricultural extensification is the expansion of agricultural land onto previously uncultivated land. U.S. EPA
(2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Environmental Protection
Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June.
317 U.S. Department of Agriculture. 2020. Summary Report: 2017 National Resources Inventory, Natural Resources
Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State University,
Ames, Iowa, at Table 18. https://www.nrcs.usda.gov/sites/default/files/2022-10/2017NRISummarv Final.pdf. See
also USD A, 2017 National Resources Inventory. The National Wetlands table shows a total area of wetlands and
aquatic habitat on water areas and non-federal land as 160,755,900 acres in 2012 and 160,731,600 acres in 2017.
318 Johnston, C. A. (2013). "Wetland losses due to row crop expansion in the Dakota Prairie Pothole Region."
Wetlands 33(1): 175-182. Johnston, C. A. (2014). "Agricultural expansion: land use shell game in the U.S. Northern
Plains." Landscape Ecology 29(1): 81-95.
214
-------
therefore wetlands). Additionally, we used a probabilistic approach in the Biological Evaluation
to estimate potential overlap between cropland changes and species critical habitats or ranges.
The study began with the area of potential land use change and overlayed that with critical
habitat and species range data. Of the estimated acreage impacts about 8% was projected to be
"idle lands". According to the National Land Cover Dataset, idle lands include emergent
herbaceous wetlands in addition to pasture/hay. With this evaluation, little change would be
anticipated on wetlands as candidate volumes would be met from alternative means.
Figure 4.3.1-1: Relative conversion rates from wetland to cropland from 2008 to 2012
Wetland to crop (%)
i i m
0 0.8 2.4 5.9 16.5
0 200 400
I ML
Km
0
Rates are relativized by ty pe of ecosystem witlvin a 3.5-mile spatial grid (modified from 319). Stars denote the
location of biorefineries in the analy sis.
4.3.2 Ecosystems Other Than Wetlands
There are many ecosystems other than wetlands that may be affected by biofuel
production and use, including grasslands, forests, and aquatic habitats downstream of corn and
soybean production areas. Impacts on aquatic habitats, such as from runoff of fertilizer and
pesticides, as well as changes in hydrology from tilling, are discussed in Chapter 4.4 and the
Biological Evaluation.
In addition to wetlands, Wright et al. (2017) also reported on the losses of grasslands,
shrublands, and forests within 50 miles of a biorefmery in their study.320 Wright et al. (2017)
estimated much larger reductions of grassland (2 million acres), forests (60,000 acres), and
shrublands (52,000 acres), than in wetland reductions (estimated 14,000 acre reduction) (Figure
4.3.2-1). The bulk of the grassland conversions occurred in South Dakota (348,000 acres), Iowa
(297,000 acres), Kansas (256,000 acres), Missouri (239,000 acres), Nebraska (213,000 acres),
and North Dakota (176,000 acres).321
319 Wright, C. K., et al. (2017), "Recent grassland losses are concentrated around US ethanol refineries."
Environmental Research Letters 12(4).
320 Id.
321 Id.
215
-------
Figure 4.3.2-1: Relative conversion rates to cropland from either (a) grassland, (b) forest,
or (c) shrubland from 2008 to 2012
Fof*M lo crop (%)
0 0.4 t.2 2.7
Rates are relativized by type of ecosystem within a 3.5-mile spatial grid (modified from 322). Stars denote the
location of biorefineries, and the 100 mile radius from all biorefineries is included in (a) for reference (purple
outline).
Net grassland change 1%)
E
-80 -5.3 -1.5 2.2
>100
The 2 million acre reduction in grassland described in Wright et al. (2017) between 2008
and 2012 is comparable to the 1.475 million acre reduction in rangeland reported in the USDA
NRi between 2007 and 2012.323 The NRI defines rangeland as a land use/land cover that is more
lightly managed than pastureland,324 and, as such, is probably the NRI land use/land cover most
comparable to the grassland in Wright et al. (2017). The biggest reduction in rangeland was from
conversion to cropland (743,400 acres), followed by developed land (535,800 acres), and then
322 Id.
323 U.S. Department of Agriculture. 2020. Summary Report: 2017National Resources Inventory, Natural Resources
Conservation Service, Washington. DC, and Center for Survey Statistics and Methodology, Iowa State University,
Ames, Iowa. https://www.nrcs.usda.gov/sites/default/files/2022-10/2017NRISummarv Final.pdf.
324 The 2020 NRI defines rangeland as "A broad land cover/use category on which the climax or potential plant
cover is composed principally of native grasses, grass-like plants, forbs or shrubs suitable for grazing and browsing,
and introduced forage species that are managed like rangeland. This would include areas where introduced hardy
and persistent grasses, such as crested wheatgrass, are planted and such practices as deferred grazing, burning,
chaining, and rotational grazing are used, with little or no chemicals or fertilizer being applied. Grasslands,
savannas, many wetlands, some deserts, and tundra are considered to be rangeland. Certain communities of low
forbs and shrubs, such as mesquite, chaparral, mountain shrub, and pinyon-juniper, are also included as rangeland."
216
-------
conversion to other land uses by smaller amounts. The NRI does not parse out individual crops
within the cropland category, making it impossible to draw specific conclusions about the impact
of crop production for biofuels on grassland habitat. This reduction in rangeland acreage between
2007 and 2012 was reported to continue between 2012 and 2017, with an additional reduction of
over 2.4 million acres of rangeland, again with the largest conversion to cropland (754,600
acres).325
The Conservation Reserve Program (CRP) is especially relevant to the land use change
and impacts to ecosystems. CRP lands are often grassland habitat that are entered into contract
for 10-15 years, and provide a range of ecosystem services over that period, including carbon
sequestration, nutrient capture, and habitat for birds.326 CRP lands are formerly agricultural
lands, and, once they have left the CRP, could be used for the production of biofuel feedstocks.
However, they are often not used for production because the lands are often of lower quality, and
the guaranteed rental rate from admission to the CRP program is more attractive to farmers than
the uncertainty of growing crop on marginal lands.327 Despite the rental payment incentive to
farmers, enrollment in the CRP has been shrinking since 2007.328 This is due to specifications in
the Farm Bills, with a reduction from 36.8 million acres in 2007 to 21.9 million acres in 2020.329
The 2020 NRI reported a net reduction of CRP land by 8.7 million acres between 2007 and 2012,
mostly to cropland (66.5%) and pastureland (38%).330 These reductions continued from 2012 to
2017, with a reduction of 7.8 million acres between 2012 and 2017, again mostly to cropland
(63%) and pasture (37%). A detailed study from a 12-state area in the Midwest found that 30%
of the CRP land that left the program between 2010 and 2013 went into five principal crops (i.e.,
corn, soybean, winter wheat, spring wheat, and sorghum), with the majority of that to corn and
soybean.331 Reconciling these studies suggests that, of the land that leaves the CRP and goes into
the generic category of cropland in the NRI, at least half of that cropland is devoted to row crops.
The change in CRP enrollment is not uniform across the country (Figure 4.3.2-2), with much of
the reduction in the western and northern plains, the same areas experiencing losses of grassland
and increases in agriculture.
325 U.S. Department of Agriculture. 2020. Summary Report: 2017 National Resources Inventory, Natural Resources
Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State University,
Ames, Iowa. https://www.nrcs.nsda.gov/sites/defanit/files/2022-10/2017NRISnmmare Final.pdf.
326 USDA Farm Services Agency (FSA). 2016. The Conservation Reserve Program: 49th Signup Results,
https://www.fsa.usda.gov/Assets/USDA-FSA-Public/usdafiles/Conservation/PDF/SU49Book State finatl.pdf
327 Gray, B. J., & Gibson, J. W. (2013). "Actor-networks, farmer decisions, and identity." Culture, Agriculture, Food
and Environment, 35(2), 82el01. Brown, J. C., etal. (2014). "Ethanol plant location and intensification vs.
extensification of corn cropping in Kansas "Applied Geography 53: 141-148.
328 Data from the USDA Farm Service Agency (FSA), https://www.fsa.iisda.gov/programs-and~
services/conservation-programs/reports-and-statistics/conservation-reserve-pro gram-statistics/index.
329 USDA FSA, FY 2020 Annual Summary. Data from the USDA Farm Service Agency (FSA),
https://www.fsa.usda.gov/programs-and-services/conservation-programs/reports-and-statistics/conservation-reserve-
program-statistics/index.
330 U.S. Department of Agriculture. 2020. Summary Report: 2017 National Resources Inventory, Natural Resources
Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State University,
Ames, Iowa, at 3-45. https://www.nrcs.nsda.gOv/sites/de:fanit/files/2022-.l.0/2017NRISnm.mare Finat.pdf.
331 Morefield, P. E., et al. (2016). "Grasslands, wetlands, and agriculture: the fate of land expiring from the
Conservation Reserve Program in the Midwestern United States." Environmental Research Letters 11(9): 094005.
217
-------
Figure 4.3.2-2: Change in CRP enrollment between 2007 and 2016.332
Note: Data as of trie end of April 2D 16.
Prepared by FSA/EPAS/NRA
Reductions in forested areas to grow corn or soybeans does not appear to be occurring in
large amounts. As noted above, Wright et al. (2017) reported a net conversion of roughly 60,000
acres of forest wi thin 50 miles of biorefineries. The NRI reported an overall increase of
forestland between 2007 and 2012 (+672,400 acres) which continued between 2012 and 2017
(+1,099,700 acres). Most of the new forest land in both periods came from conversion of
pastureland, which offset smaller losses of forest land to predominantly developed lands.333
Thus, even though some forest land did convert to cropland according to the NRI,334 these
conversions appear small and to be offset by reforestation of pastureland.
The volume increases for the 2023-2025 years compared to the No RFS baseline
described in Chapter 2 due to biofuel production from agricultural feedstock (notably soybean oil
for renewable diesel) suggests the potential for an associated increase in crop production. As
renewable diesel growth continues, it is expected that demand for soybean crush will rise. Thus
the 2023-2025 volumes will have a potential to adversely impact grassland and other non-
wetland ecosystems. In the Biological Evaluation, land conversion for crop production is
analyzed in more detail.
332 Data from the USD A Farm Services Agency (ittps://www.fsa.usda.gov/programs-and-services/conservation-
programs/reports-and-statistics/conservation-reserve-pro gram-statistics/index).
333 The increase in forest land between 2007 and 2012 came mostly from addition of pastureland (+2.5 million
acres), which offset losses to developed land (-1.4 million acres). These trends continued between 2012 and 2017,
with an increase in forestland from pasture (+2.3 million acres) offsetting losses to developed land (-1.2 million
acres). There were many other smaller changes that occurred simultaneously.
334 The 2020 NRI reports that 292,200, and 265,600 acres of forest land converted to cropland between 2007-2012,
and 2012-2017, respectively.
218
-------
A process was created in the Biological Evaluation to determine how species and their
habitats would be affected by the RFS Set Rule. This process starts with the estimation of land
use change associated with an increase in production due to the rule. It then identifies potential
locations where affected increases in biofuel production could cause land use change. We
analyzed these first two steps of the process in two different ways.
Corn and canola potential land use changes were analyzed using a probabilistic
methodology as a land selection model was unable to be developed. More information on this
can be found in the May 19 BE. This method analyzed available land in the action area with a
greater likelihood of being affected by this rule. This included four different land cover classes,
three of which were non-wetland. Corn was analyzed using ArcGIS and R5. 500,000 acres were
randomly selected from the available land within the action area. This process was repeated
hundreds of times to estimate the probability of an acre converted to crop production for
renewable fuel use. A similar process was done with canola but with a constrained analysis to the
state of North Dakota as this state is shown to be the primary affected area for this crop.
Soy potential cropland expansion was evaluated separately by contracting firm ICF. They
developed a land selection model using weighted factors deemed important in determining where
additional soy acres might be planted. This provided an estimate of which land parcels would be
most likely to be converted to additional cropland. The example given in the Biological
Evaluation shows that land near soybean fields would be weighted higher than elsewhere.
4.3.3 Wildlife
There are many subsequent potential impacts to wildlife from these changes in wetlands
and other ecosystems, which were also summarized in the Biological Evaluation. The potential
impacts and their severity vary depending on such factors as crop type, geographic location, and
land management practices. The CRP, in particular, provides incentives for maintaining many of
these habitats, including practices that target pollinators (e.g., Conservation Practice (CP) 42, and
CP2), ducks (e.g., CP 37), and other wildlife (e.g., CP4B, 4D, 33).335 Here we focus on potential
impacts to terrestrial wildlife, including primarily birds and insects, which have been the most
studied to date. Impacts to aquatic wildlife are described in Chapter 4.4.2.3 and in the May 19
BE.
There are many bird species that use patches of grassland, wetland, pasture, and other
lightly managed areas as habitat within largely agricultural areas. Conversion of wetlands to row
crops is associated with reduced duck habitat and productivity of duck food sources, including
aquatic plants and invertebrates.336 However, studies of the effects of bioenergy feedstock
production suggest that grassland bird species of conservation concern are more likely to be
335 Listed here: https://www.fsa.usda.gov/programs-aiKl-services/conservation-programs/crp-practices-librarv/index.
336 Gleason, R.A., Euliss, N.H., Tangen, B.A., Laubhan, M.K., and Browne, B.A. (2011). "USDA conservation
program and practice effects on wetland ecosystem services in the Prairie Pothole Region." Ecological Applications
21: S65-S81.
219
-------
affected by increased corn production than are more common species of birds.337 Evidence
suggests that the direct effects of increasing cultivation of corn and soybean for biofuel
production are coming mostly from the conversion of grasslands to cropland, rather than other
habitat types (e.g., wetlands, forests, shrublands). Thus, it is likely that the wildlife species with
the largest potential risk are grassland species, including bird species and various insect species.
However, other types of land use change may also occur, with evidence from the NRI suggesting
roughly 50,000 acres of wetland converted to cropland between 2012 and 2017.
While the impacts of land use and management on wildlife have been studied, such as in
Tudge et al (2021), the impacts of the RFS program specifically have not.338 Evans et al. (2015)
conducted a detailed assessment of trends in the populations of 22 grassland bird species across
an 11-state area using the USGS Breeding Bird Survey.339 The 22 species examined were a
subset of the 28 identified by the USGS as grassland birds. Six species were excluded because
their breeding ranges were outside of the 11-state study area. Evans et al. (2015) found that
observations of six species were negatively associated with primary crop area, while
observations for five species were positively associated with primary crop area.340 All of the bird
species with negative associations were on the U.S. FWS list of species of conservation concern,
while none of the species exhibiting positive responses were on the list of conservation
concern.341 Although the above results using Ordinary Least Squares regression analysis were
statistically significant, associations were weaker when random or fixed effects were included.
When using random or fixed effects, only two species of conservation concern retained a
negative association with crop area (Bobolink [Dolichonyx oryzivorus] and Henslow's Sparrow
[Ammodramus henslowii]). Furthermore, when the marginal trends from primary crop increases
were compared with overall trends, the magnitudes of effect were modest. The effects from land
use change of primary crops led to a -0.20% to +0.15% effect, compared to the overall trends
which ranged from -2.74% to +10.66%, or a 10- to 100-fold larger overall effect.342
Potential harm to insects, especially insect pollinators, has also been of particular
concern. One study estimated that bees contributed an estimated $14.6 billion toward agricultural
production in 2009, or 11% of the nation's agricultural gross domestic product.343 Roughly 20%
337 Fletcher, R.J., Robertson, B.A., Evans, J., Doran, P.J., Alavalapati, J.R.R., and Schemske, D.W. (2011).
"Biodiversity conservation in the era of biofuels: risks and opportunities." Frontiers in Ecology and the
Environment 9(3): 161-168: 10.1890/090091. Blank PJ, Sample DW, Williams CL and Turner MG (2014). "Bird
Communities and Biomass Yields in Potential Bioenergy Grasslands." PLOS ONE 9(10): el09989:
10.1371/journal.pone.0109989.
338 Tudge, S.J., Purvis, A. & De Palma, A. "The impacts of biofuel crops on local biodiversity: a global synthesis."
Biodivers Conserv 30, 2863-2883 (2021). https://doi.org/10.1007/slQ531-021-02232-5.
339 Evans, S.G. and Potts, M.D. (2015). "Effect of agricultural commodity prices on species abundance of US
grassland birds." Environmental and Resource Economics, 62(3), pp.549-565.
340 Primary crops were defined as corn, soybeans, and wheat.
341 Evans, S.G. and Potts, M.D. (2015). "Effect of agricultural commodity prices on species abundance of US
grassland birds." Environmental and Resource Economics, 62(3), pp.549-565.
342 Id.
343 Lautenbach, S., Seppelt, R., Liebscher, J., Dormann, C.F. (2012). "Spatial and temporal trends of global
pollination benefit." PLoS One 7(4):e35954. Morse, R.A., Calderone, N.W. (2000). "The value of honey bees as
pollinators of U.S. crops in 2000." Bee Culture 128:1-15. Koh, I., Lonsdorf, E.V., Williams, N.M., Brittain, C.,
Isaacs, R., Gibbs, J., and Ricketts, T.H. (2016). "Modeling the status, trends, and impacts of wild bee abundance in
the United States." Proceedings of the National Academy of Sciences 113(1): 140-145: 10.1073/pnas.l517685113.
220
-------
of these pollination services are estimated from wild populations which depend on local habitat
for food and nesting sites.344 A 2016 modeling study suggests that wild bee populations
decreased by 23% across the U.S. between 2008 and 2013.345 The causes of these reductions are
complex, but include land use change, pesticides, and disease.346 Subsequent effects from
reductions in local bee populations are possible, including reductions in pollinator-dependent
crops grown in the area,347 as well as natural pollination services provided to wild habitat and
associated ecological effects.
In the most comprehensive study to date, Hellerstein et al. (2017) found that when
averaged across the United States, the forage suitability index for pollinators increased from
1982 to 2002 and declined slightly from 2002 to 2012—though in important honey bee regions
(such as Central North and South Dakota), the decline from 2002 to 2012 was more
pronounced.348 The Dakota's are the summer grounds for many managed honey bee colonies,
and thus the reduction in forage quality in these areas may have impacts. Although the largest
stressors to honey bee populations remains the varroa mites, rather than pesticides from nearby
crops, the presence of high quality forage nearby colonies is thought to improve the resilience
and health of colonies by supplementing feeding.
In a series of recent reviews, researchers concluded that there is evidence of adverse
impacts to pollinators due to neonicotinoid pesticide exposure.349 But also that the evidence is
mixed, and major gaps remain in our understanding of how pollinator colony-level (for social
bees) and population processes may dampen or amplify the lethal or sublethal effects. EPA's
preliminary assessment of the risk to bees from imidacloprid, clothianidin, and thiamethoxam
found on-field risk to be low for these pesticides applied to corn, which is the dominant use
pattern for this crop.350 For soybeans, risks were considered uncertain at the time and are
currently undergoing re-evaluation by EPA. Neonicotinoids, like all pesticides, are approved for
344 Losey, J.E., Vaughan, M. (2006). "The economic value of ecological services provided by insects." Bioscience
56(4):311-323.
345 Koh, I., Lonsdorf, E.V., Williams, N.M., Brittain, C., Isaacs, R., Gibbs, J., andRicketts, T.H. (2016). "Modeling
the status, trends, and impacts of wild bee abundance in the United States." Proceedings of the National Academy of
Sciences 113(1): 140-145: 10.1073/pnas.l517685113.
346 Goulson, D., Nicholls, E., Botias, C., Rotheray, E.L. (2015). "Bee declines driven by combined stress from
parasites, pesticides, and lack of flowers." Science 347(6229): 1255957.
347 For example, USDA NASS data for 2017 show that even though most apples (which are highly dependent on
pollinators) are grown in Washington (165,000 acres), smaller acreages are also grown in Michigan (33,000 acres),
Ohio (4,000 acres) and Illinois (1,700 acres). If 20% of these pollination services are provided by wild insects as
estimated by Losey et al. (2006), that could have effects on local apple production.
348 Hellerstein, Daniel, Claudia Hitaj, David Smith, and Amelie Davis. Land Use, Land Cover, and Pollinator
Health: A Review and Trend Analysis, ERR-232, U.S. Department of Agriculture, Economic Research Service, June
2017.
349 Godfray, H., Charles, J., Tjeerd Blacquiere, Linda M. Field, Rosemary S. Hails, Gillian Petrokofsky, Simon G.
Potts, Nigel E. Raine, Adam J. Vanbergen, and Angela R. McLean. (2014) "A restatement of the natural science
evidence base concerning neonicotinoid insecticides and insect pollinators." Proceedings of the Royal Society B:
Biological Sciences 281, no. 1786: 20140558.
350 EPA (2016). Preliminary Aquatic Risk Assessment to Support the Registration Review of Imidacloprid. U.S.
Environmental Protection Agency Office of Chemical Safety and Pollution Prevention, EPA-HQ-OPP-2008-0844-
1086: 219 pp. Washington, DC, December 22. EPA (2017). Preliminary Bee Risk Assessment to Support the
Registration Review of Clothianidin and Thiamethoxam. Office of Pesticide Programs, EPA-HQ-OPP-2011-0865-
0173: 414 pp. Washington, DC.
221
-------
use under specific conditions that are designed to protect ecosystems and human health.
Recently, EPA expanded its pesticide risk assessment process specifically for bees to quantify or
measure exposures and relate them to effects at the individual and colony level.351 Because of the
uncertainty surrounding the impacts of neonicotinoid use in soybean cultivation on pollinators, it
is difficult to state with any certainty that the RFS standards in this action will have an impact on
pollinators.
In the associated Biological Evaluation, wildlife habitats are evaluated for many species
listed within the action area for this rule. 704 unique species are located in the potential area for
crop land expansion. These species are associated with designated habitats by the US Fish and
Wildlife Service (FWS) and National Marine Fisheries Service (NMFS). EPA evaluated critical
habitat and range of these potential affected species for each crop. Additional information on
these estimates and impacts can be found in the May 19 BE.352'353
4.3.4 Potential Future Impacts of Annual Volume Requirements
The volume increases for 2023-2025 described in Chapter 3 due to biofuels produced
from agricultural feedstocks (especially corn and soybeans) would suggest the potential for an
associated increase in crop production. As such, they may be associated with increased pressure
to convert grasslands and wetlands into cropland, and, therefore, also increased pressure on
wildlife habitats. There exists substantial uncertainty in projecting changes in land use and
management associated with corn, soybeans, and other crops. Modeling and discussion on the
estimates for land use change are further discussed in Chapter 4.2 and in the May 19 BE.
In the Biological Evaluation, a more fulsome estimate of potential land conversion for
each of the main crops for renewable fuel was projected. Corn, soybean and canola expansion
was estimated using ArcGIS Pro as a tool with data from the United States Department of
Agriculture (USDA)'s Cropland Data Layer (CDL). Mapping was conducted with this data and
an additional buffer was included to conservatively account for indirect land use change effects.
The resulting area is shown in Figure 4.3.4-1.
351 U.S. EPA (2018), "How We Assess Risks to Pollinators." https://www.epa.gov/pottinator-protectlon/tiow-we-
assess-risks-poHinators.
352 Tudge, S.J., Purvis, A. & De Palma, A. "The impacts of biofuel crops on local biodiversity: a global synthesis."
Biodivers Conserv 30, 2863-2883 (2021). https://doi.org/10.1007/sl0531-021-02232-5.
353 U.S. EPA (2018), "How We Assess Risks to Pollinators." https://www.epa.gov/pottinator-protectlon/tiow-we-
assess-risks-pottinators.
222
-------
Figure 4.3.4-1: Geographical region where additional corn, soybean, and canola might be
grown to meet biofuel volumes as established by the RFS actions covering the years 2023-
Because the size of this analysis was so large, a number of assumptions had to be made
creating some uncertainty in the final results. However, even with uncertain results, EPA found
that overlaps occurred with 712 total species and their habitats. From this information, a further
evaluation was done on the species and potential impact from a cropland expansion.
4.4 Soil and Water Quality
Soil and water quality are addressed together here as they are in many ways intertwined,
with effects on soil often directly altering water quality (e.g., soil erosion leading to
sedimentation). Soil quality, also referred to as soil health, is the capacity of a soil to function,
including the ability to sustain plant growth.355 It can be affected by biofuel feedstock production
through changes in soil erosion, soil organic matter (SOM),356 and soil nutrients, among other
354 This region was identified by extracting corn, soybean, and canola croplands from the 2020 USD A Cropland
Data Layer, applying a 15-acre minimum mapping unit filter, and applying a five-mile buffer. ICF (2022,
September). "Potential Impacts of Expanded Soybean Production on Endangered Species and Critical Habitats -
Addendum." EPA Contract No. 68HERC21D0016. Work Assignment No. 68HERC22F0305. Modification P00002.
355 Soil quality is defined by the Soil Science Society of America's Ad Hoc Committee on Soil Quality (S-581) as:
"the capacity of a specific kind of soil to function, within natural or managed ecosystem boundaries, to sustain plant
and animal productivity, maintain or enhance water and air quality, and support human health and habitation."
Karlen et aL (1997), "Soil Quality: A Concept, Definition, and Framework for Evaluation." Soil Science Society of
America Journal, 61: 4-10: https://doi.org/10.2136/sssail997.03615995006100010Q01x. In this section, "soil
quality" is used as a general term, independent of area—it is used both to describe effects on single soil types and
cumulative effects across large areas and multiple soil types.
356 Soil organic matter is defined by Brady, N. and R. Weil (2000). Elements of the Nature and Properties of Soils.
Upper Saddle River, NJ, USA, Prentice-Hall, Inc. as "[t]he organic fraction of the soil that includes plant and animal
residues at various stages of decomposition, cells and tissues of soil organisms, and substances synthesized by the
soil population." Brady N and Weil R (2000). Elements of the Nature and Properties of Soils. Upper Saddle River,
2025.354
223
-------
characteristics. Soil erosion can negatively impact soil quality by disproportionately removing
the finest soil particles generally higher in organic matter, plant nutrients, and water-holding
capacity than the remaining soil. Soil organic matter is critical to soil quality because it provides
nutrients to plants, facilitates water retention in the soil, promotes soil structure, and reduces
erosion, while also sequestering carbon from the atmosphere. Soil nutrients (e.g., nitrogen,
phosphorus) are necessary for plant growth. Too little of these nutrients can reduce crop yields;
too much can negatively affect water quality via runoff or leaching.
Water quality is the condition of water to serve human or ecological needs.357 Crop-based
biofuel feedstock production can affect water quality through associated changes in nutrients,
dissolved oxygen, sediment, and chemical loadings.358 Nutrient releases from cropland into
nearby waterways can result in excessive algal growth (i.e., algal blooms), leading to low
dissolved oxygen levels (i.e., hypoxia) in some cases. Increased sediment and total dissolved
solids can make water unsuitable for consumption and irrigation, and also have negative impacts
on aquatic species. In addition, chemical releases or biofuel leaks and spills from above-ground,
underground, and transport tanks can be detrimental to water quality leading to ground, surface,
and drinking water contamination (see Chapter 4.4.2).359 Water quality impacts are evaluated as
either proximal (i.e., geographically close) or downstream, although effects can span both. We
discuss in the Biological Evaluation, sediment and chemical loadings under proximal effects, and
nutrients and hypoxia due to algal blooms in both coastal and non-coastal waters under
downstream effects. Additionally, water quality impacts due to increased crop production are
discussed in the May 19 BE, with data from the Soil and Water Assessment Tool (SWAT) used
in the draft Third Triennial Biofuels Report to Congress (RtC3).
4.4.1 The Role of Biofuels
Corn starch ethanol and soybean oil biodiesel account for most of the biofuel volumes
produced to date. As a result, the majority of soil and water quality impacts from biofuels thus
far have come from the production of corn and soybeans. There have also been notable quantities
of biogas from landfills that is cleaned and compressed to be used in compressed natural gas
(CNG) vehicles, and waste fats, oils, and greases (FOG) that is used to produce BBD. However,
they are not sourced from crop-based feedstocks and thus have only a tenuous connection to soil
and water quality. Additionally, products such as CNG/LNG from biogas are not anticipated to
NJ, USA, Prentice-Hall, Inc. The USDA NRCS similarly defines soil organic matter as "[t]he total organic matter in
the soil. It can be divided into three general pools: living biomass of microorganisms, fresh and partially
decomposed residues (the active fraction), and the well-decomposed and highly stable organic material. Surface
litter is generally not included as part of soil organic material." (USDA-NRCS 2021).
357 EPA (2003). National Management Measures to Control Nonpoint Pollution from Agriculture. U.S.
Environmental Protection Agency Office of Water, EPA-841-B-03-004. Washington, DC, July.
EPA (2011). Biofuels and the Environment: First Triennial Report to Congress. U.S. Environmental Protection
Agency, EPA/600/R-10/183F: 220 pp. Washington, DC, December.
358 The USDA NRCS Environmental Technical Note No. MT-1 (2011) defines these water quality parameters and
their significance.
359 This section focuses on the non-point source, water quality effects of feedstock production and spills. Any direct
point source discharges from biofuel production facilities are expected to be effectively controlled by existing
environmental statutes under the Clean Water Act (EPA 2011). Biofuels and the Environment: First Triennial
Report to Congress. Office of Research and Development, National Center for Environmental Assessment,
Washington, DC; EPA/600/R-10/183F).
224
-------
have any land affecting production changes. Canola oil is also used for BBD production, though
in considerably smaller quantities than soybean oil and FOG, and it is a crop-based feedstock
that could potentially impact soil and water quality. However, few studies focus on canola oil.
Since 2007, grasslands, including CRP grasslands, have been converted to corn and
soybeans, in a process termed extensification (see Chapter 4.4.2.1). Corn and soybeans have also
replaced other kinds of cropland. By contrast, the use of other crop-based feedstocks for biofuel
production has been much more limited. For example, use of corn stover has been attempted at a
couple of locations.360 To date, other feedstocks, such as perennial grasses, woody biomass, and
algae, have generally not yet materialized, with a few exceptions (e.g., algal biofuels for the U.S.
Navy), though there is a substantial amount of literature available on the impacts of perennial
grasses on soil and water quality.361 For that reason, we have included those feedstocks in this
analysis, though they are not widely used. Finally, outside the U.S., palm oil production for
biodiesel is an established industry in countries such an Indonesia, Malaysia, and Thailand, with
production occurring mainly for export, including to the U.S. As noted in Chapter 4.3, there is
strong evidence that expanded palm oil production would adversely affect soil and water quality,
in addition to carbon sequestration, outside of the U.S.
4.4.2 Impacts to Date
4.4.2.1 Soil and Proximal Water Quality Effects
Primarily, the magnitude of the impacts to soil and water quality depend upon the
feedstock grown and land use—i.e., the type of land used for growing the biofuel feedstock and
the management implemented on that land. For a given acre of cropland, planting corn or
soybeans onto grasslands (extensification) can be expected to have greater negative effects on
soil and water quality relative to the conversion of other existing cropland, such as wheat, to corn
or soybeans (intensification). Grassland-to-annual-crop conversion typically impacts soil quality
negatively because it increases erosion and the loss of soil nutrients and SOM, including soil
carbon loss to the atmosphere.362 In a meta-analysis, Qin et al. (2016) found that replacing
grasslands with corn decreased soil carbon by approximately 20% on average.363 The effects of
converting grasslands to soybeans are likely greater on erosion, SOM, and soil carbon than
converting to corn, since corn generally inputs more organic matter and carbon into the soil than
360 81 FR 89746 (December 12, 2016).
361 Ziolkowska, J.R., and Simon, L. (2014). "Recent developments and prospects for algae-based fuels in the US."
Renewable & Sustainable Energy Reviews 29: 847-853: 10.1016/j.rser.2013.09.021.
362 Gregorich, E.G., and Anderson, D. W. (1985). "Effects of cultivation and erosion on soils of four toposequences
in the Canadian prairies." Geoderma 36(3-4): 343-354: 10.1016/0016-7061(85)90012-6. Gelfand, I., Zenone, T.,
Jasrotia, P., Chen, J.Q., Hamilton, S.K., and Robertson, G.P. (2011). "Carbon debt of Conservation Reserve
Program (CRP) grasslands converted to bioenergy production." Proceedings of the National Academy of Sciences of
the United States of America 108(33): 13864-13869: 10.1073/pnas.l017277108. Qin, Z.C., Dunn, J.B., Kwon, H.Y.,
Mueller, S., and Wander, M.M. (2016). "Soil carbon sequestration and land use change associated with biofuel
production: empirical evidence." Global Change Biology Bioenergy 8(1): 66-80: 10.1111/gcbb.l2237. Lai, R.
(2003). "Soil erosion and the global carbon budget." Environment International 29(4): 437-450: 10.1016/s0160-
4120(02)00192-7.
363 Qin, Z.C., Dunn, J.B., Kwon, H.Y., Mueller, S., and Wander, M.M. (2016). "Soil carbon sequestration and land
use change associated with biofuel production: empirical evidence." Global Change Biology Bioenergy 8(1): 66-80:
10.1111/gcbb. 12237.
225
-------
soybeans, when both crops are managed using the same tillage practice (tillage practices are
discussed in greater detail later in this section).364 Increased erosion from conversion, in turn, can
negatively impact water quality through increased sediment and nutrient loadings to
waterways.365
Corn and soybeans additionally affect water quality through increased chemical usage,
some of which moves as runoff or leaching to surface waterways or groundwater. Table 4.4.2.1-1
summarizes the most recent USDA National Agricultural Statistics Service (NASS) Agricultural
Chemical Use Survey results for domestic corn and soybean acreage, as well as domestic wheat
acreage for comparison. In general, soybean acreage receives substantially less fertilizer than
corn, particularly nitrogen, because soybeans can attain nitrogen from the atmosphere via
symbiotic nitrogen fixation whereas corn cannot. Thus, as an example, multiplying 1.94 million
acres of extensification in the U.S. attributed to corn366 by the average nitrogen fertilizer rate
corn receives (149 lbs N/acre) yields an increase of approximately 289 million pounds of
additional nitrogen added per year. Likewise, from the most recent surveys by the USDA NASS,
97% of planted corn acres were treated with herbicides, 13% with insecticides, and 17% with
fungicides (Table 4.4.2.1-1). Atrazine was the top active ingredient among herbicides applied to
the planted corn acres, applied to 65% of planted acres, followed by mesotrione, applied to 42%
of planted acres.367 For planted soybean acres, 99% were treated with herbicides, 16% with
insecticides, and 15% with fungicides.368 Glyphosate isopropylamine salt and glyphosate
potassium salt were the top active ingredients among herbicides applied to planted soybean
acres.369 Due to the widespread nutrient and pesticide usage on corn and soybeans, it can be
inferred that runoff and/or leaching of these chemicals from corn and soybean acres are
contributing in part to proximal water quality impacts. For instance, in a modeling study of the
continental U.S., Garcia et al. (2017) estimated that increased corn production (up to 18 billion
gallons of corn ethanol) between 2002 and 2022 would increase nitrate groundwater
contamination (above or equal to 5 mg/L), particularly in areas with irrigated corn on sandy or
loamy soils.370
364 Johnson, J.M.-F., Allmaras, R.R., and Reicosky, D.C. (2006). "Estimating source carbon from crop residues,
roots and rhizodeposits using the national grain-yield database." Agronomy Journal 98:622-636.
365 Yasarer, L.M.W., Sinnathamby, S., and Sturm, B.S.M. (2016). "Impacts of biofuel-based land-use change on
water quality and sustainability inaKansas watershed." Agricultural Water Management 175: 4-14:
10.1016/j.agwat.2016.05.002.
366 Lark, T.J., Salmon, J.M., and Gibbs, H.K. (2015). "Cropland expansion outpaces agricultural andbiofuel policies
in the United States." Environmental Research Letters 10(4): 10.1088/1748-9326/10/4/044003.
367 USDA NASS (2019). 2018 Agricultural Chemical Use Survey: Corn.
https://www.nass.nsda.gov/Survevs/Guide to NASS Surveys/Chemical Use/2018 Peanuts Soybeans Corn/Chem
U sell i gfali gfat s Co rn 20.1.8. pdf.
368 USDA NASS (2019). 2018 Agricultural Chemical Use Survey: Soybeans.
https://www.nass.usda.gov/Surveys/Guide to NASS Surveys/ChemicalUse/2018 Peanuts Soybeans Corn/Chem
UseHighlights Soybeans 2018.pdf.
369 Id.
370 Garcia, V., Cooter, E., Crooks, J., Hinckley, B., Murphy, M., and Xing, X. (2017). "Examining the impacts of
increased corn production on groundwater quality using a coupled modeling system." Science of The Total
Environment 586: 16-24: https://doi.Org/10.1016/i.scitotenv.2017.02.009.
226
-------
Table 4.4.2.1-1: Summary of Chemical Use for Corn, Soybeans, and Wheat Acreage in the
U.S. based on 2018 and 2019 USDA
VASS Chemical Use Surveys371'372'373
Corn
Soybeans
Winter
wheat
Spring
wheat
Durum
wheat
Nitrogen Fertilizer Applied: % of Planted
Acres
98
29
88
97
98
Average Application Rate for Year for
Acres with Nitrogen Fertilizer Applied (lbs
N/acre)
149
17
73
102
83
Phosphate Fertilizer Applied: % of Planted
Acres
79
42
63
89
84
Average Application Rate for Year for
Acres with Phosphate Fertilizer Applied
(lbs P20s/acre)
69
55
31
39
29
Atrazine Applied: % of Planted Acres
65
Not
Reported
Not
Reported
Not
Reported
Not
Reported
Average Application Rate for Year for
Acres with Atrazine Applied (lbs/acre)
1.037
Not
Reported
Not
Reported
Not
Reported
Not
Reported
Glyphosate Potassium Salt: % of Planted
Acres
Not
Reported
28
Not
Reported
Not
Reported
Not
Reported
Average Application Rate for Year for
Acres with Glyphosate Potassium Salt
Applied (lbs/acre)374
Not
Reported
1.527
Not
Reported
Not
Reported
Not
Reported
Glyphosate Isopropylamine Salt: % of
Planted Acres
34
47
Not
Reported
Not
Reported
46
Average Application Rate for Year for
Acres with Glyphosate Isopropylamine Salt
Applied (lbs/acre)375
0.993
1.202
Not
Reported
Not
Reported
0.555
There are a couple of factors that can mitigate impacts on soil and water quality, at least
in part. First, the type of CRP lands, conservation lands, or other grasslands that are converted to
cropland can affect soil quality. In a modeling study, LeDuc et al. (2017) simulated that greater
erosion and loss of soil carbon and nitrogen occurs from converting low productivity, highly
sloped CRP grasslands compared to those with higher productivity soils and lower slopes.376 In
turn, higher erosion results in greater sedimentation and nutrient loading to waterways. Second,
the effects can also depend upon land management and production practices, like different tilling
371 USDA NASS (2019). 2018 Agricultural Chemical Use Survey: Corn.
https://www.nass.nsda.gov/Survevs/Guide to NASS Surveys/Chemical Use/2018 Peanuts Soybeans Corn/Chem
U seH i ghli ght s Co rn 20.1.8. pdf.
372 USDA NASS (2019). 2018 Agricultural Chemical Use Survey: Soybeans.
https://www.nass.nsda.gov/Survevs/Guide to NASS Surveys/Chemical Use/20.1.8 Peanuts Soybeans Corn/Chem
UseHighlights Soybeans 2018.pdf.
373 USDA NASS (2020). 2019 Agricultural Chemical Use Survey: Wheat.
https://www.nass.usda.gov/Survevs/Guide to NASS Surveys/ChemicalUse/20.1.9 Field Crops/chem-highlights-
wheat-2019.pdf.
374 This is expressed in acid equivalent.
375 This is expressed in acid equivalent.
376 LeDuc SD, Zhang XS, Clark CM and Izaurralde RC (2017). Cellulosic feedstock production on Conservation
Reserve Program land: potential yields and environmental effects. Global Change Biology Bioenergy 9(2): 460-468:
10.1111/gcbb. 12352.
227
-------
practices. About 60-70% of corn and soybeans are grown using conservation tillage.377'378
Conservation tillage, including no-till, reduces soil erosion and increases SOM content relative to
conventional tillage.379'380 (Cassel, Raczkowski et al. 1995, West and Post 2002)
The soil and water quality effects of converting to corn or soybeans from other crops,
such as wheat, are generally less than those of the conversion of grasslands.381 Zuber et al.
(2015) observed similar soil effects of no-till, continuous corn rotations, and corn-soybean-wheat
rotations on fine textured soils with high organic matter content.382 From this evidence, Zuber et
al. (2015) suggests a movement from wheat to corn may not materially affect soil quality,
provided a shift from no-till to conventional tillage does not occur concomitantly. In a meta-
analysis, Qin, Dunn et al. (2016) found that corn replacing other cropland (e.g., soybean, wheat)
increased soil organic carbon, whereas the opposite occurred when corn replaced grassland or
forest land.383 Notably, the percent increase in soil organic carbon of other cropland moving to
377 Claassen, R., Bowman, M., McFadden, J., Smith, D., & Wallander, S. (2018, September). Tillage Intensity and
Conservation Cropping in the United States. (EIB-197). U.S. Department of Agriculture, Economic Research
Service.
378 Conservation tillage is defined as any tillage practice leaving at least 30% of the soil surface covered by crop
residues; whereas conventional tillage leaves less than 15% of the ground covered by crop residues Lai, R. (1997).
"Residue management, conservation tillage and soil restoration for mitigating greenhouse effect by C02-
enrichment." Soil & Tillage Research 43(1-2): 81-107. No-till management, a subset of conservation tillage, disturbs
the soil marginally by cutting a narrow planting strip. Nationally, approximately 30% and 45% of the area planted to
corn and soybeans, respectively, are under no-till Wade, T., R. Claassen and S. Wallander (2015). Conservation-
practice adoption rates vary widely by crop and region, EIB-147, US Department of Agriculture Economic Research
Service. Since 2000, there has been a general trend toward greater percent residue remaining after planting for both
crops (USDA-ERS 2018 https://data.ers.nsda.gov/reports.aspx >83). Lai R (1997). Residue management,
conservation tillage and soil restoration for mitigating greenhouse effect by C02-enrichment. Soil & Tillage
Research 43(1-2): 81-107. USDA-NRCS (2010). Assessment of the effects of conservation practices on cultivated
cropland in the Upper Mississippi River Basin. NRCS. USD A. https://www.nrcs.iisda.gov/piibIications/ceap~crop~
2010-Upper-MRB-fuil.pdf.
Wade T, Claassen R and Wallander S (2015). Conservation-practice adoption rates vary widely by crop and region,
EIB-147, U.S. Department of Agriculture Economic Research Service.
379 Cassel DK, Raczkowski CW and Denton HP (1995). Tillage effects on corn production and soil physical
conditions. Soil Science Society of America Journal 59(5): 1436-1443. West TO and Post WM (2002). Soil organic
carbon sequestration rates by tillage and crop rotation: A global data analysis. Soil Science Society of America
Journal 66(6): 1930-1946.
380 Follett RF, Varvel GE, Kimble JM and Vogel KP (2009). No-Till Corn after Bromegrass: Effect on Soil Carbon
and Soil Aggregates. Agronomy Journal 101(2): 261-268: 10.2134/agronj2008.0107.
Gelfand I, Zenone T, Jasrotia P, Chen JQ, Hamilton SK and Robertson GP (2011). Carbon debt of Conservation
Reserve Program (CRP) grasslands converted to bioenergy production. Proceedings of the National Academy of
Sciences of the United States of America 108(33): 13864-13869: 10.1073/pnas.1017277108.
381 Zuber SM, Behnke GD, Nafziger ED and Villamil MB (2015). Crop Rotation and Tillage Effects on Soil
Physical and Chemical Properties in Illinois. Agronomy Journal 107(3): 971-978: 10.2134/agronjl4.0465.
Qin ZC, Dunn JB, Kwon HY, Mueller S and Wander MM (2016). Soil carbon sequestration and land use change
associated withbiofuel production: empirical evidence. Global Change Biology Bioenergy 8(1): 66-80:
10.1111/gcbb.l2237. YasarerLMW, Sinnathamby S and Sturm BSM (2016). Impacts of biofuel-based land-use
change on water quality and sustainability in a Kansas watershed. Agricultural Water Management 175: 4-14:
10.1016/j.agwat.2016.05.002.
382 Zuber SM, Behnke GD, Nafziger ED and Villamil MB (2015). Crop Rotation and Tillage Effects on Soil
Physical and Chemical Properties in Illinois. Agronomy Journal 107(3): 971-978: 10.2134/agronjl4.0465.
383Qin ZC, Dunn JB, Kwon HY, Mueller S and Wander MM (2016). Soil carbon sequestration and land use change
associated withbiofuel production: empirical evidence. Global Change Biology Bioenergy 8(1): 66-80:
10.1111/gcbb. 12237.
228
-------
corn was exceeded in magnitude by the percent decrease in soil organic carbon by the conversion
of grassland to corn. For water quality, an increase in corn at the expense of other crops is likely
to lead to greater nutrient loadings. In a global meta-analysis, Zhou and Butterbach-Bahl (2014)
found that average nitrate losses from leaching from corn (57.4 kg N/ha) exceeded those of
wheat (29 kg N/ha), suggesting that a replacement of wheat by corn would lead to higher nitrate
leaching to waterways.384 Between 2003 and 2010, Plourde et al. (2013) found that the practice
of rotating corn and soybeans decreased, while corn mono-cropping, or continuous corn,
increased.385 In a modeling study, Secchi et al. (2011) concluded that this intensification386 of
corn would likely lead to higher nitrogen and phosphorus loads in the Upper Mississippi River
Basin.387
Beyond corn and soy, the production of cellulosic feedstocks for biofuels, such as corn
stover and perennial grasses, may also affect soil and water quality. Partial stover removal can
increase corn yields in some locations, in part by reducing nitrogen uptake from the soil by
microorganisms and potentially by increasing soil temperatures in no-till systems.388 Corn stover
collection in areas with high rates of production also facilitates no-till land management
(compared to conventional tillage), which can reduce erosion, nutrient losses, and thereby
improve soil and water quality.389 Yet too much stover removal can increase soil erosion,
decrease SOM and soil nutrients, and ultimately decrease corn yields.390 Whether corn stover can
be harvested sustainably, and at what removal rate, depends on many site-specific factors,
including yields, topography, soil characteristics, climate, and tillage practices. In a study across
multiple locations in seven states, stover harvesting increased corn grain yields slightly, although
the authors cautioned against extrapolating these results to other sites and noted that there is a
need to conduct site-specific planning with soil testing.391 Additional research is needed to
understand effects on soil and water quality if soil conservation methods are employed while
harvesting corn stover.
384 Zhou and Butterbach-Bahl (2014). "Assessment of nitrate leaching loss on a yield-scaled basis from maize and
wheat cropping systems." Plant Soil 374: 977-991: 10.1007/slll04-013-1876-9.
385 Plourde, J.D., Pijanowski, B.C., and Pekin, B.K. (2013). "Evidence for increased monoculture cropping in the
Central United States." Agriculture, ecosystems & environment 165: 50-59.
386 Agricultural intensification is the increased production from the land without an increase in acreage. U.S. EPA
(2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Environmental Protection
Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June.
387 Secchi, S., Gassman, P.W., Jha, M., Kurkalova, L., and Kling, C.L. (2011). "Potential water quality changes due
to corn expansion in the Upper Mississippi River Basin." Ecological Applications 21(4): 1068-1084.
388 Coulter, J. A. and Naftiger, E.D. (2008). "Continuous Corn Response to Residue Management and Nitrogen
Fertilization." Agronomy Journal 100(6): 1774-1780: 10.2134/agronj2008.0170. Karlen, D.L., Birrell, S.J., Johnson,
J.M.F., Osborne, S.L., Schumacher, T.E., Varvel, G.E., Ferguson, R.B., Novak, J.M., Fredrick, J.R., Baker, J.M.,
Lamb, J.A., Adler, P.R., Roth, G.W., and Nafziger, E.D. (2014). "Multilocation Corn Stover Harvest Effects on
Crop Yields and Nutrient Removal." Bioenergy Research 7(2): 528-539: 10.1007/sl2155-014-9419-7.
389 Dale, V.H., Kline, K.L., Richard, T.L., Karlen, D.L., and Belden, W.W. (2017). "Bridging biofuel sustainability
indicators and ecosystem services through stakeholder engagement." Biomass and Bioenergy.
https://doi.Org/10.10.l.6/i.biombioe.2017.09.016.
390 EPA (2011). Biofuels and the Environment: First Triennial Report to Congress. U.S. Environmental Protection
Agency, EPA/600/R-10/183F: 220 pp. Washington, DC, December.
391 Karlen, D.L., Birrell, S.J., Johnson, J.M.F., Osborne, S.L., Schumacher, T.E., Varvel, G.E., Ferguson, R.B.,
Novak, J.M., Fredrick, J.R., Baker, J.M., Lamb, J.A., Adler, P.R., Roth, G.W., and Nafziger, E.D. (2014).
"Multilocation Corn Stover Harvest Effects on Crop Yields and Nutrient Removal." Bioenergy Research 7(2): 528-
539: 10.1007/sl2155-014-9419-7.
229
-------
Perennial grasses are a potential cellulosic feedstock that is not currently used at the
commercial scale. But, like other feedstocks, their impacts on soil and water quality would likely
depend upon the type of land use replaced and the management practices employed. Replacing
grasslands with intensively managed perennial feedstocks could have negative soil and water
quality effects, while replacing annual crops would likely lead to improvements.392 The scientific
literature continues to emphasize that perennial grasses or woody biomass grown on marginal
lands (e.g., abandoned agricultural land) can help restore soil quality.393 Notably, however, the
effects of these perennial feedstocks can depend upon the plant species grown and the type of
land converted.394 Additionally, the literature definitions of what constitutes marginal land and
estimates of its extent vary widely.395 For water quality, a modeling study found partially
replacing annual crops with Miscanthus and switchgrass—two perennial grasses—could reduce
inorganic nitrogen loadings by roughly 15% and 20%, respectively, in the Mississippi-
Atchafalaya River Basin.396 Alternative feedstock production (e.g., switchgrass) requires less
fertilizer than corn, thereby reducing nutrient runoff397 One recent modeling study for the state
of Iowa estimated that converting 12% and 37% of cropland to switchgrass would reduce
leached nitrate-nitrogen (NO3-N) by 18% and 38%, respectively, statewide.398 Another modeling
study estimated cropland conversion to switchgrass and stover harvest could greatly reduce
suspended sediment, total nitrogen, and phosphorus by 54 to 57%, 30 to 32%, and 7 to 17%,
respectively, in the South Fork Iowa River (SFIR) watershed if accompanied by best
management practices (e.g., riparian buffers and cover crops).399
4.4.2.2 Downstream Water Quality Effects
Increased corn and soybean cultivation may also affect downstream surface water and
aquatic systems, which can lead to aquatic life effects (see Chapter 4.4.2.3).400 Fertilizer runoff,
392 Ha, M., Z. Zhang, M. Wu (2017). Biomass production in the Lower Mississippi River Basin: Mitigating
associated nutrient and sediment discharge to the Gulf of Mexico. Science of the Total Environment, DOI:
10.1016/j.scitotenv.2018.03.184.
393 Blanco-Canqui H (2016). Growing Dedicated Energy Crops on Marginal Lands and Ecosystem Services. Soil
Science Society of America Journal 80(4): 845-858: 10.2136/sssaj2016.03.0080.
394 Robertson GP, Hamilton SK, Barham BL, Dale BE, Izaurralde RC, Jackson RD, Landis DA, Swinton SM,
ThelenKD and Tiedje JM (2017). Cellulosic biofuel contributions to a sustainable energy future: Choices and
outcomes. Science 356(6345): 10.1126/science.aal2324.
395 Emery I, Mueller S, Qin Z and Dunn JB (2016). Evaluating the potential of marginal land for cellulosic feedstock
production and carbon sequestration in the United States. Environmental Science & Technology 51: 733-741.
396 VanLoocke A, Twine TE, Kucharik CJ and Bernacchi CJ (2017). Assessing the potential to decrease the Gulf of
Mexico hypoxic zone with Midwest US perennial cellulosic feedstock production. GCB Bioenergy 9(5): 858-875:
10.1111/gcbb. 12385.
397 Parish ES, Hilliard MR, Baskaran LM, Dale VH, Griffiths NA, Mulholland PJ, Sorokine A, Thomas NA,
Downing ME and Middleton RS (2012). Multimetric spatial optimization of switchgrass plantings across a
watershed. Biofuels, Bioproducts and Biorefining 6(1): 58-72: 10.1002/bbb.342.
398 Brandes E (2018). Targeted subfield switchgrass integration could improve the farm economy, water quality, and
bioenergy feedstock production. GCB Bioenergy 10: 199-212, doi: 10.1111/gcbb.12481.
399 Ha, M. and M. Wu (2017). Land management strategies for improving water quality in biomass production under
changing climate. Environ. Res. Lett. 12 (3), 034015.
400 LaBeau MB, Robertson DM, Mayer AS, Pijanowski BC and Saad DA (2014). Effects of future urban and biofuel
crop expansions on the riverine export of phosphorus to the Laurentian Great Lakes. Ecological Modelling 277: 27-
230
-------
in addition to other factors (e.g., temperature and precipitation) and conservation practices,
influence downstream eutrophication,401 algal blooms, and hypoxia in fresh and coastal waters.
In freshwater systems, weather conditions and agricultural activity can increase nutrient runoff,
as observed in 2011 in western Lake Erie with dissolved reactive phosphorus.402 Total nitrogen
in lake water is also strongly correlated to the probability of detecting the cyanobacterium
Microcystis in lakes, in addition to the percentage of agricultural land cover within a given lake's
403
ecoregion.
In coastal systems, nutrient loadings affect hypoxic zone size, which is also a function of
climate, weather (e.g., storms), basin404 morphology, circulation patterns, water retention time,
freshwater inflows, stratification, and mixing, as seen in the Gulf of Mexico.405 Hypoxia is
caused by excess nutrients, particularly phosphorus and nitrogen, which enter the water stream
from agricultural runoff. As discussed in the Biological Evaluation, this runoff can lead to
growth of algae which leads to additional oxygen consumption in water. Lack of oxygenated
water can lead to fish die-offs and can harm other aquatic life. Conservation practices (e.g., filter
strips, cover crops, riparian buffers) can help mitigate downstream water quality effects due to
nutrients. Additionally, studies suggest that land conversion to perennial grasses such as
switchgrass and Miscanthus, even with manure application, could significantly reduce
phosphorus runoff into water bodies.406
37: https://doi.org/.1.0. .1.016/i.ecolmodel.20.1.4.0.1..0.1.6. Jarvie HP, Sharpley AN, Flaten D, Kleinman PJA, Jenkins A
and Simmons T (2015). The pivotal role of phosphorus in a resilient water-energy-food security nexus. Journal of
Environmental Quality 44(4): 1049-1062: 10.2134/jeq2015.01.0030.
401 EPA defines eutrophication as "[a] reduction in the amount of oxygen dissolved in water. The symptoms of
eutrophication include blooms of algae (both toxic and non-toxic), declines in the health of fish and shellfish, loss of
seagrass and coral reefs, and ecological changes in food webs." EPA, Vocabulary Catalog: Acid Rain Glossary,
https://sor.epa.gov/sor internet/registrv/termreg/searchandretrieve/gk>ssariesandkevwordlists/search.do?details=&gl
ossarvName=Aeid%20Rain%20Gtossarv.
402 Michalak AM, Anderson EJ, Beletsky D, Boland S, Bosch NS, Bridgeman TB, Chaffin JD, Cho K, Confesor R,
Daloglu I, DePinto JV, Evans MA, Fahnenstiel GL, He L, Ho JC, Jenkins L, Johengen TH, Kuo KC, LaPorte E, Liu
X, McWilliams MR, Moore MR, Posselt DJ, Richards RP, Scavia D, Steiner AL, Verhamme E, Wright DM and
Zagorski MA (2013). Record-setting algal bloom in Lake Erie caused by agricultural and meteorological trends
consistent with expected future conditions. Proceedings of the National Academy of Sciences 110(16): 6448-6452:
10.1073/pnas.l216006110.
403 Taranu ZE, Gregory-Eaves I, Steele RJ, Beaulieu M and Legendre P (2017). Predicting microcystin
concentrations in lakes and reservoirs at a continental scale: A new framework for modelling an important health
risk factor. Global Ecology and Biogeography.
404 EPA defines basin as "[a]n area of land that drains into a particular river, lake, bay or other body of water. Also
called a watershed." EPA, Vocabulary Catalog: Chesapeake Bay Glossary,
https://sor.epa.gov/sor internet/registrv/termreg/searchandretrieve/gk>ssariesandkevwordlists/search.do?details=&gl
ossarvName=Chesapeake%20Bav%20Gk>ssarv.
405 Dale VH, Kling C, Meyer JL, Sanders J, Stallworth H, Armitage T, Wangsness D, Bianchi TS, Blumberg A,
Boynton W, Conley DJ, Crumpton W, David MB, Gilbert D, Howarth RW, Lowrance R, Mankin K, Opaluch J,
Paerl H, Reckhow K, Sharpley AN, Simpson TW, Snyder C and Wright. D (2010). Hypoxia in the Northern Gulf of
Mexico. New York, Springer. Turner RE and Rabalais NN (2016). 2016 forecast: Summer hypoxic zone size
Northern Gulf of Mexico. Louisiana Universities Marine Consortium: 14 pp.
406 Muenich RL, Kalcic M and Scavia D (2016). Evaluating the Impact of Legacy P and Agricultural Conservation
Practices on Nutrient Loads from the Maumee River Watershed. Environmental Science & Technology 50(15):
8146-8154.
231
-------
4.4.2.3 Aquatic Life Effects
Impacts of biofuel crop production on aquatic ecosystems is understudied compared to
the impacts on terrestrial ecosystems. 407 However, it has been shown that increased corn and
soybean cultivation may affect downstream aquatic communities, chiefly through runoff or
leaching of nutrients and pesticides, though changes in land use and land cover also impact
aquatic ecosystems, particularly through conversion of wetlands that provide ecosystem services
like improving surface water flow, groundwater recharge, and sediment control.408 The May 19
BE has made steps to elaborate on the study of aquatic ecosystems and the affects they
experience from cropland expansion near waterways. Aquatic organisms interact within a food
web and contribute to many ecosystem services. The aquatic food web includes microorganisms
(bacteria, fungi, and algae), macroinvertebrates and macrophytes (submerged and floating
aquatic plants), and larger animals such as fish and marine mammals. When increased corn and
soybean cultivation changes the flow of water, nutrients, and other chemicals to downstream
systems, aquatic communities change in assemblage composition, typically in favor of organisms
that can tolerate nutrient and chemical pollution. Sensitive organisms that decrease in abundance
in response to these changes may be important food resources or key species in aquatic chemical
and biological processes, such as nutrient uptake or fish production.
Inputs of nutrients are a leading cause of impairment of freshwater and coastal
ecosystems, in part due to corn and soybean production.409 Corn production requires greater
application of nitrogen fertilizer compared to soy production because soy plants develop root
nodules with bacteria that can fix nitrogen from the atmosphere (Table 4.4.2.1-1). EPA's
National Aquatic Resource Surveys assess the quality of the nation's freshwater and coastal
ecosystems, including biological condition usually derived from the abundance of pollution-
tolerant and pollution-sensitive benthic macroinvertebrate taxa410 and fish.411 As of 2014, nearly
half (44%) of the nation's river- and stream-miles were in poor biological condition and about
30% were in good condition based on benthic macroinvertebrate indicators, and while 37% were
in poor condition and 26% were in good condition based on fish species indicators.412 The
leading problems contributing to poor biological condition were excess nutrients (especially
phosphorus), loss of shoreline vegetation, and excess sediments.413 For rivers and streams, sites
with a condition rating of poor because of excess nutrients were most prevalent in the mid-
407 U.S. EPA (2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Environmental
Protection Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June.
408 Id.
409 Id
410 Benthic macroinvertebrate taxa are small, bottom-dwelling, aquatic animals and the aquatic larval stages of
insects. EPA, National Aquatic Resource Survey: Indicators: Benthic Macroinvertebrates,
https://www.epa.gov/national~aaiiatic~resoiirce~siirvevs/indicators~benthic~
macroinvertebrates#:~:text=What%20are%20benthic%20macroinvertebrates%3F.snails%2C%20worms%2C%20an
d%20beetles.
411 USEPA (2020). National Rivers and Streams Assessment 2013-2014: A collaborative Survey. EPA 841-R-19-
001. Washington, DC. https://www.epa.gov/national~aaiiatic~resoiirce~siirvevs/nrsa.
412 Id.
413 Id.
232
-------
continent ecoregions414 of the nation compared to the eastern and western regions.415 Agriculture
is the dominant land use in the Mississippi River basin. As of 2012, 31% of the nation's lakes
were rated as having poor biological condition, over 35% had excess nutrient concentrations, and
nearly 10% of lakes had greater concentrations of cyanobacterial cells and the algal toxin
microcystin compared to 2007.416 For lakes, disturbance by nutrients varied by ecoregion (Figure
4.4.2.3-1). Northern Plains and Southern Appalachian ecoregions had a higher proportion (67-
80%) of within-ecoregion lakes) of sites classified as most disturbed by phosphorus pollution and
there was a statistically significant increase from 2007 to 2014 in the number of most disturbed
lakes in the Northern Appalachian ecoregion. For coastal and Great Lakes nearshore waters
(Figure 4.4.2.3-1), phosphorus was again a widespread problem (rating of poor in 21% of sites)
and biological condition was poorest along the Northeast coast (rating of poor in 27% of sites),
followed by the Great Lakes nearshore waters (rating of poor in 18% of sites).417 By 2014, the
greatest reduction in number of fish species occurred in portions of the Midwest and the Great
Lakes, where several watersheds have lost more than 20 species known to occur in those
locations prior to 1970.418
414 The National Rivers and Streams Assessment 2013-2014 defines "ecoregion" as "geographic areas that display
similar environmental characteristics, such as climate, vegetation, type of soil, and geology." USEPA (2020).
National Rivers and Streams Assessment 2013-2014: A collaborative Survey. EPA 841-R-19-001. Washington, DC.
https://www.epa.gov/national-aauatic-resoiirce-siiiveYs/nrsa.
415 USEPA (2020). National Rivers and Streams Assessment 2013-2014: A collaborative Survey. EPA 841-R-19-
001. Washington, DC. https://www.epa.gov/national-aaiiatic-resoiirce-siirveys/nrsa.
416 USEPA (2016). National Lakes Assessment 2012: A Collaborative Survey of Lakes in the United States. EPA
841-R-16-113. U.S. Environmental Protection Agency, Washington, DC.
https://www.epa.gov/sites/default/files/2016-12/documents/nla report dec 2016.pdf.
417 USEPA (2015). Office of Water and Office of Research and Development. National Coastal Condition
Assessment. EPA 841-R-15-006. U.S. Environmental Protection Agency, Washington, DC.
https://www.epa.gov/sites/default/files/2016-01/documents/ncca 20.1.0 report.pdf.
418 USEPA (2015). Report on the Environment. FishFaunal Intactness.
https://cfpiib.epa.gOv/roe/i indicator. e:fm?i=84.
233
-------
Figure 4.4.2.3-1: Locations of ecoregions and coastal areas defined by the USEPA's
National Aquatic Resource Surveys.419
\r-pr
CPL: Coastal Plains
NAP; Northern Appalachians NPL: Northern Plains
SAP: Southern Appalachians
SPL: Southern Plains
tTTT
1
I
TPL: Temperate Plains
UMW: Upper Midwest WMT: Western Mountains
XER: Xeric
West coast
Gulf coast
-------
Mexico hypoxic zone was the largest size measured since 1985, spanning 8,776 square miles.421
Hypoxic zones result in the death of fish and other organisms that need oxygen to live. Along
lake shorelines, blooms of filamentous green algae such as Cladophora harbor potentially
pathogenic bacteria and foul recreational beaches when the algae proliferate and decay.422 While
fertilizer use by current agricultural practices contribute to much of the nutrient loading that
stimulates algal responses in many waterbodies, the total nutrient budgets423 of some
waterbodies also include internal nutrient recycling of legacy inputs.424
In addition to nutrients, pesticides from corn and soybeans can also have deleterious
effects on aquatic life. Toxicological studies of glyphosate on fish have measured mainly
sublethal effects, such as DNA damage425 in organ tissues and altered muscle and brain
function.426 Some bacteria can use glyphosate for growth, enhancing microbial proliferation.427
There are cyanobacteria with natural tolerance to glyphosate428 and certain concentrations of
glyphosate can stimulate photosynthesis in a common bloom-forming taxon, Microcystis
aeruginosa,429 There is a notable link between glyphosate and phosphorus because more than
18% of glyphosate acid by mass is phosphorus. Glyphosate has chemical similarities with
phosphate ions (competing for the same sorption sites in soil), and glyphosate rapidly degrades
in water and releases phosphorus compounds easily used by organisms for growth. Glyphosate-
derived phosphorus has now reached levels in aquatic systems similar to phosphorus derived
from detergents prior to legislation banning these products, in part because of negative impacts
on aquatic life.430 In 2014, 58% of U.S. rivers and streams were given a rating of poor for the
421 Louisiana Universities Marine Consortium (2017). August 2, 2017 Summary. Shelfwide Cruise: July 24 - July
31- https://gulfhypoxia.net/research/shelfwide-cruise/?v=2017&p=press release. USNOAA (2019). Large 'dead
zone' measured in Gulf of Mexico, https://www.noaa.gov/media-release/large-dead-zone-measured-in-gulf-of-
mexico.
422 Ibsen, M., Fernando, D.M., Kumar, A. and Kirkwood, A.E. (2017). Prevalence of antibiotic resistance genes in
bacterial communities associated with Cladophora glomerata mats along the nearshore of Lake Ontario. Canadian
Journal of Microbiology 63(5): 439-449.
423 "A nutrient budget quantifies the amount of nutrients imported to and exported from a system []. The budget is
considered in balance if inputs and outputs are equal. Nutrient budgets can be calculated at any scale, such as a farm,
a county, a watershed, a state, or a country." Amy L. Shober, George Hochmuth, and Christine Wiese (2011). "An
Overview of nutrient budgets for use in nutrient management planning." University of Florida IFAS Extension
SL361. https://edis.ifas.ufl.edu/pdfFiles/SS/SS56200.pdf.
424 Chen, D., Shen, H., Hu, M., Wang, J., Zhang, Y. and Dahlgren, R.A. (2018). Legacy nutrient dynamics at the
watershed scale: principles, modeling, and implications. In: Advances in Agronomy. Ed: Donald L. Sparks. 149:
237-313. Academic Press. Cambridge, MA.
425 Guilherme, S., Gaivao, I., Santos, M.A. and Pacheco, M. (2012). DNA damage in fish (Anguilla anguilla)
exposed to a glyphosate-based herbicide-elucidation of organ-specificity and the role of oxidative stress. Mutation
Research/Genetic Toxicology and Environmental Mutagenesis, 743(1-2): 1-9.
426 Modesto, K.A. and Martinez, C.B. (2010). Roundup® causes oxidative stress in liver and inhibits
acetylcholinesterase in muscle and brain of the fish I'rochtlodus lineatus. Chemosphere, 78(3): 294-299.
427 Hove-Jensen B, Zechel DL, and Jochlmsen B. (2014). Utilization of glyphosate as phosphate source:
biochemistry and genetics of bacterial carbon-phosphorus lyase. Microbiol Mol Biol R 78: 176-97.
428 Harris TD and Smith VH. 2016. Do persistent organic pollutants stimulate cyanobacterial blooms? Inland Waters
6: 124-30.
429 Qiu, H., Geng, J., Ren, H., Xia, X., Wang, X. and Yu, Y. (2013). Physiological and biochemical responses of
Microcystis aeruginosa to glyphosate and its Roundup® formulation. Journal of hazardous materials, 248:172-176.
430 Hebert, M.P., Fugere, V. and Gonzalez, A. (2019). The overlooked impact of rising glyphosate use on
phosphorus loading in agricultural watersheds. Frontiers in Ecology and the Environment, doi: 10.1002/fee. 1985.
235
-------
phosphorus indicator of EPA's National Rivers and Streams Assessment.431 While both corn and
soybean production use glyphosate, corn production can also use atrazine (Table 4.4.2.1-1). In
2012, EPA detected atrazine in 30% of lakes, but concentrations rarely reached the EPA level of
concern for plants in freshwaters (<1% of lakes).432 In 2016, EPA concluded that in areas where
atrazine use is heaviest (mainly in the Temperate Plains ecoregion, Figure 4.4.2.3-1), there are
impacts on aquatic plants and potential chronic risk to fish, amphibians, and aquatic
invertebrates; there is a high probability of changes to aquatic plant assemblage structure,
function, and primary production at a 60-day average concentration of 3.4 ug L"1 and
reproductive effects to fish exposed for several weeks to 5 ug L"1 atrazine.433 When there are
changes to aquatic plant assemblage structure, function, or productivity, other parts of the food
web become at risk because there is reduced food and altered habitat for fish, invertebrates, and
birds. Additional information on the affects to aquatic life will become available as EPA
finalizes their evaluation of the affects the RFS program has on endangered species.
4.4.3 Comparison with Petroleum
Biofuel feedstocks are not the only input to energy production affecting soil and water
quality. For comparison, petroleum used to produce gasoline and diesel fuel also impacts soil and
water quality, but at different spatial and temporal scales than corn and soy. When comparing the
two, it is necessary to consider both the spatial extent of the effects (e.g., the acreage of soil or
volume of water impacted) and the time or effort to recover from any effects. While petroleum
production may have required less land than agriculture in the U.S. between 2007 and 2011,
when considering recovery time or effort, a recent study suggested the effects of petroleum
production can be longer lasting and harder to mitigate (e.g., brine or oil contamination in soil or
groundwater) than those of biofuel feedstocks on soil and water quality.434 A full comparison
between the effects of the two fuel types of energy feedstocks would need to consider both
factors (spatial extent and recovery time or effort), but such an assessment would be expansive
and could not be performed on the timeline of this rulemaking.
4.4.4 Water Quality and Underground Storage Tanks
Releases from underground storage tank (UST) systems can threaten human health and
the environment, contaminating both soil and groundwater. As of September 2021, more than
564,767 UST releases have been confirmed across the United States, averaging about 5,400 per
431 USEPA (2020). National Rivers and Streams Assessment 2013-2014: A collaborative Survey. EPA 841-R-19-
001. Washington, DC. https://www.epa.gov/national~aaiiatic~resoiirce~siirvevs/nrsa.
432 USEPA (2016). National Lakes Assessment 2012: A Collaborative Survey of Lakes in the United States. EPA
841-R-16-113. U.S. Environmental Protection Agency, Washington, DC.
433 USEPA (2016). Refined Ecological Risk Assessment for Atrazine. EPA-HQ-OPP-2013-0266. U.S.
Environmental Protection Agency, Washington, DC.
434 Parish ES, Kline KL, Dale VH, Efroymson RA, McBride AC, Johnson TL, Hilliard MR (2012). Comparing
Scales of Environmental Effects from Gasoline and Ethanol Production. Env Management: 10.1007/s00267-012-
9983-6.
236
-------
year between 2016 and 2021.435'436 One possible cause of an UST releasing fuel to the
environment is incompatibility of the UST system with the fuel being stored.
Ensuring UST systems are compatible with the substances they store is essential because
USTs contain many components made of different materials. In certain percentages petroleum-
biofuel blends are more incompatible with certain materials used in UST system construction
than petroleum-based fuel without biofuels. The whole UST system—including the tank, piping,
pipe dopes, containment sumps, pumping equipment, release detection equipment, spill
prevention equipment, and overfill prevention equipment—needs to be compatible with the fuel
stored to prevent releases to the environment. Compatibility with the substance stored is required
for all UST systems under EPA regulations, and storing certain biofuels requires additional
actions of UST owners and operators.
Equipment or components incompatible with the fuel stored could harden, soften, swell,
or shrink, and could lead to release of fuel to the environment. Examples of observed
incompatibility between fuels stored and UST materials include equipment or components such
as tanks, piping, or gaskets and seals on ancillary equipment that have become brittle, elongated,
thinner, or swollen when compared with their condition when initially installed.
Many of the tanks, piping and ancillary components being newly introduced into the
market today have now been designed to be compatible with up to 15% ethanol or up to 20%
biodiesel. However, most currently installed UST systems have at least some components that
may not be compatible with fuel blends containing more than 10% ethanol or more than 20%
biodiesel. EPA's 2015 UST regulation includes requirements for owners and operators of UST
systems storing any regulated substances containing greater than 10% ethanol or greater than
20% biodiesel, or any other substance identified by the implementing agency, to demonstrate
their UST system is compatible with those blends of biofuels prior to storing them.437 In 2021,
EPA proposed new regulations intended to strengthen the requirements for the underground
storage of fuels to ensure compatibility of new systems with high concentrations of biofuels.438
Nevertheless, insofar as blends of biofuel with gasoline or diesel are stored in USTs that are
either incompatible with those blends or have incompatible components, the increased
consumption of biofuels could increase leaks that affect water quality.
4.4.5 Potential Future Impacts of Final Volume Requirements
Future soil and water quality impacts associated with biofuel volumes will be driven, in
large part, by any associated land use/land cover changes. Directionally, increases in production
of biofuels made from crops would likely lead to an increase in land used for agriculture globally
and in the U.S. There are inherent uncertainties in estimating the amount and type of crop-based
feedstocks needed to fulfill the candidate volumes in this action, but an increase in cropland
acreage would generally be expected to lead to more negative soil and water quality impacts. As
435 USEPA (2021). Frequent questions about underground storage tanks, fattps://www.epa.gov/ust/frequent-
questions-abont-undergronnd-storage-tanks.
436 "UST Confirmed Releases National data 2016-2021," available in the docket.
437 40 CFR Part 280.
438 86 FR 5094 (January 19, 2021).
237
-------
outlined previously, the conversion of non-cropland, such as the extensification of corn and
soybeans onto grasslands, can be expected to have greater negative effects on soil and water
quality relative to the conversion of other existing cropland (intensification).
Although effects may generally be more negative, the cumulative magnitude of such an
increase in soil and water quality impacts is uncertain. The magnitude of effects depends on the
feedstocks planted, the types of land used, and management practices, all of which are not
directly determined by the RFS standards. Additional factors, such as vegetative barriers, and
advances in biotechnology and crop yields, can lessen future impacts. Expanded use of soil
amendments (e.g., biochar, manure) also could help counterbalance the removal of organic
matter and avoid or reduce the potential negative impacts of corn stover harvesting on soil
quality.439 In the case of biogas, there are numerous soil and water quality benefits compared to a
baseline of no manure or waste management. Dairy digesters, for example, are an essential piece
of proper manure management, as once the biogas has been captured, properly aerated manure
can be applied evenly to soil as a fertilizer.440'441 The additional soil and water quality modeling
that would be needed to assess the potential cumulative impacts of future land use changes for
the candidate volumes in this action would be expansive and could not be performed on the
timeline of this rulemaking.
The volume increases for 2023-2025 described in Chapter 3 due to biofuels produced
from agricultural feedstocks (especially corn and soybeans) in comparison to the No RFS
baseline would suggest the potential for an associated increase in crop production, which in turn
may impact soil and water quality. There is substantial uncertainty in projecting changes in land
use and management associated with corn, soybean, and other crops due to the other factors
driving biofuel demand however; in the May 19 BE, we further evaluate these based on models
and assessments. Furthermore, if we consider the potential impacts relative to the current
situation in 2022 (i.e., the 2022 baseline discussed in Chapter 2.2) there would be little impact, as
the overall volume increase for biodiesel and renewable diesel is much smaller and expected to
be met with expanded waste fats, oils, and greases supply.
4.5 Water Quantity and Availability
This section assesses the impact of the production and use of renewable fuels and their
primary feedstocks on the use and availability of water in the U.S. We first review the drivers of
impacts on water use and availability of freshwater resources, summarize impacts to date,
highlight more recent work focused on groundwater supplies. Finally, we discuss the potential
future effects of the candidate volumes as increases in feedstock production such as soybeans
may lead to water impacts.
439 Blanco-Canqui H (2013). Crop Residue Removal for Bioenergy Reduces Soil Carbon Pools: How Can We Offset
Carbon Losses? Bioenergy Research 6(1): 358-371: 10.1007/sl2155-012-9221-3.
440 2010 - US Climate Action Report: Fifth National Communication of the United States of America Under the
United Nations Framework Convention on Climate Change.
441 2011 Annual Report: ENERGY STAR and Other Climate Protection Partnerships.
238
-------
4.5.1 Drivers of Impacts on Water Use and Availability
Water quantity, in the context of renewable fuels, refers to the volume of water used in
the production of biomass feedstocks (i.e., irrigation of corn, soybeans or other crops) and the
conversion of those feedstocks to biofuel (i.e., water use in the biofuel production plant itself).
The irrigation of corn and soybeans used to produce biofuels is the predominant driver of water
quantity impact and is generally orders of magnitude greater than water use in the biofuel
production process.442 The water use for the full biofuels supply chain also has been quantified
as significantly higher than the water use for petroleum-based fuels, meaning biofuels are more
water intensive on a per gallon of fuel basis. Of concern are the impacts that this water use may
have on freshwater supplies and availability. Water intensive corn and soybean production
occurs on irrigated acres in states such as Nebraska and Kansas, in particular, the western parts
of those states. These states also overlap the High Plains Aquifer (HPA)443 "where groundwater
levels have declined at unsustainable rates."444
As noted above, the primary driver of impacts to water quantity is the water used for
irrigation of the biofuel feedstocks. To the extent that feedstock production expands into regions
where irrigation is required, the demand for water will increase, whether the expansion is a direct
consequence of production specifically for biofuel feedstocks or an indirect result of increased
production for all feedstock uses. Water demand for biofuel production processes can also drive
impacts on water use and availability. Although water demands of biofuel production facilities
may be much smaller at a national scale than the water demands of irrigated feedstock
production, biofuel facility water use may be locally consequential in areas that are already
experiencing stress on water availability.
4.5.2 Life Cycle Water Use of Biofuels
In the Draft Third Triennial Report to Congress on Biofuels, the water quantity impacts
of biofuels were assessed.445 Research investigating the water quantity impacts of biofuels
started shortly after the passage of the Energy Policy Act of 2005. Several highly cited and
visible articles compared the life cycle water use of biofuels relative to petroleum-based fuels on
the basis of "gallons of water per mile" or "gallons of water per gallon of fuel."446 These early
studies characterized this issue as biofuel's water intensity,447 embodied water,448 and water
442 Wu M, Zhang Z and Chiu Y-w (2014). Life-cycle Water Quantity and Water Quality Implications of Biofuels.
Current Sustainable/Renewable Energy Reports 1(1): 3-10.
443 The High Plains Aquifer is often referred to as the Ogallala Aquifer, which is the largest formation within the
High Plains Aquifer.
444 Smidt, S. J., Haacker, E. M., Kendall, A. D., Deines, J. M., Pei, L., Cotterman, K. A.,... & Hyndman, D. W.
(2016). Complex water management in modern agriculture: Trends in the water-energy-food nexus over the High
Plains Aquifer. Science of the Total Environment, 566, 988-1001.
445 U.S. EPA (2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Environmental
Protection Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June.
446 U.S. EPA (2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Environmental
Protection Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June.
447 King, C. W., & Webber, M. E. (2008). Water intensity of transportation. Environmental Science & Technology,
42(21), 7866.
448 Chiu, Y. W., Walseth, B., & Suh, S. W. (2009). Water embodied inbioethanol in the United States.
Environmental Science & Technology, 43(8), 2688-2692.
239
-------
footprint.449 Many studies of the water footprint further divide the consumptive water use into
two components: blue water (ground and surface water) and green water (rainwater).450 Most of
the focus of life cycle analyses (LCAs) has been on blue water or irrigation requirements for crop
production, as well as other freshwater use for biofuel conversion processes. When comparing
different transportation energy sources, Scown et al. (2011) found ethanol from corn-based
feedstocks to be one of the most significant users of freshwater.451 The same study calculated the
gallons of water consumed per mile of travel and found that the full life cycle water footprint of
ethanol produced from corn grain and stover (using average irrigation rates) would require
almost seven times as much surface water consumption as any other transportation power source
and an order of magnitude more groundwater consumption when compared to other
transportation energy sources.
4.5.2.1 Feedstock Production
Researchers have continued to refine the LCA-based water footprint of biofuels—with a
focus on feedstock production—for both current biofuels crops and future feedstocks. Because
more than 90% of corn is grown in rain-fed areas where corn production is non-irrigated, Wu et
al. (2014) suggested that, at the highly aggregated level, the "national water footprint of corn is
consistently low to modest."453 However, water quantity demands depend on the crops grown,
where they are grown, and how they are grown. In terms of differences among feedstocks,
Dominguez-Faus et al. (2009) calculated the irrigation water required for corn-based ethanol at
an average of approximately 600 liters (approximately 158.5 gallons) of water per liter of ethanol
produced (liter/liter).454 Much of the focus has been on corn ethanol, due to the higher volumes
of corn ethanol produced to date. However, in the same article, Dominguez-Faus et al. estimated
that irrigated soybean based biodiesel water requirement averaged nearly 1,300 liters of water
per liter of ethanol-equivalent biodiesel (based on energy equivalence).455 These values all
represent an upper end estimate of water demands, if fuels are made from irrigated crops.
However, where and how crops are grown also matter because irrigation rates for the
same crops can vary enormously based on where they are cultivated: from no irrigation in rain-
449 Dominguez-Faus, R., Powers, S. E., Burken, J. G., & Alvarez, P. J. (2009). The water footprint of biofuels: A
drink or drive issue? Environmental Science & Technology 43(9): 3005-3010: 10.1021/es802162x; Scown, C. D.,
Horvath, A., & McKone, T. E. (2011). Water footprint of US transportation fuels. Environmental Science &
Technology. 45(7), 2541-2553.
450 Another category is the grey water footprint, which is the volume of water required to assimilate pollutant loads,
such as excess nitrogen. Topics relating to grey water are covered in the water quality section. See Hoekstra, A. Y.,
& Mekonnen, M. M. (2012). The water footprint of humanity. Proceedings of the national academy of sciences,
109(9), 3232-3237.
451 Scown CD, Horvath A and McKone TE (2011). Water Footprint of U.S. Transportation Fuels. Environmental
Science & Technology 45(7): 2541-2553: 10.1021/esl02633h.
452 Scown CD, Horvath A and McKone TE (2011). Water Footprint of U.S. Transportation Fuels. Environmental
Science & Technology 45(7): 2541-2553: 10.1021/esl02633h.
453 Wu, M., Zhang, Z., & Chiu, Y. W. (2014). Life-cycle water quantity and water quality implications of biofuels.
Current Sustainable/Renewable Energy Reports, 1(1), 3-10.
454 Dominguez-Faus R, Powers SE, Burken JG and Alvarez PJ (2009). The Water Footprint of Biofuels: A Drink or
Drive Issue? Environmental Science & Technology 43(9): 3005-3010: 10.1021/es802162x.
455 Dominguez-Faus R, Powers SE, Burken JG and Alvarez PJ (2009). The Water Footprint of Biofuels: A Drink or
Drive Issue? Environmental Science & Technology 43(9): 3005-3010: 10.1021/es802162x.
240
-------
fed acres in the Midwest to high irrigation rates in more arid regions in the West. Dominguez-
Faus et al. (2013) calculated a range of irrigation water use for corn ethanol between 350 and
1400 gal/gal.456 That study estimated that if 20% of corn production was used to produce 12
billion gallons per year of ethanol in 2011 (irrigated at a weighted average of 800 gal/gal), that
would amount to 1.8 trillion gallons of irrigation water withdrawals per year. While not an
insignificant amount, it represents only 4.4% of all irrigation withdrawals.457 Other researchers
have similarly focused on the wide range of water intensity estimates between rain-fed and
irrigated counties and among a variety of crops (see Figure 4.5.2.1-1). Gerbens-Leenes et al.
(2012) estimated Nebraska's blue water (irrigation water) footprint at three times higher than the
U.S. weighted average blue water footprint.458 Many other corn producing states have much
smaller irrigation demands relative to Nebraska. Yet, it should be noted that, after Iowa,
Nebraska is the second largest producer of corn-based ethanol in the U.S., with 25 active ethanol
facilities, many concentrated in southern Nebraska.459 Additionally, the blue water footprint in
areas that have already stressed water sources, like the HP A, could experience more severe water
quantity impacts. A report by the National Academy of Sciences (NAS 2011) highlighted the
groundwater depletion in the HP A, noting that Nebraska is "among the states with the largest
water withdrawals for irrigation, and its usage has continued to increase in recent years, largely
driven by the need to irrigate corn for ethanol."460 This suggests that the majority of groundwater
consumption would come from areas like Nebraska that are already impacted by over-pumping
due to their high blue water footprint for corn production (Gerbens-Leenes et al. 2012).461
456 Dominguez-Faus, R., Folberth, C., Liu, J., Jaffe, A. M., & Alvarez, P. J. (2013). Climate change would increase
the water intensity of irrigated corn ethanol. Environmental science & technology, 47(11), 6030-6037.
457 Dominguez-Faus R, Folberth C, Liu J, Jaffe AM and Alvarez PJJ (2013). Climate Change Would Increase the
Water Intensity of Irrigated Corn Ethanol. Environmental Science & Technology 47(11): 6030-6037:
10.1021/es400435n.
458 Gerbens-Leenes, W., & Hoekstra, A. Y. (2012). The water footprint of sweeteners and bio-ethanol. Environment
international, 40, 202-211.
459 EIA (2018). "Six states account for more than 70% of U.S. fuel ethanol production."
https://www.eia.gov/todayinenergy/detail.php?id=36892. See also EIA. (2017, February 16). "Nebraska State
Profile and Energy Estimates: Profile Analysis." https://www.eia.gov/state/anatvsis.pfap?sid=NE.
460 NAS (2011). Renewable Fuel Standard: Potential Economic and Environmental Effects of U.S. Biofuel Policy.
National Academy of Sciences. Washington, DC.
461 Gerbens-Leenes, W., & Hoekstra, A. Y. (2012). The water footprint of sweeteners and bio-ethanol. Environment
international, 40, 202-211.
241
-------
Figure 4.5.2.1-1: Estimate of the blue/irrigation, green/rainwater and grey/pollution water
footprint associated with corn grain, stover, wheat straw and soybean during the crop
growing phase.
7,000
« Regional range
* County max
-National average
Blue Water Green Water Grey Water
The national production-weighted average is represented by the horizontal bar, while the regional ranges (this
includes all USDA regions such as the Corn Belt, Southern Plains, etc.) are represented by the shaded bars. County-
level variation in feedstock water footprints, shown in dashed lines, are driven by differences in irrigation and
evapotranspiration (ET). The circles show both "County max" as well as "County inin." [Source: Cliiu and Wu
(2012)].
4.5.2.2 Biofuel Processing
Studies of water use at biofuels conversion facilities have generally quantified water
consumption as gallons of water per gallon of biofuel produced, mostly concentrating on ethanol,
especially dry mill facilities.462 Process level engineering studies and surveys of ethanol facilities
have shown declines in water requirements from 5.8 gallons of water per gallon of ethanol
(gal/gal) in 1998 to 2.7 gal/gal in 2012.463 These values are typical of a dry mill facility. Wet mill
facilities require closer to 4 gallons per gallon of ethanol.464 Some reports also point to
reductions in the water intensity of ethanol facilities through more efficient water use and
recovery, and reuse of wastewater after treatment for processes such as fermentation or possibly
462 U.S. EPA (2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Enviromnental
Protection Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June.
463 Mueller S (2010). 2008 National dry mill corn ethanol survey. Biotechnology Letters 32(9): 1261-1264:
10.1007/sl0529-010-0296-7. and Wu Y and Liu S (2012). Impacts of biofuels production alternatives on water
quantity and quality in the Iowa River Basin. Biomass and Bioenergy 36: 182-191: 10.1016/j.biombioe.2011.10.030.
See also Wu Y and Liu S (2012b). Impacts of biofuels production alternatives on water quantity and quality in the
Iowa River Basin. Biomass and Bioenergy 36: 182-191: 10.1016/j.biombioe.2011.10.030.
464 Grubert, E. A., & Sanders, K. T. (2018). Water use in the US energy system: A national assessment and unit
process inventory of water consumption and withdrawals. Enviromnental science & technology.
242
-------
cooling towers.465 Some facilities have set goals to both reduce water use and minimize
discharges.
Biodiesel conversion from oil crops such as soybeans requires water use for multiple
stages of the process (crop to oil, and oil to biodiesel). The soybean processing (crop to oil) stage
involves crushing, oil extraction and crude soybean oil refining (degumming). Water
consumption includes make-up water for cooling towers and other processes. In the biodiesel
production stage (oil to biodiesel), following the crushing and oil extraction steps, water is used
to remove residual glycerol, a by-product of the transesterification process, and other impurities,
while some water is also for additional make-up water for cooling towers.466 Tu et al. (2016)
estimated that the water footprint of soybean based biodiesel to be under 1 gal/gal biodiesel (0.17
for crop to oil and 0.31 for oil to biodiesel).467
Renewable diesel is chemically very similar to petroleum-based diesel despite being
made from renewable wastes such as fats and vegetable oils. This means that it is processed in
the same manner that petroleum diesel which is hydrotreating. With this knowledge it can be
assumed that the same about of water used for processing petroleum diesel is used to process
renewable diesel.468 As renewable diesel is a plant-based feedstock, its additional water usage
would be from the irrigation process to grow the plants used to create the oil and not from the
process itself.
There are no recently published surveys of water consumption representing all current
biofuel and renewable fuel facilities, and no comprehensive data on the type of water sources
utilized (e.g., groundwater, surface freshwater, public supply, etc.). Grubert and Sanders estimate
that the majority of the water used is freshwater. There is also some evidence that groundwater
from aquifers is being extracted for use in ethanol production in states such as Iowa and
Nebraska,469 and likely a source of water for facilities along the HPA (see Figure 4.5.3-3).
4.5.2.3 Summary and Comparison to Petroleum
Improvements in irrigation have brought down the upper range of water use, with recent
estimates of irrigation for corn production ranging from 9.7 gal/gal ethanol for USDA Region 5
(Iowa, Indiana, Illinois, Ohio and Missouri) to 220.2 gal/gal ethanol in Region 7 (North Dakota,
465 Schill, S. R. (2017) Water: Lifeblood of the Process. Ethanol Producer Magazine. January 24, 2017.
http://www.ethanolprodncer.eom/artlcles/.l.4049/water-li:feblood-of-the-process. See also Jessen, H. (2012) Dropping
Water Use. Ethanol Producer Magazine. June 12, 2012. http://www.ethanolprodncer.com/artlcles/8860/dropping-
water-iise.
466 Tu, Q., Lu, M., Yang, Y. J., and D. Scott (2016) Water consumption estimates of the biodiesel process in the US.
Clean Technologies and Environmental Policy. 18(2): 507-516.
467 Tu, Q., Lu, M., Yang, Y. J., and D. Scott (2016) Water consumption estimates of the biodiesel process in the US.
Clean Technologies and Environmental Policy. 18(2): 507-516.
468 Sun, Pinping; Estimation of U.S. refinery water consumption and allocation to refinery products; Fuel; Volume
221; June 18, 2018.
469 Schilling, K. E., Jacobson, P. J., Libra, R. D., Gannon, J. M., Langel, R. J., & Peate, D. W. (2017). Estimating
groundwater age in the Cambrian-Ordovician aquifer in Iowa: implications for biofuel production and other water
uses. Environmental Earth Sciences, 76(1), 2. See also Gerbens-Leenes, W., & Hoekstra, A. Y. (2012). The water
footprint of sweeteners and bio-ethanol. Environment international, 40, 202-211.
243
-------
South Dakota, Nebraska and Kansas) for groundwater.470 The conversion of corn to ethanol
requires 2-10 gal/gal for processing, with most dry mill plants requiring roughly 3 gal/gal. When
averaging production over all regions, and accounting for co-products of ethanol production
(such as distillers dried grain and solubles), the range for full life cycle consumptive water use
for U.S. corn ethanol is 8.7-160.0 gal/gal ethanol based on the updated analysis by Wu et al
(2018). By comparison, the most recent estimates of the net consumptive water use over the
petroleum-based fuel life cycle would be in the range of 1.4-8.6 gallons of water per gallon of
gasoline based on U.S. conventional crude, and diesel fuel would likely be a little less than
gasoline's consumption.471 472 The Wu et al (2018) analysis does not include biodiesel. The most
recent estimate for the full LCA water consumption for biodiesel provides a range of values for
each state: Missouri 21-79 gal water/gal biodiesel, Kansas and Oklahoma 80-150 gal/gal,
Nebraska and Texas 150-300 gal/gal.473 The water consumption of a biodiesel plant is well under
1 gallon of water consumed for each gallon of biodiesel produced, therefore, virtually all the
water associated with the lifecycle production of biodiesel made from vegetable oils is due to the
growing and processing of vegetable oil feedstocks.474 Assuming that renewable diesel fuel
production plants consume the same amount of water as a distillate hydrotreater, then renewable
diesel fuel production plants likely consume about 3 gallons of water for each gallon of
renewable diesel produced.475 As with biodiesel, we expect that water consumption associated
with renewable diesel made from vegetable oils is primarily associated with the production of the
underlying vegetable oil feedstocks. Since biodiesel and renewable diesel made from FOG does
not require crop-based inputs, we expect that the water usage for these biofuels is significantly
lower.
In summary, while values will vary across states and counties, ethanol, and biodiesel and
renewable diesel made from vegetable oils are substantially more water intensive than the
petroleum fuels they would displace. Of these fuels, soy biodiesel and renewable diesel is the
volume expected to expand the most as a result of the final standards. However, the increase is
expected to derive mostly from increases in soybean crushing in the U.S., not from increase
plantings of soybeans.
4.5.3 Impacts to Date
Because the majority of the growth in biofuels production has come from corn- and soy-
based biofuels, the water consumption impacts to date would have come from additional water
use for corn and soybean acreage. To our knowledge, there have been no comprehensive studies
of the changes in irrigated acres, rates of irrigation, or changes in surface and groundwater
470 Wu, M., & Xu, H. (2018,). Consumptive Water Use in the Production of Ethanol and Petroleum Gasoline—2018
Update (No. ANL/ESD/09-1 Rev. 2). Argonne National Lab.(ANL), Argonne, IL (United States).
https://piiblications.anl.gov/anlpubs/20H 343.pdf.
471 Id.
472 Sun, Pinping; Estimation of U.S. refinery water consumption and allocation to refinery products; Fuel; Volume
221; June 18, 2018.
473 Tu, Q., Lu, M., Yang, Y. J., and D. Scott (2016). Water consumption estimates of the biodiesel process in the US.
Clean Technologies and Environmental Policy. 18(2): 507-516.
474 Haas, M.J, A process model to estimate biodiesel production costs, Bioresource Technology 97 (2006) 671-678.
475 Sun, Pinping; Estimation of U.S. refinery water consumption and allocation to refinery products; Fuel; Volume
221; June 18, 2018.
244
-------
supplies attributed specifically to the increased production of corn grain-based ethanol and
soybean-based biodiesel. There are, however, studies that can give some indication of how
changes in production of these biofuels may have affected water demand and availability. The
Second Triennial Report to Congress on Biofuels highlights analyses in Lark et al. (2015) and
Wright et al. (2017) that show changes in land use, including cropland expansion in the western
Dakotas and Kansas, related to biofuels.476 These are areas unlikely to have sufficient
precipitation to support corn or soybean cultivation.477 While difficult to attribute how much
additional water use might be required as a result of the candidate volumes in this rule, there are
several lines of evidence that suggest increased production of corn-based ethanol and soybean-
based biodiesel will increase water demands and, potentially, affect limited water supplies.
The USDA Irrigation and Water Management Surveys (formerly the Farm and Ranch
Irrigation Survey or FRIS), a supplement to the Census of Agriculture completed every five
years, provide a general indication of the changes in water demands between 2013 and 2018.478
From 2013 to 2018, there was an increase in total irrigated acres of nearly 0.6 million acres in the
U.S.479 Over the same period, irrigated acres of corn for grain and seed decreased from 13.3
million acres to 11.6 million acres harvested, along with a lower irrigation rate of 0.9 acre-feet
applied in 2018 compared to 1.1 acre-feet applied in 2013.480 Over the same time period,
irrigated acres of soybeans increased from 7.4 to 8.2 million acres harvested, while average acre-
feet applied declined from 0.9 to 0.6 per acre.481 Figure 4.5.3-1 shows acres of irrigated land in
2012, the most recent year of data for which this figure is available.
476 U.S. EPA (2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Environmental
Protection Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June.
477 Lark TJ, Salmon JM, and Gibbs HK (2015). Cropland expansion outpaces agricultural and biofuel policies in the
United States. Environmental Research Letters 10(4): 10.1088/1748-9326/10/4/044003. and Wright, C. K., et al.
(2017). "Recent grassland losses are concentrated around US ethanol refineries." Environmental Research Letters
12(4).
478 USDA NASS (2018). 2018 Irrigation and Water Management Survey.
https://www.nass.nsda.gov/Piiblications/AgCensns/2Q17/Online Resources/Farm and Ranch Irrigation Survev/fri
s.pdf.
479 Id.
480 Id.
481 Id.
245
-------
Figure 4.5.3-1: Acres of irrigated land in 2012
acres
Source: USDA, National Agricultural Statistics Service, Map Atlases for the
2012 Census of Agriculture.
Based on the USDA Farm and Ranch Irrigation Survey. Source: https://www.ers.usda.gov/topics/farm-practices-
nianageinent/irrigation-water-use/background.aspx
Figure 4.5.3-2 shows corn and soybean areas and share of irrigated acres. Irrigated corn
grain/seed acres are heavily concentrated in Nebraska (4.5 million acres) followed by Kansas
(1.3 million acres). This is a decrease of 0.9 and 0.2 million acres respectively from 2012 to
2018. Irrigated soybean acres are also found in Nebraska, Kansas, particularly the more western
part of those states. Overall soybean production is generally more concentrated (as a share of
total harvested cropland) in rainfed areas, whereas corn production reaches further west. There is
also a high percentage of soybean acres in Arkansas and Mississippi, with a large share of those
soybean acres being irrigated. The top rows of Figure 4.5.3-2 show the distribution of corn and
soybean acres, as a share of total cropland acres, while the bottom rows of Figure 4.5.3-2 show
the percent of irrigated corn and soybean acres relative to total acres for those crops (measures as
harvested acres).
246
-------
Figure 4.5.3-2: Percent of irrigated corn and soybean acres relative to total acres,
respectively
t-cr.zoi**« « for Besit* HsrmWtfAaas Fewem ol Soytwans for 8mw, Haivsawl Acw» 2017
Top left: Acres of Corn Harvested as a Percent of Harvested Cropland Acreage.
Top right: Acres of Soybean Harvested as a Percent of Harvested Cropland Acreage.
Bottom left: Irrigated Corn as a Percent of Total Corn (Harvested Acres).
Bottom right: Irrigated Soybeans as a Percent of Total Soybeans (Harvested Acres).
Source: USDA Agricultural Census Web Maps.
https://www.nass.usda.gov/Publications/AgCensus/2012/Qnline Resources/Ag Census Web Maps/index.php.
Higher irrigation demands may coincide with areas of already-stressed surface and
groundwater resources, such as the HP A (also called the Ogallala Aquifer). A 2011 report by the
National Academy of Sciences highlighted the groundwater drawdown in the HP A, noting that
Nebraska is "among the states with the largest water withdrawals for irrigation, and its usage has
continued to increase in recent years, largely driven by the need to irrigate corn for ethanol."482
This suggests that the majority of groundwater consumption would come from areas like
Nebraska, which are already impacted by over-pumping due to their high blue water footprint for
corn production. Changes in irrigation practices are dependent on a number of economic and
agronomic factors that affect how land is managed, making it difficult to attribute expanded
irrigation to biofuels production and use without more detailed analysis. A study by Wright et al.
482 NAS (2011). Renewable Fuel Standard: Potential Economic and Environmental Effects of U.S. Biofuel Policy.
National Academy of Sciences. Washington. DC.
247
-------
(2017) of land use change rates noted that "along the Ogallala Aquifer, elevated rates of land use
change to corn production in Western Kansas, Oklahoma and Texas coincided with areas
experiencing groundwater depletion rates ranging from 5-20% per decade" (see Figure 4.5.3-3).
However, this correlation does not necessarily mean there is a direct, causal relationship between
biofuel production and groundwater depletion.
Figure 4.5,3-3: Relative conversion rates of arable non-cropland to cropland (2008-2012).
0 .4 .8 1.2 2.0 2.7 3.9 5.5 7.8 12.1 35
Includes conversion located along the Ogallala aquifer. Stars denote biofuel production facilities. (Source: Wright et
al. 2017)
As stated above, there have been no comprehensive studies of the changes in irrigated
acres, rates of irrigation, or changes in surface and groundwater supplies attributed specifically to
the increased production of corn grain-based ethanol and soybean-based biodiesel. In the absence
of analyses that do focus directly on crops for biofuel production, there are studies that look
more broadly at the connection between agricultural water use and groundwater levels. For
example, Smidt et al. (2016) analyzed the water-energy-food nexus over the FLPA to look at the
major drivers that have affected and will continue to affect agriculture's water use. That study
highlights that, across large portions of the HP A, "groundwater levels have declined at
unsustainable rates despite improvements in both the efficiency of water use and water
248
-------
productivity in agricultural practices."483 Figure 4.5.3-3 shows the relative conversion rates of
arable cropland to non-cropland, as well as the location of the HPA and biofuel conversion
facilities. Figure 4.5.3-4 also shows the HPA, but shows the absolute changes in groundwater
levels, from predevelopment to 2017 based on data from USGS.484 The HPA can be divided into
three geographical regions: the Northern, Central, and Southern High Plains. The Northern HPA
groundwater supplies have been relatively stable since predevelopment (with some increases
shown in green/blue), whereas the Central and Southern HP As have seen substantial declines, in
some areas over 150 ft of declines (shown in yellow/orange/red). Biofuel facility locations (from
the National Renewable Energy Laboratory485) are overlaid onto the HPA data from USGS to
highlight where biofuel production facilities are co-located with areas of changes in the
groundwater levels. Again, while the Central and Southern HP As have seen substantial declines,
the Northern HPA has remained relatively stable and even increased in some areas (as shown in
Figure 4.5.3-4). This does not demonstrate that biofuel production causes declines in
groundwater levels, but it does show that some biofuel facilities operate in areas that are
experiencing water-stressed aquifer resources.
483 Smidt, S. J., Haacker, E. M., Kendall, A. D., Deines, J. M., Pei, L., Cotterman, K. A., ... & Hyndman, D. W.
(2016). Complex water management in modern agriculture: Trends in the water-energy-food nexus over the High
Plains Aquifer. Science of the Total Environment, 566, 988-1001.
484 The predevelopment water level is defined as "the water level in the aquifer before extensive groundwater
pumping for irrigation, or about 1950. The predevelopment water level was generally estimated by using the earliest
water-level measurement in more than 20,000 wells." https://ne.water.nsgs.gov/proiects/HPA/index.html.
485 https://maps.nrel.gov/biofuels-atlas.
249
-------
Figure 4.5.3-4: Water-level changes in the High Plains Aquifer, predevelopment (about
1950) to 2015
MING
•Falls A {
C heyenne
• • • I
. • •
•\ *•*
»••• v •
J • • ' •
~ • • oUn"*' 9 •
0 NH TED'
, S T A *T E S
COLORADO
Tope.I a
V A NJ
fs- •
® 0 *£.D HILLS
FUNT HILLS
AtfTai ilo,
O ¥ . .
01 lalioina C Ity
ME •
i riegr at*d_bioi — Pini..
• giU4
La.-i3
[etnan oilplants
• <30
• £53
• £85
• £165
M<313
[btodle&el_plant
| * <20
~ <50
A -1QQ
• o
*
Midland
Odessa
hp.A'kpd
¦
«-150
£-100
<-50
<-25
£-10
£-5
£5
£ 10
¦
<25
¦
£50
¦
< 82 |
Dallas
Water levels changes in feet, yellows and reds represent decreases in groundwater levels while greens and blue
represent rises in groundwater levels. Grey indicated no substantial change. Data from U.S. Geological Survey
(USGS). Ethanol plants, biodiesel plans, and integrated biorefineries size (based on annual capacity) and location
from the Bioenergy Atlas, maintained by NREL. ttps://maps.nrel.gov/biofuels-atlas.
4.5.3.1 Crop Prices and Value of Irrigation
Recent research has also assessed the linkage between crop prices and irrigation rates to
find the irrigation values ($/ha/mm) which reflect the price incentive to irrigate. Comparing the
value of irrigation across commodities, Smidt et al (2017) found that the highest value was for
250
-------
corn. The value of irrigation was lower for soybeans.486 The high value of irrigation for corn is
due to the large yield increases that occur with irrigation for corn, as well as the high water-use
efficiency. This indicates that higher corn prices will increase the incentives to irrigation, and,
conversely, lower corn prices may lead to decreases in acres and rates of irrigation. The same
would hold true for soy- higher prices for soybeans incentivizing more irrigation, lower prices
leading to less irrigation. Though the impact would be smaller than it would be from an increase
in corn cultivation, since soybeans generally require less irrigation than does corn.
Earlier work also looked at the impact of agricultural commodity prices on irrigation
demands by taking an economic-based approach that calculated the price elasticities of irrigation
water demands.487 More recent work by Deines et al (2017) utilized satellite images to produce
annual maps of irrigation for 1999-2016 to study changes in irrigation over time.488 In addition to
looking at changes in the area and location of irrigated fields, Deines et al (2017) also did
statistical modeling to assess how factors such as precipitation and commodity prices influenced
the extent of irrigation. That study confirmed that "farmers expanded irrigation when crop prices
were high to increase crop yield and profit."489
4.5.3.2 Non-Cropland Biofuels andNon-U.S. Crops
Published research on the water quantity impacts of biofuels generally do not report or
estimate water used for production of non-cropland biofuels or impacts outside of the U.S.
However, some of the changes in volumes are associated with non-cropland biofuels, such as
biogas, or with biofuels produced from feedstocks produced in foreign countries, such as palm-
based biofuels. We will briefly discuss biogas here. Palm oil water demands are discussed in
Section 2.6 of the Second Triennial Report. In addition, as noted in Chapter 4.3, there is strong
evidence that expanded palm oil production would have adverse impacts on water quality outside
of the U.S.
Biogas does not have the irrigation requirements associated with crop-based biofuels.
Because their inventory covers all of the U.S. energy system at a high level of detail (including
126 unit processes), Grubert and Sanders (2018) examined whether there were any water
consumption and withdrawals for biogas from landfills, wastewater and animal manure
digesters.490 For biogas, they reported no water requirements. Since the biogas is a byproduct of
wastes (i.e., landfills, manure, and wastewater), none of the water used for the primary products
(e.g., the agricultural operations that produced the manure) is allocated to the produced biogas. In
486 Smidt, S. J., Haacker, E. M., Kendall, A. D., Deines, J. M., Pei, L., Cotterman, K. A.,... & Hyndman, D. W.
(2016). Complex water management in modern agriculture: Trends in the water-energy-food nexus over the High
Plains Aquifer. Science of the Total Environment, 566, 988-1001.
487 See for example, Scheierling, S. M., Loomis, J. B., & Young, R. A. (2006). Irrigation water demand: A meta-
analysis of price elasticities. Water resources research, 42(1).
488 Deines, J.M., Kendall, A.D., and Hyndman, D.W. (2017). Annual Irrigation Dynamics in the U.S. Northern High
Plains Derives from Landsat Satellite Data. Geophysical Research Letters 44, 9350-9360.
489 Deines, J.M., Kendall, A.D., and Hyndman, D.W. (2017). Annual Irrigation Dynamics in the U.S. Northern High
Plains Derives from Landsat Satellite Data. Geophysical Research Letters 44, 9350-9360.
490 Grubert, E., & Sanders, K. (2018). Water Use in the United States Energy System: A National Assessment and
Unit Process Inventory of Water Consumption and Withdrawals. Environmental Science & Technology 52(11),
6695-6703.
251
-------
the case of landfill biogas, we therefore assume no significant amounts of water are used.
Grubert and Sanders also assume negligible water requirements for the processing and
transportation of biogas, although they note that some water may be used for upgrading biogas if
water-intensive amine scrubbing is used.
4.5.4 Potential Future Impacts of Annual Volume Requirements
Most of the available research looks at the past and potential future water quantity and
availability impacts associated with increased use of corn ethanol, and in some instances,
cellulosic biofuels. Because of the high volumes of corn ethanol produced to date, the water
quantity and availability concerns have been focused on corn ethanol, with less focus on soy
biodiesel. The changes in mandatory volumes under this rule and future rules are different from
the scenarios analyzed in the literature published to date. Studies on future water quantity
impacts often project larger changes for corn ethanol491 or focus on future cellulosic feedstocks
while soy renewable diesel is the fuel with the more significant increases expected under this
final rule.492 Thus, the water quantity impacts due to this rule are difficult to quantify based on
the existing literature. That said, there are several ways to assess the impacts of the volume
scenarios, based on the studies reviewed above.
We can assess potential water demand changes based on volume changes by biofuel type
as summarized in Section 3.3 of the Second Triennial Report to Congress on Biofuels. All else
being equal, the life cycle water consumption of ethanol and biodiesel (derived from soybeans
and likely palm) is higher, sometimes orders of magnitude higher, than the petroleum-based fuels
they are intended to displace (see Chapter 2.2). However, while the life cycle approach estimates
the direction of changes in the water demands associated with shifting from petroleum to
biomass-based fuels, how much that translates into increased irrigation or changes in water
availability is more difficult to assess.
A second approach to estimate changes in water demands due to the volume changes
would rely on scenarios projecting land use changes and changes in crop management practices
with a high enough level of precision to also assess or estimate the change in irrigation
requirements. One study we reviewed attempted to project water requirements of increased
biofuels production in the U.S.493 However, the biofuel volumes modeled by Liu et al. (2017)
represented an E20 scenario for 2025 and differed greatly in their modeled expansion of crops
compared to the volumes in this rulemaking.
491 Liu, X. V., Hoekman, S. K., & Broch, A. (2017). Potential water requirements of increased ethanol fuel in the
USA. Energy, Sustainability and Society, 7(1), 18.
492 Several studies have estimated water use and availability impacts associated with future scenarios of increased
cellulosic biofuel production. These studies often project future land use/management for different scenarios of
increased production of cellulosic crops, and then estimate impacts on water use and changes in streamflow for
specific watersheds. See for example: Cibin, R., Trybula, E., Chaubey, I., Brouder, S. M., & Volenec, J. J. (2016).
Watershed-scale impacts of bioenergy crops on hydrology and water quality using improved SWAT model. Gcb
Bioenergy, 8(4), 837-848 orLe, P. V., Kumar, P., & Drewry, D. T. (2011). Implications for the hydrologic cycle
under climate change due to the expansion of bioenergy crops in the Midwestern United States. Proceedings of the
National Academy of Sciences, 108(37), 15085-15090.
493 Liu, X., Hoekman, S.K., and Broch, A. 2017. Potential water requirements of increased ethanol fuel in the USA.
Energy, Sustainability and Society, 7:18.
252
-------
A third approach to estimate the changes in water demands is based on changes in crop
prices and the associated economic value of irrigation. While the attribution of impacts due to
land use changes and associated irrigation requirements is difficult, it may be possible to assess
at a broad scale, at least in terms of directionality, the changes in irrigation that may result from
the impact of the candidate volumes on crop prices. However, we have not yet been able to
perform such an analysis and it remains an area where additional analysis and research is needed
to better understand the impacts of the promulgated volumes on water demand.
In summary, based on the approaches above, there will likely be some increased
irrigation pressure on water resources due to the candidate volumes. Specifically, the volume
increases for 2023-2025 compared to the No RFS baseline that is described in Chapter 2 due to
biofuels produced from agricultural feedstocks (especially corn and soybeans) would suggest the
potential for some associated increase in crop production, which in turn would likely increase
irrigation pressure on water resources. The increased volume requirements especially that of
renewable diesel could incent greater production of its underlying feedstock (soybeans). There is
uncertainty in projecting changes in acreage and irrigation rates associated with corn, soybeans,
and other crops. Additional information and modeling are needed to fully assess changes in
water demands and effects on water stressed regions, both for crop irrigation as well as impacts
of biofuel facility water use. Additionally, and as described in Chapter 4.4, we note that there
may be potential effects on water and soil quality with is discussed in the May 19 BE as well.
While we could not quantify these effects, as described in Chapter 4.4, the potential for negative
effects is an area of ongoing concern and research.
4.6 Ecosystem Services
Ecosystem services broadly consist of the many life-sustaining benefits humans receive
from nature, such as clean air and water, fertile soil for crop production, pollination, and flood
control.494 The United Nations Millennium Ecosystem Assessment495 categorized four different
types of ecosystem services, including:
• Provisioning Services; the provision of food, fresh water, fuel, fiber, and other goods
• Regulating Services; climate, water, and disease regulation as well as pollination
• Supporting Services; soil fermentation and nutrient cycling
• Cultural services; education, aesthetic, and cultural heritage values as well as recreation
and tourism
Several of the drivers of ecosystems loss identified in the Millennium Ecosystem
Assessment, such as climate change, pollution, and land-use change, are expected to be impacted
by the production of renewable fuels generally and may be impacted by the candidate volumes in
this rule specifically.
494 US EPA website on Ecosystem Services, https://www.epa.gov/eco-researcli/ecosvstem-services.
495 Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-being: Synthesis. Island Press,
Washington, DC.
253
-------
The previous sections in this chapter discussed the projected impacts associated with this
rule on a variety of different environmental end points such as air quality, climate change, land-
use change, soil and water quality, and water availability as required by the statute. Each of the
impacts discussed in these sections would be expected to have an impact on one or more
ecosystems services. These impacts could be positive (e.g., result in ecosystem services benefits)
or negative. We have focused our analyses in the specific factors identified in the statute and we
have not quantified all of the human well-being changes or monetized these effects. We have,
however, provided a potential framework for how the impacts on ecosystem services might be
considered (see Figure 4.6-1). Note that there are multiple frameworks for categorizing
ecosystem services in the literature. Future analyses, such as those presented in the Triennial
Biofuels and the Environment reports to Congress, may refine this approach to better capture
incremental ecosystem service benefits and costs.
254
-------
Figure 4.6-1: Framework for Considering the Impact of the RFS Volumes on Ecosystem
Services
Action: RFS Set Rule (2023-2025)
r ^
Biophysical Changes
^ J
r ^
Human Well-Being
Changes
r ^
Monetary Value
Changes
GHG Emissions
Displacing Petroleum
Based Fuels
Domestic Land Use
Change
International Land Use
Change
Air Quality
Potential changes in PM,
NOx, S02, VOC, CO, NH3
Water Quality and Aquatic
Habitats
• Fertilizer and Pesticide
Runoff
• Sediment Runoff
• Habitat and Associated
Filtration
• Leakage from
Underground Storage
Tanks
• Atmospheric
Deposition
Flydrology, Water Quantity,
and Flood Risk
Tilling
Land Use/Habitat Change
Irrigation
Wildlife and Habitat
Pollinating Insects
Commercial Species
Species of Public Interest
Pest Control Species
Social Effects from Climate
Change
Property Effects
Morbidity and Mortality Effects
Energy, Transportation, and
Drinking Water Production
Effects
Recreation Effects
Social Cost of GHGs
Property Values
Health Values
Wildlife Product Value
Wildlife Existence Value
Recreation Value
Agricultural Product Value
Soil Quality
255
-------
Chapter 5: Energy Security Impacts
The CAA directs EPA to analyze "the impact of renewable fuels on the energy security
of the United States" in using the set authority to establish volumes. U.S. energy security is
broadly defined as the uninterrupted availability of energy sources at an acceptable price.496
Most discussions of U.S. energy security revolve around the topic of the economic costs of U.S.
dependence on oil imports.497 In addition to evaluating energy security, we have also considered
energy independence, which is the idea of eliminating U.S. dependence on imports of petroleum
and other foreign sources of energy, or more broadly, as reducing the sensitivity of the U.S.
economy to energy imports and foreign energy markets.498 While energy independence is not a
statutory factor in the CAA, one goal of the RFS program is to improve the U.S.'s energy
independence.499 Energy independence and energy security are distinct but related concepts, and
an analysis of energy independence also helps to inform our analysis of energy security.500
Since renewable fuels substitute for petroleum-derived conventional fuels, changes in
renewable fuel volumes have an impact on U.S. petroleum consumption and imports. All else
being constant, a change in U.S. petroleum consumption and imports would alter both the
financial and strategic risks associated with sudden disruptions in global oil supply, thus
influencing the U.S.'s energy security position. Renewable fuels also may have some energy
security risks, for example, as a result of weather-related events (e.g., droughts). To the extent
that renewable fuel price shocks are not strongly correlated with oil price shocks, blending
renewable fuels with petroleum fuels can provide energy security benefits. However, the energy
security risks of using renewable fuels themselves are not well understood, nor well studied. This
chapter reviews the literature on energy security impacts associated with petroleum consumption
and imports and summarizes EPA's estimates of the benefits of reduced petroleum consumption
and imports that would result from the candidate volumes for 2023-2025.
The U.S.'s oil consumption has been gradually increasing in recent years (2015-2019)
before dropping dramatically as a result of the COVID-19 pandemic in 2020 and 2021.501 U.S.
oil consumption rebounded to roughly pre-COVID-19 levels in 2022 and is anticipated to
modestly decline in the 2023-2025 timeframe of this rule.502'503 The U.S. has increased its
496 IEA. Energy Security: Reliable, affordable access to all fuels and energy sources. 2019. December.
497 The issue of cyberattacks is another energy security issue that could grow in significance over time. For example,
one of the U.S.'s largest pipeline operators, Colonial Pipeline, was forced to shut down after being hit by a
ransomware attack. The pipeline carries refined gasoline and jet fuel from Texas to New York. Cyberattack Forces
a Shutdown of a Top U.S. Pipeline. New York Times. May 8th, 2021.
498 Greene, D. 2010. Measuring energy security: Can the United States achieve oil independence? Energy Policy 38,
pp. 1614-1621.
499 See Americans for Clean Energy v. Env't Prot. Agency, 864 F.3d 691, 696 (D.C. Cir. 2017) ("By mandating the
replacement—at least to a certain degree—of fossil fuel with renewable fuel, Congress intended the Renewable Fuel
Program to move the United States toward greater energy independence and to reduce greenhouse gas emissions.");
id. 697 (citing 121 Stat, at 1492).
500 Greene, D. 2010. Measuring energy security: Can the United States achieve oil independence? Energy Policy 38,
pp. 1614-1621.
501 EIA. Total Energy. Monthly Energy Review. Table 3.1. Petroleum Overview. December 2021.
502 EIA. Total Energy. Monthly Energy Review. Table 3.1. Petroleum Overview. March 2023.
503 EIA. AEO 2023. Reference Case. Table A11. Petroleum and Other Liquids Supply and Disposition.
256
-------
production of oil, particularly "tight" (i.e., shale) oil over the last decade.504 As a result of the
recent increase in U.S. oil production and to a lesser extent renewable fuels, the U.S. is projected
to be a net exporter of crude oil and refined petroleum products in 2023-2025.505 This is a
significant reversal of the U.S.'s oil trade balance position since the U.S. has been a substantial
net importer of crude oil and refined petroleum products starting in the early 1950s.506
Oil is a commodity that is globally traded and, as a result, an oil price shock is
transmitted globally. Given that the U.S. is projected to be a modest net exporter of crude oil and
refined petroleum products for 2023-2025, one could reason that the U.S. no longer has a
significant energy security problem. However, U.S. refineries still rely on significant imports of
heavy crude oil which could be subject to supply disruptions. Also, oil exporters with a large
share of global production have the ability to raise or lower the price of oil by exerting the
market power associated with a cartel—the Organization of Petroleum Exporting Countries
(OPEC)—to alter oil supply relative to demand. The degree of market power that OPEC has
during the three-year time frame of this analysis is difficult to quantify. These factors contribute
to the continued vulnerability of the U.S. economy to episodic oil supply shocks and price
spikes, even when the U.S. is projected to be a modest net exporter of crude oil and refined
petroleum products in the 2023-2025 time frame of this rule.
We recognize that because the U.S. is a participant in the world market for crude oil and
refined petroleum products, its economy cannot be shielded from worldwide price shocks.507 But
the potential for petroleum supply disruptions due to supply shocks has been diminished due to
the increase in tight oil production, and to a lesser extent renewable fuels (among other factors),
which have shifted the U.S. to being a modest net petroleum exporter in the world petroleum
market in 2023-2025.508 The potential for supply disruptions has not been eliminated, however,
due to the continued need to import petroleum to satisfy the demands of the U.S. petroleum
industry and because the U.S. continues to consume substantial quantities of oil.509
5.1 Review of Historical Energy Security Literature
Energy security discussions are typically based around the concept of the oil import
premium, sometimes also labeled the oil security premium. The oil import premium is the extra
cost/impacts of importing oil beyond the price of the oil itself as a result of: (1) potential
macroeconomic disruption and increased oil import costs to the economy from oil price spikes or
"shocks"; and (2) monopsony impacts. Monopsony impacts stem from changes in the demand
for imported oil, which changes the price of all imported oil.
504 EIA (2021). Tight oil production estimates by play.
505 EIA. AEO 2023. Reference Case. Table A11. Petroleum and Other Liquids Supply and Disposition.
506 EIA. "Oil and petroleum products explained - Oil imports and exports." April 21st, 2022.
507 Bordoff, J. 2019. The Myth of US Energy Independence has Gone Up in Smoke. Foreign Policy. September 18th.
508 Krupnick, A., Morgenstern, R., Balke, N., Brown, S., Herrara, M. and Mohan, S. 2017. "Oil Supply Shocks, U.S.
Gross Domestic Product, and the Oil Security Problem," Resources for the Future Report.
509 Foreman, D. 2018. Why the US must Import and Export Oil: American Petroleum Institute. June 14th.
257
-------
The so-called oil import premium gained attention as a guiding concept for energy policy
in the aftermath of the oil shocks of the 1970s (Bohi and Montgomery (1982), EMF (1981)).510
Plummer (1982) provided valuable discussion of many of the key issues related to the oil import
premium as well as the analogous oil stockpiling premium.511 Bohi and Montgomery (1982)
detailed the theoretical foundations of the oil import premium and established many of the
critical analytic relationships.512 Hogan (1981) and Broadman and Hogan (1986, 1988) revised
and extended the established analytical framework to estimate optimal oil import premia with a
more detailed accounting of macroeconomic effects.513 Since the original work on energy
security was undertaken in the 1980s, there have been several reviews on this topic by Leiby,
Jones, Curlee and Lee (1997) and Parry and Darmstadter (2004). 514>515
The economics literature on whether oil shocks are the same level of threat to economic
stability as they once were, is mixed. Some of the literature asserts that the macroeconomic
component of the energy security externality is small. For example, the National Research
Council (2009) argued that the non-environmental externalities associated with dependence on
foreign oil are small, and potentially trivial.516 Analyses by Nordhaus (2007) and Blanchard and
Gali (2010) question the impact of oil price shocks on the economy in the early-2000s time
frame.517 They were motivated by attempts to explain why the economy actually expanded
during the oil shock in the early-2000s, and why there was no evidence of higher energy prices
being passed on through higher wage inflation. One reason, according to Nordhaus and
Blanchard and Gali, is that monetary policy has become more accommodating to the price
impacts of oil shocks. Another reason is that consumers have simply decided that such
movements are temporary and have noted that price impacts are not passed on as inflation in
other parts of the economy.
Hamilton (2012) reviews the empirical literature on oil shocks and suggests that the
results are mixed, noting that some work (e.g., Rasmussen and Roitman (2011)) finds less
evidence for economic effects of oil shocks or declining effects of shocks (Blanchard and Gali
510 Bohi, D. and Montgomery, D. 1982. Social Cost of Imported and U.S. Import Policy, Annual Review of Energy,
7:37-60. Energy Modeling Forum, 1981. World Oil, EMF Report 6, Stanford University Press: Stanford 39 CA.
511 Plummer, J. (Ed.). 1982. Energy Vulnerability, "Basic Concepts, Assumptions and Numerical Results," pp. 13-
36, Cambridge MA: Ballinger Publishing Co.
512 Bohi, D. and Montgomery, D. 1982. Social Cost of Imported and U.S. Import Policy, Annual Review of Energy,
7:37-60.
513 Hogan, W. 1981. "Import Management and Oil Emergencies," Chapter 9 in Deese, 5 David and Joseph Nye, eds.
Energy and Security. Cambridge, MA: Ballinger Publishing Co. Broadman, H. 1986. "The Social Cost of Imported
Oil," Energy Policy 14(3):242-252. Broadman H. and Hogan, W. 1988. "Is an Oil Import Tariff Justified? An
American Debate: The Numbers Say 'Yes,'" The Energy Journal 9: 7-29.
514 Leiby, P., Jones, D., Curlee, R. and Lee, R. 1997. Oil Imports: An Assessment of Benefits and Costs, ORNL-
6851, Oak Ridge National Laboratory, November.
515 Parry, I. and Darmstadter, J. 2004. "The Costs of U.S. Oil Dependency," Resources for the Future, November 17,
2004. Also published as NCEP Technical Appendix Chapter 1: Enhancing Oil Security, the National Commission
on Energy Policy 2004 Ending the Energy Stalemate-A Bipartisan Strategy to Meet America's Energy Challenges.
516 National Research Council. 2009. Hidden Costs of Energy: Unpriced Consequences of Energy Production and
Use. National Academy of Science, Washington, DC.
517 Nordhaus, W. 2007. "Who's Afraid of a Big Bad Oil Shock?". Brookings Papers on Economic Activity,
Economic Studies Program, The Brookings Institution, vol. 38(2), pp. 219-240. Blanchard. O. and Gali. J. 2010. The
macroeconomic effects of oil price shocks: why are the 2000's so different from the 1970s. International
Dimensions of Monetary Policy. University of Chicago Press.
258
-------
(2010)), while other work continues to find evidence regarding the economic importance of oil
shocks.518 For example, Baumeister and Peersman (2012) find that an "oil price increase of a
given size seems to have a decreasing effect over time, but noted that the declining price-
elasticity of demand meant that a given physical disruption had a bigger effect on price and
turned out to have a similar effect on output as in the earlier data."519 Hamilton observes that "a
negative effect of oil prices on real output has also been reported for a number of other countries,
particularly when nonlinear functional forms have been employed" (citing as examples Kim
(2012) and Engemann, Kliesen, and Owyang (2011)).520'521 Alternatively, rather than a declining
effect, Ramey and Vine (2010) find "remarkable stability in the response of aggregate real
variables to oil shocks once we account for the extra costs imposed on the economy in the 1970s
by price controls and a complex system of entitlements that led to some rationing and
shortages."522
Some of the literature on oil price shocks emphasizes that economic impacts depend on
the nature of the oil shock, with differences between price increases caused by a sudden supply
loss and those caused by rapidly growing demand. Recent analyses of oil price shocks have
confirmed that "demand-driven" oil price shocks have greater effects on oil prices and tend to
have positive effects on the economy while "supply-driven" oil shocks still have negative
economic impacts (Baumeister, Peersman, and Robays (2010)).523 A paper by Kilian and
Vigfusson (2014), for example, assigns a more prominent role to the effects of price increases
that are unusual, in the sense of being beyond the range of recent experience.524 Kilian and
Vigfussen also conclude that the difference in response to oil shocks may well stem from the
different effects of demand- and supply-based price increases: "One explanation is that oil price
shocks are associated with a range of oil demand and oil supply shocks, some of which stimulate
the U.S. economy in the short-run and some of which slow down U.S. growth (see Kilian
2009) ,"525
The general conclusion that oil supply-driven shocks reduce economic output is also
reached in a paper by Cashin et al. (2014), which focused on 38 countries from 1979-2011.526
They state: "The results indicate that the economic consequences of a supply-driven oil-price
518 Rasmussen, T. and Roitman, A. 2011. Oil Shocks in a Global Perspective: Are We Really That Bad. IMF
Working Paper Series.
519 Baumeister, C. and Peersman, G. 2012. The Role of Time-Varying Price Elasticities in Accounting for Volatility
Changes in the Crude Oil Market. Journal of Applied Economics.
520 Kim, D. 2012. What is an oil shock? Panel data evidence. Empirical Economics, Volume 43, pp. 121-143.
521 Engemann, K., Kliesen. K. and Owyang, M. 2011. Do Oil Shocks Drive Business Cycles, Some U.S. and
International Evidence. Federal Reserve Bank of St. Louis, Working Paper Series. No. 2010-007D.
522 Ramey, V. and Vine, D. 2010. "Oil, Automobiles, and the U.S. Economy: How Much have Things Really
Changed?". National Bureau of Economic Research Working Papers, WP 16067 (June).
523 Baumeister C., Peersman, G. and Van Robays, I. 2010. "The Economic Consequences of Oil Shocks: Differences
across Countries and Time." RB A Annual Conference Volume in: Renee Fry & Galium Jones & Christopher Kent
(ed.), Inflation in an Era of Relative Price Shocks, Reserve Bank of Australia.
524 Kilian, L. and Vigfusson, R. 2014. "The role of oil price shocks in causing U.S. recessions." CFS Working Paper
Series 460, Center for Financial Studies.
525 Kilian, L. 2009. "Not All Oil Price Shocks Are Alike: Disentangling Demand and Supply Shocks in the Crude
Oil Market." American Economic Review, 99 (3): pp. 1053-69.
526 Cashin, P., Mohaddes, K., and Raissi, M. 2014. The Differential Effects of Oil Demand and Supply Shocks on
the Global Economy, Energy Economics, 12 (253).
259
-------
shock are very different from those of an oil-demand shock driven by global economic activity,
and vary for oil-importing countries compared to energy exporters." Cashin et al. continues
. .oil importers (including the U.S.) typically face a long-lived fall in economic activity in
response to a supply-driven surge in oil prices." But almost all countries see an increase in real
output caused by an oil-demand disturbance.
Considering all of the recent energy security literature, EPA's assessment concludes that
there are benefits to the U.S. from reductions of its oil imports. There is some debate as to the
magnitude, and even the existence, of energy security benefits from U.S. oil import reductions.
However, differences in economic impacts from oil demand and oil supply shocks have been
distinguished, with oil supply shocks resulting in economic losses in oil importing countries. The
oil import premium calculations in this analysis (described in Chapter 5.4) are based on price
shocks from potential future supply events. Oil supply shocks have been the predominant focus
of oil security issues since the oil price shocks/oil embargoes of the 1970s. While we project
some increase in imported renewable fuels due to this rule, the rule results in an overall reduction
by the U.S. in imported fuels (i.e., combined total of imported oil and imported renewable fuels),
moving the U.S. modestly towards the goal of energy independence and enhanced energy
security.
5.2 Review of Recent Energy Security Literature
There have also been a handful of recent studies that are relevant for the issue of energy
security. We provide a brief review and high-level summary of each of these studies below.
5.2.1 Recent Oil Energy Security Studies
The first studies on the energy security impacts of oil that we review are by Resources for
the Future (RFF), a study by Brown and two studies by Oak Ridge National Laboratory (ORNL).
The RFF study (2017) attempts to develop updated estimates of the relationship among gross
domestic product (GDP), oil supply and oil price shocks, and world oil demand and supply
elasticities.527 In a follow-on study, Brown summarized the RFF study results as well.528 The
RFF work argues that there have been major changes that have occurred in recent years that have
reduced the impacts of oil shocks on the U.S. economy. First, the U.S. is less dependent on
imported oil than in the early 2000s due in part to the 'Tracking revolution" (i.e., tight/shale oil),
and to a lesser extent, increased production of renewable fuels. In addition, RFF argues that the
U.S. economy is more resilient to oil shocks than in the earlier 2000s time frame. Some of the
factors that make the U.S. more resilient to oil shocks include increased global financial
integration and greater flexibility of the U.S. economy (especially labor and financial markets),
many of the same factors that Nordhaus and Blanchard and Gali pointed to as discussed above.
In the RFF effort, a number of comparative modeling scenarios are conducted by several
economic modeling teams using three different types of energy-economic models to examine the
impacts of oil shocks on U.S. GDP. The first is a dynamic stochastic general equilibrium model
527 Krupnick, A., Morgenstern, R., Balke, N., Brown, S., Herrara, M. and Mohan, S. 2017. "Oil Supply Shocks, U.S.
Gross Domestic Product, and the Oil Security Problem," Resources for the Future Report.
528 Brown, S. 2018. New estimates of the security costs of U.S. oil consumption", Energy Policy, 113 pp. 171-192.
260
-------
developed by Balke and Brown.529 The second set of modeling frameworks use alternative
structural vector autoregressive models of the global crude oil market.530 The last of the models
utilized is the National Energy Modeling System (NEMS).531
Two key parameters are focused upon to estimate the impacts of oil shock simulations on
U.S. GDP: oil price responsiveness (i.e., the short-run price elasticity of demand for oil) and
GDP sensitivity (i.e., the elasticity of GDP to an oil price shock). The more inelastic (i.e., the less
responsive) short-run oil demand is to changes in the price of oil, the higher the price impacts of
a future oil shock. Higher price impacts from an oil shock result in higher GDP losses. The more
inelastic (i.e., less sensitive) GDP is to an oil price change, the less the loss of U.S. GDP with
future oil price shocks.
For oil price responsiveness, RFF reports three different values: a short-run price
elasticity of oil demand from their assessment of the "new literature," -0.17; a "blended"
elasticity estimate; -0.05, and short-run oil price elasticities from the "new models" RFF uses,
ranging from -0.20 to -0.35. The "blended" elasticity is characterized by RFF in the following
way: "Recognizing that these two sets of literature [old and new] represent an evolution in
thinking and modeling, but that the older literature has not been wholly overtaken by the new,
Benchmark-E [the blended elasticity] allows for a range of estimates to better capture the
uncertainty involved in calculating the oil security premiums."
The second parameter that RFF examines is the GDP sensitivity. For this parameter,
RFF's assessment of the "new literature" finds a value of-0.018, a "blended elasticity" estimate
of-0.028, and a range of GDP elasticities from the "new models" that RFF uses that range from
-0.007 to -0.027. One of the limitations of the RFF study is that the large variations in oil price
over the last 15 years are believed to be predominantly "demand shocks" (e.g., a rapid growth in
global oil demand followed by the Great Recession and then the post-recession recovery).
There have only been two recent situations where events have led to a potential
significant supply-side oil shock in the last several years. The first event was the attack on the
Saudi Aramco Abqaiq oil processing facility and the Khurais oil field. On September 14th, 2019,
a drone and cruise missile attack damaged the Saudi Aramco Abqaiq oil processing facility and
the Khurais oil field in eastern Saudi Arabia. The Abqaiq oil processing facility is the largest
crude oil processing and stabilization plant in the world, with a capacity of roughly 7 MMBD or
about 7% of global crude oil production capacity.532 On September 16th, the first full day of
529 Balke, N. and Brown, S. 2018. "Oil Supply Shocks and the U.S. Economy: An Estimated DSGE Model." Energy
Policy, 116, pp. 357-372.
530 These models include Kilian, L. 2009. Not All Oil Price Shocks are Alike: Disentangling Demand and Supply
Shocks in the Crude Oil Market, American Economic Review, 99:3, pp., 1053-1069; Kilian, L. and Murphy, D.
2013. "The Role of Inventories and Speculative Trading in the Global Market for Crude Oil, "Journal of Applied
Economics; and Baumeister, C. and Hamilton, J. 2019. "Structural Interpretation of Vector Autoregressions with
Incomplete Identification: Revisiting the Role of Oil Supply and Demand Shocks," American Economic Review,
109(5), pp.1873-1910.
531 Mohan, S. 2017. "Oil Price Shocks and the U.S. Economy: An Application of the National Energy Modeling
System." Resources for the Future Report Appendix.
532 EIA. "Saudi Arabia crude oil production outage affects global crude oil and gasoline prices." Today in Energy.
September 23, 2019.
261
-------
commodity trading after the attack, both Brent and WTI crude oil prices surged by $7.17/bbl and
$8.34/bbl, respectively, in response to the attack, the largest price increase in roughly a decade.
However, by September 17th, Saudi Aramco reported that the Abqaiq plant was
producing 2 MMBD, and they expected its entire output capacity to be fully restored by the end
of September.533 Tanker loading estimates from third-party data sources indicated that loadings
at two Saudi Arabian export facilities were restored to the pre-attack levels.534 As a result, both
Brent and WTI crude oil prices fell on September 17th, but not back to their original levels. The
oil price spike from the attack on the Abqaiq plant and Khurais oil field was prominent and
unusual, as Kilian and Vigfusson (2014) describe. While pointing to possible risks to world oil
supply, the oil shock was short-lived, and generally viewed by market participants as being
transitory, so it did not influence oil markets over a sustained time period.
The second situation is the set of events leading to the recent world oil price spike
experienced in 2022. World oil prices rose fairly rapidly in the first half of 2022. For example,
on January 3rd, 2022, the WTI crude oil price was roughly $76/bbl.535 The WTI oil price
increased to roughly $ 124/bbl on March 8th, 2022, a 63% increase.536 Crude oil prices increased
in the first half of 2022 because of oil supply concerns. Russia's invasion of Ukraine came
during eight consecutive quarters (from the third quarter of 2020 to the second quarter of 2022)
of global crude oil inventory decreases.537 The lower inventory of crude oil was the result of
withdrawals from storage to meet the demand that resulted from rising economic activity after
pandemic-related restrictions eased.538 Oil prices have drifted downwards throughout the second
half of 2022 and early 2023. As of March 13th, 2023, the WTI crude oil price was roughly
$75/barrel.539
Geopolitical disruptions that occurred in 2022 are likely to continue to affect global trade
of crude oil and petroleum products in 2023 and beyond. In response to Russia's invasion of
Ukraine in late February 2022, the U.S. and many of its allies, particularly in Europe, announced
various sanctions against Russia's petroleum industry.540 For the European Union (EU),
petroleum from Russia had accounted for a large share of all energy imports, but the EU banned
imports of crude oil from Russia starting in December 2022 and imports of petroleum products
starting in February 2023.541 Given oil market trends in 2022, the U.S. set a new record for
petroleum product exports, up 7% from 2021.542 Since both significant demand and supply
factors are influencing world oil prices in 2022 and the early part of 2023, it is not clear how to
evaluate unfolding oil market price trends from an energy security standpoint. Thus, the attack of
533 Id.
534 Id.
535 U.S. Energy Information Administration. 2022. Petroleum and Other Liquids: Spot Prices.
https://www.eia.gov/dnav/pet/pet pri spt si d.htm.
536 Id.
537 U.S. Energy Information Administration. 2023. Today in Energy. Crude oil prices increased in the first half of
2022 and declined in the second half of 2022. January.
538 Id.
539 EIA. Petroleum and Other Liquids Spot Prices, https://www.eia.gov/dnav/pet/pet pri spt s i d.htm.
540 U.S. Energy Information Administration. 2023. Today in Energy. U.S. petroleum product exports set a record
high in 2022. March.
541 Id.
542 Id.
262
-------
the Abqaiq oil processing facility in Saudi Arabia and the events in the world oil market in 2022
and 2023 do not currently provide enough empirical evidence to provide an updated estimate of
the response of the U.S. economy to an oil supply shock of a significant magnitude.543
A second set of recent studies related to energy security are from ORNL. In the first
study, ORNL (2018) undertakes a quantitative meta-analysis of world oil demand elasticities
based upon the recent economics literature.544 The ORNL study estimates oil demand elasticities
for two sectors (transportation and non-transportation) and by world regions (OECD and Non-
OECD) by meta-regression. To establish the data set for the meta-analysis, ORNL undertakes a
literature search of peer reviewed journal articles and working papers between 2000-2015 that
contain estimates of oil demand elasticities. The data set consisted of 1,983 observations from 75
published studies. The study finds a short-run price elasticity of world oil demand of-0.07 and a
long-run price elasticity of world oil demand of-0.26.
The second relevant ORNL (2018) study from the standpoint of energy security is a
meta-analysis that examines the impacts of oil price shocks on the U.S. economy as well as many
other net oil-importing economies.545 19 studies after 2000 were identified that contain
quantitative/accessible estimates of the economic impacts of oil price shocks. Almost all studies
included in the review were published since 2008. The key result that the study finds is a short-
run oil price elasticity of U.S. GDP, roughly one year after an oil shock, of-0.021, with a 68%
confidence interval of-0.006 to -0.036.
5.2.2 Recent Studies on Tight/Shale Oil
The discovery and development of U.S. tight oil (i.e., shale oil) reserves that started in the
mid-2000s could affect U.S. energy security in at least several ways.546 First, the increased
availability of domestic supplies has resulted in a reduction of U.S. oil imports and an increasing
role of the U.S. as exporter of crude oil and petroleum-based products. In December 2015, the
40-year ban on the export of domestically produced crude oil was lifted as part of the
Consolidated Appropriations Act, 2016. Pub. L. 114-113 (Dec. 18, 20 1 5).547 According to the
GAO, the ban was lifted in part due to increases in tight (i.e., shale) oil.548 Second, due to
543 Hurricanes Katrina and Rita in 2005 primarily caused a disruption in U.S. oil refinery production, with a more
limited disruption of some crude supply in the U.S. Gulf Coast area. Thus, the loss of refined petroleum products
exceeded the loss of crude oil, and the regional impact varied even within the U.S. Hurricanes Katrina and Rita were
a different type of oil disruption event than is quantified in the Stanford EMF risk analysis framework, which
provides the oil disruption probabilities than ORNL is using.
544 Uria-Martinez, R., Leiby, P., Oladosu, G., Bowman, D., Johnson, M. 2018. Using Meta-Analysis to Estimate
World Oil Demand Elasticity, ORNL Working Paper.
545 Oladosu, G., Leiby, P., Bowman, D., Uria-Martinez, R., Johnson, M. 2018. Impacts of oil price shocks on the
U.S. economy: a meta-analysis of oil price elasticity of GDP for net oil-importing economies, Energy Policy 115.
pp. 523-544.
546 Union of Concerned Scientist, "What is Tight Oil?". 2015. "Tight oil is a type of oil found in impermeable shale
and limestone rock deposits. Also known as "shale oil," tight oil is processed into gasoline, diesel, and jet fuels—
just like conventional oil—but is extracted using hydraulic fracturing, or 'Tracking.
547 fattps://nseode.faoiise.gov/statntes/pt/1.1.4/1.1.3.pdf (see 129 stat. 2987).
548 GAO, 2020. Crude Oil Markets: Effects of the Repeal of the Crude Oil Export Ban. GAO-21-118. According to
the G AO. "Between 1975 and the end of 2015, the Energy Policy and Conservation Act directed a ban on nearly all
263
-------
differences in development cycle characteristics and average well productivity, tight oil
producers could be more price responsive than most other oil producers. However, the oil price
level that triggers a substantial increase in tight oil production appears to be higher in 2021-2022
relative to the 2010s as tight oil producers seek higher profit margins per barrel in order to
reduce the debt burden accumulated in previous cycles of production growth.549
U.S. crude oil production increased from 5.0 MMBD in 2008 to an all-time peak of 12.3
MMBD in 2019 and tight oil wells have been responsible for most of the increase.550 Figure
5.2.2-1 shows tight oil production changes from various tight oil producing regions (e.g., Eagle
Ford, Bakken, etc.) in the U.S. and the WTI crude oil spot price. Viewing Figure 5.2.2-1, one can
see that the annual average U.S. tight oil production grew from 0.6 MMBD in 2008 to 7.8
MMBD in 2019.551 Growth in U.S. tight oil production during this period was only interrupted in
2015-2016 following the world oil price downturn that began in mid-2014. The second growth
phase started in late 2016 and continued until 2020. The sharp decrease in demand that followed
the onset of the COVID-19 pandemic resulted in a 25% decrease in tight oil production in the
period from December 2019 to May 2020. U.S. tight oil production in 2020 and 2021 averaged
7.4 MMBD and 7.2 MMBD, respectively, and represents a relatively modest share (less than
10% in 2019) of global liquid fuel supply.552 Importantly, U.S. tight oil is considered the most
price-elastic component of non-OPEC supply due to differences between its development and
production cycle and that of conventional oil wells. Unlike conventional wells where oil starts
flowing naturally after drilling, shale oil wells require the additional step of fracking to complete
the well and release the oil.553 Shale oil producers keep a stock of drilled but uncompleted wells
and can optimize the timing of the completion operation depending on price expectations.
Combining this decoupling between drilling and production with the front-loaded production
profile of tight oil—the fraction of total output from a well that is extracted in the first year of
production is higher for tight oil wells than conventional oil wells—tight oil producers have a
clear incentive to be responsive to prices in order to maximize their revenues.554
exports of U.S. crude oil. This ban was not considered a significant policy issue when U.S. oil production was
declining and import volumes were increasing. However, U.S. crude oil production roughly doubled from 2009 to
2015, due in part to a boom in shale oil production made possible by advancements in drilling technologies. In
December 2015, Congress effectively repealed the ban, allowing the free export of U.S. crude oil worldwide".
549 Kemp, J. 2021. U.S. shale restraint pushes oil prices to multi-year high. Reuters. June 4, 2021.
550 EIA (2021). Crude Oil Production. https://www.eia.gov/dnav/pet/pet crd_crpdnadc rnbbl m.fatm.
551 EIA (2021). Tight oil production estimates byplay. faMs:/feww^^
552 The 2019 global crude oil production value used to compute the U.S. tight oil share is from EIA International
Energy Statistics, https://www.eia.gov/inteniational/data/world/petroleum-and-other-liauids/an.nual-petroleum-and-
other-liquids-prodiiction.
553 Hydraulic fracturing ("fracking") involves injecting water, chemicals, and sand at high pressure to open fractures
in low-permeability rock formations and release the oil that is trapped in them.
554 Bjornland. H., Nordvik, F. and Rohrer, M. 2021. "Supply flexibility in the shale patch: Evidence from North
Dakota," Journal of Applied Economics.
264
-------
Figure 5.2.2-1: U.S. Tight Oil Production (by Producing Regions) (in MMBD) and WTI
Crude Oil Spot Price (in U.S. Dollars per Barrel)
10 -I r 150
o i i i i i i i-o
2008 2010 2012 2014 2016 2018 2020 2022
Producing Regions Price
¦ Bakken ¦ Niobrara-Codell ¦| Bonespring Eagle Ford WT.
(ND & MT) ¦ (CO & WY) ¦ (TX & NM Permian) I (TX)
¦ Spraberry H Wolfcamp RpCtnfiic
(TX Permian) (TX & NM Permian)
Source: IIIA
Only in recent years have the implications of the "tight/shale oil revolution" been felt in
the international market where U.S. production of oil is rising to be roughly on par with Saudi
Arabia and Russia. Recent economic literature of the tight oil expansion in the U.S. has a bearing
on the issue of energy security as well. It could be that the large expansion in tight oil has eroded
the ability of OPEC to set world oil prices to some degree, since OPEC cannot directly influence
tight oil production decisions. Also, by effecting the percentage of global oil supply controlled
by OPEC, the growth in U.S. oil production may be influencing OPEC's degree of market
power. But given that the tight oil expansion is a relatively recent trend, it is difficult to know
how much of an impact the increase in tight oil is having, or will have, on OPEC behavior.
Three recent studies have examined the characteristics of tight oil supply that have
relevance for the topic of energy security. In the context of energy security, the question that
arises is: Can tight oil respond to an oil price shock more quickly and substantially than
conventional oil?557 If so, then tight oil could potentially lessen the impacts of future oil shocks
on the U.S. economy by moderating the price increases from a future oil supply shock.
Newell and Prest (2019) look at differences in the price responsiveness of conventional
versus shale oil wells, using a detailed data set of 150,000 oil wells, during the 2005-2017 time
frame in five major oil-producing states: Texas, North Dakota, California, Oklahoma, and
Colorado.558 For both conventional oil wells and shale oil wells (i.e., unconventional oil wells),
555 EIA. Tight oil production estimates by play, https://www.eia.gov/petroleum/data.php.
556 EIA. Petroleum and Other Liquids Spot Prices, https://www.eia.gov/dnav/pet/pet pri spt si d.htm.
557 Union of Concerned Scientist, "What is Tight Oil?". 2015. "Tight oil is a type of oil found in impermeable shale
and limestone rock deposits. Also known as "shale oil," tight oil is processed into gasoline, diesel, and jet fuels—
just like conventional oil - but is extracted using hydraulic fracturing, or "fracking."
558 Newell, R. and Prest, B. 2019. The Unconventional Oil Supply Boom: Aggregate Price Response from
Microdata, The Energy Jo urn ah Volume 40, Issue Number 3.
265
-------
Newell and Prest estimate the elasticities of drilling operations and well completion operations
with respect to expected revenues and the elasticity of supply from wells already in operation
with respect to spot prices. Combining the three elasticities and accounting for the increased
share of tight oil in total U.S. oil production during the period of analysis, they conclude that
U.S. oil supply responsiveness to prices increased more than tenfold from 2006 to 2017. They
find that tight/shale oil wells are more price responsive than conventional oil wells, mostly due to
their much higher productivity, but the estimated oil supply elasticity is still small. Newell and
Prest note that the tight oil supply response still takes more time to arise than is typically
considered for a "swing producer," referring to a supplier able to increase production quickly,
within 30-90 days. In the past, only Saudi Arabia and possibly one or two other oil producers in
the Middle East have been able to ramp up oil production in such a short period of time.
Another study, by Bjornland et al. (2021), uses a well-level monthly production data set
covering more than 16,000 crude oil wells in North Dakota to examine differences in supply
responses between conventional and tight/shale oil.559 They find a short-run (i.e., one-month)
supply elasticity with respect to oil price for tight oil wells of 0.71, whereas the one-month
response of conventional oil supply is not statistically different from zero. It should be noted that
the elasticity value estimated by Bjornland et al. combines the supply response to changes in the
spot price of oil as well as changes in the spread between the spot price and the 3-month futures
price.
Walls and Zheng (2022) explore the change in U.S. oil supply elasticity that resulted
from the tight oil revolution using monthly, state-level data on oil production and crude oil prices
from January 1986 to February 2019 for North Dakota, Texas, New Mexico, and Colorado.560
They conduct statistical tests that reveal an increase in the supply price elasticities starting
between 2008-2011, coinciding with the times in which tight oil production increased sharply in
each of these states. Walls and Zheng also find that supply responsiveness in the tight oil era is
greater with respect to price increases than price decreases. The short-run (one-month) supply
elasticity with respect to price increases during the tight oil area ranges from zero in Colorado to
0.076 in New Mexico; pre-tight oil, it ranged from zero to 0.021.
The results from Newell and Prest, Bjornland et al., and Walls and Zheng all suggest that
tight oil may have a larger supply response to oil prices in the short-run than conventional oil,
although the estimated short-run elasticity is still small. The three studies use data sets that end in
2019 or earlier. The responsiveness of U.S. tight oil production to recent price increases does not
appear to be consistent with that observed during the episodes of crude oil price increases in the
2010s captured in these three studies. Despite an 80% increase in the WTI crude oil spot price
from October 2020 to the end of 2021, Figure 5.2.2-1 shows that U.S. tight oil production has
increased by only 8% in the same period. It is a somewhat challenging period in which to
examine the supply response of tight oil to its price to some degree, given that the 2020-2021
time period coincided with the COVID-19 pandemic. However, previous shale oil production
growth cycles were financed predominantly with debt, at very low interest rates.561 Most U.S.
559 Bjornland, H., Nordvik, F. and Rohrer, M. 2021. "Supply flexibility in the shale patch: Evidence from North
Dakota," Journal of Applied Economics.
560 Walls, W. D., & Zheng, X. 2022. Fracking and Structural Shifts in Oil Supply. The Energy Journal, 43(3).
561 McLean, B. The Next Financial Crisis Lurks Underground. New York Times, September 1st, 2018.
266
-------
tight oil producers did not generate positive cashflow.562 As of 2021, U.S. shale oil producers
have pledged to repay their debt and reward shareholders through dividends and stock
buybacks.563 These pledges translate into higher prices that need to be reached (or sustained for a
longer period) than in the past decade to trigger large increases in drilling activity.
In its first quarter 2022 energy survey, the Dallas Fed (i.e., the Federal Reserve Bank of
Dallas) asked oil exploration and production (E&P) firms about the WTI price levels needed to
cover operating expenses for existing wells or to profitably drill a new well. The average
breakeven price to continue operating existing wells in the shale oil regions ranged from $23-
35/bbl. To profitably drill new wells, the required average WTI prices ranged from $48-69/bbl.
For both types of breakeven prices, there was substantial variation across companies, even within
the same region. The actual WTI price level observed in the first quarter of 2022 has been
roughly $95/bbl, substantially larger than the breakeven price to drill new wells. However, the
median production growth expected by the respondents to the Dallas Fed Energy Survey from
the fourth quarter of 2021 to the fourth quarter of 2022 is modest (6% among large firms and
15% among small firms). Investor pressure to maintain capital discipline was cited by 59% of
respondents as the primary reason why publicly traded oil producers are restraining growth
despite high oil prices. The other reasons cited included supply chain constraints, difficulty in
hiring workers, environmental, social, and governance concerns, lack of access to financing, and
government regulations.564 Given the recent behavior of tight oil producers, we do not believe
that tight oil will provide additional significant energy security benefits in 2023-2025 due to its
lack of price responsiveness. The ORNL model still accounts for U.S. tight oil production
increases on U.S. oil imports and, in turn, the U.S.'s energy security position.
Finally, despite continuing uncertainty about oil market behavior and outcomes and the
sensitivity of the U.S. economy to oil shocks, it is generally agreed that it is beneficial to reduce
petroleum fuel consumption from an energy security standpoint. The relative significance of
petroleum consumption and import levels for the macroeconomic disturbances that follow from
oil price shocks is not fully understood. Recognizing that changing petroleum consumption will
change U.S. imports, our quantitative assessment of oil costs of this rule in Chapter 5.4 focuses
on those incremental social costs that follow from the resulting changes in net imports,
employing the usual oil import premium measure.
5.3 Cost of Existing U.S. Energy Security Policies
An additional often-identified component of the full economic costs of U.S. oil imports is
the costs to the U.S. taxpayers of existing U.S. energy security policies. The two primary
examples are maintaining the Strategic Petroleum Reserve (SPR) and maintaining a military
presence to help secure a stable oil supply from potentially vulnerable regions of the world.
The SPR is the largest stockpile of government-owned emergency crude oil in the world.
Established in the aftermath of the 1973/1974 oil embargo, the SPR provides the U.S. with a
562 Id.
563 Crowley, K. and Wethe, D. Shale Bets on Dividends to Match Supermajors, Revive Sector. Bloomberg. August
2nd, 2021.
564 Federal Reserve Bank of Dallas. Dallas Fed Energy Survey. March 23rd, 2022.
267
-------
response option should a disruption in commercial oil supplies threaten the U.S. economy.565
Emergency SPR drawdowns have taken place in 1991 (Operation Desert Storm), 2005
(Hurricane Katrina), 2011 (Libyan Civil War), and 2022 (War in Ukraine). All of these releases
have been in coordination with releases of strategic stocks from other International Energy
Agency (IEA) member countries. In the first four months of 2022, using the statutory authority
under Section 161 of the Energy Policy and Conservation Act, DOE conducted two emergency
SPR drawdowns in response to ongoing oil supply disruptions.566 The first drawdown resulted in
a sale of 30 million barrels in March 2022. The second drawdown, announced in April,
authorized a total release of approximately one MMBD from May to October 2022.567 For 2023,
the DOE has announced plans to sell 26 million barrels of oil between April and June.568 While
the costs for building and maintaining the SPR are more clearly related to U.S. oil use and
imports, historically these costs have not varied in response to changes in U.S. oil import levels.
Thus, while the effect of the SPR in moderating price shocks is factored into the analysis that
EPA is using to estimate the macroeconomic oil security premiums, the cost of maintaining the
SPR is excluded.
We have also considered the possibility of quantifying the military benefits components
of energy security but have not done so here for several reasons. The literature on the military
components of energy security has described four broad categories of oil-related military and
national security costs, all of which are difficult to quantify. These include possible costs of U.S.
military programs to secure oil supplies from unstable regions of the world, the energy security
costs associated with the U.S. military's reliance on petroleum to fuel its operations, possible
national security costs associated with expanded oil revenues to "rogue states", and relatedly the
foreign policy costs of oil insecurity.
Of these categories listed above, the one that is most clearly connected to petroleum use
and is, in principle, quantifiable is the first: the cost of military programs to secure oil supplies
and stabilize oil supplying regions. There is ongoing literature on the measurement of this
component of energy security, but methodological and measurement issues—attribution and
incremental analysis—pose two significant challenges to providing a robust estimate of this
component of energy security. The attribution challenge is to determine which military programs
and expenditures can properly be attributed to oil supply protection, rather than some other
objective. The incremental analysis challenge is to estimate how much the petroleum supply
protection costs might vary if U.S. oil use were to be reduced or eliminated. Methods to address
both of these challenges are necessary for estimating the effect on military costs arising from a
modest reduction (not elimination) in oil use attributable to this action.
Since "military forces are, to a great extent, multipurpose and fungible" across theaters
and missions (Crane et al. 2009), and because the military budget is presented along regional
565 Energy Policy and Conservation Act, 42 U.S. Code § 6241(d) (1975).
566 DOE. DOE Announces Emergency Notice of Sale of Crude Oil from the Strategic Petroleum Reserve to Address
Oil Supply Disruptions. 2022. March.
567 DOE. DOE Announces Second Emergency Notice of Sale of Crude Oil From The Strategic Petroleum Reserve to
Address Putin's Energy Price Hike. 2022. April.
568 DOE. DOE Issues Notice of Congressionally Mandated Sale to Purchase Crude Oil from the Strategic Petroleum
Reserve. 2023. February.
268
-------
accounts rather than by mission, the allocation to particular missions is not always clear.569
Approaches taken usually either allocate "partial" military costs directly associated with
operations in a particular region, or allocate a share of total military costs (including some that
are indirect in the sense of supporting military activities overall) (Koplow and Martin 1998).570
The challenges of attribution and incremental analysis have led some to conclude that the
mission of oil supply protection cannot be clearly separated from others, and the military cost
component of oil security should be taken as near zero (Moore et al. 1997).571 Stern (2010), on
the other hand, argues that many of the other policy concerns in the Persian Gulf follow from oil,
and the reaction to U.S. policies taken to protect oil.572 Stern presents an estimate of military cost
for Persian Gulf force projection, addressing the challenge of cost allocation with an activity-
based cost method. He uses information on actual naval force deployments rather than budgets,
focusing on the costs of carrier deployment. As a result of this different data set and assumptions
regarding allocation, the estimated costs are much higher, roughly 4-10 times, than other
estimates. Stern also provides some insight on the analysis of incremental effects, by estimating
that Persian Gulf force projection costs are relatively strongly correlated to Persian Gulf
petroleum export values and volumes. Still, the issue remains of the marginality of these costs
with respect to Persian Gulf oil supply levels, the level of U.S. oil imports, or U.S. oil
consumption levels.
Delucchi and Murphy (2008) seek to deduct from the cost of Persian Gulf military
programs the costs associated with defending U.S. interests other than the objective of providing
more stable oil supply and price to the U.S. economy.573 Excluding an estimate of cost for
missions unrelated to oil, and for the protection of oil in the interest of other countries, Delucci
and Murphy estimated military costs for all U.S. domestic oil interests of between $24-74 billion
per year. Delucchi and Murphy assume that military costs from oil import reductions can be
scaled proportionally, attempting to address the incremental issue.
Crane et al. considers force reductions and cost savings that could be achieved if oil
security were no longer a consideration. Taking two approaches and guided by post-Cold War
force draw downs and by a top-down look at the current U.S. allocation of defense resources,
they concluded that $75-91 billion, or 12-15% of the current U.S. defense budget, could be
reduced.
Finally, an Issue Brief by Securing America's Future Energy (SAFE) (2018) found a
conservative estimate of approximately $81 billion per year spent by the U.S. military protecting
569 Crane, K., Goldthau, A., Toman, M., Light, T., Johnson, S., Nader, A., Rabasa, A. and Dogo, H. 2009. Imported
oil and US national security. RAND, 2009.
570 Koplow, D. and Martin, A. 1998. Fueling Global Warming: Federal Subsidies to Oil in the United States.
Greenpeace, Washington, D.C.
571 Moore, J., Behrens, C. and Blodgett, J. 1997. "Oil Imports: An Overview and Update of Economic and Security
Effects." CRS Environment and Natural Resources Policy Division report 98, no. 1: pp. 1-14.
572 Stern, R. 2010. "United States cost of military force projection in the Persian Gulf, 1976-2007." Energy Policy
38, no. 6. June: 2816-2825.
573 Delucchi, M. and Murphy, J. 2008. "US military expenditures to protect the use of Persian Gulf oil for motor
vehicles." Energy Policy 36, no. 6. June: 2253-2264.
269
-------
global oil supplies.574 This is approximately 16% of the recent U.S. Department of Defense's
budget. Spread out over the 19.8 million barrels of oil consumed daily in the U.S. in 2017, SAFE
concludes that the implicit subsidy for all petroleum consumers is approximately $11.25/bbl of
crude oil, or $0.28/gal. According to SAFE, a more comprehensive estimate suggests the costs
could be greater than $30/bbl, or over $0.70/gal.575
As in the examples above, an incremental analysis can estimate how military costs would
vary if the oil security mission were no longer needed, and many studies stop at this point. It is
substantially more difficult to estimate how military costs would vary if U.S. oil use or imports
were partially reduced, as is projected to be a consequence of this rule. Partial reduction of U.S.
oil use likely diminishes the magnitude of the energy security problem, but there is uncertainty
that supply protection forces and their costs could be scaled down in proportion, and there
remains the associated goal of protecting supply and transit for U.S. allies and other importing
countries, if they do not decrease their petroleum use as well.576 We are unaware of a robust
methodology for assessing the effect on military costs of a partial reduction in U.S. oil use.
Therefore, we are unable to quantify this effect resulting from the projected reduction in U.S. oil
use attributable to this rule.
5.4 Energy Security Impacts
5.4.1 U.S. Oil Import Reductions
From 2023-2025, the AEO 2023 Reference Case projects that the U.S. will be both an
exporter and an importer of crude oil.577 The U.S. produces more light crude oil than its
refineries can refine. Thus the U.S. exports lighter crude oil and imports heavier crude oil to
satisfy the needs of U.S. refineries which are configured to efficiently refine heavy crude oil.
U.S. crude oil exports are projected to be fairly stable at 3.3 MMBD in 2023 and 3.2 MMBD
from 2024-2025. U.S. crude oil imports, meanwhile, are projected to range between 6.8 MMBD
and 7.0 MMBD over the 2023-2025 time period. AEO 2023 also projects that net U.S. exports
of refined petroleum products will increase from 4.2 MMBD in 2023 to 5.5 MMBD in 2025.
Given the pattern of stable net U.S. crude oil imports, and the projected growth in the U.S.'s net
petroleum product exports, the U.S. is projected to increase its net crude oil and refined
petroleum products exports from 0.8 MMBD in 2023 to 1.9 MMBD in 2025.
U.S. oil consumption is estimated to have decreased from 19.8 MMBD in 2019 to 17.5
MMBD in 2020 and 19.1 MMBD in 2021 as a result of social distancing and quarantines that
limited personal mobility as a result of the COVID-19 pandemic.578 U.S. oil consumption is
projected to decrease modestly from 19.2 MMBD in 2023 to 18.6 MMBD in 2025.579 It is not
just U.S. crude oil imports alone, but both imports and consumption of petroleum from all
574 Securing America's Future Energy. 2018. Issue Brief. The Military Cost of Defending the Global Oil Supply.
575 Id.
576 Crane, K., Goldthau, A., Toman, M., Light, T., Johnson, S., Nader, A., Rabasa, A. and Dogo, H. 2009. Imported
oil and US national security. 2009. RAND.
577 EIA. AEO 2023. Reference Case. Table A11. Petroleum and Other Liquids Supply and Disposition.
578 EIA. Monthly Energy Review, March 2023. Calculated using series "Petroleum Consumption (Excluding
Biofuels) Annual" (Table 1.3) and "Petroleum Consumption Total Heat Content Annual" (Table A3).
579 EIA. AEO 2023. Reference Case. Table A11. Petroleum and Other Liquids Supply and Disposition.
270
-------
sources and their role in economic activity, that exposes the U.S. to risk from price shocks in the
world oil price. In 2023-2025, the U.S. is projected to continue to consume significant quantities
of oil and to rely on significant quantities of crude oil imports. As a result, U.S. oil markets are
expected to remain tightly linked to trends in the world crude oil market.
In Chapter 10.4.2.1, we estimate how increased consumption of renewable fuels caused
by this rule reduces U.S. refined product consumption (i.e., gasoline and diesel fuel). For this
energy security analysis, we undertake a detailed analysis of how the reduction in U.S. refined
product fuel consumption impacts U.S. imports/exports of crude oil and petroleum products in
the 2023-2025 timeframe. The impact of lower refined product demand on imports/exports is
estimated by comparing the AEO 2023 Low Economic Growth to the AEO 2023 Reference
Case. The Low Economic Growth Case is used since refined product demand decreases in
comparison to the Reference Case, and we assume that this reduction would be similar to the
reduction caused by increased consumption of renewable fuels.580 An oil import reduction factor
is calculated by taking the ratio of the changes in U.S. net crude oil and refined petroleum
product imports divided by the change in U.S. refined product consumption in the two different
AEO cases considered. Based on this analysis, we project that approximately 100% of the
change in refined product consumption resulting from this rule is likely to be reflected in reduced
aggregate net U.S. crude oil imports/petroleum product imports in 2023-2025. Thus, on balance,
each gallon of petroleum product reduced as a result of this rule is anticipated to reduce total net
U.S. imports of crude oil/petroleum refined products by one, energy-equivalent gallon.
Based on the changes in oil consumption estimated by EPA and the 100% oil import
reduction factor, the reductions in U.S. oil imports in 2023-2025 as a result of this rule are
estimated in Table 5.4.1-1. Included in this table are estimates of U.S. crude oil exports and
imports, net oil refined product exports, net crude oil and refined petroleum product exports, and
U.S. oil consumption for 2023-2025 based on the AEO 2023 Reference Case.581
580 We analyze how U.S. crude oil imports/exports and net petroleum products in Table 11. Petroleum and Other
Liquids Supply and Disposition of the AEO 2023, the Low Economic Growth Case, changes in comparison to that
in the Reference Case. See the spreadsheet in the Docket, "Change of product demand on imports AEO 2023.xlsx".
581 EIA. AEO 2023. Reference Case. Table A11. Petroleum and Other Liquids Supply and Disposition.
271
-------
Table 5.4.1-1: Projected Trends in U.S. Oil Exports/Imports, Net Oil Refined Product
Exports, Net Crude Oil and Refined Petroleum Product Exports, U.S. Oil Consumption
and Reductions in U.S. Oil imports Resulting From the Candidate Volumes (2023-2025)
(MMBD)
2023
2024
2025
U.S. Crude Oil Exports
3.3
3.2
3.2
U.S. Crude Oil Imports
6.8
7.0
6.9
U.S. Net Petroleum Refined Product Exports21
4.2
5.4
5.5
U.S. Net Crude Oil and Refined Petroleum
0.8
1.7
1.9
Product Exports'3
U.S. Oil Consumption0
19.2
18.7
18.6
Reduction in U.S. Oil Imports from the
Candidate Volumes
Excluding 2023 Supplemental Standard
Including 2023 Supplemental Standard
0.13
0.14
0.13
0.13
0.14
0.14
a Calculated from AEO 2023 Table A11 as Net Product Exports minus Ethanol, Biodiesel, Renewable Diesel, and
Other Biomass-derived Liquid Net Exports.
b Calculated from AEO 2023 Table A11 as Total Net Exports minus Ethanol, Biodiesel, Renewable Diesel, and
Other Biomass-derived Liquid Net Exports.
0 Calculated from AEO 2023 Table A11 as "Total Primary Supply" minus "Biofuels."
5.4.2 Oil Import Premiums Used for This Rule
In order to understand the energy security implications of reducing U.S. oil imports, EPA
has worked with ORNL, which has developed approaches for evaluating the social costs and
energy security implications of oil use. The energy security estimates provided below are based
upon a methodology developed in a peer-reviewed 2008 ORNL study.582 This ORNL study is an
updated version of the approach used for estimating the energy security benefits of U.S. oil
import reductions developed in a 1997 ORNL Report.583 This same approach was used to
estimate energy security benefits for the RFS2 final rule.584 ORNL has updated this methodology
periodically for EPA to account for updated projections of future energy market and economic
trends reported in the EIA's AEO. For this rule, EPA updated the ORNL methodology using the
AEO 2023.
The ORNL methodology is used to compute the oil import premium per barrel of
imported oil.585 The values of U.S. oil import premium components (macroeconomic
disruption/adjustment costs and monopsony components) are numerically estimated with a
compact model of the oil market by performing simulations of market outcomes using
probabilistic distributions for the occurrence of oil supply shocks, calculating marginal changes
in economic welfare with respect to changes in U.S. oil import levels in each of the simulations,
and summarizing the results from the individual simulations into a mean and 90% confidence
582 Leiby, P. 2008. Estimating the Energy Security Benefits of Reduced U.S. Oil Imports, Final Report, ORNL/TM-
2007/028, Oak Ridge National Laboratory. March.
583 Leiby, P., Jones, D., Curlee, R. and Lee, R. 1997. Oil Imports: An Assessment of Benefits and Costs, ORNL-
6851, Oak Ridge National Laboratory, November.
584 75 FR 14839-42, March 26, 2010.
585 The oil import premium concept is defined in Chapter 5.1.
272
-------
intervals for the import premium. The macroeconomic disruption/adjustment import cost
component is the sum of two parts: the marginal change in expected import costs during
disruption events and the marginal change in GDP due to the disruption. The monopsony
component is the long-run change in U.S. import costs as the level of oil import changes.
For this rule, we are using oil import premiums that incorporate the oil price projections
and energy market and economic trends, particularly global regional oil supplies and demands
(i.e., the U.S./OPEC/rest of the world), from AEO 2023 into its model.586 We only consider the
avoided macroeconomic disruption/adjustment oil import premiums (i.e., labeled
macroeconomic oil security premiums below) as costs, since the monopsony impacts stemming
from changes in renewable fuel volumes are considered transfer payments. In previous EPA
rules when the U.S. was projected by EIA to be a net importer of crude oil and petroleum-based
products, monopsony impacts represented reduced payments by U.S. consumers to oil producers
outside of the U.S. There was some debate among economists as to whether the U.S. exercise of
its monopsony power in oil markets (e.g., from the implementation of EPA's rules) was a
"transfer payment" or a "benefit." Given the redistributive nature of this monopsony impact from
a global perspective, and since there are no changes in resource costs when the U.S. exercises it
monopsony power, some economists argued that it is a transfer payment. Other economists
argued that monopsony impacts were a benefit since they partially address, and partially offset,
the market power of OPEC. In previous EPA rules, after weighing both countervailing
arguments, EPA concluded that the U.S.'s exercise of its monopsony power was a transfer
payment, and not a benefit.587
In the context of this rule, the U.S.'s oil trade balance is quite a bit different than in many
previous RFS rules. The U.S. is projected to be a net exporter of oil and petroleum-based
products in 2023-2025. As a result, reductions in U.S. oil consumption and, in turn, U.S. oil
imports, still lower the world oil price modestly. But the net effect of the lower world oil price is
now a decrease in revenue for U.S. exporters of crude oil and petroleum-based products, instead
of a decrease in payments to foreign oil producers. The argument that monopsony impacts
address the market power of OPEC is no longer appropriate. Thus, we continue to consider the
U.S. exercise of monopsony power to be transfer payments. We also do not consider the effect of
this rule on the costs associated with existing energy security policies (e.g., maintaining the SPR
or strategic military deployments), which are discussed in Chapter 5.3.
586 The oil market projection data used for the calculation of the oil import premiums came from AEO 2023,
supplemented by the latest complete EIA international projections from the Annual Energy Outlook
(AEO)/International Energy Outlook (IEO) 2021. Projections for global oil prices, U.S. GDP and all variables
describing U.S. supply and disposition of petroleum liquids (domestic supply, tight oil supply fraction, imports,
demands) as well as U.S. non-petroleum liquids supply and demand are from AEO 2023. Global and OECD Europe
supply/demand projections as well as OPEC oil production share are from IEO 2021. The need to combine AEO
2023 and IEO 2021 data arises due to two reasons: (a) EIA stopped including Table 21 "International Petroleum and
Other Liquids Supply, Disposition, and Prices" (international oil market balances) in the U.S.-focused Annual
Energy Outlook after 2019, (b) EIA does not publish complete updates of the IEO every year and IEO 2023 is not
due out until later in 2023.
587 We also discuss monopsony oil import premiums in previous EPA GHG vehicle rules. See, e.g., Section 3.2.5,
Oil Security Premiums Used for this Rule, RIA, Revised 2023 and Later Model Year Light Duty Vehicle GHG
Emissions Standards, December 2021, EPA-420-F-21-077.
273
-------
The macroeconomic oil security premiums arise from the effect of U.S. oil imports on the
expected cost of supply disruptions and accompanying price increases. A sudden increase in oil
prices triggered by a disruption in world oil supplies has two main effects: (1) it increases the
costs of oil imports in the short-run, and (2) it can lead to macroeconomic contraction,
dislocation, and GDP losses. Since future disruptions in foreign oil supplies are an uncertain
prospect, each of the disruption cost components must be weighted by the probability that the
supply of petroleum to the U.S. will actually be disrupted. Thus, the "expected value" of these
costs—the product of the probability that a supply disruption will occur and the sum of costs
from reduced economic output and the economy's abrupt adjustment to sharply higher petroleum
prices—is the relevant measure of their magnitude.
In addition, EPA and ORNL have worked together to revise the oil import premiums
based upon recent energy security literature. Based on EPA and ORNL's review of the recent
energy security literature, EPA is updating its macroeconomic oil security premiums for this
rule. The recent economics literature (discussed in Chapter 5.2) focuses on three factors that can
influence the macroeconomic oil security premiums: price elasticity of oil demand, GDP
elasticity in response to oil price shocks, and the impacts of the shale oil boom. We discuss each
factor below and provide a rationale for how we are updating the first two factors to develop new
estimates of the macroeconomic oil security premiums. We are not accounting for how U.S. tight
oil is influencing the macroeconomic oil security premiums in this rule, other than how it
significantly reduces the need for net U.S. oil imports.
First, we assess the price elasticity of demand for oil. In RFS rules prior to the 2020-2022
annual rule, EPA used a short-run elasticity of demand for oil of-0.045.588 From the recent RFF
study, the "blended" price elasticity of demand for oil is -0.05. The ORNL meta-analysis
estimate of this parameter is -0.07. We find the elasticity estimates from what RFF characterizes
as the "new literature," -0.175, and from the "new models" that RFF uses, -0.20 to -0.33,
somewhat high. Most of the world's oil demand is concentrated in the transportation sector and
there are limited alternatives to oil use in this sector, particularly in the 2023-2025 time frame of
this final rule. According to IEA, the share of global oil consumption attributed to the
transportation sector grew from 60% in 2000 to 66% in 2019.589 The next largest sector by oil
consumption, and an area of recent growth, is petrochemicals. There are limited alternatives to
oil use in this sector also, particularly in the 2023-2025 time frame. Thus, we believe it would be
surprising if short-run oil demand responsiveness has changed in a dramatic fashion.
The ORNL meta-analysis estimate encompasses the full range of the economics literature
on this topic and develops a meta-analysis estimate from the results of many different studies in a
structured way, while the RFF study's "new models" results represent only a small subset of the
economics literature's estimates. Thus, for the analysis of this rule, and consistent with the 2020-
2022 annual rule, we are increasing the short-run price elasticity of demand for oil from -0.045
588 See 75 FR 26049 (May 10, 2010).
589 IEA, Data and Statistics, https://www.iea.org/data-and-
statistics?countrv=WORLD&fuel=Oil&indicator=OiLProductsConsBv Sector
274
-------
to -0.07, a 56% increase.590 This increase has the effect of lowering the macroeconomic oil
security premium estimates undertaken by ORNL for EPA.
Second, we consider the elasticity of GDP to an oil price shock. In RFS rules prior to the
2020-2022 annual rule, a GDP elasticity to an oil shock of-0.032 was used.591 The RFF
"blended" GDP elasticity is -0.028, the RFF's "new literature" GDP elasticity is -0.018, while
the RFF "new models" GDP elasticities range from -0.007 to -0.027. The ORNL meta-analysis
GDP elasticity is -0.021. We believe that the ORNL meta-analysis value is representative of the
recent literature on this topic since it considers a wider range of recent studies and does so in a
structured way. Also, the ORNL meta-analysis estimate is within the range of GDP elasticities of
RFF's "blended" and "new literature" elasticities. For this rule and consistent with the 2020-
2022 annual rule, EPA is using a GDP elasticity of-0.021, a 34% reduction from the GDP
elasticity used previously (i.e., the -0.032 value).592 This GDP elasticity is within the range of
RFF's "new literature" elasticity, -0.018, and the elasticity EPA has used in previous rules, -
0.032, but lower than RFF's "blended" GDP elasticity, -0.028. This decrease has the effect of
lowering the macroeconomic oil security premium estimates. For U.S. tight oil, EPA has not
made any adjustments to the ORNL model, given the limited tight oil production response to
rising world oil prices in 2020 and 2021.593 Increased tight oil production still results in energy
security benefits though, through its impact of reducing U.S. oil imports in the ORNL model.
Table 5.4.2-1 provides EPA's estimates of the macroeconomic oil security prem for
2023-2025, showing that they are relatively steady over this time period.
Table 5.4.2-1: Estimated Macroeconomic Oil Security Premiums (2022$/bbl)
Year
Avoided Macroeconomic
Disruption/Adjustment Costs
(Range)
2023
$3.75
($0.86-$6.81)
2024
$3.70
($0.69-$6.87)
2025
$3.67
($0.65-$6.87)
a Top values in each cell are mean values. Values in parentheses are 95% confidence intervals.
We note that the quantified energy security benefits of this rule, while significant, are
dwarfed by the quantified costs discussed in Chapter 10, which are more than an order of
magnitude greater. Even if we were to use the lowest or highest end of the range for oil security
premiums in Table 5.4.2-1, that would continue to be the case: significant quantified energy
security benefits are far smaller than the quantified costs. In all cases, we would reach the same
590 EPA and ORNL worked together to develop an updated estimate of the short-run elasticity of demand for oil for
use in the ORNL model.
591 See 75 FR 26049 (May 10, 2010).
592 EPA and ORNL worked together to develop an updated estimate of the GDP elasticity to an oil shock for use in
the ORNL model. This slightly different value also was produced by an earlier draft of the ORNL meta-analysis.
593 The short-run oil supply elasticity assumed in the ORNL model is 0.06 and is applied to production from both
conventional and shale oil wells.
275
-------
conclusions as we factor in quantified benefits and costs with regard to the candidate volumes in
this rule.
5.4.3 Energy Security Benefits
Estimates of the total annual energy security benefits of the candidate volumes are based
on the ORNL oil import premium methodology with updated oil import premium estimates
reflecting the recent energy security literature and using AEO 2023. Annual per-gallon benefits
are applied to the reductions in U.S. crude oil and refined petroleum product imports shown in
Table 5.4.3-1. We do not consider military cost impacts or the monopsony effect of U.S. crude
oil and refined petroleum product import changes. The energy security benefits are presented in
Table 5.4.3-1.
Table 5.4.3-1: Annual Energy Security Benefits of the Candidate Volumes
Net Crude Oil
Import Reductions3
Benefits
Year
(millions of gallons)
(millions of 2022$)
2023
Excluding Supplemental Standard
2,012
$180
Including Supplemental Standard
2,151
$192
2024
1,960
$173
2025
2,141
$187
a U.S. oil import reductions used for the energy security analysis in this section are a combination of reduced U.S.
imports of gasoline, diesel fuel, and crude oil from Tables 10.4.2.1-3 and 10.4.2.1-4 converted to crude oil-
equivalent gallons.
276
-------
Chapter 6: Rate of Production and Consumption of Renewable Fuel
This chapter discusses the expected annual rate of future commercial production of
renewable fuels, including advanced biofuels in each category (cellulosic biofuel and biomass-
based diesel). For 2023-2025, we project production based on historic data and other relevant
factors. We consider both domestically produced biofuels as well as foreign produced biofuels
that are imported into and available for use in the U.S.594
We also project the use (i.e., consumption) of qualifying renewable fuels in the United
States. While not an explicit factor that we must consider under the statute, consumption is
inherent in the requisite consideration of infrastructure which is addressed in Chapter 7, and in
the cost to consumers of transportation fuel which is addressed in Chapter 10. For 2023-2025,
the projection of consumption is based on our assessment of production, exports and imports,
infrastructure constraints on distributing and using biofuels, costs, and other factors explained
below and throughout this document. Sometimes, we term this overall resulting use of biofuels
as the "supply" of biofuels. In general, we expect that all cellulosic biofuels produced in the U.S.
will be used here as they have been historically. By contrast, some quantities of domestically
produced advanced and conventional renewable fuels have historically been exported, and we
expect exports of such fuels to continue through 2025.
We discuss the production and use of each major type of biofuel in turn: cellulosic
biofuel (Chapter 6.1), biomass-based diesel (biodiesel and renewable diesel) (Chapter 6.2),
imported sugarcane ethanol (Chapter 6.3), other advanced biofuels (besides ethanol, biodiesel,
and renewable diesel) (Chapter 6.4), total ethanol (Chapter 6.5), corn ethanol (Chapter 6.6), and
conventional biodiesel and renewable diesel (Chapter 6.7).
6.1 Cellulosic Biofuel
In the past several years, production of cellulosic biofuel has continued to increase.
Cellulosic biofuel production reached record levels in 2022, driven by CNG and LNG derived
from biogas.595 Production of liquid cellulosic biofuel has remained limited in recent years (see
Figure 6.1-1). This section describes our assessment of the rate of production of qualifying
cellulosic biofuel in 2023-2025 and some of the uncertainties associated with the projected
volume for these years. These assessments address our obligation to analyze the rate of
production of renewable fuel in these years under our reset authority, CAA section
21 l(o)(2)(B)(ii)(III).
594 This is what we generally mean when we use the term biofuel "production" in this section and do not specify
whether we are discussing domestic production or imports.
595 The majority of the cellulosic RINs generated for CNG/LNG are sourced from biogas from landfills; however,
the biogas may come from a variety of sources including municipal wastewater treatment facility digesters,
agricultural digesters, separated municipal solid waste (MSW) digesters, and the cellulosic components of biomass
processed in other waste digesters.
277
-------
Figure 6.1-1: Cellulosic RINs Generated (2013-2022)
700
600
'iHl
2013 2014 2015 2016 2017 201S 2019 2020 2021 2022
¦ CNG/LNG Derived from Biogas ¦ Liquid Cellulosic Biofuels
To project the volume of cellulosic biofuel production in 2023-2025, we considered
numerous factors, including the accuracy of the methodologies used to project cellulosic biofuel
production in previous years, data reported to EPA through EMTS, available cellulosic
feedstocks, projected use of CNG, LNG as transportation fuel, and information we collected
through meetings with representatives of facilities that have produced qualifying volumes of
cellulosic biofuel in recent years or have the potential to produce qualifying volumes of
cellulosic biofuel by 2025.
To project potential production volumes of liquid cellulosic biofuel for 2023-2025 we
considered the same factors as in previous RFS rules. Based on the available information we
project that ethanol produced from corn kernel fiber will be the only type of liquid cellulosic
biofuel produced in 2023 - 2025. To project the production of cellulosic biofuel RINs for RNG
used as CNG/LNG, we used the same marketwide year-over-year growth rate methodology as in
the 2018-2022 final rules, with updated RIN generation data through March 2023. This
methodology reflects the mature status of this industry, the large number of facilities registered
to generate cellulosic biofuel RINs from these fuels, and EPA's continued attempts to refine its
methodology to yield estimates that are as accurate as possible. However, the rate of growth used
to project the production of RNG used as CNG/LNG is based on data from 2015 - 2022 rather
than data from the previous 24 months. Given the many years of steady growth with the primary
exception of the 2020 when all markets were impacts significantly by Covid-19, this longer time
period was believed to provide a more accurate estimate of potential growth in 2023-2025.
The balance of this section is organized as follows: Chapter 6.1.1 discusses our current
cellulosic biofuel industry assessment, including a review of the accuracy of EPA's projections
in prior years and the companies EPA assessed in the process of projecting qualifying cellulosic
biofuel production in the U.S. Chapters 6.1.2 and 6.1.3 discuss the methodologies used by EPA
to project cellulosic biofuel production for liquid cellulosic biofuels and RNG used as
CNG/LNG. Chapter 6.1.4 summarizes the projected rate of production and import of cellulosic
biofuel volume for 2023-2025.
278
-------
6.1.1
Cellulosic Biofuel Industry Assessment
In this section, we first explain our general approach to assessing facilities or groups of
facilities (which we collectively refer to as "facilities") that we believe are likely to generate
qualifying RINs for cellulosic biofuel in 2023-2025. We then review the accuracy of EPA's
projections in prior years. Next, we discuss the criteria used to determine whether to include
potential domestic and foreign sources of cellulosic biofuel in our projection. Finally, we provide
a summary table of all facilities that we expect to produce cellulosic biofuel by the end of 2025.
To project the rate of cellulosic biofuel production for 2023-2025, we have tracked the
progress of a number of potential cellulosic biofuel production facilities, located both in the U.S.
and in foreign countries. We considered a number of factors, including information from EMTS,
the registration status of potential biofuel production facilities as cellulosic biofuel producers in
the RFS program, publicly available information (including press releases and news reports),
information provided by representatives of potential cellulosic biofuel producers, and comments
received. As discussed in greater detail in Chapter 6.1.2 through 6.1.3, our projection of liquid
cellulosic biofuel is based on a facility-by-facility assessment of each of the likely sources of
cellulosic biofuel in 2023-2025, while our projections of RNGused as CNG/LNG is based on an
industry-wide assessment. To make a determination of which facilities are most likely to produce
liquid cellulosic biofuel and generate cellulosic biofuel RINs by the end of 2025, each potential
producer of liquid cellulosic biofuel was investigated further to determine the current status of its
facilities and its likely cellulosic biofuel production and RIN generation volumes. Both in our
discussions with representatives of individual companies and as part of our internal evaluation
process, we gathered and analyzed information including, but not limited to, the funding status of
these facilities, current status of the production technologies, anticipated construction and
production ramp-up periods, facility registration status, and annual fuel production and RIN
generation targets.
6.1.1.1 Review of EPA's Projection of Cellulosic Biofuel in Previous Years
As an initial matter, it is useful to review the accuracy of EPA's past cellulosic biofuel
projections. The record of actual cellulosic biofuel production, including both cellulosic biofuel
(which generate D3 RINs) and cellulosic diesel (which generate D7 RINs), and EPA's projected
production volumes from 2015-2022596 are shown in Table 6.1.1.1-1. These data indicate that
EPA's projection was lower than the actual number of cellulosic RINs made available in 2015,
2018, and 2022597 and higher than the actual number of RINs made available in 2016, 2017,
2019, and 2020.598 The fact that the projections made using this methodology have been
somewhat inaccurate, under-estimating the actual number of RINs made available in some years
and over-estimating in other years, reflects the inherent difficulty with projecting cellulosic
596 2022 is the last year for which complete data is available at the time of this action.
597 EPA only projected cellulosic biofuel production for the final three months of 2015, since data on the availability
of cellulosic biofuel RINs (D3+D7) for the first nine months of the year were available at the time the analyses were
completed for the final rule.
598 2021 values were set at the actuals after the fact, see 87 FR 39600 (July 1, 2022).
279
-------
biofuel production. It also emphasizes the importance of continuing to consider refinements to
our projection methodology in order to make our projections more accurate.
Table 6.1.1.1-1: Projected and Actual Cellulosic Biofuel Production (2015-2022) (million
gallons)
Projected Volume3
Actua
Production Volumeb
Liquid
Total
Liquid
Total
Cellulosic
RNG used as
Cellulosic
Cellulosic
RNG used as
Cellulosic
Year
Biofuel
CNG/LNG
Biofuel0
Biofuel
CNG/LNG
Biofuel0
2015d
2
33
35
0.5
52.8
53.3
2016
23
207
230
4.1
186.2
190.3
2017
13
298
311
11.7
239.4
251.1
2018
14
274
288
10.6
303.9
314.5
2019
20
399
418
11.1
402.8
413.9
2020
16
577
593
2.1
502.5
504.6
2021
N/A
N/A
N/A
0.7
561.8
562.5
2022
0
632
632
1.6
665.1
666.7
a Projected volumes for 2015 and 2016 can be found in the 2014-2016 Final Rule (80 FR 77506, 77508, December
14, 2015); projected volumes for 2017 can be found in the 2017 Final Rule (81 FR 89760, December 12, 2016);
projected volumes for 2018 can be found in the 2018 Final Rule (82 FR 58503, December 12, 2017); projected
volumes for 2019 can be found in the 2019 Final Rule (83 FR 63704, December 11, 2018); projected volumes for
2020 can be found in the 2020 Final Rule (85 FR 7016, February 6, 2020); projected volumes for 2022 can be found
in the 2020 - 2022 Final Rule (87 FR 39600, July 1, 2022).
b Actual production volumes are the total number of RINs generated minus the number of RINs retired for reasons
other than compliance with the annual standards, based on EMTS data.
0 Total cellulosic biofuel may not be precisely equal to the sum of liquid cellulosic biofuel and RNG used as
CNG/LNG due to rounding.
d Projected and actual volumes for 2015 represent only the final 3 months of 2015 (October-December) as EPA
used actual RIN generation data for the first 9 months of the year.
EPA's projections of liquid cellulosic biofuel were higher than the actual volume of
liquid cellulosic biofuel produced each year from 2015 to 2020. In an effort to take into account
the most recent data available and make the liquid cellulosic biofuel projections more accurate,
EPA adjusted our methodology in the 2018 final rule following the over-projections in 2015-
2016 (and anticipated over-projection in 20 1 7).599 Despite these adjustments, EPA continued to
over-project the volume of liquid cellulosic biofuel in each year from 2018 through 2020. 2020,
however, was a challenging year for the entire industry due to the impacts of COVID-19, which
was an unforeseen event that EPA could not have accounted for in projecting the volume. For the
first time since 2015 EPA under-projected liquid cellulosic biofuel volumes in 2022 using the
methodology first adopted in the 2018 final rule.
We next turn to the projection of RNG used as CNG/LNG. For 2018-2022, EPA used an
industry-wide approach, rather than an approach that projects volumes for individual companies
or facilities, to project the production of RNG used as CNG/LNG. EPA used a facility-by-facility
approach to project the production of CNG/LNG derived from biogas from 2015-2017. Notably
the facility-by-facility methodology resulted in significant over-estimates of CNG/LNG
599 82 FR 58486 (December 12, 2017).
280
-------
production in 2016 and 2017, leading EPA to develop the alternative industry wide projection
methodology first used in 2018. This updated approach reflects the fact that this industry is far
more mature than the liquid cellulosic biofuel industry, with a far greater number of potential
producers of RNG used as CNG/LNG. In such cases, industry-wide projection methods can be
more accurate than a facility-by-facility approach, especially as macro market and economic
factors become more influential on total production than the success or challenges at any single
facility. The industry-wide projection methodology slightly under-projected the production of
RNG used as CNG/LNG in 2018, 2019, and 2022 but over-projected the production of these
fuels in 2020. The accuracy of the 2020 projection, however, may have been influenced by the
unforeseen and significant impacts of COVID-19.
As further described in Chapter 6.1.3, EPA is again projecting production of RNG used
as CNG/LNG using the industry-wide approach in this final rule. We calculate a year-over-year
rate of growth in the renewable CNG/LNG industry and apply this year-over-year growth rate to
the total number of cellulosic RINs generated and available to be used for compliance with the
annual standards in 2022 to estimate the production of RNG used as CNG/LNG in 2023-
2025.600 In comments on the proposed rule, some parties claimed that the production of RNG
used as CNG/LNG was negatively impacted by the COVID-19 pandemic in 2020-2022, and that
using a growth rate based on data from these years underestimates the potential production of
this fuel in future years. During this time period the production of CNG/LNG continued to grow,
but at lower rate of growth than in previous years. As discussed in further detail below, the
growth rate we are using to project the production of RNG used as CNG/LNG for 2023 - 2025 is
based on data from 2015 - 2022, which we believe is more reflective of the potential for growth
in the production and use of these fuels in 2023 - 2025.
We applied the growth rate to the number of available 2022 RINs generated for RNG
used as CNG/LNG as data from this year allows us to adequately account for not only RIN
generation, but also for RINs retired for reasons other than compliance with the annual standards.
While more recent RIN generation data is available, the retirement of RINs for reasons other
than compliance with the annual standards generally lags RIN generation.
The production volumes of cellulosic biofuel in previous years also highlight that the
production of RNG used as CNG/LNG has been significantly higher than the production of
liquid cellulosic biofuel. This is likely the result of a combination of factors, including the
mature state of the technology used to produce RNG used as CNG/LNG relative to the
technologies used to produce liquid cellulosic biofuel, the relatively low production cost of RNG
used as CNG/LNG (see Chapter 10), and the comparatively high value of the cellulosic RIN.
These factors are unlikely to change in 2023-2025. While we project production volumes of
liquid cellulosic biofuel and RNG used as CNG/LNG separately, ultimately it is overall accuracy
of the combined cellulosic biofuel volume projection that is relevant to obligated parties.
600 To project the volume of CNG/LNG derived from biogas in 2023 - 2025, we multiply (1) the number of 2022
RINs generated for these fuels and available to be used for compliance with the annual standards by (2) the
calculated growth rate to project production of these fuels in 2023. We then multiply the projected volume of
CNG/LNG derived from biogas for 2023 by the growth rate again to project the volume of these fuels for 2024, and
repeat this process for 2025.
281
-------
6.1.1.2 Potential Domestic Producers
There are several companies and facilities located in the U.S. that have either already
begun producing cellulosic biofuel for use as transportation fuel, heating oil, or jet fuel at a
commercial scale,601 or are anticipated to be in a position to do so in 2023-2025. The RFS
program provides a strong financial incentive for domestic cellulosic biofuel producers to sell
any fuel they produce for domestic consumption.602 To date nearly all cellulosic biofuel
produced in the U.S. has been used domestically. This, along with the significant incentives
provided by the high cellulosic RIN prices, gives us a high degree of confidence that cellulosic
biofuel RINs will be generated for all cellulosic biofuel produced by such domestic commercial
scale facilities. To generate RINs, each of these facilities must be registered with EPA under the
RFS program and comply with all the regulatory requirements. This includes using an approved
RIN-generating pathway and verifying that their feedstocks meet the definition of renewable
biomass. Most of the domestic companies and facilities considered in our assessment of potential
cellulosic biofuel producers through 2023-2025 have already successfully completed facility
registration, and have successfully generated RINs.603 The remainder of this section presents a
brief description of each of the domestic companies (or group of companies for cellulosic
CNG/LNG producers, new producers of ethanol from corn kernel fiber) that EPA considered
and/or believes may produce commercial-scale volumes of RIN generating cellulosic biofuel by
the end of 2025. General information on each of these companies or group of companies
considered in our projection of the potentially available volume of cellulosic biofuel in 2023-
2025 is summarized in Table 6.1.1.4-1.
Compressed Natural Gas (CNG) and Liquefied Natural Gas (LNG) Producers
In July 2014 EPA approved, as part of the "Pathways II" rule,604 a new cellulosic biofuel
pathway for CNG and LNG derived from biogas produced at landfills, separated MSW digesters,
municipal wastewater treatment facilities, agricultural digesters, and from the cellulosic
components of biomass processed in other waste digesters. The production potential for this type
of cellulosic biofuel is large and has increased at a rapid pace since 2014 due to the fact that
many U.S.-based entities currently capture or produce biogas. This means that in many cases
both historically and in some cases in future years the construction of new facilities capable of
capturing and/or producing biogas will not be required for facilities to begin generating
cellulosic biofuel (D3) RINs. In many cases, however, new equipment is necessary to upgrade
the biogas that is currently captured or produced to meet pipeline specifications, to compress the
601 For a further discussion of EPA's decision to focus on commercial scale facilities, rather than research and
development and pilot scale facilities, see the 2019 proposed rule (83 FR 32031, July 10, 2018).
602 According to data from EMTS, the average price for a 2021 cellulosic biofuel RINs sold in 2021 was $2.75.
Alternatively, obligated parties can satisfy their cellulosic biofuel obligations by purchasing an advanced (or
biomass-based diesel) RIN and a cellulosic waiver credit. The average price for a 2021 advanced biofuel RINs sold
in 2021 was $1.61 while the price for a 2021 cellulosic waiver credit is $2.23 (EPA-420-B-22-033).
603 Most of the facilities listed in Table 5.1.1.4-1 are registered to produce cellulosic (D3 orD7) RINs with the
exception of several of the producers of CNG/LNG derived from biogas and Red Rock Biofuels. EPA is unaware of
any outstanding issues that would reasonably be expected to prevent these facilities from registering as cellulosic
biofuel producers and producing qualifying cellulosic biofuel in 2023-2025.
604 79 FR 42128, July 18, 2014.
282
-------
gas for injection into a pipeline, and to build a stub line to connect to the natural gas pipeline
system.
Corn Kernel Fiber to Ethanol Technologies
EPA is aware of several companies that have developed or are developing technologies to
enable existing corn ethanol plants to convert the cellulosic components present in the corn
kernel to ethanol. These technologies generally seek to use some combination of pretreatment
and enzymatic hydrolysis to convert the cellulose and hemicellulose present in the corn kernel to
simple sugars, and to then ferment these sugars to produce ethanol. Some of these technologies
are designed to convert the cellulosic components of the corn kernel to sugars and eventually to
ethanol simultaneously with the conversion of the corn kernel starch to ethanol. Other
technologies first convert the starch to ethanol and then separately convert the cellulosic
components remaining in the wet cake co-product of the corn starch ethanol process to sugars
and eventually to ethanol. EPA regulations currently contain a pathway (Pathway K in Table 1 to
40 CFR 80.1426(f)) that would allow ethanol produced in either manner to qualify for cellulosic
biofuel RINs, if all other regulatory requirements are satisfied. In this final rule we are projecting
production of cellulosic ethanol from CKF using both the simultaneous conversion and
sequential conversion technologies. Our projections of ethanol produce from CKF is discussed in
Chapter 6.1.2.
Fulcrum BioEnergy
Fulcrum BioEnergy has developed a technology to convert separated MSW into a
synthetic crude oil using a gasification and Fischer-Tropsch process.605 Fulcrum intends to
transport this synthetic crude oil, which EPA would consider to be a biointermediate, to an
existing petroleum refinery where it would be further processed into transportation fuel. Fulcrum
is currently constructing a facility designed to produce 11 million gallons of synthetic crude oil
in Storey County, Nevada. Construction of this facility started in May 20 1 8.606 In December
2022 Fulcrum announced that this facility had begun producing synthetic crude oil from landfill
waste.607 At this time, however, this facility has not registered as a cellulosic biofuel producer
under the RFS program.
6.1.1.3 Potential Foreign Sources of Cellulosic Biofuel
EPA's projection of cellulosic biofuel production through 2025 also considered cellulosic
biofuel that could be imported into the U.S.—specifically from all currently registered foreign
facilities under the RFS program. Currently, there are several foreign cellulosic biofuel
companies registered with EPA and with the potential to generate RINs for qualifying cellulosic
biofuel in 2025. These include facilities owned and operated by Enerkem, GranBio, and Raizen.
605 Unless otherwise noted, all information in this paragraph from Fulcrum BioEnergy website: Sierra Biofuels
Plant: https://fuicrum-bioenergy.com/facilities.
606 Fulcrum BioEnergy Completes Construction of the Sierra Biofuels Plant: https://www.fulcmm-
bioenergv.com/news-resonrces/fiilcmm-bioenergv-completes-constRiction-of-the-sierra-biofiieis-plant.
607 Fulcrum BioEnergy Successfully Produces First Ever Low-Carbon Fuel from Landfill Waste at its Sierra
BioFuels Plant. Press Release. December 20, 2022.
283
-------
All of these facilities use fuel production pathways that have been approved by EPA for
cellulosic RIN generation provided eligible sources of renewable feedstock are used, the fuel is
used as transportation fuel in the U.S., and other regulatory requirements are satisfied. Given
this, we consider imports from these companies as potential sources of cellulosic biofuel.
Nonetheless, we also note that demand for the cellulosic biofuels they produce is expected to be
high in their own local markets.
By contrast, we believe that cellulosic biofuel imports from foreign facilities not
currently registered to generate cellulosic biofuel RINs are generally highly unlikely through
2025. This is due to the strong demand for cellulosic biofuel in local markets (often driven by
mandates or incentive programs in other countries, such as Canada's recently finalized Clean
Fuels Regulations608) and the time necessary for potential foreign cellulosic biofuel producers to
register under the RFS program and arrange for the importation of cellulosic biofuel to the U.S.
For purposes of our 2023-2025 projection of the rate of production of cellulosic biofuel we have
excluded potential volumes from foreign cellulosic biofuel production facilities that are not
currently registered under the RFS program.
Cellulosic biofuel produced at three foreign facilities (GranBio's and Raizen's Brazilian
facilities, Kerry) have generated cellulosic biofuel RINs for fuel exported to the U.S. in previous
years. Another foreign facility (Enerkem's Canadian facility) has completed the registration
process as a cellulosic biofuel producer. Each of these facilities is described briefly below.
However, based on data available through EMTS no foreign facilities have generated cellulosic
(D3) RINs for imported liquid cellulosic biofuel since March 2019. Therefore, while we have
considered these facilities as potential sources of cellulosic biofuel we are not projecting any
imports of cellulosic biofuel through 2025. All of the potential cellulosic biofuel producers
through 2025 are listed in Table 6.1.1.4-1.
Enerkem
Enerkem has developed a commercial-scale technology capable of converting non-
recyclable waste to a variety of renewable chemicals and fuels, including both methanol and
ethanol.609 After feedstock preparation, Enerkem's feedstocks are gasified to produce a synthetic
gas (or syngas). Enerkem next purifies the syngas and processes it through a catalytic reactor to
convert the syngas into the desired products. Enerkem has developed their proprietary
technology over a period of 10 years before deploying it at commercial scale in Edmonton,
Canada.610 Enerkem's facility in Edmonton is designed to produce up to 13 million gallons of
cellulosic ethanol per year.611 This facility began production of methanol in 2015, with
production switching from methanol to ethanol in 2017.
608 Tuttle, Robert. Canada Releases California-Style Fuel Rules to Cut Emissions. Bloomberg, June 29, 2022.
609 "Technology," Enerkem Website, https://enerkem.com/abont-ns/technology.
610 Id.
611 "Enerkem Alberta Biofuels," Enerkem Website, https://enerkem.com/facilities/enerkem-alberta-biofiieis.
284
-------
Ensyn
Ensyn has developed a technology known as Rapid Thermal Processing (RTP) that
involves the non-catalytic thermal conversion of carbon-based solid feedstocks to liquid
products. This technology is currently being used to produce specialty chemicals and heating
fuels. The renewable fuel oil (RFO) produced using Ensysn's technology can be used for heating
and cooling applications, and Ensyn is currently exploring opportunities to sell biocrude to
petroleum refiners for co-processing with petroleum feedstocks. Ensyn is currently developing
projects in Canada,612 Brazil,613 and the United States614 with the intention of selling heating oil
and/or biocrude into the U.S. market. In November 2022 CastleRock Green Energy announced
plans to build two biofuel production facilities in Washington using Ensyn's production
technology.615
GranBio
GranBio uses a group of technologies to convert cellulosic biomass into ethanol.616
Construction of their first cellulosic ethanol production facility was announced in mid-2012, and
financing was completed in May 2013.617 In September 2014, GranBio announced that its first
cellulosic ethanol facility became operational.618 The facility uses sugarcane straw or bagasse as
a feedstock and produces both ethanol and electricity, depending on market conditions. The
facility is located in Sao Miguel dos Campos, Alagoas, Brazil and originally had a production
capacity of approximately 21.5 million gallons (82 million liters) of ethanol per year.619 Since
2016, GranBio has been implementing several equipment and technology modifications at the
plant, which will result in a production capacity of approximately 15.8 million gallons (60
million liters) of ethanol per year.
Kerry Inc.
Kerry Inc. purchased Ensyn's Renfrew, Ontario facility in December 2019.620 This
facility is designed to produce biocrude from wood residues, and is capable of producing
approximately 4 million gallons of biocrude per year. The biocrude produced from this facility is
primarily used in food ingredients but can also be used as heating oil.
612 "Cote Nord," Ensyn Website, http://www.ensvn.coni/aiiebec.html.
613 "Aracruz Project," Ensyn Website. http://www.ensyn.com/braziLhtml.
614 "Ensyn Maine Project," Ensyn Website, fattp://www.ensyn.com/maine.html.
615 "CastleRock Green Energy to build biofuel facilities in Washington," Biomass Magazine. November 10, 2022.
616 "AVAP," GranBio Website, https://www.granbio.com.br/en/oiir~technology/3vap.
617 Schill, Susanne R. "Financing Complete on Brazil's first commercial 2G Ethanol Plant," Ethanol Producer
Magazine. May 17, 2013.
618 "Our Trajectory." GranBio Website, https://www.granbio.com.br/en/aboiit-iis/oiir-traiectorv.
619 Id.
620 Mulvihill, Jonathon. "Fifth Liquid Smoke Plant: International Food Group Purchases Renfrew's Ensyn Facility."
Inside Ottawa Valley. December 19, 2019. Available Online.
285
-------
Raizen
Raizen, a joint venture between Shell and Cosan, uses a technology developed by Iogen
Energy to convert sugarcane bagasse into ethanol. Raizen has constructed a facility co-located
with a first generation ethanol production facility in Piracicaba/SP Brazil designed to be capable
of producing approximately 10.5 million gallons of ethanol a year from biomass residues from
first generation sugarcane ethanol production.621 Construction of this facility began in November
2013, and the first phase, allowing for the conversion of C6 sugars into ethanol, was completed
in July 20 1 5.622 Further construction allowing for the conversion of both C5 and C6 sugars into
ethanol was completed in May 2016. Raizen began exporting cellulosic ethanol produced at this
facility to the United States in 2017, and has exported a total of 32 million liters of cellulosic
ethanol to the U.S. through the end of 2019.
6.1.1.4 Summary of Potential Sources of Cellulosic Biofuel in 2023-2025
General information on each of the cellulosic biofuel producers (or group of producers,
for producers of RNG used as CNG/LNG, and producers ethanol from CKF) that factored into
our projection of cellulosic biofuel production through 2025 is shown in Table 6.1.1.4-1. This
table includes both facilities that have already generated cellulosic RINs, as well as those that
have not yet generated cellulosic RINs, but may do so by the end of 2025. Note that while we
believe all these facilities have the potential to produce or import cellulosic biofuel by the end of
2025, our projections of cellulosic biofuel production do not include volumes from all of the
listed facilities in all years, as we believe the most likely volume of cellulosic biofuel produced
or imported from some of these facilities is zero.
621 "Raizen Project," Iogen Website, https://www.iogen.ca/raizen-proiect.
622 Id.
286
-------
Table 6.1.1.4-
Potential Producers of Cellulosic Biofuel for U.S. Consumption in 2023-2025
Company
Name
Location
Feedstock
Fuel
Facility
Capacity (Million
Gallons per
Year)623
Construction
Start Date
First
Production624
CNG/LNG
Producers
Various
Biogas
CNG/LNG
Various
Various
Various
Enerkem
Edmonton, AL,
Canada
Separated MSW
Ethanol
10625
2012
September
2017626
Ethanol from
CKF
(registered)
Various
Corn Kernel Fiber
Ethanol
Various
Various
Various
Ethanol from
CKF (new)
Various
Corn Kernel Fiber
Ethanol
Various
Various
Various
Ensyn
Various
Woody Biomass
Heating
Oil, Diesel,
Jet
Various
Various
Various
Fulcrum/
Marathon
Storey County, NV
Separated MSW
Diesel, Jet
Fuel
11
May 2018
December 2022
GranBio
Sao Miguel dos
Campos, Brazil
Sugarcane bagasse
Ethanol
21
Mid 2012
September 2014
Raizen
Piracicaba City,
Brazil
Sugarcane bagasse
Ethanol
11
January 2014
July 2015
623 The Facility Capacity is generally equal to the nameplate capacity provided to EPA by company representatives or found in publicly available information.
Capacities are listed in physical gallons (rather than ethanol-equivalent gallons). If the facility has completed registration and the total permitted capacity is lower
than the nameplate capacity, then this lower volume is used as the facility capacity.
624 Where a quarter is listed for the first production date EPA has assumed production begins in the middle month of the quarter (i.e., August for the 3rd quarter)
for the purposes of projecting volumes.
625 The nameplate capacity of Enerkem's facility is 10 million gallons per year. However, we anticipate that a portion of their feedstock will be non-biogenic
municipal solid waste (MSW). RINs cannot be generated for the portion of the fuel produced from non-biogenic feedstocks. We have taken this into account in
our production projection for this facility (See "May 2023 Liquid Cellulosic Biofuel Projections for 2023 - 2025 CBI").
626 This date reflects the first production of ethanol from this facility. The facility began production of methanol in 2015.
287
-------
6.1.2 Projected Liquid Cellulosic Biofuel Production
For our 2023-2025 liquid cellulosic biofuel projections, we use the same general
approach as we have in projecting these volumes in previous years. After gathering updated
information on potential liquid cellulosic biofuel producers, we determined that the only
facilities likely to produce commercial scale volumes of liquid cellulosic biofuel in 2023 - 2025
were facilities intending to produce ethanol from CKF. In the proposed rule we had also
identified a few facilities that we projected could produce gasoline, jet fuel, and diesel fuel from
cellulosic biomass by 2025. Since the time of our proposal, two of these facilities have
announced project delays and/or significant revisions to their project plans. The nature of these
delays and changed plans are such that we no longer anticipate these facilities will be able to
produce commercial scale volumes of liquid cellulosic biofuel by 2025.627 One facility
considered in our proposed rule announced that they have successfully began producing
biocrude, however this facility has not registered as a cellulosic biofuel producer in the RFS
program at this time.628 We therefore believe there is significant uncertainty as to whether fuel
produced from the biocrude at this facility will generate cellulosic biofuel RINs and we have not
included production from this facility in our projections of liquid cellulosic biofuel production.
In the proposed rule we also identified ethanol produced from CKF as another potential
source of liquid cellulosic biofuel in 2023 - 2025. We noted that a significant issue that must be
resolved to register a facility to generate cellulosic biofuel RINs for ethanol when both corn
starch and corn kernel fiber are processed together is the accurate quantification of the volume of
ethanol produced from cellulosic feedstocks rather than non-cellulosic feedstocks such as starch.
In September 2022 EPA published updated guidance on how to demonstrate that an analytical
method for determining the cellulosic converted fraction of corn kernel fiber co-processed with
starch at a traditional ethanol facility.629 EPA has continued to have substantive discussions with
technology providers intending to use analytical methods consistent with the guidance document
and owners of facilities intending to register as cellulosic biofuel producers using these analytical
methods. We now project that many of these facilities will register as cellulosic biofuel
producers in 2023 - 2025 and will produce commercial scale volumes of cellulosic biofuel
during these years.
To project the volume of ethanol produced from CKF each year from 2023 - 2025 we
have developed a projection methodology that considers the size of the ethanol facilities
expected to register to produce cellulosic ethanol from CKF, the amount of cellulosic ethanol
(vs. starch ethanol) expected to be produced at each facility, and the number of facilities
expected to produce cellulosic ethanol each year. We recognize this is a different projection
methodology than what EPA has historically used to project the production of liquid cellulosic
biofuels. We believe this new projection methodology is appropriate due to the significant
differences between the production of a small volume of cellulosic ethanol at an existing corn
ethanol production facility and the production of liquid cellulosic biofuel using novel
627 May 2023 Liquid Cellulosic Biofuel Projections for 2023 - 2025 CBI.
628 May 2023 Liquid Cellulosic Biofuel Projections for 2023 - 2025 CBI.
629 Guidance on Qualifying an Analytical Method for Determining the Cellulosic Converted Fraction of Corn Kernel
Fiber Co-Processed with Starch. Compliance Division, Office of Transportation and Air Quality, U.S. EPA.
September 2022 (EPA-420-B-22-041).
288
-------
technologies at facilities that are not yet producing fuel. To produce cellulosic ethanol from corn
kernel fiber requires very few operational changes at existing ethanol production facilities, and in
some cases ethanol producers claim they are already producing ethanol from CKF. In these
cases, we would not expect a long ramp-up period to full production of cellulosic biofuel, nor
would we expect the type of construction delays or challenges related to the readiness of the
production technology that has often hindered the production of other liquid cellulosic biofuels.
The first step in projecting the volumes of cellulosic ethanol produced from CKF in 2023
- 2025 is projecting the size of the ethanol production facilities that will produce this fuel. At this
time, we do not have sufficient data that would allow us to identify precisely which ethanol
production facilities will produce ethanol from CKF. In the absence of this information, we have
projected that the total ethanol production capacity (for both ethanol from corn starch and
ethanol from CKF) will be equal to the industry average. As of January 2022 (the most data that
were available at the time of the final rule EIA reported that there were 192 ethanol production
facilities in the U.S. with a total ethanol production capacity of 17.38 billion gallons, with an
average production capacity of approximately 90 million gallons.630
The next step in projecting the production of ethanol from CKF is determining the
portion of total ethanol production from an ethanol facility is from corn starch versus CKF.
Conversations with technology providers intending to use analytical methods consistent with the
guidance document suggest that approximately 1.5% of all ethanol produced at these facilities is
from CKF.631 Taken together, these numbers indicate that an average ethanol production facility
could be capable of producing approximately 1.35 million gallons of ethanol from CKF per year.
Finally, we projected the number of ethanol facilities that would register as CKF
producers each year. To inform these projections we referenced public data from California's
LCFS program. This data lists the facilities registered as producers of ethanol from CKF under
California's LCFS program, and in nearly all cases also lists the technology the facility is using
to produce ethanol from CKF. EPA used this data, together with projections of when various
technology providers may be in a position to register facilities as cellulosic biofuel producers
under the RFS program to project the number of facilities that would register each year.632 We
projected that the facilities that register in 2023 would produce at 25% of their potential capacity
in 2023 (representative of a facility completing the registration process on October 1, 2023, with
some facilities registering earlier and other facilities registering later) since we do not expect any
facilities will register as cellulosic biofuel producers in the first half of 2023. We projected that
the facilities that register in 2024 would produce at 50% of their potential capacity in 2024
(representative of a facility completing the registration process on July 1, 2024, with some
facilities registering earlier and other facilities registering later). Our projections of cellulosic
ethanol production from CKF from new facilities is summarized in Table 6.1.2-1.
630 U.S. Fuel Ethanol Plant Production Capacity. EIA. August 2022.
631 May 2023 Liquid Cellulosic Biofuel Projections for 2023 - 2025 CBI.
632 For more information on how EPA projected the number of facilities that would register as CKF producers each
year see May 2023 Liquid Cellulosic Biofuel Projections for 2023 - 2025 CBI.
289
-------
Table 6.1.2-1: Projected Production of Ethanol from CKF from New Facilities (million
RINs)
New vs.
Number of
Total Ethanol
Potential Cellulosic
Percent of
Total
Existing
Facilities
Production
Ethanol Production
Potential
Cellulosic
Production
Ethanol
Production
2023
New
16
1,440
21.6
25%
5
2024
New
39
3,510
52.7
50%
26
Existing
16
1,440
21.6
100%
22
2025
Existing
55
4,950
74.3
100%
74
In addition to the facilities that we project will produce ethanol from CKF simultaneously
with the production of ethanol from corn starch, there are also two facilities that are currently
registered to produce ethanol from CKF after first converting the starch to ethanol (sequential
production). At this time, we are not aware of any other ethanol facilities that intend to produce
ethanol from CKF using this technology. We have projected cellulosic ethanol production from
these facilities based on the previous production rates achieved at these facilities and production
to date in 2023.633 Projected production from facilities producing ethanol from CKF sequentially
are shown in Table 6.1.2-2. Total projected production of ethanol from CKF each year from
2023 to 2025 are shown in Table 6.1.2-3. We are not projecting the production of any liquid
cellulosic biofuel other than ethanol from CKF through 2025.
Table 6.1.2-2: Projected Production from Facilities Producing Ethanol from CKF
2023
2024
2025
2
3
3
Table 6.1.2-3: Projected Production of Ethanol from CKF (Million RINs)
2023
2024
2025
Simultaneous Conversion
5
48
74
Sequential Conversion
2
3
3
Total
7
51
77
6.1.3 Projected Production of RNG used as CNG/LNG
The incentive created by the cellulosic biofuel RIN has led to rapid growth in renewable
natural gas use as CNG/LNG since 2014 (See Table 6.1.3-1). In light of this incentive, we
believe that renewable natural gas (RNG) used as CNG/LNG can continue to grow under the
influence of the RFS standards through 2025. At the same time, there are several market factors
that we expect could limit the rate of growth in the production of CNG/LNG from biogas in
633 More detail on our projections from these facilities can be found in May 2023 Liquid Cellulosic Biofuel
Projections for 2023 - 2025 CBI.
290
-------
future years that must be taken into consideration when projecting volumes. These market factors
are discussed in the following paragraphs. Several of these factors, however, are related not to
the production of CNG/LNG from biogas, but rather to the consumption of CNG/LNG from
biogas.
Table 6.1.3-1: Cellulosic RIN Generation (Million RINs) and Annual Growth Rate for
RNG used as CNG/LNG
2015
2016
2017
2018
2019
2020
2021
2022
D3 RIN Generation
139.9
188.6
240.6
304.2
404.3
503.8
567.8
666.1
Annual Growth Rate
N/A
34.8%
27.6%
26.4%
32.9%
24.6%
12.7%
17.3%
Currently, a significant volume of biogas is produced at landfills and wastewater
treatment plants across the U.S.634 Some of this biogas is currently being flared or used to
produce electricity onsite. There are also significant opportunities for increasing the production
of biogas from manure and other agricultural residues. Raw biogas from landfills, wastewater
treatment facilities, or agricultural digesters must be upgraded before it can be used as
transportation fuel in CNG/LNG vehicles, either at on-site fueling stations or transported to
fueling stations via the natural gas pipeline network. Biogas that has been upgraded and
distributed via a closed, private distribution system is called "treated biogas" while biogas that
has been upgraded and distributed via the natural gas commercial pipeline system it is referred to
as RNG. Collecting and treating the raw biogas to produce RNG requires a significant capital
investment. While the quantity of biogas produced from potentially qualifying sources exceeds
the quantity of CNG/LNG used as transportation fuel, much of this biogas is not currently being
upgraded to RNG, which is a necessary step to using this biogas in CNG/LNG vehicles.635
Along with the incentives provided by the RFS program for the use of biogas as
CNG/LNG in the transportation sector, state programs can also provide significant incentives.
Since its inception in 2011 California's LCFS program has provided credits for RNG used as
CNG/LNG that is used as transportation fuel in California. Since 2014 when RNG used as
CNG/LNG was determined to qualify as cellulosic biofuel under the RFS program, the quantity
of this fuel used with the incentives of both programs (RFS and California's LCFS) has
increased dramatically. It is likely that this rapid expansion was aided by the ability for RNG
used as CNG/LNG to generate lucrative credits under both programs and displace the fossil
CNG/LNG otherwise being used. As of 2022, however, the LCFS data indicates that the quantity
of fossil CNG/LNG generating credits under the LCFS program had decreased to only
approximately 13 million diesel gallon equivalents.636 This suggests that there is little remaining
ability for new sources of RNG used as CNG/LNG to displace CNG/LNG derived from fossil-
based natural gas in California, however the greater incentives offered for CNG/LNG from
sources with a lower carbon intensity could continue to incentivize the production and collection
634 EPA Landfill Methane Outreach Program Landfill and Project Database, available at
https://www.epa.gov/lmop/lmop-landfiH-and-proiect-database.
635 According to the American Biogas Council, there are currently over 2,200 sites producing biogas in the U.S. (see
Biogas Industry Market Snapshot - American Biogas Council, available in the docket). Approximately 860 of these
sites use the biogas they produce, and of this total 138 facilities generated RINs for CNG/LNG derived from biogas
used as transportation fuel in 2021.
636 LCFS Data Dashboard, https://www.arb.ca.gov/fiiels/lcfs/dashboard/dashboard.htin. For context, in 2021,
approximately 190 million diesel gallon equivalents of bio-CNG/LNG generated credits in the LCFS program.
291
-------
of biogas from these sources. Currently Oregon and Washington have also adopted clean fuels
programs, and there are opportunities for RNG used as CNG/LNG to generate financial
incentives in these state programs, though the size of the CNG/LNG fleets in these states are
much smaller than in California. If additional states adopt programs similar to California's LCFS
or Oregon and Washington's Clean Fuels programs, these other state programs could provide
additional incentives for the increased production and use of RNG used as CNG/LNG.
Growth of RNG used as CNG/LNG could also potentially be limited by the cost
associated with establishing a pipeline interconnect to transport RNG to CNG/LNG fueling
stations. Not all CNG/LNG vehicles will be situated such that they can refuel at the location
where the biogas is produced and upgraded to RNG. Therefore, getting the RNG to CNG/LNG
vehicles requires that it be put into common carrier pipelines. If there are no such pipelines near
the source of the biogas, then it can become cost prohibitive and/or require considerable time to
put in place a stub pipeline to connect to the common carrier pipeline. While constructing
pipelines in such cases is technically possible, the costs of doing so may result in the use of RNG
in applications other than CNG/LNG in the transportation sector.
For 2023-2025, EPA is using the same industry wide projection approach as used for
2018-2022 based on a year-over-year growth rate to project production of RNG used as
CNG/LNG used as transportation fuel.637 However, unlike in the proposed rule and previous
annual rules the rate of growth used to project the production of RNG used as CNG/LNG is
based on data from 2015 - 2022 (See Table 6.1.3-1) rather than data from the previous 24
months. While the nature of the incentives provided by state programs may be changing and
incremental volumes of CNG/LNG will likely be produced at marginally higher costs, we
believe the incentives provided by the RFS program, existing and potentially newly adopted state
programs, and the extension of the investment tax credit to qualified biogas facilities in the
IRA638 are sufficient to support growth in the production and use of CNG/LNG derived from
biogas as the rates observed in previous years. Given the many years of steady growth with the
primary exception of 2020 and 2021 when all markets were still being impacted by the economic
turmoil brought on by the COVID pandemic, we believe projecting volumes using a growth rate
calculated from a longer time period provides a more accurate estimate of potential growth in
2023-2025. The growth rate calculated using this data is 25.0%. The number of RINs generated
for CNG/LNG derived from biogas in 2015 and 2022 from which this growth rate is calculated
are shown in Table 6.1.3-1.
637 Historically RIN generation for CNG/LNG derived from biogas has increased each year. It is possible, however,
that RIN generation for these fuels in the most recent 12 months for which data are available could be lower than the
preceding 12 months. Our methodology accounts for this possibility. In such a case, the calculated rate of growth
would be negative.
638 Inflation Reduction Act Gives a Boost to Biogas Sector. The National Law Review. Volume XII, Number 279.
October 6, 2022.
292
-------
Table 6.1.3-1: Generation of Cellulosic Biofuel RINs for RNG used as CNG/LNG (million
ethanol-equivalent gallons)
RIN Generation (2015)
RIN Generation (2022)
Y ear-Over-Y ear
Increase
139.9
666.1
25.0%
EPA then applied this 25.0% year-over-year growth rate to the total number of 2022
cellulosic RINs generated and available for compliance for CNG/LNG. That is, in this rule, as in
the 2018-2022 final rules, we are multiplying the calculated year-over-year rate of growth by the
volume of CNG/LNG actually supplied in the most recent year for which data is available (in
this case 2022), taking into account actual RIN generation as well as RINs retired for reasons
other than compliance with the annual volume obligations. Since we are establishing volumes for
three future years, we do not have data which would allow for separate rates of growth to project
volumes for 2023-2025. Consequently, we applied the same rate of growth to project the
production of RNG used as CNG/LNG in 2023-2025.
Table 6.1.3-2: 2023-2025 Projection of RNG used as CNG/LNG (million ethanol-equivalent
D3 RINs generated for CNG/LGN derived
from biogas in 2022
666
RINs retired for reasons other than
compliance with annual obligations
1
Net RINs generated in 2022
665
Growth rate
25.0%
Projected production of RNG used as
CNG/LNG in 2023
831
Projected production of RNG used as
CNG/LNG in 2024
1,039
Projected production of RNG used as
CNG/LNG in 2025
1,299
In the proposal, EPA had followed the precedent from previous rules and projected the
rate of growth of RNG used as CNG/LNG based on the prior 24 months of data. This method of
projection had proven fairly reliable. However, in discussions with EPA a number of cellulosic
biogas producers highlighted that the rate of growth observed in 2020 and 2021 was negatively
impacted by relatively low cellulosic RIN prices in 2019 and 2020 and challenges developing
new cellulosic biogas production facilities in 2020 and 2020 related to the COVID pandemic.
These parties argued that the higher growth rates observed in previous years were more reflective
of the potential growth in cellulosic biogas production in 2023-2025.
In considering the appropriate growth rate to use for projecting the production of RNG
used as CNG/LNG we considered the annual growth rates each year from 2015 to 2022 (see
Table 6.1.3-1). In reviewing these data, it is apparent that the observed rate of growth in RIN
generation for RNG used as CNG/LNG was notably lower from 2020 to 2021 than in other
years. While we would also expect that as this industry matures and approaches the quantity of
CNG and LNG used as transportation fuel the rate of growth would decrease even in the absence
293
-------
of other external factors, it is also likely that the COVID pandemic was a significant factor in the
lower observed growth rates observed in 2021 and 2022. At this time we are unable to determine
how much of the decrease in the rate of growth of RNG used as CNG/LNG was due to the
COVID pandemic vs. the maturation of the industry and limitations on the quantity of these fuels
used as transportation fuel (discussed further below). However, given that the rate of increase in
2022 appears to have recovered somewhat, it would appear that the COVID pandemic likely had
a significant impact. Given the steady rate of growth over a long period of time and the apparent
anomaly in 2021, for this final rule we believe using the full period from 2015 through 2022
likely provides a more accurate projection for 2023-2025.
We then compared the resulting projected volumes with the total volume of CNG/LNG
expected to be used as transportation fuel in 2023-2025. We are aware of several estimates for
the quantity of CNG/LNG that will be used as transportation fuel in 2022 that cover a wide range
of projected volume. EIA's 2022 AEO projects that 0.12 trillion cubic feet of natural gas will be
used in the transportation sector in 2023 and 2024 (approximately 1.62 billion ethanol-equivalent
gallons), increasing to 0.13 trillion cubic feet of natural gas in 2025 (approximately 1.75 billion
ethanol-equivalent gallons).639 A paper prepared by Bates White for the Coalition for Renewable
Natural Gas presented an independent assessment of 1.53, 1.55, and 1.58 billion ethanol-
equivalent gallons used in 2023-2025.640
Separately, EPA estimated that approximately 1.36 ethanol-equivalent gallons of
CNG/LNG will be used as transportation fuel in 2022. This estimate is based on the average
throughput at CNG/LNG refueling stations in California and the number of CNG/LNG stations
in operation according to the Alternative Fuels Data Center. The data used to make this
projection is summarized in Table 6.1.3-5. Due to significant variation in the annual increase in
the number of CNG/LNG refueling stations historically we have not used this information to
project the use of CNG/LNG in 2023-2025, however consumption would be expected to
increase as the number of operational refueling stations increases.
Table 6.1.3-5: Projected Consumption of CNG/L]>
G Used as Transportation Fuel in 2022
CNG/LNG used as transportation fuel in
California in 2021
303.2 million ethanol-equivalent gallons
CNG/LNG refueling stations in California in 2021
358 stations
Average annual throughput per station in
California in 2021
0.85 million ethanol-equivalent gallons
CNG/LNG refueling stations in the U.S. in 2022
1611 stations
Projected CNG/LNG used as transportation fuel in
the U.S. in 2022
1.36 billion ethanol-equivalent gallons
These estimates of the consumption of CNG/LNG used as transportation fuel are all
fairly similar, and all are greater than the volume of qualifying RNG used as CNG/LNG
projected to be used in 2023-2025. Thus, the volume of CNG/LNG used as transportation fuel
639 These values are from the projections for Motor Vehicles, Trains, and Ships in Table 13: Natural Gas Supply,
Distribution, and Prices in the 2022 AEO.
640 Renewable Natural Gas: Transportation Demand. Bates White Economic Consulting. February 2, 2022; Updated
April 29, 2022.
294
-------
would not appear to constrain the number of RINs generated for this fuel in these years. We note,
however, that using a higher rate of growth of the production of RNG used as CNG/LNG could
exceed these estimates. Even if the production of RNG in 2023-2025 can grow at a higher rate,
RIN generation in these years may could still be limited to the quantity of CNG/LNG used as
transportation fuel. This may become a more significant constraint in years after 2025 as the
volume of RNG used as CNG/LNG approaches the total volume of CNG/LNG used as
transportation fuel.
We believe that projecting the production of RNG used as CNG/LNG using the same
industry-wide methodology as in recent years but using the data over the full period from 2015-
2022 to calculate the growth rate appropriately takes into consideration the significant incentive
created by the RFS standards as well as the real world constraints on growth. We believe that
applying the resulting 25% year-on-year growth rate takes into both the potential for future
growth and the challenges associated with increasing RIN generation from these fuels for 2023-
2025. This methodology may not be appropriate to use in the future as the projected volume of
RNG used as CNG/LNG approaches the total volume of CNG/LNG that is used as transportation
fuel, as RINs can be generated only for CNG/LNG used as transportation fuel. We do not believe
that this is yet a constraint as our projection through 2025 as the volume of RNG used as
CNG/LNG is still below the total volume of CNG/LNG that is currently used as transportation
fuel.
6.1.4 Projected Rate of Cellulosic Biofuel Production for 2023-2025
After projecting production of cellulosic biofuel from liquid cellulosic biofuels and
CNG/LNG derived from biogas, EPA combined these projections to project total cellulosic
biofuel production for 2023-2025. These projections are shown in Table 6.1.4-1. Using the
methodologies described in this section, we project that 0.84 billion ethanol-equivalent gallons of
qualifying cellulosic biofuel will be produced in 2023, 1.09 billion ethanol-equivalent gallons
will be produced in 2024, and 1.38 billion ethanol-equivalent gallons will be produced in 2025.
295
-------
Table 6.1.4-1: Projected Volume of Cellulosic Biofuel in 2023-2025
Projected Volume in 2023 (million
ethanol-equivalent gallons)
Projected Volume
Liquid Cellulosic Biofuel
7
CNG/LNG Derived from Biogas
831
Totala
840
Projected Volume in 2024 (million
ethanol-equivalent gallons)
Projected Volume3
Liquid Cellulosic Biofuel
51
CNG/LNG Derived from Biogas
1,039
Total11
1,090
Projected Volume in 2025 (million
ethanol-equivalent gallons)
Projected Volume3
Liquid Cellulosic Biofuel
77
CNG/LNG Derived from Biogas
1,299
Total11
1,380
a Rounded to the nearest 10 million gallons.
6.2 Biomass-Based Diesel
Since 2010 when the biomass-based diesel (BBD) volume requirement was added to the
RFS program, production of BBD has generally increased. The volume of BBD supplied in any
given year is influenced by a number of factors including production capacity, feedstock
availability and cost, available incentives, the availability of imported BBD, the demand for
BBD in foreign markets, and other economic factors. From 2010 through 2015 the vast majority
of BBD supplied to the U.S. was biodiesel. While biodiesel is still the largest source of BBD
supplied to the U.S. since 2015, increasing volumes of renewable diesel have also been supplied.
Production and import of renewable diesel are expected to continue to increase in future years.
There are also very small volumes of renewable jet fuel and heating oil that qualify as BBD,
however as the vast majority of BBD is biodiesel and renewable diesel we have focused on these
fuels in this section.
This section presents information on a number of factors we consider in projecting the
domestic production and net imports of BBD in 2023 - 2025. First, we present the available data
on biodiesel and renewable diesel production, import, and use in previous years (Chapter 6.2.1).
Next, we assess the current and projected future production capacity for biodiesel and renewable
diesel (Chapter 6.2.2), followed by the availability of qualifying feedstocks for biodiesel and
renewable diesel production (Chapter 6.2.3). Potential imports and exports of BBD (Chapter
6.2.4), and an analysis of the available data on BBD production, imports, and exports in 2023
(Chapter 6.2.5) are in the following sections. Finally, we describe our assessment of the rate of
production and use of qualifying biomass-based diesel biofuel in 2023-2025 based on this
information (Chapter 6.2.6), and discuss some of the uncertainties associated with those
volumes.
296
-------
6.2.1
Production and Use of Biomass-Based Diesel in Previous Years
As a first step in considering the rates of production and use of BBD in future years we
review the volumes of BBD produced domestically, imported, and exported in previous years.
Reviewing the historic volumes is useful since there are a number of complex and inter-related
factors beyond simple total production capacity that could affect the supply of advanced
biodiesel and renewable diesel. These factors include, but are not limited to, the RFS volume
requirements (including the BBD, advanced biofuel, and total renewable fuel requirements), the
availability of advanced biodiesel and renewable diesel feedstocks,641 the extension of the
biodiesel tax credit, tariffs on imported biodiesel, biofuel policies in other countries, import and
distribution infrastructure, and other market-based factors. While historic data and trends alone
are insufficient to project the volumes of biodiesel and renewable diesel that could be provided
in future years, historic data can serve as a useful reference in considering future volumes.
Production, import, export, and total volumes of BBD are shown in Table 6.2.1-1.
641 Throughout this section we refer to advanced biodiesel and renewable diesel as well as advanced biodiesel and
renewable diesel feedstocks. In this context, advanced biodiesel and renewable diesel refer to any biodiesel or
renewable diesel for which RINs can be generated that satisfy an obligated party's advanced biofuel obligation (i.e.,
D4 or D5 RINs). While cellulosic diesel (D7) can also contribute towards an obligated party's advanced biofuel
obligation, these fuels are included instead in the projection of cellulosic biofuel presented in Chapter 6.1. An
advanced biodiesel or renewable feedstock refers to any of the biodiesel, renewable diesel, jet fuel, and heating oil
feedstocks listed in Table 1 to 40 CFR 80.1426 or in petition approvals issued pursuant to 40 CFR 80.1416, that can
be used to produce fuel that qualifies for D4 or D5 RINs. These feedstocks include, for example, soybean oil; oil
from annual cover crops; oil from algae grown photosynthetically; biogenic waste oils/fats/greases; non-food grade
corn oil; camelina sativa oil; and canola/rapeseed oil (See pathways F, G, and H of Table 1 to 80.1426).
297
-------
Table 6.2.1-1: BBD (D4) Production, Imports, and Exports from 2012 to 2022642 (million
gallons)" i
2014b
2015b
2016
2017
2018
2019
2020
2021
2022
Domestic
Biodiesel
(Annual Change)
1,297
(-67)
1,245
(-52)
1,581
(+336)
1,552
(-29)
1,841
(+289)
1,706
(-135)
1,802
(+96)
1,701
(-101)
1,616
(-85)
Imported Biodiesel
(Annual Change)
130
(-23)
261
(+131)
562
(+301)
462
(-100)
175
(-287)
185
(+10)
209
(+24)
208
(-1)
240
(+32)
Exported Biodiesel
(Annual Change)
72
(-5)
73
(+1)
89
(+16)
129
(+40)
74
(-55)
76
(+2)
88
(+12)
91
(+3)
113
(+22)
Total Biodiesel
(Annual Change)c
1,355
(-85)
1,433
(+78)
2,054
(+621)
1,885
(-169)
1,942
(+57)
1,815
(-127)
1,924
(+109)
1,818
(-106)
1,743
(-75)
Domestic
Renewable Diesel
(Annual Change)
149
(+79)
169
(+20)
231
(+62)
252
(+21)
282
(+30)
454
(+172)
472
(+18)
778
(+306)
1,371
(+593)
Imported
Renewable Diesel
(Annual Change)
130
(-15)
120
(-10)
165
(+45)
191
(+26)
176
(-15)
267
(+91)
280
(+13)
362
(+82)
320
(-42)
Exported
Renewable Diesel
(Annual Change)
15
(+10)
21
(+6)
40
(+19)
37
(-3)
80
(+43)
145
(+65)
223
(+78)
241
(+18)
324
(+101)
Total Renewable
Diesel
(Annual Change)c
264
(+154)
268
(+4)
356
(+88)
406
(+50)
378
(-28)
576
(+198)
529
(-47)
899
(+370)
1,367
(+468)
Total BBDd
(Annual Change)
1,619
(-31)
1,701
(+82)
2,412
(+711)
2,293
(-119)
2,322
(+29)
2,393
(+71)
2,457
(+64)
2,720
(+263)
3,123
(+403)
" All data from EMTS. EPA reviewed all advanced biodiesel and renewable diesel RINs retired for reasons other
than demonstrating compliance with the RFS standards and subtracted these RINs from the RIN generation totals for
each category to calculate the volume in each year. This table does not include D5 or D6 biodiesel and renewable
diesel. These fuels are discussed in Chapters 6.4 and 6.7, respectively.
b RFS required volumes for these years were not established until December 2015.
0 Total is equal to domestic production plus imports minus exports.
d Total BBD includes some small volumes (<20 million gallons per year) of D4 jet fuel.
Since 2014, the year-over-year changes in the volume of advanced biodiesel and
renewable diesel used in the U.S. have varied greatly, from a low of 119 million fewer gallons
from 2016 to 2017 to a high of 711 million additional gallons from 2015 to 2016. As discussed
previously, these changes were likely influenced by multiple factors. This historical information
does not by itself demonstrate that the maximum previously observed annual increase of 709
million gallons of advanced biodiesel and renewable diesel would be reasonable to expect in a
future year, nor does it indicate that greater increases are not possible. Significant changes have
occurred in both the fuel and feedstock markets (discussed further below) that will impact the
rates of growth of biodiesel and renewable diesel production and use in future years. Rather,
642 Similar tables of biodiesel and renewable diesel production, imports, and exports presented in previous annual
rules included advanced (D5) biodiesel and renewable diesel. This table only contains volumes of biodiesel and
renewable diesel that qualifies as BBD (D4). Advanced (D5) biodiesel and renewable diesel are covered in Chapter
6.4.
298
-------
these data illustrate both the magnitude of the changes in biomass-based diesel in previous years
and the significant variability in these changes.
This data also shows the increasing importance of renewable diesel in the BBD pool. In
2014 approximately 16% of all BBD was renewable diesel, and the remaining 84% was
biodiesel. However, in the last 6 years all of the net growth has been in renewable diesel volume.
By 2022 production and imports of renewable diesel had increased not only in absolute terms
(from 264 million gallons in 2014 to 1.36 billion gallons in 2022), but also as a percentage of the
BBD pool. In 2022 approximately 44% of all BBD was renewable diesel, while the remaining
56%) was biodiesel. As discussed further in the following sections, we expect that renewable
diesel will represent an increasing percentage of total BBD in future years.
The historic data indicates that the biodiesel tax policy in the U.S. can have a significant
impact on the volume of biodiesel and renewable diesel used in the U.S. in any given year. The
availability of this tax credit also provides biodiesel and renewable diesel with a competitive
advantage relative to other biofuels that do not qualify for the tax credit.
While the biodiesel blenders tax credit has applied in each year since 2010, it has
historically only been prospectively in effect during the calendar year in 2011, 2013, 2016, and
2020-2022, while other years it has been applied retroactively. Years in which the biodiesel
blenders tax credit was in effect during the calendar year (2013, 2016, 2020, 2021, and 2022)
generally resulted in significant increases in the volume of BBD used in the U.S. over the
previous year (629 million gallons, 711 million gallons, 64 million gallons,643 2 63 million
gallons, and 403 million gallons respectively). However, following the large increases in 2013
and 2016, there was little to no growth in the use of advanced biodiesel and renewable diesel in
the following years. Data from 2018 and 2019 suggests that while the availability of the tax
credit certainly incentivizes an increasing supply of biodiesel and renewable diesel, supply
increases can also occur in the absence of the tax credit, likely as the result of the incentives
provided by the RFS program, state LCFS programs, and other economic factors.
Another important factor highlighted by the historic data is the tariffs imposed by the
U.S. on biodiesel imported from Argentina and Indonesia. In December 2017 the U.S.
International Trade Commission adopted tariffs on biodiesel imported from Argentina and
Indonesia.644 According to data from EIA,645 no biodiesel has been imported from Argentina or
Indonesia since September 2017, after a preliminary decision to impose tariffs on biodiesel
imported from these countries was announced in August 2017. As a result of these tariffs, total
imports of biodiesel into the U.S. were significantly lower in 2018 than they had been in 2016
and 2017. The decrease in imported biodiesel did not, however, result in a decrease in the
volume of advanced biodiesel and renewable diesel supplied to the U.S. in 2018. Instead, higher
domestic production of advanced biodiesel and renewable diesel, in combination with lower
exported volumes of domestically produced biodiesel, resulted in an overall increase in the
volume of advanced biodiesel and renewable diesel supplied in 2018 and subsequent years.
643 This is the volume increase in 2020, which was impacted by the COVID pandemic.
644 "Biodiesel from Argentina and Indonesia Injures U.S. Industry, says USITC,"
https://www.nsitc.gov/press room/news release/2017/erl20511876.htm.
645 See "EIA Biodiesel Imports".
299
-------
6.2.2 Biomass-Based Diesel Production Capacity and Utilization
One of the factors considered when projecting the rate of production of BBD in future
years is the production capacity. Domestic biodiesel production capacity, domestic biodiesel
production, and the utilization rate of the existing biodiesel production capacity each year is
shown in Figure 6.2.2-1. Active biodiesel production capacity in the U.S. as reported by EIA has
experienced modest growth in recent years, from approximately 2.1 billion gallons in 2012 to
just over 2.5 billion gallons in 2019.646 As of December 2022, active biodiesel production
capacity has decreased slightly since then, to approximately 2.1 billion gallons.64'' While
production of biodiesel has generally increased during this time period, significant excess
production capacity remains, with facility utilization remaining at or below 75% through 2022.
EPA data on total registered biodiesel production capacity in the U.S., which includes both
facilities that are producing biodiesel and idled facilities, is much higher, approximately 3.9
billion gallons. Active biodiesel capacity as reported by EIA is the aggregate production capacity
of biodiesel facilities that produced biodiesel in any given month, while the total registered
capacity based on EPA data includes all registered facilities, regardless of whether they are
currently producing biodiesel or not. These data suggest that domestic biodiesel production
capacity is unlikely to limit biodiesel production in future years, and that factors other than
production capacity are limit domestic biodiesel production.
Figure 6.2.2-1: U.S. Biodiesel Production Capacity, Production, and Capacity Utilization
3.00
c
o
§ 250
~o
o
S; 2.00
&
"u
ro
S- 1-50
U
o 1.00
U
o 050
s
0.00
2012 2013 2014 2015 201S 2017 2018 2019 2020 2021 2022
Production Capacity Biodiesel Production Percent Utilization
Unlike domestic biodiesel production capacity, domestic renewable diesel production
capacity has increased significantly in recent years, from approximately 280 million gallons in
2017 to approximately 2.85 billion gallons in December 2022 (Figure 6.2.2-2).648 Domestic
,yV' Biodiesel production capacity from EIA Monthly Biodiesel Production reports and the EIA Monthly Biofuels
Capacity and Feedstock Update. February 2023.
m See EIA Monthly Biofuels Capacity and Feedstock Update.
r"1x Renewable diesel capacity is based on RFS facility registration data and EIA Monthly Biofuels Capacity and
Feedstock Update.
300
-------
renewable diesel production has increased along with production capacity in recent years, and
capacity utilization at domestic renewable diesel production facilities has been high,
approximately 80% from 2017-2022. Further, much of the unused capacity was likely the result
of facilities ramping up new capacity to full production rates. Unlike the biodiesel industry, in
which unused production capacity has persisted for many years, since 2017 production of
renewable diesel neared or exceeded the production capacity from the previous year.
Figure 6.2.2-2: U.S. Renewable Diesel Production Capacity, Production, and Capacity
Utilization
2,500
100%
C
O
'4=3
(J
13
T3
O
i_
Q_
CL
ro
U
ui
C
_o
c
Q
2,000
1.500
1,000
500
|
L 1
LLlkll
i\
U
50% 3
-i-j
40% S
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
I Production Capacity
i Renewable Diesel Production
• Percent Utilization
Renewable diesel production volumes are from EMTS data. Production capacity is from EMTS from 2012-2020
and EIA Monthly Biofuels Capacity and Feedstock Update for 2021 and 2022 when EIA first reported renewable
diesel production capacity. Production capacity shown for 2021 and 2022 is the average of the monthly reported
production capacities. Capacity utilization is calculated by dividing actual production by the total production
capacity.
A number of parties have announced their intentions to build new renewable diesel
production capacity with the potential to begin production of renewable diesel by the end of
2025. These new facilities include new renewable diesel production facilities, expansions of
existing renewable diesel production facilities, and the conversion of units at petroleum
refineries to produce renewable diesel. A list of the facilities expected to begin producing
renewable diesel by 2025, as well as existing facilities expected to complete expansions by 2025,
based on publicly available data is shown in Table 6.2.2-1.
301
-------
Table 6.2.2-1: New Renewable Diesel Production Capacity in the U.S. Through 2025
Facility Name
Location
Capacity
(MGY)
Start Date (Actual
or Expected)
Bakersfield Renewables649
Bakersfield, CA
230
Q2 2023
Marathon Martinez Expansion650
Martinez, CA
470
End 2023
Phillips 66 Rodeo Phase 2651
Rodeo, CA
680
Q1 2024
Next Renewable Fuels652
Port Westward, OR
575
2024
REG Geismar Expansion653
Geismar, LA
250
2024
World Energy654
Paramount, CA
340
2025
If all these facilities were completed according to their current schedules these facilities
would increase domestic renewable diesel production capacity by approximately 2.5 billion
gallons per year, increasing total domestic renewable diesel capacity to over 5.5 billion gallons
by the end of 2025. This production capacity projection is similar to recent projections of
domestic renewable diesel capacity by 2025 from EIA (5.9 billion gallons per year by the end of
20 25).655 However, feedstock limitations (discussed in Chapter 6.2.3) are not expected to support
all of these facilities. It is also possible that some of these projects may be delayed or cancelled.
Thus, it is unlikely that the domestic renewable diesel production will reach the approximately
5.5 billion gallons implied by the sum current production capacity and the new renewable diesel
projects with the intention to begin production by 2025. Nevertheless, it appears unlikely that
domestic production capacity will limit renewable diesel production through 2025. Rather it is
more likely that the feedstock limitations discussed in Chapter 6.2.3 may limit production.
6.2.3 Availability of Biomass-Based Diesel Feedstocks
As EPA considered the rate of production of BBD through 2025, a central and critical
factor influencing final volume requirements was our assessment of the availability of qualifying
feedstocks. To assess the availability of feedstocks for producing BBD through 2025, we first
reviewed the feedstocks used in previous years. This review of feedstocks used in previous years
can provide information about the feedstocks most likely to be used in future years, as well as the
likely increase in the availability of such feedstocks in future years. A summary of the feedstocks
used to produce BBD from 2012 through 2022 is shown in Figure 6.2.3-1.
649 Cox, John. Refinery owner hunts for capital to complete past-due conversion project on Rosedale.
Bakerfield.com. November 14, 2022.
650 Id.
651 Phillips 66 Makes Final Investment Decision to Convert San Francisco Refinery to a Renewable Fuels Facility.
Phillips 66 News Release. May 11, 2022.
652 Kotrba, Ron. Oregon approves key permit for $2 billion renewable diesel project. Biobased Diesel Daily. March
25, 2022.
653 Renewable Energy Group Breaks Ground on Geismar, Louisiana Renewable Diesel Expansion and Improvement
Project. Renewable Energy Group Website, https://www.regi.com/resoiirces/press~releases/renewable~energv-
groiip~breaks~groiind~on~geisniar~toiiisiana~renewabte~dieset~expansion-and~iniprovenient~proieet.
654 Air Products Teaming Up with World Energy to Build $2 Billion Conversion of Sustainable Aviation Fuel (SAF)
Production Facility in Southern California. Air Products Website, https://www.aiiprodiicts.com/news~
center/2022/04/0422-air-products-and-world-energy-sustainable-aviation-fuel-facilitv-in-california.
655 Domestic Renewable Diesel Capacity Could More Than Double Through 2025. U.S. Energy Information
Administration. Today in Energy. February 2, 2023.
302
-------
Figure 6.2.3-1: Feedstocks Used To Produce BBD in the U.S. (2014-2022)656
iiiilllll
2014 2015 2016 2017 201S 2019 2020 2021 2022
¦ FOG ¦ Corn Oil ¦ Soybean Oil BCanolaOil
Domestic BBD production from fats, oils, and greases (FOG) in the U.S. has generally
increased from 2014 through 2022 at an average annual rate of approximately 50 million gallons
per year. These feedstocks are generally by-products of other industries. Assessments submitted
by commenters rule generally agree that the domestic supply of these feedstocks will increase
only slightly in future years.657 We expect that in future years production of BBD from FOG will
continue to increase at approximately the historical rate as the availability of FOG increases with
population. It is possible that greater demand for feedstocks for BBD production could result in
the diversion of greater quantities of FOG to BBD production at the expense of other markets
that currently use FOG feedstocks. Alternatively, it could also result in greater collection of FOG
that is currently sent to landfills or wastewater treatment systems, but we do not expect
significant increases in the collection rates of FOG for BBD production through 2025.
Production of BBD from distillers corn oil has also generally increased through 2022.
The most significant increases in the volume of BBD produced from distillers corn occurred
through 2018, as more corn ethanol plants installed equipment to produce distillers corn oil and
corn ethanol production expanded. However, production of BBD from this feedstock has been
fairly consistent at about 250 - 350 million gallons per year since 2017. Total production of
distillers corn oil in the U.S. in 2021 was approximately 2 million tons,658 or enough corn oil to
produce about 530 million gallons of BBD. This suggests that distillers corn oil could be used to
produce over 200 million gallons of additional BBD, but that would require shifting distillers
corn oil from other existing uses, which would then have to be backfilled with other new
sources.659 It is also possible that domestic production of distillers corn oil could increase in
656 Based on EMTS data
657 For example, see "The Outlook for Increased Availability & Supply of Sustainable Lipid Feedstocks in the U.S.
to 2025," LMC International. Submitted by Clean Fuels Alliance America (EPA-HQ-OAR-2021-2021-0427-0805).
658 USDA Grain Crushings and Co-Products Production 2021 Summary. March 2022.
https://www.nass.usda.gov/Publications/TodaYS Reports/reports/cagcan22.pdf.
659 For a discussion of backfilling when oil is removed from dried distillers grains, see 83 FR 37735 (August 2,
2018).
303
-------
future years for a variety of reasons, including new varieties of corn with higher oil content,
greater extraction rates, or increased ethanol production for domestic or international markets. As
with increased FOG collection, however, we do not expect these changes to significantly
increase the domestic supply of corn oil in 2025.
The remaining volume of BBD has been produced from canola oil and soybean oil.
Production of BBD from canola oil has fluctuated in recent years from a high of approximately
180 million gallons in 2016 to a low of approximately 120 million gallons in 2015. Production of
BBD from canola oil averaged approximately 160 million gallons each year from 2018 to 2022.
Total production of canola oil reached a high of approximately 1.8 billion pounds in the
2019/2020 agricultural marketing year, or enough canola oil to produce approximately 240
million gallons of BBD.660 Canola oil production has ranged between 1.5 and 2.0 billion pounds
from 2013/2014 and 2021/2022.661 Significant increases in canola oil production in the U.S.
through 2025 are unlikely due to both the relatively poor economic return on canola in many
parts of the U.S. and the lack of additional crush capacity for soft seed vegetable oil crops like
canola. An additional 4 billion pounds of canola oil, or enough to produce approximately 500
million gallons of BBD, was imported in 2020/2021. A pathway that allows for the generation of
RINs for renewable diesel produced from canola oil was also finalized in December 2022. This
new pathway could increase demand for canola oil for biofuel production.
In comments on the proposed rule, several commenters noted recent announcements of
new or expanding canola crush capacity in Canada. These commenters stated that this new crush
capacity would result in additional canola oil production that could be available to U.S. biofuel
producers. According to the Canola Council of Canada, since 2021 companies have announced
five major investments in expanded canola crushing capacity, together which would increase
total canola crushing capacity in Canada by approximately 6.7 million metric tons by 2025.662
One of these facilities has since canceled their plans, however the remaining capacity expansions
appear to be on track to complete construction by 2025. While it is unlikely that the entire
increase in canola oil production from this additional crushing capacity will be available to U.S.
biofuel producers due to demand from other markets (including Canadian biofuel producers) we
expect that increasing quantities of canola oil from Canada will be available for U.S. domestic
biodiesel and renewable diesel production in 2023 - 2025 (see Chapter 6.2.6 for more details on
our projection of the availability of imported canola oil to domestic biofuel producers).
The largest source of BBD production in the U.S. historically has been soybean oil. Use
of soybean oil to produce biodiesel increased from approximately 5.1 billion pounds in the
2013/2014 agricultural marketing year to approximately 10.3 billion pounds in the 2021/2022
agricultural marketing year.663 During this time period the percentage of all soybean oil produced
in the U.S. used to produce biodiesel increased from approximately 25% in 2013/2014 to
approximately 40% in 2021/2022. As a point of reference, if all the soybean oil produced in the
660 U.S. Canola oil production data sourced from USDA's Oil Crops Yearbook, https://www.ers.usda.gov/data-
prodncts/oil-crops-vearbook.
661 Id.
662 "The Oilseed Processing Industry." Canola Council of Canada.
663 U.S. Soybean oil production and use data sourced from USDA's March 2023 Oil Crops Yearbook.
https://www.ers.nsda.gov/data-prodncts/oil-crops-vearbook. The agricultural marketing year for soybeans runs from
September to August.
304
-------
U.S. in 2021/2022 (25 billion pounds) were used to produce BBD, this quantity of feedstock
could be used to produce approximately 3.4 billion gallons of BBD. Thus, BBD production from
soybean oil could more than double if it were all shifted from its other existing uses, including
food, and backfilled with other new sources such as palm oil, potentially impacting the GHG
benefits.
Additional soybean oil production in future years could come from several sources. The
first potential source of additional soybean oil is increased crushing of soybeans in the U.S.
Soybean crushing is the process by which whole soybeans are converted into soybean oil and
soybean meal. The percentage of U.S. soybean production that has been crushed has varied from
a low of 44% in the 2016/2017 agricultural year to a high of 61% in the 2019/2020 marketing
year.664 Most of the rest of the whole soybeans are exported to foreign countries, where the beans
are then crushed to produce soybean meal and soy oil for their own markets.
Strong demand for vegetable oil has already resulted in increasing domestic crushing of
soybeans. Recent data from USDA indicates that soybean crushing reached record levels of 65.9
million tons (approximately 2.2 billion bushels) in the 2021/2022 agricultural marketing year.665
There have also been numerous announcements of investments to increase domestic soybean
crush capacity, both through the construction of new facilities as well as the expansion of
existing facilities. Increasing the domestic soybean crushing capacity is expected to result in
increased soybean oil production in the U.S. If the increased domestic crushing capacity results
in reduced exports of whole soybeans (rather than increased soybean production) this increased
soybean oil production could be achieved with little impact on overall U.S. soybean production.
The USDA Agricultural Projections to 2032 project increasing domestic soybean oil
production through 2025, largely as a result of an increased crushing of soybeans. USDA
projects that domestic soybean oil production will increase by approximately 1.3 billion pounds
from 2022 (26.2 billion pounds) to 2025 (27.5 billion pounds).666 If this entire increase in
soybean oil production were used to produce biodiesel or renewable diesel, it would result in an
increase of approximately 170 million gallons of biofuel from 2022 to 2025, or an increase of
approximately 60 million gallons per year. These projections, however, assume that the RFS
volume requirements stay constant at 2022 levels in all future years. Absent increased demand
for soybean oil for biofuel production domestic, soybean crush, and by extension domestic
soybean oil production, is primarily driven by increasing demand for soybean meal from the
livestock industry. These projections are therefore not likely representative of potential domestic
soybean oil production with incentives for increasing biodiesel and renewable diesel in place
through 2025. Perhaps most importantly, these projections do not appear to account for the
significant investments that have been made to increase domestic soybean oil production through
2025.
664 U.S. Soybean crushing data sourced from USDA's Oil Crops Yearbook, https://www.ers.usda.gov/data-
prodncts/oil-crops-vearbook.
665 USDA's Oil Crops Yearbook (March 2023).
666 USDA Agricultural Projections to 2032. February 2023. For each year EPA converted soybean oil production
projections to calendar year prices by weighting production in the first agricultural marketing year (e.g., 2022/2023
for the 2023 price) by 0.75 and production in the second agricultural marketing year (e.g., 2023/2024 for the 2023
price) by 0.25.
305
-------
Figure 6.2.3-2: Soybean Oil Production and Soybean Oil Used for Biofuel Production
30,000
25,000
T5 20,000
c
3
O
^ 15,000
o
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
Soy Oil Production (Actual) Soy Oil Used for Biofuel (Actual)
Soy Oil Production (Projected) Soy Oil Used for Biofuel (Projected)
Actual data from USDA Oil Crops Yearbook; Projected data from USDA Agricultural Projections to 2031
In comments on the proposed rule, many stakeholders stated that EPA had significantly
under-projected the potential for domestic soybean oil production in 2023 - 2025. These
commenters generally cited the large number of announcements of investment in increasing
soybean crush capacity. The National Farmers Union identified over 24 major expansions or new
soybean crushing facilities that are announced or underway.667 The American Soybean
Association projected that the increase in domestic soybean oil production from 2023 - 2025
would be sufficient to produce approximately 700 million gallons of biodiesel and renewable
diesel.668 The Clean Fuels Alliance America submitted a study conducted by LMC international
that found that the projected growth in soybean oil production in the U.S. from 2021 - 2025
would be sufficient to produce approximately 750 - 800 million gallons of biodiesel and
renewable diesel.669 While there are some slight variations in these estimates, the data submitted
by commenters demonstrates that domestic soybean oil production is likely to increase
significantly through 2025.
In addition to increased soybean crushing, additional quantities of soybean oil could be
made available for biofuel production from decreased exports and/or increased imports of
soybean oil. From the 2011/2012 agricultural marketing year through the 2021/2022 agricultural
marketing year approximately 10% of the soybean oil produced in the U.S. was exported.670
Soybean oil exports in 2021/2022 are estimated at approximately 1.7 billion pounds, or enough
soybean oil to produce approximately 225 million gallons of biodiesel or renewable diesel.671
Soybean oil imports have been relatively small (300-400 million pounds)672 in recent years,
667 EPA-HQ-OAR-2021-0427-0595.
668 EPA-HQ-OAR-2021-0427-0579.
669 EPA-HQ-OAR-2021-0427-0805.
6711 USDA Economic Research Service. Oil Crops Data: Yearbook Tables. March 2023.
671 Id.
672 Id
306
-------
likely due to the tariff on soybean oil imports.673 As with other potential sources of BBD
feedstock with existing markets, increasing BBD production by decreasing exports and/or
increasing imports of soybean oil would require shifting these feedstocks from existing markets
including food supply in the U.S. and abroad and then backfilling with other new supplies such
as palm oil or other vegetable oils produced in foreign countries, potentially impacting the GHG
benefits.
Finally, additional vegetable oil feedstocks in future years could come from international
sources. The February 2023 WASDE report from USD A project that global production of
vegetable oils will be approximately 204 million metric tons in the 2022/2023 agricultural
marketing year.674 This quantity of vegetable oil, if converted to fuel, would result in
approximately 59 billion gallons of biodiesel and/or renewable diesel. The vast majority of this is
used for food and other purposes around the world and could not be readily used to supply
advanced biodiesel and renewable diesel to the U.S.675 Furthermore, much of this vegetable oil is
also likely to be from palm oil that does not currently have an approved pathway under the RFS
program except for the portion that could be produced under the program's grandfathering
provisions. However, the large global production of vegetable oil suggests that increased imports
of vegetable oil, or biodiesel and renewable diesel produced from vegetable oil (discussed in
Chapter 6.2.4), may be made available to markets in the U.S. in future years.
While the global production of vegetable oils far exceeds the quantity of vegetable oil
used for biofuel production, there is significant demand for vegetable oils in other markets such
as for food, animal feed, and oleochemical production. In comments on the proposed rule,
stakeholders representing the food and pet food manufacturing industries opposed increasing the
volume requirements for biodiesel and renewable diesel due to concerns that increased demand
for biofuel production would negatively impact the supply of animal fats and vegetable oils to
other markets. Some of these commenters cited recent high prices for vegetable oils as evidence
of increasing competition for these feedstocks. Recent prices for vegetable oils suggest that the
market for vegetable oils has tightened in recent years, with demand for vegetable oils increasing
relative to supply. From 2013/2014 through 2019/2020 the price for soybean oil generally ranged
from $0.30-$0.40 per pound.676 In 2021/2022 soybean oil prices increased to $0.73 per pound.677
Soybean oil prices reached a high of approximately $0.87 per pound in April 2022, before falling
to approximately $0.62 per pound in March 2023.678 Soybean oil prices are projected by USDA
to decrease from 2022/2023 ($0.69 per pound) through 2025/2026 ($0.47 per pound).679
673 Harmonized Tariff Schedule of the United States (2020) Revision 19.
674 United States Department of Agriculture World Agricultural Supply and Demand Estimates. February 8, 2023.
675 These reasons include the demand for vegetable oil in the food, feed, and industrial markets both domestically
and globally; constraints related to the production, import, distribution, and use of significantly higher volumes of
biodiesel and renewable diesel; and the fact that biodiesel and renewable diesel produced from much of the
vegetable oil available globally may not qualify as an advanced biofuel under the RFS program.
676 USDA Economic Research Service. Oil Crops Data: Yearbook Tables. March 2023.
677 Id.
678 Nasdaq Soybean Oil Price. July 14, 2022. https://www.nasdaa.com/market-activitv/commodities/zl.
679 USDA Agricultural Projections to 2032. February 2023.
307
-------
Actual data from USDA Oil Crops Yearbook; Projected data from USDA Agricultural Projections to 2032
While increased soybean oil demand for biofuel production is likely a contributing factor
to the higher soybean oil prices observed in recent years it is not the only factor. The current high
prices have also been affected by poor weather conditions in South America and Malaysia over
the past year, which has negatively impacted global vegetable oil production. In 2021 there was
drought in Argentina and Brazil (two of the largest exporters of soybeans and soybean oil).680 At
the same time, palm oil production in Malaysia was impacted by flooding caused by a
typhoon.681 Recently Argentina is experienced its worst drought in decades, with soybean
production expected to fall well below historical levels.682
Despite these current high prices for vegetable oil, the data discussed above indicate that
there will be some additional supply of vegetable oil to enable increasing production of biofuels
from vegetable oils in future years. We project increases in the availability of FOG and distillers
corn oil, consistent with historical trends. We expect increasing volumes of canola oil imported
from Canada will be available to U.S. biofuel producers as canola crushing capacity expands
through 2025. Finally, we expect significant increases in domestic soybean oil production from
increased soybean crushing capacity in the U.S. from 2023 - 2025. Taken together, these
projected increases in feedstock production are expected to enable significant growth in
renewable diesel production through 2025.
6.2.4 Imports and Exports of Biomass-Based Diesel
In evaluating the likely rate of production of BBD through 2025 we also examined BBD
imports and exports in previous years. While imports and exports of BBD may not directly
impact the rate of production of BBD in the U.S., they do impact the volume of these fuels
6811 Wilson, Nick. Oil Prices Surge - Vegetable Oil That Is. Marketplace.org. February 17, 2022.
681 Id.
682 Sigal, L. and Raszewski, E. Argentina's 'unprecedented' drought pummels farmers and economy. Reuters. March
9, 2023.
308
-------
available to obligated parties. We therefore think that the volume of these fuels that may be
imported and exported in future years is a relevant consideration as we require volumes through
2025 under the RFS program.
Since 2014 biodiesel imports have generally averaged about 200 million gallons per year,
with the exception of 2015-2017. During this time (2015-2017) biodiesel imports from Argentina
surged, with biodiesel imported from Argentina responsible for 64% of all biodiesel imports in
these three years. In August 2017, the U.S. announced preliminary tariffs on biodiesel imported
from Argentina and Indonesia.683 These tariffs were subsequently confirmed in April 2018.684
Since the time the preliminary tariffs were announced, EIA has not reported any biodiesel
imported from these countries.685 After the imposition of these tariffs, imports of biodiesel from
other countries has increased marginally; however, the biggest effect of these tariffs has been a
decrease in the total volume of imported biodiesel to approximately 200 million gallons during
2018-2022.
Figure 6.2.4-1: Biodiesel and Renewable Diesel Imports (2014-2022)
800
2014 2015 2016 2017 201S 2019 2020 2021 2022
¦ Biodiesel ¦ Renewable Diesel
Biodiesel and renewable diesel imports based on data from EMTS
Renewable diesel imports have generally increased since 2014, with larger increases
observed in recent years. A significant factor in the increasing imports of renewable diesel
appears to be the California Low Carbon Fuel Standard (LCFS), as the vast majority of the
renewable diesel consumed in the U.S. (including both domestically produced and imported
renewable diesel) has been consumed in California.686 We expect that, as the carbon intensity
683 82 FR 40748 (August 28, 2017).
684 8 3 FR 18278 (April 26, 2018).
685 See EIA data on biodiesel imports by country,
https://www.eia.gov/dnav/pet/pet move impcus a2 nus EPOORDB imO mbbl a.htm.
686 Data from California's LCFS program indicates that approximately 940 million gallons of renewable diesel were
consumed in California in 2021, the most recent year for which data are available
(https://ww3.arb.ca.gov/fuels/lcfs/dashboard/dashboard.htm). Data from EMTS indicates that 960 million gallons of
309
-------
requirements in California's LCFS program continue to decrease, and as similar LCFS programs
are taken up in other states (e.g., Oregon and Washington), these programs, in conjunction with
the RFS program and the federal tax credit, will continue to provide an attractive market for
domestically produced and imported renewable diesel.
Exports of RIN generating biodiesel, based on EMTS data, have been fairly consistent
since 2014, generally ranging between 70 and 130 million gallons per year. According to EMTS
data, renewable diesel exports increased with domestic renewable diesel production, reaching
over 300 million gallons in 2022. Increasing exports of renewable diesel reflect the existence of
biofuel mandates and significant financial incentives creating high demand in other countries that
the U.S. must compete with. As one example, Canada recently finalized new Clean Fuel
Regulations that require increasing volumes of low-carbon fuels in future years.687 At this time,
it is difficult to project whether renewable diesel exports will continue to increase in future years
or alternatively return to the low levels observed through 2017.
The fact that there are both imports and exports of BBD simultaneously also suggests that
there are efficiencies associated with importing into and exporting from certain parts of the
country as well as economic advantages associated with the use of BBD from different
feedstocks in different foreign and domestic markets. One factor likely supporting simultaneous
imports and exports of biodiesel and renewable diesel is the structure of the biodiesel tax credit.
The U.S. tax credit for biodiesel and renewable diesel applies to fuel either used or produced in
the U.S. Thus, by importing foreign produced biodiesel and renewable diesel for domestic use
and then exporting domestically produced biodiesel and renewable diesel to other countries
parties are able to claim the biodiesel tax credit on both the imported and the exported volumes.
renewable diesel were consumed in the U.S. in 2021, including both renewable diesel that generated BBD RINs and
advanced RINs.
687 Tuttle, Robert. Canada Releases California-Style Fuel Rules to Cut Emissions. Bloomberg, June 29, 2022.
310
-------
Figure 6.2.4-2: Biodiesel and Renewable Diesel Exports (2014-2022)
500
I I I I I I l l l
2014 2015 2016 2017 2018 2019 2020 2021 2022
¦ Biodiesel ¦ Renewable Diesel
Biodiesel and renewable diesel exports based on data from EMTS
6.2.5 Biomass-Based Diesel Supply in 2023
In considering the projected rate of production and use of BBD in 2023 - 2025 it is also
relevant to consider the available BBD production and import data for BBD to date for 2023. At
the time the analyses for this rulemaking were completed EPA had RIN generation data for the
first three months of 2023 (January - March). These data are summarized in Table 6.2.5-1.
Table 6.2.5-1: BBD Production, Import, and Export Data (January - March 2023) (Million
RINs)
January
February
March
Total
Domestic BBD
422.3
408.1
489.1
1,319.6
Importer/Foreign Generator BBD
100.3
106.2
129.8
336.3
Exports
N/A
N/A
N/A
134.1
Total RIN Supply
N/A
N/A
N/A
1,521.8
While RIN generation and retirement data for the first three months of 2023 are not
determinative of RIN generation and retirement though the remainder of the year, we can use this
data to inform the supply of BBD in 2023. The simplest way to project BBD RIN supply in 2023
using this data is to multiply the total RIN supply from the first quarter of the year by 4. This
would indicate that the expected supply of BBD in 2023 would be approximately 6.1 billion
RINs. This projection methodology, however, ignores the observed seasonality in BBD RIN
generation. An alternative methodology to use the available data to project the BBD RIN supply
in 2023 that takes this seasonality into account is to compare RIN generation in the first 3
months of 2023 to RIN generation during the first 3 months of 2022. From this data we can
calculate a percentage increase (or decrease) that can be applied to the total BBD RIN supply in
2022 to project the total BBD RIN supply in 2023. These calculations, and the resulting
projecting of BBD RIN supply in 2023 are shown in Table 6.2.5-2.
450
400
350
300
250
200
150
100
50
0
311
-------
Table 6.2.5-2: Projected BBD Supply for 2023 Based on Q1 20233 Data (Million RINs)
Extrapolating Q1 Production
RIN Generation (Jan.
- Mar. 2023)
RINs Retired for
Exports (Jan. - Mar.
2023)
RIN Generation (Jan.
- Mar. 2023)
BBD RIN Supply
(2023) - Projected
1,656
134
1,522
6,088
Projection based on Growth from Q1 2022 to Q1 2023
RIN Generation
(Jan. - Mar.
2022)
RIN Generation
(Jan. - Mar.
2023)
Percent Change
BBD RIN
Supply (2022)
BBD RIN
Supply (2023)-
Projected
1,241
1,656
+33.4%
4,956
6,613
In our analyses for this final rule we projected that the market would supply
approximately 6.2 billion BBD RINs to meet the implied volumes for non-cellulosic advanced
biofuel (5.1 billion RINs) and the volume of BBD needed to make up for the shortfall in the
supply of conventional renewable fuel (approximately 1.4 billion gallons, including the
supplemental volume requirement) from the proposed rule, after accounting for the projected
supply of other non-cellulosic advanced biofuels (290 ethanol-equivalent gallons).688 Data from
the first three months of 2023 indicate that the market is on track to supply this volume of BBD
in 2023. These data also indicate that both domestic production and imports have increased in the
first three months of 2023 relative to the same time period in 2022. On a percentage basis
imports have increased at a slightly higher rate (36%) than domestic production of BBD (33%).
Further, while we do not have data on the quantity of BBD produced domestically from imported
feedstocks, available data and news reports indicate that imports of soybean oil,689 canola oil,690
and FOG691 have all increased in 2023 relative to previous years. These observed increases in
imported BBD and imported feedstocks that can be used to produce BBD are consistent with our
analysis of available BBD feedstocks presented in Chapters 6.2.3 and 6.2.6, which indicated that
opportunities for increased production of BBD from domestic feedstocks in 2023 are limited.
6.2.6 Projected Rate of Production and Use of Biomass-Based Diesel
Based on the factors discussed in the preceding sections, we have projected domestic
BBD production and net BBD imports through 2025. Our analyses indicate that production
capacity and the ability to distribute and use biodiesel and renewable diesel are unlikely to
constrain BBD production through 2025 (see Chapters 6.2.2, 7.3, and 7.4 for further discussion
on biodiesel and renewable production capacity and impacts on infrastructure). Further, the
significant increase in renewable diesel production capacity projected through 2025, in
combination with the decreasing biodiesel operational capacity observed in recent years,
suggests that increases in BBD production are more likely to be renewable diesel rather than
biodiesel.
688 See Chapter 3 for more information on our projection of the candidate volumes for this rule.
689 USD A FAS Soybean Oil Import Data.
690 USD A FAS Canola Oil Import Data.
691 Xu, C., and Chipman, K. China's Used Cooking Oil is Starting to Clean Up Dirty US Diesel. Bloomberg.
February 28, 2023.
312
-------
We projected the domestic production and net imports of BBD separately by fuel type
(biodiesel and renewable diesel) and feedstock. Our projections of biodiesel production from all
feedstocks and renewable diesel production from FOG and distillers corn oil are based on linear
regressions of the quantities of these fuels supplied in previous years. We expect that changes in
the supply of these fuels will generally follow the trends observed in previous years. Our
projections of the production and net import of these fuels are presented in Chapter 6.2.6.1.
For two types of BBD, renewable diesel produced from soybean oil and canola oil, we do
not believe that the historical data provides a good indication of the future production and net
import of these fuels. Rather, we anticipate that the production and net imports of these fuels will
increase beyond what would be expected based solely on the historical data due to projected
increases in the availability of domestic soybean oil and imported canola from Canada available
for biofuel production. Our projections of the production and net imports of these fuels are based
on our projections of the increases in the quantity of North American soybean oil and canola oil
available to U.S. biofuel producers in 2023 - 2025. These projections are described in Chapter
6.2.6.2.
After projecting the production and net imports of BBD as described above, we next
compared the projected volumes to the volume of BBD we estimate would be needed to meet the
proposed RFS volumes for 2023 - 2025. The resulting projected volumes are significantly less
than the volume of BBD we project would be needed to meet the proposed RFS volumes for
2023, slightly less than the proposed volumes for 2024, and significantly higher than the
proposed volumes for 2025. This implies that in 2023 and to a lesser extent in 2024 the market
would likely need to increase BBD production by diverting feedstocks from existing uses or rely
on additional volumes of imported BBD and/or feedstocks to produce BBD to meet the proposed
RFS volumes. Available data from January - March 2023 indicates that the market is on track to
supply the volume of BBD we project would be needed to satisfy the proposed RFS volume
requirements in 2023. This data is discussed in Chapter 6.2.6.3.
Finally, Chapter 6.2.6.4 presents our projections of BBD production and net imports in
2023 - 2025. These volumes include the BBD we project would be supplied in these years based
on the historical data and projections of increased soybean oil and canola oil production in North
America and the additional volumes of BBD from imports and/or imported feedstocks that are
being supplied to U.S. markets based on data from January - March 2023.
6.2.6.1 Projected Supply of Biodiesel and Renewable Diesel Based on Historical
Data
To project the production and net imports of biodiesel from all feedstocks and renewable
diesel from distillers corn oil and FOG, we used a linear regression based on the quantities of
these fuels supplied from 2018-2022. As discussed in Chapter 6.2.3, we expect limited increases
in the availability of corn oil and FOG for biofuel production in future years, as these feedstocks
are primarily by-products of other industries. While we do project significant increases in the
production of soybean oil and canola oil in North America from 2023 - 2025, most of the
increases in the quantity of these feedstocks available for biofuel production will be used to
313
-------
produce renewable diesel rather than biodiesel. We therefore believe that regressions based on
past data are appropriate to use to project the future production and net imports of these fuels.
The domestic production and net imports of these fuels from 2018-2022, and the
equations used to project the domestic production and net imports from 2023-2025 based on a
linear regression of the historical data are shown in Table 6.2.6-1. The projected volumes of
these fuels, using the equations from Table 6.2.6-1, are shown in Table 6.2.6-2.
Table 6.2.6-1: Domestic Production and Net Imports of BBD (2016-2022; million gallons
Fuel
Feedstock
2018
2019
2020
2021
2022
Equation
Biodiesel
Canola Oil
214
223
264
269
267
15.2X-30,419
Biodiesel
DCO
255
210
181
196
130
-26.2X+ 53,108
Biodiesel
FOG
433
375
345
368
346
-18.0X +36,690
Biodiesel
Soybean Oil
1,041
1,006
1,113
954
995
-14.5X + 30,354
Renewable Diesel
DCO
59
78
61
115
209
19.7X-39,610
Renewable Diesel
FOG
304
443
421
605
859
85.OX- 171,160
Table 6.2.6-2: Projected Domestic Production and Net Imports of BBD (2023-2025; million
Fuel
Feedstock
2023
2024
2025
Biodiesel
Canola Oil
292
307
323
Biodiesel
DCO
116
89
63
Biodiesel
FOG
321
202
285
Biodiesel
Soybean Oil
982
968
953
Renewable Diesel
DCO
205
239
272
Renewable Diesel
FOG
905
1,032
1,160
6.2.6.2 Projected Supply of Renewable Diesel Based on North American
Feedstock Growth
To project domestic production and net imports of renewable diesel produced from
soybean oil and canola oil we used a different methodology. This is partially due to the fact that
there has been no discernable trend in the use of soybean oil and canola oil for renewable diesel
production from 2018-2022. More importantly, however, we expect that the production of these
fuels in future years will primarily be dependent on the availability of feedstock, and the
projected increases in soybean and canola oil in the U.S. and Canada (discussed in Chapter 6.2.3)
indicate that the availability of these feedstocks in future years will increase significantly through
2025 due to investments in soybean and canola crush capacity, and therefore is unlikely to follow
historic trends.
Our projections of renewable diesel produced from soybean oil and canola oil are
primarily based on our projections of increased soybean oil and canola oil production in the U.S.
and Canada. To project increases in soybean oil available for renewable diesel production we
started with projections of the increase in soybean oil crush capacity from the American Soybean
314
-------
Association.692 In their comments on the proposed rule, the American Soybean Association
suggested a methodology for estimating the annual increase in soybean crush based on the
projected increase in crush capacity. They suggested that because new crush capacity may come
online part way through the year that the practical crush capacity in any given year could be
estimated by averaging the available crush capacity at the end of the year with the available
crush capacity in the previous year.693 They further suggested that the practical crush capacity be
estimated at 92% of the listed nameplate capacity, and assuming each facility operates 350 days
per year.694 We used these recommended methodologies to project the increase in soybean crush
each year from 2023 - 2025 over the total soybean crush in 2022.
After projecting the increase in soybean crush each year over 2022 levels, we next
projected how the increase in crush capacity would affect the availability of soybean oil to
renewable diesel producers. We first calculated the total increase in soybean oil production
associated with the increase in soybean crush using the average yield of soybean oil per bushel of
soybeans according to data from USDA (11.7 lbs soybean oil per bushel soybeans crushed).695
We next projected that 80% of the increase in soybean oil production would be available to
renewable diesel producers, while 20% of the increased soybean oil production would be used in
other non-biofuel markets. Finally, we projected the quantity of renewable diesel that could be
produced from the increase in soybean oil production. These calculations, and the resulting
projections in the volume of renewable diesel that could be produced in 2023 - 2025 (relative to
the volume produced in 2022) are shown in Table 6.2.6-3.
Table 6.2.6-3: Projected Renewable Diesel Growth Based on Increased Soybean Oil
Production
2022
2023
2024
2025
Additional Soybean Crush Capacity (million
bushels)
59.1
94.1
276.9
142.4
Practical Crush Capacity Increase (million
bushels)
N/A
76.6
185.5
209.6
Soybean Oil Production Increase (million lbs)
N/A
896
2,170
2,453
Soybean Oil Available to RD Producers
(million lbs)
N/A
717
1,736
1,962
Projected Increase in RD Production from
Soybean Oil - Annual (million gallons)
N/A
94
228
258
Projected Increase in RD Production from
Soybean Oil - Cumulative (million gallons)
N/A
94
323
581
To project increases in the volume of renewable diesel from canola oil imported from
Canada we used a similar methodology. We first projected increases in nameplate crush capacity
each year based on public announcements of facility expansions and new canola crush facilities.
We next estimated a practical crush capacity increase using the average of the current year and
692 American Soybean Association projections of soybean crush capacity in 2023 - 2025 were provided to EPA by
USDA.
693 See comments from the American Soybean Association (EPA-HQ-OAR-2021-0427-0579).
694 Id.
695 USDA Economic Research Service. Oil Crops Data: Yearbook Tables. March 2023.
315
-------
previous year crush capacity and the factors for production based on nameplate capacity (92%)
and operational days (350 per year) described in our assessment of soybean oil production. We
then projected the increase in canola oil production based on the average yield of canola oil per
ton of canola crushed (0.43 metric tons of oil per metric ton of canola seeds crushed). Due to the
high demand for canola oil in non-biofuel markets and demand for canola oil from Canadian
biofuel producers we estimated that 50% of the increase in canola oil production in Canada
would be available to U.S. renewable diesel producers. Finally, we estimated the quantity of
renewable diesel that could be produced from the increases in canola oil production. These
calculations, and the resulting projections in the volume of renewable diesel that could be
produced in 2023 - 2025 (relative to the volume produced in 2022) are shown in Table 6.2.6-4.
Table 6.2.6-4: Projected Renewable Diesel Growth Based on Increased Canola Oil
Production
2022
2023
2024
2025
Additional Canola Crush Capacity (million
0
0
4.6
1.1
metric tons)
Practical Crush Capacity Increase (million
N/A
0
2.0
2.5
metric tons)
Canola Oil Production Increase (million metric
N/A
0
0.87
1.08
tons)
Canola Oil Available to RD Producers (million
N/A
0
0.44
0.54
metric tons)
Projected Increase in RD Production from
N/A
0
126
156
Canola Oil - Annual (million gallons)
Projected Increase in RD Production from
N/A
0
126
283
Canola Oil - Cumulative (million gallons)
After projecting increases in renewable diesel production from domestic soybean oil
production and imported canola oil from Canada we added these volumes to the volume of these
fuels produced or imported in 2022. These numbers, shown in Table 6.2.6-5, reflect our
projections of potential renewable diesel production and net imports in 2023 - 2025 from growth
in available quantities of domestic soybean oil and canola oil imported from Canada.
Table 6.2.6-5: Renewable Diesel Production and Net Imports from Soybean Oil and Canola
Oil Based on North American Feedstock Production Growth (million gallons)
2022
2023
2024
2025
Renewable Diesel from Domestic Soybean Oil
293
387
616
874
Renewable Diesel from Canadian Canola Oil
0
0
126
283
6.2.6.3 Additional Supply of BBD Imports and Imported Feedstocks
The volumes shown in Tables 6.2.6-2 and 6.2.6-5 show our projections of domestic
production and net imports of biodiesel and renewable diesel based on the historic trends (for all
biodiesel and renewable diesel from corn oil and FOG) and expected increases in the production
of soybean oil and canola oil in North America (for renewable diesel from soybean oil and
canola oil). As a point of comparison, we next considered how the total volume of BBD that
316
-------
results from these projections compares to the volume of BBD that we projected would be
needed to satisfy the implied non-cellulosic advanced biofuel and conventional renewable fuel
volumes from the proposed rule. This comparison is shown in Table 6.2.6-6. Because proposed
RFS volumes for 2023 - 2025 are in RINs (rather than physical gallons) we show the number of
RINs we project would be generated for these fuels, rather than the number of physical gallons,
in this comparison.
Table 6.2.6-6: Projected Production and Net Imports of BBD from North American
Feedstock Growth (million RINs)
2023
2024
2025
Biodiesel from Canola Oil
438
461
484
Biodiesel from DCO
173
134
95
Biodiesel from FOG
481
454
427
Biodiesel from Soybean Oil
1,473
1,451
1,430
Renewable Diesel from Canola Oil
0
215
481
Renewable Diesel from DCO
348
406
484
Renewable Diesel from FOG
1,539
1,755
1,972
Renewable Diesel from Soybean Oil
659
1,047
1,486
Total BBD Supply
5,111
5,923
6,836
Volume Need to Meet Proposed Volumes
6,215
6,205
6,481
Difference
-1,104
-282
+355
This data demonstrates that with the projected growth in these fuels from 2023 - 2025
based on the historical data and projected growth in North American canola and soybean oil
production, the projected BBD supply would be significantly less than what we project will be
needed to meet the proposed volumes for 2023 and 2024, but significantly greater than what
would be need in 2025.
However, as discussed in Chapter 6.2.5, data from January - March 2023 indicates that
the market is nevertheless on track to supply the volumes of BBD necessary to meet the
proposed RFS volumes for 2023. This is a strong indication that the market is capable of
supplying additional volumes of BBD to the U.S. beyond those projected in Chapters 6.2.6.1 and
6.2.6.2. As the volumes presented in Table 6.2.6-6 represent all of the projected increases
relative to 2022 based on historic trends (for all biodiesel and renewable diesel produced from
DCO and FOG) and the projected increase in North American canola oil and soybean oil (for
renewable diesel produced from canola oil and soybean oil).The data from January - March
2023 also indicates that the most likely source of this additional volume currently being used in
2023 is imported BBD and imported feedstocks (beyond canola oil imported from Canada from
expanded crush capacity) that can be used to produce BBD.
To reflect the data on BBD supply from the first quarter of 2023 in our projections of the
production and net imports of BBD, we have increased our projections of the production and net
imports of renewable diesel in 2023 and 2024 such that the total volume of biodiesel and
renewable diesel is equal to the volumes projected to be needed to meet the proposed RFS
volumes for these years. We projected that any additional imports of BBD and/or feedstocks
used to produce BBD will result in greater volumes of renewable diesel available to the U.S.
317
-------
market, both because of the rapidly expanding renewable diesel production capacity (discussed
in Chapter 6.2.2) and the ease of access of renewable diesel to the California, Oregon, and
Washington fuels markets where LCFS credits provide an added incentive. While we do not yet
have sufficient data to determine precisely which feedstocks are being used to produce the BBD
and/or its feedstocks imported to the U.S. in 2023, available data and news reports suggest that
imports of soybean oil, canola oil, and FOG have all increased in 2023. Therefore, for the
purposes of the analyses in this document projected that the feedstocks used to produce this
renewable diesel would be equally divided between soybean oil, canola oil, and FOG. Because,
as discussed in Preamble Sections 3 and 6, we believe that fuels produced from North American
feedstocks likely have greater benefits and fewer negative environmental impacts that fuels
produced from foreign feedstocks. We therefore did not project additional renewable diesel
imports or renewable diesel produced from imported feedstocks for 2025.
The additional volumes of imported renewable diesel and renewable diesel produced
from imported feedstocks projected for 2023 - 2025 are shown in Table 6.2.6-7. The decreasing
volumes of renewable diesel produced from foreign feedstocks from 2023 - 2025 reflect the
projected increases in BBD feedstocks produced in North America and our desire to focus the
future growth of biodiesel and renewable diesel production on these feedstocks.
Table 6.2.6-7: Additional Supply of Renewable Diesel from Foreign Feedstocks (million
RINs/gallons)
2023
2C
)24
2
025
RINs
Volume
RINs
Volume
RINs
Volume
Total Renewable Diesel
1,103
649
282
166
0
0
Renewable Diesel from Soybean Oil
368
216
94
55
0
0
Renewable Diesel from Canola Oil
368
216
94
55
0
0
Renewable Diesel from FOG Oil
368
216
94
55
0
0
6.2.6.4 Projected BBD Production and Net Imports
A summary of all of the projected volumes of BBD from 2023-2025 is shown in Tables
6.2.6-8 (in million gallons) and 6.2.6-9 (in million RINs), along with the volumes of these fuels
supplied in 2022 for context. This table includes both volumes of biodiesel and renewable diesel
projected to be produced from North American feedstocks, and feedstocks from other foreign
countries.
318
-------
Table 6.2.6-8: Projected Domestic Production and Net Imports of Renewable Diesel
Produced from Soybean Oil (million gallons)
Year
2022
2023
2024
2025
Biodiesel
Canola Oil
267
292
307
323
DCO
130
116
89
63
FOG
346
321
303
285
Soybean Oil
995
982
968
953
Renewable Diesel
Canola Oil
0
216
182
283
DCO
209
205
239
272
FOGa
859
1,122
1,088
1,160
Soybean Oil
293
604
671
874
BBD Total
All Feedstocks21
3,099
3,857
3,846
4,212
a Includes 14 million gallons of Jet Fuel projected to be supplied each year from 2023-2025
Table 6.2.6-8: Projected Domestic Production and Net Imports of Renewable Diesel
Produced from Soybean Oil (million RINs)
Year
2022
2023
2024
2025
Biodiesel
Canola Oil
400
438
461
484
DCO
195
173
134
95
FOG
519
481
454
427
Soybean Oil
1,492
1,473
1,451
1,430
Renewable Diesel
Canola Oil
0
368
309
481
DCO
356
348
406
463
FOGa
1,460
1,907
1,849
1,972
Soybean Oil
498
1,026
1,141
1,486
BBD Total
All Feedstocks21
4,921
6,215
6,205
6,836
a Includes 24 million RINs of Jet Fuel projected to be supplied each year from 2023-2025
6.3 Imported Sugarcane Ethanol
The predominant available source of advanced biofuel other than cellulosic biofuel and
BBD has historically been imported sugarcane ethanol. Imported sugarcane ethanol from Brazil
is the predominant form of imported ethanol and the only significant source of advanced ethanol.
However, data through 2022 demonstrates considerable variability in imports of sugarcane
ethanol.
319
-------
Figure 6.3-1: Historical Sugarcane Ethanol Imports
800
700
600
.= 300
200
¦ —
1
II
1 1
lllalillil
r*J(N(N(N(Nr*J(N(Nf\l(N(N(N(N(N(N(N(N(N
Source: "US Imports of Brazilian Fuel Ethanol from EIA - February 2023." Includes imports directly from Brazil
and those that are transmitted through the Caribbean Basin Initiative and Central America Free Trade Agreement
(CAFTA).
Moreover, data from EIA indicates that all 2019-2021 ethanol imports entered the U.S.
through the West Coast, as did the majority of ethanol imports in 2022. We believe that these
imports were likely used to help refiners meet the requirements of the California Low Carbon
Fuel Standard (LCFS), which provide significant additional incentives for the use of advanced
ethanol beyond the RFS.
As noted in previous annual standard-setting rulemakings, the high variability in
historical ethanol import volumes makes any projection of future imports uncertain.696 However,
import volumes for more recent years are likely to provide a better basis for making future
projections than import volumes for earlier years. To address these issues, in the final rulemaking
which established the volume requirements for 2022 we used a different methodology for
making projections of future ethanol imports than we had used in previous years.697 Specifically,
we used a weighted average of import volumes for all years where the weighting was higher for
more recent years and lower for earlier years. The weighting factor for any given year's volume
was twice as large as the weighting factor for the previous year's volume. This approach
provided a better predictor of future imports of sugarcane ethanol than either simple averages of
historical volumes or a trendline based on historical volumes.
We have again used this methodology in this action to estimate the volumes of imported
sugarcane ethanol that could be expected in the future. The volumes and weighting factors we
are using are shown in Table 6.3-1. The resulting weighted average is 95 million gallons. As we
are projecting volumes for 2023-2025 in this action, and this is the latest data available, the same
projection applies for all three years.
696 See, e.g., 85 FR 7032-33 (February 6, 2020) and 87 FR 39600 (July 1, 2022).
697 See FCC v. Fox Television Stations, Inc., 556 U.S. 502, 515 (2009).
320
-------
Table 6.3-1: Annual Advanced Ethanol Imports and Weighting Factors
Year
Imported advanced
ethanol3 (million gallons)
Weighting factor
2015
89
0.0078125
2016
34
0.015625
2017
74
0.03125
2018
78
0.0625
2019
196
0.125
2020
185
0.25
2021
60
0.5
2022
81
1
a Based on RINs generated for imported ethanol and assigned a D-code of 5 according to EMTS.
As noted above, the future projection of imports of sugarcane ethanol is inherently
imprecise, and actual imports in years 2023-2025 could be lower or higher than 95 million
gallons. Factors that could affect import volumes include uncertainty in the Brazilian political
climate, weather and harvests in Brazil, world ethanol demand and prices, constraints associated
with the El0 blendwall in the U.S., world demand for and prices of sugar, the cost of sugarcane
ethanol relative to that of corn ethanol, and the impact of the novel virus COVID-19 on
transportation fuel prices and demand.
6.4 Other Advanced Biofuel
In addition to cellulosic biofuel, imported sugarcane ethanol, and BBD, there are other
advanced biofuels that can be supplied in the years after 2022. These other advanced biofuels
include non-cellulosic CNG, naphtha, heating oil, renewable diesel co-processed with petroleum,
and domestically produced advanced ethanol. However, the supply of these fuels has been
relatively low in the last several years.
Table 6.4-1: Historical Supply of Other Advanced Biofuels (million ethanol-equivalent
Domestic
Heating
Renewable
Year
CNG/LNG
Ethanol
Oil
Naphtha
Diesel (D5)
Total
2015
0
25
1
24
8
58
2016
0
27
2
27
8
64
2017
2
25
2
32
9
70
2018
2
25
3
31
40
101
2019
5
24
3
37
58
127
2020
5
23
3
33
85
149
2021
7
26
2
33
105
173
2022
6
29
3
76
124
238
We have used the same weighted averaging approach (see Table 6.3-1) for other
advanced biofuels as we have used for sugarcane ethanol to project the supply of these other
advanced biofuels. Based on this approach, the weighted average of other advanced biofuels is
195 million RINs. This volume of other advanced biofuel is composed of 27 million RINs of
321
-------
domestic advanced ethanol, 104 million RINs of co-processed renewable diesel, and 63 million
RINs of other advance biofuels (non-cellulosic RNG, heating oil, and naphtha). We have used
these values in our candidate volumes for all three of the years addressed in this action. We do
not believe the available data and the methodology we employed can reasonably be used to
project future volumes that change over time for other advanced biofuels.
We recognize that the potential exists for additional volumes of advanced biofuel from
sources such as D5 jet fuel, liquefied petroleum gas (LPG), butanol, and liquefied natural gas (as
distinct from CNG), as well as non-cellulosic CNG from biogas produced in digesters. However,
since they have been produced, if at all, in very small amounts in the past, we do not believe the
market will make available substantial volumes from these sources in the timeframe of this
rulemaking (2023-2025).
6.5 Total Ethanol Consumption
In order to properly analyze possible future volume targets for the different categories of
renewable fuel, it was necessary to separately estimate volumes by fuel type and feedstock. For
ethanol, the process of making such estimates is complicated by the fact that there are constraints
on total ethanol consumption, a topic we discuss further in Chapter 7.5. It was therefore
necessary to estimate the total volume of ethanol that is projected to be consumed in the 2023-
2025 timeframe.
The total volume of ethanol consumed is the net result of ethanol used in El0, El 5, and
E85, while accounting for some small volume of ethanol-free gasoline (E0). In previous
rulemakings we have estimated volumes of these individual blends for the purpose of projecting
total ethanol consumption.698 However, the projection of E0, E15, and E85 for future years has
been hampered by a lack data on nationwide consumption of each individual blend. For the
purposes of this rulemaking, we have developed an alternative method that we believe is both
more accurate and avoids the need to estimate volumes separately for E0, E15, and E85. This
method, presented in Chapter 6.5.1, correlates historical pool-wide ethanol concentration derived
from EIA data with the number of stations that have offered E15 and E85.
For the purposes of estimating the costs of renewable fuel, however, it is helpful to
account for the different distribution practices required for different gasoline-ethanol blends.
Thus, for cost purposes only, we have projected volumes of E15 and E85 for 2023-2025 using
aggregate consumption data from USDA's Biofuels Infrastructure Partnership (BIP) program.
This analysis is presented in Chapter 6.5.2, and yields lower total volumes of ethanol than results
from the use of the EIA forecasts used in setting the percentage standards. The estimated
volumes of E15 and E85 are relevant for cost estimation purposes only and are not used in any
other analyses discussed in this document.
698 For instance, see "Estimates of E15 and E85 volumes in 2017," completed for use in the 2017 standards
rulemaking (81 FR 89746, December 12, 2016. See relevant discussion on page 89777).
322
-------
6.5.1 Estimation of Ethanol Consumption for Analysis of Target Volumes
The national average ethanol concentration of gasoline rose above 10.00% in 2016 and
has continued to increase since then.
Figure 6.5.1-1: Poolwide Ethanol Concentration699
10.6%
J 10.4% 10 75%y»
co 10.2% ,nnw 10.08%^. J ±0.36%
E lnny 1 10.20%
-5 iu u/o 9.83% 10.13%
19-8%
£ 9.6% * 9.15%
| 9.4% 9.33^61%
£ 9.2%
° 9.0%
8.8%
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
Source: Ethanol consumption from Table 10.3 of EIA's Monthly Energy Review, gasoline consumption from Table
3.5 of EIA's Monthly Energy Review.
As the average ethanol concentration approached and then exceeded 10.00%, the gasoline
pool became saturated with E10, with a small, likely stable volume of EO and small but
increasing volumes of El 5 and E85. The average ethanol concentration can exceed 10.00% only
insofar as the ethanol in E15 and E85 exceeds the ethanol content of E10 and more than offsets
the volume of E0. As a result, one would expect a strong correlation between ethanol
concentration and the number of retail service stations offering El 5 and E85.
To evaluate this proposition, we calculated the annual average number of stations
offering E15 and E85. For E15, annual averages were based on interpolations of the data
provided by Prime the Pump (see Figure 7.5.3-2), while for E85 annual averages were calculated
from the monthly estimates provided by DOE's Alternative Fuel Data Center (see Figure 7.5.2-
2). The results are shown in Table 6.5.1-1.
.5.1-1: Poolwide Ethanol Concentration699
10.08%
10.25%
10.36
10.02%
10.20%
9.83%
10.13%
9.65%
9.91%
9.33?^
^61%
9.75%
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
699 U.S. Energy Information Administration. https://www.eia.gov/totalenergv/data/monthlv/archive/00352303.pdf.
See, e.g., US Energy Information Administration - March 2023 Monthly Energy Review.
323
-------
Table 6.5.1-1: Annual Average Number of Stations Offering Higher Level Ethanol Blends
El5 stations
E85 stations
2013
36
2,616
2014
88
2,713
2015
145
2,932
2016
308
3,091
2017
776
3,251
2018
1,376
3,567
2019
1,838
3,717
2020
2,180
3,841
2021
2,461
4,063
2022
2,724
4,418
Examination of these sets of data suggested that the El 5 station data was nonlinearly
correlated with poolwide ethanol concentration, while the E85 station data was roughly linearly
correlated.
Figure 6.5.1-2: Correlations Between E15 and E85 Stations and Poolwide Ethanol
Concentration
10.40% j
1030% ;
10.20% ;
'
'
10.10% :
10.00%
9.90%
9.80%
1,000 1,500 2,000
E1S stations
2,500 2,700 2,900 3,100 3,300 3,500 3,700 3,900 4,100 4,300
ESS stations
Based on these observations, we applied a least-squares regression to the ethanol
concentration using the natural logarithm of the number of El 5 stations and a linear term for the
number of E85 stations as the independent variables. The result was the following equation:
324
-------
Ethanol concentration (%) = (5.721 xlO"4) x ln(E15 station count)
+ (2.028><10"6) x (E85 station count)
+ 0.09031
Given that this regression has an r squared value of 0.95, it represents a strong basis for
projecting the poolwide ethanol concentration for 2023-2025.
Using the projected number of El 5 and E85 stations discussed in Chapters 7.4.3 and
7.4.2, the regression equation above yields the ethanol concentration projections shown in Table
6.5.1-2.
Table 6.5.1-2: Projected Poolwide Ethanol Concentration
El5 stations
E85 stations
Ethanol concentration
2023
3,586
4,499
10.41%
2024
4,197
4,696
10.46%
2025
4,713
4,892
10.51%
The projected ethanol concentration can then be combined with total projected gasoline
energy demand from EIA's AEO 2023 to estimate total ethanol consumption. We use this
projection method due to a discrepancy found between EIA domestic gasoline consumption and
net gasoline exports. A further discussion of our investigation is found in Chapter 1.11.
Table 6.5.1-3: Projected Total Ethanol Consumption
Projected ethanol
concentration
Gasoline energy
demand (Quad Btu)a
Projected ethanol
consumption (million
gallons)b
2023
10.41%
16.1357
13,974
2024
10.46%
16.2350
14,128
2025
10.51%
15.9890
13,978
a See AEO2023 Table 2, "Delivered Energy Consumption, All Sectors," "Motor Gasoline"
b Based on the energy-to-volume conversion factors for denatured ethanol and BOB (Blendstock for Oxygenate
Blending) found in AEO2023 Table 68.
6.5.2 Estimation of Gasoline Blend Volumes for Cost Purposes
For the purposes of estimating costs only, we projected the volumes of E15 and E85 that
may be consumed in 2023-2025. These volume projections were based on data collected by
USDA through their BIP program and made available to EPA.700 While this data includes only a
subset of all El 5 and E85 stations, it is considerably more comprehensive than the alternatives.
For instance, the BIP data covers almost 800 retail stations in 19 states. The only other data of
which we are aware on El5 sales at retail is from two states (Iowa and Minnesota)701'702 while
700 "Communication with USDA on the BIP program 1-19-22," available in the docket.
701 Iowa Department of Revenue. https://tax.iowa.gov/report-categorv/retailers~annual~gaHons. See, e.g., "Iowa
Department of Revenue - 2021 Retailers Fuel Gallons Annual Report."
702 Minnesota Commerce Department, https://mn.gov/commerce/business/weights-
measiires/ftiel/biodiesel/ethanot.isp. See, e.g., "Minnesota Commerce Department - 2022 Minnesota E85 & Mid-
Blends Station Report."
325
-------
the data of which we are aware on E85 sales at retail is from six states (Iowa, Minnesota,
California, New York, Kansas, and North Dakota).703
USDA collected data on sales of E15 and E85 over a six-year period ending in 2020.
However, the BIP program was not completed until the end of 2018 and the largest number of
respondents to the survey occurred in 2019 and 2020. Moreover, there was a noticeable decrease
in El 5 and E85 consumption in 2020 that was consistent with the decrease in all gasoline
consumption brought about by the COVID pandemic, and which may not be representative of
future years. In the proposal, 2019 was used as a more representative value over 2020. Since
then, 2021 data has become available and presents values which are more accurate for estimating
future values to the program.
We recognize that in 2021 the 1 psi waiver applied nationwide to E15, but that it may not
apply nationwide in 2023-2025.704'705 The 1 psi waiver could have resulted in higher sales
volumes in 2021 than would have been the case if the 1 psi waiver had not applied. As a result,
the use of BIP data on El 5 sales volumes in 2019 may overestimate the potential for sales in
2023-2025 when the lpsi waiver may not apply nationwide.706 However, there are reasons to
believe that the use of data from 2021 is appropriate for 2023-2025. To begin with, El 5 sales
volumes per station have increased in previous years, and thus could continue to increase in the
future as well. The BIP demonstrates an increasing trend that is disrupted only by the results for
2015 when only 8 retail stations reported El 5 sales volumes (compared to 767 in 2019), and for
2020 when the pandemic reduced sales volumes of all fuels.
Figure 6.5.2-1: Per-Station E15 Sales Volumes from BIP Program
180,000
170,000
- 160,000 :
CL>
CT 150,000
c 140,000
.2
2 130,000
w
120,000
J 110,000
o 100,000
30,000
80,000
2014
2015 2016 2017 2018 2019 2020 2021 2022
703 See discussion of data sources in "Estimate of E85 consumption in 2020," available in the docket.
704 84 FR 26980, June 10, 2019.
705 "Court decision on 1 psi waiver for E15," available in the docket.
706 "Request from States for Removal of Gasoline Volatility Waiver" available in the docket
326
-------
This 2020 decrease in El5 volumes in the BIP data is consistent with the decrease in total
gasoline consumption brought about by the COVID pandemic. Additional BIP program data was
accessed recently showing an increase in El5 sales to pre-pandemic era trends. Given this
rebound and that of gasoline demand in 2021 and 2022, we can forgo using 2019 data for more
accurate 2021 data. Moreover, a similar upward trend is evident in per-station E15 sales volumes
in Minnesota.
Figure 6.5.2-2: Per-Station E15 Sales Volumes in Minnesota
Monthly E15 Sales per Station from Minnesota
40,000
35,000 /
r~ on ^ M
— JU,UUU
9 ^ nnn
t7> i
ai in nnn m
Gallons pi
l-» l-» r-
Ui O ui c
8 8 8 \
o o o o c
Jan/2015 "| ||
Apr/2015" J|
Jul/2015 J
Oct/2015 ^"1
Jan/2016 I
Apr/2016
Jul/2016
Oct/2016
Jan/2017 I
Apr/2017 V
Jul/2017 ^
Oct/2017
Jan/2018
Apr/2018
Jul/2018
Oct/2018
Jan/2019
Apr/2019
Jul/2019
Oct/2019
Jan/2020
Apr/2020
Jul/2020
Oct/2020
Jan/2021
Apr/2021
Jul/2021
Oct/2021
Jan/2022
Apr/2022
Source: Minnesota Commerce Department. See for example "2021 Minnesota E85 + Mid-Blends Station Report,"
available in the docket.
The reasons for such increases over time in per-station El5 sales volumes are not clear.
They may be attributable to any or all of the following: the relative price of E15 versus E10,
sustained high D6 RIN prices, changes in consumer preferences for higher octane fuel, signage
and advertising, marketing incentives on the part of the retailers, or the increasing awareness that
consumers have developed over time of their choices at retail stations.
The BIP data allowed us to estimate per-station annual sales volumes of El 5 and E85. In
combination with future projections of the number of stations offering these fuel blends (derived
and discussed in Chapters 7.5.2 and 7.5.3), we were able to project volumes of E15 and E85 as
shown in Tables 6.5.2-1 and 2.
Table 6.5.2-1: Projected Volume of E15 and Ethanol as E15
El5 stations
El5 sales volumes
Annual El5 sales
Ethanol in Excess
per year per station
volumes (mill gal)
of E10 (mill gal)
2023
3,586
172,988
535
28
2024
4,197
180,593
619
31
2025
4,713
188,198
708
35
327
-------
Table 6.5.2-2: Projected Volume of E85 and Ethanol as E85
E85 stations
E85 sales volumes
Annual E85 sales
Ethanol in Excess
per year per station
volumes (mill gal)
of E10 (mill gal)
2023
4,499
54,716
352
233
2024
4,696
47,097
368
243
2025
4,892
39,477
383
253
These are the volumes used to estimate distribution costs associated with El5 and E85 for
2023-2025 as discussed in Chapter 10.1.4.
As we noted above, we do not use these El5 and E85 estimates to assess total ethanol
volumes or for any purpose other than estimating costs.707 We note, however, that the El5 and
E85 volume projections shown in Tables 6.5.2-1 and 2 correspond to lower total ethanol
consumption than the volumes shown in Table 6.5.1-3. Below we explain how we identified this
discrepancy, why we chose to use estimates from EIA for total ethanol consumption rather than
those derived from the El5 and E85 estimates shown in Tables 6.5.4-1 and 2, and why we
nonetheless believe it is reasonable to use these El 5 and E85 estimates for purposes of our costs
analyses. See Chapter 1.11 for our investigation into the EIA forecast/projection of gasoline
demand and our adjustment to those forecasts.
Total ethanol consumption can be calculated in a bottom-up fashion by combining the
estimates of E15 and E85 with an estimate of E10. An estimate of E10 consumption, in turn, can
be back-calculated from an estimate of E0 and total gasoline energy demand derived from the
EIA's 2022 Annual Energy Outlook. As noted above, this exercise produces an amount of total
ethanol consumption that is significantly less than our projection of total ethanol consumption
shown in Table 6.5.1-3.
As for El5 and E85, there is very little available data on the consumption of E0. Iowa's
Department of Revenue has collected data on E0 sales at retail for many years, but it is unclear
how representative Iowa E0 sales are of the entire nation. For instance, the pattern of
consumption of E0 in Iowa does not appear to have followed the nationwide consumption pattern
of total gasoline since 2012. 2021 data may be the exception to this pattern with both gasoline
consumption and E0 raising from 2020 points.
707 Our approach to projecting total ethanol consumption as discussed in Chapter 6.5.1 does use a projection of the
number of stations offering E15 and E85, but does not involve any projection of separate volumes for E15 or E85.
328
-------
Figure 6.5.2-3: EO Consumption in Iowa
145,000
140,000
ro 135,000
5
o
c
C 130,000
o
to
to
aj 125,000
Q.
O
LU
&> 120,000
115,000
110,000
20
1 trcnnn
150000
145000 §
Q.
140000 |
l/l
C
o
<_>
135000
-------
in 2021 was EO.712 Insofar as this proportion is representative of the nation as a whole, 1.44%
corresponds to about 1,940 million gallons of EO, based on EIA's estimate of 134.75 billion
gallons of gasoline sold in 2021.713 Since 1,940 million gallons is very similar to 2,128 million
gallons, we had confidence that 2,209 million gallons is a reasonable estimate of E0
consumption.
Estimates of E0, El 5, and E85 consumption, combined with total gasoline energy
demand from AEO2023, allowed us to calculate the consequent volume of E10 consumed as
shown in Table 6.5.2-3.
Table 6.5.2-3: Estimating E10 Consum
ption
E0
E15a
E85b
Gasoline energy
E10c
(mill gal)
(mill gal)
(mill gal)
(Quad Btu)
(mill gal)
2023
2,209
535
352
16.1357
131,015
2024
2,209
619
368
16.2350
131,744
2025
2,209
708
383
15.9590
129,601
a Assumes that the denatured ethanol concentration of E15 is 15%.
b Assumes that the denatured ethanol concentration of E85 is 74%, consistent with the assumption made by EIA.
0 Assumes that the denatured ethanol concentration of E10 is 10.1%, based on data collected by the RFG Survey
Association indicating that the average ethanol concentration was 9.9% in 2021, and assuming 2% denaturant.
We could then derive the total volume of ethanol consumed as a function of the projected
volumes of El 5 and E85.
Table 6.5.2-4: Projected Total Ethanol Consumption Derived From E15 and E85 Volumes
El
15
E85
El
10
Total
ethanol
Fuel
Ethanol
Fuel
Ethanol
Fuel
Ethanol
2023
535
80
352
260
131,015
13,232
13,573
2024
619
93
368
272
131,744
13,306
13,671
2025
708
106
383
283
129,601
13,089
13,479
a Assumes that the denatured ethanol concentration of E15 is 15%.
b Assumes that the denatured ethanol concentration of E85 is 74%, consistent with the assumption made by EIA.
0 Assumes that the denatured ethanol concentration of E10 is 10.1%, based on data collected by the RFG Survey
Association indicating that the average ethanol concentration was 9.9% in 2021, and assuming 2% denaturant.
The total ethanol consumption calculated as a function of projected E0, E10, E15, and
E85 (Table 6.5.2-4) can then be compared to the total ethanol consumption calculated as a
function of projected pool-wide ethanol concentration (Table 6.5.1-3). The results are shown in
Table 6.5.2-5.
712 "National Fuels Survey Program Ethanol Data for the 2021 Compliance Period," available in the docket.
713 "steq jan 2022 Table 4a," available in the docket.
330
-------
Table 6.5.2-5: Comparison of Pro.
ected Total Ethanol Consumption (million gal
Based on projected
ethanol concentration
Based on projected
ethanol blend volumes
Difference
2023
13,974
13,573
-401 (1.8%)
2024
14,128
13,671
-501 (3.5%)
2025
13,978
13,479
-498 (3.5%)
We do not know why ethanol consumption based on projected ethanol concentration
differs so much from the ethanol consumption based on projected ethanol blend volumes. It
could be due to errors in estimates of EO or that the data collected through the BIP program may
not be representative of the nation as a whole. However, the magnitude of the difference suggests
that it may also be due to underestimates of the total gasoline demand. Regardless, we believe
that the BIP data represents the best available source of information on sales of E15 and E85, and
that the estimates of E15 and E85 consumption shown in Tables 6.5.2-1 and 6.5.2-2 are a
reasonable basis for estimating distribution costs for these two blends. We are not aware of better
sources of data or clearly superior methodologies to estimating El 5 and E85 use. As such, we
have done the best we can with the limited information available to us. We acknowledge the
significant limitations in the data available to us and the uncertainties this creates for our
estimates. In any event, as shown in Chapter 10, the costs unique to E15 and E85 relative to E10
(associated with distribution, including blending and retail costs) reflect only a small portion of
the costs of ethanol and a miniscule portion of the total costs associated with the candidate
volumes. Thus, even were we to estimate significantly different El 5 and E85 volumes, that
would have very limited impacts on our assessments of costs and no impact on our provisional
judgment with respect to the appropriate volumes.
6.6 Corn Ethanol
As described in more detail in Chapter 1.7, total domestic ethanol production capacity
increased dramatically between 2005 and 2010, and increased at a slower rate thereafter. In
2020, production capacity had reached 17.4 billion gallons.714'715 This production capacity was
significantly underused in 2020 due to the COVID-19 pandemic, which depressed gasoline
demand in comparison to previous years. Actual production of ethanol in the U.S. reached 12.85
billion gallons in 2020, compared to 14.72 billion gallons in 2019.716
The expected annual rate of future commercial production of corn ethanol will be driven
primarily by gasoline demand as most gasoline is expected to continue to contain 10% ethanol in
the foreseeable future. Commercial production of corn ethanol is also a function of exports of
ethanol and to a much smaller degree the demand for E0, E15, and E85. While production of
corn ethanol may be limited by production capacity in the abstract, it does not appear that
production capacity will be a limiting factor in 2023-2025 for meeting the candidate volumes.
As described in Chapter 6.5.1, we estimated total ethanol consumption for 2023-2025 by
extrapolating from historical poolwide ethanol concentration and the number of retail stations
714 "2021 Ethanol Industry Outlook - RFA," available in docket EPA-HQ-OAR-2021-0324.
715 "Ethanol production capacity - EIA April 2021," available in docket EPA-HQ-OAR-2021-0324.
716 "RIN supply as of 3-22-21," available in docket EPA-HQ-OAR-2021-0324.
331
-------
offering El 5 and E85. This total volume is a combination of corn ethanol, cellulosic ethanol, and
advanced ethanol. Our estimate of corn ethanol consumption for 2023-2025 for the purposes of
estimating the mix of biofuels that could be made available is shown in Table 6.6-1.
Table 6.6-1: Calculation of Projected Corn Ethanol Consumption for 2023-2025 (million
2023
2024
2025
Total ethanol
13,974
14,128
13,978
Imported sugarcane ethanol
95
95
95
Domestic advanced ethanol
27
927
27
Corn ethanol
13,845
13,955
13,779
Total production of corn ethanol in 2023-2025 is likely to be higher than the
consumption levels shown in Table 6.6-1 because the U.S. has exported significant volumes in
recent years. For instance, in 2021 ethanol export volumes were 1.25 billion gallons.717
6.7 Conventional Biodiesel and Renewable Diesel
While the vast majority of conventional renewable fuel supplied in the RFS program has
been corn ethanol, there have been smaller volumes of conventional biodiesel and renewable
diesel used in the U.S. in some years. Conventional biodiesel and renewable diesel can only be
produced at facilities grandfathered under the provisions of 40 CFR 80.1403 as there currently
exist no valid RIN-generating pathways for the production of conventional (D6) biodiesel or
renewable diesel. These biofuels are not required to meet the 50% GHG reduction threshold to
qualify as BBD under the statutory definition, but the feedstocks used to produce grandfathered
biodiesel or renewable diesel must still meet the regulatory definition of renewable biomass, and
the biofuel produced must meet all other statutory and regulatory requirements. The quantity of
conventional biodiesel and renewable diesel consumed each year from 2014-2022 is shown in
Table 6.7-1.
2014
2015
2016
2017
2018
2019
2020
2021
2022
Domestic D6 Biodiesel
1
0
0
0
0
0
0
0
0
Domestic D6
0
0
0
0
0
0
0
0
0
Renewable Diesel
Imported D6 Biodiesel
52
74
113
0
0
0
0
0
0
Imported D6
2
86
45
2
0
0
0
0
0
Renewable Diesel
All D6 Biodiesel and
55
160
158
2
0
0
0
0
0
Renewable Diesel
In 2014-2016 the volume of conventional biodiesel and renewable diesel used in the U.S.
was relatively small, but still significant. Use of these fuels in the U.S. dropped to very low
levels in 2017 and has been less than 1 million gallons per years from 2018-2022. Nearly all of
717 "Fuel Ethanol Exports by Destination from EIA 6-27-22," available in the docket.
332
-------
the conventional biodiesel and renewable diesel used in the U.S. has been imported, with the
only exception being one million gallons of domestically produced biodiesel in 2014. However,
conventional (D6) RINs have continued to be generated for biodiesel and renewable diesel in
recent years. From 2018 through 2022 the volumes of renewable diesel for which conventional
biofuel RINs were generated each year (in million gallons) were 107, 116, 76, 135, and 75
respectively. These RINs were retired for reasons other than compliance with the annual volume
obligations, suggesting that they were used outside of the U.S. or for purposes other than
transportation fuel.
The potential for conventional biodiesel and renewable diesel production and use in the
U.S. is far greater than the quantity of these fuels actually supplied in previous years. The total
production capacity of registered grandfathered biodiesel and renewable diesel producers is over
1.6 billion gallons in the U.S., with an additional 0.9 billion gallons internationally. While
domestic feedstock availability may be limited, worldwide feedstock availability does also not
appear to be a limiting factor, as USD A estimates that approximately 218 million metric tons of
vegetable oil will be produced globally in the 2022/2023 agricultural marketing year.718 If all of
it were to be used to produce biodiesel and renewable diesel, this quantity of vegetable oil could
be used to produce approximately 63 billion gallons.719 While the majority of this vegetable oil is
used for food and other non-biofuel purposes, any of this vegetable oil that meets the regulatory
definition of renewable biomass could be used to produce conventional biodiesel or renewable
diesel at a grandfathered facility so long as it meets all other RFS program requirements. The
quantity of conventional biodiesel and renewable diesel that could be supplied to the U.S. in
2023-2025 is not without limit, but this data suggests that large quantities of this fuel are being
or could be produced,720 and that the use of these fuels in the U.S. is largely a function of
demand for this fuel in the U.S. versus other markets.
718 USDA World Agricultural Supply and Demand Estimates. February 8, 2023.
719 This calculation assumes one gallon of biodiesel or renewable diesel can be produced from 7.6 pounds of
vegetable oil.
720 The OECD-FAO Agricultural Outlook 2021-2030 projects global biodiesel consumption to reach approximately
50 billion liters (about 13.2 billion gallons) in 2022.
333
-------
Chapter 7: Infrastructure
This chapter describes the analysis of the impact of renewable fuels on the distribution
infrastructure of the U.S. The CAA indicates that this assessment must address two aspects of
infrastructure:
1. Deliverability of materials, goods, and products other than renewable fuel.
2. Sufficiency of infrastructure to deliver and use renewable fuel.
This chapter begins by addressing the sufficiency of infrastructure to deliver and use
different types of renewable fuels. We then address how the use of renewable fuels affects the
deliverability of materials, goods, and products other than renewable fuel.
Note that while we are projecting higher volumes of renewable fuel consumption relative
to the No RFS baseline, in analyzing the impacts of the candidate volumes on infrastructure we
have considered whether the candidate volumes would require additional infrastructure relative
to the infrastructure that currently exists. We believe that the existing infrastructure is the
relevant point of reference for the No RFS baseline since it is unlikely that the infrastructure
enabling and supporting consumption of renewable fuel in 2022 would change even if we did not
establish volume requirements for future years, at least not in the 2023-2025 timeframe. The
number of vehicles that can consume particular renewable fuels, pipelines, storage tanks, fuel
delivery vehicles, and retail service stations generally change only on longer timescales, and only
insofar as the outlook for renewable fuel demand changes. Therefore, this chapter discusses
infrastructure impacts primarily in terms of the changes that might be needed or expected to
occur in 2023-2025 in comparison to their recent or current status.
7.1 Biogas
Renewable biogas infrastructure considerations differ from those for other biofuels not
only because it is a gas rather than a liquid, but also because renewable biogas can be processed
to be physically identical to natural gas, which is used for many purposes including
transportation.721 Natural gas was used in CNG/LNG vehicles for many years prior to the
introduction of renewable biogas. The RFS program allows RINs to be generated for renewable
biogas that is fungible with the wider natural gas pool, provided that a contract is in place to
demonstrate that the same volume of natural gas is used for transportation purposes and all other
regulatory requirements are met.722 As the cost of running spur pipelines for anything beyond
short distances becomes prohibitively expensive, only those biogas sources that are in relatively
close proximity to the existing natural gas pipeline infrastructure are likely to be developed.
Once connected to the natural gas pipeline network, renewable biogas uses the existing natural
gas distribution system and CNG/LNG vehicle refueling infrastructure, and is used in the same
CNG/LNG vehicle fleet as natural gas. According to data from the DOE Alternative Fuels Data
721 Growth in biogas may require investment in additional gas cleanup operations prior to pipeline injections,
particularly in California where pipeline standards currently preclude the injection of most biogas. The potential for
such biogas cleanup costs are discussed in Chapter 10.1.2.5.1.
722 See 40 CFR 80.1426(f).
334
-------
Center, there are currently approximately 1,500 public and private CNG fueling stations and
approximately 100 public and private LNG refueling stations in the U.S.723
Once the processed biogas is in the gas pipeline, it is virtually indistinguishable from
natural gas. However, expanding CNG/LNG vehicle infrastructure to support growth in the
renewable biogas beyond the current level of CNG/LNG used in the transportation sector—
estimated at 1.4-1.75 billion ethanol-equivalent gallons of CNG/LNG per year in 2023-2025—
would represent a substantial challenge.724 The incentives for increasing the use of CNG/LNG in
the transportation sector, including incentives from the RFS program and state programs such as
the California LCFS program, may be insufficient to cause a substantial increase in the
CNG/LNG vehicle fleet and refueling infrastructure. CNG/LNG vehicles are predominately used
in fleet applications where there is a unique situational advantage (e.g., a natural gas supplier's
utility fleet or landfill's waste hauler fleet). In addition, it would be more challenging to establish
the necessary contracts to demonstrate that natural gas was used in CNG/LNG vehicles outside
of fleet operations. The cost associated with removing the impurities in renewable biogas to
make it suitable for use in CNG/LNG vehicles and to facilitate its fungible transportation in the
natural gas distribution system could also be a barrier to its expanded use. Nevertheless, we do
not expect infrastructure to constrain the use of CNG/LNG derived from biogas to levels below
those projected to be available in Chapter 6.1.3.
7.2 Biodiesel
The RFS2 rule projected that 1.5 billion gallons of biodiesel would be used in 2017 and
1.82 billion gallons would be used in 2022 to meet the statutory biofuel volume requirements.725
We noted that biodiesel plants tended to be more dispersed than ethanol plants, thereby
facilitating delivery to local markets by tank truck and lessening the need to distribute biodiesel
over long distances. Biodiesel imports also helped to serve coastal markets. We projected that as
biodiesel volumes grew, there would be more need for long-distance transport of domestically-
produced biodiesel. We estimated that such long-distance transport would be accomplished by
manifest rail and, to a lesser extent, by barge, since the economy of scale would not justify the
use of unit trains. We estimated that biodiesel and biodiesel blends would not be shipped by
pipeline to a significant extent due to concerns over potential contamination of jet fuel that is
also shipped by pipeline.
In 2010, much of the biodiesel blending was taking place at facilities downstream of
terminals, such as storage facilities operated by individual fuel marketers. We projected that this
would take place to a lesser extent as volumes grew with most biodiesel being blended at
terminals to the 5% (B5) blend level that is approved for use in diesel engines by all
manufacturers for distribution to retail and fleet fueling facilities. We acknowledged that the
expansion of biodiesel volumes could pose issues for petroleum terminals, but that these issues
723 AFDC Alternative Fueling Station Locator.
https://afdc.energv.gov/stations/#/analvze?fuel=LNG&fuel=CNG&access=public&access=private&coiintrv=US.
Data current as of September 20, 2022.
724 See Chapter 6.1.3 for further discussion of the estimated use of CNG/LNG as transportation fuel in 2023-2025
and Chapter 10.1.4 for discussion of the costs associated with refueling stations.
725 See Chapter 1.2.2 of the RFS2 Regulatory Impact Analysis (EPA-420-R-10-006).
335
-------
could be resolved.726 Since vehicle refueling infrastructure is compatible with biodiesel blends
up to 20% (B20), we estimated that there would be no changes needed at retail and fleet facilities
to accommodate the projected increase in biodiesel use.
There are significant instances where actual biodiesel production and use have developed
differently than we projected in the RFS2 rule. Most importantly, biodiesel consumption reached
over 2 billion gallons in 2016 and has remained between 1.7-2 billion gallons per year from
2017-2022, often exceeding the 1.82 billion gallons that we projected would be used in 2022.727
Another significant difference is that much biodiesel blending is taking place downstream of
terminals at fuel marketer storage facilities and even at fuel retail facilities.
One factor that could somewhat ease biodiesel transportation to terminals is the fact that
in some limited cases, shipment of low-level biodiesel blends up to 5% is currently taking place
on some petroleum product pipelines that do not also carry jet fuel.728 If the transportation of
biodiesel blends via pipeline were expanded more broadly, this change could significantly reduce
the cost of biodiesel distribution. However, jet fuel is a significant product on much of the
petroleum pipeline system and concern over biodiesel contamination of jet fuel remains a
significant limitation on the ability to expand the shipment of biodiesel blends by pipeline.
Industry is currently investigating whether jet fuel can tolerate higher levels of biodiesel
contamination, which may allow low-level biodiesel blends to be shipped on pipelines that also
carry jet fuel.729 Finally, there appears to be substantial volumes of B10-B20 being used despite
the fact that a significant number of vehicle manufacturers only warranty their engines for up to
B5.730 This has resulted in an uneven distribution of biodiesel use across the nation, with some
parts using more than 5% while other locales use little or no biodiesel.731
While we are projecting that the candidate volumes for 2023-2025 would require
substantial biodiesel volumes relative to the No RFS baseline, we are also projecting small
decreases in the volume of biodiesel relative to the volume of biodiesel used in 2022. Rather, the
expansion of BBD is projected to occur through renewable diesel, as discussed in Chapter 7.4.
As such, we do not anticipate any challenges associated with the infrastructure to distribute and
use biodiesel through 2025.
However, it is possible that domestic biodiesel production and/or biodiesel imports may
increase in 2023-2025. Domestic biodiesel production capacity is significantly higher than
current production levels.732 A review of monthly biodiesel imports suggests that import
726 There is additional difficulty in storing and blending biodiesel because of the need for insulated and/or heated
equipment to prevent cold flow problems in the winter. This issue is typically not present for B5.
727 Biodiesel consumption numbers based on EMTS data.
728 Ethanol, Biofuels, and Pipeline Transportation. Association of Oil Pipelines and American Petroleum Institute.
https://www.api.Org/~/media/files/oil-and-natnral-gas/pipeline/aopl api ethanot transportation.pdf.
729 An ASTM task group is seeking additional data to address negative comments on a 2018 ballot to increase the
limit on biodiesel contamination in jet fuel from 50 mg/kg to 100 mg/kg. The ASTM limit on biodiesel
contamination of jet fuel was last revised in 2015. Revised ASTM Standard Expands Limit on Biofuel
Contamination in Jet Fuels, ASTM New Release, February 2, 2015.
730 See Pilot Flying J Fuel Offerings, Memorandum to EPA Docket EPA-HQ-OAR-2021-0427.
731 See Average Biodiesel Blend Level By State Based on EIA Data, Memorandum to EPA Docket EPA-HQ-OAR-
2021-0427.
732 See Chapter 6.2.
336
-------
infrastructure can support significantly higher volumes of imports.733 For example, over 700
million gallons of biodiesel was imported in 2016.734 Monthly import data suggests that 1.3
billion gallons per year of imports could be supported using the existing infrastructure if we were
to assume that the 112 million gallons of biodiesel imports that took place in December 2016
could be maintained year-round. Some additional expansion in import infrastructure may also
occur through 2025. Therefore, we do not believe that domestic production capacity or import
infrastructure constraints would be a substantial impediment to an expansion in biodiesel
volumes at current levels.735
We anticipate that if biodiesel production and imports increase significantly, investment
in the infrastructure to transport biodiesel from the points of production to locations where it can
be consumed would be needed. These investments would primarily be associated with securing
sufficient downstream biodiesel storage and the requisite number of rail cars and tank trucks
suitable for biodiesel transport.736
Expanding biodiesel blending infrastructure to accommodate significantly higher
biodiesel volumes may also pose challenges. Many terminals that have yet to distribute biodiesel
would likely need to install the infrastructure. All vehicle refueling infrastructure is compatible
with B20 blends, thereby easing the expansion to retail of biodiesel blends made at terminals.
However, significant infrastructure changes would be needed to biodiesel storage and blending
facilities downstream of terminals and at retail facilities if substantial additional volumes of
biodiesel blends were to be made downstream of terminals.
Further, the cold flow of petroleum-based diesel dispensed to vehicles must often be
improved in the winter through the addition of #1 diesel fuel and/or cold-flow improver
additives. Biodiesel blends tend to have poorer cold flow performance than straight petroleum-
based diesel fuel. This requires the use of additional cold-flow improvers and sometimes limits
the biodiesel blend ratio that can be used under the coldest conditions.737 Biodiesel cold flow
properties are dependent on the source of the feedstock with biodiesel produced from palm oil
being subject to wax formation at higher temperatures than soy-based biodiesel.738 Thus,
additional actions are necessary to ensure adequate cold-flow performance of palm-based
biodiesel blends compared to soy-based biodiesel. Such additional actions may be uneconomical
in some cases.739 Therefore, a substantial increase in the use of biodiesel, especially biodiesel
produced from palm oil, during the winter may be a challenge.
733 EIA, U.S. Imports of Biodiesel 2009 thorough 2021.
734 Id.
735 The expansion of biodiesel imports to the extent discussed above is for purposes of the infrastructure analyses
only. There would be significant challenges in obtaining foreign produced biodiesel volumes to approach such a
substantial increase in imported biodiesel. See Chapter 1.4.2.
736 Biodiesel rail cars and to a lesser extent tank trucks must often be insulated and or heated during the winter to
prevent cold flow problems. The use of such insulated/heated vessels is sometimes avoided by shipping pre-heated
biodiesel.
737 B5 blend levels can typically be maintained.
738 Biodiesel Cold Flow Basics, National Biodiesel Board, 2014.
739 Evaluation and enhancement of cold flow properties of palm oil and its biodiesel, Puneet Verma, et.al., Biofuel
Research Laboratory, Indian Institute of Technology, Elsevier Energy Reports, January 2016.
337
-------
7.3 Renewable Diesel
The RFS2 rule projected that the volume of "drop-in" cellulosic and renewable diesel
fuel would range from 0.15-3.4 billion gallons in 2017 and 0.15-9.5 billion gallons in 2022.740
Such fuels are referred to as drop-in fuels because their physical properties are sufficiently
similar to petroleum-based diesel to be fungible in the common diesel fuel distribution system.741
Thus, little change is needed to the fuels infrastructure system to support the use of drop-in
biofuels. The RFS2 rule projected that the distribution infrastructure could expand in a timely
fashion to accommodate that projected range of growth in drop-in cellulosic and renewable
diesel fuel.742
In practice, much of the renewable diesel produced in the U.S. has been transported by
truck, rail, and ship, rather than by pipelines. This is in part due to the location of the renewable
diesel production and demand and the lack of available pipelines to transport renewable diesel
from production sites to demand centers. Renewable diesel can generate credits under state
LCFS programs, and the magnitude of this incentive, especially in California, has caused most
renewable diesel production in the U.S. to be shipped in segregated batches to California rather
than being blended into diesel where it is produced. Regulatory challenges have also limited the
transportation of renewable diesel via pipeline. Product transfer document (PTD) requirements
for fuel shipped by pipeline and fuel pump labeling requirements often require that the blend
level be indicated, but the concentration would often be uncertain in a fungible distribution
system. Transportation of renewable diesel via common carrier pipelines can make documenting
the use and blend levels of renewable diesel difficult, if not impossible.
The projected increase in domestic renewable diesel production through 2025 is
significant both relative to the No RFS and 2022 baselines.743 We expect that much of this new
renewable diesel will also be used in California and other states with state incentive programs
(e.g., Oregon). Renewable diesel produced in California will likely be distributed locally, and
much of the renewable diesel produced on or near the Gulf Coast is likely to be transported via
ship. The remaining renewable diesel production facilities are not located near the coast, and we
therefore project that the fuel they produce will likely be transported via truck and/or rail to
markets where the fuel is used. This may require some expansion to the existing infrastructure,
such as additional rail cars to transport renewable diesel. The fact that the new or expanding
renewable diesel production facilities are generally located in the western U.S., relatively close
to California and Oregon, likely reduces the impact of distributing these fuels on the
transportation infrastructure, though this may be somewhat offset by the need to transport
feedstocks to the renewable diesel production facilities. While some adjustments will likely be
needed to accommodate the expected increase in renewable diesel production, we do not expect
that these adjustments will inhibit the growth of renewable diesel production or appreciably
impact transportation networks in the U.S. more broadly.
740 See Chapter 1.2.2 of the RFS2 Regulatory Impact Analysis (EPA-420-R-10-006).
741 Such drop-in fuels are typically blended with petroleum-based diesel prior to use.
742 See Chapter 1.6 of the RFS2 Regulatory Impact Analysis (EPA-420-R-10-006).
743 See Chapter 6.2.
338
-------
7.4 Ethanol
We are projecting that the candidate volumes for 2023-2025 would result in increased
use of higher-level ethanol blends such as E15 and E85; E10 is economical to be blended in the
absence of the RFS program.
The infrastructure needed to deliver ethanol includes that required for distribution of
denatured ethanol from production facilities to terminals, storage and blending equipment, and
distribution of gasoline-ethanol blends to retail service stations. With regard to infrastructure
needed to use ethanol, essentially all retail service stations are certified to offer E10 and all
vehicles and equipment are designed to use E10. As a result, any infrastructure-related impacts
on the use of ethanol in 2023-2025 are associated with service station storage and dispensing
equipment for higher-level ethanol blends such as El 5 and E85, and the vehicles capable of
using those blends. The majority of the El 5 and E85 volume projected to be used in 2023-2025
is already being used in 2022; consequently, the infrastructure is already in place. However, the
expanded volume in 2023-2025 would require additional infrastructure, primarily the expansion
of retail stations as discussed below.
Based on our analysis below of the sufficiency of infrastructure to deliver and use
ethanol, we have determined that there are constraints associated with El 5 and E85 that limit the
rate of future growth in their consumption. These constraints are appropriately reflected in our
projections of total ethanol consumption in Chapter 6.5 since those projections represent only
moderate changes in the nationwide average ethanol concentration in comparison to earlier
744
years.
7.4.1 Ethanol Distribution
To support the RFS2 rule, ORNL conducted an analysis of potential distribution
constraints that might be associated with attaining the statutory volume targets through 2022.745
The ORNL analysis analyzed ethanol transport pathways from production to blending facilities
at terminals by rail, waterways, and roads, and projected that most ethanol would require long-
distance shipment to demand centers. The primary mode of long-distance transport in 2010 was
via manifest rail and, to a lesser extent, by barge, although transport by unit train was beginning
to spread. ORNL projected that rail would continue to be the predominate means of long-
distance ethanol transport through 2022, with a substantial increase in the use of unit trains and
continued supplemental transport by barge. ORNL concluded that there would be minimal
additional stress on most U.S. transportation networks overall to distribute the increased biofuel
volumes.
However, ORNL stated that there would be considerable increased traffic along certain
rail corridors due to the shipment of biofuels that would require significant investment to
overcome the resulting congestion. We concluded that these investments could be made to
increase the capacity of the affected rail corridors without undue difficulty, and that therefore the
744 A nationwide average ethanol concentration above 10.00% can only occur insofar as there is consumption of E15
and/or E85.
745 "Analysis of Fuel Ethanol Transportation Activity and Potential Distribution Constraints," ORNL, March 2009.
339
-------
infrastructure system to the blending terminal could accommodate the projected increased
volume of ethanol in a timely fashion.746
To update and expand upon the analysis of distribution infrastructure upstream of retail
that was conducted for the 2010 RFS2 rule, EPA contracted with ICF International Inc. ("ICF")
to conduct a literature review, background research, and stakeholder interviews to characterize
the impacts of distributing ethanol and other biofuels.747 The 2018 ICF report determined that the
conclusions from the 2009 ORNL analysis have largely turned out to be accurate based on an
absence of indicators of distribution constraints up to and including the blending terminal. ICF
noted that there were instances when the ethanol industry went through rapid expansion where
the rail industry was not able to fully accommodate the expansion of inter-regional trade in
ethanol. However, ICF found no evidence to suggest that rail congestion from shipment of
biofuel was a persistent or common problem at the time that the study was completed. Likewise,
ICF found no evidence that marine networks, including those used for import and export, were
experiencing significant issues in accommodating increased volumes of biofuels. Consistent with
the 2010 analysis, ICF stated that the expansion of ethanol and biodiesel volumes could pose
issues for petroleum terminals, but that these issues could be resolved. While ICF indicated that
there likely had been negative impacts on rural and highway transportation networks surrounding
ethanol production facilities, ICF also determined that these impacts could be mitigated with
network infrastructure planning and increased funding for road maintenance. ICF noted these
increased costs would be small in comparison to broader maintenance costs for roads and that the
road network could accommodate substantial growth in the movement of biofuels.
Based on the ICF study and our own assessment of the implementation of the RFS
program, we conclude that the response of the ethanol distribution infrastructure system
upstream of retail has largely unfolded as we projected in the 2010 RFS2 rule. Ethanol imports to
coastal demand centers have helped to satisfy local demand. Ethanol transport over long
distances is primarily being accomplished by unit train and, to a lesser extent, by manifest rail
and barge. Materials compatibility issues continue to prevent ethanol and ethanol blends from
being shipped in petroleum product pipelines. Tank trucks are used to distribute ethanol to
markets close to the ethanol production facility and from rail receipt facilities to more distant
markets. Petroleum terminals have installed the necessary ethanol receipt, storage, and blending
infrastructure. Intermodal facilities, such as those that transfer ethanol directly from rail cars to
tank trucks, are also being used to ease the burden on terminals.
7.4.2 Infrastructure for E85
E85 is permitted to be used only in designated FFVs. As of 2020, there were about 28
million registered light-duty FFVs in the U.S., representing about 10% of all spark-ignition
vehicles.748 However, the number of registered FFVs is expected to decline in the coming years.
For instance, the total number of FFV model offerings has been declining in comparison to its
historical maximum in 2014.
746 See Chapter 1.6 of the RFS2 Regulatory Impact Analysis (EPA-420-R-10-006).
747 Impact of Biofuels on Infrastructure, Report for EPA by ICF International Inc., January 2018.
748 "FFV registrations from AFDC December 2021" and "DOT National Transportation Statistics Table 1-11,"
available in the docket.
340
-------
Figure 7.4.2-1: Light-Duty FFV Model Offerings
Source: Alternative Fuels Data Center. See "Light-Duty AFV HEV and Diesel Model Offerings by Technology-Fuel
March 2023," available in the docket.
The number of registered FFVs in the in-use fleet is changing consistent with the reduced
offerings. While the registered FFV count continued to increase during 2016-2020, the rate of
increase has slowed, as shown in Table 7.4.2-1. If FFV offerings remain at their 2022 levels or
continue to decrease, we would expect the number of FFVs in the in-use fleet to begin
decreasing after 2022.
Table 7.4.2-1: Change in Light-Duty FFV Registration Counts
% change in FFV counts
Year
compared to previous year
2017
+ 10.0%
2018
+6.4%
2019
+4.5%
2020
+1.6%
Source: Alternative Fuels Data Center. See "Change in Light-Duty Vehicle Registration Counts March 2022,"
available in the docket.
E85 is sold at retail stations where the pumps, underground storage tanks, and associated
equipment has been certified to operate safely with the high ethanol concentrations.749 As shown
in Figure 7.4.2-2, stations offering E85 have increased steadily since about 2005. By June 2022,
the total number of stations offering E85 had reached 4,476.
749 "UST System Compatibility with Biofuels," EPA 510-K-20-001, July 2020.
341
-------
Figure 7.4.2-2: Number of Public and Private Retail Service Stations Offering E85a
4500
m 4000
00
g5 3500
£
^ 3000
.1 2500
ro
5 2000
o
£ 1500
Z 1000
500
c
1996
1997
1998
1999
2000
2001
2002
2004
2005
2006
2007
2008
2009
2010
2011
2012
Z.UJ.O
2014
2015
2016
2017
2018
2019
2020
2021
a Data through 2007 is annual, whereas data for 2008 and later is monthly.
Source: Department of Energy's Alternative Fuels Data Center (AFDC). https://afdc.energy.gov/stations/states. See,
e.g., "AFDC - Alternative Fueling Station Counts by State 10-13-22," available in the docket.
Grant programs such as the USDA Biofuels Infrastructure Partnership (BIP) and the
ethanol industry's Prime the Pump program, in addition to individual company efforts, have
helped to fund the expansion of E85 offerings at retail stations. The combined effect of these
efforts ensured ongoing growth in the number of stations offering E85.
Although the total number of retail stations in the U.S. has varied, as shown in Figure
7.4.2-3, the fraction of those stations offering E85 has steadily increased. By the end of 2020, the
fraction of retail stations offering E85 had reached 2.7%, as seen in Figure 7.4.2-4.
Figure 7.4.2-3: Total Number of Retail Stations in the U.S.
155,000
130,000
125,000
OrH(Nm^Lni£r-.ooCT>OrHr\im^iniDr-ooCTio
OOOOOOOOOOrHrHrHrHrHrHrHrHrHrHfM
ooooooooooooooooooooo
(N(Nrsl(N(N(N(N(N(N(N(N(N(N(N(N(N(Nr>J(N(N(N
Source: Table 4.24, Transportation Energy Data Book, Edition 40.
342
-------
Figure 7.4.2-4: Fraction of Retail Stations Offering E85
3.0%
2.5%
2.0%
1.5%
1.0%
0.5%
0.0%
OrHr\im^Ln^r-.ooai©r-i
-------
7.4.3
Infrastructure for El 5
El5 is permitted to be used only in MY2001 and newer light-duty motor vehicles.750 The
infrastructure needed to support the use of El 5 includes blending and storage equipment at
terminals, certified storage and dispensing equipment at retail service stations, and the vehicles
that are permitted to use El5. While the majority of service stations currently offering El5 do so
through blender pumps—which can produce E15 on demand for consumers through the
combination of E10 (or EO) and E85751—the number of terminals offering preblended E15
directly to service stations has been increasing.752 The first terminals started to offer preblended
El 5 in 2016, and as of June 2022 El 5 is offered at 99 terminals, accounting for about 7% for all
U.S. terminals.753 754
As shown in Figure 7.5.3-1, the fraction of the in-use fleet that is MY2001 and newer has
increased steadily since El 5 was approved in 2011, and with it the fraction of all gasoline
consumed by highway vehicles that is consumed by MY2001 and newer vehicles.
Figure 7.4.3-1: Fraction of In-Use Fleet and In-Use Gasoline Consumption for MY2001 and
Newer
100%
95%
90%
85%
80%
75%
70%
65%
60%
55%
50%
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Source: Values calculated using annual retail vehicle sales of cars and trucks (Tables 4.6 and 4.7), survival rates
(Table 3.15), miles per year per vehicle by age (Table 9.11), and fuel economy by model year (Table 4.12) from the
Transportation Energy Data Book, Edition 40, ORNL, February 2022.
Based on the two modes of El 5 production (terminals and blender pumps at retail
stations), and the fact that the majority of in-use vehicles are legally permitted to use El5, it
750 76 FR 4662 (January 26, 2011).
751 According to Prime the Pump, 1,771 out of 2,302 stations offering E15 at the end of 2020 used blender pumps.
752 "Terminal Availability of E15 Grows as EPA Prepares to Remove RVP Barrier," available in the docket.
753 https ://growthenergy.org/resources/retailer-hub. See also "Retailer Hub Growth Energy 92421," available in the
docket.
754 Total number of active fuel terminals was 1,330 as of 3/31/22 per the Internal Revenue Service.
https://www.irs.gov/businesses/small-businesses-self-emploved/tenninal-control-number-tcn-tenninal-locations-
directory. See "Actual Fuel Tenninals as of 3-31-22," available in the docket.
Fuel volume consumed
\/phirlo rni int
344
-------
appears that the primary constraint on the consumption of E15 in the near term is likely the
number of retail stations that offer it. Since El 5 was not approved for use until 2011, there were
no retail stations offering it before 2011. Since the vast majority of the existing retail
infrastructure (including the entire system of tanks, pipes, pumps, dispensers, vent lines, and pipe
dope) is not confirmed to be entirely compatible with El 5, growth in the number of retail
stations offering E15 is dependent on investments in retail outlets to convert them to E15
compatibility or make them compatible when newly constructed. In cases wherein a retail station
already offers E85 through a blender pump, there may be little or no investments needed for new
equipment, and the decision to offer El5 may depend largely on the perceived economic benefit
of doing so. For other station owners, the costs can be substantial. Growth in the number of
stations offering El5 was slow until the BIP and Prime the Pump programs began providing
funding for station conversions in 2016.
Figure 7.4.3-2: Number of Retail Service Stations Offering E15
3000
2500
2000
1500
1000
500
0
Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan-17 Jan-18 Jan-19 Jan-20 Jan-21 Jan-22 Jan-23
Source: "Prime the Pump Infrastructure Update - Jan 2023 available in the docket.
USDA followed up its BIP program with the HBIIP program, which also provides funds
to help retail service station owners to upgrade or replace their equipment to offer biofuels.
HBIIP is composed of HBIIP 1.0 and HBIIP 2.0 which are the original and renewal of the
program. This program effectively began in 2021 and is estimated to take three years to
complete.
There may also be resistance to expanded offerings of El 5 due to concerns about
liability.755 These liability concerns fall into two areas: the use of retail storage and dispensing
equipment that is not compatible and/or not approved for El5, and consumers that refuel
vehicles and engines not designed and/or approved for its use. With regard to equipment
compatibility, even if much of the existing equipment at retail is compatible with El 5 as argued
in studies from the National Renewable Energy Laboratory (NREL)756 and Stillwater
755 See, e.g., "SIGMA NATSO NACS comments on the proposed Set rule 2-10-23," available in the docket.
756 K. Moriarty and J. Yanowitz, "E15 and Infrastructure," National Renewable Energy Laboratory, May
2015. Attachment 3 of comments submitted by the Renewable Fuels Association.
345
-------
Associates,757 compatibility with El5 is not the same as being approved for El5 use. Under EPA
regulations, parties storing ethanol in underground tanks in concentrations greater than 10% are
required to demonstrate compatibility of their tanks with the fuel through one of the following
methods:758
• A certification or listing of underground storage tank system equipment or
components by a nationally recognized, independent testing laboratory such as
Underwriter's Laboratory.
• Written approval by the equipment or component manufacturer.
• Some other method that is determined by the agency implementing the new
requirements to be no less protective of human health and the environment.
The use of any equipment to offer El 5 that does not satisfy these requirements, even if
that equipment is technically compatible with El 5, would pose potential liability for the retailer.
This issue is of particular concern for underground storage tanks and associated hardware, as the
documentation for their design and the types of materials used, and even their installation dates,
is often unavailable. As existing underground storage tank systems reach the end of their
warranties or are otherwise in need of repair or upgrade, there is an opportunity for retail station
owners to install new systems that are compatible with E15. For instance, tanks installed earlier
than 1990 have reached the end of their warranties and should be replaced to safely store fuel.
With regard to retailer concerns about litigation liability for El5 misfueling related to
vehicles not designed and/or approved for use with El5, we note that EPA regulations are
designed to address potential misfueling. These regulations require pump labeling, a misfueling
mitigation plan, surveys, PTDs, and approval of equipment configurations, providing consumers
with the information needed to avoid misfueling.759 In addition, the portion of vehicles not
designed and/or approved for El 5 use continues to decline. MY2000 and earlier light-duty
vehicles represent less than 10% of the in-use fleet, and just slightly over 5% of miles traveled.
Vehicles designed and warranted by manufacturers to be fueled on El 5 are likewise representing
an ever-increasing portion of the in-use fleet.
In sum, the relatively small, albeit growing, number of stations offering El 5 represents a
significant constraint on the expansion of El 5 through 2025. While the applicable standards
under the RFS program could theoretically provide some incentive for retail station owners to
upgrade their equipment to offer El5, there is little direct evidence that the RFS program has
operated in this capacity in the past.
In order to project the number of retail stations that may offer E15 through 2025, we first
separated the effects of USDA's BIP and HBIIP programs from all other efforts, including both
private efforts and those funded by the ethanol industry's Prime the Pump program. The BIP
757 Stillwater Associates, "Infrastructure Changes and Cost to Increase RFS Ethanol Volumes through Increased E15
and E85 Sales in 2016," July 27, 2015. Submitted with comments provided by Growth Energy.
758 "UST System Compatibility with Biofuels," available in the docket.
759 See, e.g., 40 CFR 1090.1420 and 1090.1510.
346
-------
program was responsible for the conversion of 841 retail stations from 2016-2018.760 During this
time, El 5 stations were also increasing as a result of other efforts, bringing the total number of
E15 stations to 1,630, as shown in Figure 7.4.3-2. Of this total, 841 are estimated to have been
the result of the BIP program, while the remaining 789 El 5 stations were the result of other
efforts. The HBIIP program effectively began in 2021.761 Given its similarity to the BIP program
in terms of funding levels and intended outcomes, we have assumed that it would likewise take
three years to complete and would result in 841 new El 5 stations. From these estimated impacts
of the BIP and HBIIP programs, we were able to back-calculate the growth in El 5 stations that
can be attributed to private initiatives, including Prime the Pump.
Table 7.4.3-1: Hisl
orical Breakdown of E15 Stations
Total3
BIPb
HBIIPC
PTP + private
efforts
December 2012
2
0
0
2
December 2013
70
0
0
70
December 2014
105
0
0
105
December 2015
184
0
0
184
December 2016
431
183
0
248
December 2017
1,122
563
0
559
December 2018
1,630
841
0
789
December 2019
2,045
841
0
1,204
July 2020
2,208
841
0
1,367
December 2020
2,302
841
0
1,461
September 2021
2,536
841
210
1,485
April 2022
2,667
841
372
1,454
November 2022
2,923
841
523
1,515
a "Prime the Pump Infrastructure Update - Sept 2021".
b Assumes linear growth from 2016-2018.
0 Assumes that 841 new E15 stations will ultimately be created, with linear growth from 2021-2023.
We observed that the growth due to private efforts appears to be linear after December
2016 and therefore used a least-squares regression to estimate this trend through 2025, as shown
in Figure 7.4.3-3.
760 "Biofuel Infrastructure Partnership - original grants & projections" and "Communication with USD A on the BIP
program 11-15-21," available in the docket.
761 The availability of grants and procedures for applying for them were announced in May 2020. See "USDA
Announces $100 Million for American Biofuels Infrastructure," available in the docket.
347
-------
Figure 7.4.3-3: Growth in E15 Stations Due to Private Initiatives
3,500
Least-squares regression:
Stations = 298.286 x Year-
s
601238
4
s
s
s
>
s
J
*
. . 4
ly ~
Using the available information on the BIP and HBIIP programs and the projection
shown in Figure 7.4.3-3, we were able to estimate the breakdown of E15 stations for 2023-2025,
as shown in Figure 7.4.3-4. The projected total number of E15 stations for 2023-2025 is shown
in Table 7.4.3-2.
Figure 7.4.3-4: Projected Breakdown of E15 Stations
6,000
0
2020 2021 2022 2023 2024 2025
¦ PTP +private ¦ BIP ¦ HBIIP 1.0 ¦ HBIIP 2.0
Table 7.4.3-2: Projected Average Number of Stations Offering E15a
Year
Stations
2023
3,552
2024
4,197
2025
4,713
a Annual average, not year-end.
348
-------
7.5 Deliverability of Materials, Goods, and Products Other Than Renewable
Fuel
The distribution of renewable fuels relies on the same rail, marine, and road infrastructure
networks that are used to deliver materials, goods, and products other than renewable fuels.
Therefore, we evaluated whether the use of renewable fuels would impact the deliverability of
other items that rely on these infrastructure networks.
The 2009 ORNL study of biofuel distribution for the 2010 RFS2 rule concluded that
there would be minimal additional stress on most U.S. transportation networks overall due to
increased biofuel volumes.762 This indicates that the shipment of the statutory biofuel volumes
could be accommodated without impacting the deliverability of other items. However, as
discussed in Chapter 7.5.1, ORNL noted that significant investment would be needed to
overcome congestion on certain rail corridors. The 2018 ICF study of impacts of distributing
ethanol and other biofuels determined that the conclusions from the 2009 ORNL analysis have
largely turned out to be accurate based on an absence of indicators of distribution constraints.763
However, ICF noted that there were instances when the ethanol industry went through rapid
expansion where the rail industry was not able to fully accommodate the expansion in inter-
regional trade in ethanol. During these periods, the volume of ethanol permitted to be shipped
along the sensitive rail corridors was limited to mitigate the congestion. However, ICF found no
evidence to suggest that rail congestion from shipment of biofuel was a persistent or common
problem at the time of the study's completion in 2018.
Likewise, ICF found no evidence that the shipment of biofuels has had a negative impact
on marine networks. While ICF indicated that there likely have been negative impacts on rural
and highway transportation networks surrounding ethanol production facilities, it also
determined that these impacts can be mitigated with network infrastructure planning and
increased funding for road maintenance. ICF noted these increased costs are small in relation to
broader maintenance costs for roads and that the road network can accommodate substantial
growth in the movement of biofuels.
Based on both the ORNL study and the more recent ICF study, there appears to be
minimal overall impact on transportation infrastructure from the distribution of biofuels, and the
system appears to have been able to resolve localized instances of increased stress on the system
in a timely fashion. As a result, we believe that the candidate volumes would not impact the
deliverability of materials and products other than renewable fuel.
As part of considering impacts of biofuels on the deliverability of other items, we also
considered constraints on the deliverability of feedstocks used to produce renewable fuel. We do
not anticipate constraints that would make the candidate volumes difficult to achieve. For
instance, biogas for CNG/LNG vehicles will be delivered through the same pipeline network
used to distribute natural gas.764 Since that biogas will be displacing natural gas used in
762 "Analysis of Fuel Ethanol Transportation Activity and Potential Distribution Constraints," ORNL, March 2009.
763 Impact of Biofuels on Infrastructure, Report for EPA by ICF International Inc., January 2018.
764 See Chapter 7.1.
349
-------
CNG/LNG vehicles, we do not expect a net increase in total volume of biogas + natural gas
delivered.
As shown in Table 3.1-3, there would be an increase in corn ethanol consumption in
2023-2025 in comparison to 2022. However, the projected corn ethanol volumes are about 13.8
- 14.0 billion gallons, which is less than the volume of corn ethanol consumed in 20 1 8.765 Since
the corn collection and distribution network functioned without difficulty in 2018, there is no
reason to believe that it would not function similarly in 2023-2025.
We estimate that the use of FOG for the production of biofuel will remain at
approximately 1.4 billion gallons from 2023 to 2025 (see Table 3.1-3). The projected use of FOG
for biofuel production is consistent with the trend observed from 2018-2022. FOG is collected
and distributed through a diverse network of trucking companies, and this increase would
represent a very small portion of their activities. As a result, we do not anticipate any hindrances
to the deliverability of FOG for the production of renewable diesel in 2023-2025.
Total soybean oil use for the production of BBD is projected to increase from
approximately 1.1 billion gallons in 2022 to approximately 1.8 billion gallons in 2025. This
projected increase is based on the expected expansion of soybean crushing over this time period
in the U.S.
765 "RIN Supply as of 2-17-22," available in the docket.
350
-------
Chapter 8: Other Factors
The CAA directs EPA to consider the impact of the use of renewable fuels on "other
factors" that have a more indirect relationship to volume standards, including job creation, the
price and supply of agricultural commodities, rural economic development, and food prices.766
We focus our analysis on the biofuels that are projected to have the largest changes relative to
the No RFS baseline: corn ethanol, biodiesel and renewable diesel (from soybean oil, FOG, and
corn oil), and biogas.767
8.1 Job Creation
This section provides greater detail on our evaluation of impacts of renewable fuels on
job creation. Attempting to attribute increases or decreases in employment to a single variable
such as domestic renewable fuel use is fraught with complexity. Even considering just the
biofuel production facilities themselves, there are confounding factors that include biofuel
import/export activity, shifts in agricultural commodity prices, and varying demand for
coproducts. Assessing the impacts on indirectly affected industries is even more difficult.
Recognizing this challenge, we chose to focus the analysis on the economic sectors that have the
closest association with biofuel use—biofuel production and agriculture. We acknowledge that
changes in indirect employment (e.g., service sectors, transportation, construction, etc.) can also
be associated with renewable fuel use, but due to the level of effort and uncertainty involved
with indirect effects, they were excluded from the scope of this analysis. We also recognize that
this analysis does not estimate the net employment effects, as increases in employment in some
sectors may be offset by unemployment in other sectors.
8.1.1 Fuel Production
8.1.1.1 Ethanol Production
We projected the impact of the candidate volumes on employment at ethanol production
facilities using an assessment prepared by John Urbanchuk for the Renewable Fuels Association
(RFA).768 Urbanchuk estimates that the total number of direct, full-time-equivalent jobs for
domestic corn ethanol production in 2022 was 11,635 across the 199 plants that RFA found to be
operating that year. The total nameplate capacity of those plants is reported at 17.9 billion
gallons, suggesting an average plant size of 90 million gallons per year and an average employee
concentration of 0.65 jobs per million gallons capacity.
The EIA Annual Fuel Ethanol Production Capacity Report provides plant count and total
nameplate capacity values for historical calendar years. The data currently available show a total
nameplate capacity of 17,380 million gallons of ethanol produced by 192 plants that reported
766 As we explain in Preamble Section II, we also consider several other factors besides those enumerated in the
statute.
767 The impacts evaluated in this chapter are for volume increases for 2023-2025 compared to the No RFS baseline,
as shown in Table 3.2.-2.
768 Urbanchuk, J. ABF Economics. "Contribution of the Ethanol Industry to the Economy of the United States in
2022," February 14, 2023.
351
-------
themselves as operational.769 The average plant size using these figures is 91 million gallons per
year, and using Urbanchuk's total direct employment suggests an average employee
concentration of 0.67 jobs per million gallons capacity.
In 2018, Ethanol Producer Magazine made available data on the capacity and number of
employees at each of 65 corn ethanol facilities.770 These plant capacities generally compare well
with those reported by EIA, deviating by less than 3% when averaged on a state-by-state basis.
For these 65 facilities, we examined employee concentration as a function of production
capacity. The results show a nonlinear decreasing trend in employee concentration with
production capacity, suggesting economies of scale are associated with labor in ethanol plants.
Figure 8.1.1.1-1 shows this data fit with an exponential trendline, and includes the Urbanchuk-
based estimate of employee concentration of 0.65 plotted at the national average facility size of
90 million gallons per year. The Urbanchuk value shows good agreement with the correlation fit
line.
Figure 8.1.1.1-1: Correlation Between Employee Concentration and Facility Size for Corn-
Ethanol Facilities
2.0
Employees per million gal/yr capacity
§ 1-8 = 8.30(million gal/yr capacity)-0-6
il.5 R2 = 0.72
~ 13
— *\ •
- 10 • ,
a> # S
$ 0.8 •. +*¦%
CD
>. , - *
10.5 . • j «. I.|
g Urbanchuk-based estimate of
m 0.3 0.65 plotted here at 90 million
gals/yr average plant capacity
0.0
0 25 50 75 100 125 150
Plant Capacity (Million Gallons per Year)
Since the data underlying this analysis is based on nameplate capacity and employment at
a particular point in time, we were not able to estimate the sensitivity of employment at a
particular facility to changes in production volume at that facility. Regardless, it is unlikely that
variations in production volume at a particular plant over in the short term would affect
employee headcount. Each of the unit operations (e.g., feedstock unloading, mashing,
fermentation, DDGS drying and pelletizing) must remain operational for ethanol production to
continue, and each of these areas requires trained operators. Over the longer term, we might
anticipate changes consistent with Figure 8.1.1.1-1, and demand that stretches plants above their
nameplate capacity for a sustained period could cause construction of new facilities.
769 EIA. U.S. Fuel Ethanol Plant Production Capacity as of January 1, 2022.
https://www.eia.gov/petroleuin/ethanolcapacitv/arcliive/2022/index.php.
7711 Ethanol plant employment data obtained via Ethanol Producer Magazine website in 2018
(http://www.ethanolproducer.com). A table of this information is available in the docket.
352
-------
The increases in ethanol volume evaluated in this rule generally represent increased
consumption of higher-level ethanol blends (e.g., El 5 and E85). The connection between greater
domestic consumption of ethanol and domestic production of ethanol is unclear, as significant
quantities of ethanol have been exported to foreign markets in recent years. The volume of
ethanol that would be consumed in 2023-2025 under the No RFS baseline is significantly less
than the domestic ethanol production capacity, and less than domestic ethanol production in
2022. Thus, it is possible that a decrease in ethanol consumption in the absence of the RFS
program could result in a decrease in domestic ethanol production, or alternatively, could result
in increased ethanol exports.
8.1.1.2 Biodiesel Production
To project the impact of the candidate volumes on employment at biodiesel production
facilities, we primarily relied on information from a 2022 study by LMC International on the
economic impact of the biodiesel and renewable diesel industry prepared for the Clean Fuels
Alliance America.771 The report presents economic and employment impacts of several steps in
the biofuel value chain using a methodology that combines direct, indirect, and induced jobs
based on multipliers taken from the U.S. Department of Commerce Bureau of Economic
Analysis. In Table 6, LMC presents a figure of 17,120 jobs involved in feedstock collection and
biofuel processing for production of 2.5 billion gallons in 2021. This equates to 6.8 jobs per
million gallons, a figure that represents a volume-weighted average of the biodiesel and
renewable diesel in 2021. Using these figures and the volumes in Table 3.2-2, we project an
impact of 6700-7300 long-term jobs in biodiesel production for the years 2023-2025 relative to
the No RFS Baseline. Minimal new construction is expected for biodiesel given the currently-
available capacity and the projection of flat growth relative to 2022.
8.1.1.3 Renewable Diesel Production
As described in Chapters 3 and 6, renewable diesel production has grown significantly in
recent years, and is projected to continue as shown in Table 3.3-1. The additional volume is
expected to come from expansion of existing facilities, construction of new facilities, and
conversion of process trains at petroleum refineries. These differing scenarios will result in a
wide range of employment impacts, from minimal growth for existing facility conversions to
significantly more at new sites. The 2022 LMC analysis referenced above did not break out
impact figures for renewable diesel specifically, and therefore we will use the same estimate of
6.8 operations jobs per million gallons to project an impact of 6000-7600 long-term jobs related
to renewable diesel production over the years 2023-2025 relative to the No RFS Baseline. LMC
also estimated temporary construction jobs related to volume growth. Those figures show 25 jobs
lasting up to two years per million gallons of added capacity. Considering the projected
renewable diesel growth relative to 2022, this would suggest up to 35,000 construction jobs by
2025.
771 LMC International. "Economic Impact of Biodiesel on the United States Economy." November 2022.
353
-------
8.1.1.4 RNG Production
As described in Chapters 3 and 6, we project continued growth of RNG used as
CNG/LNG as a result of the candidate volumes in 2023-2025. Sources of this fuel are expected
to be a mix of landfills and agricultural digesters. While collection of landfill gas has been
required by solid waste regulations for many years, increased credit generation under the RFS
program is expected to cause additional employment related to upgrading and maintenance of
gas cleanup and pipeline interconnect equipment.772 We also project that the construction and
operation of new agricultural digesters and digesters at wastewater treatment facilities would
result in additional employment.
An analysis by the Coalition for Renewable Natural Gas (CRNG) using 2022 data
showed 10,600 jobs across 255 operating facilities with total production of 91 million MMBtu.
773 Converting to ethanol-equivalent gallons (EGEs) gives 9 operations jobs per million EGE,
and approximately 42 workers per facility. Additional data is provided for plants under
construction in 2022, which indicates there are 67 construction jobs per million EGE. These
factors were applied to the projected volume increases of RNG used as CNG/LNG in 2022-
2025, resulting in the employment impacts shown in Table 8.1.1.4-1. These employment
estimates implicitly assume that the average employment at facilities in 2023-2025 occur at the
same ratio to EGEs as in 2022. The construction employment figures also assume that the
construction jobs occur in the year of the volume increase. The actual employment impacts for
2023-2025 may be slightly higher or lower depending on the types of new facilities (e.g.,
landfills, wastewater treatment facilities, or agricultural digesters) and the sizes of these
facilities.
Table 8.1.1.4-1: Change in Employment in RNG Production Relative to the No RFS
Baseline
Year
Construction Jobs
Operations Jobs
Total Jobs
2023
33,165
4,455
37,620
2024
46,096
6,192
52,288
2025
62,444
8,388
70,832
8.1.2 Agricultural Employment
Job creation in the agricultural sector, beyond the fuel production activities discussed
above, is expected primarily in the areas of production and transportation of crops serving as
biofuel feedstocks. Because RNG used as CNG/LNG is produced from waste or byproduct
materials (e.g., separated MSW, wastewater, and agricultural residue), we expect the projected
increases in the production of RNG used as CNG/LNG to have very little impact on employment
related to feedstock production. As noted above, we are projecting higher volumes of ethanol,
772 Jaramillo and Matthews, Environmental Science & Technology 2005 39 (19), 7365-7373.
773 Coalition for Renewable Natural Gas. Economic Analysis of the U.S. Renewable Natural Gas Industry.
December 2022.
354
-------
biodiesel, and renewable diesel production for 2023-2025 relative to the No RFS baseline. The
primary feedstocks projected to be used to produce these fuels are corn and soybean oil.774
Gauging the impact that increased use of renewable fuels has had on employment in the
agricultural sector is challenging for several reasons, including, but not limited to, seasonality,
production of a wide array of products, and the broad nature of employment in the sector, which
stretches from field hands to equipment production. To try to understand this better, we
examined available data on agricultural employment over the past several decades, with no
pretense of ascribing causation for observed trends to particular volumes of renewable fuels.
Some of the most consistently sourced data available on hired farmworkers is made
available by the National Agricultural Statistics Service (NASS).775 We used a combination of
annual and seasonal reports to track the number of harvest season (October) workers hired
directly by farm operators over the past two decades. This data is presented in Figure 8.1.2-1.
Figure 8.1.2-1: Number of Harvest Season Farm Workers
950
900
850
"D
C
| 800
o
750
700
2002
2007
2012
2017
2022
The trend in the data is that direct employment of hired farmworkers by farm operators
has been relatively stable between 750-850 thousand since 2003/4. There is variation year-over-
year, but it is difficult to conclude from this data that there has been any significant increase or
decrease in directly hired by farm labor related to increased production of renewable fuels over
the past two decades. Given the broad scope of this data, it is not possible to discern whether, for
example, an increase of workers harvesting corn in Iowa has been offset by a reduction in
employment of workers harvesting pistachios in California. Were more disaggregated
employment data available, perhaps it would be possible to discern changes in the employment
of farmworkers for the purposes of producing soy and corn.
774 We are also projecting that lesser quantities of FOG and corn oil would be used to produce biodiesel and
renewable diesel in 2023-2025, but since these feedstocks are wastes or co-products of other industries, we do not
expect their increased use to impact agricultural employment.
775 USDA NASS Agricultural Statistics and Farm Labor data,
https://usda.librarv.corneil.edu/concern/pnblications/x920fw89s and
https://usda.librarv.cornell.edu/concern/publications/i3860694x?locale=en.
355
-------
An additional set of data on agricultural employment is collected by USDA. The methods
for categorizing types of employment are slightly different than those found in the NASS data,
but the greater breadth of jobs captured by the employment in agriculture and support industries
data set provides additional insight. Figure 8.1.2-2 presents the data on employment in
agriculture and support activities for 2001-2021.776
Figure 8.1.2-2: Employment in Agriculture and Support Activities
Thousand jobs
1,400-
1,200-
Livestock support
1,000-
Livestock _
000-
Crop support
600-
400-
Crops
200-
0
i i i r i i ! i i i i r r i i i i i i
2001 2003 2005 2007 2009 2011 2013 2015 2017 2010 2021
Note: Employment is measured as the annual average number of full- and part-time jobs.
Data do rot cover smaller farm employers in those States that exempt them from
participation in (he unemployment insurance system.
Source: USDAh Economic Research Service u$ing data from U.S. Department of Labor,
Quarterly Census of Employment and Wages, June 8, 2022 retease.
This data from USDA shows that employment in crop production and crop support
activities have increased by about 3% and 20%, respectively, over the past decade. As with the
NASS data in Figure 8.1.2-1, the lack of crop-specific data makes drawing associations with
biofuel production very difficult. We observe that the lowest employment levels reported in the
USDA data for crop production workers coincide with the 2008-2010 recession and that it was
not until 2015 that the number of such jobs returned to the pre-recession levels. Looking at this
data set, it is difficult to see any clear impact of increased renewable fuel production among
broader economy-wide factors.
8.2 Rural Economic Development
Changes in biofuel production can have economic development impacts on rural
communities and financial impacts on farmers. We are projecting significantly higher
consumption of ethanol, biodiesel, renewable diesel, and RNG used as CNG/LNG in 2023-2025
relative to the No RFS baseline. As discussed in Chapter 8.1.1.1, the impact of the RFS volumes
for 2023-2025 on domestic ethanol production are uncertain. In their absence, domestic ethanol
production could continue at a level at or near current production volumes with increasing
776 USDA ERS Farm Labor, March 2023. https://www.ers.usda.gov/topics/farm-economv/farm-labor.
356
-------
ethanol exports, or alternatively, domestic ethanol production could decrease. In light of these
uncertainties, we are not projecting any significant changes in rural economic development
related to the ethanol volumes we are projecting in 2023-2025.
For biodiesel and renewable diesel, we expect that much of the rural economic impacts in
2023-2025 will be related to the production of feedstocks for these fuels. We project that almost
all of the increase in biodiesel and renewable diesel production in 2023-2025 will be produced
from soybean oil. Some of this soybean oil is expected to come from additional soybean
production and crushing, which may bring some revenue increases to rural communities.
The increased production of RNG used as CNG/LNG is expected to result in additional
rural economic activity. Using factors derived from a 2022 analysis by CRNG, we estimated that
each additional million RINs of CNG/LNG from agricultural digesters is associated with $0.76
million in economic activity related to gas capture, upgrading, and facility administration
activities.777 This figure suggests that the candidate volumes would result in $375 million, $522
million, and $707 million in economic activity in 2023, 2024, and 2025, respectively, relative to
the No RFS baseline. That analysis also indicated that 68% of facilities under construction in
2022 were agricultural waste digesters, which were likely to be located in rural areas, and that
total capital expenditure on these facilities was $1.1 billion. While the overall share of this
economic impact occurring in rural areas is unknown, the fact that the majority of CNG/LNG
facilities under construction in 2022 were agricultural waste digesters suggests that much of this
economic activity is occuring in rural areas.
8.3 Supply of Agricultural Commodities
Changes in biofuel production can have an impact on the supply of agricultural
commodities. As discussed above, we project higher volumes of ethanol, biodiesel, renewable
diesel, and RNG used as CNG/LNG in 2023-2025 relative to the No RFS baseline. These
volume increases suggest the potential for associated increases in underlying crop production;
however, the magnitude of the potential impact cannot be estimated with any certainty. Biogas is
not produced from agricultural commodities and therefore is not expected to affect their supply
or price.
For historical context, Figure 8.3-1 shows trends in corn production and uses from 1995—
2022.778 This data suggests domestic corn production has grown steadily at a 25-year average
rate of around 2%, or 250 million bushels per year, with no apparent correlation to ethanol
production volumes.
777 Coalition for Renewable Natural Gas. Economic Analysis of the U.S. Renewable Natural Gas Industry,
December 2022. See slides 35 (MMBtu/yr for agricultural digesters) and 37 (dollars per year for capture and
upgrade).
778 USDA ERS Bioenergy Statistics, April 2023 (Table 5). 2022 values are estimates.
https://www.ers.nsda.gov/data-prodncts/n-s-bioenergv-statistics.
357
-------
Figure 8.3-1: Corn Production and Usage
Between 2005-2010, additional corn required to satisfy increasing ethanol production
was sourced primarily by diversion from animal feed until overall production caught up. Supply
of corn to food uses showed modest but consistent growth at historical rates during this period,
despite increased consumption as ethanol feedstock. Exports also remained relatively steady,
except for a drop corresponding to weather-related supply disruptions and elevated prices in
2011-2012. Animal feed use began to rebound after 2014 when growth in ethanol use slowed
and prices stabilized. Another factor contributing to the longer-term shift of animal feed away
from whole corn was the increasing substitution with DDGS, a byproduct of ethanol production.
Considering historical trends over the past two decades indicating the ability of production to rise
to meet demand, the relatively modest changes in ethanol volumes associated with this rule are
likely to have minimal impact on the supply of corn to food, exports, or other uses.
Figure 8.3-2 shows that soybean production has risen steadily over time, similar to the
trend for corn production.779 Roughly 80% of this growth since 2005 has been associated with
rising exports of soybeans, which have nearly doubled over that period. Domestic crushing of
beans has grown by about 25% since 2005, which is mirrored in growth of the crush products,
soy meal and oil. This data also shows that exports of soy meal nearly doubled during this time,
which together with the growth in whole bean exports, presents a picture consistent with
expansion of meat production internationally. (Worldwide, over 95% of beans are eventually
crushed for meal and oil.)
779 USDA ERS Oilcrops Data Yearbook, Soy Tables, March 2023.
358
-------
Figure 8.3-2: Soybean Production and Usage
Figure 8.3-3: Soy Meal Production and Usage
Growth in soybean oil production and its uses are shown in Figure 8.4-2.779 The
steepened upward production trend over the past decade has been enabled by both increasing
crush capacity and increasing yields of oil per bushel of soybean input. The use of soybean oil
for biofuel production has also increased steeply since about 2014, with a further uptick since
2020. This continued expansion of biofuel demand has begun to shift the relative value
relationship between the oil and meal crush products, as discussed further in 8.4 below.
359
-------
8.4 Price of Agricultural Commodities
Agricultural commodities are bought and sold on an international market, where prices
are determined by trends and upsets in worldwide production and consumption. Renewable fuels
are only one factor among many (e.g., droughts and storm damage) in determining commodity
prices. Thus, models that attempt to project prices at specific times in the future, or in reaction to
specific demand perturbations, necessarily contain high levels of uncertainty. This section
reviews historical trends and presents key observations from the literature.
In the U.S., corn and soybeans are generally only harvested once per year, and therefore
storage is a critical factor in the supply chain. After harvest, grain stores are replenished and then
drawn down throughout the year. In recent years, 10-15% of the previous year's overall corn
production is typically still in storage at the time of the new harvest.780 If demand rises after
harvest, stocks may be drawn down faster than expected. Conversely, if demand decreases,
stocks accumulate into the next season.
Storage also has the effect of dampening price shocks in years when harvests are smaller
than expected. In 2012, a drought year, corn stocks fell to the lowest levels since 2000, putting
upward pressure on futures prices, which in turn served as a market signal to induce more corn
planting in the upcoming season. Work done by Informa Economics for RFA in 2016 examined
the historical relationship between corn usage, stocks, and futures prices.781 Figure 8.4-1 shows
the strong correlation between futures prices and the stock-to-usage ratio, illustrating that the
latter is a key driver of market signals. More generally, crop prices are influenced by an array of
factors from worldwide weather patterns to biofuel policies to international tariffs and trade
wars.
780 USDA ERS Feed Grains Data Yearbook, April 2023 (Table 4).
781 Informa Economics IEG. "The Impact of Ethanol Industry Expansion on Food Prices: A Retrospective Analysis."
November 2016. https://d35tlsvewk4d42.cloiMlfront.net/file/975/Retrospective-of-Impact-of-Ethanol-on-Food-
Prices-2016.pdf.
360
-------
Figure 8.4-1: Corn Ending Stocks / Use Ratio Versus Futures Price
$8.00
1 $7.00
Q.
W
£ $6.00
D
£ $5.00
2012
• .2011
• 2010
FF = 0.8255 \
• 2007
? $4.00
&
CO
1 $3.00
<
g $2.00
>=
2 $1.00
a
• 2013
2008
. !%V009
2006 201
""-2005
2004
$-
3% 8% 13% 18%
USCorn Ending Stocksto Use Ratio
23%
Source: USDA, Informa Economics IEG
To make more specific quantitative estimates of the impact of increased biofuel
production on corn prices, we considered two meta-studies. Condon, etal, reviewed 29
published papers in 2015 and found a central estimate of 3-5% increase in corn prices per billion
gallons of ethanol.782 Focusing only on scenarios where a supply response is included gives a
result of 3%. A supply response refers to scenarios where farmers can respond to price signals in
subsequent year(s) and plant additional crops to meet a larger demand. This is appropriate, as the
scope of the analysis is biofuel policy (rather than something unforeseen like weather shocks). A
similar meta-analysis was done in 2016 by FAPRI-Missouri that considered several newer
studies.783 This paper found an increase of $0.19 per bushel per billion gallons, or $0.15 if a
supply response is included, a figure that is generally consistent with the 3% impact above if
applied to the corn price in 2016.
We are projecting higher corn ethanol consumption in 2023-2025 (an additional 660-787
million gallons per year) than would occur under the No RFS baseline. We note, however, that in
recent years domestic ethanol production has exceeded consumption, with significant volumes
being exported. This trend appears very likely to continue during 2023-2025, as our projected
consumption volumes remain below USDA's projected production for these years.784 This
782 Condon, Nicole, Klemick, Heather and Wolverton, Ann, (2013), Impacts of Ethanol Policy on Corn Prices: A
Review and Meta-Analysis of Recent Evidence, No. 201305, NCEE Working Paper Series, National Center for
Environmental Economics, EPA, https://EconPapers.repec.org/RePEc :nev: wpaper:wp201305.
783 Food and Agricultural Policy Research Institute. Literature Review of Estimated Market Effects of U.S. Corn
Starch Ethanol, 2016. FAPRI-MU Report #01-16.
784 USDA Agricultural Projections to 2032. February 2023.
361
-------
history of significant export volumes makes it difficult to assess the impact of the projected
volumes.
It is possible that a decrease in domestic corn ethanol consumption would result in an
increase in exports and minimal change in domestic production volumes. Were this to occur, we
would expect little to no net change in domestic corn demand, and thus corn prices.
Alternatively, it is possible that a decrease in consumption would result in a decrease in domestic
corn ethanol production. In this case we would expect a decrease in corn demand and corn
prices. To illustrate the potential impact of the candidate volumes on corn prices, we have
calculated the projected impact in 2023-2025 assuming that these volumes result in an increase
in domestic corn ethanol production relative to the No RFS baseline. The projected price impacts
are calculated using a value from the literature of 3% increase per billion gallons of corn ethanol
produced, as described above. Because the USD A Agricultural Projections show corn use for
ethanol production at quantities that appear similar to the candidate volumes for 2023-2025, we
have projected lower corn prices for the No RFS baseline, rather than assuming the corn prices in
these projections represent a No RFS case and projecting higher prices for the candidate
volumes. The projected impact of the candidate volumes on corn prices relative to the No RFS
baseline are shown in Table 8.4-1.
Table 8.4-1: Projected Impact on Corn Prices Relative to the No RFS Baseline
2023
2024
2025
Corn Price Percent Increase per Billion Gallons of Ethanol
3%
3%
3%
Corn Price (Candidate Volumes); $/bushela
$6.80
$5.70
$4.90
Corn Price Increase per Billion Gallons of Ethanol; $/bushel
$0.20
$0.17
$0.15
Corn Ethanol Increase; billion gallons
0.706
0.776
0.840
Corn Price Increase; $/bushel
$0.14
$0.13
$0.12
Corn Price (No RFS Baseline); $/bushel
$6.66
$5.57
$4.78
a Corn prices are from the USDA Agricultural Projections to 2032 report. Prices represent the average price for a
calendar year. For corn, the price is calculated using 1/3 of the price for the first agricultural marketing year (e.g.,
2022/2023 for 2023) and 2/3 of the price for the second agricultural marketing year (e.g., 2023/2024 for 2023).
With biodiesel and renewable diesel production, the commodity input of interest is
soybean oil, which has an indirect link to soybean production. Oil is produced by crushing,
which also creates soy meal, and the supply and prices of these move independently from each
other. The crush quantities vary from year to year, depending on the crush margin, which is
defined as the sum of oil and meal price minus the bean price. Oversupplying either oil or meal
markets can cause prices to fall, decreasing the crush margin. Thus, the degree of passthrough of
oil price increases to bean prices, which may then influence acres planted, is not straightforward.
Figure 8.4-2 shows historical trends in soybean oil prices alongside allocation to biofuel and
other uses, based on data taken from the USDA Oilcrops Yearbook.785 Use in domestic biofuel
rose from 0.8 million tons in 2005 to 5.8 million tons in 2022. Other domestic uses also
increased steadily through 2005, decreased slightly from 2005-2010, and have remained
relatively consistent since 2010. Exports of soybean oil are a relatively minor outlet and had
remained fairly consistent for many years until dropping toward zero following the steep price
increase since 2020. Noting the lack of correlation between soybean oil price and its use in
785 USDA ERS Oilcrops Data Yearbook, Soy Tables, March 2023.
362
-------
biofuel production historically, we conclude that the price of soybean oil is influenced by a
number of factors occurring in the broader economy, including rising petroleum prices, supply
chain disruptions on a range of inputs (e.g., fertilizer), weather-related shortages of vegetable oils
internationally, as well as general price inflation. In particular, while increased soybean oil
demand for biofuel production was likely a contributing factor to the sharp price increase in
soybean oil prices in 2020 and 2021, poor weather conditions in South America and Malaysia
were also a significant factor.786
Figure 8.4-2: Soybean Oil Price and Allocation to Biofuel and Exports
60
50
40
30
20
10
0
0
1980
1990
2000
2010
2020
There are relatively few quantitative studies on the impacts of BBD production on
soybean oil and bean prices, and they show a wide range of results. This is in part because these
studies have included a variety of different policy combinations, none of which separated out just
the impact of the RFS program on BBD demand. Ethanol demand could impact the soybean
markets even in the absence of increased demand for BBD from the RFS program due to
increased competition for cropland and other inputs. The largest impacts are estimated when the
BBD obligations are modeled jointly with the conventional and cellulosic ethanol obligations.
Given that actual cellulosic ethanol volumes have been far below those modeled, we focus on the
studies that included only a conventional ethanol obligation. The range of soybean price impacts
indicated by these studies is 1.8-6.5% per billion gallons of BBD, from which we take a central
value of 4 2%.787-788-789
In our proposal, to project the impact on crude soybean oil prices, we used a value of 160
per pound of oil per billion gallons of BBD produced from soybean oil. This figure was derived
786 Wilson, Nick. "Oil Prices Surge - Vegetable Oil That Is." Marketplace.org. February 17, 2022.
787 Babcock, B. A. 2012. The impact of US biofuel policies on agricultural price levels and volatility. China
Agricultural Economic Review 4:407-426.
788 J. Huang, J. Yang, S. Msangi, S. Rozelle, and A. Weersink. 2012. Biofuels and the poor: Global impact pathways
of biofuels on agricultural markets. Food Policy 37:439-451.
789 Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis. EPA-420-R-10-006. February 2010.
363
-------
from modeling work published by Babcock, et al., and is the same figure used for other cost
estimates in this rule.790 As with corn ethanol, we have assumed that the soybean oil prices in the
USDA Agricultural Projections to 2031 represent projected prices of the candidate volumes since
they project soybean oil used for biofuel production at quantities that appear similar to the
candidate volumes for 2023-2025. We projected lower soybean oil prices for the No RFS
baseline, rather than assuming the soybean oil prices in these projections represent the No RFS
baseline and projecting higher prices for the candidate volumes. The projected impacts of the
candidate volumes on soybean oil prices at the time of proposal are shown in Table 8.4-2.1.
Table 8.4-2.1: Projected Impact on Soybean Oil Prices Relative to the No RFS Baseline
Babcock Methodology
2023
2024
2025
Soybean Oil Price (Candidate Volumes); $/pounda
$0.69
$0.57
$0,505
Soybean Oil Price Increase per Billion Gallons of Biofuel;
$/pound
$0.16
$0.16
$0.16
Soybean Oil Biofuel Increase; billion gallons
2,017
1,983
1,955
Soybean Oil Price Increase; $/pound
$0.32
$0.32
$0.31
Soybean Oil Price (No RFS Baseline); $/pound
$0.37
$0.25
$0.20
a Soybean oil prices are from the USDA Agricultural Projections to 2032 report. Prices represent the average price
for a calendar year. For soybean oil, the price is calculated using 1/4 of the price for the first agricultural marketing
year (e.g., 2022/2023 for 2023) and 3/4 of the price for the second agricultural marketing year (e.g., 2023/2024 for
2023).
However, in the time between proposal and final, research on this topic has evolved and
we have adapted our analysis to reflect this. We largely base this on Lusk, et al., which uses a
shock of 20% of current biofuel volumes (equivalent to approximately 243 million gallons of
soy-derived biodiesel) to project a price impact on soybean oil prices and other commodities.791
790 Babcock BA, Moreira M, Peng Y, 2013. Biofuel Taxes, Subsidies, and Mandates: Impacts on US and Brazilian
Markets. Staff Report 13-SR 108. Center for Agricultural and Rural Development, Iowa State University.
791 Lusk, Jayson L. (2022) Prepared for United Soybean Board, Food and Fuel: Modeling Food System Wide
Impacts of Increase in Demand for Soybean Oil.
364
-------
Table 8.4-2.2: Projected Impact on Soybean Oil Prices Relative to the No RFS Baseline
Lusk Methodology
2023
2024
2025
Soybean Oil Price (Candidate Volumes); $/pounda
$0.69
$0.57
$0,505
Soybean Oil Price Response, per Billion Gallons of Biofuelb
35.7%
35.7%
35.7%
Soybean Oil Price Increase per Billion Gallons of Biofuel;
$/pound
$0.25
$0.20
$0.18
Soybean Oil Biofuel Increase; billion gallons
2,017
1,983
1,955
Soybean Oil Price Increase; $/pound
$0.50
$0.40
$0.35
Soybean Oil Price (No RFS Baseline); $/pound
$0.19
$0.17
$0.15
Soybean Meal Price, $/ton
$390
$380
$352
Soybean Meal Price Response, %b
-7.94%
-7.94%
-7.94%
Soybean Meal Price (No RFS Baseline); $/ton
$421
$410
$380
a Soybean oil prices are from the USDA Agricultural Projections to 2032 report. Prices represent the average price
for a calendar year. For soybean oil, the price is calculated using 1/4 of the price for the first agricultural marketing
year (e.g., 2022/2023 for 2023) and 3/4 of the price for the second agricultural marketing year (e.g., 2023/2024 for
2023).
b This number is based on a modified shock from Lusk equivalent to 1 billion gallons (as opposed to approximately
240 million gallons in the Lusk paper)
The data in Table 8.4-2.2 reflects an updated methodology to soybean oil and meal price
responses based on a more recent study than Babcock, et.al. While EPA is confident of this
methodology moving forward, the price impact projections from Lusk may not be representative
of the true nature of biofuels' future effect on soybean oil, particularly for 2023-24. Vegetable oil
prices have been well above historical norms for the past couple years due to a variety of supply
and demand impacts around the world. The price impact may be more in line with numbers
shown for 2025 in Table 8.4-2.2 assuming the market begins to renormalize.
Analysis published by Irwin at the University of Illinois indicates that soybean oil prices
often move separately from meal and bean prices, and that the latter two are closely correlated.792
In recent years soybean oil prices appear to have increased significantly relative to soybean meal
prices, as shown in Figure 8.4-3.
792 Irwin, S. "The Value of Soybean Oil in the Soybean Crush: Further Evidence on the Impact of the U.S. Biodiesel
Boom." Farmdoc Daily (7): 169, Department of Agricultural and Consumer Economics, University of Illinois at
Urbana-Champaign, September 14, 2017.
365
-------
Figure 8.4-3: Relative Values of Soybean Oil and Soybean Meal
Hfli
0% L 0
2016 2017 2018 2019 2020 2021 2022
Oil share Meal share Oil value Meal value
From 2016 through the end of 2020, the value of soybean oil relative to soybean meal
was relatively stable, with soybean oil representing around 33% of the value of a soybean on
average.793 Starting in 2021, the relative value of the soybean oil has increased significantly,
averaging 47% in the 21/22 crop year. Examining the $/bushel value contribution of the
components, we see that the oil value has more than doubled while the meal has increased by
around 30%. Thus the value of the soybean is shifting more toward the oil than the meal in
recent years. This suggests that the supply of soybean oil may tightening relative to soybean
meal, with rising soybean oil prices exerting some downward pressure on soybean meal prices.
In addition to the price impacts on corn, soybean oil, and soybean meal, we also
estimated price changes for other feed grains (grain sorghum, barley, and oats) and distillers
grains. We adjusted the prices of these commodities, as they historically compete with corn in
the feed market, and to a lesser extent for acreage. The price adjustments for grain sorghum,
barley, oats, and distillers grains are based on historical price relationships of these commodities
with corn. As with corn and soybean oil, we assumed that the prices in the USD A Agricultural
Projections to 2031 represent projected prices of the candidate volumes and adjusted the
projected prices for these commodities lower in our price projections for the No RFS baseline.
The projected impact of the candidate volumes on sorghum, barley, oat, and distillers grain
prices are shown in Table 8.4-3.
793 USDA ERS Oilcrops Data Yearbook, Soy Tables, March 2023.
366
-------
Table 8.4-3: Projected Impact on Prices of Other Commodities Relative to the No RFS
Baseline
2023
2024
2025
Price Change Factor Relative to Corn Price Change3
Sorghum; $/bushel
0.93
0.93
0.93
Barley; $/bushel
0.88
0.88
0.88
Oats; $/bushel
0.72
0.72
0.72
Distillers Grains; $/ton
0.018
0.018
0.018
Projected Price Impact
Sorghum; $/bushel
$0.13
$0.12
$0.11
Barley; $/bushel
$0.12
$0.11
$0.11
Oats; $/bushel
$0.10
$0.09
$0.06
Distillers Grains; $/ton
$4.26
$4.64
$4.29
a These factors were developed in conjunction with USD A in the 2012 evaluation of the use of the general waiver
authority. See "Methodology for Estimating Impacts on Food Expenditures, CPI for Food and CPI for All Items,"
available in the docket.
8.5 Food Prices
The above impact on commodity prices may in turn have a ripple impact on food prices
and the many other products produced from these commodities. Since the candidate volumes are
projected to have a relatively small impact on the overall world commodity markets, and since
the cost of these commodities tends to be a relatively small component in the cost of food, the
projected impact of this rule on food prices is relatively modest. Further, we note that the
projected impact of the candidate volumes on food prices does not represent a cost, but rather a
transfer, since higher food prices that result from higher commodity prices represent increased
income for feedstock producers (e.g., corn and soybean farmers).794
To project the impact of the candidate volumes on food prices, we used a methodology
developed in conjunction with USDA in assessing requests from the governors of several states
to reduce the 2012 RFS volumes using the general waiver authority.795 This methodology
generally uses estimates of the impact of biofuel volumes on commodity prices (e.g., corn,
soybean oil, etc.) to calculate the estimated impacts on total food expenditures. For context, this
estimated change in food expenditures is then compared to total food expenditures. Finally, the
ratio of the estimated change in food expenditures to the total food expenditures is used to
estimate the change in food expenditures for the average consumer unit and the consumer units
in the lowest income quintile.
In Chapter 8.4, we presented estimates of the impact of the candidate volumes on
commodity prices relative to the No RFS baseline. These estimates are the starting point for our
794 In other words, food price impacts represent the movement of money within society (from consumers of foods to
the producers of foods) as opposed to additional costs that society as a whole incurs. We note that while the CAA
specifically directs EPA to calculate the impacts on "food prices," as opposed to calculating the impact on the cost
to consumers of food. We acknowledge that these market interactions are affected by deadweight losses, but we
have not estimated the proportion of deadweight losses to transfers in this rule.
795 77 FR 70752 (November 27, 2012).
367
-------
estimate of the impact of the RFS volumes on food prices. From those, we projected the impact
of commodity prices on total food expenditures, which are shown in Table 8.5-1. We assumed
that changes in commodity prices are fully passed on to consumers at the retail level, and
therefore we can project changes in total food expenditures by multiplying the quantity of these
commodities used for food and feed. Feed use is included to capture the effects of the change in
the price of the commodity on livestock producers' production costs, and ultimately the effects
on retail livestock prices.796
We recognize that projecting that the price of distillers grains increases proportionally to
the price of corn may over-state the impact of this rule on these commodities and ultimately on
food prices. It is possible increasing demand for biofuels may result in an over-supply of
distillers grains, as it is a co-product of biofuel production. Thus, while biofuel production may
increase the prices of corn and food produced from corn, it may not increase the price of
distillers grains. This could mitigate the overall impact of this rule on food prices. At this time,
we do not have sufficient data to project how increasing demand for corn for biofuel production
would impact the price of distillers grains. If the price for distillers grains increases less than the
price of corn (or if it decreases) in response to increased demand for biofuels, we would expect a
smaller impact on food prices than what we have estimated for the candidate volumes.
This methodology assumes no response by producers or consumers to changes in
commodity prices and therefore may overstate the change in food expenditures. However,
previous research suggests that demand for food is very inelastic and therefore this methodology
should provide a close approximation of the change in food expenditures.797 Our estimates of the
increase of food expenditures only reflect expenditures in the U.S. Because of the integrated
nature of agricultural commodity markets, the projected increases in agricultural commodity
prices may also impact food prices and expenditures globally. We have not attempted to quantify
these global impacts.
796 This methodology includes the expected price impact on all crops used as animal feed and does not account for
the livestock produced for the export market or imported meat or animal products.
797 Okrent, Abigail M., and Julian M. Alston. The Demand for Disaggregated Food-Away-From-Home and Food-at-
Home Products in the United States, ERR-139, USD A, Economic Research Service, August 2012.
368
-------
Table 8.5-1: Changes in Food Expenditures Relative to the No RFS I
taseline
Commodity Price
Change
Quantity Used for
Food and Feed3
Change in
Expenditures
Changes in Foot
Expenditures in 2023
Corn
$0.14 per bushel
6,725 million bushels
$942 million
Grain Sorghum
$0.13 per bushel
90 million bushels
$12 million
Barley
$0.12 per bushel
160 million bushels
$19 million
Oats
$0.10 per bushel
141 million bushels
$14 million
Soybean Oil
$0.50 per pound
13.7 billion pounds
$6,850 million
Distillers Grains
$4.26 per short ton
41.5 million short tons
$177 million
Soybean Meal
-$31 per ton
39,450 thousand short
tons
-$1223 million
Total
N/A
N/A
$6,790 million
Changes in Foot
Expenditures in 2024
Corn
$0.13 per bushel
7,150 million bushels
$930 million
Grain Sorghum
$0.12 per bushel
120 million bushels
$14 million
Barley
$0.11 per bushel
160 million bushels
$18 million
Oats
$0.09 per bushel
141 million bushels
$13 million
Soybean Oil
$0.40 per pound
13.9 billion pounds
$5,560 million
Distillers Grains
$4.64 per short ton
43.5 million short tons
$202 million
Soybean Meal
-$30 per ton
40,225 thousand short
tons
-$1,207 million
Total
N/A
N/A
$5,530 million
Changes in Foot
Expenditures in 2025
Corn
$0.12 per bushel
7,425 million bushels
$869 million
Grain Sorghum
$0.11 per bushel
120 million bushels
$13 million
Barley
$0.11 per bushel
165 million bushels
$18 million
Oats
$0.06 per bushel
142 million bushels
$9 million
Soybean Oil
$0.35 per pound
14.1 billion pounds
$4,935 million
Distillers Grains
$4.29 per short ton
44.5 million short tons
$191 million
Soybean Meal
-$28 per ton
41,025 thousand short
tons
-$ 1149 million
Total
N/A
N/A
$4,886 million
a Quantity used for food and feed calculated based on the USDA Agricultural Projections to 2032 (February 2023).
Prices represent the average price for a calendar year. Calendar year prices are calculated using a ratio based on the
number of months in the calendar year in each agricultural marketing year. In general, the quantity use for food and
feed is the sum of the quantities projected for Feed and Residual and Food, Seed & Industrial. For corn, we
subtracted the quantity used for Ethanol & by-products from this total. The quantity of distillers grains was
calculated based on the production of 17 pounds of distillers grains for every bushel of corn used to produce ethanol.
Finally, the quantity of soybean oil is equal to the amount listed for food, feed & other industrial and the quantity of
soybean meal is the total quantity of domestic disappearance.
Finally, we compared the estimated change in food expenditures to total food
expenditures as reported by the Bureau of Labor and Statistics in their 2021 survey.798 We used
798 Bureau of Labor and Statistics - Consumer Expenditures in 2021: Table 1, Quintiles of income before taxes:
Annual expenditure means, shares, standard errors, and coefficients of variation. 2021.
369
-------
the ratio of the estimated change in food expenditures to the total food expenditures to estimate
the change in food expenditures for the average consumer unit (household) and the consumer
units in the lowest and second-lowest income quintiles, as shown in Tables 8.5-2 and 3. In this
analysis we have assumed the same price effects on all foods when in fact the price impacts on
foods consumed by low and high income groups may be affected differently. Additionally, lower
price elasticities for lower-income consumers mean that the welfare effects of these changes
could be aggravated for lower-income groups.
Table 8.5-2: Percent Change in Food Expenditures Relative to the No RFS Baseline
2023 Estimate
2024 Estimate
2025 Estimate
Number of Consumer Units (thousands)
133,595
133,595
133,595
Food Expenditures per Consumer Unit
$8,289
$8,289
$8,289
Total Food Expenditures
$1,107 billion
$1,107billion
$1,107 billion
Change in Food Expenditures
$6,790 million
$5,530 million
$4,886 million
Percent Change in Food Expenditures
0.61%
0.50%
0.44%
Table 8.5-3: Change in Food Expenditures per Consumer Unit Relative to the No RFS
Baseline
2023
2024
2025
All Consumer Units
Food Expenditures
$8,289
$8,289
$8,289
Percent Impact on Food Expenditures
0.61%
0.50%
0.44%
Projected Food Expenditure Increase
$50.56
$41.45
$36.59
Lowest Quintile Income Consumer Units
Food Expenditures
$4,875
$4,875
$4,875
Percent Impact on Food Expenditures
0.61%
0.50%
0.44%
Projected Food Expenditure Increase
$29.74
$24.38
$21.52
Second-Lowest Quintile Income Consumer Units
Food Expenditures
$5,808
$5,808
$5,808
Percent Impact on Food Expenditures
0.61%
0.50%
0.44%
Projected Food Expenditure Increase
$35.43
$29.04
$25.63
370
-------
Chapter 9: Environmental Justice
Executive Orders 12898 (Federal Actions to Address Environmental Justice in Minority
Populations, and Low-Income Populations) and 14096 (Revitalizing Our Nation's Commitment
to Environmental Justice for All) establishes federal executive policy on environmental justice
(EJ). Its main provision directs federal agencies, to the greatest extent practicable and permitted
by law, to make EJ part of their mission by identifying and addressing, as appropriate,
disproportionately high and adverse human health or environmental effects of their programs,
policies, and activities on communities with EJ concerns in the U.S. EPA defines EJ as the fair
treatment and meaningful involvement of all people regardless of race, color, national origin, or
income with respect to the development, implementation, and enforcement of environmental
laws, regulations, and policies. Executive Order 14008 (86 FR 7619; February 1, 2021) also calls
on federal agencies to make achieving EJ part of their missions "by developing programs,
policies, and activities to address the disproportionately high and adverse human health,
environmental, climate-related and other cumulative impacts on disadvantaged communities, as
well as the accompanying economic challenges of such impacts." It also declares a policy "to
secure environmental justice and spur economic opportunity for disadvantaged communities that
have been historically marginalized and overburdened by pollution and under-investment in
housing, transportation, water and wastewater infrastructure and health care." EPA also released
technical guidance799 (hereinafter "EPA's Technical Guidance") to provide recommendations
that encourage analysts to conduct the highest quality analysis feasible, recognizing that data
limitations, time and resource constraints, and analytic challenges will vary by media and
circumstance.
When assessing the potential for disproportionately high and adverse health or
environmental impacts of regulatory actions on communities with EJ concerns, EPA strives to
answer three broad questions:
1. Is there evidence of potential EJ concerns in the baseline (the state of the world absent the
regulatory action)? Assessing the baseline will allow EPA to determine whether pre-
existing disparities are associated with the pollutant(s) under consideration (e.g., if the
effects of the pollutant(s) are more concentrated in some population groups).
2. Is there evidence of potential EJ concerns for the regulatory option(s) under
consideration? Specifically, how are the pollutant(s) and its effects distributed for the
regulatory options under consideration?
3. Do the regulatory option(s) under consideration exacerbate or mitigate EJ concerns
relative to the baseline? It is not always possible to quantitatively assess these questions,
though it may still be possible to describe then qualitatively.
EPA's Technical Guidance does not prescribe or recommend a specific approach or
methodology for conducting an EJ analysis, though a key consideration is consistency with the
assumptions underlying other parts of the regulatory analysis when evaluating the baseline and
regulatory options. Where applicable and practicable, EPA endeavors to conduct such an
analysis. Going forward, EPA is committed to conducting an EJ analysis for rulemakings based
799 https://www.epa.gov/sites/default/files/2016-06/documents/eitg tf.
371
-------
on a framework similar to what is outlined in EPA's Technical Guidance, in addition to
investigating ways to further weave EJ into the fabric of the rulemaking process.
9.1 Proximity Analysis of Facilities Participating in the RFS Program
As of October 2022, there were 342 registered RIN-generating facilities in the U.S. There
were also 146 petroleum refineries producing transportation fuel. These facilities are spread out
across the U.S., with the addition of 3 petroleum facilities in Hawaii and 5 petroleum facilities in
Alaska. Our analysis looks at the demographic composition of communities near these facilities
nationally, for the subset of facilities located in rural areas, and by EPA Region (Figure 9.1-1)
and major fuel type—in this case, petroleum, renewable diesel, biodiesel, ethanol, and RNG.
For data on demographic characteristics near each facility, we use block group level data
from the 2016-2020 American Community Survey. Areal apportionment is used to attribute
these data to uniform buffers of l-,3-, and 5-mile distances around each RIN-generating facility.
Because the demographic composition of urban areas dominates the national average, we also
examine facilities located in rural areas separately. We define a rural block group as one whose
centroid does not intersect with Census polygons of urban areas/clusters. We then characterize a
facility as being located in a rural area if 50% or more of the population within 3 miles live in
rural block groups.
Figure 9.1-1: Map of EPA Regions
OR
{!¦)
MT
ID
MV I 1
M UT
©
A2
MM
Guam
Trust Territories
American Samoa
Northern Mariana
l&lsnds
NO
SD
NE
MN
Ml
IA
0
KS MO S r
©
Oh
— , D£
wv VA MD
TX
,
(E) *
^ AK 1
AR
CK
J
\ MS AL
\ ^
• ft
LA
© 7-
rwi
PR 1
NC
T 'Ec]
.so
.fl
Headqiiarters
As the RFS is a national program, it is difficult to track facility-by-facility responses to
the candidate volumes, so this demographic analysis focuses on baseline characteristics of
372
-------
communities near RIN-generating production facilities. We examined near-facility demographics
in order to bring a quantitative lens to our qualitative observations. As this rule would displace
petroleum fuels—primarily gasoline and diesel—with biofuels, it is expected that communities
near facilities that produce biofuels may experience an overall increase in criteria pollutant
exposure, while those near petroleum refineries could see the opposite if refineries react to the
candidate volumes by decreasing production.3
Table 9.1-1 shows the demographic composition of communities within 1, 3 and 5 miles
of these facilities compared to the national average. As seen below, at the 5 mile buffer radius,
approximately 10 percent of the U.S. population lives near one or more RIN-generating facility.
These near-facility communities can generally be characterized as having a greater than average
percent non-white population, regardless of the distance buffer utilized. The Hispanic population
living near these facilities is nearly double the national average. The percent Black population is
1.25 times the national average. In addition, these communities tend to have a higher than
average unemployment rate, a lower median income, a higher percent with less than a high
school education, and a higher percent living lx and 2x below the federal poverty line compared
to the national average.
Table 9.1-1: Demographics Near RIN-generating Facilities Compared to National Average
Demographic
1 mi
3 mi
5 mi
Nationwide
Total Population (millions)
1.0
12.2
32.8
326.6
% Rural Population
11.0
8.1
7.8
26.6
% White
63.2
60.3
60.0
70.4
% Black
16.1
16.0
16.3
12.6
% American Indian and Alaska Native
0.8
0.7
0.7
0.8
% Asian
4.3
6.1
6.5
5.6
% Native Hawaiian and Other Pacific Islander
0.4
0.4
0.3
0.2
% Other (Including Two or More)
15.2
16.5
16.1
10.3
% Hispanic
31.2
34.7
34.1
18.2
Median Income ($2020)
$58,411
$63,945
$66,529
$73,181
% lx Poverty Line
18.5
16.7
15.9
12.5
% 2x Poverty Line
40.3
37.3
35.8
29.1
Unemployment Rate
7.0
6.5
6.3
5.4
% Less than High School Education
12.1
12.0
11.4
7.8
Table 9.1-2 presents the demographic characteristics of communities near 236 RIN-
generating facilities located in rural areas (or about 48 percent of all RIN-generating facilities).
Many biofuel facilities are located in rural areas in order to be close to feedstock crops. They
play a role in rural job creation as further discussed in Chapters 8.1 and 8.3. In general, the
demographic composition of rural communities that host RIN-generating facilities is similar to
the rural national average, with the exception of a substantially higher than average percent
Hispanic populations. People of two or more races and those living at or beneath lx and 2x the
federal poverty line are slightly higher than nationwide rural average, and the median income is
slightly lower.
373
-------
Table 9.1-2: Demographics Near Rural RIN-generating Facilities
Demographic
1 mi
3 mi
5 mi
Nationwide
Total Population (millions)
0.1
0.5
1.7
86.9
% Rural Population
85.3
82.3
61.9
100
% White
86.4
84.2
80.9
84.2
% Black
5.8
6.7
6.9
7.0
% American Indian and Alaska Native
0.5
0.6
0.7
1.5
% Asian
1.1
1.5
2.6
1.6
% Native Hawaiian and Other Pacific Islander
0.1
0.2
0.3
0.1
% Other (Including Two or More)
6.2
6.8
8.6
5.6
% Hispanic
12.4
13.9
18.5
9.0
Median Income ($2020)
$63,259
$65,346
$66,886
$68,372
% lx Poverty Line
11.3
12.0
12.6
11.3
% 2x Poverty Line
30.2
29.7
31.0
27.9
Unemployment Rate
4.6
4.4
4.6
4.8
% Less than High School Education
9.0
8.6
9.0
7.7
Tables 9.1-3.1 through 3.10 show the demographic composition of communities near
these biofuel and petroleum facilities by EPA region as shown in Figure 9.1-1. These community
demographics are compared to regional averages. We present this information using a 3 mile
distance buffer, though trends are similar at the 1 and 5 mile distances.
Table 9.1-3.1: Region 1 Near RIN-generating Facility Demographics Compared to Regional
Average
Within 3 miles
Region
Number of Facilities
7
Total Population [millions]
0.3
14.8
% Rural Population
7.5
25.0
% White
68.5
79.8
% Black or African American
14.9
6.8
% American Indian and Alaska Native
0.4
0.3
% Asian
4.0
4.9
% Native Hawaiian and Other Pacific Islander
0.1
0.0
% Other (Including Two or More)
12.1
8.1
% Hispanic
19.3
11.3
Median Income [2020$]
$63,395
$85,923
% Low Income (Below lx Poverty Line)
15.6
9.6
% Low Income (Below 2x Poverty Line)
33.6
21.8
Unemployment Rate
6.5
5.2
% Less than High School Education
7.1
6.0
374
-------
Table 9.1-3.2: Region 2 Near RIN-generating Facility Demographics Compared to Regional
Within 3 miles
Region
Number of Facilities
16
Total Population [millions]
0.6
28.4
% Rural Population
4.8
13.2
% White Alone
52.5
63.3
% Black or African American
19.2
14.8
% American Indian and Alaska Native
0.4
0.3
% Asian
3.4
8.9
% Native Hawaiian and Other Pacific Islander
0.0
0.0
% Other (Including Two or More)
24.5
12.6
% Hispanic
43.3
19.5
Median Income [2020$]
$69,340
$83,720
% Low Income (Below lx Poverty Line)
12.7
12.1
% Low Income (Below 2x Poverty Line)
33.2
26.1
Unemployment Rate
5.9
5.9
% Less than High School Education
11.8
8.3
Table 9.1-3.3: Region 3 Near RIN-generating Facility Demographics Compared to Regional
Within 3 miles
Region
Number of Facilities
21
Total Population [millions]
0.8
30.8
% Rural Population
7.4
27.7
% White
57.3
70.4
% Black or African American
33.1
17.6
% American Indian and Alaska Native
0.2
0.2
% Asian
3.8
4.8
% Native Hawaiian and Other Pacific Islander
0.0
0.0
% Other (Including Two or More)
5.5
6.9
% Hispanic
5.0
8.4
Median Income [2020$]
$53,629
$80,003
% Low Income (Below lx Poverty Line)
19.7
10.9
% Low Income (Below 2x Poverty Line)
39.0
25.0
Unemployment Rate
7.8
5.4
% Less than High School Education
7.7
6.6
375
-------
Table 9.1-3.4: Region 4 Near Facility Demographics Compared to Regional Average
Within 3 miles
Region
Number of Facilities
42
Total Population [millions]
1.0
66.3
% Rural Population
8.4
34.9
% White
47.1
68.9
% Black or African American
43.2
21.4
% American Indian and Alaska Native
0.3
0.4
% Asian
1.1
2.6
% Native Hawaiian and Other Pacific Islander
0.0
0.1
% Other (Including Two or More)
8.2
6.6
% Hispanic
26.7
13.0
Median Income [2020$]
$44,532
$61,927
% Low Income (Below lx Poverty Line)
24.1
14.1
% Low Income (Below 2x Poverty Line)
49.3
33.0
Unemployment Rate
8.5
5.7
% Less than High School Education
13.1
8.3
Table 9.1-3.5: Region 5 Near RIN-generating Facility Demographics Compared to Regional
Within 3 miles
Region
Number of Facilities
99
Total Population [millions]
1.5
52.5
% Rural Population
14.8
30.1
% White
73.4
78.1
% Black or African American
14.3
11.4
% American Indian and Alaska Native
0.4
0.4
% Asian
2.6
3.6
% Native Hawaiian and Other Pacific Islander
0.1
0.0
% Other (Including Two or More)
9.2
6.5
% Hispanic
11.3
8.3
Median Income [2020$]
$58,621
$69,979
% Low Income (Below lx Poverty Line)
17.4
12.1
% Low Income (Below 2x Poverty Line)
36.5
27.8
Unemployment Rate
5.9
5.4
% Less than High School Education
8.4
6.2
376
-------
Table 9.1-3.6: Region 6 Near RIN-generating Facility Demographics Compared to Regional
Within 3 miles
Region
Number of Facilities
105
Total Population [millions]
2.5
42.4
% Rural Population
10.3
30.4
% White
60.6
69.0
% Black or African American
22.1
13.6
% American Indian and Alaska Native
0.9
1.6
% Asian
2.7
3.9
% Native Hawaiian and Other Pacific Islander
0.1
0.1
% Other (Including Two or More)
13.5
11.8
% Hispanic
44.1
31.2
Median Income [2020$]
$57,184
$65,936
% Low Income (Below lx Poverty Line)
19.6
14.8
% Low Income (Below 2x Poverty Line)
43.1
33.9
Unemployment Rate
7.0
5.6
% Less than High School Education
14.5
9.6
Table 9.1-3.7: Region 7 Near RIN-generating Facility Demographics Compared to Regional
Within 3 miles
Region
Number of Facilities
77
Total Population [millions]
0.6
14.1
% Rural Population
18.8
40.6
% White
81.4
83.9
% Black or African American
7.7
7.6
% American Indian and Alaska Native
0.6
0.5
% Asian
2.6
2.4
% Native Hawaiian and Other Pacific Islander
0.1
0.1
% Other (Including Two or More)
7.5
5.4
% Hispanic
10.3
7.3
Median Income [2020$]
$57,183
$66,202
% Low Income (Below lx Poverty Line)
15.9
11.5
% Low Income (Below 2x Poverty Line)
35.6
28.5
Unemployment Rate
4.8
4.2
% Less than High School Education
7.2
5.8
377
-------
Table 9.1-3.8: Region 8 Near RIN-generating Facility Demographics Compared to Regional
Within 3 miles
Region
Number of Facilities
39
Total Population [millions]
1.0
12.1
% Rural Population
8.6
30.7
% White
81.5
83.8
% Black or African American
2.1
2.7
% American Indian and Alaska Native
1.7
2.3
% Asian
2.5
2.4
% Native Hawaiian and Other Pacific Islander
0.7
0.3
% Other (Including Two or More)
11.5
8.4
% Hispanic
21.5
15.2
Median Income [2020$]
$72,331
$77,609
% Low Income (Below lx Poverty Line)
11.3
10.0
% Low Income (Below 2x Poverty Line)
28.6
25.3
Unemployment Rate
4.2
4.2
% Less than High School Education
6.5
4.9
Table 9.1-3.9: Region 9 Near RIN-generating Facility Demographics Compared to Regional
Within 3 miles
Region
Number of Facilities
55
Total Population [millions]
3.6
51.0
% Rural Population
2.2
11.3
% White
50.4
58.0
% Black or African American
6.9
5.7
% American Indian and Alaska Native
0.7
1.3
% Asian
13.1
13.5
% Native Hawaiian and Other Pacific Islander
0.6
0.7
% Other (Including Two or More)
28.3
20.9
% Hispanic
56.3
36.6
Median Income [2020$]
$75,139
$82,933
% Low Income (Below lx Poverty Line)
14.4
12.5
% Low Income (Below 2x Poverty Line)
34.7
29.2
Unemployment Rate
7.0
6.3
% Less than High School Education
15.6
10.2
378
-------
Table 9.1-3.10: Region 10 Near RIN-generating Facility Demographics Compared to
Regional Average
Within 3 miles
Region
Number of Facilities
27
Total Population [millions]
0.4
14.2
% Rural Population
9.5
27.7
% White
62.1
77.5
% Black or African American
7.0
2.9
% American Indian and Alaska Native
1.6
1.9
% Asian
12.2
6.5
% Native Hawaiian and Other Pacific Islander
2.4
0.6
% Other (Including Two or More)
14.8
10.7
% Hispanic
13.5
12.7
Median Income [2020$]
$75,937
$78,170
% Low Income (Below lx Poverty Line)
12.2
10.8
% Low Income (Below 2x Poverty Line)
27.6
26.2
Unemployment Rate
5.3
5.2
% Less than High School Education
6.6
5.8
Overall, we see similar trends at the regional level as compared to the overall national
picture. In some regions, there appear to be less stark demographic disparities compared to the
regional average, while in other cases, more so. Since biofuel and petroleum facilities are
particularly concentrated in Regions 5, 6, and 7 (281 facilities) we use them to illustrate these
differences. Regions 5 and 7 have slightly elevated percent Hispanic and Black populations near
the biofuel facility compared to their regional averages, while percent Hispanic and Black
populations are 1.4 and 1.7 times the regional average in Region 6, respectively. Populations
living near these facilities also tend to have lower median incomes, a greater percent living in
poverty or with less than a high school education.
The analysis above does not differentiate by type of facility. As stated above, the effects
of this rule will not be felt evenly by different demographic groups, but the greatest contributing
factor to what communities may experience is what type of facility they are near. While the EPA
is unable to ascertain how facilities may respond to changes in required volumes of different RIN
categories, the 2023-2025 volumes are greater than those in 2020-2022. Increases in required
biofuel volumes will mean, generally, an increase in biofuel production at biofuel facilities and a
decrease in petroleum production at refineries that make gasoline or diesel, all else equal. Biofuel
directly displaces conventional transportation fuel. Communities near ethanol facilities, biodiesel
and renewable diesel facilities, and RNG facilities may see increases in criteria pollutants.
Conversely, communities near petroleum refineries may see reductions in air emissions as
producers respond to increasing RFS volumes. It is not practicable to assess what facilities may
or may not specifically experience any changes directly attributable to the RFS. In spite of these
limitations, we examine the demographic composition of communities that may be affected by
fuel type in Table 9.1-4. Results are shown for the 3 mile distance buffer.
379
-------
Regardless of facility type, nearby communities have higher percent Black population
than the national average, particularly near biodiesel and petroleum facilities; percent Black
populations are 1.7 and 2.3 times the national average, respectively. Percent Hispanic
populations near RNG, biodiesel, and petroleum facilities are also almost or more than double
than the national average. The median incomes of communities near biodiesel, ethanol, and
renewable diesel facilities are nearly $20,000 or more lower than the national median income,
while communities near RNG and petroleum facilities have a median income that are lower than
the national median by almost $10,000 and $1,000, respectively. Most of the communities near
different facility types, other than those producing RNG, have a higher unemployment rate than
the national average. All experience higher rates of poverty than the national average. A higher
proportion of these populations compared to the national average also do not have at least a high
school education.
Table 9.1-4: Facility I
Sinographies Within 2
» Miles By Fuel Production Type
Biodiesel
Facilities
Ethanol
Facilities
Petroleum
Facilities
Renewable
Diesel
Facilities
RNG
Facilities
National
Average
Number of Facilities
72
85
146
9
176
Total Population
[millionsl
1.9
0.7
6.0
0.2
3.4
326.6
% Rural Population
8.3
20.7
4.3
11.0
11.7
26.6
% White
60.8
68.9
57.8
63.7
62.5
70.4
% Black or African
American
21.2
17.9
13.9
28.3
15.9
12.6
% American Indian
and Alaska Native
0.5
0.5
0.9
0.6
0.7
0.8
% Asian
2.8
3.3
6.9
1.4
7.3
5.6
% Native Hawaiian
and Other Pacific
Islander
0.2
0.2
0.5
0.1
0.3
0.2
% Other (Including
Two or More)
14.6
9.3
20.1
5.8
13.4
10.3
% Hispanic
34.6
15.0
43.3
17.6
24.8
18.2
Median Income
T2020$l
$54,428
$55,725
$63,842
$51,606
$71,935
$73,181
% Low Income
(Below lx Poverty
Line)
19.2
17.7
17.3
19.6
13.8
12.5
% Low Income
(Below 2x Poverty
Line)
42.0
38.2
38.6
42.5
32.0
29.1
Unemployment Rate
6.9
6.7
7.0
8.9
5.4
5.6
% Less than High
School Education
12.3
8.4
13.9
10.4
9.2
7.8
380
-------
9.2 Non-GHG Air Quality Impacts
There is evidence that communities with EJ concerns are impacted by non-GHG
emissions. Numerous studies have found that environmental hazards such as air pollution are
more prevalent in areas where racial/ethnic minorities and people with low socioeconomic status
(SES) represent a higher fraction of the population compared with the general
population.800'801'802'803 Consistent with this evidence, a recent study found that most
anthropogenic sources of PM2.5, including industrial sources, and light- and heavy-duty vehicle
sources, disproportionately affect people of color.804
Emissions of non-GHG pollutants such as PM, NOx, CO, SO2, and air toxics occur
during the production, storage, transport, distribution, and combustion of petroleum-based fuels
and biofuels. Some communities with EJ concerns are located near petroleum refineries,
biorefineries, and on-road sources of pollution. For example, analyses of communities in close
proximity to petroleum refineries have found that vulnerable populations near refineries may
experience potential disparities in pollution-related health risk from that source.805 There is also
substantial evidence that people who live or attend school near major roadways are more likely
to be of a minority race, Hispanic ethnicity, and/or low SES. 806>807>808 For this rule, EPA has not
quantitatively assessed the cumulative risks to certain demographics near biorefineries, but is
evaluating the extent to which this type of analysis could be done for future rulemakings.
Although proximity to an emissions source is a useful indicator of potential exposure, it
is important to note that the impacts of emissions from both upstream and tailpipe sources are not
limited to communities in close proximity to them. As a result of regional transport and
secondary formation of pollutants in the air, the effects of both potential increases and decreases
in emissions from the sources affected by this rule might also be felt many miles away, including
in communities with EJ concerns downwind of sources. The spatial extent of these impacts from
upstream and tailpipe sources depends on a range of interacting and complex factors, including
the amount of pollutant emitted, atmospheric chemistry and meteorology.
800 Mohai, P.; Pellow, D.; Roberts Timmons, J. (2009) Environmental justice. Annual Reviews 34: 405-430.
https://doi.org/10.1146/annurev-environ-082508-094348.
801 Rowangould, G.M. (2013) A census of the near-roadway population: public health and environmental justice
considerations. Trans Res D 25: 59-67. http://dx.doi.Org/10.1016/i.trd.2013.08.003.
802 Marshall, J.D., Swor, K.R.; Nguyen, N.P (2014) Prioritizing environmental justice and equality: diesel emissions
in Southern California Environ Sci Techno1 48: 4063-4068. htlps://doi.org/.1.0..1.02l/es405.1.67f.
803 Marshall, J.D. (2000) Environmental inequality: air pollution exposures in California's South Coast Air Basin.
Atmos Environ21: 5499-5503. https://doi.Org/10.10.l.6/i.atmosenv.2008.02.005.
804 C. W. Tessum, D. A. Paolella, S. E. Chambliss, J. S. Apte, J. D. Hill, J. D. Marshall (2021). PM2.5 polluters
disproportionately and systemically affect people of color in the United States. Sci. Adv. 7, eabf4491.
805 U.S. EPA (2014). Risk and Technology Review - Analysis of Socio-Economic Factors for Populations Living
Near Petroleum Refineries. Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
January.
806 Rowangould, G.M. (2013) A census of the U.S. near-roadway population: public health and environmental
justice considerations. Transportation Research Part D; 59-67.
807 Tian, N.; Xue, J.; Barzyk. T.M. (2013) Evaluating socioeconomic and racial differences in traffic-related metrics
in the United States using a GIS approach. J Exposure Sci Environ Epidemiol 23: 215-222.
808 Boehmer, T.K.; Foster, S.L.; Henry, J.R.; Woghiren-Akinnifesi, E.L.; Yip, F.Y. (2013) Residential proximity to
major highways - United States, 2010. Morbidity and Mortality Weekly Report 62(3): 46-50.
381
-------
The manner in which biofuel producers and markets respond to the candidate volumes in
this rule could have non-GHG exposure impacts for communities living near facilities that
produce biofuels. Chapter 4.1 summarizes what is known about potential air quality impacts of
the candidate volumes assessed for this rule. We expect that small increases in non-GHG
emissions from biofuel production and small reductions in petroleum-sector emissions would
lead to small changes in exposure to these non-GHG pollutants for people living in the
communities near these facilities. This is of some concern, as we noted in Chapter 4.1 that
communities living within 10 km of biorefineries were shown to be at a higher risk of adverse
respiratory outcomes. However, we do not have the information needed to understand the
magnitude and location of facility-specific responses to the candidate volumes, and therefore we
are unable to evaluate impacts on air quality in EJ communities near these facilities. We
therefore recommend caution when interpreting these broad, qualitative observations.
9.3 Water & Soil Quality Impacts
We conducted an analysis to estimate the impacts associated with the candidate volumes
on water and soil quality in Chapter 4.4. Though soil quality is not among the statutory factors
required to be analyzed under the set authority in the CAA,809 it is discussed in conjunction with
water quality because it can have direct impacts on water quality. EPA defines water quality as
the condition of water to serve human or ecological needs, while USD A defines soil quality as
the ability of soil to function, including its capacity to support plant life. The ways in which this
rule could potentially impact water and soil is by creating an incentive for land use and
management changes, primarily through the encouragement of biofuels produced from corn and
soybeans. An increase in demand for corn and soybeans for biofuel production has historically
caused the conversion of natural grasslands to cropland.810 This land use change has negative
consequences for soil quality in that it can increase soil erosion, depletion of SOM (soil organic
matter), and loss of soil carbon. These negative impacts on soil quality then translate into
negative impacts on water quality like increased soil erosion, which causes sedimentation and
murky water conditions. Nutrient leaching can result in excessive algae growth and hypoxia (low
oxygen levels in the water), which then has negative consequences on aquatic organisms as
described in Chapter 4.4.2.3.
As discussed in Chapter 4.4, the candidate volumes have the potential to incentivize
increases in crop production, and by extension adverse impacts on soil and water quality. This
does not apply to biogas used to produce RNG, as they are making use of waste streams of
processes driven by other phenomena.811'812'813'814 97% of all RINs generated via biogas-related
pathways came from wastewater treatment plants, agricultural digesters, or landfill methane
capture. The RFS program does not affect human, animal, or solid waste production, and in fact
809 CAA section 21 l(o)(2)(B)(ii).
810 See Chapter 4.3.2.
811 Melvin, A.M.; Sarofim, M.C.; Crimmins, A.R., "Climate benefits of U.S. EPA programs and policies that
reduced methane emissions 1993- 2013," Environmental Science & Technology, 2016, in press.
http://pubs.acs.org/doi/pdf/10.1021/acs.est.6b00367. DOI 10.1021/acs.est.6b00367.
812 81 FR 59332 (August 29, 2016).
813
814 https://www.resonrcerecoveredata.org/Potential Power of Renewable Energy Generation From
Wastewater and Biosotids Fact Sheet.pdf.
382
-------
incentivizes the collection of these products, improving local soil and water quality. However,
the magnitude of both this impact and that of other biofuels is difficult to estimate as it would
require more information on the correlation between RFS-driven changes in biofuel volumes and
feedstock usage and where any increases in those feedstocks occur (e.g., domestically vs
internationally, and on what acres), the cultivation practices applied to those acres (e.g., fertilizer
and pesticide use, use of cover crops in the non-growing season, crop rotations, etc.), as well as
modeling to evaluate the magnitude of any runoff occurring from those acres. Additionally, we
would need additional information on the impacted populations in order to evaluate the EJ
concerns: where are the populations that are already being impacted most, who resides in those
areas, how are they using the water, and how are the changes in water quality and availability
impacting those uses and, thereby, those populations. For these reasons, we are unable to assess
the degree of impact the candidate volumes may have on communities with environmental
justice concerns. However, going forward, we would like to better understand the relationship
between the RFS volume standards and land use/land management decisions that impact those
concerns.
Any negative impacts on aquatic life have the potential to also negatively impact
populations that rely on fish or other aquatic life, like shrimp or crawfish, for sustenance or
income. According to a study by Beveridge, et al., fish is a very nutritious food for humans with
high quality animal protein, essential fatty acids, and micronutrients.815 Many American Indian
tribes, minority populations, and some low-income populations rely on local food sources—
including fish and other aquatic life—to supplement their diets. To better understand these high-
risk populations, we conducted a literature review to identify population groups most likely to
fall under the high-risk category for mercury exposure based on higher-than-average fish
consumption as part of the RIA for the Mercury and Air Toxics Standards rule.816 These
population groups are the same ones that would be most affected by any adverse impact the RFS
program has on fisheries and aquatic life due to their heavy reliance on fishing for sustenance.
This review included six high-risk population groups, including African-American and white
low-income recreational and subsistence fishers in the Southeast, female low-income
recreational and subsistence fishers, Hispanic and Laotian subsistence fishers, and
Chippewa/Ojibwe Tribe members in the Great Lakes area.817 American Indian tribes also rely on
recreational fisheries for income, as explained by the U.S. Department of the Interior.818 The fish
populations depend on healthy water systems to thrive. If these aquatic ecosystems are
negatively impacted by agricultural runoff and nutrient leaching, they could suffer from algae
blooms or become hypoxic, making it impossible for fish to survive and endangering the human
815 Beveridge, M. C., Thilsted, S. H., Phillips, M. J., Metian, M., Troell, M., & Hall, S. J. (2013). Meeting the food
and nutrition needs of the poor: the role of fish and the opportunities and challenges emerging from the rise of
aquaculture. Journal of fish biology, 83(4), 1067-1084. https://doi.org/ tfb.12.1.87.
816 Regulatory Impact Analysis for the Final Mercury and Air Toxics Standards. EPA-452/R-11-011. December
2011.
817 Id.
818 U.S. Department of the Interior, Working with Native American Tribes
(https://web.archive.Org/web/20210719071659/https://www.fws.gov/southeast/our-services/native-american-tribes).
See also, U.S. Department of the Interior, Native American Trust Responsibilities
(htlps://web.archive.org/web/20220/ )5/fattps://www.:fws.gov/soiitfawest/fisfaeries/native ameriean tmst.fatmt
), and U.S. Department of the Interior, Indian Affairs, Branch of Fish, Wildlife, and Recreation
(https://www.bia.gOv/bia/ots/division~natural~resources/branch~:fish~wildlife~recreation).
383
-------
populations that rely on them. Additionally, any increased use of nitrogen rich fertilizers, as are
applied to approximately 98% of corn acres (see Table 4.4.2.1-1), could result in nitrates
leaching into groundwater that may be used for human consumption, particularly in areas with
loamy and sandy soil conditions. Nitrate filtration is an expensive process that low-income
communities may not have access to. Additionally, where groundwater wells are employed in
rural areas, the concern of disproportionate impact on vulnerable populations may increase. In
this way, if and to the extent the candidate volumes adversely affect water quality, they could
potentially have disproportionately severe negative impacts on EJ communities within American
Indian tribes and other low income populations that rely on local fisheries as a source of food or
income or that may not be able to afford costly water filtration systems to address nitrate
contamination in their drinking water.
9.4 Impacts on Fuel and Food Prices
Costs are also relevant to an EJ analysis when communities are expected to face
economic challenges due to impacts of a regulation (E.O. 14008). For instance, if prices for basic
commodities such as food and fuel increase as a result of a rulemaking, lower-income
households may be differentially affected since these goods and services may make up a
relatively larger share of their income, and they are less able to adapt or substitute away from
them.
As part of the analyses conducted for this rule, we estimated the impact on food prices.
These impacts are attributed to increases in corn and soy prices associated with the candidate
volumes. Both the literature819'820 and our analysis in Chapter 10 indicate corn and soy are a
relatively small proportion of most foods purchased and consumed in the U.S., and the overall
food price impacts are relatively small as a percentage of total food expenditures. We estimate
that the candidate volumes would affect gasoline prices by 2.40/gal in 2023, 3.20/gal in 2024,
and 4.30/gal in 2025. Diesel prices would rise by 10.10/gal in 2023 and 2024, and 11.10/gal in
2025. Food prices would rise from these volumes by 0.61% in 2023, 0.55% in 2024, and 0.44%
in 2025, relative to the No RFS baseline. These impacts are discussed in greater detail in
Chapters 8.4 (price of agricultural commodities), 8.5 (food price impacts), and 10 (fuel price
impacts).
The projections of the impact associated with the candidate volumes on food and fuel
prices are ultimately derived from projections of the impact on widely traded commodities such
as corn, soybeans, gasoline, and diesel. We therefore do not expect that the impact on food and
fuel prices would vary for different parts of the country. However, changes in food and fuel
prices could have a disproportionate impact on populations that spend a larger share of their
income on food and fuel. According to data collected via the Consumer Expenditure Survey
from the Bureau of Labor and Statistics, consumer units with income in the lowest 20% spend a
819 Hayes, D.J., B.A. Babcock, J.F. Fabiosa, S. Tokgoz, A. Elobeid, T.H. Yu, F. Dong, C.E. Hart, E. Chavez, S. Pan,
M. Carriquiry, and J. Dumortier. 2009. "Biofuels: Potential Production Capacity, Effects on Grain and Livestock
Sectors, and Implications for Food Prices and Consumers." Working paper 09-WP 487. Center for Agricultural and
Rural Development, Iowa State University.
820 Taheripour, Farzad, et al. "Economic Impacts of the U.S. Renewable Fuel Standard: An Ex-Post Evaluation."
Frontiers in Energy Research, vol. 10, 2022, https://doi.org/10.3389/fenrg.2Q22.749738.
384
-------
greater portion of their total expenditures on food and fuel (see Table 9.4-1). Thus, even though
we expect that the effects on the prices for food and fuel to increase proportionally for all
consumers, we also expect that these price impacts, though small, would have a larger impact on
lower-income communities where food and fuel expenditures are a greater portion of total
expenditures.
Table 9.4-1: Proportion of Total Expenditures on Food and Fuel821
All
Lowest 20%
Second-lowest
Consumer
Consumer
20% Consumer
Units822
Unit Income
Unit Income
Total expenditures
$66,928
$30,869
$43,918
Food expenditures
$8,289
$4,875
$5,808
% of total expenditures on food
12.4%
15.8%
13.2%
Fuel expenditures
$2,148
$1,111
$1,702
% of total expenditures on fuel
3.2%
3.6%
3.9%
% Women
53%
62%
56%
% Black
13%
19%
15%
% With a High School Degree or Less
28%
45%
37%
Assuming no changes in income available to spend on goods, nor changes to the bundles
of goods consumed, the RFS program would cause the lowest quintile of consumer units to
spend $4,950.64, or 16%% of their income on food (versus 15.8% currently) by 2025, while the
second lowest quintile of consumer units would spend $5,898.10, or 13.4% of their income on
food (versus 13.2% currently), by 2025. This is shown in year by year increments below in Table
9.4-2. These consumer units would also see increases in their fuel expenditures. The lowest and
second-lowest quintile income consumer units would see cumulative increases to their fuel
expenditures of $41.66 and $63.83, respectively, by 2025. These increases would result in these
groups spending 3.7% and 4% of their total expenditures on fuel, compared to 3.3% for all
consumer units. This can be seen in Table 9.4-3
821 2021 Consumer Expenditure Survey, Bureau of Labor Statistics, published September 2022.
822 Consumer units consist of families, single persons living alone or sharing a household with others but who are
financially independent, or two or more persons living together who share major expenses. This represents an
average value.
385
-------
Table 9.4-2: Year By Year Change in Food Expenditures per Consumer Unit Relative to
the No RFS Baseline
2023
2024
2025
All Consumer Units
Food Expenditures
$8,289
$8,289
$8,289
Percent Impact on Food Expenditures
0.61%
0.50%
0.44%
Projected Food Expenditure Increase
$50.56
$41.45
$36.59
Lowest Quintile Income Consumer Units
Food Expenditures
$4,875
$4,875
$4,875
Percent Impact on Food Expenditures
0.61%
0.50%
0.44%
Projected Food Expenditure Increase
$29.74
$24.38
$21.52
Second-Lowest Quintile Income Consumer Units
Food Expenditures
$5,808
$5,808
$5,808
Percent Impact on Food Expenditures
0.61%
0.50%
0.44%
Projected Food Expenditure Increase
$35.43
$29.04
$25.63
Table 9.4-3: Year By Year Change in Fuel Expenditures per Consumer Unit Relative to the
No RFS Baseline
2023
2024
2025
All Consumer Units
Fuel Expenditures
$2,148
$2,148
$2,148
Percent Impact on Fuel Expenditures
0.79%
1.23%
1.73%
Projected Fuel Expenditure Increase
$16.97
$26.42
$37.24
Lowest Quintile Income Consumer Units
Fuel Expenditures
$1,111
$1,111
$1,111
Percent Impact on Fuel Expenditures
0.79%
1.23%
1.73%
Projected Fuel Expenditure Increase
$8.78
$13.67
$19.22
Second-Lowest Quintile Income Consumer Units
Fuel Expenditures
$1,702
$1,702
$1,702
Percent Impact on Fuel Expenditures
0.79%
1.23%
1.73%
Projected Fuel Expenditure Increase
$13.45
$20.93
$29.44
9.5 Greenhouse Gas Impacts
In 2009, under the "Endangerment and Cause or Contribute Findings for Greenhouse
Gases Under Section 202(a) of the Clean Air Act" (hereinafter the "Endangerment Finding"),
EPA considered how climate change threatens the health and welfare of the U.S. population. As
part of that consideration, we also considered risks to minority and low-income individuals and
communities, finding that certain parts of the U.S. population may be especially vulnerable based
on their characteristics or circumstances. These groups include economically and socially
disadvantaged communities; individuals at vulnerable life stages, such as the elderly, the very
young, and pregnant or nursing women; those already in poor health or with comorbidities; the
disabled; those experiencing homelessness, mental illness, or substance abuse; and/or Indigenous
386
-------
or minority populations dependent on one or limited resources for subsistence due to factors
including but not limited to geography, access, and mobility.
Scientific assessment reports produced over the past decade by the U.S. Global Change
Research Program (USGCRP),823'824 the Intergovernmental Panel on Climate Change
(IPCC),825'826'827'828 and the National Academies of Science, Engineering, and Medicine829'830
add more evidence that the impacts of climate change raise potential EJ concerns. These reports
conclude that poorer or predominantly non-white communities can be especially vulnerable to
climate change impacts because they tend to have limited adaptive capacities and are more
dependent on climate-sensitive resources such as local water and food supplies, or have less
access to social and information resources. Some communities of color, specifically populations
defined jointly by ethnic/racial characteristics and geographic location, may be uniquely
vulnerable to climate change health impacts in the U.S. In particular, USGCRP (2016) found
with high confidence that vulnerabilities are place- and time-specific, that particular life stages
823 USGCRP, 2018: Impacts, Risks, and Adaptation in the United States: Fourth National Climate Assessment,
Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C.
Stewart (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, 1515 pp. doi: 10.7930/NCA4.2018.
824 USGCRP, 2016: The Impacts of Climate Change on Human Health in the United States: A Scientific
Assessment. Crimmins, A., J. Balbus, J.L. Gamble, C.B. Beard, J.E. Bell, D. Dodgen, R.J. Eisen, N. Fann, M.D.
Hawkins, S.C. Herring, L. Jantarasami, D.M. Mills, S. Saha, M.C. Sarofim, J. Trtanj, andL. Ziska, Eds. U.S. Global
Change Research Program, Washington, DC, 312 pp. http://dx.doi.org/10.7930/J0R49NQX.
825 Oppenheimer, M., M. Campos, R.Warren, J. Birkmann, G. Luber, B. O'Neill, andK. Takahashi, 2014: Emergent
risks and key vulnerabilities. In: Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and
Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel
on Climate Change [Field, C.B., V.R. Barros, D.J. Dokken, K.J. Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatteijee,
K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, and
L.L.White (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, pp. 1039-
1099.
826 Porter, J.R., L. Xie, A.J. Challinor, K. Cochrane, S.M. Howden, M.M. Iqbal, D.B. Lobell, and M.I. Travasso,
2014: Food security and food production systems. In: Climate Change 2014: Impacts, Adaptation, and Vulnerability.
Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change [Field, C.B., V.R. Barros, D.J. Dokken, K.J. Mach, M.D. Mastrandrea,
T.E. Bilir, M. Chatteijee, K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken,
P.R. Mastrandrea, and L.L.White (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York,
NY, USA, pp. 485-533.
827 Smith, K.R., A.Woodward, D. Campbell-Lendrum, D.D. Chadee, Y. Honda, Q. Liu, J.M. Olwoch, B. Revich,
andR. Sauerborn, 2014: Human health: impacts, adaptation, and co-benefits. In: Climate Change 2014: Impacts,
Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change [Field, C.B., V.R. Barros, D.J. Dokken, K.J.
Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatteijee, K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S.
Kissel,A.N. Levy, S. MacCracken, P.R. Mastrandrea, and L.L.White (eds.)]. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, pp. 709-754.
828 IPCC, 2018: Global Warming of 1.5°C. An IPCC Special Report on the impacts of global warming of 1.5°C
above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the
global response to the threat of climate change, sustainable development, and efforts to eradicate poverty [Masson-
Delmotte, V., P. Zhai, H.-O. Portner, D. Roberts, J. Skea, P.R. Shukla, A. Pirani, W. Moufouma-Okia, C. Pean, R.
Pidcock, S. Connors, J.B.R. Matthews, Y. Chen, X. Zhou, M.I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T.
Waterfield (eds.)]. In Press
829 National Research Council. 2011. America's Climate Choices. Washington, DC: The National Academies Press.
https://doi.org/10.17226/12781.
830 National Academies of Sciences, Engineering, and Medicine. 2017. Communities in Action: Pathways to Health
Equity. Washington, DC: The National Academies Press, https://doi.org/10..1.7226/24624.
387
-------
and ages are linked to immediate and future health impacts, and that social determinants of
health are linked to greater extent and severity of climate change-related health impacts.
9.6 Effects on Specific Populations of Concern
EJ populations of concern, such as individuals living in socially and economically
disadvantaged communities (e.g., living at or below the poverty line or experiencing
homelessness or social isolation) or those who have been historically marginalized or
overburdened are at greater risk of health effects from climate change. This is also true with
respect to people at vulnerable life stages, specifically women who are pre- and perinatal, or are
nursing; in utero fetuses; children at all stages of development; and the elderly. Per the Fourth
National Climate Assessment (NCA4), "Climate change affects human health by altering
exposures to heat waves, floods, droughts, and other extreme events; vector-, food- and
waterborne infectious diseases; changes in the quality and safety of air, food, and water; and
stresses to mental health and well-being."831 Many health conditions such as cardiopulmonary or
respiratory illness and other health impacts are associated with and exacerbated by an increase in
GHGs and climate change outcomes, which is problematic as these diseases occur at higher rates
within vulnerable communities. Importantly, negative public health outcomes include those that
are physical in nature, as well as mental, emotional, social, and economic.
To this end, the scientific assessment literature, including the aforementioned reports,
demonstrates that there are myriad ways in which these populations may be affected at the
individual and community levels. Individuals face differential exposure to criteria pollutants, in
part due to the proximities of highways, trains, factories, and other major sources of pollutant-
emitting sources to less-affluent and traditionally marginalized residential areas. Outdoor
workers, such as construction or utility workers and agricultural laborers, who are frequently part
of already at-risk groups, are exposed to poor air quality and extreme temperatures without relief.
Furthermore, individuals within EJ populations of concern face greater housing and clean water
insecurity and bear disproportionate economic impacts and health burdens associated with
climate change effects. They tend to have less or limited access to healthcare and affordable,
adequate health or homeowner insurance. Finally, resiliency and adaptation are more difficult for
economically disadvantaged communities. They have less liquidity, individually and
collectively, to move or to make the types of infrastructure or policy changes to limit or reduce
the hazards they face. Finally, due to systemic challenges, affected communities may lack the
resources necessary to advocate for resources that would otherwise aid in resiliency and hazard
reduction and mitigation.
The assessment literature cited in EPA's 2009 and 2016 Endangerment Findings, as well
as USGCRP (2016), also concluded that certain populations and people in particular life stages,
including children, are most vulnerable to climate-related health effects. The assessment
literature produced from 2016 to the present strengthens these conclusions by providing more
831 Ebi, K.L., J.M. Balbus, G. Luber, A. Bole, A. Crimmins, G. Glass, S. Saha, M.M. Shimamoto, J. Trtanj, and J.L.
White-Newsome, 2018: Human Health. In Impacts, Risks, and Adaptation in the United States: Fourth National
Climate Assessment, Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel, K.L.M. Lewis, T.K.
Maycock, and B.C. Stewart (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, pp. 539-571.
doi: 10.7930/NCA4.2018.CH14
388
-------
detailed findings regarding related vulnerabilities and the projected impacts youth may
experience. These assessments—including NCA4 and USGCRP (2016)—describe how
children's unique physiological and developmental factors contribute to making them
particularly vulnerable to climate change. Impacts to children are expected from heat waves, air
pollution, infectious and waterborne illnesses, and mental health effects resulting from extreme
weather events. In addition, children are among those especially susceptible to allergens, as well
as health effects associated with storms, and floods. More generally, these reports note that
extreme weather and flooding can cause or exacerbate poor health outcomes by affecting mental
health because of stress; contributing to or worsening existing conditions, again due to stress or
also as a consequence of exposures to water and air pollutants; or by impacting hospital and
emergency services operations.832 Further, in urban areas in particular, flooding can have
significant economic consequences due to effects on infrastructure, pollutant exposures, and
drowning dangers. The ability to withstand and recover from flooding is dependent in part on the
social vulnerability of the affected population and individuals experiencing an event.833
Additional health concerns may arise in low-income households, especially those with children,
if climate change reduces food availability and increases prices, leading to food insecurity.
USGCRP (2016) also found that some communities of color, low-income groups, people
with limited English proficiency, and certain immigrant groups (especially those who are
undocumented) live with many of the factors that contribute to their vulnerability to the health
impacts of climate change. While difficult to isolate from related socioeconomic factors, race
appears to be an important factor in vulnerability to climate-related stress, with elevated risks for
mortality from high temperatures reported for Black or African American individuals compared
to white individuals after controlling for factors such as air conditioning use. Moreover, people
of color are disproportionately exposed to air pollution based on where they live, and
disproportionately vulnerable due to higher baseline prevalence of underlying diseases such as
asthma, so climate exacerbations of air pollution are expected to have disproportionate effects on
these communities.
Native American Tribal communities possess unique vulnerabilities to climate change,
particularly those communities impacted by degradation of natural and cultural resources within
established reservation boundaries and threats to traditional subsistence lifestyles. Tribal
communities whose health, economic well-being, and cultural traditions depend upon the natural
environment will likely be affected by the degradation of ecosystem goods and services
associated with climate change. The IPCC's Fifth Assessment Report of the Intergovernmental
Panel on Climate Change (AR5) indicates that losses of customs and historical knowledge may
cause communities to be less resilient or adaptable.834 NCA4 noted that while Indigenous
peoples are diverse and will be impacted by the climate changes universal to all Americans, there
are several ways in which climate change uniquely threatens Indigenous peoples' livelihoods and
832 USGCRP, 2018: Impacts, Risks, and Adaptation in the United States: Fourth National Climate Assessment,
Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C.
Stewart (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, 1515 pp. doi: 10.7930/NCA4.2018.
833 National Academies of Sciences, Engineering, and Medicine. 2019. Framing the Challenge of Urban Flooding in
the United States. Washington, DC: The National Academies Press, https://doi.org/10.17226/25381.
834 Porter et al., 2014: Food security and food production systems.
389
-------
economies.835 In addition, there can be institutional barriers (including policy-based limitations
and restrictions) to their management of water, land, and other natural resources that could
impede adaptive measures. For example, Indigenous agriculture in the Southwest is already
being adversely affected by changing patterns of flooding, drought, dust storms, and rising
temperatures leading to increased soil erosion, irrigation water demand, and decreased crop
quality and herd sizes. The Confederated Tribes of the Umatilla Indian Reservation in the
Northwest have identified climate risks to salmon, elk, deer, roots, and huckleberry habitat.836
Housing and sanitary water supply infrastructure are vulnerable to disruption from extreme
precipitation events. Confounding the general Indigenous response to natural hazards are
limitations imposed by policies such as the Dawes Act of 1887 and the Indian Reorganization
Act of 1934, which ultimately restrict Tribal peoples' autonomy regarding land-management
decisions through Federal trusteeship of certain Tribal lands and mandated Federal oversight of
management decisions.
Additionally, NCA4 noted that Indigenous peoples are subject to institutional racism
effects, such as poor infrastructure, diminished access to quality healthcare, and greater risk of
exposure to pollutants. Consequently, Indigenous people often have disproportionately higher
rates of asthma, cardiovascular disease, Alzheimer's disease, diabetes, and obesity. These health
conditions and related effects, such as disorientation and other effects, can all contribute to
increased vulnerability to climate-driven extreme heat and air pollution events, which may be
exacerbated by stressful situations, such as extreme weather events, wildfires, and other
circumstances.
NCA4 and AR5 also highlighted several impacts specific to Alaskan Indigenous Peoples.
Coastal erosion and permafrost thaw will lead to more coastal erosion, exacerbated risks of
winter travel, and damage to buildings, roads, and other infrastructure. These impacts on
archaeological sites, structures, and objects that will lead to a loss of cultural heritage for
Alaska's Indigenous people. In terms of food security, NCA4 discussed reductions in suitable ice
conditions for hunting, warmer temperatures impairing the use of traditional ice cellars for food
storage, and declining shellfish populations due to warming and acidification. While NCA4 also
noted that climate change provided more opportunity to hunt from boats later in the fall season or
earlier in the spring, the assessment found that the net impact was an overall decrease in food
security.
835 Jantarasami, L.C., R. Novak, R. Delgado, E. Marino, S. McNeeley, C. Narducci, J. Raymond-Yakoubian, L.
Singletary, and K. Powys Whyte, 2018: Tribes and Indigenous Peoples. In Impacts, Risks, and Adaptation in the
United States: Fourth National Climate Assessment, Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling,
K.E. Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C. Stewart (eds.)]. U.S. Global Change Research Program,
Washington, DC, USA, pp. 572-603. doi: 10.7930/NCA4.2018.CH15.
836 Confederated Tribes of the Umatilla, Indian Reservation, 2015. Climate Change Vulnerability Assessment.
Nasser, E., Petersen, S., Mills, P. (eds). Available online: www.ctnir.org.
390
-------
Chapter 10: Estimated Costs and Fuel Price Impacts
The statute directs EPA to assess the impact of the use of renewable fuels on the cost to
consumers of transportation fuel and on the cost to transport goods in using the set authority. In
this chapter, we assess the social costs of renewable fuels, the social costs of the petroleum fuels
which the biofuels replace, the fuel economy effect based on each fuel's energy density, and the
impacts of this rule on social costs, the costs to consumers of transportation fuel, and the costs to
transport goods.
The costs are analyzed for the renewable fuel volumes in 2023 through 2025 relative to a
No RFS baseline. Costs are also calculated for the incremental increase in renewable fuel
volumes relative to the year 2022 renewable fuel volumes established in the recent 2020-2022
final rule.837 In both cases, costs are reported in 2022 dollars. Chapter 2 contains a summary of
the baseline volumes, and Chapter 3 contains the candidate volumes analyzed. Chapters 10.4.2.1
and 10.4.3.1 contain the change in candidate volumes relative to the No RFS and 2022 baselines,
respectively, as well as the estimated change in fossil fuel volumes displaced by the change in
volume of renewable fuels.838
10.1 Renewable Fuel Costs
10.1.1 Feedstock Costs
For most renewable fuels, the feedstock costs are a primary contributing factor to the cost
to produce and use the renewable fuels. We first estimate the production cost for these feedstocks
prior to providing information for the production, distribution and blending costs for the various
renewable fuels.
In calculating feedstock costs, we used projections of feedstock prices for 2023 through
2025 from multiple sources, including EIA and USDA.839 We also made adjustments to account
for differences between these projections. Specifically, the projected feedstock prices are
adjusted to account for different crude oil prices used by USDA than those projected by EIA, and
to adjust the projected nominal prices to constant year 2022 dollars.840
837 87 FR 39600 (July 1, 2022).
838 The spreadsheet used to estimate the costs for the candidate volumes relative to the No RFS and 2022 volumes is
available in the docket for this action: "Estimated Fuel Costs for the Set Final Rule".
839 USDA Agricultural Projections to 2032; Long Term Projections Report; February 2023
840 Crude oil prices affect the cost for growing renewable fuels feedstocks, the cost to transport them to the
renewable fuels production plants, the cost for transporting the produced renewable fuels from the plant to market,
and may impact the cost for producing the renewable fuels. Because USDA agricultural price projections were based
on lower crude oil price projections than that by EIA, the USDA agricultural price projections may have
underestimated the agricultural prices that would be consistent with the EIA petroleum price projections. Therefore,
the USDA price projections for both corn and soybean oil were adjusted in an attempt to remove this potential bias
in the cost analysis.
391
-------
10.1.1.1 Corn and Corn Ethanol Plant Byproducts
The price of corn is the most important input to estimating the cost of corn ethanol. Table
10.1.1.1-1 shows the derivation of the corn prices used in this cost analysis, which adjusts the
projected prices for crude oil price differences and for inflation. To help to explain the derivation
in the discussion below, we refer to the relevant row number in Table 10.1.1.1-1.
As a starting point we used future corn price projections from USDA. We started with the
2023 through 2025 USDA projected corn prices (row #1).841 However, the USDA corn prices are
reported in nominal dollars, reflecting the inflated value of the dollars in those years. The first
adjustment we made was to convert those USDA corn prices reported in nominal dollars into the
2022 dollars used across this cost analysis (row #2). 842
Next, we made an adjustment to account for the different crude oil price projections that
USDA used (row #3) compared to those projected by EIA (row #5).843 Because EIA is the U.S.
reference organization for projecting petroleum prices, we adjusted the USDA inflation-adjusted
corn prices to put them on the same basis with the petroleum costs which are based on EIA crude
oil prices. To do so, we first adjusted the crude oil prices used by USDA (row #3) to 2022 dollars
(row #4). Then we used a regression of corn prices and crude oil prices to estimate the corn
prices at USDA crude oil prices adjusted to 2022 dollars (row #6) and the corn prices at the EIA
crude oil prices (row #6), to enable an adjustment of USDA corn prices to be consistent with the
EIA crude oil prices. The regression of corn prices and crude oil prices is based on monthly corn
prices between April 2008 and September 2017, which yielded the following equation:844
Corn Price ($/bushel) = Crude Oil Price ($/bbl) x 0.0366 + 1.81
The corn prices estimated by this regression was not used directly for the cost analysis
because farmers are more efficient at producing corn today than in the past, and corn production
is likely to be on a different supply/demand point on the corn price curve as evidenced by
today's higher corn prices. Instead, the difference in regressed corn prices (row #8) was added to
the USDA corn prices adjusted to 2022 dollars (row #2) to derive the final adjusted corn prices
(row #9) subsequently used as an input value for estimating corn ethanol costs as shown in Table
10.1.1.1-1.
841 USDA Agricultural Projections to 2032; Long Term Projections Report; February 2023.
842 USDA reports estimated future inflation rates which are used for the nominal dollar to 2021$ adjustment.
843 There seems to be an association between the renewable fuels feedstock costs and crude oil prices (regression
analysis reveals an R-squared of 0.56 for corn and crude oil). Since USDA estimated renewable fuel feedstock
prices based on lower crude oil prices, adjusting their renewable fuel feedstock prices higher to be consistent with
EIA crude oil prices better syncs the two price projections and leads to abetter estimate of costs.
844 The years from 2008 to 2017 were chosen because of the wide range in crude oil prices which existed over this
time period, and a 10-year time period was chosen to provide enough data for a quality regression.
392
-------
Table 10.1.1.1-1: Derivation of Corn Feedstock Production Costs ($/bushel for corn, $/bbl
for Crude OiF
Row #
2023
2024
2025
Corn Prices
USDA Nominal $
1
6.80
5.70
4.90
USDA 2022$
2
6.78
5.65
4.76
Crude Oil
USDA Nominal $
3
75.7
72.7
71.8
Prices
USDA 2022$
4
75.5
72.1
69.8
EIA2022$
5
83.18
89.12
83.70
Regressed
Based on USDA 2022
6
3.98
3.89
3.82
Corn Prices
Based on EIA 2022
7
4.86
5.08
4.88
Corn Prices
Difference in Regressed Corn
Prices EIA - USDA
8
0.87
1.19
1.06
Corn Prices
Adjusted USDA 2022$
9
7.65
6.84
5.82
Both the inflation and crude oil price adjustment are modest, and their effects cause
offsetting effects. Also, these adjustments are well within the recent variation in corn prices.
Since corn ethanol plants also produce byproducts which can be sold for additional value,
we also estimated the prices for those byproducts, specifically DDGS and corn oil, which is
estimated below in Chapter 10.1.1.2. Since USDA does not estimate future prices for DDGS,
these were obtained by agricultural price projections made by the University of Missouri, Food
and Agricultural Policy Research Institute (FAPRI).845 The FAPRI DDGS projected prices are
reported in nominal dollars, so we adjusted the price projections to 2022 dollars. Table 10.1.1.1-
2 summarizes DDGS prices used in the cost analysis.
Table 10.1.1.1-2: DDGS Prices (2022 dollars)
Year
DDGS Prices ($/dry ton)
2023
175.5
2024
232.4
2025
183.4
10.1.1.2 Soybean Oil, Corn Oil and Fats, Oil and Grease Prices
Soybean oil, waste fats, oils, and greases (FOG), corn oil, and canola oil were identified
in Chapter 2 as the feedstocks for producing biodiesel and renewable diesel fuel. For the cost
analysis, canola oil volumes are combined with the soybean oil volume to estimate a single
soybean oil volume. We believe it is a reasonable cost assumption to combine canola oil with
soybean oil because they are both virgin vegetable oils and would likely have similar production
costs. Soybean oil price projections made by USDA are used as a starting point for this cost
analysis.846
845 U.S. Agricultural Market Outlook, Food and Agricultural Policy Research Institute (FAPRI); FAPRI-MU Report
#02-22; March 2023.
846 USDA Agricultural Projections to 2032; Long Term Projections Report; February 2023.
393
-------
We followed the same methodology we used for corn prices described above, but for soy
oil prices this process is summarized in Table 10.1.1.2-1 and the description that follows
references the rows in that Table to aid in understanding. The first step required converting
USD A projected soy oil prices in nominal dollars (row #1) to 2022 dollars (row #2), and then
adjusting for the differences in crude oil prices (row #4 for USDA in 2022 dollars) and EIA (row
#5). When adjusting for the differences in crude oil prices, a regression of monthly soy oil and
crude oil prices between January 2012 and September 2017 yielded the following equation:847
Soy Oil Price ($/lb) = Crude Oil Price ($/bbl) x 0.259 + 19.06
The soy oil prices (row #6) based on USDA crude oil prices and soy oil prices (row #7)
based on EIA crude oil prices were not used in the cost analysis directly. Rather the difference in
regressed soy oil prices (row #8) was added to the adjusted USDA soy prices (row #2) to derive
the adjusted soy oil prices (row #9).
Table 10.1.1.2-1: Derivation of Soy Oil Feedstock Production Costs (cents/pound for soy oil,
$/bbl for crude oil)
Row #
2023
2024
2025
Soy Oil Prices
USDA Nominal $
1
69
57
50.5
USDA 2022$
2
68.8
56.5
49.1
Crude Oil
Prices
USDA Nominal $
3
75.7
72.7
71.8
USDA 2022$
4
75.5
72.1
69.8
EIA 2022$
5
83.18
89.12
83.70
Regressed Soy
Based on USDA 2022$
6
34.44
33.75
33.28
Oil Prices
Based on EIA 2022$
7
40.62
42.16
40.75
Difference in
Soy Oil Prices
Regressed Soy Oil
Prices EIA - USDA
8
6.18
8.40
7.47
Soy Oil Prices
Adjusted USDA 2022$
9
75.0
64.9
56.6
Neither USDA nor FAPRI project future corn oil or FOG prices. Instead, future prices for
these oils were estimated based on the historical differences between them and soybean oil's spot
prices.848 Corn oil and FOG spot prices were compared to soybean oil spot prices between
January 2016 and December 2022. Over that time period, soybean oil averaged about 40 cents
per pound (ranged from 29 to 68 cents per pound. Corn oil and FOG prices were compared to
soy oil prices, and these were priced at 82.7 percent and 75.4 percent of soybean oil,
respectively.
The additional demand for vegetable oils associated with this rulemaking is expected to
increase the price for those oils. A previous review of increased demand for soybean oil on
soybean oil prices found that increases of 200 million gallons of soybean oil increased the
847 There seems to be an association between the renewable fuels feedstock costs and crude oil prices (regression
analysis reveals an r-squared of 0.73 for soybean oil and crude oil). Since USDA estimated renewable fuel feedstock
prices based on lower crude oil prices, adjusting their renewable fuel feedstock prices higher to be consistent with
EIA crude oil prices better synchronizes the two price projections and leads to abetter estimate of costs.
848 USDA Yearbook Tables by the Economic Research Service, downloaded March 2023.
394
-------
soybean oil price by 0.032 dollars per pound.849 This price increase estimate was used to adjust
the soybean oil prices for this analysis based on the estimated increase of soybean oil demand
under this action. Similar adjustments are made for FOG and corn oil, the markets of which are
around 20 percent the size of the soy oil market. The projected soy oil prices in the baseline and
resulting from the increased demand for the candidate volume that are used in this cost analysis
are summarized in Table 10.1.1.2-2, along with the projected prices for FOG and corn oil.
Table 10.1.
2-2: Projected Vegetable Oil Production Costs (2022 dollars/pound)
Base Prices
Projected Vegetab
e Oil Prices
Soybean Oil
Soybean Oil
FOG
Corn Oil
2023
75
82.1
75.8
60.7
2024
65
79.1
61.2
53.1
2025
57
80.0
58.7
46.8
10.1.1.3 Biogas
For this analysis we assume that biogas is produced at landfills and collected to prevent
the release of methane gas as required by regulation, and then flared, burned to produce
electricity, or upgraded for use as natural gas. Since the biogas is a waste gas from existing
landfills, we assumed no feedstock cost for biogas. The cost of the necessary steps to collect,
purify, and distribute the biogas are all discussed under the sections discussing production and
distribution costs.
10.1.2 Renewable Fuels Production Costs
This section assesses the production costs of renewable fuels, including the feedstock
costs described above as well as the capital, fixed, and operating costs. We generally express the
production costs on a per-gallon basis for the renewable fuels being produced. The one exception
is biogas which is reported on a per-million BTU basis and also on a per ethanol-equivalent
volume basis.
Detailed cost summaries presented for each renewable fuel in this section are based on
2023 cost inputs.850 All the costs summarized in this section for all years are calculated in a
spreadsheet which is available in the docket for this rulemaking.851
10.1.2.1 Cost Factors
10.1.2.1.1 Capital and Fixed Costs
The economic assumptions used to amortize capital costs over the production volume of
renewable fuels are summarized in Table 10.1.2.1.1-1. These capital amortization cost factors are
849 Shelby, Michael; Cost Impacts of the Final 2019 Annual Renewable Fuel Standards; Memorandum to EPA Air
and Radiation Docket EPA-HQ-OAR-2018-0167.
850 Table 10.1.2.2-1, Table 10.1.2.4-1, Table 10.1.2.5-3, Table 10.1.2.5-4, Table 10.1.2.5.1-1, Table 10.1.2.5.2-2,
10.1.2.5.2-3, and 10.1.2.5.2-4.
851 Estimated Fuel Costs for SET Final Rule.xlsx.
395
-------
used in the following section for converting the one-time, total capital cost to an equivalent per-
gallon cost.852 The resulting 0.11 capital cost amortization factor is the same factor used by EPA
in the cost estimation calculations made for other rulemakings and technical
853,854,855,856,857
pdpcl
Table 10.1.2.1.1-1: Economic Cost Factors Used in Calculating Capital Amortization
Factors
Resulting
Economic
Federal and
Return on
Capital
Amortization
Depreciation
and Project
State Tax
Investment
Amortization
Scheme
Life
Life
Rate
(ROI)
Factor
Societal Cost
10 Years
15 Years
0%
7%
0.11
Capital costs were adjusted to 2022 dollars for this analysis. The Chemical Engineering
Plant Index (CEPI) capital cost index was used to adjust capital costs to 2022 dollars. Consistent
with the increased inflation observed over the past year, the CEPI capital cost index for 2022
represents a large increase in capital costs when adjusting capital costs to the year 2022.
Fixed operating costs include the maintenance costs, insurance costs, rent, laboratory
charges and miscellaneous chemical supplies.858 Maintenance costs can range from 1% to 8% for
industrial processes.859 We estimated the aggregated annual fixed operating costs to be 5.5% of
the capital costs for all renewable fuels production facilities.
10.1.2.1.2 Utility and Fuel Costs
Utility and Fuel inputs are variable operating costs incurred to run the renewable fuel
production plants on a day-to-day basis, and are based on the unit throughput. The most obvious
of the variable costs are utilities (electricity, natural gas, and water) which are required to operate
the renewable fuels plants. Natural gas is consumed for heating process streams, including
feedstocks which must be heated prior to being sent to reactors and distillation columns for
separating coproducts. Electricity is necessary to run pumps, compressors, plant controls and
852 The capital amortization factor is applied to the aggregate capital cost to create an amortized annual capital cost
which occurs each and every year for the 15 years of the economic and project life of the unit. The depreciation rate
of 10% and economic and project life of 15 years are typical for these types of calculations. The 7% return on
investment and the zeroing out of Federal and State taxes is specified by the Office of Management and Budget for
these calculations (Office of Management and Budget; Circular A-4; Regulatory Impact Analysis: A primer;
https://www.reginfo.gov/public/isp/Utilities/cirailar-a-4 regulatorv-impact-analysis-a-primer.pdf).
853 Regulatory Impact Analysis - Control of Air Pollution from New Motor Vehicles: Tier 2 Motor Vehicle
Emissions Standards and Gasoline Sulfur Control Requirements, EPA420-R-99-023, December 1999.
854 Cost Estimates of Long-Term Options for Addressing Boutique Fuels; Memorandum from Lester Wyborny to the
Docket; October 22, 2001.
855 Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control
Requirements; EPA420-R-00-028; December 2000.
856 Final Regulatory Analysis: Control of Emissions fromNonroad Diesel Engines; EPA420-R-04-007; May 2004.
857 Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis; EPA-420-R-10-006; February 2010.
858 Peters, Max S., Timmerhaus, Klaus, D.; Plant Design and Economics for Chemical Engineers 3rd Edition;
McGraw Hill; 1980.
859 McNair, Sam Budgeting for Maintenance: A Behavior-Based Approach, Life Cycle Engineering, 2011.
396
-------
other plant operations. Water can be necessary as part of the process (reaction medium), or used
in heat exchangers and cooling towers.
Projected electricity and natural gas prices are based on national average values from
Energy Information Administration's (EIA) 2023 Annual Energy Outlook.860 The cost of process
water is generally quite minimal, but a cost is estimated for it nonetheless since renewable fuels
technologies can use fairly large quantities. 861'862The utility costs used for the cost analysis are
summarized in Table 10.1.2.1.2-1.
Table 10.1.2.1.2-1: Summary of Utility Cost Factors (2022 dollars)3
Year
Natural Gas
(S/1000 cf)
Electricity
(c/kWhr)
Water
(S/1000 gals)
2023
6.76
8.18
3.0
2024
5.56
7.56
3.0
2025
4.95
7.15
3.0
a c/kWh is cents per kilowatt-hour; $/1000 cf is dollars per thousand standard cubic feet; $/1000 gallons is dollars
per thousand gallons.
10.1.2.2 Corn Ethanol Production Costs
Corn ethanol plant input and output information were based on a 2019 survey of corn
ethanol plants, although some plant information was sourced from an older analysis.863'864
Capital costs were based on a review of corn ethanol construction costs for a 100 million gallon
per year drymill corn ethanol plant in 2016. For this analysis the capital costs were scaled to the
U.S. average sized corn ethanol plant with a nameplate capacity of 85 million gallons per year
assumed to operate at 90% of nameplate capacity, therefore producing 76 million gallons of
ethanol per year.865 Since the capital cost is based on the total construction cost of already
constructed corn ethanol plants, no contingency cost factors are applied to the capital costs. Corn
prices are farm gate prices and a transportation spreadsheet was used to estimate a cost of 6 cents
per bushel to transport the corn to a corn ethanol plant.866 Of the corn ethanol plants in the 2012
survey, 74% were separating and selling corn oil, however, we believe that by now all corn
ethanol plants are separating and selling corn oil.
The quantity of dried distillers grain with solubles (DDGS) produced by corn ethanol
plants was estimated from USDA DDGS production data from February to October, 2022.
USDA reports DDGS production for four different categories of DDGS: dried distillers grain
(DDG), dried distillers grain with solubles (DDGS), distillers wet grain (DWG) with 65 or more
860 Annual Energy Outlook 2023, Energy Information Administration, March 2023.
861 Haas, M.J, A process model to estimate biodiesel production costs, Bioresource Technology 97 (2006) 671-678.
862 Water and Wastewater Annual Escalation Rates for Selected Cities across the United States, Office of Energy
Efficiency and Renewable Energy, Department of Energy; September 2017.
863 Lee, Uisung; Retrospective analysis of the U.S. corn ethanol industry for 2005 - 2019; implications for
greenhouse gas emission reductions;: Biofpr; May 4, 2021.
864 Mueller, Steffen; 2012 Corn Ethanol: Emerging Plant Energy and Environmental Technologies; April 29, 2013.
865 Irwin, Scott; Weekly Output: Ethanol Plants Remain Barely Profitable; 3/16/2018.
866 Edwards, William; Grain Truck Transportation Cost Calculator (a3-29graintransportation.xlsx version
1.4 82017); Iowa State University.
397
-------
percent moisture, and distillers wet grain (DWG) with 40 to 64 percent moisture. The production
quantity of DWG is adjusted to an equivalent of dried DDG. The DWG with 65 or more percent
moisture is assumed to have 75 percent moisture, while the DWG with 40 to 65 percent moisture
is assumed to have 52 percent moisture. Both wet distiller grain categories are adjusted to dry
distiller grain quantities assuming that dried distiller grains contains 11 percent moisture.867
Table 10.1.2.2-1 summarizes and averages the quantity of distiller grains by category, reporting
the quantity of wet distiller grains both before and after adjusting them to equivalent dry grains
amounts.
Table 10.1.2.2-1: USDA-reported DDG (tons) and Corn Ethanol (million gallons)
Production for a portion of 2022 i
February
March
April
May
June
July
August
September
October
Average
DWG 65%+
Wet
1,293,312
1,382,790
1,321,275
1,328,402
1,283,359
1,279,210
1,322,744
1,249,996
1,397,867
-
DWG 40-65
Wet
492,839
562,599
517,270
468,772
494,792
495,386
544,168
498,142
490,060
-
DWG 65%+
Dry
363,290
388,424
371,145
373,147
360,494
359,329
371,557
351,122
392,659
370,130
DWG 40-65
Dry
287,951
328,710
302,225
273,889
289,092
289,439
317,941
291,049
286,327
296,291
DDG
303,788
372,813
328,691
322,855
346,591
334,122
335,885
281,984
388,993
335,080
DDGS
1,693,253
1,877,338
1,704,698
1,896,665
1,918,611
1,934,
355
1,867,735
1,613,088
1,745,419
1,805,685
Total
DDG/DDGS
2,626,132
2,626,132
2,626,132
2,626,132
2,626,132
2,626,132
2,626,132
2,626,132
2,626,132
2,626,132
Ethanol
Volume
1,189
1,327
1,217
1,315
1,314
1,322
1,280
1,139
1,321
1,269
Pounds
DDGS/gal
ethanol
4.42
4.43
4.41
4.33
4.40
4.38
4.48
4.42
4.23
4.39
After averaging the production volume of each grain type over the 9 months, they are
totaled and divided by the average ethanol production volume. This analysis estimates DDGS
production to be 4.4 pounds per gallon of ethanol produced (12.5 pounds per bushel of corn).
Corn oil production from ethanol plants was estimated using a similar analysis as that
conducted for DDGS. The corn oil production by month is summarized and averaged in Table
10.1.2.2.-2.
Table 10.1.2.2-2: USDA-reported Corn C
il Production for a pon
tion of 2022 (tons)
February
March
April
May
June
July
August
September
October
Average
Corn Oil
154,933
174,657
163,024
177,158
184,350
187,853
180,062
159,873
186,770
174,298
To estimate the corn oil production from corn ethanol plants, the average corn oil
production is divided by the average corn ethanol production volume summarized in Table
10.1.2.2-1. Based on this analysis, corn oil production from corn ethanol plants is estimated to be
0.27 lbs per gallon of ethanol (0.79 pounds per bushel of corn).
Table 10.1.2.2-3 contains the plant demand and outputs and capital costs for corn ethanol
plants and provides an estimate of the estimated corn ethanol production cost for year 2023.
867 Shurson, Jerry; DDGS present handling and storage considerations; National Hog Farmer; May 29, 2019.
398
-------
Table 10.1.2.2-3: Corn Ethanol Plant Demands, Production Levels, and Capital Costs for
2023 (2022 dollars)
Category of Plant
Input/Output
Plant
Inputs/Outputs
Cost per Input
Cost (MM$)
Cost ($/gal)
Ethanol Yield
2.86 Gal/Bushel
7.71 $/bushel
206
2.70
DDG Yield
4.4 Lbs/Gal
175 $/ton
-29
-0.38
Corn Oil Yield
0.27 Lbs/gal
60 cents/lb
-12.5
-0.16
C02 Yield
1 lb/Gal
$12/ton
Thermal Demand
22,480 BTU/Gal
6.76 $/1000 cf
11.2
0.15
Electricity Demand
0.63 kWh/Gal
8.18 c/kwh
3.9
0.05
Water Use
2.7 Gal/Gal
$3/1000 gals
0.6
0.01
Labor Cost
0.07 $/Gal
-
5.3
0.07
Capital Cost (2022
dollars,
76 million Gals/Yr)
3.27 $/Gal Plant
32.5
0.43
Capital cost
Annual Fixed Cost
5.5% of Total
Capital Cost
16.3
0.21
Denaturant
2 volume percent
-1.2
-0.02
Total Cost
231
3.02
The projected corn ethanol social production cost for an 85 million gallon capacity
ethanol plant producing 76 million gallons per year of ethanol is $3.02 per gallon of denatured
ethanol for 2023, $2.61 for 2024, and $2.37 for 2025. The downward trend in estimated per-
gallon production costs reflect the expected downward trend in corn prices.
10.1.2.3 Biodiesel Production Costs
Biodiesel production costs for this rule were estimated using an ASPEN cost model
developed by USD A for a 38 million gallon-per-year transesterification biodiesel plant
processing degummed soybean oil as feedstock. Details on the model are given in a 2006
technical publication by Haas.868'869 Although dated, this model likely still provides
representative cost estimates because the process is fairly simple and unlikely to have changed
over time, and consequently its cost are likely to be fairly stable over time as well. Furthermore,
the biodiesel costs are primarily (>80%) determined by the feedstock prices.
The biodiesel process comprises three separate subprocesses:
1. Transesterification to produce fatty acid methyl esters (biodiesel) and coproduct
glycerol (glycerine);
2. Biodiesel purification to meet biodiesel purity specifications; and
868 Haas, M.J, A process model to estimate biodiesel production costs, Bioresource Technology 97 (2006) 671-678.
869 Since 2006 when the HAAS biodiesel plant survey was conducted, biodiesel plants may have achieved improved
energy efficiency, but also experienced increased costs to improve product quality and expand the quality of
feedstocks they can process.
399
-------
3. Glycerol recovery.870
For the transesterification process modeled by Haas, soybean oil is continuously fed
along with methanol and a catalyst sodium methoxide to a stirred tank reactor heated to 60 °C.
After a residence time of 1 hour, the contents exit the reactor and the glycerol is separated using
a centrifuge and sent to a glycerol recovery unit. The methyl ester stream, which contains
unreacted methanol and catalyst, is sent to a second reactor along with additional methanol and
catalyst. Again, the reactants reside in the second stirred tank reactor for 1 hour heated to 60 °C.
The products from of the second reactor are fed to a centrifuge which again separates the
glycerol from the other reactants. The reaction efficiency is assumed to be 90% in each reactor,
consistent with published reports, resulting in 99% combined conversion in both reactors.
The methyl ester is purified by washing with mildly acidic (4.5 pH) water to neutralize
the catalyst and convert any soaps (sodium or potassium carboxylic acids) to free fatty acids. The
solution is then centrifuged to separate the biodiesel from the aqueous phase. The remaining
water in the biodiesel is removed by a vacuum dryer to a maximum 0.05% of water by volume.
The glycerol can have a high value if it can be purified to U.S. Pharmacopia (USP) grade
to enable using this material for food or medicine. However, this purification process is
expensive. Most biodiesel plants create a crude glycerol (glycerine) grade, which is 80%
glycerol, and sell the crude glycerol for further refining by others. To create the crude glycerol,
the various glycerol streams are combined and treated with hydrochloric acid to convert the
soaps to free acids, allowing removal by centrifugation and sending to waste. The glycerol
stream is then neutralized (pH brought back up to neutral) with caustic soda. Methanol is
recovered from this stream by distillation and the methanol is recycled back into the process. The
glycerol stream is distilled to remove it from the remaining water, which is recycled back into
the process. The glycerol is now at least 80% pure, adequate to sell as crude glycerol.
We made a series of adjustments to the Haas model output. The capital cost is adjusted
from 2006 dollars to 2022 dollars using a ratio of the capital cost index from the Chemical
Engineering Cost Index. This adjustment increased installed capital cost from $11.9 million to
$14.5 million. Fixed operating costs are estimated to comprise 5.5% of the plant cost. Prices
were found on the Web for methanol,871 sodium methoxide,872 hydrochloric acid,873 sodium
hydroxide,874 and glycerine.875 876 The value of methanol is from a Methanex report plus 15
870 Haas, M.J, A process model to estimate biodiesel production costs, Bioresource Technology 97 (2006) 671-678.
871 Methanex; current North America prices plus 15 c/gal for shipping; https://www.methanex.com/onr-
business/pricing: January 31 2023.
872 Alibaba; https://www.alibaba.com/prodnct-detail/Food-Grade-Pnritv-28-31-Colorless 1600468349215.html:
February 2023.
873 Chemanalyst; https://www.chemanalvst.com/Pricing-data/tivdrochloric-acid-61: February 2023.
874 eBioChem; http://www.ebiochem.com/prodnct/canstic-soda-sodinm-hvdroxide-l.6515: February 2023.
875. Alibaba MMs://wi^
Refined 1600713799582.html?spm=a2700.7724857.0.0.4b943a616FYg3J&s=p: February 2023.
876 Irwin, Scott; 2021 Was a Devastating Year for Biodiesel Production Profits; Farmdoc Daily, February 16, 2022.
400
-------
cents per gallon distribution costs.877 Prices for sodium methoxide, hydrochloric acid, and
sodium hydroxide are all bulk prices from a chemicals supplier.878
The value of the glycerin co-product has been volatile due to a large increase in
production in biodiesel facilities that has been balanced at times by new uses. Glycerine has
traditionally been used for petrochemical-based products, but there is increased demand in
personal care and other consumer products as the standard of living increases in many parts of
the world. Some facilities are even experimenting with using it as a supplemental fuel.879 We can
expect that new uses for glycerin will continue to be found as long as it is plentiful and cheap.
We use recent cost information of about 25 cents per pound for glycerine.
Table 10.1.2.3-1 also shows the production cost allocation for the soybean oil-to-
biodiesel facility. Production cost for biodiesel is primarily a function of feedstock price, with
other process inputs, facility, labor, and energy comprising much smaller fractions.
Table 10.1.2.3-1: Biodiesel Production Cost for 2023 (year 2
022 dollars)
Thousand
Unit Demands
Cost per Unit
Dollars
$/gal
Soybean Oil Feed
76,875 (1000 lb)
82 cents/lb
63,109
6.31
Methanol
7422(1000 lb)
1.88 $/gal
2,170
0.22
Sodium Methoxide
927 (1000 lb)
$800/ton
371
0.037
Hydrochloric Acid
529 (1000 lb)
$150/MT
36.1
0.004
Sodium Hydroxide
369 (1000 lb)
$420/ton
77.5
0.008
Water
2478 (1000 lb)
$3/1000 gals
1.2
0.00
Glycerine
9000(1000 lb)
24 cents/lb
(2160)
(0.22)
Natural Gas
66.9 million cf
$6.76/1000cf
452
0.045
Electricity
1008 kW
8.18 cents/kWh
722
0.072
Labor
0.05
Capital Cost 2006$
11.35 ($million)
-
-
-
Capital Cost 2022$
18.54 ($million)
2,039
0.20
Fixed Cost
5.5%
1,019
0.10
Total Cost
67,838
6.83
As shown in Table 10.1.2.3-1, biodiesel produced from soybean oil is estimated to cost
6.83 cents per gallon in 2023. The estimated biodiesel production cost for all vegetable oil types
and for all three years is summarized in Table 10.1.2.3-2.
877 Methanex Methanol Price Sheet; US Gulf Coast; January 31,2022
878 https://www.alibaba.com.
879 Yang, Fangxia; Value-added uses for crude glycerol - a byproduct of biodiesel production; Biotechnology for
Fuels; March 14, 2012.
401
-------
Table 10.1.2.3-2: Summary of Estimated Biodiesel Production Costs ($/gal)
Year
Soy Oil
Corn Oil
FOG
2023
6.83
5.19
6.35
2024
6.59
4.59
5.21
2025
6.65
4.10
5.01
10.1.2.4 Renewable Diesel Production Costs
The renewable diesel process converts plant oils or rendered fats into diesel or jet fuel
using hydrotreating. The process reacts hydrogen over a catalyst to remove oxygen from the
triglyceride molecules in the feedstocks oils via a decarboxylation (removal of a carbon molecule
double-bonded to an oxygen molecule producing carbon dioxide) and hydro-oxygenation
reaction, yielding some light petroleum products, carbon dioxide, and water as byproducts. The
reactions also saturate the olefin bonds in the feedstock oils, converting them to paraffins, and
may also isomerize some paraffins. Depending on process operating conditions, the yield of
product which can be blended into diesel fuel is typically between 90-95% by volume, with the
rest being naphtha and light fuel gases (primarily propane). In total, the volumetric yield is
greater than 100% of the feed due to the cracking that occurs over the hydrotreating catalyst.
Besides the renewable diesel product, propane (light gas output), water and carbon dioxide are
also produced. The byproducts created from that first reactor are separated from the renewable
diesel in a separation unit.
For this cost analysis we chose to focus on stand-alone renewable diesel production. We
found a project cost estimate by Diamond Green which was $1,100 million for a standalone 400
million gallon per year facility. This large plant size and its associated capital costs were scaled
down to a 220 million gallon per year plant size which is more typical of the renewable diesel
fuel plants being built for start-up through 2022.880 The capital cost for this smaller renewable
diesel fuel plant is estimated to be $768 million.
In addition to feedstock and facility costs, another significant cost input is hydrogen. We
used an estimate provided by Duke Biofuels for our hydrogen consumption estimate for
producing renewable diesel. On average, vegetable and waste oil feedstocks require 2,000
SCF/bbl of feedstock processed.881 Hydrogen costs are estimated based on a 50 million standard
cubic feet per day steam methane reforming hydrogen plant, adjusted to represent a 32 million
cubic feet per day plant which would be the volume required for a typical sized 220 million
gallon per year renewable diesel plant.882
880 The typical renewable diesel plant size is based on volume-weighting the renewable diesel capacity data in Table
6.2.2-1. The cost for the smaller sized renewable diesel plant is scaled using a six-tenths factor which captures the
higher per gallon cost of a smaller sized plant. The cost scaling is calculated using the following equation: (new
plant size/original plant size) raised to the 0.6 power and multiplied by the capital cost of the original plant size.
881 Conversation with Mike Ackerson, Duke Biofuels, May 2020.
882 Meyers, Robert A; Handbook of Petroleum Refining Processes; 4ed., 2016; McGraw Hill.
402
-------
Table 10.1.2.4-1: Hydrogen Plant Cosl
s
Unit Demands
for a 50 mm
cf/day plant
Cost per Unit
Cost for a 32 mmcf/day plant
Million Dollars
$/thousand FT3
Feed Natural Gas
730 mmBTU/hr
6.76 $/mmBTU
27.6
2.37
Fuel Gas for Heat
150 mmBTU/hr
6.76 $/mmBTU
5.7
0.49
Power
1200 KW
8.18 c/KWh
0.6
0.05
Boiler feed water
160,000 lb/hr
$3/thousand
gallons
0.3
0.03
Cooling water
900 gal/min
$3/thousand
gallons
0.9
0.08
Export Steam
120,000 lb/hr 600
psi
-5.9
-0.50
Capital Cost
$70 MM in 2016
For a 50 mm
cf/day plant
$81 MM in 2022
For a 32 mm
cf/day plant
8.9
0.769
Fixed Cost
6.7%
5.47
0.46
Total Cost
43.5
3.70
Based on our cost analysis, hydrogen is estimated to cost $3.70 per thousand standard
cubic feet.
Our yield estimates as summarized in Table 10.1.2.4-2 were derived from material
presented by UOP and Eni at a 2007 industry conference, which describes producing renewable
diesel in a grass roots standalone production process inside a refinery.883 Despite the age of the
reference, the underlying chemistry is unlikely to have changed appreciably.
Table 10.1.2.4-2: Input and Output Streams from Renewable Diesel Plant
Vegetable Oil input
100 gal
Hydrogen
4760 SCF
Renewable diesel output (main product)
93.5 gal
Naphtha output (co-product)
5 gal
Light fuel gas output (co-product)
9 gal
We derived a cost of 6.9 cents/gallon of renewable diesel product to cover other costs:
utilities, labor, and other operating costs.884 Finally, the total cost per gallon was estimated at
$7.61. Table 10.1.2.4-3 provides more details for the process assumed in this analysis and
summarizes the total and per-gallon costs for the year 2023.
883 A New Development in Renewable Fuels: Green Diesel, AM-07-10 Annual Meeting NPRA, March 18-20, 2007.
884 Estimated based on the utility cost for an FCC naphtha hydrotreater; Control of Air Pollution from Motor
Vehicles: Tier 3 Motor Vehicle Emission and Fuel Standards Final Rule; Regulatory Impact Analysis; US
Environmental Protection Agency; March 2014.
403
-------
Table 10.1.2.4-3: Renewable Diesel Production Cost Estimate for a Greenfield 220 Million
Gallons/Yr Plant Processing Soy Oil in 2023 (2C
22 Dollars)
Stream
Estimated value
MM$/yr
$/gal
Soy Oil input
235 MMgals/yr
82c/lb
1483
6.74
Naphtha output
11.8 MMgals/yr
1.81 c/gal
(21.3)
(0.10)
Light fuel gas output
21.2 MMgals/yr
92 c/gal
(19.5)
(0.09)
Hydrogen input
4760 scf/100
gals
$3.73/thousand
standard cubic feet
41.8
0.19
Other Operating Costs
15.2
0.07
Capital Costs (2022 dollars)
$1052 million
115.7
0.53
Fixed Costs
5.5%
57.8
0.26
Total Costs
1673
7.61
A number of announced renewable diesel projects projected to start-up in 2023 through
2026 are conversions of petroleum refineries to produce renewable diesel fuel. The existing
hydrotreating units, fired heaters, heat exchangers, control and instrumentation equipment,
hydrogen plants and tank storage at these refineries is expected to be repurposed for the storage
of feedstocks and the production and storage of renewable diesel. There will likely still need to
be some additional engineering and construction costs to adapt the existing refinery equipment to
produce renewable diesel fuel. Adapting a hydrotreater to process vegetable oil requires
modifications for higher heats of reaction, increased depressurization and perhaps some changes
in metallurgy.885 These modifications are estimated to cost about one third the cost of a new
renewable diesel hydrotreater, or $315 million, instead of $768 million for a 220 million gallon
per year plant. The lower capital cost is due to the avoidance of many investments needed in a
greenfield plant, including the hydrotreater itself, the hydrogen plant, a heater and cooling,
tankage electrical switchgear, buildings, roads, fencing etc. 886>887
It is very challenging to accurately estimate the portion of the future renewable diesel
production which will be produced by these converted refineries as opposed to new greenfield
plants because of the large number of announced renewable diesel projects and the significant
uncertainty of which of these projects will move forward. Because these converted refineries will
require much less capital investment prior to producing renewable diesel fuel, these refinery
conversion projects are more likely to move forward than greenfield projects. Despite the
relatively large capital cost savings associated with the refinery conversion, the impact on the
overall cost to produce renewable diesel fuel is nevertheless modest because most of the cost of
producing renewable diesel fuel is the feedstock cost. For example, renewable diesel produced
from soybean oil by a converted petroleum refinery is estimated to cost $7.24/gallon versus
$7.61/gallon for a greenfield renewable diesel plant. Table 10.1.2.4-4 summarizes the estimated
cost information for a refinery converted to produce renewable diesel fuel.
885 Chan, Erin; Converting a petroleum diesel refinery for renewable diesel fuel; Hydrocarbon Processing; April
2021.
886 Chan, E., Converting a petroleum diesel refinery for renewable diesel; Special Focus: Clean Fuels; April 2021.
887 Lane, Robert; Renewable Diesel Interest Accelerates; August 26, 2020.
404
-------
Table 10.1.2.4-4: Renewable Diesel Production Cost Estimate for a Refinery Converted to
Produce 220 Million Gallons/Yr Plant Processing Soy Oil in 2023 (2021 Dollars
Stream
Estimated value
MM$/yr
$/gal
Soy Oil input
235 MMgals/yr
67.0c/lb
1483
6.74
Naphtha output
4.0 MMgals/yr
1.45 c/gal
(21.3)
(0.10)
Light fuel gas output
7.2 MMgals/yr
73 c/gal
(19.5)
(0.09)
Hydrogen input
4760 scf/100
gals
$2.80/thousand
standard cubic feet
41.8
0.19
Other Operating Costs
15.2
0.07
Capital Costs (2022 dollars)
$315 million
34.7
0.16
Fixed Costs
5.5%
57.8
0.26
Total Costs
1592
7.24
The difference between the low and high production costs is solely due to the difference
in capital costs and associated fixed costs. For refineries converting their refineries to produce
renewable diesel, the amortized capital cost is estimated to be only $0.16 per gallon, while the
greenfield plant's estimated capital cost is $0.53 per gallon. As a very rough estimate, half of the
future domestic renewable fuel production is estimated to be produced by these converted
refineries, and when the refinery conversions are averaged with the greenfield plants, this results
in the $7.42 estimated production cost for 2023. The estimated renewable diesel production cost
for all vegetable oil types and for all three years is summarized in Table 10.1.2.4-5.
Table 1C
1.1.2.4-5 Summary of Estimatet
Year
Soy Oil
Corn Oil
FOG
2023
7.42
5.66
6.91
2024
7.15
5.02
5.68
2025
7.23
4.50
5.47
Renewable Diesel Production Costs ($/gal)
10.1.2.5 Biogas
Biogas is the result of anaerobic digestion of organic matter, including municipal waste,
manure, agricultural waste, and food waste.888 The primary product of this anaerobic digestion of
waste is methane, which is the primary component of natural gas. Thus, once biogas is cleaned
up by removing various contaminants, it can be used by processes that normally use natural
gas.
889
The largest source of biogas, which is already being collected to avoid releasing methane
into the environment, is from landfills.890 Since landfill gas is the largest source of biogas
available for the motor vehicle fleet, this cost analysis makes the simplifying assumption that the
biogas will solely be provided by landfills.
888 Wikipedia. https://en.wikipedia.org/wi.ki/Biogas.
889 LeFevers, Daniel; Landfill Gas to Renewable Energy; Hill Briefing; April 26, 2013.
890 Biomass Explained, Landfill Gas and Biogas; US Energy Information Administration; February 1, 2019;
www.eia.gov/energyexplained/index.php?page=bioimass biogas
405
-------
While in some cases biogas can be used in local fleet vehicles which are operated at the
landfill site, in most cases, a new pipeline would need to be constructed to transport the cleaned
up biogas to a nearby common carrier pipeline. Gas is then pulled off the pipeline at downstream
locations and compressed into CNG or liquified into LNG for use in motor vehicles. Tracking
the use of the biogas in motor vehicles occurs by proxy through contracts and/or affidavits rather
than through a system designed to ensure that the same methane molecules produced at the
landfill are used in CNG/LNG vehicles.
One of the costliest aspects of using biogas is its cleanup. Biogas contains large amounts
of carbon dioxide, nitrogen, and other contaminants such as siloxanes which cannot be tolerated
if it is to be put into a natural gas pipeline or used by fleet vehicles at the landfill site. We
estimated a cost for cleaning up landfill biogas using Version 3.5 of the Landfill Gas Energy
Cost Model (LFGcost-Web).891'892 The throughput volume of landfill gas was estimated to be
600 standard cubic feet per minute based on a survey of biogas production facilities.893 The cost
estimates from the Model excluded the gas collection and control system infrastructure at the
landfill, as EPA expects that landfills that begin producing high BTU gas in 2021 are very likely
to already have this infrastructure in place, and that this infrastructure would be used regardless
to control methane emissions. The equations from the LFGcost-Web model for biogas collection
and clean-up are summarized in Table 10.1.2.5.1-1. We included a cost for biogas collection at
the landfill which amounts to $1.1 per thousand standard cubic feet.894 Distribution and retail
costs are estimated for biogas in Chapter 10.1.4.3.
Table 10.1.2.5-1: Biogas Cleanup Costs (600 scf/min)
Cost Factors (2019$)
million
dollars
(2022$)
$ per thousand
cubic feet
(2022$)
Interconnection
$400,000
0.54
0.19
Capital Costs
6,000000*e(000°3*ft3/min)
9.65
3.36
Operating and
Maintenance
250 * ft3/min +148,000
0.90/yr
2.85
Electricity Costs
0.009 kWh/ft3
0.33/yr
1.05
Total
7.46
10.1.2.6 Sugar Cane Ethanol
Unlike the starch in corn kernels which first must be depolymerized using enzymes,
sugarcane contains free sugar which, after extraction from the sugarcane, can be directly
fermented into ethanol. The fibrous portions of the sugarcane plant is typically combusted to
produce the energy needed for the process.
891 The current version of this model and user's manual dated March 2021 are downloadable from the LMOP
website: https://www.epa.gov/lmop.
892 This cost estimate does not include the cost for complying with California's more stringent natural gas pipeline
specifications designed to address harmful contaminants in some sources of biogas.
893 Economic Analysis of the US Renewable Natural Gas Industry; The Coalition of Renewable Natural Gas;
December 2021.
894 LFG Energy Project Development Handbook - Project Economics and Financing; Chapter 4.
406
-------
We estimated the cost to produce sugar cane ethanol two different ways. The first way is
based on a simple bottom-up cost estimate using some production cost information. The second
is based on price data for sugarcane ethanol received from Brazil available in the EPA
Moderated Transaction System (EMTS). Generally, ethanol from sugarcane produced in tropical
areas is cheaper to produce than ethanol from cellulose, and is similar to the cost of corn starch
ethanol. This is due to favorable growing conditions, relatively low-cost feedstock and energy
inputs, and other cost reductions gained from years of experience.
A study by OECD (2008) entitled "Biofuels: Linking Support to Performance," provides
a set of assumptions and an estimate of production costs. Our estimate of sugarcane production
costs, which is shown in Table 9.1.2.3-1, primarily relies on the analysis made for that study. The
original cost estimate reported in the RFS2 rulemaking assumes an ethanol-dedicated mill and is
based off an internal rate of return of 12%, a debt/equity ratio of 50% with an 8% interest rate
and a selling of surplus power at $57 per MWh. We revised the capital and operating costs
higher by 63% to account for the effects of inflation from 2006 to 2022. When we estimated the
amortized, per-gallon capital costs we also added a 20% capital cost contingency factor to
account for other costs not accounted for in the cost analysis and amortized the capital costs
using our capital cost amortization parameters. Table 10.1.2.6-1 provides the updated production
cost estimate for sugarcane ethanol.
Table 10.1.2.6-1: Sugar Cane Ethanol Production Cost
Cost Basis
Sugarcane Productivity
71.5 tons/hectare
Sugarcane Consumption
2 million tons/year
Harvesting days
167
Ethanol productivity
85 liters/ton feedstock (22.5 gal/ton
feedstock)
Ethanol Production
170 million liters/year (45 million gals/yr)
Surplus power produced
40 kWh/ton sugarcane
RFS2 Reported Cost
($2006)
Revised Costs
($2022)
Capital Costs
($ million)
Investment cost in
million$
97
158
Investment cost for
sugarcane production
36
59
Per Gallon Costs
($/gal)
O & M (Operating &
Maintenance) costs
0.26
0.42
Variable sugarcane
production costs
0.64
1.05
Capital costs
0.49
0.64
Total production costs
1.40
2.11
Shipping Costs to US
0.15
Delivered Cost
2.26
We also reviewed recent data on sugar cane ethanol prices which we receive in the EPA
Moderated Transaction System (EMTS). These are as-received prices, so they include the cost to
407
-------
ship the ethanol from Brazil to the US. The average of recent sugar cane ethanol prices from
EMTS was $2.73 per gallon. Other price data which EPA receives from OPIS showed a similar
average price which helps to corroborate the price data from EMTS.
The average FOB ethanol price of $2.73/gallon in Brazil is somewhat higher than the
estimated sugar cane ethanol production cost of $2.26/gallon. This cost/price difference can
mostly be attributed to the low (0.11) before-tax capital amortization factor that we use which
reflects the social cost of capital, and the shipping costs incorporated in the price data. When we
use a more typical 0.16 after-tax capital amortization factor used by industry, the per-gallon costs
increase to $2.55 per gallon. Normally we would use the bottom-up cost estimate, however, the
EMTS price data may capture some inflation effects which the bottom-up cost estimate may not
capture regardless of the applied inflation adjustment. For this reason we used the $2.73 per
gallon price data from EMTS to represent the production and distribution costs for sugar cane
ethanol.
10.1.2.7 Corn Kernel Fiber Ethanol
In addition to converting corn starch to ethanol, some of the fiber contained in the dried
distiller grains (DDGS) can also be converted to ethanol. This additional ethanol from corn fiber
is considered cellulosic ethanol and earns D3 RINs. Historically, this cellulosic conversion step
of the fiber to ethanol was thought of as a separate step than the starch to ethanol conversion, and
therefore would require a separate reactor vessel and require additional operating costs.
However, one or more companies have found it possible to convert some of the cellulosic fiber
to ethanol in the existing starch to ethanol facility, and we project that this single reactor design
would likely be producing cellulosic ethanol during the timeframe of this rulemaking.895 896
Anticipating that this cellulosic ethanol would be produced in an existing starch to ethanol
reactor provides a cost efficiency which would lower the overall production cost. But this also
presents a challenge for how to identify the quantity of ethanol produced from cellulose versus
that produced from the starch. To remedy this EPA published guidance on how to identify the
portion of ethanol being produced from cellulose.897
Anticipating that the cellulosic ethanol will be produced along with corn starch in an
existing reactor allows us to estimate the cost of producing this cellulosic ethanol. Since we
already estimate the capital, fixed and variable operating cost of producing ethanol from corn
starch, we simply apply those same cost estimates to the corn fiber ethanol. There are other cost
factors to consider, which is the potential cost for the additional enzyme added to convert corn
fiber to ethanol, and a cost savings due to increased corn oil production.898 It appears that the
cost of the additional enzyme is approximately equally offset by the cost savings of additional
895 Kacmar, Jim; Intellulose: An Innovative Approach to your Plant's Profitability; Edeniq presentation to the 2019
Distillers Grains Symposium; May 15, 2019.
896 Conversion of Corn-Kernel Fiber in Conventional Fuel-Ethanol Plants; National Corn to Ethanol Research
Center; Project No. 0340-19-03; November 11, 2018.
897 Guidance on Quantifying an Analytical Method for Determining the Cellulosic Converted Fraction of Corn
Kernel Fiber Co-Processed with Starch; EPA-420-B-22-041; September 2022.
898 Kacmar, Jim; Intellulose: An Innovative Approach to your Plant's Profitability; Edeniq presentation to the 2019
Distillers Grains Symposium; May 15, 2019
408
-------
corn oil production. Therefore, we simply use the cost for producing ethanol from corn starch for
the cost of producing ethanol from cellulosic ethanol.
It is possible that the DDGS remaining after extracting the corn fiber would have a higher
market value than regular DDGS on a mass basis of DDGS due to the higher protein content.899
However, as a conservative assumption, we do not assume any price increase for residual DDGS
due to its higher protein content.
10.1.3 Blending and Fuel Economy Cost
Certain renewable fuels, namely gasoline, biodiesel, and renewable diesel, are typically
blended into petroleum fuels. There are costs and in some cases cost savings associated with
such blending. In addition, these renewable fuels have relatively lower energy per gallon leading
to lower fuel economy (miles driven per gallon). In this section, we consider blending and fuel
economy costs for ethanol blended as E10, E15, and E85, as well as for biodiesel and renewable
diesel.
10.1.3.1 Ethanol
10.1.3.1.1 E10
Ethanol has physical properties when blended into gasoline which affect its value as a
fuel or fuel additive. Ethanol has a very high octane content, a high blending Reid Vapor
Pressure (RVP) when blended into gasoline at low concentrations, and is low in energy content
relative to the gasoline pool that it is blended into. Ethanol has essentially zero sulfur or benzene,
adding to ethanol's value because refineries must meet sulfur and benzene fuel regulations. Each
of these properties can have a different cost impact depending on the gasoline it is being blended
into (reformulated gasoline (RFG) versus conventional gasoline (CG), winter versus summer
gasoline, premium versus regular, and blended at 10% versus E15 or E85). These physical
properties are also valued differently from a refiner's perspective compared to that of the
consumer. Refiners value ethanol's octane because they can lower the octane of the gasoline the
ethanol is being blended into, reducing their refining costs. Refiners dislike ethanol's high
blending RVP when blending ethanol in gasoline (usually RFG) at 10% because they must
remove some low cost gasoline blendstock material (usually butane) to accommodate the ethanol
if the gasoline they are producing does not receive a 1 psi RVP waiver. However, refiners are not
concerned about ethanol's low energy content when blending it into gasoline since they sell
gasoline on volume, not energy content, and consumers do not appear to demand a discount for
E10. Rather, this is usually just an issue for the consumers who do not travel as far on a gallon of
fuel with lower energy content. Depending on the fuel they are purchasing, the lower energy
content will be either obvious to consumers (i.e., E85), impacting their purchase decisions, or not
(i.e., E10; most consumers do not notice its lower energy content in comparison to E0,
particularly now that almost all gasoline is E10). Since this is a social cost analysis which
899 DDGS from which the corn kernel fiber has been extracted have higher protein levels, and can be used in a wider
variety of feed markets (e.g., can be fed to poultry and swine in addition to cattle). Therefore, some marketers
suggest that this type of DDGS could have a higher value than regular DDGS. Jeremy Javers (ICM), "By-Products
to Co-Products to Products," 2017.
409
-------
incorporates all the costs to society, the fuel economy effect is included in the overall cost
estimates, although not included with the blending value estimated in this section.
Ethanol's total blending value is estimated based on the output from refinery modeling
cases conducted by ICF/Mathpro for a projected 2020 year case assuming that crude oil would
be priced at $72/bbl.900 By averaging the costs separately for conventional and reformulated
gasolines, the refinery modeling output from the first case allowed us to estimate ethanol's
volatility cost for blending ethanol into E10 reformulated gasoline.901 Due to the options
available to refiners to replace ethanol's octane, ICF/Mathpro ran two ethanol replacement cases.
In the lower per-gallon cost case, the refinery model principally relied on increased alkylate
production. But to be able to replace all of ethanol's octane, the refinery model estimates that
refiners would also increase the octane of reformate (through increased reformer severity) and
increase production of isomerate, even if the primary octane replacement is alkylate. The
refinery model estimates that for this alkylate-centric case over 7.6 million barrels per day of
new refinery unit capacity would need to be added by refiners.
ICF/Mathpro modeled a second case. Instead of relying on large butane purchases for
producing alkylate, the model increased the throughput to, and turned up the severity of, existing
reforming units to increase the octane of reformate, the product stream of the reformer. This case
still relied on other octane producing unit additions, including alkylate and isomerate, but
increased reformate volume and octane was the principal method. This second reformate-centric
refinery modeling case was less capital-intensive, but still added 3.7 million barrels per day of
additional refinery unit capacity and was more costly on a per-gallon basis. Increasing the
severity of reformers is relatively more expensive because of the cost associated with the
production of two by-products of the reforming process which increase as the severity of the
reformer is increased. Hydrogen is a by-product of reforming, but reformer-produced hydrogen
is much more expensive than hydrogen produced from natural gas because natural gas has been
priced much lower than crude oil. Fuel gas is another reformer by-product which is usually used
for refinery process heat, but displaces much cheaper purchased natural gas. For short-term
octane needs refiners would likely need to rely on increasing reformate severity to avoid or
minimize the amount of new refining unit capacity additions, but given the higher cost overall
cost, this would not be a preferable long-term solution.
Table 10.1.3.1.1-1 summarizes gasoline's marginal costs for the reference case, and
ethanol's marginal costs for two ethanol removal cases, for different gasoline types and refinery
regions. For the two ethanol removal cases the refinery modeling for both the reference case (all
gasoline with ethanol) and the low biofuel cases (conventional gasoline without ethanol), which
replaced ethanol in the gasoline pool with refinery sourced alternatives, low biofuel #1 is the
reformate-centric case while biofuel #2 is the alkylate-centric case. The lower marginal values
for PADD 1 can be explained because Mathpro forced PADD 3 refineries to satisfy PADD l's
need for replacing ethanol's volume and octane through PADD's 3 exports into the PADD 1
900 The crude oil price has a first order effect on the blending value and volatility cost for blending ethanol into
gasoline. Since the crude oil price used in the refinery modeling cost analysis is about the same as the projected
crude oil price for 2021 and 2022, it was not necessary to adjust ethanol's estimated blending cost to any other dollar
value.
901 EPA's contract was with ICF Incorporated, LLC, which in turn retained Mathpro for some aspects of the work.
410
-------
after initial refinery model runs showed PADD l's marginal costs for replacing ethanol were
exceedingly high.
Table 10.1.3.1.1-1: Gasoline Marginal Values for Reference Case and Ethanol Marginal
Values for the No-Biofuel Cases ($/bbl)
PADD
Gasoline Marginal
of
Values
Low Biofuel #1
Low Biofuel #2
Gasoline
Type
Grade
Summer
Winter
Summer
Winter
Summer
Winter
Origin
PADD 1
RFG
Prem
95.74
83.94
108.37
100.88
Reg
91.45
81.35
115.98
105.97
CG
Prem
92.68
83.89
123.02
100.87
Reg
88.93
81.35
136.43
105.88
PADD 2
RFG
Prem
88.09
81.68
132.42
110.28
113.45
96.62
Reg
84.80
79.77
145.38
116.02
122.86
101.61
CG
Prem
85.55
81.25
149.08
110.41
126.74
96.25
Reg
82.46
79.45
161.21
115.79
135.55
100.93
PADD 3
RFG
Prem
85.42
78.31
121.69
94.72
118.51
89.77
Reg
81.86
76.39
134.67
98.45
131.29
94.48
CG
Prem
83.64
78.78
133.95
95.13
129.37
89.91
Reg
79.97
76.76
146.78
98.46
142.00
94.55
PADD 4
CG
Prem
79.8
77.0
135.5
115.2
150.1
103.1
Reg
77.4
75.1
149.0
124.0
168.1
110.0
Low
Prem
94.5
136.5
151.2
RVP
Reg
98.3
150.1
169.2
PADD 5
RFG
Prem
96.89
83.68
37.68
96.05
Reg
91.61
82.01
62.46
97.37
CG
Prem
77.63
83.00
118.14
98.01
Reg
73.38
81.12
126.14
97.68
The gasoline-ethanol difference in marginal values is calculated and summarized on a
cents per gallon basis in Table 10.1.3.1.1-2.
411
-------
Table 10.1.3.1.1-2: Marginal Ethanol Replacement Cost by Gasoline Type and Season
(cents/gallon)
PADD
of
Low Biofuel #1
Low Biofuel #2
Gasoline
Reformate-centric
Alkylate-centric
Origin
Type
Grade
Summer
Winter
Summer
Winter
PADD 1
RFG
Prem
30.07
40.35
0
0
Reg
58.41
58.62
0
0
CG
Prem
72.23
40.43
0
0
Reg
113.10
58.42
0
0
PADD 2
RFG
Prem
105.56
68.08
60.39
35.55
Reg
144.23
86.31
90.61
52.00
CG
Prem
151.27
69.44
98.08
35.73
Reg
187.51
86.52
126.41
51.15
PADD 3
RFG
Prem
86.35
39.08
78.77
27.29
Reg
125.74
52.52
117.69
43.07
CG
Prem
119.78
38.93
108.86
26.50
Reg
159.07
51.68
147.69
42.38
PADD 4
CG
Prem
132.70
90.86
167.45
62.16
Reg
170.67
116.41
216.07
83.19
Low
Prem
100.19
0.00
135.02
0.00
RVP
Reg
123.27
0.00
168.77
0.00
PADD 5
RFG
Prem
-140.97
29.46
0
0
Reg
-69.39
36.56
0
0
CG
Prem
96.44
35.73
0
0
Reg
125.61
39.43
0
0
The regional ethanol replacement costs are volume-weighted together to develop
national-average ethanol replacement costs by gasoline grade and season. These costs are only
presented for the conventional gasoline pool since the ethanol was only replaced in the
conventional portion of the gasoline pool in the study. Table 10.1.3.1.1-3 summarizes these
estimated ethanol-replacement costs.
Table 10.1.3.1.1-3: National Average Ethanol Replacement Cost by Gasoline Grade and
Season (c/gal)
Gasoline
Grades
Reformate-centric
Alkylate-centric
Summer
Winter
Summer
Winter
Conventional
Premium
124.6
50.8
112.0
32.7
Regular
165.1
66.8
144.2
48.2
To estimate the volatility cost, ethanol's marginal values in Table 10.1.3.1.1-1 forRFG
are subtracted from those for CG, although the values are calculated separately for premium and
regular grade gasolines. These calculated values are summarized in Table 10.1.3.1.1-4. Although
this analysis could have separately analyzed RVP-controlled conventional gasoline without a
waiver, it did not since its gasoline volume was less than 2% of the total gasoline pool.
412
-------
Table 10.1.3.1.1-4 Ethanol's RVP Blending Cost in Reformulated Gasoline in 2020 by
PADD ($/gal)"
Gasoline PADD
Gasoline Grade
RFG-CG Marginal
Values ($/bbl)
PADD 1
Premium
9.74
Regular
10.53
PADD 2
Premium
9.59
Regular
9.32
PADD 3
Premium
9.31
Regular
9.25
PADD 5 (CA)
Premium
58.79
Regular
62.59
The ethanol RVP blending cost estimated by the refinery model are volume-weighted
together to develop national-average values, and ethanol's RVP blending costs are calculated
separately for premium and regular grades of summertime RFG, and summarized in Table
10.1.3.1.1-5. The PADD 5 RFG, which is California RFG, is modeled to have a volatility cost
which is five time higher than other RFG areas. The cost of complying with California RFG
standards may be higher than that for other RFG areas, but a factor of five seemed much too high
and was considered an outlier.902 Therefore, the modeled California RFG ethanol marginal costs,
which should reflect ethanol's volatility cost, were omitted from this analysis and the PADD 1 -
3 costs were volume-weighted together and used for all RFG areas, including California.
Table 10.1.3.1.1-5: Calculated RVP Blending Costs by Fuel Grade
Gasoline Grade
c/gal
Nationwide
Aggregated Cost
Premium
22.5
Regular
22.8
Although the ethanol replacement cost was based on a refinery modeling case when
ethanol was solely removed from conventional gasoline, it would likely be about the same for
reformulated gasoline (RFG) as well, so we assumed that they were the same for RFG.903
However, it is necessary to add in ethanol's volatility cost for RFG, which for ethanol's removal
would be a cost savings. The 23 cent per gallon volatility cost for regular and premium gasoline,
respectively, is subtracted from ethanol's replacement cost to estimate the ethanol replacement
cost for RFG. The ethanol replacement costs for both CG and RFG are shown in Table
10.1.3.1.1-6. The ethanol replacement costs are then further aggregated to national, year-round
averages for each octane replacement scenario and summarized at the bottom of the table.
902 California's relies on ethanol blended at 10 volume percent for compliance with its Low Carbon Fuel Standard,
thus, removing E10 ethanol from California gasoline is an unlikely possibility.
903 Both RFG and CG must meet many of the same gasoline property specifications, including sulfur and benzene, as
well as ASTM standards (ASTM D4814).
413
-------
Table 10.1.3.1.1-6: Aggregated Ethanol Marginal Replacement Cost (cents/gallon)
Low Biofuel #1
Low Biofuel #2
Reformate-centric
Alkylate
-centric
Summer
Winter
Summer
Winter
Prem
124.58
50.79
112.04
32.65
Conv.
Reg
165.11
66.83
144.23
48.19
Prem
105.58
50.79
93.04
32.65
RFG
Reg
144.11
66.83
123.23
48.19
82.23
68.65
Refiners would pursue the lowest cost means to produce their fuels. Therefore, for
evaluating the cost of using ethanol in gasoline at 10 volume percent, the lower cost, alkylate-
centric cost of 68.65 cents per gallon was used for ethanol's blending cost for ethanol blended as
E10. This 68.65 cents per gallon cost represents ethanol's average nationwide blending
replacement cost in U.S. gasoline. This can be thought of as the additional value or cost savings
to gasoline refiners per gallon of ethanol that results from blending 10% ethanol into gasoline
today.
10.1.3.1.2 Higher Level Ethanol Blends
While there is a considerable blending cost savings associated with blending ethanol as
E10, there currently is not a savings for blending ethanol as E15 or E85. The blending costs for
higher level ethanol blends is considerably different from that for E10 in large part due to the
inability in most instances to take advantage of the octane benefit associated with the additional
ethanol. Furthermore, the congressional 1 psi RVP waiver which applies for blending E10
gasoline in summer conventional gasoline does not apply to blending El5, requiring a lower
RVP and therefore higher cost gasoline blendstock. However, this is only an added cost in the
summer and only in conventional gasoline areas.
There have been, and there continue to be, steps taken to facilitate the blending of El 5
into summertime conventional gasoline. EPA granted El 5 a 1 psi waiver that took effect in the
summer of 2019; however, this waiver was struck down by a federal court in 2021. For summer
2022 and the beginning of summer 2023, EPA granted numerous emergency waivers to allow
El 5 to continue to be sold with a 1 psi RVP waiver. A number of states petitioned EPA to allow
them to remove the 1 psi waiver for blending E10 gasoline, and if the 1 psi waiver for El 0 were
to be removed, the same lower RVP, higher cost gasoline blendstock would be required for both
ethanol blends in summertime conventional gasoline in those states. EPA proposed to remove the
E10 1 psi waiver for those states in 2024.904
El5 could potentially realize a blending cost benefit based on the increased octane for the
additional ethanol if refiners could create and distribute a low RVP, low octane El 5 blendstock
for oxygenate blending (BOB). However, this would require a widespread shift by refineries,
pipelines, and terminals in an entire geographical region to produce and distribute another even
904 88 FR 13758 (March 6, 2023).
414
-------
lower octane BOB specially designed for producing E15 instead of ElO.905 This would most
likely only occur if El 5 becomes the predominant gasoline used in that region because of the
limitations of the distribution system and experience with the historic conversion to ElO. Since
this could not feasibly happen during the time period of this rulemaking, we have not included
any octane blending benefit for the additional ethanol blended into El 5 in excess of the ethanol
blended in ElO (the additional 5%).906 Thus, the gasoline BOB used to produce E15 in the winter
months is the same as that used for producing ElO, resulting in a higher octane fuel than what it
can be priced at. In the summer months, El 5 would also incur the additional RVP control costs.
There also is not a blending cost benefit for ethanol blended as E85 resulting from its
high octane beyond that which is already being realized when blending ElO. When producing
E85, ethanol's high octane results in significant overcompliance with the minimum octane
standard. Refiners do not produce a low octane BOB for producing E85 to realize a cost savings.
Conversely, ethanol plants produce E85 by adding a denaturant to ethanol, which typically is a
low cost, low octane, high RVP, hydrocarbon commonly called natural gas liquids (NGL). The
corn ethanol plants add an additional quantity of the NGL, above the quantity needed to denature
the ethanol, to produce the E85. Thus, E85 produced from NGLs does realize a cost savings. But
NGLs are also lower in energy density, offsetting the potential cost savings to consumers.
Regardless, there is no RVP blending cost for E85 because the high portion of ethanol results in
lower RVP instead of higher RVP; therefore, a lower RVP blendstock is not needed for
producing E85. In fact, to adjust for the lower RVP of E85 blends, E85 is actually blended at
roughly 74% ethanol on average over both the summer and winter, instead of 85%, to have
sufficiently high RVP to avoid RVP minimum limits.907
The societal cost of using ethanol must include ethanol's lower energy density (fuel
economy effect). Ethanol has about 33% lower energy density than gasoline blendstock (CBOB
and RBOB).908 Accounting for ethanol's lower energy density adds about $1 per gallon of
ethanol for all the ethanol blends to account for the additional cost to consumers for having to
purchase a greater volume of less energy dense fuel to travel the same distance.
905 Some refiners may have extra tankage available to allow producing and storing a lower octane, E15 blendstock to
enable selling E15 over its own terminal rack to local retail stations. Refinery rack gasoline sales, however, are
usually a small portion of the refinery's gasoline sales.
906 The reformulated gasoline pool always took advantage of ethanol's high octane as it was needed to cause a
reduction in aromatics to reduce the emissions of air toxics under the Complex Model - the compliance tool of the
RFG program. So when ethanol replaced methyl tertiary butyl ether (MTBE) as the oxygenate in 2005 when the
RFG oxygen requirement was rescinded, refiners took advantage of ethanol's high octane content. The CG pool,
however, could not take advantage of ethanol's high octane until an entire U.S. gasoline market (i.e., Midwest) was
blended with ethanol, and then that gasoline market shifted over all at once to a suboctane blendstock for oxygenate
blending (CBOB). Reviewing CG aromatics levels (high octane aromatics decrease when refiners produce
suboctane CBOB), refiners switched the CG pool over to low octane CBOB over the years from 2008 to 2013 which
is around the time when the U.S. reached the ElO blendwall.
907 E85 can have RVP levels which are too low which makes starting a parked car difficult. When blended at about
70% ethanol, the RVP of the ethanol-gasoline blend is a little higher than E85 blends improving cold starts.
908 Frequently Asked Questions: How much ethanol is in gasoline, and how does it affect fuel economy?; Energy
Information Administration; fattps://www.eia.gov/toots/faqs/:faq.pfap?id=27&t=.1.0.
415
-------
10.1.3.2 Biodiesel and Renewable Diesel
Biodiesel and renewable diesel fuel have properties that could cause a cost savings or
incur a cost. Both fuels have higher cetane value relative to petroleum diesel.909'910 Although
ICF/Mathpro considered the possibility of the petroleum refining industry taking advantage of
that property, they concluded that most markets are not cetane limited and that as a result refiners
likely would not take advantage of this property of biodiesel and renewable diesel.911 At this
time, we do not have any evidence that refiners are capitalizing on biodiesel and renewable
diesel's higher cetane value.
Conversely, a blending cost could be incurred for biodiesel due to the addition of
additives to prevent oxidation and lower pour or cloud point. The need to add pour point
additives is primarily a cold weather issue and likely contributes to the lower observed blending
rates of biodiesel into diesel fuel in the winter compared to the summer, particularly in northern
areas. However, for our analysis, no additive costs were included for biodiesel because we do not
have a good estimate for them.
As with ethanol, the societal cost of using biodiesel and renewable diesel must include
their lower energy density in comparison to petroleum-based diesel fuel, which impacts fuel
economy. Accounting for this fuel economy effect adds about 27 and 17 cents per gallon to the
societal cost of biodiesel and renewable diesel, respectively.
10.1.4 Distribution and Retail Costs
In this section, we evaluate the costs of distributing biofuels from the places where they
are produced to retail stations as well as the costs of dispensing these fuels at those retail stations.
10.1.4.1 Ethanol
10.1.4.1.1 Distribution Costs
Distribution costs are the freight costs to distribute the ethanol, although the total
distribution costs could also include the amortized capital costs of newly or recently installed
distribution infrastructure. A significant amount of capital has already been invested to enable
ethanol to be blended nationwide as E10, and a small amount of ethanol as E85 and E15.
Virtually all terminals, including those co-located with refineries, standalone product distribution
terminals, and port terminals, have made investments over the last 15-plus years to enable the
distribution and blending of ethanol. Thus, these capital costs are considered sunk and no
additional capital cost is explicitly included in our analysis. However, in the part of the analysis
where we estimate ethanol's distribution costs using spot ethanol prices, as described below, we
may inherently be including some distribution capital costs which are still being recovered.
909 Animal Fats for Biodiesel Production; Farm Energy; January 31, 2014.
910 McCormick, Robert; Renewable Diesel Fuel; NREL; July 18, 2016.
911 Modeling a No-RFS Case; ICF Incorporated; Work Assignment 0,1-11, EPA contract EP-C-16-020; July 17,
2018.
416
-------
As part of the effort by ICF/Mathpro to estimate use of renewable fuels in the absence of
the RFS program, ICF estimated distribution costs for ethanol and biodiesel. We used these cost
estimates for this rulemaking.912 ICF estimated ethanol's distribution costs based on ethanol spot
prices that are available from the marketplace. The spot prices likely represent the operating and
maintenance costs, and any capital costs which are being recovered. Certain publications,
including OPIS and ARGUS, publish ethanol spot prices for certain cities and these spot prices
were consulted for estimating ethanol's distribution costs. These spot prices are tracked because
they represent unit train origination and receiving locations where the custody of the ethanol
changes hands in the distribution system. Since nearly all the ethanol is being produced in the
Midwest, the ICF distribution cost analysis assumed that the ethanol is collected together in
Chicago by truck or manifest rail at an average cost of 7 cents per gallon and then moved out of
the Midwest to other areas mostly using unit trains. For the ethanol consumed in the Midwest,
the ethanol is likely to be moved by trucks directly to the terminals in the Midwest. For the areas
adjacent to the Midwest, the ethanol is assumed to be moved by truck for the areas nearest to the
Midwest (i.e., Colorado and Wyoming), and by manifest train for the adjacent areas further out
(i.e., Utah and Idaho). These various means for distributing ethanol, and their associated costs,
were accounted for when estimating the ethanol's distribution cost to and within each region.
Once the ethanol is moved to a unit train or manifest train receiving terminal, there are
many other terminals in these areas which must also receive the ethanol. Ethanol must then be
moved either by truck or, if further away, by manifest train from the unit train receiving
terminals to the other terminals. Since many of these other terminals do not have sidings for rail
car offloading, the manifest train ethanol must be offloaded to trucks at tank car-truck transfer
locations before it can be received by these other terminals. A simple analysis revealed that each
unit train receiving terminal must then service, on average, an area of 31 thousand square miles
(equivalent to a 180 x 180 miles) to make the ethanol available to the various terminals in the
area. ICF estimated that, on average, the further distribution of ethanol from these unit train
receiving terminals to the rest of the terminals would cost an additional 9 or 11 cents per gallon,
depending on the PADD. Since ICF completed its analysis, we discovered that most corn ethanol
plants are capable of sourcing unit trains from their plants. Thus, the 7 cent per gallon
transportation cost from corn ethanol plants to Chicago is not necessary and this cost was
removed from the estimated cost to each destination.913 Table 10.1.4.1.1-1 provides the estimate
of ethanol distribution costs for the various parts of the country estimated by ICF, and as revised
to remove the 70 per gallon transportation cost.
912 Modeling a No-RFS Case; ICF Incorporated; Work Assignment 0,1-11, EPA contract EP-C-16-020; July 17,
2018.
913 Rail congestion, cold weather raise ethanol spot prices; US Energy Information Administration; April 3, 2014.
417
-------
Table 10.1.4.1.1-1: Ethanol Distribution Costs for Certain Cities or Areas
Distribution Cost (^/gal) to:
Location
Hub/Terminal
Total (Wgal)
PADD
Area
To
Chicago
From
Chicago
Blending
Terminal
ICF Estimate
Revised
Estimate
Florida/Tampa
17.8
11.0
35.8
28.8
S outheast/ Atl anta
11.7
11.0
29.7
22.7
PADD 1
VA/DC/MD
9.7
11.0
27.7
20.7
Pittsburgh
6.2
11.0
24.2
17.2
New York
7.7
11.0
25.7
18.7
PADD 2
Chicago
0.0
11.0
18.0
11.0
Tennessee
7.0
9.7
11.0
27.7
20.7
PADD 3
Dallas
4.5
11.0
22.5
15.5
PADD 4
6.2
11.0
24.2
17.2
Los Angeles
16.4
9.0
32.4
25.4
PADD 5
Arizona
16.4
9.0
32.4
25.4
Nevada
12.4
9.0
28.4
21.4
Northwest
12.4
9.0
28.4
21.4
We volume-weighted the various revised regional distribution cost estimates for PADDs
1 through 5 to derive a PADD-average ethanol distribution cost for all PADDs. Table 10.1.4.1.1-
2 summarizes the estimated average ethanol distribution cost by PADD, and the average for the
U.S adjusted to 2022 dollars.
Table 10.1.4.1.1-2: Average Ethanol Distribution Cost by PADD and the U.S,
Gasoline
Average Ethanol
Volume
Distribution Cost
Region
(kgals/day)
(C/gal)
PADD 1
123,700
22.0
PADD 2
102,400
11.0
PADD 3
68,500
15.5
PADD 4
15,100
17.2
PADD 5
63,400
24.4
U.S. Average
373,100
18.1
U.S. Average 2022$
20.1
10.1.4.1.2 Retail Costs
The infrastructure at retail needed to make E10 available has been in place for many
years. As a result, no additional retail costs are assumed for E10. However, this is not the case
for E15 and E85. Additional investments are needed to make them available at retail. The E15
and E85 volumes that we are using in this costs analysis are summarized in Chapter 6.5.2.
418
-------
The retail costs for El 5 and E85 are estimated based on the investments that are needed
to be made to offer such ethanol blends. To this end, we reviewed literature and conferred with
EPA's Office of Underground Storage Tanks on what might be considered "typical" for E15 and
E85 equipment installations for a typical sized retail station selling these blends.914'915'916'917 For
the typical retail station revamp to sell El5, the station is assumed to have an underground
storage tank already compatible with El5 that it would convert over to store El5, but would still
require 4 new dispensers to dispense the E15. Each dispenser is estimated to cost $20,000 for a
total cost of $80,000 (assuming only 4 dispensers for a retail outlet), and this cost per dispenser
increases to $29,500 when adjusted to 2022 dollars.918 In addition, these retail stations are
assumed to invest in additional equipment changes to make their hardware compatible with El5
(e.g., pipes, pipe connectors, sealants including pipe dope and elastomers, pumps, and hardware
associated with underground storage tanks) at a cost of $15,000. Thus, the total investment for a
typical retail station revamp is $132,900.
The E85 stations are also assumed to have an existing underground storage it could use
for storing E85, but would require some equipment modification to allow the very high ethanol
concentration to be stored in that tank and other equipment. The E85 station would also be
required install a new E85-compatible dispenser, costing $29,500, for a total cost of $40,500
(assuming only one dispenser at a retail outlet is provided for E85).919
Retail stations can incur costs which are higher or lower than the retail revamp costs we
estimate for offering El 5 and E85. If the retail station already has dispensers, tanks and other
equipment that can offer El 5 or E85 fuel, then perhaps only a few thousand dollars would need
to be spent to make some dispenser parts compatible with the higher concentration ethanol. On
the other hand, if the retail station needs the new dispensers and also needs to install a separate
storage tank and other equipment to store and dispense El 5 or E85, then the installation costs
would be much higher. The retail revamp costs to offer higher ethanol blends estimated here
attempts to find representative costs for this large cost range.
To estimate the per-gallon cost, it is necessary to estimate the volume of E85 and E15
sold at each station which offers these blends. These per-station volume estimates were based on
data collected by USD A through their BIP program and made available to EPA.920 The total
volumes of E15 and E85 sold were divided by the estimated number of E15 and E85 retail
stations to estimate the volume per retail station. As a result, retail stations offering El 5 are
estimated to sell 181 thousand gallons of E15 per year while retail stations offering E85 are
estimated to sell 39 thousand gallons of E85 per year. Using the amortization factor shown in
Table 10.1.2.1.1-1, and amortizing these retail costs over the volume of ethanol in El 5 and E85
914 Moriarity, K.; E15 and Infrastructure; National Renewable Energy Laboratory; May 2015
915 E15's Compatibility withUST Systems; Office of Underground Storage Tanks, Environmental Protection
Agency; January 2020.
916 UST System Compatibility with Biofuels; Environmental Protection Agency; July 2020.
917 Conversations with Ryan Haerer, Office of Underground Storage Tanks; Spring 2022
918 Renkes, Robert; Scenarios to Determine Approximate Cost for E15 Readiness; Prepared by the Petroleum
Equipment Institute for the United States Department of Agriculture; September 6, 2013.
919 Because only a small percentage of the motor vehicle fleet is comprised of fuel flexible vehicles (FFVs) which
can refuel on E85, typically a retail station only offers E85 from a single dispenser at the retail station.
920 "Communication with USD A on the BIP program 1-19-22," available in the docket.
419
-------
(15% for El 5 and 74% for E85), covering the cost of capital for the retail equipment adds 54 and
19 cents per gallon to the ethanol portion of El 5 and E85, respectively. When solely amortizing
this retail cost solely over the 5% and 64% of ethanol that is incremental to E10, the cost is 1.62
and 22 cents per gallon of ethanol in E15 and E85 in excess of E10, respectively.
10.1.4.2 Biodiesel and Renewable Diesel Distribution Costs
Biodiesel distribution costs were determined by ICF under contract to EPA based on an
estimate of biodiesel being moved by rail and by truck, within each PADD, and between
PADDs.921 While biodiesel production is more spread out across the country than ethanol, a
significant amount must still be moved long distances to match the production to the demand.
The internal PADD rail costs were estimated to be 15 cents per gallon and truck movements for
shorter fuel movements were estimated based on distance moved. Movement of these fuels
between PADDs was assumed to be made by rail for most areas and also by ship from the Gulf
Coast to the West Coast. ICF relied on EIA reports for biofuel movements between PADDs.
Based on these analyses, the inter-PADD movements are estimated to cost 15 to 32 cents per
gallon, depending on the distance that the biodiesel must travel.
Renewable diesel fuel distribution costs are assumed to be the same as biodiesel. Because
renewable diesel is very similar in quality as diesel fuel, it can more readily be blended in more
places in the diesel fuel distribution system, including at refineries, where the renewable diesel
fuel would be moved by the same distribution system as diesel fuel. Thus, if renewable diesel is
used locally its distribution costs would likely be lower than biodiesel. However, much of the
renewable diesel is expected to be distributed to the West Coast to help meet the Low Carbon
Fuel Standard programs there.
Table 10.1.4.2-1 summarizes the biodiesel and renewable diesel distribution costs for
each PADD taking into account the amount of fuel that is distributed within PADDs and between
PADDs, and shows the national average distribution cost and that average cost adjusted to 2022
dollars.
Table 10.1.4.2-1: Estimated Biodiesel and Renewable Diesel Fuel Distribution Cost by
PADD
Destination
Location
PADD Total
Transportation Cost (Wgal)
PADD 1
21.6
PADD 2
15.0
PADD 3
16.0
PADD 4
25.0
PADD 5
23.8
U.S. Avg.
17.7
U.S. Average 2022$
19.7
921 Modeling a No-RFS Case; ICF Incorporated; Work Assignment 0,1-11, EPA contract EP-C-16-020; July 17,
2018.
420
-------
10.1.4.3 Renewable Natural Gas (RNG)
10.1.4.3.1 Distribution Costs
Renewable natural gas (RNG) which is gathered from landfill off-gassing and cleaned up
must then be transported to where it can be used. Typically, this RNG will end up in a nearby
natural gas pipeline, but in some rare cases it also could be compressed or liquified for
dispensing into the onboard CNG or LNG tanks of a local truck fleet at or near the landfill site.
Information on the length of pipeline needed to bring landfill gas to a nearby natural gas
pipeline is not readily available, but we made some assumptions to estimate this distance.
Landfills are generally located near to, although not in, urban areas to keep the transportation
costs lower for hauling the waste to the landfill. The landfill gas is estimated to be moved 5 miles
to access a commercial natural gas pipeline. For installing each mile of pipeline, it is estimated to
cost $1 million, and adds up to $6.7 million in 2022 dollars for the entire 5 mile pipeline.922 A
typical volume case was modeled of 600 standard cubic feet per minute to estimate the cost for a
typical sized landfill.923 When the pipeline capital costs are amortized over that typical volume
of landfill gas, the pipeline capital cost is estimated to cost $1.89 per million BTU.924
Once the RNG is transported through the new pipeline to the natural gas pipeline, it
incurs a cost for distribution through the existing natural gas pipeline. Landfills are located near
urban areas which are destination areas for natural gas pipelines. This means that the distribution
costs for RNG in the natural gas pipeline would be less than that for natural gas which is being
distributed longer distances from natural gas production areas. Natural gas will incur both
variable and fixed operating costs in the upstream pipelines, which RNG will avoid by being
injected downstream. Furthermore, the addition of biogas downstream in the natural gas pipeline
system can help the natural gas distribution system avoid capital investments that would
otherwise be necessary to debottleneck the upstream natural gas pipeline system to meet
commercial and industrial sector demand increases. If we assume that RNG would be injected
into a natural gas pipeline at least large enough to serve commercial consumers, the RNG
distribution cost can be based on commercial natural gas distribution costs which are represented
by the natural gas prices to commercial consumers. As summarized in Table 10.2.2-2,
distribution of natural gas to commercial consumers is estimated to cost $5.53 per million cubic
feet. We could not find detailed cost information for the distribution of commercial natural gas
through different parts of the distribution system that would allow us to scale the commercial
natural gas distribution costs to the portion of the natural gas pipeline used by RNG. For this
reason, half of the commercial natural gas distribution cost, or about $2.4 per million cubic feet,
is assumed to apply to biogas for distribution to the natural gas pipeline.925
922 Landfill Gas Energy Cost Model (LFGcost-Web); Version 3.5; https://www.epa.gov/lniop.
923 Economic Analysis of the US Renewable Natural Gas Industry; The Coalition of Renewable Natural Gas;
December 2021.
924 The 5.3 million capital cost is amortized over the biogas volume by first multiplying it by the capital cost
amortization factor (0.11) to derive an annual average cost, and then dividing this volume by the annual volume of
biogas which is estimated to be flowing at 600 cubic feet per minute.
925 Biogas producers tell us that they are being charged an equivalent distribution price that natural gas producers are
being charged which essentially assumes that they are using the entire natural gas pipeline. This pricing scheme,
421
-------
While this cost analysis assumes the biogas is being produced entirely at landfills, it is
worthwhile to consider the situation other RNG producers are likely to face to distribute their
biogas. Like landfills, RNG production at wastewater treatment plants and municipal waste
digesters are located near cities and thus would likely have distribution costs similar to landfills.
Conversely, agricultural waste digesters are much more likely to be located in rural areas further
away from both natural gas pipelines and urban areas. The distribution costs for RNG producers
using agricultural waste digesters would likely be higher. Some of these rural locations may be
so remote that the RNG could be considered stranded and not readily available for use as
transportation fuel, although such stranded locations could perhaps still provide RNG to local
truck fleets which distribute agricultural products.926
10.1.4.3.2 Retail Costs
Retail facilities to dispense RNG are more expensive compared to other transportation
fuel retail costs. One information source provided an estimate that a larger sized CNG retail
facility would cost about $4.61 per million BTU, so this was used for the RNG retail cost.927
When adjusted to 2022 dollars, the estimated retail cost to dispense RNG is estimated to be $6.53
per million BTU.
10.2 Gasoline, Diesel Fuel and Natural Gas Costs
10.2.1 Production Costs
As renewable fuel use increases or decreases, the volume of petroleum-based products,
such as gasoline and diesel fuel, would decrease or increase, respectively. This change in
finished refinery petroleum products results in a change in refinery industry costs. The change in
costs would essentially be the volume of fuel displaced multiplied by the cost for producing the
fuel.
In addition, there could be a situation where we may need to account for capital
investments made by the refining industry. For example, increasing renewable fuel standards
could reduce capital investments refiners would otherwise make to increase refined product
production above previous levels. In this case increased renewable fuel capital investments
would offset decreased refining industry investments. However, we have not assumed for this
analysis that there would be any reduction in refining industry investments considering the
current situation. After the economic impact of the COVID-19 pandemic, Energy Information
Administration (EIA) data shows that gasoline and diesel fuel demand are lower now and as of
early 2022, only diesel fuel is expected to increase above previous levels.928 Furthermore, light-
though, does not represent the true social cost for distributing biogas, and a separate distribution cost is estimated for
biogas.
926 The term "stranded" means the cost to recover and use the biogas is too high to justify installing the equipment
collect upgrade and distribute it for commercial use.
927 Permitting CNG and LNG Stations, Best Practices Guide for Host Sites and Local Permitting Authorities;
prepared for The California Statewide Alternative Fuel and Fleets Project by Clean Fuel Connection, Inc.
928 2023 Annual Energy Outlook; Table 12 Petroleum and Other Liquid Prices, and Table 57 Component of Selected
Petroleum Product Prices; March 3, 2023.
422
-------
duty and heavy-duty greenhouse gas standards will continue to phase-in, continuing to reduce
transportation fuel demand.929'930'931'932 Thus, we would not anticipate there to be refined product
investment regardless of the renewable fuel volumes and thus no savings that would offset
renewable fuel investments.
10.2.1.1 Gasoline and Diesel Fuel Production Costs
The production cost of gasoline and diesel fuel are based on the projected wholesale price
for gasoline and diesel fuel provided in AEO 2023.933 The projected Brent crude oil prices and
gasoline and diesel fuel wholesale prices in 2023 through 2025 are summarized in Table
10.2.1.1-1.
Table 10.2.1.1-1 Estimated Gasoline Production Costs
Gasoline
Diesel Fuel
2023
2024
2025
2023
2024
2025
Brent Crude Oil Prices ($/bbl)
91.5
92.5
87.0
91.5
92.5
87.0
Wholesale Prices - assumed to
be Production Costs ($/gal)
2.81
2.48
2.26
3.44
3.21
2.93
Since the Energy Information Administration models much of the RFS program in its
AEO modeling, some price impact of the RFS program is likely represented in these wholesale
gasoline and diesel fuel prices. The AEO models the most recent RFS standards, so these
wholesale price estimates would be optimal for modeling the final rule RFS standards
incremental to the 2022 baseline. The RFS impact on the AEO gasoline and diesel fuel prices
will slightly bias the cost analysis conducted for the No RFS Baseline, however, the impact on
the estimated costs is expected to be minimal and within the accuracy of the cost analysis.
10.2.1.2 Natural Gas Production Cost
For estimating the cost of biogas relative to natural gas, it is necessary to estimate the
production cost of fossil natural gas. The natural gas production cost can be estimated using
natural gas spot prices. In its AEO 2023, EIA projects the natural gas spot price for Henry Hub to
average $5.27 per thousand cubic feet in 2023 and decrease to $3.49 per thousand cubic feet in
929 Environmental Protection Agency, Department of Transportation; Final rule for Model Year 2012-2016 Light-
Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards; May 7, 2010.
930 Revised 2023 and Later Model Year Light-Duty Vehicle Greenhouse Gas Emissions Standards; Environmental
Protection Agency, December 30, 2021.
931 Environmental Protection Agency, Department of Transportation; Final Rule for Phase 1 Greenhouse Gas
Emissions Standards and Fuel Efficiency Standards for Medium-Duty and Heavy-Duty Engines and Vehicles;
September 15, 2011.
932 Environmental Protection Agency, Department of Transportation; Final Rule for Phase 2 Greenhouse Gas
Emissions Standards and Fuel Efficiency Standards for Medium-Duty and Heavy-Duty Engines and Vehicles;
October 25, 2016.
933 2023 Annual Energy Outlook; Table 4a. U.S. Petroleum and Other Liquids Supply, Consumption and
Inventories; March 3, 2023.
423
-------
2025.934 The Henry Hub spot price most closely represents the natural gas field price, and thus is
a proxy for its production cost.
10.2.2 Gasoline, Diesel Fuel and Natural Gas Distribution and Blending Cost
10.2.2.1 Gasoline and Diesel Fuel
Gasoline and diesel fuel distribution costs from refineries to terminals are estimated as
the difference between wholesale prices and terminal prices (which we estimated based on
historical sales-for-resale prices). This results in estimated gasoline and diesel fuel distribution
costs to the terminal of 5 and 8 cents per gallon, respectively.
We also estimated the distribution costs from terminals to retail stations, which also
contains the retailing costs. To do so, we first calculated the retail costs of gasoline, less taxes.
We calculated this by subtracting average federal and state taxes, which are 550 per gallon for
gasoline and 640 gallon for diesel fuel, from historical gasoline and diesel fuel retail prices.
Then, we calculated the difference between historical retail prices (less taxes) and historical
terminal prices (estimated as sales for resales prices) to estimate the distribution costs from the
terminal to retail and retail costs. The resulting terminal and retail distribution costs for gasoline
and diesel fuel are estimated to be 20 and 40 cents per gallon for gasoline and diesel fuel,
respectively. These various prices and estimated costs are summarized in Table 10.2.2.1-1.
Table 10.2.2.1-1: Estimated Gasoline and Diesel Fuel Distribution and Retail Costs ($/gal)
Gasoline
Diesel Fuel
2017
2018
2019
Average
2017
2018
2019
Average
Bulk Price
1.64
1.94
1.74
1.77
1.62
2.05
1.86
1.85
Sales for Resale
1.69
1.98
1.81
1.83
1.69
2.13
1.96
1.93
Retail Price
2.42
2.72
2.60
2.58
2.65
3.18
3.06
2.96
Taxes
0.55
0.55
0.55
0.55
0.64
0.64
0.64
0.64
Distribution Costs
0.05
0.04
0.07
0.05
0.07
0.08
0.08
0.08
Retail Costs
0.18
0.19
0.24
0.20
0.32
0.41
0.46
0.40
We then apply the estimated gasoline and diesel fuel distribution costs to the projected
wholesale gasoline and diesel fuel prices in Table 10.2.1.1-1 for each year to estimate the
gasoline and diesel fuel prices from refinery to retail. These gasoline and diesel fuel prices are
summarized in Table 10.2.2.1-2.
934 2023 Annual Energy Outlook; Table 13 Natural Gas Supply, Disposition, and Prices; March 3, 2023.
424
-------
Table 10.2.2.1-2:
'rojected Gasoline and Diesel Production Costs ($/
gal)
2023
2024
2025
Brent Crude Oil Prices
91.5
92.5
87.0
Gasoline
Retail Cost minus taxes
3.06
2.64
2.52
Terminal and Retail Costs
0.20
0.20
0.20
Terminal Costs
2.86
2.43
2.32
Distribution Cost
0.05
0.05
0.05
Production Cost
(from Table 10.2.1.1-1)
2.81
2.38
2.26
Diesel Fuel
Retail Cost minus taxes
3.91
3.59
3.40
Terminal and Retail Costs
0.40
0.40
0.40
Terminal Costs
3.51
3.19
3.00
Distribution Cost
0.08
0.08
0.08
Production Cost
(from Table 10.2.1.1-1)
3.44
3.11
2.93
10.2.2.2 Natural Gas
EIA projects natural gas prices downstream of natural gas production fields which can be
used to estimate natural gas distribution costs.935 The three principal natural gas consumers are
industrial, commercial, and residential. Industrial consumers consume the largest natural gas
volumes per facility, and due to the very large consumption, the distribution costs are lowest.
Commercial entities are medium sized consumers, and their distribution costs are higher than
industrial consumers. Residential consumers, because of their very low consumption, must pay a
much larger distribution cost to maintain the distribution system for much lower consumption to
each home. EIA also provides a price for natural gas sold into the transportation sector, although
this price includes road taxes which would need to be omitted for the purposes of this cost
analysis, so we did not use EIA's natural gas to transportation sector cost.936
The varying costs for these different natural gas categories permit estimating natural gas
distribution costs for natural gas consumed by motor vehicles. Natural gas produced and
distributed to retail outlets to refuel natural gas trucks and cars most likely falls in the category of
midsized consumers, or commercial users. The distribution costs of natural gas can therefore be
estimated by subtracting the projected Henry Hub prices from the projected commercial prices.
Thus, Henry Hub prices projected in AEO 2023 were subtracted from the commercial prices for
2023 through 2025. Table 10.2.2.2-1 summarizes the calculation of natural gas distribution costs.
To put the natural gas costs on the same footing as the biogas, we also add $6.53 per million
BTU for retail costs.937
935 Table 13 Natural Gas Supply, Disposition and Prices; Annual Energy Outlook 2023.
936 Taxes are not included in social cost estimates because they are not true costs, only transfer payments.
937 Permitting CNG and LNG Stations, Best Practices Guide for Host Sites and Local Permitting Authorities;
prepared for The California Statewide Alternative Fuel and Fleets Project by Clean Fuel Connection, Inc.
425
-------
Table 10.2.2.2-1: Natural Gas Distribution Cost (S/t
2023
2024
2025
Commercial Prices
10.73
9.69
9.19
Henry Hub Prices
5.27
4.07
3.49
Pipeline Distribution Costs
5.65
5.76
5.83
Retail Station Costs
6.53
6.53
6.53
Total Distribution & Retail
Station Costs
12.18
12.29
12.36
lousand cubic feet)
10.3 Fuel Energy Density and Fuel Economy Cost
To estimate the change in fossil fuel volume that would occur with these changes in
renewable fuel volumes and to estimate the fuel economy cost summarized in Chapter 10.4.1, it
was necessary to estimate the energy density of each fuel. Table 10.3-1 contains the estimated
energy densities for the various renewable fuels and petroleum fuels analyzed for this cost
analysis.
Table 10.3-1: Lower Heating Value (LHV) Energy Densities
LHV Energy Density (GREET 2017)
BTU/gal
Gasoline (E0)a
114,200
Diesel Fuel
128,450
Pure Ethanol
76,330
Natural Gas Liquids
83,686
Denatured Ethanol
76,477
E10 Gasoline
110,428
El 5 Gasoline
108,542
E85b
86,285
Biodiesel
119,550
Renewable Diesel
122,887
Crude Oil
129,670
a From Chevron Paper.938
b Assumed to contain 74% ethanol.
To account for the fuel economy effect for the cost analysis, the change in fossil fuel
volume displaced by a change in renewable fuel volume is estimated by the relative energy
content of the renewable and fossil fuels. However, if the energy density is not the same between
the fossil fuel and renewable fuel displacing it, the energy equivalent replacement is not one-for-
one on a volume basis. For example, ethanol contains about 33% lower energy per volume than
the gasoline it is displacing, such that 100 gallons of ethanol would displace 67 gallons of
gasoline. The fuel economy effect is therefore inherent in the cost analysis and is not reported
out separately.
938 Diesel Fuels Technical Review; Chevron Global Marketing; 2007.
426
-------
For the individual fuel cost summary in Chapter 10.4.1, it is desirable to report out a
specific fuel economy effect. To do so, the difference in energy density between the renewable
fuel and fossil fuel is divided by the fossil fuel energy density and then multiplied times the
fossil fuel cost at retail, before taxes, to estimate the fuel economy effect.
10.4 Costs
10.4.1 Individual Fuels Cost Summary
Table 10.4.1-1 summarizes the estimated overall societal costs (including production,
distribution, blending, and fuel economy) for the renewable fuels analyzed for this rulemaking
for the years 2023-2025. These costs do not account for the per-gallon federal cellulosic biofuel
and biodiesel tax subsidies, nor do they consider taxes or tax subsidies more generally, as these
are transfer payments which are not relevant in the estimation of societal costs. Nor do these
costs consider state or local infrastructure support funding or the funding from USDA's Blends
Infrastructure Incentive Program (HBIIP) which offsets half of the investment costs for
revamping retail stations to be compatible with E85 and El5.939 A separate line item is added for
El 5 and E85 which only adds in V2 of the retail cost to help illustrate the impact that the HBIIP
program would have on the costs for these fuels. The costs of renewable fuels, other than biogas,
are primarily influenced by the feedstock costs, which can vary significantly depending on a
wide range of factors domestically and internationally, especially since many of them are also
agricultural commodities.
To put the different fuels on an equivalent basis for the miles driven, the societal cost
analysis also needs to account for each fuel's impact on fuel economy, which is first discussed in
Chapter 10.3. While these costs may not always be reflected in the sales prices among the market
participants (e.g., if refiners sell, and consumers buy, gasoline based on volume, not energy
content), the varying impacts on fuel economy among the fuels nevertheless still result in
different costs to consumers in operating their vehicles and therefore must be accounted for in a
social cost analysis. The cost associated with the impact of renewable fuels on fuel economy
costs are determined relative to the fuels they are assumed to displace; ethanol displaces
gasoline, biodiesel and renewable diesel displace diesel fuel, and RNG displaces natural gas.940
To the extent that RINs representing RNG incentivize some incremental growth in sales of
CNG/LNG trucks at the expense of diesel fueled trucks, then some RNG could also displace
diesel fuel. However, this is expected to be a relatively minor occurrence for the volumes and
timeframe of this action, and so is not included in this cost analysis.
939 Higher Blends Infrastructure Incentive Program; USD A; https://www.rd.usda.gov/tibiip.
940 Fuel economy costs are calculated by multiplying the total of petroleum fuel production, distribution and retail
costs by the difference in energy density (BTU per gallon) between the petroleum fuel being displaced and the
renewable fuel, and the result of that operation is divided by the energy density of the petroleum fuel. For ethanol
blended as E10 as an example: (denatured ethanol production + distribution + blending cost) * (E10 gasoline energy
density - denatured ethanol energy density)/denatured ethanol energy density.
427
-------
The cost shown for RNG in two different units. The first is RNG dollars per million BTU
and dollars per ethanol-equivalent gallon. Table 10.4.1-1 is divided into two subparts, "a" and
"b."
Table 10.4.1-la: Renewable Fuels Costs estimated for 2023-2025 ($/gallon unless otherwise
noted; 2021$)
Fuel
Blending
Distribution
Retail
Economy
Production Cost
Cost
Cost
Cost
Cost
2023
2024
2025
Corn
E10
3.02
2.61
2.37
-0.69
0.40
1.02
Starch
E15 w V2
Ethanol
Retail Costs
3.02
2.61
2.37
0.40
0.81
1.02
E15
w/Retail
Costs
3.02
2.61
2.37
0.40
1.62
1.02
E85 w/1/2
Retail Costs
3.02
2.61
2.37
0.40
0.11
1.02
E85
w/Retail
Costs
3.02
2.61
2.37
0.40
0.22
1.02
Biodiesel
Soy Oil
6.83
6.59
6.65
0.59
0.27
Corn Oil
5.19
4.59
4.10
0.59
0.27
Waste Oil
6.35
5.21
5.01
0.59
0.27
Renewable
Soy Oil
7.42
7.15
7.23
0.59
0.17
Diesel
Corn Oil
5.66
5.02
4.50
0.59
0.17
Waste Oil
6.91
5.68
5.47
0.59
0.17
Other
Sugar Cane
Advanced
Ethanol
2.73
2.73
2.73
-0.69
0.40
1.02
Cellulosic
RNG ($/gal
Biofuels
Ethanol)
0.64
0.64
0.64
0.44
0.50
-
RNG
($/mmBTU)
8.43
8.43
8.43
5.82
6.53
-
Corn Kernel
Fiber E10
Ethanol
3.02
2.61
2.37
-0.69
0.40
1.02
a Fuel economy cost is per fuel being displaced—ethanol displaces gasoline, renewable diesel and biodiesel
displaces diesel fuel, and biogas displaces natural gas.
b It is important to note that in estimating the social cost for this rulemaking the fuel economy cost for ethanol
blended into E10 is included since this is a cost that consumers will bear. However, when refiners are considering
whether to blend ethanol, such as for estimating volumes for the No RFS baseline, they do not consider the fuel
economy effect and this distinction is important for understanding ethanol's relative economic viability in the
marketplace.
0 For modeling the societal costs of E15 and E85 shown in Chapters 10.4.2 and 10.4.3, the cost analysis is conducted
for the entire volume of E15 and E85, and includes the blending cost savings for the E10 BOB used to blend with
E15 and E85. For the cost analysis shown here, the cost for E15 and E85 is solely for the ethanol volume above that
428
-------
blended at 10 volume percent and therefore does not include any blending value for E10 BOBs to represent the
marginal cost for the ethanol volume above E10.
Table 10.4.1-lb: Renewable Fuels Costs estimated for 2023-2025 ($/gallon unless otherwise
noted; 2022$)
Total Cost
2023
2024
2025
Corn Starch
E10
3.76
3.35
3.10
Ethanol
El5 w/1/2 Retail Costs
5.26
4.85
4.60
El5 w/Retail Costs
6.06
5.65
5.41
E85 w/1/2 Retail Costs
4.55
4.14
3.90
E85 w/Retail Costs
4.66
4.25
4.01
Biodiesel
Soy Oil
7.70
7.45
7.51
Corn Oil
6.05
5.46
4.96
Waste Oil
7.22
6.08
5.87
Renewable
Soy Oil
8.18
7.91
7.99
Diesel
Corn Oil
6.43
5.78
5.26
Waste Oil
7.67
6.45
6.24
Other Advanced
Sugar Cane Ethanol
3.46
3.46
3.46
Cellulosic
Biogas ($/gal ethanol)
1.59
1.59
1.59
Biofuels
Biogas ($/mmBTU)
20.78
20.78
20.78
Corn Kernel Fiber El0
3.76
3.35
3.10
Ethanol
The distribution costs for the biofuels are nationwide averages, which does not capture
the substantial difference depending on the destination. For example, ethanol distribution costs
from the ethanol plants to terminals can vary from under 10 cents per gallon for local distribution
in the Midwest, to over 30 cents per gallon for moving the ethanol to the coasts. Thus, total
ethanol cost blended as E10 can vary from around 3.66 to 3.86 per gallon. Biogas distribution
includes both the amortized capital cost of transporting the biogas to a nearby pipeline as well as
the amortized retail distribution capital costs, since the retail facilities for natural gas trucks are
relatively expensive.
Table 10.4.1-2 summarizes production and distribution costs for each category of fossil
transportation fuel—gasoline, diesel fuel, and natural gas. For gasoline and diesel, production
costs are based on wholesale prices in AEO 2023.941 Projected natural gas spot prices from the
AEO 2023 are used to represent both feedstock and production costs of fossil natural gas.
The distribution costs for gasoline and diesel fuel are typical for these fuels. While they
can vary depending on the transportation distance, the differences between high and low
distribution costs for gasoline and diesel fuel are likely lower than that for renewable fuels due to
the well-established pipeline distribution system for petroleum fuels. The natural gas distribution
costs are based on the difference between the projected price for natural gas sold to commercial
941 EIA, Annual Energy Outlook 2023, Energy Information Administration, March, 2023.
429
-------
entities and the projected natural gas spot price, which reflects the price at the point of
production.
Table 10.4.1-2: Gaso
ine, Diesel Fuel, and Nal
tural Gas Costs for 2023-2025 (2021$)
Production Cost
Distribution
Cost
Retail
Cost
Total Cost
2023
2024
2025
2023
2024
2025
Gasoline ($/gal)
2.81
2.38
2.26
0.26
3.06
2.64
2.52
Diesel Fuel ($/gal)
3.44
3.11
2.93
0.47
3.91
3.59
3.40
Natural Gas $/gal
ethanol
0.51
0.42
0.38
0.51
0.50
1.52
1.43
1.39
Natural Gas
($/million BTU)
6.76
5.56
4.95
6.76
6.53
20.04
18.84
18.23
Table 10.4.1-3 compares the data from Tables 10.4.1-1 and 2 to show the relative cost of
the renewable fuels with the fossil fuels they are assumed to displace.
Table 10.4.1-3: Relative Renewable Fuel Costs for 2023-2025 ($/gal unless otherwise noted,
2022$)
Total Net Cost
2023
2024
2025
Corn Starch
Ethanol
E10
0.69
0.61
0.58
El5 w/1/2 Retail Costs
2.19
2.11
2.08
El5 w/Retail Costs
3.00
2.92
2.89
E85 w/1/2 Retail Costs
1.49
1.41
1.38
E85 w/Retail Costs
1.60
1.52
1.49
Biodiesel
Soy Oil
3.79
3.77
4.12
Corn Oil
2.14
1.77
1.56
Waste Oil
3.31
2.39
2.48
Renewable
Diesel
Soy Oil
4.27
4.23
4.59
Corn Oil
2.52
2.10
1.86
Waste Oil
3.76
2.76
2.84
Other Advanced
Sugar Cane Ethanol
0.68
1.01
1.22
Cellulosic
Biofuels
Biogas ($/gal ethanol)
0.06
0.15
0.20
Biogas ($/mmBTU)
0.74
1.93
2.54
Corn Kernel Fiber El0
Ethanol
0.69
0.61
0.58
10.4.2 Costs Relative to the No RFS Baseline
In this section, we summarize the estimated costs for the changes in renewable fuel
volumes described in Chapter 3.2 (changes relative to the No RFS baseline volumes described in
Chapter 2). For this analysis we considered all societal costs, including production, blending, and
distribution costs, and differences in energy density.
430
-------
10.4.2.1 Volumes
An important first step for the cost analysis is understanding the change in both
renewable fuel volumes and the associated change in the fossil fuel volume, which is calculated
based on its energy content relative to the renewable fuel that it is displaced by. Table 10.4.2.1-1
summarizes the renewable and fossil fuel changes relative to the No RFS baseline, and Table
10.4.2.1-2 summarizes the volumes associated with the supplemental standard for 2023.
Table 10.4.2.1-1: Renewable Fuel and Fossil Fuel Volume Changes Relative to the No RFS
Baseline (million gallons, except where noted)
Change in Renewable Fuel Volume
Change
in Fossil Fuel Volume
Fuel Type
2023
2024
2025
Fuel Type
2023
2024
2025
Cellulosic biofuel - Total
CNG - landfill biogas (MMft3)
36505
50739
68734
Natural Gas
-36505
-50739
-68734
Corn Kernel Fiber Ethanol
0
0
0
Gasoline
0
0
0
Non-cellulosic adv. - Total
Biodiesel - Soy
841
757
755
Diesel Fuel
-783
-705
-703
Biodiesel -FOG
-101
-92
-113
Diesel Fuel
94
86
105
Biodiesel - Corn Oil
46
63
20
Diesel Fuel
-43
-59
-19
Biodiesel - Canola
292
307
323
Diesel Fuel
-272
-286
-301
Renewable Diesel - Soy
457
671
729
Diesel Fuel
-437
-642
-697
Renewable Diesel - FOG
110
101
121
Diesel Fuel
-105
-97
-116
Renewable Diesel - Corn
130
-64
-20
Diesel Fuel
-124
61
19
Renewable Diesel - Canola
216
182
291
Diesel Fuel
-207
-174
-278
Sugar Cane Ethanol
0
0
0
Gasoline
0.0
0.0
0.0
Conventional - Total
Ethanol - E10
-81
-91
-101
Gasoline
55
61
68
Ethanol - El5
80
93
106
Gasoline
-54
-62
-71
Ethanol - E85
260
272
283
Gasoline
-174
-182
-190
Change in Biogas Volume
36505
50739
68734
-
-
-
-
Change in Ethanol Volume
259
274
289
-
-
-
-
Change in Biodiesel Volume
1078
1035
985
-
-
-
-
Change in Renewable Diesel Volume
913
890
1121
-
-
-
-
Change in Gasoline Volume
-
-
-
-
-174
-184
-193
Change in Diesel Fuel Volume
-
-
-
-
-1877
-1815
-1989
Change in Natural Gas Volume
-
-
-
-
-36505
-50739
-68734
431
-------
Table 10.4.2.1-2 Supplemental Standard Renewable Fuel and Petroleum Fuel Volume
Changes
Change in Renewable Fuel Volume
Change in Petroleum Fuel Volume
2023
2024
2025
2023
2024
2025
Supplemental Std. RD Soy Oil
147
0
0
Diesel Fuel
-141
0
0
The change in gasoline and diesel volume for each case is used to estimate the change in
crude oil based on its relative energy content. The change in petroleum demanded and its effect
on both imported crude oil, domestic crude oil, and imported petroleum products, is projected
based on these effects by a comparison of two separate economic cases: the Low Economic
Growth Case and the Reference Case, modeled by EIA in its AEO 2023.942 The AEO Low
Economic Growth Case estimates lower refined product demand than that of the Reference case,
and due to the reduced refined product demand the AEO estimates changes in reduced imports of
crude oil refined products. The two AEO cases project that for a volume of reduced gasoline or
diesel fuel, 92 percent of that gasoline or diesel reduction would be attributed to reduced crude
oil imports and imports of refined product would decrease by 8 percent. Based on these
correlations, Table 10.4.2.1-3 summarizes the projected change in petroleum imports expected
from the increased consumption of renewable biofuels over the years 2023 to 2025 relative to the
No RFS baseline, and Table 10.4.2.1-4 shows the same information, but also accounts for the
Supplemental Standard. In Table 10.4.2.1-4 we also consider the projected change in imported
renewable fuels in addition to changes in petroleum products. The change in crude oil volume
and imported petroleum products is used for the energy security analysis contained in Chapter 5.
Table 10.4.2.1-3: Projected Change in Petroleum Imports Due to Increased Renewable Fuel
Consumption Relative to the No RFS Base
2023
2024
2025
Change in Imported Gasoline
-14
-15
-16
Change in Imported Diesel Fuel
-151
-146
-160
Total Change in Crude Oil
-1850
-1802
-1968
Change in Domestic Crude Oil
0
0
0
Change in Imported Crude Oil
-1850
-1802
-1968
ine (million gallons)
Table 10.4.2.1-4: Projected Change in Petroleum Imports Due to Increased Renewable Fuel
Consumption Relative to the No RFS Baseline; accounts for the Renewable Fuels imports
2023
2024
2025
Change in Imported Gasoline
-14
-15
-16
Change in Imported Diesel Fuel
-96
-80
-92
Total Change in Crude Oil
-1978
-1802
-1968
Change in Domestic Crude Oil
0
0
0
Change in Imported Crude Oil
-1978
-1802
-1968
942 "Change in product demand on imports AEO 2023 for SET final rule", spreadsheet available in the docket.
432
-------
10.4.2.2 Cost Impacts Relative to the No RFS Baseline
Table 10.4.2.2-1 summarizes the component cost (production, distribution, blending
retail) of each biofuel fuel type for 2023 through 2025 compared to the fossil fuel it is displacing,
and Table 10.4.2.2-2 provides this information for the supplemental standard.
433
-------
Table 10.4.2.2-1 Renewable and Petroleum Fuel Costs for 2023 to 2025 (million dollars;
year 2022 dollars)
Renewable Fuel
Petroleum Fuel
Production
Distribution
Blending
Production
Distribution
Total
Cellulosic biofuel
CNG - landfill biogas
319
579
0
-247
-631
21
Corn Kernel Fiber Ethanol
0
0
0
0
0
0
Non-cellulosic adv.
Biodiesel - Soy
5747
499
0
-2691
-370
3185
Biodiesel -FOG
-642
-60
0
323
44
-334
Biodiesel - Corn Oil
239
27
0
-147
-20
99
Biodiesel - Canola
1995
173
0
-934
-129
1106
2023
Renewable Diesel - Soy
3391
271
0
-1503
-207
1953
Renewable Diesel - FOG
760
65
0
-362
-50
413
Renewable Diesel - Corn
675
77
0
-428
-59
265
Renewable Diesel - Canola
1603
128
0
-710
-98
923
Suaar Cane Ethanol
0
0
0
0
0
0
Conventional
Ethanol - E10
-246
-33
56
153
14
-56
Ethanol - El5
243
76
-37
-151
-14
117
Ethanol - E85
788
155
-24
-490
-45
384
Cellulosic biofuel
CNG - landfill biogas
444
805
0
-282
-877
89
Corn Kernel Fiber Ethanol
0
0
0
0
0
0
Non-cellulosic adv.
Biodiesel - Soy
4987
149
0
-2422
-333
2380
Biodiesel -FOG
-480
-18
0
294
41
-163
Biodiesel - Corn Oil
289
12
0
-202
-28
73
Biodiesel - Canola
2022
60
0
-982
-135
965
2024
Renewable Diesel - Soy
4797
143
0
-2062
-304
2575
Renewable Diesel - FOG
574
20
0
-332
-46
216
Renewable Diesel - Corn
-294
-13
0
210
29
-67
Renewable Diesel - Canola
1301
57
0
-542
-82
734
Suaar Cane Ethanol
0
0
0
0
0
0
Conventional
Ethanol - E10
-237
-37
63
151
16
-45
Ethanol - El5
243
88
-43
-154
-16
118
Ethanol - E85
712
162
-25
-452
-47
350
Cellulosic biofuel
CNG - landfill biogas
601
1091
0
-340
-1188
163
Corn Kernel Fiber Ethanol
0
0
0
0
0
0
Non-cellulosic adv.
Biodiesel - Soy
5021
149
0
-2416
-333
2421
Biodiesel -FOG
-566
-22
0
362
50
-177
Biodiesel - Corn Oil
82
4
0
-64
-9
13
Biodiesel - Canola
2148
64
0
-1033
-142
1036
2025
Renewable Diesel - Soy
5267
143
0
-2041
-330
3040
Renewable Diesel - FOG
662
24
0
-398
-55
233
Renewable Diesel - Corn
-82
-4
0
66
9
-11
Renewable Diesel - Canola
2103
57
0
-815
-132
1214
Suaar Cane Ethanol
0
0
0
0
0
0
Conventional
Ethanol - E10
-239
-41
69
153
17
-40
Ethanol - El5
252
100
-49
-161
-18
124
Ethanol - E85
672
168
-26
-430
-10
374
434
-------
Table 10.4.2.2-2 Renewable Fuel and Petroleum Fuel Costs for the 2023 Supplemental
Standard (million dollars; 2022$)
Renewable Fuel
Fossil Fuel
Production
Distribution
Blending
Production
Distribution
Total
Supplemental Std.
RD Soy Oil
1091
87
0
-483.4
-66.6
628.1
To estimate the per-gallon cost on the total gasoline, diesel, and natural gas pools, the
projected total volumes for each of these fuels was obtained from AEO 2023 and summarized in
Table 10.4.2.2-3.943
Table 10.4.2.2-3: Total Gasoline, Diesel Fuel and Natural Gas Volumes
2023
2024
2025
Units
Gasoline Volume
134.14
135.06
133.06
Billion gallons
Diesel Volume
56.11
53.35
52.89
Billion gallons
Natural Gas Volume
31.42
30.16
29.82
Trillion cubic feet
The costs are aggregated for each fossil fuel type and expressed as per-gallon and per
thousand cubic feet costs in Table 10.4.2.2-4 for 2023 through 2025.
Table 10.4.2.2-4: Total Annual Rule Cost Relative to the No RFS baseline (2022$)
Total Cost
Per-Unit
(million $)
Cost
Units
Gasoline
445
0.33
0/gal gasoline
2023
Diesel Fuel
7,610
13.56
0/gal diesel
Natural Gas
55
0.1753
$/1000 FT3 natural gas
Total
8,110
4.26
0/gal gasoline and diesel
Gasoline
445
0.33
0/gal gasoline
2023 with
Diesel Fuel
8,238
14.68
0/gal diesel
Suppl. Std.
Natural Gas
55
0.1753
$/1000 FT3 natural gas
Total
8,738
4.59
0/gal gasoline and diesel
Gasoline
423
0.31
0/gal gasoline
2024
Diesel Fuel
6,775
12.70
0/gal diesel
Natural Gas
137
0.4549
$/1000 FT3 natural gas
Total
7,352
3.90
0/gal gasoline and diesel
Gasoline
458
0.34
0/gal gasoline
2025
Diesel Fuel
7,769
14.69
0/gal diesel
Natural Gas
228
0.7645
$/1000 FT3 natural gas
Total
8,455
4.55
0/gal gasoline and diesel
943 EIA, Annual Outlook 2023, Energy Information Administration, March 3, 2023.
435
-------
10.4.3 Costs Relative to the Year 2022 Volumes
10.4.3.1 Volumes
In this section, we summarize the results of our analysis estimating the costs for changes
in the use of renewable fuels relative to the year 2022 renewable fuels volumes estimated to
occur under the 2022 RFS renewable fuel obligation (RVO). This analysis is conducted the same
way as that conducted for the No RFS baseline analysis, with the only difference being the
baseline volumes. Table 10.4.3.1-1 summarizes the cost and cost savings of each biofuel fuel
type compared to the fossil fuel it is displacing for the years 2023 to 2025.
Table 10.4.3.1-1: Renewable Fuel and Fossil Fuel Volume Changes Relative to Year 2022
Volumes (million gallons, except where noted)
Change in Renewable Fuel Volume
Change
in Fossil Fuel Volume
Fuel Type
2023
2024
2025
Fuel Type
2023
2024
2025
Cellulosic biofuel
CNG - landfill biogas (MMFT3)
12242
27582
46757
Natural Gas
-12242
-27582
-46757
Corn Kernel Fiber Ethanol
6
50
76
4
33
51
Non-cellulosic adv.
Biodiesel - Soy
-13
-28
-42
Diesel Fuel
-12
-26
-39
Biodiesel -FOG
-25
-43
-61
Diesel Fuel
-23
-40
-57
Biodiesel - Corn Oil
-15
-41
-67
Diesel Fuel
-14
-38
-62
Biodiesel - Canola
25
40
56
Diesel Fuel
23
37
52
Renewable Diesel - Soy
311
525
737
Diesel Fuel
-298
-502
-705
Renewable Diesel - FOG
249
215
295
Diesel Fuel
-238
-206
-282
Renewable Diesel - Corn
-4
30
63
Diesel Fuel
4
-29
-60
Renewable Diesel - Canola
216
182
291
Diesel Fuel
-207
-174
-278
Sugar Cane Ethanol
14
14
14
Gasoline
9
9
9
Conventional
Ethanol - E10
-512
-482
-725
Gasoline
343
323
485
Ethanol - El5
14
27
40
Gasoline
-10
-18
-27
Ethanol - E85
10
21
33
Gasoline
-6
-14
-22
Change in Biogas Volume
12242
27582
46757
-
-
-
-
Change in Ethanol Volume
-482
-384
-576
-
-
-
-
Change in Biodiesel Volume
-28
-72
-114
-
-
-
-
Change in Renewable Diesel Volume
772
952
1386
-
-
-
-
Change in Gasoline Volume
-
-
-
-
327
291
437
Change in Diesel Fuel Volume
-
-
-
-
-765
-978
-1432
Change in Natural Gas Volume
-
-
-
-
-12242
-27582
-46757
Change in Imported Gasoline
26
23
35
Change in Imported Diesel Fuel
-62
-79
-115
Total Change in Crude Oil
-422
-622
-903
Change in Domestic Crude Oil
0
0
0
Change in Imported Crude Oil
-422
-622
-903
436
-------
These volumes would need to be adjusted to account for the supplemental standard which
applies in 2022 and 2023. Since the supplemental volumes applies in 2022, the baseline year for
conducting this cost analysis, 2023 volumes would not change relative to the 2022 baseline
volumes, but the renewable diesel volumes decrease in 2024 and 2025 as summarized the
volumes in Table 10.4.3.1-2.
Table 10.4.3.1-2: Soy Renewable Diesel and Diesel Fuel Volume Changes Relative to Year
2022 Volumes due to the Supplemental Standard (million gallons)
Change in Renewable Fuel Volume
Change in Petroleum Fuel Volume
2023
2024
2025
2023
2024
2025
Supplemental Std. RD Soy Oil
0
-147
-147
Diesel Fuel
0
-141
-141
10.4.3.2 Costs
Table 10.4.3.2-1 summarizes the component cost (production, distribution, blending,
retail costs, which are costs to enable sale of the renewable fuel) for each biofuel fuel type for
2023 through 2025 compared to the fossil fuel it is assumed to displace.
437
-------
Table 10.4.3.2-1: Renewable Fuel and Petroleum Fuel Costs Relative to Year 2022 Volumes
(million dollars; 2022$) ^
Renewable Fuel
Petroleum Fuel
Production
Distribution
Blending
Production
Distribution
Total
Cellulosic biofuel
CNG - landfill biogas
107
194
0
-83
-212
7
Corn Kernel Fiber Ethanol
18
2
4
-11
-1
12
Non-cellulosic adv.
Biodiesel - Soy
-89
-8
0
42
7
-48
Biodiesel -FOG
-159
-15
0
48
8
-80
Biodiesel - Corn Oil
-78
-9
0
-80
-14
-30
Biodiesel - Canola
171
15
0
0
0
92
2023
Renewable Diesel - Soy
2308
185
0
-1023
-141
1329
Renewable Diesel - FOG
1720
148
0
13
2
907
Renewable Diesel - Corn
-21
-2
0
-710
-98
-8
Renewable Diesel - Canola
1603
128
0
0
0
923
Sugar Cane E10 Ethanol
38
6
-10
-26
-2
6
Conventional
Ethanol - E10
-1549
-207
353
963
88
-353
Ethanol - E15
43
13
-7
-27
-2
21
Ethanol - E85
29
6
-1
-18
-2
14
Cellulosic biofuel
CNG - landfill biogas
241
438
0
-153
-477
49
Corn Kernel Fiber Ethanol
131
13
34
-83
-9
86
Non-cellulosic adv.
Biodiesel - Soy
-184
-17
0
81
15
-104
Biodiesel -FOG
-224
-26
0
125
24
-101
Biodiesel - Corn Oil
-188
-24
0
119
23
-71
Biodiesel - Canola
264
24
0
-116
-22
149
2024
Renewable Diesel - Soy
3753
437
0
-1565
-238
2388
Renewable Diesel - FOG
1222
128
0
-641
-122
587
Renewable Diesel - Corn
138
18
0
-89
-17
49
Renewable Diesel - Canola
1301
173
0
-542
-82
849
Sugar Cane E10 Ethanol
38
38
38
-23
-23
9
Conventional
Ethanol - E10
-1261
-195
332
800
83
-242
Ethanol - E15
70
25
-12
-45
-5
34
Ethanol - E85
56
13
-2
-36
-4
28
Cellulosic biofuel
CNG - landfill biogas
409
742
0
-231
-808
111
Corn Kernel Fiber Ethanol
180
19
52
-115
-13
124
Non-cellulosic adv.
Biodiesel - Soy
-279
-25
0
114
23
-167
Biodiesel -FOG
-306
-36
0
166
34
-142
Biodiesel - Corn Oil
-274
-40
0
182
37
-95
Biodiesel - Canola
372
33
0
-153
-31
222
2025
Renewable Diesel - Soy
5325
437
0
-2063
-334
3366
Renewable Diesel - FOG
1615
175
0
-826
-167
796
Renewable Diesel - Corn
258
37
0
-176
-36
83
Renewable Diesel - Canola
2103
173
0
-815
-132
1329
Sugar Cane E10 Ethanol
38
38
38
-21
-21
11
Conventional
Ethanol - E10
-1719
-293
499
1099
124
-289
Ethanol - E15
95
38
-18
-61
-7
47
Ethanol - E85
77
19
-3
-49
-6
39
The costs are aggregated for each fossil fuel type and costs expressed as per-gallon
gasoline and diesel fuel, and per thousand cubic feet of natural gas, in Table 10.4.3.2-2.
438
-------
Table 10.4.3.2-2: Total Cosi
ts Relative to Year 2022 Volumes (2022$)
Total Cost
(million $)
Per-Unit
Cost
Units
2023
Gasoline
-301
-0.22
0/gal gasoline
Diesel Fuel
3,085
5.50
0/gal diesel
Natural Gas
18
0.06
0/1000 FT3 natural gas
Total
2,805
1.47
0/gal gasoline and diesel
2024
Gasoline
-85
-0.06
0/gal gasoline
Diesel Fuel
3,745
7.02
0/gal diesel
Natural Gas
75
0.25
0/1000 FT3 natural gas
Total
3,712
1.95
0/gal gasoline and diesel
2025
Gasoline
-70
-0.05
0/gal gasoline
Diesel Fuel
5,393
10.20
0/gal diesel
Natural Gas
155
0.52
0/1000 FT3 natural gas
Total
5,480
2.88
0/gal gasoline and diesel
The total costs associated with the final volumes relative to the 2022 baseline do not
include the supplemental standard that applies in 2022 and 2023. If we include these
supplemental volumes and their associated costs, the total costs after 2023 are adjusted lower
based on the cost figures in Table 10.4.3.2-3 (e.g., $291 million lower cost in 2024).
Table 10.4.3.2-3: Adjustments to the Estimated Total Costs to Account for the
Supplemental Standard (million dollars)
Renewable Fuel
Fossil Fuel
Production
Distribution
Blending
Production
Distribution
Total
2023
0
0
0
0.0
0.0
0.0
2024
-722
-87
0
452
67
-291
2025
-641
-87
0
412
67
-250
10.5 Estimated Fuel Price Impacts
In this section, we estimate the impact of the use of renewable fuels on the cost to
consumers of transportation fuel and the cost to transport goods. We have estimated cost to
consumers of transportation fuel by assessing the fuel price impacts associated with this
rulemaking. We do so based on the cost of renewable fuels (less available federal tax credits) and
accounting for the cross-subsidy implemented through the RIN system. We have also used
estimates of the fuel price impacts of this rule to estimate the cost to transport goods discussed in
Chapter 10.5.5.
10.5.1 RIN Cost and RIN Value
Before estimating fuel price impacts, we first estimated the RIN cost (i.e., the cost added
to each gallon of petroleum fuel to account for the RIN obligation on the fuel) and RIN value
(i.e., the value of the RINs associated with the renewable fuel in the fuel blend) associated with
producing petroleum and renewable fuels, respectively. Because RIN prices can be impacted by
a wide variety of different factors (including the prices of renewable fuels and petroleum-based
439
-------
fuels, oil prices, commodity prices, etc.), we are not able to project what RIN prices will be in
the future. We can, however, use the average RIN prices over the last 12 months (through April
2023) as an estimate of future RIN prices, as shown in Table 10.5.1-1.
Table 10.5.1-1: Average
UN Prices (IV
[ay 2022 - April 2023)
RFS Standard
RIN
Type
Average
RIN Price
2022
Percentage
Standards
2023
Percentage
Standard
2024
Percentage
Standard
2025
Percentage
Standard
Cellulosic
Biofuel (D3)
D3
$2.79
0.35%
0.48%
0.63%
0.81%
Biomass-Based
Diesel (D4)
D4
$1.67
2.33%
2.58%
2.82%
3.15%
Other
Advanced
BiofueP (D5)
D5
$1.82
0.48%
0.33%
0.34%
0.35%
Conventional
Renewable
Fuelb (D6)
D6
$1.52
8.57%c
8.71%
8.71%
8.82%
a Other advanced biofuel is not a fuel category for which a percentage standard is established, but is calculated by
subtracting the cellulosic biofuel and biomass-based diesel standards from the advanced biofuel standard.
b Conventional renewable fuel is not a fuel category for which a percentage standard is established, but is calculated
by subtracting the advanced biofuel standard from the total renewable fuel standard.
0 Includes the 2022 total renewable fuel supplemental standard.
d Includes the 2023 total renewable fuel supplemental standard.
We then calculated the RIN cost for petroleum fuel by weighting the RIN price for each
D code by their respective RFS standard and summing the total. The results are shown in Table
10.5.1-2.
Table 10.5.1-2: Estimated RIN Costs for Petroleum Fuel for 2022-2025
RIN Cost
Year
($/Gallon)
2022
$0.19
2023
$0.20
2024
$0.20
2025
$0.22
Finally, we calculated RIN values for fuels. For gasoline-ethanol blends, we multiplied
the average D6 RIN price by the ethanol content of each blend (i.e., 10% for E10, 15% for E15,
and an average ethanol content of 74% for E85). For biodiesel and renewable diesel, we
multiplied the average D4 RIN price by the equivalence value of each fuel (i.e., 1.5 for biodiesel
and 1.7 for renewable diesel). The results are shown in Table 10.5.1-3.
440
-------
Table 10.5.1-3: Estimated RIN Va
RIN Value
Fuel
($/Gallon)
E10
$0.15
E15
$0.23
E85
$1.13
Biodiesel
$2.50
Renewable Diesel
$2.83
ues for Fuels
10.5.2 Estimated Fuel Price Impacts (Gasoline)
In this section, we estimate the fuel price impacts of the 2023-2025 candidate volumes on
gasoline relative to the No RFS and 2022 baselines. First we estimated the total cost of gasoline-
ethanol blends for the candidate volumes. We began with the production cost for each fuel,944
added the RIN cost associated with the gasoline portion of the fuel, and then subtracted the RIN
value associated with the ethanol portion of each fuel, which gave us each fuel's net cost per
gallon. We then multiplied each fuel's net cost by its volume from Table 6.5.2-3 to get the total
cost for each fuel. Finally, we calculated the average gasoline cost by dividing the total cost of
all fuels by the total volume of all fuels. As shown in Tables 10.5.2-1 through 3, we estimate that
average gasoline costs range from $2.52-3.05 per gallon.
Table 10.5.2-1: Gasoline <
^osts - 2023 (Candidate Volumes)
E0
E10
E15
E85
Cost to Produce ($/gal)
$3.06
$3.03
$3.13
$3.40
RIN Cost ($/gal)
$0.20
$0.18
$0.17
$0.05
RIN Value ($/gal)
$0.00
-$0.15
-$0.23
-$1.13
Net Cost ($/gal)
$3.26
$3.05
$3.06
$2.32
Volume (mil gal)
2,209
131,015
535
352
Total Fuel Cost ($bil)
$7.2
$399.8
$1.6
$0.8
Average Cost ($/gal)
$3.05
Table 10.5.2-2: Gasoline Costs - 2024 (Candidate Volumes)
E0
E10
E15
E85
Cost to Produce ($/gal)
$2.64
$2.61
$2.71
$2.99
RIN Cost ($/gal)
$0.20
$0.18
$0.17
$0.05
RIN Value ($/gal)
$0.00
-$0.15
-$0.23
-$1.13
Net Cost ($/gal)
$2.84
$2.64
$2.65
$1.91
Volume (mil gal)
2,209
131,744
619
368
Total Fuel Cost ($bil)
$6.3
$347.8
$1.6
$0.7
Average Cost ($/gal)
$2.64
944 Note that for purposes of this fuel price impacts assessment, we only looked at the cost to produce and distribute
fuel to retail stations for sale to consumers (i.e., we subtracted out of the fuel economy cost for each fuel).
441
-------
Table 10.5.2-3: Gasoline <
^osts - 2025 (Candidate Volumes)
E0
E10
E15
E85
Cost to Produce ($/gal)
$2.52
$2.48
$2.57
$2.78
RIN Cost ($/gal)
$0.22
$0.19
$0.18
$0.06
RIN Value ($/gal)
$0.00
-$0.15
-$0.23
-$1.13
Net Cost ($/gal)
$2.74
$2.52
$2.52
$1.70
Volume (mil gal)
2,209
129,601
708
383
Total Fuel Cost ($bil)
$6.0
$326.3
$1.8
$0.7
Average Cost ($/gal)
$2.52
Next we estimated the cost of gasoline-ethanol blends under the No RFS and 2022
baselines. For the No RFS baseline, we began with the production cost for each gasoline-ethanol
blend and multiplied by the volume of each blend under the respective baseline to get the total
cost for each fuel.945 We then calculated the average gasoline cost by dividing the total cost of all
fuels by the total volume of all fuels. As shown in Tables 10.5.2-4 through 6, we estimate that
average gasoline costs under the No RFS baseline range from $2.48-3.03 per gallon.
Table 10.5.2-4: Gasoline <
^osts - 2023 (No RFS Baseline)
E0
E10
E15
E85
Cost to Produce ($/gal)
$3.06
$3.03
$3.13
$3.40
Volume (mil gal)
2,209
133,140
0
0
Total Fuel Cost ($bil)
$6.8
$403.1
$0.0
$0.0
Average Cost ($/gal)
$3.03
Table 10.5.2-5: Gasoline <
^osts - 2024 (No RFS Baseline)
E0
E10
E15
E85
Cost to Produce ($/gal)
$2.64
$2.61
$2.71
$2.99
Volume (mil gal)
2,209
133,970
0
0
Total Fuel Cost ($bil)
$5.8
$349.5
$0.0
$0.0
Average Cost ($/gal)
$2.61
Table 10.5.2-6: Gasoline <
^osts - 2025 (No RFS Baseline)
E0
E10
E15
E85
Cost to Produce ($/gal)
$2.52
$2.48
$2.57
$2.78
Volume (mil gal)
2,209
131,910
0
0
Total Fuel Cost ($bil)
$5.6
$326.6
$0.0
$0.0
Average Cost ($/gal)
$2.48
945 For purposes of the No RFS baseline analysis, we assumed that E0 volumes were held constant relative to the
candidate volumes scenario and that there would not be any volumes of E15 or E85. E10 volumes were calculated
by totaling ethanol production for each year from Table 2.1.5-1 and dividing by 0.1.
442
-------
For the 2022 baseline, we used the same approach described above for the 2023-2025
candidate volumes.946 As shown in Tables 10.5.2-7 through 9, we estimate that average gasoline
costs under the 2022 baseline range from $2.52-3.05 per gallon.
Table 10.5.2-7: Gasoline <
^osts - 2023 (2022 Baseline)
E0
E10
E15
E85
Cost to Produce ($/gal)
$3.06
$3.03
$3.13
$3.40
RIN Cost ($/gal)
$0.20
$0.18
$0.17
$0.05
RIN Value ($/gal)
$0.00
-$0.15
-$0.23
-$1.13
Net Cost ($/gal)
$3.26
$3.05
$3.06
$2.32
Volume (mil gal)
2,128
135,972
440
339
Total Fuel Cost ($bil)
$6.9
$414.9
$1.3
$0.8
Average Cost ($/gal)
$3
05
Table 10.5.2-8: Gasoline <
^osts - 2024 (2022 Baseline)
E0
E10
E15
E85
Cost to Produce ($/gal)
$2.64
$2.61
$2.71
$2.99
RIN Cost ($/gal)
$0.20
$0.18
$0.17
$0.05
RIN Value ($/gal)
$0.00
-$0.15
-$0.23
-$1.13
Net Cost ($/gal)
$2.84
$2.64
$2.65
$1.91
Volume (mil gal)
2,128
135,972
440
339
Total Fuel Cost ($bil)
$6.1
$358.9
$1.2
$0.6
Average Cost ($/gal)
$2
64
Table 10.5.2-9: Gasoline <
^osts - 2025 (2022 Baseline)
E0
E10
E15
E85
Cost to Produce ($/gal)
$2.52
$2.48
$2.57
$2.78
RIN Cost ($/gal)
$0.22
$0.19
$0.18
$0.06
RIN Value ($/gal)
$0.00
-$0.15
-$0.23
-$1.13
Net Cost ($/gal)
$2.74
$2.52
$2.52
$1.70
Volume (mil gal)
2,128
135,972
440
339
Total Fuel Cost ($bil)
$5.8
$342.4
$1.1
$0.6
Average Cost ($/gal)
$2
52
Finally, we calculated the fuel price impacts on gasoline for each year by subtracting the
average gasoline cost for each baseline from the average gasoline cost for the candidate volumes.
As shown in Table 10.5.2-10, we estimate that the fuel price impacts on gasoline under the No
RFS baseline range from 2.4-4.20 per gallon. As shown in Table 10.5.2-11, we estimate that the
fuel price impacts on gasoline under the 2022 baseline range are 0.00 per gallon.
946 2022 baseline gasoline-ethanol blend volumes from 2020-2022 Rule RIA Table 5.5.4-3.
443
-------
Table 10.5.2-10: Gasoline Fuel Price Impacts (No RFS Baseline)
2023
2024
2025
Average Cost (No RFS baseline) ($/gal)
$3.03
$2.61
$2.48
Average Cost (candidate volumes) ($/gal)
$3.05
$2.64
$2.52
Fuel Price Impact (0/gal)
2.40
3.20
4.30
Table 10.5.2-11: Gasoline Fuel Price Impacts (2022
Jaseline)
2023
2024
2025
Average Cost (No RFS baseline) ($/gal)
$3.05
$2.64
$2.52
Average Cost (candidate volumes) ($/gal)
$3.05
$2.64
$2.52
Fuel Price Impact (0/gal)
«...
o
o
«...
o
o
«...
o
o
10.5.3 Estimated Fuel Price Impacts (Diesel)
In this section, we estimate the fuel price impacts of the 2023-2025 candidate volumes on
diesel relative to the No RFS and 2022 baselines. First we estimated the total cost of diesel,
biodiesel, and renewable diesel for the candidate volumes. We began with the production cost for
each fuel,947 and then either added the RIN cost (for diesel) or subtracted the RIN value and tax
credit (for biodiesel and renewable diesel) associated with each fuel, which gave us each fuel's
net cost per gallon. We then multiplied each fuel's net cost by its volume from Table 3.1-4
(biodiesel and renewable diesel) or Preamble Table VII.C-1 (diesel) to get the total cost for each
fuel. Finally, we calculated the average diesel cost by dividing the total cost of all fuels by the
total volume of all fuels. As shown in Tables 10.5.3-1 through 3, we estimate that average diesel
costs range from $3.58-4.08 per gallon.
Table 10.5.3-1: Diesel Costs - 2023 (Candidate Volumes)
Diesel
Biodiesel
Renewable Diesel
Corn
FOG
Soybean/
Canola
Corn
FOG
Soybean/
Canola
Cost to Produce ($/gal)
$3.91
$5.78
$6.95
$7.43
$6.26
$7.50
$8.01
RIN Cost ($/gal)
$0.20
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
RIN Value ($/gal)
$0.00
-$2.50
-$2.50
-$2.50
-$2.83
-$2.83
-$2.83
Tax Credit ($/gal)
$0.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
Net Cost ($/gal)
$4.11
$2.28
$3.45
$3.93
$2.43
$3.67
$4.18
Volume (mil gal)
55,810
115
321
1,274
205
1,169
820
Total Fuel Cost ($bil)
$229.1
$0.3
$1.1
$5.0
$0.5
$4.3
$3.4
Total Cost ($/gal)
$4.08
947 Note that for purposes of this fuel price impacts assessment, we only looked at the cost to produce and distribute
fuel to retail stations for sale to consumers (i.e., we subtracted out of the fuel economy cost for each fuel).
444
-------
Table 10.5.3-2: Diesel Costs - 2024 (Candidate Volumes)
Diesel
Biodiesel
Renewable Diesel
Corn
FOG
Soybean/
Canola
Corn
FOG
Soybean/
Canola
Cost to Produce ($/gal)
$3.59
$5.19
$5.81
$7.18
$5.61
$6.28
$7.74
RIN Cost ($/gal)
$0.20
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
RIN Value ($/gal)
$0.00
-$2.50
-$2.50
-$2.50
-$2.83
-$2.83
-$2.83
Tax Credit ($/gal)
$0.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
Net Cost ($/gal)
$3.79
$1.69
$2.31
$3.68
$1.78
$2.45
$3.91
Volume (mil gal)
52,950
89
303
1,274
239
1,135
853
Total Fuel Cost ($bil)
$200.9
$0.2
$0.7
$4.7
$0.4
$2.8
$3.3
Total Cost ($/gal)
$3.75
Table 10.5.3-3: Diesel Costs - 2025 (Candidate Volumes)
Diesel
Biodiesel
Renewable Diesel
Corn
FOG
Soybean/
Canola
Corn
FOG
Soybean/
Canola
Cost to Produce ($/gal)
$3.40
$4.69
$5.60
$7.24
$5.09
$6.07
$7.82
RIN Cost ($/gal)
$0.22
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
RIN Value ($/gal)
$0.00
-$2.50
-$2.50
-$2.50
-$2.83
-$2.83
-$2.83
Tax Credit ($/gal)
$0.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
Net Cost ($/gal)
$3.62
$1.19
$2.10
$3.74
$1.26
$2.24
$3.99
Volume (mil gal)
52,400
63
285
1,276
272
1,215
1,174
Total Fuel Cost ($bil)
$189.5
$0.1
$0.6
$4.8
$0.3
$2.7
$4.7
Total Cost ($/gal)
$3.58
Next we estimated the total cost of diesel under the No RFS and 2022 baselines. For the
No RFS baseline, we began with the production cost for each fuel and subtracted the tax credit
(for biodiesel and renewable diesel) associated with each fuel, which gave us each fuel's net cost
per gallon. We then multiplied each fuel's net cost by its volume under the respective baseline to
get the total cost for each fuel.948 We then calculated the average diesel cost by dividing the total
cost of all fuels by the total volume of all fuels. As shown in Tables 10.5.3-4 through 6, we
estimate that average diesel costs under the No RFS baseline range from $3.46-3.98 per gallon.
948 Biodiesel and renewable diesel volumes from Table 2.1.5-2. For purposes of the No RFS baseline analysis, we
assumed that total diesel energy demand was held constant relative to the candidate volumes scenario to calculate
petroleum diesel fuel volumes.
445
-------
Table 10.5.3-4: Diesel Costs - 2023
No RFS Baseline)
Diesel
Biodiesel
Renewable Diesel
Corn
FOG
Soybean/
Canola
Corn
FOG
Soybean/
Canola
Cost to Produce ($/gal)
$3.91
$5.78
$6.95
$7.43
$6.26
$7.50
$8.01
Tax Credit ($/gal)
$0.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
Net Cost ($/gal)
$3.91
$4.78
$5.95
$6.43
$5.26
$6.50
$7.01
Volume (mil gal)
57,810
69
422
141
75
1,070
0
Total Fuel Cost ($bil)
$226.0
$0.3
$2.5
$0.9
$0.4
$7.0
$0.0
Total Cost ($/gal)
$3.98
Table 10.5.3-5: Diesel Costs - 2024
(No RFS Baseline)
Biodiesel
Renewable Diesel
Diesel
Corn
FOG
Soybean/
Canola
Corn
FOG
Soybean/
Canola
Cost to Produce ($/gal)
$3.59
$5.19
$5.81
$7.18
$5.61
$6.28
$7.74
Tax Credit ($/gal)
$0.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
Net Cost ($/gal)
$3.59
$4.19
$4.81
$6.18
$4.61
$5.28
$6.74
Volume (mil gal)
54,748
26
395
210
303
1,045
0
Total Fuel Cost ($bil)
$196.5
$0.1
$1.9
$1.3
$1.4
$5.5
$0.0
Total Cost ($/gal)
$3.64
Table 10.5.3-6: Diesel Costs - 2025 (No RFS Baseline)
Diesel
Biodiesel
Renewable Diesel
Corn
FOG
Soybean/
Canola
Corn
FOG
Soybean/
Canola
Cost to Produce ($/gal)
$3.40
$4.69
$5.60
$7.24
$5.09
$6.07
$7.82
Tax Credit ($/gal)
$0.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
Net Cost ($/gal)
$3.40
$3.69
$4.60
$6.24
$4.09
$5.07
$6.82
Volume (mil gal)
54,372
43
398
198
292
1,105
154
Total Fuel Cost ($bil)
$184.9
$0.2
$1.8
$1.2
$1.2
$5.6
$1.1
Total Cost ($/gal)
$3.46
For the 2022 baseline, we used the same approach described above for the 2023-2025
candidate volumes.949 As shown in Tables 10.5.3-7 through 9, we estimate that average diesel
costs under the 2022 baseline range from $3.58-4.08 per gallon.
949 2022 baseline biodiesel and renewable diesel volumes from Table 2.2-2. For purposes of the 2022 baseline
analysis, we assumed that total diesel energy demand was held constant relative to the candidate volumes scenario to
calculate petroleum diesel fuel volumes.
446
-------
Table 10.5.3-7: Diesel Costs - 2023 (2022 Baseline)
Diesel
Biodiesel
Renewable Diesel
Corn
FOG
Soybean/
Canola
Corn
FOG
Soybean/
Canola
Cost to Produce ($/gal)
$3.91
$5.78
$6.95
$7.43
$6.26
$7.50
$8.01
RIN Cost ($/gal)
$0.20
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
RIN Value ($/gal)
$0.00
-$2.50
-$2.50
-$2.50
-$2.83
-$2.83
-$2.83
Tax Credit ($/gal)
$0.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
Net Cost ($/gal)
$4.11
$2.28
$3.45
$3.93
$2.43
$3.67
$4.18
Volume (mil gal)
56,508
130
346
1,262
209
932
293
Total Blend Cost ($bil)
$232.0
$0.3
$1.2
$5.0
$0.5
$3.4
$1.2
Average Cost ($/gal)
$4.08
Table 10.5.3-8: Diesel Costs - 2024 (2022 Baseline)
Diesel
Biodiesel
Renewable Diesel
Corn
FOG
Soybean/
Canola
Corn
FOG
Soybean/
Canola
Cost to Produce ($/gal)
$3.59
$5.19
$5.81
$7.18
$5.61
$6.28
$7.74
RIN Cost ($/gal)
$0.20
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
RIN Value ($/gal)
$0.00
-$2.50
-$2.50
-$2.50
-$2.83
-$2.83
-$2.83
Tax Credit ($/gal)
$0.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
Net Cost ($/gal)
$3.79
$1.69
$2.31
$3.68
$1.78
$2.45
$3.91
Volume (mil gal)
56,508
130
346
1,262
209
932
293
Total Blend Cost ($bil)
$214.4
$0.2
$0.8
$4.6
$0.4
$2.3
$1.1
Average Cost ($/gal)
$3.75
Table 10.5.3-9: Diesel Costs - 2025 (2022 Baseline)
Diesel
Biodiesel
Renewable Diesel
Corn
FOG
Soybean/
Canola
Corn
FOG
Soybean/
Canola
Cost to Produce ($/gal)
$3.40
$4.69
$5.60
$7.24
$5.09
$6.07
$7.82
RIN Cost ($/gal)
$0.22
$0.00
$0.00
$0.00
$0.00
$0.00
$0.00
RIN Value ($/gal)
$0.00
-$2.50
-$2.50
-$2.50
-$2.83
-$2.83
-$2.83
Tax Credit ($/gal)
$0.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
-$1.00
Net Cost ($/gal)
$3.62
$1.19
$2.10
$3.74
$1.26
$2.24
$3.99
Volume (mil gal)
56,508
130
346
1,262
209
932
293
Total Blend Cost ($bil)
$204.3
$0.2
$0.7
$4.7
$0.3
$2.1
$1.2
Average Cost ($/gal)
$3.58
Finally, we calculated the fuel price impacts on diesel for each year by subtracting the
average diesel cost for each baseline from the average diesel cost for the candidate volumes. As
shown in Table 10.5.3-10, we estimate that the fuel price impacts on diesel under the No RFS
447
-------
baseline range from 10.1-11.10 per gallon. As shown in Table 10.5.3-11, we estimate that the
fuel price impacts on diesel under the 2022 baseline range from -0.4-0.00 per gallon.
Table 10.5.3-10: Diesel Fuel Price Impacts (No RFS Baseline)
2023
2024
2025
Average Cost (No RFS baseline) ($/gal)
$3.98
$3.64
$3.46
Average Cost (candidate volumes) ($/gal)
$4.08
$3.75
$3.58
Fuel Price Impact (0/gal)
10.10
10.10
11.10
Table 10.5.3-11: Diesel Fuel Price Impacts (2022 Baseline)
2023
2024
2025
Average Cost (2022 baseline) ($/gal)
$4.08
$3.75
$3.58
Average Cost (candidate volumes) ($/gal)
$4.08
$3.75
$3.58
Fuel Price Impact (0/gal)
o
o
-0.40
-0.10
10.5.4 Cost to Transport Goods
In this section, we consider the impact of the use of renewable fuels on the cost to
transport goods. Since most goods being transported utilize diesel fuel powered trucks (as
opposed to gasoline or natural gas vehicles), we focus on the impacts on diesel fuel prices.
Reviewing the price estimates in Table 10.5.3-10, the projected price increase for diesel fuel
relative to the No RFS baseline ranged from 10.10 per gallon in 2024 to 11.10 per gallon in
2025. As a worst case scenario, we will use the proj ected diesel fuel price increase of 11.10 per
gallon for estimating the impact on the cost to transport goods.
The impact of fuel price increases on the price of goods is based upon a study conducted
by USD A. USD A analyzed the impact of fuel prices on the wholesale price of produce from
2000 to 2009 when fuel prices ramped up because crude oil prices increased from an average of
$30 per barrel to over $90 per barrel.950 Their study found that a 100% increase in fuel prices
resulted in a 25% increase in produce prices. Assuming a baseline diesel fuel retail price of
$3.40/gal in 2025 as summarized in Table 10.2.2-1 and adding 600 per gallon state and federal
taxes to it, the projected 11.10 per gallon increase in diesel fuel price amounts to a 2.8 percent
increase in diesel fuel prices. Applying the 25% ratio from the USDA study would indicate that
the 2025 candidate volumes incremental to the No RFS baseline would then increase the
wholesale price of produce by about 0.7%. If produce being transported by a diesel truck costs
$3 per pound, the increase in that products' price due to the projected impact of the candidate
volumes would be $0.02 per pound.951 Transport of food by other means such as rail or barge
would be expected to impact food prices less than transport by truck since rail and barge
transport are both more efficient and fuel costs would likely have a lower impact those modes of
transportation costs. This estimate of the impact on food prices is only an order of magnitude
950 Volpe, Richard; How Transportation Costs Affect Fresh Fruit and Vegetable Prices; United States Department of
Agriculture; November 2013.
951 Comparing Prices on Groceries; May 4, 2021:
https://web.arc hive.org/web/202 M)805005209/fattps://www.eoiipons.eom/thegoodstiiff/eomparing~priees~on~
groceries.
448
-------
type estimate since impacts on food prices vary greatly depending on the distance that the
particular food travels by truck.
10.6 Analysis of Alternative Scenarios
In the proposed rule EPA requested comment on alternative scenarios, including reducing
the required total renewable fuel volume by 0.25 billion gallons in 2024 and 2025 and reducing
the implied conventional renewable fuel volume to below the E10 blendwall. While the volumes
we are finalizing in this rule do not reflect these alternative scenarios, for purposes of this RIA
we have analyzed the impact these alternative volume scenarios would have on costs, energy
security benefits, and fuel prices. The projected impacts of these alternative scenarios are
summarized in the following sections.
10.6.1 Alternative Scenarios Descriptions
Prior to projecting the impacts of the alternative scenarios, we first had to define these
scenarios and project the types of renewable fuel that would be used to satisfy the volume
requirements.
10.6.1.1 Alternative Scenario 1
The first alternative scenario we considered is one in which we finalize total renewable
fuel volumes for 2024 and 2025 that are 250 million gallons lower than the volumes we are
finalizing in this rule. In this scenario we also reduced the advanced biofuel volumes for 2024
and 2025 by 250 million gallons to maintain an implied conventional renewable fuel volume of
15.0 billion gallons for these years, with corresponding reductions to the BBD volume
requirement for each year. For this scenario we project that the volume of renewable diesel
produced from canola oil in 2024 and 2025 would be 147 million gallons (equivalent to 250
million RINs) less than the volumes supplied to meet the volume requirements we are finalizing
for 2024 and 2025. We projected that the reduction in the total renewable fuel volume would
result in less renewable diesel produced from canola oil because this is projected to be the most
expensive non-cellulosic biofuel supplied in these years.
Tables 10.6.1.1-1 through 3 summarize the renewable fuel volume requirements for this
scenario, the renewable fuel volumes we project would be supplied to meet the required
volumes, and the change in the supply of renewable fuels relative to the volumes we are
finalizing in this rule.
449
-------
Table 10.6.1.1-1: RFS Volume Requirements for Alternative Scenario 1 (Million RINs)
2023
2024
2025
Cellulosic Biofuel
0.84
1.09
1.38
BBDa
2.82
2.89
3.20
Advanced Biofuel
5.94
6.29
7.08
Total Renewable Fuel
20.94
21.29
22.08
Implied Conventional Renewable Fuel
15.00
15.00
15.00
Supplemental Volume Requirement
0.25
0.00
0.00
a In million gallons rather than million RINs.
Table 10.6.1.1-2: Renewable Fuel Volumes
'or Alternal
ive Scenario 1 (Million
2023
2024
2025
Cellulosic Biofuel
838
1,090
1,376
CNG/LNG from biogas
831
1,039
1,299
Ethanol from CKF
7
51
77
Total Biomass-Based Diesela
5,965
5,955
6,631
Biodiesel
2,565
2,500
2,436
Soybean oil
1,473
1,451
1,430
FOG
481
454
427
Corn oil
173
134
95
Canola oil
438
461
484
Renewable Diesel
3,376
3,431
4,171
Soybean oil
777
1,141
1,501
FOG
1,883
1,825
1,962
Corn oil
348
406
463
Canola oil
368
59
245
Jet fuel from FOG
24
24
24
Other Advanced Biofuels
290
290
290
Renewable diesel from FOG
104
104
104
Imported sugarcane ethanol
95
95
95
Domestic ethanol from waste ethanol
27
27
27
Otherb
64
64
64
Conventional Renewable Fuel
13,845
13,955
13,779
Ethanol from corn
13,845
13,955
13,779
Renewable diesel from palm oil
0
0
0
Supplemental Standard
250
0
0
Biodiesel from Soybean Oil
250
0
0
a Includes BBD in excess of the candidate volume for advanced biofuel. The excess would be used to help meet the
candidate volume for conventional renewable fuel.
b Composed of non-cellulosic biogas, heating oil, and naphtha.
450
-------
Table 10.6.1.1-3: Changes in Renewable Fuel Volumes for Alternative Scenario 1 (Million
RINs)a
2023
2024
2025
Canola Oil Renewable Diesel
0
-250
-250
Total
0
-250
-250
a Volume changes are relative to the volumes we project will be supplied to meet the volume requirements we are
finalizing in this rule. The projected volume changes for fuels not listed in this table are zero for all three years.
10.6.1.2 Alternative Scenario 2
The second alternative scenario we considered was a scenario in which we further
reduced the total renewable fuel volume requirement for each year such that the implied
conventional biofuel volume requirement was 13.0 billion gallons each year—a level considered
sufficiently low for the market to have confidence that it could be met with just E10. This
amounts to a 2.0 billion gallon reduction in the total renewable fuel volume requirement in
2023-2025 (relative to Alternative Scenario 1).
For this scenario we project that the contribution of corn ethanol towards meeting the
renewable fuel volume requirements would be limited to 13.0 billion gallons each year. We
project that obligated parties would continue to blend ethanol up to the El0 blendwall because it
is economical to do so, but that the quantity of conventional RINs obligated parties can use to
meet their RFS obligations would be limited to 13.0 billion gallons based on the difference
between the advanced biofuel and total renewable fuel standards.952
With the reduction in the conventional ethanol volumes to 13.0 billion gallons, less
advanced biofuel volume would then be necessary to backfill for conventional biofuel to meet
the total renewable fuel standard under this scenario. As such, we next reduced the projected
volumes of renewable diesel from canola oil, renewable diesel from soybean oil, and biodiesel
from canola oil (in that order) until we achieved the intended reduction in the renewable fuel
supply each year. As noted above, renewable diesel produced from canola oil is the most
expensive non-cellulosic biofuel supplied in these years, and renewable diesel produced from
soybean oil is similarly priced. For 2023, even after reducing these fuel types to zero further
reductions are necessary.
We next reduced biodiesel produced from canola oil, as we project that the incentive
provided by California's LCFS program and other similar state programs would be large enough
to prevent any reductions in the supply of renewable diesel produced from FOG or distillers corn
oil. Tables 10.6.1.2-1 through 3 summarize the renewable fuel volume requirements for this
scenario, the renewable fuel volumes we project would be supplied to meet the required
volumes, and the change in the supply of renewable fuels relative to the volumes we are
finalizing in this rule.
952 Additional ethanol would be expected to be blended as E10 based on its favorable economics up to the E10
blendwall. RINs associated for ethanol beyond the 13.0 billion gallon implied conventional biofuel volume each
year generally could not be used to meet the required volumes for that year, but could be used to satisfy deficits
from the previous year or carried over to be used in the next year.
451
-------
Table 10.6.1.2-1: RFS Volume Requirements for Alternative Scenario 2 (Million RINs)
2023
2024
2025
Cellulosic Biofuel
0.84
1.09
1.38
BBDa
2.82
2.89
3.20
Advanced Biofuel
5.94
6.29
7.08
Total Renewable Fuel
18.94
19.29
20.08
Implied Conventional Renewable Fuel
13.00
13.00
13.00
Supplemental Volume Requirement
0.25
0.00
0.00
a In million gallons rather than million RINs
Table 10.6.1.2-2: Renewable Fuel Volumes
'or Alternal
ive Scenario 2 (Million
2023
2024
2025
Cellulosic Biofuel
838
1,090
1,376
CNG/LNG from biogas
831
1,039
1,299
Ethanol from CKF
7
51
77
Total Biomass-Based Diesela
4,810
4,910
5,410
Biodiesel
2,555
2,500
2,436
Soybean oil
1,473
1,451
1,430
FOG
481
454
427
Corn oil
173
134
95
Canola oil
428
461
484
Renewable Diesel
2,231
2,386
2,950
Soybean oil
0
155
525
FOG
1,883
1,825
1,962
Corn oil
348
406
463
Canola oil
0
0
0
Jet fuel from FOG
24
24
24
Other Advanced Biofuels
290
290
290
Renewable diesel from FOG
104
104
104
Imported sugarcane ethanol
95
95
95
Domestic ethanol from waste ethanol
27
27
27
Otherb
64
64
64
Conventional Renewable Fuel
13,000
13,000
13,000
Ethanol from corn
13,000
13,000
13,000
Renewable diesel from palm oil
0
0
0
Supplemental Standard
250
0
0
Biodiesel from Soybean Oil
250
0
0
a Includes BBD in excess of the candidate volume for advanced biofuel. The excess would be used to help meet the
candidate volume for conventional renewable fuel.
b Composed of non-cellulosic biogas, heating oil, and naphtha.
452
-------
Table 10.6.1.2-3: Changes in Renewable Fuel Volumes for Alternative Scenario 2 (Million
RINs)a
2023
2024
2025
Canola Oil Biodiesel
-10
0
0
Soybean Oil Renewable Diesel
-777
-986
-976
Canola Oil Renewable Diesel
-368
-309
-495
Corn Ethanol
-845
-955
-779
Total
-2,000
-2,250
-2,250
a Volume changes are relative to the volumes we project will be supplied to meet the volume requirements we are
finalizing in this rule. The projected volume changes for fuels not listed in this table are zero for all three years. The
reduced vegetable oil volume is estimated to cause a reduction in soy oil price; for example, the 2024 soy oil price is
estimated to decrease from 80c/lb to 57c/lb under this alternative scenario.
10.6.1.3 Alternative Scenario 3
The third alternative scenario we considered is a scenario in which we increased the
advanced biofuel volumes by 2.0 billion RINs each year for 2023-2025 relative to Alternative
Scenario 1, with corresponding increases to the BBD volume requirement for each year, but with
no change to the total renewable fuel volume. This has the effect of decreasing the implied
conventional renewable fuel volume by 2.0 billion RINs each year (to a volume below the E10
blendwall).
For this scenario we project that the contribution of corn ethanol towards meeting the
renewable fuel volume requirements would be limited to 13.0 billion gallons each year, as in
Alternative Scenario 2. However, because the total renewable fuel volume requirement is only
reduced by 0.25 billion gallons in 2024 and 2025 (and not at all in 2023), additional volumes of
renewable fuel are needed to offset the reduction in corn ethanol supplied. We project that this
additional renewable fuel would be renewable diesel produced from soybean oil.
We project that in this scenario renewable diesel would be the most likely fuel type to
increase due to the excess production capacity and the relative lack of infrastructure constraints
or challenges associated with increasing the use of renewable diesel in the U.S. We further
project that this renewable diesel would be produced from soybean oil, as we project that
renewable diesel producers would have already utilized all available sources of FOG and
distillers corn oil due to the relatively high value of fuels produced from these feedstocks in
California's LCFS program.953 The additional soybean oil would be expected to be sourced from
imports and diversion from other domestic uses given that the finalized volumes are already
expected to utilize all available increases in North American feedstock supplies. Tables 10.6.1.3-
1 through 3 summarize the renewable fuel volume requirements for this scenario, the renewable
fuel volumes we project would be supplied to meet the required volumes, and the change in the
supply of renewable fuels relative to the volumes we are finalizing in this rule.
953 The additional volume of renewable diesel could also be produced from canola oil. We note that in our cost and
fuel price impact analyses we project that renewable diesel produced from canola oil and soybean oil have the same
costs, and therefore the projected impacts of this scenario would be the same whether this fuel was produced from
soybean oil or canola oil.
453
-------
Table 10.6.1.3-1: RFS Volume Requirements for Alternative Scenario 3 (Million RINs)
2023
2024
2025
Cellulosic Biofuel
0.84
1.09
1.38
BBDa
4.07
4.14
4.45
Advanced Biofuel
7.94
8.29
9.08
Total Renewable Fuel
20.94
21.29
22.08
Implied Conventional Renewable Fuel
13.00
13.00
13.00
Supplemental Volume Requirement
0.25
0.00
0.00
a In million gallons rather than million RINs
Table 10.6.1.3-2: Renewable Fuel Volumes
'or Alternal
ive Scenario 3 (Million
2023
2024
2025
Cellulosic Biofuel
838
1,090
1,376
CNG/LNG from biogas
831
1,039
1,299
Ethanol from CKF
7
51
77
Total Biomass-Based Diesela
6,810
6,910
7,410
Biodiesel
2,565
2,500
2,436
Soybean oil
1,473
1,451
1,430
FOG
481
454
427
Corn oil
173
134
95
Canola oil
438
461
484
Renewable Diesel
4,221
4,386
4,950
Soybean oil
1,622
1,846
2,030
FOG
1,883
1,825
1,962
Corn oil
348
406
463
Canola oil
368
309
495
Jet fuel from FOG
24
24
24
Other Advanced Biofuels
290
290
290
Renewable diesel from FOG
104
104
104
Imported sugarcane ethanol
95
95
95
Domestic ethanol from waste ethanol
27
27
27
Otherb
64
64
64
Conventional Renewable Fuel
13,000
13,000
13,000
Ethanol from corn
13,000
13,000
13,000
Renewable diesel from palm oil
0
0
0
Supplemental Standard
250
0
0
Biodiesel from Soybean Oil
250
0
0
a Includes BBD in excess of the candidate volume for advanced biofuel. The excess would be used to help meet the
candidate volume for conventional renewable fuel.
b Composed of non-cellulosic biogas, heating oil, and naphtha.
454
-------
Table 10.6.1.3-3: Changes in Renewable Fuel Volumes for Alternative Scenario 3 (Million
RINs)a
2023
2024
2025
Soybean Oil Renewable Diesel
+845
+705
+529
Corn Ethanol
-845
-955
-779
Total
0
-250
-250
a Volume changes are relative to the volumes we project will be supplied to meet the volume requirements we are
finalizing in this rule. The projected volume changes for fuels not listed in this table are zero for all three years. The
increased vegetable oil volume is estimated to cause an increase in soy oil price; for example, the 2024 soy oil price
is estimated to increase from 80c/lb to 93c/lb under this alternative scenario.
10.6.2 Cost Impacts of Alternative Scenarios
After determining the renewable fuel volumes projected to be used to meet the three
alternative scenarios, we next projected the costs of each alternative scenario. The methodology
used to estimate the costs of each scenario is identical to the methodology used to project the
costs of the volumes we are finalizing in this rule, discussed in greater detail in RIA Chapters
10.1 through 10.4. For these alternative scenarios we only projected costs relative to the No RFS
baseline and did not also calculate costs relative to the 2022 baseline. The projected costs for
each of the alternative scenarios are shown in Tables 10.6.2-1 through 3.954
Table 10.6.2-1: Alternative Scenario 1 Cost Relative to
the No RFS baseline (2022$)
Total Cost
(million $)
Per-Unit
Cost
Units
2023
Gasoline
445
0.33
0/gal gasoline
Diesel Fuel
7,610
13.56
0/gal diesel
Natural Gas
55
0.1753
$/1000 FT3 natural gas
Total
8,110
4.26
0/gal gasoline and diesel
2023 with
Suppl. Std.
Gasoline
445
0.33
0/gal gasoline
Diesel Fuel
8,238
14.68
0/gal diesel
Natural Gas
55
0.1753
$/1000 FT3 natural gas
Total
8,738
4.59
0/gal gasoline and diesel
2024
Gasoline
440
0.33
0/gal gasoline
Diesel Fuel
6,199
11.62
0/gal diesel
Natural Gas
137
0.4549
$/1000 FT3 natural gas
Total
6,777
3.60
0/gal gasoline and diesel
2025
Gasoline
458
0.34
0/gal gasoline
Diesel Fuel
7,156
13.53
0/gal diesel
Natural Gas
228
0.7645
$/1000 FT3 natural gas
Total
7,842
4.22
0/gal gasoline and diesel
954 More details on these cost projections can be found in the technical memorandum, "Cost Projections for
Alternative Scenarios," available in the docket for this action.
455
-------
Table 10.6.2-2: Alternative Scenario 2 Cost Relative to
the No RFS baseline (2022$)
Total Cost
(million $)
Per-Unit
Cost
Units
2023
Gasoline
0
0.00
0/gal gasoline
Diesel Fuel
4,075
7.26
0/gal diesel
Natural Gas
55
0.1753
$/1000 FT3 natural gas
Total
4,130
2.17
0/gal gasoline and diesel
2023 with
Suppl. Std.
Gasoline
0
0.00
0/gal gasoline
Diesel Fuel
4,615
8.23
0/gal diesel
Natural Gas
55
0.1753
$/1000 FT3 natural gas
Total
4,670
2.45
0/gal gasoline and diesel
2024
Gasoline
0
0.00
0/gal gasoline
Diesel Fuel
2,379
4.46
0/gal diesel
Natural Gas
137
0.4549
$/1000 FT3 natural gas
Total
2,516
1.34
0/gal gasoline and diesel
2025
Gasoline
0
0.00
0/gal gasoline
Diesel Fuel
1,984
3.75
0/gal diesel
Natural Gas
228
0.7645
$/1000 FT3 natural gas
Total
2,212
1.19
0/gal gasoline and diesel
Table 10.6.2-3: Alternative Scenario 3 Cost Relative to
the No RFS baseline (2022$)
Total Cost
(million $)
Per-Unit
Cost
Units
2023
Gasoline
0
0.00
0/gal gasoline
Diesel Fuel
10,929
19.48
0/gal diesel
Natural Gas
55
0.1753
$/1000 FT3 natural gas
Total
10,984
5.77
0/gal gasoline and diesel
2023 with
Suppl. Std.
Gasoline
0
0.00
0/gal gasoline
Diesel Fuel
11,653
20.77
0/gal diesel
Natural Gas
55
0.1753
$/1000 FT3 natural gas
Total
11,708
6.15
0/gal gasoline and diesel
2024
Gasoline
0
0.00
0/gal gasoline
Diesel Fuel
10,566
19.81
0/gal diesel
Natural Gas
137
0.4549
$/1000 FT3 natural gas
Total
10,703
5.68
0/gal gasoline and diesel
2025
Gasoline
0
0.00
0/gal gasoline
Diesel Fuel
11,960
22.61
0/gal diesel
Natural Gas
228
0.7645
$/1000 FT3 natural gas
Total
12,188
6.55
0/gal gasoline and diesel
10.6.3 Fuel Price Impacts of Alternative Scenarios
In addition to the costs of the alternative scenarios, we also projected the fuel price
impacts of each alternative scenario. The methodology used to estimate the fuel price impacts of
456
-------
each scenario is identical to the methodology used to project the fuel price impacts of the
volumes we are finalizing in this rule, discussed in greater detail in RIA Chapter 10.5. As with
the impacts on costs, we only projected fuel price impacts of the alternative scenarios relative to
the No RFS baseline and did not also calculate fuel price impacts relative to the 2022 baseline.
The projected fuel price impacts for each of the alternative scenarios are shown in Tables 10.6.3-
1 through 3.955
Table 10.6.3-1: Alternative Scenario 1 Fuel Price Impacts (No RFS Baseline)
Gasoline
Diesel
2023
2024
2025
2023
2024
2025
Average Cost (No RFS baseline)
($/gal)
$3.03
$2.61
$2.48
$3.98
$3.65
$3.46
Average Cost (candidate volumes)
($/gal)
$3.05
$2.64
$2.52
$4.08
$3.74
$3.57
Fuel Price Impact
(0/gal)
2.40
2.90
4.00
10.10
9.80
10.70
Table 10.6.3-2: Alternative Scenario 2 Fuel Price Impacts (No RFS Baseline)
Gasoline
Diesel
2023
2024
2025
2023
2024
2025
Average Cost (No RFS baseline)
($/gal)
$3.03
$2.61
$2.48
$3.98
$3.64
$3.45
Average Cost (candidate volumes)
($/gal)
$3.08
$2.67
$2.55
$3.94
$3.59
$3.39
Fuel Price Impact
(0/gal)
5.50
6.00
7.00
-3.50
-5.10
-6.00
Table 10.6.3-3: Alternative Scenario 3 Fuel Price Impacts (No RFS Baseline)
Gasoline
Diesel
2023
2024
2025
2023
2024
2025
Average Cost (No RFS baseline)
($/gal)
$3.03
$2.61
$2.48
$3.98
$3.65
$3.47
Average Cost (candidate volumes)
($/gal)
$3.10
$2.69
$2.56
$3.99
$3.67
$3.50
Fuel Price Impact
(0/gal)
7.20
7.80
8.70
1.20
1.80
3.40
10.6.4 Energy Security Benefits of Alternative Scenarios
We next projected the energy security benefits of each alternative scenario. The
methodology used to estimate the energy security benefits of each scenario is identical to the
methodology used to project the energy security benefits of the volumes we are finalizing in this
955 More details on these fuel price impacts can be found in the technical memorandum, "Fuel Price Impacts of
Alternative Scenarios," available in the docket for this action.
457
-------
rule, discussed in greater detail in RIA Chapter 5. The projected energy security benefits for each
of the alternative scenarios are shown in Tables 10.6.4-1 through 3.956
Table 10.6.4-1: Annual Energy Security I
benefits of Alternative Scenario 1
Year
Net Crude Oil
Import Reductions
(millions of gallons)
Benefits
(millions of 2022$)
2023
Excluding Supplemental Standard
Including Supplemental Standard
2,012
2,151
$180
$192
2024
1,820
$161
2025
2,001
$175
Table 10.6.4-2: Annual Energy Security I
benefits of Alternative Scenario 2
Year
Net Crude Oil
Import Reductions
(millions of gallons)
Benefits
(millions of 2022$)
2023
Excluding Supplemental Standard
Including Supplemental Standard
1,215
1,354
$109
$121
2024
1,076
$95
2025
1,151
$101
Table 10.6.4-3: Annual Energy Security I
benefits of Alternative Scenario 3
Year
Net Crude Oil
Import Reductions
(millions of gallons)
Benefits
(millions of 2022$)
2023
Excluding Supplemental Standard
Including Supplemental Standard
2,317
2,457
$207
$219
2024
2,180
$192
2025
2,257
$197
10.6.5 Consideration of Other Impacts of the Alternative Scenarios
The discussion above of the impacts of the alternative scenarios only addresses costs, fuel
prices, and energy security and does not include all the other factors also evaluated with respect
to the final volumes. We note, however, that the impact of biofuel production on each of the
statutory factors is generally proportional to the quantity of biofuel produced and used. We can
therefore approximate the expected impact of each alternative scenario by considering the
changes in the biofuel volumes projected to be used to meet the alternative scenario.
956 More details on these energy security benefits can be found in the technical memorandum, "Energy Security
Projections for Alternative Scenarios," available in the docket for this action.
458
-------
In this section we present two sets of volume changes for each of the scenarios. The first
is simply relative to the total volume of each fuel type we project will be supplied to meet the
volumes we are finalizing in this rule (Table 10.6.5-1). For example, in Alternative Scenario 1
we expect that total renewable fuel use would be approximately 1% lower in 2024 and 2025
relative to the total renewable fuel volumes projected to be used to meet the volumes we are
finalizing in this rule. We could therefore reasonably expect that employment at biofuel
production facilities under this scenario may decrease by approximately 1%. The second set of
volume changes considered is the change in volumes attributable to the rule (i.e., relative to the
projected volume increases from the No-RFS baseline) for the alternative scenario in comparison
to the change in volumes attributable to the volumes we are finalizing in this rule (Table 10.6.5-
2). We project, for example, that the incremental total renewable fuel volume use attributable to
the RFS program would decrease by about 5% in 2024 and 2025 in Alternative Scenario 1
relative to the volumes we are finalizing in this rule. Based on this information we could
reasonably project that the impacts of this rule on some of the unquantified statutory factors
would be approximately 5% lower under Alternative Scenario 1 relative to the volumes we are
finalizing in this rule. We further note that the projected impacts on particular fuel types are
much larger than the projected impacts on total renewable fuel use, and that the projected
impacts for Alternative Scenario 2 are much larger than the projected impacts of Alternative
Scenarios 1 and 3.
Table 10.6.5-1: Volume Changes of the Alternative Scenarios Relative to the Total Biofuel
Volumes Finalized (million RINs)
2023
2024
2025
Alternative Scenario 1
Canola Oil Renewable Diesel
0.0%
-80.9%
-50.5%
Total
0.0%
-1.2%
-1.1%
Alternative Scenario 2
Canola Oil Biodiesel
-2.3%
0.0%
0.0%
Soybean Oil Renewable Diesel
-100%
-86.4%
-65.0%
Canola Oil Renewable Diesel
-100%
-100%
-100%
Corn Ethanol
-6.1%
-6.8%
-5.7%
Total
-9.4%
-10.4%
-10.1%
Alternative Scenario 3
Soybean Oil Renewable Diesel
+108.8%
+61.8%
+35.2%
Corn Ethanol
-6.1%
-6.8%
-5.7%
Total
0.0%
-1.2%
-1.1%
459
-------
Table 10.6.5-2: Change in Volumes Attributable to the RFS program for the Alternative
2023
2024
2025
Alternative Scenario 1
Canola Oil Renewable Diesel
0.0%
-80.9%
-50.5%
Total
0.0%
-5.5%
-4.8%
Alternative Scenario 2
Canola Oil Biodiesel
-2.3%
0.0%
0.0%
Soybean Oil Renewable Diesel
-100%
-86.4%
-78.7%
Canola Oil Renewable Diesel
-100%
-100%
-100%
Corn Ethanol
-100%
-100%
-100%
Total
-43.1%
-49.5%
-43.5%
Alternative Scenario 3
Soybean Oil Renewable Diesel
+206.6%
+106.7%
+41.2%
Corn Ethanol
-100%
-100%
-100%
Total
0.0%
-5.5%
-4.8%
10.6.6 Summary of the Impacts of the Alternative Scenarios
Finally, we summed the projected fuel costs and energy security benefits of each
alternative scenario. The costs and energy security benefits for each scenario are described in
RIA Chapters 10.6.2 and 10.6.4. The projected costs and energy security benefits for each of the
alternative scenarios are shown in Tables 10.6.6-1 through 3. Table 10.6.6-4 shows the costs and
energy security benefits of the volumes we are finalizing in this rule and each of the three
alternative scenarios for comparison.957
Table 10.6.6-1: Cumulative Monetized Fuel Costs and Energy Security Benefits of
Discount Rate
3%
7%
Excluding Supplemental Standard
Fuel Costs
$22,081
$21,293
Energy Security Benefits
$501
$483
Including Supplemental Standard
Fuel Costs
$22,709
$21,921
Energy Security Benefits
$513
$495
957 More details on these discounted costs and energy security benefits can be found in the technical memorandum,
"Summary Cost and Energy Security Projections for Alternative Scenarios," available in the docket for this action.
460
-------
Table 10.6.6-2: Cumulative Monetized Fuel Costs and Energy Security Benefits of
Discount Rate
3%
7%
Excluding Supplemental Standard
Fuel Costs
$8,658
$8,413
Energy Security Benefits
$296
$285
Including Supplemental Standard
Fuel Costs
$9,198
$8,953
Energy Security Benefits
$308
$298
Table 10.6.6-3: Cumulative Monetized Fuel Costs and
Alternative Scenario 3 with Respect to the No RFS Ba
Discount Rate
3%
7%
Excluding Supplemental Standard
Fuel Costs
$32,864
$31,632
Energy Security Benefits
$580
$559
Including Supplemental Standard
Fuel Costs
$33,588
$32,356
Energy Security Benefits
$592
$572
Table 10.6.6-4: Cumulative Monetized Fuel Costs and Energy Security Benefits of the Final
Volumes and Alternative Scenarios with Respect to the No RFS Baseline (2022$, millions,
Final
Volumes
Alternative
Scenario 1
Alternative
Scenario 2
Alternative
Scenario 3
Fuel Costs
$23,846
$22,709
$9,198
$33,588
Energy Security Benefits
$536
$513
$308
$592
461
-------
Chapter 11: Screening Analysis
This chapter discusses EPA's screening analysis evaluating the potential impacts of the
RFS standards for 2023, 2024, and 2025 on small entities. The Regulatory Flexibility Act (RFA),
as amended by the Small Business Regulatory Enforcement Fairness Act of 1996 (SBREFA),
generally requires an agency to prepare a regulatory flexibility analysis of any rule subject to
notice and comment rulemaking requirements under the Administrative Procedure Act or any
other statute, unless the agency certifies that the rule will not have a significant economic impact
on a substantial number of small entities (referred to as a "No SISNOSE finding"). Pursuant to
this requirement, EPA has prepared a screening analysis for this rule.
We conducted the screening analyses by looking at the potential impacts on small entities
and compared the cost-to-sales ratio to a threshold of 1%.958 Specifically, we compared obligated
parties' cost of compliance (whether they acquire RINs by purchasing renewable fuels with
attached RINs and blending these fuels into transportation fuel or by purchasing separated RINs)
with the ability for the obligated parties to recover these compliance costs through higher prices
for the gasoline and diesel they sell with what would be expected in the absence of the RFS
program. Based on our recent analysis of the data, we have determined that all obligated
parties—including small refiners—fully recover the costs of RFS compliance through higher
sales prices on gasoline and diesel.959 Given this, the cost-to-sales ratio of this rule is less than
1%. Therefore, EPA finds that these standards would not have a significant economic impact on
a substantial number of small entities.
11.1 Background
11.1.1 Overview of the Regulatory Flexibility Act (RFA)
The RFA was amended by SBREFA to ensure that concerns regarding small entities are
adequately considered during the development of new regulations that affect those entities. The
RFA requires us to carefully consider the economic impacts that our rules may have on small
entities. The elements of the initial regulatory flexibility analysis accompanying a proposed rule
are set forth in 5 U.S.C. § 603, while those of the final regulatory flexibility analysis
accompanying a final rule are set forth in section 604. However, section 605(b) of the statute
provides that EPA need not conduct the section 603 or 604 analyses if we certify that the rule
will not have a significant economic impact on a substantial number of small entities.
11.1.2 Need for the Rulemaking and Rulemaking Objectives
A discussion on the need for and objectives of this action is located in Preamble Section
I. CAA section 21 l(o) requires EPA to promulgate regulations implementing the RFS program,
958 A cost-to-sales ratio of 1% represents a typical agency threshold for determining the significance of the economic
impact on small entities. See "Final Guidance for EPA Rulewriters: Regulatory Flexibility Act as amended by the
Small Business Regulatory Enforcement Fairness Act," November 2006.
959 See "April 2022 Denial of Petitions for RFS Small Refinery Exemptions," EPA-420-R-22-005, April 2022. See
also "June 2022 Denial of Petitions for RFS Small Refinery Exemptions," EPA-420-R-22-011, June 2022.
462
-------
and to annually establish renewable fuel standards that are used by obligated parties to determine
their individual RVOs.
11.1.3 Definition and Description of Small Entities
Small entities include small businesses, small organizations, and small governmental
jurisdictions. For the purposes of assessing the impacts of a rule on small entities, a small entity
is defined as: (1) a small business according to the Small Business Administration's (SBA) size
standards; (2) a small governmental jurisdiction that is a government of a city, county, town,
school district or special district with a population of less than 50,000; or (3) a small organization
that is any not-for-profit enterprise that is independently owned and operated and is not dominant
in its field.
Small businesses (as well as large businesses) would be regulated by this rule, but not
small governmental jurisdictions or small organizations as described above. As set by SBA, the
categories of small entities that would potentially be directly affected by this rulemaking are
described in Table 11.2.3-1.
Table 11.2.3-1: Small Business Definitions
Industry
Defined as small entity by
SBA if less than or equal to:
NAICS3 code
Gasoline and diesel refiners
1,500 employees'3
324110
a North American Industrial Classification System.
b EPA has included in past fuels rulemakings a provision that, in order to qualify for small refiner flexibilities, a
refiner must also produce no greater than 155,000 barrels per calendar day (bpcd) crude capacity. See 40 CFR
80.1442(a).
EPA used the criteria for small entities developed by SBA under the North American
Industry Classification System (NAICS) as a guide. Information about the characteristics of
refiners comes from sources including EIA, oil industry literature, and previous rules that have
affected the refining industry. In addition, EPA found employment information for companies
meeting the SBA definition of "small entity" using the business information database Hoover's
Inc. (a subsidiary of Dun & Bradstreet). These refiners fall under the Petroleum Refineries
category, 324110, as defined by NAICS.
Small entities that would be subject to this rulemaking include domestic refiners that
produce gasoline and/or diesel. Based on 2022 EIA refinery data,960 EPA believes that there are
about 35-40 refiners of gasoline and diesel subject to the RFS regulations. Of these, EPA
believes that there are currently 7 refiners (owning 9 refineries) producing gasoline and/or diesel
that meet the small entity definition of having 1,500 employees or fewer.
960 EIA Refinery Capacity Report. https://www.eia.gov/petroleum/refinerycapacity/archive/2022/refcap2Q22.php.
463
-------
11.1.4 Reporting, Recordkeeping, and Other Compliance Requirements
Registration, reporting, and recordkeeping are necessary to track compliance with the
RFS standards and transactions involving RINs. However, these requirements are already in
place under the existing RFS regulations.961 While EPA is making revisions to the RFS
requirements in this action, we do not anticipate that there will be any significant cost on directly
regulated small entities.
11.2 Screening Analysis Approach
We believe the most appropriate way to consider the impacts of the 2023-2025 RFS
standards on obligated parties is to compare their cost of compliance with the ability for the
obligated parties to recover these compliance costs through the higher prices for the gasoline and
diesel they sell that result from the market-wide impact of the RFS program. EPA has
determined that while there is a cost to all obligated parties to acquire RINs (including small
refiners), obligated parties recover that cost through the higher sales prices they receive for the
gasoline and diesel they sell due to the market-wide impact of the RFS standards on these
products.962 EPA has examined available market data and concluded that the costs of compliance
with the RFS program are being passed downstream, as current wholesale gasoline and diesel
prices enable obligated parties to recover the cost of the RINs.963 When viewed in light of this
data, there is no net cost of compliance with the RFS standards (cost of compliance with the RFS
standards minus the increased revenue due to higher gasoline and diesel prices that result from
implementing the RFS program) to obligated parities, including small refiners. This is true
whether obligated parties acquire RINs by purchasing renewable fuels with attached RINs or by
purchasing separated RINs.
11.3 Cost-to-Sales Ratio Result
The final step in our methodology is to compare the total estimated costs to relevant total
estimated revenue from the sales of gasoline and diesel in the U.S. in 2023-2025. Since the RFS
standards are proportional to the volume of gasoline and diesel produced by each obligated party,
all obligated parties (including small refiners) are expected to experience costs (and recover
those costs) to comply with the RFS standards that are proportional to their sales volumes. As
discussed in Chapter 11.2, all obligated parties—including small refiners—recover their RFS
compliance costs and thus they have no net cost of compliance. Therefore, the cost-to-sales ratio
for all small refiners is 0%.
961 Prior to issuing our 2009 proposal for the general RFS regulatory program regulations required to implement the
amendments enacted pursuant to EISA, we analyzed the potential impacts on small entities of implementing the full
RFS program through 2022 and convened a Small Business Advocacy Review Panel (SB AR Panel) to assist us in
this evaluation. This information is located in the RFS2 rulemaking docket (Docket ID No. EPA-HQ-OAR-2005-
0161).
962 For a further discussion of the ability of obligated parties (including small refiners) to recover the cost of RINs,
see "April 2022 Denial of Petitions for RFS Small Refinery Exemption," EPA-420-R-22-005, April 2022 and "June
2022 Denial of Petitions for RFS Small Refinery Exemption," EPA-420-R-22-011, June 2022.
963 Id.
464
-------
11.4 Conclusion
Based on our outreach, fact-finding, and analysis of the potential impacts of this rule on
small businesses, we have concluded that there is no net cost to small refiners resulting from the
RFS program. Since obligated parties have been shown to recover their RFS compliance costs
through the resulting higher market prices for their petroleum products, there are no net costs of
the rule on small businesses, resulting in a cost-to-sales ratio of 0.00%.
465
------- |