Renewable Fuel Standard (RFS) Program: RFS Annual Rules


Renewable Fuel Standard (RFS) Program:
RFS Annual Rules

Regulatory Impact Analysis

SEPA

United States
Environmental Protection
Agency

-------
Renewable Fuel Standard (RFS) Program:

RFS Annual Rules

Regulatory Impact Analysis

Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency

United States
Environmental Protection
^1	Agency

EPA-420-R-22-008
June 2022

-------
Table of Contents

Executive Summary	iii

Overview	vi

List of Acronyms and Abbreviations	viii

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	25

1.7	Ethanol	29

1.8	Other Biofuels	34

1.9	RIN System	36

Chapter 2: Baselines and Volume Scenarios	48

2.1	Mix of Renewable Fuel Types for Final Volume Requirements	48

2.2	Volume Changes Analyzed	50

Chapter 3: Environmental Impacts	59

3.1	Air Quality	59

3.2	Climate Change	65

3.3	Conversion of Wetlands, Ecosystems, and Wildlife Habitats	95

3.4	Soil and Water Quality	107

3.5	Water Quantity and Availability	123

3.6	Ecosystem Services	136

Chapter 4: Energy Security Impacts	139

4.1	Review of Historical Energy Security Literature	140

4.2	Review of Recent Energy Security Literature	143

4.3	Cost of Existing U.S. Energy Security Policies	147

4.4	Energy Security Impacts	149

Chapter 5: Rate of Production of Renewable Fuel	156

5.1	Cellulosic Biofuel	156

5.2	Biomass-Based Diesel	175

5.3	Imported Sugarcane Ethanol	191

5.4	Other Advanced Biofuel	194

5.5	Ethanol	194

5.6	Conventional Biodiesel and Renewable Diesel	204

Chapter 6: Infrastructure	206

6.1	Biogas	206

6.2	Biodiesel	207

6.3	Renewable Diesel	209

i

-------
6.4 Ethanol	210

6.5. Deliverability of Materials, Goods, and Products Other Than Renewable Fuel	220

Chapter 7: Other Factors	223

7.1	Job Creation	223

7.2	Rural Economic Development	228

7.3	Supply of Agricultural Commodities	229

7.4	Price of Agricultural Commodities	232

7.5	Food Prices	236

Chapter 8: Environmental Justice	241

8.1	Greenhouse Gas Impacts	242

8.2	Air Quality Impacts	246

8.3	Water & Soil Quality Impacts	248

8.4	Impacts on Fuel and Food Prices	249

Chapter 9: Costs	251

9.1	Renewable Fuel Costs	251

9.2	Gasoline, Diesel Fuel, and Natural Gas Costs	282

9.3	Volumes and Energy Density Related Fuel Economy Cost	285

9.4	Costs	287

Chapter 10: Biomass-Based Diesel Standard for 2022	301

10.1	Review of Implementation of the Program	301

10.2	Analysis of Other Statutory Factors	306

Chapter 11: Screening Analysis	312

11.1	Summary	312

11.2	Background	313

11.3	Approaches to the Screening Analyses	315

11.4	Cost-to-Sales Ratio Results	319

11.5	Conclusions	320

11.6	Small Refiner CBI Data	320

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 for each of the four
categories of renewable fuel. It also directs EPA to modify or establish volume targets in certain
circumstances. EPA must also translate the volume targets into compliance obligations that
obligated parties must meet every year. In this action we are establishing the applicable volumes
for cellulosic biofuel, advanced biofuel, and total renewable fuel for 2020, 2021, and 2022, and
the biomass-based diesel (BBD) applicable volume for 2022, as well as a supplemental standard
for 2022 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) (hereafter "ACE'). We are also
establishing the annual percentage standards for cellulosic biofuel, BBD, advanced biofuel, and
total renewable fuel that will apply to gasoline and diesel produced or imported by obligated
parties in 2020, 2021, and 2022, as well as the percentage standard for the 2022 supplemental
standard.

This Regulatory Impact Analysis (RIA) supports our final rulemaking in several ways.
First, this RIA addresses our statutory obligations under CAA section 21 l(o)(7)(F) for resetting
the 2020, 2021, and 2022 cellulosic biofuel, advanced biofuel, and total renewable fuel volumes.
This section of the statute directs us to modify the statutory 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 memoranda to the docket. Second, this RIA also addresses our statutory
obligations under CAA section 21 l(o)(7)(D) for the volumes that we are establishing under the
cellulosic waiver authority, i.e., the cellulosic biofuel volumes for 2020, 2021, and 2022, as well
as the advanced biofuel and total renewable fuel volumes for 2020. As we explain further in
Preamble Section II and the Response to Comments (RTC) document, the analyses for the
cellulosic waiver authority are subsumed into the analysis for the reset authority. Our analysis of
the "projected volume available" of cellulosic biofuel is chiefly set forth in Chapter 5.1. Support
for exercising the discretionary portion of the cellulosic waiver authority to reduce the 2022
advanced and total volumes is found throughout the RIA. Third, this RIA addresses our statutory
obligations under CAA section 21 l(o)(2)(B)(ii) for setting the 2022 BBD volume. This section
of the statute requires a similar analysis as our reset authority. Chapter 10 of this document
specifically focuses on the 2022 BBD volume; however, other chapters are also relevant as they
more generally address BBD and its constituent fuels (advanced biodiesel and renewable diesel).
Finally, this RIA supports our establishment of the 2022 supplemental standard in response to
the ACE remand. Among other things, Chapters 2 and 5 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. 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. A full discussion of each
impact, including the uncertainties associated with estimating the impact, is contained in the RIA
Chapter under the "More Information" column. EPA compiled this table to provide additional
information to the public regarding this rulemaking and to comply with Circular A4.

iv

-------
Table ES-1: Potential <

Quantified and Unquantified Impacts Associated with the Volumes in this Rule

Potential impacts associated
with the volumes in this rale

Effect

Effect Quantified

Effect
Monetized

More
Information

Impacts on air quality from
biofiiel production and use

Increases in CO, NH3, NOx, PM10, PM2.5, SO2 and VOC emissions associated with
biorefinery production and product transport.

Emission inventory impacts

-

3.1

Higher ambient concentrations of NOx, HCHO and SO2 downwind of production facilities

Emission inventory impacts

-

3.1

Varying emission impacts from vehicles running on ethanol blends

Emission inventory impacts

-

3.1

Decrease for THC, CO, and PM2.5, but increase slightly for NOx emissions from pre-2007
diesels running on biodiesel

Emission inventory impacts

-

3.1

Impacts on climate change

from biofiiel feedstocks
production and displacement
of petroleum fuels

Reduced GHG Emissions

Illustrative

Illustrative

3.2

Impacts on wetlands,
ecosystems, and wildlife
habitat from land use change

Increased conversion of grasslands, wetlands, forests, and other habitats to cropland

Qualitative

-

3.3

Decreased plant diversity and decreased natural forage for wildlife, particularly for birds
and insects

Qualitative

-

3.3

Increased use of pesticides and reduced access to natural forage leading to reduced insect
biodiversity, especially in pollinators

Qualitative

-

3.3

Impacts on soil and water
quality from biofuel feedstock
production

Increased erosion, fertilizer and pesticide runoff and/or leachate

Qualitative

-

3.4

Depletion of natural soil organic matter, thereby depleting the soils nutrients

Qualitative

-

3.4

Chemical contamination from releases and spills

Qualitative



3.4

Increased erosion from tilling and other land management practices

Qualitative

-

3.4

Increased chance of a cyanobacterium bloom occuring

Qualitative



3.4

Increased turbidity and sedimentation in aquatic ecosystems; Nutrient loading in
waterways; Increased stress on aquatic organisms

Qualitative

-

3.4

Impacts on water quantity and
availability from biofuel and
feedstock production

Aquifer depletion

Qualitative

-

3.5

Use of limited water resources for irrigation instead of meeting human needs

Qualitative

-

3.5

Energy security

Increased energy security

Energy security benefits

$294
million

4

Production and use of
renewable fuels

Increased production and use of renewable fuels

Increased production and use
of renewable fuels

-

5

Infrastructure

Increased development of infrastructure of deliver and use renewable fuels

Qualitative



6

No adverse impact on deliverability of materials, goods, and products other than
renewable fuel

Qualitative



6

Jobs

Increased employment

Qualitative

-

7.1

Rural economic development

Support for rural economic development associated with biofuel and feedstock production

Qualitative





Commodity supply and price
impacts

Higher corn, soybean, and soybean oil prices

Commodity price increases

-

7.3

Increased supply of certain agricultural commodities

Qualitative





Higher food prices

Food price increases

-

7.5

Costs

Increased societal cost

Fuel cost increases

$7.0
billion

9.4

Changes to costs to consumers of transportation fuel

Cost changes

-

9.4

Increased costs to transport goods

Cost increases

-

9.4

V

-------
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 and Volume Scenarios

This chapter identifies the specific biofuel types and associated feedstocks that are projected to
be used to meet the final volume requirements, and the appropriate baselines for comparison.

Chapter 3: Environmental Impacts

This chapter discusses the environmental factors EPA analyzed in developing the final volume
requirements.

Chapter 4. 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 5: Rate of Production 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 6: Infrastructure

This chapter analyzes the impact of renewable fuels on the distribution infrastructure of the U.S.
Chapter 7: 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 8: Environmental Justice

This chapter describes potential environmental justice impacts associated with the production
and use of renewable fuels.

Chapter 9: Costs

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 10: Biomass-Based Diesel Standard for 2022

This chapter analyzes the required statutory factors specifically with respect to the final 2022
BBD volume requirement, in comparison to other advanced biofuels.

-------
Chapter 11: Screening Analysis

This chapter discusses EPA's screening analysis evaluating the potential impacts of the final
RFS standards on small entities.

Note: Unless otherwise stated, all documents cited in this document are available in the docket
for this action (EPA-HQ-OAR-2021-0324). 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.

vii

-------
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

CAA

Clean Air Act

CAFE

Corporate Average Fuel Economy

CBI

Confidential Business Information

CBOB

Blendstock for Oxygenate Blending

CG

Conventional Gasoline

CNG

Compressed Natural Gas

CO

Carbon Monoxide

CWC

Cellulosic Waiver Credit

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

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

GDP

Gross Domestic Product

GHG

Greenhouse Gas

GREET

Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model

LCA

Lifecycle Analysis

LCFS

Low Carbon Fuel Standard

LNG

Liquified Natural Gas

MSW

Municipal Solid Waste

MTBE

Methyl Tertiary Butyl Ether

MY

Model Year

NEMS

National Energy Modeling System

NHTSA

National Highway Transportation Administration

NOx

Nitrogen Oxides

OPEC

Organization of Petroleum Exporting Countries

OPIS

Oil Price Information Service

ORNL

Oak Ridge National Laboratory

PHEV

Plug-in Hybrid Electric Vehicle

viii

-------
PM

Particulate Matter

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

RVO

Renewable Volume Obligation

RVP

Reid Vapor Pressure

SOx

Sulfur Oxides

SRE

Small Refinery Exemption

STEO

Short Term Energy Outlook

ULSD

Ultra-Low-Sulfur Diesel

USD A

U.S. Department of Agriculture

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
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 RFS program began. 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 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 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
and 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
EPAct. While 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,
refiners now had increased disincentives to continue using MTBE after 2005. In addition,
although the oxygen requirement for RFG was removed in EPAct, the emission standards for

1	"Timeline - A Very Short History of MTBE in the US."

2	"Clinton-Gore Administration Acts To Eliminate MTBE, Boost Ethanol."

1

-------
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 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 were amended by the Energy Independence and Security
Act of 2007 (EISA), replacing the single total renewable fuel standard under the RFS 1 program
with four nested standards (cellulosic biofuel, BBD, advanced biofuel, and total renewable fuel).
EPA implemented these changes through 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, available at https://www.epa.gov/fuels-registration-reporting-and-compIiance-

heip/gasoline-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 RIA, "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 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
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 on only one occasion, in the 2016 annual rule
based on a finding of inadequate domestic supply.9 However, the D.C. Circuit Court of Appeals

7 CAA section 21 l(o)(7).

8 Obligated parties are producers and importers of gasoline and diesel. See 40 CFR 80.1406.

9 80 FR 77420 (December 14, 2015).

-------
vacated EPA's use of this waiver authority in 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, which EPA is acting on 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. The first is that the percentage standards are based on projected volumes of
gasoline and diesel consumption which typically deviate to some degree from what actually
occurs. EPA relies on forecasts provided by the U.S. Energy Information Administration (EIA)
in the Short Term Energy Outlook (STEO).10 In the context of the RFS program, the sum of non-
renewable gasoline and diesel demand is relevant. Since the first percentage standard was
applied in 2007, this forecast has sometimes overpredicted and sometime underpredicted actual
consumption as shown below.

Figure 1.2-1: Percent Change from Projected to Actual Sum of Non-Renewable Gasoline +

Diesela'b'c

a From 2007 to 2009, the RFS1 regulatory structure was in effect, and the applicable percentage standards were
based on non-renewable gasoline projections only. Therefore, the values for these three years represent non-
renewable gasoline.

b For purposes of demonstrating the error in forecasts, the blue line uses projected volumes derived from the October
edition of the STEO for the following year. The orange line, in contrast, uses forecasted volumes derived from the
version of the STEO that was actually used to calculate the percentage standards, which in some years was not the
October edition of the preceding year (e.g., the 2014 standards were established in December 2015, well after the
2014 compliance year was over).

0 See data and calculations in "Calculation of Percent Change from Projected to Actual Gasoline and Diesel,"
available in the docket.

d Represents the 2020 rulemaking which established the original 2020 standards, not this rulemaking which revises
those standards.

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 (negative values in
the figure above), 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

10 CAA section 21 l(o)(3)(A).

-------
used to set the applicable percentage standards (positive values in the figure above), the actual
volumes of renewable fuel used as transportation fuel will be higher. 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.11 These credits are called Renewable Identification Numbers, or "RINs."
Obligated parties have the flexibility to use RINs representing renewable fuel produced in the
previous year, often called "carryover RINs" or "banked RINs," to demonstrate compliance
rather than by using RINs representing current year renewable fuel production.12'13 The
nationwide total of banked RINs grew dramatically in the early years of the RFS program, and
obligated parties have at times drawn down this bank to help fulfill their obligations. For
instance, in 2013 consumption of renewable fuels fell more than 500 million ethanol-equivalent
gallons short of the applicable volume requirement, and obligated parties used banked RINs to
make up the shortfall. Similarly, while compliance demonstrations for 2019 have not yet fully
occurred, we estimate that obligated parties will use a significant number of banked RINs to
make up for a shortfall in actual consumption.14

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 the LCFS.

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 program, renewable fuel use
significantly outpaced the RFS requirements spurred by the transition from MTBE to ethanol as

11	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.

12	This flexibility is a function of the two-year life of RINs as discussed more fully in Chapter 1.9.

13	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) as indicated in 40 CFR 80.1427(a)(5).

14	"Carryover RIN Bank Calculations for 2020-2022 Final Rule," available in the docket for this action.

15	80 FR 77506 (December 14, 2015).

16	80 FR 77457 (December 14, 2015).

5

-------
an oxygenate. We discuss numerous other, non-RFS economic drivers for renewable fuel use
throughout this section.

Finally, exemptions for small refineries 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. These exemptions are permitted 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
generally not adjusted them to account for SREs that were subsequently granted. Rather, from
the start of the RFS program through the 2019 annual rule, EPA's standard-setting only
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 1.2-2: Volume of SREs Granted After the Applicable Percentage Standards Were
Set, By Compliance Year3

2000
1800
1600
1400

{/)

^ 1200

§ 1000

^ 800
600
400
200
0

Compliance year

a No SREs have been granted after 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.

As shown in Figure 1.2-2, 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. Other factors also played a role in determining whether and

17 75 FR 76805 (December 9, 2010).

¦ Advanced biofuel

¦ Conventional renewable fuel

I....ill

2011 2012 2013 2014 2015 2016 2017 2018

-------
when actual consumption was affected by SREs. 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 volumes of SREs were granted. 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/or the
following years, rather than in the compliance year to which they applied, as shown in Figure
1.2-3. This was often due to EPA granting the SREs after the compliance year had passed.

Figure 1.2-3: 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, it was not always the case that SREs affected the demand for RINs only in the
calendar year in which they were granted and/or 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/or 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 carryover RIN bank 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.18 Given that EPA has made a decision
not to exempt any volumes of gasoline or diesel for 2020, the original 2020 percentage standards

18 85 FR 7016 (February 6, 2020).

7

-------
were applicable to a larger volume of gasoline and diesel, effectively increasing the total
requirement for renewable fuel.

In sum, due to the many factors that affect renewable fuel consumption, actual
consumption was sometimes higher and sometimes lower than the volumes that EPA originally
intended to be achieved in setting the percentage standards.

Figure 1.2-4: Intended3 Versus Actual Consumption of Total Renewable Fuel

20,000
18,000
16,000
Z 14,000

CtL

.9. 12,000
5 10,000
8,000
6,000
4,000

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020b

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 original 2020 standards, not this rulemaking which revises
those standards.

The total volume of renewable fuel that was intended to be used between 2007 and 2020
(i.e., the volume that was used to calculate the applicable percentage standards) was almost 210
billion ethanol-equivalent gallons. In comparison, actual consumption was about 212 billion
ethanol-equivalent gallons over the same time period. Thus, actual consumption actually
exceeded what was intended over the life of the RFS program through 2020. 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 E10 blendwall and
blending ethanol as E10 was economically attractive relative to gasoline. In more recent years,
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.
Economic factors are coupled with the use of carryover RINs for compliance, the size of the
carryover RIN bank, 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

Actual consumption
Intended consumption

8

-------
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. Although 2019 compliance has not yet occurred, the E10 blendwall was
again lower than the implied conventional renewable fuel volume requirement, and we estimate
that the carryover RIN bank will be drawn down by 1.6 billion RINs from 3.43 to 1.83 billion
RINs.

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.19 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.

19 75 FR 14670 (March 26, 2010)

-------
Figure 1.3.1-1: Comparison of Statutory Volume Targets to Actual Consumption

Cellulosic Biofuel

18,000

Biomass-Based Diesel3

3,000

— — Statutory target
2/500 Actual consumption

£ 2,000 /

16,000 - ~ -Statutorytarget ^

14,000 Actual consumption j ^
w 12,000 7

z '

cc 10,000 *

o '

¦= 8,000 ~

"i 6 ooo —

J? 1,500 /

4,000 *
2,000 ^ —

500 /

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

Advanced Biofuel

Total Renewable Fuel

40,000

35,000 ~ ^Statutorytarget ^

— — Statutory target

20,000 ; ^

Actual consumption ^

w ~

? 15,000 X

O r

~ 10,000 ^ '

5,000 ^ *

Actual consumption ^ ^

w 30,000 ~

z
E

0 25,000 ^

1 *

20,000 3-^

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

10,000

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

6,000

r „ _— — Implied statutory target

Conventional Renewable Fuelb

16,000

Actual consumption ^ _ ~

-------
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 carryover RIN bank 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)

12,000

10,000

8,000

§ 6,000
c
o

ii 4,000
2,000
0

-2,000

Source for actual consumption: EMTS

1.3.2 Relative Proportions of Ethanol and Non-Ethanol Renewable Fuel

In the 2010 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 2010 rule that established the RFS2 program.

-Shortfall in cellulosic biofuel
-Shortfall in advanced biofuel
-Shortfall in non-cellulosic advanced biofuel

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

-Shortfall in cellulosic biofuel
-Shortfall in total renewable fuel
-Shortfall in conventional renewable fuel

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

-------
Figure 1.3.2-1: Volumetric Proportions of Each Fuel Type in Non-Cellulosic Advanced
Biofuel

2010 Projection Actual Consumption

2012 2013 2014 2015 2016 2017 2018 2019 2020 2012 2013 2014 2015 2016 2017 2018 2019 2020

Source: "2010 projection" is from the 2010 final rule which created the RFS2 program. See Table 1.2-3 in the RIA
for that rule. "Actual consumption" is fromEMTS.

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
„ 16,000

2? 15,000

5 14,000
13,000
12,000

2012 2013 2014 2015 2016 2017 2018 2019 2020

Source: "2010 projection" is from the 2010 final rule which created the RFS2 program. See Table 1.2-3 in the RIA
for that rule. "Actual consumption" is from EMTS.

a The 2010 projection of ethanol shown here represents the "primary control case." EPA also analyzed a "low
ethanol control case" and a "high ethanol control case."

-------
Figure 1.3.2-3: Actual Versus 2010 Projection of Biodiesel + Renewable Diesel
Consumption in Non-Cellulosic Renewable Fuel

4,500

4,000

3,500

1 3,000

2,500

2,000

1,500

Actual

1,000

2012 2013 2014 2015 2016 2017 2018 2019 2020

Source: "2010 projection" is from the 2010 final rule which created the RFS2 program. See Table 1.2-3 in the RIA
for that rule. "Actual consumption" is from EMTS.

This pattern of lower ethanol and higher non-ethanol in comparison to expectations
appears to be linked to the El0 blendwall and the difficulty that the market has had in increasing
sales of higher ethanol blends such as El 5 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
„ 16,000

o
"to

® 15,000

5 14,000
13,000
12,000

2012 2013 2014 2015 2016 2017 2018 2019 2020

Source: "2010 projection" is from the 2010 final nile which created the RFS2 program. See Table 1.2-3 in the RIA
for that 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 El5 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
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

1 19.0000

18.0000
17.0000
16.0000
15.0000

¦2010 projection
•Actual

¦AE02000
•AE02002
•AE02005
•AE02007
¦AE02010

OrH(NfO'^-iri(£ir~.CO<7lO

oooooooooo
oooooooooo

rsJcNrslrslrNirvirsirvirvl

-------
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 2010 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).

Figure 1.3.2-6: 2010 Projected Versus Actual Ethanol Concentration3

16%

15%

6PA-Ff4mafv-ease y—

14%

EPA non-cellulosic and conventional ethanol

Actual —

13%

ElO blendwall

12%

03
_£=

11%

10%

—

o/o

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

a Here and elsewhere in this RIA, "ethanol concentration" refers to the concentration of denatured ethanol in
gasoline.

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 fuels feedstock prices and
gasoline and diesel fuel prices tend to increase as well, although gasoline and diesel fuel prices
generally increase more relative to renewable fuels feedstock prices. Thus, higher crude oil
prices generally improve the economics of renewable fuels relative to gasoline and diesel fuel.
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,20 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 fuel prices projected by EIA in
AEO 2008, which projects 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
crude oil).21'22'23'24 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.

20 75 FR 14670 (March 26, 2010).

21 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 fuel sulfur specifications - thus sour crude oils are typically priced lower than sweet crude oils.

22 AEO 2008 - Petroleum Product Prices; Reference Case; EIA; June 2008.

23 AEO 2008 - Petroleum Product Prices; High Price Case; EIA: June 2008.

24 Petroleum and Other Liquids - Spot Prices WTI - Cushing and Brent - Europe; EIA;
https: //www. eia. gov/dnav/pet/pet_pri_spt_s 1 a. htm.

-------
Figure 1.4.1-1: AEO 2008 Projected Crude Oil Prices and Actual Crude Oil Prices (2007
dollars)3

120

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.

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 than the prices in 2008 attributed to crude oil production
not keeping up with demand.25 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, 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 4, 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

25 Hirsch, Robert L.; Peaking of World Oil Production: Impacts, Mitigation & Risk Management; Report to the
Department of Energy; February 2005.

-------
causing the U.S. economy to slide into a recession.26 It also led to Congress banning the export
of U.S. oil from 1975 to 2015.27

At the time Congress passed EPAct and EISA and EPA was promulgating the RFS1 and
RFS2 rulemakings, the U.S. was importing a large portion of its crude oil and refined petroleum
products. That trend was expected to continue because the eventual increase in U.S. crude oil
production due to fracking 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.28 The EIA-reported net imports of petroleum values account for
imports and exports of crude oil, petroleum products, and biofuels.29 For the net imports figures
shown below 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 to 2018. Because the production volume of
U.S. tight oil (fracked 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
fuel.

Figure 1.4.2-1: U.S. Net Petroleum Imports and U.S. Tight Oil Production

2020

¦ Net Petroleum Imports

¦ Distillate Imports

¦Gasoline Imports
¦Tight Oil Production

(O
T3

IP
_Q
_Q

c
o
'+-»

¦a

O
_c

Figure 1.4.2-1 shows that net petroleum imports increased from just over 10 million
barrels per day in 2000 to a maximum of about 12.5 million barrels per day in 2005. After

26 Verrastro, Frank A., The Arab Oil Embargo-40 Years Later; Center for Strategic & International Studies; October
16, 2013.

27 1975 Energy Policy and Conservation Act; Consolidated Appropriations Act of 2016.

28 U.S. Net Imports by Country; Petroleum and Other Liquids; EIA;
https://www.eia.gov/dnav/pet/pet move neti a EP00 IMN mbblpd m.htm

29 To calculate net petroleum imports EPA subtracted net biofuel imports from the U.S. Net Imports reported by
EIA.

-------
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 barrels per day.

Increased tight oil production and changes in gasoline and distillate (comprised largely of
diesel fuel) 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
barrels per day 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 barrels per day initially, trended
downward starting in 2006 to a negative 1.1 million barrels per day in 2017. Gasoline net
imports reached a maximum of over 1 million barrels per day 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 barrels per day lower than in 2007.

Renewable fuels likely contributed to reducing net petroleum imports by a relatively
modest amount. Corn ethanol volume increased from about 2 billion gallons per year in 2000 to
over 14 billion gallons per year in 2015.30'31 Biodiesel consumption increased from 10 million
gallons per year in 2001 to over 1 billion gallons in 2013, and biodiesel and renewable diesel
consumption totaled over 2 billion gallons in 2019.32'33 Assuming that this total renewable fuel
volume displaced an energy-equivalent volume of petroleum imports, in 2019 corn ethanol and
biodiesel/renewable diesel fuel combined would have displaced about 0.75 million barrels per
day of petroleum equivalent volume - which is equivalent to 6% of the highest import volume.

Not only do petroleum imports create energy security vulnerabilities, they also contribute
to a monetary trade imbalance. What made the continued increase of net petroleum imports until
2005 a particular concern was that crude oil prices were increasing at the same time.34 Crude oil
spot prices (both WTI and Brent) had doubled in 2005 to over $50/bbl compared to the average
of 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.

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 2008.35 The increasing crude oil prices on
top of the increasing petroleum imports contributed to this increasing trade imbalance. The 12
million barrel per day net petroleum import volume combined with the approximately $70 per

30 EIA, Monthly Energy Review, Table 10.3, Fuel Ethanol Review;

https://www.eia.gov/totalenergy/data/monthiv/pdf/sec.1.0 7.pdf

31 Note that "corn ethanol" also includes small amounts of ethanol produced from other sources of starch such as
wheat and grain sorghum.

32 Biodiesel consumption data from EIA, Monthly Energy Review, Table 10.4, Biodiesel and Other Renewable
Fuels Overview; https://www.eia.gov/totaienergv/data/monthiv/pdf/seclO 8.pdf

33 Renewable consumption data from Public Data for the Renewable Fuel Standard; EPA Moderated Transaction
System; https://www.epa.gov/fiieis-registration-reporting-and-compliance-heip/pnblic-data-renewable-fiiei-standard

34 Spot Prices - Petroleum and Other Liquids; EIA; https://www.eia.gov/dnav/pet/pet pri spt si a.fatm

35 U.S. Trade in Goods with World, Seasonally Adjusted; United States Census Bureau;
https://www.censiis.gov/foreign-trade/balance/c0004.html

-------
barrel 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 prices also increased the prices of many other goods that were imported into the
U.S., which likely indirectly contributed to the trade imbalance.36 In 2009, the U.S. trade
imbalance dropped to $500 billion. Since then, the U.S. trade imbalance increased back into the
$600 to $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.

However, we recognize that because the U.S. is a participant in the world market for
petroleum products, our economy cannot be shielded from world-wide price shocks.37 But, 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 the time frame of this rule, 2021-2022. The potential for supply disruptions has not
been eliminated.38 We discuss these topics further in Chapter 4.

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 crude oil cracking refinery; and (3) a medium crude oil hydrocracking refinery in
Singapore.39 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.

36 U.S. Trade Deficit and the Impact of Changing Oil Prices; Congressional Research Service; February 24, 2020;
https://fas.org/sgp/crs/misc/RS22204.pdf

37 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-u-s-energy-independence-has-gone-up-in-smoke

38 Foreman, Dean; Why the US must Import and Export Oil; American Petroleum Institute; June 14, 2018;
https://www.api.org/news-policy-and-issiies/blog/2018/06/14/wltv-the-us-must-import-and-export-oil

39 Oil Refinery Margins - Regional; NASDAQ Data Link; https://data.nasdaa.com/data/BP/OIL REF MARG-oil-
refi ni n g - ma r g i n s - re g i o na 1.

-------
Figure 1.4.3-1: Refinery Margins in Three Different Regions ($/bbl)

I SINGAPORE MEDIUM SOUR HYDROCRACKING

Figure 1.4.3-1 shows that from 1993 to 2004, refinery margins were modest, and the
Singapore refinery particular experienced zero or near-zero margins over much of the 1998 to
2004 time period. In the period from 2004 to 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. 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 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 to 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 represented in the
figure due to the pandemic. Refinery margins returned to their pre-pandemic levels in 2022, but
then increased significantly starting in March 2022 due to geopolitical factors.40

1.4.4 Transportation Fuel Demand

At the time the RFS2 program was being enacted 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

4" McGurty. Janet; Refinery Margin Tracker: Russian crude cargoes taper off as margins rise; S&P Global; April 4,
2022.

-------
2000 to 2018, as well as the projected demand of gasoline and distillate if transportation fuel
demand growth had continued at the historic rate based on EIA's AEO 2008.41 42

Figure 1.4.
renewable

250.0

200.0

= 150.0
o

O. 100.0
a.

50.0

4-1: Actual and Projected Transportation Fuel Demand (Source: EIA; includes
fuel volumes)

o.o

2000

Gasoline

EIA Projectec

Gasoline
Distillate

Jet Fue

2005

2010
Year

2015

2020

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
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.

41 Product Supplied; Petroleum and Other Liquids, Energy Information Administration,
https://www.eia. gov/dnav/pet/pet_cons_psup_dc_nus_mbbl_a. htm.

42 Annual Energy Outlook 2008; Energy Information Administration; June 2008;
https://www.eia.gov/outlooks/arcliive/aeo08/index.html.

-------
• 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 to $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 to 2025 and thereafter.43 EPA and NHTSA
also established GHG/CAFE standards for new heavy-duty vehicles and their
trailers.44 The phase 1 and phase 2 heavy-duty GHG standards began to phase-in in
2014 and will continue to do so through 2027.45 The GHG standards only affect new
internal combustion vehicles; thus, as consumers purchase new motor vehicles, these
new vehicles consume less gasoline and diesel fuel compared to the vehicles sold in
previous years, reducing overall petroleum demand.

• Electric vehicle penetration and fuel 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 that EVs and PHEVs displaced an estimated 5 million gallons of fuel in
2011 and that this increased to over 400 million gallons in 2019.46

1.5 Cellulosic Biofuel

Actual production of cellulosic biofuel through 2021 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 2010 RFS2 final 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.47 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)

43 75 FR 25324 (May 7, 2010) and 86 FR 74434 (December 30, 2021).

44 76 FR 57106 (September 15, 2011).

45 81 FR 73478 (October 25, 2016).

46 Transportation Research Center at Argonne National Laboratory, https://www.anl.gov/es/light~diitv-electric~drive~
vehicles-monthly-sales-updates

47 Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis (EPA-420-R-10-006). February 2010.

-------
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 5.1, 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 has been produced using technologies not discussed in that rule. The
production of 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 pathway48 for the production
of biogas from landfills, sewage and waste treatment plants, and manure digesters to generate
advanced biofuel (D5) RINs.49 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.50 Following this decision, production of
CNG/LNG derived from biogas increased rapidly, from approximately 33 million RINs in 2014
to over 560 million RINs in 2021.51 Through 2021, 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 2022. Actual cellulosic biofuel production for each year
from 2014 through 2021 is shown in Figure 1.5-1.

48 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).

49 75 FR 14872 (March 26, 2010).

50 79 FR 42128 (July 18, 2014).

51 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.

-------
Figure 1.5-1: Cellulosic RINs Generated (2013-2021)

600

2013 2014 2015 2016 2017 201S 2019 2020 2021

¦ CNG/LNG Derived from Biogas ¦ Liquid Cellulosic Biofue!

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's RIA, EPA projected that in 202152 1.62 billion
gallons of biodiesel and 0.15 billion gallons of renewable diesel would be supplied, all of which
was projected to be produced in the U.S. The actual supply of biodiesel and renewable diesel in
2021 was 1.82 billion gallons and 0.99 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 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.

52 2021 is the most recent year for which data are available for comparison.

-------
Figure 1.6-1: 2010 Projected vs. Actual Biodiesel and Renewable Diesel Supply (2014-2021)

3.00

2.50

c 2.00
o

^ 1.50

0.50

0.00

2014 2015 2016 2017 2018 2019 2020 2021

¦ Projected Renewable Diesel ¦ Actual Renewable Diesel

¦ Projected Biodiesel ¦ Actual Biodiesel

Projected volumes are from the 2010 RFS2 final rale. Actual volumes are from EMTS data.

Figure 1.6-2: Source of Biodiesel and Renewable Diesel Consumed in the U.S. (2014-2021)

3.00
2.50

C 2.00

<3 1.50

1 1-00

0.50
0.00

¦ II

2014 2015 2016

¦ Biodiesel Net Imports

¦ Domestic Biodiesel

l ll ll

2017 2018 2019 2020 2021
¦ Renewable Diesel Net Imports
n 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 and 2021 (12.70 and 13 .88
billion gallons, respectively) that was lower than the projected ethanol consumption volume in
2022 even under the low ethanol case from the 2010 final rule (17.04 billion gallons).53 Since the

53 Ethanol consumption volume are from EIA's Monthly Energy Review, while the ethanol projections are from the
2010 RFS2 final rale. Ethanol consumption in 2020 was significantly impacted by the COVID-19 pandemic.

-------
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, a topic we discuss further in
Chapter 10.

The feedstocks used to produce biodiesel and renewable diesel each year from 2014
through 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 (2014-2021)

2.50

2014 2015 2106 2017 2018 2019 2020 2021

¦ FOG ICorn Oil ISoybeanOil ¦ Canola Oil ¦ Grandfathered (Unknown)
Source: EMTS

Ethanol 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.

-------
Figure 1.6-3: Imported Biodiesel and Renewable Diesel Feedstocks (2014-2021)

1.00

2014 2015 2106 2017 201S 2019 2020 2021

¦ FOG ¦CornOil ¦Soybean Oil ¦ Canola Oil ¦ Algal Oil ¦ 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 2021. Domestic biodiesel production in 2021 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 RIA, 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
projections being over-ambitious, and demand for this feedstock in animal feed and other
sectors.

Domestic renewable diesel production in 2021 was significantly higher than projected in
the RFS2 rule, in which EPA projected that all of the renewable diesel would be produced
domestically from FOG. While the majority of domestic renewable diesel was produced from
FOG in 2021, 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
2021, as well as in previous years. By 2021, 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.54

54 Source: EMTS

-------
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 biofuels consumed in the
U.S. transportation sector.55 Over the following years, 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 2019.56 57
By 2010, ethanol use in the U.S. was approaching the "E10 blendwall" for the nation as a whole,
as represented by the nationwide average ethanol concentration, and actually exceeded 10.00%
in 2016. By 2021, ethanol accounted for 80% of the 17.2 billion gallons of biofuels consumed in
the U.S.58

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

2006200720082009201020112012201320142015201620172018201920202021 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Source: EMTS

As shown in Figure 1.7-2, actual consumption of ethanol in the U.S. was very close to
domestic production through 2009. Thereafter, domestic production began exceeding domestic
consumption, indicative of an increase in exports.

55 EIA's Monthly Energy Review, April 2021, Tables 10.3 and 10.4. Comparison is based on ethanol-equivalence.

56 Id.

57 In 2020 and 2021, total ethanol consumption dropped significantly as a result of the COVID-19 pandemic.

58 "RIN supply as of 2-17-22," available in the docket.

-------
Figure 1.7-2: Domestic Production and Consumption of Ethanol

18,000
16,000
14,000
12,000

= 10,000

QO
c

£ 8,000

6,000
4,000
2,000
0

— ^

—

• Domestic

production

—Domestic

consurr

iptic

> ElOblendwal

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

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 deciding 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. Ethanol production exceeded domestic consumption through exports.

a Here and elsewhere in this RIA, "ethanol concentration" refers to the concentration of denatured ethanol in
gasoline.

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
E0 (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 flex-fuel
vehicle (FFV) was produced that could operate on fuel containing up to 85% denatured ethanol
(83% ethanol).59 Starting in 2007, the American Society for Testing and Materials (ASTM)
limited the maximum ethanol content of E85 to 83% in specification D5798, with a minimum
content of 51%. EIA assumes that the annual, nationwide average ethanol concentration of E85
is 74%.60

E85 is not considered gasoline under EPA's regulations, and as such is permitted to be
used only in designated FFVs. However, FFVs can operate on either gasoline or E85. Under
basic economic 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).61'62'63 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.

59 "Alternative Fuel Ford Taurus" available in the docket for this action.

60 "AEO2021 Table 2," available in the docket. See footnote 11.

61 74% is the average E85 ethanol concentration used by EIA (e.g., see footnote 1 in Table 6.36 of AEO 2020).

62 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.

63 A comparison to E0 would be more relevant prior to 2010 when there remained significant volumes of E0 for sale
at retail stations.

-------
Figure 1.7.1-1: Volumetric Price Reduction of E85 Compared to E10a

¦ E85prices.com
» gasprices.aaa.com
» Energy equivalence with E10

a The 21% energy equivalence level of E85 compared to E10 assumes that E85 contains 74% ethanol.

Note: While E85prices.com is a decentralized system consisting of voluntary submissions from motorists, the
American Automobile Association (AAA) data is based on a daily collection of credit card swipe and direct feed
price data from up to 130,000 retail stations. Moreover, the data collection by AAA is done in cooperation with the
Oil Price Information Service (OPIS) and Wright Express to ensure reliability of the results.

EPA has estimated the nationwide volume of E85 consumed in recent years using two
different methods.64 The results of those analyses are shown in Figure 1.7.1-2.

Figure 1.7.1-2: Estimated E85 Consumption3

350
300
250

S 150
o

100

306 311

I 297

239

205

186 _ .

r 225 224

t | 206

—* 20

192

4 m

16(

# Method #1
—Method #2a

• Method #2b

2014

2015

2016

2017

2018

2019

2020

2021

a The ethanol concentration of E85 is assumed to be 74% on average.

As discussed in Chapter 5.5.1, we do not need estimates of E85 (or El 5) 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 above. These estimates cover
2020, 2021, and 2022. See further discussion in Chapter 5.5.4.

64 "Estimate of E85 consumption in 2020," memorandum from Akshav Delity to the docket.

-------
1.7.2 E15

Following a 2011 waiver action, gasoline containing up to 15% ethanol was permitted to
be used in model year (MY) 2001 and newer vehicles.65 El5 has since been offered at an
increasing number of retail service stations (see Chapter 6.4.3). There is currently no publicly
available data on actual El5 sales volumes for the nation as a whole.

Sales of El 5 prior to 2019 were seasonal due to the fact that El 5 did not qualify for a 1
psi Reid Vapor Pressure (RVP) volatility waiver for summer gasoline in CG areas that has been
permitted for E10 since the summer volatility standards were implemented in 1989.66 Monthly
E15 sales in Minnesota from 2015 to 2018, presented below in Figure 1.7.2-1, demonstrate that
sales volumes of El 5 in summer months were notably lower than in non-summer months in this
time period.

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 RVP waiver to El5.67 EPA estimated that the annual
average El5 sales per station in Minnesota would have been 16% higher than they were in reality
had the 1 -psi waiver been in place in 2015 to 2018.68 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 E15 was based on an
impermissible reading of the statute and vacated it. Insofar as the 1 -psi waiver for El 5 had an
impact on summer sales of E15, therefore, it did so only for 2019 - 2021 and would not generally
have an impact in 2022 or later years.69

65 76 FR 4662 (January 26, 2011).

66 54 FR 11883(Marcli 22, 1989).

67 84 FR 26980 (June 10, 2019).

68 "Estimating the impacts of the lpsi waiver for E15," memorandum from David Korotney to EPA docket EPA-
HQ-0 AR-2019-0136.

69 We note that EPA issued an emergency waiver for 2022 that allows E15 to take advantage of the 1-psi waiver to
address issues related to fuel price and supply.

-------
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 1.8-2, biofuels other than corn ethanol and
B BD represented between 2% and 5% of total renewable fuel between 2012 and 2021 /0

Figure 1.8-1: Contribution of Biofuels to Total Renewable Fuel Consumption"-1'

19,000

¦ Imported conventional ethanol

18,000

¦ Imported conventional renewable diesel

¦ Imported conventional biodiesel

¦ Domestic conventional biodiesel

17,000

¦ Domestic advanced renewable diesel

¦ Domestic advanced gasoline/naphtha

(/>
z

¦ Domestic advanced heating oil

cc
c

16,000
15,000

¦ Imported advanced ethanol

¦ Domestic advanced ethanol

¦ Domestic advanced biogas

¦ Domestic advanced jet fuel

¦ Imported cellulosic ethanol

14,000

¦ Domestic cellulosic ethanol

¦ Imported cellulosic biogas

¦ Domestic cellulosic biogas

13,000

¦ Domestic com ethanol + BBD

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

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 dicscl. 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.

70 Detailed data prior to 2012 on RIM generation, adjustments to account for invalid RINs. and exports is less robust
and is therefore not presented here.

-------
Figure 1.8-2: Biofuels Other than Corn Ethanol and BBD

1,000

900

800

700

600

500

400

300

200

100

¦ 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

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

As illustrated by 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

¦ Imported advanced ethanol

l/>

400

¦ Domestic advanced renewable diesel

C£
C

¦ Domestic advanced gasoline/naphtha

300
200
100

¦ Domestic advanced ethanol

¦ Domestic advanced biogas

¦ Domestic advanced heating oil

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Source: EMTS

-------
These sources of advanced biofuel varied widely in both their overall contributions to the
advanced biofuel pool from 2012 to 2021, 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 5.3. Production of domestic advanced renewable diesel,71
gasoline/naphtha, and ethanol were of approximately similar magnitude and demonstrated no
consistent increasing or decreasing trends between 2012 and 2021. Domestic advanced biogas
fell to near zero in 2015 after biogas from landfills was recategorized as cellulosic biofuel in
2014.72 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.

1.9 RIN System

Renewable Identification Numbers (RINs) were created by EPA under the authority of
CAA section 21 l(o)(5) as a flexible mechanism to enable obligated parties under the RFS
program across the country to meet their renewable fuel blending obligations without having to
blend the renewable fuels themselves.73 RINs allow: (1) The refining industry to comply with the
RFS program without producing, purchasing, or blending the renewable fuels themselves; (2)
The non-obligated blenders of renewable fuels 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 fuels, 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 fuels they produce. These RINs are generally sold together
with the renewable fuel to refiners or fuel 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.

In the RFS regulations, 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).74 The type of RIN that can be generated for each gallon of
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 fuels. The
obligations under the RFS regulations are nested, so that some RIN types can be used to satisfy
obligations in multiple categories. A graphic of the nested nature of the RFS obligations is shown
in Figure 1.9-1.

71 Small quantities of renewable diesel are not BBD but are nonetheless advanced biofuel.

72 79 FR 42128 (July 18, 2014).

73 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).

74 40 CFR 80.1425(g).

-------
Figure 1.9-1: Nested Structure of the RFS Program

Total renewable fuel

Advanced biofuel

D3/7

t t

CeNulosic BBD "Other" advanced Conventional

biofuel (sugarcane ethanol, (mostly corn-ethanol)

Non-cel lulosic advanced

1.9.1 Carryover RIN Bank

CAA section 21 l(o)(5) requires that EPA establish a credit program as part of its RFS
regulations, and that the credits be valid for obligated parties to show compliance for 12 months
as of the date of generation. EPA implemented this requirement through the use of RINs, which
can be used to demonstrate compliance for the year in which they are generated or the
subsequent compliance year. Obligated parties can obtain more RINs than they need in a given
compliance year, allowing them to "carry over" these surplus RINs for use in the subsequent
compliance year. In order to ensure reasonably consistent demand for new renewable fuel use in
the ensuing year, however, our regulations limit the use of these carryover RINs to 20% of an
obligated party's renewable volume obligation (RVO). For the bank of carryover RINs to be
preserved from one year to the next, individual carryover RINs are used for compliance before
they expire and are essentially replaced with newer vintage RINs that are then held for use in the
next year. For example, vintage 2020 carryover RINs must be used for compliance in 2021, or
they will expire. However, vintage 2021 RINs can then be "banked" for use in 2022.

In the context of setting the annual volume standards, the relative number of carryover
RINs projected to be available compared to the projected volume requirement for each category
has helped inform EPA's decisions regarding the extent to which it should exercise its waiver
authorities. During the first several years of the RFS2 program, the total number of RINs
generated far exceeded the total number of RINs needed for obligated parties to demonstrate
compliance. This resulted in a dramatic increase in the carryover RIN bank, up to an estimated
2.67 billion total carryover RINs in 2013, which represented over 16% of the total renewable
fuel volume standard for that year.75 As a result, EPA determined that sufficient carryover RINs
existed such that it was not necessary for EPA to use its cellulosic or general waiver authority to

75 EPA first began projecting the size of the carryover RIN bank in the 2013 RFS annual rule.

-------
reduce the total or advanced biofuel volume requirements specified in the statute.76 At the time,
EPA recognized that this decision may result in a reduction in the carryover RIN bank, and this
in fact occurred, as the total number of carryover RINs dropped by over 900 million RINs to an
estimated 1.74 billion total carryover RINs in 2014.77 As seen in Table 1.9.1-1, at the time the
annual standards for individual years were established (i.e., the information available to EPA
when the standards were finalized for the subsequent year), the relative size of the projected total
carryover RIN bank compared to the projected total renewable fuel volume requirement ranged
from a high of over 16% in 2013 to a low of 8% in 2017. However, the magnitude of the
carryover RIN bank deviated from these projections sometimes significantly based on the
decisions of the market players that only became known after the rule were finalized. With the
benefit of hindsight, EPA can calculate the number of carryover RINs for each category that
were actually available for compliance for a given year. As seen in Table 1.9.1-1, the actual size
of the total carryover RIN bank compared to the actual volume obligation for 2013 through 2019
has ranged from a high of nearly 17% in 2018 to a low of 9% in 2016.78 In absolute terms, the
carryover RIN bank reached its highest historical levels going into the 2019 compliance year, at
3.43 billion RINs. However, as we further discuss in Preamble Section III, we expect 2019
compliance to significantly draw down the carryover RIN bank to 1.83 billion RINs. We also
present the size of the carryover RIN bank for the other biofuel categories below. As with the
total RIN bank, the advanced and BBD RIN banks are also expected to be drawn down
significantly for 2019 compliance.

76 78 FR 49820-22 (August 15, 2013).

77 80 FR 77482-87 (December 14, 2015).

78 Similar comparisons can also be made for the advanced biofuel, BBD, and cellulosic biofuel categories, and are
presented in Tables 1.9.1-2 through 4.

-------
Table 1.9.1-1: Total Renewable Fuel Carryover RINs Compared to Total Renewable Fuel
Volume Requirement"

Compliance
Year

Total Renewable Fuel
Carryover RINs Available
(billion RINs)

Total Renewable Fuel
Volume Requirement
(billion gal)

Carryover RINs as
% of Volume
Requirement

Projectedb

Actual0

Projectedb

Actual0

Projected

Actual

2013

2.67

2.47

16.55

16.92

16.1%

14.6%

2014

1.74

1.58

16.28

16.31

10.7%

9.7%

2015

1.74

1.69

16.93

17.00

10.3%

10.0%

2016

1.74

1.65

18.11

17.93

9.6%

9.2%

2017

1.54

2.48

19.28

18.49

8.0%

13.4%

2018

2.22

3.13

19.29

18.51

11.5%

16.9%

2019

2.59

3.43

19.92

21.03

13.0%

16.3%

2020

1.83

n/a

17.13

n/a

10.7%

n/a

2021

1.83

n/a

18.84

n/a

9.7%

n/a

2022

1.83

n/a

20.63

n/a

8.9%

n/a

a For further discussion of these calculations, see "Carryover RIN Bank Calculations for 2020, 2021, and 2022 Final
Rule," available in the docket for this action.

b Projected volume requirements and number of carryover RINs reflect the values projected in the rules establishing
the standards for those years.

0 Data current as of March 10, 2022, and compiled from Tables 2 and 3 at https://www.epa.gov/fiiels-registration-
reporting-and-compliance-heip/aiinnal-compliance-data-obligated-parties-and. Actual Volume Requirement =
Reported Volume Obligation (Table 2) + Total End-of-Year Compliance Deficit (Table 5).

-------
Table 1.9.1-2: Advanced Biofuel Carryover RINs Compared to Advanced Biofuel Volume
Requirement"

Compliance
Year

Advanced Biofuel
Carryover RINs Available
(million RINs)

Advanced Biofuel
Volume Requirement
(million gal)

Carryover RINs as
% of Volume
Requirement

Projectedb

Actual0

Projectedb

Actual0

Projected

Actual

2014

580

540

2,670

2,680

21.7%

20.1%

2015

580

460

2,880

2,900

20.1%

15.7%

2016

580

410

3,610

3,560

16.1%

11.4%

2017

390

860

4,280

4,100

9.1%

21.0%

2018

810

690

4,290

4,110

18.9%

16.8%

2019

600

660

4,920

5,240

12.2%

12.5%

2020

n/a

4,630

n/a

0.9%

n/a

2021

n/a

5,050

n/a

0.8%

n/a

2022

n/a

5,630

n/a

0.7%

n/a

a For further discussion of these calculations, see "Carryover RIN Bank Calculations for 2020, 2021, and 2022 Final
Rule," available in the docket for this action.

b Projected volume requirements and number of carryover RINs reflect the values projected in the rules establishing
the standards for those years.

0 Data current as of March 10, 2022, and compiled from Tables 2 and 3 at https://www.epa.gov/fuels-registration-

reporting-and-compliance-heip/aiinnal-compliance-data-obligated-parties-and. Actual Volume Requirement =
Reported Volume Obligation (Table 2) + Total End-of-Year Compliance Deficit (Table 5).

Table 1.9.1-3: BBD Carryover RINs Compared to BBD Volume Requirement1

Compliance
Year

BBD Carryover RINs
Available
(million RINs)

BBD Volume
Requirement
(million RINs)d

Carryover RINs as
% of Volume
Requirement

Projectedb

Actual0

Projectedb

Actual0

Projected

Actual

2014

690

500

2,450

2,500

28.2%

19.9%

2015

690

390

2,600

2,660

26.5%

14.6%

2016

690

270

2,850

2,800

24.2%

9.5%

2017

270

830

3,000

2,850

9.0%

29.1%

2018

780

720

3,150

3,020

24.8%

23.7%

2019

640

600

3,150

3,340

20.3%

17.9%

2020

n/a

3,650

n/a

1.6%

n/a

2021

n/a

3,650

n/a

1.6%

n/a

2022

n/a

4,330

n/a

1.4%

n/a

a For further discussion of these calculations, see "Carryover RIN Bank Calculations for 2020, 2021, and 2022 Final
Rule," available in the docket for this action.

b Projected volume requirements and number of carryover RINs reflect the values projected in the rules establishing
the standards for those years.

0 Data current as of March 10, 2022, and compiled from Tables 2 and 3 at https://www.epa.gov/fiiels-registration-
reporting-and-compliance-heip/annnal-compliance-data-obligated-parties-and. Actual Volume Requirement =
Reported Volume Obligation (Table 2) + Total End-of-Year Compliance Deficit (Table 5).
d Projected BBD volume requirements have been converted to ethanol-equivalent gallons (i.e., RINs). Actual BBD
requirements are reported in RINs.

-------
Table 1.9.1-4: Cellulosic Biofuel Carryover RINs Compared to Cellulosic Biofuel Volume
Requirement"

Compliance
Year

Cellulosic Biofuel
Carryover RINs Available
(million RINs)

Cellulosic Biofuel
Volume Requirement
(miillion gal)

Carryover RINs as
% of Volume
Requirement

Projectedb

Actual0

Projectedb

Actual0

Projected

Actual

2014

0.02

0.1%

0.6%

2015

0.02

123

120

0.0%

9.7%

2016

0.02

230

0.0%

17.0%

2017

311

310

8.7%

11.1%

2018

288

280

10.8%

3.0%

2019

418

440

0.2%

11.1%

2020

n/a

510

n/a

7.5%

n/a

2021

n/a

560

n/a

6.8%

n/a

2022

n/a

630

n/a

6.0%

n/a

a For further discussion of these calculations, see "Carryover RIN Bank Calculations for 2020, 2021, and 2022 Final
Rule," available in the docket for this action.

b Projected volume requirements and number of carryover RINs reflect the values projected in the rules establishing
the standards for those years.

0 Data current as of March 10, 2022, and compiled from Tables 2 and 3 at https://www.epa.gov/fuels-registration-

reporting-and-compliance-heip/aiinnal-compliance-data-obligated-parties-and. Actual Volume Requirement =
Reported Volume Obligation (Table 2) + Total End-of-Year Compliance Deficit (Table 5).

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 below.79 While there are a
wide variety of factors that impact RIN prices, including both market-based and regulatory
factors, a review of RIN prices reveals several notable aspects of the RFS program.

79 RIN prices are reported publicly on EPA's website (https://www.epa.gov/fuels-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 entry errors).
Several other services 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.

-------
Figure 1.9.2-1: RIN Prices

$3.50
$3.00

liik IKhBi

JiffV u|V|

$2.50 ¦M l Ml I

.In' rrLki Mil

u vuu j|i i

$2.00

D6 RIN Price D5 RIN Price D4 RIN Price D3 RIN Price

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 through 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 (the difference between the total renewable fuel
requirement and the advanced biofuel requirement) exceeded the volume of ethanol that could be
consumed as E10. While it has generally been cost-effective to blend ethanol as E10, higher-
level ethanol blends (such as E15 or 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 able to be optimized to take advantage of the very
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 through 2016,
D6 RIN prices remained close to, but slightly less than, D4 and D5 RIN prices. During this time
period 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 them to
meet their total renewable fuel obligation. 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
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, reducing the quantity 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 $1.00 and $1.50 by the end of 2021. 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/gal 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 2010-2021.
Instead, higher D6 RIN prices have resulted in lower effective prices (the price after subtracting
the RIN value from the price of the fuel with the attached RIN) for ethanol after the RINs have
been separated and sold. Higher D6 RIN prices have thus served to subsidize fuel blends that
contain higher proportions of conventional biofuels (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, BO).80

80 Burkholder, Dallas. A preliminary Assessment of RIN Market Dynamics, RIN Prices, and Their Effects. US EPA
Office of Transportation and Air Quality. May 2015.

-------
Figure 1.9.2-2: Ethanol Prices and D6 RIN Prices

$4.00

¦tA

O
C
ra
sz

$3.50
$3.00
$2.50
$2.00
$1.50
$1.00
$0.50

$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, N

i/1

LMto

til

•Ethanol Price (Iowa)

•06 RIN Price

D5 RINs were priced at a level between D4 and D6 RINs from 2010 through 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 El0
blendwall. This is because there are two primary fuel types that have been used to satisfy the
advanced biofuel requirements: sugarcane ethanol and BBD. From 2010 through 2012, obligated
parties generally met their implied requirements for "other advanced biofuel" (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) with sugarcane
ethanol. 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.81 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 the market 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

81 See Chapters 5.3 and 5.2 for volumes of sugarcane ethanol and BBD used in the U.S., respectively.

-------
remaining total renewable fuel requirements. RIN prices responded, and since 2013 the price 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 diesel fuel (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/gal federal tax subsidy and state LCFS
credits).82 Most recently, in 2021, D4 RIN prices have increased quite significantly, tracking
with an increase in feedstock (e.g., soybean oil) commodity prices, 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 to increase BBD production and use.
A December 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 credit is in effect."83 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).84 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

82 A $1 per gallon biodiesel blenders tax credit has been available to biodiesel blended every year from 2010 through
the end of 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 2018 and put in place
prospectively through 2022.

83 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.

84 Pursuant to CAA section 21 l(o)(7)(D)(ii), 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).

-------
as expected, as CWCs can be purchased directly from EPA when obligated parties demonstrate
compliance and carry no risk of RIN invalidity.8' This discount tends to be larger at the
beginning of the year, before narrowing near the end of the year as the RFS compliance date
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.

Figure 1.9.2-3: D3 RIN Prices and D5 RIN Price Plus CWC Price86

$4.50

$4.00

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
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 and protects these parties from excessively high cellulosic RIN prices. The
CWC price is also information to potential cellulosic biofuel producers. Potential cellulosic
biofuel producers can use the CWC price, along with the price of the petroleum fuel displaced by

85 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.

• EPA offers cellulosic waiver credits for years in which we reduce the cellulosic biofuel volume from the statutory
target. Cellulosic waiver credit prices for 2021 and 2022 will be posted on EPA's website shortly following the
publication of this rule. For the purposes of this table we have calculated the 2021 cellulosic waiver credit price
using the statutory formula and publicly available data.

-------
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 fuel
they produce. Knowing this price can help potential cellulosic biofuel producers determine
whether their cellulosic biofuel production processes are economically viable under 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 since 2010 to set the required volume of cellulosic biofuel at the
volume expected to be produced each year,87 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 generation of this type of cellulosic
biofuel.88 Unlike other RIN costs that are generally transfered within the liquid fuel pool (e.g.,
from consumers of fuels with relatively low renewable fuel content such as E0 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 fuel to parties outside of the liquid fuel pool (e.g., landfill owners).89 For
example, in 2021 the average cellulosic RIN price was $2.75.90 Thus, the total cost associated
with the 560 million cellulosic RINs required for compliance in 2021 was approximately $1.54
billion. Therefore, the cellulosic biofuel requirement likely added approximately $0.01 per gallon
of gasoline and diesel sold in the U.S. in 2021.91

87 CAA section 21 l(o)(7)(D).

88 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.

89 E0 and B0 respectively refer to gasoline and diesel without any renewable fuels blended in, that is, gasoline with
0% ethanol content and diesel with 0% biodiesel content. E85 refers to gasoline blended with 51%-83% ethanol.
B20 refers to diesel blended with 20% biodiesel.

90 Average D3 RIN price in 2021 according to EMTS RIN price data.

91 In the January 2022 STEO EIA forecasted gasoline and diesel consumption in 2021 at 8.79 million barrels per day
(134.5 billion gallons per year) and 3.25 million barrels per day (49.9 billion gallons per year) respectively. Dividing
the total cost of cellulosic RINs in 2021 ($1.46 billion) by the total consumption of gasoline and diesel (184.4
billion) results in an estimated cost of $0,008 per gallon of gasoline and diesel as a result of the cellulosic biofuel
requirement.

-------
Chapter 2: Baselines and Volume Scenarios

This RIA contains a collection of analyses prescribed by the CAA as well as other
analyses EPA conducted to evaluate the impacts of this rule. For analyses in which we have
quantified the impacts of the final volume requirements for 2020, 2021, and 2022 (including the
2022 supplemental standard), we have identified the specific biofuel types and associated
feedstocks that are projected to be used to meet the volume requirements, and the appropriate
baselines for comparison. While we acknowledge that there is significant uncertainty about the
types of renewable fuels that would be used to meet the final volume requirements in this rule,
we believe that the mix of biofuel types described in this chapter are reasonable projections of
what could be supplied in response to the volume requirements for the purpose of assessing the
potential impacts. We also acknowledge that the choice of baseline affects the estimated impacts
of the final volume requirements. This section describes both the methodology for identifying the
mix of biofuels that could result from the volume requirements and the baseline that we have
chosen as a reference point for comparison, as well as a discussion of alternative baselines.

2.1 Mix of Renewable Fuel Types for Final Volume Requirements

The volume requirements that we are establishing for 2020, 2021, and 2022 (including
the 2022 supplemental standard) are presented in Preamble Section III. We estimate the
constituent mix of renewable fuel types and feedstocks that could be used to meet those volume
requirements in Table 2.1-1.

-------
Table 2.1-1: Volumes Assessed for 2020, 2021, and 2022 (million RINs)

2020 Volume

2021 Volume

2022 Volume3

Cellulosic Biofuel

505

561

630

CNG/LNG derived from biogas

503

561

630

Liquid Cellulosic Biofuel

Total Biomass-Based Diesel

3,792

4,260

5,555

Biodiesel

2,886

2,726

2,780

Renewable Diesel

900

1,530

2,770

Otherb

Other Advanced Biofuels

335

227

256

Renewable Diesel

100

Imported Sugarcane Ethanol

185

110

Domestic Ethanol

Other

Total Advanced Biofuel

4,632

5,048

6,441a

Conventional Renewable Fuelc

12,493

13,787

14,189

Ethanol

12,493

13,787

14,175

Imported Renewable Diesel

Imported Renewable Diesel to
address the Supplemental Standard

250

Other

Total Renewable Fuel

17,125

18,835

20,630

Total Renewable Fuel with the
Supplemental Standard

17,125

18,835

20,880

a Includes 811 million advanced biofuel RINs in excess of the advanced biofuel volume requirement of 5,630
million RINs.

b Nearly all of this "other" BBD is jet fuel, sometimes referred to as sustainable aviation fuel (SAF). Production of
jet fuel that qualifies as BBD has been relatively stable in recent years. For 2022 we projected production of this fuel
would continue at the average volume produced from 2018-2021.

0 Does not include the Supplemental Standard.

The analyses leading to the mix of renewable fuel types and feedstocks shown in Table
2.1-1 are presented in later parts of this RIA as follows:

• Chapter 5.1: Cellulosic biofuel

• Chapter 5 .2: BBD

• Chapter 5.3: Sugarcane ethanol

• Chapter 5.4: Other advanced biofuels

• Chapter 5.5.3: C orn ethanol

The volume of conventional renewable diesel for 2022 was projected based on the
difference between the sum of our projections of all the other types of renewable fuel discussed
in Chapter 5 and our volume requirements for 2022. That is, we determined that 14 million RINs
of conventional renewable diesel would be needed in order to meet the total renewable fuel
volume requirement of 20,630 million RINs after consideration of the volumes of other
renewable fuels that can be supplied in 2022. We based this conclusion on our determination that

-------
grandfathered renewable diesel can be produced from cheaper feedstocks (e.g., palm oil) that are
not eligible to be used to produce BBD under the RFS program. The production capacity of
foreign grandfathered renewable diesel production facilities is also significant. In some previous
years when the implied conventional biofuel volume requirement could not be fully satisfied
with ethanol, we have seen some volumes of grandfathered renewable diesel to the U.S. (e.g.,
2016 and 2017). Moreover, because of domestic production limitations, we project that all of this
conventional renewable diesel would be imported and sourced from palm oil, as palm oil has
historically been the least expensive feedstock available for renewable diesel production.

2.2 Volume Changes Analyzed

In order to analyze the volumes associated with this rulemaking, EPA must first establish
a baseline for analysis. The statute does not specify a baseline for analysis, indicating that
Congress entrusted this technical determination to EPA. We considered several possible
baselines: (1) the volumes that would be consumed absent the RFS program (a "No RFS"
baseline), (2) the statutory volume targets found in CAA section 21 l(o)(2)(B), (3) the required
RFS volumes from the prior RFS annual rule (the 2020 RFS rule), and (4) actual renewable fuel
use in 2020. We ultimately selected a baseline of actual renewable fuel use in 2020. This is
equivalent to the final applicable volume requirements for 2020. We discuss our consideration of
each baseline in turn.

In assessing the impact of an action EPA's preferred baseline for comparison is generally
the world as it would exist in the absence of the rulemaking action, or a No RFS baseline.
However, in the case of the RFS program, determining the volume of renewable fuel that would
be produced and consumed in the absence of RFS volume requirements is particularly
challenging given the many other market and policy drivers for the use of renewable fuels, which
we further discuss later in this chapter. Consequently, we have not been able to determine a No
RFS baseline for this rulemaking. We have, however, included a brief description of what a No
RFS baseline for 2020-2022 might look like at the end of this section.

We also considered using the statutorily mandated volumes as the baseline for this rule.
The statutory baseline is one of two different baselines used in recent annual rules.92
Nonetheless, we deemed this inappropriate for this rule given that the statutory volumes are
infeasible and therefore any assessment of the impacts relative to them would be hypothetical
and of very limited relevance. We acknowledged this problem in recent annual rules as well.93 In
those rules, we used the statutory baseline merely to calculate what we ourselves termed
"illustrative" costs that were not intended to be comprehensive. In this rule, by contrast, EPA is
required to analyze the impacts of renewable fuels for numerous factors. While the statute does
not specify how we are to assess those factors, we found it appropriate to conduct quantitative
analyses of multiple factors, including a much more rigorous analysis of costs than the
"illustrative" analyses we performed in recent annual rules. We do not believe it is technically
sound to conduct such analysis using a baseline representing the extremely large hypothetical
statutory volume targets. For example, it would be nonsensical to suggest that this rule results in

92 See, e.g., 85 FR 7042 (2020 final rule).

93 See, e.g., 85 FR 7042 (noting that our estimate was purely "illustrative" and that "[estimating the cost savings
from renewable fuel volumes that are not projected to be produced is inherently challenging").

-------
sharp decreases in energy security or costs relative to the statutory volumes, when the statutory
volumes are impossible to achieve.

Another baseline we considered was the required RFS volumes from the most recent
prior annual rule, which in this case would be the original 2020 RFS rule published on February
6, 2020. We have also used the prior year volume requirements as the baseline in recent annual
rules, but we determined that it would not be suitable for this rule for several reasons. For one,
the 2020 rule did not specify the particular fuel types, feedstocks, and their quantities that would
be used to meet the renewable fuel requirements. It was not necessary to have this information in
previous annual standard-setting rules due to the illustrative and thus simpler analyses conducted
for those prior rules, but they are necessary in this rule in order to conduct the more detailed
analyses. Moreover, as we explain in Preamble Section III, there was a significant,
unprecedented, and unforeseen shortfall in renewable fuel use in 2020 relative to the volumes
that we required in the original 2020 final rule. Using the original 2020 rule's volume
requirements as a baseline would thus involve using much higher volumes than what we now
know, after the fact, to have actually materialized. Like the statutory baseline, this would also
result in analytical oddities, such as analyzing large decreases in renewable fuel use in 2021,
when in reality renewable fuel use in 2021 was significantly higher than in 2020. We also do not
think that using the original volume requirements established for 2020 as a baseline is
appropriate given that we are changing those volume requirements in this rule.94

Finally, we considered and adopted a baseline of actual biofuel use in 2020, which is also
equivalent to the revised volume requirements for 2020. Choosing this baseline results in zero
impacts in 2020, but significant impacts in 2021 and 2022 given the large increases in renewable
fuel use in those years relative to 2020. This approach has several pros and cons. However, on
balance, we found this approach to be most appropriate and feasible for this rulemaking. We
explain each of our considerations for use of the revised 2020 volume requirements as a baseline
below.

First, the use of actual biofuel consumption 2020 as a baseline is analytically feasible. By
contrast, as we noted above, we were not able to determine a No RFS baseline during the
timeframe for this rulemaking; the statutory volumes are impossible to achieve and thus not
amenable to rigorous analysis; and past annual rules also did not project which biofuels and
feedstocks would be used to satisfy the volume requirements. However, we have available
historical data on the types of biofuels and feedstocks actually used in 2020, and so we can
readily use this baseline for analysis.

Second, as we show in Table 2.1-1 above, the volumes we are finalizing today reflect
significant year-over-year increases during the 2020-2022 time period. Using the baseline of
actual biofuel use in 2020 thus accurately reflects the increase in the volumes over time and
provides a meaningful assessment of the impacts associated with them. We acknowledge that a
No RFS baseline would provide different, but also meaningful, assessment of the impacts of the

94 Relatedly, we also considered using a baseline of the 2019 rule volume requirements. However, some of the same
concerns apply: the 2019 final rule did not estimate which specific fuel types would be used to satisfy the
requirements, thus making analysis impractical.

-------
volumes; however, as noted above, we were not able to quantify the No RFS baseline during the
timeframe for this rulemaking.

Third, this approach is largely, but not entirely, consistent with the only prior EPA rule
when we simultaneously promulgated both retrospective and prospective standards, namely the
rulemaking that established the volume requirements for 2014, 2015, and 2016.95'96 In that rule,
the impacts of the volume requirements for 2014 and 2015 were assessed against a baseline
consisting of the applicable volume requirements for 2014 (which in that rule we were also
setting to actual consumption) and the impacts of the 2016 volumes were assessed against a
baseline of the 2015 volumes. As in the 2014-2016 rule we are in this action setting applicable
volume requirements for the first year (2020) at the actual consumption of renewable fuel in that
year and are using these volumes as a baseline. The baseline we are using for this rule is slightly
different in that we are using the 2020 actual consumption volumes as the baseline for all three
years, including for 2022. We believe this approach is preferable to the approach used in the
2014-2016 rule as it provides a consistent baseline for the volumes in all three years covered by
this rule (2020-2022). Generally, we believe that using a consistent baseline provides a clearer
perspective on the costs and impacts of the volume requirements. That is, even though the
incremental increase in volume requirements from 2020 to 2021 is similar to the increase from

2021 to 2022, using a consistent baseline accurately characterizes the fact that the impacts of the

2022 volume requirements are higher than the impacts of the 2021 volume requirements.97

Fourth, we are cognizant that analysis conducted against the 2020 baseline does not yield
the impacts caused by this rulemaking. As we note in Preamble Section III, the volume
requirements for 2020 and 2021 are entirely retrospective, such that this final rule cannot impact
biofuel use or production in those years. Nonetheless, the use of actual consumption in 2020 as
the baseline results in significant impacts for 2021. Similarly, while this rulemaking is expected
to drive significant additional renewable fuel use in 2022, much of the increases in 2022 relative
to the 2020 baseline are due to factors other than this rulemaking. We describe some of these
other factors below.

With regard to ethanol, its predominant use is as E10. As discussed in Chapter 9, since
E10 would be used regardless of the RFS program, total consumption of ethanol rises and falls
with gasoline demand, which is independent of the RFS program. Much of the increase in the
ethanol volume analyzed for 2021 and 2022 relative to 2020 is associated with the market
recovering from the impacts of the COVID-19 pandemic in 2020. Furthermore, even where the
RFS standards may directly drive increased renewable fuel use, it is unclear to what extent those

95 80 FR 77420 (December 14, 2015).

96 We acknowledge that the DRIA stated that in the 2014-16 rule, "the impacts of the volume requirements for all
three years were assessed against a baseline consisting of the actual consumption of renewable fuel in 2014." DRIA
52. That statement, as we explain above, is not correct with respect to 2016. Nonetheless, for the reasons we explain
in the text, we think that our choice of baseline is well-justified.

97 We recognize that were we to use the 2021 volumes as the baseline for analyzing the 2022 volumes, the volume
changes analyzed and resulting quantified impacts would be different. Namely, the volume changes analyzed for
2022 would be smaller, as would be the resulting costs and other impacts. We would, however, recognize such
changes as an artifact of our baseline selection, not a reflection of any substantive difference in the fuels markets or
the impacts of the volumes. As such, even if we took the same approach as in the 2014-2016 rule, we would draw
the same conclusions and reach the same final judgment with regard to the appropriate volumes.

-------
standards may drive changes in domestic biofuel production. Increased consumption in the U.S.
could come from a reduction in exports, which have been relatively high in recent years, or from
increased imports. It is also unclear to what extent additional biofuel production drives feedstock
production in the U.S. or abroad given the many factors that impact those decisions. This is
especially an issue when the feedstocks used (e.g., soybean oil) have not been the primary source
of value for the agricultural product from which it is derived (e.g., soybeans). This is particularly
an important consideration when evaluating the many statutory factors discussed in Chapter 3
that relate to impacts on crop production and associated land and water use. As a result of these
non-RFS factors and drivers, there are numerous uncertainties associated with the environmental
and economic impacts we have analyzed for the volume requirements we are establishing in this
action. The quantitative impact assessments carried out in subsequent chapters simply assume
that the standards are driving the volume changes described below.

Nonetheless we believe the quantitative analysis that we perform using the 2020 baseline
provides a meaningful assessment of the impacts associated with the 2021 and 2022 volume
requirements in the absence of a No RFS baseline. Notably, this analysis provides information on
the impact of renewable fuel use in 2021 and 2022, regardless of whether renewable fuel use in
these two years was caused by the RFS program. This information allows EPA and stakeholders
to assess how renewable fuel production and use is changing during the 2020-2022 timeframe
and how such changes may affect the environment, economy, and energy security of the U.S.
even if such changes were not driven entirely by the final standards.

Moreover, such an analysis, focusing on the impacts of renewable fuel production and
use, also complies with our statutory mandate to analyze the impacts of renewable fuels and their
production and use.98 We note that the statute does not actually specifically require EPA to
assess the impacts of the rulemaking, but rather to assess the "impact of renewable fuels" (or
their production and use) more broadly. This expressly indicates that Congress conferred on EPA
significant discretion in determining the baseline for analysis.

Finally, while we performed quantitative analysis against the 2020 actual volumes
baseline, this is by no means the only perspective for our analytical work. Much of the analyses
in this rule are qualitative, and we have considered the qualitative analyses along with the
quantitative ones in balancing the statutory factors and arriving at our final judgment of the
applicable volume requirements for 2020 - 2022. Moreover, much of the quantitative analysis
considers per unit impacts (e.g., the GHG impacts and the costs of producing and using
renewable fuels relative to the petroleum fuels they displace), which does not rely on any
particular volumes baseline.

With the above understanding, for those factors for which we quantified the impacts of
the volume requirements for 2020, 2021, and 2022, the impacts were based on the difference in
the volumes of specific renewable fuel types between the volume requirements for 2021 and
2022 and the 2020 baseline. These differences are shown in Tables 2.2-1 and 2.2-2 in terms of
both RINs and physical volumes, respectively.

98 CAA section 21 l(o)(2)(B)(ii)(I)-(VI).

-------
Table 2.2-1: Volume Changes for 2020, 2021, and 2022 Relative to the Baseline (million

RINs)

2020

2021

2022

Cellulosic Biofuel

+56

+125

CNG/LNG derived from biogas

+58

+127

Liquid Cellulosic Biofuel

-2

Total Biomass-Based Diesel

+468

+1,763

Biodiesel

-160

-106

Renewable Diesel

+630

+1,870

Other

-2

-1

Other Advanced Fuels

-108

-79

Renewable Diesel

+ 14

-5

Sugarcane Ethanol

-125

-75

Domestic Ethanol

Other

-1

Total Advanced Biofuel

+416

+1,809

Conventional Renewable Fuela

+1,294

+1,696

Ethanol

+1,294

+1,682

Imported Renewable Diesel

+14

Imported Renewable Diesel to
address the Supplemental Standard

+250

Other

Total Renewable Fuel

+1,710

+3,755

Total Renewable Fuel with the Supplemental
Standard

+1,710

+4,005

a Does not include the Supplemental Standard.

-------
Table 2.2-2: Volume Changes for 2020, 2021, and 2022 Relative to the Baseline (million
gallons)

2020

2021

2022

Cellulosic Biofuel

+56

+125

CNG/LNG derived from biogas

+58

+127

Liquid Cellulosic Biofuel

-2

Total Biomass-Based Diesel

+263

+1,029

Biodiesel

-107

-71

Renewable Diesel

+371

+1,100

Other

-1

Other Advanced Fuels

-114

-77

Renewable Diesel

-3

Sugarcane Ethanol

-125

-75

Domestic Ethanol

Other

-0

-1

Total Advanced Biofuel

+205

+1,077

Conventional Renewable Fuela

+ 1,294

+1,690

Ethanol

+ 1,294

+1,682

Imported Renewable Diesel

Imported Renewable Diesel to
address the Supplemental Standard

+147

Other

Total Renewable Fuel

+ 1,499

+2,914

Total Renewable Fuel with the Supplemental
Standard

+ 1,499

+3,061

a Does not include the Supplemental Standard.

For the purposes of our analyses, we made some simplifications to the volume changes
shown in Tables 2.2-1 and 2.2-2. We generally grouped fuels with very small changes in
volumes with similar fuels with much larger volume changes. We did this both because of the
more limited data on the impacts of those renewable fuel types with smaller volume changes, and
because we expect small volume changes to have little material impact on the analyses. These
simplifications were to assume that:

• All changes in cellulosic biofuel were CNG/LNG derived from biogas.

• All changes in "Other BBD" were renewable diesel

• All changes in advanced renewable diesel were from domestic facilities

• All changes in "Other advanced fuels" were sugarcane ethanol.

We also made the simplifying assumption for analysis purposes that the volume increases
in 2021 and 2022 will result in increases in renewable fuel production on a gallon-for-gallon
basis. For example, we have assumed that increases in corn ethanol used to meet the 2021 and
2022 volume requirements result in an equivalent increase in ethanol production. For some of the
fuel types, the impacts of increasing (or decreasing) volumes of the fuel vary depending on the
feedstock used to produce the fuel. It was therefore necessary to define the feedstocks used to
produce the renewable fuels expected to increase (or decrease) in 2021 and 2022. In cases where

-------
the vast majority of the fuel type has historically been produced from a single feedstock, as is the
case for many of the fuels we are analyzing changes in 2021 and 2022, we assumed that the
entire volume change was produced from that feedstock. The fuels that have historically been
produced from a variety of feedstocks for which we are assessing appreciable volume changes in
2021 and 2022 are biodiesel and renewable diesel that qualify as BBD. For these fuels we
projected the feedstocks that would be used to produce increasing volumes in 2021 and 2022
according to the trends we have observed in previous years." The changes in renewable fuel
production that we assessed in this RIA (after making the simplifications described above) and
the feedstocks used to produce these fuels, are shown in Table 2.2-3.

Table 2.2-3: Renewable Fuel Volume Changes and Feedstocks Analyzed Relative to the
Baseline (million gallons)

Fuel

Feedstock

Change in 2021

Change in 2022

CNG/LNG derived from biogas

Separated MSW

+58

+127

Domestic Biodiesel

Soybean Oil

-107

-71

Domestic Renewable Diesel

FOG

+40

+80

Domestic Renewable Diesel

Soybean Oil

+331

+ 1,020

Imported Ethanol

Sugarcane

-125

-75

Domestic Ethanol

Corn

+1,294

+ 1,680

Imported Renewable Diesel

Palm oil

+155a

a Includes 8 million gallons for the 2022 total renewable fuel volume requirement and 147 million gallons for the
2022 supplemental volume requirement.

In performing the quantitative analyses associated with these changes, we have generally
attempted to use the most recent public data available, while accounting for the lead time
necessary to perform the complex and multi-layered analyses supporting this rule. For the final
rule, we updated the most of our data using sources available in January 2022, including the
January 2022 STEO, MER, and WASDE reports. We acknowledge that recent events since then
(e.g., inflation, geopolitical factors, and international restrictions on palm oil exports) may affect
the U.S. agricultural and fuels markets, with potential consequences for the analyses. We have
generally not quantitatively accounted for these more recent events as doing so would require
significant revisions to the analyses and further delay this rule, during which time still further
events could occur.

We also acknowledge that in some of the analyses for this rule, we used different data
and projections than in other analyses (e.g., prices for commodities such as corn and soybean oil
or petroleum fuels such as gasoline and diesel). These differences are generally due to the fact
that different analyses required different information, and the different EPA technical staff
working on each analysis made reasonable, albeit potentially different, technical choices. We
generally preferred to use data and projections from the same or consistent sources for each of
the analyses, at the expense of (in certain cases) the use of differing data and projections across
different analyses. In some cases, a longer lead time for the analysis also required the use of
slightly older data sets. In still other cases, we found it appropriate to confirm our work by
considering more recent data sets. Where the differences were more significant, we have

99 These feedstock trends for BBD production are discussed in greater detail in Chapter 5.2.4. In assessing trends for
feedstocks used to produce BBD we considered both domestically produced BBD and imported BBD.

-------
described them explicitly in the RIA or other technical documents supporting the rule. In all
other instances, it is EPA's technical judgment that these differences do not have a sufficiently
significant impact on the analytical results to warrant changing the final volumes.

As noted above, in assessing the impact of an action EPA's preferred baseline for
comparison is generally the world as it would exist in the absence of the rulemaking action.100
Such a baseline addresses many of the challenges discussed above associated with using the
2020 volumes as the baseline. A No RFS baseline would better estimate the impacts of the action
in comparison to not taking the action, including the potential benefits, costs and impacts on
other statutory factors. Such a baseline would better assess the changes actually driven by the
rule, as opposed to merely assessing changes relative to the 2020 actual volumes, many of which
are driven by other factors as noted above. A No RFS baseline would also avoid the
complication of a comparison to previous years that are themselves impacted by past RFS
standard-setting rules. Despite these advantages, we have not been able to precisely quantify a
No RFS baseline to serve as the basis for analysis for this rulemaking due to the complex market
and regulatory dynamics associated with biofuels. Nonetheless, we present our preliminary
views on what such a baseline might look like as many aspects of the qualitative analyses in this
RIA discuss impacts relative to a No RFS baseline.

We believe that the No RFS baseline for ethanol may be approximated by using the
volumes associated with the E10 blend wall (e.g. ethanol volumes representing 10% of the
gasoline pool). In the absence of this or future RFS standard-setting rules, we believe that the
fuel market would continue to produce and blend 10 percent ethanol for a variety of economic
reasons out into the future, including demand for oxygenates and other high octane blendstocks.
Such an approach would not directly account for the relatively small volume of ethanol
consumed in higher level ethanol blends such as El 5 or E85, but it would implicitly assume
some ongoing use of these higher-level ethanol blends that is offset by the continued use of
gasoline that does not contain ethanol (E0). We expect that the vast majority of the ethanol
volume would be conventional biofuel produced from corn starch, with much smaller quantities
of sugarcane ethanol.

In regards to biodiesel and renewable diesel, we believe that the volume that would be
produced, imported, and used without any RFS mandates would be substantially lower than the
approximately 2.5 billion gallons used in 2020. Since 2010, biodiesel and renewable diesel prices
have consistently been far higher than the price of petroleum-based diesel, even after accounting
for the $1 per gallon federal tax credit. This suggests that little if any biodiesel and renewable
diesel would be driven by relative fuel prices. However, several states have adopted regulations
to incentivize the use of biodiesel and renewable diesel. These incentives range from mandates
for biodiesel and renewable diesel use, to tax credits or tax exemptions, to carbon intensity-based
incentives such as California's LCFS. While we expect that the use of biodiesel and renewable
diesel in the U.S. would decrease in the absence of the RFS program, these state incentives, in
combination with the federal tax credit, would likely result in the continued use of significant
volumes of these fuels even in the absence of the RFS program.

100 68 FR 58366 (October 9, 2003).

-------
As for other advanced biofuels and cellulosic biofuel (e.g., RNG), we expect that in the
absence of the RFS program some volumes of these fuels would continue to be used, but likely a
lower volume than was used in 2020. State programs such as California's LCFS and Oregon's
Clean Fuels Program are expected to support the continued use of advanced and cellulosic
biofuels, especially in light of the generally low carbon intensities assigned to these fuels, but
volumes would likely decrease without the significant financial incentive provided by the RFS
program.

-------
Chapter 3: Environmental Impacts

The statute requires EPA to analyze a number of environmental factors in its
determination of the appropriate volumes to establish under the reset 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 Report to Congress on Biofuels, which provides additional information on
environmental impacts.101

3.1 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.102 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.103

In addition to the type of biofuel, other factors affect air quality, including but not limited
to the blend level, the vehicle technology, emissions control technology, and operating
conditions. 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 in favor of higher or
lower volumes. This conclusion is based on analyses of impacts from much higher volumes. For
example, EPA's "anti-backsliding study" (ABS), required under CAA section 21 l(v)(l),
examined the impacts on air quality as a result of changes in vehicle and engine emissions
resulting from required renewable fuel volumes under the RFS, relative to approximately 2005
levels.104Hoekman et al. (2018) also reviewed available literature on potential air quality impacts
for E10 versus E0 across the entire lifecycle.105 Both studies found potential increases and

101 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=34.1.49.1.

102 EPA (2022). "Health and environmental effects of pollutants discussed in Chapter 3 of Regulatory Impact
Analysis (RIA) supporting RFS Annual Rules for 2020, 2021 and 2022." Memorandum from Rich Cook to Docket
No. EPA-HQ-OAR-2021-0324, May 26, 2022.

103 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.

104 EPA (2020). Clean Air Act Section 211(V)(1) Anti-Backsliding Study.
https://nepis.epa. gov/Exe/ZvPDF.cgi?Dockev=P100ZBYl.pdf

105 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.

-------
decreases in ambient pollutant levels of pollutants, but none of them were large, even when they
considered much greater changes in biofuel volumes than are being finalized in this rule.

Table 2.2-3 summarizes the changes in renewable production volume assessed for this
rule. The largest volume changes are for corn ethanol and renewable diesel, primarily produced
from soybean oil, with a small volume from fats, oil and grease (FOG) and biogas. The
discussion below focuses on potential impacts for these fuel/feedstock combinations.

3.1.1 Corn Ethanol Emissions

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 cumulative impacts of processes upstream of
emissions from corn ethanol. A 2009 analysis using the GREET model concluded that criteria
pollutant emissions for corn ethanol production are substantially higher than for gasoline on a
mass per gasoline equivalent gallon basis.106 A significant source of upstream emissions from
corn ethanol is production facilities.107'108 Table 3.1.1-1 summarizes corn ethanol plant
emissions, using data from the 2017 National Emissions Inventory (NEI) where available.109 For
facilities not found in the 2017 NEI, we used data from the 2016 emissions modeling platform
version 1110 or inventory estimates using facility-level volume data from the EPA moderated
transaction system.111 Only a few plants used coal or coal in combination with other energy
sources, although the wet mill plants contributed disproportionately to emissions, especially
sulfur dioxide.

106 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://ecommons.cornell.edu/bitstream/handle/1813/46218/scope. 1245782010.pdf?sequence=2

107 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://cfpub.epa.gov/si/si_public_record_report.cfm?Lab=IO&dirEntryId=341491

108 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/10.1002/2015JD023138

109 https://www.epa.gov/air-emissions-inventories/2017-national-emissions-inventory-nei-data

110 https://www.epa.gov/air-emissions-modeiing/2016vl-platfonii

mhttps://www.epa.gov/fuels-registration-reporting-and-compliance-help/reporting-rfs-rin-transactions-epa-
moderated

-------
Table 3.1.1-1: Pollutant Emissions (short tons) From Biodiesel and Corn Ethanol

Biorefineries in

J.S. in 20

Finished
Fuel

Number
of

Facilities

MI;

NOx

PM10

PM25

SO2

VOCs

Corn

Ethanol

(total)

176

7362.8

278.7

9045.5

5218.7

4088.5

1854.4

8908.7

Coal; Dry
Mill

75.3

55.8

20.7

20.0

n.a.

39.7

Coal; Wet
Mill

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

251.8

9.0

142.2

184.5

102.7

74.7

270.5

Unknown;
Unknown

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-emissions-inventories/2017-national-emissions-inventory-nei-
data and EPA 2016 version 1 modeling platform (https://www.epa.gov/air-emissions-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
during storage and transport. The largest emission contribution is for VOC due to evaporation.
Table 3.1.1-2 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.112 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.113

112 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.eom/science/article/pii/Sl.35223.l.010007375

113 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), 4385-4397. https://doi.org/10.1002/2015JD023.1.38

-------
Table 3.1.1-2: Emissions From Transportation of Et

lanol (short 1

ons)

nh3

NOx

PMio

pm25

so2

voc

4,225

19,270

630

533

340

660,674

Source: EPA 2016 version 1 modeling platform (lit!ps://www.epa.gov/air-emissions-modeling/20.1.6v.1.-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 to
estimate the impacts associated with ethanol production and transport (Table 3.1.1-3). In doing
so, we assumed that additional ethanol use in the U.S. is associated with ethanol production in
the U.S. We describe this assumption further in Chapter 2. However, we note that volumes used
domestically could be sourced from imports. Thus, it is unclear what overall impacts would be
on domestic production and therefore 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.

Table 3.1.1-3: Pollutant Emission Impact Estimates for Production and Transport of Corn

Ethanol of the 2021 ant

2022 Ethanol Volumes (short

tons), Relative to a 202

0 Baseline

nh3

NOx

PMio

pm25

so2

VOC

Biorefinery
Emissions

7,142

270

8,775

5,062

3,996

1,7984

8,643

Transport
Emissions

4,225

19,270

630

533

340

660,674

Total

Emissions

11,588

305

28,316

5,849

4,622

2194

669,583

Impacts per
Million
Gallons
Ethanol11

0.75

0.02

1.83

0.38

0.30

0.14

43.45

2021

Annual Std
Impacts

715

1,744

362

286

133

41,407

2022

Annual Std
Impacts

1,102

2,688

558

440

206

63,828

a Emissions per million gallons ethanol is calculated using total domestic ethanol production in 2016 as reported in
the EI A Monthly Energy Review (15.41 billion gallons)

After distribution to the retail outlet stations, end use at the vehicle occurs. This step
includes both evaporative losses during dispensing the fuel, and exhaust emissions from
combustion during vehicular use. 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 Tier 2 and Tier 3 compliant

-------
vehicles).114'115'116'117 However, because the projected changes in volume of ethanol are much
smaller than the total amount of fuel consumed across the country, and given the magnitude of
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 minor increases in E15 and E85 use, as we discuss in Chapter 5.5.3, and thus
emission changes due to increased use of these fuels are also anticipated to be very minor.

3.1.2 Biodiesel/Renewable Diesel Emissions

Although biodiesel is sourced from a variety of feedstocks, domestic soybean and
domestic FOGs made up nearly 70% of the biodiesel in 2019, with most of that being domestic
soybean. Data are lacking on emission and air quality impacts of either soybean 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).118

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 are expected to be much lower than
for soybean biodiesel.

Table 3.1.1-1 provides estimated emissions from biodiesel refineries in the U.S. End-use
emissions when biodiesel is burned in diesel engines may be different from petroleum diesel
fuel, primarily in older engines. EPA's MOVES3 model assumes no emission impacts of
biodiesel fuel for engines meeting 2007 and later standards. However, the model does estimate
criteria pollutant emission impacts for pre-2007 engines running on 20 volume percent soybean-
based biodiesel (Table 3.1-1; EPA, 2020, Tale 8-1).119 The biodiesel effects implemented in

114 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).

115 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.

116 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/2019/05/CRC~E~94~3_Final~Report_2018-06-26.pdf

117 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

118 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

119 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 1010M6C.pdf

-------
MOVES are obtained from an analysis conducted in the course of the 2010 Renewable Fuel
Standard Program.120

Table 3.1.2-1: 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 Name

Percent Change in Emissions

THC (Total Hydrocarbons)

-14.1

-13.8

NOx

+2.2.

PM2.5

-15.6

Given the limited impact of this rule on biodiesel use, national-scale impacts are expected
to be small. However, there could be localized impacts. Impacts in years past (i.e., prior to the
turnover of the fleet to 2007 and newer engines) may have been larger. Hums et al (2016)
compared life cycle emissions for low sulfur diesel (LSD), soybean-based biodiesel, and grease
trap waste (GTW) based biodiesel.121 This study relied on GREET-2014 for soybean-based
biodiesel impacts.122 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.

While biodiesel is the predominant advanced biofuel used in diesel engines, renewable
diesel is projected to increase significantly during the 2021/22 timeframe. Though research on
renewable diesel impacts is limited, research in California concluded that it reduced vehicle
exhaust emissions of PM, NOx, hydrocarbons, and CO. Toxics test results also show reductions
in most PAHs and VOCs. 123>124 However, these data were obtained on only two pre-2007
engines. There are also likely to be air quality impacts from renewable diesel production. Figure
5.2.2-1 provides new renewable diesel production capacity in the U.S. through 2022. 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. Thus, given the limited research available on renewable diesel

120 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).

121 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/10.1021/acs.est.5b02667

122 Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation Model; Argonne National
Laboratory: Argonne, IL, 2014. https://greet.es.anl.gov/

123 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

124 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, 307-314.

-------
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.

This rule also projects increases in biogas use. We are not aware of research comparing
life cycle air pollution emissions of biogas versus compressed natural gas and have also not been
able to quantify such impacts associated with this rule.

3.1.3 Air Quality Modeling

As mentioned at the beginning of Chapter 3.1, previous air quality analyses based on
much larger volumes indicate that air quality impacts from this rule at the geographic scale used
in photochemical air quality modeling are expected to be minimal. Thus, no air quality modeling
was done. In general, fossil fuels are assumed to be displaced by biofuel use due to the
renewable fuel standard. These fossil fuels also produce air emissions at the production,
distribution, and combustion stages. Relative to emissions from biofuels, emissions from fossil
fuels differ based on the fuel, production process, transport, and other factors. Geographic
distribution of emissions also varies, and a comprehensive evaluation of offsetting impacts is
very complex. Furthermore, to the extent that the 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.

3.2 Climate Change

The statute directs EPA to analyze "The impact of the production and use of renewable
fuels on the environment, including on.. .climate change..." This rulemaking may affect climate
change by altering the amount of greenhouse gas (GHG) emissions to the atmosphere. This
section discusses our evaluation of the potential effects of this rulemaking on GHG emissions.

3.2.1 GHG Emissions

One of the goals of the RFS program is to reduce greenhouse gas (GHG) emissions by
replacing petroleum-based transportation fuels such as gasoline and diesel with renewable fuels
such as ethanol and biodiesel. Renewable fuels composed of biogenic carbon recently
sequestered from the atmosphere reduce GHGs and mitigate climate change if their use displaces
petroleum derived fuels, provided that the full lifecycle GHG emissions associated with biofuels
do not exceed those of the displaced petroleum fuels.

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 approved feedstocks and
have lifecycle GHG emission that are at least 20% less than the baseline petroleum-based
gasoline and diesel fuels.125 Advanced biofuels and biomass-based diesel are required to have
lifecycle GHG emissions that are at least 50% less than the baseline fuels, while cellulosic

125 See 42 USC 7545(o)(l), (2)(A)(i)

-------
biofuel is required to have lifecycle emissions at least 60% less than the baseline fuels.126
Congress also allowed for facilities that existed or were under construction when EISA was
passed to be grandfathered into the RFS program and exempt from the lifecycle GHG emission
reduction requirements.127

In the early years of the RFS program the vast majority of the biofuel volume
requirement was comprised of renewable fuel that was not required to be advanced biofuel.128 In
the latter years of the program the vast majority of the growth in the volume obligations were in
the cellulosic biofuel and advanced biofuel categories, which are required to achieve greater
GHG emission reductions. For instance, over the time period covered by this rulemaking, 2020-
22, the statute provides for an 8 billion gallon increase in total renewable fuel relative to 2019,
the entirety of which must come from advanced biofuel.129

In the March 2010 RFS2 rule (75 FR 14670), EPA estimated lifecycle GHG emissions of
different biofuel production pathways; that is, the emissions associated with the production and
use of a biofuel, including indirect emissions, on a per-unit energy basis. Since the RFS2 rule,
EPA has also conducted numerous (over 140) analyses of new pathways and their lifecycle GHG
emissions. Over time, EPA has approved additional pathways for participation in the RFS
program. These pathways rely on novel feedstocks (e.g., camelina oil,130 distillers sorghum
oil131) and novel production processes involving existing feedstocks (e.g., catalytic pyrolysis and
upgrading of cellulosic biomass,132 gasification and upgrading of crop residues133).

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."134 Our lifecycle analyses of various pathways are also published online.135 A list of
pathways that have been approved by regulation can also be found at 40 CFR 80.1426(f)(1).
Today's rulemaking also expands the pathways that can participate in the RFS program to
include renewable fuels produced from biointermediates, as discussed further in the preamble to
this rulemaking.

126 See 42 USC 7545(o)(l)(B)-(E)

127 42 USC 7545(o)(2)(A)(i).

128 42 USC 7545(o)(2)(B)(i)

129 Compare 42 USC 7545(o)(2)(B)(i)(I), with id. (II). In Preamble Section III.A, we further discuss the potential
GHG benefits of different types of biofuels.

130 March 2013 Pathways I rule. 78 FR 14190. https://www.epa.gov/renewabte-fiiet-standard-program/finat-Rite-
additional-analifving-renewable-fiiel-pathwavs-under

131 August 2018 sorghum oil rule. 83 FR 37735. https://www.gpo.gov/fdsvs/pkg/FR-2018-08-02/pdf/2018-
16246.pdf

132 March 2013 Pathways I rule. 78 FR 14190. https://www.epa.gov/renewable-fiiel-standard-program/final-riile-
additional-qualify ing-renewable-fiiel-pathways-under

133 "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

134 This document is available on EPA's website at: https://www.epa.gov/fiiels-registration-reporting-and-
compliance-help/lifecvcle-greenhouse-gas-resiilts. This summary is also available in docket EPA-HQ-OAR-2021-
0324.

135 See https://www.epa.gov/renewable-fiiel-standard-program/approved-pathwavs-renewable-fiiel and
https://www.epa.gov/renewable-fiiel-standard-program/other-actions-renewable-fiiel-standard-program

-------
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 will 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 virgin vegetable oils, such as soybean or canola
oil.136 In addition, most cellulosic biofuels—which are required to meet the highest statutory
lifecycle GHG reduction threshold of 60%—are currently produced from wastes, residues, or by-
products, including landfill biogas.137

The quantity of energy used to produce the renewable fuel and the source of this energy
also affects its lifecycle GHG emissions. For instance, the use of natural gas, biomass, or biogas
for process energy can potentially result in lower lifecycle GHG emissions than the use of coal or
oil for process energy. This can help ensure that a biofuel production process meets certain
lifecycle GHG reduction thresholds.138

In the past decade, estimation of biofuel lifecycle GHG emissions has been an active area
of research. Despite extensive research on the lifecycle GHG emissions of biofuels since the start
of the RFS program, considerable uncertainty regarding the GHG emission impacts of renewable
fuel use remains, and fuel-specific lifecycle emissions estimates published in prior EPA actions
remain within ranges found in more recent studies.139 Sources of uncertainty in biofuel lifecycle
emissions are discussed below.

136 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/fuels-registration-reporting-and-
compliance-help/lifecycle-greenhouse-gas-results.

137 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.

138 See, e.g., 40 CFR 80.1426(f)(1) (describing pathways that require the use of natural gas, biomass, and/or biogas
for process energy, e.g., at rows A-C).

139 The following references report a significant range of lifecycle GHG estimates based on modeling, data and other
uncertainties: CARB (2014). Detailed Analysis for Indirect Land Use Change. California Air Resources Board.
Sacramento, CA: 113. https://www.arb.ca.gov/fiiels/lcfs/ilnc assessment/itue analvsis.pdf: Carriquirv. M., et al.
(2019). "Incorporating Sub-National Brazilian Agricultural Production and Land-Use into U.S. Biofuel Policy
Evaluation." Applied Economic Perspectives and Policy, doi: 10.1093/aepp/ppy033; 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; Dunn, J. B., et al. (2013). "Land-use change and greenhouse gas
emissions from corn and cellulosic ethanol." Biotechnology for Biofuels 6(1): 51; ICAO (2021). CORSIA Eligible
Fuels - Life Cycle Assessment Methodology. Volume 3. March 2021. 155 pages.
https://www.icao.int/enviromnental-

protection/CORSIA/Documents/CORSIA Supporting Document CQRSIA%20Eligibte%20Fiiels LCA Metfaodot
ogv V3.pdf: Laborde. D., et al. (2014). Progress in estimates of I LUC with MIRAGE model. JRC Scientific and
Policy Reports. Italy, European Commission Joint Research Centre Institute for Energy and Transport: 46.

-------
Estimates of the lifecycle GHG emissions associated with an increase in biofuel use are
often broken down into different components, which can be categorized into direct and indirect
emissions. Direct emissions are the emissions directly attributable to the production and use of
renewable fuel, including emissions throughout the supply chain, from agricultural production
for feedstock crops, feedstock and fuel transport, energy use for fuel production, and fuel
combustion. Indirect emissions are the emissions that occur as a consequence of indirect effects
of increased use of renewable fuel. These effects are often mediated through prices of goods
which are traded in global markets.

While direct emissions can, in theory, be directly measured, data availability and
measurement methodologies can introduce uncertainty in estimates of direct emissions. For
example, accurately estimating the emissions attributable to production of feedstock crops
requires detailed data on not only the energy and chemical inputs of agricultural production, but
on soil organic carbon (SOC) and carbon fluxes on cropland. There can be significant
heterogeneity in SOC measurements even within small geographic areas, and when combined
with different land management practices, estimates of carbon sequestered in or emitted from
cropland vary widely. Furthermore, in order to estimate SOC changes directly attributable to
feedstock production, a baseline SOC needs to be defined representing what the SOC would
have been in the absence of the biofuel feedstock production. Cropland tends to be depleted in
SOC relative to unmanaged land.140 Cropland taken out of production and allowed to revert to
grassland or forest would sequester CO2 at different rates compared with cropland left in
production.141 Certain land cover types are better carbon sinks than others. Certain agricultural
practices can also allow for more sequestration of CO2 and/or mitigation of CO2 release to the
atmosphere, either by increasing the carbon inputs into the soil, or decreasing the decomposition

http://www.doc-transition-energetiaue.info/GED CDM/194819191209/ifpri-irc report.pdf: Lewandrowski, J., et al.
(2019). "The greenhouse gas benefits of corn ethanol - assessing recent evidence." Biofuels: 1-15; 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; Qin, Z., et al. (2016). "Influence of spatially
dependent, modeled soil carbon emission factors on life-cycle greenhouse gas emissions of corn and cellulosic
ethanol." Global Change Biology Bioenergy 8(6): 1136-1149; Scully, M. J., et al. (2021). Carbon intensity of corn
ethanol in the United States: state of the science. Environmental Research Letters, 16(4), doi: 10.1088/1748-
9326/abde08; 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; Valin H, P. D., et al.
(2015). The land use change impact of biofuels consumed in the EU: Quantification of area and greenhouse gas
impacts. Utrecht, Netherlands, Ecofys: 261; Woltjer, G., et al. (2017). Study Report on Reporting Requirements on
Biofuels and Bioliquids Stemming from the Directive (EU) 2015/1513, European Commission: 124.
https://dspace.librarv.uu.nl/handle/1874/358650; CA LCFS summary of current fuel pathways.xlsx" in docket EPA-
HQ-OAR-2021-0324. Retrieved on 6/11/2021 from https://ww2.arb.ca.gov/resonrces/documents/lcfs-pathwav-
certified-carbon-intensities: Spawn-Lee, S. A., Lark, T. J., Gibbs, H. K., Houghton, R. A., Kucharik. C. J., Malins,
C., Pelton, R. E. O., & Robertson, G. P. (2021). Comment on 'Carbon Intensity of corn ethanol in the United States:
state of the science. Environmental Research Letters, 16(4). doi:10.1088/1748-9326/ac2e35; Lark, T. J., Hendricks,
N. P., Smith, A., Pates, N., Spawn-Lee, S. A., Bougie, M., Booth, E. G., Kucharik, C. J., & Gibbs, K. G. (2022).
Environmental outcomes of the US Renewable Fuel Standard. Proceedings of the National Academy of Sciences
119(9) e2101084119. doi:10.1073/pnas.2101084119.

140 Paustian, K., et al. (2019). Soil C Sequestration as a Biological Negative Emission Strategy. Frontiers in Climate,
1. https://doi.org/10.3389/fclim.2019.000Q8

141 Post, W.M. and Kwon, K.C. (2000). Soil carbon sequestration and land-use change: processes and potential.
Global Change Biology, 6: 317-327. https://doi.Org/10.1046/i.1365-2486.2000.00308.x

-------
rate of SOC.142 As is the case with other categories of emissions, uncertainty is amplified when
extending estimates based on local measurements, such as measurements of SOC, to larger
geographic regions or timeframes.

Indirect impacts, by definition, cannot be directly measured, and are often global in
nature.143 Indirect effects are often mediated through prices of goods which are traded in global
markets. Examples include the effects of increased crop demand and coproduct supply on animal
feed markets and the consequent effects on livestock and crop production emissions, price effects
on energy markets including "rebound" effects,144 and increased agricultural product demand
driving expansion of cropland into natural land types. Despite significant research on the
lifecycle emissions of biofuels since the start of the RFS program, uncertainty regarding direct
and indirect emissions associated with biofuel use remains.

Estimating indirect categories of emissions is particularly challenging since indirect
emissions are driven by market behaviors. Quantifying indirect emissions caused by land use
change (LUC) has been an especially active area of research over the last decade, with methods
typically relying on statistical studies and comparing simulated scenarios.145 LUC emissions
have the potential to constitute a large portion of the total emissions attributable to crop-based
biofuels, but such estimates also have significant uncertainty.146 Given that biofuel feedstocks
are produced around the world in a wide variety of countries, climates, and market conditions,
and given that these feedstocks have many different uses with many different substitutes, there
are significant uncertainties associated with estimating the indirect impacts of biofuel
consumption or biofuel policies.

The time period over which land use change emissions are quantified also influences
whether renewable fuel use is estimated to increase or decrease net GHG emissions relative to
petroleum fuels. For example, if increased demand for biofuels leads to land conversion, an
initial pulse of emissions would likely be released in the first year, and there would also be
foregone sequestration over time based on the carbon that would have been sequestered through
plant growth absent the biofuel induced land conversion.147 Over time, if the biofuel production
continues, the GHG benefits of displacing fossil fuels may eventually "pay back" the initial

142 Paustian, K., et al. (2019). Soil C Sequestration as a Biological Negative Emission Strategy. Frontiers in Climate,
1- https://doi.org/10.3389/fclim.2019.000Q8

143 National Research Council 2011. Renewable Fuel Standard: Potential Economic and Environmental Effects of
U.S. Biofuel Policy. Washington, DC: The National Academies Press, https://doi.org/.1.0.17226/13.1.05

144 Rajagopal, D., Hochman, G., and Zilberman, D. (2011), Indirect fuel use change (IFUC) and the lifecycle
environmental impact of biofuel policies. Energy Policy, v. 39, iss. 1, 6 p.

145 Plevin, R. J. (2017). Assessing the Climate Effects of Biofuels Using Integrated Assessment Models, Part I:
Methodological Considerations. Journal of Industrial Ecology, 21(6), 1478-1487. doi:10.1111/jiec. 12507

146 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.

147 The initial pulse of emissions may actually 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.

-------
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.

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. In this rulemaking, EPA is not reopening the related aspects of the 2010 RFS2
rule or any prior EPA lifecycle greenhouse gas analyses, methodologies, or actions. That is
beyond the scope of this rulemaking. With respect to estimating the GHG impacts of this
rulemaking specifically, the application of a 30-year time period may or may not be the most
appropriate analytical time period over which to evaluate the impact of a rule that covers only
three years of volume requirements. On one hand, the annual volume rules are 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, it seems likely that land will continue to be used for agricultural
purposes in the future for a period of time. On the other hand, the volumes in this rule do not
extend beyond 2022, and making projections about future policies, volume requirements, and
renewable fuel use are inherently uncertain. In the illustrative GHG scenario presented in
Chapter 3.2.2, the analyses used make the assumption 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.

The guidance to federal agencies on development of regulatory analysis discusses the
importance of quantifying the benefits of regulatory actions to the extent possible.148 We
recognize this guidance and agree that providing a quantitative estimate of the effects of this
final rule on GHG emissions could provide valuable information. However, quantifying the
GHG benefits of the final renewable fuel volumes includes a number of complexities. In addition
to the uncertainties discussed above related to estimating the GHG emissions associated with
individual biofuel pathways, there are additional complexities associated with estimating the
GHG emissions associated with producing multiple biofuels concurrently (e.g., crop-based
biofuels compete for land and other resources). Extensive modeling would be necessary to
accurately address the uncertainties discussed above and the complexities associated with
evaluating the GHG impacts associated with setting standards for multiple biofuels concurrently.
Due to limited time and resources available for undertaking this analysis, we are unable to
perform such extensive modeling of the GHG effects of the final volumes. For these reasons, we
are not presenting modeled estimates of the GHG impacts of the combined volumes in this final
rule. Nevertheless, in order to provide some indication of the potential impacts, in the following

148 Office of Management and Budget (OMB). Circular A-4. September 17, 2003.

-------
section we include an illustrative analysis of GHG emissions that represent a potential stream of
annual GHG emissions, using existing EPA lifecycle analyses for each individual fuel based on
lifecycle analysis framework developed for the March 2010 RFS2 rule.

Assessing the greenhouse gas impacts of biofuels, including analyzing the lifecycle GHG
emissions of individual fuels, is an area of ongoing research. The fact that considerable
uncertainty exists in lifecycle analysis (LCA), particularly with respect to indirect or induced
LUC impacts, does not mean that further progress on research in this area cannot be made. EPA
recognizes that subsequent to the publishing of our original lifecycle GHG emissions estimates
in 2010, various academic and institutional studies on this topic have been published - in
addition to work that EPA has done - and that new and relevant data is now available.
Accordingly, EPA has initiated work to develop a revised modeling framework of GHG impacts
from biofuels. The modeling framework could potentially be used to assess other impacts too.
EPA will share more information about the framework and its development when appropriate.
Our next assessment will conduct a robust modeling comparison exercise and present results
from that exercise. Following that, EPA will finalize the framework and use it in future
assessments.

In order to help inform future assessments, EPA held a public workshop on the GHG
impacts of land-based biofuels on February 28th and March 1st, 2022.149 At this workshop,
speakers within and outside of the federal government presented on available data, models,
methods, and uncertainties of assessing the GHG impacts of land-based biofuels. In tandem with
the workshop, EPA opened a public docket (independent from the docket for this rulemaking)
and sought comment on the following questions: 1) what sources of data exist and how can they
be used to inform the assumptions that drive GHG estimates; 2) how best to characterize the
sources of uncertainty associated with quantifying the GHG emissions associated with biofuels;
and 3) what model(s) are available to evaluate the lifecycle GHG emissions of land-based
biofuels, and do the model(s) meet the Clean Air Act requirements for quantifying the direct and
significant indirect emissions from biofuels. EPA continues to review the information discussed
at this workshop and submitted to the public docket, and will continue to engage with
stakeholders outside of this specific rulemaking action on how best to improve future
assessments of the GHG impacts of biofuels.

3.2.2 Illustrative GHG Scenario

As noted above, EPA has not provided a primary estimate of the greenhouse gas impacts
of the standards in this rulemaking. Rather, this section provides an illustrative assessment of the
potential GHG impacts of a set of assumed biofuel volume changes, and under other
assumptions, as described below. This illustrative scenario is not EPA's assessment of the likely
greenhouse gas impacts of this rulemaking. As the reset provision does not require EPA to
quantify climate impacts, this illustrative scenario is also not strictly needed to comply with our
statutory obligations. Nor is it needed to support the qualitative analysis set forth above or in the
preamble. While this section does further our analysis of the impact of the production and use of
renewable fuels on climate change, we would reach the same judgment as to the appropriate

149 The workshop agenda, information about the associated docket, and presented slides are available on EPA's
website at https://www.epa.gov/renewable-fiiel-standard-program/workshop-biofiiel-green.honse-gas-modeling.

-------
volumes regardless. We are chiefly providing this illustrative GHG scenario for the purpose of
providing useful information to the public regarding climate change and to comply with E.O.
12866.

For this illustrative scenario, we rely on previously published EPA biofuel lifecycle GHG
emissions estimates and straightforward adjustment of these published analyses where necessary.
The specific biofuel volumes and methods of assessment for each fuel / feedstock combination
are discussed in Chapter 3.2.2.1. EPA's existing lifecycle GHG assessments consider a 30-year
time horizon following an initial change in biofuel production in order to assess temporal
dynamics of land use change emissions.150 In applying these analyses, we assume that for the 30
years following a change in standards, those levels of biofuel consumption are maintained. For
this illustrative scenario we assess the standards in each year, 2021 and 2022, separately, and
then present the effect on GHGs in Chapter 3.2.2.2 as combined streams of impacts from 2021
through 2051. In Chapter 3.2.2.3 we present the implied annual global social benefit for each
year of emissions. All calculations are available in a spreadsheet in the docket for this
rulemaking.151

For this illustrative scenario, we evaluate the simplified biofuel volume changes for
biofuels and volumes presented in Table 2.2-3. This includes assumed changes in corn and
sugarcane ethanol, changes in domestic soybean oil biodiesel and soy- and FOG-based
renewable diesel, changes in grandfathered renewable diesel, and changes in CNG/LNG derived
from biogas. Assumptions about the fuels and quantities analyzed are discussed in Chapter 2.2.
We do not include quantified GHG impacts for the projected increase in grandfathered
renewable diesel in 2022 for either the annual or supplemental standards. We note that this
grandfathered volume may be produced from a mix of feedstocks, as determined by market
conditions, for which we have not published lifecycle GHG estimates.152 For these reasons we do
not quantify the GHG impacts of the grandfathered renewable diesel volume as part of this
illustrative scenario.

For this illustrative scenario, each biofuel volume change is assumed to displace an
energy-equivalent quantity of either baseline gasoline or baseline diesel fuel.153 These volumes
and displacements are carried forward for 30 years following the initial change in volume; that

150 As noted in Chapter 3.2.1, the 30-year time horizon used for biofuel lifecycle GHG emissions was determined
after considering public comments and the input of an expert peer review panel in the March 2010 RFS2 rule (75 FR
14670). In this rulemaking, EPA is not reopening the related aspects of the 2010 RFS2 rule or any prior EPA
lifecycle greenhouse gas analyses, methodologies, or actions.

151 See "GHG Scenario for 2020-22 RVO Rule.xlsx," available in the docket for this rule.

152 Grandfathered fuels also do not have a GHG reduction requirement under the RFS program, increasing the
uncertainty of the GHG emissions of these fuels.

153 We note that this assumption is a source of uncertainty. Increased production and use of renewable fuel may not
result in a permanent reduction in an energy equivalent amount of conventional fossil fuels. The RFS's impact on
conventional fossil fuel extraction and refining is necessarily indirect: to the extent the RFS causes conventional
fossil fuel use to decline, it will depress prices for gasoline and diesel, which would be expected to reduce petroleum
prices. Faced with this price reduction, petroleum extraction and refining would be expected to decrease. However,
consumption of fossil fuels, particularly outside the U.S. where the RFS does not apply, would be expected to
increase in response to lower prices. We have not estimated the net effect of these market mediated indirect effects.
Instead, we use the simplifying assumption of energy equivalent displacement, as has been done in EPA's past
biofuel lifecycle GHG analyses.

-------
is, for the 2021 standard we account for the effects of the change in biofuel volumes from the

2020 baseline through 2050, and we account for the effects of the 2022 standard compared to the

2021 standard through 2051.154 We then sum these impacts and account for the effects of both
standards over a thirty year time horizon. Annual changes in emissions are calculated by
comparing the emissions attributable to the increased biofuel consumption with the displaced
emissions associated with consumption of a baseline petroleum-based fuel.

3.2.2.1 Carbon Intensity Analyses Used

EPA evaluates the lifecycle GHG emissions associated with biofuel pathways to
determine which pathways satisfy the statutory GHG reduction requirements to qualify for the
RFS renewable fuel categories. These lifecycle GHG estimates are often presented as grams of
CChe emissions per megajoule of additional renewable fuel consumed. These estimates are often
called the "carbon intensity" or "CI" of the fuel. As described in more detail in the March 2010
rule, EPA's CI estimates for agriculture-based fuels are calculated over 30 years, assuming a
fixed volume increase of biofuel consumption persists over those 30 years. Under these
assumptions, land converted to agricultural use in Year 1 as a result of an assumed Year-1
increase in biofuel consumption would remain in agricultural production for the ensuing 29
years, and the assumed volumes of biofuels would continue to displace fossil-based fuels for the
same time period.155 Thus, EPA's methodology for calculating CI values sums the land use
change and other emissions over a 30-year period, and then divides those emissions by 30 years
of renewable fuel production volumes to calculate an averaged emission factor per megajoule of
renewable fuel consumed. In the discussion of each fuel in the following sections, we provide the
estimated CI value as a basis for comparing EPA's past evaluations of different biofuels on a
per-megajoule basis. For this illustrative scenario, we rely on the same modeling used to evaluate
the CI of individual fuels. However, instead of applying the CI values directly, we show the
annual stream of emissions from which each of the CI estimates was computed. We do this for
two reasons: 1) because the social cost of greenhouse gases (SC-GHG, see Chapter 3.2.2.3.1)
differs based on the year in which a gas is emitted, providing a monetized estimate requires an
annual stream of emission; and 2) using EPA's CI values to compare renewable fuels implicitly
assumes that a 30-year time horizon is the relevant period over which to analyze that fuel.
Presenting the annual stream of emissions makes these assumptions explicit to the reader.

Baseline Gasoline and Diesel

For this analysis, we used the baseline assessments of 2005 gasoline (93.1 gCChe/MJ)
and 2005 diesel (91.9 gCChe/MJ) which were used, as required by the Energy Independence and
Security Act of 2007, in EPA's assessment of biofuel lifecycle GHG emissions percent
reductions for the March 2010 RFS2 rule (75 FR 14670).156 For comparison, Cooney et al. 2017

154 This is the same baseline we used for other analyses in the RIA, and consistent with the baselines used in the
illustrative cost analyses in previous RFS annual rules. As discussed further in Chapter 2.2, this baseline is different
than current guidance to federal agencies on the development of regulatory analysis states that a baseline "should be
the best assessment of the way the world would look absent this action.

155 75 FR 14670 (March 26, 2010)

156 U.S. EPA, (2010). 2005 Petroleum Baseline Lifecycle GHG (Greenhouse Gas) Calculations. U.S. Environmental
Protection Agency, EPA-HQ-OAR-2005-0161-3151. Washington DC, January. Available at
https://www.regulations.gov/document/EPA-HO-QAR-2005-0 lo I .1 I

-------
reported emissions intensities of 98.1 gCChe/MJ for gasoline and 94.1 gCChe/MJ for diesel
based on an updated assessment of the 2005 baseline, and estimated 2014 average emissions
intensities of 96.2 gCChe/MJ for gasoline and 92.0 gCChe/MJ for diesel.157 Based on GREET-
2021 we calculated intensities of 91.7 gCChe/MJ for average U.S. gasoline and 89.6 gCChe/MJ
for average U.S. diesel.158'159 While a source of some uncertainty, these results are all relatively
similar, and bracket the estimates used in the March 2010 RFS2 rule.

Starch Ethanol, Sugarcane Ethanol & Soybean Biodiesel

To assess the impacts of volume changes of corn starch ethanol, we directly use EPA's
prior emissions intensity analysis published in the March 2010 RFS2 rule for projected corn
starch ethanol, assuming that all such ethanol is produced using a natural gas fired dry mill
process.160'161 For sugarcane ethanol and soybean oil biodiesel, we similarly use EPA's prior
emissions intensity analyses published in the March 2010 RFS2 rule.162 These RFS2 analyses
resulted in emissions intensity estimates of 73.2 gCChe/MJ for corn starch ethanol, 36.1
gCChe/MJ for sugarcane ethanol, and 40.0 gCChe/MJ for soybean oil biodiesel. These RFS2
analyses estimated annual emissions per megajoule of each biofuel starting with an initial change
in fuel volume in Year 1 and continuing for 30 years into the future. In this illustrative scenario,
we apply these 30-year annual streams of emissions factors, as estimated in the analyses
supporting the 2010 RFS2 rule, to volume changes in corn ethanol, sugarcane ethanol, and
soybean oil biodiesel consumption.

Soybean Oil Renewable Diesel

To determine the lifecycle GHG intensity of soybean oil based renewable diesel we used
a straightforward combination of numerical estimates that EPA has published in previous
rulemakings to arrive at an overall carbon intensity value. The lifecycle GHG emissions
associated with using soybean oil as a feedstock for biodiesel and renewable diesel production
were estimated for the March 2010 RFS2 rule. In the 2012 Pathways I rule (78 FR 14190), we
evaluated renewable diesel from camelina oil and reported the GHG emissions associated with
the hydrotreating process used to convert the camelina oil to renewable diesel. The same
hydrotreating process is used for other virgin vegetable oils and is thus applicable for soybean oil

157 Cooney, G., Jamieson, M., Marriott, J., Bergerson, J., Brandt, A., & Skone, T. J. (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. doi:10.1021/acs.est.6b02819

158 GREET1-2021 Fuel Cycle Model available for download at https://greet.es.aiiLgov/greet.models

159 We use SAR2 global warming potential factors in order to better compare estimates in GREET-2020 with the
2005 gasoline and diesel baselines.

160 See "EPA-HQ-OAR-2005-0161-3173_Results_6_Corn_EtOH_noCHP.xls," available in the docket for this rule.
The vast majority of corn starch ethanol production in the U.S. (approximately 90% of corn ethanol production in
2020 according to EMTS data) use natural gas as the primary energy source. We note, however, that some of the
projected increase in corn ethanol use could be met by grandfathered facilities which do not have a GHG reduction
requirement under the RFS program.

161 Analyses of lifecycle greenhouse gas emissions completed for the March 2010 RFS2 rule included multiple
scenarios that evaluated an increase in biofuel production starting in different years. For the illustrative scenario
presented in this chapter, we use scenarios from the RFS2 analyses which had a start year of 2022. These are the
same scenarios EPA used for threshold determinations in the RFS2 rulemaking.

162 See "EPA-HQ-OAR-2005-0161-3173_Results_ll_Sugarcane_EtOH_NoTrash.xls," and "EPA-HQ-OAR-2005-
0161-3173_attachment_9_Soybean_Oil_Biodiesel.xls" available in the docket for this rule.

-------
renewable diesel. We also published our analysis of hydrotreating in the 2018 distillers sorghum
oil rule (83 FR 37735). We combined the annual stream of GHG emissions per pound of soybean
oil as estimated for the March 2010 RFS2 rule with the hydrotreating process modeled for the
2018 distillers sorghum oil rule to estimate an annual stream of emissions for renewable diesel
produced from soybean oil.163 This was a straightforward calculation as the output from the
soybean oil analysis (GHGs per pound soybean oil) was readily input into the hydrotreating
modeling to estimate the GHGs per MJ of renewable diesel. The methodology used in the 2012
Pathways I rule and the 2018 distillers sorghum oil rule, assumes the naphtha and LPG co-
products replace conventional gasoline and natural gas, respectively, and assign the displacement
credit to the renewable diesel. Using the above methodology, we calculate that renewable diesel
from soybean oil has a lifecycle emissions intensity of 45.9 gC02e/MJ.

FOG Renewable Diesel

To determine the lifecycle GHG intensity of renewable diesel produced from waste fats,
oils, and greases, we rely upon an existing EPA analysis of a similar pathway. In the March 2010
RFS2 rule, we determined that renewable diesel produced from biogenic waste FOG through a
hydrotreating process meets the 50% GHG reduction threshold to qualify as biomass-based
diesel and advanced biofuel. However, we did not publish a specific estimate of waste FOG
renewable diesel lifecycle GHGs; instead, we made the threshold determination based on a
logical extension of our estimate that waste FOG biodiesel results in a greater than 80% lifecycle
GHG reduction compared to conventional diesel. In 2018, we approved a facility-specific
petition for naphtha and liquefied petroleum gas (LPG) produced from biogenic waste FOG at a
hydrotreating facility in Louisiana and included our estimate of the GHG emissions associated
with this fuel.164 Although that petition determination was for naphtha and LPG, the results for
renewable diesel would be the same except for small differences for fuel distribution and use.
Although the petition determination considered data, claimed confidential, for one individual
facility, we believe the results, 21.8 gC02e/MJ, provide a reasonable estimate for the purposes of
this illustrative scenario. For comparison, GREET-2021 estimates lifecycle GHG emissions for
tallow-based renewable diesel of 17.7 gC02e/MJ. For these reasons, we are using the LPG
results from the 2018 petition approval as the basis for evaluating the GHG emissions associated
with waste FOG renewable diesel in this illustrative scenario. We note that since FOG renewable
diesel is produced from waste oil rather than an agricultural commodity, this analysis assesses no
land use change impacts and thus does not provide a temporal stream of emissions; using this
intensity factor for FOG renewable diesel volumes in 2022, for example, does not imply future
levels of FOG renewable diesel consumption after 2022.165 However, for consistency with other
biofuels analyzed in this illustrative scenario, we assess 30 years of consumption of FOG
renewable diesel at levels set by the 2021 and 2022 standards.

163 See "Soy RD LCA for 2021-22 RVO Rule.xlsx," available in the docket for this rule.

164 U.S. EPA, (2018). REG Geismar Approval (PDF). Available at https://www.epa.gov/renewable~fuel~standard~

program/reg~geismar~approval~0

165 These "waste oils," such as used cooking oil, are currently collected and sold as commodities.

-------
Landfill Biogas CNG

The July 2014 pathways II rule (79 FR 42128) included a technical memo to the docket
explaining EPA's conclusion that CNG and LNG produced from landfill biogas, when compared
to a flaring baseline, results in more than an 80% reduction in GHG emissions compared to
petroleum-based diesel.166 For this illustrative scenario, we use this estimate of an 80% reduction
in emissions compared to the 2005 diesel baseline to assess GHG reductions for changed
volumes of landfill biogas CNG. The implied CI used in this illustrative scenario is 18.4
gC02e/MJ for landfill biogas CNG. For comparison, GREET-2021 estimates a CI of 11.7 to
27.3 gC02e/MJ for landfill biogas CNG depending on whether renewable natural gas or
conventional natural gas are used to power equipment for landfill gas processing.

We note that renewable CNG is currently displacing primarily fossil natural gas used in
existing CNG/LNG fleets, but it may also be displacing some petroleum-based diesel as its use
provides an incentive for the growth in new CNG/LNG vehicle sales. However, since we are
relying on existing analyses for this scenario, and past considerations of biogas under the RFS
program were for the purpose of lifecycle analysis for pathway approval, we continue to
compare against the 2005 diesel baseline for this illustrative scenario. An alternative approach
could compare biogas to a fossil natural gas GHG intensity, such as the GHG intensity estimates
in the 2016 Phase 2 rule for heavy-duty vehicles (81 FR 73478).167 Since most estimates for
emissions intensities of natural gas are less than the 2005 diesel baseline (91.9 gCChe/MJ), the
result of using this alternative approach would be similar in scale but would likely result in a
smaller reduction in the GHG emissions.168

3.2.2.2 Greenhouse Gas Impacts

For each of the 2021 and 2022 standards, we separately estimate a 30-year stream of
changes in GHG emissions for each analyzed fuel using the carbon intensity analyses discussed
above.169 While the annual standards finalized in this rulemaking 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. Emissions streams based on the 2021 standards and 2022 standards are
presented in Tables 3.2.2.2-1 and 3.2.2.2-2, respectively. These annual sequences of emissions

166 The pathways II rule (79 FR 42128) included a technical memo to the docket (Docket Item No. EPA-HQ-OAR-
2012-0401-0243) that discusses EPA's analysis of the GHG emissions of landfill biogas. While that memo did not
provide a specific point estimate in terms of gCChe/MJ, it did conclude that waste derived biogas would result in
over an 80% reduction in GHG emissions compared to the petroleum baseline. For this illustrative scenario analysis,
we assume that landfill biogas CNG/LNG reduces GHG emissions by 80% when compared to the 2005 diesel
baseline. For comparison, landfill biogas emissions calculated in GREET-2021 achieve up to an 87% reduction
when compared with EPA's 2005 diesel baseline.

167 U.S. EPA, (2016). Greenhouse Gas Emissions and Fuel Efficiency Standards for Medium- and Heavy-Duty
Engines and Vehicles - Phase 2: Regulatory Impact Analysis. U.S. Environmental Protection Agency, EPA-420-R-
16-900: Table 13-19, Washington, DC, August.

168 For comparison, the GHG intensity of landfill biogas CNG used for this analysis would achieve a 76% reduction
if it were compared against the fossil CNG intensity used in the 2016 Phase 2 rule for heavy-duty vehicles (81 FR
73478). When compared against the 2005 diesel baseline, as is done in our analysis for this rule, landfill biogas
CNG achieves an 80% reduction.

169 Renewable fuel volume changes assessed in this illustrative scenario are presented in Table 2.2-3.

-------
for the 2021 and 2022 standards are then summed, resulting in a combined stream of estimated
annual emissions from 2021 through 2051. This is presented in Table 3.2.2.2-3 below. The table
includes positive net GHG emissions for the corn ethanol and soy renewable diesel volumes in
2021 and 2022 due to the initial pulse of land use change emissions in the estimates used for this
illustrative scenario. The net GHG emissions for these fuels then become negative over time as
biofuel production remains in place and continues to replace petroleum-based fuels. As noted
above, this scenario assumes that the biofuel production continues for 30 years, irrespective of
the RVO mandate in future years. Based on the analysis described above, we estimate that in this
illustrative scenario the GHG benefit from biofuel production displacing fossil fuel use is larger
than the GHG emissions associated with biofuel production and use, including land use change
emissions relative to the baseline scenario.

The renewable fuel volume changes assessed in this illustrative scenario are presented in
Table 2.2-3. Note that, relative to the baseline used for this analysis, soybean oil biodiesel
consumption decreases by 107 million gallons in 2021 and sugarcane ethanol consumption
decreases by 125 million gallons in 2021. For this reason, the scenario includes negative
emissions associated with the sugarcane ethanol and soybean oil biodiesel in 2021.170 Values for
FOG renewable diesel and landfill biogas CNG have no land-use change impact and therefore
are always negative. The illustrative GHG emissions reductions attributable to landfill biogas in
this scenario had a calculation error in the proposed rule. This error has been corrected, which
results in a larger GHG emissions reduction.171

170 We note that the general trend is of increasing sugarcane area in Brazil, the world's primary producer of
sugarcane ethanol. As long as sugarcane area is increasing in the baseline, then reducing sugarcane ethanol
production would result in reduced crop expansion. Reducing crop expansion means avoiding a pulse of land
clearing emissions.

171 The analysis in this illustrative scenario assumes that landfill biogas achieves an 80 percent reduction in GHG
emissions compared to the 2005 diesel baseline. The illustrative scenario for this rule instead incorrectly assessed a
20 percent reduction in GHG emissions for landfill biogas compared to the 2005 diesel baseline.

-------
Table 3.2.2.2-1: 30-year stream of emissions for 2021 standards using lifecycle analysis for
individual biofuels from EPA's 2010 analysis and subsequent EPA actions, presented in
millions of metric tons CC>2e.a

Year

Corn
Starch
Ethanol

Sugarcane
Ethanol

Soybean

Oil
Biodiesel

Soybean
Oil
Renewable
Diesel

FOG
Renewable
Diesel

Landfill
Biogas
CNG

Total

2021

42.7

(2.2)

(10.5)

34.3

(0.4)

(0.3)

63.6

2022

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2023

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2024

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2025

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2026

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2027

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2028

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2029

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2030

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2031

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2032

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2033

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2034

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2035

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2036

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2037

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2038

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2039

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2040

(6.0)

0.6

1.1

(3.1)

(0.4)

(0.3)

(8.2)

2041

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2042

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2043

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2044

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2045

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2046

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2047

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2048

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2049

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2050

(7.4)

0.5

1.1

(3.4)

(0.4)

(0.3)

(9.9)

2051

-------
Table 3.2.2.2-2: 30-year stream of emissions for 2022 standards using lifecycle analysis for
individual biofuels from EPA's 2010 analysis and subsequent EPA actions, presented in
millions of metric tons CC>2e.a

Year

Corn Starch
Ethanol

Sugarcane
Ethanol

Soybean Oil
Biodiesel

Soybean Oil
Renewable
Diesel

FOG
Renewable
Diesel

Landfill
Biogas CNG

Total

2021

2022

12.7

0.9

3.5

71.5

(0.4)

87.8

2023

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2024

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2025

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2026

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2027

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2028

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2029

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2030

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2031

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2032

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2033

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2034

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2035

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2036

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2037

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2038

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2039

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2040

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2041

(1.8)

(0.2)

(0.4)

(6.6)

(0.4)

(9.8)

2042

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

2043

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

2044

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

2045

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

2046

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

2047

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

2048

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

2049

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

2050

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

2051

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

-------
Table 3.2.2.2-3: 30-year stream of emissions for combined 2021 and 2022 standards using
lifecycle analysis for individual biofuels from EPA's 2010 analysis and subsequent EPA
actions, presented in millions of metric tons CQ2e.a ^

Year

Corn Starch
Ethanol

Sugarcane
Ethanol

Soybean Oil
Biodiesel

Soybean Oil
Renewable
Diesel

FOG
Renewable
Diesel

Landfill
Biogas CNG

Total

2021

42.7

(2.2)

(10.5)

34.3

(0.4)

(0.3)

63.6

2022

6.8

1.5

4.6

68.3

(0.8)

79.6

2023

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2024

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2025

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2026

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2027

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2028

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2029

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2030

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2031

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2032

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2033

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2034

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2035

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2036

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2037

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2038

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2039

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2040

(7.8)

0.4

0.7

(9.7)

(0.8)

(18.0)

2041

(9.2)

0.3

0.8

(9.9)

(0.8)

(19.6)

2042

(9.6)

0.3

0.8

(10.4)

(0.8)

(20.5)

2043

(9.6)

0.3

0.8

(10.4)

(0.8)

(20.5)

2044

(9.6)

0.3

0.8

(10.4)

(0.8)

(20.5)

2045

(9.6)

0.3

0.8

(10.4)

(0.8)

(20.5)

2046

(9.6)

0.3

0.8

(10.4)

(0.8)

(20.5)

2047

(9.6)

0.3

0.8

(10.4)

(0.8)

(20.5)

2048

(9.6)

0.3

0.8

(10.4)

(0.8)

(20.5)

2049

(9.6)

0.3

0.8

(10.4)

(0.8)

(20.5)

2050

(9.6)

0.3

0.8

(10.4)

(0.8)

(20.5)

2051

(2.2)

(0.2)

(0.4)

(7.0)

(0.4)

(10.6)

" 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.

-------
3.2.2.3 Monetized Impacts

3.2.2.3.1 Social Cost of Greenhouse Gases

For assessing GHG impacts in this illustrative scenario, we rely upon past biofuel
emissions reductions estimates which are available in CChe - carbon equivalent emissions using
the global warming potentials utilized in those analyses.172 We estimate the social benefits of
GHG reductions in this illustrative scenario using estimates of the social cost of greenhouse
gases (SC-GHG)173, specifically using the SC-CO2 estimates presented in the Technical Support
Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under
Executive Order 13990 (hereafter, "February 2021 TSD").174 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. The SC-GHG 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 of 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.

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.

EPA and other federal agencies began regularly incorporating SC-C02 estimates in
benefit-cost analyses conducted under Executive Order (E.O.) 12866175 in 2008, following a
court ruling in which an agency was ordered to consider the value of reducing C02 emissions in

172 It would be preferable to use estimates for each gas (e.g., CO2, CH4, N20), but we use C02e estimates for this
illustrative scenario as they are the most readily available biofuel carbon intensity estimates.

173 Estimates of the social cost of greenhouse gases are gas specific (e.g., social cost of carbon (SC-CO2), social cost
of methane (SC-CH4), social cost of nitrous oxide (SC-N20)), but collectively they are referenced as the social cost
of greenhouse gases (SC-GHG).

174 Interagency Working Group on Social Cost of Greenhouse Gases (IWG). 2021. Technical Support Document:
Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990. February.
United States Government. Available at: https://www.whitehoiise.gov/briefing-rooni/blog/2021/Q2/26/a-retiirn-to-
science-evidence-based-estimates-of-the-benefits-of-reducing-cli mate-pollution/.

175 Under E.O. 12866, agencies are required, to the extent permitted by law and where applicable, "to assess both the
costs and the benefits of the intended regulation and, recognizing that some costs and benefits are difficult to
quantify, propose or adopt a regulation only upon a reasoned determination that the benefits of the intended
regulation justify its costs."

-------
a rulemaking process. The SC-C02 estimates presented here 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 the 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 IAM.176"111'178 In August 2016 the IWG published estimates of the social cost of methane
(SC-CH4) and nitrous oxide (SC-N2O) using methodologies that are consistent with the
methodology underlying the SC-CO2 estimates. 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-CO2 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.179 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-CO2 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-CO2 estimates attempted to focus on the U.S. specific share of
climate change damages as estimated by the models and were calculated using two discount rates
recommended by Circular A-4, 3 percent and 7 percent. All other methodological decisions and
model versions used in SC- CO2 calculations remained the same as those used by the IWG in
2010 and 2013, respectively.

On January 20, 2021, President Biden issued Executive Order 13990, which re-
established the IWG and directed it to develop updated estimates of the social cost of carbon and
other greenhouse gases that reflect the best available science and the recommendations of the
National Academies. The IWG was tasked with first reviewing the SC-GHG estimates currently
used in Federal analyses and publishing interim estimates within 30 days of the E.O. that reflect
the full impact of GHG emissions, including by taking global damages into account.

176 Dynamic Integrated Climate and Economy (DICE) 2010 (Nordhaus 2010).

177 Climate Framework for Uncertainty, Negotiation, and Distribution (FUND) 3.8 (Anthoff and Tol 2013a, 2013b)

178 Policy Analysis of the Greenhouse Gas Effect (PAGE) 2009 (Hope 2013).

179 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.

-------
As noted above, EPA participated in the IWG but has also independently evaluated the interim
SC-C02 estimates published in the February 2021 TSD and determined they are appropriate to
use here to estimate the climate benefits associated with this illustrative scenario. 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.
The 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.

Therefore, in this illustrative analysis, 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 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 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.180'181'182'183'184 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
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% and 7% 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% discount rate is not
appropriate to apply to value the social cost of greenhouse gases in the analysis presented in this
analysis. 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

180 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.

181 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.

182 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. United
States Government.

183 Interagency Working Group on Social Cost of Greenhouse Gases (IWG). 2016a. Technical Support Document:
Technical Update of the Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866.
August. United States Government.

184 Interagency Working Group on the Social Cost of Greenhouse Gases. 2016b. 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 Social Cost of Methane and the Social Cost of Nitrous Oxide. August. United
Stated Government. Available at: https://www.epa.gov/sites/production/files/2016-12/documents/addendum_to_sc-
ghg_tsd_august_2016.pdf (accessed February 5, 2021).

-------
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 provide information on potentially higher-than-expected economic impacts
from climate change, conditional on the 3 percent estimate of the discount rate. 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 process.

Table 3.2.2.3.1-1 summarizes the interim SC-CO2 estimates for the years 2021 - 2051.185
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. For purposes of capturing uncertainty around the SC-
C02 estimates in analyses, the February 2021 TSD emphasizes the importance of considering all
four of the SC-C02 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.

185 The February 2021 TSD provides SC-GHG estimates through emissions year 2050. In this illustrative scenario
we held the 2051 SC-CO2 value for 2051 at the same level as the 2050 value.

-------
Table 3.2.2.3.1-1: Interim Social Cost of Carbon Values, 2020-2051 (2021$/Metric Tonne

CO2)186

Emissions
Year

Discount E

late and Statistic

5%
Average

3%
Average

2.5%
Average

3% 95th
Percentile

2020

$15

$52

$77

$153

2021

$15

$53

$79

$157

2022

$16

$54

$80

$160

2023

$16

$55

$81

$164

2024

$17

$56

$82

$167

2025

$17

$57

$84

$171

2026

$18

$58

$85

$174

2027

$18

$59

$86

$178

2028

$19

$60

$88

$181

2029

$19

$61

$89

$185

2030

$20

$62

$90

$189

2031

$20

$64

$92

$192

2032

$21

$65

$93

$196

2033

$21

$66

$95

$200

2034

$22

$67

$96

$204

2035

$23

$68

$97

$208

2036

$23

$69

$99

$212

2037

$24

$70

$100

$216

2038

$24

$72

$101

$219

2039

$25

$73

$103

$223

2040

$25

$74

$104

$227

2041

$26

$75

$105

$231

2042

$27

$76

$107

$234

2043

$27

$77

$108

$238

2044

$28

$79

$110

$241

2045

$29

$80

$111

$245

2046

$29

$81

$112

$248

2047

$30

$82

$114

$252

2048

$31

$83

$115

$255

2049

$31

$84

$116

$259

2050

$32

$86

$118

$263

2051

$32

$86

$118

$263

Note: The 2020-2050 SC-CO2 values are identical to those reported in the 2016 TSD (IWG 2016a) adjusted for
inflation to 2021 dollars using the annual GDP Implicit Price Deflator values in the U. S. Bureau of Economic

186 Interagency Working Group on Social Cost of Greenhouse Gases (IWG). 2021. Technical Support Document:
Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990. February.

-------
Analysis' (BEA) NIPA Table 1.1.9 (U.S. BEA 2021). 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.whitehouse.gOv/omb/information-regulatory-affairs/regulatory-matters/#scghgs.

The estimates were extended to 2051 using the same value for 2051 as the 2050 value. The values are stated in

$/metric tonne CO2 and vary depending on the year of C02 emissions.

There are a number of limitations and uncertainties associated with the SC- CO2
estimates presented in Table 3.2.2.3.1-1. 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 3.2.2.3.1-1, presents the quantified sources of uncertainty in the form of frequency
distributions for the SC-CO2 estimates for emissions in 2030 (in 2018$). 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.

United States Government. Available at: https://www.whitehouse.gov/briefing-room/blog/2021/02/26/a-return-to-
science-evidence-based-estimates-of-the-benefits-of-rediicing-cli mate-pollution

-------
Figure 3.2.2.3.1-1: Frequency Distribution of SC-CO2 Estimates for 2030

187

If)

c
o

Ira

E
to

-------
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.188 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.189'190'191'192'193'194'195'196 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 2000, while not ruling out even more extreme
outcomes. 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.
EPA has reviewed and considered the limitations of the models used to estimate the interim SC-
GHG estimates, and concurs with the February 2021 TSD's assessment that, taken together, the

188 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.

189 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.

190 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.)].

191 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. vanDiemen, 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.)].

192 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.)].

193 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.

194 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.

195 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.

196 National Academies of Sciences, Engineering, and Medicine (National Academies). 2019. Climate Change and
Ecosystems. Washington, DC: The National Academies Press. https://doi.Org/.l.0.17226/25504.

-------
limitations suggest that the interim SC-C02 estimates likely underestimate the damages from
C02 emissions. The IWG, of which EPA is a member, 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, and public
comments received on the February 2021 TSD.

Table 3.2.2.2.2-1 and Table 3.2.2.2.2-2 show the estimated changes in CChe for the
volume changes analyzed for 2021 and 2022. 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 2021 through 2051 by
applying the SC-CO2 estimates, shown in Table 3.2.2.3.1-1, to the estimated changes in GHG
emissions inventories resulting from the final volumes. EPA then calculated the present value
and annualized benefits from the perspective of 2021 by discounting each year-specific value to
the year 2021 using the same discount rate used to calculate the SC-CO2.

3.2.2.3.2 Results

For this illustrative scenario, the interim estimates for carbon dioxide from the February
2021 TSD, presented in Table 3.2.2.3.1-1, were used to estimate the social benefits of the
estimated 30-year stream of GHG impacts presented in Chapter 3.2.2.2. For each year, the total
of emissions changes presented in Tables 3.2.2.2-1 and 3.2.2.2-2 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 in 2022 using the corresponding discount rates. The resulting streams
of estimated social benefits of the biofuel volume changes assumed in this illustrative scenario
for the 2021 standards and the 2022 standards are presented in Tables 3.2.2.3.2-1 and 3.2.2.3.2-2
respectively.197 Note that in the following tables, volume changes for the 2021 standards are

197 According to OMB's Circular A-4 (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 separately. Circular A-4 also
reminds analysts that "[d]ifferent regulations may call for different emphases in the analysis, depending on the
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 E.O. 13783
were a limited approximation of some of the U.S. specific climate damages from GHG emissions (e.g., $7/mtC02
(2021 dollars) and $12/mtC02 using a 3% discount rate for emissions occurring in 2020 and 2050, respectively).
Applying the same estimate (based on a 3% discount rate) to the CChe emission reduction expected under this
illustrative analysis would yield benefits from climate impacts, expressed as annualized net present value, of $114
million (2021 dollars) for the combined volumes assessed for both the 2021 and 2022 standards. As discussed at
length in the February 2021 TSD, these estimates are an underestimate of the benefits of C02 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 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" (U.S.
GAO 2020, p. 29). 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

-------
relative to the 2020 standards, while the volume changes for the 2022 standard are relative to the
2021 standards. For this illustrative analysis we summarize the results in two different ways, a
net present value and an annualized benefit, both of which assume that, in each of the 29 years
following the standards finalized in this rule, aggregate renewable fuel consumption for each
category exceeded baseline levels by the same volume associated with the rule. The total net
present value of the 30-year stream of monetized impacts is calculated by summing the annual
stream of discounted benefits presented in each table. An annualized value is also calculated for
each discount rate following the methodology presented in EPA guidance document Guidelines
for Preparing Economic Analyses19*'199 This annualized benefit is the constant value such that
the sum of all thirty years of benefits in present value terms equals the net present value,
calculated as described above.200 Combined monetized impacts for the 2021 and 2022 standards
are presented in Table 3.2.2.3-3; this also reflects the monetized impacts of the 2022 standards
relative to the 2020 baseline. All calculations are available in a spreadsheet in the docket for this
rule.201 As discussed above, net and annualized impacts are sensitive to both the discount rate
and the time horizon used. We note that it would not be appropriate to directly compare these
benefits to the costs or energy security benefits, presented in Chapters 9 and 4, respectively,
because this illustrative scenario for GHG analysis covers a different time horizon than those
analyses.

model all relevant interactions among regions" (U.S. GAO 2020, p. 26). 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.

198 U.S. Environmental Protection Agency. "Chapter 6: Discounting Future Benefits and Costs," equation 4.
Guidelines for Preparing Economic Analyses. United States Government. Retrieved from:

https://www.epa.gOv/environ.mentai-econoniics/giiideiines-preparing-economic-anaiyses#download October 26th,
2021.

199 Annualized values are calculated using 30-year NPV totals, and so are also sensitive to the 30-year time horizon.
Thus, annualized climate benefits are also incomparable to other monetized costs and benefits presented in this RIA.

200 This is true assuming there is a benefit or cost in the initial time period, as is the case in this analysis. This
corresponds with equation 4 in Chapter 6 of the cited EPA Guidelines for Preparing Economic Analyses.

201 See "GHG Scenario for 2020-22 RVO Rule.xlsx," available in the docket for this rule.

-------
Table 3.2.2.3.2-1: Present value of 30-year stream of climate benefits for 2021 standards,

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2021

(1,010)

(3,453)

(5,122)

(10,271)

2022

129

443

657

1,319

2023

126

438

652

1,309

2024

124

434

646

1,298

2025

122

429

641

1,287

2026

119

425

635

1,276

2027

117

420

629

1,264

2028

114

415

623

1,251

2029

112

410

617

1,239

2030

109

405

611

1,226

2031

107

401

605

1,214

2032

105

396

599

1,203

2033

103

392

593

1,191

2034

100

387

587

1,178

2035

382

581

1,166

2036

377

575

1,153

2037

372

569

1,139

2038

367

562

1,126

2039

363

556

1,112

2040

358

550

1,099

2041

102

424

653

1,301

2042

100

418

645

1,283

2043

412

637

1,264

2044

406

629

1,246

2045

399

621

1,227

2046

393

614

1,209

2047

388

606

1,190

2048

382

598

1,172

2049

376

590

1,153

2050

370

583

1,135

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 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.

-------
Table 3.2.2.3.2-2: Present value of 30-year stream of climate benefits for 2022 standards,

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%

95th
percentile

2022

(1,371)

(4,720)

(7,009)

(14,069)

2023

150

519

772

1,550

2024

147

514

765

1,538

2025

144

509

759

1,525

2026

141

503

752

1,511

2027

138

498

745

1,497

2028

135

492

738

1,482

2029

132

486

731

1,467

2030

129

480

723

1,452

2031

127

475

716

1,438

2032

124

469

709

1,424

2033

122

464

702

1,410

2034

119

458

695

1,396

2035

116

453

688

1,381

2036

114

447

681

1,365

2037

111

441

673

1,350

2038

108

435

666

1,334

2039

106

429

659

1,318

2040

103

424

651

1,301

2041

101

418

644

1,283

2042

107

449

694

1,380

2043

105

443

685

1,360

2044

102

436

677

1,340

2045

430

669

1,320

2046

423

660

1,300

2047

417

652

1,280

2048

410

644

1,260

2049

404

635

1,241

2050

398

627

1,221

2051

386

612

1,185

-------
Table 3.2.2.3.2-3: Sum of present value of 30-year stream of climate benefits for 2021 and
2022 standards, using four values for the social cost of carbon (SC-CO2) (millions of
2021$).a

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%

95th
percentile

2021

(1,010)

(3,453)

(5,122)

(10,271)

2022

(1,242)

(4,277)

(6,352)

(12,750)

2023

276

957

1,424

2,859

2024

271

948

1,412

2,836

2025

265

938

1,399

2,812

2026

260

928

1,387

2,787

2027

255

918

1,374

2,761

2028

249

907

1,361

2,734

2029

244

896

1,347

2,706

2030

238

886

1,334

2,677

2031

234

876

1,321

2,653

2032

229

866

1,308

2,627

2033

224

856

1,295

2,601

2034

219

845

1,282

2,574

2035

215

835

1,269

2,546

2036

210

824

1,256

2,518

2037

205

814

1,242

2,489

2038

200

803

1,228

2,460

2039

195

792

1,215

2,430

2040

190

781

1,201

2,400

2041

203

841

1,296

2,585

2042

207

867

1,339

2,663

2043

202

854

1,322

2,624

2044

197

842

1,306

2,586

2045

192

829

1,290

2,547

2046

187

817

1,274

2,509

2047

182

804

1,258

2,470

2048

177

792

1,242

2,432

2049

172

780

1,226

2,394

2050

167

767

1,210

2,356

205 lb

386

612

1,185

-------
climate benefits calculated using lower discount rates are also warranted when discounting intergenerational
impacts.

b Combined impacts presented in Table 3.2.2.3.2-3 are the sum of the two thirty-year streams of impacts for the
2021 and 2022 standards presented in Tables 3.2.2.3.2-1 and 3.2.2.3.2-2. Because we assess thirty years of impacts
for each year standards, the period of analysis for the 2022 standards extends to 2051, while the period of analysis
for the 2021 standards does not.

We note that the methodology underlying the SC-C02 estimates used in this analysis has
been subject to public comment in the context of dozens of proposed 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.202 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.

3.3 Conversion of Wetlands, Ecosystems, and Wildlife Habitats

The Second Triennial Report to Congress on Biofuels203 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. However, at this time we cannot quantify the amount of
land with increased intensity of cultivation nor confidently estimate the portion of crop land
expansion that is due to the market for biofuels. (see Second Triennial Report to Congress on
Biofuels Sections 2 and 4.2). Often these changes are ascribed to agricultural expansion for
biofuel production, and in some cases even to the RFS program itself, but, in reality, such a
causal connection is difficult to make with confidence (see Second Triennial Report to Congress
on Biofuels Section 2). Moreover, as discussed in Section 2 of the Second Triennial Report to
Congress on Biofuels, these land use change studies vary widely in approach and scope, making
comparison inherently difficult. This section focuses on impacts related to the domestic
production of renewable 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 3.1), climate effects (Chapter
3.2), and possible leakage from underground storage tanks (Chapter 3.4.4). Insofar as there are
impacts of renewable fuel on the conversion of wetlands, ecosystems, and wildlife habitats, they

202 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. See
https://www.federalregister.gOv/documents/2022/01/25/2022-01387/request-for-nominations-of-experts-for-the-
review-of-technical-support-document-for-the-social-cost. EPA has a webpage where additional information
regarding the peer review process will be posted as it becomes available: https://www.epa.gov/environmental-
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.

203 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.

-------
are associated with crop-based feedstocks rather than waste oils, fats, and greases. We note that
to the extent the RFS standards in this rule 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
degredation 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 GHG
benefit provided by biofuel displacement of conventional fuels.204

3.3.1 Wetlands

There are several federal reports that describe the status and trends of U.S. wetlands,205
including the U.S. Fish and Wildlife Service (USFWS) Status and Trends of Wetlands in the
Conterminous United States,206 the USFWS and NOAA Status and Trends of Wetlands in the

204 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; Austin, K. G., A. Schwantes, Y. Gu and P. S.
Kasibhatla (2019). "What causes deforestation in Indonesia?" Environmental Research Letters 14(2): 024007.

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.

205 Summarized and listed here: https://www.epa.gov/wetlands/how~does~epa~keep~track~status~and~trends~
wetlands-iis

206 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.

-------
Coastal Watersheds of the Coterminous United States,207 the USFWS Status and Trends of
Prairie Wetlands in the United States,208 EPA's National Wetland Condition Assessment209
(NWCA), and USDA's Natural Resources Inventory (NRI).210 The USGS NWALT (National
Water-Quality Assessment (NAWQA) Program's Wall-to-Wall Anthropogenic Land Use
Trends) series does not model changes in wetlands.211 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.212 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 3.3.1-1. 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
developed and water areas.213 The report does not provide the information needed to determine
the portion of wetland acres lost in order to grow feedstocks for biofuels.

207 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.)

208 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).

209 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

210 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/wps/portal/nrcs/main/national/techiiicai/nra/nri/resnits.

211 Falcone JA (2015). U.S. conterminous wall-to-wall anthropogenic land use trends (NWALT), 1974-2012. U.S.
Geological Survey: 33 pp. Washington, DC.

212 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/wps/portal/nrcs/main/national/techiiical/nra/nri/resnlts.

213 "Water areas" are defined in the USDA NRI as "[a] broad land cover/use category comprising water bodies and
streams that are permanent open water."

-------
Table 3.3.1-1: Changes in palustrine214 and estuarine215 wetlands on different land
use/cover types between 2007, 2012, and 2017216

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.217 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., and a significant portion of U.S. ethanol production is exported and therefore cannot be
attributed to the RFS program.

Also, there are 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

214 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, available at

https://www.nrcs.usda.gov/wps/portal/nrcs/detail/national/technical/nra/nri/processes/?cid=nrcsl43 014.1.27 (last
accessed May 6, 2021).

215 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, available at

https://www.nrcs.usda.gov/wps/portal/nres/detail/national/technieal/nra/nri/processes/?cid=nrcsl43 0.1.4.1.27 (last
accessed May 6, 2021).

216 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/wps/portal/nrcs/main/national/techiiical/nra/nri/resnits.

217 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.

-------
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,218 but any losses are relatively small compared with the total amount of wetland and
compared with losses of grassland.219 Moreover, as stated above, these studies were directed at
the potential impacts of biofuels generally, not the RFS program or the volumes being finalized
today specifically. Furthermore, while we are evaluating the impact of the final volumes relative
to a 2020 baseline as discussed in Chapter 2, in reality there may be little change in ethanol
volume compared to what would have occurred in the absence of the RFS program (the "No
RFS" baseline) as the vast majority of ethanol is used as E10 and such uses would continue in
the absence of the RFS.

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 2017.220 The bulk of the wetland losses were in the
Prairie Pothole region, as reported elsewhere,221 with some very high rates (i.e., >15%, Wright et
al. 2017). The conversion reported by Wright 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. However, additional information is needed in order to draw any
conclusions with confidence as to whether biofuel production is a driving factor in that loss, as
well as the extent to which the annual volume requirements under the RFS program cause
changes in biofuel production and in particular the final volume requirements. The volume
increases for 2021 and 2022 compared to 2020 that are described in Chapter 2 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 causal impacts of this rule (as opposed to the volume changes in Chapter 2), the
2020 and 2021 volumes will have no impact, since these volumes will be entirely retroactive.
The 2022 volume requirements could have some impact largely due to their potential ability to
incent greater production of biofuels and their underlying crop-based feedstocks. Thus, the 2022

218 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.

219 See Chapter 3.4.2.1 for a discussion of grassland losses.

220 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.iisda.gov/wPs/portal/nrcs/main/national/technical/nra/nri/resiilts/. See
also USD A, 2017 National Resources Inventory, at

https://www.nrcs.nsda.gov/Internet/NRCS RCA/reports/nri wet nat.fatnit. (Last accessed on April 12, 2021). 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.

221 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.

-------
volumes have the potential to adversely impact wetland ecosystems. More information is needed
to assess the degree to which the RFS volume requirements impact land use and management
decisions in order to estimate the magnitude of their impacts on wetland loss. However, such
analysis would be expansive and could not be performed on the timeline of this rulemaking.

Figure 3.3.1-1: Relative conversion rates from wetland to cropland from 2008 to 2012.

0 0.8 2.4 5.9 16.5 100 0 400 800

Rates are relativized by type of ecosystem within a 3.5-mile spatial grid (modified from 2"). Stars denote the
location of biorefineries in the analysis.

3.3.2 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 3.4. As with
other land use changes and associated environmental effects, attributing the fraction of these
changes to biofuels is not currently possible with any degree of confidence. Consequently,
attribution to the RFS annual volumes is also not currently possible, and such an analysis would
be expansive and could not be conducted in time to be included in this rulemaking. The
conversion of these ecosystems to other uses, including agriculture, is summarized below.

In addition to wetlands, Wright et al. (2017) also reported on the losses of grasslands,
shrublands, and forests within 50 miles of a biorefinery in their study.223 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

322 Wright, C. K., et al. (2017), "Recent grassland losses are concentrated around US ethanol refineries."
Environmental Research Letters 12(4).

: '5 Id.

100

-------
3.3.2-1). 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).224

Figure 3.3.2-1: Relative conversion rates to cropland from either (a) grassland, (b) forest,
or (c) shrubland from 2008 to 2012

Net grassland change(%)

-80 -5.3 -1.5 2.2

>100

225

Rates are relativized by type of ecosystem within a 3.5-mile spatial grid (modified from ). Stars denote the location of
biorefineries, and the 100 mile radius from all biorefineries is included in (a) for reference (purple outline).

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 2012.226 The NRI defines rangeland as a land use/land cover that is more
lightly managed than pastureland,227 and, as such, is probably the NRI land use/land cover most

234 id.

225 Id

'226 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/wps/portal/nrcs/main/national/teclTnical/nra/nri/results
227 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

101

-------
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
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).228

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.229 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.230 Despite the rental payment incentive to
farmers, enrollment in the CRP has been shrinking since 2007.231 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 2020.232
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%).233 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.234 Reconciling these studies suggests that, of the land that leaves the CRP and goes into

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."

228 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/wps/portal/nrcs/main/national/techiiicai/nra/nri/resnits/.

229 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_finall.pdf

230 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.". Ipplied Geography 53: 141-148.

231 Data from the USDA Farm Service Agency (FSA), available at https://www.fsa.usda.gov/programs-and-
services/conservation-programs/reports-aiKl-statistics/conse rvation-reserve-pro gram-statistics/index (last accessed
on April 12, 2021).

232 USDA FSA, FY 2020 Annual Summary. Data from the USDA Farm Service Agency (FSA), available at

https://www.fsa.iisda.gov/programs-and-services/conservation-programs/reports-aiKl-statistics/conservation-reserve-
program-statistics/index (last accessed on May 5, 2021).

233 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.iisda.gov/wps/portal/nrcs/main/national/techiiicai/nra/nri/resiiits/.

234 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.

102

-------
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 3.3.2-2), with much of
the reduction in the western and northern plains, the same areas experiencing losses of grassland
and increases in agriculture.

Figure 3.3.2-2: Change in CRP enrollment between 2007 and 2016.235

Change in CRP Enrollment 2007 - 2016

Note: Data as of the end of April 2018.

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 within 50 miles of biorefineries. The NRI reported an overall increase of
forestland between 2007 and 2012 (-1-672,400 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.236

235 Data from the from the USDA Farm Services Agency (https://www.fsa.usda.gov/programs-and-
services/conservation-programs/reports-and-statistics/conservation-reserve-program-statistics/index).

236 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.

103

-------
Thus, even though some forest land did convert to cropland according to the NRI,237 these
conversions appear small and to be offset by reforestation of pastureland.

The volume increases for 2021 and 2022 compared to 2020 that are described in Chapter
2 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 other non-wetland ecosystems into
cropland. However, if we consider the causal impacts of this rule (as opposed to the volume
changes in Chapter 2), the 2020 and 2021 volumes will have no impact, since these volumes will
be entirely retroactive. The 2022 volume requirements could have some impact due to their
potential ability to incent greater production of biofuels and their underlying crop-based
feedstocks. Thus, the 2022 volumes have the potential to adversely impact grasslands and other
non-wetland ecosystems. Additional information on the factors driving grassland to cropland
conversion is needed in order to estimate the direction and magnitude of any impact the RFS
volumes may have on those land use/land cover changes.

3.3.3 Wildlife

There are many subsequent potential impacts to wildlife from these changes in wetlands
and other ecosystems, which were also summarized in the Second Triennial Report to Congress
on Biofuels.238 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).239 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 3.4.2.3.

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.240 However, studies of the effects of bioenergy feedstock
production suggest that grassland bird species of conservation concern are more likely to be
affected by increased corn production than are more common species of birds.241 Evidence
suggests that the direct effects of increasing cultivation of corn and soybean for biofuel

237 The 2020 NRI reports that 292,200, and 265,600 acres of forest land converted to cropland between 2007-2012,
and 2012-2017, respectively.

238 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.

239 Listed here: https://www.fsa.usda.gov/programs-and-services/conservation-programs/crp-practices-library/index.

240 Gleason, R.A., Euliss, N.H., Tangen, B.A., Laubhan, M.K., and Browne, B.A. (2011). "USD A conservation
program and practice effects on wetland ecosystem services in the Prairie Pothole Region." Ecological Applications
21: S65-S81.

241 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.

104

-------
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.242 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.243 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.244 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.245 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.24''

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.247 Roughly 20%
of these pollination services are estimated from wild populations which depend on local habitat
for food and nesting sites.248 A 2016 modeling study suggests that wild bee populations

242 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

243 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.

244 Primary crops were defined as corn, soybeans, and wheat.

245 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.

246 Id.

247 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.

248 Losey, J.E., Vaughan, M. (2006). "The economic value of ecological services provided by insects." Bioscience
56(4):311-323.

105

-------
decreased by 23% across the U.S. between 2008 and 2013.249 The causes of these reductions are
complex, but include land use change, pesticides, and disease.250 Subsequent effects from
reductions in local bee populations are possible, including reductions in pollinator-dependent
crops grown in the area,251 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.252 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.253 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.254 For soybeans, risks were considered uncertain at the time and are
currently undergoing re-evaluation by EPA. Neonicotinoids, like all pesticides, are approved for
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

249 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.

250 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.

251 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.

252 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.

253 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.

254 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.

106

-------
measure exposures and relate them to effects at the individual and colony level.255 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.

At present it is not possible to confidently estimate the fraction of wildlife habitat loss
that is attributable to biofuel production or use. Thus, we cannot confidently estimate the impacts
to date on wildlife from biofuels generally nor from the annual volume requirements,
specifically. Attributing such impacts to RFS annual volume rules specifically, let alone the
changes in renewable fuel volumes resulting from the RFS standards in this action, is even more
difficult (see Second Triennial Report to Congress on Biofuels Sections 3.4.3.2 and 2.4.4).

3.3.4 Potential Future Impacts of Annual Volume Requirements

The volume increases for 2021 and 2022 compared to 2020 that are described in Chapter
2 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. However, if we consider the causal
impacts of this rule (as opposed to the volume changes in Chapter 2), the 2020 and 2021 volumes
will have no impact since they will be entirely retroactive. The 2022 volume requirements could
have some impact largely due to their potential ability to incent greater production of biofuels
and their underlying crop-based feedstocks. Thus, the 2022 volumes have the potential to impact
wildlife, wetlands, and other ecosystems. There exists substantial uncertainty in projecting
changes in land use and management associated with corn, soybeans, and other crops. Additional
information and modeling are needed to fully assess changes in habitat areas and effects on
wildlife, both for crop expansion and pesticide use.

3.4 Soil and Water Quality

Soil and water quality are addressed here in one combined section because 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.256 It can be affected by biofuel
feedstock production through changes in soil erosion, soil organic matter (SOM),257 and soil

255 U.S. EPA (2018), "How We Assess Risks to Pollinators." Available at https://www.epa.gov/poHinator~

protection/how-we-assess-risks-poHinators (last accessed April 12, 2021).

256 The USD A Natural Resources Conservation Service (NRCS) defines soil health or soil quality 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. In short, the
capacity of the soil to function" (USDA-NRCS 2021). 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.

257 Soil organic matter is defined by Brady and Weil (2010) 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.

107

-------
nutrients, among other 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.258 Crop-based
biofuel feedstock production can affect water quality through associated changes in nutrients,
dissolved oxygen, sediment, and chemical loadings.259 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 3.4.2).260 Water quality impacts are presented as
either proximal (i.e., geographically close) or downstream, although effects can span both. We
discuss 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.

3.4.1 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 biodiesel.
However, they are not sourced from crop-based feedstocks and thus have only a tenuous
connection to soil and water quality. Furthermore, production of these biofuel feedstocks is very
unlikely to increase as a result of the 2022 RFS standards in this action, as biofuel use will
provide another outlet for their use but not incent increases in their production. Canola oil is also
used for biodiesel production, though in considerably smaller quantities than soybean oil and

Upper Saddle River, 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).

258 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.

259 The USDA NRCS Environmental Technical Note No. MT-1 (2011) defines these water quality parameters and
their significance.

260 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).

108

-------
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 3.4.2.1). 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.261 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.262 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 3.3, 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.

3.4.2 Impacts to Date

3.4.2.1 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.263 In a meta-analysis, Qin et al. (2016) found that replacing
grasslands with corn decreased soil carbon by approximately 20% on average.264 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
soybeans, when both crops are managed using the same tillage practice (tillage practices are

261 81 FR 89746 (December 12, 2016).

262 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.

263 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.

264 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.

109

-------
discussed in greater detail later in this section).265 Increased erosion from conversion, in turn, can
negatively impact water quality through increased sediment and nutrient loadings to
waterways.266

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 3.4.2.1-1
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 corn267 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 3.4.2.1-1). 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.268 For planted soybean acres, 99% were treated with herbicides, 16% with
insecticides, and 15% with fungicides.269 Glyphosate isopropylamine salt and glyphosate
potassium salt were the top active ingredients among herbicides applied to planted soybean
acres.270 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.271

265 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.

266 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.

267 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.

268 USDA NASS (2019). 2018 Agricultural Chemical Use Survey: Corn. Available at

https://www.nass.nsda.gov/Survevs/Guide to NASS Surveys/Chemical Use/2018 Peanuts Soybeans Corn/Chem
UseHighlights Corn 2018.pdf (last accessed April 13, 2021).

269 USDA NASS (2019). 2018 Agricultural Chemical Use Survey: Soybeans. Available at

https://www.nass.usda.gov/Surveys/Guide to NASS Surveys/ChemicalUse/2018 Peanuts Soybeans Corn/Chem
UseHighiiehts Soybeans 20.1.8.pdf (last accessed April 13, 2021).

270 USDA NASS (2019). 2018 Agricultural Chemical Use Survey: Soybeans. Available at

https://www.nass.nsda.gov/Survevs/Guide to NASS Surveys/Chemical Use/2018 Peanuts Soybeans Corn/Chem
UseHighlights Soybeans 2018.pdf (last accessed April 13, 2021).

271 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/j.scitotenv.2017.02.009.

110

-------
Table 3.4.2.1-1: Summary of Chemical Use for Corn, Soybeans, and Wheat Acreage in the

U.S. based on 2018 and 2019 USDA

VASS Chemical Use Surveys272'273'274

Corn

Soybeans

Winter
wheat

Spring
wheat

Durum
wheat

Nitrogen Fertilizer Applied: % of Planted
Acres

Average Application Rate for Year for
Acres with Nitrogen Fertilizer Applied (lbs
N/acre)

149

102

Phosphate Fertilizer Applied: % of Planted
Acres

Average Application Rate for Year for
Acres with Phosphate Fertilizer Applied
(lbs P20s/acre)

Atrazine Applied: % of Planted Acres

Not
Reported

Average Application Rate for Year for
Acres with Atrazine Applied (lbs/acre)

1.037

Not
Reported

Glyphosate Potassium Salt: % of Planted
Acres

Not
Reported

Average Application Rate for Year for
Acres with Glyphosate Potassium Salt
Applied (lbs/acre)275

Not
Reported

1.527

Not
Reported

Glyphosate Isopropylamine Salt: % of
Planted Acres

Not
Reported

Average Application Rate for Year for
Acres with Glyphosate Isopropylamine Salt
Applied (lbs/acre)276

0.993

1.202

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.277 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

272 USD A NASS (2019). 2018 Agricultural Chemical Use Survey: Corn. Available at

https://www.nass.nsda.gov/Survevs/Guide to NASS Surveys/Chemical Use/2018 Peanuts Soybeans Corn/Chem
UseHighlights Corn 2018.pdf (last accessed April 13, 2021).

273 USD A NASS (2019). 2018 Agricultural Chemical Use Survey: Soybeans. Available at

https://www.nass.nsda.gov/Survevs/Guide to NASS Surveys/Chemical Use/2018 Peanuts Soybeans Corn/Chem
UseHighlights Soybeans 2018.pdf (last accessed April 13, 2021).

274 USD A NASS (2020). 2019 Agricultural Chemical Use Survey: Wheat. Available at

https://www.nass.usda.gov/Survevs/Guide to NASS Surveys/ChemicalUse/2019 Field Crops/chem-highlights-
wheat-2019.pdf (last accessed April 13, 2021).

275 This is expressed in acid equivalent.

276 This is expressed in acid equivalent.

277 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.

Ill

-------
practices. About 60-70% of corn and soybeans are grown using conservation tillage.278'279
Conservation tillage, including no-till, reduces soil erosion and increases SOM content relative to
conventional tillage.280'281

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.282 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.283 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 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.284 Notably, the percent increase in soil organic carbon of other cropland moving to
corn was exceeded in magnitude by the percent decrease in soil organic carbon by the conversion

278 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.

279Conservation 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 1997). 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 et al. 2015). 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.asDx 583; data accessed

2/15/2018). 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. N. R. C. S. United
States Department of Agriculture. Available at

https://www.nrcs.usda.gov/wps/portal/nres/detail/national/technieal/nra/ceap/piib/?cid=nresl43 01416.1. (last
accessed April 13, 2021).

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.

280 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.

281 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.

282 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. 12237. Yasarer LMW, 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.

283 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.

284Qin 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.

112

-------
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.285 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.286 In a modeling study, Secchi et al. (2011) concluded that this intensification287 of
corn would likely lead to higher nitrogen and phosphorus loads in the Upper Mississippi River
Basin.288

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.289 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.290 Yet too much stover removal can increase soil erosion,
decrease SOM and soil nutrients, and ultimately decrease corn yields.291 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.292 Additional research is needed to
understand effects on soil and water quality if soil conservation methods are employed while
harvesting corn stover.

285 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.

286 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.

287 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.

288 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.

289 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.

290 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. Also available at
https://doi.Org/10.10.l.6/i.biombioe.2017.09.016 (last accessed April 13, 2021).

291 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.

292Karlen, 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.

113

-------
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.293 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.294 Notably, however, the
effects of these perennial feedstocks can depend upon the plant species grown and the type of
land converted.295 Additionally, the literature definitions of what constitutes marginal land and
estimates of its extent vary widely.296 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.297 Alternative feedstock production (e.g., switchgrass) requires less
fertilizer than corn, thereby reducing nutrient runoff298 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.299 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).300

3.4.2.2 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 3.4.2.3).301 Fertilizer runoff,

293 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.

294 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.

295 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.

296 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.

297 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.

298 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.

299 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.

300 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.

301 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-
37: https://doi.0rg/lO.lOl6/j.ecolmodel.2Oi4.Ol.Ol6. Jarvie HP, Sharpley AN, FlatenD, KleinmanPJA, 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.

114

-------
in addition to other factors (e.g., temperature and precipitation) and conservation practices,
influence downstream eutrophication,302 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.303 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

304

ecoregion.

In coastal systems, nutrient loadings affect hypoxic zone size, which is also a function of
climate, weather (e.g., storms), basin305 morphology, circulation patterns, water retention time,
freshwater inflows, stratification, and mixing, as seen in the Gulf of Mexico.306 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.307

3.4.2.3 Aquatic Life Effects

According to the Second Triennial Report to Congress on Biofuels, the impacts of biofuel
crop production on aquatic ecosystems is understudied compared to the impacts on terrestrial
ecosystems.308 However, it has been shown that increased corn and soybean cultivation may

302 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 coal reefs, and ecological changes in food webs." EPA, Vocabulary Catalog: Acid Rain Glossary,
available at

https://sor.epa.gov/sor interne t/registrv/termreg/searchandretrieve/gIossariesandkevwordIists/search.do?detaiIs=&gI
ossarvName=Acid%20Rain%20Glossarv (last accessed April 14, 2021).

303 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. 1216006110.

304 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.

305 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, available at

https://sor.epa.gov/sor internet/registrv/termreg/searchandretrieve/gk>ssariesandkevwordlists/search.do?details=&gl
QSSMvMMtg^ (last accessed April 14, 2021).

306 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.

307 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.

308 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.

115

-------
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.309 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.310 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 3.4.2.1-1). 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 taxa311 and fish.312 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.313 The
leading problems contributing to poor biological condition were excess nutrients (especially
phosphorus), loss of shoreline vegetation, and excess sediments.314 For rivers and streams, sites
with a condition rating of poor because of excess nutrients were most prevalent in the mid-

309 Ibid.

310 Ibid

311 Benthic macroinvertebrate taxa are small, bottom-dwelling, aquatic animals and the aquatic larval stages of
insects. EPA, National Aquatic Resource Survey: Indicators: Benthic Macroinvertebrates, available at

1iIMs://ww;m^

d%20beettes (last accessed April 14, 2021).

312 USEPA (2020). National Rivers and Streams Assessment 2013-2014: A collaborative Survey. EPA 841-R-19-
001. Washington, DC. Available at https://www.epa.gov/national-aaiiatic-resoiirce-siirveys/nrsa (last accessed April
14, 2021).

313 USEPA (2020). National Rivers and Streams Assessment 2013-2014: A collaborative Survey. EPA 841-R-19-
001. Washington, DC. Available at https://www.epa.gov/national-aauatic-resoiirce-siirvevs/nrsa (last accessed April
14, 2021).

314 USEPA (2020). National Rivers and Streams Assessment 2013-2014: A collaborative Survey. EPA 841-R-19-
001. Washington, DC. Available at https://www.epa.gov/national-aaiiatic-resoiirce-siirvevs/nrsa (last accessed April
14, 2021).

116

-------
continent ecoregions315 of the nation compared to the eastern and western regions.316 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 2007.317 For lakes, disturbance by nutrients varied by ecoregion (Figure
3.4.2.3-1). 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 3.4.2.3-1), 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).318 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 1970.319

315 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.
Available at https://www.epa.gov/national-aauatic-resoiirce-siiiveYs/nrsa (last accessed April 14, 2021).

316 USEPA (2020). National Rivers and Streams Assessment 2013-2014: A collaborative Survey. EPA 841-R-19-
001. Washington, DC. Available at https://www.epa.gov/national-aaiiatic-resoiirce-siirveYs/nrsa (last accessed April
14, 2021).

317 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. Available at

https://www.epa.gov/sites/default/files/2016-12/documents/nla report dec 2016.pdf (last accessed September 14,
2021).

318 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. Available at

TOM) NCCA 20.1.0 Summary Report Review Draft ES 20101029 (epa.gov) (last accessed September 14, 2021).

319 USEPA. (2015). Report on the Environment. Fish Faunal Intactness.

https://cfpiib.epa.gOv/roe/i indicator. e:fm?i=84

117

-------
Figure 3.4.2.3-1: Locations of ecoregions and coastal areas defined by the USEPA's
National Aquatic Resource Surveys.320

CPL: Coastal Plains

NAP: Northern Appalachians NPL: Northern Plains

SAP: Southern Appalachians

SPL: Southern Plains TPL: Temperate Plains

UMW: Upper Midwest WMT: Western Mountains

West coast

XER: Xeric

Great Lakes Southeast coast Northeast coast

Excess nutrients (both nitrogen and phosphorus)321 in waterbodies can also result in
harmful algal blooms, some of which can produce toxins. Algal blooms, especially
cyanobacteria, can create surface scums that block sunlight and reduce the growth of other algae
and aquatic plants. Because of potential toxin production and composition of fatty acids in their
cells, cyanobacteria are lower quality food for aquatic insects and fish compared to other algae
such as diatoms. Lakes and reservoirs with excess nutrient loads are susceptible to recurring

320 Figures modified from Figure 5.1 in Ecoregions at a Glance in the National Lakes Assessment 2012. USEPA
(2016). National Lakes Assessment 2012. EPA 841-R-16-113 and from USEPA (2015) National Coastal Condition
Assessments 2010. EPA 841-R-15-006. U.S. Environmental Protection Agency. Washington. DC

321 Paerl, H.W., Scott J.T.. McCarthy, M.J., Newell, S.E., Gardner, W.S.. Havens, K.E., Hoffman. D.K., Wilhelm,
S.W. and Wurtsbaugh, W. A. (2016). It takes two to tango: When and where dual nutrient (N & P) reductions are
needed to protect lakes and downstream ecosystems. Environmental science & technology, 50(20): 10805-10813.

118

-------
algal blooms. The western Lake Erie is a good example as it receives nutrient loads from a
drainage area dominated by agricultural land use. Larger streams and rivers are often associated
with nutrient loading from nearby agricultural activities, as well as slower water flow rates and
longer residence times favorable for algal blooms.

In both freshwater and coastal marine systems, algal blooms terminate with microbial
decomposition of algal cells resulting in oxygen depletion or hypoxic zones. The 2017 Gulf of
Mexico hypoxic zone was the largest size measured since 1985, spanning 8,776 square miles.322
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.323 While
fertilizer use by current agricultural practices contribute to much of the nutrient loading that
stimulates algal responses in many waterbodies, the total nutrient budgets324 of some
waterbodies also include internal nutrient recycling of legacy inputs.325

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 damage326 in organ tissues and altered muscle and brain
function.327 Some bacteria can use glyphosate for growth, enhancing microbial proliferation.328
There are cyanobacteria with natural tolerance to glyphosate329 and certain concentrations of
glyphosate can stimulate photosynthesis in a common bloom-forming taxon, Microcystis
aeruginosa,330 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

322 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. Available at https://www.noaa.gov/media~release/Iarge~dead~zone~measured~in~
gulf-of-mexico (last accessed April 14, 2021).

323 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.

324 "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. Available at https://edis.ifas.nfl.edu/pdEfiles/SS/SS56200.pdf (last accessed April 14, 2021).

325 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.

326 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.

327 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.

328 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.

329 Harris TD and Smith VH. 2016. Do persistent organic pollutants stimulate cyanobacterial blooms? Inland Waters
6: 124-30.

330 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.

119

-------
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.331 In 2014, 58% of U.S. rivers and streams were given a rating of poor for the
phosphorus indicator of EPA's National Rivers and Streams Assessment.332 While both corn and
soybean production use glyphosate, corn production can also use atrazine (Table 3.4.2.1-1). 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).333 In 2016, EPA concluded that in areas where
atrazine use is heaviest (mainly in the Temperate Plains ecoregion, Figure 3.4.2.3-1), 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.334 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.

3.4.3 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.335 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.

331 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.

332 USEPA (2020). National Rivers and Streams Assessment 2013-2014: A collaborative Survey. EPA 841-R-19-
001. Washington, DC. Available at https://www.epa.gov/national-aauatic-resoiirce-siirveYs/nrsa (last accessed April
14, 2021).

333 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.

334 USEPA (2016). Refined Ecological Risk Assessment for Atrazine. EPA-HQ-OPP-2013-0266. U.S.
Environmental Protection Agency, Washington, DC.

335 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.

120

-------
3.4.4 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
year between 2016 and 2021.336'337 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.338 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.339
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.

336 USEPA (2021). Frequent questions about underground storage tanks. Available at

https://www.epa.gov/nst/freanent-anestions-abont-niidergroniid-storage-tanks (last accessed April 14, 2021).

337 "UST Confirmed Releases National data 2016-2021," available in the docket.

338 40 CFR Part 280.

339 86 FR 5094 (January 19, 2021).

121

-------
3.4.5 Potential Future Impacts of Annual 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 volume requirements we are finalizing in this action, but an
increase in cropland acreage would generally be expected to lead to more negative soil and water
quality impacts. As 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.340 The additional soil and water quality modeling that would be needed to assess the
potential cumulative impacts of future land use changes for the volume requirements being
finalized in this action would be expansive and could not be performed on the timeline of this
rulemaking.

The volume increases for 2021 and 2022 compared to 2020 described in Chapter 2 due to
biofuels produced from agricultural feedstocks (especially corn and soybeans) would suggest the
potential for an associated increase in crop production, which in turn may impact soil and water
quality. However, if we consider the causal impacts of this rule (as opposed to the volume
changes in Chapter 2), the 2020 and 2021 volumes will have no impact since these volumes will
be entirely retroactive. The 2022 volume requirements could have some impact largely due to
their potential ability to incent greater production of biofuels and their underlying crop-based
feedstocks. However, the magnitude of this potential impact cannot be estimated at this time, as
more information is needed regarding the other factors and the magnitude of the impacts on land
use and management changes. 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. Additional information and modeling are needed to fully assess the degree to
which the annual volume requirements drive land use and management changes that would
impact soil and water quality. Such analysis would be expansive and could not be performed on
the timeline of this rulemaking.

340 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.

122

-------
3.5 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, and finally discuss the potential
future effects of the final RFS volumes.

3.5.1 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.341 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)342 "where groundwater
levels have declined at unsustainable rates."343

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.

3.5.2 Life Cycle Water Use of Biofuels

In the Second Triennial Report to Congress on Biofuels, the water quantity impacts of
biofuels were assessed.344 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

341 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.

342 The High Plains Aquifer is often referred to as the Ogallala Aquifer, which is the largest formation within the
High Plains Aquifer.

343 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.

344 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.

123

-------
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."345 These early
studies characterized this issue as biofuel's water intensity,346 embodied water,347 and water
footprint.348 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).349 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.350 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

o c 1

transportation energy sources.

3.5.2.1 Life Cycle Water Use: 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."352 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).353 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

345 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.

346 King, C. W., & Webber, M. E. (2008). Water intensity of transportation. Environmental Science & Technology,
42(21), 7866.

347 Chiu, Y. W., Walseth, B., & Suh, S. W. (2009). Water embodied inbioethanol in the United States.

Environmental Science & Technology, 43(8), 2688-2692.

348 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.

349 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.

350 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.

351 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.

352 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.

353 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.

124

-------
that irrigated soybean based biodiesel water requirement averaged nearly 1,300 liters of water
per liter of ethanol-equivalent biodiesel (based on energy equivalence).354 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-
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.355 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.356 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 3.5.2.1-1). 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.357 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.358 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."359 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).360

354 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.

355 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.

356 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.

357 Gerbens-Leenes, W., & Hoekstra, A. Y. (2012). The water footprint of sweeteners and bio-ethanol. Environment
international, 40, 202-211.

358 EIA (2018). "Six states account for more than 70% of U.S. fuel ethanol production." Available at
https://www.eia.gov/today inenergy/detait.pfap?id=36892 (last accessed April 14, 2021). See also EIA. (2017,
February 16). "Nebraska State Profile and Energy Estimates: Profile Analysis." Retrieved June 2, 2017, from
https://www.eia. gov/state/analysis.php?sid=NE.

359 NAS (2011). Renewable Fuel Standard: Potential Economic and Environmental Effects of U.S. Biofuel Policy.
National Academy of Sciences. Washington, DC.

360 Gerbens-Leenes, W., & Hoekstra, A. Y. (2012). The water footprint of sweeteners and bio-ethanol. Environment
international, 40, 202-211.

125

-------
Figure 3.5.2.1-1: 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

6.000

5,000

L-

4.000

3,000

2,000

1,000

a Regional range
• County max
-National average

o
o
u.

CD
rtl
?

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)].

3.5.2.2 Life Cycle Water Use: Biofuel Conversion

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.361 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 2012.362 These values are typical of a dry mill facility. Wet mill
facilities require closer to 4 gallons per gallon of ethanol.363 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
cooling towers.364 Some facilities have set goals to both reduce water use and minimize
discharges.

361 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.

362 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.

363 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.

364 Scliill, S. R. (2017) Water: Lifeblood of the Process. Ethanol Producer Magazine. January 24, 2017.

http://www.ethanolproducer.com/articles/14049/water-lifeblood-of-the-process. See also Jessen, H. (2012) Dropping

126

-------
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.365 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).366

There are no recently published surveys of water consumption representing all current
biofuel 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,367
and likely a source of water for facilities along the HPA (see Figure 3.5.3-3 below).

3.5.2.3 Life Cycle Water Use: 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,
South Dakota, Nebraska and Kansas) for groundwater.368 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.369 370 The Wu et al (2018) analysis does not include biodiesel. The most

Water Use. Ethanol Producer Magazine. June 12, 2012. Available at

http://www.ethairolprodiicer.com/articles/8860/dropping~water~use (last accessed September 16, 2021).

365 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.

366 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.

367 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.

368 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/anlpnbs/201! 043.pdf

369 Id.

370 Sun, Pinping; Estimation of U.S. refinery water consumption and allocation to refinery products; Fuel; Volume
221; June 18, 2018.

127

-------
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.371 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.372 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.373 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. It is likely, though, that FOG-based biodiesel and
renewable diesel are more similar in water consumption to petroleum diesel fuel which it
displaces.

3.5.3 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
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.374 These are areas unlikely to have sufficient
precipitation to support corn or soybean cultivation.375 While difficult to attribute how much
additional water use might be required as a result of the volumes finalized 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.

371 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.

372 Haas, M.J, A process model to estimate biodiesel production costs, Bioresource Technology 97 (2006) 671-678.

373 Sun, Pinping; Estimation of U.S. refinery water consumption and allocation to refinery products; Fuel; Volume
221; June 18, 2018.

374 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.

375 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).

128

-------
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 20 1 8.376
From 2013 to 2018, there was an increase in total irrigated acres of nearly 0.6 million acres in the
U.S.3// 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 2013.378 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.379 Figure 3.5.3-1 shows acres of irrigated land in
2012, the most recent year of data for which this figure is available.

Figure 3.5.3-1: Acres of irrigated land in 2012

Acres of irrigated land, 2012

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-
management/irrigation-water-use/background.aspx

376 USDA NASS (2018). 2018 Irrigation and Water Management Survey. Available at

https://www.nass.usda.gov/Publications/AgCensus/2017/Qnline Resources/Farm and Ranch Irrigation Survev/fri
s.pdf (last accessed April 14, 2021).

3 USDA NASS (2018). 2018 Irrigation and Water Management Survey. Available at

https://www.nass.usda.gov/Publications/AgCensus/2017/Online Resources/Farm and Ranch Irrigation Survev/fri
s.pdf (last accessed April 14, 2021).

378 USDA NASS (2018). 2018 Irrigation and Water Management Survey. Available at

https://www.nass.usda.gov/Publications/AgCensus/2017/Qnline Resources/Farm and Ranch Irrigation Survev/fri
s.pdf (last accessed April 14, 2021).

379 USDA NASS (2018). 2018 Irrigation and Water Management Survey. Available at

https://www.nass.usda.gov/Publications/AgCensus/2017/Qnline Resources/Farm and Ranch Irrigation Survev/fri
s.pdf (last accessed April 14, 2021).

129

-------
Figure 3.5.3-2 shows corn and soybean areas and share of irrigated acres. Irrigated com
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 3.5.3-2 below show the distribution of corn
and soybean acres, as a share of total cropland acres, while the bottom rows of Figure 3.5.3-2
show the percent of irrigated corn and soybean acres relative to total acres for those crops
(measures as harvested acres).

Figure 3.5.3-2: Percent of irrigated corn and soybean acres relative to total acres,
respectively

Top Left: Acres of Corn Harvested as a Percent of Harvested Cropland Acreage; Top Right: (b) Acres of Soybean
Harvested as a Percent of Harvested Cropland Acreage: Bottom left: Irrigated Corn as a Percent of Total Com
(Harvested Acres), Irrigated Soybeans as a Percent of Total Soybeans (Harvested Acres).

Source: USDA Agricultural Census Web Maps (Accessed April 14, 2021).

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 ITPA (also called the Ogallala Aquifer). A 2011 report by the

130

-------
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."380
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.
(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 3.5.3-3).
However, this correlation does not necessarily mean there is a direct, causal relationship between
biofuel production and groundwater depletion.

Figure 3.5.3-3: Relative conversion rates of arable non-cropland to cropland (2008-2012).

0 .4 .8 1.2 2.0 2.7 3.9 5.5 7.3 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

380 NAS (2011). Renewable Fuel Standard: Potential Economic and Environmental Effects of U.S. Biofuel Policy.
National Academy of Sciences. Washington. DC.

131

-------
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 HP A 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
productivity in agricultural practices."381 Figure 3.5.3-3, above, shows the relative conversion
rates of arable cropland to non-cropland, as well as the location of the HPA and biofuel
conversion facilities. Figure 3.5.3-4, below, also shows the HPA, but shows the absolute changes
in groundwater levels, from predevelopment to 2017 based on data from USGS.382 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 Laboratory383) 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 3.5.3-4). 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.

381 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.

382 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

383 https://maps.nrel.gov/biofuels-atlas/

132

-------
Figure 3.5.3-4: Water-level changes in the High Plains Aquifer, predevelopment (about
1950) to 2015.

••

•• ' I
* •

•*¦»?« pt*t*

ir v v *

IEBRASI . * * a.

o»- ©« •

9- • ' * *

» 0 • • <,Linc°ln # *

COLORADO

-------
corn. The value of irrigation was lower for soybeans.384 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.385 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.386 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."387

3.5.3.2 Non-Cropland Biofuels andNon-U.S. Crops

The Second Triennial Report to Congress on Biofuels and the 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 3.3, 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.388 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

384 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.

385 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).

386 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.

387 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.

388 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.

134

-------
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.

3.5.4 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 ethanol389 or focus on future cellulosic
feedstocks.390 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.391 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.

389 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.

390 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.

391 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.

135

-------
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 finalized 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 finalized volumes. Specifically, the volume
increases for 2021 and 2022 compared to 2020 that are described in Chapter 2 due to biofuels
produced from agricultural feedstocks (especially corn and soybeans) would suggest the
potential for an associated increase in crop production, which in turn are likely to increase
irrigation pressure on water resources. However, if we consider the causal impacts of this rule (as
opposed to the volume changes in Chapter 2), the 2020 and 2021 volumes will have no impact,
since these volumes will be entirely retroactive. The 2022 volume requirements could have some
impact largely due to their ability to incent greater production of biofuels and their underlying
crop feedstocks. There is uncertainty in projecting changes in acreage and irrigation rates
associated with corn, soybean, 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 3.4, we note that there may be potential effects on water and soil quality. While we
could not quantify these effects, as described in Chapter 3.4, the potential for negative effects is
an area of ongoing concern and research.

3.6 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.392 The United Nations Millennium Ecosystem Assessment393 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 renewable fuel
volumes obligations we are finalizing in this rule specifically.

392 US EPA website on Ecosystem Services. Available at: https://www.epa.gov/eco-research/ecosystem-services

393 Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-being: Synthesis. Island Press,
Washington, DC.

136

-------
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 3.6-1). 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.

137

-------
Figure 3.6-1: Framework for Considering the Impact of the RFS Volumes on Ecosystem
Services

Action: RFS Annual Rule (2020-2022)

Biophysical Changes

Human Weil-Being
Changes

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

Hydrology, 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

Agricultural Product Value

Wildlife Product Value

Wildlife Existence Value

Recreation Value

Soil Quality

138

-------
Chapter 4: 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 reset authority to establish volumes. U.S. energy security is
broadly defined as the continued availability of energy sources at an acceptable price.394 Most
discussions of U.S. energy security revolve around the topic of the economic costs of U.S.
dependence on oil imports.395 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. 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.396 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.397

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 will 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 (i.e., droughts). To the extent that renewable fuel
price shocks are not strongly correlated with oil price shocks, blending renewable fuels with oil
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 final volumes for 2021 and 2022.398

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 2020.399 The U.S. has
increased its production of oil, particularly "tight" (i.e., shale) oil over the last decade. As a result
of the recent increase in U.S. oil production, the U.S. became a net exporter of the aggregate of
crude oil and refined petroleum products in 2020. The U.S. is projected to be a modest net
importer of crude oil and refined petroleum products through 2021-2022 before becoming a net

394 Greene, D. 2010. Measuring energy security: Can the United States achieve oil independence? Energy Policy 38,
pp. 1614-1621.

395 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.

396 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).

397 Greene, D. 2010. Measuring energy security: Can the United States achieve oil independence? Energy Policy 38,
pp. 1614-1621.

398 Our decision to assess the 2021 and 2022 volumes against the baseline of 2020 volumes for purposes of our
quantitative analyses, including the energy security analysis presented here, is set forth in Chapter 2.

399 U.S. EIA. Total Energy. Annual Energy Review. Table 3.1. Petroleum Overview. April 2021.

139

-------
exporter of the aggregate of crude oil and petroleum products for the foreseeable future.400 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.401

Given that the U.S. was a net exporter of crude oil and refined petroleum products in
2020 and is projected to be a modest net importer through 2021-2022, one could reason that the
U.S. does not have a significant energy security problem anymore. However, U.S. refineries still
rely on significant imports of heavy crude oil from potentially unstable regions of the world.
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 time frame of this analysis is hard to quantify. These factors contribute
to the continued vulnerability of the U.S. economy to episodic oil supply shocks and price
spikes.

We recognize that because the U.S. is a participant in the world market for crude oil and
refined petroleum products, our economy cannot be shielded from world-wide price shocks.402
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 projected modest net petroleum importer in the
world petroleum market in 2021-2022. The potential for supply disruptions has not been
eliminated, though, due to the continued need to import petroleum to satisfy the demands of the
U.S. petroleum industry.403

4.1 Review of Historical Energy Security Literature

Energy security discussions are typically based around the concept of the oil import
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.

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 1982).404
Plummer (1982) provided valuable discussion of many of the key issues related to the oil import

400 U.S. Energy Information Administration. 2021 .Annual Energy Outlook 2021. Reference Case. Table A11.
Petroleum and Other Liquids Supply and Disposition.

401 See EIA https://www.eia.gov/energyexplained/oil-and-petroleum-products/imports-and-exports.php

402 Bordoff, J. 2019. The Myth of US Energy Independence has Gone Up in Smoke. Foreign Policy. September 18.
https://foreignpolicy.com/2019/09/18/the-myth-of-u-s-energy-independence-has-gone-up-in-smoke/

403 Foreman, D. 2018. Why the US must Import and Export Oil; American Petroleum Institute. June 14.
https://www.api.org/news-policy-and-issues/blog/2018/06/14/why-the-us-must-import-and-export-oil

404 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.
https://emf.stanford.edu/publications/emf-6-world-oil.

140

-------
premium as well as the analogous oil stockpiling premium.405 Bohi and Montgomery (1982)
detailed the theoretical foundations of the oil import premium and established many of the
critical analytic relationships.406 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.407'408 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). 409>410

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.411 Analyses by Nordhaus (2007) and Blanchard and
Gali (2010) question the impact of oil price shocks on the economy in the early 2000 time
frame.412 They were motivated by attempts to explain why the economy actually expanded
during the oil shock in the early-2000s timeframe, 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
(2010)), while other work continues to find evidence regarding the economic importance of oil
shocks.413 For example, Baumeister and Peersman (2011) 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

405 Plummer, J. (Ed.). 1982. Energy Vulnerability, "Basic Concepts, Assumptions and Numerical Results," pp. 13 -
36, Cambridge MA: Ballinger Publishing Co.

406 Bohi, D. and Montgomery, D. 1982. Social Cost of Imported and U.S. Import Policy, Annual Review of Energy,
7:37-60.

407 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.

408 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.

409 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.

410 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.

411 National Research Council. 2009. Hidden Costs of Energy: Unpriced Consequences of Energy Production and
Use. National Academy of Science, Washington, DC.

412 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.

413 Rasmussen, T. and Roitman, A. 2011. Oil Shocks in a Global Perspective: Are We Really That Bad. IMF
Working Paper Series.

141

-------
turned out to have a similar effect on output as in the earlier data."414 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)).415'416 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."417

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)).418 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.419 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
2009a)."420

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 1979-2011.421
They state: "The results indicate that the economic consequences of a supply-driven oil-price
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.

414 Hamilton, J. 2012. Oil Prices, Exhaustible Resources, and Economic Growth. In Handbook of Energy and
Climate Change. Retrieved from http://econweb.ucsd.edu/~jhamilto/handbook_climate.pdf

415 Kim, D. 2012. What is an oil shock? Panel data evidence. Empirical Economics, Volume 43, pp. 121-143.

416 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.

417 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).

418 Baumeister C., Peersman, G. and Van Robays, I. 2010. "The Economic Consequences of Oil Shocks: Differences
across Countries and Time," RBA Annual Conference Volume in: Renee Fry & Galium Jones & Christopher Kent
(ed.), Inflation in an Era of Relative Price Shocks, Reserve Bank of Australia.

419 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.

420 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.

421 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).

142

-------
EPA's assessment of the energy security literature finds that there are benefits to the U.S.
from reductions in its oil imports. But there is some debate as to the magnitude, and even the
existence, of energy security benefits from U.S. oil import reductions. Over the last decade,
differences in economic impacts from oil demand and oil supply shocks have been distinguished.
The oil security premium calculations in this analysis (described in 4.4) are based on price
shocks from potential future supply events only. Oil supply shocks, which reduce economic
activity, 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 as a result
of this final RFS rule, the final 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.

4.2 Review of Recent Energy Security Literature

There have also been a handful of recent studies that are relevant for the issue of energy
security: one by Resources for the Future (RFF), a study by Brown, two studies by Oak Ridge
National Laboratory (ORNL), and a couple of studies, Newell and Prest and Bjornland et al., on
the responsiveness of U.S. tight oil (i.e., shale oil) to world oil price changes. We provide a brief
review and high-level summary of each of these studies below.

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.422 In a follow-on study, Brown summarized the RFF study results as well.423 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 timeframe. 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
developed by Balke and Brown.424 The second set of modeling frameworks use alternative

422 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.

423 Brown, S. 2018. New estimates of the security costs of U.S. oil consumption", Energy Policy, 113 pp. 171-192.

424 Balke, N. and Brown, S. 2018. "Oil Supply Shocks and the U.S. Economy: An Estimated DSGE Model." Energy
Policy, 116, pp. 357-372.

143

-------
structural vector autoregressive models of the global crude oil market.425 The last of the models
utilized is the National Energy Modeling System (NEMS).426

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 will be 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 fifteen years are believed to be predominantly "demand shocks": for example, 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 percent of global crude oil production capacity.427 On September 16th, the first full day
of commodity trading after the attack, both Brent and WTI crude oil prices surged by

425 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, https://doi.org/10.1002/jae.2322; 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.

426 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.

427 U.S. Energy Information Administration. September 23, 2019. "Saudi Arabia crude oil production outage affects
global crude oil and gasoline prices." Today in Energy.

144

-------
$7.17/barrel and $8.34/barrel, 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.428 Tanker loading estimates from third-party data sources indicated that loadings
at two Saudi Arabian export facilities were restored to the pre-attack levels.429 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 current world oil price spike being
experienced in 2022. World oil prices have risen fairly rapidly since the beginning of 2022. For
example, as of January 3rd, 2022, the WTI crude oil price was roughly $76 per barrel. The WTI
oil price increased to roughly $123 per barrel on March 8th, 2022, a 62 percent increase.430 High
and volatile oil prices in 2022 are a result of a combination of several factors: supply not rising
fast enough to meet rebounding world oil demand from increased economic activity as Covid-19
recedes, reduced supply from some leading oil-producing nations and geopolitical concerns. It is
not clear to what extent the current oil price volatility will continue, or even increase, or be
transitory. Since both significant demand and supply factors are influencing world oil prices in
2022, it is not clear how to evaluate unfolding oil market price trends from an energy security
standpoint. Thus, the attack of the Abqaiq oil processing facility in Saudi Arabia and the
unfolding events in the world oil market in 2022 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.431

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.432 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 dataset for the meta-analysis, ORNL undertakes a
literature search of peer reviewed journal articles and working papers between 2000 and 2015
that contain estimates of oil demand elasticities. The dataset consisted of 1,983 observations

428 Ibid.

429 Ibid.

430 U.S. Energy Information Administration daily spot prices, available at:
https://www.eia.gov/dnav/pet/pet_pri_spt_sl_d.htm (accessed March 2022)

431 The Hurricanes Katrina/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. The Katrina/Rita Hurricanes 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.

432 Uria-Martinez, R., Leiby, P., Oladosu, G., Bowman, D., Johnson, M. 2018. Using Meta-Analysis to Estimate
World Oil Demand Elasticity, ORNL Working Paper.

145

-------
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.433 Nineteen 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
percent confidence interval of-0.006 to -0.036.

Only in recent years have the implications of the "tight 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 economics literature of the tight (i.e., shale/unconventional) 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 shale oil has eroded the ability of OPEC to set world oil prices to some
degree, since OPEC cannot directly influence shale 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 shale oil expansion is a
relatively recent trend, it is difficult to know how much of an impact the increase in shale oil is
having, or will have, on OPEC behavior.

Two 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?434 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 for oil wells,
using a detailed dataset of 150,000 oil wells, during the time frame of 2005-2017 in five major
oil-producing states: Texas, North Dakota, California, Oklahoma, and Colorado.435 For both
conventional oil wells and shale oil wells, 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 shale 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 unconventional 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 shale oil supply

433 Oladosu, G., Leiby, P., Bowman, D., Una-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.

434 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."

435 Newell, R. and Prest, B. 2019. The Unconventional Oil Supply Boom: Aggregate Price Response from
Microdata, The Energy Journal, Volume 40, Issue Number 3.

146

-------
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 oil/shale oil.436 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. Both the results from
the Newell and Prest and Bjornland et al. 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.

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 (set forth in 4.4) focuses
on those incremental social costs that follow from the resulting changes in net imports,
employing the usual oil import premium measure.

4.3 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
response option should a disruption in commercial oil supplies threaten the U.S. economy. 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.

EPA also has considered the possibility of quantifying the military benefits components
of energy security but has 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 hard to quantify. These include possible costs of U.S.
military programs to secure oil supplies from unstable regions of the world, the energy security

436 Bjornland, H., Nordvik, F. and Rolirer, M. 2021. "Supply flexibility in the shale patch: Evidence from North
Dakota," Journal of Applied Economics, February.

147

-------
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 an 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 final rule.

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
accounts rather than by mission, the allocation to particular missions is not always clear.437
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).438

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).439 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.440 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 to 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.

437 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.

438 Koplow, D. and Martin, A. 1998. Fueling Global Warming: Federal Subsidies to Oil in the United States.
Greenpeace, Washington, D.C.

439 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.

440 Stern, R. 2010. "United States cost of military force projection in the Persian Gulf, 1976-2007." Energy Policy
38, no. 6. June: 2816-2825. http://linkinghub.elsevier.com/retrieve/pii/S0301421510000194.

148

-------
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.441 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 and $74
billion annually. 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 percent 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
global oil supplies.442 This is approximately 16 percent 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 per barrel of crude oil, or $0.28 per gallon. According to SAFE, a more comprehensive
estimate suggests the costs could be greater than $30 per barrel, or over $0.70 per gallon.443

As in the examples above, an incremental analysis can estimate how military costs would
vary if the oil security mission is 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
are partially reduced, as is projected to be a consequence of this final rule. Partial reduction of
U.S. oil use surely 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.444 EPA is 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 final rule.

4.4 Energy Security Impacts

4.4.1 U.S. Oil Import Reductions

Over the time frame of this quantitative energy security analysis, 2021-2022, the U.S.
Department of Energy's (DOE) Energy Information Administration's (EIA) Annual Energy

441 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.

442 Securing America's Future Energy. 2018. Issue Brief. The Military Cost of Defending the Global Oil Supply.

443 Ibid.

444 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.

149

-------
Outlook (AEO) 2021 (Reference Case) projects that the U.S. will be both an exporter and an
importer of crude oil.445 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 oils 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 3.1 MMBD in 2021 and 2.8 MMBD in 2022. U.S. crude oil imports, meanwhile,
are projected to be 7.6 MMBD in 2021 and 7.8 MMBD in 2022. The AEO 2021 also projects
that U.S. net oil refined product exports will increase from 3.6 MMBD in 2021 to 4.6 MMBD in
2022. Given the pattern of U.S. crude oil exports/imports, and the U.S.'s net oil refined product
exports, the U.S. is projected to be a modest net crude oil and refined product importer of 0.9
MMBD in 2021 and 0.3 MMBD in 2022. After the 2021-2022 time period, the AEO2021
projects that the U.S. will become a net exporter of crude oil and refined oil products for the
foreseeable future.

U.S. oil consumption is projected to have decreased significantly in 2020 to 17.2 MMBD
as a result of social distancing and quarantines that limited personal mobility as a result of the
Covid-19 pandemic. U.S. oil consumption is projected to rebound to 18.9 MMBD in 2021 and
18.7 MMBD in 2022. It is not just U.S. crude oil imports alone, but both imports and
consumption of petroleum from all sources and their role in economic activity, that exposes the
U.S. to risk from price shocks in the world oil price. During the 2021-2022 time frame, 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 9.3, EPA estimates changes in U.S. petroleum consumption and U.S. oil
imports attributable to this final rule. A 99 percent oil import reduction factor is calculated by
examining changes in petroleum demand and its effect on both imported versus domestic crude
oil, and imported petroleum refined products considering two separate economic cases, the
Reference Case and the Low Economic Growth Case, modeled by EIA in its 2021 Annual
Energy Outlook.446 The 99 percent oil import 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. oil consumption in the different AEO cases. Based upon the changes in oil consumption
estimated by EPA and the 99 percent oil import factor, the reduction in U.S. oil imports as a
result of the final RFS fuel volume cases are estimated in Table 4.4.1-1 for the 2021-2022 time
frame. Included in Table 4.4.1-1 are estimates of U.S. crude oil exports and imports, net oil
refined product exports, net crude oil and refined product imports, and U.S. oil consumption.
These figures place the reduction in U.S. oil imports in the context of U.S. oil market trends
during the 2021-2022 time frame.

While this final RFS annual rule is projected to reduce U.S. oil imports modestly, it also
results in some changes in U.S. renewable fuel imports. We do not have a methodology to
estimate the energy security impacts of an increased use of imported renewable fuels. To the
extent that renewable fuel price shocks are uncorrected with oil price shocks, blending
renewable fuels with oil can provide energy security benefits. This is so because most renewable

445 U.S. Energy Information Administration. Annual Energy Outlook 2021. Reference Case. Table All. Petroleum
and Other Liquids Supply and Disposition.

446 AEO 2021 Change in product demand on imports; docketed spreadsheet.

150

-------
fuel price shocks are usually weather-related (i.e., droughts), while oil price shocks tend to result
when oil markets are tight, or due to geopolitical events. When including the Supplemental
Volumes from this final rule, the estimated cumulative increase in net imports of renewable fuels
is less than one percent of the estimated cumulative change in U.S. oil import reductions. When
excluding the Supplemental Volumes from this final rule, EPA estimates that there will be a net
decrease in renewable fuel imports, due to the projected decrease in imported sugarcane ethanol
offsetting a small increase in imported biodiesel.

Table 4.4.1-1: Projected Trends in U.S. Crude Oil Exports/Imports, Net Refined Oil
Product Exports, Net Crude Oil and Refined Product Imports, Oil Consumption and U.S.
Oil Import Reductions Resulting from the Final RFS Annual Rule from 2021 to 2022
(MMBD)a

2021

2022

U.S. Crude Oil Exports

3.1

2.8

U.S. Crude Oil Imports

7.6

7.8

U.S. Net Oil Refined Product Exports'3

3.6

4.6

U.S. Net Crude Oil and Refined Product Imports

0.9

0.3

U.S. Oil Consumption0

18.9

18.7

Reduction in U.S. Oil Imports from the Final
Renewable Fuel Standards'1

0.06

0.12d

a The AEO 2021 Reference Case, Table A11. Values have been rounded off from the AEO 2021, so the totals may
not add up to the AEO estimates.

b We subtract out renewable fuel net exports from U.S. Net Oil Refined Product Exports to account for only
petroleum-based product exports.

0 Calculated from AEO Table A11 as "Total Primary Supply" minus "Biofuels".
d Estimated changes in U.S. imported crude oil/refined oil products are presented in Table 9.3-4.
e Excludes supplemental standard.

4.4.2 Oil Security Premiums Used for this Final Rule

In order to understand the energy security implications of reducing U.S. oil imports, EPA
has worked with Oak Ridge National Laboratory (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 study
entitled, "The Energy Security Benefits of Reduced Oil Use, 2006-2015," completed in 2008.447
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.448 This same approach
was used to estimate energy security benefits for the March 2010 RFS2 final rule (75 FR 14839-
14842).449 ORNL has updated this methodology periodically for EPA to account for updated
projections of future energy market and economic trends reported in the U.S. EIA's AEO.

447 Leiby, Paul N. 2008. Estimating the Energy Security Benefits of Reduced U.S. Oil Imports, Final Report,
ORNL/TM-2007/028, Oak Ridge National Laboratory, Rev. March 14.

448 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.

449 See 40 CFR Part 80, Regulation of Fuels and Fuels Additives: Changes to the Renewable Fuel Standard Program;
Final Rule, March 26, 2010. https://www.govinfo.gov/content/pkg/FR-2010-03-26/pdf/2010-3851.pdf

151

-------
The ORNL methodology is used to compute the oil import premium (concept defined in
Chapter 4.1) per barrel of imported oil. The values of U.S. oil 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
percent confidence 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 gross domestic product 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 final rule, EPA is using oil security 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 the AEO 2021 into its model.450 EPA only
considers the avoided macroeconomic disruption/adjustment oil security 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.451 EPA also does not
consider the effect of this final rule on the costs associated with existing energy security policies
(e.g., maintaining the Strategic Petroleum Reserve or strategic military deployments), which are
discussed in Chapter 4.3.

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.

450 The oil market projection data used for the calculation of the oil security premiums came from EIA's Annual
Energy Outlook (AEO) 2021, supplemented by the latest EIA international projections from the Annual Energy
Outlook (AEO)/International Energy Outlook (IEO) 2019. Global oil prices 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 2021. Global and OECD Europe supply/demand
projections as well as OPEC oil production share are from IEO 2019. The need to combine AEO 2021 and IEO 2019
data arises due to two reasons: (a) EIA stopped including Table 21 "International Petroleum and Other Liquids
Supply, Disposition, and Prices" in the U.S.-focused Annual Energy Outlook after 2019, (b) EIA does not publish
complete updates of the IEO every year.

451 See the most recent EPA GHG vehicle rule for a discussion of why monopsony oil security premiums are
considered transfer payments, 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, pp. 14841—
14842.

152

-------
In addition, EPA and ORNL have worked together to revise the oil security premiums
based upon recent energy security literature. Based upon EPA and ORNL's review of the recent
energy security literature, EPA is updating its macroeconomic oil security premiums for this
final RFS annual rule. The recent economics literature (discussed in Chapter 4.2) 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 shale oil is influencing the macroeconomic oil security premiums in this final
rule, other than how it significantly reduces net U.S. oil imports.

First, we assess the price elasticity of demand for oil. In previous EPA RFS rulemakings,
EPA has used a short-run elasticity of demand for oil of-0.045.452 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. According to the IEA, the share of global oil consumption
attributed to the transportation sector grew from 60 percent in 2000 to 66 percent in 2018.453 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, particularly in the time frame of this final rule. 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 final rule, we are increasing the
short-run price elasticity of demand for oil from -0.045 to -0.07, a 56 percent increase.454 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. For previous EPA
rulemakings, a GDP elasticity to an oil shock of-0.032 was used.455 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.
Thus, for this final rule, EPA is using a GDP elasticity of-0.021, a 34 percent reduction from the

452 For example, see Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program;
Proposed Rule, 40 CFR Part 80 (2010).

453 IEA, Data and Statistics, https://www.iea.org/data-and-

statistics?country=WORLD&fuel=Oil&indicator=OilProductsConsBySector (Accessed March 2021)

454 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.

455 Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program; Proposed Rule, 40 CFR
Part 80. 2010.

153

-------
GDP elasticity used previously (i.e., the -0.032 value).456 This GDP elasticity is within the range
of RFF's "new literature" elasticity, -0.018, and the elasticity EPA has used in previous
rulemakings, -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. Finally, we have not
factored in how increases in U.S. tight oil might influence U.S. oil security values, other than
how they significantly reduce net U.S. oil imports, given the complexity of this issue. There is
not currently a robust methodology for estimating the energy security benefits of tight oil in
comparison to conventional oil.

Table 4.4.2-1 below provides estimates of EPA's macroeconomic oil security premium
estimates for 2021 and 2022. The macroeconomic oil security premiums are relatively steady
over the time period of this final rule at $3.17/barrel in 2021 (7.55 cents/gallon) and $3.30/barrel
(7.86 cents/gallon) in 2022 (in 2021 U.S. dollars).

Table 4.4.2-1: Macroeconomic Oil Security Premiums for 2021-2022 (2021$/barrel)a

Year

Avoided Macroeconomic
Disruption/Adjustment Costs
(Range)

2021

$3.17
($1.10 - $5.68)

2022

$3.30
($1.04 - $5.91)

a Top values in each cell are mean values. Values in parentheses are 90 percent confidence intervals.

We note that the quantified energy security benefits of this rule, while significant, are
dwarfed by the quantified costs, which are more than an order of magnitude greater. Even if EPA
were to use the lowest or highest end of the range for oil security premiums (described in the
above table), 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 conclusions as we
factor in quantified benefits and costs with regard to the proper volumes in this final rule.

4.4.3 Energy Security Benefits

Estimates of the total annual energy security benefits of the final 2021 and 2022 RES
standards are based on the ORNL oil security premium methodology with updated oil security
premium estimates reflecting the recent energy security literature and using the AEO 2021.
Annual per-gallon benefits are applied to the reductions in U.S. crude oil/refined oil products
imports, 893 millions of gallons for 2021, and 2,042 millions of gallons, including the
Supplemental Volumes, and 1904 millions of gallons, excluding the Supplemental Volumes, in
2022, respectively.457 We do not consider military cost impacts or the monopsony effect of U.S.

456 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.

457 See Tables 9.3-4 and 9.3-5. Energy security benefits are assessed for volume changes relative to final 2020
volumes. Thus, the 2022 volume change is assumed to be the sum of the U.S. 2021/2020 crude oil/refined oil
products imports change and the U.S. 2022/2021 crude oil/refined oil products imports change. Note that we do not

154

-------
crude oil import/refined oil products changes. There is currently no available methodology to
quantify the energy security benefits of imported renewable fuels, so we do not account these
benefits. The energy security benefits are estimated to be 67 million dollars in 2021 and 160
million dollars in 2022 (in 2021 dollars), including the Supplemental Volumes, from this final
rule and 150 million dollars in 2022 (in 2021 dollars), excluding the Supplemental Volumes
from this final rule, as presented in Table 4.4.3-1.

Table 4.4.3-1: Annual Energy Security Benefits of the Final 2021/2022 RFS Standards
(Millions of 2021$)

U.S. Crude Oil

/Refined Oil Products

Import Reductions21

Benefits

Year

(Millions of Gallons)

(Millions of 2021$)

2021

893

2022

Excluding Supplemental Volumes

1,904

150

Including Supplemental Volumes

2,042

160

2021 and 2022 Combined

Excluding Supplemental Volumes

2,797

217

Including Supplemental Volumes

2,935

227

a U.S. import reductions used for the energy security analysis in this section are a combination of reduced U.S.
imports of crude oil, gasoline and diesel fuels from Tables 9.3-4 and 9.3-5 converted to crude oil-equivalent gallons.

separately assess the energy security implications of renewable diesel/sugar cane ethanol volumes which are
expected to be imported by the U.S.; we only consider the energy security impacts of the petroleum-based fuel that
renewable fuel is assumed to displace.

155

-------
Chapter 5: Rate of Production 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 2020 and 2021, we have at the time of this final rule actual data of past
production.458 We thus focus on our discussion on 2022, where 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.459

We also project the use (i.e., consumption) of qualifying renewable fuels in the United
States. Sometimes, we term this the "supply" of biofuels. For 2020 and 2021, we have actual
data of past consumption. For 2022, this projection 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 the RIA. 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 2022.

We discuss the production and use of each major type of biofuel in turn: cellulosic
biofuel (5.1), biodiesel and renewable diesel (5.2), advanced ethanol (5.3), other advanced
biofuels (besides ethanol, biodiesel, and renewable diesel) (5.4), and corn ethanol (5.5).

5.1 Cellulosic Biofuel

In the past several years, production of cellulosic biofuel has continued to increase.
Cellulosic biofuel production reached record levels in 2021, driven by CNG and LNG derived
from biogas.460 The projected volumes of cellulosic biofuel production in 2022 is even higher
than the volume produced in 2021. Production of liquid cellulosic biofuel has remained limited
in recent years (see Figure 5.1-1). This chapter describes our assessment of the rate of production
of qualifying cellulosic biofuel in 2020 - 2022 and some of the uncertainties associated with the
projected volume for 2022. These assessment address both our obligation to project the amount
of cellulosic biofuel under the cellulosic waiver authority, CAA section 21 l(o)(7)(D)(i), and 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). We also discuss the potential availability of
cellulosic biofuels produced from other sources that are not included in our volume projections
for 2022.

458 In the proposed rule, which was released in late 2021, we also projected 2021 volumes. However, since actual
data for 2021 is available now, we are using that data in this final rule.

459 This is what we generally mean when we use the term biofuel "production" in this chapter and do not specify
whether we are discussing domestic production or imports.

460 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.

156

-------
Figure 5.1-1: Cellulosic RINs Generated (2013-2021)

600

2013 2014 2015 2016 2017 201S 2019 2020 2021

¦ CNG/LNG Derived from Biogas ¦ Liquid Cellulosic Biofuel

To project the volume of cellulosic biofuel production in 2022, 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, 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
in 2022. We also based our projection of cellulosic biofuel production in 2022 on EIA's
projection.461

To project the range of potential production volumes of liquid cellulosic biofuel for 2022
we used the same general methodology as the methodology used in the 2018 - 2020 RFS annual
rules. We have adjusted the percentile values used to select a point estimate within a projected
production range for each group of companies based on updated information (through 2021) with
the objective of improving the accuracy of the projections. To project the production of
cellulosic biofuel RINs for CNG/LNG derived from biogas, we used the same general year-over-
year growth rate methodology as in the 2018 - 2020 final rules, with updated RIN generation
data through December 2021. 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.

Our discussion of the rate of production of cellulosic biofuel also discusses the
availability of potential cellulosic biofuels produced from two other sources: cellulosic ethanol
derived from corn kernel fiber (beyond those facilities currently registered with EPA) and
renewable electricity used as transportation fuel. At this time, we are not projecting that
significant volumes of cellulosic biofuel will be produced using these pathways.

461 Letter from Stephen Nalley, Acting Administrator, EIA to Michael Regan Administrator, U.S. EPA. October 29,
2021.

157

-------
The balance of this chapter is organized as follows. Chapter 5.1.1 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. Chapter 5.1.2 discusses the methodologies used by EPA to project
cellulosic biofuel production in 2022. Chapter 5.1.3 discusses other potential sources of
cellulosic biofuel. Chapter 5.1.4 discusses the projected rate of production and import of
cellulosic biofuel volume for 2022.

5.1.1 Cellulosic Biofuel Industry Assessment

In this chapter, 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 2022. 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 2022.

To project the rate of cellulosic biofuel production for 2022, 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 the
comments received on the proposed rule. As discussed in greater detail in Chapter 5.1.3.1, our
projection of liquid cellulosic biofuel is based on a facility-by-facility assessment of each of the
likely sources of cellulosic biofuel in 2022, while our projection of CNG/LNG derived from
biogas 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 2022, 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.

5.1.1.1 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 2015-2021462 are shown in Table 5.1.1.1-1. These data indicate that
EPA's projection was lower than the actual number of cellulosic RINs made available in 2015

462 2021 is the last year for which complete data is available at the time of this final rule..

158

-------
and 2018463 and higher than the actual number of RINs made available in 2016, 2017, 2019, and
2020. 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 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 5.1.1.1-1: Projected and Actual Cellulosic Biofuel Production (2015-2021) (million
gallons)

Projected Volume3

Actua

Production Volumeb

Liquid

CNG/LNG

Total

Liquid

CNG/LNG

Total

Cellulosic

Derived

Cellulosic

Derived

Cellulosic

Year

Biofuel

from Biogas

Biofuel0

Biofuel

from Biogas

Biofuel0

2015d

0.5

52.8

53.3

2016

207

230

4.1

186.2

190.3

2017

298

311

11.7

239.4

251.1

2018

274

288

10.6

303.9

314.5

2019

399

418

11.1

402.8

413.9

2020

577

593

2.1

502.5

504.6

2021

N/A

0.7

560.4

561.1

'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).

bActual 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.

Total cellulosic biofuel may not be precisely equal to the sum of liquid cellulosic biofuel and CNG/LNG derived
from biogas due to rounding.

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 2017).464 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. Given
this and the fact the liquid cellulosic biofuel volume is a small fraction of the total cellulosic
biofuel volume, we are again applying the same general approach we first used in the 2018 final
rule: using percentile values based on actual production in previous years, relative to the
projected volume of liquid cellulosic biofuel in these years. We believe that the use of the

463 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.

464 82 FR 58486 (December 12, 2017).

159

-------
methodology (described in more detail in Chapter 5.1.3.1), results in a projection that reflects a
neutral aim at accuracy since it accounts for expected growth in the near future by using
historical data.

We next turn to the projection of CNG/LNG derived from biogas. For 2018 - 2020, EPA
used an industry-wide approach, rather than an approach that projects volumes for individual
companies or facilities, to project the production of CNG/LNG derived from biogas. 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 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 CNG/LNG derived from biogas. 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 CNG/LNG derived from biogas in 2018 and 2019 but over-projected the
production of these fuels in 2020. The accuracy of the 2020 projection, however, was likely
influenced by the unforeseen and significant impacts of COVID-19.

As further described in Chapter 5.1.3.2, EPA is again projecting production of CNG/LNG
derived from biogas 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 2021 to estimate the production of CNG/LNG derived
from biogas in 2022.465 We have applied the growth rate to the number of available 2021 RINs
generated for CNG/LNG derived from biogas 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 CNG/LNG derived from biogas 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 CNG/LNG derived from biogas relative to the
technologies used to produce liquid cellulosic biofuel, and the relatively low production cost of
CNG/LNG derived from biogas (see Chapter 9). These factors are unlikely to change in 2022.
While we project production volumes of liquid cellulosic biofuel and CNG/LNG derived from
biogas separately, ultimately it is overall accuracy of the combined cellulosic biofuel volume
projection that is relevant to obligated parties.

465 To project the volume of CNG/LNG derived from biogas in 2022, we multiply (1) the number of 2021 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 2022..

160

-------
5.1.1.2 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,466 or are anticipated to be in a position to do so at by the end of 2022. The
RFS program provides a strong financial incentive for domestic cellulosic biofuel producers to
sell any fuel they produce for domestic consumption.467 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 2022 have already successfully completed facility
registration, and have successfully generated RINs.468 The remainder of this chapter presents a
brief description of each of the domestic companies (or group of companies for cellulosic
CNG/LNG producers and the facilities using Edeniq's technology) that EPA considered and/or
believes may produce commercial-scale volumes of RIN generating cellulosic biofuel by the end
of 2022. General information on each of these companies or group of companies considered in
our projection of the potentially available volume of cellulosic biofuel in 2022 is summarized in
Table 5.1.1.4-1.

Ace Ethanol

Ace Ethanol has installed technology developed by D3MAX designed to convert the
cellulosic portion of the distillers grains produced at their Stanely, Wisconsin facility to ethanol.
The D3MAX technology uses heat, acid, and enzymes to convert the cellulosic biomass in the
wet cake produced by a corn ethanol production facility to sugars, which are subsequently
converted to ethanol.469 Ace Ethanol began construction to integrate the D3MAX technology
into their existing corn ethanol production facility in October 2018.470 Production of ethanol

466 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).

467 According to data from EMTS, the average price for a 2020 cellulosic biofuel RINs sold in 2020 was $1.35.
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 2020 advanced biofuel RINs sold
in 2020 was $0.59 while the price for a 2020 cellulosic waiver credit is $1.80 (EPA-420-B-20-005). EPA has not yet
announced the cellulosic waiver credit price for 2021, however the average cellulosic RIN price in 2021 increased to
$2.75 per RIN.

468 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.

469 "Technology," D3MAX website. Accessed 4/27/2020. https://www.d3maxllc.com/technology

470 "Ace Ethanol Begins D3MAX Plant Construction at Stanley Wisconsin Facility," Ace Ethanol Website. October
17, 2018. https://www.aceethanol.com/news/ace-ethanol-begins-d3max-plant-construction-at-stanley-wisconsin-
facility/

161

-------
from the distillers grains began in January 2020.471 In April 2020 this facility completed the
process to register as a cellulosic biofuel producer under the RFS program.472

Compressed Natural Gas (CNG) and Liquefied Natural Gas (LNG) Producers

In July 2014 EPA approved, as part of the "Pathways II" rule,473 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 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 and to compress the gas for
injection into a pipeline.

Corn Kernel Fiber to Ethanol Technologies

EPA is aware of several companies that 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.

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.

Edeniq

Edeniq has developed a proprietary process technology (Edeniq Pathway) that allows
corn ethanol producers to co-produce conventional ethanol from corn starch and cellulosic
ethanol from corn kernel fiber at the corn ethanol producers' existing production facilities. This
is the same general production process as described in the Corn Kernel Fiber to Ethanol

471 "Ace Approved to Generate D3 RINs From Corn Kernel Fiber Ethanol," BBI International. April 1, 2020.

472 Id.

473 79 FR 42128, July 18, 2014.

162

-------
Technologies described above, however several facilities have registered as cellulosic biofuel
producers prior to EPA's May 2019 guidance474 using the Edeniq pathway. Their process
involves the addition of cellulase enzymes to convert the cellulosic material in the corn kernel
into sugars and ultimately cellulosic ethanol. Edeniq tested the Edeniq Pathway technology,
including both their proprietary Cellunator milling technology and the cellulase enzymes, at a
demonstration-scale facility in Visalia, California and announced the successful performance of
the Edeniq Pathway technology in June 2014. In the July 2014 "Pathways II" rule EPA clarified
that processes, such as the one developed by Edeniq, that convert corn kernel fiber into
transportation fuel can be approved to produce fuel that qualifies for cellulosic RINs provided all
applicable requirements are met.475 In September 2016, EPA approved Pacific Ethanol's
registration update to allow this facility to generate cellulosic biofuel RINs for fuel produced
from cellulosic biomass using Edeniq's technology.476 EPA has since registered other facilities
to generate cellulosic RINs for fuel produced from cellulosic biomass using Edeniq's
technology.477

Fulcrum BioEnergy

Fulcrum BioEnergy has developed a technology to convert separated MSW into a
synthetic crude oil using a gasification and Fischer-Tropsch process.478 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 2018.479 In July 2021
Fulcrum announced they had completed construction, and that fuel production from this facility
was expected to begin in the fourth quarter of 2021.480

Quad County Corn Processors/Syngenta

Quad County Corn Processors (QCCP) have been operating their bolt-on technology
called Cellerate™ since July 2, 2014, at their Galva, Iowa facility. The Cellerate Process
Technology enables QCCP to produce cellulosic ethanol from corn kernel fiber. The Cellerate
Process Technology is a second process, separate from the process that converts corn starch to

474 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. Environmental
Protection Agency. EPA-420-B-19-022 (May 2019).

475 79 FR 42128, July 18, 2014.

476 Schroeder, Joanna. "EPA Approves Edeniq Cellulosic Pathway." Energy.Agwired.com. September 13, 2016.
Available at: http://energv.agwired.eom/20.l.6/09/13/epa-approves-eden.ia-ceilniosic-pathwav

477 "Edeniq Pathway Technology Renamed to Intellulose"

https://www.bnsinesswire.com/news/home/20180503005977/en/Edeniq-Pathway-Technology-Renamed-to-
Intellulose

478 Unless otherwise noted, all information in this paragraph from Fulcrum BioEnergy website: Sierra Biofuels
Plant: https://ftilcRim-bioenergy.com/facilities

479 Fulcrum BioEnergy Breaks Ground on Sierra Biofuels Plant: https://fiilernm-bioenergv.com/wp-
eontent/nptoads/20.1.8/05/20.1.8-05-16-Sierra-Gronndb reaking-FINAL-l.pdf

480 Fulcrum BioEnergy Completes Construction of the Sierra Biofuels Plant: https://fulcrum-bioenergy.com/wp-
content/uploads/2021/07/2021-07-06-Sierra-Construction-Completion-Press-Release-FINAL.pdf

163

-------
ethanol. The Cellerate Process Technology pre-treats the remaining cellulosic portion of the corn
kernel to produce sugars, and then ferments these sugars to produce ethanol.

Red Rock Biofuels

Red Rock Biofuels intends to use an integration of existing technologies to convert forest
residues to diesel and jet fuel using a Fischer-Tropsch process.481 Construction of their first
commercial scale production facility began on July 18, 2018.482 This facility, located in
Lakeview, Oregon, is designed to produce approximately 16 million gallons of cellulosic biofuel
per year.483 In the proposed rule, EPA considered Red Rock Biofuels as a potential producer in
2022. However, we are no longer doing so in this final rule, as in January 2022 Red Rock
Biofuels announced that further construction of the facility was on hold until 2023.484

5.1.1.3 Potential Foreign Sources of Cellulosic Biofuel

EPA's projection of cellulosic biofuel production through 2022 also includes cellulosic
biofuel that is projected to be imported into the U.S. This means the potential imports 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 2022. These include facilities owned and operated by Enerkem,
GranBio, and Raizen. 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
2022. This is due to the strong demand for cellulosic biofuel in local markets 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 2022
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 two foreign facilities (GranBio's and Raizen's Brazilian
facilities) 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

481 Red Rock Biofuels website: https://www.redrockbio.com/. Accessed April 30, 2019.

482 Red Rock Biofuels: Lakeview Site: https://www.redrockbio.com/lakeview-site.html. Accessed April 30, 2019.

483 Red Rock Biofuels: Lakeview Site: https://www.redrockbio.com/lakeview-site.html. Accessed April 30, 2019.

484 Jester, Danielle and Winter, Kevin. Red Rock biofuel company says Lake County work on hold until 2023. Lake
County Examiner. January 12, 2022. Available Online:

https://www.heraldandnews.com/news/local_news/business/red-rock-biofuel-company-says-lake-county-work-on-
hold-until-2023/article_e4bb4f94-flel-5c02-b9fe-fae6e8194bfb.html

164

-------
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 included these
facilities as potential sources of cellulosic biofuel we are not projecting any imports of cellulosic
biofuel through 2022. All of the potential cellulosic biofuel producers for 2022 are listed in Table
5.1.1.4-1.

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.485 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.486 Enerkem's facility in Edmonton is designed to produce up to 13 million gallons of
cellulosic ethanol per year.487 This facility began production of methanol in 2015, with
production switching from methanol to ethanol in 2017.

GranBio

GranBio uses a hydro-thermal pretreatment and enzymatic hydrolysis process to convert
cellulosic biomass into ethanol.488 Construction of their first cellulosic ethanol production
facility was announced in mid-2012, and financing was completed in May 20 1 3.489 In September
2014, GranBio announced that its first cellulosic ethanol facility became operational.490 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.491 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.

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

485 "Technology," Enerkem Website. Accessed 5/15/2018. https://enerkem.com/about-us/technology/

486 Ibid

487 "Enerkem Alberta Biofuels," Enerkem Website. Accessed 5/15/2018. https://enerkem.com/facilities/enerkem-
alberta-biofuels/

488 "Who We Are," GranBio Website. Accessed May 15, 2018. http://www.granbio.com.br/en/conteudos/who-we-
are/

489 Schill, Susanne R. "Financing Complete on Brazil's first commercial 2G Ethanol Plant," Ethanol Producer
Magazine. May 17, 2013.

490 "Who We Are," GranBio Website. Accessed May 15, 2018. http://www.granbio.com.br/en/conteudos/who-we-
are/

491 Ibid

165

-------
of producing approximately 10.5 million gallons of ethanol a year from biomass residues from
first generation sugarcane ethanol production.492 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 2015.493 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 32 million liters of cellulosic ethanol to the
U.S. through the end of 2019.

5.1.1.4 Summary of Potential Sources of Cellulosic Biofuel in 2022

General information on each of the cellulosic biofuel producers (or group of producers,
for producers of CNG/LNG derived from biogas and producers of liquid cellulosic biofuel using
Edeniq's technology) that factored into our projection of cellulosic biofuel production through
2022 is shown in Table 5.1.1.4-1. 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 2022. Since we are finalizing regulations that allow for the production of qualifying
renewable fuels from biointermediates in this rule, we have considered facilities intending to
produce cellulosic biofuel from biointermediates in this table, however we do not project any
cellulosic biofuel production from these facilities in 2022. Note that while we believe all these
facilities have the potential to produce or import cellulosic biofuel by the end of 2022, our
projections of cellulosic biofuel production do not include volumes from all of the listed
facilities, as we believe the most likely volume of cellulosic biofuel produced or imported from
some of these facilities in 2022 is zero.

492 "A track record of innovation." Iogen website. Accessed May 1, 2018.

493 Ibid

166

-------
Table 5.1.1.4-

Potential Producers of Cellulosic Biofuel for U.S. Consumption in 2022

Company
Name

Location

Feedstock

Fuel

Facility
Capacity (Million
Gallons per

Year)494

Construction
Start Date

First
Production495

Ace Ethanol

Stanley, WI

Corn Kernel Fiber

Ethanol

October 2018

January 2020

CNG/LNG
Producers

Various

Biogas

CNG/LNG

Various

Edeniq

Various

Corn Kernel Fiber

Ethanol

Various

October 2016

Enerkem

Edmonton, AL,
Canada

Separated MSW

Ethanol

10496

2012

September
2017497

Fulcrum/
Marathon

Storey County, NV

Separated MSW

Diesel, Jet
Fuel

May 2018

2022

GranBio

Sao Miguel dos
Campos, Brazil

Sugarcane bagasse

Ethanol

Mid 2012

September 2014

QCCP/
Syngenta

Galva, IA

Corn Kernel Fiber

Ethanol

Late 2013

October 2014

Raizen

Piracicaba City,
Brazil

Sugarcane bagasse

Ethanol

January 2014

July 2015

494 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.

495 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.

496 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).

497 This date reflects the first production of ethanol from this facility. The facility began production of methanol in 2015.

167

-------
5.1.2 Projected Cellulosic Biofuel Production in 2022

5.1.2.1 Liquid Cellulosic Biofuel

For our 2022 liquid cellulosic biofuel projections, we use the same general approach as
we have in projecting these volumes in previous years. We begin by first categorizing potential
liquid cellulosic biofuel producers in 2022 according to whether or not they have achieved
consistent commercial scale production of cellulosic biofuel to date. We refer to these facilities
as consistent producers and new producers, respectively. Next, we define a range of likely
production volumes for 2022 for each group of companies. Finally, we use a percentile value to
project from the established range a single projected production volume for each group of
companies in 2022. As in the 2018 - 2020 final rules, we then calculated percentile values for
each group of companies based on the past performance of each group relative to our projected
production ranges.

We first separate the list of potential producers (listed in Table 5.1.1.4-1) into two groups
according to whether the facilities have achieved consistent commercial-scale production and
cellulosic biofuel RIN generation. Unlike in previous years, we have not listed the names of the
companies that fall into each group due to the small number of companies in some of the
categories and the fact that some of the data used to calculate the ends of the range is considered
confidential business information by these companies.498 We next defined a range of likely
production volumes for each group of potential producers. The low end of the range for each
group of producers reflects actual RIN generation data over the last 12 months for which data
were available at the time our technical assessment was completed (January 2021 - December
2021).499 For potential producers that have not yet generated any cellulosic RINs, the low end of
the range is zero. For the high end of the range, we considered a variety of factors, including the
expected start-up date and ramp-up period, facility capacity, and the number of RINs the
producer expects to generate in 2022.500 The projected range for each group of companies is
shown in Table 5.1.2.1-1.501

498 More information on the companies included in each group is contained in the memos "February 2022 Liquid
Cellulosic Biofuel Projections for 2022 CBI".

499 We recognize that in some years cellulosic biofuel production from facilities that have achieved consistent
commercial scale production may be lower than the volume achieved in the previous 12 months. This has happened
several times since 2016. In these cases the methodology would suggest that using a negative percentile value
(indicating production in the coming year that is lower than the volume produced in the previous 12 months). By
considering the use of negative percentile values for facilities that have achieved consistent commercial scale
production we believe this methodology adequately accounts for this possibility.

500 As in our 2015-2020 projections, EPA calculated a high end of the range for each facility (or group of facilities)
based on the expected start-up date and a six-month straight-line ramp-up period. The high end of the range for each
facility (or group of facilities) is equal to the value calculated by EPA using this methodology, or the number of
RINs the producer expects to generate in 2022, whichever is lower.

501 More information on the data and methods EPA used to calculate each of the ranges in these tables in contained
in "February 2022 Liquid Cellulosic Biofuel Projections for 2022 CBI" memorandum from Dallas Burkholder to
EPA Docket EPA-HQ-OAR-2021-0324. We have not shown the projected ranges for each individual company. This
is because the high end of the range for some of these companies are based on the company's production
projections, which they consider confidential business information (CBI). Additionally, the low end of the range for
facilities that have achieved consistent commercial scale production is based on actual RIN generation data in the
most recent 12 months, which is also claimed as CBI.

168

-------
Table 5.1.2.1-1: 2022 Liquid Cellulosic Biofuel Projected Production Ranges (million

Category

Low End of
the Range

High End of
the Range

New Producers

Consistent Producers

a Rounded to the nearest million gallons.

After defining likely production ranges for each group of companies, we next determined
the percentile values to use in projecting a production volume for each group of companies. We
calculate the percentile values by comparing actual production data from 2016 through 2020
with the production ranges projected in the annual rules for those years. We chose the 2016-20
time period because the first full year in which EPA used the current methodology for
developing the range of potential production volumes for each company was 2016, while 2020 is
the most recent year for which we have both cellulosic RIN generation data and a prospective
projection of liquid cellulosic biofuel production.502

As in previous years, to calculate the percentiles within the projected ranges used to
project liquid cellulosic biofuel production for 2022 we used the average percentile values for
each group of companies from 2016 - 2020. We considered weighting recent years more heavily
than previous years, but we have not done so. The disruptions in the cellulosic biofuel industry
caused by the COVID pandemic in 2020 suggest that we should not more heavily weigh 2020,
even though this is the most recent year for which we have data. Alternatively, we considered
only considering the percentile values from 2016-2019 and excluding 2020 due to the unforeseen
impacts of the COVID pandemic. Excluding the percentiles from 2020 results in slightly higher
percentile values than those we calculate when including the 2020 data (the 9th percentile for new
producers and the 6th percentile for consistent producers). However, using the slightly higher
percentile values for consistent producers that results from excluding consideration of 2020 has a
very small impact (1 million gallons) on our projection of liquid cellulosic biofuel production
and no impact on the cellulosic biofuel volume for 2022 (which is rounded to the nearest 10
million gallons).

For each group of companies and for each year from 2016-2020, Table 5.1.2.1-2 shows
the projected ranges for liquid cellulosic biofuel production (from relevant annual rules), actual
production, the percentile values that would have resulted in a projection equal to the actual
production volume, and the weighting factor used in each year.

502 While we did project liquid cellulosic biofuel production for 2021 in the proposed rule, we did not finalize our
projection of cellulosic biofuel production for 2021 with the benefit of updated data and public comment. We have
therefore not included 2021 in our consideration of the percentile values used to project cellulosic biofuel production
in 2022.

169

-------
Table 5.1.2.1-2: Projected and Actual Liquid Cellulosic Biofuel Production in 2016-2020

(million gallons)

Actual

Low End of

High End of

Production

Actual

the Range

503

Percentile

New Producers504

2016

1.06

1st

2017

8.79

27th

2018

2.87

6th

2019

0.00

0th

2020

1.53

5th

Average21

N/A

8th

Consistent Producers505

2016

3.28

43rd

2017

3.5

3.02

-14th

2018

7.74

4th

2019

11.13

-10th

2020

0.52

-36th

Average21

N/A

-3rd

a We have not averaged the low and high ends of the ranges, or actual production, as we believe it is more
appropriate to consider the averages of the actual percentiles from 2016 - 2020 rather than calculating a percentile
value for 2016 - 2020 in aggregate.

Based upon this analysis, EPA has projected cellulosic biofuel production from new
producers at the 8th percentile of the calculated range and from consistent producers at the -3rd
percentile.506 These percentiles are calculated by averaging the percentiles that would have
produced cellulosic biofuel projections equal to the volumes produced by each group of
companies in 2016 - 2020. Prior to 2016, EPA used different methodologies to project available
volumes of cellulosic biofuel and thus believes it inappropriate to calculate percentile values
based on projections from those years.507

503 Actual production is calculated by subtracting RINs retired for any reason other than compliance with the RFS
standards from the total number of cellulosic RINs generated.

504 Companies characterized as new producers in the 2014-2016, 2017, 2018, 2019, and 2020 final rules were as
follows: Abengoa (2016), CoolPlanet (2016), DuPont (2016, 2017), Edeniq (2016, 2017), Enerkem (2018, 2019,
2020), EnsynPort Cartier (2018, 2019, 2020), GranBio (2016, 2017), IneosBio (2016), and Poet (2016, 2017) Red
Rock Biofuels (2020).

505 Companies characterized as consistent producers in the 2014-2016, 2017, 2018, and 2019 final rules were as
follows: Edeniq Active Facilities (2018, 2019, 2020), Ensyn Renfrew (2016 -2020), GranBio (2018-2020), Poet
(2018, 2019), Quad County Corn Processors/Syngenta (2016 - 2020), and Raizen (2019-2020).

506 The negative percentile we are using to project cellulosic biofuel production from consistent producers in 2022
means that we are projecting less cellulosic biofuel will be produced from these facilities than they produced over
the last 12 months for which data is available. We observed a similar pattern in 2017 and 2020, where liquid
cellulosic biofuel production from consistent producers fell from the prior year. This is generally because producing
liquid cellulosic biofuel at a commercial scale remains challenging, and many producers have gone out of business
not long after they began production.

507 EPA used a similar projection methodology for 2015 as in 2016-2018, however we only projected cellulosic
biofuel production volume for the final 3 months of the year, as actual production data were available for the first 9
months. We do not believe it is appropriate to consider data from a year for which 9 months of the data were known
at the time the projection was made in determining the percentile values used to project volume over a full year.

170

-------
We then used these percentile values, together with the ranges determined for each group
of companies discussed above, to project a volume for each group of companies in 2022. These
calculations are summarized in Table 5.1.2.1-3.

Table 5.1.2.1-3: Projected Volume of Liquid Cellulosic Biofuel in 2022 (million ethanol-
equivalent gallons)

Low End of

High End of

Projected

the Range3

Percentile

Volume3

2022

Liquid Cellulosic Biofuel Producers;

Producers without Consistent

8th

Commercial Scale Production

Liquid Cellulosic Biofuel Producers;

Producers with Consistent

-3rd

Commercial Scale Production

Total

N/A

a Volumes rounded to the nearest million gallons.

b Note that though these volumes round to zero million gallons we do project a small volume of liquid cellulosic
biofuel will be produced in 2022.

5.1.2.2 CNG/LNG Derived from Biogas

For 2022, EPA is using the same industry wide projection approach as used for 2018 -

2020 based on a year-over-year growth rate to project production of CNG/LNG derived from
biogas used as transportation fuel.508 EPA calculated the year-over-year growth rate in
CNG/LNG derived from biogas by comparing RIN generation from January 2021 to December

2021 (the most recent 12 months for which data are available) to RIN generation in the 12
months that immediately precede this time period (January 2020 to December 2020). The growth
rate calculated using this data is 12.6 percent. These RIN generation volumes are shown in Table
5.1.2.2-1.

Table 5.1.2.2-1: Generation of Cellulosic Biofuel RINs for CNG/LNG Derived from Biogas

(ethanol-equivalent gallons

RIN Generation
(January 2020 -
December 2020)

RIN Generation (January
2021 - December 2021)

Y ear-Over-Y ear
Increase

503,821,039

567,082,572

12.6%

EPA then applied this 12.6% year-over-year growth rate to the total number of 2021
cellulosic RINs generated and available for compliance for CNG/LNG. That is, in this rule, as in
the 2018 - 2020 final rules, we are multiplying the calculated year-over-year rate of growth by

508 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.

171

-------
the volume of CNG/LNG actually supplied in the most recent year for which data is available (in
this case 2021), taking into account actual RIN generation as well as RINs retired for reasons
other than compliance with the annual volume obligations. This provides a projection of the
production of CNG/LNG derived from biogas in 2022. This results in a projection of 632 million
ethanol-equivalent gallons of CNG/LNG derived from biogas in 2022.

Table 5.1.2.2-2: 2022 Projection of CNG/LNG Derived from Biogas (ethanol-equivalent

D3 RINs generated for CNG/LGN derived
from biogas in 2021

567,082,572

RINs retired for reasons other than
compliance with annual obligations

5,225,692

Net RINs generated

561,856,880

Growth rate

12.56%

Projected production of CNG/LNG derived
from biogas in 2022

632,405,596

We then compared these projected volumes with the total volume of CNG/LNG expected
to be used as transportation fuel in 2022. 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 January 2022 STEO forecasts that 0.16 billion cubic feet per day of natural gas
will be used as transportation fuel in 2022 (approximately 770 million ethanol-equivalent gallons
respectively, assuming 1015 BTU per cubic foot of natural gas and 77,000 BTU per gallon of
ethanol).509 EIA's 2021 AEO projects that 0.1 trillion cubic feet of natural gas will be used in the
transportation sector in 2022 (approximately 1.3 billion ethanol-equivalent gallons).510 A paper
prepared by Bates White for the Coalition for Renewable Natural Gas511 Fuels Institute North
America projects that 2.0 billion ethanol-equivalent gallons of natural gas will be used as
transportation fuel in 2022. A more recent independent assessment prepared by Bates White for
the Coalition for Renewable Natural Gas projected approximately 1.9 billion ethanol-equivalent
gallons of natural gas will be used as transportation fuel in 2022.512 All of these estimates are
greater than the volume of qualifying CNG/LNG derived from biogas projected to be used in
2022. In addition, as discussed in the Bates White report, the forecast in the STEO is based on
industry surveys, and is highly unlikely to reflect all of the CNG/LNG used as transportation
fuel.513 While there is considerable uncertainty about the quantity of CNG/LNG that will be used
in transportation fuel in 2022, we believe that the higher estimates are likely more representative
of the industry capacity to use CNG/LNG as transportation fuel in 2022. These projections of
natural gas used as transportation fuel are significantly greater than the volume of qualifying
CNG/LNG derived from biogas projected to be used in 2022. Thus, the volume of CNG/LNG

509 These values are from the forecasts for Vehicle Use in Table 5a: U.S. Natural Gas Supply, Consumption, and
Inventories in the January 2022 STEO.

510 These values are from the projections for Motor Vehicles, Trains, and Ships in Table 13: Natural Gas Supply,
Distribution, and Prices in the 2021 AEO.

511 Renewable Natural Gas Supply and Demand for Transportation. Bates White Economic Consulting. April 5,
2019.

512 Renewable Natural Gas: Transportation Demand. Bates White Economic Consulting. February 2, 2022.

513 Renewable Natural Gas Supply and Demand for Transportation. Bates White Economic Consulting. April 5,
2019.

172

-------
used as transportation fuel would not appear to constrain the number of RINs generated for this
fuel in these years.

We believe that projecting the production of CNG/LNG derived from biogas in this
manner appropriately takes into consideration the actual recent rate of growth of this industry,
and that this growth rate accounts for both the potential for future growth and the challenges
associated with increasing RIN generation from these fuels in future years. This methodology
may not be appropriate to use as the projected volume of CNG/LNG derived from biogas
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 for 2022 is below the total volume of CNG/LNG that is currently
used as transportation fuel.

5.1.3 Additional Potential Sources of Cellulosic Biofuel

As explained in Preamble Section III, EPA has consistently interpreted the statutory
phrase "projected volume available" to refer to the volume of qualifying cellulosic biofuel
projected to be produced in the U.S. or imported and available for use in the U.S. in that year.
This can also be expressed as the quantity of likely cellulosic RINs generated in 2022 (i.e., from
cellulosic biofuel production that meets statutory and regulatory requirements) that are available
for obligated parties to use for compliance. This is the approach we have taken in projecting
biogas and liquid cellulosic biofuel in this rule and in prior annual rules.514

Although CAA section 21 l(o)(7)(D)(i) does not expressly connect the "projected volume
available" with RIN generation, this interpretation is consistent with the statutory context,
structure, and purpose.515 Notably, in API v. EPA, the D.C. Circuit explained that CAA section
21 l(o)(7)(D)(i) "serves as a non-discretionary safety valve when the refiners and importers of
transportation fuel subject to [§ 21 l(o)'s] mandate would otherwise be put in an impossible
position, or at least a highly punitive one—that is, forced to purchase volumes of cellulosic
biofuel greater than total production, or pay fines for failing to do so."516 The court further held
that EPA's methodology for making its cellulosic biofuel projection must take a "neutral aim at

514 For example, consider our projection in the 2020 final rule. There as here, we considered RINs generated in 2020
and available for obligated parties to use for compliance in determining the projected volume available. See 85 FR
7023/1 (projecting "the production of cellulosic biofuel RINs for CNG/LNG derived from biogas"), 7024/2
(considering "which facilities are most likely to produce liquid cellulosic biofuel and generate cellulosic biofuel
RINs in 2020"); 2020 RTC 32, 31 (declining to include renewable electricity in projected volume available because
technical and regulatory obstacles precluded timely registration of facilities "to generate RINs for the production of
renewable electricity"; thus, such facilities were not "reasonably likely to produce qualifying cellulosic biofuel in
2020"). Similarly, our projection included an adjustment for RINs not available for compliance. See 85 FR 7028/1
n.53 (in projecting liquid cellulosic biofuel, considering historical actual production, which excludes RINs retired
for reasons other than compliance), 7029/1 (in projecting biogas, considering the number of "cellulosic RINs
generated and available for compliance"). RINs may be retired for various non-compliance purposes, e.g., they are
erroneously or unlawfully generated, associated with spilled biofuel, etc. In addition, RINs may be retired to satisfy
the exporter renewable volume obligation and thus not available for compliance by obligated parties, see 40 CFR
80.1430; however, the vast majority of domestically produced cellulosic biofuel is used domestically.

515 See Chevron USA, Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984).

516 706 F.3d 474, 479 (D.C. Cir. 2013).

173

-------
accuracy."517 Thus, the projected volume of cellulosic biofuel available in a given year should
reflect the volume that can actually be used to satisfy obligated parties' renewable volume
obligations. Obligated parties demonstrate compliance via the procurement and retirement of
RINs; a fuel that does not generate RINs cannot be used towards compliance and cannot be said
to be "available."

In other words, cellulosic RINs are generated for fuels that meet all the applicable
statutory and regulatory requirements to qualify as cellulosic biofuel. A fuel that cannot be
demonstrated to meet those requirements, whether due to technical or to regulatory obstacles, is
not a qualifying cellulosic biofuel, cannot generate RINs, and is therefore not available. In the
case of ethanol produced from corn kernel fiber co-processed with corn starch and renewable
electricity, technical, regulatory, and/or implementation issues have revolved around
demonstrating that the fuel claimed as cellulosic biofuel satisfies the statutory and regulatory
requirements to qualify as such. For ethanol produced from corn kernel fiber, EPA and facilities
are still working to identify and implement technical approaches for demonstrating that the fuel
is produced from renewable biomass that is derived from cellulose, hemicellulose or lignin.518
For renewable electricity, outstanding technical and regulatory issues include ensuring that the
volumes of the fuel claimed as being used to replace or reduce the quantity of fossil fuel present
in transportation fuel are in fact used for that purpose without double counting volumes
claimable by others.519

For renewable electricity, EPA believes that the outstanding issues surrounding RIN
generation cannot be resolved in a sufficiently timely manner such that significant amounts of
fuel from these sources will be available as cellulosic biofuel in 2022. We have previously
stated, and continue to believe, that these issues are best resolved through a notice and comment
rulemaking process. We are therefore not including fuel from this source in our projections of the
available volume of cellulosic biofuel for 2022. For cellulosic ethanol produced from corn kernel
fiber, technical work towards implementing approaches for measuring cellulosic content
remains, and there is considerable uncertainty about whether resolution of extant issues will
allow for significant additional volume of cellulosic biofuel to be available in 2022. We therefore
have not included volumes from additional facilities that intend to produce cellulosic ethanol
from corn kernel fiber co-processed with corn starch in our projections of cellulosic biofuel
production in 2022.

517 Id. at 476.

518 "[C]ellulosic biofuel means renewable fuel derived from any cellulose, hemicellulose, or lignin that is derived
from renewable biomass and that has lifecycle greenhouse gas emissions, as determined by the Administrator, that
are at least 60% less than the baseline lifecycle greenhouse gas emissions." CAA section 21 l(o)(l)(E). See also 40
CFR 80.1401, 80.1450.

519 81 FR 80828, 80890-91 ("Vehicle charging data demonstrate the use of electricity as transportation fuel, one of
the two main requirements for RIN generation (production from renewable biomass being the other). However, there
are several sources of charging data that could be provided to verify the use of electricity as transportation fuel....
Any of these sources of data could conceivably be used as the basis for generating RINs for renewable electricity.
Although multiple types of data can be used to demonstrate the use of electricity as transportation fuel, allowing
them to be used simultaneously would almost certainly result in the generation of RINs by multiple parties for the
same charging event (i.e., double counting).").

174

-------
5.1.4 Projected Rate of Cellulosic Biofuel Production for 2022

After projecting production of cellulosic biofuel from liquid cellulosic biofuel production
facilities and producers of CNG/LNG derived from biogas, EPA combined these projections to
project total cellulosic biofuel production for 2022. These projections are shown in Table 5.1.4-
1. Using the methodologies described in this chapter, we project that 0.63 billion ethanol-
equivalent gallons of qualifying cellulosic biofuel will be produced in 2022 We believe that
projecting overall production in 2022 in the manner described above results in a neutral estimate
(neither biased to produce a projection that is too high nor too low) of likely cellulosic biofuel
production in 2022.

Table 5.1.4-1: Projected Volume of Cellulosic Biofuel in 2022

Projected Volume in 2022

Projected
Volume3

Liquid Cellulosic Biofuel Producers; Producers without Consistent Commercial
Scale Production (million gallons)

Liquid Cellulosic Biofuel Producers; Producers with Consistent Commercial
Scale Production (million gallons)

CNG/LNG Derived from Biogas (million gallons)

632

Total (billion gallons)

0.63

a Rounded to the nearest million gallons.

As discussed in Chapter 5.1.3, this projection does not include any volume of cellulosic
ethanol produced from corn kernel fiber or renewable electricity used as transportation fuel. We
believe the projection represents a "neutral aim at accuracy" consistent with the court's direction
in API v. EPA.

While there is currently uncertainty related to the demand for transportation fuel in 2022,
we do not expect these overall transportation fuel demand to significantly impact the supply of
cellulosic biofuel in 2022. The vast majority of fuel we project will be produced is CNG/LNG
derived from biogas, much of which is sourced from landfills. We do not expect that lower
transportation fuel demand will impact the production of this fuel, nor do we expect that the use
of CNG/LNG as transportation fuel will decrease to a level below the volume we are projecting
will be available in 2022.

5.2 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 and 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,

175

-------
however as the vast majority of BBD is biodiesel and renewable diesel we have focused on these
fuels in this chapter. This chapter presents available data on biodiesel and renewable diesel
production, import, and use in 2020 and 2021, describes our assessment of the rate of production
and use of qualifying biomass-based diesel biofuel in 2022, and describes some of the
uncertainties associated with those volumes.

5.2.1 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 availability of
advanced biodiesel and renewable diesel feedstocks,520 the extension of the biodiesel tax credit,
tariffs on imported biodiesel, 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 5.2.1-1.

520 Throughout this chapter 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 5.1. 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).

176

-------
Table 5.2.1-1: BBD (D4) Production, Imports, and Exports from 2014 to 2021521 (million
gallons)"

2014b

2015b

2016

2017

2018

2019

2020

2021

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)

1,699
(-103)

Imported Biodiesel
(Annual Change)

130
(-23)

261
(+131)

562
(+301)

462
(-100)

175
(-287)

185
(+10)

209
(+24)

208
(-1)

Exported Biodiesel
(Annual Change)

72
(-5)

73
(+1)

89
(+16)

129
(+40)

74
(-55)

76
(+2)

88
(+12)

90
(+2)

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)

1,817
(-107)

Domestic
Renewable Diesel
(Annual Change)

149
(+79)

169
(+20)

231
(+62)

252
(+21)

282
(+30)

454
(+172)

472
(+18)

780
(+308)

Imported
Renewable Diesel
(Annual Change)

130
(-15)

120
(-10)

165
(+45)

191
(+26)

176
(-15)

267
(+91)

280
(+13)

362
(+82)

Exported
Renewable Diesel
(Annual Change)

15
(+10)

21
(+6)

40
(+19)

37
(-3)

80
(+43)

145
(+65)

223
(+78)

241
(+18)

Total Renewable

Diesel
(Annual Change)c

264
(+154)

268
(+4)

356
(+88)

406
(+50)

378
(-28)

576
(+198)

529
(-47)

900
(+371)

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)

2,721
(+264)

a 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 in the table above to calculate the volume in each year.
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 (<10 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
2022, 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,

521 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
5.4. The reason we are presenting this information somewhat differently in this rulemaking is because the statute
specifically directs EPA to analyze "the expected annual rate of future commercial production of renewable fuels,
including ... biomass-based diesel."

177

-------
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. By 2021 production and imports of renewable diesel had increased not only in
absolute terms (from 264 million gallons in 2014 to 927 million gallons in 2021), but also as a
percentage of the BBD pool. In 2021 approximately 34% of all BBD was renewable diesel,
while the remaining 66% was biodiesel. As discussed further in the following chapters, 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, 2020,
and 2021, 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, and 2021) 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, 63 million gallons,522 and 291 million gallons
respectively). There is a substantial probability that with the tax credit in place prospectively for
2022, the increase in the supply of biodiesel and renewable diesel will be greater than the
average historical increase observed. 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.523 According to data from EIA,524 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

522 This is the volume increase in 2020, which was impacted by the COVID pandemic.

523 "Biodiesel from Argentina and Indonesia Injures U.S. Industry, says USITC," Available at:

https://www.nsitc.gov/press room/news release/2017/er.l.20511876.htm

524 See "EIA Biomass-Based Diesel Import Data" available in docket EPA-HQ-OAR-2021-0324.

178

-------
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.

5.2.2 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 5.2.2-1. 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
approximately 2.4 billion gallons in 2021.525 While production of biodiesel has increased during
this time period, significant excess production capacity remains, with facility utilization
remaining below 75% through 2021. 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 currently limiting domestic biodiesel
production.

Figure 5.2.2-1: U.S. Biodiesel Production Capacity, Production, and Capacity Utilization

3.00 80%

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Production Capacity Production Volume ^^—Utilization

Unlike domestic biodiesel production capacity, domestic renewable diesel production
capacity has increased significantly in recent years, from approximately 280 million gallons in

525 Biodiesel production capacity from EIA Monthly Biodiesel Production reports and the EIA U.S. Biodiesel Plant
Production Capacity report.

179

-------
2017 to approximately 910 million gallons in 2021 (Figure 5.2.2-2).52b Domestic 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 84%
from 2017-2021. 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. These high utilization rates
and rapid capacity expansion in recent years suggest that increased renewable diesel production
will likely be dependent on the construction of new renewable diesel production capacity, and
that renewable diesel production is likely to increase along with increases in renewable diesel
production capacity.

Figure 5.2.2-2: U.S. Renewable Diesel Production Capacity, Production, and Capacity
Utilization

1,000 1 00%

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Production Capacity Production Volume ^^—Utilization

Renewable diesel production capacity and actual production values are from EMTS data. 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
2022. 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 2022, as well as existing facilities expected to complete expansions by 2022,
based on publicly available data is shown in Table 5.2.2-1. In addition to the facilities listed in
Table 5.2.2-1, a significant number of renewable diesel facilities have been announced with
intended start-up dates after 2022. These facilities are not discussed here since they are not likely

526 Renewable diesel capacity is based on RFS facility registration data.

180

-------
to impact renewable diesel production in 2022, however they do indicate that there is high
interest in expanding renewable diesel production in future years.

Table 5.2.2-1: New Renewab

e Diesel Production Capacity in the U.S. Through 2022

Facility Name

Location

Capacity (MGY)

Start Date (Actual
or Expected)

Bakersfield Renewable
Fuels

Bakersfield, CA

230

Q1 2022527

CVR Energy Inc. -
Wynnewood

Wynnewood, OK

100

April 2022528

Diamond Green Diesel

Norco, LA

Expansion (275 to
675)

Q4 2021529

HollyFrontier - Artesia

Artesia, NM

110

Q2 2022530

HollyFrontier - Cheyenne

Cheyenne, WY

End 2021531

Viking Energy

Reno, NV

2022532

5.2.3 Availability of Biomass-Based Diesel Feedstocks

Another key factor in considering the rate of production of BBD in 2022 is the
availability of qualifying feedstocks. To assess the availability of feedstocks for producing BBD
through 2022 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 these feedstocks in future years.
A summary of the feedstocks used to produce BBD from 2012 through 2021 is shown in Figure
5.2.3-1.

527 Global Clean Energy Holdings, Inc. Letter to Shareholders from Richard Palmer, President and CEO (October
11,2021).

528 Voegele, Erin. CVR to move forward with Wynnewood conversion in early 2022. Biomass Magazine. November
3,2021.

529 Kotrba, Ron. Diamond Green Diesel Expansion Complete, In Start-Up Phase. Biobased Diesel Daily. October
21,2021.

530 McGurty, Janet. HollyFrontier readies renewable diesel project as Q3 refining margins jump. S&P Global
Platts. November 3, 2021.

531 Ibid

532 Viking Energy Signs Agreement to Acquire Renewable Diesel Facility. December 1, 2021.

181

-------
Figure 5.2.3-1: Feedstocks Used To Produce BBD in the U.S. (2014-2021)533

3,500
3,000
2,500

^ 2,000

c
o

= 1,500
1,000
500

¦ FOG ¦ Soybean Oil ¦ Corn Oil ¦ Canola Oil

BBD production from fats, oils, and greases (FOG) in the U.S. has generally increased
from 2014 through 2021 at an average annual rate of approximately 35 million RINs per year.
These feedstocks are generally by-products of other industries, and assessments submitted by
commenters generally agree that the domestic supply of these feedstocks will increase only
slightly in future years.534 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 2022.

Production of BBD from distillers corn oil has also generally increased through 2021.
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 400 - 500 million RINs per year since 2017. Total production of
distillers corn oil in the U.S. in 2020 was approximately 1.7 million tons,535 or enough corn oil to
produce about 500 million gallons (approximately 750 million RINs) of BBD. This suggests that
distillers corn oil could be used to produce as much as 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.536 It is also possible that domestic production of

533 Based on EMTS data

534 See comments from the Advanced Biofuels Association (EPA-HQ-OAR-2021-0324-0476) and the Clean Fuels
Alliance America (EPA-HQ-OAR-221-0324-0458).

535 USDA Grain Crushings and Co-Products Production 2020 Summary. March 2021. Available at:
https://www.nass.usda.gov/Publications/Todavs Reports/reports/cagcan21 .pdf

536 For a discussion of backfilling when oil is removed from dried distillers grains, see 83 FR 37735 (August 2,
2018).

2014 2015 2016 2017 201S 2019 2020 2021

182

-------
distillers corn oil could increase in 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 2022.

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 over 300 million
RINs in 2016 to a low of approximately 170 million RINs in 2015. Production of BBD from
canola oil averaged approximately 240 million RINs each year from 2018 to 2021. Total
production of canola oil reached a high of approximately 1.9 billion pounds in the 2020/2021
agricultural marketing year, or enough canola oil to produce approximately 250 million gallons
of BBD.537 Significant increases in canola oil production in the U.S. through 2022 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, were
imported in 2020/2021. It is unclear how much of this imported canola oil would be able to
qualify as renewable biomass under the statutory definition, and thus available to be used to
produce qualifying BBD under the RFS program.538 Further, increasing biofuel demand in
Canada is likely to impact the quantity of canola oil available to U.S. biofuel producers. Any
increased use of canola oil for biofuel production would require shifting it from its other existing
uses which would then have to be backfilled with other new sources, the lowest cost of which is
currently palm oil, potentially impacting the GHG benefits.

The largest source of BBD production in the U.S. historically has been soybean oil. Due
to the potential limitations on the increased availability of the other BBD feedstocks discussed
above, we also expect that soybean oil is the most likely source of feedstock used to produce
higher volumes of BBD in future years. Use of soybean oil to produce biodiesel increased from
approximately 5.1 billion pounds in the 2013/2014 agricultural marketing year to approximately
8.3 billion pounds in the 2020/2021 agricultural marketing year.539 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 35% in 2020/2021. As a point of reference, if
all the soybean oil produced in the U.S. in 2020/2021 (25.6 billion pounds) were used to produce
BBD, this quantity of feedstock could be used to produce approximately 3.3 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

537 U.S. Canola oil production data sourced from USDA's Oil Crops Yearbook (https://www.ers.usda.gov/data-
products/oil-crops-yearbook/).

538 CAA section 21 l(o)(l)(I).

539 U.S. Soybean oil production and use data sourced from USDA's Oil Crops Yearbook

(https://www.ers.iisda.gov/data-prodiicts/oil-crops-yearbook/'). The agricultural marketing year for soybeans runs
from September to August.

183

-------
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.540 Higher soybean crushing rates produce greater quantities of soybean oil from the same
soybean crop.

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 in
December 2021.541 USDA estimates of soybean crush capacity, after accounting for necessary
down time at crush facilities, range from 2.45 - 2.6 billion bushels per year.542 The high end of
this range represents an increase of 0.4 billion bushels of soybean crush relative to the estimated
quantity of soybeans crushed in the 2020/2021 agricultural marketing year.543 Increasing
soybean crushing by 0.4 billion bushels per year would increase soybean oil production by
approximately 4.6 billion pounds, assuming an oil yield of 11.6 pounds per bushel, consistent
with the soybean oil yield per bushel of soybeans crushed in recent years.544 This increase in
soybean oil production could be used to produce approximately 600 million gallons of biodiesel
and renewable diesel if this entire quantity of soybean oil was used for this purpose. In addition,
a number of companies have recently announced plans to build new soybean crushing facilities,
or expand existing facilities.545 In comments on the proposed rule the American Soybean
Association noted at least 13 announcements for the expansion of soybean crush facilities or new
facilities.546 They estimated that these projects could increase soybean crushing by 15%.547 Any
soybean oil production from these facilities if they come online in 2022 would further increase
the quantity of soybean oil available for biofuel production or other uses. Increasing the domestic
crush of soybeans would likely result in a decrease in the quantity of soybeans exported to other
countries and could also result in increased soybean cultivation in the U.S.

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 2020/2021 agricultural
marketing year approximately 10% of the soybean oil produced in the U.S. was exported.548
Soybean oil exports in 2020/2021 are estimated at approximately 2.6 billion pounds, or enough
soybean oil to produce approximately 340 million gallons of biodiesel or renewable diesel.549
Soybean oil imports have been relatively small (300 - 400 million pounds)550 in recent years,
likely due to the tariff on soybean oil imports.551 As with other potential sources of BBD

540 U.S. Soybean crushing data sourced from USDA's Oil Crops Yearbook (https://www.ers.usda.gov/data-
products/oil-crops-vearbook/).

541 Ates, Aron M. and Bukowski, Maria. Oil Crops Outlook: February 2022. USDA Economic Research Service.
February 11, 2022.

542 Ibid

543 USDA Economic Research Service. Oil Crops Data: Yearbook Tables. March 26, 2021.

544 Ibid.

545 Demarre-Saddler, Holly. Cargill Plans US Soy Processing Operations Expansion. World-Grain.com. March 4,
2021.

546 See comments from the American Soybean Association (EPA-HQ-OAR-2021-0324-0471).

547 Ibid.

548 USDA Economic Research Service. Oil Crops Data: Yearbook Tables. March 26, 2021.

549 Ibid

550 Ibid

551 Harmonized Tariff Schedule of the United States (2020) Revision 19.

184

-------
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 most recent WASDE report from USD A project that global production of vegetable
oils will be approximately 214 million metric tons in the 2021/2022 agricultural marketing
year.552 This quantity of vegetable oil, if converted to fuel, would result in approximately 61
billion gallons of biodiesel and/or renewable diesel. The vast majority of this is used for food and
other purposes and could not be readily used to supply advanced biodiesel and renewable diesel
to the U.S..553 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 the next chapter), may be made
available to markets in the U.S. in future years.

EPA received several comments on our proposed rule suggesting we should finalize
lower volumes for BBD and advanced biofuel due to concerns that increasing demand for
vegetable oils to produced biodiesel and renewable diesel would result in continued high prices
for vegetable oils and insufficient supply to other markets.554 Some commenters raised particular
concerns over the availability of refined vegetable oil, which is used in some biofuel production
processes.555 We recognize that vegetable oil prices have increased significantly in recent
years556, and that these high prices affect both the biodiesel and renewable diesel industries and
others industries that use vegetable oil, such as the food industry. While increased soybean oil
demand for biofuel production is likely a contributing factor, we believe the current high prices
are largely the result of poor weather conditions in South America and Malaysia over the past
year. In 2021 there was drought in Argentina and Brazil (two of the largest exporters of soybeans
and soybean oil).557 At the same time, palm oil production in Malaysia was impacted by flooding
caused by a typhoon.558

Despite these current high prices for vegetable oil, the data discussed above indicate that
there will be a sufficient supply of vegetable oil to produce the volumes of biodiesel and

552 United States Department of Agriculture World Agricultural Supply and Demand Estimates. January 12, 2022.

553 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.

554 For example, see comment from the American Bakers Association (EPA-HQ-OAR-2021-0324-0433).

555 Ibid.

556 According to the January 2022 WASDE soybean oil prices increased from 29.65 cents per pound in the
2019/2020 agricultural marketing year to 56.87 cents per pound in the 2020/2021 agricultural marketing year, and
are expected to reach 65 cents per pound in the 2021/2022 agricultural marketing year.

557 Wilson, Nick. Oil Prices Surge - Vegetable Oil That Is. Marketplace.org. February 17, 2022.

558 Ibid.

185

-------
renewable diesel we project will be used to meet the volumes we are finalizing for 2022 in this
rule. We received several comments supporting this conclusion.559 Better harvest conditions
could result in an increased supply of soybeans and vegetable oil from Argentina, Brazil, and
Malaysia. As discussed earlier, soybean crushing facilities have already increased their crush
rates, and there is likely some ability to increase their production of soybean oil further.
Additionally, several new soybean crushing facilities are currently under construction, and
expansions of existing facilities are also underway. Increased domestic crushing of soybeans
would supply greater quantities of soybean oil without requiring increased soybean production.
While we do not believe that the availability of vegetable oil will limit biodiesel and renewable
diesel production in 2022, we do recognize that the high price of vegetable oil does have an
impact on both food prices and the cost of renewable fuels produced from vegetable oils (see
Chapters 7 and 9 for a further discussion of the impact on food prices and fuel costs).

5.2.4 Imports and Exports of Biomass-Based Diesel

In evaluating the likely rate of production of BBD through 2022 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
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
2022 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.560 These tariffs were subsequently confirmed in April 2018.561
Since the time the preliminary tariffs were announced, EIA has not reported any biodiesel
imported from these countries.562 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
2018-2021.

559 For example, see comment from Minnesota Soybean Producers (EPA-HQ-OAR-2021-0321-0442).

560 82 FR 40748 (August 28, 2017).

561 83 FR 18278 (April 26, 2018).

562 See EIA data on biodiesel imports by country, available at:

https://www.eia.gov/dnav/pet/pet move impctis a2 tins EPOORDB imO rnbbt a.htm

186

-------
Figure 5.2.4-1: Biodiesel and Renewable Diesel Imports (2014-2021)

800

2014 2015 2016 2017 2018 2019 2020 2021

¦ Biodiesel imports ¦ Renewable Diesel Imports
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.563 We expect that, as the carbon intensity
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.

Biodiesel exports have been fairly consistent since 2014, generally ranging between 70
and 100 million gallons per year. Renewable diesel exports increased with domestic renewable
diesel production, reaching over 200 million gallons in 2020 and 2021. Increasing exports of
renewable diesel reflect the existence of biofuel mandates and significant financial incentives
creating high demand in other countries as well that the U.S. must compete with. 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.

563 Data from California's LCFS program indicates that approximately 590 million gallons of renewable diesel were
consumed in California in 2020, the most recent year for which data are available

(https://ww3.arb.ca.gov/fuels/lcfs/dashboard/dashboard.htm). Data from EMTS indicates that 580 million gallons of
renewable diesel were consumed in the U.S. in 2020, including both renewable diesel that generated BBD RINs and
advanced RINs.

187

-------
Since the U.S. tax credit for biodiesel and renewable diesel applies to fuel either used or
produced in the U.S. it applies equally to fuel whether it is used in the U.S., or exported.
Furthermore, 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 increase claim the biodiesel tax credit on both the imported and the exported volumes.

Figure 5.2.4-2: Biodiesel and Renewable Diesel Exports (2014-2021)

350
300

¦ ¦I

¦lllllll

2014 2015 2016 2017 201S 2019 2020 2021
¦ Biodiesel Exports ¦ Renewable Diesel Exports

Biodiesel and renewable diesel exports based on data from EMTS

5.2.5 Projected Rate of Production and Use of Biomass-Based Diesel

Based on the factors discussed in the preceding chapters we have projected BBD
production in the U.S. through 2022. We have also included projections of BBD imports and
exports, as imports and exports of BBD impact the quantity of BBD available to obligated
parties. Table 5.2.5-1 shows our projections for 2022, along with the observed volumes from
2018-2021 to provide context for our projections and a basis for the volumes we are finalizing
for 2020 and 2021. We have focused on data from 2018 through 2021 as this is the most recent
data available, and we believe the trends observed in these years are most likely to be predictive
of growth in BBD production through 2022. Notably, data from 2017 and earlier years does not
reflect the impact of the tariffs on biodiesel from Argentina and Indonesia.

188

-------
Table 5.2.5-1: BBD (D4) Biodiesel and Renewable Diesel Production, Imports, and Exports

Projected Through 2022 (Mi

lion Gallons

2018

2019

2020

2021

2022

Domestic Biodiesel
Production (million RINs)

1,841
(2,762)

1,706
(2,559)

1,802
(2,703)

1,699
(2,549)

1,760
(2,640)

Domestic Renewable Diesel
Production (million RINs)

282
(479)

454
(772)

472
(802)

780
(1,326)

1,500
(2,550)

Biodiesel Imports
(million RINs)

175
(263)

185
(277)

209
(314)

208
(313)

190
(290)

Renewable Diesel Imports
(million RINs)

176
(300)

267
(454)

280
(477)

362
(614)

420
(710)

Biodiesel Exports
(million RINs)

74
(111)

76
(115)

88
(132)

90
(136)

100
(150)

Renewable Diesel Exports
(million RINs)

80
(136)

145
(247)

223
(379)

241
(409)

290
(490)

Total BBD Available
(million RINs)a

2,322
(3,559)

2,393
(3,705)

2,456
(3,792)

2,721
(4,260)

3,480
(5,550)

a Total BBD includes some small volumes (<10 million gallons per year) of D4 jet fuel.

Our projections for each of the various potential sources of BBD in 2022 in Table 5.2.5-
1, and described below, are based on the information presented in the preceding chapters and
further discussed below.

As discussed in Chapter 5.2.3, since 2016 there has been a consistent relationship
between domestic renewable diesel production capacity and domestic renewable diesel
production, with consistently high production capacity utilization rates for renewable diesel. A
number of new renewable diesel production facilities began production in 2021, and several
additional facilities are currently being commissioned or are under construction with the
intention to begin producing fuel in 2022. Several of these facilities are conversions of existing
refineries, which are good candidates for conversion to renewable diesel production facilities due
to the available infrastructure and the similarities between some of the refinery processing units
and the equipment needed to produce renewable diesel. Market factors, such as restrictions on
selling biodiesel blends that contain more than 5% biodiesel in states such as California with
significant biofuel incentives through their LCFS program are also favoring increased production
of renewable diesel over biodiesel.

At this time we are unaware of any other factors that are likely to inhibit these facilities
from being able to produce renewable diesel at or near their intended facility capacity after a
period of commissioning is complete. If domestic renewable diesel production capacity
continues to expand at this rate it is possible that the availability of feedstock could be a limiting
factor in future years, but we do not expect that the availability of feedstock will limit renewable
diesel production through 2022. As noted above, we believe that the primary feedstock used to
supply this increased renewable diesel production will be increased use of soybean oil, but some
increased reliance on other feedstocks (such as FOG, canola oil, and distillers corn oil) is also
probable, and that there are sufficient quantities of these feedstocks to support significantly more
renewable diesel production. Moreover, the surge of additional investments in renewable diesel

189

-------
production capacity through 2022 also suggests that market participants believe there are
sufficient feedstocks to enable the productive use of these new production facilities. We have
therefore projected that domestic renewable diesel production increases consistent with the
increases in domestic renewable diesel production capacity through 2022 (shown in Table 5.2.2-
1), after accounting for the announced start-up dates for these facilities and a period of facility
commissioning564 and production ramp-up.565

Domestic biodiesel production dropped significantly in 2019 from its peak in 2018, only
to recover much of its losses in 2020 and dropping again in 2021, providing no discernable trend
in recent years. Similarly, there is no discernable trend in biodiesel imports from 2018-2021.
Biodiesel imports increased from 2018-2020, and then decreased slightly in 2021. As discussed
above, going forward into 2022, it would appear that future growth in BBD production is more
likely to come from increased renewable diesel production rather than increased biodiesel
production. In light of current feedstock prices and limitations it is possible that some of the
expansion in renewable diesel production may come at the expense of biodiesel production.
Nevertheless, we are not projecting a decrease in biodiesel production or imports in 2022 since
biodiesel producers can take advantage of existing facility capacity and previously established
feedstock supply agreements and downstream markets. Given the high value of BBD after
accounting for the incentives provided by the RFS program, state LCFS programs, and the
federal tax credit, and lacking information to the contrary, we have projected that biodiesel
production and imports will continue at the average observed volume from 2018-2021 (1,760
million gallons and 190 million gallons respectively) through 2022.

For the remaining three categories of potential sources of BBD through 2022 (renewable
diesel imports and biodiesel and renewable diesel exports) the observed data from 2018-2021
suggested trends for each of these potential sources of BBD. At this time we are unaware of any
information that would suggest these trends are unlikely to continue through 2022. For each of
these categories of fuel we used linear regressions of the data from 2018 - 2021 to project the
volumes in 2022. As demonstrated by the observed data presented in Chapter 5.2.5, projecting
imports and exports of BBD is challenging as these volumes can be significantly impacted by
both market conditions in various countries as well as changing incentives and trade policies in
the U.S. and around the world. It is possible that higher RFS volumes, in conjunction with the
federal tax credit and other state incentive programs, could result in higher imported volumes of
BBD (or lower exports from the U.S. to other countries) than shown in Table 5.2.5-1. In
particular, significant expansion of renewable diesel production in foreign countries has also

564 Commissioning is the process of planning, documenting, scheduling, testing, adjusting, verifying, and training, to
provide a facility that operates as a fully functional system.

565 Renewable diesel production in 2022 from domestic renewable diesel facilities that are currently producing fuel
was based on production in recent months for which data are available. To project domestic renewable diesel
production in 2022 from facilities that have not yet started production we have projected production based on the
projected start of production at each facility and a six-month ramp-up period. This assumes that new facilities are
producing at 50% capacity 3 months after the start of production and 100% capacity 6 months after the start of
production. This results in an average production rate of 50% of their listed facility capacity over this 6-month ramp-
up period. For facilities that begin production on January 1 it would result in production at 75% of the nameplate
capacity, which is lower than the observed average utilization rate for renewable diesel facilities in recent years.
Facilities that begin production later in the year would have significantly lower utilization rates. More details on
domestic renewable diesel production can be found in "Projected U.S. Renewable Diesel Production," available in
docket EPA-HQ-OAR-2021-0324.

190

-------
been announced.566 While the U.S. (particularly California) is likely to be an attractive market
for increasing volumes of imported renewable diesel other countries are also providing
significant incentives for biofuel use or have established ambitious renewable fuel targets, and in
some cases mandates. There is likely to be significant competition for renewable diesel from a
number of potential markets.

5.3 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 imported
advanced ethanol. In setting the 2020 standards, we estimated that 70 million gallons of imported
sugarcane ethanol would be imported in 2020.567 This was based on a combination of data from
recent years demonstrating relatively low import volumes and older data indicating that higher
volumes were possible. We also noted the high variability in ethanol import volumes in the past
(including of Brazilian sugarcane ethanol), increasing gasoline consumption in Brazil, and
variability in Brazilian production of sugar as reasons why it would be inappropriate to assume
that sugarcane ethanol imports would reach the much higher levels suggested by some
stakeholders.

At the time of the 2020 standards final rule, we also noted that while sugarcane ethanol
imports had stabilized at a relatively low level after 2013, averaging 67 million gallons, partial
data from 2019 indicated a notable increase. Nevertheless, at that time we had little evidence that
the increase exhibited in 2019 would continue into 2020 as there was no consistent upward or
downward trend after 2013, and several factors created disincentives for increasing imports
above the levels in recent years. We therefore projected that 70 million gallons of sugarcane
ethanol could be supplied in 2020.

Import data for 2019 - 2021 is now available and indicates that imports of sugarcane
ethanol were sometimes higher than in previous years.

566 por exampie see https://www.futurebridge.com/industry/perspectives-energy/renewable-diesel-the-fuel-of-the-
future

567 85 FR 7016, 7032-34 (February 6, 2020).

191

-------
Figure 5.3-1: Historical Sugarcane Ethanol Imports

800

700

600

500

' 400

300

200

100

I I

i I ¦ i i I I i

rNfNrNfNrNfNrNfNfNfNrslfNrslfNPvlfNrNfN

Source: "US Imports of Brazilian Fuel Ethanol from EIA - February 2022." 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. 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.

It is not certain whether the increase in imports in 2019 - 2021 in comparison to previous
years represents the beginning of an upward trend that could reasonably be expected to continue
in 2022. It is likely that the LCFS standards will continue to become more stringent and provide
an incentive to increase imports of sugarcane ethanol into California, but the reinstatement of the
biodiesel tax credit for 2020 through 2022 could provide a preferential incentive for biodiesel
and renewable diesel over that of sugarcane ethanol imports for meeting the advanced biofuel
volume requirement.568 On balance, we believe that the increase in ethanol imports in 2019 -
2021 may be indicative of future import volumes and thus should be taken into account in
estimating imports that could occur in 2022.

As noted in previous annual standard-setting rulemakings, the high variability in
historical ethanol import volumes makes any projection of future imports uncertain.569 However,
import volumes for more recent years are likely to provide a better basis for making future
projections than import volumes for earlier years, and the higher volumes for 2019 - 2021
disrupted the trend of lower volumes that had occurred between 2014 and 2018. To address these
issues, we are using a different methodology for making projections of future ethanol imports
than we have used in previous years.570

Specifically, we are using a weighted average of import volumes for all years where the
weighting is higher for more recent years and lower for earlier years. The weighting factor for

568 "ARB summary of LCFS program." See slide 6.

569 See, e.g., 85 FR 7032-33 (February 6, 2020).

5711 See FCC v. Fox Television Stations, Inc., 556 U.S. 502, 515 (2009).

192

-------
any given year's volume is twice as large as the weighting factor for the previous year's volume.
This approach provides a better predictor of future imports of sugarcane ethanol than either
simple averages of historical volumes or a trendline based on historical volumes. While this
approach could in theory be used to account for import data from all years, in practice this would
not be necessary as data from more than eight years ago would collectively contribute only 0.4%
to the weighted average. Given the levels of the volumes in the 2014 - 2021, the addition of 2013
and earlier data has an impact on the weighted average of less than 1 million gallons, which is
not meaningful in the broader context of the assessment of the volume requirements that we
believe are appropriate to establish.

The volumes and weighting factors we are using are shown below. The resulting
weighted average is 110 million gallons. For 2022, this is the volume of sugarcane ethanol that
we project can be supplied.

Table 5.3-1: Annual Advanced Ethanol Imports and Weighting Factors

Year

Imported advanced
ethanol3 (million gallons)

Weighting factor

2014

0.0078125

2015

0.015625

2016

0.03125

2017

0.0625

2018

0.125

2019

196

0.25

2020

185

0.5

2021

a Based on RINs generated for imported ethanol and assigned a D-code of 5 according to EMTS.

In comparison to the straight average approach used in previous annual standard-setting
rules, this new approach generates a similar projection. However, the similarity depends on the
years over which a straight average would be calculated. For instance, in the rulemaking that
established the original 2020 standards, we used an average of the import volumes for the five-
year period 2014 - 2018 as a projection for 2020. A straight average for all years since 2014
(2014 - 2021) would yield 98 million gallons, while a straight average for the most recent five
years would yield 119 million gallons. These values bracket the 110 million gallons derived from
the weighted approach described above.

As noted above, the future projection of imports of sugarcane ethanol is inherently
imprecise, and actual imports in 2022 could be lower or higher than 110 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 E10
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.

193

-------
5.4 Other Advanced Biofuel

In addition to cellulosic biofuel, imported sugarcane ethanol, and BBD, there are other
advanced biofuels that can be supplied in 2022. These other advanced biofuels include non-
cellulosic CNG, naphtha, heating oil, co-processed renewable diesel, and domestically produced
advanced ethanol. However, the supply of these fuels has been relatively low in the last several
years.

Table 5.4-1: Historical Supply of Other Advanced Biofuels (million ethanol-equivalent

gallons)

Domestic

Heating

Renewable

Year

CNG/LNG

Ethanol

Oil

Naphtha

Diesel (D5)

Total

2014

2015

2016

2017

2018

101

2019

127

2020

150

2021

164

We have used the same weighted averaging approach (see Table 5.3-1) for other
advanced biofuels as we have used for sugarcane ethanol to project the supply of these other
advanced biofuels in 2022. Based on this approach, the weighted average of other advanced
biofuels is 146 million gallons.571 This volume of other advanced biofuel is composed of 25
million gallons of advanced ethanol and 121 million gallons of non-ethanol advanced biofuel.

We recognize that the potential exists for additional volumes of advanced biofuel from
sources such as 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 2022.572

5.5 Ethanol

In order to properly analyze the volume requirements for 2020 - 2022, 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

571 Under our prior approach, we would have taken a straight average of 2015-2021, all years following the change
in the categorization of landfill biogas from D5 to D3. This approach would yield a volume of 105 million gallons,
somewhat smaller than the 146 million gallons derived from the new approach. Moreover, for both sugarcane
ethanol and other advanced biofuels, given the relatively small volumes we are projecting, even a significant
deviation in their actual availability would likely have negligible impact on the market's ability to meet the
advanced biofuel volumes.

572 Less than 1 million gallons of these other renewable fuels were produced in the years prior to 2021. In 2021, 8
million gallons of D4 jet fuel was produced, but it is difficult to project volumes of jet fuel for 2022 based on data
from this single year.

194

-------
topic we discuss further in Chapter 6. It was therefore necessary to estimate the total volume of
ethanol that is projected to be consumed, which we discuss in Chapter 5.5.1.

In Chapter 5.5.2, we project the volume of ethanol production for 2022. Together with
our other estimates of biofuel production rates in this chapter, this contributes to fulfilling the
statutory requirement that we analyze "the expected annual rate of future commercial production
of renewable fuels." CAA section 21 l(o)(2)(B)(ii)(III).

In Chapter 5.5.3, we project the volume of corn ethanol used in 2022.

In Chapter 5.5.4, we estimate volumes of E15 and E85 for the singular purpose of
estimating costs. We do not use the estimated volumes of El 5 and E85 in any other context in
the RIA, including to estimate total ethanol consumption, which instead is based on EIA
projections as discussed in Chapter 5.5.1.

5.5.1 Total Ethanol Consumption

In the 2020 final rule, we stated that "we believe that the market is capable of reaching an
ethanol concentration in 2020 that is at least as high as the highest level that has actually been
reached by the market in the past."573 Were we to apply that approach in this rule, we would
project a nationwide average ethanol concentration in 2022 of 10.36%, equal to the historical
highest level, which occurred in 2021.

However, in analyzing the volumes of ethanol consumption that may be achievable in
2022, we note that the ethanol concentrations actually achieved in 2020 and 2021 may not be
representative of what is achievable in 2022 due to very atypical market conditions in those years
caused by the COVID-19 pandemic. Indeed, the projected nationwide average ethanol
concentration for 2022 derived from EIA's ethanol and gasoline projections is somewhat lower
than it was in 2021 as shown below.

Table 5.5.1-1: Average Ethanol Concentration from EIA

Year

Nationwide average
ethanol concentration

EIA Report574

2017

10.13%

January 2022 MER

2018

10.08%

January 2022 MER

2019

10.20%

January 2022 MER

2020

10.25%

January 2022 MER

2021

10.36%

January 2022 MER

2022

10.30%

January 2022 STEO

As a result, we have used the projection of ethanol concentration derived from EIA
reports for 2022 as a reasonable estimate of what level can be achieved in 2022, rather than the
actual ethanol concentration in 2021.

573 See id. at 3.

574 STEO = Short Term Energy Outlook. MER = Monthly Energy Review. Ethanol concentration in % was
calculated to two decimal places from the ethanol volume and total gasoline volume available in these reports.

195

-------
By using the ethanol concentration shown in Table 5.5-1 for 2022, we are by definition
using EIA's projections of total ethanol consumption since the ethanol concentration is derived
from ethanol volume and gasoline volume. We use EIA's estimates of total ethanol consumption
in 2020 - 2022 for our quantitative analyses throughout this rule. These volumes are shown
below.

Table 5.5.1-2: Total Ethanol Consumption (million gallons)

2020

12,704

2021

13,878

2022

14,310

5.5.2 Domestic Ethanol Production

We begin our analysis of domestic ethanol production by considering production
capacity. As described in more detail in Chapter 1.6.1.1, 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.575'576 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 denatured ethanol in the
U.S. reached 12.85 billion gallons in 2020, compared to 14.72 billion gallons in 2019.577It rose
to 14.11 billion gallons in 2021, still below pre-pandemic levels.578

The expected annual rate of future commercial production of ethanol will be driven
primarily by gasoline demand as most gasoline is expected to continue to contain 10% ethanol in
2022. Commercial production of 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 ethanol may be limited by
production capacity in the abstract, it does not appear that production capacity (at least 17.4
billion gallons as described above) will be a limiting factor in 2022 for satisfying domestic
consumption, which as shown in Table 5.5.1-2 we estimate will include 14.3 billion gallons of
ethanol.

Next, we turn to ethanol production. When combined with potential imports of ethanol,
domestic ethanol production provides an indication of what volumes could be made available for
use in transportation fuel. Constraints on ethanol consumption, as discussed in Chapter 5.5.1,
have historically constrained consumption well below the level of production and are anticipated
to do so through 2022, as other biofuels, specifically biodiesel and renewable diesel have tended
to outcompete ethanol used at higher blend concentrations (as El5 or E85).

Historical domestic ethanol production is estimated by EIA in its Monthly Energy
Review (MER). As of March, 2022, the MER only provided annual ethanol production volumes

575 "2021 Ethanol Industry Outlook - RFA," available in the docket.

576 "Ethanol production capacity - EIA April 2021," available in the docket.

577 "RIN supply as of 2-17-22," available in the docket.

578 Ibid.

196

-------
through 2020. These are provided below, along with estimated consumption and net exports of
ethanol.

Table 5.5.2-1: Historical Domestic Ethanol Production and Net Exports (million gallons)

Production

Net Exports

Consumption
(production - exports)

2011

13,929

1,036

12,893

2012

13,218

336

12,882

2013

13,293

13,216

2014

14,313

869

13,444

2015

14,807

861

13,947

2016

15,413

1,057

14,356

2017

15,936

1,451

14,485

2018

16,091

1,671

14,420

2019

15,778

1,227

14,552

2020

13,941

1,260

12,681

Source: MER March 2022 Table 10.3

EIA estimates 2021 and 2022 ethanol production in the January 2022 edition of its Short-
Term Energy Outlook. For 2021 it estimates that 14,982 billion gallons was produced, while for
2022 it forecasts that 15,672 billion gallons will be produced.

We also considered a separate estimate of 2021-2022 ethanol production. We can
separately estimate domestic ethanol production volumes for 2021 and 2022 by combining EIA
estimates of ethanol consumption for these two years with a projection of net ethanol exports.
We applied the weighting methodology described in Chapter 5.3, which we believe is
appropriate in this case for substantially the same reasons, to the historical net export data in
Table 5.5.2-1 to determine that future exports could be 1,295 million gallons. Although the
March 2022 edition of EIA's MER does not include annual ethanol consumption data for 2021,
the January 2022 edition of the STEO includes an estimate of 2021 consumption. The STEO also
includes a forecast of ethanol consumption for 2022 that we are using in this rulemaking, as
described more fully in Chapter 5.5.1. The resulting estimates of domestic ethanol production for
2021 and 2022 are shown below.

Table 5.5.2-2: Estimates of Domestic Ethanol Production for 2021 and 2022 (million

gallons)

Year

Consumption

Exports

Production
(consumption + exports)

2021

13,878

1,295

15,173

2022

14,310

1,295

15,605

These numbers are similar to those provided directly by EIA in the January 2022 edition
of its STEO. In any event, we reiterate that ethanol consumption in the U.S. is currently
constrained by factors other than production. These higher production numbers (regardless of
whether we use the EIA or EPA estimate) do not indicate that the market can use this quantity of
ethanol or that we should assume such use in our analyses.

197

-------
5.5.3

Corn Ethanol

The total volume of ethanol consumed is a combination of corn ethanol, cellulosic
ethanol, and advanced ethanol. Our estimate of corn ethanol consumption for 2022 for the
purposes of estimating the mix of biofuels that could be made available is shown below. We
derived this estimate by subtracting from total ethanol consumption (described in 5.5.1) our
projections of cellulosic ethanol (5.1), imported sugarcane ethanol (5.3), and domestic advanced
ethanol (5.4).

Table 5.5.3-1: Calculation of Projected Corn Ethanol Consumption for 2022 (million

Total ethanol

14,310

Cellulosic ethanol

Imported sugarcane ethanol

110

Domestic advanced ethanola

Corn ethanol

14,175

a Derived from values in Table 5.4-1 using the weighted average approach described in Chapter 5.3.

For 2020 and 2021, we have historical data on each of the types of ethanol consumption.
We present this information in Chapter 2.1.

5.5.4 Estimation of Gasoline Blend Volumes for Cost Purposes

In some earlier rulemakings, we estimated volumes of E0, El5, and E85 for the purpose
of projecting total ethanol consumption. However, as we have noted in recent annual rules, the
projection of these blends for future years has been hampered by a lack data on nationwide
consumption of each individual blend.579 Data on these blends remains limited at this time. As
such, we believe it is more technically sound to project total ethanol consumption based on the
estimates provided by EIA discussed in Chapter 5.5.1.

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. As
we discuss in Chapter 9, the costs of distributing El 5 and E85 vary considerably from costs of
distributing E10. Thus for cost purposes only, we have estimated the volumes of E15 and E85
that were consumed in 2020 and 2021 and may be consumed in 2022. These volume estimates
were based on data collected by USD A through their BIP program and made available to
EPA.580 While this data includes only a subset of all El5 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 E15 sales at retail is from two

579 See, e.g., "Updated market impacts of biofuels in 2020" 3-4.

580 "Communication with USDA on the BIP program 1-19-22," available in the docket.

198

-------
states (Iowa and Minnesota)581'582 while 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).583

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. As a result, we would expect that the data
collected for 2019 and 2020 would be more representative of the country as a whole than the
data collected for 2015 - 2018.

Between these two years, we chose to use BIP data from 2019, as opposed to 2020, for
our analyses for 2020-2022. The BIP per-station sales volumes for 2019 reflect rising volumes
over the prior several years and are somewhat higher than for 2020.

Figure 5.5.4-1: Per-Station E15 Sales Volumes from BIP Program

=¦ 130,01

1 120,0(
11G,QC

i/>

1 1QO.OC
o

90.CX

80,CX

This 2020 decrease in El5 and E85 volumes in the BIP data is consistent with the
decrease in total gasoline consumption brought about by the COVID pandemic. Given the
rebound in gasoline demand in 2021 and 2022, as well as the overall increase in El 5 and E85 use
leading up to 2019, we believe that using the higher 2019 BIP volumes for all three years reflects
a reasonable choice. Moreover, a similar upward trend is evident in per-station El 5 sales
volumes in Minnesota.

581 Iowa Department of Revenue, https://tax.iowa.gov/report-category/retailers-annual-gallons

582 Minnesota Commerce Department, https://mn.gov/commerce/consumers/your-vehicle/clean-energy.jsp

583 See discussion of data sources in "Estimate of E85 consumption in 2020," available in the docket.

199

-------
Figure 5.5.4-2: Per-Station E15 Sales Volumes in Minnesota

35,000

Source: Minnesota Commerce Department, https://mn.gov/commerce/consumers/your-veliicle/clean-energy.jsp. 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,
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. Regardless, the use of BIP data from 2019 as the
basis for making a projection for 2022 is nevertheless consistent with the increasing trend shown
above. Indeed, that trend suggests that per-station El5 sales volumes for 2022 could be even
higher than they were in 2019.

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 6.4.2 and 6.4.3), we were able to project volumes of E15 and E85 as
shown below.

Table 5.5.4-1: Estimated Volume of E15

El5 stations

El5 sales volumes

Annual El5 sales

Ethanol in Excess

per year per station

volumes (mill gal)

of E10 (mill gal)

2020

2,180

146,857

320

2021

2,458

146,857

361

2022

2,998

146,857

440

Table 5.5.4-2: Estimated Volume of E85

E85 stations

E85 sales volumes

Annual E85 sales

Ethanol in Excess

per year per station

volumes (mill gal)

of E10 (mill gal)

2020

3,979

78,342

312

206

2021

4,156

78,342

326

215

2022

4,333

78,342

339

224

These are the volumes used to estimate distribution costs associated with El5 and E85 for
2020 - 2022 as discussed in Chapter 9.

200

-------
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. We note, however, that the E15 and E85
volume estimates shown in Tables 5.5.4-1 and 5.5.4-2 correspond to lower total ethanol
consumption than the volumes shown in Table 5.5.1-2. Below we explain how we realized this
discrepancy, why we chose to use estimates from EIA for total ethanol consumption rather than
those derived from the E15 and E85 estimates shown in Tables 5.5.4-1 and 5.5.4-2, and why we
nonetheless believe it is reasonable to use these El 5 and E85 estimates for purposes of our costs
analyses.

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, was
back-calculated from an estimate of EO and total gasoline energy demand derived from the
January 2022 edition of EIA's STEO. As noted above, this exercise produces an amount of total
ethanol consumption that is somewhat less than the EIA projection of total ethanol consumption
shown in Table 5.5.1-2.

As for El 5 and E85, there is very little available data on the consumption of EO. Iowa's
Department of Revenue has collected data on EO sales at retail for many years, but it is unclear
how representative Iowa EO sales are of the entire nation. For instance, the pattern of
consumption of EO in Iowa does not appear to have followed the nationwide consumption pattern
of total gasoline since 2012.

Figure 5.5.4-3: EO Consumption in Iowa

ion

155,000

¦ 150,000

TO
QO

¦ 145,000 ^

- 140,000 '£

Q.
£
3

270

"ro

= 250

«5

o 230
_c
c
o

t 210

l/>

o 190

170

150
20

(

¦ 135,000 £

¦ 130,000 J

¦ 125,000 &
-o

¦ 120,000 =
.2

¦ 115,000 2

¦L 1U,UUU

11 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

EO data source: See, for instance, "2020 Retailers Fuel Gallons Annual Report," available in the docket.
https://tax.iowa.gov/reports.

Gasoline consumption data source: "MER March 2022 Table 3.5," available in the docket.

Moreover, EO sales in Iowa in 2020 represented about 15% of total gasoline sales, a
proportion that is much too high to be representative of the nation as a whole. For instance, 15%
of total gasoline consumption in 2020 would mean that 18.6 billion gallons of gasoline would
have been sold as EO.584 Given that total ethanol consumption was 12.7 billion gallons in 2020

584 According to the January 2022 edition of EIA's STEO, total gasoline consumption in 2020 was 123.73 billion
gallons.

201

-------
(see Table 5.5.2-1), 18.6 billion gallons of EO would have required that 37.9 billion gallons of
E15 be consumed.585 As there were about 2,180 retail stations offering E15 in 2020 (see Table
6.4.3-2), the sale of 37.9 billion gallons of E15 would have required each station to sell 17
million gallons of E15 in 2020. This is far more than the total volume of gasoline sold by the
largest retail stations.

Nevertheless, the Iowa data represents the only available estimate of actual E0 sales
volumes. Therefore, we used the most recently available Iowa data to estimate that the average
per-station sales of E0 were 128,642 gal/year.586 For the total number of retail stations in the
U.S. that offer E0, we used the estimate provided by Pure-gas.org on January 7, 2022, which was
16,544 stations. The result was an estimate of 2,128 million gallons of E0.

We were able to confirm that 2,128 millon gallons of E0 is consistent with data collected
by the RFG Survey Association. That data indicated that 1.44% of sampled gasoline blends sold
in 2021 was EO.587 Insofar as this proportion is representative of the nation as a whole, 1.44%
corresponds to about 1,940 million gallons of E0, based on EIA's estimate of 134.75 billion
gallons of gasoline sold in 2021.588 Since 1,940 million gallons is very similar to 2,128 million
gallons, we had confidence that 2,128 million gallons is a reasonable estimate of E0
consumption.

Estimates of E0, El 5, and E85 consumption, combined with total gasoline energy
demand from EIA, allowed us to calculate the consequent volume of E10 consumed as shown
below.

Table 5.5.4-3: Estimating E10 Consumption

E15a

E85b

Gasoline energy

E10c

(mill gal)

(Quad Btu)

(mill gal)

2020

2,128

320

312

14.836

120,532

2021

2,128

361

326

16.201

131,884

2022

2,128

440

339

16.700

135,972

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 can then be compared to the total ethanol consumption estimated by EIA (Table 5.5.1-2).
The results are shown below.

585 Assumes that E85 consumption was 312 million gallons in 2020. See Table 5.5.4-2.

586 "2020 Retailers Fuel Gallons Annual Report," available in the docket.

587 "National Fuels Survey Program Ethanol Data for the 2021 Compliance Period," available in the docket.

588 "STEO Jan 2022 Table 4a," available in the docket.

202

-------
Table 5.5.4-4: Comparison of Estimated Total Ethanol Consumption (million gallons)

EIA Estimates

Based on projected
ethanol blend volumes

Difference

2020

12,704

12,453

-251 (-2.0%)

2021

13,878

13,610

-268 (-1.9%)

2022

14,310

14,041

-269 (-1.9%)

Table 5.5.4-5: Comparison of Estimated Ethanol Concentration (volume percent)

EIA Estimates

Based on projected
ethanol blend volumes

Difference

2020

10.25%

10.10%

-0.15%

2021

10.36%

10.11%

-0.25%

2022

10.30%

10.12%

-0.18%

We note that if we had used the 2020 results for per-station El 5 and E85 sales volumes
from the BIP program to estimate total 2020 ethanol consumption, rather than using the 2019
results from the BIP program, the difference between that estimate of total ethanol consumption
and the EIA estimate would have been larger at 321 million gallons instead of the 251 million
gallons shown in Table 5.5.4-4. As noted above, this provides an additional reason for using
2019 BIP to represent sales of E15 and E85 for all years 2020 - 2022.

We believe that the total ethanol consumption estimates provided by EIA are likely to be
more accurate than those derived from our estimates of E0, E10, El 5, and E85 volumes for two
reasons:

1. The EIA estimates of total historical ethanol consumption are based on better data.
They are derived from direct assessments of ethanol volumes produced, imported,
exported, and stored, nationwide. By contrast, as explained above, there are
significant limitations on the availability of data regarding volumes of specific
ethanol blends.

2. The total ethanol consumption estimates from EIA are very similar to those from the
EPA-Moderated Transaction System (EMTS). Over the 10-year period 2012 - 2021,
the two differed by an average of only about 90 million gallons, or about 0.6%.

EMTS data reflects an independent estimate of total ethanol use in the U.S. This data
is collected based on an entirely different methodology than that which EIA uses, as it
is based on reports provided by regulated parties for RFS compliance, as opposed to
responses to EIA surveys. As such, the similarity between the total ethanol
consumption volume across these two sources provides significant corroboration that
the EIA data is accurate.

We do not know for certain why the EIA data on projected ethanol consumption differs
from the ethanol consumption calculated based on projected ethanol blend volumes. We
hypothesize that these differences may be the result of errors in estimates of E0, the possibility
that the data collected through the BIP program may not be representative of the nation as a
whole, underestimates of the total gasoline demand, or some combination of the three.

203

-------
Regardless, we believe that the estimates of E15 and E85 consumption shown in Tables 5.5.4-1
and 5.5.4-2 are a reasonable basis for estimating distribution costs. 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 final
estimates. In any event, as shown in Chapter 9, 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 annual
standards. Thus, even were we to estimate significantly different El5 and E85 volumes, that
would have very limited impacts on our assessments of costs and no impact on our final
judgment with respect to the appropriate volumes.

5.6 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 - 2021 is shown in
Table 5.6-1.

Table 5.6-1: Conventional Biodiesel and Renewable Diesel Used in the U.S. (million gallons)

2014

2015

2016

2017

2018

2019

2020

2021

Domestic D6 Biodiesel

Domestic D6 Renewable

Diesel

Imported D6 Biodiesel

113

Imported D6 Renewable

Diesel

All D6 Biodiesel and

160

158

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 2018-2021. Nearly all of
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 2021 the volumes of renewable diesel for which conventional
biofuel RINs were generated each year (in million gallons) were 107, 116, 76, and 135
respectively. These RINs were retired for reasons other than compliance with the annual volume

204

-------
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
3 billion gallons in the U.S., with an additional 0.6 billion gallons internationally. Feedstock
availability does also not appear to be a limiting factor, as USDA estimates that approximately
214 million metric tons of vegetable oil will be produced globally in the 2021/2022 agricultural
marketing year.589 This quantity of vegetable oil could be used to produce approximately 61
billion gallons of biodiesel and renewable diesel.590 While much of this vegetable oil is currently
used in non-biofuel markets, 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. While the quantity
of conventional biodiesel and renewable diesel that could be supplied to the U.S. in 2022 is not
without limit, this data suggests that large quantities of this fuel are being or could be
produced,591 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.

After accounting for the available volumes of other types of renewable fuel (cellulosic
biofuel, advanced biodiesel and renewable diesel, other advanced biofuels, and corn ethanol) we
projected that an additional 155 million gallons (264 million RINs) of conventional biodiesel and
renewable diesel may be used in 2022 to satisfy both the 2022 total renewable fuel volume
requirement and the supplemental volume requirement. This is slightly lower than the volume of
conventional biodiesel and renewable diesel used in the U.S. in 2015 and 2016. The historical
production, together with the available data on the production capacity at grandfathered biodiesel
and renewable diesel production facilities and the quantity of feedstock available globally,
indicate that this volume of conventional biodiesel and renewable diesel can be supplied to the
U.S. in 2022.

589 USDA World Agricultural Supply and Demand Estimates. January 12, 2022.

590 This calculation assumes one gallon of biodiesel or renewable diesel can be produced from 7.6 pounds of
vegetable oil.

591 The OECD-FAO Agricultural Outlook 2021-2030 projects global biodiesel consumption to reach approximately
50 billion liters (about 13.2 billion gallons) in 2022.

205

-------
Chapter 6: Infrastructure

This chapter analyzes the impact of renewable fuels on the distribution infrastructure of
the United States. The statute 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.

6.1 Biogas

Infrastructure considerations regarding the use of renewable biogas were not evaluated in
the 2010 RFS2 rule because it was only after that time that renewable biogas was used in
significant volumes under the RFS program. 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.592 Natural gas was used in compressed natural
gas (CNG) and liquified natural gas (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.593 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, however, renewable biogas utilizes the existing CNG/LNG vehicle fleet, natural gas
distribution system, and CNG/LNG vehicle refueling infrastructure. According to data from the
Department of Energy Alternative Fuels Data Center, there are currently 1,506 public and private
CNG fueling stations and 103 public and private LNG refueling stations in the U.S.594

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 1.3 to 2.0 billion ethanol-equivalent gallons of CNG/LNG per year in 2022, would
represent a substantial challenge.595 The incentives for increasing the use of CNG/LNG in the
transportation sector, including incentives from the RFS program and state programs such as
California's LCFS, may be insufficient to cause a substantial increase in the CNG/LNG vehicle

592 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 9.1.2.6.

593 See generally 40 CFR 80.1426(f).

594 Data accessed from the AFDC website (https://afdc.enerev. gov/stations/) on February 24, 2022.

595 See Chapter 5.1.3.2 for further discussion of the estimated use of CNG/LNG as transportation fuel in 2022.

206

-------
fleet and refueling infrastructure. CNG/LNG vehicles are predominately used in fleet
applications where there is a unique situational advantage such as a natural gas supplier's utility
fleet or landfill's waste hauler fleet. In addition, it would be more challenging to establish the
necessary contacts 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 5.1.

6.2 Biodiesel

The 2010 RFS2 final 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.596 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 to 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 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
could be resolved.597 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 2010 RFS 2 rule. Most importantly, biodiesel consumption
reached over 2 billion gallons in 2016 and has remained between 1.8 and 2 billion gallons per
year from 2017-2020, largely exceeding the 1.82 billion gallons that we projected would be used
in 2022.598 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.

596 See Chapter 1.2.2 of the RFS2 Regulatory Impact Analysis (EPA-420-R-10-006).

597 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 biodiesel blends.

598 Biodiesel consumption numbers based on EMTS data.

207

-------
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.599 If the transportation of
biodiesel blends via pipeline were expanded more broadly this change could significantly reduce
the cost of biodiesel distribution. Jet fuel is a significant product on much of the petroleum
pipeline system. Therefore, 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,
however this effort will likely not be completed until after the 2022 timeframe of this rule.600
Finally, there appears to be substantial volumes of BIO - B20 being used despite the fact that a
significant number of vehicle manufacturers only warranty their engines for up to B5.601 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.602

In this rule we are not projecting significant increases in the volume of biodiesel
produced and imported in the U.S. As such, we do not anticipate any challenges associated with
the infrastructure to distribute and use biodiesel through 2022. However, it is possible that
domestic biodiesel production and/or biodiesel imports may increase in 2022 and in future years.
Domestic biodiesel production capacity is discussed in Chapter 5.2, and is significantly higher
than current production levels. A review of monthly biodiesel imports suggests that import
infrastructure can support significantly higher volumes of imports.603 For example, 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 of 2016
could be maintained year-round. Some additional expansion in import infrastructure may also
occur through 2022. 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.604

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 will be needed. These investments would primarily be associated with securing

599 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

600 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.

601 See Pilot Flying J Fuel Offerings, Memorandum to EPA Docket EPA-HQ-OAR-2021-0324.

602 See Average Biodiesel Blend Level By State Based on EIA Data, Memorandum to EPA Docket EPA-HQ-OAR-
2021-0324.

603 Energy Information Administration, U.S. Imports of Biodiesel 2009 thorough 2018.

604 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.

208

-------
sufficient downstream biodiesel storage and the requisite number of rail cars and tank trucks
suitable for biodiesel transport.605

Significantly expanding biodiesel blending infrastructure to accommodate the increase in
biodiesel volume may also pose significant 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. While many commenters
stated that there are no infrastructure constraints to expanding the consumption of biodiesel in
the U.S., significant biodiesel expansion would likely require further investment in the
distribution system.

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.606 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.607 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.608 Therefore, a
substantial increase in the use of biodiesel, especially biodiesel produced from palm oil, during
the cold weather season may be a challenge.

6.3 Renewable Diesel

The 2010 RFS2 rule projected that the volume of "drop-in" cellulosic and renewable
diesel fuel would range from 0.15 billion gallons to 3.4 billion gallons in 2017 and 0.15 billion
gallons to 9.5 billion gallons in 2022.609 Such fuels are referred to as drop-in fuels because their
physical properties are sufficiently similar to petroleum-based diesel fuel to be fungible in the
common diesel fuel distribution system.610 Thus, little change is needed to the fuels
infrastructure system to support the use of drop-in biofuels. The 2010 RFS 2 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.611

605 Biodiesel rail cars and to a lesser extent tank trucks must often be insulated and or heated during the cold season
to prevent cold flow problems. The use of such insulated/heated vessels is sometimes avoided by shipping pre-
heated biodiesel.

606 B5 blend levels can typically be maintained.

607 Biodiesel Cold Flow Basics, National Biodiesel Board, 2014.

608 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.

609 See Chapter 1.2.2 of the RFS2 Regulatory Impact Analysis (EPA-420-R-10-006).

610 Such drop-in fuels are typically blended with petroleum-based diesel fuel prior to use.

611 See Chapter 1.6 of the RFS2 Regulatory Impact Analysis (EPA-420-R-10-006).

209

-------
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. Regulatory challenges have also limited the
transportation of renewable diesel via pipeline. Regulatory programs such as the RFS program
require renewable diesel to be used as transportation fuel to be eligible to generate RINs. In
addition, renewable diesel can generate credits for California and other State biofuel programs
only if it is used in those States. Finally, product documentation requirements for product
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 2022 is
significant (see Chapter 4.2). We expect that much of this renewable diesel will be used in states
with state incentive programs for fuels with low carbon intensity such as California and Oregon.
As both these markets are located on the Pacific coast, 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.

6.4 Ethanol

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 2022 are associated with service station storage and dispensing
equipment for higher ethanol blends such as El 5 and E85, and the vehicles that are permitted to
use those blends.

Based on our analysis of the sufficiency of infrastructure to deliver and use ethanol as
discussion below, we have determined that there are constraints associated with El5 and E85
that limit the rate of future growth in their consumption. These constraints are appropriately
reflected in our use of projected ethanol consumption from EIA as discussed in Chapter 5.5 since

210

-------
those projections represent only moderate changes in the nationwide average ethanol
concentration in comparison to earlier years.612

6.4.1 Ethanol Distribution

To support the 2010 RFS2 rule, Oak Ridge National Laboratory (ORNL) conducted an
analysis of potential distribution constraints that might be associated with attaining the statutory
volume targets through 2022.613 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 in the 2010 RFS2 rule that these investments
could be made to increase the capacity of the effected rail corridors without undue difficulty.614
Therefore, the 2010 RFS2 rule concluded that the infrastructure system to the blending terminal
could accommodate the projected increased volume of ethanol in a timely fashion.

To update and expand upon the analysis of distribution infrastructure upstream of retail
that was conducted for the 2010 RFS2 rule, we 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.615 ICF's 2018 report determined that the
conclusions from ORNL's 2010 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

612 A nationwide average ethanol concentration above 10.00% can only occur insofar as there is consumption of E15
and/or E85. The ethanol volume projection from EIA for 2022 from the January 2022 STEO represents an ethanol
concentration of 10.30%.

613 "Analysis of Fuel Ethanol Transportation Activity and Potential Distribution Constraints," ORNL, March 2009.

614 2010 Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis.

615 Impact of Biofuels on Infrastructure, Report for EPA by ICF International Inc., January 2018.

211

-------
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 program's implementation, 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.

6.4.2 Infrastructure for E85

E85 is permitted to be used only in designated flex fuel vehicles (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.616 617 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.

Figure 6.4.2-1: Light-Duty FFV Model Offerings

Source: Alternative Fuels Data Center. See "Light-Duty AFV HEV and Diesel Model Offerings by Technology-Fuel
March 2022," available in the docket.

The number of registered FFVs in the fleet is changing consistent with the reduced offerings.
While the registered FFV count continued to increase in the 2016 - 2020 timeframe, the rate of
increase has slowed.

616 "FFV registrations from AFDC December 2021," available in docket EPA-HQ-OAR-2021-0324.

617 "DOT National Transportation Statistics Table 1-11," available in docket EPA-HQ-OAR-2021-0324.

212

-------
Table 6.4.2-1: 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.

If FFV offerings remain at their 2020 levels or continue to decrease, we would expect the
number of FFVs in the fleet to begin decreasing after 2020.

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.618 As shown
below in Figure 6.4.2-2, stations offering E85 have increased steadily since about 2005. By
January of 2022, the total number of stations offering E85 had reached 4,377.

Figure 6.4.2-2: Number of Public and Private Retail Service Stations Offering E85a

5000 |
4500
^ 4000

g? 3500

1 3000

.2 2500
n

| 2. 2000
o

.o 1500
E

2 1000

0 1

o r-* » c

s s s s

S 8 8 8 8 8 § 8 i

5 22 2 o »
I 8 8 S 5

\i
-------
end of 2020, the fraction of retail stations offering E85 had reached 2.7% as seen in Figure 6.4.2-
4.

Figure 6.4.2-3: Total Number of Retail Stations in the U.S.

155,000
150,000
145,000
140,000
135,000
130,000
125,000

oo*-io

OOOOOOOOOOtHtHt—It—It—It—It—It—It—It—l(N

ooooooooooooooooooooo

(NN(N(NN(N(N(NlNN(NN(N(N(NN(NlNN(N(N

Source: Table 4.24, Transportation Energy Data Book, Edition 39.

Figure 6.4.2-4: Fraction of Retail Stations Offering E85

3.0%

2.5%

2.0%

1.5%

1.0%

0.5%

0.0%

oOrHo
OOOOOOOOOOr-It—It—It—ItHt—IrHrHT—It—l(N
OOOOOOOOOOOOOOOOOOOOO
lNlN(NlNlN(N(N(NNlN(N(NlN(N(N(NNN(NlNN

The above two factors - the small and declining fraction of vehicles capable of
consuming E85 and the low, albeit modestly growing, number of retail stations that offer E85 -
represent significant infrastructure constraints on the market's ability to deliver and use E85
through 2022. While theoretically the applicable standards under the RFS program could provide
some incentive for retail station owners to upgrade their equipment to offer E85, there is little
direct evidence that the RFS program has operated in this capacity in the past.

The BIP program was in effect from 2016 -2018 while the HBIIP program effectively
began at the beginning of 2021. While a higher growth rate in the number of E85 stations is not
readily apparent in these years compared to previous years in Figure 6.4.2-2, an analysis of
growth rates on shorter timescales suggests that these programs did have a moderate impact on
growth rates. Therefore, for purposes of making projections of future growth in E85 stations, we
applied a least-squares regression to a weighted dataset wherein each successive year was given

214

-------
greater weight than the previous year: 2021 data was given a weighting of 12, 2020 data was
given a weighting of 11, and so on back to 2010. This approached led to an average growth rate
of 178 E85 stations per year. Using this growth rate, we estimated the total number of retail
stations offering E85 as shown below.

Table 6.4.2-1: Estimated Average Number of Stations Offering E85

Year

Stations

2020

3,979

2021

4,156

2022

4,333

a Annual average, not year-end.

6.4.3 Infrastructure for E15

The CAA waiver which permitted all spark ignition highway vehicles of MY 2001 and
later to use E15 was published on January 26, 2011.619 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 El 5.

While the majority of service stations offering El5 today do so through blender pumps
which can produce El5 on demand for consumers through the combination of E10 (or E0) and
E85,620 the number of terminals offering preblended El 5 directly to service stations has been
increasing.621 The first terminals started to offer E15 in 2016, and today E15 is offered at 262
terminals, accounting for 20% for all U.S. terminals.622

The fraction of the in-use highway fleet that is MY2001 or later 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 or later vehicles (see Figure 6.4.3-1).

619 Partial Grant of Clean Air Act Waiver Application Submitted by Growth Energy To Increase the Allowable
Ethanol Content of Gasoline to 15 Percent; Decision of the Administrator, 76 FR 4662 (January 26, 2011).

620 According to Prime the Pump, 1,771 out of 2,302 stations offering E15 at the end of 2020 used blender pumps.

621 "Terminal Availability of E15 Grows as EPA Prepares to Remove RVP Barrier," available in docket EPA-HQ-
OAR-2021-0324

622 https://growthenergv.org/resonrces/retailer-hnb. See also "Retailer Hub Growth Energy 92421," available in the
docket.

215

-------
Figure 6.4.3-1: MY2001 Or Later Fraction of In-Use Vehicle Fleet and MY 2001 or Later
Fraction of In-Use Gasoline Consumption

100%

95%

90%

85%

80%

75%

70%

65%

60%

Vehicle count

55%

50%

2011 2012 2013 2014 2015 2016 2017 2018 2019

Source: Values calculated using annual retail vehicle sales of cars and trucks (Tables 4.6 and 4.7), survival rates
(Table 6.3.13), miles per year per vehicle by age (Table 6.3.12), and fuel economy by model year (Table 6.4.12)
from the Transportation Energy Data Book, Edition 38, published by Oak Ridge National Laboratory in August
2020.

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
appears that in the near term the primary constraint on the consumption of El 5 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 USDA's Biofuels Infrastructure Partnership (BIP) program
and the ethanol industry's Prime the Pump program began providing funding for station
conversions in 2016.

216

-------
Figure 6.4.3-2: 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 - Sept 2021," available in the docket.

USD A has followed up its BIP program with a similar grant program called the Higher
Blends Infrastructure Incentive Program (HBIIP) which will also provide funds to help retail
service station owners to upgrade or replace their equipment to offer biofuels. This program was
initiated in 2020. A similar program will follow HBIIP in 2022.

Based on comments received in 2016 from the Petroleum Marketers Association of
America in response to previous annual standard-setting proposals, there may also be resistance
to expanded offerings of E15 due to concerns about liability.623 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 El5 as argued in studies from the National Renewable Energy
Laboratory (NREL) and Stillwater Associates, compatibility with El5 is not the same as being
approved for El5 use.624 625 Recently-amended EPA regulations require that parties storing
ethanol in underground tanks in concentrations greater than 10% demonstrate compatibility of
their tanks with the fuel, through one of the following methods:626

• 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

623 "PMAA comments on proposed 2014 - 2016 standards rule 7-27-15," available in the docket.

624 K. Moriarty and J. Yanowitz, "E15 and Infrastructure," National Renewable Energy Laboratory, May
2015. Attachment 3 of comments submitted by the Renewable Fuels Association.

625 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.

626 "UST System Compatibility with Biofuels," available in docket EPA-HQ-OAR-2021-0324.

217

-------
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 El 5. 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 were
designed to address potential misfueling. These regulations require pump labeling, a misfueling
mitigation plan, surveys, product transfer documents, and approval of equipment configurations,
providing consumers with the information needed to avoid misfueling.627 In addition, the portion
of vehicles not designed and/or approved for El 5 use continues to decline. Vehicles
manufactured before EPA's threshold of 2001 represent less than 10% of the vehicle stock, and
just slightly over 5% of miles travelled. Vehicles designed and warranted by manufacturers to be
fueled on E15 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 2022. While theoretically the applicable
standards under the RFS program could 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 estimate the number of retail stations that may offer El 5 for 2020 - 2022, we
first separated the effects of USD As 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
program was responsible for the conversion of 841 retail stations.628'629 This occurred over the
years 2016 through 2018, during which time E15 stations were also increasing as a result of
other efforts, bringing the total number of E15 stations to 1,630 as shown in Figure 6.4.3-2. Of
this total, 841 are estimated to have been the result of the BIP program while the remaining 789
E15 stations were the result of other efforts. The HBIIP program effectively began in 2021.630
Given its similarity to the BIP program, we have assumed that it would likewise take 3 years to
complete and would result in 841 new E15 stations. From these estimates of the impacts of the
BIP and HBIIP programs we could back-calculate the growth in El5 stations that can be
attributed to private initiatives including Prime the Pump.

627 See 40 CFR Part 80, Subpart N: Sections 1500 to 1509.

628 "Biofuel Infrastructure Partnership - original grants & projections," available in the docket.

629 "Communication with USDA on the BIP program 11-15-21," available in the docket.

630 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.

218

-------
Table 6.4.3-1: Hisl

orical Breakdown of E15 Stations

Total3

BIPb

HBIIPC

PTP + private efforts

December 2012

December 2013

December 2014

105

December 2015

184

December 2016

431

183

248

December 2017

1122

563

559

December 2018

1630

841

789

December 2019

2045

841

1,204

July 2020

2208

841

1,367

December 2020

2302

841

1,461

September 2021

2536

841

1,620

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. We therefore used a least-squares regression to estimate this trend. The result is shown
below.

Figure 6.4.3-3: Growth in E15 Stations Due to Efforts Other Than the BIP Program

3,500
3,000
2,500
2,000
1,500

Least-squares regression:

Stations = 298.286x Year- 601238

3 S
s s

Using the available information on the USDA grant programs and the projection shown
in Figure 6.4.3-3, we were able to estimate the breakdown of E15 stations for the 2020 - 2022
time period addressed by this rule as shown below.

219

-------
Figure 6.4.3-4: Projected Breakdown of E15 Stations

riO riO aO rft riO nO o/S* ryV A*V rvV nN rQs aV aV rCV aV aV

Krsy jy .«ov ~ov v0ov ~ov ^ jy jy ~ov jS* ~ov ^ ^
AV- a\ A n4 a\ a\ .<,V a\ .<,\^ ,A a\ a\ n4 AV AV a\

J? of if AV o,\^ jf J? of if AV cf^f •$? of
-------
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.

Given the overall minimal impact on transportation infrastructure from the distribution of
biofuels and the ability of the system to resolve localized instances of increased stress on the
system in a timely fashion, we believe that the volumes under consideration in this final rule will
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 volume requirements difficult to achieve. For
instance, as described in Chapter 6.1 above, biogas for CNG/LNG vehicles will be delivered
through the same pipeline network used to distribute natural gas. Since that biogas will be
displacing natural gas used in CNG/LNG vehicles, we would not expect a net increase in total
volume of biogas + natural gas delivered.

The largest volumes of renewable fuel increases in 2022 will be for corn ethanol and
soybean oil renewable diesel. However, although corn ethanol will increase in 2022 under our
final volume requirements as shown in Table 2.3-3, the total volume consumed in both years will
remain below that consumed in 2019. Since the corn collection and distribution network
functioned without difficulty in 2019, there is no reason to believe that it would not function
similarly in 2022.

We estimate that fats, oils, and greases (FOG) for the production of renewable diesel in
2022 will increase by about 80 million gallons in comparison to 2020 (see Table 2.2-3),
equivalent to about 10.2%.631 The increase in FOG use from 2021 to 2022, which is a more
relevant comparison for the purposes of analyzing the ability of the market to collect and
transport additional FOG for BBD production above what has already occurred, is 40 million
gallons. This increase is a very small fraction of the total amount of FOG produced. According to
LMC International, the total production volume of all animal fats and used cooking oil was about
2.7 billion gallons in 2020.632 An increase of 40 million gallons represents only 1.5% of total
2020 production. 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 2022.

631 Total domestic consumption of FOG for domestic biodiesel and renewable diesel production was 784 million
gallons in 2020. See "Biodiesel and Renewable Diesel Feedstocks (2014-2021)," available in the docket.

632 "The Outlook for Increased Availability & Supply of Sustainable Lipid Feedstocks in the U.S. to 2025,"
attachment to comments submitted by Clean Fuels Alliance America. For 2020, Diagram 10 indicates about 7.55
mill tonnes of total animal fats was produced, while Diagram 11 indicates that about 1.85 mill tonnes of UCO was
produced. Conversion factor of 7.66 lb/gal leads to a total of 2.7 bill gal.

221

-------
As shown in Table 2.3-3, total soybean oil use for the production of BBD is projected to
increase by 949 million gallons in 2022 in comparison to 2020. While we expect a portion of this
soybean oil to be diverted to BBD production from existing uses, we believe that a significant
portion will be due to increased soybean oil production and greater imports of soybean oil and/or
biofuels produced from soybean oil. Such an increase in soybean oil production could be the
result of an increase in the fraction of soybeans crushed with no increase in soybean production.

However, in order to evaluate the ability of existing infrastructure to collect and deliver
sufficient feedstocks to produce an additional 949 million gallons of BBD in 2022, we analyzed
an extreme scenario wherine all 949 million gallons of soybean oil was derived from more
soybean production. In this case, about 673 million additional bushels of soybeans would be
needed over two years.633 In comparison, over the time period 2012 - 2018 when the growth in
soybean production was relatively consistent, domestic production increased at an average rate
of about 237 million bushels per year, or about 474 million bushels every two years.

Figure 6.5-1: Historical Production of Soybeans

5,000,000
4,500,000
4,000,000

-{= 3,500,000

-Q

-o 3,000,000

o 2,500,000

-C

I-

2,000,000
1,500,000
1,000,000

Source: USDA's Economic Research Service. See "Oil Crops Yearbook 2021," available in the docket.

The 673 million bushels that would be needed over two years is considerably higher than
the average two-year growth rate of 474 million bushels. However, the market has been able to
mobilize the needed collection and distribution infrastructure for considerably larger increases in
soybean production in the past. For instance, in 2014 soybean production increased by 571
million bushels in comparison to 2013. This is a larger annual increase than would be needed
even in the extreme scenario where the entire increase in BBD production was supplied by
increased planting of soybeans (337 million bushels/year). Also, as with FOG, the increase in
soybean oil use from 2021 to 2022 is a more relevant comparison for the purposes of analyzing
the ability of the market to collect and transport additional soybeans for BBD production above
what has already occurred. From Table 2.3-3, the increase in soybean oil consumption from 2021
to 2022 is projected to be 725 million gallons, equivalent to 514 million bushels of soybeans in
the extreme case where the entire volume is produced from additional soybean production. This
is less than the single-year increase in soybean production from 2013 to 2014. Since the market
has demonstrated that it can collect and transport 571 million additional bushels in one year, we
believe that it could also do so for 514 million additional bushels in one year.

633 Assumes 60 pounds per bushel of soybeans, 7.66 pounds of oil per gallon of oil, and that soybeans are 18% oil by
weight.

222

-------
Chapter 7: Other Factors

The statute directs EPA to consider the impact of the use of renewable fuels on "other"
factors, including job creation, the price and supply of agricultural commodities, rural economic
development, and food prices. This chapter addresses the enumerated "other" factors.634 We
focus our analysis on the biofuels where we are projecting the largest relative changes from the
baseline year of 2020, namely, corn ethanol, renewable diesel made from soybean oil, and
biogas. While we also expect increases in renewable diesel from palm, we expect that the palm
will be produced in foreign countries and thus will have little impact on the domestic rural
economy. The impacts evaluated in this chapter are for volume increases for 2021 and 2022
compared to 2020 baseline volumes that are described in Chapter 2. However, if we consider the
causal impacts of this rule (as opposed to the volume changes in Chapter 2), the 2020 and 2021
volumes will have no impact, since these volumes will be entirely retroactive. The 2022 volume
requirements could have some impact largely due to their potential ability to incent greater
production of biofuels.

7.1 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, namely biofuel production and agriculture. We acknowledge
that changes in indirect employment (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 our analysis. We 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.

7.1.1 Fuel Production

This subsection provides analysis and discussion on employment aspects of renewable
fuel production as they relate to the volumes projected in this rulemaking.

7.1.1.1 Ethanol Production

Urbanchuk estimates that the total number of direct, full-time-equivalent jobs for
domestic corn ethanol production in 2021 was 8,942 across the 208 plants that the Renewable
Fuels Association found to be operating that year.635 The total nameplate capacity of those plants

634 As we explain in Preamble Section II, we also consider several other factors besides those enumerated in the
statute.

635 From "Contribution of the Ethanol Industry to the Economy of the United States in 2021," prepared for the
Renewable Fuels Association by JohnM. Urbanchuk, ABF Economics. February 3, 2022.

223

-------
is reported at 17.7 billion gallons, suggesting an average plant size of 85 million gallons per year,
and an average employee concentration of 0.51 jobs per million gallons capacity.

The U.S. Energy Information Administration (EIA) Annual Fuel Ethanol Production
Capacity Report provides plant count and total nameplate capacity values for historical calendar
years. The latest data currently available is from January 2021, where they show 17,546 million
gallons (MMgal) produced by 197 plants that reported themselves as operational.636 The average
plant size using these figures is 89 million gallons per year, which gives the same average
employee concentration of 0.51 jobs per million gallons capacity using Urbanchuk's total direct
employment.

In 2018, we were able to collect data on the number of employees at each of 65 corn
ethanol facilities through the website of Ethanol Producer magazine.637 A table of this
information is provided in the docket. The plant capacities provided by Ethanol Producer
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 7.1.1.1-1 shows this data fit with an exponential trendline, and includes the
Urbanchuk-based estimate of employee concentration of 0.51 plotted at the national average
facility size of 85 million gallons per year. The Urbanchuk value shows good agreement with the
correlation fit line.

636 U.S. EIA. U.S. Nameplate Fuel Ethanol Production Capacity, January 2021. As of March 2022, this is currently
the latest data available, https://www.eia.gov/petroleum/ethanolcapacitv/archive/2021/index.php

637 Ethanol plant employment data obtained via Ethanol Producer Magazine website. Current issue and archives
available at http://www.ethanolprodncer.com

224

-------
Figure 7.1.1.1-1: Correlation between employee concentration and facility size for corn-
ethanol facilities

2.0
1.8

ro 1.5
e?

.2 1.3

r i.o

-------
We are also projecting significant growth of CNG/LNG derived from biogas in the RFS
program, by 58 and 127 million additional RINs for 2021 and 2022, respectively. 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 is expected to cause additional employment related to upgrading and maintenance
of gas cleanup and pipeline interconnect equipment.638 We also project that the construction and
operation of new agricultural digesters and digesters at wastewater treatment facilities will result
in additional employment.

In comments on the proposed rule, the Coalition for Renewable Natural Gas submitted an
analysis of 2021 data showing that on average 24 jobs were created for every 1 million ethanol-
equivalent gallons increase in production of CNG/LNG derived from biogas.639 The data
indicated that 30% of these jobs were related to the operation of these facilities and 70% of the
jobs were related to construction.639 Applying these numbers to the projected volume increases of
CNG/LNG derived from biogas in 2021 and 2022 results in 1392 additional jobs related to the
production of CNG/LNG from biogas in 2021 (418 jobs in operations and 974 jobs in
construction) and 3048 jobs in 2022 (914 jobs in operations and 2134 jobs in construction).

These numbers are based on the average facility producing CNG/LNG derived from biogas in
2021, and the employment estimates therefore implicitly assume that the average employment at
facilities in 2021 and 2022 are equal to the average employment at facilities in 2021. The actual
employment impacts for 2022 may be slightly higher or lower depending on the types of new
facilities that produce CNG/LNG derived from biogas (e.g., landfills, wastewater treatment
facilities, or agricultural digesters) and the sizes of these facilities.

7.1.2 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 CNG/LNG derived from biogas is produced from waste or
byproduct materials such as separated MSW, wastewater, and agricultural residue we expect the
projected increases in the production of CNG/LNG derived from biogas to have very little
impact on employment related to feedstock production. As noted above, we are projecting
significant growth in ethanol and renewable diesel production for 2021 and 2022 (relative to the
2020 baseline), with nearly all this feedstock being corn and soybean oil respectively. This
excludes imported renewable diesel, which we expect will be made outside the U.S. from palm
oil or other feedstocks.

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 couple of decades, with no
pretense of ascribing causation for observed trends to particular volumes of renewable fuels.

638 Jaramillo and Matthews, Environmental Science & Technology 2005 39 (19), 7365-7373.

639 Economic Analysis of the U.S. Renewable Natural Gas Industry, December 2021. Document submitted in
comments from the Coalition for Renewable Natural Gas. (EPA-HQ-OAR-2021-0324-0485).

226

-------
Some of the most consistently sourced data available on hired farmworkers is made
available by the National Agricultural Statistics Service (NASS).640 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 7.1.2-1
below.

Figure 7.1.2-1: Harvest Season Farm Workers, 2003-2021

950

900

to
JD

o 850

~u
c
ro

*3 800
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 and 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.

An additional set of data on agricultural employment is collected by the United States
Department of Agriculture (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 dataset provides additional insight. Figure
7.1.2-2 presents the data on employment in agriculture and support activities for 2001-2020.641

6411 USDA NASS Farm Labor reports, available at https://usda.librarv.cornell.edu/concern/publications/x920fw89s

andhttps://usda.librarv.cornel!edu/concern/publications/i3860694x?locale=en

641 USDA ERS Farm Labor webpage, March 2022. Available at https://www.ers.usda.gov/topics/farm-

economv/farm-labor

227

-------
Figure 7.1.2-2: Employment in Agriculture and Support Activities, 2001-2020

U.S. employment in agriculture and support industries, 2001-20

Thousand jobs

1,400 -

1,200-

Livestock support

1,000 -

Livestock ___

800-

600-

Crop support

400 -

Crops

200-

n -

i i i i i i i i i i i i i i i i i i

2001 I 2003 I 2005 I 2007 I 2009 I 2011 I 2013 I 2015 I 2017 I 2019

2002 2004 2006 2008 2010 2012 2014 2016 2018 2020

Note: Employment is measured as the annual average number of full- and part-time jobs.

Data do not cover smaller farm employers in those States that exempt them from

participation in the unemployment insurance system.

Source: USDA, Economic Research Service using data from U.S. Department of Labor,

| Quarterly Census of Employment and Wages, as of June 2, 2021.

This data from USI)A shows that employment in crop production and crop support
activities have increased about 3% and 20%, respectively, over the past decade. As with the
previous dataset presented in Figure 7.1.2-1, the lack of crop-specific data makes drawing
associations with biofuel production very difficult. We see that the lowest employment levels
reported in the USD A data for crop production workers coincide with the 2008-2009 recession
and that it wasn't until 2015 that the number of such jobs returned to the pre-recession levels.
Looking at this dataset, it is difficult to see any clear impact of increased renewable fuel
production among broader economy-wide factors.

7.2 Rural Economic Development

Changes in biofuel production can have economic development impacts on rural
communities and financial impacts on farmers. We have seen significant growth in renewable
diesel volumes for 2021 and are projecting significant additional growth for 2022, with the vast
majority of this feedstock coming from soybean oil. Some of this oil is expected to come from
additional soybean production and crushing, which may bring some revenue increases to rural
communities.

We are also projecting modest growth in ethanol volumes, though this is measured from a
2020 baseline. Thus, much of this volume is getting back to consumption levels before the

228

-------
economic contraction associated with the COVID-19 pandemic, and therefore could be
considered more as economic restoration rather than development.

The increased production of CNG/LNG derived from biogas is expected to result in
additional rural economic activity. Using factors derived from an analysis in comments by the
Coalition for Renewable Natural Gas, the 2021 and 2022 volumes in this rule would be
associated with economic activity of $35 million in 2021 and $76 million in 2022 related to gas
upgrading and facility administration activities.639 That analysis also indicates that 76% of
facilities under construction in 2021 were agricultural waste digesters, suggesting $84 million in
additional rural economic activity in 2021 and 2022.

7.3 Supply of Agricultural Commodities

Changes in biofuel production can have an impact on the supply of agricultural
commodities. As discussed above, we are projecting modest growth in corn ethanol volumes for
2021-2022. Growth in renewable diesel volumes will be more substantial, with additional
feedstock likely coming from additional soybean crushing as well as diversion from other
domestic uses. As we discuss in Chapter 3, these volume increases suggest the potential for
associated increases in underlying crop production; however, the magnitude of the potential
impact cannot be estimated at this time. Biogas is not produced from agricultural commodities
and therefore is not expected to affect their supply or price.

For historical context, Figure 7.3-1 shows trends in corn production and uses between
1995-2021.642 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.

642 USDA ERS Feed Grains Data Yearbook, March 2022 (Tables 4 and 31).

229

-------
Figure 7.3-1: Corn Production and Usage for Years 1995-2021
10 |

Ethanol Food

Feed Exports

Total Corn /

1995 2000 2005 2010 2015 2020

Between 2005-2010, additional corn required to satisfy increasing ethanol production
was apparently largely filled by diversion from animal feed in the short term until overall
production caught up. Supply of corn to food uses continued to grow steadily during this period,
despite increased consumption as ethanol feedstock. Exports also remained steady, except for a
drop in exports corresponding to weather-related supply disruptions and elevated prices in 2011-
2012. Animal feed use began to rebound after 2014 when growth in ethanol production slowed
and prices stabilized. Another factor contributing to the longer-term shift of animal feed away
from whole corn was the increasing substitution with dried distillers' grains (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 modest increases in ethanol
volumes associated with this rulemaking are likely to have no or only a small short-term impact
on the supply of corn to food, exports, or other uses.

Figure 7.3-2 shows that soybean production has risen steadily over time, similar to the
trend for corn production.643 Roughly 80% of this growth since 2005 is associated with rising
exports of soybeans, which have nearly doubled over that period. Domestic crushing of beans
has grown by 24% since 2005, which mirrors the same relative growth in production of the crush
products, soy meal and oil (shown in Figure 7.4-2). As shown in Figure 7.3-3, exports of soy
meal also nearly doubled during this time, which together with the growth in whole bean exports,
presents a picture consistent with expansion of meat production internationally.644 (Worldwide,
over 95% of beans are eventually crushed for meal and oil.)

643 USD A ERS Oilcrops Data Yearbook, Soy Tables, March 2022.

644 USD A ERS Oilcrops Data Yearbook, Soy Tables, March 2022.

230

-------
Figure 7.3-2: Soybean Production and Usage for Years 1980-2021

Figure 7.3-3: Soy Meal Production and Usage for Years 1980-2021

total prod

domestic use

export

1980 1985 1990 1995 2000 2005 2010 2015 2020

Overall, these data indicate that biofuel production represents a major use of domestic
soy oil. Soybean oil production has generally increased since 2005 with much of this increase
being used for biodiesel production. This increase in soybean oil production has been due to both
increasing domestic soybean crush and increasing yields of soybean oil per bushel of soybeans
crushed. Increases in soybean exports have also increased, at a much faster rate than the increase
in soybean crushing. In terms of overall soybean production, the primary driver for growth over

231

-------
the past 15 years in soybean production and planted acres has clearly been rising exports, with
crushing of beans for meal and oil being a distant second.

7.4 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. Drought and
storm damage to crops may strike entire regions with only hours' notice and have a major effect
on harvest yields. Renewable fuels are only one factor in determining their 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, about 15% of the previous year's overall corn
production is typically still in storage at the time of the new harvest.645 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 that induces more
corn planting in the upcoming season. Work done by Informa Economics for the Renewable
Fuels Association in 2016 examined the historical relationship between corn usage, stocks, and
futures prices.646 Figure 7.4-1 illustrates 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.

645 USDA ERS Feed Grains Data Yearbook, March 2022 (Table 4).

646 Informa Economics IEG. The Impact of Ethanol Industry Expansion on Food Prices: A Retrospective Analysis.
November 2016. Available at https://ethanolrfa.org/impact-of-ethanol-expansion-on-food-ps_informa_2016/

232

-------
Figure 7.4-1: Corn Ending Stocks / Use Ratio versus Futures Price

$8.00

| $7.00
a.

| $6.00
£ $5.00

2012
• .2011
*2010

FF = 0.8255 \

• 2007

a $4.00

1 $3.00
<

| $2.00

g $1.00
0

• 2013

2008

. •'%H#009
2006 2015 -...

'"¦-2005

2004

3% 8% 13% 18%
USCorn Ending Stocks to 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 consider 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.647 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 (not something unforeseen like weather shocks). A similar meta-
analysis was done in 2016 by FAPRI-Missouri that considered several newer studies.648 This
paper found $0.19 per bushel per billion gallons or $0.15 if a supply response is included, a
figure that is consistent with the 3% impact above.

With biodiesel and renewable diesel production, the commodity input of interest is soy
oil, which has an indirect link to bean 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 cnish margin. Thus, the degree of pass-through of oil price

6'r Condon, Nicole, Klemick, Heather and Wolverton, Ann, (2013), Impacts of Ethanol Policy on Com Prices: A
Review and Meta-Analysis of Recent Evidence. No 201305. NCEE Working Paper Series, National Center for
Environmental Economics, U.S. Environmental Protection Agency,
https://EconPapere.repec.Org/RePEc:nev:wpaper:wp201305.

6:48 Food and Agricultural Policy Research Institute. Literature Review of Estimated Market Effects of U.S. Com
Starch Ethanol, 2016. FAPRI-MU Report #01-16.

233

-------
increases to bean prices, which may then influence acres planted, is not straightforward. Figure
7.4-2 shows historical trends in soy oil prices alongside allocation to biofuel and other uses,
based on data taken from the USDA Oilcrops Yearbook.649 Use in domestic biofuel rose from
0.8 million tons in 2005 to over 4.4 million tons in 2020. Other domestic uses also increased
steadily through 2005, decreased slightly from 2005-2010, and have remained relatively
consistent since 2010. Exports of soy oil play a minor role and have remained fairly consistent
over the past decade. Noting the lack of correlation between soy oil price and its use in biofuel
production historically, we conclude that the price of soybean oil is influenced by a number of
factors occurring 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, we believe the current high prices are largely the result of
poor weather conditions in South America and Malaysia over the past year.650

Figure 7.4-2: Soy oil price and allocation to biofuel and exports, 1980-2021

70
60

50 u*

ro
c

40 E
o
c

30 8
\_

20
10
0

There are relatively few quantitative studies on the impacts of biodiesel production on
soy oil and bean prices, and they show a wide range of results. This is in part because these
studies have used a variety of different policy combinations, none of which separated out just the
impact of the RFS program on biomass-based diesel demand. Ethanol demand could impact the
soybean markets even in the absence of increased demand for biomass-based diesel from the
RFS program due to increased competition for cropland and other inputs. The largest impacts are
estimated when the biomass-based diesel 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

649 USDA ERS Oilcrops Data Yearbook, Soy Tables, March 2022.

6511 Wilson, Nick. Oil Prices Surge - Vegetable Oil That Is. Marketplace.org. February 17, 2022.

234

-------
obligation. The range of soybean price impacts indicated by these studies is 1.8% to 6.5% per
billion gallons of biomass-based diesel, from which we derive a single central value of 4.2%.651'
652'653 We have used this value to estimate the price impact on soybean meal. However, as
discussed further below we recognize that this may over-estimate the impact of this rule on
soybean meal. Considering data presented here and in 7.3 above, we see the primary impacts of
increases in renewable fuel production have historically been short-term price increases for both
corn and soybeans, which have evened out after a period of market and production adjustments.

To project the impact on crude soybean oil prices, we used a value of 16 cents per pound
of oil per billion gallons of biomass-based diesel produced from soybean oil. This figure was
derived from modeling work published by Babcock, et al., and is the same figure used for other
cost estimates in this rule.654 Analysis published by Irwin at the University of Illinois indicates
that soy oil prices often move separately from meal and bean prices, and that the latter two are
closely correlated.655 From this we made the simplifying assumption that the price impact on soy
meal would match the 4.2% per billion gallons described above for beans. Table 7.4-1
summarizes these price impacts for corn, soy oil, and soy meal for the finalized volumes in 2021
and 2022.

651 Babcock, B. A. 2012. The impact of US biofuel policies on agricultural price levels and volatility. China
Agricultural Economic Review 4:407-426.

652 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.

653 U.S. EPA. 2010. Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis. Assessment and
Standards Division. EPA Office of Transportation and Air Quality.

654 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, Ames,
Iowa.

655 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.

235

-------
Table 7.4-1: Projected Impact of

the 2021-2022 Volumes on Commodity Prices

Corn

Soybean Oil

Soybean Meal

Percent increase per billion
gallons of biofuel

N/A

4.2%

Impact of the 2021 Volumes Relative to Baseline

Commodity Pricea

$4.21 per bushel

$0.50 per pound

$369 per short ton

Price increase per billion gallons
of biofuel

$0.13 per bushel

$0.16 per pound

$16 per short ton

Volume Increase in 2021
(million gallons)13

1,294

224

Price increase in 2021

$0.16 per bushel

$0.04 per pound

$3.47 per short ton

Impact of the 2022 Volumes Relative to Baseline

Commodity Pricea

$4.21 per bushel

$0.50 per pound

$369 per short ton

Price increase per billion gallons
of biofuel

$0.13 per bushel

$0.16 per pound

$16 per short ton

Volume Increase in 2022
(million gallons) b

1,680

949

Price increase in 2022

$0.21 per bushel

$0.15 per pound

$14.71 per short
ton

a Commodity prices are from the January 2022 WASDE. Prices represent the average price in the 2020 calendar
year. For corn, the price is calculated using 1/3 of the price for the 2019/20 agricultural marketing year and 2/3 of
the price for the 2020/21 agricultural marketing year. For soybean oil and soybean meal the prices are calculated

using 1/4 of the 2019/20 agricultural marketing year and 3/4 of the price for the 2020/21 agricultural marketing year.
b Volume increases are derived from table 2.2-3. The soybean oil and meal volume increases are the sum of volume
increases for soybean oil biodiesel and renewable diesel.

7.5 Food Prices

The above impacts on corn and soy commodity prices may in turn have a ripple impact
on food prices and the many other products produced from these commodities. Since the volume
changes would have a relatively small impact on the overall world commodity markets and since
the cost of the commodity itself tends to be a relatively small component in the cost of food, the
impacts are likewise relatively modest. Further, the impact of the volume changes 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).656

To project the impact of the renewable fuel volumes on food prices we used a
methodology developed in conjunction with USDA in assessing requests from the Governors of

656 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 statute
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.

236

-------
several states to reduce the 2012 RFS volumes using the general waiver authority.657 This
methodology generally uses estimates of the impact of biofuel volumes on commodity prices
(corn, soybean oil, and soybean meal) to calculate the estimated impacts on total food
expenditures. For context this estimated change in food expenditures is 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.

The previous section presented a discussion of literature estimates of changes in biofuel
volumes on commodity prices. These estimates are the starting point for our 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 7.5-1. 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.658 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.659

We recognize that projecting that the prices of distillers grains and soybean meal increase
proportionally to the prices of corn and soybeans, respectively, 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 and soybean meal as these are co-
products of biofuel production and are not converted to biofuels. Thus, while biofuel production
may increase the prices of corn, soybeans, and food produced from corn and soybeans, it may
not increase the price of the co-products from the biofuel production and food produced from
these co-products. 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 and soybean oil for
biofuel production will impact the price of distillers grains and soybean oil. If the prices for these
commodities increase less than the price of corn and soybeans (or if they decrease) in response to
increased demand for biofuels we would expect a smaller impact on food prices than what we
have estimated from the renewable fuel volumes we are finalizing in this rule.

657 77 FR 70752 (November 27, 2012).

658 Other commodity price impacts were estimated as follows: change in the price of grain sorghum = 0.93 times the
change in the price of corn; change in the price of barley = 0.88 times the change in the price of corn; change in the
price of oats = 0.72 times the change in the price of corn; change in the price of distillers grains = 1/56 times the
change in the price of corn. These factors are from 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 in EPA-HQ-
OAR-2021-0324.

659 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.

237

-------
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.660 We also note that
the estimates on food price impacts presented here are relative to the volumes of renewable fuel
consumed in 2020, consistent with the other analyses in this RIA. The food price impacts
estimated in this rule for 2022 are likely reflected in current food prices, and this final rule is
unlikely to further increase food prices.661

660 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, U.S. Department of Agriculture, Economic Research Service,

August 2012.

661 As discussed further in RTC Section 6.3.4, renewable fuel use in the U.S. in the first three months of 2022
(January - March) is consistent with the rates needed to meet the volumes we are finalizing in this rule for 2022
based on observed seasonal trends in 2021. Thus, the market appears to have anticipated the volumes are finalizing
in this rule, which are very similar to proposed volumes. Any increases in demand for agricultural commodities that
result from this rule are therefore likely already reflected in the current market prices for these commodities and are
not expected to change significantly with the finalization of this rule.

238

-------
Table 7.5-1: Changes in Food Expenditures (Relative to 2020)

Commodity Price
Change

Quantity Used for
Food and Feed3

Change in
Expenditures

Changes in Foot

Expenditures in 2021

Corn

$0.16 per bushel

7,040 million bushels

$1150 million

Grain Sorghum

$0.15 per bushel

99 million bushels

$15 million

Barley

$0.14 per bushel

173 million bushels

$25 million

Oats

$0.12 per bushel

146 million bushels

$17 million

Soybean Oil

$0.04 per pound

14.5 billion pounds

$519 million

Soybean Meal

$3.47 per short ton

37.6 million short tons

$130 million

Distillers Grains

$5.83 per short ton

42.7 million short tons

$249 million

Total

N/A

$2,105 million

Changes in Foot

Expenditures in 2022

Corn

$0.21 per bushel

7,085 million bushels

$1,502 million

Grain Sorghum

$0.20 per bushel

125 million bushels

$25 million

Barley

$0.19 per bushel

130 million bushels

$24 million

Oats

$0.15 per bushel

125 million bushels

$19 million

Soybean Oil

$0.15 per pound

14.2 billion pounds

$2,149 million

Soybean Meal

$14.71 per short ton

37.9 million short tons

$558 million

Distillers Grains

$7.57 per short ton

45.3 million short tons

$343 million

Total

N/A

$4,619 million

a Quantity used for food and feed calculated based on the January 2022 WASDE Projections. Quantities listed for
2021 are based on the 2020/2021 agricultural marketing year and quantities listed for 2022 are based on the
2021/2022 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.

The estimate of the impact of this rule on food prices is higher than the estimate from the
proposed rule. There are two primary reasons for this increase. The first is that we are projecting
higher volumes of corn ethanol will be used in 2021 and 2022 than in the proposed rule. These
higher volumes increase demand for corn in these years, and thus lead to higher projected corn
prices. The second main reason for the higher estimate of food price impacts is that USDA's
projections for the prices of agricultural commodities have increased since the proposed rule.
Because the estimated impacts on the price of corn and soybean meal relative to the 2020
volumes are percentage increases, higher commodity prices mean that the same percentage
increase results in a larger price impact per bushel of corn and ton of soybean meal. We note that
the food price impacts estimated here are for the volume increases for 2021 and 2022 compared
to 2020 that are described in Chapter 2. However, if we consider the causal impacts of this rule
(as opposed to the volume changes in Chapter 2), the 2020 and 2021 volumes will have no
impact, since these volumes will be entirely retroactive. The 2022 volume requirements could
have some impact largely due to their potential ability to incent greater production of biofuels in
2022.

239

-------
Finally, we compared the estimated change in food expenditures to total food
expenditures as reported by the Bureau of Labor and Statistics in their 2019 survey.662 We used
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 income quintile. These estimates are shown in Tables 7.5-2 and 7.5-3.

Table 7.5-2: Percent Change in Food Expenditures (Relative to 202C

2021 Estimate

2022 Estimate

Number of Consumer Units (thousands)

131,234

Food Expenditures per Consumer Unit

$7,316

Total Food Expenditures

$960 billion

Change in Food Expenditures

$2,105 million

$4,619 million

Percent Change in Food Expenditures

0.22%

0.48%

Table 7.5-3: Change in Food Expenditures per Consumer Unit (Relative to 2020)

2021

2022

All Consumer Units

Food Expenditures

$7,316

Percent Impact on Food Expenditures

0.22%

0.48%

Projected Food Expenditure Increase

$16.04

$35.20

Lowest Quintile

ncome Consumer Units

Food Expenditures

$4,099

Percent Impact on Food Expenditures

0.22%

0.48%

Projected Food Expenditure Increase

$8.99

$19.72

662 Bureau of Labor and Statistics Consumer Expenditure Survey, 2020: Table 1, Quintiles of income before taxes:
Annual expenditure means, shares, standard errors, and coefficients of variation, 2020.

240

-------
Chapter 8: Environmental Justice

Executive Order 12898 (59 FR 7629; February 16, 1994) establishes federal executive
policy on environmental justice. Its main provision directs federal agencies, to the greatest extent
practicable and permitted by law, to make environmental justice 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 minority populations and low-
income populations in the United States. EPA defines environmental justice 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.663 Executive Order 14008 (86 FR 7619; February 1, 2021) also calls on
federal agencies to make achieving environmental justice 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 its "Technical Guidance for Assessing Environmental
Justice in Regulatory Analysis" (U.S. EPA, 2016) 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 minority populations, low-income populations,
tribes, and/or indigenous peoples, EPA strives to answer three broad questions: (1) Is there
evidence of potential environmental justice (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? And, (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.

663 Fair treatment occurs when "no group of people should bear a disproportionate burden of environmental harms
and risks, including those resulting from the negative environmental consequences of industrial, governmental, and
commercial operations or programs and policies" (U.S. EPA, 2011). Meaningful involvement occurs when "1)
potentially affected populations have an appropriate opportunity to participate in decisions about a proposed activity
[i.e., rulemaking] that will affect their environment and/or health; 2) the population's contribution can influence
[EPA's] rulemaking decisions; 3) the concerns of all participants involved will be considered in the decision-making
process; and 4) [EPA will] seek out and facilitate the involvement of population's potentially affected by EPA's
rulemaking process" (U.S. EPA, 2015). A potential EJ concern is defined as "actual or potential lack of fair
treatment or meaningful involvement of minority populations, low-income populations, tribes, and indigenous
peoples in the development, implementation and enforcement of environmental laws, regulations and policies" (U.S.
EPA, 2015). See also fattps ://www.epa.gov/enviroiimentatiiistiee.

241

-------
EPA's 2016 Technical Guidance does not prescribe or recommend a specific approach or
methodology for conducting an environmental justice 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, the Agency
endeavors to conduct such an analysis. Going forward, EPA is committed to conducting
environmental justice analysis for rulemakings based on a framework similar to what is outlined
in EPA's Technical Guidance, in addition to investigating ways to further weave environmental
justice into the fabric of the rulemaking process.

8.1 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" (hereafter "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 lifestages, 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
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),664'665 the Intergovernmental Panel on Climate Change

664 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. Kuiikel, K.L.M. Lewis, T.K. May cock, and B.C.
Stewart (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, 1515 pp. doi: 10.7930/NCA4.2018.
665: 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

242

-------
(IPCC),666'667'668'669 and the National Academies of Science, Engineering, and Medicine670'671
add more evidence that the impacts of climate change raise potential environmental justice
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 United States. In particular,
the 2016 scientific assessment on the Impacts of Climate Change on Human Health2 found with
high confidence that vulnerabilities are place- and time-specific, that particular lifestages 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.

8.1.1 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 lifestages, specifically women who are pre- and perinatal, or are

666 Oppenheimer, M., M. Campos, R.Warren, J. Birkmann, G. Luber, B. O'Neill, and K. 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.

667 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.

668 Smith, K.R., A.Woodward, D. Campbell-Lendrum, D.D. Chadee, Y. Honda, Q. Liu, J.M. Olwoch, B. Revich,
and R. 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.

669 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.

670 National Research Council. 2011. America's Climate Choices. Washington, DC: The National Academies Press.
https://doi.org/10.17226/12781

671 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

243

-------
nursing; in utero fetuses; children at all stages of development; and the elderly. Per the Fourth
National Climate Assessment, "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."672 Many health conditions such as cardiopulmonary or respiratory illness and
other health impacts are associated with and exacerbated by an increase in greenhouse gases 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 environmental justice ("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 Impacts of Climate Change on Human Health, 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 detailed findings regarding related vulnerabilities and the
projected impacts youth may experience. These assessments - including the Fourth National
Climate Assessment (2018) and The Impacts of Climate Change on Human Health in the United
States (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

672 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

244

-------
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.673 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.674 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.

The Impacts of Climate Change on Human Health (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 indicates that losses of customs and historical
knowledge may cause communities to be less resilient or adaptable.675 The Fourth National
Climate Assessment (2018) 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 economies.676 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

673 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.

674 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.

675 Porter et al., 2014: Food security and food production systems.

676 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

245

-------
Confederated Tribes of the Umatilla Indian Reservation in the Northwest have identified climate
risks to salmon, elk, deer, roots, and huckleberry habitat. 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 (e.g., disorientation, heightened exposure to PM2.5, etc.) 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 IPCC AR5677 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, the NCA 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 the
NCA 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.

8.1.2 Potential Future Impacts Associated with the Annual Volume
Requirements

As described in Chapter 3.2, the biofuel volumes associated with this rulemaking may
help mitigate the impacts of climate change by potentially reducing GHG emissions. Future
impacts of climate change are still expected and will likely be unevenly distributed in ways that
uniquely impact communities with EJ concerns.

8.2 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

677 Porter et al., 2014: Food security and food production systems.

246

-------
population.678'679'680'681 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.682

Emissions of non-GHG pollutants associated with this rule, including, for example, PM,
NOx, CO, SO2 and air toxics, occur during the production, storage, transport, distribution, and
combustion of petroleum-based fuels and biofuels. Communities with EJ concerns may be
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.683 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
socioeconomic status.684'685'686 For this rule, EPA has not quantitatively assessed the
demographic characteristics of populations living 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 depend on a range of interacting and complex factors including the
amount of pollutant emitted, atmospheric chemistry and meteorology.

The manner in which biofuel producers and markets respond to the provisions in this rule
could have non-GHG exposure impacts for communities living near facilities that produce
biofuels. Chapter 3.1 summarizes what is known about potential air quality impacts of the
changes in renewable fuel volumes assessed for this rule. We expect that small increases in non-
GHG emissions from biofuel production and small reductions in petroleum-sector emissions

678 Mohai, P.; Pellow, D.; Roberts Timmons, J. (2009) Environmental justice. Annual Reviews 34: 405-430.
https://doi.org/10.1146/annurev-environ-082508-094348

679 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

680 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

681 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

682 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.

683 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.

684 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.

685 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. JExposure Sci Environ Epidemiol 23: 215-222.

686 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.

247

-------
would lead to small changes in exposure to these non-GHG pollutants for people living in the
communities near these facilities. We do not have the information needed to understand the
magnitude and location of facility-specific responses to the biofuel volumes associated with this
rule, 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.

8.3 Water & Soil Quality Impacts

We conducted an analysis to estimate the impacts associated with the finalized volumes
on water and soil quality in Chapter 3.4. Though soil quality is not among the statutory factors
required to be analyzed under the reset authority in the Clean Air Act,687 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, see Chapter 3.3.2. This land
use change has negative consequences for soil quality in that it can increase soil erosion,
depletion of SOM, 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, and 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 3.4.2.3.

As discussed in Chapter 3.4, the volumes associated with this rule have the potential to
incentivize increases in crop production, and by extension adverse impacts on soil and water
quality. The magnitude of this impact is difficult to estimate as it would require more
information on the extent to which the RFS standards drove changes in which biofuel volumes,
the impact this might have on feedstock production (e.g., corn, soy, palm), 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 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 environmental justice 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
annual volume requirements may have. However, going forward, we would like to better
understand the relationship between the annual RFS volume standards and land use/land
management decisions.

Any negative impacts on aquatic life have the potential to also negatively impact
populations who rely on fish or other aquatic life, like shrimp or crawfish, for sustenance or

687 CAA § 211(o)(2)(B)(ii).

248

-------
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.688 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.689 These
population groups are the same ones who would be most affected by any adverse impact the RFS
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.690 American Indian tribes also rely on recreational fisheries for
income, as explained by the U.S. Department of the Interior.691 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 populations who
rely on them. Additionally, any increased use of nitrogen rich fertilizers, as are applied to
approximately 98% of corn acres (see Table 3.4.2.1-1), 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 volumes in this rule adversely affect water quality, they could potentially have
disproportionately severe negative impacts on environmental justice 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.

8.4 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

688 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 offish and the opportunities and challenges emerging from the rise of

aquaculture. Journal of fish biology, 83(4), 1067-1084. https://doi.org/ .2.1.87

689 EPA, Regulatory Impact Analysis for the Final Mercury and Air Toxics Standards, Dkt. No. EPA-HQ-OAR-
2009-0234: tPA-HQ-0AR-201T-0044 (Dec. 2011).

690 Id.

691 U.S. Department of the Interior, Working with Native American Tribes, available at
https://www.fws.gov/policv/510fwLpdf (last accessed May 28, 2021), see also U.S. Department of the Interior,

Native American Trust Responsibilities, available at

https://www.doi.gov/sites/doi.gov/files/migrated/news/pressreleases/upload/Signed-SO-3335.pdf (last accessed May
28, 2021), and U.S. Department of the Interior, Indian Affairs, Branch of Fish, Wildlife, and Recreation, available at

https://www.bia.gov/bia/ots/division-natnral-resonrces/branch-fish-wildlife-recreation (last accessed May 28, 2021).

249

-------
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.
We estimated that the volumes would be associated with impacts on food prices (increases of
total food expenditures of 0.22% and 0.48% in 2021 and 2022 respectively). These impacts are
attributed to increases in corn and soy prices associated with the volumes. Because corn and soy
are a relatively small proportion of most foods purchased and consumed in the United States, the
overall food price impacts are relatively small as a percentage of total food expenditures. We
also projected relatively small price impacts on the prices of transportation fuel, with slight
increases in the price of gasoline (0.09 and 0.14 cents per gallon in 2021 and 2022 respectively)
and increases in the price of diesel (1.3 and 5.4 cents per gallon in 2021 and 2022 respectively).
These impacts are discussed in greater detail in Chapter 7.5 (food price impacts), Chapter 7.4
(price of agricultural commodities), and Chapter 9 (fuel price impacts).

The projections of the impact associated with the finalized volumes on food and fuel
prices for this rule 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 will vary for different parts of the country. However, changes in
food and fuel prices could have a disproportionate impact on populations who 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 units692 with income in
the lowest 20% spend a greater portion of their total expenditures on food and fuel (see Table
8.4-1). 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 8.4-1: Proportion of Total Expenditures on Food and Fuel

Lowest 20%

All Consumer

Income

Units

Consumer Units

Total expenditures

$61,334

$28,724

Food expenditures

$7,316

$4,099

Percent of total expenditures on food

11.9%

14.3%

Fuel expenditures

$1,568

$817

Percent of total expenditures on fuel

2.6%

2.8%

692 Consumer unit members are either related by blood, marriage, or other legal arrangement, or they are financially
dependent (i.e., live together and share the responsibility for some main expenditures in housing, food, and other
living expenses). In 2019 the average consumer unit was comprised of 2.5 people, of which 0.6 people were children
under 18. For more information see Bureau of Labor and Statistics Consumer Expenditure Survey, 2020.

250

-------
Chapter 9: Costs

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 reset authority. In
this chapter, we assess the social costs of renewable fuels (9.1), the social costs of the petroleum
fuels which the biofuels replace (9.2), the fuel economy effect based on each fuel's energy
density (9.3), and the impacts of this rule on social costs, the costs to consumers of transportation
fuel, and the costs to transport goods (9.4).

The costs are analyzed on the incremental volumes between the renewable fuel volumes
in 2021 and 2022 required by this final rule and the actual renewable fuels volumes in 2020.
Chapter 2 contains a summary of the final renewable fuels volumes as well as the actual
renewable fuels volumes in 2020 which serves as the baseline for establishing costs. Table 9.3-2
contains the change in renewable fuels volumes, as well as the estimated change in fossil fuel
volumes displaced by the increased volume of renewable fuels.

As noted in Chapter 2.2 of the RIA, for this cost analysis we generally used data sources
available in January 2022. We acknowledge that recent events since then (e.g., inflation and
geopolitical factors) have affected the U.S. agricultural and fuels markets, with potential
consequences for the analyses. We have generally not been able to quantitatively account for
these recent events in our cost analysis. In general, higher prices for gasoline and diesel make
renewable fuels more cost competitive which would lower the estimated costs of this action,
while higher prices for biofuel feedstocks such as corn and soybean oil make renewable fuels
less cost competitive and would increase the estimated cost of this action. Higher or lower
volumes of renewable fuels can also have a small impact on the price of gasoline and diesel by
increasing or decreasing the available supply of transportation fuel. Due to these potentially
offsetting effects and the varying market shifts in recent months in both crude oil and agricultural
commodity prices, it is difficult to estimate what their overall impacts might be on the estimated
costs presented in RIA Chapter 9.4. We included a sensitivity cost assessment (see RIA Chapter
9.4.4) to demonstrate the impact of higher crude oil prices on the cost of this rule, however we
note that this sensitivity did not consider the very recent further increase in biofuel feedstock
prices and is therefore not directly comparable to current market prices.

9.1 Renewable Fuel Costs

9.1.1 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 both 2021 and
2022. Although the year 2021 has passed, at the time of estimating the costs for the final rule, the
historical feedstock prices had not yet been compiled for 2021. Therefore, this analysis still relies
on recent projections for 2021 feedstock prices. In addition, we chose to use constant 2021

251

-------
dollars in our analysis. Since the projected prices of renewable fuel feedstocks are reported in
nominal year dollars, we converted the projected 2022 feedstock prices back to 2021 year dollars
using an inflation factor of about 2%.693

As we describe further below, we used data from multiple sources, including EIA and
USD A, in making the projections. We also made adjustments to account for differences between
these data. Specifically, the projected feedstock prices are adjusted to account for different crude
oil prices used by USDA than those projected by EIA.694 In addition to these adjustments for
inflation and crude oil price differences, the projected corn and soybean oil prices for this final
rule also reflect much higher corn and vegetable prices compared to what was used for the
proposed rule cost analysis. After the proposed rule, corn and vegetable oil prices increased by
50% to 100%. The USDA price projections that we use generally reflect these price increases.

Finally, recently geopolitical factors have caused petroleum prices to spike well above
their recent price norms, but these recent petroleum price increases are not reflected in the cost
analysis since they are not reflected in the underlying price projections available at this time. We
provide a sensitivity analysis at a higher crude oil price at the end of this chapter to demonstrate
the potential impact of higher crude oil prices. The supply issues related to the pandemic and
these other geopolitical factors increases the uncertainty that price projections precisely reflect
the cost of producing and using these renewable fuels and ultimately increases the uncertainty in
conducting this cost analysis.

9.1.1.1 Corn and Corn Ethanol Plant Byproducts

The price of corn is the most important input to enable estimating the cost of corn
ethanol. The starting corn prices and the adjustments made to them to put them on the same basis
as the EIA projected prices are summarized in Table 9.1.1.1-1 which helps to illustrate these
adjustments—the relevant row in the table that corresponds with the description here is indicated
in the text by row # in the table.

As a starting point we used corn price projections from the United States Department of
Agriculture (USDA) which makes projections of future corn prices. We started with the 2021
and 2022 USDA projected corn prices, which were $59.7 and $59.3 per bushel, respectively (row
#1).695 However, the USDA corn prices are reported in nominal dollars, reflecting the inflated
value of the dollar in the year 2022. The first adjustment we made was to convert those year
2022 USDA corn prices reported in nominal dollars into 2021 dollars (row #2). 696

693 USDA Agricultural Projections to 2031; Long Term Projections Report; February 2022

694 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.

695 USDA Agricultural Projections to 2031; Long Term Projections Report; February 2022.

696 USDA reports estimated future inflation rates which are used for the nominal dollar to 2021$ adjustment.

252

-------
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).697 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 2022 crude oil prices used by USDA (row #3) to 2021
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 2021$ (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: Corn
Price ($/bushel) = Crude Oil Price ($/bbl) x 0.0366 + 1.81.698 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 on a different supply/demand
point on the corn price curve as evidenced by todays higher corn prices. Instead, the difference in
regressed corn prices (row #8) was added to the USDA corn prices adjusted to 2021$ (row #2) to
derive the final adjusted corn prices used in the cost analysis (row #9).

Table 9.1.1.1-1: Derivation of Corn Feedstock Production Costs ($/bushel for corn, $/bbl
for Crude Oil)

Row #

2021

2022

Corn Prices

USDA Nominal $

4.53

5.45

USDA 2021$

4.53

5.33

Crude Oil
Prices

USDA Nominal $

59.7

60.6

USDA 2021$

59.7

59.27

EIA2021$

67.67

69.75

Regressed Corn
Prices

Based on USDA 2022

4.00

3.98

Based on EIA 2022 (Jan STEO)

4.29

4.37

Corn Prices

Difference in Regressed Corn
Prices EIA - USDA

0.29

0.38

Corn Prices

Adjusted USDA 2021$

4.82

5.71

Since corn ethanol production results not only in ethanol, but also byproducts which can
be sold for additional value, we also estimated the prices for those byproducts, specifically dried
distiller grains with solubles (DDGS) and inedible corn oil. Since USDA does not estimate future
prices for DDGS and corn oil, these were obtained by agricultural price projections made by the
University of Missouri, Food and Agricultural Policy Research Institute (FAPRI).699 The FAPRI
DDGS and corn oil projected prices are reported in nominal (2022) dollars, so we adjusted the

697 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.

698 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.

699 U.S. Agricultural Market Outlook 2021, Food and Agricultural Policy Research Institute (FAPRI); September
2021.

253

-------
price projections adjusted to 2021 dollars. Table 9.1.1.1-2 summarizes DDGS and corn oil prices
used in the cost analysis.

Table 9.1.1.

-2: Corn, Dried Distiller Grains Soluble and Inedible Corn Oil Prices (2021$)

Year

DDGS Prices ($/dry ton)

Inedible Corn Oil Prices (c/lb)

2021

182.6

42.78

2022

161.7

40.9

9.1.1.2 Soybean and Palm Oil Prices

Soybean oil and palm oil were identified in Chapter 2 as the primary feedstocks for
producing renewable diesel fuel. Soybean oil price projections made by USDA are used as a
starting point for this cost analysis.700

We then followed the same methodology we used for corn prices described above, but for
soy oil prices this process is summarized in Table 9.1.1.2-1 and is enumerated for linking the
following text to that table. The first step required converting USDA projected soy oil prices in
nominal dollars (row #1) to 2021 dollars (row #2), and then adjusting for the differences in crude
oil prices (row $4 for USDA in 2021$) 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: Soy Oil Price ($/lb) = Crude Oil Price ($/bbl) x
0.259 + 19.06.701 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 soy oil prices (row #9) incorporating these inflation and crude oil price
adjustments, which are then adjusted further below to account for increased demand.

Table 9.1.1.2-1: Derivation of Soy Oil Feedstock Production Costs (cents/lb for soy oil,
$/bbl for crude oil)

Row #

2021

2022

Soy Oil Prices

USDA Nominal $

0.569

0.65

USDA 2021$

0.569

0.636

Crude Oil Prices

USDA Nominal $

59.7

60.6

USDA 2021$

59.7

59.27

EIA2021$

67.67

69.75

Regressed Soy Oil
Prices

Based on USDA 2021

34.53

34.42

Based on EIA 2021

36.60

37.14

Soy Oil Prices

Difference in Regressed Soy Oil
Prices EIA - USDA

2.07

2.73

Soy Oil Prices

Adjusted USDA 2021$

59.0

66.3

700 USDA Agricultural Projections to 2031; Long Term Projections Report; February 2022.

701 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 syncs the two price projections and leads to a better estimate of costs.

254

-------
Palm oil is expected to be the other feedstock used for producing a portion of renewable
diesel, but neither USDA nor FAPRI project future palm oil prices. Instead, future palm oil
prices for this analysis were estimated based on the historical differences between palm oil spot
prices and soybean oil's spot prices.702 Palm oil spot prices were compared to soybean oil spot
prices between January 2012 and September 2017. Over that time period, soybean oil averaged
37.6 cents per pound, and palm oil averaged 29.4 cents per pound, respectively. The palm oil to
soybean oil ratio is 0.78, and this was used for establishing palm oil prices relative to USDA
projected 2021 and 2022 soybean oil prices.

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
soybean oil price by 0.032 dollars per pound.703 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 final rulemaking, which is estimated to be 286 million gallons in 2021 and 819 million
gallons in 2022. A similar adjustment was made for the estimated increased demand for palm oil.
The volume of the global palm oil market is nearly 6 times greater than the U.S. soybean oil
market, so the adjustment to the palm oil prices is assumed to be one sixth that for soy, or 0.005
dollars per pound for each 200 gallon increase in palm oil demand. The projected soy and palm
oil prices in the baseline and resulting from the increased demand due to the rule that are used in
this cost analysis are summarized in Table 9.1.1.2-2.

Table 9.1.1.2-2: Projected Vegetable Oil Production Costs (2021$/lb)

Year

Modeled Ve

getable Oil Prices

Base Oil Prices

Adjusted for Increased Demand
Due to RFS Volumes

Soybean Oil

Palm Oil

Soybean Oil

Palm Oil

2021

59.0

61.9

2022 (2021$)

66.3

52.9

75.4

53.2

As discussed in Chapter 2, we are also projecting that a small volume of renewable diesel
will be produced from waste fats, oils, and greases (FOG). In the DRIA, EPA separately
estimated costs for FOG renewable diesel. As we discussed in the DRIA, these feedstocks have
historically been priced lower than soybean oil, but may have slightly higher processing costs.
Specifically, the DRIA estimated a 0.78 ratio for FOG to soybean oil feedstock costs.

In performing the analysis for this final rule, EPA inadvertently estimated the entire
increase of domestic renewable diesel in 2021 and 2022 as being produced from soybean oil for
the purposes of this cost analysis, rather than primarily from soybean oil with a much smaller
volume produced from waste fats, oils, and greases. As such, the cost estimates presented in this
section somewhat over-estimate the cost of renewable diesel. However, given that the volume of
FOG renewable diesel is an order of magnitude smaller than the volume of soybean oil
renewable diesel, and given that their costs are relatively similar, the impact on the overall costs

702 USDA Yearbook Tables by the Economic Research Service, downloaded March 2021.

703 Babcock, Bruce, et. al. Biofuel Taxes, Subsidies, and Mandates: Impacts on US and Brazilian Markets (May
2013) http://www.card.iastate.edu/publications/dbs/pdffiles/13srl08.pdf.

255

-------
for this rulemaking is expected to be quite small and would not impact the volumes we are
finalizing in this rule.704

9.1.1.3 Biogas

For this analysis we assume that biogas is the byproduct by landfills and collected by the
landfills to prevent the emissions of methane gas as required by regulation and flared, burned to
produce electricity, or cleaned up 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 subsections discussing
production and distribution costs.

9.1.1.4 SugarCane

The sugar cane feedstock costs are not estimated separately from the rest of the sugar
cane processing costs, which are discussed below in Chapter 9.1.2.3.

9.1.2 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
2021 cost inputs705 - similar detailed costs for 2022 are not presented. All the costs information
summarized in this section are calculated in a spreadsheet which available in the docket for this
action.

9.1.2.1 Cost Factors
9.1.2.1.1 Capital and Fixed Costs

The economic assumptions used to amortize capital costs over the production volume of
renewable fuels are summarized below in Table 9.1.2.1.1-1. These capital amortization cost

704 We also performed a sensitivity analysis by assessing the costs of FOG renewable diesel separately, as we did in
the proposed rule. In the proposed rule, we estimated a 0.78 ratio for FOG to soybean oil feedstock costs. If we
apply this ratio to the updated price of soybean oil and further make the simplifying assumption (as we did in the
cost analysis for the proposed rule) that FOG feedstocks do not require additional processing to remove impurities,
we estimate FOG feedstocks could be priced as low as 52 cents per pound compared to 66 cents per pound estimated
for soybean oil for 2022. This results in a FOG renewable diesel cost of $5.02 per gallon, which is $1.96 per gallon
less than the cost of soybean oil renewable diesel. In Chapter 2, we estimated 80 million gallons of additional FOG
renewable diesel in 2022 relative to 2020. If we apply the $1.96 per gallon differential to 80 million gallons, that
would reduce the cost of this rulemaking by $160 million in 2022. This equates to 2.6% of the total costs in 2022
compared to 2020.

705 Table 9.1.2.2-1, Table 9.1.2.3-1, Table 9.1.2.4-1, Table 9.1.2.5-3 and Table 9.1.2.6-1

256

-------
factors are used in the following section for converting the one-time, total capital cost to an
equivalent per-gallon cost.706 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
papers.707 708 709 710 711

Table 9.1.2.1.1-1: Economic Cost Factors Used in Calculating Capital Amortization Factors

Amortization
Scheme

Depreciation
Life

Economic
and Project
Life

Federal and
State Tax
Rate

Return on
Investment
(ROI)

Resulting
Capital
Amortization
Factor

Societal Cost

10 Years

15 Years

0.11

Capital costs were adjusted to 2021 year dollars for this analysis. The cost of installing
capital decreased in 2020 due to the pandemic, and then seemed to have increased dramatically
in 2021, based on only part year data, due to factors associated with inflation. Basing capital cost
on either 2020 or 2021 capital cost factors would likely misrepresent the cost of capital moving
forward.712 Instead, a regression of Chemical Engineering Plant Index (CEPI) capital cost factors
from 2015 to 2019 was used to estimate a capital cost inflation factor for 2021, and this capital
cost factor for 2021 is consistent with historical increases in capital cost inflation factors - about
a 2% per year increase.

Fixed costs include the maintenance costs, insurance costs, rent, laboratory charges and
miscellaneous chemical supplies.713 Maintenance costs can range from 1% to 8% for industrial
processes.714 The fixed costs are lowest for the best designed production facilities for well-
established industries. For the renewable fuels technologies which have not been operating as
long (for a decade or two, as opposed to the refining industry which is better established and has
been operating for many decades), we assumed that maintenance costs are at the higher end of

706 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/jsp/Utilities/circular-a-4_regulatory-impact-analysis-a-primer.pdf).

707 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.

708 Cost Estimates of Long-Term Options for Addressing Boutique Fuels; Memorandum from Lester Wyborny to the
Docket; October 22, 2001.

709 Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control
Requirements; EPA420-R-00-028; December 2000.

710 Final Regulatory Analysis: Control of Emissions fromNonroad Diesel Engines; EPA420-R-04-007; May 2004.

711 Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis; EPA-420-R-10-006; February 2010.

712 For newly constructed renewable diesel plants for 2021 and 2022, the ordering of equipment or equipment
purchases may have occurred in 2020 when capital costs are low, and other aspects of the capital costs such as its
installation could have occurred in 2021, so a highly inflated capital cost inflation factor for 2021 would likely not
accurately represent the average of capital installation costs for a new facility being installed for 2021 or 2022. For
this reason, a more average cost of capital adjustment is more appropriate.

713 Peters, Max S., Timmerhaus, Klaus, D.; Plant Design and Economics for Chemical Engineers 3rd Edition;
McGraw Hill; 1980.

714 McNair, Sam Budgeting for Maintenance: A Behavior-Based Approach, Life Cycle Engineering, 2011.

257

-------
this range comprising 5.5% of the capital costs.

9.1.2.1.2 Utility and Fuel Costs

Variable operating costs are those 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. Utilities (electricity, natural gas and water) 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
other plant operations. Water can be necessary as part of the process (reaction medium), or used
in cooling towers.

Projected electricity and natural gas prices are based on national average values from
Energy Information Administration's (EIA) Jan. 2022 Short Term Energy Outlook (STEO).715
The STEO-projected electricity and natural gas prices for industrial users in the midwest are used
because it is the most common location for renewable fuels plants.716 The cost of process water
is generally quite minimal, but a cost is estimated for process water as well since renewable fuels
technologies can use fairly large quantities.717 718The utility costs used for the cost analysis are
summarized in Table 9.1.2.1.2-1.

Table 9.1.2.1.2-1: Summary of Utility Cost Factors (nominal dollars)1

Year

Natural Gas
(S/1000 cf)

Electricity
(c/kWhr)

Water
(S/1000 gals)

2021

5.46

6.67

3.0

2022

5.18

7.20

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.

9.1.2.2 CornEthanol

Corn ethanol plant demand and production output information were based on a 2012
survey of corn ethanol plants.719 Capital costs are based on a review of corn ethanol construction
costs for a 100 million gallon per year 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.720 Since the capital cost is based on the total construction

715 Short Term Energy Outlook 2021, Energy Information Administration, January 2022;
https: //www. eia. gov/outlooks/steo/

716 The ratios of the average Midwest natural gas and electricity prices to the national natural gas and electricity
prices are applied to the AEO projections of national natural gas and electricity prices to derive Midwest projections
for natural gas and electricity prices.

717 Haas, M.J, A process model to estimate biodiesel production costs, Bioresource Technology 97 (2006) 671-678.

718 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.

719 Mueller, Steffen; 2012 Corn Ethanol: Emerging Plant Energy and Environmental Technologies; April 29, 2013.

720 Irwin, Scott; Weekly Output: Ethanol Plants Remain Barely Profitable; 3/16/2018.

258

-------
cost of already constructed corn ethanol plants, no contingency cost factors are applied to the
capital costs. The survey information estimates the quantity of DDGS produced by corn ethanol
plants. 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.721 Of the corn ethanol
plants in the 2012 survey, 74% were separating and selling corn oil so selling corn oil was
assumed for 70% of the plant capacity. Table 9.1.2.2-1 contains the plant demand and outputs
and capital costs for corn ethanol plants.

Table 9.1.2.2-1: Corn Ethanol Plant Demands, Production Levels, and Capital Costs (2021

Category of Plant
Input/Output

Plant
Inputs/Outputs

Cost per Input

Cost (MM$)

Cost ($/gal)

Ethanol Yield

2.83 Gal/Bushel

4.82 $/bushel

131

1.72

DDG Yield

15.7 Lbs/Bushel

183 $/ton

-38

-0.50

Corn Oil Yield

0.53 Lbs/Bushel

43 cents/lb

-4.5

-0.06

Thermal Demand

23,860 BTU/Gal

5.46 $/1000 cf

9.6

0.13

Electricity Demand

0.75 kWh/Gal

7.2 c/kwh

4.1

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 (2020
dollars,

76 million Gals/Yr)

2.34 $/Gal Plant

25.5

0.33

Capital cost

Annual Fixed Cost

5.5% of Total
Capital Cost

12.8

0.17

Denaturant

2 volume percent

0.4

0.01

Total Cost

146

1.92

The projected corn ethanol social production cost for a 85 million gallon capacity ethanol
plant is $1.92 per gallon of denatured ethanol for 2021, and $2.28 per gallon of denatured
ethanol for 2022 when the year 2022 utility and corn prices are used.

9.1.2.3 Imported Sugarcane Ethanol Production Costs

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 fermented into
ethanol. The fibrous portions of the sugarcane plant is typically combusted to produce the energy
needed for the process.

We based our imported ethanol fuel costs on a simple cost estimate and compared it to
price data for sugarcane ethanol in Brazil. Generally, ethanol from sugarcane produced in
tropical areas is cheaper to produce than ethanol from cellulose, and approaches 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.

721 Edwards, William; Grain Truck Transportation Cost Calculator (a3-29graintransportation.xlsx version
1.4 82017); Iowa State University.

259

-------
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 operating costs higher by 28% to
account for the effects of inflation from 2006 to 2021 using the CEPI capital cost index. 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 9.1.2.3-1 provides the
updated production cost estimate for sugarcane ethanol.

Table 9.1.2.3-1: Cost of Production for a Standard Sugarcane Ethanol Project in Brazil
(2021$)

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
($/gal; $2006)

Revised Costs
($/gal; $2021)

Capital Costs
($ million)

Investment cost in mill

124

Investment cost for
sugarcane production

Per Gallon Costs
($/gal)

O & M (Operating &
Maintenance) costs

0.26

0.33

Variable sugarcane
production costs

0.64

0.82

Capital costs

0.49

0.50

Total production costs

1.40

1.77

We also reviewed Brazil's market price of the sugar cane ethanol which should
approximate the production cost of sugar cane ethanol. The Argus Americas report provides the
sugarcane ethanol free on board (FOB) price at Santos, the port of San Paulo, Brazil, from which
sugarcane ethanol is typically shipped to the U.S. We reviewed daily price data from the Argus
Americas Biofuels Report from September 2015 to August 2017 and taking a weekly sample of
prices, sugar cane ethanol averaged $1.98 per gallon, including the shipping costs to the U.S.,
over that two year period. The ethanol prices varied from about $1.60 to $2.20 per gallon. The
average FOB ethanol price of $1.98/gallon in Brazil is somewhat higher than the estimated sugar
cane ethanol production cost of $1.77/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 0.17

260

-------
after-tax capital amortization factor used by industry, the per-gallon costs increase to $2.05 per
gallon which is about the middle of the range for the FOB price data.

9.1.2.4 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.722723 This data is likely still representative of biodiesel plant costs
because the process is fairly simple and its cost are likely to be fairly stable over time.
Furthermore, the biodiesel costs are mainly (>80 percent) 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 section to meet biodiesel purity specifications, and 3) glycerol recovery section.724
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 60C. 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 60C.
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 USP grade, however this
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.

722 Haas, M.J, A process model to estimate biodiesel production costs, Bioresource Technology 97 (2006) 671-678.

723 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.

724 Haas, M.J, A process model to estimate biodiesel production costs, Bioresource Technology 97 (2006) 671-678.

261

-------
We made a series of adjustments to the Haas model output. The capital cost is adjusted
from 2006 dollars to 2021 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,725 sodium methoxide,726 hydrochloric acid,727 sodium
hydroxide,728 and glycerine.729 The value of methanol is from a Methanex report plus 15 cents
per gallon distribution costs.730 Prices for sodium methoxide, hydrochloric acid and sodium
hydroxide are all bulk prices from a chemicals supplier.731

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.732 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 10 cents per pound for glycerine.733

Table 9.1.2.4-1 also shows the production cost allocation for the soybean oil-to-biodiesel
facility as modeled for 2021 and 2022. Production cost for biodiesel is primarily a function of
feedstock price, with other process inputs, facility, labor, and energy comprising much smaller
fractions.

725 Methanex; current North America prices plus 15 c/gal for shipping; https://www.methanex.com/onr-
business/pricing: August 2018.

726 Alibaba;

fattps://www.alibaba.eom/trade/searefa?rsb=v&IndexArea=prodiict en&CatId=&SearchText=sodium+methoxide:
August 2018.

727 Alibaba;

https://www.alibaba.com/trade/search?IndexArea=prodnct en&CatId=&fsb=v&SearchText=hvdrochloric+acid:
August 2018.

728 eBioChem; http://www.ebiochem.eom/SearcIi/searcIi/cate2/tiame/cate/0/kevwords/sodinm%2520hvdroxide/:
August 2018.

729 Perez, Leela Landress; US crude glycerine prices could dip as spring nears;

https://www.icis.com/explore/resources/news/2018/02/14/10193613/us-crude-glycerine-prices-could-dip-as-spring-
nears: February 14, 2018.

730 Methanex Methanol Price Sheet; US Gulf Coast; May 31, 2018

731 https://www.alibaba.com.

732 Yang, Fangxia; Value-added uses for crude glycerol - a byproduct of biodiesel production; Biotechnology for
Fuels; March 14, 2012.

733 US crude glycerine prices could dip as spring nears; ICIS News; February 14, 2018.

262

-------
Table 9.1.2.4-1: Biodiesel Production Cos!

(2021$)

Unit Demands

Cost per Unit

Thousand
Dollars

$/gal

Soybean Oil Feed

74,152 (1000 lb)

62 cents/lb

47,800

4.78

Methanol

7422(1000 lb)

1.48 $/gal

1,700

0.17

Sodium Methoxide

927 (1000 klb)

$2000/ton

927

0.09

Hydrochloric Acid

529 (1000 lb)

$200/MT

48.1

0.005

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)

10 cents/lb

(900)

(0.09)

Natural Gas

66.9 million cf

$5.46/1000cf

365

0.037

Electricity

1008 kW

7.20 cents/kWh

635

0.064

Labor

0.05

Capital Cost 2006$

11.35 ($million)

Capital Cost 2020$

15.12 ($million)

1,600

0.16

Fixed Cost

5.5%

800

0.08

Total Cost

53,080

5.36

9.1.2.5 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 $1100 million for a standalone 400
million gallon per year facility.734 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 2022.735 The capital cost for this smaller
renewable diesel fuel plant is estimated to be $768 million.

734 Sapp, Meghan; Diamon Green Diesel to Invest in $1.1 billion expansion with UOP's Ecofining tech; October I,
2019.

735 The typical renewable diesel plant size is based on volume-weighting the renewable diesel capacity data in Table
5.2.2-1. 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.

263

-------
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 require 2,000 SCF/bbl of
feedstock processed.736 Hydrogen costs are estimated based on hydrogen production by a steam
methane reforming hydrogen plant.737

Table 9.1.2.5-1: Hydrogen Plant Costs in 2021$ (Based on 1000 standard cubic feet of

Hydrogen; Hydrog

en plant capacity oi

'32 million cubic

eet per day)

Unit Demands

Cost per Unit

Cost

Million Dollars

$/thousand ft3

Feed Natural Gas

0.247 mmBTU

5.66 $/mmBTU

40.2

3.84

Fuel Gas for Heat

0.14 mmBTU

5.66 $/mmBTU

22.8

2.17

Power

0.798 KWh

7.16 c/KWh

5.7

0.55

Catalyst

4.8 cents

0.04

Capital Cost

$50 MM in 2006

$80 MM in 2021

8.7

0.84

Fixed Cost

5.5 %

4.4

0.42

Total Cost

7.85

Based on our cost analysis, hydrogen is estimated to cost $7.85 per thousand standard
cubic feet in 2021. The estimate assumes a dedicated hydrogen plant producing for the
renewable diesel facility of 220 million gallons per year being modeled for this cost analysis.

Our yield estimates as summarized in Table 9.1.2.5-2 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.738 Despite the age of the
reference, the underlying chemistry is unlikely to have changed appreciably.

Table 9.1.2.5-2: Input and Output from Renewable Diesel Plant

Fat 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

gal is gallons; SCF is standard cubic feet

We derived a cost of 6.9 cents/gallon of renewable diesel product to cover other costs:
utilities, labor, and other operating costs.739 Finally, the total cost per gallon was estimated at

736 Conversation with Mike Ackerson, Duke Biofuels, May 2020.

737 Gary, JH, Handwerk, GE; 2001 Petroleum Refining Technology and Economics, 4th edition, Marcel Dekker.

738 A New Development in Renewable Fuels: Green Diesel, AM-07-10 Annual Meeting NPRA, March 18-20, 2007.

739 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.

264

-------
$5.87. Table 9.1.2.5-3 provides more details for the process assumed in this analysis and
summarizes the total and per-gallon costs for the year 2021.

Table 9.1.2.5-3: Renewable Diesel Production Cost Estimate for a Greenfield 220 million
gallons/yr Plant Processing Soy Oil (2021$)

Stream

Estimated value

MM$/yr

$/gal

Soy Oil input

235 MMgals/yr

62.0c/lb

1124

5.11

Naphtha output

4.0 MMgals/yr

1.45 c/gal

(17.1)

(0.08)

Light fuel gas output

7.2 MMgals/yr

73 c/gal

(15.5)

(0.07)

Hydrogen input

4760 scf/100
gals

$7.85/thousand
standard cubic feet

88.0

0.40

Other Operating Costs

15.2

0.07

Capital Costs (2021 dollars)

$768 million

54.09

0.25

Fixed Costs

5.5%

42.2

0.19

Total Costs

1291

5.87

A number of announced renewable diesel projects projected to start-up in 2021 and 2022
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.740 These modifications are estimated to cost about one third the cost of a new
renewable diesel hydrotreater, or $230 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. 741 742

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 $5.73/gallon versus
$6.01/gallon for a greenfield renewable diesel plant. The difference between the low and high

740 Chan, Erin; Converting a petroleum diesel refinery for renewable diesel fuel; Hydrocarbon Processing; April
2021.

741 Chan, E., Converting a petroleum diesel refinery for renewable diesel; Special Focus: Clean Fuels; April 2021.

742 Lane, Robert; Renewable Diesel Interest Accelerates; August 26, 2020.

265

-------
production costs is solely due to the difference in capital costs. For refineries converting their
refineries to produce renewable diesel, the amortized capital cost is estimated to be only $0.12
per gallon, while the greenfield plant's estimated capital cost is $0.38 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 $5.87 estimated production cost for 2021 as summarized in Table
9.1.2.5-3.

9.1.2.6 Biogas

Biogas is the result of anaerobic digestion of organic matter, including municipal waste,
manure, agricultural waste and food waste.743 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 it can be used by processes that normally use natural gas.744

The largest source of biogas, which is already being collected to avoid methane
emissions, is from landfills.745 Since landfill gas is the most likely source of biogas for the motor
vehicle fleet, to estimate the cost of biogas, this analysis assumes that the biogas will solely be
provided by landfills. To use the biogas for transportation purposes, it must first be cleaned up to
remove contaminants. While in some cases, the gas can be used in local fleet vehicles which
would be refueled at the landfill, 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 through contracts and/or
affidavits rather than using the same methane molecules produced at the landfill.

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 by the transportation sector using Version 3.5 of
the Landfill Gas Energy Cost Model (LFGcost-Web).746 747 The throughput volume of landfill
gas was estimated to be 600 standard cubic feet per minute based on a survey of biogas
production facilities.748 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 9.1.2.6-1.

743 Wikipedia. Accessed April 2021; https://en.wikipedia.org/wiki/Biogas.

744 LeFevers, Daniel; Landfill Gas to Renewable Energy; Hill Briefing; April 26, 2013.

745 Biomass Explained, Landfill Gas and Biogas; US Energy Information Administration; February 1, 2019;

www.eia.gov/ene rgvexplained/index.php?page=bioimass biogas

746 The current version of this model and user's manual dated March 2021 are downloadable from the LMOP
website: https://www.epa.gov/lmop

747 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.

748 Economic Analysis of the US Renewable Natural Gas Industry; The Coalition of Renewable Natural Gas;
December 2021.

266

-------
We included an operating cost for biogas collection at the landfill.749 While the capital
costs for collecting the biogas are assumed to be sunk at existing sites, the operating costs are
ongoing and should be attributed to this new use. Distribution and retail costs are estimated for
biogas in Chapter 9.1.4.3.

Table 9.1.2.6-1: Biogas Col

ection and Cleanup Costs (60

0 cf/min)

Cost Factors (2019$)

million
dollars
(2021$)

$ per thousand
cubic feet
(2021$)

Interconnection

$400,000

0.42

0.15

Capital Costs

6,000000*e(000°3*ft3/min)

7.57

2.64

Operating and
Maintenance

250 * ft3/min +148,000

0.31/yr

1.00

Electricity Costs

0.009 kWh/ft3

0.20/yr

0.65

Total

4.43

9.1.3 Blending and Fuel Economy Cost

Certain renewable fuels, namely gasoline, biodiesel, and renewable diesel, are typically blended
into petroleum fuels. These renewable fuels also 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.

9.1.3.1 Ethanol
9.1.3.1.1 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) at low concentrations, and is low in energy content relative to the gasoline pool
that it is blended into. Ethanol has about zero sulfur and no benzene which also has 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 (RFG versus CG,
winter versus summer gasoline, premium versus regular, and blended at 10% or blended as El5
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 refining costs. Refiners dislike
ethanol's high blending RVP when blending ethanol in RFG at 10% because they must remove
some low value 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. Rather, this is usually just an issue for the consumers

749 LFG Energy Project Development Handbook - Project Economics and Financing; Chapter 4.

267

-------
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, particularly now when almost all gasoline is E10). Since this is a social cost
analysis which 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 in two different steps based on the output from
two different refinery modeling cases conducted by ICF/Mathpro. In the first step, ethanol's
blending value is estimated while blended into E10 conventional and reformulated gasolines.
The blending value (primarily octane, but also includes its low sulfur and absence of benzene)
and RVP cost of ethanol at 10% was estimated. The refinery modeling cases were for a 2020
year case assuming that crude oil would be priced at $72/bbl.750 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. The refinery modeling output from the second case, which modeled
ethanol's removal from the conventional gasoline pool, allowed us to estimate ethanol's
replacement cost - replacing both its volume and octane and includes the cost of installing the
capital necessary to replace ethanol's octane and volume.

The refinery modeling output from the first modeling case is summarized in Tables
9.1.3.1.1-1 and 2, which summarize the refinery model's marginal blending values (also termed
shadow prices) for both ethanol and gasoline.

750 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.

268

-------
Table 9.1.3.1.1-1: Refinery Model's Ethanol Value (c/gal)a

PADD 1

PADD 2

PADD 3

PADD 4

PADD 5

Ethanol

Summer

Winter

Summer

Winter

Summer

Winter

Summer

Winter

Summer

Winter

RFG

Premium

188

196

193

199

193

184

102

202

Regular

197

202

203

206

203

190

108

205

CG with

Waiver

Premium

218

196

214

199

213

184

211

190

225

201

Regular

233

205

CG No

Waiver

Premium

190

Regular

200

7.8 RVP

with Waiver

Premium

217

215

216

274

Regular

225

226

268

7.8 RVP No

Waiver

Premium

189

222

193

184

Regular

198

142

202

189

7.0 RVP

with Waiver

Premium

218

Regular

226

7.0 RVP No

Waiver

Premium

Regular

a Premium means premium grade gasoline; regular means regular gasoline grade. Values are reported for each
gasoline grade/type produced in each PADD - blank cells means that the gasoline type/grade is not produced in that
PADD.

269

-------
Table 9.1.3.1.1-2: Refinery Model's Gasoline Value (c/gal)a

PADD 1

PADD 2

PADD 3

PADD 4

PADD 5

Summer

Winter

Summer

Winter

Summer

Winter

Summer

Winter

Summer

Winter

Gasoline

RFG

Premium

196

182

191

179

190

171

199

183

Regular

189

178

186

175

183

168

190

180

CG with

Waiver

Premium

192

182

188

179

187

172

178

167

173

182

Regular

185

178

182

175

181

168

172

163

164

179

CG No

Waiver

Premium

192

181

Regular

185

177

7.8 RVP

with Waiver

Premium

190

178

189

171

180

167

192

Regular

184

174

183

168

174

163

182

7.8 RVP No

Waiver

Premium

195

189

172

Regular

188

183

168

7.0 RVP

with Waiver

Premium

191

179

Regular

186

175

7.0 RVP No

Waiver

Premium

Regular

a Premium means premium grade gasoline; regular means regular gasoline grade. Values are reported for each
gasoline grade/type produced in each PADD - cells are blank because that gasoline type/grade is not produced in
that PADD.

Ethanol's blending value is calculated by subtracting gasoline's marginal value from
ethanol's marginal value for the same category of gasoline in Tables 9.1.3.1.1-1 and 2. For
example, ethanol's value for blending with regular grade summertime CG with the waiver is
96.14 dollars per barrel, compared to the gasoline which is 77.89 dollars per barrel. Thus, the
refinery model values ethanol 18.25 dollars per barrel more than gasoline, or about 0.43 dollars
per gallon. Table 9.1.3.1.1-3 summarizes the calculated blending values for ethanol in summer
gasoline after the blending values for the various PADDs were volume-weighted averaged
together and the costs were converted over to cents per gallon.751

751 The summertime ethanol blending values for California RFG were very negative, which the contractor thought
could be due to low butane prices. Regardless of the cause, these negative ethanol blending costs were considered
outliers and, thus, were not included in our average of ethanol blending costs.

270

-------
Table 9.1.3.1.1-3: Ethanol's Blending Value in Summer Gasoline ($/gal assuming a crude
oil price of $72/bbl)

Gasoline Type

Gasoline Grade

Summer

Regular

0.32

Premium

0.20

RFG

Regular

0.11

Premium

0.01

Ethanol's volatility cost for blending ethanol into RFG is estimated by subtracting
ethanol's blending value in reformulated gasoline from ethanol's blending value in conventional
gasoline. Thus, ethanol's volatility's cost is 21 and 19 cents per gallon in regular and premium
grade gasolines, respectively.

The results from the first modeling case demonstrate ethanol's considerable blending
value on a day-to-day basis in the marketplace. However, it does not really capture the full
blending value of ethanol were a refiner to be faced with the cost of replacing it in the gasoline
pool - i.e., if it were not part of the fuel and octane supply and other sources of volume and
octane supply had to replace it. This situation was modeled in the second refinery modeling case
described above in which ethanol is removed from the conventional gasoline pool, and thus
provided an estimate of ethanol's replacement cost. As shown in 9.1.3.1.1-1 through 3, this
second case revealed that ethanol's very high octane and volume would require significant
investment at refineries to replace ethanol and this is reflected in ethanol's much higher blending
value when accounting for these additional costs.

Table 9.1.3.1.1-4 summarizes ethanol's marginal costs by different gasoline types and
refinery regions 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. Two different ethanol replacement cases were modeled: low biofuel #1 is a
reformate-centric case while low biofuel #2 is an alkylate centric case.752 The much lower
marginal values for PADD 1 can be explained because Mathpro allowed PADD 3 refineries to
satisfy PADD l's need for replacing ethanol's volume and octane through its exports into the
PADD 1 after initial refinery model runs showed PADD l's marginal costs for replacing ethanol
were exceedingly high.

752 Reformate is a gasoline blendstock produced by a refinery unit named the reformer. Reformers react a low-octane
stream from the crude distillation tower which boils in the gasoline boiling range over a catalyst to convert (reform)
low octane hydrocarbons to high octane aromatic compounds. Alkylate is a gasoline blendstock produced by a
refinery unit named the alkylation unit. The alkylation unit reacts isobutylene with isobutane and normal butane in
acid to branched chain paraffins which are high in octane.

271

-------
Table 9.1.3.1.1-4: Gasoline Marginal Values for Reference Case and Ethanol Marginal
Values for the Low-Biofuel Cases ($/bbl assuming crude oil priced at $72/bbl)a

PADD of
Gasoline
Origin

Type

Grade

Gasoline Marginal
Values

Ethanol Marginal Values

Low-Biofuel #1
Reformate Centric

Low-Biofuel #2
Alkylate Centric

Summer

Winter

Summer

Winter

Summer

Winter

PADD 1

RFG

Prem

95.737

83.936

108.37

100.88

Reg

91.452

81.349

115.98

105.97

Conv

Prem

92.680

83.890

123.02

100.87

Reg

88.927

81.346

136.43

105.88

PADD 2

RFG

Prem

88.087

81.685

132.42

110.28

113.45

96.62

Reg

84.803

79.771

145.38

116.02

122.86

101.61

Conv

Prem

85.549

81.248

149.08

110.41

126.74

96.25

Reg

82.457

79.447

161.21

115.79

135.55

100.93

PADD 3

RFG

Prem

85.424

78.308

121.69

94.72

118.51

89.77

Reg

81.863

76.395

134.67

98.45

131.29

94.48

Conv

Prem

83.644

78.784

133.95

95.13

129.37

89.91

Reg

79.975

76.755

146.78

98.46

142.00

94.55

PADD 4

RFG

Prem

79.756

77.014

135.5

115.2

150.1

103.1

Reg

77.367

75.066

149.0

124.0

168.1

110.0

Conv

Prem

81.785

82.073

136.5

151.2

Reg

81.702

81.993

150.1

169.2

PADD 5

RFG

Prem

96.890

83.676

37.68

96.05

Reg

91.607

82.013

62.46

97.37

Conv

Prem

77.631

82.999

118.14

98.01

Reg

73.384

81.116

126.14

97.68

a Low biofuel case means no biofuel in the conventional gasoline pool.

The difference between ethanol and gasoline marginal values is calculated, converted to
cents per gallon and then volume-weighted to summarize them on a national average basis by
gasoline grade and season in Table 9.1.3.1.1-5.

Table 9.1.3.1.1-5: Marginal Ethanol Replacement Cost by Gasoline Grade and Season
(cents/ethanol gallon, crude oil priced at $72/bbl)

Low Bi(
Reformat

)fuel #1
e-centric

Low Biofuel #2
Alkylate-centric

Summer

Winter

Summer

Winter

Conv.

Prem

124.58

50.79

112.04

32.65

Reg

165.11

66.83

144.23

48.19

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

272

-------
reformulated gasoline (RFG) as well, so we assumed that they were the same for RFG.753
However, it is necessary to add in ethanol's volatility cost, which for ethanol's removal would be
a cost savings. The 21 and 19 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 9.1.3.1.1-
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 Table 9.1.3.1.1-
6.

Table 9.1.3.1.1-6: Aggregated Ethanol Marginal Replacement Cost (cents/gallon)

Low Biofuel #1
Reformate-centric

Low Biofuel #2
Alkylate-centric

Summer

Winter

Summer

Winter

Conv.

Prem

124.58

50.79

112.04

32.65

Reg

165.11

66.83

144.23

48.19

RFG

Prem

105.58

50.79

93.04

32.65

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.

9.1.3.1.2 Higher Ethanol Blends

Compared to E10, the blending costs are different for ethanol blended into E15 and E85
because these fuels are blended differently. The 1 psi RVP waiver only applies to E10 in
summertime conventional gasoline. Although El 5 could potentially realize a blending benefit
based on the increased octane for the additional ethanol if refiners could create and distribute a
low RVP, low octane El 5 BOB, this would require a widespread shift by refineries, pipelines,
and terminals in an entire geographical region to produce and distribute another even lower
octane blendstock for oxygenate blending (BOB) specially for blending El 5 instead of El 0.754
This would most likely only occur if E15 becomes the predominant gasoline used in that region
based on the limitations of the distribution system and experience with the historic conversion to
E10. 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 over the ethanol

753 Both RFG and CG must meet many of the same gasoline property specifications, including sulfur and benzene, as
well as ASTM standards (ASTM D4814).

754 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.

273

-------
blended in E10 (the additional 5%).755 Thus, the gasoline BOB used to blend up E15 in the
winter months is the same as that used for blending into E10 resulting in a higher octane fuel
than what it can be priced at. In the summer months, El5 also incurs the additional RVP control
costs.

We also did not assign a blending credit for E85 for its high octane beyond that credit
already being realized when blending E10. At higher ethanol blends, the very high octane
resulting from the high ethanol content results in significant overcompliance with the minimum
octane standard regardless of the octane level of the gasoline portion of the blend. As a result,
there is no additional octane value associated with the additional ethanol. But there is no RVP
cost for E85 either because the high portion of ethanol results in lower RVP instead of higher
RVP in such mixtures, therefore, a lower RVP blendstock is not needed for blending 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.756

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).757 Accounting for ethanol's lower energy density adds about 80 cents per each
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 (see
Chapter 6.3 for additional discussion of the fuel economy effect).

9.1.3.2 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.758 759 Although
ICF/Mathpro considered the possibility of the oil industry taking advantage of that property, they
concluded that most markets are not cetane limited and the oil industry likely would not take
advantage of this property of biodiesel and renewable diesel.760 At this time, we do not have any

755 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 E10 blendwall.

756 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.

757 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

758 Animal Fats for Biodiesel Production; Farm Energy; January 31, 2014.

759 McCormick, Robert; Renewable Diesel Fuel; NREL; July 18, 2016.

760 Modeling a No-RFS Case; ICF Incorporated; Work Assignment 0,1-11, EPA contract EP-C-16-020; July 17,
2018.

274

-------
evidence that the oil industry is 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 these pour point
additives is primarily a cold weather wintertime issue and likely contributes to the lower
observed blending rates in the winter compared to the summer, particularly in northern areas. For
the analysis, no additive costs were included for biodiesel because we do not have a good
estimate for additive costs for biodiesel, nor how biodiesel additive costs compares to that for
petroleum diesel.

As with ethanol, the societal cost of using biodiesel and renewable diesel must include
their lower energy density (fuel economy effect). Biodiesel and renewable diesel have somewhat
lower energy density than petroleum diesel and when accounting for this fuel economy effect
adds about 18 and 11 cents per gallon to their societal cost, respectively.

9.1.4 Distribution and Retail Costs

In this section, we evaluate the costs of distributing biofuels from the places where they
are produced to retail stations well as the costs of dispensing these fuels at the retail stations.

9.1.4.1 Ethanol

9.1.4.1.1 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, 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. Thus,
those capital costs are still being accounted for in these market prices that we use.

As part of the effort by ICF/Mathpro (which we also term ICF or Mathpro in this chapter)
to estimate use of renewable fuels, ICF estimated distribution costs for ethanol and biodiesel.
These cost estimates were also used for the distribution cost analysis for this rulemaking.761 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

761 Modeling a No-RFS Case; ICF Incorporated; Work Assignment 0,1-11, EPA contract EP-C-16-020; July 17,
2018.

275

-------
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
distribution cost analysis assumes 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, that 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.
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.762 Table 9.1.4.1.1-1 provides the estimate of ethanol distribution costs for the
various parts of the country estimated by ICF, and as revised to remove the 7 c/gal transportation
cost.

762 Rail congestion, cold weather raise ethanol spot prices; US Energy Information Administration; April 3, 2014.

276

-------
Table 9.1.4.1.1-1: Ethanol Distribution Costs for Certain Cities or Areas

Location

Distribution Cost (0/g) to:

Total (c/gal)

Hub/Terminal

Blending
Terminal

PADD

Area

To
Chicago

From
Chicago

ICF Estimate

Revised
Estimate

PADD 1

Florida/Tampa

7.0

17.8

11.0

35.8

28.8

S outheast/ Atl anta

11.7

11.0

29.7

22.7

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

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

PADD 5

Los Angeles

16.4

9.0

32.4

25.4

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 PADD 1
and 5 to derive a PADD-average ethanol distribution cost for those PADDs. Table 9.1.4.1.1-2
summarizes the estimated average ethanol distribution cost by PADD, and the average for the
U.S adjusted to 2021 dollars.

Table 9.1.4.1.1-2: Average Ethanol Distribution Cost by PADD and the U.S

Gasoline

Average Ethanol

Volume

Distribution Cost

Region

(kgals/day)

(Wgal)

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 2021$

19.7

9.1.4.1.2 Retail Costs

Retail costs for El0 have been sunk for many years and no additional retail costs are
assumed for E10. However, this is not the case for El 5 and E85. Additional investments are
needed to make them available at retail. The El5 and E85 volumes that we are using in this costs
analysis are summarized in Chapter 5.5.4.

277

-------
The retail costs for El 5 and E85 are estimated based on the investments that are needed
to be made to offer a higher volume of such ethanol blends. For estimating retail costs, the
necessary retail station changes and their associated installation costs for offering El5 and E85
must be estimated. To do so, we conferred with EPA's Office of Underground Storage Tanks on
what might be considered "typical" for El5 and E85 installations for a typical sized retail station
selling higher ethanol blends. The El 5 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 El 5 - each dispenser is estimated to cost $20,000 for a total cost
of $80,000 (assuming only 4 dispensers for El5 at a retail outlet, with other dispensers still
devoted to E10 and other fuels).763 The E85 stations are also assumed to have an existing
underground storage it could use for storing E85, but would require some equipment costing
$14,500 to allow the very high ethanol concentration to be stored in that tank. The E85 station
would also be required install a new E85-compatible dispenser, costing $20,000, for a total cost
of $34,500 (assuming only one dispenser at a retail outlet is provided for E85).764 Service
stations can incur costs which are higher or lower than the costs we estimate for offering El 5 and
E85. If the service station already has dispensers that can offer an additional fuel, then perhaps
only a few thousand dollars would need to be spent to make the dispenser parts compatible with
ethanol. On the other hand, if the service station needs the new dispensers and also needs to
install a separate storage tank to store the El5 or E85, then the installation costs would be much
higher.

To estimate the per-gallon cost, it is necessary to estimate the volume of E85 and E15
sold at each station which can be estimated based on El 5 and E85 throughput volumes sold in
Iowa.765 Retail stations offering E15 are estimated to sell 187 thousand gallons of E15 per year
while each retail station offering E85 are estimated to sell 64 thousand gallons of E85 per
year.766 Using the amortization factor shown in Table 9.1.2.1.1-1, and amortizing these retail
costs over the incremental volume of ethanol in El 5 and E85 above the E10 they displace, which
is 5% for El 5 and 64% for E85, covering the cost of capital for the retail equipment adds 41 and
7 cents per gallon to the ethanol portion of E15 and E85, respectively, for both 2021 and 2022.

763 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.

764 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.

765 Awan, Hamid, Internal Services Division, Iowa Department of Revenue; 2020 Retailers Fuel Gallons Annual
Report; April 2021.

766 For this portion of the costs analysis only (volumes of E15 and E85 per station used solely for amortizing the
retail costs), we used the approach described in this paragraph, which is different from the approach we used to
derive the total E15 and E85 volumes in Chapter 5.5.4. We acknowledge this discrepancy, but we find that this
ultimately is likely to have only very small impacts on the cost analysis and no impact on our final judgment in this
rulemaking. Specifically, this aspect of the cost analysis is based on E15 and E85 throughput data from a single
state. However, the most comprehensive data available (data from 19 states versus 1) for E15 and E85 throughput
volumes at retail stations is the Biofuel Infrastructure Partnership (BIP) data we obtained from USD A which is
summarized in Chapter 5.5.4. The BIP data shows lower throughput volumes for E15, and higher throughput
volumes for E85, than the values used for estimating per-gallon retail costs for the cost analysis. Thus, if the BIP
data had been used for estimating E15 and E85 retail costs, E15 per-gallon retail costs would be somewhat higher,
and E85 per-gallon retail costs would be somewhat lower, than those used for the cost analysis. On average, these
adjustments would offset each other and combined with the fact that E15 and E85 volumes are small to begin with,
the net effect on costs would be very small.

278

-------
The total volumes of E15 and E85 for estimating the total capital cost is from Tables 5.5.3-1 and
2.

Table 9.1.4.1.2-1 summarizes the increased volume of E15 and E85 and the associated
estimated capital cost.

Table 9.1.4.1.2-1: Volumes and Capital Costs for E15 and E85 for New Retail
Infrastructure

2021 vs
2020

2022 vs
2021

2021 & 2022
vs 2020

Volume (MMgals/yr)a

E15

361

440

E85

326

339

Change in Volume (MMgals/yr)

E15

120

E85

Capital Cost ($million) (2021$)

E15

142

186

E85

a Fuel volumes, not ethanol volumes.

9.1.4.2 Biodiesel and Renewable Diesel Distribution Costs

The biodiesel and renewable diesel distribution costs were determined by ICF under
contract to EPA based on an estimate of how much biodiesel and renewable diesel was moved by
rail and by truck, within each PADD, and between PADDs.767 While biodiesel and renewable
diesel 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. Table 9.1.4.2-1 summarizes the biodiesel and
renewable diesel distribution costs for each PADD taking into account the amount of fuel which
is distributed within PADDs and between PADDs, and shows the national average distribution
cost and that average cost adjusted to 2021 prices.

767 Modeling a No-RFS Case; ICF Incorporated; Work Assignment 0,1-11, EPA contract EP-C-16-020; July 17,
2018.

279

-------
Table 9.1.4.2-1: 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 2021$

19.2

The 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,
because much of the renewable diesel is expected to be distributed to the West Coast to help
meet the Low Carbon Fuel Standard programs there. Since renewable diesel is most likely to be
moved by ships, distribution costs are estimated to be 4 cents per gallon less than that for
biodiesel, or about 19.2 cents per gallon on average.

9.1.4.3 Biogas
9.1.4.3.1 Distribution Costs

Biogas which is gathered from the landfill off-gassing and cleaned up must then be
transported to where it can be used. Typically, this gas will end up in a nearby natural gas
pipeline, but it also could be, in some rare cases, 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 have to be moved
5 miles to access commercial natural gas pipeline. For installing each mile of pipeline, it is
estimated to cost $1 million, and adds up to $5.3 million in 2021 dollars for the entire 5 mile
pipeline.768 A typical volume case was modeled of 600 standard cubic feet per minute to estimate
the cost for a typical sized landfill.769 When the pipeline capital costs are amortized over that

768 Landfill Gas Energy Cost Model (LFGcost-Web); Version 3.5; https://www.epa.gov/lniop

769 Economic Analysis of the US Renewable Natural Gas Industry; The Coalition of Renewable Natural Gas;
December 2021.

280

-------
typical volume of landfill gas, the pipeline capital cost is estimated to cost $1.84 per million
BTU.770

Once the biogas is transported through the new pipeline to the natural gas pipeline, it
incurs a cost for distribution through an existing natural gas pipeline. Landfills are located near
urban areas which are destination areas for natural gas pipelines. This means that the distribution
costs in the natural gas pipeline would be less than that for natural gas which is being distributed
long distances from natural gas production areas. Natural gas will incur both variable and fixed
operating costs in the upstream pipelines, which biogas 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 biogas would be injected into a natural gas pipeline at least
large enough to serve commercial consumers, the biogas 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 9.2.2-2, distribution of natural gas to
commercial consumers is estimated to cost $4.72 and $5.10 per million BTU in 2021 and 2022,
respectively. 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
biogas. For this reason, half of the commercial natural gas distribution cost, or about $2.3 per
million BTU, is assumed to apply to biogas for distribution to the natural gas pipeline.

While this cost analysis assumes the biogas is being produced at landfills, it is
worthwhile to consider the situation other biogas producers are likely to be faced with to
distribute their biogas. Like landfills, biogas 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 biogas producers using ag waste digesters would likely be higher. Some of these rural
locations may be so remote that the biogas could be considered stranded and not readily
available for use as transportation fuel, although such stranded locations could perhaps still
provide biogas to local truck fleets which distribute agricultural products.771

9.1.4.3.2 Retail Costs

Retail facilities to dispense biogas are by comparison 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 $4.78 per million BTU, so this was used for the biogas
retail cost.772

770 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 a 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.

771 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.

772 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.

281

-------
9.2 Gasoline, Diesel Fuel, and Natural Gas Costs
9.2.1 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 pandemic, Energy Information
Administration (EIA) data shows that gasoline and diesel fuel demand are lower now and as of
early 2021, only diesel fuel is expected to increase above previous levels.773 Furthermore, light-
duty and heavy-duty greenhouse gas standards will continue to phase-in, continuing to reduce
transportation fuel demand.774 775 776 777 Thus, we would not anticipate there to be refined
product investment regardless of the RFS volumes and thus no savings that would offset
renewable fuel investments.

9.2.1.1 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 EIA's Short-Term Energy Outlook.778 The estimated
production costs in 2022 are adjusted to 2021 dollars and summarized in Table 9.2,1,1-1. Table
9.2.1.1-1 summarizes the projected gasoline and diesel fuel wholesale prices which we equated
to their production costs.

773 Short-Term Energy Outlook; Table 4a. U.S. Petroleum and Other Liquids Supply, Consumption and Inventories;
January 6, 2022.

774Environmental 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.

775 Revised 2023 and Later Model Year Light-Duty Vehicle Greenhouse Gas Emissions Standards; Environmental
Protection Agency, December 30, 2021.

776 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.

777 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.

778 Short-Term Energy Outlook; Table 4a. U.S. Petroleum and Other Liquids Supply, Consumption and Inventories;
January 2022.

282

-------
Table 9.2.1.1-1: Estimatet

Gasoline Production Costs

Gasoline Costs

Diesel Fuel Costs

2021

2022

2022 in
2021$

2021

2022

2022 in
2021$

Brent Crude Oil Prices ($/bbl)

70.89

74.95

73.47

70.89

74.95

73.47

Wholesale Prices assumed to
be Production Costs ($/gal)

2.18

2.20

2.16

2.10

2.25

2.21

9.2.1.2 Natural Gas Production Cost

For estimating the relative cost of biogas 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 Short-Term Energy Outlook, EIA projects the natural gas spot price
for Henry Hub to average $4.06 and $3.94 per thousand cubic feet for 2021 and 2022,
respectively.779 When adjusted to 2021 dollars, the 2022 natural gas price decreases to $3.86 per
thousand cubic feet. The Henry Hub spot price most closely represents the natural gas field price,
and thus is a proxy for its production cost.

9.2.2 Gasoline, Diesel Fuel and Natural Gas Distribution and Blending Cost

Gasoline and diesel fuel distribution costs to terminals are estimated as the difference
between wholesale prices (see above Table 9.2.1.1-1) and terminal prices (which we estimated
based on 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 costs of distribution from the terminal and retail stations. 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 55 c/gal for gasoline and 64 c/gal for diesel fuel, from
gasoline and diesel fuel retail prices. Then, we took the difference between retail prices (less
taxes) and terminal prices (again, estimated as sales for resales prices) to estimate the costs for
terminal and retail distribution. The terminal and retail distribution costs for gasoline and diesel
fuel are estimated to be 20 and 40 cents per gallon for both gasoline and diesel fuel, respectively.
The 2022 estimated gasoline and diesel cost are adjusted to 2021 dollars. These costs are
summarized in Table 9.2.2-1

779 Short-Term Energy Outlook; Table 5b. US Reginal Natural Gas Prices; Wholesale Spot; January , 2021

283

-------
Table 9.2.2-1: Projected Gasoline and Diesel Production Costs ($/gal)

Gasoline Costs

Diesel Fuel Costs

2021

2022

2022 in
2021$

2021

2022

2022 in
2021$

Brent Crude Oil Prices

70.89

74.95

73.47

70.89

74.95

73.45

Retail Cost minus taxes

2.44

2,46

2.42

2.58

2.72

2.68

Terminal and Retail

0.20

0.40

Costs

Terminal Costs

2.23

2.25

2.21

2.18

2.33

2.28

Distribution Cost

0.05

0.08

Production Cost

2.18

2.20

2.16

2.10

2.25

2.21

(from Table 9.2.1.1-1)

EIA projects natural gas prices downstream of natural gas production fields which can be
used to estimate natural gas distribution costs. 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.780

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 Henry Hub prices from the commercial prices. Thus, the Short-
Term Energy Outlook projected Henry Hub prices were subtracted from the commercial prices
for 2021 and 2022. Table 9.2.2-1 summarizes the calculation of natural gas distribution cost. To
put the natural gas costs on the same footing as the biogas, we also add $4.61 per million BTU
for retail costs.781

Table 9.2.2-2: Natural Gas Distribution Cost ($/thousand cubic feet)

2021

2022

2022 in 2021$

Commercial Prices

8.78

9.14

8.96

Henry Hub Prices

4.06

3.94

3.86

Pipeline Distribution Costs

4.72

5.20

5.10

Retail Costs

+4.78

Total Distribution Costs

9.50

9.98

9.88

780 Taxes are not included in social cost estimates because they are not true costs, only transfer payments.

781 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.

284

-------
9.3 Volumes and Energy Density Related Fuel Economy Cost

To estimate the change in petroleum fuel volume that would occur with these changes in
renewable fuels volumes and to estimate the fuel economy cost for the summary of individual
renewable fuel costs summarized in Chapter 9.4.1, it was necessary to estimate the energy
density of each fuel. Table 9.3-1 contains the estimated energy densities for the various
renewable fuels and petroleum fuels analyzed for this cost analysis.

Table 9.3-1: 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

E85 (E74) Gasoline

86,285

Biodiesel

119,550

Renewable Diesel

122,887

Crude Oil

129,670

a From Chevron Paper.782

When accounting for the difference in energy density, changes in ethanol volume affect
gasoline volume, changes in renewable diesel and biodiesel volumes affect diesel fuel volume,
and changes in biogas volume affect natural gas. 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. The renewable fuel volumes are from Table 2.2-3. The gasoline
and diesel volumes are estimated based on their energy content relative to the energy content and
volumes of renewable fuel which are displacing them, as shown in Table 9.3-2, and in Table 9.3-
3 for the supplemental standard. For example, E10 gasoline contains about 3% less energy
content than EO gasoline, and the cost of lower energy dense gasoline is paid by consumers
through lower fuel economy and more frequent refueling. However, the lower energy density of
E10 is due to the lower energy density of ethanol, which contains about 33% lower energy per
volume than the gasoline it is displacing. While this energy cost is not separately calculated for
the cost of the program, the costs are separately calculated and shown as its own line item for
each renewable fuel in Chapter 9.4.1.

782 Diesel Fuels Technical Review; Chevron Global Marketing; 2007.

285

-------
Table 9.3-2: Renewable Fuel and Fossil Fuel Volume C

langes (million gallons)

Change in Renewable Fuel
Volume

Fossil Fuel
Displaced

Chang

e in Fossil Fuel Volume

2021

2022

2021 & 2022

2021

2022

2021 & 2022

Cellulosic biofuel - Total

4269

5079

9348

-4269

-5079

-9348

CNG - landfill biogas
(MMFT3)

4269

5079

9348

Natural Gas

-4269

-5079

-9348

Non-cellulosic adv. - Total

138

816

954

-171

-765

-936

Biodiesel - Soy Oil

-107

-71

Diesel Fuel

100

-34

Renewable Diesel - Soy Oil

371

729

1100

Diesel Fuel

-355

-698

-1052

Ethanol - sugarcane

-125

-75

Gasoline

-33

Conventional - Total

1221

440

1660

-817

-297

-1114

Ethanol - Corn

1221

432

1652

-817

-289

-1107

Ethanol - Corn - E10

1141

410

1551

Gasoline

-764

-274

-1038

Ethanol - Corn - E15

Gasoline

-26

-8

-34

Ethanol - Corn - E85

Gasoline

-28

-7

-35

Renewable Diesel -Palm

Diesel Fuel

-8

Change in Biogas Volume
(MMFT3)

4269

5079

9348

Change in Ethanol Volume

1016

460

1476

Change in Biodiesel Volume

-107

-71

Change in Renewable Diesel
Volume

371

737

1108

Change in Gasoline Volume

-734

-323

-1056

Change in Diesel Fuel Volume

-255

-740

-994

Change in Natural Gas Volume
(MMFT3)

-4269

-5079

-9,348

Table 9.3-3 Supplemental Standard Renewable Fuel and Petroleum Fuel Volume Changes

Change in Renewable Fuel Volume

Change in Diesel Fuel Volume

2021

2022

2021 &
2022

2021

2022

2021 &
2022

Supplemental Std. Renewable
Diesel - Palm

147

Diesel Fuel

-141

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 2021 Annual Energy Outlook.783
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, almost 87 percent of that gasoline or diesel reduction would be
attributed to reduced crude oil imports and 13 percent of the reduction would be attributed to
reduced imported refined product. Based on these correlations, Table 9.3-4 summarizes the
projected change in petroleum imports expected from the increased consumption of renewable
biofuels from the 2020-22 standards, and Table 9.3-5 shows the same information but also

783 AEO 2021 Change in product demand on imports; docketed spreadsheet.

286

-------
accounts for the Supplemental Standard. In both tables, we consider the projected change in
imported petroleum products only as well as the projected change in all imported fuels, including
imported renewable diesel and imported sugar cane ethanol. The change in crude oil volume and
imported petroleum products is used for the energy security analysis contained in Chapter 4.

Table 9.3-4: Projected Change in Petroleum Imports Due to Increased Renewable Fuel
Consumption (million gallons)

2021

2022

2021 &
2022

Projected Change in Imported
Petroleum Only

Change in Imported Gasoline

-93

-41

-134

Change in Imported Diesel Fuel

-32

-94

-126

Total Change in Crude Oil

-785

-888

-1,673

Change in Domestic Crude Oil

-5

-6

-11

Change in Imported Crude Oil

-779

-882

-1,661

Projected Change in Imported
Petroleum accounting for Imported
Renewable Diesel Fuel and Sugar
Cane Ethanol3

Change in Imported Gasoline and Imported
Sugarcane Ethanol

-177

-7

-184

Change in Imported Diesel Fuel and
Imported Renewable Diesel

-32

-86

-118

Total Change in Crude Oil

-785

-888

-1,673

Change in Domestic Crude Oil

-5

-6

-11

Change in Imported Crude Oil

-779

-882

-1,661

a When accounting for the imported renewable diesel and sugar cane ethanol volumes along with the imported
diesel and gasoline volumes, the imported renewable diesel and sugar cane ethanol volumes are diesel and gasoline-
equivalent volumes.

Table 9.3-5: Projected Change in Petroleum Imports Due to Increased Renewable Fuel

Consumption; accounts for the Supplemental Standard imports (mil

ion gallons)

2021

2022

2021 &
2022

Projected Change in Imported
Petroleum Only

Change in Imported Gasoline

-93

-41

-134

Change in Imported Diesel Fuel

-32

-111

-144

Total Change in Crude Oil

-785

-1010

-1,795

Change in Domestic Crude Oil

-5

-7

-12

Change in Imported Crude Oil

-779

-1003

-1,782

Projected Change in Imported
Petroleum accounting for Imported
Renewable Diesel Fuel and Sugar
Cane Ethanol3

Change in Imported Gasoline and Imported
Sugarcane Ethanol

-177

-7

-184

Change in Imported Diesel Fuel and
Imported Renewable Diesel

-32

Total Change in Crude Oil

-785

-1,010

-1,795

Change in Domestic Crude Oil

-5

-7

-12

Change in Imported Crude Oil

-779

-1003

-1,782

9.4 Costs

9.4.1 Individual Fuels Cost Summary

Table 9.4.1-1 summarizes the estimated overall societal costs (including production,
distribution, blending, and fuel economy) for the renewable fuels analyzed for this rulemaking

287

-------
for the year 2021.784 These costs do not include the per-gallon federal cellulosic and biodiesel tax
subsidies, nor do they consider taxes or tax subsidies more generally, as these are transfer
payments which are not included in these costs. Nor do these costs consider federal state or local
infrastructure support funding (e.g., the USDA Higher Blends Infrastructure Incentive Program;
HBIIP) supporting E85 and El 5 retail station equipment, as these subsidies are also transfer
payments.785 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 fuels impact on fuel economy. While these costs may not
always be reflected in the sales prices among the market participants (e.g., refiners sell gasoline
based on volume, not energy content), the varying impacts on fuel economy among the fuels
result in different costs to consumers in operating their vehicles. The cost associated with the
impact of the renewable fuel on fuel economy costs are determined relative to the fuels they are
assumed to displace; the ethanol fuels are assumed to displace gasoline, the biodiesel and
renewable diesel are assumed to displace diesel fuel, and biogas is assumed to displace natural
gas.786 To the extent that RINs from biogas incentivize some incremental sales of trucks capable
of refueling on biogas (essentially CNG/LNG trucks) at the expense of diesel fueled trucks, then
some biogas could also displace diesel fuel, but this is expected to be a relatively minor portion
for the volumes and timeframe of this rule. Nevertheless, we also provide a comparison of the
costs of biogas to the costs of diesel fuel.

784 Since this is solely for illustrative purposes, the 2022 costs are not calculated and shown in this format.

785 Higher Blends Infrastructure Incentive Program; United States Department of Agriculture (USDA);
https://www. rd.usda. eov/fabiip

786 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.

288

-------
Table 9.4.1-1: Renewable Fuels Costs Estimated for 2021 ($/gal unless otherwise noted;
2021$)

Production
Cost

Blending
Cost

Distribution
Cost

Retail
Cost

Fuel
Economy
Cost3

Total
Cost

Corn

Starch

Ethanol

E10

1.92

-0.65

0.40

0.8 lb

2.48

E15 with retail
costc

1.92

0.40

1.24

0.81

4.37

E15 with Zi
retail costc

1.92

0.40

0.62

0.81

3.75

E85 with retail
costc

1.92

0.40

0.09

0.81

3.22

E85 with '/i
retail costc

1.92

0.40

0.05

0.81

3.17

Renewable
Diesel

Soybean Oil

5.87

0.59

0.11

6.57

Palm Oil

5.10

0.59

0.11

5.81

Biodiesel

Soy Oil

5.36

0.59

0.18

6.13

Sugarcane
Ethanol

E10

1.77

-0.65

0.40

0.81

2.33

Cellulosic

Biogas

($/gal ethanol)

0.38

0.31

0.35

0.99

2.03

Biogas

($/thousand ft3)

5.15

4.28

4.78

0.00

14.21

a Fuel Economy cost is per fuel being displaced - ethanol and FT naphtha displaces gasoline, renewable and FT
diesel displaces diesel fuel, and biogas represented as ethanol-equivalent volume is relative to diesel fuel.
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, they do not consider the fuel economy effect and this distinction is important for
understanding ethanol's relative economic viability in the marketplace.

c For modeling the social cost of E15 and E85 shown in Table 9.4.2-1, 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 in Table 9.4.1-1, the cost for E15 and E85 does not include any blending value for E10
BOBs to represent the marginal cost for the ethanol volume above E10.

The distribution costs for the biofuels are 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 2.38 to 2.58 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 9.4.1-2 summarizes production and distribution costs for each category of fossil
transportation fuel—gasoline, diesel fuel, and natural gas. For gasoline and diesel, production

289

-------
costs are based on prices in the January 2022 STEO.787 Natural gas spot prices from the January
2022 STEO are used to represent both feedstock and production costs.

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
entities and the projected natural gas spot price, which reflects the price where it is produced.

Table 9.4.1-2: Gasoline, Diese

Fuel, and Natura

Gas Costs for 2021 (2021$)

Production Cost

Distribution Cost

Total Cost

Gasoline ($/gal)

2.18

0.26

2.44

Diesel Fuel ($/gal)

2.11

0.47

2.58

Natural Gas ($/million BTU)

4.06

9.79

13.85

Table 9.4.1-3 compares the data from Tables 9.4.1-1 and 2 to show the relative cost of the
renewable fuels with the petroleum fuels they displace.

Table 9.4.1-3: Comparison of Renewable Fuel Costs with Petroleum Fuel Costs for 2021
($/gal unless otherwise noted; 2021$)

Fossil
Fuel Total
Cost

Renewable
Fuel Total
Cost

Relative
Cost

E10

Relative to
Gasoline

2.44

2.48

0.04

El 5 with retail cost

2.44

4.37

1.93

El 5 with 1/2 retail cost

2.44

3.75

1.31

E85 with retail cost

2.44

3.22

0.78

E85 with l/2retail cost

2.44

3.17

0.74

Sugarcane E10

2.44

2.33

-0.11

Soybean Oil Renewable Diesel

Relative to
Diesel Fuel

2.58

6.57

3.99

Palm Oil Renewable Diesel

2.58

5.81

3.22

Soybean Oil Biodiesel

2.58

6.13

3.54

Biogas ($/million BTU)

Relative to
Natural Gas

13.85

14.21

0.36

9.4.2 Total Estimated Cost Impacts

In this section, we summarize the results of our analysis estimating the costs for changes
in the use of renewable fuels which displace gasoline, diesel fuel and natural gas. For this
analysis we considered the production, blending, distribution, and differences in energy density.

787 EIA, Short Term Energy Outlook 2021, Energy Information Administration, January, 2022;
https://www.eia.gov/oiitlooks/steo/archives/Jan22.pdf

290

-------
Table 9.4.2-1 summarizes the cost and cost savings of each biofuel fuel type compared to the
fossil fuel it is displacing, and Table 9.4.2-2 provides this information for the supplemental
standard.

291

-------
Table 9.4.2-1: Renewable Fuel and Petroleum Fuel Costs (million dollars; 2021$)

Renewable Fuel

Petroleum Fuel

Production

Distribution

Blending

Production

Distribution

Total

Cellulosic biofuel - Total

22.0

49.4

0.0

-17.3

-41.8

32.6

CNG - landfill biogas

22.0

49.4

0.0

-17.3

-41.8

32.6

Non-cellulosic adv. - Total

1377.9

105.2

81.6

-354.8

-99.1

1110.7

Biodiesel - soy oil

-575.1

-63.3

0.0

210.8

47.3

-380.3

Renewable Diesel - soy oil

2174.8

218.5

0.0

-748.1

-167.8

1477.4

2021

Ethanol - sugarcane

-221.7

-50.0

81.6

182.5

21.4

13.7

Conventional - Total

2343.6

508.2

-747.7

-1781.9

-209.0

113.2

Ethanol - corn

2343.6

508.2

-747.7

-1781.9

-209.0

113.2

Ethanol - Corn - E10

2190.7

456.8

-744.6

-1665.7

-195.3

42.0

Ethanol - Corn - E15

73.2

31.0

-2.4

-55.7

-6.5

39.7

Ethanol - Corn - E85

79.6

20.4

-0.8

-60.5

-7.1

31.6

Renewable Diesel - palm

0.0

Cellulosic biofuel - Total

26.1

34.4

-0.1

-19.6

-49.7

OO
00

CNG - landfill biogas

26.1

34.4

-0.1

-19.6

-49.7

OO
OO

Non-cellulosic adv. - Total

5395.7

471.7

-32.6

-1686.5

-355.0

3793.2

Biodiesel - soy oil

234.4

21.6

0.0

-75.3

-16.2

164.6

Renewable Diesel - soy oil

5072.6

430.0

0.0

-1539.0

-330.3

3633.3

2022

Ethanol - sugarcane

88.7

20.0

-32.6

-72.2

-8.6

-4.7

Conventional - Total

1025.3

183.5

-271.9

-640.5

-77.6

218.8

Ethanol - Corn

984.5

178.8

-271.9

-623.6

-73.9

193.7

Ethanol - Corn - E10

933.9

164.0

-267.3

-591.6

-70.1

168.9

Ethanol - Corn - E15

27.1

9.7

-4.4

-17.2

-2.0

13.2

Ethanol - Corn - E85

23.4

5.1

-0.2

-14.9

-1.8

11.7

Renewable Diesel - palm

40.8

4.7

0.0

-16.9

-3.6

25.0

Cellulosic biofuel - Total

48.1

83.8

-0.1

-36.9

-91.5

23.8

CNG - landfill biogas

48.1

83.8

-0.1

-36.9

-91.5

23.8

Non-cellulosic adv. - Total

6773.6

576.8

48.9

-2041.3

-454.1

4903.9

Biodiesel- soy oil

-340.7

-41.7

0.0

135.5

31.1

-215.7

Renewable Diesel - soy oil

7247.4

648.5

0.0

-2287.1

-498.1

5110.6

2021
&

2022

Ethanol - sugarcane

-133.0

-30.0

48.9

110.3

12.8

9.0

Conventional - Total

3368.9

691.7

-1019.7

-2422.4

-286.5

332.0

Ethanol - Corn

3328.0

687.0

-1019.7

-2405.5

-282.9

307.0

Ethanol - Corn - E10

3124.6

620.8

-1011.9

-2257.3

-265.5

210.8

Ethanol - Corn - E15

100.3

40.7

-6.8

-72.9

-8.6

52.9

Ethanol - Corn - E85

103.0

25.5

-1.0

-75.4

-8.9

43.3

Renewable Diesel - palm

40.8

4.7

0.0

-16.9

-3.6

25.0

292

-------
Table 9.4.2-2: Renewable Fuel and Petroleum Fuel Costs for the Supplemental Standard
(million dollars; 2021$)

Renewable Fuel

Petroleum Fuel

Year 2022

Production

Distribution

Blending

Production

Distribution

Total

Supplemental Std. RD
Palm

750.3

86.7

0.0

-310.2

-66.6

460.2

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 EIA's Short Term Energy
Outlook and summarized in Table 9.4.2-3.788

Table 9.4.2-3: Total Gasoline, Diesel Fuel and Natural Gas Volumes

2021

2022

Units

Gasoline Volume

134.75

138.89

Billion gallons

Diesel Volume

60.55

62.39

Billion gallons

Natural Gas Volume

82.96

82.77

Trillion cubic feet

The costs are aggregated for each fossil fuel type and costs expressed as per-gallon and
per thousand cubic feet costs in Table 9.4.2-4 for 2021, and the 2022 per-volume costs calculated
and shown incremental to 2021 and the combined 2021 and 2022 costs are also shown
incremental to the 2020, the baseline year. Table 9.4.2-5 includes similar information, but also
includes the supplemental standard costs.

Table 9.4.2-4: Total Annual Rule Cost without the Supplemental Standard (2021$)

Total Cost
(Million $)

Per Unit
Costs

Units

2021

Gasoline

127

0.09

c/gallon gasoline

Diesel Fuel

1097

1.81

c/gallon diesel fuel

Natural Gas

0.00108

$/1000cf natural gas

Total

1257

2022
relative
to 2021

Gasoline

189

0.14

c/gallon gasoline

Diesel Fuel

3,823

6.13

c/gallon diesel fuel

Natural Gas

-8.8

-0.00029

$/1000cf natural gas

Total

4,003

2022
relative
to 2020

Gasoline

316

0.23

c/gallon gasoline

Diesel Fuel

4920

7.89

c/gallon diesel fuel

Natural Gas

0.00079

$/1000cf natural gas

Total

5260

788 EIA, Short Term Energy Outlook 2021, Energy Information Administration, January, 2022;
https://www.eia.gov/oiitlooks/steo/archives/Jan22.pdf

293

-------
Table 9.4.2-5: Total Annual Rule Cost wit

Total Cost
(Million $)

Per Unit
Costs

Units

2021

Gasoline

127

0.09

c/gallon gasoline

Diesel Fuel

1097

1.81

c/gallon diesel fuel

Natural Gas

0.00108

$/1000cf natural gas

Total

1257

2022
relative
to 2021

Gasoline

189

0.14

c/gallon gasoline

Diesel Fuel

4,283

6.86

c/gallon diesel fuel

Natural Gas

-8.8

-0.00029

$/1000cf natural gas

Total

4,463

2022
relative
to 2020

Gasoline

316

0.23

c/gallon gasoline

Diesel Fuel

5380

8.62

c/gallon diesel fuel

Natural Gas

0.00079

$/1000cf natural gas

Total

5720

1 the Supplemental Standard (2021$)

The estimated costs in 9.4.2-4 and 5 are much higher than those in the proposed rule. The
primary reason for this is the much higher corn and vegetable oil prices. Also, renewable diesel
plant capital costs were increased based on more recent information and the hydrogen costs for
renewable diesel plants were updated higher based on a better estimate of hydrogen plant
production costs.

9.4.3 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 using two different methodologies: based on the cost of renewable fuels
less available federal tax credits and based on the cross-subsidy implemented through the RIN
system.

The RFS program operates as a cross-subsidy, effectively increasing the price of
petroleum-based fuels through the renewable volume obligations and subsidizing the price of
renewable fuels by allowing renewable fuel producers to generate a RIN that can be separated
from the renewable fuel and sold under specified conditions. Retail fuel prices are a combination
of many inter-related factors, including the cost of the component fuels, the RIN price increase to
the petroleum-based fuel and the RIN price decrease to the renewable fuel for the particular
blends, the price discounts required by consumers to purchase increasing volumes of renewable
fuels, and many others. We have therefore been unable to precisely and comprehensively project
how the volumes we are finalizing in this rule will impact retail fuel prices but have presented
estimates of what potential price impacts might be using the two methodologies described in the

294

-------
previous paragraph. We have also used estimates of the fuel price impacts of this rule to estimate
the cost to transport goods.789

9.4.3.1 Fuel Price Impacts Based on Renewable Fuel Costs

In this section we have estimated the component of fuel price impacts attributable to the
use of renewable fuels. In this section we have assumed that the cost of increasing biodiesel and
renewable diesel are reflected in the price of diesel fuel and the cost of increasing ethanol is
reflected in the price of gasoline. We have not accounted for the fuel price impacts attributable to
the RIN obligations on petroleum-based fuels or the separable RIN value of qualifying
renewable fuels. The fuel price impacts estimated in this section use the individual fuel costs
from Chapter 9.4.1. Since the costs to consumers of transportation fuel are affected by the
cellulosic and biodiesel subsidies, we adjusted the fuel costs to account for the cellulosic biofuel
(1.01 dollar per gallon) and biodiesel (1.00 dollar per gallon) fuel subsidies. We also note that
the estimates on fuel price impacts presented here are relative to the volumes of renewable fuel
consumed in 2020, consistent with the other analyses in this RIA. The fuel price impacts
estimated in this rule for 2022 are likely reflected in current fuel prices, and this final rule is
unlikely to further increase fuel prices.790

Table 9.4.3.1-1 summarizes the contribution to retail price impacts arising from the cost
of the incremental biofuels (less the available federal tax credits) associated with the volume
requirements for 2021 and 2022relative to the 2020 renewable fuel volumes. The renewable fuels
costs are expressed relative to the petroleum fuel each renewable fuel is blended into - ethanol
costs relative to gasoline, renewable diesel and biodiesel relative to diesel fuel, and biogas
relative to natural gas. When we total the costs, we express the total renewable fuel costs over
both gasoline and diesel fuel since refiners may spread all their renewable fuels costs over their
entire gasoline and diesel fuel pools. This methodology does not consider RIN price impacts,
but rather estimates fuel price impacts based on the estimated cost impacts less the available
federal tax credits.

789 In the DRIA, we only calculated costs to consumers based on the fuel costs, and fuel cost plus federal subsidies,
methodologies. In the final rule, to support our response to comments regarding RIN price impacts on costs to
consumers, we have added a separate calculation of costs based on the RIN price impacts methodology.

790 As discussed further in RTC Section 6.3.4, renewable fuel use in the U.S. in the first three months of 2022
(January - March) is consistent with the rates needed to meet the volumes we are finalizing in this rule for 2022
based on observed seasonal trends in 2021. Thus, the market appears to have anticipated the volumes we are
finalizing in this rule, which are very similar to the proposed volumes. Any increases in demand for renewable fuel
that result from this rule are therefore likely already reflected in the current market prices for fuel and are not
expected to change significantly with the finalization of this rule.

295

-------
Table 9.4.3.1-1: Estimated Effect of 2021 and 2022 Biofuel Volume Requirements
(including supplemental standard) on Fuel Costs including Subsidies (per-unit costs may

represent

prices, 2021$)

Total Cost
(Million $)

Per Unit
Costs

Units

2021
relative
to 2020

Gasoline

127

0.091

0/gallon gasoline

Diesel Fuel

834

1.336

0/gallon diesel fuel

Natural Gas

-26

-0.00086

$/1000cf natural gas

Total

935

0.48

0/gallon gasoline and diesel fuel

2022
relative
to 2021

Gasoline

189

0.136

0/gallon gasoline

Diesel Fuel

3370

5.401

0/gallon diesel fuel

Natural Gas

-79

-0.0026

$/1000 cf natural gas

Total

3481

1.73

0/gallon gasoline and diesel fuel

2022
relative
to 2020

Gasoline

316

0.228

0/gallon gasoline

Diesel Fuel

4204

6.738

0/gallon diesel fuel

Natural Gas

-104

-0.00346

$/1000 cf natural gas

Total

4415

2.19

0/gallon gasoline and diesel fuel

9.4.3.2 RIN Price Impacts on Blended Fuel Prices

In this section we have estimated the impact of the final volumes on the price of various
blends of transportation fuel based on the cross-subsidy implemented through the RIN system.
For this analysis we have used the average RIN prices in 2021 along with the RVOs we are
establishing in this rule to estimates fuel price impacts of this rule. We have estimated the price
impacts relative to 2020, as well as the price impacts relative to a No RFS baseline where there
were not RFS requirements in 2021 or 2022 (in which case there would be no RVOs and no
RINs) so as to show the direct contribution of RIN prices on various fuel blends.791 Consistent
with our analysis of the impact of RIN prices on transportation fuel we have assumed that both
the cost and the value of the RIN are fully passed through to consumers.792

To estimate the fuel price impacts for various fuel blends using the RVOs and RIN prices
we first considered the increase in the percentage standards we are establishing in this rule.

These increases are shown in Table 9.4.3.2-1.

Table 9.4.3.2-1: Increases in RVOs by Category

791 As discussed in Chapter 2, we have been unable to determine the renewable fuel volumes that would occur in the
absence of the RFS program and thus for most of our analyses have been unable to compare the impacts to a "No
RFS" baseline. However, in the case of the fuel price impacts based on RIN prices the No RFS baseline is simple -
there would be no renewable volume obligations and the price for all RINs would be zero.

792 See "Denial of Petitions for RFS Small Refinery Exemptions," EPA-420-R-22-005, April 2022. We recognize
that for some fuels with smaller fuel volumes, such as E85, the wholesale and retail markets may not be completely
competitive, and thus the full value of the RIN may not be passed through to consumers. In this analysis we have
assumed that the full cost and value of the RIN is passed through to consumers for all fuel blends.

296

-------

Cellulosic
Biofuel

BBD

Other
Advanced

Conventional

2020 RVO

0.32%

2.30%

0.31%

7.89%

2021 RVO

0.33%

2.16%

0.51%

8.19%

2022 RVO

0.35%

2.33%

0.48%

8.57%

Increase from 2020 to 2021

0.01%

(0.14)%

0.20%

0.30%

Increase from 2020 to 2022

0.03%

0.17%

0.68%

We next calculated the price increase per gallon of petroleum-based fuel in 2021 and
2022 relative to a 2020 baseline, and for all three years relative to a scenario where there was no
RFS program in each year. These price increases were calculated by multiplying the RVO
increase in each category by the RIN price for that category of RIN. Because we cannot project
RIN prices in 2022 for this analysis we used the average RIN price in 2021 as the RIN price for
all three years. The average RIN price for each RIN category was based on data from the EPA
Moderated Transaction System (EMTS). The average RIN prices and the cost increases per
gallon of petroleum fuel are shown in Tables 9.4.3.2-2 and 9.4.3.2-3.

Table 9.4.3.2-2: Average RIN Prices in 2021793

Cellulosic (D3)

BBD (D4)

Other Advanced (D5)

Conventional (D6)

$2.75

$1.46

$1.44

$1.35

Table 9.4.3.2-3: RIN Price Impact Per Gallon of Obligated Petroleum Based Fuel (Gasoline
and Diesel Fuel)

Cellulosic

BBD

Other
Advanced

Conventional

Total

2021 relative to 2020

$0,000

$(0,002)

$0,003

$0,004

$0,005

2022 relative to 2020

$0,001

$0,000

$0,002

$0,009

$0,013

2020 relative to No RFS

$0,009

$0,034

$0,004

$0,107

$0,153

2021 relative to No RFS

$0,009

$0,032

$0,007

$0,111

$0,159

2022 relative to No RFS

$0,010

$0,034

$0,007

$0,116

$0,166

Finally, we estimated the RIN price impact for each of the comparisons presented in
Table 9.4.3.2-3 on a number of different fuel blends. To calculate the price impact on each fuel
blend we summed the price impact on the petroleum component of the fuel and the price impact
on the renewable component of the fuel. For the estimates of the price impacts in 2021 and 2022
relative to the 2020 baseline there was no change in the price impact on the renewable
component of the fuel because we are assuming the same RIN price for all three years and the
quantity of ethanol in E10, for example, does not change from 2020 to 2022. In these scenarios
the estimated price impact is simply the price impact per gallon of petroleum fuel multiplied by
the quantity of petroleum fuel in each fuel blend (e.g., 0.9 gallons for E10, 0.85 gallons for E15,
etc.). For the estimates of the price impacts relative to the No RFS baselines we considered both
the price impact on the petroleum and renewable components of the fuel blend. For example, to

793 Prices shown are the volume weighted averages of all 2021 RINs traded in 2021. RIN trades reported at less than
$0.05 were excluded from these averages as in most cases they likely reflect placeholder values for internal
company transfers.

297

-------
calculate the RIN price impact on El0 in 2020 relative to the No RFS baseline we added the
price impact on the petroleum blendstock ($0.153 per gallon x 0.9 gallons petroleum blendstock)
to the price impact on the ethanol ($1.35 per D6 RIN x 0.1 RINs) for a total price impact of
$0.00 per gallon. For the diesel blends, we assumed that the biofuel was biodiesel and that each
gallon of biodiesel generates 1.5 D4 RINs. The results for various fuel blends for each scenario
are shown in Table 9.4.3.2-4.

Table 9.4.3.2-4: RIN Price Impacts on Various Fuel Blends

E10

E15

E85a

B20

2021 relative to 2020

$0.01

$0.00

$0.01

$0.00

2022 relative to 2020

$0.01

$0.00

$0.01

2020 relative to No RFS

$0.15

$0.00

$(0.07)

$(0.96)

$0.15

$0.04

$(0.32)

2021 relative to No RFS

$0.16

$0.01

$(0.07)

$(0.96)

$0.16

$0.04

$(0.31)

2022 relative to No RFS

$0.17

$0.01

$(0.06)

$(0.96)

$0.17

$0.05

$(0.31)

aAssumes E85 contains 74% ethanol, consistent with EIA assumptions

The projected fuel price impacts based on RIN prices differs from the projected fuel price
impacts based on renewable fuel costs in several important ways. Most importantly, the RIN
price impacts presented in this section only consider the impact on individual fuel blends (E10,
B5, etc.). These price impacts on various fuel blends do not address the price of increasing
renewable fuel use, for example by increasing the quantity of B5 consumed relative to the
volume of B0 consumed. The projection based on RIN prices also show similar price impacts for
all fuel blends in 2021 and 2022 relative to the 2020 volumes, while the projection based on
renewable fuel costs shows significantly higher impacts for diesel than for gasoline. This is
because the projection based on RIN prices accounts for the fact that all obligated fuels (gasoline
and diesel) are subject to the same obligations, which include obligations for all four categories
of renewable fuel. Conversely the projection based on renewable fuel costs attribute the entire
price impact for increased renewable fuel blending to the fuel that is blended with the renewable
fuel. For example, the entire price impact of increased use of ethanol is attributed to gasoline,
while the entire price impact of increased use of biodiesel and renewable diesel is attributed to
diesel.

9.4.3.3 Cost to Transport Goods

We next considered 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. We focused on
the estimated fuel price impacts presented in Tables 9.4.3.1-1 as these price impact estimates
incorporate the estimated impact across the entire diesel fuel pool, rather than just the RIN price
impacts for individual fuel blends. To account for the fact that the RFS obligations apply equally
to gasoline and diesel we have estimated a range of impacts of this rule on the cost to transport
goods. The range provided considers a case where the entire cost of increasing biodiesel and
renewable diesel (less available federal subsidies) is realized in the diesel fuel pool ($0.0674 per
gallon of diesel) and a case where the entire price cost of the volumes we are finalizing in this

298

-------
rule (less available federal subsidies) is divided between the gasoline and diesel fuel pools on a
per gallon basis ($0.0219 per gallon of gasoline and diesel).

We further considered the impact of fuel price increases on the price of goods based upon
a study conducted by the United States Department of Agriculture (USDA). The USDA 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 $30 per barrel to over $90 per barrel.794 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 $2.72/gal in 2022 as summarized in Table 9.2.1.1-
1 and adding 60 cents per gallon taxes to it, the $0.0674 per gallon of diesel and the $0.0219 per
gallon of gasoline and diesel increases projected in Table 9.4.3.1-1 amounts to a 2% and 0.7%
increases in the price of diesel fuel respectively. Applying the 25% ratio from the USDA study
would indicate that the incremental 2022 RFS volumes would then increase the wholesale price
of produce by 0.5% and 0.2% respectively. If produce being transported by a diesel truck costs
$3 per pound, the increase in that products' prices would be $0,015 per pound and $0,005 per
pound respectively.795 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 type estimate since
impact on food prices varies greatly depending on the distance that the particular food travels by
truck, and some food would travel less than the average estimated by USDA and other foods
travel more than average.

9.4.4 Petroleum Cost Sensitivity Analysis

A simple sensitivity cost analysis is conducted to provide a sense of the impact of how
higher crude oil prices affects the social costs of this program. For this sensitivity case, the
renewable fuels feedstock prices, the renewable fuel byproduct and utility prices are assumed to
be the same. The only difference is that crude oil is assumed to be priced at $110 per barrel, and
this is modeled by increasing both gasoline and diesel fuel prices by 85 cents per gallon, which
approximates the impact of this higher crude oil price on gasoline and diesel fuel wholesale
prices. No change in natural gas prices is assumed for this analysis. Table 9.4.4-1 summarizes
the social costs of this final rule based on these assumptions. As one would expect, increasing the
crude oil price from about $70 per barrel to $110 per barrel and holding other prices constant
reduces the relative cost of renewable fuels, in this case by an estimated 12.5 percent. In reality
other prices are typically also impacted by crude oil prices to varying degrees, but this sensitivity
analysis nevertheless provides some sense of the impact of crude oil price changes.

794 Volpe, Richard; How Transportation Costs Affect Fresh Fruit and Vegetable Prices; United States Department of
Agriculture; November 2013.

795 Comparing Prices on Groceries; May 4, 2021: http://www.conpons.com/thegoodstnff/comparing-prices-on-
groceries

299

-------
Table 9.4.4-1: Total Annual Rule Sensitivity Cost at $110/bbl Crude Oil Price with the

Supplemental Standard (2021$)

Total Cost
(Million $)

Per Unit
Costs

Units

2021

Gasoline

-496

-0.37

0/gallon

Diesel Fuel

881

1.45

0/gallon

Natural Gas

0.00108

$/1000cf

Total

418

2022

Gasoline

-85

-0.06

0/gallon

Diesel Fuel

3536

5.67

0/gallon

Natural Gas

-8.8

-0.00029

$/1000 cf

Total

3442

2021 &
2022

Gasoline

-581

-0.42

0/gallon

Diesel Fuel

4417

7.08

0/gallon

Natural Gas

-0.00079

$/1000 cf

Total

3860

300

-------
Chapter 10: Biomass-Based Diesel Standard for 2022

In setting the BBD volumes in recent annual rules, EPA has recognized that the advanced
biofuel volume drives the use of BBD beyond what is required by the BBD volume.

Nonetheless, we have continued to ensure space for "other" advanced biofuels (beyond BBD and
cellulosic biofuels) to participate in the RFS program given the desire to support biofuels that
might be able to further increase biofuel volumes over the longer term and at potentially lower
cost. At the same time, given the relatively low rate of participation of other advanced biofuels,
we previously maintained approximately the same space for other advanced biofuels in 2021 as
in 2020.796 As we explain in Preamble Section III, this action maintains this same space in 2022
through commensurate increases in both the 2022 BBD and advanced volumes.

Elsewhere in this RIA, we analyze the required statutory factors in CAA section
21 l(o)(2)(B)(ii) for all the biofuel volume requirements we are establishing today, including for
BBD. Specifically, we analyze biodiesel and renewable diesel fuels used for RFS compliance,
the vast majority of which are BBD. In this section, we conduct further analyses of the statutory
factors specifically with respect to the BBD volume, in comparison to other advanced biofuels.
This analysis provides additional support for the BBD volume and for EPA's policy of
maintaining space for other advanced biofuels.

In general, we think that the size of this space for other advanced biofuels is a policy
choice that Congress entrusted to EPA. To date, the BBD industry is the single largest
contributor to the advanced biofuel pool, and has been largely responsible for the growth in
advanced biofuels envisioned by Congress. However, in establishing the BBD and cellulosic
standards as nested within the advanced biofuel standard, Congress clearly intended to support
development of BBD and cellulosic biofuels, while also providing an incentive for the growth of
other non-specified types of advanced biofuels. That is, the advanced biofuel standard provides
an opportunity for other advanced biofuels (advanced biofuels that do not qualify as cellulosic
biofuel or BBD) to be used to satisfy the advanced biofuel standard after the cellulosic biofuel
and BBD standards have been met. Indeed, since Congress specifically directed growth in BBD
only through 2012, leaving development of volume targets for BBD to EPA for later years while
also specifying substantial growth in the cellulosic biofuel and advanced biofuel categories, we
believe that Congress clearly intended for EPA to evaluate the appropriate rate of participation of
BBD within the advanced biofuel standard.

10.1 Review of Implementation of the Program

One of the considerations in determining the BBD volume for 2022 is a review of the
implementation of the program to date, as it affects BBD. This review is required by CAA
21 l(o)(2)(B)(ii), and also provides insight into the capabilities of the industry to produce, import,
export, distribute, and use BBD. It also helps us to understand what factors, beyond the BBD
standard, may incentivize the availability of BBD. In reviewing the program, we assess
numerous regulatory, economic, and technical factors, including the availability of BBD in past
years relative to the BBD and advanced standards; the prices of BBD, advanced, and

796 The 2021 BBD volume was set in the 2020 final rule, while the 2020 BBD volume was set in the 2019 final rule.

301

-------
conventional RINs; the competition between BBD and other advanced biofuels in meeting the
portion of the advanced standard not required to be met by BBD or cellulosic RINs; the
maturation of the BBD industry over the course of the RFS program; and the effects of the BBD
standard on the production and development of both BBD and other advanced biofuels.

The BBD industry in the U.S. and abroad has matured since EPA first increased the
required volume of BBD beyond the statutory minimum in 2013.797 To assess the maturity of the
biodiesel industry, EPA compared information on BBD RIN generation by facility in 2012 and
2021 (the most recent year for which complete RIN generation by company is available). In
2012, the annual average RIN generation per facility producing BBD was about 7.2 million
gallons with approximately 53% of facilities producing less than 1 million gallons of BBD a
year.798 Since that time, the BBD industry has matured in a number of critical areas, including
growth in the size of facilities, the consolidation of the industry, and more stable funding and
access to capital. In 2021, the average BBD RIN generation per facility had climbed to over 32
million gallons annually, more than a 4-fold increase. Only 18% of the facilities produced less
than 1 million gallons of BBD in 2020.799

Table 10.1-1 shows, for 2011-2022, the number of BBD RINs generated, the number of
RINs retired due to export, the number of RINs retired for reasons other than compliance with
the annual BBD standards, and the consequent number of available BBD RINs; and for 2011-
2022, the BBD and advanced biofuel standards.

797 See also generally 84 FR 36794-95 (further explaining our approach in establishing the 2013 BBD volume and
our experience since that time).

798 "BBD RIN Generation by Company in 2012 and 2021," available in EPA docket EPA-HQ-OAR-2021-0324.

799 Id.

302

-------
Table 10.1-1: Biomass-Based Diesel (D4) RIN Generation and Advanced Biofuel and

Biomass-Based Diesel Standarc

s in 2011-2022 (million R

Ns or gallons)800

Year

BBD

RINs
Generated

Exported
BBD
(RINs)

BBD RINs
Retired,
Non-
Compliance
Reasons

Available
BBD

RINsa

BBD
Standard
(Gallons)b

BBD
Standard
(RINs)b

Advanced
Biofuel
Standard
(RINs)b

2011

1,692

102

1,542

800

1,200

1,350

2012

1,738

102

1,545

1,000

1,500

2,000

2013

2,740

125

101

2,514

1,280

1,920

2,750

2014

2,710

134

2,477

1,630

2,490c

2,670

2015

2,796

145

2,606

1,730

2,655c

2,880

2016

4,009

203

121

3,685

1,900

2,850

3,610

2017

3,849

257

115

3,477

2,000

3,000

4,280

2018

3,873

247

3,559

2,100

3,150

4,290

2019

4,147

361

3,705

2,100

3,150

4,920

2020

4,491

512

187

3,792

2,430

3,645c

4,630

2021

4,871

549

4,260

2,430

3,645c

5,050

2022

N/A

2,760

4,140c

5,630

a Available BBD RINs may not be exactly equal to BBD RINs Generated minus Exported RINs and BBD RINs
Retired, Non-Compliance Reasons, due to rounding.

b The volumes for each year are those used as the basis for calculating the percentage standards in the final rule.

They have not been retroactively adjusted for subsequent events, such as differences between projected and actual
gasoline and diesel use and exempted small refinery volumes.

0 Each gallon of biodiesel qualifies for 1.5 RINs due to its higher energy content per gallon than ethanol. Renewable
diesel qualifies for between 1.5 and 1.7 RINs per gallon, but generally has an equivalence value of 1.7. While some
fuels that qualify as BBD generate more than 1.5 RINs per gallon, EPA has in the past multiplied the required
volume of BBD by 1.5 in calculating the percent standard per 80.1405(c). In 2014 and 2015 however, the number of
RINs in the BBD Standard column is not exactly equal to 1.5 times the BBD volume standard as these standards
were established based on actual RIN generation data for 2014 and a combination of actual data and a projection of
RIN generation for the last three months of the year for 2015, rather than by multiplying the required volume of
BBD by 1.5. Some of the volume used to meet the BBD standard in these years was renewable diesel, with an
equivalence value higher than 1.5.

In reviewing historical BBD RIN generation and use, we see that the number of RINs
available for compliance purposes exceeded the volume required to meet the BBD standard in
2011-2013 and 2016-2021.801 Additional production and use of BBD was likely driven by a
number of factors, including demand to satisfy the advanced biofuel and total renewable fuels
standards, the biodiesel tax credit,802 and various other State and local incentives and mandates

800 Available BBD RINs Generated, Exported BBD RINs, and BBD RINs Retired for Non-Compliance Reasons
information from EMTS.

801 The number of RINs available in 2014 and 2015 was approximately equal to the number required for compliance
in those years, as the standards for these years were finalized at the end of November 2015 and EPA's intent at that
time was to set the standards for 2014 and 2015 to reflect actual BBD use. See 80 FR 77490-92, 77495 (December
14, 2015).

802 The biodiesel tax credit was reauthorized in January 2013. It applied retroactively for 2012 and for the remainder
of 2013. It was once again extended in December 2014 and applied retroactively to all of 2014 as well as to the
remaining weeks of 2014. In December 2015 the biodiesel tax credit was authorized and applied retroactively for all

303

-------
allowing for favorable blending economics. Moreover, additional production of BBD, beyond
the volumes shown in the above table, was exported without generating any RINs.

The prices paid for advanced biofuel and BBD RINs beginning in early 2013 through
December 2021 (the last month for which data is available) also support the conclusion that the
advanced biofuel, and in some periods the total renewable fuel standards, provide a sufficient
incentive for additional BBD volume beyond what is required by the BBD standard. Because the
BBD standard is nested within the advanced biofuel and total renewable fuel standards, and
therefore can help to satisfy three RVOs, we would expect the price of BBD RINs to exceed that
of advanced and conventional renewable RINs.803 If, however, BBD RINs are being used (or are
expected to be used) by obligated parties to satisfy their advanced biofuel obligations, above and
beyond the BBD standard, we would expect the prices of advanced biofuel and BBD RINs to
converge.804 Further, if BBD RINs are being used (or are expected to be used) to satisfy
obligated parties' total renewable fuel obligation, above and beyond their BBD and advanced
biofuel requirements, we would expect the price for all three RIN types to converge.

When examining RIN price data from 2011 through December 2022, shown in Figure
10.1-1, we see that beginning in early 2013 and through December 2022 the advanced RIN (D5)
price and BBD (D4) RIN prices were approximately equal. Similarly, from early 2013 through
late 2016 the conventional renewable fuel (D6) RIN and BBD RIN prices were approximately
equal. This demonstrates that the advanced biofuel standard, and in some periods the total
renewable fuel standard, are capable of incentivizing increased BBD volumes beyond the BBD
standard. The advanced biofuel standard has incentivized additional volumes of BBD since 2013,
while the total standard had incentivized additional volumes of BBD from 2013 through 2016.805
We do note, however, that in 2011-2012 the BBD RIN price was significantly higher than both
the advanced biofuel and conventional renewable fuel RIN prices. At this time, the E10
blendwall had not yet been reached, and it was likely more cost effective for most obligated
parties to satisfy the portion of the advanced biofuel requirement that exceeded the BBD and
cellulosic biofuel requirements with advanced ethanol.

of 2015 as well as through the end of 2016. In February 2018 the biodiesel tax credit was authorized and applied
retroactively for all of 2017. On December 17, 2019, the biodiesel tax credit was authorized and applied
retroactively for all of 2018, 2019, and the majority of 2020; and prospectively for the remainder of 2020 through
2022.

803 This is because when an obligated party retires a BBD RIN (D4) to help satisfy their BBD obligation, the nested
nature of the BBD standard means that this RIN also counts towards satisfying their advanced and total renewable
fuel obligations. Advanced RINs (D5) count towards both the advanced and total renewable fuel obligations, while
conventional RINs (D6) count towards only the total renewable fuel obligation.

804 We would still expect D4 RINs to be valued at a slight premium to D5 and D6 RINs in this case (and D5 RINs at
a slight premium to D6 RINs) to reflect the greater flexibility of the D4 RINs to be used towards the BBD, advanced
biofuel, and total renewable fuel standard. This pricing has been observed over the past several years.

805 Although we did not issue a rule establishing the final 2013 standards until August of 2013, we believe that the
market anticipated the final standards, based on EPA's July 2011 proposal and the volume targets for advanced and
total renewable fuel established in the statute. (76 FR 38844, 38843 July 1, 2011). Similarly, for 2014 and 2015,
although we issued the final standards in late 2015, the proposed rule incentivized the market to use BBD volumes
exceeding the proposed BBD standard to help satisfy the proposed advanced and total standards. See 80 FR 33100
(2014-16 standards proposed June 10, 2015); 78 FR 71732 (2014 standards proposed Nov. 29, 2013).

304

-------
D4RINS D5 RINs D6 RINs

RIN Price Source: EMTS Data

Figure 10.1-1: D4, D5, and D6 RIN Prices (July 2010 - December 2022)

$2.00
$1.80

qj $1.60 ij ljjj&

£ $1.40 flyu JR

1$120 tfir\—i

$1.00 II fl' jl t (V""N /

^ $0.80
Q. $0.60
Q $0.40
$0.20
$0.00

A -vj'j r^p A rtO r

^ /«/ / f f #

We also examined the opportunity for advanced biofuels other than BBD and cellulosic
biofuels to participate in the RFS program, as shown in Table 10.1-2. We believe it is important
to preserve this opportunity for other advanced biofuels, and we are conscious of public
comments claiming that BBD volume requirements that are a significant portion of the advanced
volume requirements effectively disincentivize the future development of other promising
advanced biofuel pathways.806 A variety of different types of advanced biofuels, rather than a
single type such as BBD, would increase energy security (e.g., by increasing the diversity of
feedstock sources used to make biofuels, thereby reducing the impacts associated with a shortfall
in a particular type of feedstock) and increase the likelihood of the development of lower cost
advanced biofuels that meet the same GHG reduction threshold as BBD.

8116 See, e.g.. Comments from Advanced Biofuel Association, available in EPA docket EPA-HQ-OAR-2018-0167-
1277.

305

-------
Table 10.1-2: Opportunity for and RIN Generation of "Other" Advanced Biofuels (million
RINs)

Opportunity for
"Other" Advanced

Available Advanced

Available BBD (D4)
RINs in Excess of the

Year

Biofuels3

(D5) RINs

BBD Requirementb

2011

150

223

342

2012

500

597

2013

829

548

594

2014

147

143

-13

2015

102

147

-49

2016

530

835

2017

969

144

477

2018

852

178

415

2019

1,352

322

555

2020

475c

334

146

2021

724

227

493

2022

860

N/A

a The opportunity for "other" advanced biofuel is calculated by subtracting the number of cellulosic biofuel and
BBD RINs required each year from the number of advanced biofuel RINs required. This portion of the advanced
standard can be satisfied by advanced (D5) RINs, BBD RINs in excess of those required by the BBD standard, or
cellulosic RINs in excess of those required by the cellulosic standard.

b The available BBD (D4) RINs in excess of the BBD requirement is calculated by subtracting the required number
of BBD RINs (from Table 10.1-1) from the number of BBD RINs available for compliance in thatyear. This
number does not include carryover RINs, nor do we account for factors that may impact the number of BBD RINs
that must be retired for compliance, such as differences between the projected and actual volume of obligated
gasoline and diesel. The required BBD volume has not been retroactively adjusted for subsequent events, such as
differences between projected and actual gasoline and diesel use and exempted small refinery volumes.

0 In 2020, the prospective opportunity for other advanced biofuels was 855 million gallons, as established in the
2020 annual rule. However, after adjusting the 2020 advanced biofuel volume in this action to those actually
consumed, the space is reduced to 475 million gallons.

In each year since 2016, there has been significant space for other advanced biofuels, but
this space has nonetheless been dominated by excess BBD. While the RFS volumes created the
opportunity for up to 530 million, 969 million, 852 million, 1,352 million, and 855 million
gallons of "other" advanced for 2016, 2017, 2018, 2019, 2020, and 2021 respectively to be used
to satisfy the advanced biofuel obligation, only 98 million, 144 million, 178 million, 322 million,
334 million, and 227 million gallons of "other" advanced biofuels were generated. This is
significantly less than the volumes of "other" advanced available in 2012-2013. Despite creating
space within the advanced biofuel standard for "other" advanced, in recent years, only a small
fraction of that space has been filled with "other" advanced, and BBD continues to fill most of
the gap between the BBD standard and the advanced standard. Thus, there does not appear to be
a compelling reason to increase the "space" maintained for "other" advanced biofuel volumes for
2022.

10.2 Analysis of Other Statutory Factors

In this section, we discuss our consideration of the other statutory factors (besides review
of the implementation of the program). As already noted, this analysis supplements the analyses

306

-------
elsewhere in this RIA and focuses on a comparison between BBD and other advanced biofuels.
Our analysis here focuses on the impacts on the statutory factors in 2022 and in the longer term.

Consistent with the historical trends described above as well as the expected availability
of BBD in 2022 (see Chapter 5.2), we expect that the actual use of BBD in 2022 will exceed the
BBD standard and that its use will be driven by the 2022 advanced biofuel standard, and
potentially the 2022 total renewable fuel standard. Thus, EPA continues to believe that
approximately the same overall volume of BBD would likely be used in 2022 even if we were to
establish a somewhat lower or higher BBD volume for 2022. Despite this, we think the BBD
volume requirement can still have a positive impact on the future development and marketing of
BBD by providing a base guaranteed level for investment certainty in BBD.

At the same time, we believe that, over the long-term, maintaining a significant space for
other advanced biofuels could result in greater production and use of these other advanced
biofuels, which may in turn have positive impacts on several of the statutory factors.

Specifically, when viewed in a long-term perspective, BBD can be seen as competing for
research and development dollars with other types of advanced biofuels for participation as
advanced biofuels in the RFS program. Preserving space within the advanced biofuel standard
for advanced biofuels that do not qualify as BBD or cellulosic biofuel provides the appropriate
incentives for the continued development of these types of fuels. The increased use of other
advanced biofuels in the long-term (which would be less likely if higher BBD volumes were
established leaving a smaller market share for other advanced biofuels within the advanced
biofuel requirement) leads to potentially favorable outcomes for some statutory considerations,
including wetlands, ecosystems, and wildlife habitats; water quality and water supply; energy
security; and costs. These potential positive impacts, which are the focus of our discussion in this
section, support our continued practice of providing space for other advanced biofuels.

We note that while preserving space for other advanced biofuels facilitates the
commercialization of those fuels, whether such commercialization is actually realized or not
depends on market and other factors beyond the RFS program. In contrast to BBD, non-
cellulosic advanced biofuel volumes other than BBD have not grown to the same extent in recent
years. This may or may not be the case in the future. One reason this trend may continue is that
in contrast to some prior years the biodiesel blender's tax credit is in place prospectively through
2022. The biodiesel blender's tax credit provides a competitive advantage to BBD over other
advanced biofuels. Given this and also because we have not seen significant increases in other
advanced biofuels in recent years, we are increasing the BBD volume by the same energy-
equivalent amount as the increase in the implied volume for non-cellulosic advanced biofuels in
2022. This is similar to what we did in establishing the 2020 BBD volume and is also consistent
with our action setting the 2021 BBD volume. This preserves a space of roughly 850 million
RINs for other advanced biofuels.807

At present, the other advanced biofuels that could fill any gap between the advanced
biofuel volume and the BBD volume include the following: sugarcane ethanol; ethanol from
grain sorghum using certain processing technologies; naphtha; compressed natural gas/liquefied
natural gas (CNG/LNG) from non-cellulosic sources; liquified petroleum gas; renewable diesel

807 Due to rounding, the space in 2022 is roughly 860 million RINs.

307

-------
co-processed with petroleum diesel fuel; renewable jet fuel; renewable heating oil; and cellulosic
biofuel. EPA also continues to consider other potential pathways for producing advanced
biofuels. To this end we have published analyses of greenhouse gas emissions attributable to the
production and transport of feedstocks used to produce advanced biofuels, including analyses of
pennycress oil, carinata oil, cottonseed oil, jatropha oil, and sugar beets.808 These analyses could
help influence the submission of new petitions seeking EPA approval of additional advanced
biofuel pathways, and other advanced biofuels that may be affected by the use of BBD.

Following is a discussion of the six sets of factors listed under CAA section
21 l(o)(2)(B)(ii)(I)-(VI).

10.2.1 Impact on the Environment

10.2.1.1 Air Quality

We describe the impacts of biofuels, including BBD, on air quality in Chapter 3. As we
explain further there, some emission increases or decreases might be associated with a change in
the use of BBD and other advanced biofuels. Depending on which of the competing advanced
biofuels is compared to BBD, the emissions impacts of requiring a larger or smaller portion of
the advanced biofuel standard to BBD will vary. However, given that other advanced biofuels
also result in air quality impacts (both positive and negative), and that the volume at issue is
relatively small when considered in the context of transportation fuel use throughout the
geographic area in which the RFS program applies (the continental United States and Hawaii),
the overall impacts of marginal shifts between BBD and other advanced biofuels is expected to
be small, and air quality impacts do not appear to provide a good reason for setting a higher or
lower BBD volume requirement.

10.2.1.2 Climate Change

We describe the greenhouse impacts of biofuels, including BBD, in Chapter 3. As we
detail in Chapter 5, the vast majority of increases in both BBD and advanced biofuels in 2022 is
expected to come from renewable diesel made from soy, which is a form of BBD. Other
advanced biofuels could also contribute to satisfying the increased 2022 BBD and advanced
biofuel volumes. Some of these other biofuels may have superior GHG profiles relative to soy
renewable diesel.

A somewhat higher or lower 2022 BBD volume is not expected to significantly change
the availability of advanced biofuels with superior GHG profiles in 2022. In the long-term,
however, setting the nested BBD volume requirement sufficiently below the advanced biofuel
requirement should provide a continuing incentive for the further development and marketing of
other advanced biofuels, and may result in more GHG reductions.

808 See EPA's webpage, Analysis of Lifecycle Greenhouse Gas Emissions Attributable to Production and Transport
of Certain Feedstocks; https://www.epa.gov/renewable-fuel-standard-program/other-actions-renewable-fuel-
standard-program (last accessed June 6, 2019).

308

-------
10.2.1.3 Wetlands, Ecosystems, Wildlife Habitats, Water Quality, and Water
Supply

In Chapter 3, we describe the impacts of biofuels, including BBD, on wetlands,
ecosystems, wildlife habitats, water quality, and water supply. Our comparative assessment of
BBD and other advanced biofuels is similar to the above assessment for climate change. Namely,
there are some advanced biofuels (e.g., ethanol from food waste, CNG/LNG from non-cellulosic
feedstock) that are not made directly from crops and would therefore likely have significantly
lower adverse impacts on some or all of these environmental factors relative to soy renewable
diesel. By contrast, some other advanced biofuels are made from planted crops (e.g., biodiesel
from canola, ethanol from sugarcane) and therefore may have impacts on these environmental
factors roughly comparable to, yet still different from, soy renewable diesel.

Setting the BBD volume requirement sufficiently below the advanced biofuel
requirement as we are for 2022 should provide a continuing incentive for the further
development and marketing of fuels with diminished adverse impacts on these environmental
factors. To the extent such fuels become increasingly significant players in the advanced biofuel
market, there may be long-term reductions in such adverse impacts as compared to a scenario
where the BBD volume requirement is set at a level that does not incentivize production of such
fuels.

10.2.2 Impact on Energy Security

In Chapter 4, we assess the impacts of replacing petroleum with biofuels, including BBD.
In this section, we further consider the energy security impacts of BBD relative to other
advanced biofuels. Having a diversity of biofuels available to displace petroleum is likely to
reduce the occurrences and severity of supply disruptions or "supply shocks," since a diversity of
fuel supplies means that a disruption to any single supply of fuel will not have as severe of an
impact. Setting a BBD standard that requires all (or virtually all) of the anticipated 2022
advanced biofuel requirement be made up of BBD would reduce diversity of the fuel supply and
therefore reduce energy security. By contrast, providing space for other advanced biofuels, such
as sugarcane ethanol and renewable naptha, contributes to a more diverse fuel mix and thus
energy security. In the long-term, preserving this space could increase production and use of
other advanced biofuels, which would further enhance energy security.

10.2.3 Rate of Production

We assess the expected rate of production of biofuels, including BBD, in Chapter 5.

10.2.4 Impact of Renewable Fuels on Infrastructure

In Chapter 6, we describe the impacts of biofuels, including BBD, on infrastructure. With
respect to the sufficiency of infrastructure to deliver and use renewable fuels, all of the
renewable fuels have some infrastructure issues and challenges associated with them to varying
degree, some of which are dependent on the feedstock and feedstock location. We acknowledge
that there are constraints associated with higher level ethanol blends (El 5 and E85) that limit

309

-------
their use in 2022. By contrast, while there may be some infrastructure challenges on the expected
increase in BBD use through 2022, we anticipate that the market will be able to adjust
accordingly. Other advanced biofuels/feedstock combinations may also present differing
infrastructure considerations. To the extent that infrastructure considerations favor BBD,
especially relative to ethanol, the market can choose BBD over competing products. However, to
the degree that non-BBD advanced biofuels face lesser infrastructure challenges, preserving a
space for other advanced biofuels allows the market to choose those advanced biofuels in lieu of
BBD. With respect to potential impacts on deliverability of materials other than renewable fuels,
we do not anticipate any significant impacts of devoting a larger or smaller portion of the
advanced biofuel standard to BBD for 2022.

As discussed in Chapter 6, we believe that the volumes under consideration in this final
rule will not impact the deliverability of materials and products other than renewable fuel.

10.2.5 Impact of the Use of Renewable Fuels on the Cost to Consumers of
Transportation Fuel and on the Cost to Transport Goods

In Chapter 9, we assess the impacts of biofuels, including BBD, on costs. Generally, the
market will determine which advanced biofuels are used to meet the portion of the advanced
standard that is not dedicated to BBD. To the extent that cost considerations favor BBD, the cost
benefit can be obtained either through setting a higher BBD nested standard or allowing the
market to choose BBD over competing products in meeting the advanced and total RFS
standards. However, to the degree that non-BBD advanced biofuels can be supplied at a lower
cost than BBD, requiring the use of a higher volume of BBD would result in higher
transportation fuels costs and higher costs to transport goods.

In the long-term, leaving room within the advanced standard for non-BBD advanced
biofuels can incentivize the development and production of non-BBD advanced biofuels which
may eventually provide a cost benefit as compared to BBD and support the growth in their
volumes that will enhance the longer term success of the RFS program. In addition, this
approach may temper to some extent BBD prices, by providing competition between the
marginal BBD production volumes and other types of advanced biofuels. We note that in
general, BBD remains more costly than diesel fuel and some other types of advanced biofuel,
primarily driven by feedstock costs, and that this cost represents a significant portion of the costs
of the RFS program in general. By preserving space for other advanced biofuels, other advanced
biofuels may offer greater potential for costs reductions in the long-term. Therefore, we believe
that this factor supports maintaining the BBD volume requirement at a level sufficiently lower
than the advanced biofuel volume requirement to incentivize production of non-BBD advanced
biofuels, consistent with today's rule.

10.2.6 Impact of Renewable Fuels on Other Factors
10.2.6.1 Job Creation and Rural Economic Development

In Chapter 7, we assess the impacts of biofuels, including BBD, on job creation and rural
economic development. Overall, we believe that BBD will have small but positive impacts on

310

-------
job creation and rural economic development, including due to the construction and operation of
new renewable diesel facilities anticipated through 2022. Depending on which other advanced
biofuel is compared to BBD, the job creation and rural economic development opportunities will
vary. Notably, if we compare advanced biofuels that are predominantly imported (such as
imported sugarcane ethanol) with BBD, which is predominantly produced in the U.S., BBD will
likely have superior domestic job creation and rural economic benefits. By contrast, if we
compare other domestically produced advanced biofuels with BBD, both fuels may provide
domestic job creation and economic benefits, albeit potentially for different sectors of the
economy or kinds of workers. In general, we think that the marginal shifts in job creation and
rural economic benefits do not provide a good reason for setting a higher or lower BBD
requirement.

10.2.6.2 Price and Supply of Agricultural Commodities and Food Prices

In Chapter 7, we assess the impacts of biofuels, including BBD, on the price and supply
of agricultural commodities and food prices. We do not expect that the volumes would cause
adverse impacts on the supply of agricultural commodities. We do expect that the volumes, as
analyzed in Chapter 7, could cause increases in soybean oil and food prices. Depending on which
competing advanced biofuel is compared to BBD, the impacts on agricultural commodities and
food prices would vary. To the extent that BBD is replaced by biofuels made from other
agricultural products (e.g., sugarcane ethanol), there will be different impacts on the associated
agricultural commodities and food prices (e.g., increases in the price of sugar as opposed to
soybean oil). By contrast, if the market uses advanced biofuels made from non-agricultural
products (e.g., ethanol from food waste) in lieu of BBD, we would expect less impacts on
agricultural commodities and food prices. If the use of BBD decreases as a result of replacement
by other advanced biofuels, then this may also cause decreases in soybean oil and food prices.

311

-------
Chapter 11: Screening Analysis

11.1 Summary

This chapter discusses EPA's screening analysis evaluating the potential impacts of the
RFS standards for 2020, 2021, and 2022 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
using three different methods and compared the cost-to-sales ratio for each method to a threshold
of 1%. For our first method and the one we consider to be most appropriate, we considered the
annual RFS standards as a subset of the overall Renewable Fuel Standard program finalized in
2010809 (RFS2); for the remainder of our methods, we considered the annual RFS standards
(including the 2022 supplemental standard) to be a separate action.810 For our second method, we
compared obligated parties' cost of compliance (whether they acquire renewable identification
numbers (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 fuel they sell
than 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 fuel.811

In the DRIA, we also assessed three additional methods, which we described as
hypothetical worst-case scenarios assuming obligated parties do not recover the RFS compliance
costs. Two of those methods assessed the costs for obligated parties as a whole (i.e., all refiners
and importers subject to RFS obligations). The final methods assessed the costs for small refiners
specifically. Particularly given our recent work demonstrating the ability of all obligated parties
to fully recover their costs of RFS compliance, we generally do not believe continuing to
reiterate these purely hypothetical scenarios is necessary or conducive to a realistic evaluation of
the impacts of this rule on small entities. In addition, the methods assessing the costs for
obligated parties as a whole were of limited relevance since the regulatory flexibility analysis is
meant to analyze the impacts on small entities, not on the entire market.

As such, in this final rule, we are no longer presenting the two methods assessing costs
for obligated parties as a whole. We also believe the small refiner-specific analysis is not
necessary either, as small refiners fully recover their RIN costs. Nonetheless, we have continued
to include it in this final rule. We do so because we included this hypothetical scenario in the

809 75 FR 14670 (March 26, 2010).

810 Note that throughout the rest of this chapter that when we refer to the 2022 RFS standards, we are also including
the 2022 supplemental standard.

811 See "Denial of Petitions for RFS Small Refinery Exemptions," EPA-420-R-22-005, April 2022.

312

-------
proposed rule to respond to prior concerns from small refiners that they could not recover their
RIN costs. Moreover, relative to the two methods assessing impacts on all obligated parties that
we are excluding, this method seems more relevant to addressing small refiner-specific concerns.
This method emphasizes that even erroneously assuming no RIN cost recovery by small refiners
as suggested by some parties, a NO SISNOSE finding would still be appropriate.

Viewed from the perspective of the first method, the annual RFS standards impose no
additional cost, since they simply represent the ongoing implementation of a program put into
place in 2010. As can be seen in Table 11.1-1 below, all of the methods that assume this rule is a
separate action result in a cost-to-sales ratio of less than 1%.812 Therefore, EPA finds that these
standards would not have a significant economic impact on a substantial number of small
entities.

Chapter 11.2 provides background on the RFA and this rule, including the regulated
small entities. For each of the three methods described above, Chapter 11.3 describes EPA's
calculations of the costs of the rule, and Chapter 11.4 sets forth the cost-to-sales ratios. Chapter
11.5 concludes.

Table 11.1-1: Estimated Cost-to-Sales Ratios of the 2021 and 2022 RFS Standards

Method

Screening Analysis

Cost-to-Sales Ratio

2021

2022

Cost as Part of RFS2 Rule

N/A

Market Cost Recovery

0.00%

Full RIN Price as Cost for Small Refiners
(No Cost Recovery)

0.03%-0.17%

0.06% - 0.43%

11.2 Background

11.2.1 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.

11.2.2 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,

812 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.

313

-------
and to annually establish renewable fuel standards that are used by obligated parties to determine
their individual renewable volume obligations (RVOs).

11.2.3 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 the 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 rulemaking, 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 the table below.

Small Business Definitions

Industry

Defined as small entity by SBA if less
than or equal to:

NAICS" code

Gasoline and diesel fuel 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 the Small Business Administration
under the North American Industry Classification System (NAICS) as a guide. Information about
the characteristics of refiners comes from sources including the Energy Information
Administration (EIA) within the U.S. Department of Energy, oil industry literature, and previous
rulemakings 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 2020 EIA refinery data,813 EPA believes that there are
about 35 refiners of gasoline and diesel fuel subject to the RFS regulations. Of these, EPA
believes that there are currently 8 refiners (owning 12 refineries) producing gasoline and/or
diesel that meet the small entity definition of having 1,500 employees or fewer.

813 Data available at https://www.eia.gov/petrolenm/refinerecapacitv/archive/2020/refcap2020.php.

314

-------
11.2.4 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. While EPA is making minor revisions to the RFS
requirements in this action, we do not anticipate that there will be any significant cost on directly
regulated small entities.

11.3 Approaches to the Screening Analyses

This section concerns EPA's screening analyses performed for the 2020, 2021, and 2022
RFS standards only. However, prior to issuing our 2009 proposal for the general RFS regulatory
program regulations required to implement the amendments enacted pursuant to the Energy
Independence and Security Act of 2007, we analyzed the potential impacts on small entities of
implementing the full RFS program through 2022 and convened a Small Business Advocacy
Review Panel (SBAR Panel, or 'the Panel') to assist us in this evaluation. This information is
located in the RFS2 rulemaking docket (Docket ID No. EPA-HQ-OAR-2005-0161). We believe
that it is most appropriate to consider the annual RFS standards as a subset of the overall RFS2
program finalized in 2010. Viewed from that perspective, the 2020, 2021, and 2022 RFS
standards impose no additional cost, since they simply represent the ongoing implementation of a
program put into place in 2010. The annual rulemaking standards in fact reflect reductions from
the statutory volumes that were evaluated in 2010. However, for the purposes of this screening
analysis, we also estimated the costs of the 2020, 2021, and 2022 RFS standards relative to a
"baseline" of the final 2020 RFS standards (i.e., the amount of biofuel actually used in 2020).

The changes between 2020 and 2022 in required volumes of the four categories of
renewable fuel in this action are shown in Table 11.3-1. In this table we have included volume
increases that result from both the annual RFS volume requirements and the 2022 supplemental
standard. In addition, we have also referenced in Table 11.3-1 the implied volumes of
conventional renewable fuel (total renewable fuel minus advanced biofuel) and non-cellulosic
advanced biofuel (advanced biofuel minus cellulosic biofuel) that may be used to satisfy the
2020, 2021, and 2022 RFS standards.

315

-------
Table 11.3-1: Changes in RFS Standards from 2020 to 2022 (million RINs)1

2020

2021

2022

Final

RFS Standard

Volumesb

Volumes0

Volumes

Cellulosic Biofuel

510

560

630

120

Biomass-Based Diesel (BBD)d

3,650

4,140

490

Advanced Biofuel

4,630

5,050

420

5,630

1,000

Non-Cellulosic Advanced Biofuel6

4,120

4,490

370

5,000

880

Total Renewable Fuel

17,130

18,840

1,710

20,630

3,500

Conventional Renewable Fuelf

12,500

13,790

1,290

15,000

2,500

Supplemental Standard8

n/a

250

a All volumes have been rounded to the nearest 10 million gallons, consistent with our practice in previous annual
rules. A represents the change relative to the baseline of the final 2020 RFS standards.

b The 2020 BBD volume requirement was established in the 2019 final rule (83 FR 63704, December 11, 2018).
0 The 2021 BBD volume requirement was established in the 2020 final rule (85 FR 7016, February 6, 2020).
d For ease of comparison, the BBD standards shown in this table are in ethanol-equivalent RINs, rather than physical
volume. To convert from BBD volume to ethanol-equivalent RINs, a multiplier of 1.5 was used.
e Non-cellulosic advanced biofuel is not a fuel category for which a percentage standard is established but is
calculated by subtracting the cellulosic biofuel applicable volume from the advanced biofuel applicable volume.
f Conventional renewable fuel is not a fuel category for which a percentage standard is established but is calculated
by subtracting the advanced biofuel applicable volume from the total renewable fuel applicable volume.
g The 2022 supplemental standard is a total renewable fuel requirement and can be satisfied with RINs from any
category. For the purpose of this analysis we have assumed that obligated parties use conventional renewable fuel
RINs to satisfy their supplemental volume obligation.

As shown in Table 11.3-1, because we are using a baseline of the final 2020 RFS
standards, there are no impacts of the final 2020 RFS standards themselves, and hence,
regardless of the method used, there are no costs on obligated parties. Therefore, for the
remainder of this analysis, we have only considered the impacts of the 2021 and 2022 RFS
standards.

We next considered various methods for estimating the cost of the 2021, and 2022 RFS
standards to obligated parties using the baseline of the final 2020 RFS standards. If, as has been
demonstrated, obligated parties recover the costs of RFS compliance through higher prices in the
marketplace for the petroleum products they sell, there is no net cost to obligated parties.
However, because various parties, including several small refiners, have continued to claim that
they are not able to recover the cost of RFS compliance in the marketplace, we also estimated the
cost-to-sales ratios for each of the 8 small refiners that are obligated parties under the RFS
program using refinery-specific data under the assumption that they could not recover RIN costs.

11.3.1 Method 1: Cost as Part of RFS2 Rule

We continue to believe that it is most appropriate to consider the impacts of the annual
RFS standards on small businesses as a subset of the overall RFS2 program finalized in 2010,
rather than as an unrelated or separate action. When viewed from this perspective, the 2021 and
2022 RFS standards are simply the ongoing implementation of a program finalized in 2010 and
therefore add no additional burden. More precisely, to the extent the RFS2 program imposes any
net costs on small entities, this rule only reduces net costs relative to the standards analyzed in

316

-------
the RFS2 rule. This is because the RFS2 rule considered the economic impacts of standards
based on the statutory volumes, while this rule reduces those volumes using available statutory
authorities, with corresponding reductions to the RFS standards.814

11.3.2 Method 2: Market Cost Recovery Method

One way, and we believe the most appropriate way to consider the impacts of the 2021
and 2022 RFS standards on obligated parties if the 2021 and 2022 RFS standards are considered
as a separate action, 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
fuel 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 entities,
obligated parties recover that cost through the higher sales prices they receive for the gasoline
and diesel fuel they sell due to the market-wide impact of the RFS standards on these
products.815 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 generally enable obligated parties to recover the cost of the RINs.816 When viewed in light
of this data, there is no net cost of compliance with the RFS standards to obligated parties (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). This is true whether obligated
parties acquire RINs by purchasing renewable fuels with attached RINs or by purchasing
separated RINs.

11.3.3 Method 3: Full RIN Price as Cost for Small Refiners Method (No Cost
Recovery)

For our extreme hypothetical case, we estimated the actual change in the total cost of the
2021 and 2022 RFS standards to the 8 obligated parties that are small refiners if they acquired
the RINs necessary for compliance by purchasing separated RINs. We note, however, that in
doing so we ignored the fact that these parties automatically recover the cost of the RINs they
purchase through the higher market-wide prices they receive for the petroleum-based gasoline
and diesel fuel that they sell, as discussed in Method 2. This would then reflect the cost they
would have to pay for compliance at the end of the year if they had spent the added revenue
received from the higher gasoline and diesel prices for other purposes. Because RIN prices can
be impacted by a wide variety of different factors (including the prices of renewable fuels and
petroleum-based fuels, oil prices, commodity prices, etc.), EPA is 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 February 2022) as an estimate of future RIN prices, as shown in Table 11.3.3-1.

814 Although the BBD standards in this action exceed the statutory minimum of 1.0 billion gallons, we do not expect
any additional burden due to the BBD standard, as it is nested within the advanced biofuel standard and it is that
standard that is driving BBD use.

815 For a further discussion of the ability of obligated parties to recover the cost of RINs see "Denial of Petitions for
Rulemaking to Change the RFS Point of Obligation," EPA-420-R-17-008, November 2017. See also "Denial of
Petitions for RFS Small Refinery Exemptions," EPA-420-R-22-005, April 2022.

816 Id.

317

-------
Table 11.3.3-1: Average RIN Prices and RFS Standards for 2020, 2021, and 2022"

RFS
Standard

RIN
Type

Average RIN

Price
(March 2021
- Feb 2022)

2020
Standard

2021

2022

Standard

Cellulosic
Biofuel

$2.86

0.32%

0.33%

0.01%

0.35%

0.03%

Biomass-
Based Diesel

$1.47

2.30%

2.16%

-0.14%

2.33%

0.03%

Other

Advanced

Biofuelb

$1.47

0.31%

0.51%

0.20%

0.48%

0.17%

Conventional

Renewable

Fuelc

$1.29

7.89%

8.19%

0.30%

8.43%

0.54%

Supplemental
Standard

$1.29

n/a

0.14%

a A represents the change relative to the baseline of the final 2020 RFS standards.

b 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.
0 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.

Using 2019 compliance data and small refinery exemption (SRE) petition materials
where available, and assuming that the total gasoline and diesel fuel production for each of these
small refiners remains unchanged, we estimated their RVOs for 2020, 2021, and 2022. The
difference between the estimated RVOs for each year multiplied by the estimated RIN price for
each standard then gives us the estimated cost of the 2021 and 2022 RFS standards for each
small refiner that chooses to meet their obligations by purchasing separated RINs. The actual
calculations for each small refiner are provided in Chapter 11.6; a non-CBI example of these
calculations is shown below in Tables 11.3.3-2 and 3.

Table 11.3.3

-2: Example Small Refiner Costs Calculation for 2021

Company

Gas/Diesel
Production
(gal)

Celli
(I

jlosic
>3)

BBD
(D4)

(RINs)

Cost
($)

(RINs)

Cost
($)

Example

100,000,000

10,000

$28,600

(140,000)

($205,800)

Company

Other Advanced
Biofuel
(D5)

Conve
Renewa
(E

ntional
ble Fuel
6)

Total
Cost
($)

(RINs)

Cost
($)

(RINs)

Cost
($)

Example

200,000

$294,000

300,000

$387,000

$503,800

318

-------
Table 11.3.3

-3: Example Small Refiner Costs Calculation for 2022

Company

Gas/Diesel
Production
(gal)

Cellulosic
(D3)

B]
(I

BD
>4)

Other Advanced
Biofuel
(D5)

(RINs)

Cost
($)

(RINs)

Cost
($)

(RINs)

Cost
($)

Example

100,000,000

30,000

$85,800

30,000

$44,100

170,000

$249,900

Company

Conventional
Renewable Fuel
(D6)

Supplt
Stan
(I

omental
dard

>6)

Total Cost
($)

(RINs)

Cost
($)

(RINs)

Cost
($)

Example

540,000

$696,600

140,000

$180,600

$1,257,000

11.4 Cost-to-Sales Ratio Results

The final step in our methodology is to compare the total estimated costs from each of the
methods above to relevant total estimated revenue from the sales of gasoline and diesel in the
U.S. in 2021 and 2022. 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. To calculate total revenue for all obligated parties we
multiplied the total expected sales volume of gasoline and diesel used in the U.S. in 2021 and
2022 from the EIA's January 2022 STEO by the projected prices for each of these fuels. These
calculations are shown in Table 11.4-1 below.

Table 11.4-1: Estimated Revenue from the Sale of Gasoline and Diesel in the U.S. (2021$)

Projected Sales Volume

Total Revenue (Billion

(Billion Gallons)

Price Per Gallon

Dollars)

Fuel Type

2021

2022

2021

2022

2021

2022

Gasoline

134.5

138.6

2.18

2.20

$293

$305

Diesel

49.9

56.2

2.11

2.25

$105

$126

Total

N/A

$398

$431

a Gasoline and diesel volumes and prices from EIA's January 2022 STEO.

For Methods 1 and 2, we then divided the estimated costs from each of the methods by
the total projected revenue for the sale of gasoline and diesel in the U.S. in 2021 and 2022. For
Method 3, we divided the estimated costs for each small refiner by its total estimated annual
sales.817 The resulting cost-to-sales ratios for each method are shown in Table 11.4-2, along with
a non-CBI example of these calculations for Method 3 using the data from Tables 11.3.3-2 and 3.

817 Estimated annual sales data gathered from Hoovers, Inc. (available at https://www.dnb.com/products/marketing~
sales/dnb-hoovers.htmD and SRE petition materials.

319

-------
Table 11.4-2: Estimated Cost-to-Sa es Ratios of the 2021 and 2022 RFS Standards

Method

Screening Analysis

Total Cost
(Million Dollars)

Total Sales
(Billion Dollars)

Cost-to-Sales Ratio

2021

2022

2021

2022

2021

2022

Cost Relative to the
Statutory Baseline

Volume Reduction
- No Net Cost

$398

$431

N/A

Market Cost
Recovery

$398

$431

0.00%

Full RIN Price as
Cost for Small
Refiners (Actual)a

—

0.03% -
0.17%

0.06% -
0.43%

Full RIN Price as
Cost for Small
Refiners (Example)

$0.50

$1.26

$0.25

0.13%

0.42%

' The actual calculations for Method 3 for each small refiner are provided in Chapter 11.6.

11.5 Conclusions

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. When considered as a subset of the overall RFS2 program finalized in 2010, the
2020-2022 RFS standards are simply the ongoing implementation of the program and therefore
add no additional burden. Even assuming a baseline of the final 2020 RFS standards, since
obligated parties have been shown to recover their RFS compliance costs through the resulting
higher market prices for their petroleum products, there are still no net costs of the rule on small
businesses. However, we also conducted worst-case sensitivity analyses that ignored the fact that
obligated parties recover their costs. Under these extreme assumptions we were able to estimate
costs of this rule and then use a cost-to-sales ratio test (a ratio of the estimated annualized
compliance costs to the value of sales per company) to assess whether the costs were significant.
Costs were analyzed using the methods described above.

Under any of the three methods, the cost-to-sales analyses indicated that the 8 small
refiners would be affected at less than 1% of their sales (i.e., the estimated costs of compliance
with the rule would be less than 1% of their sales). The cost-to-sales percentages estimated using
the various methods described in the previous section ranged from 0.03% to 0.43%.

11.6 Small Refiner CBI Data
[Information Redacted - Claimed as CBI]

320

-------