-------
Figure 4.1.3-3: eGRID Methane Emissions By Source Type, 2020

100000 f

10000

100

-C

5
5

10

v «•:;

"• * • —• %•••*"/•

0.1

l|0	10.0 100.0 1,000.0 10,000.0 100,000.0 1,000,000.010,000,000.000,000,000.0

• • * ¦»

Annual Production (MWh)

• Biogas • Coal • Natural Gas

Figure 4.1.3-4: eGRID N2O Emissions By Source Type, 2020

10000

o

fNi

1000

100

10

i ••

" " • «•
»• . r * •
. • %•**••••

* **#/

* Vc•• . • - .'.r

.. >?*: ¦¦•¦•-•r.v-yifMfi-r."- •

1.0

0.1

10,0 100.0 1,000.0 JO,OOD.QLOO,OOOigOOO,OQ£CtaOO,QDI]OpDOO;000.0

0.01

0.001

Annual Production (MWh)

• Biogas • Coai • Natural Gas

The lack of data on these many small facilities makes it difficult to quantify the emission
impacts of biogas EGUs. However, they can be characterized by category using a mixture of
modelled and estimated data. We used data from the 2020 Emissions and Generation Resource
Integrated Database (eGRID), prepared by the Clean Air Markets Division of the Office of
Atmospheric Programs in the EPA. We also obtained some data from stakeholders that represent
monitored projects within the US. In California, where biogas electricity generators have been
earning credit under the LCFS since 2018, engines using biogas emit a variety of criteria
pollutants, and offer some of the best available data on emissions from EGUs powered by

104

-------
biogas.170 It is important to note that EGUs often contain multiple engines and/or turbines, and
that the numbers displayed in Table 4.1.3-1 represent a single turbine. Facilities can have
between 1-48 turbines or engines on site\

Table 4.1.3

Pollutant

Average
(lb/MWh)

voc

2.23

CO

6.96

NOx

1.67

SO2

0.07

PM10

0.14

CO2

1,441

: California Biogas EGU Individual Turbine Modelled Emissions3

a Results from CA GREET Model

Figure 4.1.3-5: NOx Emissions From Biogas EGUs

I BO ^

140

*

170
£ 100

J 80

0.0	50,000.0	100,000.0	150,000.0	200,000,0	250,000.0

Annual Production (MWh)

300,000,0

170 LCFS Guidance 19-06: Determining Carbon Intensity of Dairy and Swine Manure Biogas to Electricity
Pathways, https://ww2.arb.ca.gov/sites/default/files/classic/fuels/lcfs/guidance/lcfsguidance 19-06.pdf

105

-------
Figure 4.1.3-6: SOx Emissions From Biogas EGUs

•























•

- *











tfim'

«<% *

• •
». .

•

•	

-Jk	

•

0.0	50,000.0	100,000.0	150,000.0	200,000.0	250,000.0	300,000.0

Annual Production (MWhf

Figure 4.1.3-7: Methane Emissions From Biogas EGUs













•











p











•



































	*

•• * # %

0

I

m

0,0	50,000 0	lOG/JOO.O	150.000.0	200,000.0	250,0000	300,000,0

Annual Production {MWii)

106

-------
Figure 4.1.3-8: N2O Emission From Biogas EGUs













•











a











•























•

•

> >







'	¦*			

• ~

• • * V

#

#

•

0.0	SO,GOO.0	100,000.0	150,000.0	200,000.0	250,000.0	300,0000

Annual Production (MWh)

As shown in Figures 4.1.3-1 through 4, we estimate emission rates per kWh of electricity
produced for key pollutants for various biogas to electricity facilities in comparison to the rates
for natural gas and coal fired EGUs. While for the largest biogas EGUs the rates are similar, for
the vast majority of biogas EGUs, the rates are considerably higher. When it comes to overall
facility emissions, however, these higher emission rates tend to be offset by the smaller size of
biogas EGUs. Data shown in Table Figures 4.1.3-5 through 8 displays weighted averages of
natural gas, coal, and biogas EGUs against their nameplate capacity. Some of these EGUs
compare favorably against coal and natural gas plants,171 while others exceed emissions from
those plants, which can be seen in Figures 4.1.3-1 through 4.

4.1.3.1 Landfills

Landfill gas (LFG) is estimated to be 50% methane and 50% carbon dioxide and water
vapor. However, it also contains less than 1% of non-methane organic components, including
hazardous air pollutants, volatile organic compounds, and nearly 30 other hazardous air
pollutants.172 Using this gas to power gensets in turn causes the release of a variety of pollutants,
including some criteria pollutants. Some pollutants, such as sulfur dioxide, are solely based on
the sulfur content of the raw LFG, while others are based on gas usage and genset configuration.
Table 4.1.3.1-1 shows data of emissions from three different landfill biogas EGUs.173 While this
is a small sample, it nevertheless provides both an indication of the magnitude of the emissions
and the variability depending on the landfill and EGU configuration.

171	eGRID Database, Clean Air Markets Division EPA.

172	https://www.epa.gov/lmop/basic-information-about-landfill gasfmethane

173	Data derived from a CBI source provided to us by a biogas developer/electric utility company.

107

-------
Table 4.1.:

1.1-1: Landfill Biogas E

ataset







Annual















Annual

Power















Landfill

Production











Facility
Type

Size
(MW)

Gas Usage
(MMscf/yr)

to Grid
(MW/yr)

NOx
(tons)

CO
(tons)

voc

(tons)

S02
(tons)

PM
(tons)

Turbine

24.5

4,038.4

132,228

27.2

10.9

5.1

35.6

3.4

Engine

8

1,271.2

57,627.9

39.8

265.5

6.4

2.9

14.6

Engine

1.6

261.4

11,870.1

8.2

54.6

1.3

0.6

3.3

A broader set of emissions data modelled by the EPA Clean Air Market Division (Figures
4.1.3.1-1 through 4) is illustrative of the broad range of potential nitrogen dioxide, sulfur
dioxide, methane, and nitrous oxide emitted from landfills. The emissions modelling in eGRID
for landfills is adjusted based on the assumption that landfills would flare the gas if they did not
combust it for electricity generation. It is therefore assumed that the gas would have been
combusted in a flare and produced some amount of the pollutants tracked in eGRID.174

Figure

120
100
80

-C

> 60
O
z

40
20
0

0	50,000	100,000	150,000	200,000	250,000	300,000

Annual Production (MWh)

4.1.3.1-1: eGRID Landfill NOx Emissions Data

•















































>j,:



•









.Wis.

f •• •••



•

%

•

174 The Emissions & Generation Resource Integrated Database Technical Guide with Year 2020 Data,
https://www.epa.gov/svstem/files/documents/2022-01/egrid202Q technical guide.pdf

108

-------
Figure 4.1.3.1-2: eGRID Landfill SO2 Emissions Data

0.0018
0.0016
0.0014
0.0012

.c

^ 0.001

rsl

O 0.0008

-Q

0.0006
0.0004
0.0002
0

r

50,000	100,000	150,000	200,000

Annual Production (MWh)

250,000

300,000

Figure 4.1.3.1-3: eGRID Landfill Methane Emissions Data

0.2
0.18
0.16
0.14

g 0.12

S

^ 0-1

I

U

£ 0.08
0.06
0.04
0.02
0













•



























































1











• ^

> m m 9













%

•

•
•



EMI











0

50,000	100,000	150,000	200,000

Annual Production (MWh)

250,000

300,000

109

-------
Figure 4.1.3.1-4: eGRID Landfill N2O Emissions Data

0.02

0.018
0.016
0.014
5 0.012
§ 0.01
1 0.008
0.006
0.004
0.002
0

0	50,000	100,000	150,000	200,000	250,000	300,000

Annual Production (MWh)

4.1.3.2 Agricultural Digesters

Agricultural digesters are often under 10 MW and thus are lightly regulated by air permits.
Literature and modeling results from eGRID offer varying estimates of the amount of criteria
pollutant emissions, as seen in Figures 4.1.3.2-1 through 4.

Figure 4.1.3.2-1: eGRID Agricultural Digester NOx Emissions Data

160

140

120

g 100
2

> 80

O

2

_q 60

40
20
0

0























•































4

i















•













•



*• •

•

•





•





•



























.0 5,000.0 10,000.0 15,000.0 20,000.0 25,000.0 30,000.0 35,000.0 40,000.0
Annual Production (MWh)













•



























































~

L «











• W











iW



%

r

•

•
•



•











110

-------
Figure 4.1.3.2-2: eCRID Agricultural Digester SO2 Emissions Data

0.00045

0.0004

0.00035

g 0.0003

2 0.00025

o 0.0002
1/1

£ 0.00015
0.0001
0.00005
0

0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000
Annual Production (MWh)

•















































































































• 1



















• •••

•

•





•

Figure 4.1.3.2-3: eGRID Agricultural Digester Methane Emissions Data

















*















































































m 1
•• •



••••

•

•





•

0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000
Annual Production (MWh)

Figure 4.1.3.2-4: eGRID Agricultural Digester N2O Emissions Data



0.035



0.03



0.025

g



2

0.02

O"



(N

0.015

~Z.



-O





0.01



0.005

0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000
Annual Production (MWh)

111

-------
4.1.3.3 Wastewater Treatment Plants

Figure 4.1.3.3-1 through 4 contains the EPA Clean Air Market Division's modelled data
on wastewater treatment plants. These facilities have higher nameplate capacities than
agricultural digesters and tend to be located nearer to urban areas, and are therefore more often to
have their emission regulated. These facilities are most likely to also be using CHP/on-site
electricity generation as a cost savings measure as they currently supply their own operational
electricity needs.

Figure 4.1.3.3-1: eGRID Wastewater Treatment Plant NOx Emissions Data

140
120

c 100

X

g 60

-Q

~~ 40

20
0

0 20,000 40,000 60,000 80,000 100,000 120,000 140,000 160,000 180,000
Annual Production (MWh)



















•



































•

¦



»













tiK

•















¦ % *
••••

• •
•

















•

•









•

•

Figure 4.1.3.3-2: eGRID Wastewater Treatment Plant SO2 Emissions Data

0.0007

0.0006

c 0.0005

| 0.0004

O 0.0003
1/1
-Q

0.0002
0.0001
0

0 20,000 40,000 60,000 80,000 100,000120,000 140,000 160,000180,000
Annual Production (MWh)





















•

































•





















































•

« ••

> •









•

•

112

-------
Figure 4.1.3.3-3: eGRID Wastewater Treatment Plant Methane Emissions Data

0.25

0.2

0.15

5 o.i

-Q

0.05























































•

•

















ft
••

» •









•

•

0 20,000 40,000 60,000 80,000 100,000 120,000 140,000 160,000 180,000
Annual Production (MWh)

Figure 4.1.3.3-4: eGRID Wastewater Treatment Plant N2O Emissions Data

0.025

0.02

0.015

O

r\l

0.01

0.005

0 20,000 40,000 60,000 80,000 100,000 120,000 140,000 160,000 180,000
Annual Production (MWh)

4.1.4 Air Quality Modeling

As mentioned at the beginning of Chapter 4.1, air quality impacts resulting from this
proposal are expected to be relatively minor. Any significant impacts are likely to be highly
localized, and thus, would likely not be captured at the geographic scale used in photochemical
air quality modeling. Thus, no air quality modeling was done. The largest impacts from this
proposal are expected to be increased use of renewable diesel and biogas for electricity.
Available data limit our ability to quantify renewable diesel impacts, and increases in biogas to
electricity emissions are not expected from this proposal. Geographic distribution of emissions
also varies, and a comprehensive evaluation of offsetting impacts is very complex. Furthermore,
to the extent that this rule is associated with reductions in imported refined petroleum products,
those upstream emissions and the adverse impacts they cause would occur in foreign countries.
Such upstream international impacts are typically considered outside the scope of an RIA or
other analysis used to support a rulemaking.

113

-------
4.2 Climate Change

CAA section 211 (o) (2) (B) (ii) states that the basis for setting applicable renewable fuel
volumes after 2022 must include, among other things, "an analysis of.. .the impact of the
production and use of renewable fuels on the environment, including on.. .climate change."

While the statute requires that EPA base its determinations, in part, on an analysis of the climate
change impact of renewable fuels, it does not require a specific type of analysis. While the
impacts of climate change include rising temperatures and sea levels, ocean acidification,
increased occurrence and intensity of wildfires and extreme weather events, and other impacts,175
these impacts are driven by changes in greenhouse gas emissions. Since the CAA requires
evaluation of lifecycle greenhouse gas (GHG) emissions as part of the RFS program, we believe
the CAA gives us the discretion to use lifecycle GHG emissions estimates as a reasonable proxy
for climate change impacts.

Our assessment of the climate change impacts of the candidate volumes relies on an
extrapolation of lifecycle analyses (LCA) of GHG emissions.176 As we did in the 2020-2022
RVO rulemaking, this approach involves multiplying LCA emissions of individual fuels by the
change in the consumption of each fuel in the candidate volumes scenario relative to the No RFS
baseline to quantify the GHG impacts. We repeat this process for each fuel (e.g., corn ethanol,
soybean biodiesel, landfill biogas CNG) to estimate the overall GHG impacts of the candidate
volumes. In the 2020-2022 RVO rulemaking, we applied the LCA estimates that we developed
in the March 2010 RFS2 rule (75 FR 14670) and in subsequent agency actions. In this
rulemaking, we are updating our approach to use a range of LCA emissions estimates that are in
the literature. Instead of providing one estimate of the GHG impacts of each candidate volume,
we provide a high and low estimate of the potential GHG impacts, which is inclusive of the
values we estimated in the 2010 RFS final rule and subsequent agency actions. We then use this
range of values for considering the GHG impacts of the candidate renewable fuel volumes that
change relative to the No RFS baseline.

This section of the DRIA discusses our evaluation of the potential effects of the candidate
volumes on GHG emissions. We start with background on our LCA of the GHG emissions
associated with biofuels since the beginning of the RFS2 program in 2010. We then discuss
advances in the modeling science since 2010, including a brief summary of currently available
models that can be used to estimate biofuel GHG emissions. We then discuss how these models
compare across a number of important characteristics. We review available biofuel LCA
estimates and make observations about why some estimates are higher or lower than others. For
example, we discuss the potential influence of model choice relative to input assumptions. Based

175	Wuebbles, D.J., D.W. Fahey, K.A. Hibbard, B. DeAngelo, S. Doherty, K. Hayhoe, R. Horton, J.P. Kossin, P.C.
Taylor, A.M. Waple, and C.P. Weaver, 2017: Executive summary. In: Climate Science Special Report: Fourth
National Climate Assessment, Volume I [Wuebbles, D.J., D.W. Fahey, K.A. Flibbard, D.J. Dokken, B.C. Stewart,
and T.K. Maycock (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, pp. 12-34, doi:
10.7930/J0DJ5CTG.

176	In this chapter, we use a range of terminology, consistent with the scientific literature, to describe the concept of
lifecycle GHG emissions. We sometimes call lifecycle GHG emissions "LCA emissions," "LCA ranges," LCA
values," "LCA estimates," "carbon intensity (CI)," or some combination of these terms. For purposes of this
discussion, the meaning of these terms is the same, namely the GHG emissions associated with all stages of fuel
production and use, including significant indirect GHG emissions.

114

-------
on our literature review, we produce a range of LCA carbon intensity estimates for each biofuel
pathway affected by the candidate volumes. We use these ranges along with the volume
scenarios discussed in Chapter 3 to produce a range of potential GHG emissions impacts.

Finally, we monetize this range of GHG emissions to produce an estimate of the monetized GHG
benefits associated with the candidate volumes.

For the final rule, we intend to conduct a model comparison exercise that will produce
new estimates of crop-based biofuel GHG emissions from multiple models to expand upon what
is already in the literature for the types of volume changes required. By running common
scenarios and aligning results, the model comparison exercise should allow us to compare model
estimates more directly, giving us further understanding of the best available science on the
GHG impacts associated with biofuels. As we consider updates to our LCA methodology, we
will consider the broad range of new science related to biofuel LCA, including the insights from
the model comparison exercise.

As discussed elsewhere in this document, the science associated with the lifecycle
assessment of biofuels is a much discussed topic. Significant analytical work has been
undertaken since EPA laid out its lifecycle methodology in the 2010 RFS rulemaking, with work
in this area continuing. For example, in October 2022, the National Academies of Science,
Engineering and Medicine published a report titled "Current Methods for Life Cycle Analyses of
Low-Carbon Transportation Fuels in the United States (2022)." This report assesses the current
methods of estimating lifecycle GHG emissions associated with transportation fuels used in a
potential national low-carbon fuels program. While this report does not endorse any particular
numerical result or model, it provides useful insights into estimations of GHG emissions over
each part of the lifecycle of a given fuel, indirect GHG emissions, and data quality and quantity.
EPA is carefully reviewing this report as it adds to the feedback EPA received on lifecycle
assessment through its LCA workshop held earlier this year. We also note the Administration, as
part of its SAF Grand Challenge, has created a workgroup between DOE, EPA, FAA, and USDA
to look at LCA methodologies and data needs specifically related to renewable aviation fuel,
which will also be a useful platform in assessing LCA capabilities and uncertainties. As EPA
uses LCA models not just for RFS analysis of program performance and feedstock assessment
but also broader policy analysis, the Agency would benefit from updating its existing set of
analytical methodologies. Data and findings from recent science and the modeling comparison
exercise will help inform EPA's specific next steps on updating its methodology as part of a
separate action.

4.2.1 Background on Renewable Fuel GHG Analysis for the RFS Program

A primary policy goal of the RFS program is to reduce GHG emissions by increasing the
use of 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. Depending on the
LCA of the fuel, renewable fuels can provide a substantial GHG emission reduction.

115

-------
4.2.1.1	Clean Air Act Requirements

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.177 The CAA specifically defines the term "lifecycle greenhouse gas
emissions" to mean "the aggregate quantity of greenhouse gas emissions (including direct
emissions and significant indirect emissions such as significant emissions from land use
changes), as determined by the Administrator, related to the full fuel lifecycle, including all
stages of fuel and feedstock production and distribution, from feedstock generation or extraction
through the distribution and delivery and use of the finished fuel to the ultimate consumer, where
the mass values for all greenhouse gases are adjusted to account for their relative global warming
potential."178 In the March 2010 RFS2 rule (75 FR 14670), EPA interpreted the provision
"including direct emissions and significant indirect emissions" as requiring our LCA to consider
the consequential, or market-mediated, impacts of increased demand for renewable fuels.

Indirect emissions, by definition, cannot be directly measured in the way that direct emissions
can be calculated. Indirect emissions result from changes in prices (e.g., agricultural
commodities or petroleum prices) that ripple through the economy. For example, if increased
consumption of renewable fuel in the U.S. diverts U.S. exports of corn from the global markets,
the market-mediated impact could be for other countries, such as Argentina, to produce more
corn to supply the global demand for cereal grains. While all of the corn used to produce ethanol
in the scenario may have been grown in the U.S., the land use changes emissions in Argentina
would be considered "indirect" land use change emissions. Other examples of market-mediated
impacts include changes in livestock production that result from increased production of
renewable fuel co-products such as soybean meal. If the increased production of soybean meal
leads to a decrease in feed prices for cattle, an indirect impact could include the increased
production of beef.

While the term "significant indirect emissions" requires some analytical judgement, prior
modeling work has indicated that the indirect impacts from land use change, livestock, and crop
production can result in emissions that can have a large impact on the lifecycle analysis.
Therefore, to be consistent with the CAA requirements, our lifecycle analysis must take into
account global agricultural and livestock markets, since many biofuel feedstocks use globally
traded commodities. In addition, the increasing interdependence of the energy and agricultural
markets suggests that capturing indirect energy sector impacts could have important implications
for lifecycle analysis. We consider these statutory requirements when describing the differences
in modeling frameworks in Chapter 4.2.2.7.

4.2.1.2	Lifecycle Analysis Under the RFS Program

As part of the March 2010 RFS2 rule, 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. At the time of the

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

178	See 42 USC 7545(o)(l)(H).

116

-------
analysis for the 2010 RFS2 rule, there were no models available off the shelf that could perform
the type of lifecycle analysis required by EISA. Thus, EPA developed a new modeling
framework to perform the required analysis. The framework we developed used multiple models
and data sources.179 We used the Forest and Agricultural Sector Optimization Model with
Greenhouse Gases model (hereinafter referred to as "FASOM") and the FAPRI-CARD model
(Food and Agricultural Policy Research Institute international model; hereinafter referred to as
"FAPRI") developed at the Center for Agriculture and Rural Development at Iowa State
University. We ran aligned scenarios in both models and used FASOM to estimate domestic
agricultural and forestry sector impacts, and the FAPRI model to estimate international
agricultural sector impacts. Our framework included data from many other sources, including
emissions factors and other data from the GREET model (see below for a description of
GREET). We proposed this new framework for public comment, organized four peer reviews of
different aspects of it and held a public workshop. Based on all of this input we refined the
modeling for the final March 2010 RFS2 rule. We also estimated the uncertainty associated with
the land use change satellite data and emissions factors used in our analysis. The framework we
developed in 2010 used the best science, data and models available at the time.

Since the 2010 RFS2 rule, we have used the RFS2 modeling framework to conduct
numerous (over 140) analyses of new pathways and their lifecycle GHG emissions. Based on
these analyses, we have approved additional pathways for participation in the RFS program.
These pathways rely on novel feedstocks (e.g., canola oil, grain sorghum, camelina oil,180
distillers sorghum oil181) and novel production processes involving existing feedstocks (e.g.,
catalytic pyrolysis and upgrading of cellulosic biomass,182 gasification and upgrading of crop
residues183). 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."184 Our lifecycle analyses of various pathways are also published online.185 A list of
pathways that have been approved by regulation can also be found at 40 CFR 80.1426(f)(1).

Depending on the renewable fuel, the feedstocks used to produce it, the amount of fossil
energy used in growing the feedstocks and producing the fuel, land use change and associated
agricultural emissions, and other factors, the GHG emission reductions will vary considerably. In
general, we have found that renewable fuels that are not expected to have significant impacts on

179	EPA (2010). Renewable fuel standard program (RFS2) regulatory impact analysis. Washington, DC, US
Environmental Protection Agency Office of Transportation Air Quality. EPA-420-R-10-006. Chapter 2.4.

180	March 2013 Pathways I rule. 78 FR 14190. ittps://www. epa.gov/renewable-fuel-standard-prograni/final-rule-
additional-qualifying-renewable-fuel-pathways-under

181	August 2018 sorghum oil rule. 83 FR 37735. https://www.gpo.gov/fdsys/pkg/FR-2018-08-02/pdf/2Q18-

18248.pdf

182	March 2013 Pathways I rule. 78 FR 14190. ittps://www. epa.gov/renewable-fuel-standard-prograni/final-rule-
additional-qualifying-renewable-fuel-pathways-under

183	"San Joaquin Renewables Fuel Pathway Determination under the RFS Program." May 11, 2020.
https://www.epa.gov/renewab1e-fue1-standard-prograni/san-joaquin-renewab1es-approva1

184	ThLs document Is available on EPA's website at: https://www.epa.gov/fue1s-registration-reporting-and-
conipliance-help/1 ifecycle-greenhouse-gas-results. This summary is also available in docket EPA-HQ-OAR-2021 -
0324.

185	See https://www.epa.gov/renewab1e-fue1-standard-program/approved-pathways-renewab1e-fue1 and
https://www.epa.gov/renewab1e-fue1-standard-prograni/other-actions-renewab1e-fue1-standard-program

117

-------
land use, such as fuels produced from wastes, residues, or by-products, have greater GHG
emission reductions than renewable fuels produced from crops intended to be used as feedstock
for renewable fuel production. For instance, with respect to biodiesel and renewable diesel
production, the use of waste fats, oils, and greases (FOG) as feedstocks typically results in lower
lifecycle GHG emissions compared to use of vegetable oils, such as soybean or canola oil.186 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.187

Since the existing LCA methodology was developed for the March 2010 RFS2 rule, there
has been more research on the lifecycle GHG emissions associated with transportation fuels in
general and crop-based biofuels in particular. New models have been developed to evaluate
biofuels and more models have been developed for other purposes have been modified to
evaluate the GHG emissions associated with biofuel production and use. There has also been
rapid growth in available data on land use, farming practices, crude oil extraction and many other
relevant factors. While our existing LCA estimates for the RFS program remain within the range
of more recent estimates, we acknowledge that our previously relied on biofuel GHG modeling
framework is comparatively old and an updated framework is needed. Accordingly, EPA has
initiated work to develop a revised modeling framework of the GHG impacts associated with
biofuels. In consultation with our interagency partners at USDA and DOE, we hosted a virtual
public workshop on biofuel GHG modeling on February 28 and March 1, 2022.188 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 response to the public docket we opened for the workshop, we received approximately
30 comments, with the majority coming from industry representatives.189 A large majority of the
comments expressed support for the workshop. Many comments asked EPA to change the way it
models biofuel GHG emissions, with updated data and/or a different model or combination of
models. Most of these latter commenters asked EPA to consider using The Greenhouse Gases,
Regulated Emissions, and Energy Use in Technologies Model (GREET), though at least one
commentor asked EPA to maintain its current model structures. Several other commenters asked
EPA to consider applying a risk-based approach when considering biofuel targets and pathway
approvals. Other commenters pointed to the importance of using real-world data. All of the
commenters supported ongoing engagement with EPA on the topics discussed during the
workshop.

186	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/fue1s-registration-reporting-arid-
coiiipliance-help/1 ifecvcle-greenhouse-gas-results.

187	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.

188	For more information see the Federal Register Notice, "Announcing Upcoming Virtual Meeting on Biofuel
Greenhouse Gas Modeling." 86 FR 73756. December 28, 2021. More information is also available on the workshop
webpage: https://www.epa.gov/renewable-fuel-standard-prograni/workshop-btofuel-greenhouse-gas-modeling.

189	Docket number: EPA-HQ-OAR-2021-0921

118

-------
Based on the workshop presentations and public input, it is clear that there continues to
be significant uncertainty and a wide range of estimates on the climate effects of biofuels,
especially when it comes to biofuel-induced land use change emissions. Uncertainties in land use
change emissions estimates stem from both economic modeling of market-mediated effects as
well as biophysical modeling of soil carbon and other biological systems. The workshop
proceedings, including the workshop presentations and the comments submitted to the workshop
docket, touch on a broad and complex set of topics. A general theme that emerged from this
process is that, in support of a better understanding of the lifecycle GHG impacts of biofuels, it
would be helpful to compare available models, identify how and why the model estimates differ,
and evaluate which models and estimates align best with available science and data. The rest of
this section makes progress in this direction by reviewing available models and published LCA
estimates. To further explore the differences between the available models, for the final rule we
intend to conduct new modeling and compare the results.

4.2.2 Review of Available Models for Renewable Fuel GHG Analysis

There are many factors that influence biofuel GHG estimates, including model
framework choice, data inputs and assumptions and other methodological decisions. In this
section we discuss available models. To the extent possible based on available information, we
also discuss how the data and assumptions used in these models influence the biofuel GHG
estimates they produce. The February 2022 biofuel GHG modeling workshop included
presentations on five models (The Greenhouse Gases, Regulated Emissions, and Energy Use in
Technologies Model (GREET), Global Biosphere Management Model (GLOBIOM), Global
Change Analysis Model (GCAM), Global Trade Project (GTAP), and Applied Dynamic
Analysis of the Global Economy (ADAGE) used for biofuel GHG analysis. In this section we
provide a summary of each of these models including their history, sectoral representation,
spatial coverage and resolution, temporal representation, and GHG emissions representation.

This selection of models provides a broad cross-section of the most common types of modeling
frameworks used to assess biofuels, as discussed in the following paragraph. We chose to
highlight these models based on discussions with our partners at USDA and DOE and our
experience reviewing scientific literature on the lifecycle GHG emissions of biofuels. In
addition, our choice to highlight these particular models is also informed by the statutory
requirement to evaluate significant indirect emissions, including indirect land use change
emissions. Furthermore, we are guided by our decision in the 2010 RFS2 rule to include
significant indirect emissions occuring anywhere in the world (i.e., international impacts) given
that GHG emission impacts are global. We also include a brief summary of other models that
have been used for biofuel analysis. Models that exclude indirect emissions or are limited in
geographic scope receive less attention in our review, except insofar as they can inform a broader
analysis that meets the statutory requirements. We then compare the characteristics of all of these
models and make some observations about what may be contributing to the different biofuel
GHG estimates they have produced. Our goal is not to provide a comprehensive accounting of
any one of these models or the differences between them. Rather, our objective is to summarize
each model at a high level and highlight some of the differences between them that we intend to
explore further as part of the model comparison exercise for the final rule.

119

-------
There are four general types of models commonly used for biofuel GHG analysis:
lifecycle inventory (LCI) models, partial equilibrium (PE) models, computable general
equilibrium (CGE) models and integrated assessment models (IAM). LCI models, such as
GREET, are designed to estimate in detail the inputs and outputs of a product supply chain, using
rule-based methods (i.e., allocation or displacement) to account for co-products. PE models, such
as GLOBIOM,190 equate supply and demand in one or more markets such that prices stabilize at
their equilibrium level. PE models offer highly detailed representations of one or a few sectors of
the economy, such as the agricultural sector, but lack linkages to other sectors of the economy. In
contrast, CGE models, such as GTAP and ADAGE, are comprehensive in their representation of
the economy, reflecting feedback effects among all economic sectors and factors of production,
such as capital and labor. IAMs, such as GCAM, integrate knowledge from several disciplines,
for example, biogeochemistry, economics, engineering, and atmospheric science, to evaluate
how changes in any of these areas affect the others. While it is hard to state the specific criteria
for identifying an IAM, we might distinguish them from PE and CGE models by their deeper
integration of human economic systems with Earth (biosphere and atmosphere) systems and
GHG emissions into one modelling framework. LCI models, such as GREET, are designed to
estimate in detail the inputs and outputs of a product supply chain, using rule-based methods
(i.e., allocation or displacement) to account for co-products. PE, CGE and IAM models can all be
called economic models. LCI models are categorically different from the other three model types
as they do not simulate economic behavior or prices. Across the four model types there tends to
be a tradeoff between scope and detail, which are discussed in more detail in section 4.2.2.7.

4.2.2.1 The GREET Model

The Greenhouse gases, Regulated Emissions, and Energy use in Technologies (GREET)
Model is a lifecycle analysis model. It provides well-to-wheels lifecycle energy, water, GHG,
and other air emissions results intended to evaluate the impacts of various vehicle and fuel
combinations. The developer is Argonne National Laboratory (ANL), and the project is
sponsored by the U.S. Department of Energy (DOE). Initially made available in 1995, it was
developed with the purpose of evaluating the energy and emission impacts of new fuels and
vehicles for use in the transportation sector.191

GREET includes a suite of models and tools. It includes a fuel cycle model of vehicle
technologies and transportation fuels (GREET 1) and a vehicle manufacturing model of vehicle
technologies (GREET2). Given that our focus is on renewable fuels, we are primarily concerned
with GREET1. GREET is available in two platforms, a large Excel workbook and a ".net"
version. The Excel version of GREET provides transparency while the .net version offers a
modular user interface with a structured database. There are several derivates of the core GREET
model, such as CA-GREET developed with the California Air Resources Board (CARB) and
used in support of the California Low Carbon Fuels Standard (CA-LCFS), and ICAO-GREET
developed with the International Civil Aviation Organization in support of the Carbon Offsetting
and Reduction Scheme for International Aviation (CORSIA). New versions of GREET are

190	The FASOM and FAPRI models EPA used for the March 2010 RFS2 rule biofuel GF1G analysis are also
categorized as PE models.

191	Elgowainy, A. and Wang, M. (2019) 'Overview of Life Cycle Analysis (LCA) with the GREET Model', p. 21.
Available at: https://gfeet.es.anl.gov/files/workshop 2019 overview.

120

-------
released in October of each year, with the latest version as of the time of this writing being
GREET-2021. GREET includes more than 100 fuel production pathways including fuels used in
road, air, rail, and marine transportation. It also examines more than 80 on road vehicle/fuel
systems for both light and heavy-duty vehicles. The model reports lifecycle energy use, air
pollutants, GHGs and water consumption. It includes detailed representations of the petroleum,
electric, natural gas, hydrogen, and renewable energy sectors.

The GREET modeling framework is a process-based LCA approach (sometimes referred
to as attributional LCA).192 GREET can be used to estimate the carbon intensity of individual
supply chains and the benefits of specific supply chain adjustments, such as reducing fertilizer
application rates or switching to more efficient fuel distribution modes. Fundamentally, GREET
is most closely related to other lifecycle inventory (LCI) models such as SimaPro, GaBi, and
OpenLCA. In general, GREET assumes that additional production of a fuel commodity entails a
linear increase in the activities from the associated supply chain with no market-mediated effects
on other supply chains or economic sectors (e.g., diverting certain crops to biofuels may lead to
new or more land area devoted to agriculture, increased use of fertilizers in addition to other
energy inputs, and higher food or feed prices that in turn change what agricultural products are
consumed by both people and livestock). Figure 4.2.2.1-1 provides a schematic overview of how
the biofuel lifecycle is represented in GREET. GREET can be used to estimate the carbon
intensity of individual supply chains and the benefits of specific supply chain adjustments, such
as reducing fertilizer application rates or switching to more efficient fuel distribution modes.

Figure 4.2.2.1-1: Schematic of Biofuel Supply Chain Representation in GREET193

¦*	Fuel Production (Well to Pump)	~

GREET primarily estimates fuel carbon intensities using data for average resource and
energy production in the United States. For example, GREET by default models electricity based

192	Wang, M. (2022). "Biofuel Life-cycle Analysis with the GREET Model." Presentation at the EPA Biofuel
Modeling Workshop. Argonne National Laboratory. March 1, 2022.

https://www.epa.gov/svstem/files/documents/2022-03/biofuel ghg-model-workshop-biofuel-lifecvcle-analvsis-
greet-model-2022-03-01.pdf. Slide 5.

193	Copied from Wang (2022), slide 9.

121

-------
on data for average U.S. electricity generation. However, GREET includes some pathways
representing foreign fuel production (e.g., Brazilian sugarcane ethanol) and in some cases users
can choose to model some supply chains located in particular regions of the U.S. (e.g., states or
electricity grid regions). A user with enough data on their supply chain could in certain cases
customize GREET to estimate the carbon intensity of their fuel considering regional details and
particular suppliers of energy and material inputs.

In general, GREET is not a dynamic or temporal model. Specifically, GREET does not
take into account significant indirect emissions associated with increased biofuel demand, such
as through market-mediated impacts on the agriculture, livestock, or energy sectors. GREET
accounts for important biofuel co-products such as distillers grains and soybean meal through
allocation or displacement rules. Furthermore, GREET is designed to estimate the carbon
intensity of a fuel in a particular year, not modeling cascading changes across the economy and
annualizing that stream of associated emission changes over time. However, the input data can
be customized to estimate carbon intensities in any given year. Thus, it can be used to show how
the estimated carbon intensity of a fuel changes over time based on changes in technological
efficiency and other factors. For example, Lee et al. (2021) used data on U.S. ethanol production
efficiencies and corn yields to estimate the carbon intensity of U.S. corn ethanol each year from
2005 to 2019.194

Although GREET does not endogenously estimate indirect emissions such as those
resulting from indirect land use change, GREET incorporates a static module called the Carbon
Calculator for Land Use Change from Biofuels Production (CCLUB) to account for indirect land
use change emissions.195 The opposing terms "static" and "dynamic" are used in different ways
in modeling literature. In this instance, we describe CCLUB as a static module because it
estimates land use area changes for one time period based on simulations with the GTAP-BIO
model (discussed more below). In contrast, dynamic models, such as GCAM and GLOBIOM
(discussed below), simulate land use changes and resulting GHG emissions over multiple
decades. CCLUB relies on a selection of land use change estimates from GTAP-BIO studies
conducted between 2011-2018 (see Table 4.2.2.1-1), combined with emissions factors from
other sources to estimate land use change GHG emissions per unit of biofuel production.196
Thus, the well-to-wheel emissions for crop-based pathways are estimated as the process-based
emissions plus the induced land use change estimates from CCLUB. The data sources and
calculations in CCLUB are summarized in Figure 4.2.2.1-2, reproduced from the CCLUB user
manual.

194	Lee, U., et al. (2021). "Retrospective analysis of the US corn ethanol industry for 2005-2019: implications for
greenhouse gas emission reductions." Biofuels, Bioproducts and Biorefining.

195	Dunn, J. B., et al. (2017). Carbon calculator for land use change from biofuels production (CCLUB) users'
manual and technical documentation, Argonne National Lab, Argonne, IL.

196	Hoyoung Kwon and Uisung Lee (2019) 'Life Cycle Analysis (LCA) of Biofuels and Land Use Change with the
GREET Model'. Available at: https://greet.es.an1.gov/fi1es/workshop 2019 biofuel luc.

122

-------
Figure 4.2.2.1-2: Schematic of Data Sources and Calculations in CCLUB197

Models & Data
Sources

Output

CCLUB
Calculations

Results

Domestic &
International

Domestic
Only

Parameterized
CENTURY

GTAP

Biofue) production
scenarios

land transitions by area *r>d
type







Winrock &
Woods Hole
Datasets



Abowigrotmd carbon
stocks (for forest ancfcr
grassland)

lielowground or soil carbon
stocks



Adjust US forest area baseline
with US Forest Service data

Carton emission factors

LUC scenarios

LMC scenarto*

Carbon Online
Estimator





Beiowground or soil carbon

stocks

Abovegrourtd carbon
stocks (only Cot for«st)

IPCC NjO emission factors

IPCC CH« emission factors —

So»l carton emission factors

N,0 emission factors

Carbon tfflltSbn fatlori of
Harvested wood product

<3HG emissions (g CO^e'MJ)
by combining la^d arsa
changes with emissions
factors arid applying
assumptions

Soil carboo emissions

{& C03ertij)

by- applying assumptions

Spatial coverage

County
AEZ
CountTv/Biomc

CCLUB includes land use change area estimates from nine different GTAP-BIO
scenarios: four soy biodiesel shocks, two corn ethanol shocks, and one shock each for ethanol
from corn stover, miscanthus and switchgrass. The corn ethanol and soy biodiesel scenarios
included in CCLUB are described in Table 4.2.2.1-1. The two corn ethanol scenarios are similar
except that the "Corn Ethanol 2013" estimate was produced with a version of GTAP-BIO with
regionally differentiated land transformation elasticities and a modified land nesting structure to
that makes it more costly within the model to convert forest to cropland relative to converting
pasture to cropland.

197 Kwon, Hoyoung, Liu, Xinyu, Dunn, Jennifer B., Mueller, Steffen, Wander, Michelle M., and Wang, Michael.
(2020). Carbon Calculator for Land Use and Land Management Change from Biofuels Production (CCLUB). United
States: N. p., 2020. Web. doi: 10.2172/1670706. Copy of Figure 1.

123

-------
Table 4.2.2.1-1: Corn Starch and Soybean Oil Based Biofuel Scenarios Modeled in
CCLUB198

Case Description

Shock Size
(Billion
Gallons)

Source

"Corn Ethanol 2011." An increase in corn ethanol
production from its 2004 level (3.41 billion gallons
[BG]) to 15 BG

11.59

Taheripour et al.

(2011)199

"Corn Ethanol 2013." An increase in corn ethanol
production from its 2004 level (3.41 billion gallons
[BG]) to 15 BG

11.59

Taheripour and
Tyner (2013)200

Increase in soy biodiesel production by 0.812 BG
(CARB case 8)

0.812

Chen et al.
(2018)201

Increase in soy biodiesel production by 0.812 BG
(CARB average proxy)

0.812

Chen et al. (2018)

Increase in soy biodiesel production by 0.8 BG
(GTAP 2004)

0.8

Taheripour et al.
(2017)202

Increase in soy biodiesel production by 0.5 BG
(GTAP 2011)

0.5

Taheripour et al.
(2017)

For each case, the estimates CCLUB uses from GTAP-BIO are the area of changes in
cropland, forest, pasture in each agro-ecological zone (AEZ) and region, and cropland pasture in
the U.S. and Canada. Land use change GHG emissions are estimated based on these land
conversion areas using data from a few different sources. Based upon user selections, CCLUB
ultimately combines a given GTAP scenario's expected land use change impacts with the user-
selected emission factor data to provide domestic and international GHG emissions per
functional unit of analyzed biofuel to GREET as the land use change emissions component for a
given biofuel of interest.

A module called the Feedstock Carbon Intensity Calculator (FD-CIC) was recently added
to GREET.203 FD-CIC is designed to examine carbon intensity variations of different corn,
soybean, sorghum, and rice farming practices at the farm level. The FD-CIC uses county level
data and allows users to input their own farm level data on energy and chemical farming inputs,
tillage, cover cropping and other crop management practices. Based on these input data, the FD-

198	Adapted from Table 1 in Dunn, J. B., et al. (2017). Carbon calculator for land use change from biofuels
production (CCLUB) users' manual and technical documentation, Argonne National Lab.(ANL), Argonne, IL
(United States).

199	Taheripour, F., et al. (2011). Global land use change due to the U.S. cellulosic biofuels program simulated with
the GTAP model, Argonne National Laboratory: 47.

200	Taheripour, F. and W. E. Tyner (2013). "Biofuels and land use change: Applying recent evidence to model
estimates." Applied Sciences 3(1): 14-38.

201	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.

202	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.

203	Liu, X., et al. (2020). "Shifting agricultural practices to produce sustainable, low carbon intensity feedstocks for
biofuel production." Environmental Research Letters 15(8): 084014.

124

-------
CIC estimates the farm level emissions from energy, fertilizers, herbicide, and insecticide, as
well as effects on soil organic carbon relative to the baseline assumptions in GREET. The FD-
CIC may be useful to estimate the soil carbon benefits of reduced tillage and cover cropping, and
to examine regional differences in feedstock carbon intensity.

GREET is used by a variety of academic, commercial, and government users.

California's Low Carbon Fuel Standards (LCFS) program relies in part on a customized version
of GREET called CA-GREET to provide state-specific fuel pathways and carbon intensity
values.204 Oregon uses a similar approach for their LCFS program.205 The International Civil
Aviation Organization (ICAO) used GREET among several models to provide carbon intensities
for specific aviation fuel pathways.206 Most of these programs use the non-land use change GHG
estimates from GREET and add their own land use change estimates instead of those derived
from CCLUB to calculate biofuel carbon intensities. Among other applications, EPA has used
GREET since the inception of the RFS program to provide data for rulemakings and pathway
support as part of our suite of tools in addition to FASOM and FAPRI.

4.2.2.2 The GLOBIOM Model

The Global Biosphere Management Model (GLOBIOM) was developed and continues to
be managed by the International Institute for Applied Systems Analysis (IIASA). The model was
developed in the late 2000s originally to conduct impact assessments of climate change
mitigation policies of biofuels and other land-based efforts.207 It was developed on the basis of
the U.S. Forest and Agricultural Sector Optimization Model (FASOM model).208 There are
several model versions of GLOBIOM available for different contexts. A sample of GLOBIOM
code is available to the public, and an open source version is in the works.209

204	California Air Resources Board. LCFS Life Cycle Analysis Models and Documentation. Available at:
https://ww2.arb.ca.gov/resources/docunients/lcfs-Hfe-cycle-analysis-models-and-documentation (Accessed: 29 April
2022).

205	Oregon Department of Environmental Quality. Carbon Intensity Values: Oregon Clean Fuels Program. Available
at: https://www.oregon.gov/deq/ghgp/cfp/Pages/Clean-Fuel-Pathways.aspx (Accessed: 29 April 2022).

206	ICAO. Models and Databases. Available at: https://www.icao.int/environniental-protection/pages/modeinng-and-
d cUb h ases.a s p x (Accessed: 29 April 2022).

207	International Institute for Applied Systems Analsyis, "GLOBIOM,"
https://previous.iiasa.ac.at/web/home/research/GLUBIUM/GLUBIUM.html.

208	Frank, Stefan, et al. "The Global Biosphere Management Model,"

https://www.epa.gov/systeni/files/docunients/2022-03/biofuel-ghg-niodel-workshop-global-biosphere-nignit-niodel-
2022-03-01.pdf. And, Valin, Hugo et al. The Land Use Change Impact of Biofuels Consumed in the EU:
Quantification of Area Greenhouse Gas Impacts. August 27, 2015, pg. 128.

https://ec.europa.eu/energy/sites/ener/files/docunients/Final%20Report GLOBIOM publication.pdf

209	See, GLOBIUM, "Model Code," https://iiasa.github.io/GLOBIOM/model code.html.

125

-------
Figure 4.2.2.3-1: GLOBIOM Regional Mapping"10

GLOBIOM is a PE model that captures the agricultural, forest, and bioenergy sectors.
The model solves recursively dynamic using a spatial equilibrium modeling approach with
detailed grid cell coverage. The model finds market equilibriums that maximize the sum of
producer and consumer surplus subject to resource, technological, demand and policy
constraints. Producer surplus is defined as the difference between market prices at a regional
level and the product's supply curve. The supply curve takes into account labor, land, capital and
other purchased input. Consumer surplus is based on the level of consumption of each market
and is arrived at by integrating the difference between the demand function of a good and its
market price. The model uses linear programming to solve, although it also contains some non-
linear functions that have been linearized using stepwise approximation.211 The global trunk
version of the model features global coverage with 37 regions (see Figure 4.2.2.3-1) and
simulates for the years 2000-2100 using ten-year time-steps. As a PE model, GLOBIOM does
not have feedback from labor, capital, or other parts of the economy, however, the model can be
linked to other models, such as IIASA's energy sector model MESSAGE.

210IIASA. (2020). "GLOBIOM regional and country level modeling." SUPREMA GLOBIOM-MAGNET Training.
December 4, 2020. https://iiasa.github.io/GLOBIOM/training material/GLOBIOM/GLOBIOM
Topic RegionalApplications APalazzo Nov2020.pdf
211 IIASA, "GLOBIOM Documentation_20180604.pdf,"
https://iiasa.github.io/GLOBIOM/GLOBIOM Documentation 20180604.pdf.

126

-------
Figure 4.2.2.3-2: Schematic Overview of the GLOBIOM Model212

~D
C
TO

E
o

a}
J*
k.

TO

s

c
o

u
3
T3
O

OJ
1/1

=3

TD
C
TO



Food

Fibers Energy

Industry



1
t

1 - _



MARKET & TRADE: EU + WORLD PRICES

i

ewe

Crop model

RUMINANT

Digestibility model

BIOCNERGY
Processing

G4M
Global Forest model

World wide:
18 crops (FAO *5PAMJ
Management systems:
low/high input® irrigated

EU28:
9 additional crop^
crop rotati-ons.
Management opOons
fertiliser, irrigation & tillage

-> Feed intake
-> Animal production
GHG emissions

7 animate
If AO t  MJ bfofuel
-> MJ bioelectric
-> Coproducts

*4 Harvestabie wood
Harvesting costs

Perennial crop*
Short rotation topples

Conversion technologies
Rrat generation biofuels
Second generation
blofuals

Biamass power plants

Short .rotation

Down scaled F AO FftA at
grid level

Area
Carter stcci
Age
TniasiW
Species

Rotation time
Thinrvng

Grtdded representation of world land use

Managed forest

Natural
forest.

Other
natural land

The detailed grid-cell level coverage for GLOBIOM includes more than 10,000 spatial
units worldwide. The model represents 18 crops globally using FAOSTAT as the primary
database for crop statistics. Crop modeling includes differentiation in management systems and
multi-cropping.

GLOBIOM also features highly detailed livestock representation, based on FAOSTAT
data. The model includes 7 animals, which can be raised on differentiated production systems.
For ruminants there are 8 production system possibilities, including grazing systems in different
climatic locations such as arid and humid, mixed crop-livestock systems, and others. Pigs and
poultry are classified under either small holder or industrial systems. Based on the production
system, animal species, and region, GLOBIOM differentiates diets, yields, and GHG emissions.
For instance, dairy and meat herds are modeled separately and their diets are differentiated.
Poultry in industrial systems is split into laying hens and broilers, again with different dietary
needs.

Apart from monogastrics, livestock production is modeled spatially in GLOBIOM's
gridded cell structure. At the cell level, animal yields for bovine and small ruminants are

212IIASA. GLOBIOM Online Documentation. https://iiasa.github.io/GLOBIOM/introduction.html

127

-------
estimated using the GLOBIOM module, RUMINANT. RUMINANT calculates a production
yield that matches plausible feed rations and checks this against regional-level data of livestock
production. Feed for animals is also differentiated in the RUMINANT model and can be
composed of feed crops, grass, stover, and other feed. Monogastric productivities are calculated
based on FAOSTAT and assumptions of potential productivities of smallholder and industrial
systems. Livestock production is allowed to intensify or extensify thereby altering the amount of
feed or grass consumed.213 Since for ruminants this is modeled spatially, any changes in
grassland consumed due to changes in production systems, animal type, yield, and GHGs is
captured in the spatially-relevant areas. Each final livestock product is considered a
differentiated good with its own specific market (apart from bovine and small ruminant milk).

Forestry in GLOBIOM is captured through the G4M module214 and includes detailed
representation of the sector and its supply chain and a differentiation between managed and
unmanaged forest areas. GLOBIOM includes bilateral trade for agricultural and wood products.
These products are assumed to be homogenous and traded based on least expensive production
costs though transportation costs and tariffs are also included.

The model also includes a bioenergy sector with first and second generation biofuels and
biomass power plants. Perennial crops and short-rotation coppice are included as inputs to the
bioenergy sector. In 2014, GLOBIOM added biofuel co-products including distillers grains,
oilseed meals, and sugar beet fibers. These co-products can be traded either in their processed or
whole form. Co-products that can be used for livestock feed are incorporated into the livestock
RUMINANT module and can substitute other forms of feed depending on protein and
metabolizable energy content.215

There are nine land cover types in GLOBIOM, and 6 of these are modeled dynamically:
cropland, grassland, short rotation plantations, managed forests, unmanaged forests, and other
natural vegetation land. The final three land cover categories are represented in the model but
kept constant, they include other agricultural land, wetlands, and not relevant (ice, water bodies
etc). Greenhouse gas emission coverage includes 12 sources of emissions that cover crop
cultivation, livestock, above and below-ground biomass, soil-organic carbon, and peatland.
Although GLOBIOM does not track terrestrial carbon stocks dynamically, carbon fluxes from
land use change are calculated with equations, following IPCC guidelines, that estimate changes
over time and allocate the average annual emissions to the time period in which the land use
change occurs.

Land use in GLOBIOM allows for both intensification and extensification. When land is
converted, this is endogenously determined in the model based on conversion costs, and the
profitability of primary, co-products, and final products. Costs increase as the area converted

213	Intensifying involves increasing livestock output without expanding the area of pasture land by grazing more
livestock per area of land, increasing feed relative to grazing, or using feedlots. Extensifying is the opposite - it
involves expanding pasture area in order to increase livestock production.

214	International Institute for Applied Systems Analysis, "Global Forest Model (G4M)https://iiasa.ac.at/models-
and-data/global-forest-model (accessed May 12, 2022).

215	Valin, Hugo, et al., September 17, 2014, "Improvements to GLOBIOM for Modelling of Biofuels Indirect Land
Use Change," http://www.alobioni-iluc.eu/wp-content/uploads/2014/12/GLUBIUM All improvements Septl4.pdf
pg. 38.

128

-------
expands. Additionally, there are biophysical land suitability and production potential restrictions.
Land use change is determined at the grid cell level, aka a one-by-one hectare. There is a land
transition matrix that sets the options for land conversion for each cell and is based on land
conversion patterns specific to that region and conversion costs depending on the type of land
converted.216

In policy settings, GLOBIOM is used for both modeling the European Union's biofuel
mandates and for estimating induced land use change impacts of biofuels for the International
Civil Aviation Organization's Carbon Offsetting and Reduction Scheme for Civil Aviation
(CORSIA). In research contexts, the model has regularly participated in AgMIP, an agricultural
model intercomparison and improvement project.217 One result of this project was an article
published in the journal, Nature, on the key determinants of global land use projections.218
GCAM, discussed in Section 4.2.2.3, was also part of the AgMIP study. GLOBIOM has been
used to assess other topics in the academic literature, publishing work on topics such as reducing
greenhouse gas emissions from the agricultural sector, food security, and climate mitigation of
livestock system transitions.

4.2.2.3 The Global Change Analysis Model (GCAM)

The Global Change Analysis Model (GCAM) is a partial equilibrium, integrated
assessment modeling framework which explores human and earth dynamics. The model includes
representation of energy, economy, land, water, and physical earth systems and interactions
between these systems within a fully integrated computational system. The model includes all
human systems and economic sectors which produce or consume energy, or which emit GHGs.
GCAM operates as a recursive dynamic framework, generally in 5-year time steps. In practice,
the model is often run from a base year in the recent past through the years 2050 or 2100.
However, time step and scenario length are flexible input assumptions to GCAM, and the
framework can support scenario analysis across a wide range of time scales. By default, the
model base year is currently 2015, but other historical base periods may be specified. For each
modeled time period, GCAM iterates until it finds a vector of prices that clears all markets and
satisfies all consistency conditions. The model is designed to explore different "what-if"
scenarios, assessing the implications of different futures on a wide range of outcomes, such as
energy supplies and demands, land allocation, or commodity prices.

216	IIASA, "Spatial Resolution and Land Use Representation,"

https://iiasa.gUhu b.io/GLUBIUM/documentation.html#spatial-resolution-and-land-use-representation. Accessed
May 19, 2022.

217	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 or Le, 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.

218	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.

129

-------
The core GCAM is developed and maintained at the Joint Global Change Research
Institute, a partnership between Pacific Northwest National Lab (PNNL) and the University of
Maryland (UMD) in College Park, Maryland. PNNL is the primary steward of the model, though
members of a larger GCAM Community also contribute to development of the framework.219
GCAM was originally developed in the early 1980s to assess the magnitude of GHG emissions
from fossil fuel C02 through the mid-21st Century. Over time, the model has expanded in scope
to serve a wide set of scientific modeling applications. The model has now been in continuous
development for over 40 years and has been applied in several studies and model inter-
comparison activities, including the IPCC's Representative Concentration Pathways220 and
Shared Socioeconomic Pathways.221 GCAM is an open-source community model that can be
downloaded from a public repository.222 The model documentation is also publicly available223
and includes a partial list of GCAM publications.224

Economic systems in GCAM are divided into sectors and, within each sector, specific
technologies. Figure 4.2.2.4-1 provides an overview of the sectors represented in GCAM, along
with the inputs and outputs of the model. Each sector of GCAM is structured with a multi-level
nesting approach that allows competition between different nodes at each level, and any number
of levels. This nested competition follows a discrete logit225 or modified logit model226,
depending on the object. The market share of each discrete technology is determined by a) a
share-weight parameter that reflects the specific preferences for a particular choice, b) the cost,
which includes fuel and non-fuel costs, and c) an exogenous logit exponent that determines the
price responsiveness of the competition. In most cases the share-weights are derived from base-
year calibration when market shares are known. Technologies that are introduced in future time
periods are assigned exogenous share-weights in each model time period. The market shares are
therefore influenced by a number of endogenous and exogenous parameters, including fuel and
non-fuel costs, efficiency or input-output coefficients, share-weights, and logit exponents. These
parameters are documented and can be consulted in online repository.227

219	For more information, see https://gcinis.pnnl.gov/community.

220	Thomson AM, Calvin KV, Smith SJ, Kyle GP, Volke A, Patel P, et al. RCP4. 5: a pathway for stabilization of
radiative forcing by 2100. Clim Change 2011:109:77.

221	Calvin K, Bond-Lamberty B, Clarke L, Edmonds J, Eom J, Hartin C, et al. The SSP4: A world of deepening
inequality. Glob Environ Change 2017:42:284-96.

222	See https://github.com/IGCRI/gcam-core.

223	See http://jgcri.github.io/gcam-doc/index.html

224	See more specifically http://jgcri.github.io/gcam-doc/references.html

225	McFadden D. Conditional logit analysis of qualitative choice behavior 1973.

226	Clarke JF, Edmonds JA. Modelling energy technologies in a competitive market. Energy Econ 1993:15:123-9.

227	See Calvin et al. 2019. GCAM v5.1: Representing the linkages between energy, water, land, climate, and
economic systems. Geoscientific Model Development 12, 1-22. See also the online documentation
(https://github.com/IGCRI/gcam-doc/blob/gh-pages/ssp.md) for the specific quantification of the inputs and
parameters to the model.

130

-------
Figure 4.2.2.4-1: GCAM diagram of model inputs, sectors, and outputs228

INPUTS	GCAM

OUTPUTS

GCAM includes detailed representations of the energy sector, inclusive of liquid biofuels,
and the agriculture and land sectors. The energy sector module in GCAM consists of depletable
and renewable resources229, energy transformation and distribution sectors (electricity, refining,
gas processing, hydrogen production, and district services), and final energy demand sectors
(buildings, industry, and transportation).230 For transportation biofuels specifically (referred to in
the GCAM documentation as "biomass liquids"), by default the model includes a total of 11
biofuel production technologies. These include four "first generation" technologies, representing
ethanols and biodiesels produced from agricultural commodity crops, and seven "second
generation" technologies representing fuels produced from a variety of feedstocks, including
energy crops and residues. By default, the technology assumptions for second generation
represent the inputs and outputs of cellulosic ethanol and Fischer-Tropsch fuels. However, the
input assumptions for these technologies can be modified to represent other fuel production

228	See http://igcri.github.io/gcam-doc/index.htnil.

229	Depletable resources are based on graded supply curves for coal, oil, gas and uranium. Renewable resources
include annual flows of wind, solar, geothermal, hydropower, and biomass.

230	More detailed information on the GCAM energy system can be found in online documentation, see
http://igcri.github.io/gcam-doc/index.html. and also in previous studies (see Clarke L, Eom J, Marten EH, Horowitz
R, Kyle P, Link R, et al. Effects of long-term climate change on global building energy expenditures. Energy Econ
2018:72:667-77: Muratori M, Ledna C, Mcjeon H, Kyle P, Patel P, Kim SH, et al. Cost of power or power of Cost:
A US modeling perspective. Renew Sustain Energy Rev 2017:77:861-74.)

131

-------
pathways. Secondary outputs such as dried distillers grains with solubles (DDGS) and electricity
produced from lignin can be considered, as can the potential for carbon capture and storage.
Further description of these technological representations is available in the online GCAM
documentation.231

The agriculture and land use module differentiates 384 land use regions globally,
generated as the intersection of 32 socioeconomic regions with 235 water basins (see Figure
4.2.2.2-2). Within each land use region, up to 25 land use types compete for land share based on
the relative profitability of each use, using a nested land allocator tree structure.232 This
competition follows a similar profit-based logit structure, so economic land use decisions are
based on the relative profitability of using land for competing purposes. Profitability of lands that
are not in commercial production in the historical calibration period are inferred from the
profitability of proximate lands used for agriculture and forestry. Land use types include
exogenous land types (tundra, desert, urban), commercial and non-commercial pasture and forest
lands, grasslands and shrublands, and a detailed set of agricultural crop commodities, including
bioenergy crops, classified by irrigation type and fertilizer use.233 Terrestrial carbon stocks and
flows are modeled for each land type in each water basin.234 The agricultural sector of the model
primarily relies on input data from the UN Food and Agriculture Organization (FAO) historical
data sets, and includes all crops for which FAO reports area and production data for the model
base year of 20 1 5.235 Major global commodity crops, such as corn, rice, soybeans and wheat are
modeled individually, while all other crops are modeled as a series of thematic aggregations.

231	See http ://j gcr L githu bio/gcam- doc/su pply energy .html.

232	See Wise M, Calvin K, Kyle P, Luckow P, Edmonds J. Economic and physical modeling of land use in GCAM
3.0 and an application to agricultural productivity, land, and terrestrial carbon. Clim Change Econ 2014;5:1450003,
and Zhao X, Calvin KV, Wise MA. The critical role of conversion cost and comparative advantage in modeling
agricultural land use change. Clim Change Econ 2020; 11.

233	A complete description of the land use module can be found in the online documentation (see
http://igcrLgithub.io/gcani-doc/toc.html) and in Kyle GP, Luckow P, Calvin KV, Emanuel WR, Nathan M, Zhou Y.
GCAM 3.0 agriculture and land use: data sources and methods. Pacific Northwest National Lab.(PNNL), Richland,
WA (United States): 2011.

234	Input assumptions related to terrestrial carbon and land transitions are documented at http://jgcrL aithu b. io/gcam-

d o c/inp u ts 1 a nd. h I ml.

235	See http://jgcri.github.io/gcam-doc/inputs land.html for further data on land inputs to the model.

132

-------
igure 4.2.2.4-2: GCAM Regional Mapping236

Energy/Socioeconomics: 32 regions

Land: 384 regions

y* *,

Water. 235 regions

"X

s' *4;

Climate: 1 global region

In addition to the core GCAM described in this section, there exist several other
subversions and downscaling tools which can be used to examine regions and systems at a finer
grain of resolution. These include, among others, GCAM-USA237, which models each US state
as an individual region, Tethys?:'"s. which allows for the downscaling of modeled GCAM water
impacts, and Demeter239, which allows for the downscaling of modeled land allocation impacts.
Numerous additional tools are in various stages of development at JGCRI and other research
groups which participate in the GCAM Community.240 One of these, GCAM-T, was used in a
recent study of corn ethanol impacts by Plevin et al. The results of that study are discussed in
greater detail later in this chapter.241

4.2.2.4 The Global Trade Project (GTAP) Model

The GTAP model is a multisector, multiregion, general equilibrium model developed at
the Center for Global Trade Analysis, Purdue University (Hertel, 1997 Ed.). It is widely used for
analyzing international trade, agriculture, and environmental policy issues. It offers a framework
to address the complex interactions of biofuels and other sectors of an economy.

Different applications of the GTAP model have resulted in establishing substitution
among energy types,242 and explicit global competition among different land categories
classified at the agro-ecological zone (AEZ) level.243 Based on these applications, a version of

236	See http://jgcri. github.io/gcam- doc/overview.htm 1

237	See http://igcri.github.io/gcam-doc/gcam-usa.html

238	https://github.com/TGCRI/tethvs,

239	https://github.com/TGCRI/demeter

240	For more information, see httus://gcims.pnnl.gov/cominunitv.

241	Plevin, R. J.y et al. (2022). "Choices in land representation materially affect modeled biofuel carbon intensity
estimates." Journal of Cleaner Production: 131477.

242	Burniaux, Jean Marc, andTruong P. Truong. "GTAP-E: an energy-environmental version of the GTAP model."
GTAP Technical Papers (2002): 18

Lee, Huey-Lin. "The GTAP Land Use Data Base and the GTAPE-AEZ Model: incorporating agro-ecologically
zoned land use data and land-based greenhouse gases emissions into the GTAP Framework." (2005).

133

-------
the model called GTAP-BIO was developed to assess the economy-wide impact of first-
generation biofuels production,244 and later developments add the capability to model dedicated

?45

energy crops.

Conceptually, the standard GTAP model assumes perfect competition in all markets with
price adjustments to ensure that all markets clear simultaneously. The regional household
collects all the income in its region and spends it over three expenditure types: private household
(consumer), government, and savings, over a Cobb-Douglas utility function. A representative
firm maximizes profits in nested CES functions in a perfectly competitive market for each
industry/sector in each region and pays income to the regional household for using the
endowment commodities (i.e., land, labor, capital, and natural resources). Consequently, firms
sell the final goods produced by combining the endowments with the intermediates to the private
households and the government, and the investment goods to the regional household. In an open
economy, firms also export the tradable commodities and import the intermediate inputs from the
rest of the world. The model follows Armington assumptions to account for product
heterogeneity for outputs produced in different regions.

Figure 4.2.2.2-1: GTAP Standard Model Analytical Framework246

The land endowment in the GTAP model is imperfectly mobile, whereas labor and
capital are perfectly mobile within a region but imperfectly mobile across regions. Government
spending is modeled using a Cobb-Douglas subutility function, which maintains constant

244	Birur D, Hertel T, Tyner WE. Impact of biofuel production on world agricultural markets: a computable general
equilibrium analysis. GTAP Working Paper No. 53. Purdue University; 2008; Hertel TW, Tyner WE, Birur DK. The
global impacts of biofuel mandates. Energy J. 2010;30:75-100.

245	Taheripour F, Tyner WE. Introducing first and second generation biofuels into GTAP data base version 7. In:
GTAP, editor. GTAP Research Memorandum No. 21. Purdue University; 2011.

246	Taheripour, F. (2022). "GTAP-BIO Model and Data Base: Main Components and Improvements." Presentation
for EPA Biofuels Modeling Workshop. March 1, 2022. Slide 7.

134

-------
expenditure shares across all the sectors. The private household consumption is modeled by
adopting a nonhomothetic Constant Difference of Elasticity implicit expenditure function, which
allows for differences in income elasticities across commodities. Taxes (and subsidies) go as net
tax revenues (subsidy expenditures) to the regional household from the private households, the
government, and the firms.

The rest of the world gets revenues by exporting to private households, firms, and the
government. These revenues are spent on export taxes and import tariffs, which eventually go to
the regional household. A closure, defining the set of endogenous and exogenous variables in the
model, is a typical GE closure in the GTAP model, which allows for equilibrium in all the
markets, all firms earn zero profits, the regional household is on its budget constraint, global
investment equals global savings, and sum of global exports and imports is zero. The global
trade balance condition determines the world price of a given commodity. The Cobb-Douglas
utility function of the regional household allows for maintaining constant budget shares.

The standard GTAP database does not disaggregate biofuels from other energy products.
The biofuel linkages in the GTAP-BIO model are introduced into the version of the GTAP-E
model to capture the implications of biofuels mandates on global energy and agricultural
markets. The substitution of biofuels is represented by intermediate demand substitution as well
as household substitution, by modifying the production and consumption structures, respectively.
For analyzing the LUCs, GTAP-BIO features the GTAP 9 land use and land cover database
developed by Baldos (2017), which aggregates 108 region-Agro-Ecological Zones (AEZs) into
18 global AEZs. The AEZs characterize the biophysical growing conditions and land use for
crops and forestry.

135

-------
Figure 4.2.2.2-2: Comparison of GTAP LULC v.6 and v.9 AEZs24i

One of the key components of the GTAP model is the GTAP database, which is
constructed based on national input-output tables, bilateral trade, protection, and macroeconomic
data to give the consistent representation of the global economy in a given base year. The GTAP-
BIO model was constructed based on the GTAP database version 6 database pertaining to the
base year 2001 and has since been updated with GTAP database version 9, which represents the
world economy in 2011.248 Based on this database, the GTAP I3IO model was enhanced with
three types of biofuels, corn-ethanol, oilseed-based biodiesel, and sugar-based ethanol, along
with the coproducts, DDGS and oil meal.

GTAP-BIO is a static model that simulates scenarios for one time period. GTAP BIO can
currently run simulations using the 2011 GTAP database. Given that GTAP databases are now
available for 2014 and 2017, GTAP-BIO is likely to be able to simulate these time periods in the
near future. Researchers at Purdue have the ability to manually project a GTAP database forward
in time based on macro-economic projections in order to simulate future time periods.249

GTAP-BIO has been updated multiple times to add features that are relevant for biofuel
GHG modeling. Tyner et al. (2010) included marginal lands and productivity estimates for the
potential new cropland based on a biophysical model. Taheripour et al. (2012) used a biophysical
model (TEM) and estimated a set of extensification parameters which represent productivity of

247	Uris, B. L. (2017) Development of GTAP 9 Land Use and Land Cover Data Base for years 2004, 2007 and 2011.
GTAP Research Memorandum No. 30

248	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. This study includes a summary of
GTAP-BIO developments over time.

249	Dhoubhadel, S., Taheripour, F. and Stockton, M. (2016) Livestock Demand, Global Land Use Changes, and
Induced Greenhouse Gas Emissions. Journal of Environmental Protection, 7, 985-995. doi: 10.4236/jep.2016.77087.
https://file.scirp.org/Html/2-6702993 67110.htm.

136

-------
new cropland versus the existing land by AEZ region.250 Taheripour and Tyner (2013) used a
tuning process to differentiate land transformation elasticities by region based on FAO data.251
Taheripour and Tyner (2013b) modified the land supply tree putting cropland pasture and
dedicated energy crops (e.g., switchgrass) in one nest and all other crops in another nest, "to
make greater use of cropland pasture (a representative for marginal land) to produce dedicated
energy crops."252 Taheripour et al. (2016) altered the land use module of GTAP-BIO to include
cropland intensification due to multiple cropping or returning idled cropland production, defined
a new set of regional intensification parameters and determined, and defined regional yield
responses to price based on analysis of regional changes in crop yields.253 Taheripour et al.
(2017) brought all of these modifications into one version of GTAP-BIO using the GTAP
database representing 2011.254

GTAP-BIO has been used in biofuel GHG analysis to estimate the areas and types of land
use change by region. Given that GTAP-BIO does not endogenously estimate land use change
GHG emissions, they are estimated using either the AEZ-EF tool or the CCLUB module of
GREET, which produce significantly different estimates. The resulting land use change GHG
emissions have been added to direct emissions estimates from LCI models such as GREET.
Given that GTAP-BIO models all sectors of the economy it may be possible to conduct a full
consequential lifecycle GHG analysis with GTAP-BIO, but such an analysis has yet to be
published.

Our focus in this section is on the GTAP-BIO model as that has been the most
extensively applied GTAP model for biofuel analysis, but there are other versions of GTAP that
have been or could potentially be used for biofuel modeling. GDyn-BIO is a recursive-dynamic
version of GTAP that has been used to model U.S. corn ethanol land use change GHG
impacts.255 GTAP-DEPS is another recursive-dynamic version of GTAP that has been used to
simulate corn ethanol effects, though it does not report GHG emissions.256 ENVISAGE is
another dynamic version of GTAP complemented by an emissions and climate module that links
changes in temperature to impacts on economic variables such as agricultural yields. To our
knowledge ENVISAGE has not been used for biofuel GHG analysis.257 Further exploring and

250	Taheripour, F., et al. (2012). "Biofuels, cropland expansion, and the extensive margin." Energy, Sustainability
and Society 2(1): 25.

251	Taheripour, F. and W. E. Tyner (2013). "Biofuels and land use change: Applying recent evidence to model
estimates." Applied Sciences 3(1): 14-38.

252	Taheripour, F. and W. E. Tyner (2013). "Induced Land Use Emissions due to First and Second Generation
Biofuels and Uncertainty in Land Use Emission Factors." Economics Research International 2013: 12.

253	Taheripour, F., et al. (2016). An Exploration of Agricultural Land Use Change at Intensive and Extensive
Margins. Bioenergy and Land Use Change: 19-37.

254	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.

255	Golub, A. A., et al. (2017). Global Land Use Impacts of U.S. Ethanol: Revised Analysis Using GDyn-BIO
Framework. Handbook of Bioenergy Economics and Policy: Volume II: Modeling Land Use and Greenhouse Gas
Implications. M. Khanna and D. Zilberman. New York, NY, Springer New York: 183-212.

256	Oladosu, Gbadebo, and Keith Kline. "A dynamic simulation of the ILUC effects of biofuel use in the USA."
Energy policy 61 (2013): 1127-1139.

257	Van der Mensbrugghe, Dominique. "The environmental impact and sustainability applied general equilibrium
(ENVISAGE) model." The World Bank, January (2008): 334934-1193838209522.

137

-------
comparing the capabilities of these other GTAP models for biofuel analysis is a potential area for
future research.

4.2.2.5 The ADAGE Model

The Applied Dynamic Analysis of the Global Economy (ADAGE) model is a multi-
region, multi-sector computable general equilibrium (CGE) model developed and maintained by
RTI International. The original ADAGE model was a forward-looking model.258 It was
originally developed to examine impacts of climate change mitigation policies and was used to
analyze economy-wide impacts of the Waxman-Markey climate change bill and the American
Clean Energy and Security Act of 2009. More recently, the ADAGE model has been developed
to have additional sectoral detail, particularly in agriculture, bioenergy, and transportation.259
This version of the ADAGE model (hereinafter referred to as "ADAGE" or "the ADAGE
model") is global, rather than national, and is recursive-dynamic, which means that decisions
about production, consumption, savings, and investment are based on previous and current
economic conditions. There are plans to make ADAGE publicly available.

ADAGE represents the entire economy, including private and public consumption,
production, trade, and investment, and follows the classical Arrow-Debreu general equilibrium
framework.260 The model uses nested constant elasticity of substitution (CES) production
functions. As illustrated in Figure 4.2.2.5-1, ADAGE includes representative households and
firms, and economic flows among households, firms, and government are considered. Bilateral
trade is represented using an Armington approach.261 Dynamics in ADAGE are represented by 1)
growth in the available effective labor supply from population growth and changes in labor
productivity; 2) capital accumulation through savings and investment; 3) changes in stocks of
natural resources; and 4) technological change from improvements in manufacturing, energy
efficiency and land productivity, and advanced technologies that become cost competitive.

258	Ross, M. 2009. Documentation of the Applied Dynamic Analysis of the Global Economy (ADAGE) Model.
Working paper 09 01. Research Triangle Park, NC: RTI International.

259	Cai Y., Beach R., Woollacott J., Daenzer K., 2022. Documentation of the Applied Dynamic Analysis of the
Global Economy (ADAGE) model. Technical Report.

260	Arrow, K.J., and G. Debreu. 1954. Existence of an equilibrium for a competitive economy. Econometrica 22:265
290.

261	Armington, P. S. (1969). A Theory of Demand for Products Distinguished by Place of Production. Staff Papers -
International Monetary Fund, 16(1), 159-178.

138

-------
Figure 4.2.2.5-1: Representation of Economic Flows in a CGE model262

Region A	Region B



Primary Factors

Income

Taxes



17

Households Government
#

Social	Govt.

Transfers Purchases

Firms

Itid

¦A.

Expenditures	/ J

14^

Trade
flows

I

I I

Goods & Services

Region C

n



f_J\

I I li



(rid

in

ADAGE includes energy, industry, food, agriculture, and transportation sectors. It runs in
5 year intervals from 2010 through 2050, and includes 8 global regions (Africa, Brazil, China,
EU 27, United States, Rest of Asia, Rest of South America, and Rest of World; Figure 4.2.2.5-2).
ADAGE is built off the GTAP v7.1 database,263 with additional data from other sources such as
the International Energy Agency, U.S. Energy Information Administration, and United Nations
Food and Agriculture Organization. Many CGE models only track inputs and outputs in
monetary units, but ADAGE also tracks physical units (such as energy units of fuel consumption
and area of land).

262	Cai Y., Beach R., Woollacott J., Daenzer K., 2022. Documentation of the Applied Dynamic Analysis of the
Global Economy (ADAGE) model. Technical Report.

263	Narayanan, G. B., and T. L. Walmsley (Eds.). 2008. Global Trade, Assistance, and Production: The GTAP 7
Data Base. West Lafayette, IN: Center for Global Trade Analysis, Purdue University.
http://www.gtap.agecon.purdue.edu/databases/v7/v7 doco.asp

139

-------
Figure 4.2.2.5-2: ADAGE Regional Mapping

ADAGE includes several agricultural commodities: wheat, corn, soybean, sugarcane,
sugar beet, rest of cereal grains, rest of oilseeds, and rest of crops, in addition to one livestock
category and one forestry category. The agricultural sector in the underlying GTAP v7.1
database is more aggregated, so creating these commodities in ADAGE required disaggregation
using information on trade shares, consumption shares, cost shares, and own use shares.264 This
disaggregation was done with software called SplilConr(i > and data from the United Nations
Food and Agricultural Organization FAOSTAT database and the United Nations Comtrade
Database.266'267 The "cereal grains" sector in GTAP v7.1 was split into corn and rest of cereal
grains, the oil seeds sector was split into soybean and rest of oilseeds, and the combined
sugarcane and sugar beet sector was split into sugarcane and sugar beet.

Agricultural sector details in ADAGE enable it to model several kinds of biofuels.
ADAGE includes 8 types of first generation biofuels (corn ethanol, wheat ethanol, sugarcane
ethanol, sugar beet ethanol, soy biodiesel, rape-mustard biodiesel, palm kernel biodiesel, and
corn oil biodiesel) and 5 types of advanced biofuels (ethanol from switchgrass, miscanthus,
agricultural residue, forest residue, and forest pulpwood). These biofuels are not included in the
GTAP 7.1 database and were split from GTAP v7.1 sectors using the SplitCom software and
secondary data from USDA's Economic Research Service, DOE's Energy Information

264	Beach, R.I [.. D.K. Birur, L.M. Davis, and M.T. Ross. 2011. A dynamic general equilibrium analysis of U.S.
biofuels production. AAEA & NAREA Joint Annual Meeting, Pittsburgh, PA.
https://ageconsearch.umn.edU/bitstream/103965/2/ADAGE-Biofuels AAEA Conference Paper.pdf

265	Horridge, M., J. Madden, and G. Wittwer. 2005. The impact of the 2002-2003 drought on Australia. Journal of
Policy Modeling 27(3):285-308.

268 Food and Agriculture Organization of the United Nations. 2012. FAOSTAT Database. Rome, Italy: FAO.
http://www.fao.Org/faostat/en/#data

United Nations. 2012. UN Comtrade Database, http://comtrade.un.org

140

-------
Administration, and the United Nations Comtrade database.268,269,270 Corn ethanol and wheat
ethanol were split from the "food products sector" in GTAP v7.1, which receives inputs from
corn and wheat. Sugarcane ethanol and sugar beet ethanol were split from the chemicals sector.
Biodiesel from soybean, rapeseed, and palm oil were split from the vegetable oils and fats sector.
Distillers grains with solubles (DGS) and corn oil biodiesel are coproducts of corn ethanol
production. An oil meal byproduct was split from the vegetable oil sector in GTAP v7.1.

Because ADAGE does not explicitly represent rapeseed and palm oil production, the input shares
of "rest of oilseeds" is based on region-specific palm oil and rapeseed biodiesel yields (gallon of
biodiesel per ton of feedstock). Advanced biofuels were not included in the 2010 base year in
ADAGE but are allowed to enter the market in future years.

The energy sectors of the ADAGE model include coal, natural gas, crude oil, and refined
oil, and several categories of electricity generation technologies (conventional coal, conventional
natural gas, conventional oil, combined-cycle natural gas, nuclear, hydropower, geothermal,
wind, solar, and biomass). The supply of fossil fuels is limited by the availability of natural
resources, which is represented as a fixed factor in the model. Crude oil is used as an input for
refined oil and enters the production function in a fixed proportion. Electricity generation
technologies are combined into a single electricity output.

The transportation sector in ADAGE has been developed to include light duty vehicles,
freight trucks, buses, marine, aviation, freight rail and passenger rail. Biofuels can be consumed
in on-road transportation (light duty vehicles, buses, and trucks). Alternative fuel options
(hybrid, battery electric, fuel cell, and natural gas) are available for on-road vehicles. The GTAP
v7.1 database includes three types of transportation (air, water, and rest of transportation) and
was disaggregated using data from several sources.271

ADAGE includes six land types (cropland, pasture, managed forest, natural forest,
natural grassland, and other land272). Land use change is represented using a constant elasticity
of supply. Each land type has its own endowment, land rent, and usage. Willingness to convert
land is represented by an elasticity, and the conversion cost is equal to the difference in land rent
between land types. ADAGE models land in physical as well as monetary quantities. Emissions
from land use change are based on the differences in carbon stocks (vegetative and soil carbon)
between the land types, and emission factors (one for vegetative carbon, and one for soil carbon)
that represent the fraction of the change in carbon stock that would occur over 20 years after land
conversion. Land use change emissions and sequestration are all reported in the model year in

268	U.S. Department of Agriculture (USDA), Economic Research Service (ERS). 2012. U.S. Bioenergy statistics.
Washington, DC: U.S. Department of Agriculture, https://www.ers.usda.gov/data-products/us-bioenergy-statistics/

269	U.S. Department of Energy, Energy Information Administration (EIA). 2012. Petroleum & other liquids.
Washington, DC: U.S. Department of Energy.

https://www.eia.gov/dnav/pet/pet move impcus a2 nus epooxe imO mbbl a.htm

270	United Nations. 2012. UN Comtrade Database, http://comtrade.un.org/

271	Data sources include GCAM 4.2, the Bureau of Economic Analysis, the Bureau of Transportation Statistics, the
International Energy Agency, and the Energy Information Administration. For more details, see Cai Y., Beach R.,
Woollacott J., Daenzer K., 2022. Documentation of the Applied Dynamic Analysis of the Global Economy (ADAGE)
model. Technical Report.

272	"Other land" includes bare ground, wetlands, mangroves, salt marsh, glaciers, and lakes, and is assumed to be
constant over time.

141

-------
which the land use change occurs. Vegetative and soil carbon stocks are based on data from
GCAM 3.2, which were aggregated to ADAGE regions using weighted land area.

ADAGE includes six types of greenhouse gases: carbon dioxide (CO2), methane (CH4),
nitrous oxide (N2O), perfluorocarbons (PFCs), hydrofluorocarbons (HFCs), and sulfur
hexafluoride (SFe). CO2 emissions from fossil fuel combustion are based on emissions factors
(kg CCh/mmBtu) for coal, gas, and oil. The emission factors are differentiated by region and
based on data from EIA's International Energy Statistics. CO2 emission factors from sources
other than fossil fuel combustion and land use change are based on data from the Emissions
Database for Global Atmospheric Research (EDGAR) version 4.2.273 Non-C02 emission factors
are based on data from EPA.274

CGE models typically do not have the sectoral detail available in other types of models.
However, because CGE models capture the entire economy, they can be useful for determining
impacts of environmental policies across sectors and on GDP. In one study, the ADAGE model
was used to analyze projected impacts of the RFS on land use, crop production, crop prices,
fossil energy use, GHG emissions, and GDP.275 ADAGE has also been used to study the impact
of oil prices on biofuel expansion.276 In model comparison studies, ADAGE was used to analyze
the GHG abatement potential in Latin America,277 and the impacts of climate policy and
agriculture, forestry, and land use emissions.278

4.2.2.6 Other Models and Approaches

Other models, besides the ones discussed above, have also been used to estimate land use
change and GHG emissions associated with biofuels. We briefly describe some of these models
here and focus our discussion on models that were used to produce at least one of the estimates
that appear in our literature review. Austin et al. (2022) conducted a literature review of studies
that examined U.S. biofuels and land use change.279 They review studies published before 2019
and, after applying a filtering process, identified 15 economic simulation modeling studies and
14 empirical studies for detailed assessment. In this section, we discuss relevant studies and

273	Joint Research Centre at European Commission. 2013. Emission Database for Global Atmospheric Research.
http://edgar.jrc. ec.europa.eu/overview.php?v=42FT2010

274	U.S. Environmental Protection Agency (EPA). 2012. Global Non-C02 GHG Emissions: 1990-2030.

Washington, DC: EPA. https://www.epa.gov/global-niitigation-non-co2-greenhouse-gases/global-non-co2-ghg-

emissions-1990-2030

275	Cai, Y., D.K. Birur, R.H. Beach, and L.M. Davis. (2013, August). Tradeoff of the U.S. Renewable Fuel Standard,
a General Equilibrium Analysis. Presented at 2013 AAEA & CAES Joint Annual Meeting, Washington, D.C.

276	Cai, Y., R.H. Beach, and Y. Zhang. (2014, March). Exploring the Implications of Oil Prices for Global Biofuels,
Food Security, and GHG Mitigation. Presented at 2014 AAEA Annual Meeting, Minneapolis, MN.

277	Clarke L., McFarland J., Octaviano C., van Ruijven B., Beach R., Daenzer K., Herreras Martinez S., Lucena
A.F.P., Kitous A., Labriet M., Loboguerrero Rodriguez A.M., Mundra A., van der Zwaan B., 2016. Long-term
abatement potential and current policy trajectories in Latin American countries. Energy Econ. 56, 513-525.
http://dx.doi.org/10.1016/i.eneco. 2016.01. Oil

278	Calvin K.V., Beach R., Gurgel A., Labriet M., Loboguerrero Rodriguez A.M., 2016. Agriculture, forestry, and
other land-use emissions in Latin America. Energy Econ. 56, 615-624.

http://dx.doi.org/10.1016/i.eneco. 2015.03.020

279	Austin, K., et al. (2022). "A review of domestic land use change attributable to US biofuel policy." Renewable
and Sustainable Energy Reviews 159: 112181.

142

-------
models reviewed by Austin et al. (2022). The purpose of this brief overview is to highlight that
there are other models, besides the five discussed above, that have some capabilities to evaluate
biofuel GHG impacts. We are providing a brief discussion of other models for informational
purposes, but we do not think they meet our statutory requirements under the CAA to evaluate
all significant direct and indirect emissions. For example, some of the models do not have global
coverage, some are not spatially explicit and cannot model land use change emissions, and some
do not include GHG emissions. Furthermore, some of these models are difficult to access given
their proprietary nature (e.g., they are not open source models).

In their literature review, Austin et al. (2022) considered 14 empirical studies that "derive
a statistical relationship between an observed land use conversion response (e.g., non-crop to
crop conversion) and a treatment (e.g., ethanol refinery location or capacity)." All of these
empirical studies were limited in spatial extent to the contiguous U.S. or a subset of U.S. states
(generally the corn belt). Only two of the 14 empirical studies estimated the effects of corn
ethanol production on land use for the contiguous U.S. and considered the effects of corn and
crop prices changes on land use. The other 12 studies were either limited to a subset of U.S.
states, did not consider prices effects, or evaluated price effects in general but not biofuel
induced effects. These limitations make these 12 studies potentially useful for informing our
understanding, but the models and methods have limited relevance for our purposes. We briefly
discuss each of the two studies that met these criteria.

Lark et al. (20 2 2)280 was widely cited by commenters on our biofuel GHG modeling
workshop. This study combines econometric analyses, land use observations, and biophysical
models to estimate the effects of increasing U.S. corn ethanol production by 5.5 billion gallons
since 2007 compared to a counterfactual scenario. Austin et al. (2022) categorized this as an
empirical study as it does not develop an economic simulation model that can be
straightforwardly applied to a wide range of scenarios. For GHG emissions attributable to corn
ethanol, Lark et al. (2022) report net changes in CO2 and N2O emissions from land use and land
management changes in the U.S. The reported CO2 emissions are the result of ecosystem carbon
changes (e.g., plowing grassland to produce corn) and the reported N2O emissions are from
increased fertilizer application. Lark et al. (2022) added their estimates of U.S. land use change
to the corn ethanol LCA estimates from the 2010 RFS2 rule, GREET and the CARB's analysis
for the CA-LCFS. The fact that this study only estimates historical U.S. land use change GHG
emissions means that we can only do a limited comparison with estimates from other models that
evaluate all lifecycle stages and/or project scenarios into the future.

Brandao (2022) uses a consequential approach to estimate the GHG emissions associated
with ramping up U.S. corn ethanol production to 15 billion gallons from 1999 to 2018. This
study estimates market-mediated effects without an economic model. Rather it looks at the
difference in corn supply over this time period and uses a rules based approach to estimate
diversion, direct land use change, and intensification effects of increasing the amount of corn
used for ethanol. It then estimates the resulting crop production and land use change emissions.
We include this study in the range of estimates for corn ethanol presented below as it provides a
full lifecycle estimate.

280 Lark, T. J., et al. (2022). "Environmental outcomes of the US Renewable Fuel Standard." Proceedings of the
National Academy of Sciences 119(9): e2101084119.

143

-------
Li et al. (2019) estimated the effects of ethanol capacity, corn prices and crop prices on
corn and cropland area at a county level.281 Unlike Lark et al. (2022), this study did not estimate
the influence of ethanol production on corn and crop prices, nor did they estimate GHG
emissions or other environmental effects. Combined with exogenous estimates of the effects of
ethanol production on corn and crop prices, it is possible to use Li et al. (2019) to estimate corn
and cropland area changes at county-scale resolution for comparison with economic modeling
estimates. Furthermore, Li et al. (2019) estimated corn and crop area elasticities in response to
ethanol capacity and prices, which may be useful as inputs to economic simulation models.

In addition to the economic simulation models discussed above (i.e., GTAP-BIO,
GLOBIOM, GCAM, ADAGE), Austin et al. (2022) identified studies that used models called
REAP, PEEL-Co, BEPAM, AEPE and an unnamed multi-market equilibrium model. All of these
models limit their spatial extent to the U.S., and thus we give them brief treatment. REAP is a
price-endogenous mathematical programming model of the U.S. agricultural sector that can
simulate changes in soil carbon.282 PEEL-Co is a stochastic PE land use change allocation model
that can estimate domestic land use change emissions associated with corn ethanol production.283
BEPAM is a PE model that simulates biofuel effects on Conservation Reserve Program and
cropland pasture acres, and currently excludes GHG emissions.284 AEPA is a PE model that
examines interactions between agricultural and energy markets and evaluates the consequences
of changes in biofuel policies.285

In addition to the models discussed in the Austin et. al. paper, another model has been
used to evaluate biofuels is the Bio-based circular carbon economy Environmentally-extended
Input-Output Model (BEIOM). BEIOM is an economy-wide environmentally extended input-
output model using economic transactions data together with emissions inventories to provide a
high-level snapshot of the U.S. economy at different points in time.286 As such it performs a
"top-down" assessment linking national-level economic transactions with emissions inventories
for specific points in time. However, BEIOM only includes emissions in the U.S., therefore we
do not discuss it in more extensive detail.

Finally, while most studies in the literature focus on estimating indirect/induced land use
change emissions, other studies estimate "direct" land use change emissions (i.e., the emissions
associated with converting the land required for biofuel feedstock production). A relatively early

281	Li, Y., et al. (2019). "Effects of Ethanol Plant Proximity and Crop Prices on Land-Use Change in the United
States." American Journal of Agricultural Economics.

282	Johansson, Robert C. Regional environment and agriculture programming model. (Technical bulletin (United
States. Dept. of Agriculture) ; no. 1916); Malcolm, Scott A., Marcel P. Aillery, and Marca Weinberg. Ethanol and a
changing agricultural landscape. No. 1477-2016-121116. 2009.

283	Elliott, J., et al. (2014). "A Spatial Modeling Framework to Evaluate Domestic Biofuel-Induced Potential Land
Use Changes and Emissions." Environmental Science & Technology 48(4): 2488-2496.

284	Chen, X. and M. Khanna (2018). "Effect of corn ethanol production on Conservation Reserve Program acres in
the US." Applied Energy 225: 124-134.

285	Taheripour, F., et al. (2022). "Economic Impacts of the U.S. Renewable Fuel Standard: An Ex-Post Evaluation."
Frontiers in Energy Research 10.

286	Lamers, P., et al. (2021). "Potential Socioeconomic and Environmental Effects of an Expanding U.S.
Bioeconomy: An Assessment of Near-Commercial Cellulosic Biofuel Pathways." Environmental Science &
Technology.

144

-------
study of direct land use change emissions (Fargione et al. 2008) estimated the "land clearing
carbon debt" associated with converting a range of natural ecosystems (rainforests, grasslands,
savannas) to food crop-based biofuel feedstock production.287 ICAO (2022) provides a
methodology for calculating direct land use change emissions for an "event where feedstocks
were sourced from land obtained through land use conversion after 1 January 2008."288 Given
difficulties with measuring direct land use change emissions in cases where the land was
converted many years ago, Searchinger et al. (2022) use an alternate approach called "carbon
opportunity cost" estimated as either: (1) "the global carbon loss from plants and soils generated
by producing each crop to date (the numerator), divided by the global production (the
denominator)", or (2) "the quantity of carbon that could be sequestered annually if the average
productive capacity of land used to produce a kilogram of each food globally were instead
devoted to regenerating forest."289 Direct land use change does not factor into the rest of our
review as it has not been used in recent LCA studies; however, it may deserve additional
consideration in the future as it can be estimated with empirical measurements instead of
counterfactual modeling.

4.2.2.7 Comparison of Model Characteristics

In this section we compare the characteristics of the five models highlighted above in
Chapter 4.2.2.1-5. While there are many models that could potentially be used for biofuel GHG
analysis (see Chapter 4.2.2.8), we focus our discussion on these five models for the reasons
discussed at the beginning of Chapter 4.2.2. We sometimes bring other models and empirical
studies into the discussion as comparison points, but we otherwise set them aside to focus on
models that are designed to evaluate hypothetical scenarios and project future effects. We
compare the models across several characteristics that are important for biofuel analysis. In the
next section (Chapter 4.2.2.8) we also compare published results from these models and discuss
the importance of input assumptions relative to model choice. We then outline the goals of the
model comparison exercise that we intend to conduct for the final rule.

Table 4.2.2.7-1 summarizes some of the key characteristics of the five models featured in
section 4.2.2. Although there are many ways to compare these models, we choose six key
characteristics that we believe will help us determine whether the modeling framework meets the
statutory requirements for evaluating biofuel lifecycle GHG emissions, including direct and
indirect emissions.290 The models that are the most comprehensive across these key
characteristics are more likely to satisfy the statutory requirements, as discussed in Section
4.2.2.1. For example, models that represent all relevant sectors, regions, time periods, GHG
emissions sources, and land categories are most likely to capture significant indirect emissions.
These six characteristics provide a good starting point for understanding the primary differences
across these frameworks. We start our discussion based on these six characteristics before
touching on other key aspects of these models for biofuel GHG analysis. While we are not ruling

287	Fargione, J., et al. (2008). "Land Clearing and the Biofuel Carbon Debt." Science 319(5867): 1235-1238.

288	ICAO (2022). CORSIA Methodology for Calculating Actual Life Cycle Emissions Values, International Civil
Aviation Organization. June 2022.

289	Searchinger, T. D., et al. (2018). "Assessing the efficiency of changes in land use for mitigating climate change."
Nature 564(7735): 249-253.

290	While not explicitly required as part of the CAA, we also note that open access to the models is an important
consideration.

145

-------
out consideration or future use of other models, based on the biofuel GHG modeling workshop
and our review of the literature, we believe the models listed in the table are the most likely to
meet our statutory requirements for evaluating lifecycle GHG emissions. In addition, the models
selected provide a broad representation of the types of models that can be used for lifecycle
analysis.

Table 4.2.2.7-1 Comparison of Key Characteristics Across Models

Characteristic

GREET

GLOBIOM

GTAP-BIO

ADAGE

GCAM

Type of Model

Lifecycle
inventory (LCI)

Partial

equilibrium

(PE)

Computable
general
equilibrium
(CGE)

Computable
general
equilibrium
(CGE)

Integrated
assessment model
(IAM)

Sectoral
Coverage

Fuel supply
chains
including
energy resource
and material
inputs

Agriculture,
forestry and
bioenergy

Economy-wide
with 57 sectors

Economy-
wide with 36
sectors

Energy

(conventional and
renewable),
agriculture,
forestry, water

Temporal
Representation

Static

Recursive
dynamic (10-
year time steps)

Comparative
static

Recursive
dynamic (5-
year time
steps)

Recursive
dynamic (5-year
time steps)

Regional
Coverage

Customizable
(typically U.S.
average)

37 economic
regions; 10,000
spatial units
(grid cell)

19 economic
regions; 18
agro-ecological
zones

8 economic
and spatial
regions

32 economic
regions; 235
spatial regions
(water basins)

GHG Emissions
Coverage

Direct supply-
chain emissions
+ indirect land
use change
from CCLUB
module

Crop

production,
livestock and
land use change

Land use
change GHGs
calculated with
CCLUB or
AEZ-EF
modules

Economy-
wide GHGs
including land
use change

Global GHGs
including land
use change

Land

Representation
(Arable land
categories
considered in
biofuel land use
change analysis)

Exogenous
(Land use
change

estimates from
GTAP-BIO and
CCLUB)

Cropland, other
agricultural
land, grassland,
commercial and
non-
commercial
forest,

wetlands, other
natural land

Cropland
(including
cropland-
pasture) ,
livestock
pasture,
"accessible"
forestry land

Cropland,
pasture,
commercial
forest, non-
commercial
forest, natural
grassland,
other land

Cropland,
commercial
pasture and
forest, non-
commercial
pasture and
forest, shrubland,
grassland,
"protected" non-
commercial land

Across the four model types there tends to be a trade-off between scope and detail. The
LCI models have the most detailed technological representations but the most limited scope. For
example, the GREET model includes detailed representations of many biofuel and energy
production processes, but includes no price-induced interactions between supply chains or
economic sectors. PE models also tend to have a high level of detail in the agricultural sector, but
limited interactions with other sectors. For example, the GLOBIOM model has a detailed
representation of crop production, livestock, and land use, but does not include economic
interactions between the agricultural and energy sectors (e.g., fertilizer prices are exogenous).
CGE models are the broadest in economic scope but tend to lack detail in their physical and

146

-------
technological representations. For example, CGE models are designed to track resources in terms
of their monetary value and require subsequent accounting methods to estimate physical
quantities. IAMs are the broadest in their representation of the interactions between human (e.g.,
economic) and Earth (e.g., biophysical) systems but tend to lack detail in their representation of
particular sectors (e.g., finance, labor) and technologies (e.g., oil refining is represented with one
or two generic technologies).

The modeling frameworks differ substantially in the scope of economic interactions that
they represent. Since the Clean Air Act requires us to consider all significant indirect impacts as
part of our lifecycle analysis, capturing a wide range of economic interactions is important for
understanding the overall GHG impacts of crop-based biofuel production. Based on economic
theory, we expect increased consumption of crop-based biofuels to have complex ripple effects
through the entire world economy. For example, as the demand for feedstocks increase, we
expect the price of these commodities to increase with consequences for food and feed markets
not only in the U.S., but around the world. These interactions are complicated by the fact that the
major crop-based biofuel feedstocks have co-products (e.g., distiller grains, soybean meal) that
are used as livestock feed. Given that producing biofuels requires material (e.g., fertilizer) and
energy (e.g., natural gas), increased biofuel production may affect these input commodity
markets as well. When biofuels displace gasoline or diesel in the U.S., this change may affect
consumer fuel prices and crude oil prices which may in turn affect other sectors of the economy.
LCI models such as GREET ignore most of these economic interactions. However, GREET
includes agricultural sector interactions to a limited extent through the exogenous addition of
land use change GHG estimates. GLOBIOM models economic interactions within and between
the agricultural (including crops and livestock) and forestry sectors. GLOBIOM also includes a
bioenergy sector with limited economic interactions other than through its consumption of
feedstocks from the agricultural and forestry sectors. GCAM models economic interactions
within and between the energy, agriculture, forestry, and water systems. The energy system in
GCAM is highly developed, including energy production from a broad range of technologies and
resources, and energy consumption in the industrial, commercial, residential, transportation,
agriculture, and forestry sectors. As CGE models, GTAP-BIO and ADAGE model interactions
across the entire economy. Thus, CGE models include economic interactions that the other
modeling frameworks take as exogenous or ignore but do so at a highly aggregated level.

Temporal representation, or the treatment of time dynamics, is another important
characteristic that differentiates the modeling frameworks. As a general matter, the ability to
model temporal dynamics is an important feature given that biofuel land use change emissions
occur over time (e.g., soil carbon levels change over multiple decades following land conversion)
and biofuel-induced effects are dependent on factors that change over time, such as crop yields
and overall demands on land to produce food, feed, and fiber. GREET does not represent time as
it is not designed to simulate temporal changes.291 GTAP-BIO is a comparative static model,
meaning it models only one time period and does not project changes over time.292 GLOBIOM,
GCAM and ADAGE are recursive dynamic models whereby production, consumption, and

291	However, as discussed above, if provided with sufficient data, GREET can estimate supply chain emissions for
different time periods

292	GTAP-BIO can model different time periods if the GTAP database is first manually projected forward (or
backward) based on assumptions.

147

-------
investment decisions are made on the basis of market conditions in each period with dependence
on previous model periods through capital and/or resource stocks.

Consistent with the statutory definition of lifecycle GHG emissions, we need to consider
significant indirect emissions, and there are many scientific studies showing that such indirect
emissions can occur in many different regions of the globe. Thus, models need to represent all of
the relevant regions in order to satisfy our needs. Furthermore, regional representation is
important due to differing regional conditions related to terrestrial carbon stocks, agricultural
yields, energy resources and other factors. PE, CGE and IAM models often distinguish between
economic regions and spatial regions. These models use algorithms to find market clearing
conditions in, and trade between, each of the economic regions. Spatial regions are often defined
separately to allocate the economic activities to physical locations. GTAP-BIO models 19
economic regions and 18 non-contiguous AEZs (see Figure 4.2.2.2-1). GLOBIOM models 37
economic regions and uses a grid-cell approach to represent 10,000 spatial units worldwide.
GCAM models 32 economic regions and 235 global water basins—the intersection of the
economic regions and water basins produces 384 spatial subregions. A spatial downscaling
model called Demeter is able to present GCAM results at higher spatial resolution
(0.05° x 0.05°).293 ADAGE models 8 economic and spatial regions. In contrast, GREET is not a
spatial or regional model, but it can be customized to represent biofuel production conditions for
particular regions or supply chains. GREET also has modules that are designed to estimate soil
carbon and land use change emissions at a regional level. The FD-CIC module allows users to
estimate feedstock production emissions at county level, and the CCLUB module estimates
indirect land use change emissions based on the spatial regions represented by GTAP-BIO.

There are major differences across the models in their coverage of GHG emissions
sources. As such, the biofuel GHG emissions estimates produced from these models are often
quite different in their scopes. As mentioned previously, GREET estimates direct GHG
emissions from a biofuel production supply chain and ignores indirect market-mediated
emissions from other sources and sectors, with the exception of indirect land use change
emissions which are added exogenously through the CCLUB module. There are versions of
GTAP models that endogenously track GHG emissions from energy and land use. It may be
possible to endogenously track GHG emission in GTAP-BIO, but to our knowledge this
approach has never been used in peer-reviewed studies on biofuel GHG impacts.294 GTAP-BIO
models the entire economy but, at least in terms of peer-reviewed studies on biofuels, changes in
emissions from other sectors of the economy are generally not reported. As discussed above,
land use change GHG emissions are estimated from GTAP-BIO's land use change area estimates
through the application of static land conversion emissions factors. GLOBIOM endogenously
calculates GHG emissions from agriculture, including crop and livestock production, forestry,
and land use change. ADAGE endogenously calculates GHG emissions from the entire
economy, including land use change, while GCAM endogenously calculates global GHG
emissions from the energy, agriculture, forestry and water systems, including from land use
changes. Of the five highlighted models, ADAGE, GCAM, and GTAP are the only models that

293	Chen, M., Vernon, C.R., Graham, N.T. et al. Global land use for 2015-2100 at 0.05° resolution under diverse
socioeconomic and climate scenarios. Sci Data 7, 320 (2020). hi:t:ps://doLorg/10.1038/s41597-020-00889-x

294	See for example: https://www.gtap.agecon.purdue.edu/models/energy/default.asp

148

-------
capture GHG emissions from market-mediated changes within the energy system, though energy
sector emissions have not historically been reported in GTAP-BIO LCA calculations.

It is important to note that although GREET, GLOBIOM, ADAGE and GCAM seem to
overlap in their coverage of GHG emissions, they estimate GHG impacts associated with very
different phenomena. For example, GREET and GLOBIOM both estimate GHG emissions from
crop production, but they do so in fundamentally different ways. GREET estimates the GHG
emissions associated with producing the crops that are directly used in the biofuel supply chain
under evaluation. In contrast, GLOBIOM estimates the GHG emissions associated with the
market-mediated marginal changes in crop production stemming from a biofuel shock (i.e., the
difference in crop production emissions from a scenario with a given amount of biofuel relative
to a scenario absent that biofuel). ADAGE and GCAM represent a further departure from the
GREET approach as they include market mediated GHG impacts from yet more economic
sectors. A notable example is the inclusion of GHG emissions from transportation fuel market
effects in ADAGE and GCAM. When these models are shocked to consume more biofuels in a
particular region, they estimate the effects of the shock on transportation fuel prices and
consumption, both in the region where the shock occurs and all other regions around the world.
Thus, instead of assuming that biofuels displace gasoline or diesel on an energy-equivalent basis,
they estimate the global market-mediated changes in gasoline and diesel consumption associated
with the biofuel shock and report the resulting GHG emissions changes.

Representation of land is an important, but often overlooked, consideration for land use
change modeling. By land representation we mean the way that land is categorized and how
much of it is assumed to be unavailable for commercial use. GREET does not explicitly
represent land. The other four models estimate interactions between cropland, pasture, and
forestry. GLOBIOM, ADAGE and GCAM also model the expansion of commercial cropland,
pasture and forestry activities into grassland and forests that are not otherwise used for
commercial production. In contrast, GTAP-BIO only allows managed lands in the U.S. and
around the world to be used for productive uses, excluding the possibility for "unmanaged" land
such as rainforests to be brought into production. As shown in Figure 4.2.2.7-1, this assumption
applies to a relatively large share of arable land and means that GTAP-BIO employs a much
different representation of land than the other models. Additionally, the share of non-commercial
land that is assumed to be protected or unavailable for protected use is also an important
assumption. For example, if the modeling assumes that policies will be implemented and
enforced to protect natural forests with high carbon stocks, this will likely reduce the land use
change GHG estimates by a significant amount.

149

-------
Figure 4.2.2.7-1: Land available for productive use in five models used to estimate biofuel-
induced land use change295

Cropland

Forest

Grassland & Pasture

01
QJ

to

t> 3

CD
JT
C

M2
£

Cropland
Foraslry
Natural Forest
Pasture
Grassland

l_;
CD
<

O LU
<

Q
ra
O

O b

O
m

Cl.
<

UJ	<	5

O	0.	<

<	a.	o

Q	m	(3

<

o

m

CL
<

o

CO

o

o b

LU	<

0	a.

<	0-

Q	UU

<

Q

CO

cn	a.

O	<
—J

o	o

There are also significant differences in the data and parameter inputs that are used within
models used for biofuel GHG analysis. There have been very few efforts to compare the many
assumptions across these models or evaluate how they influence the results. As part of the model
comparison exercise that we intend to conduct for the final rule, we will strive, time permitting,
to compare some of the key input assumptions. For now, we can make a limited set of
observations about which assumptions are likely to be important for biofuel GHG modeling and
how they compare across the models.

Assumptions related to crop yields and crop intensification are important for biofuel
GHG modeling. Global crop yield data is readily available from FAO; however, there may be
differences in how the models map this historical data to the crops and regions they represent.
Assumptions about how crop yields may change in the future are also influential and inherently
uncertain. Perhaps even more important for biofuel modeling are assumptions about how crop
yields may change in response to price changes. Plevin et al. (2015) performed a sensitivity
analysis of biophysical and economic inputs to the GTAP-BIO+AEZ-EF modeling framework,
and found the elasticity of crop yield with respect to price (YDEL) to be "by far" the most
influential parameter in terms of its effect on the estimated ILUC emissions associated with corn
ethanol, sugarcane ethanol and soybean oil biodiesel.296 Later studies confirmed that YDEL is
influential in GTAP-BIO and found that other parameters related to crop intensification, such as
the parameters that control multi-cropping (i.e., multiple harvest per year on the same land)
instead of crop expansion to meet demands, also have a significant effect.2' However, a

295	Figure 1 from Plevin, R. J., et al. (2022). "Choices in land representation materially affect modeled biofuel
carbon intensity estimates." Journal of Cleaner Production: 131477. For simplicity, shrubland and some other land
types (e.g., wetlands) are excluded. Note: GREET is not included in this chart since it does not explicitly model land
use change.

296	pieVjn; 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.

297	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

150

-------
sensitivity analysis with GCAM did not identify crop yield assumptions to be among the most
influential parameters determining corn ethanol land use change GHG emissions.298 This
suggests that input parameters that are influential in one model might not be very influential in
another model due to structural differences between them. Unfortunately, we do not have
sensitivity analyses from GLOBIOM or ADAGE that estimate the most influential parameters on
biofuel GHG estimates.

The parameters that control land competition and land transitions within the models may
also be important. A sensitivity analysis with GCAM found the parameter controlling ease of
transition between cropland, forest, and grassland to be an influential parameter. A sensitivity
analysis with GTAP-BIO also found that the assumed elasticity of transformation between
managed forest, cropland, and pasture is influential for corn ethanol LUC GHG estimates.299

Sensitivity analysis using GCAM found other assumptions to be influential when
estimating corn ethanol land use change GHG emission, including the soil carbon density of
cropland, ease of transition between crop types, the soil carbon density of grassland and the soil
carbon density of other arable land.300 Other influential assumptions identified through
sensitivity analysis with GTAP-BIO include the relative productivity of newly converted
cropland, trade elasticities (i.e., ease of substitution among products imported from other
countries) and emissions from conversion of cropland pasture.301

Sensitivity analyses have shown that other influential assumptions within GTAP-BIO
include, but are not limited to, tropical peat soil oxidation and the share of palm oil expansion on
peatland for vegetable oil based biofuel modeling, and the share of vegetable oil biofuel
feedstock that is supplied through expanded vegetable oil production versus reduced demand and
substitutions with other products.302

These aspects of vegetable oil market dynamics have led to variation in how different
models capture emission impacts of vegetable oils. Most of the mass and value of soybeans for
instance comes from the plant matter and goes towards livestock feed. Only about 20 percent of
a soybean's mass goes towards oil. This means that modeling the livestock sector and associated
data and parameters are also important considerations for modeling oilseeds. Likewise on the oil
side, the number of vegetable oil substitutes and the degree to which people can or want to
substitute between these oils for both food and fuel is another important consideration in oilseed
modeling. Furthermore, the vegetable oil market dynamics have not been static; over the last

298	Plevin, R. J., et al. (2022). "Choices in land representation materially affect modeled biofuel carbon intensity
estimates." Journal of Cleaner Production: 131477. Figure 7.

299	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. Table S9 in the Supplemental
Information.

3°° pievin; R J ; et ai (2022). "Choices in land representation materially affect modeled biofuel carbon intensity
estimates." Journal of Cleaner Production: 131477. Figure 7.

301	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. Table S9 in the Supplemental
Information.

302	ICAO (2021). CORSIA Eligible Fuels - Lifecycle Assessment Methodology. CORSIA Supporting Document.
Version 3: 155. Section 6.2

151

-------
decade the relative and absolute prices of vegetable oils have grown and changed from their
2010 values. The data and market structure that underlies how these market interactions are
captured differs from model to model.

Another influential assumption in biofuel GHG modeling is the choice of data sets for
soil carbon and biomass carbon stocks, and how these data are mapped to land categories and
regions to determine the GHG emissions from converting an acre of land from one use to
another. The soil and biomass carbon data sources used in each model are discussed in the model
descriptions above. Soil carbon data and analysis are active areas of research, and higher
resolution datasets have recently been produced using statistical methods and remote sensing
data.303 For example, the SoilGrids250m version 2.0 dataset provides soil carbon estimates for
the globe with quantified spatial uncertainty,304 and Spawn et al. (2020) developed global maps
of above and below ground biomass carbon density in the year 20 1 0.305 With few exceptions,306
these newer data sets have not yet been incorporated into published estimates of biofuel land use
change.

4.2.2.8 Review of Land Use Change GHG Estimates

Land use change GHG estimates continues to be a large source of variation between
lifecycle GHG estimates for crop-based fuels. In order to further our understanding of available
models, we reviewed studies that estimate the land use change GHG emissions associated with
corn ethanol and soybean oil biodiesel, which account for the large majority of crop-based
biofuel supply in the U.S. This review included journal articles, major reports and studies that
informed biofuel-related policies. We reviewed studies that were published after the March 2010
RFS2 rule, as that rule considered the available science at the time. In cases where there were
multiple studies that include updates to the same general model and approach, we included only
the most recent study. However, we include older estimates from the GTAP-BIO model that are
still used for the CA-LCFS or the default assumptions in GREET.

We focused our review on estimates of the average type of each fuel produced in the
United States. Many of the studies we reviewed include sensitivity analysis, where many
parameters are varied to produce a large number of estimates. In these cases, we include
representative high and low estimates. For example, when studies report a 95% confidence
interval, we use only the central estimate (usually the default, mean or median estimate) and the
estimates at the top and bottom of the confidence interval. This approach simplifies the
presentation of results relative to including every estimate in between. We intentionally do not
calculate or present any statistics (e.g., mean, median) derived from the estimates, as we do not
believe such statistics would be meaningful or appropriate based on the design and purpose of
our literature review.

303	Spawn-Lee, Seth. (2022). "Carbon: Where is it and how can we know?" Presentation for EPA Biofuel GHG
Modeling Workshop. February 28, 2022. EPA-HQ-OAR-2021-0921-0022

304	Poggio, L., de Sousa, L. M., Batjes, N. H., Heuvelink, G. B. M., Kempen, B., Ribeiro, E., and Rossiter, D.:
SoilGrids 2.0: producing soil information for the globe with quantified spatial uncertainty, SOIL, 7, 217-240, 2021.

305	Spawn, S. A., et al. (2020). "Harmonized global maps of above and belowground biomass carbon density in the
year 2010." Scientific Data 7(1): 112.

306	Lark, T. J., et al. (2022). "Environmental outcomes of the US Renewable Fuel Standard." Proceedings of the
National Academy of Sciences 119(9): e2101084119.

152

-------
Figure 4.2.2.8-1 summarizes the land use change GHG estimates from our literature
review for corn ethanol and soybean oil biodiesel.307 All of the estimates in this chart report land
use change GHG estimates as carbon dioxide-equivalent (COze) emissions per megajoule (MJ)
of fuel consumed. All C02e estimates are based on 100-year global warming potential (GWP)
from the IPCC,308 This allows us to compare all of the estimates on a gCCtee/MJ of fuel basis.
However, we stress that many of the studies in this chart do not align in terms of their time
horizon, year of analysis, or other factors. Therefore, the estimates reported in this figure give us
a sense for the range of estimates, but caution is needed when comparing the points in this figure
as the estimates are from evaluations of divergent scenarios and methodologies.

Figure 4.2.2.8-1: Land Use Change GHG Emissions Estimates by Pathway

75

a>

CM

O
o
s


(/>

E

LU

CD
X
C5

O 25

50



*

+*

•+i*

i ++eT+*

a++*+++-

+

+



*

+¦,+,
+.++'

LUC Model

• FASOM-FAPRI

GCAM-T
¦ GLOBIOM
GTAP-BIO
fx] MIRAGE
Other

Corn Starch Ethanol

Soybean Oil Biodiesel

307	Details on the sourcing of each estimate from our literature review are available in a memo to the docket for this
proposed rule titled "Notes on Literature Review of Transportation Fuel Greenhouse Gas (GHG) Lifeeycle Analysis
(LCA)."

308	The reviewed estimates use GWP values from the IPCC Second Assessment Report (SAR), Fourth Assessment
Report (AR4) or Fifth Assessment Report (AR5). We did not attempt to harmonize GWP assumptions across studies
as many studies only reported COge results and not emissions by gas.

153

-------
Figure 4.2,2.8-1 shows a relatively wide range of land use change GHG estimates,
especially for soybean oil-based biodiesel.3"9 These findings are consistent with the conclusions
of another recent literature review presented at our February 2022 workshop/'1" In order to
highlight the influence of model choice on the land use change estimates, Figure 4.2.2.8-2
arranges the data by model. Once again, caution is needed when comparing the estimates in
Figure 4.2.2.8-2, as the estimates were produced for different scenarios and contexts. With these
qualifications in mind, we can make some general observations from Figure 4.2.2.8-2 about the
effect of model choice on land use change estimates.

Figure 4.2.2.8-2: Land Use Change GHG Emissions Estimates by Model311

Com Starch Ethanol	Soybean Oil Biodiesel

Other -	• • •

MIRAGE-	•	•

• • ¦ •* • •

GTAP-BIO- • • ••«•	• •	••••	••

• •

GLOBIOM •	•	•	•	•	•

GCAM-T	•	•	•

FASOM+GTAP-BIO -
FASOM+FAPRI -
FAPRI -

25	50	75	0	25	50	75

Land Use Change GHG Emissions (gC02e/MJ)

GTAP-BIO is the model with the largest number of estimates in our literature review.
This is partly because GTAP-BIO has been used for multiple publications and purposes. For
example, GTAP-BIO is used to estimate land use change GHG emissions for the CA-LCFS and
the ICAO CORSIA programs. It is also partly due to the way that we filtered and selected
estimates for our review. For example, GLOBIOM recently produced 300 different estimates for
each pathway, but for practical reasons Figure 4.2.2.8-2 includes only the default estimate and
the 2.5% and 97.5% quantile estimates from the Monte Carlo sensitivity analysis.31' As another
example, analysis by Plevin et al (2022) using GCAM-T produced 3,000 estimates of corn

309	Some of the estimates in this figure come from studies that only estimate land use change GHG emissions or do
not conduct full lifecycle estimates for corn ethanol or soybean oil biodiesel and thus do not factor into the LC A
ranges developed below in Chapter 4.2.3. For example, the highest estimate for soybean oil from GLOBIOM does
not factor into the range of LCA estimates for soybean oil biodiesel or renewable diesel summarized in Chapter
4.2.3.12 (see Chapter 4.2.3.4 for more information).

310	Daioglou, V., et al. (2020). "Progress and barriers in understanding and preventing indirect land-use change."
Biofuels, Bioproducts and Biorefining 14(5): 924-934.

311	Details on the sourcing of each estimate from our literature review are available in a memo to the docket for this
proposed rule titled "Notes on Literature Review of Lransportation Fuel Greenhouse Gas (GHG) Lifecycle Analysis
(LCA)."

312	ICAO (2021). CORSIA Eligible Fuels - Lifecycle Assessment Methodology. CORSIA Supporting Document.
Version 3: 155. Lable 67.

154

-------
ethanol land use emissions in a recent publication, but Figure 4.2.2.8-3 includes only the mean,
5th and 95th percentile results from the Monte Carlo sensitivity analysis conducted for this
study.313 The FASOM+FAPRI estimates in this figure are from the 2010 RFS2 rule; although
these estimates are over 10 years old, the figure illustrates that they continue to be within the
range of more recent studies. The estimates from the model labeled "Other" in this figure are
from Lark et al. (2022). Lark et al. (2022) developed a new model to estimate U.S. land use
change GHG emissions, and pairs these estimates with estimates of non-U.S. land use change
emissions from other models.

Since Figure 4.2.2.8-2 only reports land use change emissions, we cannot say, based on
the estimates in this figure, which models produce the highest or lowest lifecycle GHG estimates.
Although many of these models can estimate GHG emissions from other sources (and in some
cases all sources), with few exceptions only the land use change GHG emissions have been
reported from these models. As part of our model comparison exercise for the final rule, we
intend to examine a broader range of GHG emissions estimates from these models. For example,
we intend to compare estimates from these models of the GHG emissions associated with
changes in crop production, livestock production, and energy supply and consumption. For now,
we focus our discussion on the land use change results and what factors may explain differences
between estimates.

While there is a large amount of overlap between the range of land use change GHG
estimates from each model, many of the GTAP-BIO estimates are clustered at the relatively low
end of the range of estimates. GCAM tends to have the highest land use change GHG estimates
for corn ethanol, and GLOBIOM tends to have the highest estimates for soybean oil biodiesel
(there are no modeling results from GCAM for soybean oil biodiesel available in the literature).
Without harmonizing scenarios and assumptions it isn't possible to fully explain the source of
the differences. However, based on our current understanding of how these models compare (see
previous section), we can identify some of the potential reasons.

There are three broad elements that contribute to land use change GHG estimates per
gallon of biofuel production: 1) acres of cropland expansion, 2) types of land displaced by
cropland expansion, and 3) GHG emissions per acre of land use change. Comparing these
elements for the published estimates in Figure 4.2.2.8-2 helps us better understand the underlying
differences between the studies. The next two figures compare cropland area impact estimates
across studies for corn ethanol and then soybean oil biodiesel. These studies compared a
reference scenario to a scenario with increased corn ethanol or soybean oil biodiesel production.
The figures show the change in cropland area between these two scenarios. To facilitate
comparison, we normalized the units from all studies to million acres of cropland per billion
gallons of biofuel production (Mac/Bgal). For dynamic models, results are reported for the peak
year of the biofuel shock. We made no other efforts to harmonize these estimates, and we note
that caution is needed when interpreting these figures for many reasons including the divergent
scenarios modeled and differences in the definition of cropland between models.

313 pievin; r j ; (2022). "Choices in land representation materially affect modeled biofuel carbon intensity
estimates." Journal of Cleaner Production: 131477.

155

-------
Figure 4.2.2.8-3: Cropland Area Change Estimates by Study for Corn Ethanol

RFS2 rule (aOIOJ/FASOM-FAPRIWIean LUC -











[ 1.3]















Plevin et al (2022)/GCAM-T/Median LUC-



¦T25J

















CARB (2014)/GTAP-BIO+AEZ- E F /Mea ri LUC-



[ 0.78 |









GREET (2021)/GTAP-BIO+CCLUB -

Talieripour et al. (2017)/GTAP-BIO+AEZ-EF/Central LUC-

ICAO (2021 )/GLOSIOM/Central LUC -









[ 0 46 )







Crop Region
World
I Non-USA
F USA

I"1 I

ICAO (2021VGTAP-BIO+AEZ-EF/Centrsl LUC -

Laborde elal (2014VMIRAGE-

0 0	0.5	1.0

Corn Starch Ethanol Crop Area Change (M Acres / Bgal)

Notes: The name on they-axis for each bar/estimate includes multiple descriptors separated by In order, these descriptors are
the author or other name (e.g. , RFS2 rule), the name of the model used to estimate land use change impacts, and a brief descriptor
of the scenario modeled. For studies that did not report disaggregated estimates by USA and Non-USA we only report the World
total. Scenarios modeled and definitions of cropland differ across studies. ICAO (2021) estimates for corn ethanol to jet fuel were
adjusted based on the assumed jet fuel yield.

For corn ethanol, we see relatively wide variation across studies in the amount of
cropland expansion estimated per gallon of biofuel production, ranging from approximately 0.2
to 1.5 Mac/Bgal ethanol. The largest estimate comes from the FASOM-FAPRI modeling for the
2010 RFS2 rule. The next largest estimate comes from GCAM-T (Plevin et al. 2022). Estimates
from GTAP-BIO straddle the most recent estimate from GLOBIOM. The lowest estimate comes
from a 2014 study for the European Commission using the MIRAGE model. Estimates from the
same model also show large variations. The estimates from GTAP-BIO range from 0.25 to 0.8
Mac/Bgal. For studies that report cropland area changes by region, there are also large
differences in how much of the estimated cropland impacts occur in the U.S. versus other
regions. Among these studies, GCAM-T estimates the largest share of cropland impacts in the
U.S. (Plevin et al. 2022) and the GTAP-BIO modeling by Taheripour et al. (2017) estimates the
lowest share of cropland impacts in the U.S.

The empirical studies reviewed in Austin et al. (2021) provide a point of comparison for
the U.S. cropland area estimates in Figure 4.2.2.8-3 which come from economic models. Lark et

156

-------
al. (2022) estimates increased cropland area of 0.9 Mac/Bgal of corn ethanol (95% confidence
interval of 0.8 to 1.1). Li et al. (2019) estimates a direct effect of 0.6 Mac/Bgal of corn ethanol
capacity (95% confidence range of 0.4 to 0.8), excluding price effects. If we use the Li et al.
(2019) elasticity of cropland area to crop price estimated (0.07 +/- 0.02) and assume crop prices
increase 1.8% (+/- 0.7%) per Bgal of corn ethanol based on the empirical study by Roberts and
Schlenker (2013)314, we derive an estimate from Li et al. (2019) including crop price effects of
1.0 Mac/Bgal (range of 0.4 to 1.4). This range of 0.4 to 1.4 Mac/Bgal for U.S. cropland
encompasses the Lark et al. (2022) estimate and the four highest of the economic model
estimates in the figure above. Other estimates from the Austin et al. (2021) review include
estimates from a PE model called BEPAM of 0.47 Mac/Bgal (Chen and Khanna 2018),315 and
0.14 Mac/Bgal (Khanna et al. 2020).316 Using an unnamed PE model, Bento et al. (2015)
estimated 0.33 Mac/Bgal for the period from 2009-2012 and 0.49 Mac/Bgal for 2012-2015.
Other estimates reviewed in Austin et al. (2021) either predate the 2010 RFS2 rule analysis or
failed to consider price-induced effects. Overall, our review of estimates from PE, CGE, IAM
and empirical studies includes a range for the effect of corn ethanol on U.S. cropland spanning
from 0.14 to 1.4 Mac/Bgal. There is overlap between empirical and modeled estimates in the
range of 0.4 to 1.0 Mac/Bgal, although this includes only two empirical studies. Below we
discuss some of the modeling assumptions that influence these cropland area estimates and their
implications for land use change GHG estimates.

314	Roberts, M. J. and W. Schlenker (2013). "Identifying Supply and Demand Elasticities of Agricultural
Commodities: Implications for the US Ethanol Mandate." American Economic Review 103(6): 2265-2295.

315	Chen, X. and M. Khanna (2018). "Effect of corn ethanol production on Conservation Reserve Program acres in
the US." Applied Energy 225: 124-134.

316	Khanna, M., et al. (2020). "Assessing the Additional Carbon Savings with Biofuel." BioEnergy Research 13(4):
1082-1094.

157

-------
Figure 4.2.2.8-4: Cropland Area Change Estimates by Study for Soy Biodiesel

CARB (2014)/GTAP'BIOAEZ-EF/Meari LUC -

ICAO (2021 yGLOBIOM/Central LUC -









GREET (2021 )/GT AP-B 10+CC LUB/Avg. Proxy-



¦jM3j









GREET (2021 VGTAP-8IO+CCLUB/Case 8 -

Laborde etal. (2014VMIRAGE-

GREET (2021)K3TAP-BIO+CCLUB/GTAP 2011 •

ICAO (2021 VGTAP-BICHAEZ-EF/Ceritral LUC -

Taheripour et al. (2017 )/GT AP-B I O+AEZ-ERCentral LUC -

Crop Region
[ Wold
I Non-USA
I USA

0 0	0.5	1 0	1.5

Soybean Oil Biodiesel Crop Area Change (M Acres / Bgal)

Notes: The name on they-axis for each bar/estimate includes multiple descriptors separated by "if". In order, these descriptors are
the author or other name (e.g., CARB), the name of the model used to estimate land use change impacts, and in some cases a
brief descriptor of the scenario modeled. For studies that did not report disaggregated estimates by USA and Non-USA we only
report the World total. Scenarios modeled and definitions of cropland diffei across studies. ICAO (2021) estimates for soybean
oil to jet fuel were adjusted based on the assumed jet fuel yield relative to biodiesel.

Figure 4.2.2.8-4 summarizes estimated crop area changes associated with soybean oil
biodiesel production. The estimates from the 2010 RFS2 rule are excluded from this chart to
improve legibility as they project a much larger amount of cropland change (6.6 M Acres/Bgal).
As discussed below, the relatively large cropland area impact from FASOM-FAPRI is not
accompanied by similarly large land use change GHG estimates due to the way these models
estimate other land types to shift in response.31. There is a group of estimates between 1.2 and
1.5 million acres per billion gallons, including estimates with GLOBIOM, MIRAGE and the
GTAP-BIO modeling done with CARB for the CA-LCFS. The lowest group of estimates, 0.2 to
0.5 million acres per billion gallons, all come from the GTAP-BIO model. The most recent

317 For example, for the U.S. FASOM projected an increase in cropland of 3.5 M acres.Bgal, but the GHG impacts
were muted because the cropland increases were accompanied by decreases in idle land (-5.7 M aCres/Bgal) and
rangeland ( 5.4 M atres/Bgal) and increases in forest-pasture land (6.9 M acres/Bgal). For the rest of the world,
FAPRI projected a large increase in cropland (3.1 M acres/Bgal) accompanied by a reduction in pasture area
(associated with the market-mediated effects of increased soybean meal supplies on livestock producers) and
accompanying increases in forest area (0.7 M acres/Bgal) especially in Brazil.

158

-------
GTAP-BIO estimates are the lowest among the group, at 0.24-0.25 million acres per billion
gallons. Between the GTAP-BIO modeling for CARB and the more recent estimates, a number
of updates were made to the GTAP-BIO model that lowered the cropland area estimates,
including revising the assumptions that determine crop intensification in response to price
changes and multi-cropping (Taheripour et al. 2017). Unlike corn ethanol, we did not identify
any empirical estimates of cropland area changes attributable to soybean oil biodiesel
production. Thus, we do not have any empirical comparisons with the modeled estimates in
Figure 3.2.2.8-4. Below we discuss some of the factors that influence these modeled estimates of
cropland area impacts and their implications for land use change GHG emissions estimates.

The second broad element that contributes to land use change GHG estimates is the type
of land impacted. The next two figures compare area changes in cropland, pasture, forest and
other land types across studies for corn ethanol and then soybean oil biodiesel.318 Similar to the
figures above for cropland area impacts, we normalized the estimates per billion gallons of
biofuel to facilitate comparison. In some cases, we aggregated or slightly modified the categories
of land to facilitate comparison. For example, we aggregated GCAM-T's commercial and non-
commercial forest categories and report them as forest. Once again, caution is needed comparing
results as we made no other efforts to align scenario design or other factors.

318 The MIRAGE estimates from Laborde et al. (2014) are also excluded as that study does not report area changes
by land type.

159

-------
Figure 4.2.2.8-5: Land Use Change Estimates by Study for Corn Ethanol

-2	-10	12

Corn Starch Ethanol Area Change (M Acres / Bgal)

Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by In order, these descriptors are
the author or other name (e.g.. CARB). the name of the model used to estimate land use change impacts, and in some cases a
brief descriptor of the scenario modeled. In some cases, we have aggregated or relabeled land Categories to facilitate comparison
across estimates. For example, forest pasture in FASOM is reported as forest in this figure. ICAO (2021) estimates for corn
ethanol to jet fuel were adjusted based on the assumed jet fuel yield.

All of the estimates in the above figure reported cropland area increases and decreases in
other land types in response to additional corn starch ethanol production. However, the types of
land displaced by increased cropland differs among the estimates. The estimates with GTAP-
BIO have cropland pasture decreasing more than any other land type. As discussed more below,
cropland pasture plays an important role in land use change modeling both in terms of how it is
represented, and the estimated soil carbon changes associated with its use for crop production.
Pasture is modeled separately from cropland pasture, it decreases in area in all of the estimates,
and it is the land type that decreases the most in the GCAM-T estimate from Plevin et al. (2022).
Most of the estimates include a relatively small amount of forest area decreases, with the
GCAM-T estimate reporting the largest decrease. The GCAM-T and GLOBIOM estimates
include decreases in other natural land types such as grassland, savannah, and shrubland; the
GTAP-BIO estimates do not report decreases in these land types as the model only represents
commercially managed land.

160

-------
Figure 4.2.2.8-6: Land Use Change Estimates by Study for Soybean Oil Biodiesel

Taheripourelal. (2017)/GTAP-BIO-

-1	0	1

Soybean Oil Biodiesel Area Change (M Acres / Bgal)

Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by In order, these descriptors are
the author or other name (e.g., RFS2 rule), the name of the model used to estimate land use change impacts, and in some cases a
brief descriptor of the scenario modeled. In some cases we have aggregated or relabeled land categories to facilitate comparison
across estimates. ICAO (2021) estimates for soybean oil to jet fuel were adjusted based on the assumed jet fuel yield relative to
biodiesel.

The land use change estimates for soy biodiesel in Figure 4.2.2.8-6 come from GTAP-
B10 and GLOBIOM. There are several estimates from GTAP-BIO as the GREET model gives
users multiple choices of which GTAP-BIO estimates to use when estimating soybean oil
biodiesel land use change GHG emissions. The GTAP-BIO estimates conducted for the CA-
LCES have much larger areas of cropland expansion than the other more recent GTAP-BIO
estimates. However, the pattern of land use change is similar in all the GTAP-BIO estimates,
with cropland pasture decreasing the most followed by pasture and managed forest. In contrast,
the GLOBIOM estimates have other natural land (e.g., grassland, savannah, shrubland) as the
largest category of decline. GLOBIOM does not include cropland pasture as a categoiy, but it
includes "abandoned" land that GTAP-BIO does not represent. In all the estimates in this figure,
forest declines by a relatively small amount, with the most recent GTAP-BIO estimates by
Taheripour et al. (2017) and ICAO (2021) showing very small areas of forest loss. We note once
again that caution is needed interpreting these results as the figure above aggregates land
categories for comparison across models to facilitate comparison. Land categories are defined
and modeled differently across the models.

161

-------
The RFS2 (2010) estimates are not included in Figure 4.2.2.8-6 because they are different
in their size and pattern and would reduce legibility. In the RFS2 (2010) estimates, cropland area
increases 6.6 M acres per Bgal, and pasture declines by about the same amount. Although RFS2
(2010) estimates much larger cropland area increases than the other studies, the associated land
use change GHG emissions are more similar in size due to the pattern of other land use changes.
For example, the RFS2 (2010) results include an increase in forest pasture area in the U.S. (6.9
M acres per Bgal) and an increase in forest area (0.7 M acres per Bgal) outside of the U.S.
associated with a reduction in pasture area particular in Brazil. These forest area increases offset
the GHG impacts of expanding cropland in the RFS2 (2010) analysis.

The third broad element that contributes to land use change GHG estimates are the
emissions factors and formulas that are used to convert areas of land use change to GHG
emissions. These emissions factors are developed from data on biomass (above and below
ground) and soil carbon for multiple land categories for regions. We do not have sufficient
information at this time to summarize these emission factors in a figure the directly and
meaningfully compares these inputs assumptions. However, we can make some observations
based on our review of the available studies.

At the most basic level, we can clearly say that the land use change emissions factors are
an influential part of biofuel GHG modeling. For the GTAP-BIO studies reviewed, two different
tools are used to estimate land use change GHG emissions based on the estimated land area
changes: the AEZ-EF model319 and the CCLUB model.320 These tools essentially apply different
emission factors to the land area change outputs from GTAP-BIO. Based on cases where both
tools have been applied to the same GTAP-BIO estimates we can see that they have a relatively
large effect on the resulting land use change GHG estimates. Our literature review includes 12
corn ethanol land use change GHG estimates from GTAP-BIO+AEZ-EF ranging from 12 to 37
gC02e/MJ, and 9 estimates from GTAP-BIO+AEZ-EF ranging from -1 to 9 gC02e/MJ. Chen et
al. (2018) applied both CCLUB and AEZ-EF to four sets of GTAP-BIO results. Using AEZ-EF
the land use change GHG emissions estimates are 22, 26, 17 and 18 gC02e/MJ. Applying
CCLUB to the same GTAP-BIO results gives land use change GHG estimates of 8, 10, 4 and 6.
In these cases, using CCLUB reduced the estimates by 12-16 gC02e/MJ, or about 67% on
average, relative to using AEZ-EF. The data and assumptions within each of these tools are also
important. Plevin et al. (2015) conducted a Monte Carlo simulation and found that varying key
assumptions within AEZ-EF could increase or decrease corn ethanol and soy biodiesel land use
change GHG estimates by about 20% in either direction (95% confidence interval).321

In this section (Chapter 4.2.2), we have reviewed available models for biofuel GHG
analysis, the main characteristics of these models, and GHG estimates they have produced for
corn ethanol and soybean oil biodiesel. Based on this review we have found that there are four
different types of models (PE, CGE, IAM, LCI) with fundamentally different structures. These
models have large differences in the sectors and GHG emission they cover, the way they

319	Plevin, R., et al. (2014). Agro ecological Zone Emission Factor (AEZ EF) Model (v52): A model of greenhouse
gas emissions from land use change for use with AEZ based economic models.

320	Kwon, H., et al. (2020). Carbon Calculator for Land Use and Land Management Change from Biofuels
Production (CCLUB).

321	Based on visual inspection of Figure SI 1.

162

-------
represent land and time, and many other differences. Our review shows that land use change is
an important and uncertain aspect of crop-based biofuel GHG modeling. There continues to be a
wide range of available estimates of land use change GHG emission associated with corn
ethanol, soybean oil biodiesel and soybean oil renewable diesel. We have made some
observations about key factors that are contributing to the differences in these estimates, but we
do not have a systematic quantification of which factors are most important across all of the
models. Furthermore, in many areas where the models disagree in their assumptions (e.g., crop
intensification, soil carbon effects, livestock market effects, fuel market effects) there is
disagreement or a lack of information on the most scientifically justifiable values to use for this
type of modeling.

A major challenge in our review is that available estimates come from studies that
evaluated different scenarios and reported results in different formats and units. In order to
provide additional information to inform the final rule, we intend to conduct a model comparison
exercise that will address some of these challenges. For the model comparison exercise, we will
use a common set of scenarios and align model outputs to the extent possible. We will also
attempt to align input assumptions and time permitting conduct sensitivity analyses. We believe
this will give us a much greater level of information about the best available biofuel GHG
modeling to inform the final rule. Given the complexity of crop-based biofuel GHG modeling, it
is likely the model comparison exercise will raise further important questions. While this
exercise will focus on comparing economic and lifecycle simulation models, we will also
continue to review relevant empirical studies (see Chapter 4.2.2.6) and other science to inform
and compare with the modeling.

4.2.3 Range of LCA Estimates by Fuel Pathway for Illustrative Scenario

As discussed at the beginning of Chapter 4.2, our assessment of the climate change
impacts of the proposed rule relies on an extrapolation of lifecycle analyses (LCA) of GHG
emissions. As we did in the 2020-2022 RVO rulemaking, this approach involves multiplying
LCA emissions of individual fuels by the change in the candidate volumes of that fuel to
quantify the GHG impacts. We repeat this process for each fuel (e.g., corn ethanol, soybean
biodiesel, landfill biogas CNG) to estimate the overall GHG impacts of the candidate volumes.
In the 2020-2022 RVO rulemaking, we applied the LCA estimates that we developed in the
March 2010 RFS2 rule (75 FR 14670) and in subsequent agency actions. In this rulemaking, we
are updating our approach to use a range of LCA emissions estimates that are in the literature.
Instead of providing one estimate of the GHG impacts of each candidate volume, we provide a
high and low estimate of the potential GHG impacts, which is inclusive of the values we
estimated in the 2010 RFS final rule and subsequent agency actions. We then use this range of
values for considering the GHG impacts of the candidate renewable fuel volumes that change
relative to the No RFS baseline.

In this chapter we present the range of LCA estimates contained within the published
literature for each fuel pathway.322 We conducted a high-level review of relevant literature for

322 Details on the sourcing of each estimate from our literature compilation are available in a memo to the docket for
this proposed rule titled "Notes on Literature Review of Transportation Fuel Greenhouse Gas (GHG) Lifecycle
Analysis (LCA)."

163

-------
the biofuel pathways (combination of biofuel type, feedstock, and production process) that would
be most likely to satisfy the candidate renewable fuel volumes. Our literature review was broad
and includes studies that estimate the lifecycle GHG emissions associated with the relevant
biofuel pathways and the petroleum-based fuels they replace.

We believe the presentation of ranges based on literature review provides important
information for readers about the variety of estimates in the scientific literature and illustrates the
level of uncertainty in these estimates. In Chapter 4.2.4 we describe how the LCA estimates for
each pathway are used to estimate a potential range of GHG impacts associated with the
candidate volumes relative to the No RFS baseline for an illustrative 30-year scenario. The
chapter that follows (Chapter 4.2.5) describes how the GHG estimates are used to estimate the
potential monetized benefits of the GHG impacts associated with the candidate volumes relative
to the No RFS baseline for an illustrative 30-year scenario.

We reviewed relevant literature and identified a range of lifecycle GHG estimates for the
biofuel pathways with increased consumption in the candidate volumes scenario relative to the
No RFS baseline scenario (see Chapter 3 for description of these scenarios). We also identified a
range of lifecycle GHG estimates for the conventional fossil-based fuels that the biofuels are
likely to replace. Our compilation includes journal articles, major reports and studies that inform
biofuel-related policies. We include estimates from the March 2010 RFS2 rule and studies
published after the March 2010 RFS2 rule, as that rule considered the available science at the
time. We do not claim that our compilation is fully comprehensive, but we attempted to include
relevant studies published before Spring 2022. In cases where there were multiple studies that
include updates to the same general model and approach, we included only the most recent
study. However, we include a subset of older estimates that are still used for major regulatory
programs or that continue to be widely cited for other reasons. We focused our compilation on
estimates of the average type of each fuel produced in the United States (e.g., natural gas-fired
corn ethanol plants), though we include a discussion at the end of this section (Chapter 4.2.3.12)
about how advanced technologies could lead to more significant emissions reductions in the
future. In this section (Chapter 4.2.3) we focus on studies that estimate full lifecycle (or "well-to-
wheel") GHG emissions. For crop-based biofuels, there are many studies that only estimate land
use change GHG emissions; these studies are discussed in Chapter 4.2.2.8.

Many of the studies we compiled include sensitivity analysis, where many parameters are
varied to produce a large number of estimates. In these cases, we include representative high and
low estimates. For example, when studies report a 95% confidence interval, we use only the
central estimate (usually the default, mean or median estimate) and the estimates at the top and
bottom of the confidence interval. This approach simplifies the presentation of results relative to
including every estimate in between. We believe this approach is appropriate given that the
primary purpose of our literature review is to produce a range (high and low estimate) for each
pathway. We intentionally do not calculate or present any statistics (e.g., mean, median) derived
from the estimates included in our literature review, as we do not believe such statistics would be
meaningful or appropriate based on the design of our literature compilation. As discussed below,
in some cases we remove outlier estimates in order to form a range that we believe is
representative of the likely upper and lower GHG impacts of each biofuel pathway on a national
average basis. We believe this is appropriate as the purpose of our review is to consider national

164

-------
average fuel production, not regional variation or unique conditions that are unlikely to represent
the impacts of the candidate volumes relative to the No RFS baseline.

Figure 4.2.3-1 provides an overview of the lifecycle GHG estimates in our literature
compilation. This chart only includes studies that report the full well-to-wheel emissions
associated with each pathway. All of the pathways in our compilation are included with the
exception of compressed natural gas (CNG) produced from manure digester biogas, as some of
the estimates for this pathway (e.g., 533 gCChe/MJ) are so low that they skew the rest of the
chart. All of the estimates in this chart report lifecycle GHG estimates as carbon dioxide-
equivalent (CChe) emissions per megajoule (MJ) of fuel consumed. All CChe estimates are based
on 100-year global warming potential (GWP) from the IPCC.323 This allows us to compare all of
the estimates on a gCChe/MJ of fuel basis. However, we stress that many of the studies in this
chart do not align in terms of their scope, system boundaries, time horizon, year of analysis, or
other factors. Therefore, the estimates reported in this figure give us a sense for the range of
estimates for each pathway, but we must exercise caution when comparing estimates and
drawing conclusions.

323 The reviewed estimates use GWP values from the IPCC Second Assessment Report (SAR), Fourth Assessment
Report (AR4) or Fifth Assessment Report (AR5). We did not attempt to harmonize GWP assumptions across studies
as many studies only reported C02e results and not emissions by gas.

165

-------
Figure 4.2.3-1 Lifecycle GHG Emissions Estimates by Pathway

Petroleum Gasoline -

• •V

Petroleum Diesel -

•V

Natural Gas CNG -

••••*

Corn Starch Ethanol -

Soybean Oil Biodiesel -

Soybean Oil Renewable Diesel-

• • •
W • (I
• •

• •
• • •

¦ »
: ••• •

• • •• •
• • • ••• • HH

• • •

• • •

• • • • •

UCO Biodiesel -

•	a

•	• • •

UCO Renewable Diesel -

• • •« •
» ••• ••

Tallow Biodiesel-



Tallow Renewable Diesel - •

• • •

DCO Biodiesel -

• • •• •

DCO Renewable Diesel -

LFG CNG -

• •

0	40	80	120

Lifecycle GHG Emissions (gC02e/MJ)

Notes: LFG = landfill gas, CNG = compressed natural gas, DCO = distillers corn oil, UCO = used cooking oil.
Other than reporting all estimates in gC02/MJ no effort has been made to harmonize estimates. Corn starch ethanol
and LFG electricity are reported following gasoline as these pathways are more likely to displace gasoline than
diesel. The other pathways are reported after petroleum diesel as they are more likely to displace diesel.

Figure 4.2.3-1 shows that, in general, the CI estimates for biofuel pathways tend to be
lower than those for petroleum gasoline, diesel, and natural gas. GHG emissions for biofuels
produced from corn and soybeans tend to be higher than those produced from used cooking oil,
tallow or landfill gas. However, there are some high estimates for the tallow-based pathways

166

-------
when animal production emissions are allocated to the tallow.324 Most lifecycle GHG estimates
are presented as grams of CChe emissions per megajoule (gCChe/MJ) of additional renewable
fuel consumed. These estimates are often called the "carbon intensity" (CI) of the fuel or the
well-to-wheel (WTW) GHG emissions associated with the fuel. Below, we summarize the results
of this literature review for each of the relevant biofuel pathways. For better legibility we provide
a list of references at the end of this section rather than using footnotes.

4.2.3.1 Length of Time Period for Analysis

The time period over which land use change emissions are quantified influences the GHG
estimates for crop-based renewable fuels. If increased demand for biofuels leads to land
conversion, an initial pulse of emissions would likely be released 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 use change.325 Over time, if the
biofuel production continues, the GHG benefits of displacing fossil fuels may eventually "pay
back" the initial increase in GHG emissions from the first year. Thus, when increased biofuel
production is expected to result in land conversion, longer analytical time horizons should result
in greater GHG reduction estimates than shorter time horizons, provided other assumptions are
met. The question of the appropriate time horizon over which to evaluate the net emissions can
depend on many factors (e.g., the lifetime of the project, the goals of the program, future
projections of renewable fuels use). After considering public comments and the input of an
expert peer review panel, in the March 2010 RFS2 rule (75 FR 14670), EPA determined that our
lifecycle greenhouse gas emissions analysis for renewable fuels would quantify the GHG
impacts over a 30-year period. One of the reasons for using 30 years as a reasonable time horizon
for analysis is that biofuel production facilities last multiple decades after they are constructed.

EPA continues to believe that 30 years is an appropriate timeframe for evaluating the
lifecycle GHG emissions of renewable fuels for purposes of determining which fuel pathways
satisfy the statutory GHG reduction thresholds for qualification under each of the four categories
of renewable fuel. With respect to estimating the GHG impacts of this rulemaking specifically,
the CAA gives us discretion to choose the appropriate analytical time period. On one hand, this
Set rule is part of the broader RFS program that has been in existence since 2005, so there have
been long-term market impacts of standards that were set in past individual years. Furthermore,
once the cost of clearing and converting land is incurred, and given that global cropland areas are
expected to continue expanding, it seems likely that land will continue to be used for agricultural

324	Seber, G., et al. (2014). "Environmental and economic assessment of producing hydroprocessed jet and diesel
fuel from waste oils and tallow." Biomass and Bioenergy67: 108-118.

325	The initial pulse of emissions may take longer than one year depending on the fate of the biomass cleared from
the land. For example, if the biomass is burned, the emissions will indeed occur in the first year. If it is left on the
ground or landfilled the emissions associated with biomass decay may occur over several years. The lifecycle GHG
analyses for the March 2010 RFS2 rule allocated international biomass clearing emissions to the first year. We said
at the time that this was a simplification that was appropriate for the purposes of the analysis. EPA (2010).
Renewable fuel standard program (RFS2) regulatory impact analysis. Washington, DC, US Environmental
Protection Agency Office of Transportation Air Quality. EPA-420-R-10-006. Section 2.4.4.2.6.8.

167

-------
purposes in the future for a period of time.326 On the other hand, the volumes in this rule do not
extend beyond 2025 and making projections about future policies, volume requirements, and
renewable fuel use are inherently uncertain. However, since we have chosen to use LCA GHG
estimates as the proxy for evaluating the climate change impacts of this rule, it is consistent to
use the same 30-year analytical time period for the purposes of estimating the GHG impacts of
this rule. Therefore, in the illustrative GHG scenario presented in Chapter 4.2.4, the analyses
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.

In order to estimate the monetized social cost or benefit of the candidate biofuel volumes
in Chapter 4.2.5, annual streams of emissions are required. In the literature review described
below, all the studies that include land use change annualize these emissions over a time period
of 20 to 30 years. However, many of these studies do not report annual streams of emissions.
Rather, they report average annualized emissions over a 20-30-year period, and it is not possible
to produce a credible annual stream of emissions from these estimates based on the information
provided. Thus, to develop ranges for purposes of estimating the monetized GHG impacts we
must rely on the high and low estimates that report annual streams of emissions. This is
discussed further below for each of the crop-based pathways that involve land use change
emissions.

4.2.3.2 Petroleum Gasoline and Diesel

The net GHG impacts of the production and use of biofuels depends on the GHG
emissions associated with the conventional fuels they displace. For the purposes of conducting
the lifecycle GHG emissions analysis and determining which biofuels meet the GHG
requirements, CAA Section 211 (o) (1) (C) defines baseline lifecycle greenhouse gas emissions as
"the average lifecycle greenhouse gas emissions, as determined by the Administrator, after notice
and opportunity for comment, for gasoline or diesel (whichever is being replaced by the
renewable fuel) sold or distributed as transportation fuel in 2005." As the baseline lifecycle GHG
emissions are used for a specific purpose under the RFS program, we are not required to use it
for evaluating the GHG impacts of this proposed rule. Given that this rule involves biofuel
production and use in 2022 and beyond, we believe it is appropriate to consider LCA estimates
for gasoline and diesel production that occurred more recently than 2005. Furthermore, given
that we are developing a range LCA estimates from literature for biofuels, we believe a similar

326 Globally cropland areas have been expanding, suggesting that once land is put into cultivation it is likely to stay
under production. Potapov et al. (2022) report that cropland area increased by 9% globally from 2003 to 2019.
Furthermore, integrated assessment modeling of future scenarios suggests that global cropland areas, including
bioenergy cropland, are expected to increase through 2100. See for example the figure on page 32 of IPCC (2019).
Potapov, P., Turubanova, S., Hansen, M.C. et al. Global maps of cropland extent and change show accelerated
cropland expansion in the twenty-first century. Nat Food 3, 19-28 (2022). hi:t:ps://doLorg/10.1038/s43018-021-
00429-z; IPCC, 2019: Summary for Policymakers. In: Climate Change and Land: an IPCC special report on climate
change, desertification, land degradation, sustainable land management, food security, and greenhouse gas fluxes in
terrestrial ecosystems [P.R. Shukla, J. Skea, E. Calvo Buendia, V. Masson-Delmotte, H.- O. Portner, D. C. Roberts,
P. Zhai, R. Slade, S. Connors, R. van Diemen, M. Ferrat, E. Haughey, S. Luz, S. Neogi, M. Pathak, J. Petzold, J.
Portugal Pereira, P. Vyas, E. Huntley, K. Kissick, M. Belkacemi, J. Malley, (eds.)]. In press.

168

-------
approach is appropriate for the conventional fuels they replace. Thus, our literature review for
this proposed rule includes studies that estimate the lifecycle GHG emissions associated with
petroleum-based gasoline and diesel.

For the March 2010 RFS2 rule, EPA estimated the lifecycle GHG emissions associated
with average 2005 gasoline and diesel. Our review includes the 2010 RFS2 rule and studies that
estimated lifecycle GHG emissions for average U.S. gasoline and diesel that were published
following the 2010 RFS2 rule. Studies that estimate only the GHG emissions associated with
crude oil extraction327 or refining328 are excluded from our review, as we require estimates of the
full lifecycle GHG emissions. It is not appropriate to simply sum crude oil extraction estimates
with refining estimates as the properties of the crude oil from different wells significantly effects
the refining emissions.

We recognize that the CI of the average gallon of gasoline and diesel replaced by biofuels
may be different than the CI of the marginal gallons replaced. While there is one study that
suggests the CI of the marginal oil supplies may be higher than average oil supplies,329 we did
not identify any studies that estimate the full lifecycle GHG emissions of the marginal volumes,
including oil extraction, oil transport, refining, fuel distribution and use.

327	See for example, Masnadi, M. S., et al. (2018). "Global carbon intensity of crude oil production." Science
361(6405): 851-853.

328	See for example, Jing, L., et al. (2020). "Carbon intensity of global crude oil refining and mitigation potential."
Nature Climate Change 10(6): 526-532.

329	Masnadi, M. S., et al. (2021). "Carbon implications of marginal oils from market-derived demand shocks."
Nature 599(7883): 80-84.

169

-------
Figure 4.2.3.2-1: Petroleum Gasoline Lifecycle GHG Estimates

0	25	50	75	100

Petroleum Gasoline GHG Emissions (gC02e/MJ)

Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by "/". In order, these descriptors are
the author or other name (e.g., RFS2 rule) and a brief descriptor of the scenario modeled. The Upstream stage includes all of the
emissions associated with extracting, handling and delivering crude oil to the refinery gate. The. Conversion stage includes
emissions associated with refining. The Downstream stage includes emissions associated with gasoline distribution and tailpipe
combustion emissions. The gasoline baseline estimate in the March 2010 RFS2 rule used SAR GWP values. All values in this
chart use 100-year AR5 GWP values.

170

-------
Figure 4.2.3.2-2: Petroleum Diesel Lifecycle GHG Estimates

0	25	50	75

Petroleum Diesel GHG Emissions (gC02e/MJ)

Notes: The name on the y-axis for eaich bar/estimate includes multiple descriptors separated by 7". In order, these descriptors are
the author or other name (e.g., RFS2 rule) and a brief descriptor of the scenario modeled. The Upstream stage includes all of the
emissions associated with extracting, handling and delivering crude oil to the refinery gate. The. Conversion stage includes
emissions associated with refining. The Downstream stage includes emissions associated with diesel distribution and tailpipe
combustion emissions. The diesel baseline estimate in the March 2010 RFS2 rule used SAR GWP values, but all the values in
this chart use 100-year AR5 GWP values.

The 2010 RFS2 estimates were largely based on a study by the National Energy
Technology Laboratory (NETL).330 A team of NETL researchers published new estimates of the
lifecycle GHG emissions associated with 2005 and 2014 average U.S. gasoline and diesel
(Cooney et al. 2017). For 2005 average diesel the Cooney et al. (2017) estimates are very similar
to our estimates for the 2010 RFS2 rule, For 2005 average gasoline the Cooney et al. (2017)

330 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-OAR 2005-0161-3151

171

-------
estimates are higher by approximately 4 gC02e/MJ. The Cooney et al. (2017) estimates for 2005
average gasoline and diesel values represent the high end of the range used in our illustrative
GHG impacts assessment. The GREET-2021 model estimates lifecycle GHG emissions for
average U.S. gasoline and diesel. The GREET estimates for average gasoline and diesel are
lower than the estimates from RFS2 rule (2010) and Cooney et al. (2014) but all of these
estimates are all within 5 gCChe/MJ. The figures above also report results from a study with the
BEIOM model (Avelino et al. 2021), an "environmentally extended input-output model." The
BEIOM estimates are significantly lower than those from the other studies. BEIOM's
methodology differs significantly from the other studies in our review and is limited in
geographic scope to the United States, which may explain its lower estimates for the carbon
intensity of gasoline and diesel. Based on our review of published estimates, we use a range of
84 to 98 gC02e/MJ for gasoline and for diesel we use a range of 84 to 94 gC02e/MJ.

4.2.3.3 Corn Starch Ethanol

More studies have been published on the GHG emissions associated with corn starch
ethanol than any of the other biofuel pathways considered for this rule. Our literature review
includes 9 studies that estimate the lifecycle GHG emissions associated with corn ethanol. Many
of these studies include multiple emissions estimates based on different assumptions about the
energy efficiency of dry mill ethanol production, co-products and other factors. Some of these
studies report a large number of estimates. The figure below includes 19 estimates from these
studies that are representative of the range of results that each of them reports.

172

-------
Figure 4.2.3.3-1: Corn Starch Ethanol Lifecycle Greenhouse Gas Estimates

Lark et al. (2022)/Other/RFS2 RIA -
Brandao (2022)-
Lark et al. (2022)/Other/CA-LCFS -
CARB (2018)/GTAP-BIO+AEZ-EF/Dry Mill/High LUC-
RFS2 rule (2010)/FASOM-FAPRI/NG Dry DDGS/High LUC -
Lark et al. (2022)/Other/GREET -
CARB (2018)/GTAP-BIO+AEZ-EF/Dry Mill/Mean LUC -
RFS2 rule (2010)/FASOM-FAPRI/2022 Avg NG Dry Mill/Mean LUC-
BEIOM (2021)/Avg. Dry Mill -
CARB (2018)/GTAP-BIO+AEZ-EF/Dry Mill/Low LUC-
Scully et al. (2021)/GTAP-BIOCCLUB/High LUC -
GREET (2021 )/GTAP-BIO+CCLUB/NG Dry Mill DDGS -
GREET (2021 )/GTAP-BIO+CCLUB/Avg Plant -
GREET (2021 )/GTAP-BIO+CCLUB/Gen 1.5 w/ DCO -
Lewandrowski et al. (2019)/FASOM+GTAP-BIO/2022 BAU -
Scully et al. (2021 )/GTAP-BIO+CCLUB/Central LUC -
RFS2 rule (2010)/FASOM-FAPRI/Adv. NG Dry Mill/Low LUC-
GREET (2021 )/GTAP-BIO+CCLUB/NG Dry Mill WDGS-
Lee et al. (2021 )/GTAP-BIO+CCLUB/2019 -
Scully et al. (2021)/GTAP-BIO+CCLUB/Low LUC-

Notes: The name on the y-axis for eaieh bar/estimate includes multiple descriptors separated by 7". In order, these descriptors are
the author or other name (e.g., RFS2 rule) and year of the study; the model used to estimate the LUC emissions; the type of
natural gas-fired diy mill used for ethanol production (e.g., 2022 Avg. NG Diy Mill); the LUC estimate case (e.g., Low LUC);
and the non-LUC estimate case (e.g.. Low CI). The Upstream stage includes all of the emissions associated with corn production
and transport upstream of the ethanol production facility . The Conversion stage includes emission associated with fuel production
at the ethanol production facility. The Downstream stage includes emissions associated with ethanol transport and non-C02
combustion emissions. The LUC stage includes emissions from induced land use changes. For studies that do not report
disaggregated results, results are reported as LUC and Non-LUC emissions.

We include estimates for different natural gas-fired dry mill configurations from RFS2
(2010) and GREET-2021 (see Chapter 4.2.2 for more information on the 2010 RFS2 rule
modeling and GREET). We include the high, low and mean land use change GHG estimates
from RFS2 (2010). The CARB (2018) estimates are based on the default assumptions in the most
recent version of the CA-GREET model (version 3.0), a version of GREET developed by CARB
for the CA-LCFS. We include a range of land use change emissions for CARB (2018) based on
the CARB (2014) report describing the indirect land use change modeling that continues to be
used for CA-LCFS implementation. Lewandrowski et al. (2019) is a study that attempts to
update the RFS2 (2010) estimates based on more recent data and swaps in land use change

173

-------
estimates from GTAP-BIO. Lee et al. (2021) uses GREET to estimate U.S. corn ethanol carbon
intensity from 2005 to 2019. We include the estimate for 2019 and add the default land use
change estimate from GREET. The lowest estimate is from Scully et al. (2021), a review paper
that developed a range of LCA estimates by combining elements of prior studies. The highest
estimates are from Lark et al. (2022), a study that modeled historical U.S. land use change GHG
emissions attributable to corn ethanol and added these estimates to the LCA estimates from RFS
(2010), CARB (2018) and GREET.331 We also include the estimate from BEIOM (2021) which
uses economic input-output methodology (see Chapter 4.2.2.6 for more information). Finally, we
include the estimate from Brandao (2022), a consequential lifecycle analysis of the GHG
emissions associated with ramping up ethanol production to 15 billion gallons from 1999-2018.

Among the estimates in the above figure, upstream emissions range from 9 to 51
gC02e/MJ. These include emissions associated with feedstock production and transport,
including non-LUC market-mediated impacts in the agricultural sector. BEIOM reports the
highest upstream emissions, with the next highest estimate being 25 gC02e/MJ from GREET
(2021). The lowest estimate comes from Scully et al. (2021), which includes a relatively large
credit (-13.5 gC02e/MJ) for DGS displacing other sources of livestock feed.

We include estimates for ethanol produced at a U.S. dry mill facility using natural gas
and electricity for energy,332 as dry mills produce over 90% of U.S. fuel ethanol and natural gas
and electricity account for almost all of the energy use at these facilities.333 Among the studies in
Figure 4.2.3.2-1, conversion emissions range from 13 to 33 gC02e/MJ.334 The highest estimates
are from CARB (2018) based on the default assumptions used in the CA-GREET3.0 model. The
GREET-2021 estimate for "industrial average dry mill corn ethanol production is 21 gC02e/MJ,
and the RFS2 (2010) estimate for a projected 2022 natural gas dry mill facility is 26 gC02e/MJ.
The lowest GHG estimate of 13 gC02e/MJ for the fuel production stage comes from the most
advanced natural gas-fired dry mill facility evaluated in the 2010 RFS2 rule. This advanced
facility includes wet DGS, corn oil fractionation, combined heat and power (CHP) and
membrane separation technologies.335

The largest source of variation between estimates are the land use change emissions,
ranging from -1 to 65 gC02e/MJ in Figure 4.2.3.3-1. The highest land use change estimates are
from Lark et al. (2021), which produced new estimates of U.S. land use change attributable to
corn ethanol and added the U.S. emission to non-US land use change estimates from RFS2

331	Lark et al. caveat that incorporating their U.S. land use change emissions into other fuel program estimates is
only a partial analysis, and that to accurately assess the carbon intensity of corn ethanol, a full reanalysis is needed
to ensure consistent treatment and systems boundaries.

332	In accordance with CAA 211 (o) (2) (A) (i) renewable fuel production from facilities that commenced construction
prior to December 19. 2007, are exempt from the 20% GHG reduction requirement to qualify as renewable fuel
under the RFS program. Our review in this section focuses on average dry mill corn ethanol production in the U.S.
regardless of facility status pursuant to this "grandfathering" exemption.

333	Lee, U., et al. (2021). "Retrospective analysis of the US corn ethanol industry for 2005-2019: implications for
greenhouse gas emission reductions." Biofuels, Bioproducts and Biorefining

334	This excludes the studies that do not report disaggregated non-LUC emissions.

335	Including this facility in our review also allows us to include a wider range of RFS2 (2010) estimates which is
beneficial for the 30-year illustrative scenario as the RFS2 (2010) estimates are the only ones that report a
differentiated 30-year stream of annual emissions. Without this estimate the range used for the illustrative scenario
would be further from the full range identified in the literature.

174

-------
(2010), GREET and CARB. Scully et al. (2021) reports negative land use change emissions in
part because they assumed that planting annual crops on land categorized as cropland pasture
would result in net sequestration of soil carbon.336 For more discussion of corn ethanol land use
change estimates see Chapter 4.2.2.

Downstream emissions range from 1 to 7 gC02e/MJ. Downstream emissions are
associated with fuel distribution from ethanol production facilities to retail gasoline stations and
tailpipe emissions. Some studies also include emissions from production and use of denaturant
which is added to ethanol in small volume percentages to render it undrinkable. The highest
downstream estimate is from Scully et al. (2021) including emissions associated with denaturant.
All of the other estimates are 4 gC02e/MJ or less.

Overall, our literature review of estimates representative of average corn ethanol
production at natural gas-fired U.S. dry mills produces a range from 38 to 116 gC02e/MJ. The
largest source of variation across studies continues to be estimated emissions associated with
direct and indirect land use change. Although this is already a wide range, corn ethanol can have
higher or lower GHG emissions depending on farm specific or facility specific factors. For
example, ethanol produced from facilities fired with coal or from corn grown on marginal lands
with lower yields produce emissions that are greater than the top end of this range. On the other
hand, corn ethanol produced with the adoption of advanced technologies or climate smart
agricultural practices can have lower LCA emissions. Corn ethanol facilities produce a highly
concentrated stream of C02 that lends itself to carbon capture and sequestration (CCS). CCS is
being deployed at ethanol plants and has the potential to result in negative emissions at the
ethanol production facility, especially if mills with CCS use renewable sources of electricity and
other advanced technologies to lower their needs for thermal energy. Climate smart practices are
being adopted at the feedstock production stage. For example, planting cover crops between corn
rotations can build soil organic carbon stocks. Collecting data on and evaluating these trends in
corn and ethanol production are areas for additional effort that will inform future LCA estimates
for corn ethanol.

4.2.3.4 Soybean Oil Biodiesel

Relative to corn ethanol, there have been fewer studies published on the GHG emissions
associated with soybean oil biodiesel. Our literature review includes 8 studies that estimate the
lifecycle GHG emissions associated with soybean oil biodiesel. Given that soybeans are
approximately 20% oil and 80% meal by dry mass, these studies often include several estimates
based on different allocation approaches for the soybean meal coproduct. The figure below
includes 20 estimates of the lifecycle GHG emissions associated with soybean oil biodiesel
production and use.

336 Spawn-Lee, S. A., et al. (2021). "Comment on 'Carbon Intensity of corn ethanol in the United States: state of the
science'." Environmental Research Letters 16(11): 118001.

175

-------
Figure 4.2.3.4-1: Soybean Oil Biodiesel Lifecycle Greenhouse Gas Estimates

Stage

¦

LUC



Non-LUC



Downstream



Conversion



Upslrearn

CARB (2018)/GTAP-B IO+AEZ-EF/High LUC-
RFS2 rule (2010>/FASOM-FAPRI/H»gh LUC-
CARB (2018)/GTAP-BIOAEZ-EF/Mean LUC-
BEIOM (2021J/USA Avg .-
Xu etal (2021 VGTAP-BfO+AEZ-EF/USA avg./Mass/CARB (2014)-
Knoope et al (2018)/Higli -
Chen et al. (2018)/GTAP-B10+A£2-EF/GTAP 2011/High Non-LUC -
GREET (2021 VGTAP-BIO+CCLUB/Energy/Avg Proxy -
GREET (2021 yGTAP-BIQ+CCLUB/MM/Case 8-
CARB (2018>/GTAP-BIQ+AEZ-EF/Low LUC-
Chen el al (2018J/GTAP-B IO+AEZ-EF/GTAP 2011/Central Non-LUC -
RFS2 rule (2010VFASOM-FAPR I/Mean LUC-
Knoope el al (2018)/Average-
Xw et al (2021VGTAP-BIO+CCLUB/USA avg ,/Mass/CCLUB -
GREET (2021 )/GTAP-B!OCCLU Brassy Case 8 -
Chen et al (2018}/GTAP-BlO+CCLUB/GTAP 2011 /Central Non-LUC *

GREET (2021 )/GTAP-BIO+CCLUB/Mas&/GTAP 2011 -
Knoope el al <2018)/Low-
Chen et al (2018)/GTAP-BICHCCLUB/GTAP 2011/Low Non-LUC -

RFS2 rule (2010)/FASOM.F APR I/Low LUC-

0	20	40	60

Soybean Oil Biodiesel GHG Emissions (gC02e/MJ)

Notes: The name on they-axis for each bar/estimate includes multiple descriptors separated by In order, these descriptors are
the author or other name (e.g., RFS2 rule] and year of the study; the model used to estimate the LUC emissions; the allocation
approach used for soybean meal (e.g., mass, energy); the LUC estimate case (e.g.. Low LUC); and the non-LUC estimate case
(e.g., Low CI). The Upstream stage includes all of the emissions associated with soybean oil production and transport upstream
of the biodiesel production facility. The Conversion stage includes emission associated with fuel production at the biodiesel
production facility. The Downstream stage includes emissions associated with biodiesel transport and non-C02 combustion
emissions. The LUC stage includes emissions from induced land use changes. For studies that do not report disaggregated results,
results are reported as LUC and Non-LUC emissions.

RFS2 (2010) estimated uncertainty in land use change GHG emissions and reported a
relatively wide range of estimates. The only estimate in our review that is outside the range of
estimates from the RFS2 (2010) estimate is from CARB (2018) using CARB's high estimate for
soy biodiesel land use change emissions from CARB (2014). GREET-2021 allows users to



¦ 1

f^4 ]



p7]

^¦1

B



; 1

- QcT)





jgg



13

'ifiTj



m





176

-------
choose from land use change results from three different GTAP-BIO runs, and four different
allocation approaches to account for soybean meal coproduct. We include a range of estimates
from GREET-2021 based on different combinations of these factors. Chen et al. (2018) used
GREET and the GTAP-BIO model to estimate soy biodiesel carbon intensity. This study uses a
range of estimates based on sensitivity analysis on the GREET input parameters as well as
multiple prior land use change GHG estimates based on GTAP-BIO. Knoope et al. (2018) is a
GREET-style LCA study that excludes land use change GHG emissions and reports a range of
estimates based on sensitivity analysis of input parameters. We also include the estimate from
BEIOM (2021) which uses economic input-output methodology (see Chapter 4.2.2.6 for more
information).

Among the estimates in the above figure, upstream emissions range from 7 to 37
gC02e/MJ. These include emissions associated with feedstock production and transport,
including non-LUC market-mediated impacts in the agricultural sector. Upstream emissions
estimates depend on the methodology used to estimate them and the co-product accounting
methods applied to the soybean meal co-product. By default, GREET uses mass allocation for
the meal co-product, but GREET allows users to select market, energy or displacement
approaches. We include the mass, energy and market-based allocation approaches in the figure
above.337 The highest estimate (19 gC02e/MJ) uses the market-based allocation approach, and
the lowest estimate (9 gC02e/MJ) uses the mass-based allocation. Market-based allocation in
GREET produces higher estimates based on the assumed prices for soybean oil and meal. The
lowest overall estimate for upstream emissions is from RFS2 (2010). RFS2 (2010) is the only
study in our review that uses a consequential modeling approach for non-land use change
emissions, whereby the GHG impacts were modeled using economic models. The highest
estimate is from BEIOM which is also unique in its modeling approach. All of the other studies
use an attributional approach to estimates upstream GHG emissions, and most of them apply a
mass-based allocation approach to account for the soybean meal co-product.

We include estimates for biodiesel produced at U.S. facilities that use a transesterification
process. The range of conversion emissions in Figure 4.2.3.4-1 range from -1 to 15 gC02e/MJ.
Most of the fuel production estimates are from 8-12 gC02e/MJ. The RFS2 (2010) estimate is -1
gC02e/MJ based on the assumption that the glycerin co-product from biodiesel production is
burned for thermal process energy displacing the use of petroleum residual oil. Most of the other
studies use energy, market, or mass-based allocation to account for the glycerin co-product
which results in a larger estimate for fuel production GHG emissions.

Similar to corn ethanol, the largest source of variation between soybean oil biodiesel
LCA estimates are the land use change emissions, ranging from 5 to 64 gC02e/MJ in Figure
4.2.3.4-1.338 The highest and lowest land use change GHG estimates are from RFS2 (2010)
based on the upper and lower bounds of the reported 95% confidence interval. The land use
change uncertainty analysis for RFS2 (2010) considered uncertainty in land conversion types and
emissions factors but did not consider uncertainty in economic model parameters. As discussed,

337	Using displacement results in upstream emissions of -17 gC02e/MJ and total LCA emissions of 38 gC02e/MJ.
Although the inclusion of the displacement method provides interesting variation in the estimates for each stage the
overall estimate is near the middle of the range, thus we exclude it from Figure 4.2.3.3 to improve legibility.

338	This excludes Knoope et al. (2021) and BEIOM (2021) which exclude land use change emissions.

177

-------
in Chapter 4.2.2.8, above, the range of soybean oil biodiesel land use change GHG estimates in
the literature is wider when we consider studies that only estimate land use change emissions,
ranging from 5 to 80 gC02e/MJ.

Particular characteristics of soybean oil biodiesel production introduce greater potential
for uncertainty relative to corn ethanol. For example, the quantity of biofuel that can be produced
from an acre of U.S. soybeans is substantially smaller than that produced from an acre of US
corn. Based on data from USDA and GREET, an average acre of U.S. farmland yields about four
times as much corn ethanol as soybean oil biodiesel on an energy basis.339 This difference in per-
acre fuel yields means that soybean biodiesel modeling results are far more sensitive to tradeoffs
between cropland extensification and other means of obtaining additional soybean oil. For
example, every acre of cropland extensification projected in a given soybean oil biodiesel
scenario represents four times as much new cropland per megajoule of biodiesel relative to a
corn ethanol scenario.

Another key sensitivity for soybean oil is the impact on livestock markets. While soybean
oil represents about 19 percent, by mass, of the crush product of soybeans, meal represents about
80 percent of that crush. Soybean meal is an important source of protein in livestock feed diets.
Corn ethanol also has a livestock feed co-product in the form of distillers grains. On a weight
basis, one MMBTU of soybean oil biodiesel is associated with approximately 4 times as much
feed coproduct as one MMBTU of corn ethanol.340 Soybean meal and distillers grains are used
differently in feed rations and their nutritional contents are also not identical. However, even this
general comparison demonstrates that the quantity of feed product associated with a given
quantity of soybean oil biodiesel is substantially greater than that associated with the same
quantity of corn ethanol. The impact of biofuel feed coproducts on GHG emissions is highly
complex. As a brief example, to the extent greater production of feed coproducts allows cattle
producers to intensify production, reducing the use of grazing lands, these feed coproducts may
mitigate LUC emissions. Conversely, to the extent that greater availability of feed products
reduces costs for livestock producers, this may lead to increased livestock-related emissions. It is
unclear which of these market dynamics may prove dominant in the future, making the net signal
of livestock emissions highly uncertain. The larger quantities of feed coproducts associated with
the production of soybean oil biodiesel relative to corn ethanol amplify this uncertainty at both
ends of the emissions range, contributing to the wider overall range of GHG impacts we observe
in the literature.

Another key sensitivity present in the literature is uncertainty about the type of land that
will be converted in response to increased soybean oil demand, either to increase production of

339	Assumes soybean yield of 3,084 lbs/acre and corn yield of 9,912 lbs/acre, based on 2021 average yields from
USDA NASS. Assumes 93.6 lbs of soybean oil per MMBTU of biodiesel output and 163.7 lbs of corn per MMBTU
of ethanol based on GREET-2021.Thus, one acre yields 6.26 MMBTU (128.6 gallons ethanol-equivalent) of
soybean oil biodiesel, or 60.6 MMBTU (506.2 gallons of ethanol) of corn ethanol. USDA NASS data from
QuickStats database, https://quickstats.nass.usda.gov/ (accessed August 3rd, 2022).

340	For every lb of soybean oil produced, approximately 5.26 lbs of soybean meal are produced. At GREET-2021
average biodiesel yields, production of the one MMBTU of soybean oil biodiesel is associated with the production
of approximately 251.4 lbs of soybean meal. For comparison, according to GREET-2021, production of one
MMBTU of corn ethanol using the aforementioned dry mill ethanol process coproduces about 60.4 lbs of dried
distillers grains (assuming 100 percent drying).

178

-------
soybeans or through market-mediated impacts on other vegetable oils. Soybean oil is part of the
larger global market for vegetable oils and is a good substitute for palm oil in many use cases.
Given the potential for marginal palm oil extensification into carbon-rich peat lands in Southeast
Asia,341 the extent to which palm oil backfills for soybean oil diverted to biofuel production from
other uses also substantially impacts LUC emissions. In addition, increased demand for soybean
oil in the U.S. could lead to land conversion in other large soybean-producing countries such as
Argentina and Brazil that have carbon-dense forests and grasslands. Therefore, the potential for
these types of impacts on sensitive high-carbon lands creates additional uncertainty in soybean
oil biodiesel GHG modeling.

Downstream emissions range from 1 to 4 gC02e/MJ. Downstream emissions are
associated with fuel distribution from biodiesel production facilities to retail gasoline stations
and tailpipe emissions. The highest downstream estimates are from GREET and lowest are from
RFS2 (2010).

Overall, our literature review of estimates representative of average U.S. soybean oil
biodiesel production provides a range from 14 to 73 gC02e/MJ. The largest source of variation
across studies continues to be estimated emissions associated with direct and indirect land use
change. Although this is a wide range, biodiesel produced under particular conditions may
produce emissions that are outside of this range on a per MJ basis. For example, LCA emissions
may be higher if economic conditions result in soybean oil used for biodiesel to be backfilled
with palm oil or soybeans grown in tropical regions with high rates of deforestation. It may also
be possible to produce soybean oil biodiesel with lower LCA emissions with the adoption of
climate smart agricultural practices. For example, planting cover crops between soybean
rotations has the potential to build soil organic carbon stocks. Collecting data on and evaluating
these trends in soybean production and vegetable oil markets are areas for additional research
that will inform future LCA estimates for soybean oil biodiesel.

4.2.3.5 Soybean Oil Renewable Diesel

Relative to soybean oil biodiesel, there have been fewer studies published on the GHG
emissions associated with soybean oil renewable diesel. Lifecycle GHG estimates for soybean
oil renewable diesel Our literature review includes 5 sources that estimate the GHG emissions
associated with soybean oil renewable diesel. These studies include numerous estimates based on
different scenarios for land use change and assumptions related to co-product accounting. The
figure below includes 13 LCA estimates from 5 studies.

341 See for example: Austin, K. G., et al. (2017). "Shifting patterns of oil palm driven deforestation in Indonesia and
implications for zero-deforestation commitments." Land Use Policy 69: 41-48.

179

-------
Figure 4.2.3.5-1: Soybean Oil Renewable Diesel Lifecycle Greenhouse Gas Estimates







J 54







1 "|S3







1 (52]







B 50)

Stage



¦

LUC



Non-LUC



Downstream



Conversion



Upslream

RFS2 rale {2010)/FASOM-FAPRI/CA-LCFS Avg /High LUC-
CARB (2022VGTAP-BIO+AEZ-EF/Mass/H'igh LUC/H»gh Ch
CARB {20t8)/GTAP-BIO»AEZ-EF/Mass/High LUC -
RFS2 rule (2Q10)/FASOM-FAPRI/CA-LCFS Avg /Mean LUC-
Xu et al (2021KGTAP-BIO+AEZ-EF/USA avg /Mass/CARB (2014)-
CARB (2Q18)/GTAP-BIQ+AEZ-EF/Mass/Mean LUC-
GREET (2021 )/GTAP-BICHCCLUB/Mkt./Avg Proxy-
Xu el af. <2021 VGTAP-BtG+AEZ-EF/USA avg,/Mass/ICAO (2019)-
CARB (2022)/GTAP- BIO+AEZ-E F/Mass/Low LUC/Low CI -
CARB (2018)/GTAP-B10+AEZ-EF/Mass/ L ow LUC-
Xu et al (2021VGTAP-BIO+CCLUB/USA avg /Mass/CCLUB-
GREET (2021 J/GTAP-BlO+CCLUB/Mass/Case 8 -
GREET (2021 J/GTAP-BIO+CCLUB/Mass/GTAP 2011 -

RFS2 rule (2010J/FASOM-FAPRI/CA-LCFS Avg ./Low LUC - 	

0	25	50	75

Soybean Oil Renewable Diesel GHG Emissions (gC02e/MJ)

Notes: The name on the y-axis for each bar/estimate includes multiple descriptors separated by 7". In order, these descriptors are
the author or other name (e.g., RFS2 rule) and year of the study; the model used to estimate the LUC emissions; the allocation
approach used for soybean meal (e.g., mass, energy); the LUC estimate case (e.g., Low LUC); and the non-LUC estimate case
(e.g., Low CI). The Upstream stage includes all of the emissions associated with soybean oil production and transport upstream
of the renewable diesel production facility. The Conversion stage includes emission associated with fuel production at the
renewable diesel production facility. The Downstream stage includes emissions associated with renewable diesel transport and
non-C02 combustion emissions. The LUC stage includes emissions from induced land use changes. For studies that do not report
disaggregated results, results are reported as LUC and Non-LUC emissions.

The estimates from RFS2 (2010) in the figure above are based on the "upstream" GHG
modeling for soybean oil from the 2010 RFS2 rule combined with estimates that EPA published
more recently for renewable diesel production and downstream fuel distribution and use.342
Similar to the review for soybean oil biodiesel, CARB (2018) provides a range of land use
change GHG estimates and GREET (2021) includes multiple land use change scenarios and co-
product allocation approaches for soybean meal. Given the relative scarcity of LCA estimates for
soybean oil renewable diesel, we also include the highest and lowest carbon intensities for
individual U.S. facilities as certified by CARB for the CA-LCFS (CARB 2022) using their
central land use change GHG estimates.

342 April 2022 Canola Oil Pathways NPRM (87 FR 22823): https://www.govinfo.gov/content/pkg/FR-2022-04-
18/pdf/2022-07598.pdf

180

-------
Among the estimates in the above figure, upstream emissions range from 8 to 20
gC02e/MJ. Upstream emissions vary for the same reasons discussed for soybean oil biodiesel,
including the methodology used to estimate them and the co-product accounting methods applied
to the soybean meal co-product. The lowest upstream emissions estimates are from RFS2
(2010).343 The highest upstream emissions are from GREET using a market-based allocation
approach to account for the meal co-product. Market-based allocation in GREET produces
higher estimates for soybean oil than mass- or energy-based allocation based on the assumed
prices for soybean oil and meal. Using mass-based allocation, CARB (2018) estimates higher
emissions (14 gC02e/MJ) than GREET-2021 (9 gC02e/MJ). The lowest upstream estimates are
from RFS2 (2010). As discussed above for soybean oil biodiesel, the RFS2 (2010) analysis uses
an entirely different methodology than GREET or CARB for estimating GHG emissions
associated with feedstock production.

Our review includes estimates representative of U.S. renewable diesel production via a
hydrotreating process. The range of conversion emissions in Figure 4.2.3.5-1 range from 8 to 19
gC02e/MJ. This range does not include the facility-specific carbon intensities from CARB
(2022) as this source does not report carbon intensity disaggregated into lifecycle stages. The
lowest estimate is from CARB (2018) and the highest estimate is from GREET-2021 using
market-based allocation. The RFS2 (2010) estimates in the figure above use hydrotreating
processing data provided by CARB representing the average of renewable diesel production
facilities registered for the CA-LCFS as of June 20 21.344 The estimate uses an energy allocation
approach to account for co-products of renewable diesel production. The lowest estimates come
from CARB (2018) based on the default assumptions in CA-GREET version 3.0.

Similar to soybean oil biodiesel, the largest source of variation between soybean oil
biodiesel LCA estimates are the land use change emissions. The same factors, discussed above,
that introduce additional complexity into LUC modeling for soybean oil biodiesel also apply to
soybean oil renewable diesel. For renewable diesel the land use change GHG estimates range
from 6 to 67 gC02e/MJ in Figure 4.2.3.4-1.345 The highest and lowest land use change GHG
estimates are from RFS2 (2010) based on the upper and lower bounds of the reported 95%
confidence interval. The land use change uncertainty analysis for RFS2 (2010) considered
uncertainty in land conversion types and emissions factors but did not consider uncertainty in
economic model parameters. As discussed, in Chapter 4.2.2.8, above, the range of soybean oil
biodiesel land use change GHG estimates in the literature is wider when we consider studies that
only estimate land use change emissions, ranging from 5 to 80 gC02e/MJ.

Downstream emissions are associated with fuel distribution from renewable diesel
production facilities to retail gasoline stations and tailpipe emissions. For renewable diesel, there
is little variation in the review estimates, as they range from 0.5 to 1 gC02e/MJ.

343	The renewable diesel upstream emissions from RFS2(2010) are lower than those for soybean oil biodiesel,
because we have updated the soybean oil upstream estimates for renewable diesel using more recent emissions
factors from GREET and AR5 GWP values. More details are provided in a technical memo to the docket titled
"Notes on Literature Review of Transportation Fuel Greenhouse Gas (GHG) Lifecycle Analysis (LCA)."

344	For more information on hydrotreating process data evaluated by EPA, see April 2022 Canola Oil Pathways
NPRM (87 FR 22823), Section II.C.9.

345	This excludes Knoope et al. (2021) and BEIOM (2021) which exclude land use change emissions.

181

-------
Overall, our literature review of estimates representative of average U.S. soybean oil
renewable production provides a range from 26 to 87 gCChe/MJ. This is a relatively wide range,
and the largest source of variation between studies continues to be estimated emissions
associated with direct and indirect land use change. Although this is a wide range, renewable
diesel produced under particular conditions may produce emissions that are outside of this range
on a per MJ basis. Some of the main factors that could result in emissions higher or lower than
the literature range are the same as those discussed above for soybean oil biodiesel. Renewable
diesel production requires a relatively large amount of hydrogen. Renewable diesel carbon
intensities could be reduced by consuming less hydrogen or sourcing the hydrogen from low
carbon sources.

It is worth noting that the International Civil Aviation Organization (ICAO) has been
conducting similar lifecycle GHG analysis in support of the Carbon Offsetting and Reduction
Scheme for International Aviation (CORSIA). Given ICAO's focus on jet fuel, they have not
specifically released an LCA value for soybean oil renewable diesel. However, the lifecycle
analysis for soybean oil jet fuel is almost identical to an LCA for soybean oil renewable diesel
since both are produced through the same hydrotreating process. While most current
hydrotreating processes yield renewable diesel with small amounts of naptha and LPG co-
products, these facilities can be configured to produce a separate jet fuel stream from the rest of
the products produced. Producing jet fuel requires additional refining, therefore jet fuel LCA
estimates tend to be slightly more GHG intensive than producing renewable diesel alone. The
soybean oil jet fuel results from ICAO (2021) are summarized in the table below.

Table 4.2.3.5-1: U.S. Soybean Oil Jet Fuel Estimates from ICAO (2021) (gCChe/MJ)

Estimate

Core346

Land Use Change

LCA Value

GLOBIOM LUC

40

14

54

(Low end of 95% CI)







GTAP-BIO LUC

40

20

60

ICAO Default

40

25

65

GLOBIOM LUC

40

50

91

GLOBIOM LUC

40

92

132

(High end of 95% CI)







Notes: For their default land use change estimate, ICAO uses the GTAP-BIO estimate plus 4.45 gC02e/MJ, see ICAO (2021) p.
149 for explanation. The GTAP-BIO and central GLOBIOM land use change estimates are from ICAO (2021) Table 67. The low
and high GLOBIOM estimates are from ICAO (2021) Table 72. The low estimate is the 2.5% quantile and the high estimate is
the 97.5% quantile from sensitivity analysis (300 runs). LCA values in table might not be the sum or core and LUC values due to
rounding.

Given the similarities in the hydrotreating process, it is a relatively straightforward
adjustment to modify a soybean oil jet fuel LCA to a soybean oil renewable diesel LCA.347 If the
ICAO soybean oil jet fuel LCA was adjusted for the lower energy needs for renewable diesel, the

346	ICAO (2021) includes estimates from GREET of the direct, or "core," GHG emissions associated withjet fuel
produced from soybean oil through a hydrotreating process.

347	ICAO's core GHG estimates are based on analysis with GREET using an energy allocation approach for co-
products. The GREET-2021 core GHG estimate for soybean oil renewable diesel using energy allocation is 36
gC02e/MJ. This GREET-2021 estimate can be substituted for the 40 gC02e/MJ core jet fuel value to produce an
LCA range for soybean oil renewable diesel.

182

-------
ICAO estimates for soybean oil renewable diesel would be 50 to 128 gC02e/MJ. If the ICAO
LCA values were included in Figure 4.2.3.5-1 and Table 4.2.3.12, the overall range of values for
soybean oil renewable diesel would be wider (26 gCOze/MJ to 128 gCChe/MJ).

4.2.3.6 FOG Biodiesel

We reviewed literature on the GHG emissions associated with biodiesel produced from
fats, oils and greases (FOG). Specifically, we reviewed estimates for biodiesel produced from
used cooking oil (UCO) and animal tallow. Figure 4.2.3.6-1 includes the LCA estimates for
IJCO biodiesel, and Figure 4.2.3.6-2 includes the LCA estimates for animal tallow biodiesel.

Figure 4.2.3.6-1: UCO Biodiesel Lifecycle Greenhouse Gas Estimates

CARB (2022)/Higriest C! -

RFS2 (2010) + O'Malley (2021)-

CARB (2018)-

Xuelal (2021VUSA avg ¦

RFS2 rule (2010) -

Stage

LUC

Non-LUC
Downstream
Conversion
Upstream

CARB (2022)/LoweSl C< -

0

a	10	20	30

UCO Biodiesel GHG Emissions (gC02e/MJ)

Notes: The Upstream stage includes all of the emissions associated with UCO pre treatment/rendering and transport upstream of
the biodiesel production facility. O'Malley et al. (2021) also includes indirect GHG emissions in the Upstream stage. The
Conversion stage includes: emission associated with fuel production at the biodiesel production facility. The Downstream stage
includes emissions associated with biodiesel transport and non-C02 combustion emissions. Estimates that only report total
lifecycle GHG emissions are depicted only with a label and no bars.

Estimates for UCO biodiesel range from 12 to 32 gC02e/MJ. Given the relative scarcity
of LCA studies on UCO biodiesel, we include the highest and lowest certified carbon intensities
for individual biodiesel facilities under the CA-LCFS in our review. The highest and lowest
estimates come from the CA-LCFS range (CARB 2022). The CARB and RFS2 estimates assume

183

-------
that the only upstream emissions for supplying UCO are associated with rendering/cooking the
raw UCO and transporting it to biodiesel production facilities. O'Malley et al. (2021) looked at
the current uses for UCO apart from biofuel production and evaluated a case study where UCO
diverted from livestock feed and oleochemical uses is backfilled with corn, soybean oil and palm
oil. Based on this case study, they estimated potential indirect emissions of 12.2 gC02e/MJ
associated with UCO use for biodiesel. In the figure above, we include the potential indirect
emissions estimate from O'Malley et al. (2021) added to the RFS2 (2010) estimates.

Figure 4.2.3.6-2: Animal Tallow Biodiesel Lifecycle Greenhouse Gas Estimates

GREET (2021) + O'Malley et al <2021)/Energy-
CARB (2018)-
CARB (2022)/Higbest CI -
Chen et al (2018)/Htgh-
CARB (2022)/Lowest CI -
GREET (2021 VEnergy -
GREET (2021 yMarket -
Chen et al (2018VNBB (2016)/Central-
Xu et al (2021 )/USA avg -

GREET (2021)/Displacement-

0	20	40	60

Tallow Biodiesel GHG Emissions (gC02e/MJ)

Notes: The Upstream stage includes all of the emissions associated with animal tallow pre-treatment/rendering and transport
upstream of the biodiesel production facility. The Conversion stage includes emission associated with fuel production at the
biodiesel production facility. The Downstream stage includes emissions associated with biodiesel transport and non-C02
combustion emissions. For studies that do not report disaggregated results, results arc reported as LUC and Non-LUC emissions.
Estimates that only report total lifecycle GHG emissions are depicted only with a label and no bars.

Estimates for tallow biodiesel range from 15 to 58 gC02e/MJ. Most of the estimates
assume that tallow is a byproduct of meat production and assume zero upstream emissions from
livestock production allocated to the tallow. For these estimates the ranges are primarily based
on different assumptions about the energy requirements for rendering, as well as different
assumptions about the co-products from rendering and the accounting methods for these co-
products. The exception is the case study by O'Malley et al. (2021) which estimates emissions of
34.8 gC02e/MJ associated with backfilling tallow used in livestock feed and oleochemical

184

-------
production with corn, soybean oil and palm oil. To inform our range of estimates we add this
backfill emissions estimate to the estimates from GREET-2021.

4.2.3.7 FOG Renewable Diesel

We reviewed literature on the GHG emissions associated with renewable diesel produced
from FOG. Specifically, we reviewed estimates for renewable diesel produced from used
cooking oil (UCO) and animal tallow. Figure 4.2.3.7-1 includes the LCA estimates for UCO
biodiesel, and Figure 4.2.3.7-2 includes the LCA estimates for animal tallow biodiesel.

Figure 4.2.3.7-1: UCO Renewable Diesel Lifecycle Greenhouse Gas Estimates

RFS2 (2010) ~ O'Malley (2021J/CA-LCFS Avg •

I®

CAR8 (2022)/Highest CI -

@

RFS2 rule (2010)/CA-LCFS Avg -

RFS2 rule (201 OyPearlson et al Energy -

CARB (2018)-

CARB (2022VLowest CI -

Xuetal (2021 )/USA avg -

0

Stage

LUC

Non-LUC
Downstream
Conversion
Upstream

Seber et al (2014)/Pear1son et al (2013)/Low -

0	10	20	30

UCO Renewable Diesel GHG Emissions (gC02e/MJ)

Notes: The Upstream stage includes all of the emissions associated with UCO pre-treatment/rendering and transport upstream of
the biodiesel production facility. O'Malley et al. (2021) also includes indirect GHG emissions in the Upstream stage. The
Conversion stage includes emission associated with fuel production at the biodiesel production facility. The Downstream stage
includes emissions associated with biodiesel transport and non-C02 combustion emissions. Estimates that only report total
lifecycle GHG emissions are depicted only with a label and no bars.

Estimates for UCO renewable diesel range from 12 to 37 gC02e/MJ. The CARB and
RFS2 estimates assume that the only upstream emissions for supplying UCO are associated with
rendering/cooking the raw UCO and transporting it to biodiesel production facilities. O'Malley
et al. (2021) looked at the current uses for UCO apart from biofuel production and evaluated a
case study where UCO diverted from livestock feed and oleochemical uses is backfilled with

185

-------
corn, soybean oil and palm oil. Based on this case study, they estimated potential indirect
emissions of 12.2 gC02e/MJ associated with UCO use for biodiesel. In the figure above, the
highest estimate is based on the sum of the potential indirect emissions estimate from O'Malley
et al. (2021) added to the RFS2 (2010) estimate. The lowest estimates are from Seber et al.
(2014), which do not include any backfilling emissions.

^¦80





J 63 )









Stage

Figure 4.2.3.7-2: Animal Tallow Renewable Diesel Lifecycle Greenhouse Gas Estimates

Seber el al. (2014)ff»earlson et al {20t3)/CopFOducl High -
Seber el al (2014)/Pearlson el al (2013)/Coproducl Base -
Seber et al {2014VPear>son et al. (2013}/Coprotluct Low-
GREET (2021) + O'Malley et al (2021 VEnergy -
Riazi {2020)/Aspen Pius/Poullry High/Mass -
CARB (2022)/Highest CI -
CARS (2018)-
Riazl (2020VA$pen Plus/Beef Highu'Mass -
Seber et al (2014VPearlson el al. (2013)/Byproducl High -
R(azi (2020)/Aspen Plus/Poultry Low/Mass-
Seber et al. (2014yPearlson el al (2013)/Byproducl Base -
R»az» (202Q)/Aspen Plus/Beef Lew/Mass -
GREET (2021 )/GREET-202VEnergy -
GREET (2021 )/GREET-2021/Market -
Sebef et al (2014VPear1son et al (201 SJ/Byproctuct Low -
CARB (2022VLowesl CI -
GREET (2021 yGREET-20217Mass -
Xu et al (202iyUSA avg -
GREET (2021VGREET-2021/Displacemenl -

LUC

Non-LUC
Downstream
Conversion
Upslrearn

60

@0

Tallow Renewable Diesef GHG Emissions (gC02e/MJ)

Notes: The Upstream stage includes all of the emissions associated with tallow pre-treatment/rendering and transport upstream of
the biodiesel production facility. Seber et al. (2021) also includes livestock production GHG emissions in the Upstream stage.
The Conversion stage includes emission associated with fuel production at the biodiesel production facility. The Downstream
stage includes emissions associated with biodiesel transport and non-CO2 combustion emissions. Estimates that only report total
lifecycle GHG emissions are depicted only with a label and no bars.

Estimates for tallow renewable diesel range from 14 to 80 gC02e/MJ. The highest
estimates are from Seber et al. (2014). Seber et al. (2014) is the only study that includes
scenarios where GHG emission associated with livestock raising and meat production are
allocated the tallow. In other words, in these scenarios the tallow is considered a co-product of
meat production rather than a byproduct. Our review also includes a case study by O'Malley et
al. (2021) that evaluates emission associated with backfilling tallow used in livestock feed and
oleochemical production with corn, soybean oil and palm oil. The lowest estimate is from
GREET-2021 using a displacement approach for the co-products from tallow rendering.

186

-------
4.2.3.8 Distillers Corn Oil Biodiesel

We reviewed published estimates of the GHG emissions associated with biodiesel
produced from distillers corn oil (DCO). DCO is a co-product from ethanol production whereby
oil is removed the DGS before it is sold as livestock feed. The DCO can then be used as a biofuel
feedstock or added back into livestock feed at desired levels. Figure 4.2.3.8-1 includes the LCA
estimates for DCO biodiesel.

Figure 4.2.3.8-1: DCO Biodiesel Lifecycle Greenhouse Gas Estimates

CARB (2022Wishes! CI-

0

RFS2 rule (2020)-

CARB (2018>-

CARB (2022VLowesl CI -

Stage

LUC

Non-LUC
Downstream
Conversion
Upstream

Xuetal {202iyUSAavg ¦

GREET (20211-

10	20

DCO Biodiesel GHG Emissions (gC02e/MJ)

Notes: The Upstream stage includes all of the emissions associated with DCO extraction and in some cases backfilling with corn
in livestock feed. The Conversion stage includes emission associated with fuel production at the biodiesel production facility.
The Downstream stage includes emissions associated with biodiesel transport and non-CO2 combustion emissions. Estimates that
only report total lifecycle GHG emissions are depicted only with a label and no bars.

DCO biodiesel estimates range from 10 to 37 gC02e/MJ. Most of the estimates assume
that DCO is a byproduct and assign none of the emissions associated with corn or ethanol
production to it. For the 2020 RFS2 rule, we estimated the emissions associated with corn
backfilling for DCO in animal feed. As discussed in that rule and a prior rule on distillers
sorghum oil, we determined that DCO is used as a source of energy/calories in feed diets and that
corn is a likely product to backfill when DCO is used as a biofuel feedstock. The lowest estimate
is from GREET-2021.

187

-------
4.2.3.9 Distillers Corn Oil Renewable Diesel

We reviewed published estimates of the GHG emissions associated with renewable diesel
produced from DCO. Figure 4.2.3.9-1 includes the LCA estimates for DCO renewable diesel.

Figure 4.2.3.9-1: DCO Renewable Diesel Lifecycle Greenhouse Gas Estimates

CARB (2Q22)/Highest CI -

0

RFS2 rule (2020)-

CARB (2018)-

CARB (2022VLowest CI -

0

Stage

LUC

Non-LUC
Downstream
Conversion
Upstream

GREET (2021)-

Xuelal (2021J/USA avg -

0	10	20	30	40

DCO Renewable Diesel GHG Emissions (gC02e/MJ)

Notes: The Upstream stage includes all of the emissions associated with DCO extraction and in some cases backfilling with corn
in livestock feed. The Conversion stage includes emission associated with fuel production at the renewable diesel production
facility. The Downstream stage includes emissions associated with biodiesel transport and non-C02 combustion emissions.
Estimates that only report total lifecycle GHG emissions are depicted only with a label and no bars.

DCO renewable diesel estimates range from 12 to 46 gC02e/MJ. Most of the estimates
assume that DCO is a byproduct and assign none of the emissions associated with corn or
ethanol production to it. For the 2020 RFS2 rule, we estimated the emissions associated with
corn backfilling for DCO in animal feed. The lowest estimate is from GREET-2021.

4.2.3.10 Natural Gas CNG

As discussed above for petroleum gasoline and diesel, for the purposes of conducting the
lifecycle GHG emissions analysis and determining which biofuels meet the GHG requirements,
CAA Section 211 (o) (1) (C) defines baseline lifecycle greenhouse gas emissions as "the average

188

-------
lifecycle greenhouse gas emissions, as determined by the Administrator, after notice and
opportunity for comment, for gasoline or diesel (whichever is being replaced by the renewable
fuel) sold or distributed as transportation fuel in 2005." While the baseline lifecycle GHG
emissions are used for a different specific purpose under the RFS program, we are not required
to use it here in this analysis for evaluating the GHG impacts of the candidate volumes.

To inform a range of potential GHG impacts associated with renewable CNG we
consider two scenarios for the conventional fuels it displaces. In the first scenario we assume that
the candidate volumes of renewable CNG, relative to the No RFS baseline, cause some miles
traveled with diesel vehicles to be replaced with miles traveled with vehicles that run on
renewable CNG. This scenario assumes that the candidate volumes make CNG vehicles more
economically attractive than diesel vehicles in some cases, leading to a marginal increase in
CNG vehicle miles traveled relative to diesel vehicle miles traveled. In the second scenario, we
assume the candidate volumes of renewable CNG do not shift the relative miles traveled for
diesel vehicles relative to CNG vehicles, but instead cause CNG vehicles to be fueled with
renewable CNG instead of conventional CNG.

Thus, our literature review for this proposed rule includes studies that estimate the
lifecycle GHG emissions associated with natural gas CNG. Figure 4.2.3.9-1 includes the natural
gas CNG from our review of the literature. Based on our review, LCA estimates for diesel are
higher than those for natural gas CNG on a per MJ of fuel basis. For the illustrative 30-year
GHG scenario discussed in the next section (Chapter 4.2.4), the scenario where renewable CNG
replaces diesel fuel produces a high estimate of the GHG benefits of renewable CNG. The low
estimate of renewable CNG GHG benefits is based on the scenario that assumes renewable CNG
displaces conventional CNG.

189

-------
Figure 4.2.3.10-1: Natural Gas CNG Well-to-Wheel Greenhouse Gas Estimates

CARB (2022)/Highest CI -	[sT|

CARB (2022)/Lowest CI -

( 79 ]

L—J Stage

GREET (2021 )/Default CH4 Leak-

GREET (2021 /'EPA" CH4 Leak -

EPA and NHTSA(2016)-

Upstream

Downstream

Conversion

Non-LUC

LUC

0

20

40

60

80

Natural Gas CNG GHG Emissions (gC02e/MJ)

Notes:: The name on the y-axis for each bar/estimate includes multiple descriptors separated by "t". In order, these descriptors are
the author or other name (e.g., RFS2 rule) and a brief descriptor of the scenario modeled. The Upstream stage includes all of the,
emissions associated with extracting, processing and delivering natural gas to a compression facility. The Conversion stage
includes emissions associated compressing natural gas to CNG. The Downstream stage includes emissions associated with
fueling a CNG vehicle and tailpipe combustion emissions. The gasoline baseline estimate in the March 2010 RFS2 rule used
SAR GWP values, All values in this chart use 100-year AR5 GWP values, Estimates that only report total lifecycle GHG
emissions are depicted only with a label and no bars.

The natural gas CNG estimates in our review range from 72 to 81 gCChe/MJ. Our review
did not identify many applicable studies, as there are many more studies on the lifecycle
emissions associated with natural gas production than natural gas for CNG vehicles. The EPA
and NHTSA (2016) estimate is from the R1A for the Greenhouse Gas Emissions and Fuel
Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles - Phase 2 rule. It
represents lifecycle emissions for CNG used in a 2014 or later dedicated CNG vehicle. The
lowest estimates are from GREET-2021, By default, GREET-2021 assumes a set of assumptions
for methane leakage during natural gas production. The model also gives users the option of
choosing methane leakage assumptions derived from the EPA GHG Inventory. The highest
estimates are the natural gas CNG pathways certified under the CA-LCFS program. CNG
produced from natural gas that is in turn produced from wells or systems with high leakage rates
may have much greater carbon intensity than the estimates in our review. This is an area where
additional LCA research would be helpful.

190

-------
4.2.3.11 Landfill Biogas CNG

Our literature review did not identify many studies on the lifecycle GHG emissions
associated with CNG produced from landfill gas (LFG). Our review is limited to estimates
derived from the GREET model and estimates by CARB as part of their implementation of the
CA-LCFS program. Figure 4.2.3.11-1 includes the LCA estimates for CNG produced from
landfill biogas.

Figure 4.2.3.11-1: Landfill Biogas CNG Lifecycle Greenhouse Gas Estimates

CARB (2018)/1% Leak/NG/Offsite-

CARB (2022)/Highest Cl/Offsite -

CARB (2022)/Lowest Cl/Offsite -

GREET (2021 V3% Leak/RNG/Offsite -

GREET (2021 )/2% Leak/RNG/Offsite -

GREET (2021 )/2% Leak/RNG/Onsite -

GREET (2021 )/1% Leak/RNG/Offsite -

Q

Stage

-40	0	40

LFG CNG GHG Emissions (gCQ2e/MJ)

80

Non-LUC
Downstream
Conversion
Upstream

Notes: The Upstream stage includes all of the emissions associated with capturing LFG. processing it to pipeline quality, and
transporting it to the fueling location. Downstream emission include tailpipe emissions including C02 emissions. In most studies
upstream emissions are negative because they assume LFG will beflared in the counterfactual baseline scenario. Because
reductions in C02 emissions are included in the Upstream emissions, C02 tailpipe emissions are included in the Downstream
emissions. CARB (2018) reports more disaggregated results and excludes tailpipe C02 emissions. Estimates that only report total
lifecycle GF1G emissions are depicted only with a label and no bars.

The range of estimates in Figure 4.2.3.11-1 range from 6 to 70 gC02e/MJ. The higher
estimates are from CARB and the lower estimates are from GREET-2021. We varied two
parameters in GREET-2021 to provide a range of estimates. By default, GREET-2021 assumes a
2% methane leakage rate associated with processing landfill gas to pipeline quality. We include
estimates assuming 1% and 3% methane leakage and show that each 1% of additional methane
leakage increases the LCA estimates by 6 gC02e/MJ. By default, GREET-2021 assumes CNG
fueling occurs offsite from where the landfill gas is produced, and the majority of CA-LCFS
certified LFG CNG pathways are for offsite CNG fueling. Based on GREET-2021 onsite fueling
reduces the LCA estimate by approximately 3 gC02e./MJ. The highest estimates are from CARB,

191

-------
with the very highest estimate coming from the default CA-GREET version 3.0 model. GREET -
2021 includes negative upstream emissions based on reduced GHG emissions in a baseline
scenario absent the use of the LFG as fuel. The CA-GREET model has much larger upstream
GHG emissions than GREET-2021 as it does not include the large emissions reductions relative
to the baseline. Landfills with high leakage rates associated with capturing and cleaning up the
biogas may have much greater carbon intensity than the estimates in our review. This is an area
where additional I.CA research would be helpful.

4.2.3.12 Manure Digester CNG

Our literature review did not identify many studies on the lifecycle GHG emissions
associated with CNG produced from manure digester biogas. Our review is limited to estimates
derived from the GREET model and estimates by CARB as part of their implementation of the
CA-LCFS program. Figure 4.2.3.12-1 includes the LCA estimates for CNG produced from
landfill biogas.

Figure 4.2.3.12-1: Manure Digester CNG Lifecycle Greenhouse Gas Estimates

GREET (2021 )/BroiIer/Offsite -

GREET (2021 )/Layers/Offsite -

GREET (2021 VBeef/Offsite-

GREET (2021 )/Beef Heifer/Offsite -

GREET (2021 yDairy/Offslte -

CARB (2022)/Highest CI/Dairy-

Stage

Q

LUC

Non-LUC
Downstream
Conversion
Upstream

GREET (2021 )/Swine/0(fsite -

-256



CARB (2022(/Lowest CI/Dairy-[ -533 )

-400	-200	0

Manure Biogas CNG GHG Emissions (gC02e/MJ)

Notes: The Upstream stage includes all of the emissions associated with capturing LFG. processing it to pipeline quality, and
transporting it to the fueling location. Downstream emission include tailpipe emissions including C02 emissions. In most studies
upstream emissions are negative because they assume LFG will be flared in the counteffactual baseline scenario. Because
reductions in C02 emissions are included in the Upstream emissions, C02 tailpipe emissions are included in the Downstream
emissions. CARB (2018) reports more disaggregated results and excludes tailpipe C02 emissions. Estimates that only report total
lifecycle GF1G emissions are depicted only with a label and no bars.

192

-------
There are relatively few studies but a very large range of LCA estimates (-533 to 44
gCChe/MJ) for biogas CNG produced from manure digesters. CARB has certified pathways for
CNG produced from over 50 different sources of manure biogas. All of these pathways have
negative carbon intensities meaning they reduce GHG emissions even before displacing any
conventional transportation fuels. The negative emissions are due to the assumed high methane
and nitrous oxide emissions in the baseline scenario absent the collection and treatment of animal
manure in anaerobic digesters. Consequently, the biggest area of uncertainty in the LCA for
manure digester GHG emissions is what level of control is assumed in the baseline. Based on
estimates from GREET-2021, CNG produced from poultry manure biogas has positive carbon
intensities, as high as 44 gC02e/MJ.

4.2.3.13 Summary of LCA Ranges

Based on the literature review for each pathway discussed above, the range of LCA
estimates are summarized in Table 4.2.3.12-1.

Table 4.2.3.13-1: Lifecycle GHG Ranges Based on Literature Review (gCChe/MJ)

Pathway

LCA Range

Petroleum Gasoline

84 to 98

Petroleum Diesel

84 to 94

Corn Starch Ethanol

38 to 116

Soybean Oil Biodiesel

14 to 73

Soybean Oil Renewable Diesel

26 to 87

Used Cooking Oil Biodiesel

12 to 32

Used Cooking Oil Renewable Diesel

12 to 37

Tallow Biodiesel

15 to 58

Tallow Renewable Diesel

14 to 81

Distillers Corn Oil Biodiesel

10 to 37

Distillers Corn Oil Renewable Diesel

12 to 46

Natural Gas CNG

72 to 81

Landfill Gas CNG

9 to 70

Manure Biogas CNG

-533 to 44

In the sections that follow we present a range of monetized climate benefits associated
with the candidate volumes for an illustrative 30-year scenario. In order to appropriately
monetize GHG impacts over this period an annual stream of net GHG emissions is required. For
the non-crop based fuel pathways we assume a constant stream of GHG emissions per MJ over
the 30-year period. The land use change emissions associated with crop-based biofuels are highly
dynamic, as the majority of emission increases associated with land use changes occur relatively
quickly (e.g., in the first few years) with the reduced emissions associated with the biofuel use
occuring over time. Thus, for the 30-year illustrative scenario, we use estimates for crop-based
biofuels that report an annual stream of land use change emissions. The majority of the land use
change GHG estimates in the literature do not report an annual stream. In many cases, these LUC
estimates are derived by estimating land conversions induced by the crop-based biofuel
production and then multiplying these conversions by emissions factors that estimate the

193

-------
resulting total emissions over a 20-30 year period. The only study identified in our review that
does report an annual stream of land use change emissions is the analysis for the 2010 RFS2 rule.
The reasons that no other studies report annual emissions are not entirely clear, but many studies
use static models to estimate land use change that are not conducive to reporting annual
emissions. Other studies use models that have the capability to estimate an annual stream but did
not report them for reasons that were not discussed in the publication. Thus, for the illustrative
GHG scenario we use the highest and lowest LCA estimates from the 2010 RFS2 rule for the
crop-based biofuel pathways. The LCA ranges used for the illustrative 30-year scenario are
summarized in the following table. The results for the 30-year scenario are described in the
following section.

Table 4.2.3.13-2: Lifecycle GHG Ranges for Illustrative 30-Year Scenario (gCChe/MJ)

Pathway

LCA Range

Petroleum Gasoline

84 to 98

Petroleum Diesel

84 to 94

Corn Starch Ethanol

49 to 91

Soybean Oil Biodiesel

14 to 72

Soybean Oil Renewable Diesel

24 to 78

Used Cooking Oil Biodiesel

12 to 32

Used Cooking Oil Renewable Diesel

12 to 37

Tallow Biodiesel

15 to 44

Tallow Renewable Diesel

14 to 81

Distillers Corn Oil Biodiesel

10 to 37

Distillers Corn Oil Renewable Diesel

12 to 46

Natural Gas CNG

72 to 90

Landfill Gas CNG

6 to 70

Manure Biogas CNG

-533 to 44

4.2.3.14 References for LCA Literature Review

For ease of reference, the following is the list of references cited in Chapter 4.2.3, unless
otherwise cited with the footnote:

•	BEIOM. (2021). Avelino, A. F. T., et al. "Creating a harmonized time series of
environmentally-extended input-output tables to assess the evolution of the US
bioeconomy - A retrospective analysis of corn ethanol and soybean biodiesel." Journal of
Cleaner Production 321: 128890.

•	Brandao, M. (2022). "Indirect Effects Negate Global Climate Change Mitigation
Potential of Substituting Gasoline With Corn Ethanol as a Transportation Fuel in the
USA." Frontiers in Climate 4.

•	CARB (2014). Detailed Analysis for Indirect Land Use Change. California Air Resources
Board. Sacramento, CA. 113 pages

•	CARB (2018). CA-GREET3.0 Model. Sacramento, CA, California Air Resources Board.

194

-------
•	CARB (2022). CA-LCFS Current Pathways Certified Carbon Intensities. California Air
Resources Board. Sacramento, CA. https://ww2.arb.ca.gov/resources/documents/lcfs-
pathway-certified-carbon-intensities

•	Carriquiry, M., et al. (2019). "Incorporating Sub-National Brazilian Agricultural
Production and Land-Use into U.S. Biofuel Policy Evaluation." Applied Economic
Perspectives and Policy.

•	Chen, R., et al. (2018). "Life cycle energy and greenhouse gas emission effects of
biodiesel in the United States with induced land use change impacts." Bioresource
Technology 251: 249-258.

•	Cooney, G., et al. (2017). "Updating the U.S. Life Cycle GHG Petroleum Baseline to
2014 with Projections to 2040 Using Open-Source Engineering-Based Models."
Environmental Science & Technology 51(2): 977-987.

•	EPA (2010). Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis.
U.S. Environmental Protection Agency, Office of Transportation Air Quality.
Washington, DC. EPA-420-R-10-006.

•	EPA (2020). Renewable Fuel Standard Program: Standards for 2020 and Biomass-Based
Diesel Volume for 2021 and Other Changes. 85 FR 7016.

•	EPA and NHSTA (2016). Greenhouse Gas Emissions and Fuel Efficiency Standards for
Medium- and Heavy-Duty Engines and Vehicles - Phase 2: Regulatory Impact Analysis.

•	EPA (2014). July 2014 Pathways II final rule. 79 FR 42128.

•	GREET (2021). Wang, Michael, Elgowainy, Amgad, Lee, Uisung, Bafana, Adarsh,
Banerjee, Sudhanya, Benavides, Pahola T., Bobba, Pallavi, Burnham, Andrew, Cai, Hao,
Gracida, Ulises, Hawkins, Troy R., Iyer, Rakesh K., Kelly, Jarod C., Kim, Taemin,
Kingsbury, Kathryn, Kwon, Hoyoung, Li, Yuan, Liu, Xinyu, Lu, Zifeng, Ou, Longwen,
Siddique, Nazib, Sun, Pingping, Vyawahare, Pradeep, Winjobi, Olumide, Wu, May, Xu,
Hui, Yoo, Eunji, Zaimes, George G., and Zang, Guiyan. Greenhouse gases, Regulated
Emissions, and Energy use in Technologies Model ® (2021 Excel). Computer Software.
USDOE Office of Energy Efficiency and Renewable Energy (EERE). 11 Oct. 2021.
Web. doi: 10.11578/GREET-Excel-202l/dc.20210902.1.

•	ICAO (2021). CORSIA Eligible Fuels — Lifecycle Assessment Methodology. CORSIA
Supporting Document. Version 3: 155.

•	Knoope, M. M. J., et al. (2019). "Analysing the water and greenhouse gas effects of soya
bean-based biodiesel in five different regions." Global Change Biology Bioenergy 11 (2):
381-399.

•	Laborde, D., et al. (2014). Progress in estimates of ILUC with MIRAGE model. JRC
Scientific and Policy Reports. Italy, European Commission Joint Research Centre
Institute for Energy and Transport: 46.

•	Lark, T. J., et al. (2022). "Environmental outcomes of the US Renewable Fuel Standard."
Proceedings of the National Academy of Sciences 119(9)

•	Lee, U., et al. (2021). "Retrospective analysis of the US corn ethanol industry for 2005—
2019: implications for greenhouse gas emission reductions." Biofuels, Bioproducts and
Biorefining.

•	Lewandrowski, J., et al. (2019). "The greenhouse gas benefits of corn ethanol - assessing
recent evidence." Biofuels: 1-15.

195

-------
•	O'Malley, J., et al. (2021). Indirect Emissions from Waste and Residue Feedstocks: 10
Case Studies from the United States, The International Council of Clean Transportation:
49.

•	Plevin, R. J., et al. (2015). "Carbon Accounting and Economic Model Uncertainty of
Emissions from Biofuels-Induced Land Use Change." Environmental Science &
Technology 49(5): 2656-2664.

•	Plevin, R. J., et al. (2022). "Choices in land representation materially affect modeled
biofuel carbon intensity estimates." Journal of Cleaner Production: 131477.

•	Riazi, B., et al. (2020). "Renewable diesel from oils and animal fat waste: implications of
feedstock, technology, co-products and ILUC on life cycle GWP." Resources,
Conservation and Recycling 161: 104944.

•	Scully, M. J., et al. (2021). "Carbon intensity of corn ethanol in the United States: state of
the science." Environmental Research Letters.

•	Seber, G., et al. (2014). "Environmental and economic assessment of producing
hydroprocessed jet and diesel fuel from waste oils and tallow." Biomass and Bioenergy
67: 108-118.

•	Taheripour, F., et al. (2017). "The impact of considering land intensification and updated
data on biofuels land use change and emissions estimates." Biotechnology for Biofuels
10(1): 191.

•	Xu, H., et al. (2022). "Life Cycle Greenhouse Gas Emissions of Biodiesel and Renewable
Diesel Production in the United States." Environmental Science & Technology 56(12):
7512-7521.

4.2.4 GHG Results for Illustrative Scenario

For each of the 2023, 2024, and 2025 standards, we estimate a 30-year stream of changes
in GHG emissions for renewable fuel volumes above the No RFS baseline for each analyzed fuel
using the carbon intensity analyses discussed above. While the standards proposed 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.

Table 4.2.4-2 summarizes the annual low biofuel emission estimates and high petroleum
baseline emission estimates in grams CChe per megajoule (Table 4.2.4-4 does the same for high
biofuel and low petroleum baseline estimates). Table 4.2.4-3 presents the high petroleum
subtracted from the low biofuel emission estimates to show net emissions from displacing
petroleum fuels with biofuels on a per fuel-equivalent megajoule basis (Table 4.2.4-5 does the
same for high biofuel and low petroleum baseline estimates). GHG benefits from biofuels
displacing fossil fuel use include the GHG emissions associated with biofuel production and use,
including land use change emissions relative to the baseline scenario.

Emissions streams based on the 2023 through 2025 standards are presented in Tables
4.2.4-6 through 4.2.4-8 for low biofuel/high petroleum lifecycle analysis estimates, and Tables
4.2.4-10 through 4.2.4-12 for high biofuel/low petroleum lifecycle analysis estimates
respectively, both compared to the No RFS baseline. These are derived by first converting the
net emission streams presented in Tables 4.2.4-3 and 4.2.4-5 from grams CChe per megajoule to

196

-------
million metric tons CChe per megajoule, then multiplying these streams of emissions factors by
changes in renewable fuel volumes. The volume changes in 2023 reflect the difference between
the target volumes and the No RFS baseline as presented in Table 4.2.4-1. As discussed in
Section 4.2.3.1, our GHG analysis of the 2023 standard assumes that the target volumes will
produce GHG benefits for the subsequent 29 years due to ongoing use of renewable fuels (and
their consequent displacement of fossil fuels). In analyzing the GHG impacts of the 2024
standard we only consider the difference in volumes between the 2024 standard and 2023
standard because the emissions benefits of the increase in use of renewable fuels to meet the
2023 standards are already accounted for in the 29 years following 2023 (i.e., 2023-2053). Thus,
we only attribute emissions to the 2024 standard for the volumes that have changed compared to
the previous year (2023). Similarly, for 2024, we only include the emission impact for volumes
that have changed from the 2024 levels. These resulting annual sequences of emissions for the
2023 through 2025 standards are then summed, resulting in a combined stream of estimated
annual emissions from 2023 through 2054. These are presented in Tables 4.2.4-9 and 4.2.4-13
respectively below.

Table 4.2.4-

: Volume Changes Used for Illustrative GHG Scenario





Landfill
Biogas
CNG/
LNG348

Agricultural
Digester
Biogas
CNG/LNG

Wood

Waste/

MSW

Diesel/Jet

Fuel

Soybean/

Canola

Oil

Biodiesel

Fats/Oils/

Greases

Biodiesel

Corn Oil
Biodiesel

Soybean Oil

Renewable

Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil

Renewable

Diesel

Corn

Starch

Ethanol

Volume
Changes
Relative to No
RFS Baseline
[Table 3.2-3]
(million
gallons)

2023

44

44

-

968

200

120

901

275

80

706

2024

91

91

3

935

200

120

1,048

329

86

776

2025

145

145

6

901

200

120

1,054

388

91

840

Volume
Changes
Relative to
Previous Year
(million
gallons)

2023

44

44

-

968

200

120

901

275

80

706

2024

48

48

3

(33)

-

-

147

53

6

70

2025

54

54

3

(33)

-

-

6

60

6

65

Table 4.2.4-5 shows positive net GHG emissions for the corn ethanol and soybean oil
renewable diesel and biodiesel volumes in 2023 due to the initial pulse of land use change
emissions in the estimates used for this illustrative scenario. For corn ethanol, volumes relative to
the No RFS baseline increase in years 2024 and 2025, therefore there are positive emissions in
all three years. Conversely as shown in Table 4.2.4-6, soybean oil renewable diesel and biodiesel
volumes are negative in 2024 because those volumes decrease relative to the previous year's
(2023) volume increases from the No RFS baseline. As noted above, this scenario assumes that
the biofuel production continues for 30 years, irrespective of volume mandates in future years.

We separately estimate a 30-year stream of changes in GHG emissions for renewable fuel
volumes from the supplemental volume requirement proposed in this rulemaking as described in

348 Table 3.2-3 presents total volume changes for CNG/LNG from biogas. We assume for purposes of this
illustrative GHG scenario that half of that biogas is sourced from landfills and half from agricultural digesters.

197

-------
Chapter 3.3. As shown in Table 3.3-1, the supplemental volume requirement of 250 million
ethanol-equivalent gallons is represented by an energy-equivalent 147 million soybean oil
renewable diesel gallons in 20 2 3.349 Table 4.2.4-14 shows the illustrative GHG scenario using
the same process described above for both low biofuel/high petroleum and high biofuel/low
petroleum lifecycle analysis estimates for the supplemental volumes.

349 See Chapter 3.3 for more details.

198

-------
Table 4.2.4-2: Gross low biofuel/high petroleum annual lifecycle analysis estimates for



Landfill
Biogas
CNG/
LNG

Ag-
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel351

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean

Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch

Ethanol

Gasoline

(High
Estimate)

Diesel
(High
Estimate)

Year 0

8.7

(532.7)

37.6

710.3

13.0

9.9

755.9

12.8

12.4

395.0

98.1

94.1

Year 1

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 2

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 3

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 4

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 5

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 6

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 7

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 8

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 9

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 10

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 11

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 12

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 13

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 14

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 15

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 16

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 17

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 18

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 19

8.7

(532.7)

37.6

(20.0)

13.0

9.9

(8.7)

12.8

12.4

38.9

98.1

94.1

Year 20

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

Year 21

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

Year 22

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

Year 23

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

Year 24

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

Year 25

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

Year 26

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

Year 27

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

Year 28

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

Year 29

8.7

(532.7)

37.6

7.5

13.0

9.9

19.9

12.8

12.4

34.5

98.1

94.1

35° Parentheses indicate a reduction in GHG emissions.

351 Wood waste/MSW diesel and jet fuels are comprised of a wide variety of feedstocks and represent a small
volume of fuels in this rule. We have made a simplifying assumption that these fuels meet a 60% GHG reduction
(equal to the cellulosic threshold) compared to the diesel GHG estimate shown in this table.

199

-------
Table 4.2.4-3: Net low biofuel/high petroleum (low biofuel minus high petroleum baseline)
annual lifecycle analysis estimates for individual biofuels, presented in grams CChe per
mega joule of fuel.352 								



Landfill

Biogas
CNG/LNG

Agricultural
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch
Ethanol

Year 0

(85.4)

(626.8)

(56.5)

616.2

(81.1)

(84.2)

661.8

(81.3)

(81.7)

296.9

Year 1

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 2

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 3

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 4

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 5

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 6

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 7

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 8

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 9

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 10

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 11

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 12

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 13

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 14

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 15

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 16

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 17

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 18

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 19

(85.4)

(626.8)

(56.5)

(114.1)

(81.1)

(84.2)

(102.8)

(81.3)

(81.7)

(59.2)

Year 20

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

Year 21

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

Year 22

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

Year 23

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

Year 24

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

Year 25

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

Year 26

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

Year 27

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

Year 28

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

Year 29

(85.4)

(626.8)

(56.5)

(86.6)

(81.1)

(84.2)

(74.2)

(81.3)

(81.7)

(63.6)

352 Parentheses indicate a net reduction in GHG emissions.

200

-------
Table 4.2.4-4: Gross high biofuel/low petroleum annual lifecycle analysis estimates for

individual biofuels, presented in grams CChe per mega joule of fuel.



Landfill
Biogas
CNG/
LNG

Ag-
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet

Fuel 353

Soybean/
Canola

Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean

Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch

Ethanol

Gasoline

(Low
Estimate)

Diesel
(Low
Estimate)

Natural
Gas
(Low
Estimate)

Year 0

69.8

44.0

33.4

1,044.0

42.3

36.6

1,102.7

54.2

46.3

665.1

83.6

83.5

71.9

Year 1

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 2

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 3

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 4

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 5

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 6

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 7

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 8

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 9

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 10

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 11

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 12

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 13

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 14

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 15

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 16

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 17

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 18

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 19

69.8

44.0

33.4

54.2

42.3

36.6

68.4

54.2

46.3

79.8

83.6

83.5

71.9

Year 20

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

Year 21

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

Year 22

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

Year 23

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

Year 24

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

Year 25

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

Year 26

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

Year 27

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

Year 28

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

Year 29

69.8

44.0

33.4

8.4

42.3

36.6

20.9

54.2

46.3

53.6

83.6

83.5

71.9

353 Wood waste/MSW diesel and jet fuels are comprised of a wide variety of feedstocks and represent a small
volume of fuels in this rule. We have made a simplifying assumption that these fuels meet a 60% GHG reduction
(equal to the cellulosic threshold) compared to the diesel GHG estimate shown in this table.

201

-------
Table 4.2.4-5: Net high biofuel/low petroleum (high biofuel minus low petroleum baseline)
annual lifecycle analysis estimates for individual biofuels, presented in grams CChe per
mega joule of fuel.354 								



Landfill

Biogas
CNG/LNG

Agricultural
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch
Ethanol

Year 0

(2.1)

(27.8)

(50.1)

960.5

(41.2)

(46.9)

1,019.2

(29.4)

(37.2)

581.5

Year 1

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 2

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 3

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 4

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 5

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 6

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 7

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 8

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 9

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 10

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 11

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 12

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 13

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 14

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 15

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 16

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 17

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 18

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 19

(2.1)

(27.8)

(50.1)

(29.3)

(41.2)

(46.9)

(15.1)

(29.4)

(37.2)

(3.8)

Year 20

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

Year 21

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

Year 22

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

Year 23

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

Year 24

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

Year 25

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

Year 26

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

Year 27

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

Year 28

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

Year 29

(2.1)

(27.8)

(50.1)

(75.1)

(41.2)

(46.9)

(62.6)

(29.4)

(37.2)

(30.0)

354 Parentheses indicate a net reduction in GHG emissions.

202

-------
Table 4.2.4-6: 30-year stream of emissions for 2023 standards using low biofuel/high
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS



Landfill
Biogas
CNG/LNG

Agricultural
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch
Ethanol

2023

(0.3)

(2.2)

-

75.2

(2.0)

(1.3)

77.3

(2.9)

(0.8)

16.9

2024

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2025

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2026

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2027

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2028

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2029

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2030

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2031

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2032

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2033

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2034

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2035

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2036

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2037

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2038

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2039

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2040

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2041

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2042

(0.3)

(2.2)

-

(13.9)

(2.0)

(1.3)

(12.0)

(2.9)

(0.8)

(3.4)

2043

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2044

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2045

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2046

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2047

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2048

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2049

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2050

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2051

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2052

(0.3)

(2.2)

-

(10.6)

(2.0)

(1.3)

(8.7)

(2.9)

(0.8)

(3.6)

2053

-

-

-

-

-

-

-

-

-

-

2054

-

-

-

-

-

-

-

-

-

-

355 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.

203

-------
Table 4.2.4-7: 30-year stream of emissions for 2024 standards using low biofuel/high
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS



Landfill
Biogas
CNG/LNG

Agricultural
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch
Ethanol

2023

-

-

-

-

-

-

-

-

-

-

2024

(0.3)

(2.4)

(0.0)

(2.6)

-

-

12.6

(0.6)

(0.1)

1.7

2025

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2026

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2027

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2028

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2029

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2030

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2031

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2032

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2033

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2034

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2035

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2036

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2037

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2038

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2039

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2040

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2041

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2042

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2043

(0.3)

(2.4)

(0.0)

0.5

-

-

(2.0)

(0.6)

(0.1)

(0.3)

2044

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2045

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2046

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2047

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2048

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2049

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2050

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2051

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2052

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2053

(0.3)

(2.4)

(0.0)

0.4

-

-

(1.4)

(0.6)

(0.1)

(0.4)

2054

-

-

-

-

-

-

-

-

-

-

356 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.

204

-------
Table 4.2.4-8: 30-year stream of emissions for 2025 standards using low biofuel/high
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS



Landfill
Biogas
CNG/LNG

Agricultural
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch
Ethanol

2023

-

-

-

-

-

-

-

-

-

-

2024

-

-

-

-

-

-

-

-

-

-

2025

(0.4)

(2.7)

(0.0)

(2.6)

-

-

0.5

(0.6)

(0.1)

1.5

2026

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2027

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2028

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2029

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2030

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2031

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2032

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2033

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2034

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2035

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2036

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2037

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2038

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2039

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2040

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2041

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2042

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2043

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2044

(0.4)

(2.7)

(0.0)

0.5

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2045

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2046

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2047

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2048

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2049

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2050

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2051

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2052

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2053

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

2054

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

357 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.

205

-------
Table 4.2.4-9: 30-year stream of emissions for combined 2023-2025 standards using low
biofuel/high petroleum lifecycle analysis estimates for individual biofuels, relative to the No
RFS baseline, presented in millions of metric tons C02e.358				



Landfill
Biogas
CNG/LNG

Agricultural
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch
Ethanol

2023

(0.3)

(2.2)

-

75.2

(2.0)

(1.3)

77.3

(2.9)

(0.8)

16.9

2024

(0.6)

(4.6)

(0.0)

(16.5)

(2.0)

(1.3)

0.6

(3.5)

(0.9)

(1.7)

2025

(1.0)

(7.3)

(0.0)

(16.0)

(2.0)

(1.3)

(13.5)

(4.1)

(1.0)

(2.2)

2026

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2027

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2028

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2029

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2030

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2031

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2032

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2033

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2034

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2035

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2036

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2037

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2038

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2039

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2040

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2041

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2042

(1.0)

(7.3)

(0.0)

(13.0)

(2.0)

(1.3)

(14.0)

(4.1)

(1.0)

(4.0)

2043

(1.0)

(7.3)

(0.0)

(9.6)

(2.0)

(1.3)

(10.7)

(4.1)

(1.0)

(4.3)

2044

(1.0)

(7.3)

(0.0)

(9.7)

(2.0)

(1.3)

(10.2)

(4.1)

(1.0)

(4.3)

2045

(1.0)

(7.3)

(0.0)

(9.8)

(2.0)

(1.3)

(10.1)

(4.1)

(1.0)

(4.3)

2046

(1.0)

(7.3)

(0.0)

(9.8)

(2.0)

(1.3)

(10.1)

(4.1)

(1.0)

(4.3)

2047

(1.0)

(7.3)

(0.0)

(9.8)

(2.0)

(1.3)

(10.1)

(4.1)

(1.0)

(4.3)

2048

(1.0)

(7.3)

(0.0)

(9.8)

(2.0)

(1.3)

(10.1)

(4.1)

(1.0)

(4.3)

2049

(1.0)

(7.3)

(0.0)

(9.8)

(2.0)

(1.3)

(10.1)

(4.1)

(1.0)

(4.3)

2050

(1.0)

(7.3)

(0.0)

(9.8)

(2.0)

(1.3)

(10.1)

(4.1)

(1.0)

(4.3)

2051

(1.0)

(7.3)

(0.0)

(9.8)

(2.0)

(1.3)

(10.1)

(4.1)

(1.0)

(4.3)

2052

(1.0)

(7.3)

(0.0)

(9.8)

(2.0)

(1.3)

(10.1)

(4.1)

(1.0)

(4.3)

2053

(0.7)

(5.1)

(0.0)

0.7

-

-

(1.5)

(1.2)

(0.1)

(0.7)

2054

(0.4)

(2.7)

(0.0)

0.4

-

-

(0.1)

(0.6)

(0.1)

(0.3)

358 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.

206

-------
Table 4.2.4-10: 30-year stream of emissions for 2023 standards using high biofuel/low
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS
baseline, presented in millions of metric tons C02e.359

2023

2024

2025

2026

2027

2028

2029

2030

2031

2032

2033

2034

2035

2036

2037

2038

2039

2040

2041

2042

2043

2044

2045

2046

2047

2048

2049

2050

2051

2052

2053

2054

Landfill
Biogas
CNG/LNG

(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)
(0-0)

(0.0)

Agricultural
Digester
Biogas
CNG/LNG

(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)
(0-1)

(0.1)

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

117.3
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(3.6)
(9-2)
(9-2)
(9.2)
(9.2)
(9.2)
(9.2)
(9.2)
(9.2)
(9.2)
(9-2)

Fats/Oils/
Greases
Biodiesel

1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)

1.0)

Corn Oil
Biodiesel

(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0-7)
(0.7)

Soybean Oil
Renewable
Diesel

119.0

(7.3)
(7.3)
(7.3)
(7.3)
(7.3)
(7.3)
(7.3)
(7.3)
(7.3)
(7.3)

Fats/Oils/
Greases
Renewable
Diesel

1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)
1.0)

1.0)

Corn Oil
Renewable
Diesel

(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0-4)
(0.4)

359 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.

207

-------
Table 4.2.4-11: 30-year stream of emissions for 2024 standards using high biofuel/low
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS

360



Landfill
Biogas
CNG/LNG

Agricultural
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch
Ethanol

2023

-

-

-

-

-

-

-

-

-

-

2024

(0.0)

(0.1)

(0.0)

(4.0)

-

-

19.4

(0.2)

(0.0)

3.3

2025

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2026

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2027

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2028

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2029

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2030

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2031

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2032

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2033

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2034

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2035

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2036

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2037

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2038

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2039

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2040

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2041

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2042

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2043

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.3)

(0.2)

(0.0)

(0.0)

2044

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2045

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2046

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2047

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2048

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2049

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2050

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2051

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2052

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2053

(0.0)

(0.1)

(0.0)

0.3

-

-

(1.2)

(0.2)

(0.0)

(0.2)

2054

-

-

-

-

-

-

-

-

-

-

360 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.

208

-------
Table 4.2.4-12: 30-year stream of emissions for 2025 standards using high biofuel/low
petroleum lifecycle analysis estimates for individual biofuels, relative to the No RFS



Landfill
Biogas
CNG/LNG

Agricultural
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch
Ethanol

2023

-

-

-

-

-

-

-

-

-

-

2024

-

-

-

-

-

-

-

-

-

-

2025

(0.0)

(0.1)

(0.0)

(4.0)

-

-

0.8

(0.2)

(0.0)

3.0

2026

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2027

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2028

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2029

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2030

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2031

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2032

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2033

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2034

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2035

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2036

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2037

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2038

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2039

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2040

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2041

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2042

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2043

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2044

(0.0)

(0.1)

(0.0)

0.1

-

-

(0.0)

(0.2)

(0.0)

(0.0)

2045

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

2046

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

2047

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

2048

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

2049

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

2050

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

2051

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

2052

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

2053

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

2054

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

361 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.

209

-------
Table 4.2.4-13: 30-year stream of emissions for combined 2023-2025 standards using high
biofuel/low petroleum lifecycle analysis estimates for individual biofuels, relative to the No
RFS baseline, presented in millions of metric tons C02e.362				



Landfill
Biogas
CNG/LNG

Agricultural
Digester
Biogas
CNG/LNG

Wood
Waste/
MSW
Diesel/Jet
Fuel

Soybean/
Canola Oil
Biodiesel

Fats/Oils/
Greases
Biodiesel

Corn Oil
Biodiesel

Soybean Oil
Renewable
Diesel

Fats/Oils/
Greases
Renewable
Diesel

Corn Oil
Renewable
Diesel

Corn
Starch
Ethanol

2023

(0.0)

(0.1)

-

117.3

(1.0)

(0.7)

119.0

(1.0)

(0.4)

33.1

2024

(0.0)

(0.2)

(0.0)

(7.6)

(1.0)

(0.7)

17.7

(1.3)

(0.4)

3.1

2025

(0.0)

(0.3)

(0.0)

(7.5)

(1.0)

(0.7)

(1.3)

(1.5)

(0.4)

2.8

2026

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2027

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2028

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2029

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2030

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2031

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2032

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2033

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2034

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2035

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2036

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2037

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2038

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2039

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2040

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2041

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2042

(0.0)

(0.3)

(0.0)

(3.3)

(1.0)

(0.7)

(2.1)

(1.5)

(0.4)

(0.3)

2043

(0.0)

(0.3)

(0.0)

(8.9)

(1.0)

(0.7)

(7.6)

(1.5)

(0.4)

(1.7)

2044

(0.0)

(0.3)

(0.0)

(8.7)

(1.0)

(0.7)

(8.5)

(1.5)

(0.4)

(1.9)

2045

(0.0)

(0.3)

(0.0)

(8.5)

(1.0)

(0.7)

(8.6)

(1.5)

(0.4)

(2.0)

2046

(0.0)

(0.3)

(0.0)

(8.5)

(1.0)

(0.7)

(8.6)

(1.5)

(0.4)

(2.0)

2047

(0.0)

(0.3)

(0.0)

(8.5)

(1.0)

(0.7)

(8.6)

(1.5)

(0.4)

(2.0)

2048

(0.0)

(0.3)

(0.0)

(8.5)

(1.0)

(0.7)

(8.6)

(1.5)

(0.4)

(2.0)

2049

(0.0)

(0.3)

(0.0)

(8.5)

(1.0)

(0.7)

(8.6)

(1.5)

(0.4)

(2.0)

2050

(0.0)

(0.3)

(0.0)

(8.5)

(1.0)

(0.7)

(8.6)

(1.5)

(0.4)

(2.0)

2051

(0.0)

(0.3)

(0.0)

(8.5)

(1.0)

(0.7)

(8.6)

(1.5)

(0.4)

(2.0)

2052

(0.0)

(0.3)

(0.0)

(8.5)

(1.0)

(0.7)

(8.6)

(1.5)

(0.4)

(2.0)

2053

(0.0)

(0.2)

(0.0)

0.6

-

-

(1.2)

(0.4)

(0.1)

(0.3)

2054

(0.0)

(0.1)

(0.0)

0.3

-

-

(0.0)

(0.2)

(0.0)

(0.2)

362 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.

210

-------
Table 4.2.4-14: 30-year stream of lifecycle analysis estimates for volume changes from 2023
supplemental volume requirement, represented by soybean renewable diesel compared to

363



Low Biofuel/High Petroleum

High Biofuel/Low Petroleum

2023

12.6

19.4

2024

(2.0)

(0.3)

2025

(2.0)

(0.3)

2026

(2.0)

(0.3)

2027

(2.0)

(0.3)

2028

(2.0)

(0.3)

2029

(2.0)

(0.3)

2030

(2.0)

(0.3)

2031

(2.0)

(0.3)

2032

(2.0)

(0.3)

2033

(2.0)

(0.3)

2034

(2.0)

(0.3)

2035

(2.0)

(0.3)

2036

(2.0)

(0.3)

2037

(2.0)

(0.3)

2038

(2.0)

(0.3)

2039

(2.0)

(0.3)

2040

(2.0)

(0.3)

2041

(2.0)

(0.3)

2042

(2.0)

(0.3)

2043

(1.4)

(1.2)

2044

(1.4)

(1.2)

2045

(1.4)

(1.2)

2046

(1.4)

(1.2)

2047

(1.4)

(1.2)

2048

(1.4)

(1.2)

2049

(1.4)

(1.2)

2050

(1.4)

(1.2)

2051

(1.4)

(1.2)

2052

(1.4)

(1.2)

2053

-

-

2054

-

-

363 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.

211

-------
4.2.5 Monetized GHG Impacts

4.2.5.1 Social Cost of Greenhouse Gases

For assessing GHG impacts in this illustrative scenario, we rely upon past biofuel
emissions reductions estimates that are available in CChe—carbon equivalent emissions using
the global warming potentials utilized in those analyses.364 We estimate the social benefits of
GHG reductions in this illustrative scenario using estimates of the social cost of greenhouse
gases (SC-GHG)365, 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 (hereinafter the "February 2021 TSD").366 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 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-CO2 estimates in
benefit-cost analyses conducted under Executive Order (E.O.) 128 6 6 367 in 2008, following a
court ruling in which an agency was ordered to consider the value of reducing CO2 emissions in a

364	It would be preferable to use estimates for each gas (e.g., CO2, CH4, N2O), but we use C02e estimates for this
illustrative scenario as they are the most readily available biofuel carbon intensity estimates.

365	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-N2O)), but collectively they are referenced as the social cost

of greenhouse gases (SC-GHG).

366	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.whitehouse.gov/briefing-room/b1og/2021/02/26/a-return-to-

science-evidence- based-estimates-of-the-benefits-of-reducing-climate-pollu Hon/.

367	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."

212

-------
rulemaking process. The SC-CO2 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 JAM.368,369,370 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.371 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 that 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.

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

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

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

371	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.

213

-------
As noted above, EPA participated in the IWG but has also independently evaluated the interim
SC-CO2 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 that accounts for the
myriad of ways that global climate change reduces the net welfare of U.S. populations does not
currently exist in the literature. As explained in the February 2021 TSD, existing estimates are
both incomplete and an underestimate of total damages that accrue to the citizens and residents
of the U.S. because they do not fully capture the regional interactions and spillovers discussed
above, nor do they include all of the important physical, ecological, and economic impacts of
climate change recognized in the climate change literature, as discussed further below. EPA, as a
member of the IWG, will continue to review developments in the literature, including more
robust methodologies for estimating the magnitude of the various damages to U.S. populations
from climate impacts and reciprocal international mitigation activities, and explore ways to
better inform the public of the full range of carbon impacts.

Second, the IWG concluded that the use of the social rate of return on capital (7 percent
under current OMB Circular A-4 guidance) to discount the future benefits of reducing GHG
emissions inappropriately underestimates the impacts of climate change for the purposes of

214

-------
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.372,373,374,375,376 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 an
appropriate value to apply to the social cost of greenhouse gases 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

372	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.

373	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.

374	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.

375	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.

376	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/addenduni to sc-
ghg tsd august 2016.pdf (accessed February 5, 2021).

215

-------
on how SC-GHG estimates can "be combined in RIAs with other cost and benefits estimates that
may use different discount rates." The National Academies reviewed "several options," including
"presenting all discount rate combinations of other costs and benefits with [SC-GHG] estimates."

While the IWG works to assess how best to incorporate the latest, peer reviewed science
to develop an updated set of SC-GHG estimates, it recommended the interim estimates to be the
most recent estimates developed by the IWG prior to the group being disbanded in 2017. The
estimates rely on the same models and harmonized inputs and are calculated using a range of
discount rates. As explained in the February 2021 TSD, the IWG has concluded that it is
appropriate for agencies to revert to the same set of four values drawn from the SC-GHG
distributions based on three discount rates as were used in regulatory analyses between 2010 and
2016 and subject to public comment. For each discount rate, the IWG combined the distributions
across models and socioeconomic emissions scenarios (applying equal weight to each) and then
selected a set of four values for use in agency analyses: an average value resulting from the
model runs for each of three discount rates (2.5 percent, 3 percent, and 5 percent), plus a fourth
value, selected as the 95th percentile of estimates based on a 3 percent discount rate. The fourth
value was included to 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 4.2.5.1-1 summarizes the interim SC-CO2 estimates for the years 2023-2054.377
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-
CO2 estimates in analyses, the February 2021 TSD emphasizes the importance of considering all
four of the SC-CO2 values. The SC-CO2 increases over time within the models (i.e., the societal
harm from one metric ton emitted in 2030 is higher than the harm caused by one metric ton
emitted in 2025) because future emissions produce larger incremental damages as physical and
economic systems become more stressed in response to greater climatic change, and because
GDP is growing over time and many damage categories are modeled as proportional to GDP.

377 The February 2021 TSD provides SC-GHG estimates through emissions year 2050. Estimates were extended for
the period 2051 to 2054 using the IWG methods, assumptions, and parameters identical to the 2020-2050 estimates.
Specifically, 2051-2054 SC-GHG estimates were calculated in Mimi.jl, an open-source modular computing platform
used for creating, running, and performing analyses on IAMs (www.iiiiiiiifraniework.org). For CO2, the 2051-2054
SC-GHG values were calculated by linearly interpolating between the 2050 TSD values and the 2055 Mimi-based
values.

216

-------
Table 4.2.5.1-1: Interim Social Cost of Carbon Values, 2023-2054 (2021$/Metric Ton

C02)378

Emissions
Year

Discount Rate and Statistic

5% Average

3% Average

2.5% Average

3% 95th Percentile

2023

$16

$55

$81

$164

2024

$17

$56

$82

$167

2025

$17

$57

$84

$171

2026

$18

00
LO

LO
00

$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

$33

$86

$119

$263

2052

$33

$87

$120

$264

2053

$34

$88

$121

$265

2054

$35

$89

$123

$266

Note: The 2023-2050 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

378 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.whitehouse.gov/briefing-room/blog/2021/02/26/a-return-to-
science-evidence-based-estimates-of-the-benefits-of-reducing-climate-pollution

217

-------
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-regulatorv-affairs/regulatorv-matters/#scghgs.

The estimates were extended for the period 2051 to 2054 using methods, assumptions, and parameters identical to

the 2020-2050 estimates. The values are stated in $/metric ton C02 and vary depending on the year of C02

emissions.

There are a number of limitations and uncertainties associated with the 8C CO2
estimates presented in Table 4.2.5.1-1. Some uncertainties are captured within the analysis, while
other areas of uncertainty have not yet been quantified in a way that can be modeled. Figure
4.2.5.1-1 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.

Figure 4.2.5.1-1: Frequency Distribution of SC-CO2 Estimates for 20303' 9

5% Average = $19

E

U3

Discount Rate
S.0%
D 3.0%

n 2.5%

3%

95th Pet =3181

I

5 - 95" Percentile
of Simulations

rrTrriTiTn 1111 rnmTrn 1»' 1 < 1 > n rnrvn 1 1 rhffTii

60 80 100 120 140 160 180 200 220 240 260 280 300 320

Social Cost of Carbon in 2030 [2018S I metric ton COJ

In addition, the interim SC-C02 estimates presented in Table 4.2.5.1-1 have a number of
other limitations. First, the current scientific and economic understanding of discounting

379 Although the distributions and numbers are based on the full set of model results (150,000 estimates for each
discount rate and gas), for display purposes the horizontal axis is truncated with 0.02 to 0.68 percent of the estimates
falling below the lowest bin displayed and 0.12 to 3.11 percent of the estimates falling above the highest bin
displayed, depending on the discount rate and GHG.

218

-------
approaches suggests discount rates appropriate for intergenerational analysis in the context of
climate change are likely to be less than 3 percent, near 2 percent or lower. Second, the IAMs
used to produce these interim estimates do not include all of the important physical, ecological,
and economic impacts of climate change recognized in the climate change literature and the
science underlying their "damage functions" (i.e., the core parts of the IAMs that map global
mean temperature changes and other physical impacts of climate change into economic (both
market and nonmarket) damages) lags behind the most recent research. For example, limitations
include the incomplete treatment of catastrophic and non-catastrophic impacts in the integrated
assessment models, their incomplete treatment of adaptation and technological change, the
incomplete way in which inter-regional and intersectoral linkages are modeled, uncertainty in the
extrapolation of damages to high temperatures, and inadequate representation of the relationship
between the discount rate and uncertainty in economic growth over long time horizons.

Likewise, the socioeconomic and emissions scenarios used as inputs to the models do not reflect
new information from the last decade of scenario generation or the full range of projections.

The modeling limitations do not all work in the same direction in terms of their influence
on the SC- CO2 estimates. However, as discussed in the February 2021 TSD, the IWG has
recommended that, taken together, the limitations suggest that the SC- CO2 estimates used in this
rule likely underestimate the damages from GHG emissions. In particular, the Intergovernmental
Panel on Climate Change (IPCC) Fourth Assessment Report, which was the most current IPCC
assessment available at the time when the IWG decision over the ECS input was made,
concluded that SC-CO2 estimates "very likely.. .underestimate the damage costs" due to omitted
impacts.380 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

380 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 9169122 4.

219

-------
assessments.381,382'383'384'385'386'387,388 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
limitations suggest that the interim SC-CO2 estimates likely underestimate the damages from
CO2 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.

Tables 4.2.4-6 through Tables 4.2.4-13 show the estimated changes in CChe for the
volume changes analyzed in each year, 2023-2025. This analysis portrays what might be

381	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.

382	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.)].

383	Intergovernmental Panel on Climate Change (IPCC). 2019a. Climate Change and Land: an IPCC special report
on climate change, desertification, land degradation, sustainable land management, food security, and greenhouse
gas fluxes in terrestrial ecosystems [P.R. Shukla, J. Skea, E. Calvo Buendia, V. Masson-Delmotte, H.-O. Portner, D.
C. Roberts, P. Zhai, R. Slade, S. Connors, R. van Diemen, M. Ferrat, E. Haughey, S. Luz, S. Neogi, M. Pathak, J.
Petzold, J. Portugal Pereira, P. Vyas, E. Huntley, K. Kissick, M. Belkacemi, J. Malley, (eds.)].

384	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.)].

385	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.

hl:l:t)s://dx.doi. org/10.7930/I0R49NQX.

386	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.

387	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://d0i.0rg/l 0.17226/21852.

388	National Academies of Sciences, Engineering, and Medicine (National Academies). 2019. Climate Change and
Ecosystems. Washington, DC: The National Academies Press, https://doi.org/10.17226/25504.

220

-------
expected if, in each of the ensuing 29 years, aggregate renewable fuel consumption for each
category exceeded baseline levels by the same volume associated with the rule. EPA estimated
the dollar value of these GHG-related effects for each analysis year between 2023 through 2054
by applying the SC-CO2 estimates, shown in Table 4.2.5.1-1, to the estimated changes in GHG
emissions inventories resulting from the candidate volumes. EPA then calculated the present
value and annualized benefits from the perspective of each year by discounting each year-
specific value to that year using the same discount rate used to calculate the SC-CO2.

4.2.5.2 Results

For this illustrative scenario, the interim estimates for carbon dioxide from the February
2021 TSD, presented in Table 4.2.5.1-1, were used to estimate the social benefits of the
estimated 30-year stream of GHG impacts presented in Chapter 4.2.4. For each year, the total of
emissions changes presented in Tables 4.2.4-6 through 4.2.4-13 are multiplied by each of the
four SC-CO2 values for the same year. Values for each year and discount rate statistic are then
converted to present value using the corresponding discount rates. The resulting streams of
estimated social benefits of the biofuel volume changes assumed in this illustrative scenario for
the 2023-2025 standards are presented in Tables 4.2.5.2-1 through 4.2.5.2-8.389 Note that in these
tables, volume changes for the 2023-2025 standards are relative to the No RFS baseline. We
separately estimate the social benefits of the biofuel volume changes assumed in this illustrative
scenario from the supplemental volume requirement proposed in this rulemaking as described in

389 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" [different 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). As
discussed at length in the February 2021 TSD, these estimates are an underestimate of the benefits of CO2 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 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.

221

-------
Section 3.3. We present the results for these supplemental volumes in Tables 4.2.5.2-9 and
4.2.5.2-10. All calculations are available in a spreadsheet in the docket for this rule.390

Table 4.2.5.2-1: Present value of 30-year stream of climate benefits for 2023 standards,
using low biofuel/high petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2021$)a	

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$(2,778)

$(9,461)

$(14,001)

$(28,257)

2024

$645

$2,216

$3,285

$6,633

2025

$633

$2,193

$3,256

$6,576

2026

$620

$2,170

$3,227

$6,517

2027

$607

$2,146

$3,197

$6,456

2028

$594

$2,121

$3,167

$6,393

2029

$581

$2,096

$3,136

$6,328

2030

$568

$2,071

$3,105

$6,261

2031

$557

$2,048

$3,075

$6,203

2032

$546

$2,025

$3,045

$6,143

2033

$535

$2,001

$3,015

$6,082

2034

$523

$1,977

$2,984

$6,019

2035

$512

$1,952

$2,953

$5,954

2036

$500

$1,927

$2,922

$5,888

2037

$488

$1,902

$2,890

$5,821

2038

$477

$1,877

$2,859

$5,752

2039

$465

$1,852

$2,827

$5,683

2040

$453

$1,827

$2,795

$5,613

2041

$443

$1,801

$2,762

$5,534

2042

$432

$1,776

$2,729

$5,455

2043

$350

$1,451

$2,235

$4,457

2044

$341

$1,430

$2,208

$4,391

2045

$332

$1,408

$2,180

$4,326

2046

$324

$1,387

$2,153

$4,261

2047

$315

$1,366

$2,126

$4,195

2048

$307

$1,345

$2,099

$4,130

2049

$298

$1,324

$2,071

$4,066

2050

$290

$1,303

$2,044

$4,001

2051

$284

$1,273

$2,023

$3,898

2052

$275

$1,250

$1,991

$3,797

2053

$-

$-

$-

$-

2054

$-

$-

$-

$-

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle

390 See "GHG Scenario for 2023-25 Set Rule (NPRM).xlsx," available in the docket for this rule.

222

-------
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
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.

223

-------
Table 4.2.5.2-2: Present value of 30-year stream of climate benefits for 2024 standards,
using low biofuel/high petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2021$)a	

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$-

$-

$-

$-

2024

$69

$237

$352

$710

2025

$50

$173

$257

$520

2026

$49

$171

$255

$515

2027

$48

$170

$253

$510

2028

$47

$168

$250

$505

2029

$46

$166

$248

$500

2030

$45

$164

$245

$495

2031

$44

$162

$243

$490

2032

$43

$160

$241

$485

2033

$42

$158

$238

$481

2034

$41

$156

$236

$476

2035

$40

$154

$233

$470

2036

$40

$152

$231

$465

2037

$39

$150

$228

$460

2038

$38

$148

$226

$455

2039

$37

$146

$223

$449

2040

$36

$144

$221

$444

2041

$35

$142

$218

$437

2042

$34

$140

$216

$431

2043

$33

$138

$213

$425

2044

$34

$142

$220

$437

2045

$33

$140

$217

$431

2046

$32

$138

$214

$424

2047

$31

$136

$212

$418

2048

$31

$134

$209

$411

2049

$30

$132

$206

$405

2050

$29

$130

$204

$398

2051

$28

$127

$201

$388

2052

$27

$124

$198

$378

2053

$27

$122

$195

$368

2054

$-

$-

$-

$-

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th

224

-------
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.

225

-------
Table 4.2.5.2-3: Present value of 30-year stream of climate benefits for 2025 standards,
using low biofuel/high petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2021$)a	

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$-

$-

$-

$-

2024

$-

$-

$-

$-

2025

$67

$233

$346

$698

2026

$56

$196

$292

$590

2027

$55

$194

$289

$584

2028

$54

$192

$287

$579

2029

$53

$190

$284

$573

2030

$51

$187

$281

$567

2031

$50

$185

$278

$561

2032

$49

$183

$276

$556

2033

$48

$181

$273

$551

2034

$47

$179

$270

$545

2035

$46

$177

$267

$539

2036

$45

$174

$264

$533

2037

$44

$172

$262

$527

2038

$43

$170

$259

$521

2039

$42

$168

$256

$514

2040

$41

$165

$253

$508

2041

$40

$163

$250

$501

2042

$39

$161

$247

$494

2043

$38

$158

$244

$487

2044

$37

$156

$241

$479

2045

$37

$159

$246

$487

2046

$36

$156

$243

$480

2047

$35

$154

$239

$473

2048

$35

$152

$236

$465

2049

$34

$149

$233

$458

2050

$33

$147

$230

$451

2051

$32

$143

$228

$439

2052

$31

$141

$224

$428

2053

$30

$138

$221

$417

2054

$29

$136

$217

$406

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th

226

-------
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.

227

-------
Table 4.2.5.2-4: Present value of 30-year stream of climate benefits for the combined 2023-
2025 standards, using low biofuel/high petroleum lifecycle analysis estimates, relative to the
No RFS baseline, presented with four values for the social cost of carbon (SC-CO2)

(millions of 2021$)a	

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$(2,778)

$(9,461)

$(14,001)

$(28,257)

2024

$714

$2,454

$3,637

$7,342

2025

$750

$2,600

$3,859

$7,794

2026

$725

$2,538

$3,774

$7,622

2027

$710

$2,510

$3,739

$7,550

2028

$695

$2,481

$3,703

$7,476

2029

$679

$2,452

$3,667

$7,400

2030

$664

$2,422

$3,631

$7,323

2031

$651

$2,395

$3,596

$7,255

2032

$638

$2,368

$3,561

$7,185

2033

$625

$2,340

$3,526

$7,113

2034

$612

$2,312

$3,490

$7,039

2035

$598

$2,283

$3,454

$6,963

2036

$585

$2,254

$3,417

$6,886

2037

$571

$2,225

$3,380

$6,807

2038

$558

$2,196

$3,343

$6,728

2039

$544

$2,166

$3,306

$6,647

2040

$530

$2,136

$3,269

$6,565

2041

$518

$2,107

$3,230

$6,473

2042

$506

$2,077

$3,192

$6,380

2043

$421

$1,748

$2,692

$5,368

2044

$412

$1,728

$2,669

$5,308

2045

$403

$1,707

$2,643

$5,244

2046

$392

$1,682

$2,610

$5,165

2047

$382

$1,656

$2,577

$5,086

2048

$372

$1,630

$2,544

$5,007

2049

$361

$1,605

$2,511

$4,928

2050

$351

$1,580

$2,478

$4,850

2051

$344

$1,543

$2,452

$4,725

2052

$334

$1,515

$2,414

$4,603

2053

$57

$260

$416

$785

2054b

$29

$136

$217

$406

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th

228

-------
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
b Combined impacts presented in Table 4.2.5.2-4 are the sum of the three thirty-year streams of impacts for the 2023
through 2025 standards presented in Tables 4.2.5.2-1, 4.2.5.2-2, and 4.2.5.2-3. Because we assess thirty years of
impacts for each year standards, the period of analysis for the 2023 standards extends to 2052.

229

-------
Table 4.2.5.2-5: Present value of 30-year stream of climate benefits for 2023 standards,
using high biofuel/low petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2021$)a	

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$(4,599)

$(15,662)

$(23,178)

$(46,778)

2024

$144

$496

$735

$1,484

2025

$142

$491

$728

$1,471

2026

$139

$485

$722

$1,458

2027

$136

$480

$715

$1,444

2028

$133

$475

$708

$1,430

2029

$130

$469

$701

$1,415

2030

$127

$463

$694

$1,400

2031

$125

$458

$688

$1,387

2032

$122

$453

$681

$1,374

2033

$120

$448

$674

$1,361

2034

$117

$442

$668

$1,346

2035

$114

$437

$661

$1,332

2036

$112

$431

$654

$1,317

2037

$109

$426

$647

$1,302

2038

$107

$420

$639

$1,287

2039

$104

$414

$632

$1,271

2040

$101

$409

$625

$1,256

2041

$99

$403

$618

$1,238

2042

$97

$397

$610

$1,220

2043

$234

$972

$1,497

$2,985

2044

$228

$958

$1,479

$2,941

2045

$223

$943

$1,460

$2,897

2046

$217

$929

$1,442

$2,854

2047

$211

$915

$1,424

$2,810

2048

$205

$901

$1,406

$2,766

2049

$200

$887

$1,387

$2,723

2050

$194

$873

$1,369

$2,680

2051

$190

$853

$1,355

$2,611

2052

$184

$837

$1,334

$2,543

2053

$-

$-

$-

$-

2054

$-

$-

$-

$-

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th

230

-------
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.

231

-------
Table 4.2.5.2-6: Present value of 30-year stream of climate benefits for 2024 standards,
using high biofuel/low petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2021$)a	

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$-

$-

$-

$-

2024

$19

$67

$99

$200

2025

$4

$14

$21

$43

2026

$4

$14

$21

$42

2027

$4

$14

$21

$42

2028

$4

$14

$21

$42

2029

$4

$14

$20

$41

2030

$4

$13

$20

$41

2031

$4

$13

$20

$40

2032

$4

$13

$20

$40

2033

$3

$13

$20

$39

2034

$3

$13

$19

$39

2035

$3

$13

$19

$39

2036

$3

$13

$19

$38

2037

$3

$12

$19

$38

2038

$3

$12

$19

$37

2039

$3

$12

$18

$37

2040

$3

$12

$18

$36

2041

$3

$12

$18

$36

2042

$3

$12

$18

$35

2043

$3

$11

$18

$35

2044

$2

$9

$14

$28

2045

$2

$9

$14

$28

2046

$2

$9

$14

$27

2047

$2

$9

$14

$27

2048

$2

$9

$13

$26

2049

$2

$8

$13

$26

2050

$2

$8

$13

$25

2051

$2

$8

$13

$25

2052

$2

$8

$13

$24

2053

$2

$8

$12

$24

2054

$-

$-

$-

$-

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th

232

-------
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.

233

-------
Table 4.2.5.2-7: Present value of 30-year stream of climate benefits for 2025 standards,
using high biofuel/low petroleum lifecycle analysis estimates, relative to the No RFS
baseline, presented with four values for the social cost of carbon (SC-CO2) (millions of
2021$)a	

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$-

$-

$-

$-

2024

$-

$-

$-

$-

2025

$10

$35

$51

$103

2026

$5

$17

$25

$50

2027

$5

$16

$25

$50

2028

$5

$16

$24

$49

2029

$4

$16

$24

$49

2030

$4

$16

$24

$48

2031

$4

$16

$24

$48

2032

$4

$16

$23

$47

2033

$4

$15

$23

$47

2034

$4

$15

$23

$46

2035

$4

$15

$23

$46

2036

$4

$15

$22

$45

2037

$4

$15

$22

$45

2038

$4

$14

$22

$44

2039

$4

$14

$22

$44

2040

$3

$14

$21

$43

2041

$3

$14

$21

$42

2042

$3

$14

$21

$42

2043

$3

$13

$21

$41

2044

$3

$13

$20

$41

2045

$3

$12

$19

$37

2046

$3

$12

$19

$37

2047

$3

$12

$18

$36

2048

$3

$12

$18

$36

2049

$3

$11

$18

$35

2050

$3

$11

$18

$35

2051

$2

$11

$18

$34

2052

$2

$11

$17

$33

2053

$2

$11

$17

$32

2054

$2

$10

$17

$31

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th

234

-------
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.

235

-------
Table 4.2.5.2-8: Present value of 30-year stream of climate benefits for the combined 2023-
2025 standards, using high biofuel/low petroleum lifecycle analysis estimates, relative to the
No RFS baseline, presented with four values for the social cost of carbon (SC-CO2)

(millions of 2021$)a	

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$(4,599)

$(15,662)

$(23,178)

$(46,778)

2024

$164

$563

$834

$1,684

2025

$156

$539

$801

$1,617

2026

$148

$516

$768

$1,550

2027

$144

$510

$760

$1,536

2028

$141

$505

$753

$1,521

2029

$138

$499

$746

$1,505

2030

$135

$493

$738

$1,489

2031

$132

$487

$731

$1,475

2032

$130

$482

$724

$1,461

2033

$127

$476

$717

$1,447

2034

$124

$470

$710

$1,432

2035

$122

$464

$702

$1,416

2036

$119

$458

$695

$1,401

2037

$116

$453

$687

$1,385

2038

$113

$447

$680

$1,368

2039

$111

$441

$672

$1,352

2040

$108

$434

$665

$1,335

2041

$105

$428

$657

$1,316

2042

$103

$422

$649

$1,298

2043

$240

$997

$1,535

$3,061

2044

$234

$980

$1,513

$3,010

2045

$228

$964

$1,493

$2,962

2046

$222

$950

$1,474

$2,918

2047

$216

$936

$1,456

$2,873

2048

$210

$921

$1,437

$2,829

2049

$204

$907

$1,419

$2,784

2050

$199

$892

$1,400

$2,740

2051

$194

$872

$1,385

$2,669

2052

$188

$856

$1,364

$2,600

2053

$4

$18

$29

$56

2054b

$2

$10

$17

$31

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th

236

-------
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.
b Combined impacts presented in Table 4.2.5.2-8 are the sum of the three thirty-year streams of impacts for the 2023
through 2025 standards presented in Tables 4.2.5.2-5, 4.2.5.2-6, and 4.2.5.2-7. Because we assess thirty years of
impacts for each year standards, the period of analysis for the 2023 standards extends to 2052.

237

-------
Table 4.2.5.2-9: Present value of 30-year stream of climate benefits for 2023 supplemental
volume requirement, using low biofuel/high petroleum lifecycle analysis estimates,

presented with

'our values for the social cost of carbon (SC-CO2) (millions of 2021$)a

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$(203)

$(692)

$(1,024)

$(2,066)

2024

$31

$106

$158

$318

2025

$30

$105

$156

$316

2026

$30

$104

$155

$313

2027

$29

$103

$153

$310

2028

$29

$102

$152

$307

2029

$28

$101

$150

$304

2030

$27

$99

$149

$300

2031

$27

$98

$148

$298

2032

$26

$97

$146

$295

2033

$26

$96

$145

$292

2034

$25

$95

$143

$289

2035

$25

$94

$142

$286

2036

$24

$92

$140

$283

2037

$23

$91

$139

$279

2038

$23

$90

$137

$276

2039

$22

$89

$136

$273

2040

$22

$88

$134

$269

2041

$21

$86

$133

$266

2042

$21

LO
00

$131

$262

2043

$15

$61

$93

$186

2044

$14

$60

$92

$184

2045

$14

$59

$91

$181

2046

$14

00
LO

$90

$178

2047

$13

$57

$89

$175

2048

$13

$56

$88

$173

2049

$12

$55

$87

$170

2050

$12

$54

LO
00
&

$167

2051

$12

$53

LO
OO

$163

2052

$12

$52

$83

$159

2053

$-

$-

$-

$-

2054

$-

$-

$-

$-

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated

238

-------
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.

Table 4.2.5.2-10: Present value of 30-year stream of climate benefits for 2023 supplemental
volume requirement, using high biofuel/low petroleum lifecycle analysis estimates,

presented with

'our values for the social cost of carbon (SC-CO2) (millions of 2021$)a

Year

Rate: 5%

Rate: 3%

Rate: 2.5%

Rate: 3%
95th percentile

2023

$(313)

$(1,065)

$(1,577)

$(3,182)

2024

$5

$16

$23

$47

2025

$4

$15

$23

$46

2026

$4

$15

$23

$46

2027

$4

$15

$23

$46

2028

$4

$15

$22

$45

2029

$4

$15

$22

$45

2030

$4

$15

$22

$44

2031

$4

$14

$22

$44

2032

$4

$14

$21

$43

2033

$4

$14

$21

$43

2034

$4

$14

$21

$42

2035

$4

$14

$21

$42

2036

$4

$14

$21

$41

2037

$3

$13

$20

$41

2038

$3

$13

$20

$41

2039

$3

$13

$20

$40

2040

$3

$13

$20

$40

2041

$3

$13

$19

$39

2042

$3

$13

$19

$38

2043

$12

$51

$79

$157

2044

$12

$50

$78

$155

2045

$12

$50

$77

$153

2046

$11

$49

$76

$150

2047

$11

00

$75

$148

2048

$11

$47

$74

$146

2049

$11

$47

$73

$143

2050

$10

$46

$72

$141

2051

$10

$45

$71

$138

2052

$10

$44

$70

$134

2053

$-

$-

$-

$-

2054

$-

$-

$-

$-

aThis analysis portrays what might be expected if, in each of the ensuing 29 years, aggregate renewable fuel
consumption for each category exceeded baseline levels by the same volume as required by the rule. EPA's lifecycle
analysis methodology includes GHG impacts for biofuels over a 30-year period based on public comment and the

239

-------
input of an expert peer review panel as described in the March 2010 RFS2 rule (75 FR 14670). Parentheses indicate
negative values. Climate benefits are based on changes (reductions) in CO2 emissions and are calculated using four
different estimates of the SC-CO2 (model average at 2.5 percent, 3 percent, and 5 percent discount rates; and 95th
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated
using all four estimates. As discussed in the Technical Support Document: Social Cost of Carbon, Methane, and
Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a consideration of climate benefits
calculated using lower discount rates are also warranted when discounting intergenerational impacts.

We note that the methodology underlying the SC-CO2 estimates used in this analysis has
been subject to public comment in the context of dozens of 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.391 As part of that separate process, the EPA
welcomes the opportunity to continually improve its understanding through public input on the
analytical issues associated with the presentation of anticipated costs, benefits, and other impacts
of its actions, as done through RIAs.

4.3 Conversion of Wetlands, Ecosystems, and Wildlife Habitats

The Second Triennial Report to Congress on Biofuels392 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. It can also be seen in section 4.2.2 of this section the wide range
of estimates for the area and types of land use change depending on feedstock, model choice,
scenario design and input assumptions. 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

391	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/docunients/2022/01/25/2022-Q1387/request-for-noniinations-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.

392	U.S. EPA (2018). Biofuels and the Environment: Second Triennial Report to Congress. U.S. Environmental
Protection Agency, EPA/600/R-18/195: 159 pp. Washington, DC, June 2018.

240

-------
vehicles and engines) are mostly from air quality effects (Chapter 4.1), climate effects (Chapter
4.2), and possible leakage from underground storage tanks (Chapter 4.4.4). Insofar as there are
impacts of renewable fuel on the conversion of wetlands, ecosystems, and wildlife habitats, they
are associated with crop-based feedstocks rather than waste fats, oils, and greases, biogas, or
biogas electricity. Discussion of the impacts of the candidate volumes in many of these chapters
will be related to the 2022 baseline as a substitute to the No RFS baseline as land use change
would not be drastically affected by the removal of the RFS program as referenced in Chapter 2.
Additionally, as stated in Chapter 4.1.3 the proposed electricity volume requirements are
expected to be met with existing renewable electricity production. We note that to the extent the
RFS standards in this proposed 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 degradation
of wetlands, ecosystems and wildlife habitats outside of the U.S., and other adverse
environmental impacts on air quality, soil quality and water quality outside of the U.S. Tropical
forests are carbon sinks, and their conversion for oil palm production results in both sequestered
carbon emissions and foregone future carbon sequestration. These impacts mitigate the potential
GHG benefit otherwise provided by biofuel displacement of conventional fuels.393

393 Austin, K. G., A. Schwantes, Y. Gu and P. S. Kasibhatla (2019). "What causes deforestation in Indonesia?"
Environmental Research Letters 14(2): 024007; Austin, K. G., M. Gonzalez-Roglich, D. Schaffer-Smith, A. M.
Schwantes and J. J. Swenson (2017). "Trends in size of tropical deforestation events signal increasing dominance of
industrial-scale drivers." Environmental Research Letters 12(5); Austin, K. G., A. Mosnier, J. Pirker, I. McCallum,
S. Fritz and P. S. Kasibhatla (2017). "Shifting patterns of oil palm driven deforestation in Indonesia and implications
for zero-deforestation commitments." Land Use Policy 69: 41-48; Babel, M. S., B. Shrestha and S. R. Perret (2011).
"Hydrological impact of biofuel production: A case study of the Khlong Phlo Watershed in Thailand." Agricultural
Water Management 101(1): 8-26.; Carlson, K. M., L. M. Curran, G. P. Asner, A. M. Pittman, S. N. Trigg and J. M.
Adeney (2013). "Carbon emissions from forest conversion by Kalimantan oil palm plantations." Nature Climate
Change 3(3): 283-287; Gatto, M., M. Wollni and M. Qaim (2015). "Oil palm boom and land-use dynamics in
Indonesia: The role of policies and socioeconomic factors." Land Use Policy 46: 292-303; Gaveau, D. L. A., D.
Sheil, M. A. Salim, S. Arjasakusuma, 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 USA 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; Mukherjee, 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): eO 159668.

241

-------
4.3.1 Wetlands

There are several federal reports that describe the status and trends of U.S. wetlands,394
including the U.S. Fish and Wildlife Service (USFWS) Status and Trends of Wetlands in the
Conterminous United States,395 the USFWS and NOAA Status and Trends of Wetlands in the
Coastal Watersheds of the Coterminous United States,396 the USFWS Status and Trends of
Prairie Wetlands in the United States,397 EPA's National Wetland Condition Assessment398
(NWCA), and USDA's Natural Resources Inventory (NRI).399 The USGS NWALT (National
Water-Quality Assessment (NAWQA) Program's Wall-to-Wall Anthropogenic Land Use
Trends) series does not model changes in wetlands.400 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.401 Wetlands are not an independent land
cover class in the NRI, but are overlaid on other land cover types (e.g., wetlands on forested
lands made up 66,053,800 acres in 2007). The changes in wetland acres between 2007 and 2017
are shown in Table 4.3.1-1. There was an overall reduction by roughly 52,800 acres between
2007 and 2012, and a further reduction of 64,300 acres between 2012 and 2017. Over the full
2007 to 2017 timeframe, these changes represent a reduction of 0.11%. These reductions were
mostly from losses of wetlands on cropland and rangeland, which were partly offset by gains in

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

wetlands-ns

395	Dahl, T.E. 2011. Status and trends of wetlands in the conterminous United States 2004 to 2009. U.S.
Department l'lc Interior; Fish and Wildlife Service, Washington, D.C. 108 pp.

396	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.)

397	Dahl, T.E. 2014. Status and trends of prairie wetlands in the United States 1997 to 2009. U.S. Department of
t]lc Interior; Fish and Wildlife Service, Ecological Services, Washington, D.C. (67 pages).

398	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

399	U.S. Department of Agriculture. 2020. Summary Report: 2017 National Resources Inventory, Natural
Resources Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State
University, Ames, Iowa. https://www.nrcs.usda.gov/sites/defauit/files/2022-10/2017NRISummary Finai.pdf
(accessed November 30, 2022).

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

401	See Table 7 - Changes in land use/cover between 2012 and 2017, U.S. Department of Agriculture. 2020.
Summary Report: 2017 National Resources Inventory, Natural Resources Conservation Service, Washington, DC,

and Center for Survey Statistics and Methodology, Iowa State University, Ames, Iowa.

https://www.nrcs.nsda.gov/sites/defanit/files/2022-10/2017NRISnm.mare Finai.pdf (accessed November 30, 2022).

242

-------
developed and water areas.402 The report does not provide the information needed to determine
the portion of wetland acres lost in order to grow feedstocks for biofuels, nor does it attempt to
identify the portion of lost wetland acres attributable to the RFS program.

Table 4.3.1-1: Changes in palustrine403 and estuarine404 wetlands on different land
use/cover types between 2007, 2012, and 2Q17405	



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.406 In the only other national assessment to date, Wright
et al. (2017) found that within 50 miles of an ethanol biorefineiy there was a 14,000-acre loss of
wetland between 2008 and 2012. While one might infer a causal connection between proximity
to an ethanol biorefinery and loss of wetlands (a question that was not investigated directly), this
study nevertheless does not demonstrate a connection to the RFS program specifically. As
discussed in Chapter 1, there are and have been numerous other drivers for ethanol use in the
U.S., most significantly the economic benefits of using ethanol in E10 blends. Additionally, a

402	"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."

403	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=nrcs 143 014127 (last
accessed May 6, 2021).

404	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/nrcs/detail/national/technical/nra/nri/processes/?cid=nrcs 143 014127 (last
accessed May 6, 2021).

405	U.S. Department of Agriculture. 2020. Summary Report: 2017National Resources Inventory, Natural Resources
Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State University,
Ames, Iowa, https://www.nrcs.usda.gov/wps/portal/nrcs/main/national/technical/nra/nri/results.

406	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.

243

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

There are also many differences between Wright et al. (2017) and the NRI that make
direct comparison of these two studies not relevant. These differences stem from numerous
sources, including the geographic extent (the entire contiguous U.S. for the NRI versus only
areas in the contiguous U.S. within 100 miles of a biorefineiy in Wright et al. (2017)), and
source data (fixed random points in the NRI versus satellite-derived data from the USDA's
Cropland Data Layer in Wright et al. (2017)). Reconciling these estimates is beyond the scope of
this rulemaking. Nonetheless, when we consider these two national assessments and the other
studies cited above overall, they demonstrate that agricultural extensification may affect
wetlands,407 but any losses are relatively small compared with the total amount of wetland.
Moreover, as stated above, where these studies were directed at the potential impacts of biofuels
they considered the impact of increased biofuel production generally, not the incremental impact
in biofuel production attributable to the RFS program or the volumes being proposed in this
proposed rule. As discussed in further detail in Chapter 2, much of the biofuel projected to be
used in 2023-2025 would be expected to be used even in the absence of the RFS volume
requirements. Any land use change associated with biofuels that would be used in the absence of
the RFS volume requirements is therefore not attributable to this proposed rule.

In the most recent NRI, the USDA 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 20 1 7.408 The bulk of the wetland losses were in the
Prairie Pothole region, as reported elsewhere,409 with some very high rates (i.e., >15%, Wright et
al. 2017). The conversion reported by Wright, Larson et al. (2017) explicitly included only lands
that had not been in cropland for at least 20 years; although these areas may not represent
pristine habitats, they are expected to represent habitats that are in a relatively natural state.

The studies discussed above show that total wetland acres in the contiguous U.S. have
been decreasing since 2007. 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 candidate volumes. The volume increases for
2023-2025 compared to the No RFS baseline that are described in Chapter 3 due to biofuels
produced from agricultural feedstocks (especially corn and soybeans) would suggest the

407	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.

408	U.S. Department of Agriculture. 2020. Summary Report: 2017National Resources Inventory, Natural
Resources Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State

University, Ames, Iowa, at Table 18. https://www.nrcs.nsda.gov/sites/defanlt/files/2022-10/

2017NRISummary Finatpdf (accessed November 30, 2022). See also USDA, 2017 National Resources Inventory,

at https://www.nrcs.usda.gov/Internet/NRCS	RCA/repoits/nil	wet	nathlml. (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.

409	Johnston, C. A. (2013). "Wetland losses due to row crop expansion in the Dakota Prairie Pothole Region."
Wetlands33(1): 175-182. Johnston, C. A. (2014). "Agricultural expansion: land use shell game in the UTS.

Northern Plains." Landscape Ecology 29(1): 81-95.

244

-------
potential for an associated increase in crop production. As such, they may be associated with
increased pressure to convert wetlands into cropland or otherwise impact wetlands. However, if
we consider the potential impacts relative to the current situation in 2022 (i.e., the 2022 baseline
discussed in Chapter 2.2) there would be much less potential impact. Additional information on
land use change from corn and soybean can be located in Chapter 4.2. The main proposed
volume changes are from biofuel produced from biogas (CNG/LNG and electricity) which are
anticipated to be supplied from existing facilities. Therefore, no effect from this change would be
anticipated on wetlands as no additional land is needed to meet these candidate volumes. More
information is needed to assess the degree to which the 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 4.3.1-1: 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 410). Stars denote the
location of biorefineries in the analysis.

4.3.2 Ecosystems Other Than Wetlands

There are many ecosystems other than wetlands that may be affected by biofuel
production and use, including grasslands, forests, and aquatic habitats downstream of corn and
soybean production areas. Impacts on aquatic habitats, such as from runoff of fertilizer and
pesticides, as well as changes in hydrology from tilling, are discussed in Chapter 4.4. 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.

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

245

-------
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.411 Wright et al. (2017)
estimated much larger reductions of grassland (2 million acres), forests (60,000 acres), and
shrublands (52,000 acres), than in wetland reductions (estimated 14,000 acre reduction) (Figure
4.3.2-1). The bulk of the grassland conversions occurred in South Dakota (348,000 acres), Iowa
(297,000 acres), Kansas (256,000 acres), Missouri (239,000 acres), Nebraska (213,000 acres),
and North Dakota (176,000 acres).412

Figure 4.3.2-1: Relative conversion rates to cropland from either (a) grassland, (b) forest,
or (c) shrubland from 2008 to 2012

0 04 1.2 2 7 7.8 too	0 1.2 i5 82 21.2 100

Rates are relativized by type of ecosystem within a 3.5-mile spatial grid (modified from 41^). 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 2012.414 The NRI defines rangeland as a land use/land cover that is more

411	Id.

412	Id.

413	Id.

414	U.S. Department of Agriculture, 2020. Summary Report: 2017 National Resources Inventory, Natural
Resources Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State
University, Ames, Iowa. https://www.nrcs.usda.gov/sites/default/files/2022-10/2017NRISummaA' Final.pdf
(accessed November 30, 2022).

246

-------
lightly managed than pastureland,415 and, as such, is probably the NRI land use/land cover most
comparable to the grassland in Wright et al. (2017). The biggest reduction in rangeland was from
conversion to cropland (743,400 acres), followed by developed land (535,800 acres), and then
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).416

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.417 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.418 Despite the rental payment incentive to
farmers, enrollment in the CRP has been shrinking since 2007.419 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 20 20.420
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%).421 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.,

415	The 2020 NRI defines rangeland as "A broad land cover/use category on which the climax or potential plant
cover is composed principally of native grasses, grass-like plants, forbs or shrubs suitable for grazing and browsing,

and introduced forage species that are managed like rangeland. This would include areas where introduced hardy

and persistent grasses, such as crested wheatgrass, are planted and such practices as deferred grazing, burning,

chaining, and rotational grazing are used, with little or no chemicals or fertilizer being applied. Grasslands,

savannas, many wetlands, some deserts, and tundra are considered to be rangeland. Certain communities of low

forbs and shrubs, such as mesquite, chaparral, mountain shrub, and pinyon-juniper, are also included as rangeland."

416	U.S. Department of Agriculture. 2020. Summary Report: 2017National Resources Inventory, Natural Resources
Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State University,

Ames, Iowa. https://www.nrcs.usda.gov/sites/default/files/2022-10/2017NRISummarv Final.pdf (accessed

November 30, 2022).

417	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

418	Gray, B. J., & Gibson, J. W. (2013). "Actor-networks, farmer decisions, and identity." Culture, Agriculture,
Food311^Environment, 35(2), 82el0l. Brown, J. C., et al. (2014). "Ethanol plant location and intensification vs.
extensification of corn cropping in Kansas." Applied Geography 53: 141-148.

419	Data from the USDA Farm Service Agency (FSA). available at https://www.fsa.usda.gov/proarams-and-

services/ctmservation-programs/reports-and-statistics/amseivation-liijifv^

on April 12, 2021).

420	USDA FSA, FY 2020 Annual Summary. Data from the USDA Farm Service Agency (FSA), available at^E^
www.fsa.usda.gov/programs-and-services/conservation-programs/reports-and-statistics/conservation-reserve-

program-statistics/index (last accessed on May 5, 2021).

421	U.S. Department of Agriculture. 2020. Summary Report: 2017National Resources Inventory, Natural Resources
Conservation Service, Washington, DC, and Center for Survey Statistics and Methodology, Iowa State University,

Ames, Iowa, at 3-45. https://www.nrcs.usda.gov/wps/portal/nrcs/main/national/technical/nra/nri/results/.

247

-------
corn, soybean, winter wheat, spring wheat, and sorghum), with the majority of that to corn and
soybean.42z Reconciling these studies suggests that, of the land that leaves the CRP and goes into
the generic category of cropland in the NRI, at least half of that cropland is devoted to row crops.
The change in CRP enrollment is not uniform across the country (Figure 4.3.2-2), with much of
the reduction in the western and northern plains, the same areas experiencing losses of grassland
and increases in agriculture.

Figure 4.3.2-2: Change in CRP enrollment between 2007 and 2016.423

Note. Data as of the end of April 2010.

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 (+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.424

422	Morefield, P. I".., 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.

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

424	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.

248

-------
Thus, even though some forest land did convert to cropland according to the NRI,425 these
conversions appear small and to be offset by reforestation of pastureland.

The volume increases for the 2023-2025 years compared to the No RFS baseline
described in Chapter 2 due to biofuel production from agricultural feedstock (notably soybean oil
for renewable diesel) suggests the potential for an associated increase in crop production. As
renewable diesel growth continues, it is expected that demand for soybean crush with rise. Thus
the 2023-2025 volumes will have a potential to adversely impact grassland and other non-
wetland ecosystems. However, if we consider the potential impacts relative to the current
situation in 2022 (i.e., the 2022 baseline discussed in Chapter 2.2) there would be little impact, as
the overall volume increase for biodiesel and renewable diesel is much smaller and expected to
be met with expanded waste fats, oils, and greases supply. 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.

4.3.3 Wildlife

There are many subsequent potential impacts to wildlife from these changes in wetlands
and other ecosystems, which were also summarized in the Second Triennial Report to Congress
on Biofuels.426 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).427 Here we focus on potential impacts to terrestrial wildlife, including primarily birds
and insects, which have been the most studied to date. Impacts to aquatic wildlife are described
in Chapter 4.4.2.3.

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.428 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.429 Evidence
suggests that the direct effects of increasing cultivation of corn and soybean for biofuel
production are coming mostly from the conversion of grasslands to cropland, rather than other

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

426	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.

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

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

429	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^(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 ONE9(10): el09989:

10.1371/journal.pone.0109989.

249

-------
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.430 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.431 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.432 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.433 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 oryzivoras] and Henslow's Sparrow
[Anunodra/nus 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.434

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.435 Roughly 20%
of these pollination services are estimated from wild populations which depend on local habitat
for food and nesting sites.436 A 2016 modeling study suggests that wild bee populations
decreased by 23% across the U.S. between 2008 and 20 1 3.437 The causes of these reductions are

430	Tudge, S.J., Purvis, A. & De Palma, A. "The impacts of biofuel crops on local biodiversity: a global synthesis."
Biodivers Conserve, 2863-2883 (2021). https://doLorg/10.1007/sl0531-021-02232-5

431	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.

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

433	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.

434	Id.

435	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. 1517685113.

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

437	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. 1517685113.

250

-------
complex, but include land use change, pesticides, and disease.438 Subsequent effects from
reductions in local bee populations are possible, including reductions in pollinator-dependent
crops grown in the area,439 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.440 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.441 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.442 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
measure exposures and relate them to effects at the individual and colony level.443 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.

438	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.

439	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.

440	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.

441	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.

442	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.

443	U.S. EPA (2018), "How We Assess Risks to Pollinators." Available at https://www.epa.gov/po11inator-
protection/how-we-ass ess-risks-pollinators (last accessed April 12, 2021).

251

-------
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 candidate volumes, specifically.
Attributing such impacts to the RFS program generally, let alone the specific candidate volumes
being analyzed in this action, is even more difficult (see Second Triennial Report to Congress on
Biofuels Sections 4.4.3.2 and 2.4.4). EPA is in the process of conducting a Biological Evaluation
which will evaluate impacts on endangered species from the RFS Program. More information on
the estimated impact to species in the affected region on the RFS program will be available when
the evaluation is concluded.

4.3.4 Potential Future Impacts of Annual Volume Requirements

The volume increases for 2023-2025 described in Chapter 3 due to biofuels produced
from agricultural feedstocks (especially corn and soybeans) would suggest the potential for an
associated increase in crop production. As such, they may be associated with increased pressure
to convert grasslands and wetlands into cropland, and, therefore, also increased pressure on
wildlife habitats. There exists substantial uncertainty in projecting changes in land use and
management associated with corn, soybeans, and other crops. Additional information and
modeling are needed to fully assess changes in habitat areas and effects on wildlife, both for crop
expansion and pesticide use. Modeling and discussion on the estimates for land use change are
further discussed in Chapter 4.2. Changes as a result of biofuel to electricity are not excepted to
impact grasslands and wetlands as it is believed that existing CNG/LNG production will be
transferred to meet these volumes.

4.4 Soil and Water Quality

Soil and water quality are addressed together here as they are in many ways intertwined,
with effects on soil often directly altering water quality (e.g., soil erosion leading to
sedimentation). Soil quality, also referred to as soil health, is the capacity of a soil to function,
including the ability to sustain plant growth.444 It can be affected by biofuel feedstock production
through changes in soil erosion, soil organic matter (SOM),445 and soil nutrients, among other
characteristics. Soil erosion can negatively impact soil quality by disproportionately removing

444	The USDA 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. (2017). "Soil health glossary." Retrieved 5/1, 2017, from
https://www.nrcs.usda.gov/wps/portal/nrcs/detailfull/soils/health/?cid=nrcsl42p2 053848. 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.

445	Soil organic matter is defined by Brady, N. and R. Weil (2000). Elements of the Nature and Properties of Soils.
Upper Saddle River, NJ, USA, Prentice-Hall, Inc. as "[t]he organic fraction of the soil that includes plant and animal
residues at various stages of decomposition, cells and tissues of soil organisms, and substances synthesized by the
soil population." Brady N and Weil R (2000). Elements of the Nature and Properties of Soils. Upper Saddle River,
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).

252

-------
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.446 Crop-based
biofuel feedstock production can affect water quality through associated changes in nutrients,
dissolved oxygen, sediment, and chemical loadings.447 Nutrient releases from cropland into
nearby waterways can result in excessive algal growth (i.e., algal blooms), leading to low
dissolved oxygen levels (i.e., hypoxia) in some cases. Increased sediment and total dissolved
solids can make water unsuitable for consumption and irrigation, and also have negative impacts
on aquatic species. In addition, chemical releases or biofuel leaks and spills from above-ground,
underground, and transport tanks can be detrimental to water quality leading to ground, surface,
and drinking water contamination (see Chapter 4.4.2).448 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.

4.4.1 The Role of Biofuels

Corn starch ethanol and soybean oil biodiesel account for most of the biofuel volumes
produced to date. As a result, the majority of soil and water quality impacts from biofuels thus
far have come from the production of corn and soybeans. There have also been notable quantities
of biogas from landfills that is cleaned and compressed to be used in compressed natural gas
(CNG) vehicles, and waste fats, oils and greases (FOG) that is used to produce biodiesel. As of
this proposed rule, electricity from biogas will also be biofuel source. However, they are not
sourced from crop-based feedstocks and thus have only a tenuous connection to soil and water
quality. Additionally, products such as CNG/LNG and electricity from biofuel are not anticipated
to have any land affecting production changes. Canola oil is also used for biodiesel production,
though in considerably smaller quantities than soybean oil and FOG, and it is a crop-based
feedstock that could potentially impact soil and water quality. However, few studies focus on
canola oil.

446	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.

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

448	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).

253

-------
Since 2007, grasslands, including CRP grasslands, have been converted to corn and
soybeans, in a process termed extensification (see Chapter 4.4.2.1). Corn and soybeans have also
replaced other kinds of cropland. By contrast, the use of other crop-based feedstocks for biofuel
production has been much more limited. For example, use of corn stover has been attempted at a
couple of locations.449 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.450 For that reason, we have included those feedstocks in this
analysis, though they are not widely used. Finally, outside the U.S., palm oil production for
biodiesel is an established industry in countries such an Indonesia, Malaysia, and Thailand, with
production occurring mainly for export, including to the U.S. As noted in Chapter 4.3, there is
strong evidence that expanded palm oil production would adversely affect soil and water quality,
in addition to carbon sequestration, outside of the U.S.

4.4.2 Impacts to Date

4.4.2.1 Soil and Proximal Water Quality Effects

Primarily, the magnitude of the impacts to soil and water quality depend upon the
feedstock grown and land use—i.e., the type of land used for growing the biofuel feedstock and
the management implemented on that land. For a given acre of cropland, planting corn or
soybeans onto grasslands (extensification) can be expected to have greater negative effects on
soil and water quality relative to the conversion of other existing cropland, such as wheat, to corn
or soybeans (intensification). Grassland-to-annual-crop conversion typically impacts soil quality
negatively because it increases erosion and the loss of soil nutrients and SOM, including soil
carbon loss to the atmosphere.451 In a meta-analysis, Qin et al. (2016) found that replacing
grasslands with corn decreased soil carbon by approximately 20% on average.452 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
discussed in greater detail later in this section).453 Increased erosion from conversion, in turn, can

449	81 FR 89746 (December 12, 2016).

450	Ziolkowska, J.R., and Simon, L. (2014). "Recent developments and prospects for algae-based fuels in the US."
Renewable & Sustainable Energy Reviews I'd: 847-853: 10.1016/j .rser.2013.09.021.

451	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.1017277108. 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. Ill 1/gcbb. 12237. Lai, R.
(2003). "Soil erosion and the global carbon budget." Environment Internationally^)'. 437-450: 10.1016/s0160-
4120(02)00192-7.

452	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.

453	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.

254

-------
negatively impact water quality through increased sediment and nutrient loadings to
waterways.454

Corn and soybeans additionally affect water quality through increased chemical usage,
some of which moves as runoff or leaching to surface waterways or groundwater. Table 4.4.2.1-1
summarizes the most recent USDA National Agricultural Statistics Service (NASS) Agricultural
Chemical Use Survey results for domestic corn and soybean acreage, as well as domestic wheat
acreage for comparison. In general, soybean acreage receives substantially less fertilizer than
corn, particularly nitrogen, because soybeans can attain nitrogen from the atmosphere via
symbiotic nitrogen fixation whereas corn cannot. Thus, as an example, multiplying 1.94 million
acres of extensification in the U.S. attributed to corn455 by the average nitrogen fertilizer rate
corn receives (149 lbs N/acre) yields an increase of approximately 289 million pounds of
additional nitrogen added per year. Likewise, from the most recent surveys by the USDA NASS,
97% of planted corn acres were treated with herbicides, 13% with insecticides, and 17% with
fungicides (Table 4.4.2.1-1). Atrazine was the top active ingredient among herbicides applied to
the planted corn acres, applied to 65% of planted acres, followed by mesotrione, applied to 42%
of planted acres.456 For planted soybean acres, 99% were treated with herbicides, 16% with
insecticides, and 15% with fungicides.457 Glyphosate isopropylamine salt and glyphosate
potassium salt were the top active ingredients among herbicides applied to planted soybean
acres.458 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.459

454	Yasarer, L.M.W., Sinnathamby, S., and Sturm, B.S.M. (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.

455	Lark, T.J., Salmon, J.M., and Gibbs, H.K. (2015). "Cropland expansion outpaces agricultural and biofuel policies
in the United States." Environmental Research Letters 10(4): 10.1088/1748-9326/10/4/044003.

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

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

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

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

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

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

459	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
Environment586: 16-24: https://doi.Org/10.1016/j.scitotenv.2017.02.009.

255

-------
Table 4.4.2.1-1: Summary of Chemical Use for Corn, Soybeans, and Wheat Acreage in the

U.S. based on 2018 and 2019 USDA

MASS Chemical Use Surveys460'461'462



Corn

Soybeans

Winter
wheat

Spring
wheat

Durum
wheat

Nitrogen Fertilizer Applied: % of Planted
Acres

98

29

88

97

98

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

149

17

73

102

83

Phosphate Fertilizer Applied: % of Planted
Acres

79

42

63

89

84

Average Application Rate for Year for
Acres with Phosphate Fertilizer Applied
(lbs PzCVacre)

69

55

31

39

29

Atrazine Applied: % of Planted Acres

65

Not
Reported

Not
Reported

Not
Reported

Not
Reported

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

1.037

Not
Reported

Not
Reported

Not
Reported

Not
Reported

Glyphosate Potassium Salt: % of Planted
Acres

Not
Reported

28

Not
Reported

Not
Reported

Not
Reported

Average Application Rate for Year for
Acres with Glyphosate Potassium Salt
Applied (lbs/acre)463

Not
Reported

1.527

Not
Reported

Not
Reported

Not
Reported

Glyphosate Isopropylamine Salt: % of
Planted Acres

34

47

Not
Reported

Not
Reported

46

Average Application Rate for Year for
Acres with Glyphosate Isopropylamine Salt
Applied (lbs/acre) 464

0.993

1.202

Not
Reported

Not
Reported

0.555

There are a couple of factors that can mitigate impacts on soil and water quality, at least
in part. First, the type of CRP lands, conservation lands, or other grasslands that are converted to
cropland can affect soil quality. In a modeling study, LeDuc et al. (2017) simulated that greater
erosion and loss of soil carbon and nitrogen occurs from converting low productivity, highly
sloped CRP grasslands compared to those with higher productivity soils and lower slopes.465 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

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

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

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

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

462	USDA NASS (2020). 2019 Agricultural Chemical Use Survey: Wheat. Available at

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

463	This is expressed in acid equivalent.

464	This is expressed in acid equivalent.

465	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.

256

-------
practices. About 60-70% of corn and soybeans are grown using conservation tillage.466,467
Conservation tillage, including no-till, reduces soil erosion and increases SOM content relative to
conventional tillage.468,469 (Cassel, Raczkowski et al. 1995, West and Post 2002) (Cassel, Raczkowski
et al. 1995, West and Post 2002) (Cassel, Raczkowski et al. 1995, West and Post 2002) (Cassel,
Raczkowski et al. 1995, West and Post 2002) (Cassel, Raczkowski et al. 1995, West and Post 2002)
(Cassel, Raczkowski et al. 1995, West and Post 2002) (Cassel, Raczkowski et al. 1995, West and Post
2002) (Cassel, Raczkowski et al. 1995, West and Post 2002) (Cassel, Raczkowski et al. 1995, West and
Post 2002) (Cassel, Raczkowski et al. 1995, West and Post 2002) (Cassel, Raczkowski et al. 1995, West
and Post 2002) (Cassel, Raczkowski et al. 1995, West and Post 2002) (Cassel, Raczkowski et al. 1995,
West and Post 2002) (Cassel, Raczkowski et al. 1995, West and Post 2002)

The soil and water quality effects of converting to corn or soybeans from other crops,
such as wheat, are generally less than those of the conversion of grasslands.470 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.471 From this evidence, Zuber et

466	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.

467	Conservation tillage is defined as any tillage practice leaving at least 30% of the soil surface covered by crop
residues; whereas conventional tillage leaves less than 15% of the ground covered by crop residues Lai, R. (1997).
"Residue management, conservation tillage and soil restoration for mitigating greenhouse effect by C02-
enrichment." Soil & Tillage Research 43(1-2): 81-107. No-till management, a subset of conservation tillage, disturbs
the soil marginally by cutting a narrow planting strip. Nationally, approximately 30% and 45% of the area planted to
corn and soybeans, respectively, are under no-till Wade, T., R. Claassen and S. Wallander (2015). Conservation-
practice adoption rates vary widely by crop and region, EIB-147, US Department of Agriculture Economic Research
Service. Since 2000, there has been a general trend toward greater percent residue remaining after planting for both
crops (USDA-ERS 2018 https://data.ers.usda.gov/reports.aspx?ID=17883; 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/nrcs/detail/national/technical/nra/ceap/pub/?cid=nrcsl43 014161
(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.

468	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.

469	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.

470	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 with biofuel 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.

471	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.

257

-------
al. (2015) suggests a movement from wheat to corn may not materially affect soil quality,
provided a shift from no-till to conventional tillage does not occur concomitantly. In a meta-
analysis, Qin, Dunn et al. (2016) found that corn replacing other cropland (e.g., soybean, wheat)
increased soil organic carbon, whereas the opposite occurred when corn replaced grassland or
forest land.472 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
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.473 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.474 In a modeling study, Secchi et al. (2011) concluded that this intensification475 of
corn would likely lead to higher nitrogen and phosphorus loads in the Upper Mississippi River
Basin.476

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.477 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.478 Yet too much stover removal can increase soil erosion,
decrease SOM and soil nutrients, and ultimately decrease corn yields.479 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

472Qin ZC, Dunn JB, Kwon HY, Mueller S and Wander MM (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.

473	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/sl 1104-013-1876-9.

474	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.

475	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.

476	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.

477	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.

478	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/l0.1016/j.bioriibioe.2017.09.016 (last accessed April 13, 2021).

479	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.

258

-------
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.480 Additional research is needed to
understand effects on soil and water quality if soil conservation methods are employed while
harvesting corn stover.

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.481 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.482 Notably, however, the
effects of these perennial feedstocks can depend upon the plant species grown and the type of
land converted.483 Additionally, the literature definitions of what constitutes marginal land and
estimates of its extent vary widely.484 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.485 Alternative feedstock production (e.g., switchgrass) requires less
fertilizer than corn, thereby reducing nutrient runoff.486 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.487 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).488

480	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/sl 2155-014-9419-7.

481	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.

482	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.

483	Robertson GP, Hamilton SK, Barham BL, Dale BE, Izaurralde RC, Jackson RD, Landis DA, Swinton SM,

Thelen KD and Tiedje JM (2017). Cellulosic biofuel contributions to a sustainable energy future: Choices and
outcomes. Science 356(6345): 10.1126/science.aal2324.

484	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.

485	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.

486	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.

487	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.Ill 1/gcbb. 12481.

488	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.

259

-------
4.4.2.2 Downstream Water Quality Effects

Increased corn and soybean cultivation may also affect downstream surface water and
aquatic systems, which can lead to aquatic life effects (see Chapter 4.4.2.3).489 Fertilizer runoff,
in addition to other factors (e.g., temperature and precipitation) and conservation practices,
influence downstream eutrophication,490 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.491 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

49?

ecoregion.

In coastal systems, nutrient loadings affect hypoxic zone size, which is also a function of
climate, weather (e.g., storms), basin493 morphology, circulation patterns, water retention time,
freshwater inflows, stratification, and mixing, as seen in the Gulf of Mexico.494 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

489	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.Org/10.1016/j.ecolmodel.2014.01.016. Jarvie HP, Sharpley AN, Flaten D, Kleinman PJA, Jenkins A
and Simmons T (2015). The pivotal role of phosphorus in a resilient water-energy-food security nexus. Journal of
Environmental Quality 44(4): 1049-1062: 10.2134/jeq2015.01.0030.

490	EPA defines eutrophication as " [a] reduction in the amount of oxygen dissolved in water. The symptoms of
eutrophication include blooms of algae (both toxic and non-toxic), declines in the health of fish and shellfish, loss of
seagrass and coral reefs, and ecological changes in food webs." EPA, Vocabulary Catalog: Acid Rain Glossary,
available at

https://sor.epa.gov/sor internet/registry/termreg/searchandretrieve/glossariesandkeywordlists/search.do?details=&gl
ossaryName=Acid%20Rain%20Glossary (last accessed April 14, 2021).

491	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.

492	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.

493	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/registry/termreg/searchandretrieve/glossariesandkeywordlists/search.do?details=&gl
ossaryName=Chesapeake%20Bay%20Glossary (last accessed April 14, 2021).

494	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.

260

-------
grasses such as switchgrass and Miscanthus, even with manure application, could significantly
reduce phosphorus runoff into water bodies.495

4.4.2.3 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.496 However, it has been shown that increased corn and soybean cultivation may
affect downstream aquatic communities, chiefly through runoff or leaching of nutrients and
pesticides, though changes in land use and land cover also impact aquatic ecosystems,
particularly through conversion of wetlands that provide ecosystem services like improving
surface water flow, groundwater recharge, and sediment control.497 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.498 Corn production requires greater
application of nitrogen fertilizer compared to soy production because soy plants develop root
nodules with bacteria that can fix nitrogen from the atmosphere (Table 4.4.2.1-1). EPA's
National Aquatic Resource Surveys assess the quality of the nation's freshwater and coastal
ecosystems, including biological condition usually derived from the abundance of pollution-
tolerant and pollution-sensitive benthic macroinvertebrate taxa499 and fish.500 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

495	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.

496	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.

497	Ibid.

498	Ibid

499	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
https://www.epa.gov/nationa1-aquatic-resource-surveys/indicators-benthic-

macro invertebrates#:~:text=What%20are%20benthic%20macroinvertebrates%3F.snai1s%2C%20worms%2C%20an
d%20beet1es (last accessed April 14, 2021).

500	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-aquatic-resource-surveys/nrsa (last accessed April
14, 2021).

261

-------
in poor condition and 26% were in good condition based on fish species indicators.501 The
leading problems contributing to poor biological condition were excess nutrients (especially
phosphorus), loss of shoreline vegetation, and excess sediments.502 For rivers and streams, sites
with a condition rating of poor because of excess nutrients were most prevalent in the mid-
continent ecoregions503 of the nation compared to the eastern and western regions.504 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 2007.505 For lakes, disturbance by nutrients varied by ecoregion (Figure
4.4.2.3-1). Northern Plains and Southern Appalachian ecoregions had a higher proportion (67-
80% of within-ecoregion lakes) of sites classified as most disturbed by phosphorus pollution and
there was a statistically significant increase from 2007 to 2014 in the number of most disturbed
lakes in the Northern Appalachian ecoregion. For coastal and Great Lakes nearshore waters
(Figure 4.4.2.3-1), phosphorus was again a widespread problem (rating of poor in 21% of sites)
and biological condition was poorest along the Northeast coast (rating of poor in 27% of sites),
followed by the Great Lakes nearshore waters (rating of poor in 18% of sites).506 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 1970.507

501	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/nationa1-aquatic-resource-surveys/nrsa (last accessed April
14, 2021).

502	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-aquatic-resource-surveys/nrsa (last accessed April
14, 2021).

503	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-aquatic-resource-surveys/nrsa (last accessed April 14, 2021).

504	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-aquatic-resource-surveys/nrsa (last accessed April
14, 2021).

505	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).

506	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

TOIO NCCA 2010 Summary Report Review Draft ES 20101029 (epa.gov) (last accessed September 14, 2021).

507	USEPA (2015). Report on the Environment. Fish Faunal Intactness. hl:l:ps://cfpub.epa.gov/roe/indical:orxfm?i=84

262

-------
Figure 4.4.2.3-1: Locations of ecoregions and coastal areas defined by the USEPA's
National Aquatic Resource Surveys.508

CPL: Coastal Plains

NAP: Northern Appalachians NPL: Northern Plains



I

m

SAP: Southern Appalachians	SPL: Southern Plains	TPL: Temperate Plains

UMW: Upper Midwest

%

WMT: Western Mountains

liJJX
XER: Xeric



West coast

Gulf coast

<

i

Great Lakes Southeast coast Northeast coast

Excess nutrients (both nitrogen and phosphorus)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
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

508 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.

5°9 paeri; 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.

263

-------
Mexico hypoxic zone was the largest size measured since 1985, spanning 8,776 square miles.510
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.511 While
fertilizer use by current agricultural practices contribute to much of the nutrient loading that
stimulates algal responses in many waterbodies, the total nutrient budgets512 of some
waterbodies also include internal nutrient recycling of legacy inputs.513

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 damage514 in organ tissues and altered muscle and brain
function.515 Some bacteria can use glyphosate for growth, enhancing microbial proliferation.516
There are cyanobacteria with natural tolerance to glyphosate517 and certain concentrations of
glyphosate can stimulate photosynthesis in a common bloom-forming taxon, Microcystis
aeruginosa,518 There is a notable link between glyphosate and phosphorus because more than
18% of glyphosate acid by mass is phosphorus. Glyphosate has chemical similarities with
phosphate ions (competing for the same sorption sites in soil), and glyphosate rapidly degrades
in water and releases phosphorus compounds easily used by organisms for growth. Glyphosate-
derived phosphorus has now reached levels in aquatic systems similar to phosphorus derived
from detergents prior to legislation banning these products, in part because of negative impacts
on aquatic life.519 In 2014, 58% of U.S. rivers and streams were given a rating of poor for the

510	Louisiana Universities Marine Consortium (2017). August 2, 2017 Summary. Shelfwide Cruise: July 24 - July
31. https://gulfhypoxia.net/research/shelfwide-cruise/?y=2017&p=press release. USNOAA (2019). Large 'dead
zone' measured in Gulf of Mexico. Available at https://www.noaa.gov/niedia-release/large-dead-zone-measured-in-
gulf-of-mexico (last accessed April 14, 2021).

511	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.

512	"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.ufl.edu/pdffiles/SS/SS56200.pdf (last accessed April 14, 2021).

513	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.

514	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.

515	Modesto, K.A. and Martinez, C.B. (2010). Roundup® causes oxidative stress in liver and inhibits
acetylcholinesterase in muscle and brain of the fish Prochilodus lineatus. Chemosphere, 78(3): 294-299.

516	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.

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

518	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.

519	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.

264

-------
phosphorus indicator of EPA's National Rivers and Streams Assessment.520 While both corn and
soybean production use glyphosate, corn production can also use atrazine (Table 4.4.2.1-1). In
2012, EPA detected atrazine in 30% of lakes, but concentrations rarely reached the EPA level of
concern for plants in freshwaters (<1% of lakes).521 In 2016, EPA concluded that in areas where
atrazine use is heaviest (mainly in the Temperate Plains ecoregion, Figure 4.4.2.3-1), there are
impacts on aquatic plants and potential chronic risk to fish, amphibians, and aquatic
invertebrates; there is a high probability of changes to aquatic plant assemblage structure,
function, and primary production at a 60-day average concentration of 3.4 ug L 1 and
reproductive effects to fish exposed for several weeks to 5 ug L 1 atrazine.522 When there are
changes to aquatic plant assemblage structure, function, or productivity, other parts of the food
web become at risk because there is reduced food and altered habitat for fish, invertebrates, and
birds. Additional information on the affects to aquatic life will become available as EPA
finalizes their evaluation of the affects the RFS program has on endangered species.

4.4.3	Comparison with Petroleum

Biofuel feedstocks are not the only input to energy production affecting soil and water
quality. For comparison, petroleum used to produce gasoline and diesel fuel also impacts soil and
water quality, but at different spatial and temporal scales than corn and soy. When comparing the
two, it is necessary to consider both the spatial extent of the effects (e.g., the acreage of soil or
volume of water impacted) and the time or effort to recover from any effects. While petroleum
production may have required less land than agriculture in the U.S. between 2007 and 2011,
when considering recovery time or effort, a recent study suggested the effects of petroleum
production can be longer lasting and harder to mitigate (e.g., brine or oil contamination in soil or
groundwater) than those of biofuel feedstocks on soil and water quality.523 A full comparison
between the effects of the two fuel types of energy feedstocks would need to consider both
factors (spatial extent and recovery time or effort), but such an assessment would be expansive
and could not be performed on the timeline of this rulemaking.

4.4.4	Water Quality and Underground Storage Tanks

Releases from underground storage tank (UST) systems can threaten human health and
the environment, contaminating both soil and groundwater. As of September 2021, more than
564,767 UST releases have been confirmed across the United States, averaging about 5,400 per

520	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/nationa1-aquatic-resource-surveys/nrsa (last accessed April
14, 2021).

521	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.

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

523	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.

265

-------
year between 2016 and 2021.524,525 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.526 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.527
Nevertheless, insofar as blends of biofuel with gasoline or diesel are stored in USTs that are
either incompatible with those blends or have incompatible components, the increased
consumption of biofuels could increase leaks that affect water quality.

4.4.5 Potential Future Impacts of Proposed Volume Requirements

Future soil and water quality impacts associated with biofuel volumes will be driven, in
large part, by any associated land use/land cover changes. Directionally, increases in production
of biofuels made from crops would likely lead to an increase in land used for agriculture globally
and in the U.S. There are inherent uncertainties in estimating the amount and type of crop-based
feedstocks needed to fulfill the candidate volumes in this action, but an increase in cropland
acreage would generally be expected to lead to more negative soil and water quality impacts. As

524	USEPA (2021). Frequent questions about underground storage tanks. Available at
https://www.epa.gov/ust/frequent-questions-about-undergfound-storage-tanks (last accessed April 14, 2021).

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

526	40 CFR Part 280.

527	86 FR 5094 (January 19, 2021).

266

-------
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.528 In the case of biogas, there are numerous soil and water quality benefits compared to a
baseline of no manure or waste management. Dairy digesters, for example, are an essential piece
of proper manure management, as once the biogas has been captured, properly aerated manure
can be applied evenly to soil as a fertilizer.529,530 The additional soil and water quality modeling
that would be needed to assess the potential cumulative impacts of future land use changes for
the candidate volumes in this action would be expansive and could not be performed on the
timeline of this rulemaking.

The volume increases for 2023-2025 described in Chapter 3 due to biofuels produced
from agricultural feedstocks (especially corn and soybeans) in comparison to the No RFS
baseline would suggest the potential for an associated increase in crop production, which in turn
may impact soil and water quality. 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. Furthermore, if we consider the potential
impacts relative to the current situation in 2022 (i.e., the 2022 baseline discussed in Chapter 2.2)
there would be little impact, as the overall volume increase for biodiesel and renewable diesel is
much smaller and expected to be met with expanded waste fats, oils, and greases supply.
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.

4.5 Water Quantity and Availability

This section assesses the impact of the production and use of renewable fuels and their
primary feedstocks on the use and availability of water in the U.S. We first review the drivers of
impacts on water use and availability of freshwater resources, summarize impacts to date,
highlight more recent work focused on groundwater supplies. Finally, we discuss the potential

528	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 92213.

529	2010 - US Climate Action Report: Fifth National Communication of the United States of America Under the
United Nations Framework Convention on Climate Change.

530	2011 Annual Report: ENERGY STAR and Other Climate Protection Partnerships.

267

-------
future effects of the candidate volumes as increases in feedstock production such as soybeans
may lead to water impacts.

4.5.1	Drivers of Impacts on Water Use and Availability

Water quantity, in the context of renewable fuels, refers to the volume of water used in
the production of biomass feedstocks (i.e., irrigation of corn, soybeans or other crops) and the
conversion of those feedstocks to biofuel (i.e., water use in the biofuel production plant itself).
The irrigation of corn and soybeans used to produce biofuels is the predominant driver of water
quantity impact and is generally orders of magnitude greater than water use in the biofuel
production process.531 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) 532 "where groundwater
levels have declined at unsustainable rates."533

As noted above, the primary driver of impacts to water quantity is the water used for
irrigation of the biofuel feedstocks. To the extent that feedstock production expands into regions
where irrigation is required, the demand for water will increase, whether the expansion is a direct
consequence of production specifically for biofuel feedstocks or an indirect result of increased
production for all feedstock uses. Water demand for biofuel production processes can also drive
impacts on water use and availability. Although water demands of biofuel production facilities
may be much smaller at a national scale than the water demands of irrigated feedstock
production, biofuel facility water use may be locally consequential in areas that are already
experiencing stress on water availability.

4.5.2	Life Cycle Water Use of Biofuels

In the Second Triennial Report to Congress on Biofuels, the water quantity impacts of
biofuels were assessed.534 Research investigating the water quantity impacts of biofuels started
shortly after the passage of the Energy Policy Act of 2005. Several highly cited and visible
articles compared the life cycle water use of biofuels relative to petroleum-based fuels on the
basis of "gallons of water per mile" or "gallons of water per gallon of fuel." 535 These early

531	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.

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

533	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.

534	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.

535	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.

268

-------
studies characterized this issue as biofuel's water intensity,536 embodied water,537 and water
footprint.538 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).539 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.540 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

CA 1

transportation energy sources.

4.5.2.1 Feedstock Production

Researchers have continued to refine the LCA-based water footprint of biofuels—with a
focus on feedstock production—for both current biofuels crops and future feedstocks. Because
more than 90% of corn is grown in rain-fed areas where corn production is non-irrigated, Wu et
al. (2014) suggested that, at the highly aggregated level, the "national water footprint of corn is
consistently low to modest." 542 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).543 Much of the focus has been on corn ethanol, due to the higher volumes
of corn ethanol produced to date. However, in the same article, Dominguez-Faus et al. estimated
that irrigated soybean based biodiesel water requirement averaged nearly 1,300 liters of water

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

537	Chiu, Y. W., Walseth, B., & Suh, S. W. (2009). Water embodied in bioethanol in the United States.
Environmental Science & Technology, 43(8), 2688-2692.

538	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.

539	Another category is the gre/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.

540	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.

541	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.

542	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.

543	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.

269

-------
per liter of ethanol-equivalent biodiesel (based on energy equivalence).544 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.545 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.546 Other researchers
have similarly focused on the wide range of water intensity estimates between rain-fed and
irrigated counties and among a variety of crops (see Figure 4.5.2.1-1). Gerbens-Leenes et al.
(2012) estimated Nebraska's blue water (irrigation water) footprint at three times higher than the
U.S. weighted average blue water footprint.547 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.548 Additionally, the blue water footprint in
areas that have already stressed water sources, like the HPA, could experience more severe water
quantity impacts. A report by the National Academy of Sciences (NAS 2011) highlighted the
groundwater depletion in the HPA, 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." 549 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. 20 1 2).550

544	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.

545	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.

546	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.

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

548	EIA (2018). "Six states account for more than 70% of U.S. fuel ethanol production." Available at
https://www.eia.gov/todayinenergy/detail.php ?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.

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

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

270

-------
Figure 4.5.2.1-1: Estimate of the blue/irrigation, green/rainwater and grey/pollution water
footprint associated with corn grain, stover, wheat straw and soybean during the crop
growing phase.

« Regional range
* County max
-National average

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 min." [Source: Chiu and Wu
(2012)].

4.5.2.2 Biofuel Processing

Studies of water use at biofuels conversion facilities have generally quantified water
consumption as gallons of water per gallon of biofuel produced, mostly concentrating on ethanol,
especially dry mill facilities.551 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 20 1 2.552 These values are typical of a dry mill facility. Wet mill
facilities require closer to 4 gallons per gallon of ethanol.553 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

551	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.

552	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 a/so 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.

553	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. Environmental science & technology.

271

-------
cooling towers.554 Some facilities have set goals to both reduce water use and minimize
discharges.

Biodiesel conversion from oil crops such as soybeans requires water use for multiple
stages of the process (crop to oil, and oil to biodiesel). The soybean processing (crop to oil) stage
involves crushing, oil extraction and crude soybean oil refining (degumming). Water
consumption includes make-up water for cooling towers and other processes. In the biodiesel
production stage (oil to biodiesel), following the crushing and oil extraction steps, water is used
to remove residual glycerol, a by-product of the transesterification process, and other impurities,
while some water is also for additional make-up water for cooling towers.555 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).556

Renewable diesel is chemically very similar to petroleum-based diesel despite being
made from renewable wastes such as fats and vegetable oils. This means that it is processed in
the same manner that petroleum diesel which is hydrotreating. With this knowledge it can be
assumed that the same about of water used for processing petroleum diesel is used to process
renewable diesel.557 As renewable diesel is a plant-based feedstock, its additional water usage
would be from the irrigation process to grow the plants used to create the oil and not from the
process itself.

There are no recently published surveys of water consumption representing all current
biofuel and renewable fuel facilities, and no comprehensive data on the type of water sources
utilized (e.g., groundwater, surface freshwater, public supply, etc.). Grubert and Sanders estimate
that the majority of the water used is freshwater. There is also some evidence that groundwater
from aquifers is being extracted for use in ethanol production in states such as Iowa and
Nebraska,558 and likely a source of water for facilities along the HPA (see Figure 4.5.3-3).

4.5.2.3 Summary and Comparison to Petroleum

Improvements in irrigation have brought down the upper range of water use, with recent
estimates of irrigation for corn production ranging from 9.7 gal/gal ethanol for USD A Region 5
(Iowa, Indiana, Illinois, Ohio and Missouri) to 220.2 gal/gal ethanol in Region 7 (North Dakota,

554	Schill, S. R. (2017) Water: Lifeblood of the Process. Ethanol Producer Magazine. January 24, 2017.
http://www.ethanolproducer.coni/artLcles/14049/water-Hfeblood-of-the-process. See also Jessen, H. (2012) Dropping
Water Use. Ethanol Producer Magazine. June 12, 2012. Available at

http://www.ethanolproducer.com/artLcles/8860/droppLng-water-use (last accessed September 16, 2021).

555	Tu, Q., Lu, M., Yang, Y. J., and D. Scott (2016) Water consumption estimates of the btodLesel process In the US.
Clean Technologies and Environmental Policy. 18(2): 507-516.

556	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.

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

558	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.

272

-------
South Dakota, Nebraska and Kansas) for groundwater.559 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.560 561 The Wu et al (2018) analysis does not include biodiesel. The most
recent estimate for the full LCA water consumption for biodiesel provides a range of values for
each state: Missouri 21-79 gal water/gal biodiesel, Kansas and Oklahoma 80-150 gal/gal,
Nebraska and Texas 150-300 gal/gal.562 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.563 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.564 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.

4.5.3 Impacts to Date

Because the majority of the growth in biofuels production has come from corn- and soy-
based biofuels, the water consumption impacts to date would have come from additional water
use for corn and soybean acreage. To our knowledge, there have been no comprehensive studies
of the changes in irrigated acres, rates of irrigation, or changes in surface and groundwater
supplies attributed specifically to the increased production of corn grain-based ethanol and

559 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://publications.anl.gov/anlpubs/2019/01/148043.pdf

56°

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

562	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.

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

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

273

-------
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.565 These are areas unlikely to have sufficient
precipitation to support corn or soybean cultivation.566 While difficult to attribute how much
additional water use might be required as a result of the candidate volumes in this rule, there are
several lines of evidence that suggest increased production of corn-based ethanol and soybean-
based biodiesel will increase water demands and, potentially, affect limited water supplies.

The USD A Irrigation and Water Management Surveys (formerly the Farm and Ranch
Irrigation Survey or FRIS), a supplement to the Census of Agriculture completed every five
years, provide a general indication of the changes in water demands between 2013 and 2018.567
From 2013 to 2018, there was an increase in total irrigated acres of nearly 0.6 million acres in the
U.S.568 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 20 1 3.569 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.570 Figure 4.5.3-1 shows acres of irrigated land in
2012, the most recent year of data for which this figure is available.

565	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.

566	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).

567	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 Survey/fri
s.pdf (last accessed April 14, 2021).

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

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

569	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 Survey/fri
s.pdf (last accessed April 14, 2021).

570	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 Survey/fri
s.pdf (last accessed April 14, 2021).

274

-------
Figure 4.5.3-1: Acres of irrigated land in 2012

management/irrigation water-use/background.aspx

Figure 4.5.3-2 shows corn and soybean areas and share of irrigated acres. Irrigated corn
grain/seed acres are heavily concentrated in Nebraska (4.5 million acres) followed by Kansas
(1.3 million acres). This is a decrease of 0.9 and 0.2 million acres respectively from 2012 to
2018. Irrigated soybean acres are also found in Nebraska, Kansas, particularly the more western
part of those states. Overall soybean production is generally more concentrated (as a share of
total harvested cropland) in rainfed areas, whereas corn production reaches further west. There is
also a high percentage of soybean acres in Arkansas and Mississippi, with a large share of those
soybean acres being irrigated. The top rows of Figure 4.5.3-2 show the distribution of corn and
soybean acres, as a share of total cropland acres, while the bottom rows of Figure 4.5.3-2 show
the percent of irrigated corn and soybean acres relative to total acres for those crops (measures as
harvested acres).

275

-------
Figure 4.5.3-2: Percent of irrigated corn and soybean acres relative to total acres,
respectively

Acr.&o'Garn Hnrastec flcrOrcln a: PfircsYta* Ha\wlf..1 CroplaidAc^a^: >317

Aarar cf S^tftBra Hnivfivwl for aeora *i Pcroci; of	Groptanfl Acreopr ?D I7

¦ n^iUc "arn fa-Urar. harvt,&fc»J Ao« At FwctuK. a' Com kM Gcaii, Ha^ssted Aoai UU1 /

Irrigaseti Soybeansfor Bean*. HsrvwiWAerw.« P«ic«r5ol Soytwaire for Beats. Harvsefed Aero. 201 r

Top left: Acres of Corn Harvested as a Percent of Harvested Cropland Acreage,

Top right: Acres of Soybean Harvested as a Percent of Harvested Cropland Acreage.

Bottom left: Irrigated Corn as a Percent of Total Corn (Harvested Acres).

Bottom right: Irrigated Soybeans as a Percent of Total Soybeans (Harvested Acres).

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

https://www.nass.usda.gov/Publications/AgCensus/2Q12/Online Resources/Ag Census Web Maps/index.php

Higher irrigation demands may coincide with areas of already-stressed surface and
groundwater resources, such as the HPA (also called the Ogallala Aquifer). A 2011 report by the
National Academy of Sciences highlighted the groundwater drawdown in the HPA, 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."1,1
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.

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

276

-------
(2017) of land use change rates noted that "along the Ogallala Aquifer, elevated rates of land use
change to corn production in Western Kansas, Oklahoma and Texas coincided with areas
experiencing groundwater depletion rates ranging from 5-20% per decade" (see Figure 4.5.3-3).
However, this correlation does not necessarily mean there is a direct, causal relationship between
biofuel production and groundwater depletion.

(2008-2012).

H Ogallala Aquifer

Non-crop to crop (%}

— I I H

0 ,4 .8 1.2 2.0 2.7 3.9 5.5 7.8 12.1 35

Includes conversion located along the Ogallala aquifer, Stars denote biofuel production facilities, (Source: Wright et
al. 2017)

As stated above, there have been no comprehensive studies of the changes in irrigated
acres, rates of irrigation, or changes in surface and groundwater supplies attributed specifically to
the increased production of corn grain-based ethanol and soybean-based biodiesel. In the absence
of analyses that do focus directly on crops for biofuel production, there are studies that look
more broadly at the connection between agricultural water use and groundwater levels. For
example, Smidt et al. (2016) analyzed the water-energy-food nexus over the HPA 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 HPA, "groundwater levels have declined at
unsustainable rates despite improvements in both the efficiency of water use and water

-Minnesota

1 /

Jmtfi bat

.* Nebraslto

Indiani

Color tide

JVf/sjOiirf

Virginia

r

COxtatobmp-1

Tennessee ^	NotthCahfha

Mi- I-	. Sv /

- ' V v	Carofifa -

"J t •	if ¦. /

f'Pf»Wfado.mo\	j*

I ' ' -"1 • h ••	,*'A 1-*^

Arkansas

Figure 4.5.3-3: Relative conversion rates of arable non-cropland to cropland

277

-------
productivity in agricultural practices."572 Figure 4.5.3-3 shows the relative conversion rates of
arable cropland to non-cropland, as well as the location of the HPA and biofuel conversion
facilities. Figure 4.5.3-4 also shows the HPA, but shows the absolute changes in groundwater
levels, from predevelopment to 2017 based on data from USGS.573 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 Laboratory574) 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 HPAs have seen substantial declines,
the Northern HPA has remained relatively stable and even increased in some areas (as shown in
Figure 4.5.3-4). This does not demonstrate that biofuel production causes declines in
groundwater levels, but it does show that some biofuel facilities operate in areas that are
experiencing water-stressed aquifer resources.

572	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.

573	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.usgs.gov/projects/HPA/index.htnil

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

278