Document Rj1m66ygROz7668b9OZYmDMqv

Comments of URBAN AIR INITIATIVE, CLEAN FUELS DEVELOPMENT COALITION, 25x`25 ALLIANCE, NEBRASKA ETHANOL BOARD, and NEBRASKA ETHANOL INDUSTRY COALITION On the U.S. Environmental Protection Agency's Proposed Rule RENEWABLE FUEL STANDARD PROGRAM: STANDARDS FOR 2018 AND BIOMASS-BASED DIESEL VOLUME FOR 2019 Docket ID No. EPA-HQ-OAR-2017-0091 82 Fed. Reg. 34,206 (July 21, 2017) August 31, 2017 by C. Boyden Gray Adam R.F. Gustafson Andrew R. Varcoe James R. Conde Boyden Gray & Associates PLLC 801 17th Street NW, Suite 350 Washington, DC 20006 ........Ex. 6 ........ i____________________________________ i gustafson@boydengrayassociates.com Sierra Club v. EPA 18cv3472 NDCA Tier 3/4 ED 002061 00161533-00001 Executive Summary Urban Air Initiative, Clean Fuels Development Coalition, 25x`25 Alliance, Nebraska Ethanol Board, and Nebraska Ethanol Industry Coalition (Commenters) respectfully submit these comments on the United States Environmental Protection Agency's Proposed Rule: Renewable Fuel Standard Program: Standardsfor 2018. In the Proposed Rule, EPA continues to ignore new data concerning ethanol's lifecycle emissions of greenhouse gases (GHG). EPA last conducted a lifecycle analysis (LCA) in its regulatory impact analysis accompanying the 2010 Renewable Fuel Standard (RFS) Rule. Seven years later, EPA continues to rely on its outdated 2010 LCA to meet its cost-benefit analysis obligations and to approve pathways under the RFS. Despite EPA's recognition that the Proposed Rule is "an economically significant regulatory action," EPA admits that it "ha[s] not quantified benefits for the 2018 proposed standards." EPA is required by Executive Order to "use the best available techniques to quantify anticipated present and future benefits and costs as accurately as possible." But the Proposed Rule offers merely an "illustrative" analysis of costs limited to wholesale fuel costs and justifies its failure to conduct a full cost-benefit analysis by pointing to the 2010 LCA. This does not satisfy EPA's cost-benefit obligation, because the Agency has failed to update the 2010 LCA, despite "committing" in 2010 "to further reassess. . . the lifecycle estimates." In addition, EPA's continued reliance on its outdated 2010 LCA increases RFS compliance costs by making it harder for existing ethanol producers to qualify under the 20% threshold needed to generate non-grandfathered RINs. EPA's continued reliance on the 2010 LCA is improper. The best available science shows that blending ethanol into gasoline reduces emissions of GHGs far more than EPA projected in 2010. In particular, new evidence shows that: Increased demand for corn causes much less land-use change and related emissions than EPA predicted in 2010. This evidence includes improved economic models and newly available land-use data from periods of increasing corn ethanol production, which show significant increases in yield but no significant increases in forest conversion. Sierra Club v. EPA 18cv3472 NDCA l Tier 3/4 ED 002061 00161533-00002  Improved agricultural practices and technologies are substantially reducing the carbon intensity of ethanol by increasing the soil carbon that is captured from the atmosphere by the corn plant and retained deep below ground. This evidence includes a growing body of science demonstrating that conservation tillage practices sequester more carbon in the soil than previously thought. In fact, the evidence suggests that many com fields are net carbon "sinks," capturing more carbon than land-use change and corn farming releases. More efficient agricultural practices and technologies have reduced nitrogen fertilizer losses of the greenhouse gas nitrous oxide (N20 ), and updated guidance has reduced the weight given to N 20 compared to other GHG pollutants. Ethanol plants have become much more efficient, as yields have continued to increase. Ethanol plants are also producing new co products that reduce the carbon intensity of ethanol. In addition to distillers' grains, used as animal feed, ethanol plants now produce corn oil, which replaces soy-based biodiesel. By contrast, petroleum-based fuels are becoming increasingly carbon intensive. As a result, the gasoline carbon intensity baseline is higher than EPA suggested, increasing the comparative benefit of com ethanol. k k k A review of the scientific literature confirms that EPA fundamentally erred in the conclusions it reached in 2010 about the lifecycle GHG emissions of corn ethanol. A recent study by the Department of Agriculture estimates that corn ethanol produces 43% and 48% less greenhouse gas emissions than EPA's gasoline baseline, in 2014 and 2022, respectively, without fully accounting for soil carbon sequestration. But despite a growing body of updated scientific studies, EPA continues to rely on its 2010 EGA in the Proposed Rule. We urge EPA to correct its 2010 LCA or adopt USDA's updated model and to conduct a new cost-benefit analysis in light of the best available science. Sierra Club v. EPA 18cv3472 NDCA h Tier 3/4 ED 002061 00161533-00003 Table of Contents Executive Summary........................................................................................................... i Table of Contents............................................................................................................. iii Introduction........................................................................................................................ 1 I. The Commenters' Interest in EPA's 2010 L C A ................................................. 3 II. EPA Continues to Rely on its Outdated 2010 LCA............................................ 4 III. EPA's Greenhouse Gas Emissions EstimatesAre Erroneous........................... 6 A. Corn Production...................................................................................... 8 1. International Land-Use Change Emissions........................................9 2. Domestic Land-Use Change Emissions...........................................13 3. Domestic Farm Inputs and Fertilizer N 20 ...................................... 15 4. International F arm Inputs and Fertilizer N20 .................................17 B. Ethanol Fuel Production.......................................................................18 1. Ethanol Plant Yields........................................................................... 18 2. Corn O il...............................................................................................18 C. Gasoline Lifecycle Emissions............................................................... 19 C o n c lu s io n ....................................................................................................................... 21 Appendix.......................................................................................................................... 23 Sierra Club v. EPA 18cv3472 NDCA iii Tier 3/4 ED 002061 00161533-00004 I n t r o d u c t io n In 2010, EPA conducted a comprehensive lifecycle analysis of com ethanol and gasoline in support of its RFS program (2010 LCA).1EPA's 2010 LCA included GHG emission inventories based on future industry projections and the scientific evidence available at the time.2As EPA noted, that data was subject to many uncertainties.3EPA "recognizejd] that as the state of scientific knowledge continues to evolve in this area, the lifecycle GHG assessments for a variety of fuel pathways will continue to change."4EPA therefore committed to "further reassess. . . the lifecycle estimates" on an ongoing basis,5and to incorporate "any updated information we receive into a new assessment of the lifecycle GHG emissions performance of the biofuels being evaluated in [the 2010] rule."6 As EPA predicted in 2010, new science now shows that its past projections no longer represent "the best available information."7As summarized in a recent lifecycle analysis report commissioned by the Department of Agriculture (USDA), "a large body of information has become available since 2010--including new data, 1See Renewable Fuel Standard Program, Regulatory Impact Analysis (2010) [hereinafter 2010 RFS RIA], The Energy Independence and Security Act requires EPA to estimate lifecycle emissions, including emissions from land-use change. See 42 U.S.C. 7545(o)(l)(H). 2Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program, 75 Fed. Reg. 14,670, 14,785 (Mar. 26, 2010) [hereinafter 2010 RFS Rule] (representing that the 2010 LCA included the "most up to date information currently available on the GHG emissions associated with each element of the full lifecycle assessment"). 3Id. at 14,677, 14,765, 14,785. To illustrate the magnitude ofEPA's scientific uncertainty, while EPA estimated a GHG reduction of 21% for corn ethanol in 2022 using advanced pathways, EPA's "95% confidence interval" ranged from a 7% to a 32% reduction. Id. at 14,786. 4I d at 14,765. 5Id. at 14,765 ("Therefore, while EPA is using its current lifecycle assessments to inform the regulatory determinations for fuel pathways in this final rule, as required by the statute, the Agency is also committing to further reassess these determinations and lifecycle estimates."); accord id. at 14,785. 6 Id. 1 2010 RFS Rule, supra note 2, 75 Fed. Reg. at 14,785. Sierra Club v. EPA 18cv3472 NDCA 1 Tier 3/4 ED 002061 00161533-00005 scientific studies, industry trends, technical reports, and updated emission coefficients--that indicates th a t. . . actual emissions. . . differ significantly from those projected" by EPA's 2010 LCA.8As the USDA study demonstrates, corn ethanol results in less GHG emissions than EPA predicted in its 2010 LCA. Thus, the best available science demonstrates that blending ethanol into gasoline lowers GHG emissions. But despite this growing body of evidence, and despite EPA's assurances that it would reassess its initial estimates as the science evolved, the Proposed Rule fails to update EPA's 2010 cost-benefit analysis to include updated lifecycle emissions information. Instead of performing a comprehensive cost-benefit analysis based on the best available science, the Proposed Rule provides an "illustrative cost analysis for the proposed reductions" based solely on wholesale fuel costs.9And the Proposed Rule attempts to justify its omission by pointing out that the relevant costs and benefits, including "GHG emissions," "were analyzed in the 2010 [LCA]."10 EPA's failure to update its lifecycle analysis affects more than the cost-benefit analysis of the present rule. EPA also continues to evaluate corn ethanol producer pathway petitions based on the same "feedstock modeling . . . done as part of the March 2010 [LCA]."11This makes it harder for new producers of renewable fuel to demonstrate their eligibility for RINs under the RFS. EPA should update its lifecycle analysis to reflect the best available science. Part I of these comments describes the commenters' interest in the accuracy of EPA's lifecycle analysis of ethanol and gasoline. Part II explains how EPA continues to rely on its outdated 2010 LCA. Part III summarizes the best available science on IGF, A Life-Cycle Analysis of the Greenhouse Gas Emissions of Corn-Based Ethanol 4-5 (Jan. 12, 2017) [hereinafter 2017 USD A LCA], 9Renewable Fuel Standard Program: Standards for 2018 and Biomass-Based Diesel Volume for 2019, 82 Fed. Reg. 34,206, 34,237 (July 21, 2017) [hereinafter Proposed Rule]. 11 EPA, Al-Corn Clean Fuel Pathway Determination under the RFS Program 7 (Aug. 15, 2017) [hereinafter EPA, Al-Corn Determination].2 2 Sierra Club v. EPA 18cv3472 NDCA Tier 3/4 ED 002061 00161533-00006 the GHG emission effects of corn ethanol and gasoline and explains why EPA's 2010 LCA is inaccurate. I. T he C o m m e n t e r s' Interest in EPA's 2010 LCA Urban Air Initiative is a non-profit organization dedicated to improving air quality and protecting public health by reducing vehicle emissions. UAI is focused on increasing the use of clean burning ethanol in our gasoline supply to replace harmful aromatic compounds in gasoline. UAI is helping meet public policy goals to lower emissions and reduce carbon in the environment through scientific studies and real-world data to promote new fuels, engine design, and public awareness. The Clean Fuels Development Coalition was established in 1988 and works with auto, agriculture, and biofuels interests in support of a broad range of energy and environmental programs. 25x`25 Alliance is a national coalition united behind the goal of securing 25 percent of the nation's energy needs from renewable sources by the year 2025. The 25x`25 goal has been endorsed by nearly 1,000 partners, 35 current and former governors, 15 state legislatures and the U.S. Congress through the Energy Independence and Security Act of 2007. The Nebraska Ethanol Board is a state agency supporting ethanol development programs throughout the state, and assisting the industry with a range of technical, marketing, and regulatory issues. The Nebraska Ethanol Industry Coalition is a statewide non-profit organization working together on issues of common interest to their members with a particular focus on market development and expansion. Because the best available science shows that ethanol is cleaner and gasoline dirtier than EPA believed in 2010, EPA's continued use of its 2010 LCA frustrates the commenters' mutual interest in advancing a clean, low-carbon energy future while reducing harmful air pollution. Sierra Club v. EPA 18cv3472 NDCA 3 Tier 3/4 ED 002061 00161533-00007 II. EPA Continues to Rely on its Outdated 2010 LCA. EPA correctly classifies the Proposed Rule as "an economically significant action" subject to regulatory review under the relevant Executive Orders.12 Therefore, "in deciding . . . how to regulate[,]" EPA "should assess all costs and benefits of available regulatory alternatives" and "select those approaches that maximize net benefits."13In assessing the costs and benefits of the Proposed Rule, EPA must "use the best available techniques to quantify anticipated present and future benefits and costs as accurately as possible."14Moreover, EPA must "ensure the objectivity of any scientific and technological information and processes used to support the agency's regulatory actions."15In short, the Proposed Rule must be "based on the best available science."16 But in the Proposed Rule, EPA's analysis "do[es] not take into consideration the benefits of the program."17Eschewing "comprehensive estimates" of the Proposed Rule's costs and benefits, EPA offers only simplistic analyses of the cost of producing the additional volumes of ethanol required to be blended with the cost of producing an energy-equivalent amount of gasoline.18EPA provides these estimates "solely for the purpose of illustrating how the cost to produce a gallon of `representative' renewable fuels could compare to the costs of producing petroleum fuels"19--not to assess the actual costs and benefits of its annual standard. For 12Proposed Rule, supra note 9, 82 Fed. Reg. at 34,243. 13Exec. Order No. 12,866 1(a), 3 C.F.R. 638 (1994), reprinted as amended in 5 U.S.C. 601 app. at 45-49 (2006). 14Exec. Order No. 13,563 1(c), 76 Fed. Reg. 3,821, 3,821 (Jan. 18, 2011). 15Id. 5, 76 Fed. Reg. at 3,822. 16Id. 1, 76 Fed. Reg. at 3,821. 17Proposed Rule, supra note 9, 82 Fed. Reg. at 34,238. 18Id. at 34,237. Sierra Club v. EPA 18cv3472 NDCA 4 Tier 3/4 ED 002061 00161533-00008 analysis of the impacts of the RFS generally, the Proposed Rule directs the reader to EPA's cost-benefit analysis performed "in the 2010 final rulemaking."20 But EPA's cost-benefit analysis was premised on its erroneous 2010 LCA. EPA must update its analysis of the lifecycle emissions of ethanol and gasoline to enable a comprehensive evaluation of the costs and benefits of the Proposed Rule and the RFS program as whole. In addition, EPA continues to rely on its 2010 LCA to implement the RFS. Under the RFS program, non-grandfathered ethanol fuel must "achieve[] at least a 20 percent reduction in in lifecycle greenhouse gas emissions compared to baseline lifecycle greenhouse gas emissions."21 In its March 2010 rule, EPA finalized pathways that corn ethanol producers could use to generate com ethanol renewable identification number credits under the RFS.22These pathways are based on EPA's 2010 LCA, which concluded that by 2022, com ethanol plants using natural gas and com oil fractionation technology would achieve annual lifecycle greenhouse gas (GHG) emissions savings of only 21% compared to EPA's 2005 gasoline carbon intensity baseline of 93 grams of carbon dioxide equivalent per megajoule (g C 0 2e/M J).23By contrast, EPA predicted that grandfathered ethanol plants that do not use advanced technologies would 21 42 U.S.C. 7545(o)(2)(A)(i). 22 40 C.F.R. 80.1426(f)(1). 23 2010 RFS Rule, supra note 2, 75 Fed. Reg. at 14,786 ("The results for this corn ethanol scenario are that the midpoint of the range of results is a 21% reduction in GFIG emissions compared to the gasoline 2005 baseline."); 2010 RFS RIA, supra note 1, at 469-70. EPA's central estimate of com ethanol's carbon intensity in 2022 using these technologies was 79 kg C 0 2e/mmBTU (million British thermal units), id. at 14,788, which is equivalent to 74.9 g C 0 2e/MJ. EPA reported the carbon intensity baseline for 2005 gasoline at 98.2 kg C 0 2e/mmBTU, which is equivalent to 93.1 g C 0 2e/MJ. 2010 RFS RIA, supra note 1, at 467. The conversion factor used to convert kg C 0 2e/mmBTU to g C 0 2e/M J is 0.947817. All carbon intensity numbers are rounded to a single decimal. Sierra Club v. EPA 18cv3472 NDCA 5 Tier 3/4 ED 002061 00161533-00009 achieve only a 16.8% reduction in GHG emissions relative to 2005 gasoline.24EPA continues to use its 2010 LCA to evaluate ethanol producer petitions.25 Over half of the assessed GHG emissions of the typical grandfathered ethanol plant are estimated "upstream emissions" over which these ethanol plants have no control--the emissions from changes in international land-use patterns, as well as domestic farm inputs and emissions from fertilizer.26This high estimate of upstream emissions makes it more difficult for new com ethanol producers to qualify to generate non-grandfathered RINs, and this in turn increases the cost of RINs. EPA should update its 2010 LCA to reduce these compliance costs. III. E P A 's G r e e n h o u s e G a s E m issio n s E stim ates A re E r r o n e o u s . While EPA's findings were doubtful in 2010, they are now demonstrably erroneous, given the wealth of newly available scientific and economic data that undermines EPA's 2010 LCA. As a lifecycle analysis of com ethanol GHG emissions commissioned by USDA recently found, "a large body of information has become available since 2010--including new data, scientific studies, industry trends, technical reports, and updated emission coefficients--that indicates th a t. . . actual emissions. . . differ significantly from those projected" by EPA's 2010 LCA.27Using this updated information, USDA's study--which largely tracks the methodology of EPA's 2010 LCA--estimates that in 2014 com ethanol was 43% less carbon intensive than EPA's 2005 gasoline baseline, and that corn ethanol's advantage will grow to 48% by 2022.28This is a much greater benefit that EPA's median estimate 24EPA, Al-Corn Determination, supra note 11, at 11, Table 2. 25 Id. 26Id.] 2010 RFS RIA, supra note 1, at 470. 27 2017 USDA LCA, supra note 8, at 4-5. 28 Id. at 166. The study estimated corn ethanol's lifecycle emissions at 55,731 g C 0 2e/MMBtu in 2014, equivalent to 52.8 g C 0 2e/MJ, id. at 151, and at 50,553 g C 0 2e/MMBtu in 2022, id. at 166, equivalent to 47.9 g C 0 2e/MJ. Sierra Club v. EPA 18cv3472 NDCA 6 Tier 3/4 ED 002061 00161533-00010 that corn ethanol will produce only 21% less greenhouse gas emissions than baseline gasoline in 2022.29 Studies by the Department of Energy confirm that EPA's 2010 LCA understates com ethanol's carbon reduction benefit. The Department of Energy's influential model of transportation sector GHG emissions (the GREET model) estimated a 35% lifecycle GHG emissions reduction for corn ethanol produced in 2015 compared to 2005 gasoline.30And Department of Energy scientists have suggested that further improvements in corn ethanol production "could render com ethanol as having a 50% reduction in life-cycle GHG emissions as compared to gasoline."31 Com ethanol's relative carbon intensity is even lower than these numbers suggest, because the carbon intensity of gasoline has increased since 2005, even as ethanol's carbon intensity has steadily fallen.32 EPA should evaluate the costs and benefits of ethanol blending in light of the best available science concerning the lifecycle emissions of ethanol and gasoline. Section A will address upstream emissions from corn production, including indirect emissions from land-use change, emissions from domestic land-use change, and emissions from domestic and international farm input and fertilizer nitrous oxide 29 2010 RFS Rule, supra note 2, 75 Fed. Reg. at 14,786. 30 See Zhichao W ang et al., Influence o f Corn Oil Recovery on Life-Cycle Greenhouse Gas Emissions o f Corn Ethanol and Corn Oil Biodiesel, 8 Biotechnol. Biofuels 178, 178, 183, Fig. 3 (2015) (using GREET2015 to estimate an average corn-ethanol carbon intensity of 62 to 59 g C 0 2e/MJ); Susan Boland & Stefan Unnasch, Life Cycle Associates, GHG Emissions Reductions Due to RFS, LCA.6075.11.2015, at 9 (2015) (using GREET2015 to estimate an average corn ethanol carbon intensity of 59.2 g C 0 2/MJ). 31 Wang et al., supra note 30, at 186. 32 Amgad Elgowainy et al., Energy Efficiency and Greenhouse Gas Emission Intensity o fPetroleum Products a t U.S. Refineries, 48 Envtl. Sci. & Tech. 7612, 7623 (2014) (estimating that the "total life-cycle GHG emissions for gasoline" are 94 g C 0 2e/MJ); see also Hao Cai et al., Well-to- Wheels Greenhouse Gas Emissions o f Canadian Oil Sands Products: Implicationsfor U.S. Petroleum Fuels, 49 Envtl. Sci. & Tech. 8219 (2015) (predicting greater emissions due to the growing share of Canadian oil sands gasoline in the U.S. market). Sierra Club v. EPA 18cv3472 NDCA 7 Tier 3/4 ED 002061 00161533-00011 (N20 ) emissions.33Section B will address biorefmery emissions. Section C will discuss gasoline's lifecycle emissions. A. Com Production EPA's estimate of "upstream emissions" from corn production (and its alleged indirect effects), accounts for the majority of the GHG emissions that the 2010 LCA attributes to corn ethanol.34Within upstream emissions, international land-use change emissions (ILUC) account for the greatest fraction (40%) of EPA's estimate of com ethanol's carbon intensity, followed by domestic farm input and fertilizer emissions (13%) and international farm input and fertilizer emissions (7%).33 See Figure 1. New evidence has exposed wx:::::: Fuel Prod uctio n international Land Use Change .............In ternational F arm Inputs and Fed 8888888888 urtw ii- (fuel and feedstock significant flaws in EPA's estimate of corn ethanol's upstream GHG emissions. Updated models and empirical evidence of actual land-use patterns demonstrate that 2005 G asoline Baseline 2022 A vg Dry M ill MG Figure 1:2010R F S RIA (Figure2.6-2) SSSSSSSSSS: D om estic Land Use Change 8888888888International Livestock 8888888881international R ice M ethane Tailpipe Dom estic Rice Methane 18888888881D om estic Faim inputs and Fert N20 8888888888D om estic Livestock " - 20% Reduction Threshold i Net Em issions carbon emissions from land-use change are much lower than the estimate in EPA's 2010 LCA. EPA's assessment of domestic and international farm input and fertilizer N 20 emissions, are also outdated and in need of correction. As explained below, correcting these upstream emission estimates based on the updated science noted in 33 These comments do not address all GHG emission categories included in EPA's 2010 LCA. For a comprehensive, updated analysis, see 2017 USDA LCA, supra note 8. 342010 RFS RIA, supra note 1, at 470, Figure 2.6-2. 35 See id. 8 Sierra Club v. EPA 18cv3472 NDCA Tier 3/4 ED 002061 00161533-00012 the USDA 's study would reduce EPA's estimate of corn ethanol's upstream emissions in these categories from 45.5 gto 11.1 g C 0 2e/M J in 2022--a 76% reduction. See Table 1. Table 1: EPA Upstream Emissions Compared to Updated USDA Upstream Emissions EPA Upstream 2022 Emissions (gC02e/MJ) ILUC 30.3 Dom. Farm 9.8 Inti. Farm 5.4 Total 45.5 USDA 2014 (g C02e/MJ) 1.3 8.6 2.1 12.0 A (g C02e/MJ) -29.1 -1.2 -3.3 -33.6 USDA 2022 BAU (g C02e/MJ) 1.3 7.8 2.1 11.1 A (g C02e/MJ) -29.1 -2 -3.3 -34.4 In addition, EPA's analysis of domestic land-use change does not account for the adoption of crop management techniques that improve soil carbon sequestration in corn croplands, particularly when combined with com ethanol's high yields. Accounting for these practices would further reduce corn ethanol's emissions. 1. International Land-Use Change Emissions EPA's 2010 LCA estimated ILUC emissions for corn ethanol in 2022 at 30.3 g C 0 2e/M J, accounting for 40% of corn ethanol's estimated carbon intensity.36As EPA explained, "the majority of international land use change emissions originate in Brazil. . . . This is largely as a consequence of projected pasture expansion . . . especially in the Amazon region where land clearing causes substantial GHG emissions."37Indeed, in EPA's 2010 LCA, more than two-thirds of corn ethanol's predicted ILUC emissions were attributable to predicted land-use changes in Brazil.38 36 20 10 RFS RIA, supra note 1, at 407, Table 2.4-47 (estimating ILUC at 31.8 kg C 0 2e/mmBTU); 2010 RFS Rule, supra note 2, 75 Fed. Reg. at 14,788. 3' 2010 RFS RIA, supra note 1, at 470. op Id. at 470 (showing that 22 out of 31.8 kg C 0 2e/mmBTU are attributable to Brazil). Sierra Club v. EPA 18cv3472 NDCA 9 Tier 3/4 ED 002061 00161533-00013 At the time, EPA acknowledged that these results were subject to great uncertainty.39In fact, the estimates reported in these early analyses were never accurate, and they have since been refuted by the best available science. Parameters related to intensification, yield improvement, land displacement, and the type of land converted are key drivers of ILUC emissions, but EPA's models failed to accurately reflect these complexities. For example, EPA's ILUC model does not "distinguish what types of land will be affected by a given shock to the agricultural system."40More recent models of indirect land-use change have included "a more detailed assessment of yield improvement, land cover type, soil carbon stocks, and other parameters," resulting in significantly lower estimates of land-use change emissions.41 EPA's ILUC assessment in 2010 relied on outdated economic models developed by the Food and Agricultural Policy and Research Institute, maintained by the Center for Agricultural and Rural Development (FAPRI-CARD).42EPA also "opted to use the GTAP [Global Trade Analysis Project] model to inform the range of potential GH G emissions associated with land use change resulting from an increase in renewable fuels."43 Since 2010, more accurate land-use change models have shown that EPA's initial estimates were too high.44As one recent study explained, "prior to the last couple of years, there was insufficient data on global land-use change during the 3Q 2010 RFS Rule, supra note 2, 75 Fed. Reg. at 14,765 ("The indirect, international emissions are the component of our analysis with the highest level of uncertainty."). 40 2017 USDA LCA, supra note 8, at 121. 41 Boland & Unnasch, supra note 30, at 20. 42 The agency used FAPRI-CARD to model international land-use emissions, and FASOM to model domestic emissions. 2010 RFS Rule, supra note 2, 75 Fed. Reg. at 14,768. 43 Id. at 14,781. 44 See, e.g., Jennifer B. Dunn et al., Land-use change and greenhouse gas emissionsfrom corn and cellulosic ethanol, 6 Biotech, for Biofuels 51 (2013). Sierra Club v. EPA 18cv3472 NDCA 10 Tier 3/4 ED 002061 00161533-00014 biofuels boom era. However, now we have that data, and it can be used to better calibrate prior estimates of land-use change."45Accordingly, Purdue's agricultural economists recalibrated the GTAP model in 2013.46As a result of these changes, the GTAP model now projects "less expansion in global cropland due to ethanol expansion"; a "lower U.S. share in global cropland expansion"; and a "lower forest share in global cropland expansions."47More recently GTAP analysts have also refined the land carbon stock estimates used by the model.48Department of Energy scientists now say that, in light of GTAP model refinements, a more realistic estimate of com ethanol's ITUC emissions is 5.1 g C 0 2e/M J.49EPA's ILUC estimate should be corrected using the updated GTAP model to accord with the Department of Energy's estimate. Even more importantly, EPA failed to account for the intensification of agriculture in its ILUC estimate. Empirical data cited in USDA's new study has discredited EPA's predicted ILUC emissions in Brazil and other countries: corn ethanol has not significantly increased deforestation in the Amazon region or elsewhere.50Contrary to EPA's FAPRI-CARD model predictions, empirical evidence shows that during the period of corn ethanol expansion, Brazilian deforestation actually fell significantly, and farmers responded to changes in price primarily by using available land resources more efficiently--mostly by harvesting 45 See, e.g., Farzad Taheripour & Wallace E. Tyner, Biofuels and Land-use Change: Applying Recent Evidence to Model Estimates, 3 Appl. Sci. 14, 15 (2013). 46 Id. 47 Id. 48 See, e.g., Holly Gibbs et al., New Estimates of Soil and Biomass Carbon Stocks for Global Economic Models, Global Trade Analysis Project (GTAP) Tech. Paper No. 33, at 21 (2014), available at http://bit.ly/lTuJq98. 49 See Jennifer B. Dunn et al., DOE Argonne Nat'l Lab., Carbon Calculator for Land Use Change from Biofuels Production, A N L/ESD /12-5, at 25 (2016), available at http://L usa.gov/lM 84W IT. 50 See 2017 USDA LCA, supra note 8, at 60-66. Sierra Club v. EPA 18cv3472 NDCA 11 Tier 3/4 ED 002061 00161533-00015 land more often ("double cropping")--not expanding acreage.51 That is particularly true for Brazil.52 EPA's 2010 LCA, however, does not take into account the "non-yield" intensification of cropland through techniques like double cropping.53Thus, EPA overstated the carbon intensity of corn ethanol.54As the USDA's recent lifecycle analysis shows, when the updated 2013 GTAP model is further adjusted to account for this new empirical evidence, ILUC emissions for com ethanol fall to an almost insignificant 1.3 g C 0 2e/M J.55 Despite this new evidence, EPA's 2016 response to a Request for Correction of Information (RFC) submitted by Urban Air Initiative stated that no correction to its ILUC estimate for corn ethanol was required. The Agency claimed that because "[sjtudies published between 2011 and 2015 vary" widely and EPA's estimate "is still within the range.56Six of the twelve studies cited by EPA, however, are European biofuel studies of no apparent relevance to ILUC emissions from corn ethanol produced in the United States.57Another study cited by EPA is based on a 2009 working paper that uses the same erroneous FAPRI-CARD model as EPA's 51 Id. (citing Bruce A. Babcock & Zabid Iqbal, Using Recent Land-use Changes to Validate Landuse Models, 14-SR 109 (2014)). 52 See id. at 63 (showing that 76% of the increase in harvested land in Brazil is due to changes in double cropping). 53 See Babcock & Iqbal, supra note 51, at 20-22 (criticizing the FAPRI-CARD model). 54 See id. ("The pattern of recent land use changes suggests that existing estimates of greenhouse gas emissions caused by land conversions due to biofuel production are too high because they are based on models that do not allow for increases in non-yield intensification of land use. Intensification of land use does not involve clearing forests or plowing up native grasslands that lead to large losses of carbon stocks.). 55 20 17 USDA LCA, supra note 8, at 125, Table 3.43 (estimating ILUC emissions at 1,326 g C 0 2e/mmBTU). 56EPA, Response to RFC 16003, at 1 (Dec. 8, 2016), available at https://www.epa.gov/ sites/production/files/2016-12/documents/epa,,response__to,,rfc__16003.pdf. 57 See id. at 1, nn. 4, 6, 8, 9, 10, 11 (citing studies). Sierra Club v. EPA 18cv3472 NDCA 12 Tier 3/4 ED 002061 00161533-00016 2010 LCA analysis.58In the other studies EPA cited in its response to the RFC, the mean ILUC emissions are lower than EPA's 2010 estimate.59 USDA's study shows that a plausible range of ILUC emissions from corn ethanol based on recent scientific estimates extends from 18.9 g C 0 2e/M J to 1.3 g C 0 2e/M J, significantly below EPA's 2010 LCA estimate of 30.3 g C 0 2e/M J.60 2. Domestic Land-Use Change Emissions In its 2010 LCA, EPA estimated that com ethanol's domestic land use change emissions would reduce com ethanol's carbon intensity by 3.8 g C 0 2e/M J.61EPA developed its estimate using the Forest and Agricultural Sector Optimization Model designed by Texas A&M.62 This estimate may be too low, because EPA's model assumes corn ethanol is grown with conventional tilling practices.63 Since EPA's 2010 LCA, new evidence has demonstrated that reduced tillage practices--particularly no-till agriculture--significantly increase soil organic carbon in corn soils. A multiyear study of South Dakota surface soil samples (0-15 cm in depth), led by soil scientist David Clay, found clear evidence that no-tillage practices (and higher corn yields) increase soil carbon sequestration.64The study used laboratory surface soil samples submitted by agricultural producers. From the 58 Id. at 1 n.3 (citing Jerome Dumortier et al., Sensitivity of Carbon Emission Estimates from Indirect Land-Use Change, Working Paper, 09-WP 493 (July 2009), http:/ / www.card.iastate.edu/products/publications/pdf/09wp493 .pdf.) 59 See id. at 1, n. 5, 7, 12, 13, 14 (citing studies). 60 2017 USDA LCA, supra note 8, at 127, Figure 3-4. 61 2010 RFS RIA, supra note 1, at 362, Figure 2.4-19. 62Id. at 355. 63 2017 USDA LCA, supra note 8, at 155. 64 See David E. Clay et al., Corn Yields and No- Tillage Affects Carbon Sequestration and Carbon Footprints, 104 Agron. J. 763 (2012) [hereinafter Clay et al., Carbon Sequestration]; see also David Clay et al., Tillage and Corn Residue Harvesting Impact Surface and Subsurface Carbon Sequestration, 44 J. Environ. Qual. 803 (2015) [hereinafter Clay et al., Tillage and Corn Residue]. Sierra Club v. EPA 18cv3472 NDCA 13 Tier 3/4 ED 002061 00161533-00017 laboratory results, Clay concluded that the soils studied were now net "carbon sinks," thanks, in part, to the adoption of reduced tillage and no-tillage practices, as well as increased corn crop yields over the years.65Over a period of three years, Clay found that the average carbon sequestration rate was 341 kg of carbon per hectare per year.66Over a longer period of 25 years, Clay concluded that the average carbon sequestration rate was 386 kg of carbon per hectare per year.67This is equivalent to an annual carbon intensity credit of 18.2 C 0 2e/M J for that time period.68 Studies of deeper soil samples have shown even greater increases in soil carbon from reduced tillage. For example, a 2012 USDA study collected soil samples from as deep as 150 cm below the surface of experimental no-till fields in Nebraska, measuring changes in soil organic content over nine years.69The study found that improved agricultural management practices can double or even quadruple total soil organic carbon when deep soil is taken into account.70The study found average annual increases of more than 2 metric tons of soil organic carbon per hectare, with over 50% of the carbon sequestered deeper than 30 cm in the soil profile.71The sequestration rates found by the study "greatly exceed the soil carbon credits that have been used in modeling studies to date for maize and switchgrass grown for bioenergy."72Other recent USDA studies have reached similar results.73 65 Clay et al., Carbon Sequestration, supra note 64, at 769. 66Id. at 768. 61 Id. 68 See Appendix, infra p. 23. 69 Ronald F. Follett et al., Soil Carbon Sequestration by Switchgrass and N o Till M aize Grown for Bioenergy, 5 Bioenerg. Research 866, 867 (2012), available at http://bit.ly/ lQIHAPv. 70Id. at 867. 71 Id. at 873. 12 Id. 73 See Ardel D . Halvorson & Catherine E . Stewart, Stover Removal Affects No- Till Irrigated Corn Yields, Soil Carbon, and Nitrogen, 107 Agron. J. 1504 (2015). Sierra Club v. EPA 18cv3472 NDCA 14 Tier 3/4 ED 002061 00161533-00018 In light of these studies, EPA should update its lifecycle analysis to include a pathway for corn cultivated with reduced tillage practices.74 3. Domestic Farm Inputs and Fertilizer N20 According to EPA's 2010 LCA, domestic farm inputs accounted for 9.8 g C 0 2e/M J of corn ethanol's lifecycle emissions in 2022, or 13% of total lifecycle emissions.75A significant fraction of these emissions result from N20 emissions from the application of nitrogen fertilizer to com fields, as the applied nitrogen is released as N 20 through a biochemical process of microbial "nitrification" and "denitrification" that is stimulated when nitrogen fertilizer application exceeds plant needs.76 EPA's estimate for domestic farm inputs needs correction for at least two reasons. First, it uses outdated U.N. Intergovernmental Panel on Climate Change (IPCC) guidelines to calculate the effect of N 20 emissions on global warming. Second, it ignores available technologies that reduce N20 emissions by reducing fertilizer losses. a. The 2010 LCA Uses Outdated IPCC Guidelines. Because a molecule of N 20 contributes more to climate change than a molecule of C 0 2, a conversion factor, known as a global warming potential (GWP), is used to convert N20 emissions to a C 0 2-equivalent.77For its 2010 LCA, EPA used the GWP from the IPCC's Second Assessment Report, which was 310.78This value is outdated. The IPCC's Fourth and Fifth Assessments both recommend a lower GWP of 298 for N 20 , and in 2013, the UN updated its GHG reporting guidelines to 74 See Appendix, infra p. 55 (estimating carbon intensity credits from several studies). 75 2010 RFS RIA, supra note 1, at 334, Table 2.4-13. 76Id. at 330, Table 2.4-8. 11 Id. at 313. 78Id. at 313, Table 2.3 3. Sierra Club v. EPA 18cv3472 NDCA 15 Tier 3/4 ED 002061 00161533-00019 require EPA to use a GWP of 298 for N20 emissions.79Applying this updated GWP would reduce the contribution of N 20 emissions to corn ethanol's lifecycle emissions. b. The 2010 LCA Ignores Technologies that Reduce Farm N20 Emissions. Second, because the EPA's 2010 LCA does not include updated USDA data on farm practices, EPA fails to account for "an increase in crop and nutrient management strategies" that greatly decrease N 20 losses.80The most important technologies that EPA's 2010 LCA ignores are the increased use of nitrification inhibitors to delay the nitrification process, and the use of precision agriculture to optimize fertilizer application and minimize losses to the environment.81 Studies show that the use of nitrification inhibitors alone can reduce N 20 emissions from fertilizer by 19% to 60%.82But because EPA's 2010 LCA does not include the latest USDA data, it does not include "changes in emissions caused by these increasingly common practices."83 In its recent response to Urban Air Initiative's Request for Correction, EPA stated that no correction to its N 20 emissions estimate for corn ethanol was required because its projected fertilizer application rate for 2022 was not inconsistent with 2010 data.84But application rate is a separate issue from the GHG reductions 19 EPA, Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2015, 1-9, 1-10 (Apr. 2017). 8020 17 USDA LCA, supra note 8, at 15-16. 81 Id. at 15. Id. at 15-16 (collecting studies). 83Id. at 16. 84EPA, Response to RFC# 16003, at 2 (Dec. 8, 2016) (emphasis added). Sierra Club v. EPA 18cv3472 NDCA 16 Tier 3/4 ED 002061 00161533-00020 achieved by nitrification inhibitors, and EPA's data still does not account for the effect of nitrification inhibitors on N 20 losses.85 According to the USDA's recent study, an updated lifecycle analysis would yield a domestic farm inputs and fertilizer N 20 emissions value of 8.6 g C 0 2e/M J in 2014.86By 2022, the USDA study estimates these emissions will be even lower, at 7.8 g C 0 2e/M J, a significant reduction relative to EPA's estimate of 9.8 g C 0 2e/M J in 2022.87 4. International Farm Inputs and Fertilizer N20 In its 2010 RIA, EPA estimated that international farm inputs and fertilizer emissions resulting from its projected increase in corn ethanol would be 5.4 g C 0 2e/M J,88or 7% of total corn ethanol lifecycle emissions, mostly as a result of increased N20 losses resulting from an increase in crop acreage abroad.89 This estimate is too high, for at least two reasons. First, as already mentioned, EPA applied an outdated GWP for N 20 emissions that was too high. Second, the international land-use changes on which EPA's estimate was predicated are based on outdated models, and have not in fact occurred.90The USDA's recent lifecycle analysis estimates a more realistic 2.1 g C 0 2e/M J for international farm inputs and N 0 2emissions from fertilizer, significantly below EPA's 2010 EGA estimate. QC ' 2017 USDA LCA, supra note 8, at 15 (stating that while USDA data "already reflect the effects of precision agriculture through the reduced fertilizer use per bushel of corn harvest. . . use of nitrification inhibitors is not reflected in estimation of N20 emissions."). 86Id. at 95, Table 3-10. 87Id. at 157, Table 4-3. 88 20 17 USDA LCA, supra note 8, at 95, Table 3-47 (reporting EPA's value at 5,720 g C 0 2/mmBTU). QQ See 2010 RFS RIA, supra note 1, at 342, Table 2.4-16 (estimating com ethanol international N20 emissions at 3.38 kg C 0 2e/mmBTU). 90 2017 USDA LCA, supra note 8, at 95. Sierra Club v. EPA 18cv3472 NDCA 17 Tier 3/4 ED 002061 00161533-00021 B. Ethanol Fuel Production In its 2010 LCA, EPA estimated that ethanol fuel production at biorefmeries would account for 28.4 g C 0 2/M J, or 38% of EPA's estimated carbon intensity for corn ethanol.91This value needs to be corrected because it underestimates ethanol plant yields and it fails to fully account for corn ethanol co-products. 1. Ethanol Plant Yields EPA's estimate of ethanol fuel production emissions is in part a result of its underestimation of the ethanol yield--the amount of ethanol that biorefmeries produce from each bushel of corn. EPA's 2010 LCA predicted a yield of "2.71 gallons per bushel for dry mill plants and 2.5 gallons per bushel for wet mill plants."92 This implies a weighted average yield of 2.63 gallons per bushel for ethanol plants.93 Based on recent data from the Energy Information Administration and USDA, the current average yield for both wet and dry mill ethanol plants is 2.84 gallons per bushel, significantly above the yields built into EPA's models.94 Correcting the 2010 LCA's yield assumption would significantly reduce EPA's estimate of ethanol fuel production emissions. 2. Com Oil A proper lifecycle analysis of corn ethanol would fully account for biorefmery co-products that displace GHG emissions elsewhere. EPA's based its 2010 LCA emission estimated in part on the assumption that "70% of dry mill ethanol plants" 91 Id. at 145, Table 3-63 (reporting EPA's value). 92 2010 RFS RIA, supra note 1, at 425. Q3EPA estimated ethanol plants in 2022 would consist of 63% dry mill and 37% wet mill. Id. at 471, Figure 2.6-3. 94 See Renewable Fuels Ass'n, Industry Statistics: Monthly Implied Average Ethanol Yield (Gallons per Bushel) (last updated August 1, 2017), available at http://www.ethanolrfa.org/resources/industry/statistics/# 1461259890924-697180ef-b2a8 (reporting an average yield for 2016 of 2.84). Sierra Club v. EPA 18cv3472 NDCA 18 Tier 3/4 ED 002061 00161533-00022 would extract com oil for use as biodiesel in 2022." More recently, Department of Energy scientists estimated that as of 2014, over 80% of the dry mill ethanol plants now generate corn oil for biodiesel plants.96A bushel of com currently produces about 0.55 pounds of com oil.97And com oil displaces soy oil used as a feedstock for biodiesel, reducing GHG emissions.98GREET has been updated to include a one-toone displacement credit to account for the displacement of soy oil.99But EPA has not updated its 2010 LCA to reflect the increase in corn oil co-products. C. Gasoline Lifecycle Emissions Since EPA's 2010 LCA, petroleum-based fuels have become more carbon intensive. As a result, the baseline gasoline carbon intensity value that EPA relied upon in the 2010 RFS Rule is inaccurate. Even if EPA is obligated to use an arbitrary 2005 petroleum baseline to approve renewable fuel pathways,100EPA is not obligated to use that baseline to calculate the GHG benefits of the program. As the National Academy of Sciences noted in 2011, a proper "comparison scenario" for ethanol should include marginal GHG emissions "resulting from any change in the use of oil sands and other nonconventional sources of petroleum."101 Because gasoline's carbon 95 See 2010 RFS RIA, supra note 1, at 428. 96 See Zhichao Wang et al., Argonne Nat'l Lab., Updates to Corn Ethanol Pathway and Development of an Integrated Corn and Corn Stover Ethanol Pathway on the GREET Model, ARGONNE/ESD-14/11 (2014). 97 See Scott Irwin, The Profitability o fEthanol Production in 2015, 6 Farmdoc Daily, Department of Agricultural and Consumer Economics, University of Illinois at Urbana-Champaign, (Jan. 6, 2016), available a t http://bit.ly/ lphwLdh. 98 Wang, supra note 96, at 4. 99Id. at 5. 100 See 42 U.S.C. 7545(o)(l)(C), 7545(o)(2)(A)(i). 101 NRC, Renewable Fuel Standard, Potential Economic and Environmental Effects of U.S. Biofuel Policy 195 (2011). Sierra Club v. EPA 18cv3472 NDCA 19 Tier 3/4 ED 002061 00161533-00023 intensity has increased, the corresponding GHG benefits of the RFS have also increased. Gasoline GHG emissions are trending upwards because of increased "use of oil sands and other nonconventional sources of petroleum."102Unlike renewable fuel producers, which are required to achieve lifecycle reduction benefits to qualify for the RFS, EPA does not hold gasoline producers accountable for their increased lifecycle GHG emissions.103 Methane flares from shale oil extraction have increased GHG emissions from oil production.104Tar sand recovery often requires carbon-intensive steam injection, additional carbon-intensive processing to separate bitumen from tar sands, and chemicals to reduce the viscosity of the product for transportation, increasing extraction emissions.105Emissions associated with refining a barrel of tar sand oil are also higher.106And even conventional oil is becoming more carbon-intensive. Oil 102 Jeremy Martin, Union of Concerned Scientists, Fueling a Clean Transportation Future, at 1 (2016) ("As oil companies increasingly go after unconventional, hard-to-reach sources such as tar sands and use more intense extraction techniques such as hydraulic fracturing (fracking), dirtier sources of oil have become an increasingly large part of the mix, and wasteful practices are needlessly increasing emissions."). Oil is the largest fossil fuel contributor to global warming in the United States, contributing more than coal and natural gas. Id. at 8. For other studies on the high marginal emissions of unconventional oil sources, see Deborah Gordon et al., Know Your Oil: Creating a Global Oil-Climate Index, Carnegie Endowment for Global Peace (2015); Susan Boland & Stefan Unnasch, Life Cycle Associates, Carbon Intensity of Marginal Petroleum and Corn Ethanol Fuels, LCA.6075.83.2014 (2014). 103 See Martin, supra note 102, at 5 ("[Electricity and biofuels are getting cleaner because producers are subject to careful scrutiny of the global warming emissions associated with the fuels' production, and public policy is holding producers accountable to reduce these emissions. However, the same level of scrutiny is not being applied to the different sources and methods of producing gasoline. In addition, oil companies are not obligated to reduce emissions from their supply chains. For the United States to avoid the worst consequences of climate change, all fuel producers have to minimize their global warming pollution."). While regulation might help mitigate GHG emissions from tight oil, "[t]he most obvious way for the United States to reduce the problems caused by oil use is to steadily reduce oil consumption through improved efficiency and by shifting to cleaner fuels." Id. at 7, 12. 104Id. at 16-17. 105Id. at 19-20. 106Id. at 20. Sierra Club v. EPA 18cv3472 NDCA 20 Tier 3/4 ED 002061 00161533-00024 producers are injecting additional steam, chemicals, and gases (including methane) to enhance oil recovery, increasing the energy and carbon intensity of conventional oil extraction.107 EPA's 2010 LCA understates the carbon intensity of gasoline. W hen EPA's skewed carbon intensity baseline for gasoline is corrected, com ethanol is an even more attractive substitute. Because of tight oil, the Department of Energy estimated that carbon intensity of gasoline in 2014 was 94 g C 0 2e/M J, higher than EPA's 2005 baseline value.108 C o n c l u sio n In 2010, EPA predicted that blending com ethanol into gasoline would reduce GHG emissions. The Agency was right about that, but ethanol is even better at cutting carbon emissions than EPA gave it credit for. In the 2010 RFS Rule, EPA estimated com ethanol would have a carbon intensity of 74.9 g C 0 2e/M J in 2022.109 The USDA's recent estimate is 36% lower--47.9 g C 0 2e/M J.110And when adjusted for the soil carbon sequestration of the corn plant, the carbon intensity of corn ethanol may fall by 18.2 g or more, depending on soil conditions, tillage practices, and corn crop yield, resulting in a carbon intensity of 29.7 g C 0 2e/M J or less.111At that rate, ethanol would generate at least 68% less lifecycle GHG pollution than 2005 baseline gasoline on an energy-equivalent basis. The GHG benefits of ethanol will only grow as ethanol production becomes increasingly efficient, and gasoline production continues to get dirtier. 107 ITdj . at. ,15c . 108 See Elgowainy et al., supra note 32, at 7623 (estimating that the "total life-cycle GHG emissions for gasoline" are 94 g C 0 2e/MJ). 1092010 RFS Rule, supra note 2, 75 Fed. Reg. 14,788. 1102017 USDA LCA, supra note 8, at 166. 111 See Clay et al., Carbon Sequestration, supra note 64, at 769; Appendix, infra p. 23. Sierra Club v. EPA 18cv3472 NDCA 21 Tier 3/4 ED 002061 00161533-00025 This analysis does not account for the fuel efficiency gains that would be possible if ethanol were blended above the 10% level of most U.S. gasoline. By enabling the auto industry to produce engines with higher compression ratios and more fuel-efficient vehicles, high-octane mid-level ethanol fuel blends could achieve significant downstream, as well as upstream, GHG reductions.112 The data and studies that were available to EPA in 2010 were inaccurate, and they are now obsolete. The 2010 EGA is not a sound basis for estimating the costs and benefits of the Proposed Rule or for evaluating the carbon intensity of new ethanol producers. EPA must either adopt USDA's updated estimate and allow for situation-specific soil carbon adjustments, or correct the inaccuracies in its outdated lifecycle analysis to reflect the best available science. 112 See Control o fA ir Pollution from M otor Vehicles: Tier 3 M otor Vehicle Emission and Fuel Standards, 79 Fed. Reg. 23,414, 23,528-29 (Apr. 28, 2014) ("E30 or higher ethanol blends. . . could help manufacturers who wish to raise compression ratios to improve vehicle efficiency as a step toward complying with the 2017 and later light-duty greenhouse gas and CAFE standards. This in turn could help provide a market incentive to increase ethanol use beyond E 10."). 22 Sierra Club v. EPA 18cv3472 NDCA Tier 3/4 ED 002061 00161533-00026 Study & Year Soil Depth ! illtlgC Study Length (years) SOL' gain (Nig. Ila. Vr.)' Avg. Corn Yield in Study (Bushels/Ha./Yr.)vi i ihanol Yield (l iallons Bushel)'" Ethanol Energy Yield (MJ, Gallon)'"5 (rams Soil Cui hon ML Cto C02conversion (CO= C*3.664)x Credit in Grams CO: eq./MJ"' A p p e n d ix Clav et al (2012 Long- Term)' 0-15 cm \ arious 25 0.308 334 92 \ 74,144 4.9o 3.664 18.2 Clay et al (2015)" 0-30 cm \o -1ill & Chisel 5 0.53 -1-19 1240 99,826 5.3 : 3.664 19.5 Foiled et al (2012)"' 0-150 cm So-1 ill 9 2.6 240 663 53,378 48.7: 3.664 178.5 Halvorsou & Stewart (20l5f 0-60 cm Si- ! ill 7 0.856 347 959 77,214 11.09 3.664 40.6 Sierra Club v. EPA 18cv3472 NDCA 23 Tier 3/4 ED 002061 00161533-00027 I Clay, Carbon Sequestration, supra note 64. The 2012 Clay paper includes two studies. The first, a seven-year study, estimated that surface soil carbon sequestration reduces the carbon intensity of com ethanol by as much 19.6g C 0 2e/M J in the North-Central and Southeast regions of North Dakota. Id. at 769. The data in this study is based on the second study, a twenty-five year study. II Clay et al., Tillage and Corn Residue, supra note 64. IIIFollett et al., supra note 69. IVHalvorson & Stewart, supra note 73. v Soil Organic Carbon (SOC) gain is expressed in annual Megagrams (Mg.) (1 Mg. = 1,000 Kg.) of carbon sequestered per year, per hectare (ha.). The .368 Mg. SOC for Clay's 2012 study is based on the reported average over the 25 years of the study. Clay et al., Carbon Sequestration, supra note 64, at 768 ("[D]uring the past 25 yr, surface SOC amounts have increased at an average rate of 368 kg C (ha x yr).'1"). The 2.65 Mg. SOC gain for Clay's 2015 study is based on the average SOC gain, with no stover removal. Clay et al., Tillage and Corn Residue, supra note 64, at 808 ("[I]n the combined 0- to 15- and 15- to 30-cm soil zones. . . 2.65 Mg SOC ha'1were sequestered . . . in the 0% residue removal treatment!]."). The 2.6 Mg. SOC gain for Follett's study is based on the observed gain applying 120 kg/ha of nitrogen fertilizer, with no stover removal. Follett et al., supra note 69, at 873 ("At the 120 kg ha"1N fertility rate with no stover harvest, the annual increase in soil C was 2.6 Mg ha 1year.'1!.]"). The .856 Mg. SOC gain figure for Halvorson & Stewart's study is based on the annual average, with no stover removal. Halvorson & Stewart, supra note 73, at 1510 ("The estimated annual rate of SOC gain from the FR [full stover retained] treatments over the 7yr of this study would have been . . . 856 kg C h 1from the . . . 0 to 60-cm soil depths."). V1One bushel equals 25.40 kg of corn grain. See Iowa State, Ag Decision Maker Metric Conversions, C6-80 (May 2013), available at http://bit.ly/lVxnEks. The average yield for Clay's 2012 study is based on USDA historical data for the counties tested. Nat'l Agric. Research Serv., Quick Stats, available a t http://www.nass.usda.gov/Quick_Stats/; see also Clay et al., Carbon Sequestration, supra note 64, at 768 & fig. 6. The average yield for Clay's 2015 study is based on the reported yield of 11,408 kg. per ha., with no stover removal. Clay et al., Tillage and Corn Residue, supra note 64, at 806, Table 1. The average yield for Follett's study is based on the reported figure for corn grain using 120 kg of nitrogen fertilizer per ha., with no stover removal. Follett 2012, supra note 69, at 873. The average yield for Halvorson & Stewart's study is 8,824 kg. per ha., with no stover removal. Halvorson & Stewart, supra note 73, at 1507. TMThe ethanol yield is conservatively based on the USDA's average yield of 2.76 gallons per bushel in 2010, multiplied by the number of bushels produced every year. USDA, 2015 Energy Balance for the Corn Ethanol Industry, Table 1 (Feb. 2016). vmThe ethanol energy yield is based on multiplying the ethanol yield by the heating value of undenatured ethanol used by CARB: 80.53 MJ per gallon of ethanol. CARB, Calculation of Denatured Ethanol Cl and CA RFG, http://bit.ly/ loCEj9k. " Grams of soil carbon are derived by converting Mg. SOC gain into grams and dividing it by the ethanol energy yield. xThe carbon to C 0 2conversion factor is based on a molecular weight conversion from carbon to C 0 2: 1gram of carbon = 3.664g C 0 2. See Carbon Dioxide Information Analysis Center, Conversion Tables, Oak Ridge Nat'l Lab., Table 3, http://cdiac.oml.gov/pns/convert.html. X1The carbon intensity credit is arrived at by multiplying the carbon conversion factor by grams of soil carbon per MJ. Sierra Club v. EPA 18cv3472 NDCA 24 Tier 3/4 ED 002061 00161533-00028