Document 6b11ev6k6ynQEJOkMBEo64jwd

BEFORE THE UNITED STATES ENVIRONMENTAL PROTECTION AGENCY In re: ) ) Cross-State Air Pollution Rule Update ) for the 2008 Ozone NAAQS ) ) EPA Docket No. EPA-HQ-OAR-2015-0500 Western Farmers Electric Cooperative's Petition for Reconsideration of the Cross-State Air Pollution Rule Update for the 2008 Ozone NAAQS Western Farmers Electric Cooperative ("Western Farmers") respectfully petitions the U.S. Environmental Protection Agency ("EPA" or "Agency") for reconsideration of the final rule entitled "Cross-State Air Pollution Rule Update for the 2008 Ozone NAAQSf Docket No. EPAHQ-OAR-2015-0500, 81 Fed. Reg. 74,504 (Oct. 26, 2016) ("CSAPR Update Rule" or "Final Rule").1 Specifically, Western Farmers requests that EPA recalculate (1) unit-level allocations of seasonal nitrous oxide ("NOx") allowances for 2017 and beyond for Western Farmers' Anadarko Plant, using data-substitution called for by the Final Rule; and (2) Oklahoma's NOx seasonal emission budget for 2017 and beyond, using the proposed limited adjustment to EPA's budget calculation methodology to account for arbitrary modeling results. Western Farmers is a not-for-profit company, and its primary mission is to provide affordable and reliable power to its consumers. Western Farmers is a generation and transmission cooperative that is owned by, and serves, 21 member distribution cooperatives in Oklahoma and New Mexico. Western Farmers also is proud to serve Altus Air Force Base. Western Farmers owns and maintains approximately 3,700 miles of transmission that is operated and administered by the Southwest Power Pool, and its diverse generating mix includes a single 450 megawatt ("MW") coal-fired generation unit at the Hugo plant and 14 natural gas-fired steam cycle, simple cycle, and combined cycle generating units at the Anadarko and Mooreland plants in Oklahoma.2 Additionally, Western Farmers has been a leader in Oklahoma and the west in developing low- and zero-carbon emitting generation, with almost 600 MW of wind 1 Along with an electronic submission of this petition, Western Farmers is supplying EPA with a CD-Rom containing the six Appendices (Appendices A-l to C-2). The Appendices are Excel datasheets and workbooks demonstrating the calculations and modeling described in this petition. The CD-Rom contains both a "locked" (i.e., password-protected, read-only) version and an "unlocked" courtesy working copy of each file. Appendices B-l and C-l-2 are modifications of EPA's own tools (Appendix E and AQAT, respectively) and use EPA's own cell formulas. Appendices A-1-2 and B-2 are Western Farmers' original files with cell formulas removed, but with all necessary data and explanation of the methodology underlying the calculations. 2 As well as the 42 MW Wartsila plant in Lovington, New Mexico. Active 30698243.11 1 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00001 energy resources currently, and another 30 MW wind facility planned for 2017. Also, Western Farmers added solar to its portfolio mix in 2016 (to come online in early 2017), including 25 MW of solar energy under contract and just over 20 MW of owned solar energy resources.3 BACKGROUND CSAPR addresses the "good neighbor" provision of the Clean Air Act, which prohibits sources within each state "from emitting any air pollutant in an amount which will. . . contribute significantly" to any other state's nonattainment, or interference with maintenance of, any NAAQS. 42 U.S.C. 7410(a)(2)(D)(i)(I). EPA defines "significant contribution" by reference to (1) a state's "linkage" to downwind receptors (i.e., emissions of approximately 1 percent of compliant ambient levels);4 and (2) the ability of the state to achieve emission reductions at the relevant cost threshold.5 CSAPR achieves emissions reductions through annual and ozone season emissions trading programs. EPA finalized CSAPR in 2011 to address three NAAQS, including the 1997 ozone NAAQS. 76 Fed. Reg. 48,208 (Aug. 8, 2011) ("Original CSAPR Rule" or "CSAPR"). In December 2011, EPA finalized a supplemental CSAPR rule that added several states, including Oklahoma, to the ozone season NOx program. 76 Fed. Reg. 80,760 (Dec. 27, 2011) ("Supplemental CSAPR Rule"). Various groups challenged CSAPR, resulting in a stay of the rule. See EME Homer City Generation, L.P. v. EPA, No. 11-1302 (D.C. Cir. Dec. 30, 2011). CSAPR and Supplemental CSAPR Rule requirements and deadlines were effectively tolled pending resolution of the litigation. See 79 Fed. Reg. 71,663 (Dec. 3, 2014). In April 2014, the U.S. Supreme Court generally upheld the rule, see EPA v. EME Homer City Generation LP, 134 S. Ct. 1584 (2014), and the stay was lifted in October 2014. See EME Homer City, No. 11-1302 (Oct. 23, 2014). Phase 1 of the CSAPR ozone season program began in May 2015. EPA published the Final CSAPR Update Rule in October 2016 to address good neighbor provisions with respect to the 2008 ozone NAAQS. The Agency had published a proposed rule 3 WFEC purchases or produces energy from various wind sources, and, in the near future, solar resources. However, WFEC has not historically, nor may not in the future, retain or retire all of the renewable energy certificates associated with the energy production from these facilities. 4 EPA used an ozone air quality assessment tool ("AQAT") to determine which states contribute to ozone concentrations at maintenance and non-attainment downwind receptors, and to estimate the impact of upwind states' NOx emissions reductions on downwind ozone pollution concentrations. The AQAT was developed specifically for the CSAPR Update Rule. EPA, Regulatoiy Impact Analysis of the Cross-State Air Pollution Rule (CSAPR) Update for the 2008 National Ambient Air Quality Standards for Ground-Level Ozone (Sept. 2016) at 3-10, available at https://www3.epa.gov/ttn/ecas/docs/ria/transport ria final-csapr-update 2016-09.pdf ("Regulatory Impact Analysis"). 5 EPA used the Integrated Planning Model v.5.15 ("IPM") to predict how many emissions reductions are available at the relevant cost thresholds. IPM is a dynamic, linear programming model used to project power sector behavior under current and future conditions. IPM's primary objective is to provide estimates of least-cost capacity expansions, electricity dispatch, and emission control strategies while meeting energy demand and environmental, transmission, dispatch, and reliability constraints. See https://www.icf.com/solutions-and-apps/ipm. Active 30698243.11 2 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00002 about 10 months earlier. See 80 Fed. Reg. 75,706 (Dec. 3, 2015) ("Proposed CSAPR Update Rule" or "Proposed Rule"). In the Final Rule, EPA found that 22 eastern states, including Oklahoma, had failed to submit a state implementation plan to meet their good neighbor obligations. For these 22 states, EPA issued federal implementation plans that generally update existing CSAPR NOx ozone season emission budgets, and that implement these budgets through modifications to the existing CSAPR NOx ozone season allowance trading program. The Final Rule addresses only emission reductions from electric generating units ("EGUs"). Implementation will start in the 2017 ozone season (May - September 2017). EPA made several significant changes between the Proposed and Final CSAPR Update Rules. Most notably for Western Farmers and other Oklahoma generators: (1) In calculating allocations for Oklahoma units in the Final Rule, EPA did not use data-substitution. Both the Proposed and Final Rules call for EPA to use up to five years of historic heat input data and up to eight years of historic NOx emissions data. To accomplish this, both rules also call for EPA to use a data-substitution method (so that if data is not available from EPA's preferred data source for a given year, EPA can pull equivalent data from an alternative source). EPA followed this method in the Proposed Rule. But in the Final Rule, EPA switched to using a single year of historic baseline data without any data-substitution for the Anadarko Plant and other Oklahoma units. EPA's error in the Final Rule led to a drastic (over 80 percent) reduction in the Anadarko Plant units' allocations from the Proposed Rule to the Final Rule.68 7 (2) In calculating Oklahoma's budget in the Final Rule, EPA used a revised formula with new inputs. Specifically, EPA introduced new variables, including (1) a NOx emission rate "delta" (equal to the difference between an IPM 2017 Final Base Case and IPM 2017 Final $ 1,400/ton Cost Case emission rate); and (2) an "adjusted" historic emission rate, based on a newly-developed dataset.9 For seven states, including Oklahoma, the IPM 2017 Base Case 6 The two data sources EPA says it will look to are: (1) EPA's Clean Air Markets Division ("CAMD"), for years for which it is available; and (2) U.S. Energy Information Administration ("EIA"), for years for which CAMD data is unavailable. 7 With some caveats: EPA did not include 2014 EIA data because it was unavailable at the time the Proposed Rule was published. Thus, for units without 2014 CAMD data, EPA did not "backfill" or substitute any EIA data. Additionally, EPA did not include data for years for which a unit was not yet operating (e.g., if a unit that began operating in 2012, EPA used available 2012-2014 heat input and NOx emissions data). See EPA, Proposed Rule, Unit Level Allocations and Underlying Data for the CSAPR for the 2008 Ozone NAAQS ("Proposed Allocation Spreadsheet TSD"). Excel file available at https://www.epa.gov/airmarkets/proposed-cross-state-air-pollutionupdate-rule. Further, as discussed below, it appears that EPA inappropriately relied on annual, not monthly, EIA data and made arbitrary downward adjustments to reported EIA data for cogeneration and combined cycle units. EPA should use monthly, reported (i.e., unadjusted) EIA data in the revised Final Rule allocations. 8 See Part I below. 9 Both the Proposed and Final Rules call for EPA to set state budgets as the minimum of either (1) historic emissions or (2) IPM-predicted 2017 emissions. EPA's changes pertain to the formula for calculating IPM-predicted 2017 Active 30698243.11 3 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00003 emission rate is significantly higher than the state's historic actual or adjusted 2015 emission rates--an unrealistic and arbitrary outcome (the "Perverse IPM Result''). The Perverse IPM Result creates an arbitrarily high "delta" value, causing EPA to overestimate the amount of available emission reductions in the state. Carried through the rest of the formula, this error results in an unrealistically low state budget. Oklahoma's budget is approximately 28 percent lower in the Final Rule than it was in the Proposed Rule--the most drastic reduction for any st.at.e. 10 REQUEST FOR RECONSIDERATION The Clean Air Act requires that EPA grant reconsideration of a regulation if a petitioner "can demonstrate to the Administrator that it was impracticable to raise [an] objection [during the period for public comment] or if the grounds for such objection arose after the period for public comment. . . and if such objection is of central relevance to the outcome of the rule." 42 U.S.C. 7607(d)(7)(B). In such a situation, reconsideration is mandatory: EPA "shall convene a proceeding for reconsideration of the rule and provide the same procedural rights as would have been afforded had the information been available at the time the rule was proposed." Id. (emphasis added). The notice-and-comment requirements of the Clean Air Act and the Administrative Procedure Act further require that EPA's "proposed rule and its final rule . . . differ only insofar as the latter is a 'logical outgrowth' of the former." Envtl. Integrity Project v. EPA, 425 F.3d 992, 996 (D.C. Cir. 2005). A "final rule is a 'logical outgrowth' of a proposed rule only if interested parties should have anticipated that the change was possible, and thus reasonably should have filed their comments on the subject during the notice-and-comment period." Id. at 998. EPA must grant reconsideration of the CSAPR Update Rule because: (1) In calculating allocations for Oklahoma units, EPA made a clear technical error by failing to use EIA data-substitution for the Anadarko Plant and other units. As described further below, the grounds for this objection arose after the public comment period. EPA did not make a similar error in the Proposed Rule, nor did EPA give notice of, or seek comment on, the option of not using data-substitution--so that some units would receive allocations based on a single baseline year--in the Final Rule. This objection is of central relevance to the outcome of the Final Rule. EPA's error threatens Western Farmers' ability to comply with the Final Rule, especially given the extremely short compliance timeframe for the 2017 ozone season. It has also created an unfair and unrepresentative distribution of allowances within Oklahoma, making it harder for the state to comply efficiently with the rule. Compliance feasibility and efficiency are central to the outcome of the CSAPR Update Rule.10 emissions. EPA used a historic emissions baseline period of 2014 in the Proposed Rule and 2015 in the Final Rule. See Part II below. 10 The primary driver of this drop in Oklahoma's budget appears to be the new IPM-predicted 2017 emissions formula, not the switch from a 2014 to 2015 historic baseline year. See Part II below. Active 30698243.11 4 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00004 (2) In calculating Oklahoma's budget, EPA used a new formula with inaccurate and perverse data inputs that overestimated the available emissions reductions in the state. As described further below, the grounds for this objection arose after the public comment period. EPA's broad request in the Proposed Rule for comment on "all aspects" of quantifying state budgets did not give Western Farmers reasonable notice of the highly complex changes EPA would adopt in the Final Rule--changes that would negatively impact Oklahoma more than any other state. This objection is of central relevance to the outcome of the Final Rule. Oklahoma's current Final Rule budget is unrealistically and arbitrarily low. This threatens the ability of all covered Oklahoma units to comply with the rule. I. UNIT-LEVEL ALLOWANCE ALLOCATIONS A. EPA Should Follow its Own Methodology and Use ElA Data-Substitution EPA's unit-level allowance allocation calculations in the Final Rule are arbitrary and capricious as applied to Western Farmers' Anadarko Plant units. EPA's failure to perform EIA data-substitution--and use of a single historic baseline year for some units--goes against the Final Rule's own prescribed methodology. It also puts units on unequal footing: some units received allocations based on one year of operation, while other units received allocations based on multiple years of operation. This effectively penalizes units, like the Anadarko Plant units, that had no obligation to report data to CAMD in prior years. EPA must recalculate Western Farmers' allocations for the Anadarko Plant units to avoid this unfair and arbitrary outcome. Specifically, EPA must use (1) five years of reported heat input data (2011-2015) and eight years of reported NOx emissions data (2008-2015); and (2) EIA data-substitution, where CAMD data is unavailable. Therefore, the allocations should be based on 2011-2014 EIA and 2015 CAMD heat input data and 2008-2014 EIA and 2015 CAMD NOx emissions data. EPA should use monthly (not annual), reported (not adjusted) data. See Appendix A-l for detailed proposed calculations.11 12 1. What EPA Said it Would Do Western Farmers' request to EPA is straightforward: simply do what you said you would do (but did not actually do). The Final Rule calls for using a five-year (2011-2015) historic baseline period for heat input data, and an eight-year (2008-2015) historic baseline period for NOx emissions data. See Allocation TSD at 6-7. 12 11 Where reported unit-level heat input EIA data is unavailable, EPA should make the best estimate based on reported plant-level EIA heat input data and reported unit-level heat input data (e.g., from CAMD). Appendix A-2 includes a detailed proposed estimation technique. 12 EPA's allocation methodology has two main phases: (1) Heat Input Data Phase: EPA "uses the average of the three highest years of heat input data out of a consecutive five-year period [2011-2015] to establish the heat input baseline for each unit." 81 Fed. Reg. at 74,564. EPA then calculates initial heat input-based allowance allocations based on each unit's percentage share of the state's total ozone season heat input; and (2) NOx Emissions Data Phase: EPA then "constrains the unit-level allocations so as not to exceed the maximum historical baseline emissions, calculated as the highest year of emissions out of a consecutive eight-year period [2008-2015]." Id. Active 30698243.11 5 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00005 EPA's rationale for using multiple historic baseline years is sound: it helps ensure that outlier data from a single unrepresentative year (e.g., where heat input or NOx emissions levels were significantly lower than usual) does not skew the calculations.13 As EPA recognizes, the power sector is susceptible to a range of variables affecting fuel use and emissions, including equipment failures, changing market forces, and weather patterns. See 81 Fed. Reg. at 74,566. One year of historic data cannot capture the inherent variability in a unit's operations from year to year, or ozone season to ozone season. EPA says it will to look to two possible sources to obtain this historic data: (1) First, EPA's CAMD database, which contains data from units already reporting under CSAPR or other trading programs, such as the Clean Air Interstate Rule or the Acid Rain Program; and (2) Second, EIA forms, "where EPA data are unavailable." Id. at 6-7. Under this approach, the only baseline years for which a unit should have a "missing" value are baseline years "for which a unit has no data on heat input (e.g., for a baseline year before the year when a unit started operating)." Id. at 7 (emphasis added). In other words, if CAMD data is available for a unit for a particular year within the relevant historic baseline period, EPA should use that CAMD data. If CAMD data is unavailable but EIA is available, EPA should use the EIA data.14 If and only if neither CAMD nor EIA data is available (e.g., because the unit was not yet operating), EPA should use no data--i.e., "zero" values. Indeed, for many units EPA must look to a historic data source other than CAMD to populate the multi-year baseline periods called for in the Final Rule. Prior to Phase I of the CSAPR program in 2015, many units, including the Anadarko Plant units, had no obligation to report data to CAMD under any EPA program. The Anadarko Plant units have only a single year of reported CAMD data (2015), but multiple years of EIA data (back to before 2008, the earliest relevant baseline year under the Final Rule). Under such circumstances, EPA's methodology calls for the Agency to use a combination of 2015 CAMD data and 2008-2014 EIA data to establish unit allocations. Overall, this methodology "bases a unit's allocation on the unit's historical heat input but limits any unit's allocation to its historical maximum emissions." EPA, Allowance Allocation Final Rule TSD at 6 (Aug. 2016) ("Allocation TSD"). Available at https://www3.epa.gOv/ainnarkets/CSAPR.U/CSAPR.%20Allowance%20Allocations%20Final%20Rxile%20TSD.PD F. 13 See Allocation TSD at 7. EPA chose a multi-year heat input baseline because "[sjelecting the three highest, non zero ozone season heat input values within the five-year baseline reduces the likelihood that any particular single year's operations (which might be negatively affected by outages or other unusual events) determine a unit's allocation." Id. EPA chose a multi-year NOx emissions baseline "in order to capture the unit-level emissions before and after the promulgation of the original CSAPR." Id. 14 For certain years and facilities, EIA heat input data is reported only on a plant-level, not unit-level, basis. As noted above, EPA should make its best estimate of unit-level heat input data. See Appendices A-l and A-2. Active 30698243.11 6 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00006 EPA also states it will look to publicly available, reported EIA data.15 EPA does not state, in either the Final Rule or the Allocation TSD, that the Agency will adjust EIA data from what is reported directly on the EIA forms.16 EPA therefore should use monthly (not annual), reported (not downward-adjusted for combined cycle and cogeneration units) EIA data whenever EIA data is called for in allocation calculations. Monthly data is better than annual data with a crude 5/12 multiplier because monthly data more accurately reflects actual ozone season operations.17 Reported heat input data is better than downward-adjusted heat input data for cogeneration and combined cycle units because reported data more accurately reflects these units' actual fuel use, as well as their highly efficient processes.18 Using monthly, reported EIA data also is consistent with EPA's treatment of CAMD data, which EPA did not adjust when incorporating it into the Final Rule's current allocation calculations. Treating EIA and CAMD data differently would be arbitrary and capricious. 2. What EPA Actually Did In the Final Rule, EPA inexplicably relied on only CAMD data, without backfilling EIA data, in its Oklahoma unit allocations. See EPA, Final Rule, Unit-Level Allocations and Underlying Data for the CSAPR Update for the 2008 Ozone NAAQS ("Final Allocation Spreadsheet TSD").19 Several units, including the Anadarko Plant units, had only a single year (2015) of reported CAMD data, because these units were not required to report data to CAMD 15 Specifically, EPA "used historical heat input and emissions data [EIA] forms, 860, 906, 920, and 923. These data are publicly available at http://www.eia.doe.gov/cneaf/electricity/page/data.html." Allocation TSD at 6. 16 In the Proposed Rule, EPA appears to have altered Western Fanners' reported EIA data in two ways: (1) instead of looking at monthly data for ozone season months, EPA took annual data and multiplied it by 5/12 to estimate ozone season operations (because the ozone season comprises 5 months out of the year); and (2) instead of using total reported heat input values, EPA applied a downward "proportional nameplate capacity adjustment" to cogeneration and combined cycle units, which have both a gas turbine and a steam turbine that produce electricity (basically, EPA multiplied the gas turbine's share of the plant's total nameplate capacity by the plant's total heat input). Both of these data manipulations are inappropriate. EPA historically has used reported, monthly EIA data-- as it should again here. 17 The 5/12 ozone season multiplier incorrectly assumes that units operate more or less the same during all 12 months of the year. In fact, the Anadarko Plant units operate significantly more during the ozone season because they are highly-efficient natural gas-fired units that are called on during peak periods of electricity demand, which often coincide with ozone season summer months. Further, monthly EIA data for May-September for 2008-2014 is available for these units; there is no reason not to use it. 18 The "proportional nameplate capacity adjustment" for cogeneration and combined cycle units misunderstands how these units operate and punishes these units for being efficient. All fuel consumed in combined cycle units is used in electricity generation: fuel is combusted in the gas turbine to either produce power or produce steam, which then produces power (in the steam turbine). Therefore, the adjustment punishes units for employing a more efficient process that reuses gas turbine exhaust to produce steam and more electricity, rather than just venting it off. 19 Excel file available at https://www.epa.gov/airmarkets/final-cross-state-air-pollution-rule-update. Active 30698243.11 7 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00007 under any EPA program prior to 2015.20 However, the Anadarko Plant and other Oklahoma units do have multiple prior years of available EIA data. EPA could easily backfill this EIA data into its allocation calculation spreadsheet. EPA does not explain why the Agency did not perform this data-substitution in the Final Rule's allocation calculations. B. Corrected Calculations Show the Anadarko Plant Units Should Receive Over 120 Additional Allowances EPA's approach arbitrarily reduced the allowances allocated to the Anadarko Plant units. As shown in the figures and tables below, the Final Rule's allocations for the Anadarko Plant units are more than 75 percent below what they should be under EPA's allocation method that incorporates EIA data-substitution,22102a2nd over 80 percent below what they were under the Proposed Rule. To correct the allocations for the Anadarko Plant units, Western Farmers used monthly, unadjusted EIA heat input data from EIA forms 923, 860, 906, and 920.23 See Appendix A-2.24 Western Farmers also pulled publicly available, reported EIA heat input data and performed data-substitution for all Oklahoma units. However, Figures 1-4 below show the results for only the Anadarko Plant units. Figure 1 shows the historic heat input data EPA relied on in performing the initial heat input-based allocation calculations (Steps 1-6) for the Anadarko Plant units in the current Final Rule, while Figure 2 shows the historic heat input data EPA should have relied on: 20 Western Farmers' analysis indicates that other Oklahoma units with only a single year of CAMD data include AES Shady Point, Mustang, Oklahoma Cogeneration LLC, and Weleetka units. 21 Under the Final Rule, the Anadarko Plant units have allocations of 9 tons, 14 tons, and 17 tons, respectively (40 tons combined). Using the corrected method, the Anadarko Plant units would receive allocations of 52 tons, 53 tons, and 56 tons, respectively (161 tons combined). See Figures 1-4 and Table 1 below. 22 Under the Proposed Rule, the Anadarko Plant units would have had allocations of 77 tons each (231 tons combined). 23 For facilities reporting EIA heat input data at a plant-level, not unit-level, an estimation technique was used to arrive at unit-level heat input data. First, the total reported plant-level heat input (EIA) was reduced by the total reported unit-level heat input for all units for which reported data was available (from EIA or CAMD). Then, the remaining heat input was distributed evenly among the unreported units. See Appendix A-2. 24 For the NOx emission data shown in Figure 4, Western Farmers relied on the NOx emissions values used in the Proposed Rule, due to the company's limited resources and the time it would take to calculate updated NOx emissions values based on monthly, unadjusted EIA data (instead of the annual, adjusted EIA data used in the Proposed Rule). Because only the maximum historic NOx emissions affect the unit allocations, re-calculating the NOx emissions values is not expected to impact the final allocations. Active 30698243.11 8 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00008 Figure 1. Historic Heat Input Data (Current Final Rule Allocations) Plant Name Calculation Anadarko Plant Anadarko Plant Anadarko Plant ORIS Boiler ID ID 3006 4 3006 5 3006 h 2011 Ozone Season CAMD Heat Input Unit ID (mmBtu) 2012 Ozone Season Heat Input (mmBtu) 2013 Ozone Season Heat Input [mmBtu) 2014 Ozone Season Heat Input (mmBtu) - . ' Unit Level Averageof 3 . Highest Non-Zero 2015 Ozone Ozone Season Season Heat Inputs from Heat Input 2011 to 2015 (mmBtu) (mmBtu) StateLevel Unit's Summation of Percentage Ozone Unit Level Share of SeasonNOx Three Year State's 2017 State AverageOzone Ozone Budget for SeasonHeat Season Existing Input (mmBtu) Heatlnput Units (tons) Initial Heat Input Based 2017 Ozone Season NOx Allocation (tons) ' \\ crugi'tif titm* highest ii<>u-/cni Sum column Is values in columns 1' - values lo gel Stale Ihilk-il by 1 lev dl ofnls ciiiumtil. A'-::;.''--..'.-; Column Ms column IS 90966 736,757 736,757 339,851,789 0.002168 25 90967 1,118,066 1,118,066 339,851,789 0.003290 38 90968 1,335,829 1,335,829 339,851,789 0.003931 45 Figure 2. Historic Heat Input Data (Corrected Final Rule Allocations). Plant Name Calculation Anadarko Plant Anadarko Plant Anadarko Plant ORIS Boiler ID ID 3006 y4 3006 '5 3006 h Slop 1SIt'ps 2 & 3Slop 4Slop 5StopG Unit Level StateLevel Unit's Averageof 3 Summation of Percentage Ozone Initial Heat ' " Highest Non-Zero Unit Level 2011 Ozone 2012 Ozone 2013 Ozone 2014 Ozone 2015 Ozone Ozone Season Three Year Share of State's SeasonNOx Input Based 2017 State 2017 Ozone Season Season Season Season Season Heat Inputs from AverageOzone Ozone Budget for Season NOx CAMD Heat Input Heat Input Heat Input Heat Input Heat Input 2011 to 2015 SeasonHeat Season Existing Allocation Unit ID (mmBtu) [mmBtu) [mmBtu) (mmBtu) [mmBtu) (mmBtu) Input (mmBtu) Heatlnput Units (tons) (tons) \v etugcoflhree highest nn-/ert> Sum column Is valuesiu columns 1-'- values in get St ill e ilhiilti! by 1 Icvellnials colii mill. Column M s ciilu mulS 90966 1.310.612 1 426.485 1 081.944 736,757 1,256,347 341,127,969 0.003683 42 90967 1.310.612 1.426.485 1,118,066 1,285,054 341,127,969 0.003767 43 90968 1 310.612 1.426.485 1,335,829 1,357,642 341,127,969 0.003980 45 In the current Final Rule, heat input data for 2011-2014 is effectively treated as "zero." This is historically inaccurate and distorts the calculations. Under this method, the Step 6 initial heat input-based allowance allocations for the Anadarko Plant units are 25 tons, 38 tons, and 45 tons, respectively. Under the corrected method where available EIA heat input data from the years 2011-2014 is backfilled, the Step 6 initial heat input-based allowance allocations are 42 tons, 43 tons, and 45 tons, respectively. Figure 3 shows the historic NOx emissions data EPA relied on in performing the final allowance allocation calculations (Steps 7-10) for the Anadarko Plant units in the current Final Rule, while Figure 4 shows the historic NOx emissions data EPA should have relied on: Figure 3. Historic NOx Emissions Data (Current Final Rule). Plant Name Calculation Anadarko Plant Anadarko Rant Anadarko Rant ORIS Boiler ID ID 3006 4 3006 5 3006 6 CAMD Unit ID 2008 Ozone Season NOx 2009 Ozone Season NOx 2010 Ozone Season NOx 2011 Ozone Season NOx 2012 Ozone Season NOx 2013 Ozone Season NOx 2014 Ozone Season NOx 2015 Ozone Season NOx Ozone FirialTransport Season NOx Rule Unit Level Maximum NOx Ozone Historic Season Emissions Emissions Emissions Errassions Emissions Emissions Emissions Emissions Baseline Allocation 2017 (tons) (tons) (tons) (tons) (tons) (tons) (tons) (tons) (tons) (tons) [Lesser tifcolumns Highest \ ulue X ami O of columns \V n \ 90968 17 7 Active 30698243.11 9 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00009 Figure 4. Historic NOx Emissions Data (Corrected Final Rule). Plant Name Calculation Anadarko Rant Anadarko Rant Anadarko Rant ORIS Boiler ID ID 3006 '4 3006 5 3006 6 CAMD Unit ID 2008 Ozone Season NOx (tons) 2009 Ozone Season NOx Emissions (tons) 2010 Ozone Season NOx Emissions (tons) 2011 Ozone Season NOx Emissions (tons) 2012 Ozone Season NOx Emissions (tons) 2013 Ozone Season NOx Emissions (tons) 2014 Ozone Season NOx Emissions (tons) 2015 Ozone Season NOx Emissions (tons) Ozone Final Transport Season NOx Rule Unit Level Maximum NOx Ozone Historic Season Baseline Allocation 2017 (tons) [Lesser iil'eoiuiims Highest value \ ami O >nf columns 1* - rcupporlitiinmntf \\ 112 102 99 110 52 112 102 99 110 53 112 102 110 56 In the current Final Rule, EPA's approach--which effectively assumes the units had "zero" emissions for 2008-2014--sets Step 8 maximum emissions "upper bound" limits of 9 tons, 14 tons, and 17 tons, respectively. These values serve as "caps" on final allocations, and effectively override and replace the initial Step 6 allocations from Figure 1. Thus, final allocations under Steps 9 & 10 are 9 tons, 14 tons, and 17 tons, respectively. Under the corrected method where available EIA NOx emissions data for 2008-2014 is backfilled,25 the Step 8 maximum emissions "upper bound" limits are significantly higher, at 119 tons for all three Anadarko Plant units. This is a much more generous cap, allowing for final allocations under Steps 9 & 10 that are based on taking the initial Step 6 allocations from Figure 2 and adding a reapportionment adjustment. This amounts to final allocations of 52 tons, 53 tons, and 56 tons, respectively (161 combined)--or an additional 43 tons, 39 tons, and 39 tons, respectively (an additional 121 tons combined). Table 1 summarizes the impact of the corrected calculations on the Anadarko Plant units: Table 1. Impact of Performing EIA Data-Substitution on Anadarko Plant Unit Allocations. Plant Name AnadarkoPlant AnadarkoPlant AnadarkoPlant ORIS II) Boilerll) Proposed/linal 3006 '4 3006 5 3006 '6 O/nm- Si-asnn \<>N M:i vim imi llislorii- IIum-Mik(InnO lYnpiiM'd ( s\ru l-inal ( s \I*U 145 9 145 14 145 17 I'inal 1 rans|HirlKnU 1 nil 1 r\rl \Ox <>/mih ( haniH 11-inal Ni-nsun Mlm-alinn^O1 " (inn\) ( N\Plt In PrupiiM d ( S\|*|< l-inal < Mlnralinn w l.l \ D:il:i Nralin^l-iM Mi.no 77 9 III 4.3 77 14 .3') 77 17 .3') In sum, under the corrected method, the Anadarko Plant units should receive a combined total of approximately 161 additional allowances (39-43 additional allowances each). See Appendix A-l for a complete version of Table 1 showing the impact of EIA data-substitution on all Oklahoma unit allocations. 25 NOx emissions data for 2014 was not readily available to Western Farmers. Excluding the 2014 data is not expected to impact these calculations because 2014 emissions for the Anadarko Plant units are likely to be lower than 2008-2013 emissions. Active 30698243.11 10 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00010 C. Western Farmers Could Not Have Reasonably Anticipated EPA's Error The grounds for Western Farmer's objection arose after the public comment period. Before the Final Rule was issued, Western Farmers could not have expected or anticipated that EPA would set the Anadarko Plant unit allocations based on a single historic baseline year and only CAMD data. Both the Proposed and Final Rules call for EPA to set unit allocations using multiple historic baseline years and EIA data-substitution. EPA did this in the Proposed Rule, but not in the Final Rule. What EPA did in the Final Rule looks like a clear technical error. It would be impracticable, if not impossible, for Western Farmers to have anticipated and submitted comment on an error that EPA had not yet committed during the public comment period, but committed for the first time in the Final Rule. Western Farmers therefore had no meaningful opportunity to comment on the final allocations. EPA did not provide notice of or seek comment on the possibility of switching from multiple to single historic baseline years, or from using both CAMD and EIA data to only CAMD data. Had EPA done so, Western Farmers would have submitted comments that these changes would arbitrarily and unfairly penalize the Anadarko Plant units. As it stands, Western Farmers was caught by surprise by the new unit-level allocation data inputs. EPA may not "use the rulemaking process to pull a surprise switcheroo on regulated entities." Envtl. Integrity Project, 425 F.3d at 998. EPA's data errors have left Western Farmers with a major and unexpected allowance shortfall for the 2017 ozone season. Correcting these errors is critical to Western Farmers' ability to plan for and achieve compliance with the CSAPR Update Rule, especially given the extremely near-term compliance timeframe for the 2017 ozone season. D. EPA Should Correct These Technical Errors Through a Direct Final Rule EPA should issue these technical corrections through a direct final rule as an alternative to full notice-and-comment rulemaking.2276 Direct final rules rely on the Administrative 26 The only difference between the Proposed and Final Rule was a one-year shift in the relevant baseline periods. Because 2015 data was not yet available, the Proposed Rule calculated allowance allocations based on a 2010-2014 heat input baseline period, and a 2007-2014 NOx emissions baseline period. The Final Rule uses a 2011-2015 heat input baseline period and a 2008-2015 NOx emissions baseline period. See Allocation TSD at 6. 27A direct final rule is "a rule that is issued in final form, without prior notice and comment, that becomes effective on a particular date unless adverse comment is submitted within a specified period of time." EPA Office of General Counsel, Direct Final Rulemaking Guidance for EPA Rule Writers, Attachment 1, 2 (1998). Available at https://vosemite.epa.gov/oaqps/rdms.nsf/591 caf4ab 155e210852566de00539f57/c92ad 1453ad5de6885256728006a0f 30!QpenDocument ("Direct Final Rulemaking Guidance"). In conjunction with a direct final rule, EPA's typical practice is to simultaneously publish a separate, parallel proposed rule. If EPA receives significant adverse comments on the direct final rule, the Agency will withdraw the direct final rule and address the public comments in a subsequent final rule based on the parallel proposed rule. Id. at 4. Western Farmers encourages EPA to act on this parallel tract to ensure the most efficient and timely resolution of EPA's data errors. Active 30698243.11 11 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00011 Procedure Act's "good cause" exemption from notice-and-comment rulemaking,28 while giving the Agency "the benefit of any public input that may unexpectedly surface." Direct Final Rulemaking Guidance at 2 (citation omitted). EPA uses direct final rules for "noncontroversial rules where [the Agency does not] expect adverse comment," including routine or minor actions. Id. at 4, 6. EPA previously has issued direct final rules under the CSAPR program. See, e.g.. Revisions to Federal Implementation Plans To Reduce Interstate Transport of Fine Particulate Matter and Ozone, 77 Fed. Reg. 10,342 (Feb. 21, 2012). Flere, the corrections to EPA's allocation calculations are a non-controversial, minor action unlikely to attract adverse comment because they (1) would affect only a subset of units within a single state; and (2) are necessary to correct clear technical errors in EPA's application of the Final Rule's allocation methodology. Further, time is of the essence. A direct final rule generally is a more efficient procedural mechanism than notice-and-comment rulemaking. The 2017 ozone season begins in just over five months. Western Farmers needs to get the correct amount of allowances on the books for their units as soon as possible in order to plan, and carry out, its compliance strategy for the 2017 ozone season. II. OKLAHOMA STATE BUDGET A. EPA's New Budget Calculation Method Arbitrarily Slashed Oklahoma's Budget Oklahoma's budget decreased more than that of any other state between the Proposed and Final Rules, dropping by more than 25 percent, from 16,215 tons to 11,641 tons. Oklahoma's Final Rule budget is also about 49 percent lower than its 2016 ozone season budget, decreasing from 22,694 tons to 11,641 tons.2298 In contrast, most (15 out of 22 Group 2) CSAPR Update states saw their budgets increase in the Final Rule. Only six other states saw their budgets decrease, but none as significantly as Oklahoma. The drastic reduction in Oklahoma's budget was driven by significant revisions EPA made to its budget calculation formula. Most critically, EPA introduced a NOx emission rate "delta" value that is derived from IPM. For certain states, including Oklahoma, the IPMpredicted emission rates are blatantly arbitrary. When plugged into EPA's new formula, they significantly over-predict available emissions reductions and significantly under-predict emissions levels. This results in an unrealistically low state budget. Western Farmers proposes a limited technical fix to EPA's formula that would help avoid this outcome while preserving EPA's Final Rule methodology. 28 5 U.S.C. 553(b)(3)(B) (Section 553's notice-and-comment requirement does not apply "when the agency for good cause finds . . . that notice and public procedure thereon are impracticable, unnecessary, or contrary to the public interest."). 29 Under the Proposed Rule, this decrease was approximately 29 percent, a less drastic change. Active 30698243.11 12 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00012 1. EPA's Changes to the State Budget Formula EPA changed its state budget calculation formula significantly between the Proposed and Final Rules. Specifically, EPA introduced additional variables, including (i) an IPM-derived NOx emission rate "delta;" and (ii) an adjusted historic emission rate based on a newlydeveloped dataset. In both the Proposed and Final Rule, EPA set state budgets as the minimum of either: (1) Flistoric actual emissions (2014 for the Proposed Rule; 2015 for the Final Rule); or (2) IPM-predicted 2017 emissions. See 81 Fed. Reg. 74,548. What changed--significantly--between the Proposed and Final Rules is EPA's process for calculating IPM-predicted 2017 emissions. For the Proposed Rule, EPA's method can be represented by the following formula: IPM-predicted 2017 Emissions (tons) = (Historic Heat Input, 2014) x (IP.\I Emission Rule, 201 7 SI.200 ion C 'osi ('use)!0 For the Final Rule, EPA's new method requires a much more complex formula: IPM-predicted 2017 Emissions (tons) = (Historic Heat Input, 2015) x \(.ldjusied 20/5 Emission Rule) - \ flPM Emission Rule. 20I~ Ruse ('use) (IP.\I Emission Rule. 2017 SI.400 ion ('osi ('use) \ \ According to EPA, both formulas reflect a basic method of "multiplying historical statelevel heat input [i.e., first parentheses on right-hand side of the equations] by state-level emission rates that reflect EGU NOX reduction potential [i.e., second parentheses on right-hand side of the equations, highlighted in yellow]." Id. at 74,547. In the Proposed Rule, the "NOx reduction potential" emission rate was represented simply by an IPM Cost Case emission rate. In the Final Rule, the "NOx reduction potential" emission rate is represented by a more complex calculation30 31 30 See id. at 74,547 ("[T]he proposed CSAPR Update put forward a methodology to set emission budgets by multiplying monitored historical state-level heat input by model-projected 2017 state-level emission rates. The monitored historical data were based on 2014, which was the most recent complete ozone season dataset at the time of the proposal."). 31 See id. ("For the final CSAPR Update rule, the EPA is refining its methodology for establishing emission budgets that reflect EGU NOX reduction potential by using historical state-level NOX emission rates adjusted by modeled NOX reduction potential. Specifically, the final rule's approach applies the change in modeled 2017 state-level emission rates (the budget-setting base case 2017 projected rates minus the cost threshold modeling 2017 projected rates) to historical 2015 state-level NOX emission rates.") Active 30698243.11 13 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00013 involving three variables: an adjusted historic emission rate, an IPM Base Case emission rate, and an IPM Cost Case emission rate. The only other step in the formulas is to multiply the "NOx reduction potential" emission rate by historic heat input to derive IPM-predicted 2017 emissions. This updated formula reflects a new, multi-step process for calculating the NOx reduction potential emission rate that entails (1) first, calculating a NOx emission rate "delta," based on the difference between an IPM Base Case emission rate and an IPM $ 1,400/ton Cost Case emission rate; and (2) second, applying this "delta" to an adjusted historic emission rate. See id. at 74,548. This latter step requires assuming that "the potential of each state to improve its historical NOX rate by the same degree that it is projected to improve its NOX rate when moving between the budget-setting base case 2014 projection and cost threshold projection." Id. at 74,547. Finally, this "NOx reduction potential" emission rate is multiplied by historic heat input. The result is the final IPM-predicted 2017 emissions. As under the Proposed Rule, if this result is lower than historic actual emissions, then EPA bases the final state budget on the IPM-predicted 2017 emissions. If this result is higher than historic actual emissions, then EPA bases the final state budget on actual historic emissions. Notably, EPA introduced several entirely new variables in the final state budget formula. EPA introduced an IPM Base Case emission rate value into the new formula, while the Proposed Rule included only an IPM $ 1,300/ton Cost Case emission rate. The new formula also required the Agency to develop an "adjusted historic dataset" in order to derive the adjusted 2015 emission rate values. See id. This adjusted dataset started with actual historic 2015 heat input and NOx emissions data. EPA then adjusted the actual historic data for three categories of "known changes in the power sector occurring between 2015 and 2017": (1) announced new selective catalytic reduction ("SCR") at existing EGUs; (2) announced coal-to-gas conversions; and (3) announced retirements. Id. According to the Agency, these adjustments "ensure that the emission budgets established by this rule reflect EGU NOX reductions both from already announced power sector changes and further EGU NOX reductions quantified in the EPA's EGU NOX reduction potential analysis." Id.32 2. Perverse IPM Result EPA's new Final Rule budget-setting methodology yields the following results for Oklahoma: 32 EPA may have relied on inaccurate and over-simplified assumptions in calculating adjusted historic 2015 emission rates. For example, EPA's assumption that retirements necessarily will lead to a reduction in the state's average emission rate may not adequately account for the effects of intrastate or interstate load-shifting. Active 30698243.11 14 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00014 Table 2. Comparison of 2015 Actual, 2015 IPM-Predicted, and 2017 IPM Base Case and Cost Case Results for Oklahoma Under the Final Rule. 2015 Actual 2015 Actual. Adjusted 2017 IPM Rase Case 2017 IPM (Final SI,400/ton) CSAPR budget (Final Rule) Heat Input (MMRtu) 250,108,. 90 243,267,181 208,776,019 207,804,804 256,168,790 N()\ (Ions) 13,922 13,055 16,506 14,720 11,641 Emission Rate (Ih/MMRlu) 0.109 0.107 0.158 0.142 0.091 The Perverse IPM Result is highlighted. Importantly, IPM predicts that the 2017 Base Case emission rate for Oklahoma would be substantially higher than the historic actual or adjusted 2015 emission rate. In other words, IPM predicts that in a world without the CSAPR Update Rule, Oklahoma's NOx emission rate would increase from 0.107 lb/MMBtu in 2015 to 0.158 lb/MMBtu in 2017--an increase of over 47 percent in just two years, without any predicted increase in heat input. Similarly, IPM predicts that the 2017 $ 1,400/ton Cost Case emission rate would be substantially higher than the historic emission rate. So, even in a world with the CSAPR Update Rule, Oklahoma's NOx emission rate would increase from 0.107 lb/MMBtu in 2015 to 0.142 in 2017--an increase of over 32 percent in just two years, again without any predicted increase in heat input. The IPM Base Case prediction that the 2017 NOx emission rate will increase to 0.158 lb/MMBtu is perverse, arbitrary, and demonstrably flawed. First, 0.158 lb/MMBtu is a significant increase from both the actual (0.109 lb/MMBtu) and adjusted (0.107 lb/MMBtu) 2015 emission rates. Second, the fact that EPA's adjusted 2015 emission rate is lower than the actual 2015 emission rate shows that EPA anticipates that announced SCRs, conversions, and retirements will lower the emission rate by approximately 0.002 lb/MMBtu between 2015 and 2017. This cuts against the IPM prediction of a significant rate increase. Third, CAMD data for 2016 shows a NOx emission rate of 0.101 lb/MMBtu for Oklahoma,33 which is comparable to (even lower than) the actual or adjusted 2015 emission rates and nowhere near the 2017 IPM Base Case rate of 0.158 lb/MMBtu. Fourth, the IPM prediction would require rapid degradation and/or shut-down of existing NOx control technology in Oklahoma. The effectiveness of existing NOx controls is highly unlikely to degrade at the fast pace IPM predicts. Indeed, EPA has provided no apparent explanation for why the NOx emission rate would increase so significantly in just two years. 33 See https://ampd.epa.gov/ampd/. Active 30698243.11 15 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00015 For similar reasons, the IPM Cost Case prediction that the 2017 NOx emission rate will increase to 0.142 lb/MMBtu also is arbitrary and demonstrably flawed. Further, comparing the IPM Cost Case emission rate with the emission rate resulting under Oklahoma's final CSAPR budget indicates that EPA does not trust its own model. If IPM were relatively accurate, the IPM Cost Case emission rate (0.142 lb/MMBtu) would be closer to the final budget value (0.091 lb/MMBtu), not more than 55 percent greater. The fact that EPA does not rely directly on the IPM Cost Case emission rate to calculate state budgets (consistent with the Proposed Rule) shows that EPA is aware of the questionable reliability of direct use of IPM emission rate outputs. Nonetheless, EPA relied on these IPM outputs--which are patently arbitrary in Oklahoma's case--in setting state budgets. 3. Oklahoma's Budget Calculation The Perverse IPM Result led to an unrealistically and arbitrarily low budget for Oklahoma. Critically, the Perverse IPM Result overestimates the amount of emissions reductions available in the state. That is, because the IPM Base Case emission rate is unrealistically high, the emission rate "delta" between the 2017 IPM Base Case and Cost Case also is unrealistically high.34 This error--a NOx emission rate "delta" that is too high--then carries through the entire calculation, ultimately resulting in final IPM-predicted 2017 emissions that are too low. For Oklahoma, these IPM-predicted 2017 emissions were lower than its historic 2015 emissions. Accordingly, EPA set Oklahoma's final budget equal to these arbitrary and inaccurate IPM-predicted emissions. Below is a step-by-step overview of EPA's calculation for Oklahoma. Values affected by the Perverse IPM Result are highlighted: Values that arc too high because of the Perverse IPM Result are highlighted in green. Values that are too low because of the Perverse IPM Result are highlighted in red. Step 1. Calculate NOx Emission Rate "Delta" = (IPMEmission Rate, 2017 Base Case) (IPM Emission Rate, 2017 $1,400/ton Cost Case) 34 Even if IPM accurately predicted the relative difference between the Base Case and Cost Case emission rates, it would be inappropriate to apply this absolute "delta" value directly to an actual emissions rate. To take a simplified, extreme example: State O has an actual emissions rate of 0.1 lb/MMBtu. IPM predicts a Base Case emission rate of 0.3 lb/MMBtu and a Cost Case Emission rate of 0.2 lb/MMBtu. The "delta" between the IPM rates is 0.1 lb/MMBtu. Applying this "delta" to the historic emissions rates leaves you with an emission rate of 0.0 lb/MMBtu! The moral of the story is that if IPM-predicted rates are unrealistically high, the "delta" between these values is not something that can just be applied to a real emission rate. Indeed, in extreme cases, this "delta" can even be equal to or even bigger than the real emission rate. But if the IPM Base Case rate was effectively "capped" at the historic emission rate (0.1 lb/MMBtu in our example), there is no way the "delta" between the two IPM rates could be greater than the historic rate (the "delta" could only range from 0.0 lb/MMBtu to 0.1 lb/MMBtu in our example). This "capping" is effectively what the proposed scaling fix described below would accomplish. Active 30698243.11 16 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00016 = (0.158 Ib/MMBtu) -10-142 Ib/MMBtu) = 0.010 Ib/MMBtu In Step 1, the calculated "delta" value (0.016 lb/MMBtu) is arbitrarily high due to the Perverse IPM Result. Western Farmer's analysis, discussed in the following section, indicates that a more realistic value would be closer to 0.011 lb/MMBtu--over 30 percent lower than the current value. This means that EPA likely has significantly overestimated the amount of emissions reductions available in Oklahoma. Step 2. Calculate "NOx Reduction Potential" Emission Rate = (Adjusted 2015 Emission Rate) - (NOx Emission Rate "Delta) " = (0.107 lb/MMBtu) - (Mil lb/MMBtu) 0.001 // \l\lBln In Step 2, applying the too-high "delta" value reduces the adjusted 2015 emission rate by too much--i.e., by an amount greater than what is achievable in reality. This means the resulting "NOx Reduction Potential" emission rate (0.091 lb/MMBtu) is lower than what it should be had EPA relied on a more realistic "delta." Indeed, an emission rate of 0.091 lb/MMBtu is below even the rate achieved by new SCRs (0.10 lb/MMBtu) and NOx combustion controls (0.15 lb/MMBtu, at best). Achieving this rate on a statewide average basis likely would require shutting down fossil-fuel fired EGUs and constructing a sufficient number of new alternative energy sources to replace generation capacity by 2017. Step 3. Calculate IPM-predicted 2017 emissions = (Historic Heat Input, 2015) x ("NOx Reduction Potential" Emission Rate (256,168,790 MMBtu) x <0.091 lb/MMBtu) / 2000 11.641 loii1' In Step 3, multiplying the too-low "NOx reduction potential" emission rate by historic heat input results in IPM-predicted 2017 emissions that are too low--again, lower than what is achievable in reality. Oklahoma's final state budget of 11,641 tons likely is unattainable. To correct the errors associated with the Perverse IPM Result, EPA must adjust the IPM emission rates in Step 1 to more realistic values. Active 30698243.11 17 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00017 B. Proposed State Budget Fix Western Farmers proposes a limited adjustment to EPA's budget calculation for Oklahoma and other states affected by the Perverse IPM Result.35 The proposal retains the Final Rule's state budget calculation formula. Flowever, it calls for an additional "scaling" step for states with the Perverse IPM Result ("Scaling Fix"). This Scaling Fix allows EPA to keep the Final Rule's current methodology, but ensures that the methodology is not applied in an arbitrary manner that achieves unrealistic results. See Appendix B-l for detailed calculations. The proposed Scaling Fix would be applied during Step 1 of EPA's formula (calculation of NOx emission rate "delta"). It has two components: (1) Scaling 2017 IPM Base Case Emission Rate to Reflect Actual 2015 Data. Both 2017 IPM Base Case NOx emissions and 2017 IPM Base Case heat input values would be scaled to actual 2015 values. For Oklahoma, NOx emissions would be scaled from 16,506 tons to 13,922 tons. Similarly, heat input would be scaled from 208,776,019 MMBtu to 256,168,790 MMBtu. (Currently, IPM illogically predicts that emissions will increase while heat input will decrease between 2015 and 2017. This makes predicted NOx emission rates unachievable if fuel consumption remains constant in reality. The Scaling Fix reflects relatively flat levels of fuel consumption and emissions over the next two years.) Oklahoma's scaled 2017 IPM Base Case emission rate would be 0.109 lb/MMBtu. (2) Scaling 2017 IPM Cost Case Emission Rate, Based on a Scaling Factor Derived From Part 1. Both 2017 IPM Cost Case NOx emissions and 2017 IPM Cost Case heat input values would be scaled by a scaling factor derived from part one--i.e., by the same relative amount that NOx emissions and heat input were scaled for the IPM Base Case emission rate. For Oklahoma, the scaling factor would be 0.84 for NOx emissions and 1.23 for heat input. Accordingly, NOx emissions (for the 2017 IPM $ 1,400/ton Cost Case) would be scaled from 14,720 tons to 12,353 tons. Heat input would be scaled from 207,804,804 MMBtu to 255,435,708 MMBtu. Oklahoma's scaled 2017 IPM $1,400/ton Cost Case emission rate would be 0.095 lb/MMBtu. Then, the NOx emission rate "delta" (Step 1 of EPA's formula) would be calculated based on these scaled IPM emission rates. For Oklahoma, this scaled "delta" value would be 0.011 lb/MMBtu, about 31 percent lower than the "delta" value used in the Final Rule (0.016 lb/MMBtu). The rest of EPA's formula (Steps 2 and 3) would then be applied as currently described in the Final Rule. This should result in a final budget for Oklahoma of 12,309 tons, or 35 Western Farmers' analysis indicates that the Perverse IPM Result negatively affects the budgets for six other states: Indiana, Kansas, Michigan, Missouri, Mississippi, and Texas. Active 30698243.11 18 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00018 about 668 additional tons compared to the Final Rule budget. See Table 4 below; see also Appendix B-l.36 C. The Proposed State Budget Fix Would Preserve Air Quality Benefits The proposed Scaling Fix would preserve the Final Rule's air quality benefits. Western Farmers analyzed the impact of applying the Scaling Fix to all seven states whose Final Rule budgets are negatively affected by the Perverse IPM Result on reductions of ozone design values (average and maximum) in 2017 at 19 downwind nonattainment and maintenance receptors.37 As Table 3 shows, these results indicate no change in average reduction of ozone average design values, and a minimal decrease in average reduction of ozone maximum design values, from the Final Rule. See Appendix C-l for detailed underlying calculations. If the Scaling Fix is applied only to Oklahoma, then there is no change in average reduction of ozone average or maximum design values. See Appendix C-2 for detailed underlying calculations. Table 3. Impact of Scaling Fix on Average Ozone Reduction at 19 Nonattainment and Maintenance Receptors, Based on AQAT. Historic Actual 2015 (7 States) Final Rule Scaling Fix (7 Stales) Scaling Fix (Oklahoma Only) Avg. Reduction of O/.onc Average Design Values (ppb) 0.15 0.28 0.28 0.28 Avg. Reduction of O/.onc Maximum Design Values (ppb) 0.16 0.29 0.28 0.29 Further, the Scaling Fix would not dramatically increase state budgets. Table 4 shows the impact of applying the Scaling Fix to the seven affected states. As shown, the Scaling Fix generally would raise state budgets to levels somewhere in-between current Final Rule budget and historic emissions levels. For four states, the Scaling Fix would increase Final Rule budgets by less than 100 tons. Oklahoma is the only state whose budget would increase by more than 36 Appendix B-2 shows updated Oklahoma unit allocation calculations using the EIA data-substitution described in Part I and the proposed state budget Scaling Fix. Under this method, the Anadarko Plant units would receive 55, 56, and 60 units each (171 units combined). 37 Western Farmers used the same AQAT-based methodology that EPA describes in the Regulatory Impact Analysis (consistent with the Ozone Transport Analysis Final Rule TSD and the preamble to the CSAPR Update Rule). See Regulatory Impact Analysis at 3-9 to 3-10. Active 30698243.11 19 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00019 500 tons. Overall, the Scaling Fix would increase all seven state budgets by a combined total of just over 1,400 tons:38 Table 4. Impact on State Budgets of Applying the Proposed Scaling Fix. Stale Historic Actual 2015 1.missions (tons) Final Rule Budget (tons) Scaling Fix Budget (Ions) Change (Final Rule and Scaling Fix Budget) (tons) Indiana 36,353 23,303 23,377 74 Kansas 8,136 8,027 8,027 0 Michigan 21,530 17,023 17,377 354 Missouri 18,855 15,780 15,798 18 Mississippi 6,438 6,315 6,336 21 Oklahoma 13,922 11,641 12,309 668 Texas 55,409 52,301 52,591 290 Total 160,643 134,390 135,815 1,425 D. Western Farmers Could Not Have Reasonably Anticipated the Changes in EPA's Methodology Western Farmers' could not have reasonably anticipated that EPA would pull a "surprise switcheroo" and change its state budget calculation methodology so significantly between the Proposed and Final Rules, slashing Oklahoma's budget by more than 25 percent. In particular, EPA did not provide notice that the Agency would change its formula for calculating modelpredicted emissions to include (1) adjusted historic emission rates based on a newly-developed dataset; or (2) IPM-derived emission rate "deltas" that rely on the difference between IPM Base 38 Western Farmers' preliminary analysis indicates that applying the Scaling Fix to all 22 CSAPR Update states likely would decrease the budgets for 14 out of the 15 states without the Perverse IPM Result (all states but Virginia). However, further state-by-state scrutiny would be needed to ascertain the impact of the Scaling Fix on all CSAPR Update states. Further, Western Farmers notes that the Scaling Fix is specifically designed to address a modeling flaw that is common to a small subset of states. The Scaling Fix may not be appropriate to apply to states without the Perverse IPM Result. Active 30698243.11 20 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00020 Case and IPM Cost Case emission rates.39 Nor did EPA provide notice that it would finalize a methodology that entailed applying IPM-derived emission rate "deltas" to historic emission rates.3409 Had EPA done so, Western Farmers would have analyzed and submitted comments on the impact on Oklahoma's budget of EPA's unrealistic and arbitrary assumptions, including the Perverse IPM Result. EPA characterizes the changes in the Final Rule's formula as a "refinement" to the proposed approach. See 81 Fed. Reg. 74,547.41 However, the alterations go far beyond mere refinements. First, as EPA acknowledges, the Agency had to develop an adjusted historical dataset to derive the adjusted historic emission rates. Id. Second, EPA had to add at least two additional steps to its calculation methodology. The Proposed Rule's methodology consisted of one basic step: multiply historic heat input by an IPM Cost Case emission rate. The Final Rule's methodology is a complex, three-step process: (1) calculate an emission rate "delta" by subtracting an IPM Cost Case emission rate from an IPM Base Case emission rate; (2) calculate a "NOx reduction potential" emission rate by subtracting this "delta" from an adjusted historic emission rate; and (3) calculate IPM-predicted 2017 emissions by multiplying the "NOx reduction potential" emission rate by historic heat input. Id. at 74,548. This three-step process, based in part on a newly-developed historical dataset, is no logical outgrowth of the Proposed Rule's single-step process that did not rely on "adjusted" historic data. For these reasons, EPA must reconsider, at least in the proposed limited fashion, the CSAPR Update Rule. CONCLUSION In sum, Western Farmers requests that EPA recalculate (1) unit-level allowance allocations for the Anadarko Plant units, using the data-substitution method called for in Final Rule and based on monthly (not annual), reported (not adjusted) data; and (2) Oklahoma's state budget, using the proposed limited "scaling" adjustment to the Final Rule's existing calculation methodology to ensure arbitrary IPM outputs do not distort the calculation. 39 Notably, the Proposed Rule include did not include an IPM 2017 Base Case emission rate in its formula, only an IPM Cost Case emission rate. The IPM 2017 Base Case emission rate is the primary driver of the Perverse IPM Result. 40 While EPA sought comment "on all aspects of quantifying state emissions budgets reflecting upwind obligations," 80 Fed. Reg. 75,739, it is unreasonable to expect that commenters could have anticipated and commented on all possible permutations of EPA budget calculation formula. EPA's blanket request for comment does not amount to notice that the Agency would introduce new variables, complex calculation steps, or underlying datasets--much less all three. 41 EPA further asserts that the "final rule methodology for setting emission budgets reflects the CSAPR Update proposal in that it retains the approach of multiplying historical state-level heat input by state-level emission rates that reflect EGU NOX reduction potential." Id. As EPA seems to implicitly acknowledge, any methodological similarities between the two approaches exist at only a very high level. Critically, as discussed above, EPA's approach for determining "state-level emission rates that reflect EGU NOX reduction potential" changed dramatically between the Proposed and Final Rule. Active 30698243.11 21 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00021 Dated: December 22, 2016 Respectfully submitted, /x/ Brian Hobbs Enclosures (CD-Rom): Appendix A-l: Appendix A-2: Appendix B-l: Appendix B-2: Appendix C-l: Appendix C-2: Brian Hobbs VP, Legal & Corporate Svcs. Western Farmers Electric Cooperative Unit-Level Allocations with EIA Substitution (Final Rule) EIA Data Substitution Calculations Revised Appendix E State Budgets (Scaling Fix) Revised Unit-Level Allocations with EIA Substitution (Scaling Fix) AQAT (Scaling Fix -- 7 States) AQAT (Scaling Fix - Oklahoma Only) Active 30698243.11 22 17cv1906 Sierra Club v. EPA - 6/22 Production ED 001523 00005329-00022