Document XOY89QO1qZEpB1VLdGLVMd83B

TALEN I1 MONTANA practicable and cannot sustain emissions this low. In 2022, Colstrip achieved approximately 0.022 lb/MM13tu on an annual basis, far above EPA's assumption of the 99:". percentile of the lowest quarter. EPA should use all data available from coal-fired EGUs except as noted below with respect to units co-firing natural gas and units with an early retirement date to provide a full picture of achieved fPM emission rates. At the least, EPA should provide justification for its selection of the data, why reliance on the selected data is appropriate, and why certain quarterly data from 2017, 2019, and 2021 were excluded, so that interested stakeholders can verify the accuracy and representativeness of the underlying unit-specific ciata.32 Among other issues in the evaluation, EPA: Included some units that will be converted to gas in 2025. Did not include data for all quarters but instead selected only quarters with the lowest emissions for some units and excluded other quarters with higher emissions (peaking for some units, ramping for others). Excluded some units with no current plans to retire or switch to gas. Included some units that have a federally enforceable requirement to cease coal combustion by December 31, 2028 (despite stating that the evaluation excluded coal-tired ECilis that will retire by that date). Used the last day of a quarter in some cases and the average of 30-day averages for others. Included only certain test runs in conducting its distribution analysis.i3 As to the last point above, EPA should use a historical data pool that encompasses data from different times of year and operating conditions. EPA should include all affected units and all operating quarters in its analysis. Without a more comprehensive data pool, it is difficult to see how EPA could conduct a proper statistical analysis to justify the proposed fPM limit. Talon Montana strongly recommends that EPA correct the deficiencies identified above, as well as make its statistical analysis or Python code used for the fPM evaluation available for public review, to ensure that the proposed fPM limit is not deemed arbitrary and capricious. It is confusing as to which units EPA includcd'excludcd, and as to which quarterly data sets wcrc includcd'excludcd EPA failcd to explain its rationale for determining which units and data sets should be included or excluded The lack of explanation, coupled with the large number of supporting documents in the docket, makes it extremely difficult to identify the unit-specific data complicit analy7ed, and ultimately relied upon by EPA and, more importantly, to meaningfully review EPA's evaluation For the same reasons articulated in in 32, it is confusing as to which test runs EPA included/ excluded in its distribution analysis, and EPA's lack of rationale for how it determined which test runs to use 11 Sierra Club FOIA 2025-EPA-04883 ED_018388_00000307-00012 SC_EVERSPLIT0006108