Document 3NBDbEkDpNgvbE5NQMbaXx0rE

USCA Case #24-1287 Document #2077531 Filed: 09/30/2024 Page 11 of 81 not equipped with mercury controls, such as a PAC injection system. 27. EPA's lack of data is compounded by another problem. In the Final Rule, EPA admitted that setting numerical MACT floor limits for emissions with test runs consistently below the detection limit (BDL) "is not economically and technically feasible." 89 Fed. Reg. at 55707-08. Yet it did exactly that here. For example, EPA adopted MACT floor limits for acid gases in SunCoke's HRSG main stacks and polycyclic aromatic hydrocarbons (PAHs) for SunCoke's pushing operations--even though these HAPs tested "below the detection Emit" 50% of the time. 89 Fed. Reg. at 55707. In the pushing category, mercury tested BDL 40% of the time. This means EPA had even fewer data to work with when setting the MACT floor limits. 28. To calculate the MACT floors, EPA applied its UPL statistical method to its limited dataset to account for variability and uncertainty in emissions data. The UPI, calculation estimates the true average and true variance. Estimating the true average is possible with a small number of samples, but the same is not true of the estimation of the variance, which requires a substantially larger number of samples. Specifically, samples that cover the range of varying factors are required (Page 11 of Total) - 11- Sierra Club FOIA 2025-EPA-04883 ED_018388_00000109-00110 SC_EVERSPLIT0005784