Technical Support Document (TSD) Addressing Big Escambia Data Requirements Rule (DRR) Modeling for the Purpose of Evaluating Interstate Transport
1.	Purpose:
      The following provides further information and analysis to support EPA's proposed conclusions in section III.C.1.b. of the notice of proposed rulemaking (NPRM) pertaining to Alabama's August 20, 2018, state implementation plan (SIP) revision addressing prongs 1 and 2 of CAA section 110(a)(2)(D)(i)(I) for the 2010 1-hour sulfur dioxide (SO2) National Ambient Air Quality Standard (NAAQS).  For the purposes of assessing interstate transport of SO2 from Alabama into Florida, EPA performed additional analyses to evaluate supplemental information from the Alabama Department of Environmental Management (ADEM) that addresses the issues with the original modeling submitted by ADEM in January of 2017 for the Escambia Operating Company - Big Escambia Creek Plant (Big Escambia) located in Escambia County, Alabama, performed under the DRR.  This TSD summarizes the issues with the original DRR modeling for Big Escambia  and how ADEM has addressed them for the purpose of assessing interstate transport of SO2.   
2.  Background
      This TSD explains EPA's evaluation and proposed conclusions regarding ADEM's August 20, 2018, interstate transport SIP submission as it relates to the Big Escambia Supplement.  EPA's analysis of any potential impacts of Big Escambia on other states described in section III.C.1.b. of the NPRM is to assess whether Alabama's SO2 emissions will contribute significantly to nonattainment or interfere with maintenance of the 2010 1-hour SO2 NAAQS in neighboring states.
      EPA's proposed and final TSDs for the SO2 round 3 designations for the the 2010 1-hour SO2 NAAQS identified the following issues with the Big Escambia modeling:
 Lack of adequate documentation to support the American Meteorological Society/Environmental Protection Agency Regulatory Model Meteorological Processor (AERMET) modeling used to generate the surface and upper air meteorology files;  
 No documentation regarding how the surface characteristics of the area of analysis were developed;  
 Possible contribution to SO2 modeled concentrations from a nearby source (Georgia-Pacific Brewton LLC) not included in the modeling;  
 Lack of adequate documentation to demonstrate that appropriate SO2 emissions rates were used in the modeling for Big Escambia; and, 
 The inappropriate use of background concentrations from the Centreville Southeastern Aerosol Research and Characterization (SEARCH) network and Mammoth Cave National Park (Mammoth Cave) ambient monitoring sites. 
		In a letter dated October 19, 2017, Alabama provided a response to the air dispersion modeling issues that EPA identified in the proposed designations TSD.  The responses provided by Alabama indicate that the meteorological data was processed consistent with the newest version of AERMET guidelines and in consultation with EPA; that an analysis using American Meteorological Society/Environmental Protection Agency Regulatory Model Land Cover Processor (AERSURFACE) was performed in accordance with the newest version of the AERSURFACE User's Guide and provided by Alabama to the facility; that Georgia-Pacific Brewton LLC, a nearby source in Alabama, was not included in the modeling analysis because it was greater than 20 kilometers (km) away, which was the distance that Alabama used to screen nearby sources for inclusion in the modeled analysis; and that the background concentration issue was addressed sufficiently in previous correspondence.  Alabama did not submit a revised modeling analysis for Big Escambia at that time.
		The responses provided by Alabama in the October 19, 2017, submittal did not resolve any of the issues that EPA identified in the proposed TSD for designations.  EPA was still unable to confirm that the procedures used for the processing of the meteorology were appropriate given that the AERMET and AERSURFACE files were not provided.  Alabama did not address whether the nearby source that was not modeled, Georgia-Pacific Brewton LLC, could contribute to SO2 modeled concentrations within the vicinity of the Big Escambia facility.  No additional information was provided to demonstrate that appropriate SO2 emissions rates were used in the modeling for Big Escambia.  A comparison of the emissions used in the modeling with the emissions for the facility contained in the 2014 National Emissions Inventory (NEI) showed a discrepancy and further clarification was needed from the State.  Finally, no additional information was provided to address the inappropriate use of the Centreville SEARCH and Mammoth Cave ambient monitoring sites for the background concentrations.  Additional rationale explaining EPA's concerns is contained in EPA's proposed and final designations TSDs, Chapter 3.
3.  EPA Analysis of the Big Escambia Supplement
		In the Big Escambia Supplement, Alabama provided additional information to address the previously identified issues with the Big Escambia modeling.  Alabama submitted the AERMET and AERSURFACE files used in the modeling so that EPA could confirm that the procedures used for the processing of the meteorology are appropriate.  EPA has confirmed that the processing of the meteorology files for Big Escambia was done appropriately for the purpose of assessing interstate transport of SO2 from Alabama into Florida.  Alabama indicated that the concern over the exclusion of nearby source Georgia-Pacific Brewton LLC, located approximately 24 km from Big Escambia, had been addressed in prior communication with EPA.  Although Georgia-Pacific Brewton LLC is located more than 20 km from Big Escambia, EPA identified this source as a possible contributor to Big Escambia's SO2 modeled concentrations based on 2014 SO2 emissions of 972 tons per year (tpy).  EPA reviewed more recent (2017) emissions for Georgia-Pacific Brewton LLC.  The 2017 SO2 emissions are 103 tpy.  Based on the 2017 SO2 emissions for Georgia-Pacific Brewton LLC and the distance of 24 km from Big Escambia, EPA believes that the exclusion of Georgia-Pacific Brewton LLC from the model does not render the model invalid for use in assessing interstate transport of SO2 into the neighboring state of Florida.
		To address the issue regarding the modeled emissions rates, Alabama provided additional documentation to explain why the modeled emissions from Big Escambia differ from the NEI emissions values for the modeled years.  Alabama indicated that the difference in emissions is the result of startups, shutdowns, and malfunctions (SSM) of the thermal oxidizer unit and that per draft EPA guidance, SSM emissions are exempted from the modeling.  Alabama indicated that the discrepancy in emissions for 2014 was due to an unusual amount of SSM events for the unit.  Alabama provided documentation on December 2, 2019, that indicated the discrepancy in emissions for each of the modeled years was due to acid gas being diverted to a flare, unit FL-02, when the thermal oxidizer was down during SSM/upset events.  The information provided by ADEM explains the discrepancy between the modeled emissions and the emissions values contained in the 2014 NEI.  However, the information is inadequate for demonstrating that the SSM/upset emissions are not continuous enough or frequent enough to contribute significantly to the annual distribution of maximum daily 1-hour concentrations, which would enable them to be classified as intermittent, and therefore excluded from the modeling per EPA's draft guidance.[9]   While the emissions due to SSM/upset events should have been included in the modeling analysis, EPA believes that the nature of the flare and the fact that the modeled SO2 concentrations from all of the Big Escambia sources combined are relatively low at receptors across the border in Florida make it highly unlikely that the additional SSM/upset emissions from the flare, had they been included in the model, would have increased modeled concentrations in Florida to a level above the 2010 1-hour SO2 NAAQS.  Therefore, EPA does not believe that the omission of the flare renders the model invalid for use in assessing interstate transport of SO2 into the neighboring state of Florida.  
		To augment the information provided in the Big Escambia Supplement, EPA is providing an assessment of the most recent emissions for the years 2016 and 2017 for Big Escambia compared to the modeled emissions for the years 2013, 2014, and 2015, as shown in Table 1.  The first two rows of Table 1 show the emissions modeled at Big Escambia for round 3 designations, and the third row provides annual emissions from the source from EPA's Emissions Inventory System.  The 2016 SO2 emissions are lower than the modeled emissions and the 2017 emissions are slightly higher than the 2015 modeled emissions but lower than the 2013 and 2014 modeled emissions.  The third row in Table 1 shows the emission trends for Big Escambia from 2012-2017.  Because the recent actual emissions are less overall than the modeled emissions inputs for 2013, 2014, and 2015, EPA believes that it is acceptable to use these inputs in the model for assessing interstate transport of SO2 into the neighboring state of Florida.  
                  Table 1:  Big Escambia SO2 Emissions (tons)
                                Alabama Source
                                     2012
                                     2013
                                     2014
                                     2015
                                     2016
                                     2017
                                     2018
Big Escambia (modeled - Source ID S1201)*
                                       
                                     4,079
                                     3,885
                                     3,574
                                       
                                       
                                       
Big Escambia (all other modeled units - potential to emit emissions)*
                                       
                                     17.4
                                       
                                       
                                       
Big Escambia (actual emissions)
                                    14,644
                                     4,908
                                     5,478
                                     4,776
                                     3,573
                                     3,793
                                     N/A**
*  Modeled emissions for Big Escambia are available at:  https://www.epa.gov/sites/production/files/2017-08/documents/3_al_so2_rd3-final.pdf.
** 2018 actual emissions are not available yet for Big Escambia. 
		To address the inappropriate use of background concentrations from the Centreville SEARCH and Mammoth Cave ambient monitoring sites, Alabama reran the round 3 designations modeling for Big Escambia using the most recent three years of valid data from the Mobile County, Alabama, SO2 monitor (AQS ID: 01-097-0003) located north of Mobile, Alabama.  By rerunning the model with recent data from the Mobile County monitor, EPA believes that ADEM has resolved the issue regarding background concentrations for the purposes of assessing interstate transport of SO2 into the neighboring state of Florida.  The model predicted a 99[th] percentile daily maximum 1-hour SO2 concentration of 58.8 ppb in Florida.  Alabama provided modeling files on September 20, 2019, and the associated hourly varying emissions on September 25, 2019.
		EPA proposes to find that for the purpose of assessing interstate transport of SO2 into the neighboring state of Florida, the Big Escambia Supplement has adequately addressed the modeling issues identified in the background section above.  As shown in Table 2 of Section III.C.1.b of the NPRM, Big Escambia's DRR modeling extended into a portion of Florida and included a source in Florida, Breitburn Operating, L.P (Breitburn).  The modeling results indicate that the maximum impacts of 58.8 ppb in Florida did not exceed the level of the 2010 1-hour SO2 NAAQS.  EPA believes that this modeling result, when considered in conjunction with EPA's additional analysis described in footnote 10 above, supports EPA's proposed conclusion that sources in Alabama, including Big Escambia, will not significantly contribute to nonattainment or interfere with maintenance of the 2010 1-hour SO2 NAAQS in any other state.
