
Responses to EPA regarding issues associated with the transport provisions in Alabama's SO2 ISIP:

 Lack of adequate documentation to support the AERMET modeling used to generate the surface and upper air meteorology files; and, 
   
   The following language was taken from ADEM's response to comments document submitted to EPA in October 2017.  In addition to this documentation, ADEM is attaching the surface and upper air files for the Evergreen Middleton Field Airport.  As more reliable meteorological databases are developed, ADEM has begun evaluating smaller airports as representative meteorological datasets, based on climatological analyses.  The Evergreen dataset (GZH) for the 2013-2015 period was 99.3% complete and the airport is located approximately 50 km northeast of the Escambia Operating Company- Big Escambia Creek facility.  This data is considered representative of the BEC facility.
   
   This data has historically been used to characterize modeling for this facility for the past few years.  There have not been any geographical changes in the area that would deem this NWS site unrepresentative.  There are no other new datasets nearby that would better represent this location.  NWS surface and upper air sites are limited in this area. Furthermore the data map below has been used to determine met data for PSD for decades. This data is typically determined on an application by application basis. Below is ADEMs section of the guidance document that addresses representativeness. 
   
   Use the following Meteorological PSD Data Map to identify the area of the State in which the proposed new source or modified source will be located to determine which National Weather Service (NWS) station data to use in the modeling. The station identification numbers are also indicated:
   
   The map of Alabama modeling domains was broken out into 12 sections.  These sections were determined by average monthly precipitation, average monthly mean temperature and topography.  In each county, a COOP weather station was chosen and a 30 year (some stations less than 30) monthly average rainfall and monthly mean temperature was compared to the 12 surrounding NWS stations monthly data.  The NWS station that correlated the closest to the COOP station was linked to that county.  Once all the counties were looked at, they were grouped together by NWS station.  The regions were adjusted to account for the various topographical differences across the state of Alabama.
   
   
   
   
   
   
   
   
   
   
   
   
   
   
 No documentation regarding how the surface characteristics of the area of analysis were developed; and,
    
   The following language was taken from ADEM's response to comments document submitted to EPA in October 2017.  In addition to this documentation, ADEM is attaching the AERSURFACE files generated by ADEM for the Evergreen Middleton Field Airport.  The DRR Modeling Guidance provided flexibility in the modeling of sources, and ADEM directed Alabama subject DRR sources to model using the surface characteristics at the representative meteorological station.  As to the analysis, the comment below from the 2017 Response to Comments Document contained documentation of the Evergreen station.
   
   Modeling of BEC was performed using the surface characteristics around the Evergreen NWS (National Weather Service) in conjunction with Evergreen NWS meteorological data that was used in AERMOD.  No surface characteristics analysis was performed around the BEC facility.  See attached documents for surface characteristics analysis of Evergreen NWS.

 Possible contribution to SO2 modeled concentrations from a nearby source not included in the modeling; and, 
   
   This concern has been addressed in prior communication with EPA Region IV.

 Lack of adequate documentation to demonstrate that appropriate SO2 emissions rates were used in the modeling for the Big Escambia Creek Plant; and, 
   
   
   EPA questioned the lack of documentation that the appropriate SO2 emissions rates were used in the modeling for Big Escambia Creek.  In subsequent calls with EPA to clarify this question, ADEM was informed that the issue in question was the difference in the 2014 NEI emissions and what emissions were modeled for the facility.  This difference in emissions is the result of startup, shut downs and malfunctions of the unit.  As per the DDR guidance SSM emissions are exempted from modeling.  The 2014 NEI shows a total of 4,776 TPY of SO2 versus the 3,902.4 TPY of SO2 that was modeled.  This difference of 873.6 TPY of SO2 is explained by an unusual amount of SSM events for the unit.

 The inappropriate use of background concentrations from the Centreville SEARCH and Mammoth Cave ambient monitoring sites 

   The issue of the background monitor(s) used in the DRR modeling for Alabama sources was discussed at length during the development of the DRR modeling protocols and subsequent modeling analyses.  Specifically, EPA did not accept the use of the two sites above, even after ADEM performed an analysis of the available valid background monitors in the southeastern U.S. that would best represent background in Alabama.  ADEM believed at the time that the use of these ambient monitoring sites best represented the purpose of background concentrations, due to poor data completeness and/or the undue influence of urban industrial sources not representative of the area at the other monitoring sites.  However, after recent discussions with EPA concerning the transport provisions of ADEM's SO2 ISIP, ADEM reran the modeling using the most recent 3 years of valid data from the Chickasaw SO2 monitor, located north of Mobile, Alabama.  The High 4[th] High 1 hour SO2 concentration reported in the 2017 DRR modeling was 169.3 ug/m[3].  Using the Chickasaw monitoring dataset from 2016-2018, the modeling was rerun, and the resulting High 4[th] High 1 hour SO2 predicted concentration increased to 177.3 ug/m[3].  Additionally, EPA requested that the county wide SO2 emissions from both Mobile and Escambia Counties be provided to evaluate the appropriateness of the background monitor.  This information is provided below.  As can be seen, the emissions in Mobile County in both 2014 and 2017 are well above those in Escambia County.  It can be implied, therefore, that this is a conservative background monitor and that the increase in the predicted High 4[th] High 1 hour concentration is reflective of the higher monitoring values.  It does not, however, imply that the use of this monitor is appropriate for the BEC facility due to the industrial profile around Chickasaw when compared to the BEC facility.  It is, however, a conservative analysis and does not change the conclusion previously concluded in the DRR modeling; that is that impacts from the BEC facility, in addition to nearby sources, were expected to comply with the SO2 1 hour NAAQS. 
   
                   SO2 Actual Emissions (TPY) for Mobile and Escambia Counties in Alabama
                                         


                                    County
                             2014 Actual Emissions
                                     (tpy)
                             2017 Actual Emissions
                                     (tpy)
                             2014 Difference (tpy)
                             2017 Difference (tpy)
                                  Escambia Co
                                    7414.86
                                    4055.07
                                    8682.21
                                    3552.21
                                   Mobile Co
                                   16097.07
                                    7607.28
                                       
                                       
    


