                                       
                                       
                                       
                        Technical Support Document for
                                       
                       Kentucky Regional Haze Submittal
                                       
                                 November 2011
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                            U.S. EPA  -  Region 4 
                                  Atlanta, GA
                     Technical Support Document (TSD) for 
                     Kentucky Regional Haze SIP Submittal
                                       
                               Table of Contents

     I. Background
     II. Summary of EPA's Action
     III. Evaluation of Regional Haze Submittal
         A. Identification of Areas
         B. Reasonable Progress Goals (RPGs) Development
                 1. Calculations of Baseline Visibility Conditions
                 2. Calculations of Natural Visibility Conditions
                 3. Uniform Rate of Progress (UROP or "Glidepath") for Reasonable Progress
         C. Long-Term Strategy (LTS)
                 1. Technical Basis for Strategy
                 2. Identification of Sources and Factors to be Considered for Reasonable Progress
                 3. Source or Source Category Specific Analyses for Reasonable Progress
                 4. Enforceability of Reasonable Progress Measures
         D. Best Available Retrofit Technology (BART) Requirements
                 1. Identification of all BART-Eligible Sources
                 2. Modeling or other Demonstrations for BART Exempt Sources
                 3. Identification of Sources Addressed by Clean Air Interstate Rule (CAIR)
                 4. Source Specific BART Determinations
                 5. Enforceability of BART
         E. Coordination of Regional Haze and Reasonably Attributable Visibility Impairment (RAVI) Requirements
         F. Monitoring Strategy and Other Implementation Requirements
         G. Emission Inventory
     IV. Conclusions
__________________________________________________________________
Note:  Italicized text in this document that appears under the "State Regional Haze Submittal" headings are excerpts from relevant portions of the Commonwealth's regional haze submittal as cited.  Figure and table citations under these headings reflect the actual references as listed in the Commonwealth's SIP submittal.  Page numbers refer to the SIP Narrative, unless otherwise specified.

                 TSD for Kentucky Regional Haze SIP Submittal 
   
I. Background

In Section 169A of the 1977 Amendments to the Clean Air Act (CAA), Congress set forth a program for protecting visibility in Federal Class I areas which calls for the "prevention of any future, and the remedying of any existing, impairment of visibility in mandatory Federal Class I areas which impairment results from manmade air pollution." Congress adopted the visibility provisions to protect visibility in 156 Federal Class I areas, which include certain national parks, memorial parks, and wilderness areas over a certain size, and all international parks.  These areas are defined at 40 CFR 81.400, and listed by state at 40 CFR 81.401-81.437.  On December 2, 1980, EPA promulgated regulations to address visibility impairment that is "reasonably attributable" to a single source or small group of sources.  45 FR 80084.  These regulations represented the first phase in addressing visibility impairment and deferred action on regional haze that emanates from a variety of sources until monitoring, modeling and scientific knowledge about the relationships between pollutants and visibility impairment improved. 

In the 1990 Amendments to the CAA, Congress added Section 169B and called on EPA to issue regional haze rules. The Regional Haze Rule (RHR) that EPA promulgated on July 1, 1999 (64 FR 35713) revised the existing visibility regulations to integrate provisions addressing regional haze visibility impairment and establish a comprehensive visibility protection program for Federal Class I areas.  States are required to submit state implementation plans (SIPs) to EPA that set out each state's plan for complying with the RHR, including consultation and coordination with other states and with Federal Land Managers (FLMs).  States must submit a regional haze SIP to EPA within three years after the date of designation of areas under the National Ambient Air Quality Standard (NAAQS) for fine particulate matter (PM2.5).  EPA promulgated PM2.5 designations on December 17, 2004, and thus, regional haze SIPs were to be submitted to EPA by December 17, 2007, which is also specified at 40 CFR 51.308(b).

The RHR addressed the combined visibility effects of various pollution sources over a wide geographic region.  Consequently, all 50 states, including those without Class I areas, Washington, D.C., and the Virgin Islands, are required to submit regional haze SIPs. (40 CFR 51.300(b)(3)).  EPA designated five Regional Planning Organizations (RPOs) to assist with the coordination and cooperation needed to address the visibility issue. The RPO that makes up the southeastern portion of the contiguous United States is known as VISTAS (Visibility Improvement  -  State and Tribal Association of the Southeast).  VISTAS members include:  the Eastern Band of the Cherokee Indians, other Federally Recognized Tribes and Local Air Programs within the geographic area, and the following states: Alabama, Florida, Georgia, Kentucky, Mississippi, North Carolina, South Carolina, Tennessee, Virginia, and West Virginia. The Commonwealth of Kentucky through the KYDAQ chose to participate in VISTAS to develop the technical analyses needed to fulfill the requirements of the RHR.

                                        
       Figure 1. Geographical Areas of Regional Planning Organizations 
                                       
                                       

            
                 Figure 2. Class I Areas in the VISTAS Region 

II. Summary of Federal Register Action

EPA is proposing a limited approval and a limited disapproval of two revisions to the Kentucky SIP submitted by the Commonwealth of Kentucky through the Kentucky Energy and Environment Cabinet, Division of Air Quality (KYDAQ), on June 25, 2008, and May 28, 2010, that address regional haze for the first implementation period.  These revisions address the requirements of the CAA and EPA's rules that require states to prevent any future and remedy any existing anthropogenic impairment of visibility in mandatory Class I areas caused by emissions of air pollutants from numerous sources located over a wide geographic area (also referred to as the "regional haze program").  States are required to assure reasonable progress toward the national goal of achieving natural visibility conditions in Class I areas.  EPA is proposing a limited approval of these SIP revisions to implement the regional haze requirements for Kentucky on the basis that the revisions, as a whole, strengthen the Kentucky SIP.  Also in this action, EPA is proposing a limited disapproval of these same SIP revisions because of the deficiencies in the Commonwealth's regional haze SIP submittal arising from the remand by the U.S. Court of Appeals for the District of Columbia Circuit (D.C. Circuit) to EPA of the Clean Air Interstate Rule (CAIR).  
                                       
Notes:  
   1. This TSD provides further explanation of certain technical elements addressed in the Federal Register Action.
      
   2. Throughout this document, references to Kentucky's (or KYDAQ's or the Commonwealth's) "regional haze SIP" refer to Kentucky's original June 25, 2008, regional haze SIP submittal, as later amended in a SIP revision submitted May 28, 2010.



III. Evaluation of State Regional Haze Submittal
      A. Identification of Areas 
      
The SIP must identify each mandatory Class I Federal area located within the State and in each mandatory Class I Federal area located outside the State that may be affected by emissions from within the State.  Where a mandatory Class I Federal area is in multiple states, a description of how the states have assigned responsibilities for the area needs to be included.

State Regional Haze Submittal
KY SIP Narrative, Page 4:

Kentucky has one Class I area within its borders: Mammoth Cave National Park. The Kentucky Energy and Environment Cabinet's Division for Air Quality (KYDAQ) is responsible for developing the Kentucky's Regional Haze SIP. This SIP establishes reasonable progress goals for visibility improvement at its Class I area, and a long-term strategy that will achieve those reasonable progress goals within the first regional haze planning period.



EPA Findings: Kentucky has adequately addressed this provision.  
III. Evaluation of Regional Haze Submittal 
      B. Reasonable Progress Goals Development 
B.1.Calculations of Baseline Visibility Conditions: ((40 CFR 51.308(d)(2)(i))

The period for establishing baseline visibility conditions is 2000 to 2004.  For each mandatory Class I Federal area located within the State, the State must determine the average degree of visibility impairment (expressed in deciviews) for the most impaired and least impaired days (i.e. the 20% most impaired days and 20% least impaired days) for each of these years.  Baseline conditions are the average of these annual values.  For each mandatory Class I Federal area located within the State, the SIP needs to calculate the number of deciviews by which baseline conditions exceed natural visibility conditions for the most impaired and least impaired days. 

Where onsite monitoring is not available, the baseline visibility conditions must be established using the most representative monitoring available for the 2000-2004 period in consultation with EPA.

State Regional Haze Submittal 
KY SIP Narrative, Pages 9-10:

Baseline visibility conditions at Kentucky's Class I area is estimated using sampling data collected at the IMPROVE monitoring site at Mammoth Cave National Park. A five year average (2000 to 2004) was calculated for each of the 20 percent worst and 20 percent best visibility days in accordance with 40 CFR 51.308(d)(2) and Guidance for Tracking Progress Under the Regional Haze Rule, EPA-454-03-004, September 2003. IMPROVE data records for Mammoth Cave for the period 2000 to 2004 meet USEPA requirements for data completeness (75 percent for the year and 50 percent for each quarter). The light extinction and deciview visibility values for the 20 percent worst and 20 percent best visibility days at Mammoth Cave are based on data and calculations included in Appendix B of this SIP.

KY SIP Narrative, Page 10 - Excerpt of Table 2.3-1:

Natural Background Conditions for Kentucky's Class I Area 

Baseline Visibility Conditions 2000-2004
Class 1 area
Average for 20 percent Worst Days (deciviews)
Average for 20 percent Best Days (deciviews)
Bext (Mm-1) Average for 20 percent Worst Days
Bext (Mm-1) Average for 20 percent Best Days
Mammoth Cave
                                     31.4
                                     16.5
                                     241.4
                                     53.0

EPA Findings:  Kentucky has adequately addressed this provision.
III. Evaluation of Regional Haze Submittal 
      B. Reasonable Progress Goals Development 
B.2. Calculations of natural visibility conditions: ((40 CFR 51.308(d)(2)(iii))

For each mandatory Class I Federal area located within the State, natural visibility conditions must be calculated by estimating the degree of visibility impairment existing under natural conditions for the most impaired and least impaired days, based on available monitoring information and appropriate data analysis techniques as specified in EPA-454/B-03-005, Guidance for Estimating Natural Visibility Conditions Under the Regional Haze Program or an approved alternative technique.

State Regional Haze Submittal 
KY SIP Narrative, Pages 8-10:

Natural background visibility, as defined in Guidance for Estimating Natural Visibility Conditions Under the Regional Haze Program, EPA-454/B-03-005, September 2003, is based on annual average concentrations of fine particle components. The same annual average natural background visibility is assumed for all Class I areas in the eastern United States (separate values are estimated for the western United States). Natural background visibility for the 20 percent worst days is estimated by assuming that fine particle concentrations for natural background are normally distributed and the 90[th] percentile of the annual distribution represents natural background visibility on the 20 percent worst days. In the 2003 guidance, USEPA also provided that states may use a "refined approach" to estimate the values that characterize the natural visibility conditions of the Class I areas. The purpose of such a refinement would be to provide more accurate estimates with changes to the extinction algorithm that may include the concentration values, factors to calculate extinction from a measured particular species and particle size, the extinction coefficients for certain compounds, geographical variation (by altitude) of a fixed value, and the addition of visibility pollutants. 
In 2005 the IMPROVE Steering Committee made recommendations for a refined equation that modifies the terms of the original equation to account for the most recent data. The choice between use of the old or the new equation for calculating the visibility metrics for each Class I area is made by the state in which the Class I area is located. 
The new IMPROVE equation accounts for the effect of particle size distribution on light extinction efficiency of sulfate, nitrate, and organic carbon. The mass multiplier for organic carbon (particulate organic matter) is increased from 1.4 to 1.8. New terms are added to the equation to account for light extinction by sea salt and light absorption by gaseous nitrogen dioxide. Site-specific values are used for Rayleigh scattering to account for the site-specific effects of elevation and temperature. Separate relative humidity enhancement factors are used for small and large size distributions of ammonium sulfate and ammonium nitrate and for sea salt. The elemental carbon (light-absorbing carbon), fine soil, and coarse mass terms do not change between the original and new IMPROVE equation. 
Natural background conditions using the new IMPROVE equation are calculated separately for each Class I area. The calculation starts with the annual average values for natural background for each component of PM2.5 mass from the EPA 2003 guidance (default values). The annual frequency distribution of values of each PM2.5 component for current conditions (2000-2004) is then defined. This species-specific frequency distribution is applied to the default annual average values for that PM2.5 component to calculate natural conditions on the 20% worst days. The current variability in each component is retained while also retaining the same annual average background condition for that component as defined in the 2003 guidance. The new calculation of natural background allows Rayleigh scattering to vary with elevation. Current sea salt values are used for natural background levels of sea salt. 
The VISTAS states chose to use the new IMPROVE equation as the basis for the conceptual description because it takes into account the most recent review of the science and because it is recommended by the IMPROVE Steering Committee. For more detailed discussion of the two IMPROVE equations, see Appendix B.

KY SIP Narrative, Page 10 - Excerpt of Table 2.3-1:

Natural Background Conditions for Kentucky Class I Area

Class I area 
Average for 20 percent Worst Days
(deciviews) 
Average for 20 percent Best Days 
(deciviews) 
Average for 20 percent Worst Days 
Bext (Mm-1) 
Average for 20 percent Best Days 
Bext (Mm-1) 
Mammoth Cave 
                                     11.1 
                                     5.0 
                                     30.7 
                                     16.5 


EPA Findings:  EPA Region 4 has approved the use of the VISTAS methodology with the application of the new IMPROVE equation.  Kentucky has adequately addressed this provision and documented its basis in Appendix B.2 of its submittal.


III. Evaluation of Regional Haze Submittal 
      B. Reasonable Progress Goals Development 
B.3. Uniform Rate of Progress (UROP or "Glidepath") for Reasonable Progress:  (40 CFR 51.308(d)(1) and 40 CFR 51.308(d)(1)(i)(B))

For each mandatory Class I Federal area located within the State, the State must establish goals (expressed in deciviews) that provide for reasonable progress towards achieving natural visibility conditions.  For each Class I area within the state, the state must analyze and determine the rate of progress needed to attain natural visibility conditions by the year 2064 and determine the uniform rate of visibility improvement (measured in deciviews) that would need to be maintained during each implementation period in order to attain natural visibility conditions by 2064. The State must show that the reasonable progress goal for each mandatory Class I Federal area represents at least as much visibility improvement as is expected to result from implementation of other requirements of the CAA during the applicable implementation period (40 CFR 51.308(d)(1)(vi)).  The reasonable progress goals must provide for an improvement in visibility for the most impaired days over the period of the implementation plan and ensure no degradation in visibility for the least impaired days over the same period.  Modeling is not used to determine reasonable progress goals and does not determine whether reasonable progress has been met.  Modeling is one part of the reasonable progress analysis.  The modeling results are used to determine if future year visibility at Class I areas are estimated to be on a (glide) path towards reaching natural background.  This is called a uniform rate of progress analysis or "glidepath" analysis.
 
State Regional Haze Submittal
KY SIP Narrative, Pages 45-47, 84:

Using 2000 - 2004 IMPROVE monitoring data, the deciview values for the 20 percent best days in each year are averaged together, producing a single average deciview value for the best days. Similarly, the deciview values for the 20 percent worst days in each year are averaged together, producing a single average deciview value for the worst days. The average values represent the current visibility conditions. 

The predicted visibility improvement is calculated by comparing the 2002 typical year modeling results for the 12-km grid to the 2018 12-km modeling results to develop a relative reduction factor. This factor is then applied to the current visibility condition values to estimate the future visibility. Detailed discussions about how the relative reduction factors are calculated can be found in Appendix G. For the 20% worst days in Kentucky's Class I area, Figure 7.2.4-1 graphically compares the visibility which would result with the area's uniform rate of progress (red line) to the predicted visibility from 2004 to 2018 due to modeled emission reductions expected by federal and state control programs (purple line).

Similarly, for the 20% best days in the area, Figure 7.2.4-2 graphically compares visibility with no degradation over the first planning period (red line) to the predicted visibility from 2004 to 2018 due to modeled emission reductions expected by federal and state programs (purple line).

Note that for Mammoth Cave, visibility improvements on the 20 percent worst days are expected to be better than the uniform rate of progress glidepath by 2018 based solely on reductions from existing and planned emissions controls. For Mammoth Cave, a 4.73 dv improvement in visibility would meet uniform rate of progress in 2018; expected emissions reductions by 2018 are projected to achieve a 5.81 dv improvement.



                                       
                                       



KY SIP Appendix G, Pages 6-12:
2. Reasonable Progress Goals
For each Class I area, there are three metrics of visibility that are part of the determination of reasonable progress:
1) Natural conditions,
2) Baseline conditions
3) Current conditions.
Each of the three metrics includes the concentration data of the visibility impairing pollutants as different terms in the light extinction algorithm, with respective extinction coefficients and relative humidity factors. Total light extinction is then converted to deciviews, and an average value for both the 20 percent best and 20 percent worst visibility days is calculated.
After the natural, baseline, and current conditions are calculated, a uniform rate of progress, or glidepath, can be created for the 20% worst days for each Class I area. Next, an analysis of potential control measures is performed to determine what level of control would be achievable in the given timeframe and what the potential improvement to visibility would be as a result of these controls. After the potential control analysis, reasonable progress goals for the Class I area are determined after consultation with surrounding states. The entire RPG determination for Kentucky's Class I area is described in Appendix H.
3. Model Application
The purpose of a modeling assessment of uniform rate of progress is to determine if a proposed control strategy will result in uniform rate of progress being met when measured concentrations of particulate matter are used to estimate visibility impairment at some future date. The modeling is applied in a relative sense, similar to the fine particulate matter (PM2.5) attainment test. The following sections outline the process to determine 2018 projections from regional modeling, as suggested in the "Guidance on the Use of Models and Other Analyses for Demonstrating Attainment of Air Quality Goals for Ozone, PM2.5, and Regional Haze"(from this point forward, referred to as "Attainment Guidance").
3.1 Determination of 20% Best and worst Days
The first step in developing model projections is to rank visibility (in deciviews) on each day with observed speciated PM2.5 data for each of the 5 years comprising the base period (2000-2004) for each Class I area. Day-specific observations for mass associated with sulfate (SO4), nitrates (NO3), organic carbon (OC), elemental carbon (EC), soil, and coarse matter (CM), should be used to calculate bext for each day using climatological relative humidity adjustment factor(s) (f(rh)). Total light extinction (bext) for all components should be converted to deciviews for each day to get a daily deciview value.
3.2 Determine Baseline Conditions
The ranked visibility days from the first step are then used to calculate the average baseline deciviews for the 20% best and worst days. First, the mean deciview value for the identified 20% worst- and best visibility days in each year is calculated (dvcurrent). The 5 year mean deciview values reflecting mean visibility on the days with best visibility and then averaged to determine the baseline values for the best days, or the dvbaseline values. Then the 5 year mean deciview values reflecting worst visibility are averaged together. This represents the value subject to improvement to meet the glidepath for regional haze.
For Kentucky's Class I area, the baseline visibility conditions were estimated using sampling data collected at Interagency Monitoring of Protected Visual Environments (IMPROVE) monitoring site at four of the Class I area in Kentucky. A five year average (2000 to 2004) was calculated for each of the 20 percent worst and 20 percent best visibility days in accordance with 40 CFR 51.308(d)(2) and the USEPA's Guidance for Tracking Progress Under the Regional Haze Rule. IMPROVE data records for Mammoth Cave for the period 2000 to 2004 meet the USEPA's requirements for data completeness (75 percent for the year and 50 percent for each quarter). The light extinction and deciview visibility values for the 20 percent worst and 20 percent best visibility days at Mammoth Cave is based on data and calculations included in Appendix B of this SIP. Table 3.2-1 summarizes the mean yearly visibility conditions and the baseline visibility conditions for Kentucky's Class I area on both the 20% best and 20% worst days.


3.3 Relative Response Factor Calculation
The next step is to use base year and future year modeling results to estimate a relative response factor (RRF) for each component of PM2.5 on the 20% best and worst monitoring days. Simply put, the RRF is the average future year concentration near the monitor divided by the average base year concentration or: 

To accomplish this step, concentration for each component of PM2.5 and coarse matter (the difference between particles less than 10 micrometers in diameter (PM10) and PM2.5) are extracted from the base year and future year modeling output near IMPROVE monitoring sites for the monitored best and worst days. The component concentrations are then averaged to develop base and future year average component concentrations. The future year average component concentrations are then divided by their respective base year component concentrations to develop RRFs for each component of PM2.5 and coarse matter.
For a numerical example, consider Table 3.3-1, which has all the pertinent data for the Mammoth Cave IMPROVE monitor. Looking at the top table, the first column identifies the 20% best days identified in step 1 (Section 3.1). The next six columns have the 12km modeled base year (2002) concentration values near the monitor corresponding to the 20% best days. The next six columns have the 12km modeled future year (2018) component concentrations near the monitor for the same set of days. The final six columns contain the daily RRF values for each component species, which is simply the future concentration divided by the base year concentration. The bottom row of the table contains the average values for the base year components, the future year components, and the RRFs for each of the average component values. The bottom table contains the same data for the 20% worst days. Tables for the IMPROVE monitor of interest for Kentucky's Class I area are provided, and abide by the same format.


3.4. Future Year Daily Concentration Calculation
The fourth step in using model output to estimate the future visibility impairment is to multiply the relative response factors (determined in the previous step, Section 3.3) by the measured species concentration data during the base period (for the measured 20% best and worst days). This results in daily future year species concentrations data.
This is similar to the final step in the ozone attainment test, where base year concentrations, or the current year design value (DVC) is multiplied by the RRF to determine the future design value (i.e. DVF = RRF * DVC). For regional haze, this calculation is performed for each best and worst day for the five-year base line period (calculated in step 2, Section 3.2) using an RRF for each PM component, for both the best days and the worst days. This produces future concentration estimates for SO4, NO3, OC, EC, Soil and CM, for each of the "worst" and "best" visibility days.
3.5. Future Year Daily bext Calculation
Once future year concentrations are calculated (Section 3.4) the future daily extinction coefficients are then calculated for the 20% best and 20% worst visibility days in each of the five base years. This is achieved by applying the new IMPROVE equation (see Appendix B for a more information on the new IMPROVE equation):
bext ≈ 2.2 * fs(RH) * [Small Sulfate] + 4.8 * fL(RH) * [Large Sulfate]
+ 2.4 * fs(RH) * [Small Nitrate] + 5.1 * fL(RH) * [Large Nitrate]
+ 2.8 * {Small Organic Mass] + 6.1 * [Large Organic Mass]
+ 10 * [Elemental Carbon]
+ 1 * [Fine Soil]
+ 1.7 * fss(RH) * [Sea Salt]
+ 0.6 * [Coarse Mass]
+ Rayleigh Scattering (site specific)
+ 0.33 * [NO2 (ppb)]
Model estimated daily concentrations are used with the corresponding base line daily relative humidity adjustment factor(s) (f(rh)) identified in step 1 (Section 3.1).
3.6. Future Mean Deciview Values Calculation
The final step is to calculate daily deciview values from the total daily extinction calculated in the previous step, then compute the future average mean deciviews for the worst and best days for each year. The 5 years are then averaged together to get the final future mean deciview value for the worst and best days.
The first step in this process is to convert the total daily bext for each day (calculated in the previous step, Section 3.5) to deciviews. This future year daily deciview value for each of the best and worst days is then averaged to determine the future deciview value for the "worst" and "best" visibility days for each year. This leads to 5 future estimated mean deciview values for the  "worst" and future estimated mean deciview values for the "best" visibility days. The five yearly values for the best and worst days are then averaged to produce the final future mean deciview value, or dvfuture. Table 3.2-1 summarizes the mean yearly future visibility conditions and the baseline visibility conditions for Kentucky's Class I area on both the 20% best and 20% worst days.
The resulting dvfuture values for the "worst" and "best" visibility days can be compared to the dvbaseline values calculated in step 2 (Section 3.2). If the resulting change in deciviews is a negative number (future - base), this represents an improvement in visibility. Table 3.6-2 summarizes these calculation for the IMPROVE monitor associated with Kentucky's Class I area. From the table, it is apparent that Kentucky's Class I area is showing visibility improvement on both the 20% best and 20% worst days.



EPA Findings: Kentucky has adequately addressed this provision.
III. Evaluation of Regional Haze Submittal 
      C. Long-Term Strategy 
C.1. Technical Basis for Strategy: (40 CFR 51.308(d)(3)(iii)) 

The SIP must document the technical basis, including modeling, monitoring and emissions information, on which the State is relying to analyze regional haze.  Modeling is not used to determine reasonable progress goals and does not determine whether reasonable progress has been met.  Modeling is one part of the reasonable progress analysis.  The modeling results are used to determine if future year visibility at Class I areas are estimated to be on a (glide) path towards reaching natural background.  This is called a uniform rate of progress analysis or "glidepath" analysis.  The State may meet this requirement by relying on technical analyses developed by the regional planning organization and approved by all State participants.

State Regional Haze Submittal
The following is taken and/or summarized from Kentucky's June 25, 2008, regional haze SIP submittal, Appendix H  -  RPG Evaluation and LTS (page H-4):

EPA Note:  The RHR requires states to establish RPGs for visibility improvement at each affected Class I area covering each approximately 10-year period until 2064.  The Commonwealth's first set of RPGs will comprise a long term strategy (LTS) covering the period from the present until 2018.  Appendix H provides an overview of this first LTS.  

Kentucky's LTS contains the following components:

   I.       Inventory of all controls required or expected under all federal and state regulations by 2009 and 2018.
   II.       Modeled expected visibility improvement and Comparison of modeling results for 2018 to the uniform rate of progress, or "glidepath" for Kentucky's Class I area.  (See also EPA TSD Section III.B.3)
   III.       Discussion of emissions sensitivity model runs to help determine which pollutants and source categories are contributing the most to visibility impairment at Kentucky's Class I area (runs conducted by the Georgia Institute of Technology).
   IV.       Examination of SO2 Area of Influence (AOIs) projections for each of the State's Class I areas, and also for any surrounding states' Class I areas whose AOI includes any Kentucky counties, to determine which specific sources are most likely to be impairing visibility at specific Class I areas.
   V.       Application of the four "statutory factors," listed in 40 CFR 51.308(d)(1)(i)(A) to those sources identified in the previous step.  (See also EPA TSD Section III.C.3)
   VI.       Using the results of the four factor analyses, identification of control determinations for specific sources.
   VII.       Identification of BART-eligible sources.  (See also EPA TSD Section III.D.1)
   VIII.       BART exemption modeling for BART-eligible sources.  (See also EPA TSD Section III.D.2)
   IX.       BART determinations for sources subject to BART.  (See also EPA TSD Section III.D.3)
   X.       Modeling demonstration of visibility improvement after inclusion of BART and reasonable progress control determinations. 
   XI.       Determination of RPGs based on the visibility improvement expected by 2018.

From KY SIP Narrative starting at page 22:

Modeling Demonstration
Modeling for regional haze was performed by VISTAS for the ten southeastern states, including Kentucky. The sections below outline the methods and inputs used by VISTAS for the regional modeling. Additional details are provided in Appendices C, D and M. 
5.1 Analysis Method 
The modeling analysis is a complex technical evaluation that begins by selection of the modeling system. VISTAS decided to use the following modeling system: 
   ::  Meteorological Model: The Pennsylvania State University/National Center for Atmospheric Research (PSU/NCAR) Mesoscale Meteorological Model (MM5) is a nonhydrostatic, prognostic meteorological model routinely used for urban- and regional-scale photochemical, fine particulate matter, and regional haze regulatory modeling studies. 
   
   ::   Emissions Model: The Sparse Matrix Operator Kernel Emissions (SMOKE) modeling system is an emissions modeling system that generates hourly gridded speciated emission inputs of mobile, nonroad mobile, area, point, fire and biogenic emission sources for photochemical grid models. 
   
   ::   Air Quality Model: The USEPA's Models-3/ Community Multiscale Air Quality (CMAQ) modeling system is an `One-Atmosphere' photochemical grid model capable of addressing ozone, particulate matter (PM), visibility and acid deposition at regional scale. 
   .....
5.4 Modeling Domains 
....
Figure 5.4.1-1 shows the MM5 horizontal domain as the outer most, blue grid with the CMAQ 36-km domain nested in the MM5 domain.  Figure 5.4.1-2 shows the 12-km grid, modeling domain for the VISTAS region.  

Figure 5.4.1-1. The MM5 horizontal domain is the outer most, blue grid, with the CMAQ 36-km domain nested in the MM5 domain. 


Figure 5.4.1-2. A more detailed view of the 12-km grid over the VISTAS region.

Emissions Controls Modeled (SIP Narrative, Pages 40-43):

There are significant control programs being implemented between the baseline period and 2018. These programs are described in more detail below.
Federal and State Control Requirements
NOx SIP Call or state equivalent. Phase I of the NOx SIP call applies to certain EGUs and large non-EGUs, including large industrial boilers and turbines, and cement kilns. Those states affected by the NOx SIP call in the VISTAS region have developed rules for the control of NOx emissions that have been approved by the USEPA. The NOx SIP Call has resulted in a 66 percent reduction in summertime NOx emissions from large stationary combustion sources in Kentucky. 
CAIR. CAIR will permanently cap emissions of SO2 and NOx from EGUs in the eastern United States by 2015. When fully implemented, CAIR will reduce SO2 emissions from EGUs in these states by more than 70 percent, and NOx emissions by more than 60 percent, from 2003 levels. 
One-hour ozone SIPs (Atlanta / Birmingham / Northern Kentucky). New SIPs have been submitted to the USEPA to demonstrate attainment of the one-hour ozone NAAQS. These SIPs require NOx reductions from specific coal fired power plants and address transportation plans in these cities.
Heavy Duty Diesel (2007) Engine Standard (for on-road trucks and buses). The USEPA set a PM emissions standard for new heavy-duty engines of 0.01 grams per brake-horsepower-hour (g/bhp-hr), to take full effect for diesel engines in the 2007 model year. This rule also includes standards for NOx and non-methane hydrocarbons (NMHC) of 0.20 g/bhp-hr and 0.14 g/ bhp-hr, respectively. These NOx and NMHC standards will be phased in together between 2007 and 2010, for diesel engines. Sulfur in diesel fuel must be lowered to enable modern pollution-control technology to be effective on these trucks and buses. The USEPA requires a 97 percent reduction in the sulfur content of highway diesel fuel from its previous level of 500 parts per million (low sulfur diesel, or LSD) to 15 parts per million (ultra-low sulfur diesel, or ULSD). 
Tier 2 Tailpipe (On-road vehicles). The USEPA mobile source rules include the Tier 2 fleet averaging program, modeled after the California LEV II standards. Manufacturers can produce vehicles with emissions ranging from relatively dirty to zero emissions, but the mix of vehicles a manufacturer sells each year must have average NOx emissions below a specified value. Tier 2 standards became effective in the 2005 model year. 
Large Spark Ignition and Recreational Vehicle Rule. The USEPA has adopted new standards for emissions of NOx, hydrocarbons, and carbon monoxide from several groups of previously unregulated nonroad engines. Included in these are large industrial spark-ignition engines and recreational vehicles. Nonroad spark-ignition engines are those powered by gasoline, liquid propane gas, or compressed natural gas rated over 19 kilowatts (kW) (25 horsepower). These engines are used in commercial and industrial applications, including forklifts, electric generators, airport baggage transport vehicles, and a variety of farm and construction applications. Nonroad recreational vehicles include snowmobiles, off-highway motorcycles, and all-terrain-vehicles. These rules were initially effective in 2004 and will be fully phased-in by 2012. 
Nonroad Diesel Rule. This rule sets standards that will reduce emissions by more than 90 percent from nonroad diesel equipment, and reduce sulfur levels by 99 percent from current levels in nonroad diesel fuel starting in 2007. This step will apply to most nonroad diesel fuel in 2010 and to fuel used in locomotives and marine vessels in 2012. 
Industrial Boiler/Process Heater/RICE MACTs. The USEPA issued final rules to substantially reduce emissions of toxic air pollutants from industrial, commercial and institutional boilers, process heaters and from stationary reciprocating internal combustion engines (RICE). These rules reduce emissions of a number of toxic air pollutants, including hydrogen chloride, manganese, lead, arsenic and mercury by 2009. This rule also reduces emissions of SO2 and PM in conjunction with the toxic air pollutant reductions. The applied Maximum Achievable Control Technology (MACT) control efficiencies were 4 percent for SO2 and 40 percent for PM10 and PM2.5. The USEPA's industrial boiler MACT rules were vacated on June 8, 2007. The VISTAS states decided to leave these controls in the modeling since it is believed that by 2018 the USEPA will have re-promulgated a boiler MACT rule or states will have addressed the issue through state rule making.
Combustion Turbine MACT. The projection inventories do not include the NOx co-benefit effects of the MACT regulations for Gas Turbines or stationary Reciprocating Internal Combustion Engines, which the USEPA estimates to be small compared to the overall inventory. 
VOC 2-, 4-, 7-, and 10-year MACT Standards. Various point source MACTs and associated emission reductions were implemented. Reductions occurring before 2002 were assumed to be accounted for in the 2002 base year inventory. 
Consent Agreements. Under a settlement agreement, by 2008, Tampa Electric will install permanent emissions-control equipment to meet stringent pollution limits; implement a series of interim pollution-reduction measures to reduce emissions while the permanent controls are designed and installed; and retire pollution emission allowances that Tampa Electric or others could use, or sell to others, to emit additional NOx, SO2, and PM. 
Virginia Electric and Power Co. (VEPCO) agreed to spend $1.2 billion by 2013 to eliminate 237,000 tons of SO2 and NOx emissions each year from eight coal-fired electricity generating plants in Virginia and West Virginia. 
A 2002 agreement calls for Gulf Power to upgrade its operation to cut NOx emission rates by 61 percent at its Crist generating plant by 2007, with major reductions beginning in early 2005. The Crist plant is a significant source of NOx emissions in the Pensacola area. 
East Kentucky Power Cooperative (EKPC), a coal-fired electric utility, has agreed to spend approximately $650 million on pollution controls and pay a $750,000 penalty to resolve violations of the Clean Air Act at its three plants in Kentucky known as Spurlock, Dale, and Cooper. The agreement will reduce harmful air pollutants by more than 60,000 tons per year according to EPA. The company will install state-of-the-art pollution control equipment to reduce emissions of pollutants that cause acid rain and smog. The controls will reduce annual emissions of smog-forming nitrogen oxides by approximately 8,000 tons and sulfur dioxide by more than 54,000 tons per year from EKPC's Spurlock, Dale, and Cooper plants when the controls are fully implemented. This consent decree will facilitate that SO2 scrubbers are installed for EKPC's Spurlock Units 1 and 2 and Cooper Units 1 and 2 for BART. Per IPM, SO2 scrubbers for EKPC Spurlock Units 1 and 2 and Cooper Unit 2 have been included in this SIP's regional haze modeling results for 2018. 
American Electric Power (AEP) has agreed to cut 813,000 tons of air pollutants annually (654,000 tons of SO2 and 159,000 tons of NOx) at an estimated cost of more than $4.6 billion, pay a $15 million penalty, and spend $60 million on projects to mitigate the adverse effects of its past excess emissions. The agreement imposes caps on emissions of pollutants from 16 plants located in five states. The facilities are located in Moundsville (2 facilities), St. Albans, Glasgow, and New Haven (2 facilities), West Virginia; Louisa, Kentucky; Glen Lyn and Carbo, Virginia; Brilliant, Conesville, Cheshire, Lockburne, and Beverly, Ohio; and Rockport and Lawrenceburg, Indiana. AEP will install pollution control equipment to reduce and cap sulfur dioxide and nitrogen oxide emissions by more than 813,000 tons per year when fully implemented. By installing these pollution control measures, the plants will emit 79 percent less sulfur dioxide and 69 percent less nitrogen oxides, as compared to 2006 emissions. This consent decree will facilitate that a SO2 scrubber is installed for Kentucky's AEP Big Sandy Unit 2 for BART. Per IPM, a SO2 scrubber for AEP Big Sandy Unit 2 has been included in this SIP's regional haze modeling results.

Sources and Pollutants relied on for LTS (SIP Narrative, Pages 54-56):

7.4 Relative Contributions to Visibility Impairment: Pollutants, Source Categories, and Geographic Areas  
An important step toward identifying potential reasonable control measures is to identify the key pollutants contributing to visibility impairment at each Class I area. To understand the relative benefit of further reducing emissions from different pollutants, source sectors, and geographic areas, VISTAS engaged the Georgia Institute of Technology to perform emission sensitivity model runs using CMAQ. Emissions sensitivities were initially performed for three episodes representing winter and summer conditions: Jan 2002, July 2001, and July 2002. These runs used the initial 2018 projections inventory and considered 30 percent reductions from specific pollutants, source categories, and geographic areas. As part of a separate effort, emissions sensitivities were repeated using a preliminary 2009 projection inventory and two, month-long episodes from 2002: Jun 1  -  Jul 10 and Nov 19  -  Dec 19. The emissions in 2009 were reduced by 30 percent for each pollutant sensitivity run. The pollutant contributions that were evaluated were: 
:: SO2 from EGU sources in each VISTAS state, other RPOs in the VISTAS 12 km grid, and Boundary Conditions from outside the 12 km domain. 
:: SO2 from non-EGU point sources in each VISTAS state, other RPOs, and Boundary Conditions 
:: NOx from ground level sources (on-road plus off-road plus area) in each VISTAS state and other RPOs. In the VISTAS states, these reductions were only applied to specific counties that were of concern for 8-hour ozone nonattainment. 
:: NOx from point (EGU plus non-EGU) sources in each VISTAS state and other RPOs 
:: NH3 from all sources in VISTAS and other RPOs 
:: Volatile Organic Compounds from anthropogenic and biogenic sources in the 12 km modeling domain 
:: Primary Carbon from all ground level sources in each VISTAS state and other RPOs. In the VISTAS states, these reductions were only applied to specific counties that were of concern for PM2.5 nonattainment. 
:: Primary Carbon from all point sources in each VISTAS state and other RPOs 
:: Primary Carbon from all fires in each VISTAS state and other RPOs 

...

As Figure 7.4-1 illustrates, the greatest visibility benefits on the 20 percent worst days for the Kentucky's Class I area are projected to result from further reducing SO2 from EGUs. At the mountain Class I areas, benefits are projected from SO2 reductions from EGUs in several VISTAS states including Alabama, Georgia, Kentucky, North Carolina, South Carolina, Tennessee, Virginia, and West Virginia. Contributions from other RPOs and SO2 coming from outside the boundary are also significant. The greatest benefit would likely be from further reductions from other VISTAS states, the MRPO, and outside the boundary. Additional, smaller benefits are projected from additional SO2 emission reductions from non-utility, industrial point sources. The pattern of relative SO2 contributions from non-EGUs among the various VISTAS states is similar to the pattern of relative SO2 contributions from EGUs....

KYDAQ concludes that reducing SO2 emissions from EGUs in the Midwest RPO would have the greatest visibility benefit for Mammoth Cave. Contributions from other VISTAS states are also significant for this area.

Model Performance Evaluation (KY SIP Narrative, Pages 31-39): 
 6.0 MODEL PERFORMANCE EVALUATION 
The initial modeling effort focused on evaluating previous regional air quality modeling applications and testing candidate model configurations for the SMOKE emissions and CMAQ model for the VISTAS 36-km and 12-km modeling domains. This effort resulted in a report recommending the model configuration for the annual emissions and air quality modeling, which is included as part of the VISTAS Emissions and Air Quality Modeling Protocol. The evaluation of the meteorological modeling configuration can be found in Appendix F.1, with a summary of the final meteorological and air quality modeling configuration in the modeling protocol contained in Appendix E and Appendix C, respectively. 
Air quality model performance for the 2002 modeling year was initially tested in 2004 using an early version of the VISTAS emissions inventory. In keeping with the one-atmosphere objective of the CMAQ modeling platform, model performance was evaluated based on measured ozone, fine particles, and acid deposition in the Air Quality System (AQS), IMPROVE, Speciated Trends Network (STN), Southeastern Aerosol Research and Characterization (SEARCH), National Acid Deposition Program (NADP) and Clean Air Status and Trends Network (CASTNet) monitoring networks (Figure 6.0-1). A detailed examination of the results was published in 2005 in the Journal of Air and Waste Management (see Appendix B.3) as well as being summarized in Appendix B.1. 
... 
6.3 Kentucky's Class I Area Performance
...A cursory view of the stacked bars charts reiterates that sulfates are a large contributor to light extinction in the Kentucky Class I area on both 20% best days and 20% worst days. The bar charts also suggest that organic carbon and nitrates are important on the 20% best days at the IMPROVE site at Mammoth Cave for Kentucky. The bar charts for the 20% best days indicate an over prediction of the nitrate and a slight under prediction of the sulfate on many of the 20% best days. 
Comparing the 20% best day charts to the 20% worst days charts, one notices that the various components of particle pollution play a more prominent role in the 20% best days than with the 20% worst days. Also, the species make up on the 20% best days varies more widely compared to the 20% worst days. This suggests accurately modeling each species is especially important on the 20% best days. 
With the bar chart for the 20% worst days, you can see model performance does improve across the sites. Light extinction due to sulfate prediction is better, but still falls short on some days. Light extinction due to organic carbon also becomes more important to total light extinction. Much like the sulfate component, the organic carbon component accuracy has improved performance over the 20% best day series, thought some days are still under predicted. Overall, the KYDAQ found model performance to fall within acceptable limits for model performance. The KYDAQ further asserts the one atmosphere modeling performed by the VISTAS contractors is representative of conditions in the southeastern states and is applicable for use in attainment demonstrations.

EPA Findings:  The technical analyses and modeling to assess uniform rate of progress and to support the LTS were successfully developed consistent with EPA's interim and final modeling guidance.  EPA accepts the VISTAS technical modeling to support the LTS and determine visibility improvement for uniform rate of progress because the modeling system was chosen and simulated according to EPA modeling guidance.  EPA agrees with the VISTAS model performance procedures and results, and that the CMAQ is an appropriate tool for the regional haze assessments for the Kentucky LTS and regional haze SIP.  

      However, CAIR is an element of Kentucky's LTS.  Kentucky is subject to CAIR and has rules addressing the CAIR requirements adopted into its SIP (72 FR 46388 (August 20, 2007) and 74 FR 61535 (November 25, 2009)).  The State demonstrated that CAIR satisfies reasonable progress for SO2 for its EGUs for this first implementation period.  On July 11, 2008, the D.C. Circuit issued its decision to vacate and remand both CAIR and the associated CAIR FIPs in their entirety.  North Carolina v. EPA, 531 F.3d 836 (D.C. Cir. 2008).  However, in response to EPA's petition for rehearing, the Court issued an order remanding CAIR to EPA without vacating either CAIR or the CAIR FIPs.  North Carolina v. EPA, 550 F.3d 1176 (D.C. Cir. 2008).  The Court thereby left CAIR in place in order to "temporarily preserve the environmental values covered by CAIR" until EPA replaces it with a rule consistent with the Court's opinion.  Id. at 1178. The Court directed EPA to "remedy CAIR's flaws"' consistent with its July 11, 2008, opinion, but declined to impose a schedule on EPA for completing that action.  Id.  Although the D.C. Circuit found CAIR to be flawed, the rule was remanded without vacatur and thus remains in place.  In response to the Court's decision, EPA has issued a new rule to address interstate transport of NOX and SO2 in the eastern United States (i.e., the Transport Rule, also known as the Cross-State Air Pollution Rule).  See 76 FR 48208 (August 8, 2011).  EPA explained in that action that EPA is promulgating the Transport Rule as a replacement for (not a successor to) CAIR's SO2 and NOx emissions reduction and trading programs.  In other words, the CAIR and CAIR Federal Implementation Plan (FIP) requirements only remain in force to address emissions through the 2011 control periods.  As part of the Transport Rule, EPA finalized regulatory changes to sunset the CAIR and CAIR FIPs for control periods in 2012 and beyond.  See 76 FR 48322.  
      Prior to the CAIR remand by the D.C. Circuit, EPA believed the Commonwealth's reliance on CAIR for the specific BART and reasonable progress provisions affecting its EGUs was adequate.  The regional haze provisions specify that a state may not adopt a RPG that represents less visibility improvement than is expected to result from other CAA requirements during the implementation period (40 CFR 51.308(d)(1)(vi)).  As such, Kentucky took into account emission reductions anticipated from the CAIR in determining its 2018 RPGs.  Due to the need to address the remand of CAIR to EPA by the D.C. Circuit, states may not permanently rely upon the emissions reductions predicted by CAIR.  For this reason, EPA cannot fully approve regional haze SIP revisions that rely on CAIR.
      
      Also, on June 8, 2007, and effective on July 30, 2007, the D.C. Circuit mandated the vacatur and remand of the Industrial Boiler MACT Rule.  Notwithstanding the vacatur of the this rule, the VISTAS states, including Kentucky, decided to leave these controls in the modeling for their regional haze SIPs since it is believed that by 2018, EPA will have re-promulgated an industrial boiler MACT rule or the states will have addressed the issue through state-level case-by-case MACT reviews in accordance with section 112(j) of the CAA.  EPA finds this approach acceptable for the following reasons.  EPA proposed a new Industrial Boiler MACT rule to address the vacatur on June 4, 2010 (75 FR 32006), and issued a final rule on March 21, 2011 (76 FR 15608), giving Kentucky time to assure the required controls are in place prior to the end of the first implementation period in 2018.  In the absence of an established MACT rule for boilers and process heaters, the statutory language in section 112(j) of the CAA specifies a schedule for the incorporation of enforceable MACT-equivalent limits into the title V operating permits of affected sources.  Should circumstances warrant the need to implement section 112(j) of the CAA for industrial boilers, EPA would expect, in this case, that compliance with case-by-case MACT limits for industrial boilers would occur no later than January 2015, which is well before the 2018 RPGs for regional haze.  In addition, the RHR requires that any resulting differences between emissions projections and actual emissions reductions that may occur will be addressed during the five-year review prior to the next 2018 regional haze SIP.  The expected reductions due to the original, vacated Industrial Boiler MACT rule were relatively small compared to the State's total SO2, PM2.5, and coarse particulate matter (PM10) emissions in 2018 (i.e., 0.5 to 1.5 percent, depending on the pollutant, of the projected 2018 SO2, PM2.5, and PM10 inventory), and not likely to affect any of Kentucky's modeling conclusions.  Thus, if there is a need to address discrepancies such that projected emissions reductions from the now-vacated Industrial Boiler MACT were greater than actual reductions achieved by the replacement MACT, EPA would not expect that this would affect the adequacy of the existing Kentucky regional haze SIP.

III. Evaluation of Regional Haze Submittal 
      C. Long-Term Strategy 
C.2. Identification of Sources and Factors to be Considered:  (40 CFR 51.308(d)(3)(iv-v))  

The State must identify all anthropogenic sources of visibility impairment considered by the State in developing its long-term strategy. The State should consider major and minor stationary sources, mobile sources, and area sources. The State must consider, at a minimum, the following factors in developing its long-term strategy: (A) emission reductions due to ongoing air pollution control programs, including measures to address reasonably attributable visibility impairment; (B) measures to mitigate the impacts of construction activities; (C) emissions limitations and schedules for compliance to achieve the reasonable progress goal; (D) source retirement and replacement schedules; (E) smoke management techniques for agricultural and forestry management purposes including plans as currently exist within the State for these purposes; measures to mitigate the impacts of construction activities; (F) enforceability of emission limitations and control measures; and the (G) anticipated net effect on visibility due to projected changes in point, area, and mobile source emissions over the period.  


State Regional Haze Submittal
KY SIP Narrative, Pages 79-83:

The following summarizes the process for determining reasonable progress for Kentucky sources.  For a detailed discussion of the reasonable progress assessments for all units with a contribution of greater than one percent to visibility impairment at the Class I area in Kentucky or in neighboring states, please see Appendix H. 

Step 1: Determine pollutants of concern. 
VISTAS evaluated the species contribution on the 20 percent worst visibility days in the baseline period and concluded that sulfate accounted for greater than 70 percent of the visibility impairing pollution. The VISTAS States concluded that controlling SO2 emissions was the appropriate step in addressing the reasonable progress assessment for 2018. The VISTAS findings were consistent with the findings of SAMI. As you recall, SAMI confirmed that sulfate particles account for the greatest portion of the haze affecting Class I areas in the Southern Appalachian region and that these sulfates were produced in large part from SO2 emissions from coal combustion. 

Step 2: Determine which source sectors should be evaluated for reasonable progress. 
Since the pollutant of primary concern was determined to be SO2, the emissions inventory was assessed to determine the source categories that contribute the most SO2 emissions. Since point source emissions in 2018 are projected to represent greater than 95 percent of the total SO2 emissions inventory, the VISTAS States concluded that the focus should be on electric generating unit (EGU) and non-EGU point sources of SO2 emissions. 

Step 3: Determine if the Clean Air Interstate Rule is sufficient for reasonable progress for subject EGUs. 
The KYDAQ evaluated the amount of SO2 reduction from the EGU sector resulting from the implementation of the CAIR. The EGUs in Kentucky are expected to reduce their 2002 SO2 emissions by an estimated 54 percent by 2018. Much of that reduction is the result of requirements that are predicted by the IPM to meet CAIR. 

To further support EGUs subject to CAIR is sufficient for reasonable progress, a discussion in the CAIR rule highlighted below (See Appendix H for 70 FR 25197-25198) addresses the reasonable progress factors of cost and time necessary for compliance for these EGUs, and provide the necessary support for a State's four factor reasonable progress analysis that must accompany a State's assertion that CAIR is sufficient for reasonable progress for subject EGU's during the first planning period. 

From past experience in examining multi-pollutant emissions trading programs for SO2 and NOX, EPA recognized that the air pollution control retrofits that result from a program to achieve highly cost-effective reductions are quite significant and can not be immediately installed. Such retrofits require a large pool of specialized labor resources, in particular, boilermakers, the availability of which will be a major limiting factor in the amount and timing of reductions. 

Also, EPA recognized that the regulated industry will need to secure large amounts of capital to meet the control requirements while managing an already large debt load, and is facing other large capital requirements to improve the transmission system. Furthermore, allowing pollution control retrofits to be installed over time enables the industry to take advantage of planned outages at power plants (unplanned outages can lead to lost revenue) and to enable project management to learn from early installations how to deal with some of the engineering challenges that will exist, especially for the smaller units that often present space limitations. 

Based on these and other considerations, EPA determined in the NPR that the earliest reasonable deadline for compliance with the final highly cost-effective control levels for reducing emissions was 2015 (taking into consideration the existing bank of title IV SO2 allowances). First, the Agency confirmed that the levels of SO2 and NOX emissions it believed were reasonable to set as annual emissions caps for 2015 lead to highly cost- effective controls for the CAIR region. 

Once EPA determined the 2015 emissions reductions levels, the Agency determined a proposed first (interim) phase control level that would commence January 1, 2010, the earliest the Agency believed initial pollution controls could be fully operational (in today's final action, the first NOX control phase commences in 2009 instead of in 2010, as explained in detail in section IV.C). The first phase would be the initial step on the slope of emissions reductions (the glide-path) leading to the final (second) control phase to commence in 2015. The EPA determined the first phase based on the feasibility of installing the necessary emission control retrofits, as described in section IV.C. 

Although EPA's primary cost-effectiveness determination is for the 2015 emissions reductions levels, the Agency also evaluated the cost effectiveness of the first phase control levels to ensure that they were also highly cost effective. Throughout this preamble section, EPA reports both the 2015 and 2010 (and 2009 for NOX) cost-effectiveness results, although the first phase levels were determined based on feasibility rather than cost effectiveness. The 2015 emissions reductions include the 2010 (and 2009 for NOX) emissions reductions as a subset of the more stringent requirements that EPA is imposing in the second phase. 

The KYDAQ intends to re-evaluate the IPM predictions of SO2 reductions for CAIR at the time of the next periodic report in 2012 to ensure that the reductions currently predicted by IPM for CAIR are in fact taking place where they were expected and needed. If KYDAQ's assessment for the periodic report indicates that its emissions are likely not to meet 2018 projections then the KYDAQ may re-evaluate the four factors to re-assess the Long-Term Strategy. Based on the controls currently being installed for CAIR, required by BART, consent decrees, and predicted by IPM under CAIR, the KYDAQ has concluded that at this time these existing regulatory programs constitute reasonable control measures for Kentucky EGUs during this first assessment period (between baseline and 2018). 

Step 4: Determine which emission units would be evaluated based on impact. 
The KYDAQ calculated the fractional contribution from all emission units within the SO2 Area of Influence for a given Class I area and identified those emission units with a contribution of one percent or more to the visibility impairment at that Class I area. A full description of this process and a list of sources considered in the reasonable progress evaluation can be found in Appendix H. 


. 

Step 5: Evaluate the four factors. 
Each emission unit identified in Step 4 above was considered for evaluation using the statutory and regulatory factors of 1) cost of compliance, 2) time necessary for compliance, 3) the energy and non-air quality environmental impacts of compliance, and 4) the remaining useful life of the emissions unit. If any control measure for an emission unit was found reasonable after assessing the four factors, modeling would be performed to determine if the controls would result in a visibility improvement at any Class I area. 

For the limited purpose of evaluating the cost for the reasonable progress assessment in this first regional haze SIP, the KYDAQ believes it is not equitable to require Non-EGUs to bear a greater economic burden than EGUs for a given control strategy. The KYDAQ used EPA's CAIR EGU cost analysis to establish a cost/ton of SO2 removed threshold. During the current reasonable progress assessment, no units in Kentucky were identified for additional control since no measures were found to be below the cost threshold. Below is a summary of the analysis. The detailed analysis is included in Appendix H

KY SIP Narrative, Pages 82-83:

Section 308(d)(3)(v) of the regional haze rule lists several factors that must be addressed in each SIP. These factors include the role of fire at Class I areas and status of state planning for smoke management, the role of dust and fine soil at Class I areas and status of state plans to mitigate emissions from construction activities and the role of NH3 and potential benefits if emissions from agricultural sources were mitigated. 

As discussed in Section 2.4 and demonstrated in Figures 2.4-1 and 2.4-2, elemental carbon (sources include agriculture, prescribed wildland fires, and wildfires) is a relatively minor contributor to visibility impairment at the Class I area in Kentucky. However, KYDAQ has an open burning regulation, 401 KAR 63:005, which establish requirements for the control of open burning in Kentucky. The KYDAQ believes that 401 KAR 63:005, which is already incorporated into Kentucky's SIP, provides additional support to aid the Commonwealth with meeting its reasonable progress goals (RPGs) for this first planning period. A copy of KYDAQ's open burning regulation can be obtained at www.lrc.ky.gov. The exact benefits from the reduction in open burning emissions can not be quantified at this time and will not be included in Final VISTAS modeling.
 
Also as discussed in Section 2.4 and demonstrated in Figures 2.4-1 and 2.4-2, fine soils are a relatively minor contributor to visibility impairment at the Class I areas in Kentucky. Nevertheless, in regard to construction activities, KYDAQ has a fugitive emissions regulation, 401 KAR 63:010. Fugitive emissions, that provides for the control of fugitive emissions in Kentucky. The KYDAQ believes that 401 KAR 63:010, which is already incorporated into Kentucky's SIP, provides additional support to aid the Commonwealth with meeting its RPGs for this first planning period. A copy of KYDAQ's fugitive dust regulation can be obtained at www.lrc.ky.gov. 

In regard to agricultural ammonia, Kentucky, per its CMAQ regional haze modeling with VISTAS, is focused on obtaining additional SO2 emissions reductions to address CAIR, BART, and consent decrees in Kentucky. The reduction in large amounts of SO2 emissions will lessen the formation of Ammonium sulfate, (NH4)2SO4 emissions.  

Additional Kentucky EGU controls for CAIR per IPM, consent decrees, and BART that can be quantified have been included by VISTAS in a final modeling run to address the cumulative benefits from the emission controls discussed in Section 7.2.1, any controls resulting from BART determinations within the VISTAS states and any other controls resulting from the states within VISTAS to address reasonable progress that can be quantified. If the final modeling run and analyses are completed within a reasonable period of time prior to the final Regional Haze SIP submittal on December 17, 2007, the KYDAQ would consider incorporating these findings in the final SIP, including revising reasonable progress goals if needed. If the modeling runs and analyses are not completed in time, the KYDAQ will review the information as it becomes available and determine if a SIP revision is necessary. Furthermore, if the addition of some additional controls or changes in a final VISTAS model run does not significantly change the current VISTAS regional haze modeling results for Kentucky, as presented in Section 8, then it is not likely that KYDAQ will modify its regional haze SIP for this reason. 

EPA Findings:  
Kentucky has adequately addressed this provision.   

KYDAQ identified 10 emissions units at five facilities in Kentucky with SO2 emissions that were above the Commonwealth's minimum threshold for reasonable progress evaluation because they were modeled to fall within the sulfate AOI of any Class I area and have a one percent or greater contribution to the sulfate visibility impairment to at least one Class I area.  Nine of these 10 emissions units were already subject to CAIR.  KYDAQ determined that the only unit not subject to CAIR that falls within the sulfate AOI of any Class I area and contributes one percent or more to visibility impairment is located at Century Aluminum of KY LLC.

Kentucky made an adequate demonstration it is SIP to show that no additional controls for SO2 beyond CAIR are reasonable for reasonable progress for its EGUs for this first implementation period.  As detailed in the action accompanying this TSD, KYDAQ considered the four reasonable progress factors set forth in EPA's RHR as they apply to the Commonwealth's entire EGU sector.  Since EPA made the determination in CAIR that the earliest reasonable deadline for compliance for reducing emissions was 2015, KYDAQ concluded that the emissions reductions required by CAIR constitute reasonable measures for Kentucky EGUs during this first assessment period (between baseline and 2018).  This conclusion is bolstered by the fact that visibility improvement at Mammoth Cave National Park is projected to exceed the uniform rate of progress in this first implementation period.  KYDAQ also stated in its SIP that the Commonwealth intends to re-evaluate the IPM predictions of SO2 reductions for CAIR at the time of the next periodic report to ensure that the reductions predicted by IPM for CAIR are taking place where expected and needed.  

Prior to the CAIR remand by the D.C. Circuit, EPA believed the State's demonstration that no additional controls for SO2 beyond CAIR are reasonable for reasonable progress for its EGUs for this first implementation period to be acceptable.  In this instance, EPA considered the visibility improvement at Class I areas in Kentucky and affected nearby states, the time necessary for compliance, the cost of compliance, and available reasonable controls, and EPA's belief that the CAIR requirements reflect the most cost-effective controls that can be achieved over the CAIR SO2 compliance timeframe, which spans out to 2015 and overlaps most of the first regional haze implementation period.  However, as explained in Section III.C.1 - EPA Findings, the State's demonstration regarding CAIR and reasonable progress for EGUs, and other provisions in this SIP revision, are based on CAIR and thus, the Agency proposes to issue a limited approval and a proposed limited disapproval of the State's regional haze SIP revision.


III. Evaluation of Regional Haze Submittal
      C. Long-Term Strategy 
C.3. Source or Source Category Specific Analyses for Reasonable Progress:  (40 CFR 51.308(d)(1)(i)(A)

For the significant sources and source categories located within the State and determined to be significantly impacting a mandatory Class I Federal area, the State must document that it considered the costs of compliance, the time necessary for compliance, the energy and non-air quality environmental impacts of compliance, the remaining useful life of any potentially affected sources, and how these factors were taken into consideration in establishing the reasonable progress goal. 

State Regional Haze Submittal
KY SIP Narrative, Pages 81-82:

Results of four-factor analysis 
The following is a brief summary of the Non-EGU four-factor analysis. Additional detail is included in Appendix H. The KYDAQ used the cost of compliance as a screening tool to determine the universe of sources to perform the full four-factor evaluation. Therefore, the summary is focused on the cost of control. The dollar per ton of SO2 removed threshold that the KYDAQ used to determine if the cost was reasonable was based on EPA's CAIR cost analysis for implementing CAIR (See EPA's CAIR cost analysis in 70 FR 25201-25208 12May2005 available in Appendix H). After a review of EPA's CAIR cost analysis, the KYDAQ determined that the CAIR SO2 control costs vary by year of analysis (2010 vs. 2015) and may range from $400 to $3,400 per ton of SO2 removed. Ultimately, EPA found a consistent marginal cost for both years at $2000 per ton, which KYDAQ believes establishes an appropriate threshold against which cost-effectiveness may be evaluated for reasonable progress. During the current reasonable progress assessment, no Non-EGU units in Kentucky were identified for additional control because no measures were found to be cost-effective. 
For Non-EGUs, KYDAQ found that emissions from the following facility contributed one percent or more to visibility impairment in a Class I area, and therefore focused the reasonable progress assessments on specific units at this facility: 

   :: Century Aluminum of Kentucky for impacts at Mammoth Cave. 

The KYDAQ also looked at what sources in Kentucky may be impacting Class I areas located outside of the Kentucky, as well as what sources located outside of Kentucky may be impacting Kentucky's Class I area. KYDAQ, based on its Q/d times RTMax analysis identified eight EGUs, six from Indiana and two from Tennessee, with a one percent or more contribution for the Mammoth Cave area of influence. KYDAQ sent letters to Indiana and Tennessee indicating that no additional controls are requested at this time since Mammoth Cave is currently exceeding the uniform rate of progress and the EGUs are being addressed by CAIR (See copies of the letters in Appendix J). In addition, based on the KYDAQ Q/d times RTMax analysis, no Kentucky sources were identified with a contribution of one percent or more to the visibility impairment at Class I areas in other states. The list of sources identified by the KYDAQ's Q/d times RTMax analysis for given Class I areas are available in Appendix H. 

Cost of Compliance 
VISTAS evaluated control options and costs for sources within the AoI for the Class I areas of concern. VISTAS used EPA's AirControlNet software to evaluate control options and costs for controls. The SO2 control suggested by the VISTAS control cost spreadsheet for Century Aluminum is a sulfuric acid plant at a cost of $14,207, $23,020, and $43,281 per ton of SO2 removed for potlines 1-4, potline 5, and the anode baking furnace respectively (See the VISTAS control cost spreadsheet for Century Aluminum in Appendix H). Therefore, since the cost of compliance for the control option ranges from 7 to 22 times greater than the cost-effectiveness threshold, the KYDAQ concludes that there are no cost-effective controls available for these Century Aluminum units at this time within the cost threshold established for this reasonable progress assessment. 

Time Necessary for Compliance, Energy and Non-Air Impacts, and Remaining Useful Life 
The three remaining statutory factors: 1) time necessary for compliance, 2) the energy and non-air quality environmental impacts of compliance, and 3) the remaining useful life of the emissions unit, while required to be considered, were deemed not applicable, since there were no cost-effective controls to evaluate.


EPA Findings:  Kentucky has adequately addressed this provision. KYDAQ concluded, based on their evaluation of VISTAS control costs spreadsheet, that no further controls were warranted for Century Aluminum.  After reviewing KYDAQ's methodology and analyses, EPA finds Kentucky's conclusions acceptable.  EPA also finds that all currently available control technologies at the time of analysis and applicable to this facility were evaluated fully in terms of the four reasonable progress factors, and the Commonwealth consistently applied its criteria for reasonable compliance costs.  Although the use of a specific threshold for assessing costs means that a state may not fully consider available emissions reduction measures above its threshold that would result in meaningful visibility improvement, EPA believes that the Kentucky SIP still ensures reasonable progress.  In proposing to approve Kentucky's reasonable progress analysis, EPA is placing great weight on the fact that there is no indication in the SIP submittal that Kentucky, as a result of using a specific cost effectiveness threshold, rejected potential reasonable progress measures that would have had a meaningful impact on visibility in its Class I area.  EPA notes that given the emissions reductions resulting from CAIR, Kentucky's BART determinations, and the measures in nearby states, the visibility improvements projected for the affected Class I area are in excess of that needed to be on the uniform rate of progress glidepath.III. Evaluation of Regional Haze Submittal 
      C. Long-Term Strategy 
C.4. Enforceability of Reasonable Progress Measures: [40CFR51.308(d)(3)]
The plan needs to provide enforceable emissions limitations and schedules for compliance for each newly adopted measure needed to achieve the reasonable progress goals. 

State Regional Haze Submittal
KY SIP Narrative, Pages 81-82:
No newly adopted source specific measures were required to achieve the reasonable progress goals. 

EPA Findings:  Kentucky has adequately addressed this provision. 

III. Evaluation of Regional Haze Submittal 
      D. Best Available Retrofit Technology (BART) Requirements 
D.1 Identification of all BART Eligible Sources: The SIP should include a list of all BART-eligible sources within the State.

State Regional Haze Submittal
KY SIP Narrative, Page 57:

The following is a list of BART-eligible sources in Kentucky. See Appendix L for detailed information regarding each of the BART-eligible sources: 

   :: American Electric Power Big Sandy Plant 
   :: AK Steel Corporation. - Coke Mfg Plant 
   :: AK Steel Corporation - Steel Plant 
   :: Alcan Primary Products Corporation 
   :: Arch Chemicals Inc. 
   :: Calgon Carbon Corporation 
   :: Century Aluminum 
   :: Commonwealth Aluminum Lewisport LLC 
   :: Duke Energy East Bend Station 
   :: E.ON U.S Brown Station 
   :: E.ON U.S Cane Run Station 
   :: E.ON U.S Ghent Station 
   :: E.ON U.S Mill Creek Station 
   :: East Kentucky Power Cooperative Cooper Station 
   :: East Kentucky Power Cooperative Spurlock Station 
   :: Henderson Power and Light 
   :: Marathon Petroleum Company Refinery 
   :: Martin County Coal Corporation 
   :: NewPage Corporation Wickliffe Paper Company 
   :: Owensboro Municipal Utilities 
   :: Pinnacle Processing Inc. 
   :: TVA Paradise Plant 
   :: Western Kentucky Energy Coleman Station 
   :: Western Kentucky Energy Green Station 
   :: Western Kentucky Energy Reid/Henderson Station 
   :: Westlake Vinyls Inc. 

The BART-eligible sources were identified using the methodology in the BART Guidelines. 
    *  One or more emissions units at the facility fit within one of the 26 categories listed in the BART Guidelines;
    *  The emission unit(s) were in existence on August 7, 1977 and began operation at some point on or after August 7, 1962; and
    *  The potential emissions considering enforceable limits from all emission units identified in the previous two bullets emission units were 250 tons or more per year of any of these visibility-impairing pollutants: SO2, NOx, and PM10. 
 
The BART Guidelines recommend addressing these visibility-impairing pollutants: SO2, NOx, and particulate matter, and suggest that States use their best judgment in determining whether to address VOC or ammonia emissions. The KYDAQ addressed SO2 and NOx, and used particulate matter less than 10 microns in diameter (PM10) as an indicator for particulate matter to identify BART-eligible units, as the BART Guidelines recommend. As discussed in detail in Appendix L, VISTAS modeling demonstrated that VOCs and ammonia from point sources are not visibility-impairing pollutants. For this reason, the KYDAQ did not evaluate emissions of VOCs and ammonia in BART determinations. Additional BART modeling information and BART related information regarding KYDAQ BART-eligible sources is available in Appendix L.

EPA Findings: Kentucky has adequately addressed this provision.
 
III. Evaluation of Regional Haze Submittal 
      D. Best Available Retrofit Technology (BART) Requirements 
D2. Modeling or other Demonstrations for BART Exempt Sources: For each BART-eligible source in the State exempted from BART, a demonstration, using an acceptable approach, that the source or group of sources does not violate the established threshold. The established exemption threshold for determining if a BART-eligible source contributes to regional haze cannot exceed 0.5 deciviews.

State Regional Haze Submittal
KY SIP Narrative, Pages 58-61:

All of Kentucky's twenty-six BART-eligible sources had BART exemption-modeling demonstrations performed; nine of the sources that had Q/d of less than 10 for actual 2002 SO2 emissions had exemption modeling performed through VISTAS and 17 sources performed the BART exemption modeling with their own contractor. Twenty-one of the twenty-six sources were able to demonstrate exemption from BART (< 0.5 dv) either with 12 km or 4 km modeling. Results of these demonstrations are summarized in Table 7.5.2-1 as follows. Additional details are available in Appendix L. Facilities found to be subject to BART were required to complete a BART determination analysis. 

Table 7.5.2-1 represents the facilities that were able to demonstrate exemption from BART based on CALPUFF modeling conducted using the VISTAS modeling protocol and either the old IMPROVE equation or new IMPROVE equation. The KYDAQ is proposing to exempt the units listed in Table 7.5.2-1. For further details about the BART exemption modeling, please refer to Appendix L. 


      
KY SIP Appendix L, Page 15:

Kentucky opted to consider its BART-eligible sources subject to BART unless the source demonstrated exemption via modeling. BART-eligible sources can be excluded from BART determinations by demonstrating that the source cannot be reasonably expected to cause or contribute to visibility impairment in a Class I area. The threshold for determining that a source causes visibility impairment is set at 1.0 dv change from natural conditions over a 24 hour averaging period. The BART guidelines also propose that the threshold at which a source may "contribute" to visibility impairment should not be higher than 0.5 deciviews; however, depending on factors affecting a specific Class I area it may be set lower than 0.5 deciviews. 
As stated in the BART regulation EPA's preferred approach for determining cause or contribution is an assessment with an air quality model such as CALPUFF or other appropriate model followed by comparison of the estimated 24-hour visibility impacts against a threshold above estimated natural conditions to be determined by the State. EPA recommends that the 98th percentile value from the modeling be compared to the State's chosen contribution threshold to determine if a source does not contribute to visibility impairment and thus is not subject to BART. Comparison of the 98th percentile value to the threshold must be made for each Class I area. For an annual period, this implies the 8th highest 24-hr value at a particular Class I area is compared to the contribution threshold. For a 3-year modeling period, the 98th percentile value may be interpreted as the highest of the three annual 98th percentile values at a particular Class I area or the 22nd highest value in the combined three year record, whichever is more conservative. 
Kentucky worked with the regional planning organization (RPO) VISTAS on development of the VISTAS Protocol for the Application of the CALPUFF Model for Analyses of Best Available Retrofit Technology (BART) (available in Appendix L.5). The common protocol was established to provide the basis for a common understanding among the organizations performing BART analyses or reviewing BART modeling results in the VISTAS region. 
The VISTAS protocol describes common procedures for carrying out air quality modeling to support BART determinations that are consistent with the 40 CFR Part 51 Appendix Y guidelines. The protocol provides a consistent model, CALPUFF, and modeling guidelines for BART determinations, clearly delineated modeling steps, a common CALPUFF configuration, guidance for site-specific modeling, and common expectations for reporting model results. Details of the CALPUFF system can be found in Chapter 3 of the VISTAS protocol, specific recommendations for its application for BART purposes are found in Chapter 4, and specific information that should be included in site-specific protocols is found in Chapter 5. 

KY SIP Appendix L, Pages 8-9:

4. Contribution Threshold 
Determining whether a source causes or contributes to visibility impairment is one step in the BART review process. The Guidelines for BART Determinations Under the Regional Haze Rule (40 CFR 51, Appendix Y, Section III.A.1) state that "A single source that is responsible for a 1.0 deciview change or more should be considered to `cause' visibility impairment." The guideline document also states that "the appropriate threshold for determining whether a source `contributes to visibility impairment' may reasonably differ across states," but, "As a general matter, any threshold that you use for determining whether a source `contributes' to visibility impairment should not be higher than 0.5 deciviews." The rationale for these instructions is provided in the preamble to the BART guidance, in the statement, "If `causing' visibility impairment means causing a humanly perceptible change in visibility in virtually all situations (i.e. a 1.0 deciview change), then `contributing' to visibility impairment must mean having some lesser impact on the conditions affecting visibility that need not rise to the level of human perception." (70 FR 39120, footnote 31). The guidance document itself also states that, "States remain free to use a threshold lower than 0.5 deciviews if they conclude that the location of a large number of BART-eligible sources within the State and in proximity to a Class I area justify this approach." 
The EPA's documents strive to set these thresholds in the context of the human perception of visibility change. As noted above, the EPA considers a 1.0-deciview change in visibility to be humanly perceptible "in virtually all situations." Also, the preamble to the BART guidance (70 FR 39119, Footnote 28) cites an analysis in an appendix of a NAPAP (National Acid Precipitation Assessment Program) report, which asserts that "changes in light extinction of 5 percent will evoke just noticeable changes in most landscapes."2 (A 5% change is approximately 0.5 dv.) But, as noted above, the preamble also states that perceptibility is not a prerequisite for choosing a contribution threshold. Putting all this together, it appears that "causing" visibility impairment means having a humanly perceptible impact (for which EPA considers the practical threshold to be 1.0 dv) while "contributing" to visibility impairment means having a smaller impact (for which EPA considers the threshold to be 0.5 dv or some smaller value) that may or may not be perceptible. 
The EPA argues that a contribution threshold of less than 0.5 dv impact per source is appropriate when multiple sources contribute, in order to limit the combined effect of these sources. As an example, EPA asserts that if there were 100 sources, each affecting visibility by 0.1 dv (presumably an imperceptible amount), their total impact would be 10 dv, which can be expected to be quite perceptible (70 FR 39121. 1st column). The point remains that multiple sources can cause a larger impact than a single one. For BART purposes, visibility impacts are calculated as 24-hr averages of 1-hr plume impacts, so if the plumes from the various sources each impact the point of interest at some time during a 24-hr period (not necessarily all at the same hour) then the 24-hr average will reflect their combined impact. 
Appendix L Draft Kentucky Regional Haze SIP 8 
KYDAQ concluded that the EPA suggested contribution threshold of 0.5 dv was appropriate in this situation since there are a limited number of in and out of state sources that impact the various Class I areas in the state. In addition there are a limited number of sources in close proximity to each of the Class I areas. Considering results of the visibility impacts modeling conducted (see Section 7), a 0.5 dv threshold was appropriate and a lower threshold was not warranted since the majority of the visibility impacts were well below 0.5 deciviews. Also even though several sources impacted each Class I area, the overall impacts were low from the sources.

KY SIP Appendix L,  Page 12:

The results show that the maximum impact from eliminating all point source VOC emissions in the VISTAS 12 km domain is less than a 0.5 dv for all Class I areas in the VISTAS domain. Given that the fraction of the total point source VOC emissions that are also BART-eligible in the state of Kentucky is just 8%, the expected impact of controlling VOCs from a BART source would be much less than the 0.5 dv threshold. VISTAS and Kentucky conclude that VOCs from point sources are not a visibility impairing pollutant for BART purposes and that BART-eligible sources do not need to consider VOC emissions.

Similar to its treatment of VOCs, EPA guidance allows States the discretion to decide whether or not ammonia emissions are to be considered for BART purposes based on evaluations of the contributions of the emissions to haze at Class I areas in their areas of influence. One approach a State can use to determine whether applying BART will be needed is to evaluate the haze impacts of all current emissions from all BART-eligible sources in the State. If the impact from all sources in the state is less than the contribution threshold established by the State, 0.5 dv for Kentucky, then source by source analysis for BART is not needed. Kentucky has determined through modeling that with one exception ammonia (NH3) emissions from point sources are not anticipated to cause or contribute significantly to any impairment of visibility in Class I areas and should be exempt for BART purposes.


EPA Findings: Kentucky has adequately addressed this provision.  The State has adequately identified the BART-eligible sources and those sources subject to BART controls.  The SIP successfully developed and provided the technical analyses to support the BART assessment provision for the non-CAIR sources.  Kentucky opted to have CAIR satisfy BART for SO2 and NOx for affected CAIR EGUs, as allowed under the regional haze regulations.  As explained in Section III.C.1 - EPA Findings, the State's demonstration regarding CAIR and reasonable progress for EGUs, and other provisions in this SIP revision, are based on CAIR and thus, the Agency proposes to issue a limited approval and a proposed limited disapproval of the State's regional haze SIP revision.

      The CALPUFF modeling system was adequately applied by KYDAQ in the assessment of visibility impairment in Class I areas for specific sources subject to the BART regulations.  The BART regulations indicate that CALPUFF or other appropriate models can be used to determine if an individual source is anticipated to cause or contribute to impairment of visibility in Class I areas.  It is important to note that the current regulatory niche for the CALPUFF modeling system in 40 CFR part 51, Appendix W: Guideline on Air Quality Models (Appendix W) is for compliance with non-attainment New Source Review (NNSR) and Prevention of Significant Deterioration (PSD) permitting requirements for long range transport (i.e., greater than 50 KM) applications at Class I areas.  There is no preferred model for assessing visibility impact under BART.  However, it is appropriate for the Commonwealth of Kentucky to use the CALPUFF modeling system for its BART exemption and control determination modeling.
      
      VISTAS developed a "Protocol for the Application of CALPUFF for BART Analyses."  The final VISTAS BART Modeling Protocol (Revision 3.2, dated Aug. 6, 2006) states that VISTAS will use CALPUFF ver. 5.754 and CALMET ver. 5.7 (aka, the VISTAS version of CALPUFF). No formal approval of the CALPUFF VISTAS version (either ver. 5.754 or 5.756) has been made by EPA Region 4.  However, there has been an implied acceptance of the VISTAS CALPUFF version as EPA Region 4 has reviewed and commented on the VISTAS BART Modeling Protocol and the source-specific modeling protocols and reports that were developed by the Kentucky specific sources performing BART modeling.   The VISTAS BART modeling protocol underwent an extensive stakeholder development and review process, that involved EPA, FLMs, industrial sources, trade groups, lawyers and other interested parties, before the Commonwealth decided to require CALPUFF modeling to evaluate the BART requirements.  
      
      The Commonwealth required all BART-eligible sources to perform BART exemption modeling and to develop source-specific protocols that complied with the VISTAS protocol.  EPA Region 4 provided reviews and comments on the source-specific protocols and their subsequent exemption modeling reports.  The exemption protocols and modeling reports are presented in Appendix L of Kentucky's regional haze SIP. 
      
      The CALPUFF model was simulated with the default inputs as recommended in the March 16, 2006 EPA memorandum, "Dispersion Coefficients for Regulatory Air Quality Modeling in CALPUFF. "  A new methodology was used by the VISTAS states to process the mass concentrations from CALPOST (post-CALPUFF processor) using the new IMPROVE equation so that the BART analyses could consider both the old and new IMPROVE equations.  EPA Headquarters, EPA regional offices, and the Federal Land Managers (FLMs) from the National Park Service and the U.S. Fish and Wildlife Service consulted on the VISTAS innovative approach that was developed to utilize the new IMPROVE equation in the BART assessment and to determine the change in visibility impacts.  It should be noted that while these other parties were consulted in preparation of EPA Region 4's response, the response represented the EPA Region 4 position on the new methodology.  Use of the methodology by BART sources in States located in other EPA Regions would require consultation with the appropriate regional office.  EPA required the Commonwealth of Kentucky to submit a request to the EPA Region 4 Regional Administrator requesting the use of this new methodology for the BART CALPUFF-based modeling.  A September 13, 2007, letter justifying the need for this approach was sent and approval from the Regional Administrator was provided to the Commonwealth on October 5, 2007.  The methodology and Commonwealth's and Region 4 letters are located in KYDAQ's TSD.
       
      The Commonwealth was required to justify their choice of the 0.5 dv contribution threshold for exempting sources from BART eligibility.  EPA agrees with KYDAQ's rationale for choosing this threshold value since there are a limited number of in and out of state sources that impact the various Class I areas in the State.  In addition there are a limited number of sources in close proximity to each of the Class I areas.  Considering results of the visibility impacts modeling conducted (see Section 7 of Appendix L), a 0.5 dv threshold was appropriate and a lower threshold was not warranted since the majority of the visibility impacts were well below 0.5 deciviews and the sources are distributed across the Commonwealth.  Also, even though several sources impacted each Class I area, the overall visibility impacts were low from the sources.  As stated in the BART Guidelines, where a state concludes that a large number of these BART-eligible sources within proximity of a Class I area justify a lower threshold, it may warrant establishing a lower contribution threshold.  See 70 FR 39161-39162 (July 6, 2005).  In addition, EPA has not made any adverse comments on the use of the CALPUFF VISTAS versions. 
      
      Note that EPA's reference to CALPUFF encompasses the whole CALPUFF modeling system, including CALMET/CALPUFF/CALPOST and other pre and post processors.  The different versions of CALPUFF have corresponding versions of CALMET, CALPOST, etc. which may not be compatible with previous versions (e.g., the output from a newer version of CALMET may not be compatible with an older version of CALPUFF).  The different versions of the CALPUFF modeling system are available from the model developer on the following website:  http://www.src.com/verio/download/download.htm

References used in the EPA review: 

   1. 40 CFR Part 5, Appendix W: Guideline on Air Quality Model
   
   2. October 5, 2007 letter from James I. Palmer, Regional Administrator, US EPA Region 4, to Barry Stephens, TDEC-APC,  the use of the VISTAS methodology for use of the New IMPROVE equation with CALPUFF model results.  
   3. BART Guidelines
   
   4. IWAQM. 1998. Interagency Workgroup on Air Quality Modeling (IWAQM) Phase 2 Summary Report and Recommendations for Modeling Long-Range Transport and Impacts on Regional Visibility.  EPA-454/R-98-019. U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC. 
   
   5. VISTAS, 2006. Tombach, I., P. Brewer, T. Rogers, and C. Arrington, Protocol for the Application of the CALPUFF Model for Analysis of Best Available Retrofit Technology (BART) Revision 2  -  3/9/06 developed through consultation with EPA, the FLMs, VISTAS state and local Agencies and industry stakeholders VISTAS, 2006. Tombach, I., P. Brewer, T. Rogers, and C. Arrington, Located at http://www.vistas-sesarm.org/documents/BARTModelingProtocol_rev3.2_31Aug06.pdf.

   6. March 16, 2006 EPA memo from Dennis Atkinson to Kay T. Prince on Dispersion Coefficients for Regulatory Air Quality Modeling in CALPUFF
   
   7. Additional Regional Haze Questions - September 27, 2006 Revision.  Informal questions and answers developed by EPA Regional Offices Workgroup and shared with States and Tribal Agencies, FLMs and industry stakeholders.

   8. Appendix Y to 40 CFR Part 51, the Guidelines for BART Determinations Under the Regional Haze Rule



III. Evaluation of Regional Haze Submittal 
      D. Best Available Retrofit Technology (BART) Requirements 
D.3. Identification of Sources Addressed by CAIR: States which adopt the CAIR cap and trade program for SO2 and NOx are allowed to apply CAIR controls as a substitute for controls required under BART. The SIP should include a list of all BART-eligible sources within the State for which the CAIR rule is being applied.

State Regional Haze Submittal
KY SIP Narrative, Page 61:

Fourteen of Kentucky's twenty-six BART-eligible sources are EGUs that are subject to CAIR. 
The USEPA has determined that, as a whole, the CAIR cap-and-trade program improves visibility more than implementing BART for individual sources in states affected by CAIR. A State that opts to participate in the CAIR program under 40 CFR 96.201-.224 (Subpart AAA through EEE) need not require affected BART-eligible EGUs to install, operate, and maintain BART for SO2 or NOx emissions. Given that most BART-eligible units have existing or are installing scrubbers and NOx controls, and since Kentucky is participating in CAIR and accepts the USEPA's overall finding that CAIR "substitutes" for BART for NOx and SO2, Kentucky's EGUs were allowed to submit BART exemption modeling demonstrations for PM emissions only. Nine of the fourteen Kentucky EGUs demonstrated that they do not contribute to visibility impairment in any Class I area. 

EPA Findings: Kentucky has adequately addressed this provision.  The State has adequately identified the BART-eligible sources and those sources subject to BART controls.  Prior to the CAIR remand, EPA believed the Commonwealth's demonstration that CAIR satisfies BART for SO2 and NOX for affected EGUs for the first implementation period to be approvable and in accordance with 40 CFR 51.308(e)(4).  However, the Commonwealth's demonstration regarding CAIR and BART for EGUs, and other provisions in its regional haze SIP revision, are based on CAIR and thus, the Agency proposes to issue a limited approval and a limited disapproval of the Commonwealth's regional haze SIP revision. 



III. Evaluation of Regional Haze Submittal 
      D. Best Available Retrofit Technology (BART) Requirements 
D.4. Source Specific BART Determinations: For each BART-eligible source in the State subject to BART, a determination of BART based on an analysis of the best system of continuous emission control technology available and associated emission reductions achievable (unless the State demonstrates that an emissions trading program or other alternative will achieve greater reasonable progress toward natural visibility conditions). To address the requirements for BART, the State must take into consideration all of the  "Five Factors" from 51.308(e)(1)(ii)(A):  (1) the technology available, (2) the costs of compliance, (3) the energy and non-air quality environmental impacts of compliance, any pollution control equipment in use at the source, (4) the remaining useful life of the source, and (5) the degree of improvement in visibility that may reasonably be anticipated to result from the use of such technology. 

State Regional Haze Submittal
KY SIP Narrative, Pages 62-63:

Table 7.5.3-1 presents BART determination modeling results for the five Kentucky EGU sources that were unable to demonstrate a contribution of less than 0.5 dv at all Class I areas within 300 km from their source location. These five sources are considered to be "subject to BART" and were required to submit BART determination modeling containing their evaluation of potential BART control options and proposed BART determinations. Each of these sources has agreed to install emission controls to address inorganic condensable particulate emissions (SO3/H2SO4), which is causing the sources to be subject to BART. The BART determination resulting controls are provided in the Table 7.5.3-1 that follows and they were taken to public hearing concurrent with the public hearing on Kentucky's Regional Haze SIP. Table 7.5.3-2 that follows, in addition to the emission controls, provides the source's BART emission limits and timeframes for compliance. Applicable BART controls and emission limits will be incorporated into the sources' Title V permits as appropriate or upon renewal. In addition, since TVA had previously indicated to the KYDAQ its plans to install hydrated lime injection controls on TVA Paradise Units 1-3 to mitigate opacity due to SO3 emissions and that additional controls are not cost-effective at this time, the KYDAQ has determined BART to be no control for TVA Paradise Units 1-3. However, as related by TVA, the hydrated lime injection controls for TVA Paradise Units 1-3 will be in place well before the BART controls are required; will achieve the reduction in visibility impacts listed in the Draft Implementation Plan (Kentucky Regional Haze SIP); and will be included in TVA Paradise's Title V permit. Specifically, regarding the installation of hydrated lime injection controls for TVA Paradise Units 1-3, TVA has communicated to KYDAQ its proposed plan that provides for permitting activities to proceed in July 2008; for construction to begin in mid-2009 on Unit 3 with construction for Unit 1 and 2 to follow; and for controls to be operating on all three TVA Paradise units possibly by the fall of 2010. Also, as indicated in the E.ON U.S. Mill Creek BART determination submittal, the average cost for installing sorbent controls on all four Mill Creek units is about the same (an estimated 5.1 million $/dv). However, sorbent injection at all four units would mean an additional total capital investment of $8.8 million as compared to controls only on the larger Units 3 and 4. Therefore, E.ON U.S. concluded that BART should be the installation of sorbent injection controls on the larger Mill Creek Units 3 and 4 since they can achieve an estimated 70 percent of the total dv improvement achieved by controlling all four units. Given the extra cost for the lesser additional dv improvement for Units 1 and 2, the Cabinet agreed that BART for Mill Creek is the installation of sorbent injection controls on the larger Units 3 and 4. For further details about the BART determination modeling for the five BART-Subject sources, please refer to Appendix L. 

Updated May 28, 2010 submittal:



KY SIP Appendix L11, Pages 14-18:

EKPC BART Assessment for Spurlock and Cooper Stations






Updated May 28, 2010 submittal:

A. Control Technology Comparison
In the 2007 BART Submittal, EKPC determined that a WFGD/WESP control train capable of achieving 0.030 lb/mmBtu filterable PM and 0.052lblmmBtu total PM was BART for Cooper Units I and 2. EKPC is requesting that it be allowed to substitute a
DFGD/FF control train capable of achieving 0.030 lb/mmBtu filterable PM and 0.045
Lb/mmBtu total PM for the WFGD/WESP control train previously approved.

The previously approved WFGD/WESP technology utilizes a lime or limestone slurry- based scrubbing medium for SOx removal in the absorber, followed by a wet ESP, which
Provides removal of scrubber-generate PM as well as sulfuric acid mist. As discussed in the 2007 BART Submittal, WFGD/WESP can provide control of filterable PM to levels of 0.030 lb/mmBtu. Total PM emissions from the WFGD/WESP control train were estimated using the National Park Service (NPS PM speciation spreadsheet (see Attachment A).  At a filterable PM emission rate of 0.030 lb/mmBtu, the NPS speciation
Spreadsheet for PC boilers with WFGD/WESP estimates total PM emissions to be 0.052
lb/mmBtu. These values were used in the previous BART submittal.

The proposed DFGD/FF control train utilizes a pebble lime-based dry scrubber system in combination with a fabric filter for control of SO2, PM, and sulfuric acid mist. The
DFGDIFF can also attain filterable PM levels meeting the previously accepted 0.030
lb/mmBtu filterable PM emission rate. Condensable PM emissions from the DFGD/FF as specified by Burns & McDonnell Engineering Company, Inc. equate to 50To of the filterable emissions, providing a condensable PM emission rate of 0.015 lb/mmBtu for a total PM rate of 0.045 lb/mmBtu. Based upon this information, at equal filterable PM emission levels, the DFGD/FF actually provides better overall control for total PM than the approved WFGD/WESP, primarily due to lower sulfuric acid emissions. The NPS spreadsheet for WFGD/WESP determines sulfuric acid emissions (listed as inorganic
condensable PM) to be 0.018 lb/mmBtu, while the NPS spreadsheet for DFGD/FF
(Attachment A) determines sulfuric acid emissions to be 0.012lblmmBtu.

Table 1 provides a summary of PM emission rates for the two control options. Based on
these rates, the proposed DFGD/FF is more effective at controlling PM emissions than
the WFGD/WESP control train identified in the SIP.

B. BART Modeling
In order to confirm and to demonstrate the equivalence of the DFGD/FF control train to
the previously accepted WFGD/WESP control train, EKPC performed additional CALPUFF modeling to evaluate the visibility impacts in Class I areas from Cooper Units 1 and 2. PM emissions were speciated in the same manner as described in the 2007
BART Submittal, except that a maximum filterable PM limit of 0.030 and a total PM limit of 0.045 were input into the NPS speciation spreadsheets for dry-bottom PC boilers employing FGD and fabric filtration. Emissions were then calculated using the spreadsheets and a maximum continuous rate (heat input) (MCR) of 1,350 mmBtu/hr for
Unit 1 and 2,400 mmBtu/hr for Unit 2. The completed NPS spreadsheets for Cooper
Units I and 2 are included in Attachment A. Table 2 summarizes the speciated emissions after retrofit.

Stack parameters for the two Cooper units subject to the BART analysis are presented below in Table 3. The exit velocity and temperature parameters are different than those used in the previous modeling and reflect the use of the DFGD/FF control train.

Having specified exhaust parameters and calculated the speciated emission rates, the
CALPUFF model was run as described previously for Units 1 and 2 at Cooper for each Class I area within 300 km.

Table 4 shows the top eight 24-hour changes in light extinction (deciviews, dv) from the 20 % best days for each of the Class I areas subject to analysis after application of retrofit control for PM. Table 5 presents a summary of the results of the revised BART modeling for the Cooper sources and each C lass I area with the number of days and receptors in each Class I area where dv > 0.5. These tables demonstrate compliance with the Regional Haze Rule since the 98th percentile modeled values (8tn highest) or the 22nd highest predictions over the three years modeled, whichever are higher, are below the exemption threshold of 0.5 dv in each Class I area.

In fact, no days in any of the modeled Class I areas out of the three years modeled was predicted where the change in light extinction was greater than 0.5 dv. Table 6 shows a comparison of the 2007 BART Submittal to this submittal for the 98th percentile and maximum 24-hour change in light extinction in deciviews. As shown, with the exception
of Linville Gorge, where the 98th percentile change in light extinction goes up slightly from 0.046 to 0.047, the visibility impacts decrease for both the 98th percentile and the maximum 24-hour.

Therefore, application of DFGD/FF controls to Cooper Units 1 and 2, with a filterable
PM limit of 0.030 lb/mmBtu, mitigates any adverse visibility/impact in Class I areas within 300 km of each source and fulfills the BART requirements.

TVA Paradise Plant
KY SIP, Appendix L.11, Pages L.11-30 to L.11-43:

II. BART Engineering Analysis
Based on emissions testing at PAF, Units 1 and 2 emit approximately 17.7 ppmdv SO3, and Unit 3 emits approximately 17.6 ppmdv SO3 (both corrected to 3%02). These values were used as the baselines for the engineering analysis.

Two options were identified as technically feasible for controlling SO3 emissions at PAF: wet ESPs and hydrated lime injection. An ESP controls PM by imparting a negative charge to fly ash particles entering the device and collecting them on positively charged plates. In a wet ESP, the fly ash is removed by flushing the collection plates with water.
Wet ESPs are very expensive, requiring large capital investments. They also require large volumes of water for cleaning the plates. This technology is capable of about 90%
SO3 removal, so it was evaluated at an emission level of 1.8 ppmdv. It should be noted, however, that the vendor guarantee accompanying the capital cost quote that TVA received for this equipment equated to about 3 ppmdv SO3 (3%O2).
Lime injection involves injecting hydrated lime in the convective pass or ductwork of the boiler. The lime reacts with SO3 to form calcium sulfate, which is removed downstream by the PM control device. TVA plans to install lime injection systems on all three units at PAF to mitigate opacity. These systems will be installed well before the regulatory deadline for implementing BART. However, because they are not currently installed and operating, lime injection is included as a control option in the BART analysis. Lime injection testing on Unit 2 demonstrated an achievable SO3 emission level of approximately 7 ppmdv (3%O2). Because Unit 1 is similar to Unit 2, the same emission level would apply. Testing on Unit 3 has not been completed, so an achievable emission level has not been established. For the purposes of this analysis, an SO3 emission level of
7 ppmdv was assumed. The analysis for Unit 3 may have to be revised after testing is completed if that emission rate cannot be achieved without negative impacts to existing particulate control equipment.



Cost of Compliance
Costs were estimated for both control options. Capital equipment costs for the wet ESP were obtained from a vendor quote ("Budgetary Proposal, Paradise Units 1 & 2, Babcock & Wilcox Company, July 3, 2007"). This quote applies to Units 1 and 2, but it was scaled using engineering assumptions for use with Unit 3. Capital costs for lime injection systems and annual reagent costs were based on TVA experience with similar equipment at other TVA locations. Other capital and annual costs were based on information in the EPA Air Pollution Control Cost Manual, Sixth Edition.

A summary of results from the cost analysis are presented in Table 2. Details can be found in the Appendix to this report. The total capital investment for a wet ESP ranges from about $100 million for Unit 1 or 2 to almost $156 million for Unit 3. Total annual costs range from about $29 million to $44 million per year. The corresponding total cost effectiveness ranges from $27,594 to $39, 263 per ton of SO3/H2SO4. Incremental cost effectiveness compared to lime injection ranges from $77, 821 to $82,940 per ton. Based on these results, a wet ESP is clearly economically infeasible for PAF and should, therefore, be eliminated from consideration as a basis for BART.

The total capital investment for hydrated lime injection ranges from $4.2 million for Unit 1 or 2 to $8.4 million for Unit 3. Total annual costs range from about $2.3 million to $4.4 million per year. The corresponding cost effectiveness ranges from $3,265 to $6,776 per ton of SO3/H2SO4. Although considerably less expensive than a wet ESP, the cost effectiveness values for lime injection are still too high to be considered as an acceptable cost of compliance for BART.

Non-air Environmental Impacts
As stated previously, a wet ESP is a water intensive system. According to the EPA Air
Pollution Control Cost Manual, a wet ESP uses at least 5 gal/min of water per 1,000 cfm of flue gas. Based on this ratio, operating wet ESPs on all three units at PAF would consume about 16 billion gallons per year of additional water and require treatment of an equivalent volume of wastewater.


BART Modeling Analysis
TVA conducted BART modeling for PAF following both the general VISTAS protocol and the revised TVA BART protocol submitted to the Kentucky Division for Air Quality on July 17, 2006. Modeling was conducted on the VISTAS 4-km domain for the state of KY (dom 3). Two Class I areas are located within 300 km of PAF: Mammoth
Cave (MACA) at 63.3 km and Mingo (MING) at 283 km.
....
For Unit 1 or 2, lime injection is projected to reduce the visibility impact at MACA from 1.285 dv to 0.606 dv (2003). A wet ESP is projected to further reduce the impact to 0.166 dv. For Unit 3, lime injection is projected to reduce the visibility impact at MACA from 1.842 dv to 0.836 dv (2003). A wet ESP is projected to further reduce the impact to 0.264 dv.

For Unit 1 or 2, lime injection is projected to reduce the visibility impact at MING from 0.251 dv to 0.116 dv (2002). A wet ESP is projected to further reduce the impact to 0.031 dv. For Unit 3, lime injection is projected to reduce the visibility impact at MING from 0.381 dv to 0.166 dv (2002). A wet ESP is projected to further reduce the impact to 0.051 dv.
....
Both control options reduce visibility impacts at MACA and MING. However, the excessive costs associated with wet ESPs, as well as the large volumes of water required, make that option infeasible as a basis for BART. Although lime injection is less expensive, the cost effectiveness results associated with that option are too high, making the technology unacceptable as a basis for BART. As a result, TVA proposes that no additional control be required under the BART regulations.

As stated previously, TVA plans to install lime injection on all three units at PAF to mitigate stack opacity. These controls will be in place well before BART controls are required. So visibility impacts at MACA and MING will continue to be reduced, regardless of any decision related to BART.

KY SIP Appendix L.11, Pages L.11-65 to L.11-74:

AEP Big Sandy Plant
BART SCENARIOS CONSIDERED
Once, it was determined that Big Sandy Plant did not pass the BART Exemption modeling, AEP examined the potential options to meet BART. The selected options were initially focused on controlling the inorganic condensable fraction and included the following control options:
1. Inject Ammonia full time on both units to reduce the inorganic condensable fraction by up to 10 ppm
2. Inject trona on both units to reduce the inorganic condensable fraction by up to 30 
ppm.
3. Install a wet FGD system on Unit 2 with no action taken on Unit 1
4. Install a wet FGD system on Unit 2 and inject ammonia on Unit 1
5. Install a wet FGD system on Unit 2 and inject trona on Unit 1
Unit 1 is not considered a viable candidate for retrofit pollution controls that require a 15 to 20 year amortization period due to its age and size. This unit will be 50 years old in
2013. While a specific retirement date has not yet been established for this unit, the likelihood of this unit continuing operations in its present form until the 2033-2035 time frame is low. Therefore, this unit is not considered a candidate for retrofit of an FGD system, new electrostatic precipitator, or bag filter. Unit 2 is currently expected to run until at least the 2033-2035 timeframe, so retrofit controls are considered a viable option for this unit.

In addition, the options to inject Trona on either unit at Big Sandy were rejected for technical reasons. Based on the experience of AEP at units where sorbents are injected
for the reduction of inorganic condensable, particulate control problems are expected where the electrostatic precipitators are below 400 SCA (a measure of square footage of collection area per million acfm of flue gas). In the case of both Big Sandy units, the
SCA values of the existing electrostatic precipitators are below 200 SCA, which renders them unsuitable for sorbent injection without the replacement of the particulate control devices.

While developing the analysis following the completion of the modeling of these options, which were modeled on each unit individually, it was determined that a primary particulate option should also be examined. The case estimated for this study was to install a new primary particulate collector on Unit 2 to meet essentially a control level that a new unit would achieve. Enhancements to the existing particulate controls were not considered since both units currently have SO3 flue gas conditioning, Unit 2 has an ammonia conditioning system available to it, and both existing ESP boxes have optimal electrical systems for their current configuration already installed. The following sections discuss the options considered and the results of the analysis of each of these options.

PRIMARY PARTICULATE
In performing the BART modeling, we consciously chose to model Big Sandy using the permit limit and not its 2003 tested values. Based on the testing performed in 2003 on
Units 1 and 2, the tested values for these units were approximately 0.11 lb/MMBtu for Unit 2 and 0.13 lb/MMBtu for Unit 1, roughly half of the value used in the modeling.
This makes the actual impact from the primary particulate emissions from Big Sandy less than that modeled. Due to the layout of the plant, it is not possible to expand the existing particulate control equipment in its current location. Further, as a result of the space limitations at Big Sandy, the Unit 2 electrostatic precipitator was overhauled in conjunction with the SCR installation on that unit.

Based on AEP's knowledge of the particulate control systems on these units, the only potentially viable option would be to retrofit a new control device on one or both of these units. These units are both already equipped with SO3 flue gas conditioning systems on both units and an ammonia injection system on Unit 2. The electrostatic precipitators are equipped with optimal electrical systems for their design and size.

In considering the primary particulate control options, we have evaluated the installation of a bag house solely to Unit 2 with no changes to Unit 1. Based on the specification for the Southwestern Electric Power Company Turk Plant currently being permitted in Arkansas Based which has a bag house with a guaranteed emission rate of 0.012 lb/MMBtu, we would expect a similar emission rate should such a device be retrofitted at Big Sandy. 

While this alternative was not explicitly modeled with CALPUFF, the CALPOST processor allows us to adjust the individual primary particulate species to the levels that we would expect if emissions were at the rate of 0.012 lb/MMBtu. We determined the primary particulate speciated emission rates using the same spreadsheet referenced earlier with the lower anticipated emission rate for a new particulate control device. This new emission rate generated in adjustment factors (multipliers) of 0.053 for coarse particulate, 0.047 for fine particulate, and 0.052 for elemental carbon, which were input into CALPOST as multipliers to the respective pollutant species and applied only to Unit 2.
 ....
On the peak deciview impact, this results in a projected reduction in primary particulate impact to 0.124 dV from 0.466 dV at an annualized cost of $8,595,090. When spread over the total primary particulate deciviews modeled and completely attributable to the change in Unit 2 primary particulate emissions, a cumulative 76.14 dV, on days where the impact exceeded 0.5 dV, the $/dV value comes out to $112,187. In looking at the total number of exceedances of the 0.5 dV threshold, a primary particulate strategy would resolve 77 of the 350 modeled exceedances of the threshold, assuming that all of the
change from the modeled level would be attributable to the controls added to Unit 2.

SECONDARY PARTICULATE
Due to the large contribution of secondary particulates in the modeling results that contribute to the formation of regional haze, all of these methods show improvements that far exceed those found in a primary particulate strategy. The modeling results for
each option are presented in the same format as the exemption modeling for the cases examined.

Ammonia Injection on Both Units
The first of the alternatives examined for control of the condensable portion of the fine particulates at Big Sandy Plant was the use of ammonia injection on both of the units. An ammonia system currently exists on Unit 2 for purposes of precipitator conditioning
during periods of SCR operation (which will go full time in 2009 due to CAIR requirements). It was assumed that this system would be used to control up to 10 ppm of the condensable fraction of the fine particulates on Unit 2 and a feed from the current
Ammonia on Demand System would be used to feed Unit 1 along with the installation of the necessary piping and injection equipment. In addition, it has been determined that the
Unit 1 ESP would require a similar overhaul to that done to the Unit 2 ESP when the
SCR was installed on that unit. Based on our current experience with similar overhaul work at other units, it is estimated that this work on Unit 1 would have a capital cost of an additional $14,000,000 beyond the installation of the necessary ammonia equipment.
A reduction level of 10 ppm of condensable particulate was selected by AEP's Engineering Department as a level that can be achieved without risking the generation of sticky deposits that would potentially jeopardize the operation of the units. Table 4 shows the stack parameters used in the air quality modeling portion of the analysis of this option.

Table 5 shows the results of this strategy in a format identical to that used for the exemption modeling. In examining the data in Table 5, we find that there are improvements in the visibility impacts on the various Class I areas. At a cursory look, we note an improvement in peak dV impact from 2.13 dV at Otter Creek in the base case to 1.44 dV in this control case. This strategy also reduces the number of days with impacts greater than 0.5 dV by 169 over all years and Class I areas analyzed. There is also an aggregate reduction of 164.84 dV over all days in the base case where the impacts exceeded 0.5dV. Since the Ammonia on Demand system already is in place at Big Sandy
Plant, the only additional costs that would be absorbed in implementing this strategy would be the capital costs to extend the system to Unit 1 and operate in on a year around basis; this option is expected to be fairly low cost. When the costs are examined and an annual cost is determined, we find that the annualized cost is estimated at $2,595,966, resulting in a $/dV of improvement value of $15,748.

No controls on Unit 1 with Wet FGD on Unit 2
The second alternative examined for reducing the condensible fraction of particulate was the option of installing a wet FGD system on Unit 2 only with no action taken with Unit
1. This option is considered reasonable due to the likelihood that Unit 2 will have a wet FGD system installed in the future for CAIR compliance reasons. As was discussed previously, Unit 1 is not considered a candidate for controls of this nature due to the age of the unit and the probability of the unit operating for approximately 20 years following the likely installation date for a control of this nature. Based on the current AEP system experience with installing wet FGD systems on similar units to Big Sandy 2 at this time, we estimate the all in capital cost to install a wet FGD system on Unit 2 at $350/KW.
For purposes of the CALPUFF modeling analysis, Table 9 shows the model inputs used.
The Unit 2 condensables were calculated as discussed in the section of this report on
Stack Parameters and Emissions using the FGD Equipped Coal Fired Utility Unit profile spreadsheet with adjustments to the fuel profile, stack temperature, and exit velocity based on the FGD systems being installed at the John Amos Plant. Two of the units receiving wet FGD systems at John Amos Plant are from the same 800 MW series as is Big Sandy Plant Unit 2. The Amos profile used for this data is included on the enclosed DVD.
The results of this portion of the analysis shown in Table 8 indicate that 246 days with dV impacts greater than the 0.5 dV goal are eliminated over the three year period by this alternative and the peak impact is reduced from 2.13 dV to 1.27 dV. Over the days that had dV impacts greater than the goal of 0.5 dV, there is a reduction of 225.28 dV with
this strategy. When taken with the total annual estimated cost for this strategy of $97,108,704, we obtain a $/dV of improvement value of $431,058.

....
Ammonia on Unit 1 with Wet FGD on Unit 2
The third alternative examined was the option of installing a wet FGD system on Unit 2 with an ammonia system added to Unit 1 as described in the section on ammonia system option previously. The costs assumed for this option were based on the costs used for the ammonia option and the wet FGD option for Unit 2 combined. These individual options have been discussed at greater length previously. Table 10 shows the CALPUFF inputs used to model the visibility impacts for this option.
The results of this portion of the analysis shown in Table 11 indicate that 295 days with
dV impacts greater than the 0.5 dV goal are eliminated over the three year period by this alternative and the peak impact is reduced from 2.13 dV to 1.13 dV. Over the days that had dV impacts greater than the goal of 0.5 dV, there is a reduction of 260.41 dV with this strategy. When taken with the total annual estimated cost for this strategy of
$99,465,904, we obtain a $/dV of improvement value of $381,959.

OTHER ECONOMIC AND ENVIRONMENTAL CONSIDERATIONS
In performing the analyses discussed in the previous sections, the cost of disposal of any additional particulate material was not considered. For the case of injecting ammonia it was assumed that the existing wet fly ash system would be utilized along with disposal in the existing Horseford Hollow fly ash impoundment. While the addition of more material to the existing waste stream would reduce the time until a new ash disposal facility would be needed to some extent, this was not considered in this analysis for the primary particulate case or any of the secondary pollutant cases. The addition of a wet FGD system would result in the immediate need to develop a new much larger waste disposal site prior to the start of operation of the FGD system at a cost of several million dollars along with the additional cost to convert one or both units ESP(s) to a dry fly ash handling system. Depending on the exact site and size of the landfill, it may be necessary to develop an additional site or sites prior to the ultimate retirement of Unit 2.  It should be noted that there may be a possibility of generating wall board grade gypsum with a wet FGD system and then selling the synthetic gypsum to a wall board manufacturer. However, there are several issues that suggest that this option may not be a viable option for Big Sandy. First and foremost, AEP is currently unable to market all of the gypsum being generated thus far by its current FGD retrofit program, which will cover five units by early 2008. All of the units selling their gypsum are also located on the Ohio or Kanawha Rivers, making barge shipment a reasonable alternative to reach the wall board manufacturers. Big Sandy, on the other hand is not located on a navigable waterway, so any gypsum would need to be shipped either by truck or rail to reach either a wall board factory or a navigable waterway. This additional shipping would increase the shipping cost to a point that Big Sandy would not likely be a competitive supplier due solely to the shipping cost.  From this perspective, any of the options that add FGD to the plant would have a high cost to develop a new FGD waste landfill and large waste disposal foot print that is not present or currently anticipated with the other options.

BART CONCLUSION
Based on the data examined, the costs, and the other environmental impacts considered, it is concluded that BART for Big Sandy Plant should focus on the reduction of the secondary inorganic pollutants. The most cost effective and environmentally benign means to achieve this solution is the injection of ammonia on both units along with the required overhaul to the Unit 1 electrostatic precipitator to reduce up to 10 ppm of the inorganic condensable portion of the flue gas.

KY SIP Appendix L11, Pages 118- 155:

EON.US Mill Creek Station
In Step 1 of the BART analysis, all available retrofit control technologies that have a practical potential for application at Mill Creek Station were identified. These technologies are considered as available technologies. The technology considered could be a change in plant operation method, addition/modification of emissions control system, or a combination of these options for control of a pollutant. Control technologies for the two major pollutants, SO3 and PM, and a description of the technology are presented in this section. Information on the working principle, retrofit considerations, advantages, and disadvantages of the technology are provided in the descriptions contained in Sections 3.1 through 3.2.
3.1 SO3 Control Technologies
SO3 control technologies that were identified as available for retrofit at Mill Creek Station are listed below. A short description of each technology is included in the following subsections:
:: Lime-based semi-dry FGD.
:: Wet FGD.
:: Sorbent injection.
:: WESP.
:: Low SO2 to SO3 oxidation SCR catalyst.
:: Coal additives.
....
3.2 Particulate Matter Control Technologies
PM control technologies that were identified as available for retrofit at Mill Creek station are listed below. Only post combustion control is available for PM control. A short summary of each technology is included in the following subsections:
:: Dry ESP enhancements.
:: PJFF.
:: Compact hybrid particulate collector.
:: Max-9 electrostatic fabric filter.
....
4.1 Technically Infeasible SO3 Control Technology
4.1.1 Lime-Based Semi-Dry FGD
All four units at Mill Creek Station are already equipped with wet FGD systems for SO2 removal. It should be noted that the wet FGD system has a higher SO2 removal capability than the semi-dry FGD, since Mill Creek Station combusts high-sulfur bituminous coal. It has been established in Step 1 that the SO3 removal achieved in a semi-dry FGD system is a co-benefit of the SO2 removal technology installed. Based on these two points, the lime-based semi-dry FGD is not a technically feasible control technology for SO3, since it will result in a lower controlled level of SO2 at Mill Creek Station.
4.1.2 Wet FGD
Section 1.1 provided details of the current unit configuration for all four units at Mill Creek Station, which includes limestone-based wet FGD systems. Therefore, the wet FGD is technically feasible as a SO3 control technology and is already in place. SO3 reduction in the flue gas is approximately 25 percent through the wet FGD (based on information in the Sargent & Lundy [S&L] SO3 Mitigation Report, 3/29/2006). Also, in recent testing, it was determined that in Units 3 and 4; the SO3 reduction achieved through the wet FGD is approximately 50 percent. While this technology is technically feasible, it will not be subjected to further BART analysis steps since it is an existing system.
4.1.3 Low SO2 to SO3 Conversion SCR Catalyst
This technology is only available to Units 3 and 4 at Mill Creek Station, since both units are equipped with SCR systems for NOx reduction. Referencing the basis established in the S&L SO3 Mitigation Report issued on 3/29/2006, the SO2 to SO3 conversion rate for the catalyst currently installed in Mill Creek Units 3 and 4 is 1.40 percent and 1.20 percent, respectively. This performance characteristic indicates that the SCR catalyst currently installed is of the standard formula/composition. However, it is noted that the long-term catalyst management plan that is in place for Units 3 and 4 includes the replacement of existing catalyst layers in the SCR with new low SO2 to SO3 conversion catalyst compositions. The latest layer that was recently installed was of the low SO2 to SO3 conversion type. With the use of this new catalyst composition, it is anticipated that catalyst life will be reduced. The use of a low SO2 to SO3 conversion catalyst will not reduce the quantity of SO3 already in the flue gas from the combustion process; it will only reduce the addition of SO3 into the flue gas after passing through the SCR catalysts. Based on this reasoning, this option is not technically feasible for consideration in this BART analysis.
4.1.4 Coal Additives
The addition of alkaline material as a coal additive in the precombustion stage is not technically feasible for SO3 reduction at the Mill Creek Station because of potential impacts to boiler operations such as the slagging and fouling tendencies of ash, increase in LOI content in the ash, and changes in the economizer outlet flue gas temperature.

4.2 Technically Infeasible PM Control Technology
4.2.1 Dry ESP Enhancements
The existing cold-side ESPs at all four units are already demonstrating high removal efficiencies of 99 percent. The enhancements listed in Subsection 3.2.1 will not improve the removal efficiency of PM in the flue gas more than what is currently being achieved. Therefore, the dry ESP enhancements available are not technically feasible.
4.2.2 Compact Hybrid Particulate Collector
The COHPAC system is technically feasible but will be evaluated as a design variant of the PJFF, since both systems are identical in design and operational characteristics; the key difference is in the design A/C ratio, which determines the total amount of filtration area available. Therefore, the subsequent BART analysis steps performed for the PJFF will apply to this high A/C ratio fabric filter.
4.2.3 Max-9 Electrostatic Fabric Filter
The Max-9 electrostatic fabric filter has been recently installed commercially in a smaller-sized utility boiler (up to 80 MW). However, there are no current commercial installations in similarly sized units as the units at Mill Creek Station. Therefore, the
Max-9 electrostatic fabric filter will not be considered as technically feasible for further BART analysis steps.
4.3 Technical Feasibility Summary
After the completion of the screening process (Step 2 of the BART determination), the following technologies were identified as feasible control technologies in addition to those currently achieved for SO3 reduction:
:: Sorbent injection.
:: WESP.
The PM control technology identified that would have better emissions reduction than the existing cold-side ESPs available at Mill Creek Station is a PJFF.
....



....
Cost-Effectiveness Comparison
This section presents a brief summary of the evaluated cost-effectiveness for the control of SO3 and PM at all four Mill Creek Station units. For the two SO3 control technologies evaluated, sorbent injection cost-effectiveness ranges from 4,293 to 5,017 $ per ton for sorbent injection and 20,385 to 44,766 $ per ton for WESP. The large range in the WESP cost-effectiveness is due to the costs of operating the new induced draft (ID) fans and higher powered ID fans if WESP is implemented on Units 1 and 2. Another observation is the incremental cost-effectiveness when comparing the sorbent injection to the WESP, which ranges from 60,586 to 105,316 $ per additional ton of SO3 removed. For the lone PM control technology evaluated, the cost-effectiveness to install PJFF ranges from $20,950 to $52,190 per ton. Again, the large range in the PJFF cost effectiveness is due to the same reasons as that for WESP discussed above.


....


....
After completing all five steps of the BART analysis, a BART control scenario for all four units at Mill Creek Station was selected and evaluated for the visibility improvement at affected Federal Class I areas. The visibility improvement modeling is summarized in Section 7.0 of this report.

The recommended BART control scenario was based on the evaluated control technologies to meet the prescribed emission limits performed as detailed in Sections 3.0 to 6.0 of this report. The recommended BART control scenario for the units at Mill Creek Station consists of sorbent injection for the reduction of SO3 emission in the flue gas for Units 3 and 4 only. The control scenario also includes utilizing the existing ESP to control PM emissions.
The recommended control scenario is the most cost-effective solution that will allow Mill Creek Station to have a visibility improvement that is greater than a 0.5 dv change.
The visibility improvement modeled for the BART control scenario, as described in Section 7.0, indicates a maximum visibility improvement of 1.180 dv when the sorbent injection technology is applied to each unit (Scenario SO3 - 1) at Mill Creek Station and
0.825 dv in the recommended scenario when sorbent injection is applied to only Units 3 and 4. The difference in the maximum visibility improvement is 0.355 dv.
Based on the visibility improvement modeled and the total annual cost evaluated in the impact analysis stage (Step 4), the cost-effectiveness for visibility improvement (which was defined as annual cost per improvement in visibility [$/dv]) was determined.
The total annual cost for the implementation of the recommended control technologies is approximately 4.2 million $/yr, as shown in Table 7-8. Based on the maximum modeled visibility improvements at the affected Federal Class I of 0.825 dv, the visibility improvement cost-effectiveness by using sorbent injection at Units 3 and 4 is 5.1 million
$/dv. This value is approximately the same if sorbent injection was applied to all four units at Mill Creek Station (Scenario SO3 - 1), as detailed in Table 7-8. However, sorbent injection at all four units would mean an additional total capital investment of $8.8 million. Additionally, the incremental cost between the recommended scenario of sorbent injection at Units 3 and 4 versus sorbent injection at all four units is 5,000 $/additional ton of SO3 removed, and 5.0 million $/additional dv improvement.

Based on the relatively similar visibility cost-effectiveness and the incremental cost effectiveness, sorbent injection at Units 3 and 4 is the recommended BART technology applicable at Mill Creek Station to obtain visibility improvement.

Updated May 28, 2010 submittal:
Louisville Gas and Electric (LG&E) and the Kentucky Energy and Environment Cabinet entered into an Agreed Order on October 20, 2008 (File No. DAQ-29458-039). As specified in paragraph I of the order, LG&E is herein providing the Kentucky Division for Air Quality with information identifying emission rates utilized for the modeling conducted in conjunction with LG&E's September 24,2007 submittal of Mill Creek Station's Best Available Retrofit Technology analysis. Additionally, this correspondence provides written explanation of the infeasibility of incorporating a sulfur trioxide (SO3) emission limitation of 0.015 lb/mmBtu into the Mill Creek Station Title V permit.

As described in the submitted analysis, for the purpose of SO3 emission determination, all H2SO4 particles determined from the CALPUFF modeling were assumed to be SO3 emissions. As shown in "Table7-l CALPUFF Modeling Parameters" of the report, the baseline 2zSO4 primary particle information are in units of grams per second (g/s), 25.39 g/s and 27 .67 g/s for Mill Creek Units 3 and 4, respectively. When input into the CALPUFF model, they were converted to pounds per hour (lb/hr) values of 20L5 lb/hr and 2l9.6lb/hr for Mill Creek Units 3 and 4, respectively. In determining the effectiveness of sorbent injection technology, a stack exit concentration of five (5) parts per million (ppm) SO3 was used to evaluate the technology's effect on visibility impacts. For input into the CALPUFF model, five (5) ppm SO3equates to H2SO4 emission rates of 64.3lb/hr and 76.5lb/hr for Mill Creek Units 3 and 4, respectively.  These values are displayed in Appendix D on pages D-4 and D-5 of the previously submitted analysis. The lb/hr values were the primary model input values utilized in our CALPUFF modeling. As such, these are the values that are appropriate for incorporation into Mill Creek Station's Title V permit.

For illustrative purposes only, the lb/hr values were converted to lb/mmBtu values in Table 7-4 of the September 24, 2007 submittal. While we apologize for the confusion which resulted from including the value in our submittal, we wish to clarify that we never intended to suggest that it is appropriate for inclusion as an emission limit in our permit.

The 0.015 lb/mmBtu value provided in Table 7-4 is based on the design heat input (i.e. maximum) value for the two units. However, in the course of normal utility operations, the units operate at a wide range of heat inputs, some of which are substantially lower than the design heat input value. Although the SO: mitigation system proposed by LG&E can meet the specified lb/hr emissions values necessary to achieve BART, it is not capable of controlling H2SO4 emission to 0.015 lb/mmBtu at all heat input levels. Indeed, the measurement accuracy of H2SO4 emissions is questionable when operating at the level of lower heat inputs that would be encountered in the course of normal operations.

Consequently, adding a 0.015 lb/mmBtu limit to the permit would place severe constraints on our operational flexibility which could effectively preclude us from operating the unit at certain heat inputs. This would have major financial implications for us. In conclusion, it is not technically feasible for our proposed SO3 mitigation system to meet the 0.015 lb/mmBtu H2SO4 target at all operating loads and we do not believe that the value reflects BART for our units.

Therefore, if the Division desires to incorporate specific BART emission limits into the Title V permit for Mill Creek, the appropriate values would be 64.3 and 7 6.5 lb/hr H2SO4 for Units 3 and 4 respectively.

EPA Findings:  Kentucky has adequately addressed this provision. The Commonwealth adequately documented the technical analysis to assess the need and implementation of BART controls as applicable for those non-CAIR EGU emission units of PM, SO2 and NOx precursor emissions.  This involved identification of existing controls on applicable emission units, potential control options for applicable emission precursors, and the successful application of the five statutory factors to determine the feasibility of controls that constituted BART.  

EPA finds these assessments and conclusions for BART acceptable for these facilities because the analyses were conducted consistent with EPA's BART Guidelines and EPA's Air Pollution Control Cost Manual, and the conclusions reflect a reasonable application of EPA's guidance to these sources. 
 
                                                                                                                                                                            

III. Evaluation of Regional Haze Submittal 
      D. Best Available Retrofit Technology (BART) Requirements 
D.5. Enforceability of BART: For each BART-eligible source in the State subject to BART, the SIP should include emission limitations representing BART and schedules for compliance with BART.  For each BART-eligible source in the State subject to BART, the SIP must include a requirement that each source subject to BART be required to install and operate BART as expeditiously as practicable, but in no event later than five years after approval of the implementation plan revision and a requirement that each source subject to BART maintains the control equipment required by this subpart and establishes procedures to ensure such equipment is properly operated and maintained. 
      
State Regional Haze Submittal 
Appendix N, Page N-8 to N-9:

Appendix N - Public Hearing Notice and Summary of Comments Received and Cabinet Responses 
1. Comment: In the prehearing package, the proposed BART controls for five facilities are summarized in Table 7.5.3-1 in Section 7.5.3, p.62-64 of the SIP narrative, and in Table 9.1 of SIP Appendix L on pages 24-26. The BART limits and associated compliance schedules are required to be in the SIP per 40 CFR 51.308(e): "The State must submit an implementation plan containing emission limitations representing BART and schedules for compliance with BART . . .". Therefore, the SIP must include the specific emission limits (and supporting information for those limits) and compliance schedules for BART for the two facilities not currently subject to federal consent decrees (CDs). These limits must be in a form that can be enforced by both the Commonwealth and EPA.
For the three facilities installing controls for BART to meet federally enforceable emission limits contained in CDs, the Commonwealth may rely on the federal CDs to establish emission limits that are federally enforceable. The BART limits established pursuant to federal CDs, and the conditions necessary to support those limits, do not need to be incorporated into the SIP, unless any changes are made to the limits. Emission limits from the CDs must be placed in each facility's title V permit prior to the expiration date of its CD.
Richard A. Schutt, USEPA
Response: The Cabinet concurs and the applicable BART emission limitations and compliance schedules have been more clearly identified in the SIP and BART emission limitations will be added to the source Title V permit as appropriate or on renewal. Also, in reviewing the applicable consent decrees in response to the USEPA's comments, it was discovered that the East Kentucky Power Cooperative (EKPC) consent decree summarized in the SIP on page 42 was improper and as a result the Cabinet has provided the proper EKPC consent decree summary for the EKPC 7/2/2007 consent decree.

3. Comment: Legal Authority: In the final SIP submittal, please include evidence as required per Appendix V of 40 CFR part 51 that the Commonwealth has the necessary legal authority under Kentucky law to adopt and implement the Regional Haze implementation plan, including the BART limits.
Richard A. Schutt, USEPA
Response: The Cabinet acknowledges this comment and has provided language in the preface on page ii of the SIP narrative to indicate that the Cabinet, per KRS 224.10-
100(5), has the statutory legal authority to adopt and implement the Regional Haze State
Implementation Plan (SIP), including the BART emission limits.
 
Pages 65-66 as modified by May 28, 2010 submittal:




KY SIP, Page ii:

Preface: This document contains summaries of the technical analyses that will be used by the Kentucky Division for Air Quality (KYDAQ) to support the regional haze State Implementation Plan (SIP) pursuant to §§107(d)(3)(D) and (E) of the Clean Air Act, as amended. KRS 224.10-100(5) provides the Kentucky Energy and Environment Cabinet (formerly the Environmental and Public Protection Cabinet) with the statutory authority to adopt and implement its regional haze SIP. A link to KRS 224.10-100 is as follows: 
http://www.lrc.ky.gov/KRS/224-10/100.PDF.

EPA Findings: Kentucky has adequately addressed this provision. Rather than adopt a rule, Kentucky has chosen to directly adopt BART emission limits in its SIP for those sources not addressed by a federally enforceable consent decree.  These limits will become effective upon the EPA's final approval of the submittal.  These limits will subsequently be added to the appropriate Part 70 permits..

III. Evaluation of Regional Haze Submittal 
      E. Coordination of Regional Haze and RAVI Requirements
The State must revise its plan to provide for review and revision of a coordinated long-term strategy for addressing reasonably attributable and regional haze visibility impairment.
   
State Regional Haze Submittal
Kentucky did not explicitly address this requirement regarding a coordinated RAVI and regional haze long-term strategy.  However, KYDAQ has included provisions that updated the elements subject to the RAVI long-term strategy requirements (including monitoring and new source review (NSR)):

KY SIP Narrative, Pages 86-88:

Monitoring Strategy

In addition to the IMPROVE measurements, some ongoing long-term limited monitoring supported by Federal Land Managers provides additional insight into progress toward regional haze goals. Kentucky benefits from the data from these measurements, but is not responsible for the funding decisions to maintain these measurements into the future. Such measurements include web cameras operated by the National Park Service in Mammoth Cave National Park. 

KYDAQ and the local air agencies in the State operate a comprehensive PM2.5 network of the filter based Federal reference method monitors, continuous mass monitors (TEOMs), and filter based speciated monitors. A map of the various locations around the State is included in Figure 9.0-1. These PM2.5 measurements help the KYDAQ characterize air pollution levels in areas across the state, and therefore aid in the analysis of visibility improvement in and near the Class I areas. 

The IMPROVE measurements are central to Kentucky's regional haze monitoring strategy, and it is difficult to visualize how the objectives listed above could be met without the monitoring provided by IMPROVE. Any reduction in the scope of the IMPROVE network in Kentucky would jeopardize the State's ability to demonstrate reasonable progress toward visibility improvement in some of its Class I areas. In particular, Kentucky's regional haze strategy relies on emission reductions that will result from the CAIR, which occur on different time scales and will most likely not be spatially uniform. Monitoring at every Class I area is important to document the different air quality responses to the emissions reductions. 
Because each of the current IMPROVE monitor in Mammoth Cave represents a different airshed, reduction of the IMPROVE network by shutting down one of these monitoring sites impedes tracking progress at reducing haze at the affected Class I area. In the event this occurs, Kentucky, in consultation with the USEPA and relevant Federal Land Managers, will develop an alternative approach for meeting the tracking goal, perhaps by seeking contingency funding to carry out limited monitoring or by relying on data from nearby urban monitoring sites to demonstrate trends in speciated PM2.5 mass. 
Data produced by the IMPROVE monitoring network will be used nearly continuously for preparing the 5-year progress reports and the 10-year SIP revisions, each of which relies on analysis of the preceding five years of data. Consequently, the monitoring data from the IMPROVE sites needs to be readily accessible and to be kept up to date. Presumably, IMPROVE will continue to process information from its own measurements at about the same pace and with the same attention to quality as it has shown in the recent past. The VIEWS web site has been maintained by VISTAS and the other Regional Planning Organizations to provide ready access to the IMPROVE data and data analysis tools. KYDAQ is encouraging VISTAS and the other RPOs to maintain VIEWS or a similar data management system to facilitate analysis of the IMPROVE data.

KY SIP Narrative, Page 90:

Ten Year Plan and 5 Year Plan Update
The requirements listed in 51.308(g) include the following: 
   1. Description of the status of implementation; 
   2. Summary of emission reductions achieved thus far, including especially the status of implementation of the CAIR compliance plans for EGUs compared to the control assumed in the modeling; 
   3. Assessment of changes in visibility conditions at each Class I area (current vs. baseline), expressed as 5-year averages of annual values for 20 percent best and worst days; 
   4. Analysis of emission changes over the 5-year period, identified by source or activity; 
   5. Analysis of any significant changes in or out of the State which have impeded progress; 
   6. Assessment of the sufficiency of the implementation plan to meet Reasonable progress goals (RPGs); and 
   7. Review and any modifications to our visibility monitoring plan. 

All requirements listed in 51.308(g) shall be addressed in the SIP revision for reasonable progress. In particular, the KYDAQ recognizes that the 2018 projections of EGU controls from the IPM runs represent one solution to how the CAIR requirements will be met. By the time of the first periodic report, the KYDAQ anticipates that the actual compliance strategy for the various utility companies will be much more defined. An assessment of those actual compliance plans will be done for the first periodic report. 

NSR/PSD
The KYDAQ believes that its New Source Review (NSR) regulation for nonattainment areas as well as its Prevention of Significant Deterioration (PSD) regulation for attainment areas will address emissions from new sources that may locate near a Class I area or increased emissions from major modifications to existing sources. In addition to the KYDAQ regulations that would govern these sources, consultation with the FLMs is also required for sources that are subject to KYDAQ's NSR/PSD regulations. 

Consultation
KYDAQ also plans for continued consultation with the FLMs throughout the implementation process, including discussion of the implementation process and the most recent IMPROVE monitoring and VIEWS data. Consultation between KYDAQ and the FLMs will include early involvement of FLMs in the periodic review process and FLMs will receive copies of the revised regional haze SIP for comment prior to finalization.

EPA Findings: Under the RAVI regulations, a State's RAVI SIP must address any integral vistas identified by the FLMs pursuant to 40 CFR 51.304.  An integral vista is defined in 40 CFR 51.301 as a view perceived from within the mandatory Class I Federal area of a specific landmark or panorama located outside the boundary of the mandatory Class I Federal area.  Visibility in any mandatory Class I Federal area includes any integral vista associated with that area.  Kentucky does not have any integral vistas identified by the FLMs, none of its Class I areas are experiencing RAVI, nor are any of its sources affected by the RAVI provisions.  Thus, the June 28, 2008, Kentucky regional haze SIP submittal does not explicitly address the two provisions regarding coordination of regional haze and RAVI long-term strategies and monitoring requirements.  However, the Commonwealth's previous commitment to address RAVI should the FLM certify visibility impairment from an individual source remains unchanged as do the previously approved PSD/NSR visibility provisions (July 12, 1988, at 53 FR 26256; September 1, 1989, at 54 FR 36308; and July 11, 2006, at 71 FR 38990).  EPA finds that this regional haze submittal adequately supplements and augments Kentucky's RAVI visibility provisions to address regional haze, updating the monitoring and LTS provisions as summarized below in this section.
      
	In the June 25, 2008, submittal, KYDAQ updated its monitoring program commitments and developed a LTS to address regional haze.  KYDAQ also commits to review and revise its regional haze implementation plan and submit a plan revision to EPA by July 31, 2018, and every ten years thereafter, as required by the RHR.  In accordance with the requirements listed in 51.308(g) of the federal rule for regional haze, KYDAQ commits to submitting a report on reasonable progress to EPA every five years following the initial submittal of the SIP, as required by the RHR.  The report will be in the form of a SIP revision.  The reasonable progress report will evaluate the progress made towards the reasonable progress goal for each mandatory Class I area located within Kentucky and in each mandatory Class I area located outside Kentucky which may be affected by emissions from within Kentucky.  In addition, consistent with the monitoring provisions for RAVI and regional haze under 40 CFR 51.305 and 40 CFR 51.308(d)(4), Kentucky will rely on the IMPROVE network for complying with the regional haze monitoring requirement in the RHR.  Finally, the Kentucky NSR rules continue to provide a framework for review and coordination with the FLM on new sources which may have an adverse impact on visibility in either form (i.e., RAVI and/or regional haze) in any Federal Class I area.

Although Kentucky submitted and EPA approved visibility rules addressing nonattainment NSR (July 11, 2006, at 71 FR 38990), the EPA approvals of these rules neglected to remove the federally promulgated provisions in 40 CFR 52.936.  EPA will correct this omission in a separate, subsequent rulemaking.




III. Evaluation of Regional Haze Submittal 
      F. Monitoring Strategy and Other Implementation Requirements

Monitoring Strategy:
The plan must include a monitoring strategy for measuring, characterizing, and reporting of regional haze visibility impairment that is representative of all mandatory Class I Federal areas within the State and/or summarize monitoring strategy of States with affected mandatory Class I Federal areas. Compliance with this requirement may be met through participation in the Interagency Monitoring of Protected Visual Environments (IMPROVE) network. 

Other Requirements:
The implementation plan must also provide for the following:
   A. The establishment of any additional monitoring sites or equipment needed to assess whether reasonable progress goals to address regional haze for all mandatory Class I Federal areas within the State are being achieved.
   B. Procedures by which monitoring data and other information are used in determining the contribution of emissions from within the State to regional haze visibility impairment at mandatory Class I Federal areas both within and outside the State.
   C. The reporting of all visibility monitoring data at least annually for each mandatory Class I Federal area in the State. To the extent possible, the State should report monitoring data for visibility electronically and use of the Visibility Information Exchange Web System (VIEWS) is encouraged.
   D. Other elements, including reporting, recordkeeping, and other measures, necessary to assess and report on visibility.
 
State Regional Haze Submittal
KY SIP Narrative, Pages 85-87:

The primary monitoring network for regional haze, both nationwide and in Kentucky, is the IMPROVE network. Given that IMPROVE monitoring data from 2000-2004 serve as the baseline for the regional haze program, the future regional haze monitoring strategy must necessarily be based on, or directly comparable to, IMPROVE. The IMPROVE measurements provide the only long-term record available for tracking visibility improvement or degradation and therefore Kentucky intends to rely on the IMPROVE network for complying with the regional haze monitoring requirement in the Regional Haze Rule. 
There is currently one IMPROVE site in Kentucky's Mammoth Cave National Park as provided in Table 9.0-1. 
Class I Area 
                           IMPROVE Site Designation 
Mammoth Cave National Park 
                                  MACA1 (KY) 
Table 9.0-1. Kentucky Class I Area and Representative IMPROVE Monitor. 

   In addition to the IMPROVE measurements, some ongoing long-term limited monitoring supported by Federal Land Managers provides additional insight into progress toward regional haze goals. Kentucky benefits from the data from these measurements, but is not responsible for the funding decisions to maintain these measurements into the future. Such measurements include web cameras operated by the National Park Service in Mammoth Cave National Park. 

KYDAQ and the local air agencies in the State operate a comprehensive PM2.5 network of the filter based Federal reference method monitors, continuous mass monitors (TEOMs), and filter based speciated monitors. A map of the various locations around the State is included in Figure 9.0-1. These PM2.5 measurements help the KYDAQ characterize air pollution levels in areas across the state, and therefore aid in the analysis of visibility improvement in and near the Class I areas. 

Figure 9.0-1 PM2.5 Monitoring Network in Kentucky.

The IMPROVE measurements are central to Kentucky's regional haze monitoring strategy, and it is difficult to visualize how the objectives listed above could be met without the monitoring provided by IMPROVE. Any reduction in the scope of the IMPROVE network in Kentucky would jeopardize the State's ability to demonstrate reasonable progress toward visibility improvement in some of its Class I areas. In particular, Kentucky's regional haze strategy relies on emission reductions that will result from the CAIR, which occur on different time scales and will most likely not be spatially uniform. Monitoring at every Class I area is important to document the different air quality responses to the emissions reductions. 
Because each of the current IMPROVE monitor in Mammoth Cave represents a different airshed, reduction of the IMPROVE network by shutting down one of these monitoring sites impedes tracking progress at reducing haze at the affected Class I area. In the event this occurs, Kentucky, in consultation with the USEPA and relevant Federal Land Managers, will develop an alternative approach for meeting the tracking goal, perhaps by seeking contingency funding to carry out limited monitoring or by relying on data from nearby urban monitoring sites to demonstrate trends in speciated PM2.5 mass. 
Data produced by the IMPROVE monitoring network will be used nearly continuously for preparing the 5-year progress reports and the 10-year SIP revisions, each of which relies on 
analysis of the preceding five years of data. Consequently, the monitoring data from the IMPROVE sites needs to be readily accessible and to be kept up to date. Presumably, IMPROVE will continue to process information from its own measurements at about the same pace and with the same attention to quality as it has shown in the recent past. The VIEWS web site has been maintained by VISTAS and the other Regional Planning Organizations to provide ready access to the IMPROVE data and data analysis tools. KYDAQ is encouraging VISTAS and the other RPOs to maintain VIEWS or a similar data management system to facilitate analysis of the IMPROVE data.

EPA Findings: Kentucky has adequately addressed this provision.

III. Evaluation of Regional Haze Submittal 
      G. Emission Inventory (40 CFR 51.308(d)(4)(v) (See also TSD Section III.C.))
      
The SIP must include a statewide inventory of emissions of pollutants that are reasonably anticipated to cause or contribute to visibility impairment in any mandatory Class I Federal area within the State. The inventory must include emissions for a baseline year, emissions for the most recent year for which data are available, and estimates of future projected emissions, and a commitment to update the inventory periodically.

State Regional Haze Submittal
KY SIP Narrative, Page 16:

The Regional Haze Rule at 51.308(d) (4) (v) requires a statewide emissions inventory of pollutants that are reasonably anticipated to cause or contribute to visibility impairment in any mandatory Class I area. An inventory was developed for the baseline year 2002 and projected to 2009 and 2018. The pollutants inventoried include volatile organic compounds, nitrogen oxides, fine particulate (PM2.5), coarse particulate (PM10), ammonia and sulfur dioxide. The baseline emissions inventory for 2002 was developed for Kentucky following the methods described in Appendix D. 
There are five different emission inventory source classifications: stationary point and area sources, off-road and on-road mobile sources, and biogenic sources. Stationary point sources are those sources that emit greater than a specified tonnage per year, with data provided at the facility level. Electric generating utilities and industrial sources are the major categories for stationary point sources. Stationary area sources are those sources whose individual emissions are relatively small, but due to the large number of these sources, the collective emissions from the source category could be significant (i.e., dry cleaners, service stations, agricultural sources, fire emissions, etc.). These types of emissions are estimated on a countywide level. Non-road (or off-road) mobile sources are equipment that can move but do not use the roadways, i.e., lawn mowers, construction equipment, railroad locomotives, aircraft, etc. The emissions from these sources, like stationary area sources, are estimated on a countywide level. On-road mobile sources are automobiles, trucks, and motorcycles that use the roadway system. The emissions from these sources are estimated by vehicle type and road type and are summed to the countywide level. Biogenic sources are the natural sources like trees, crops, grasses and natural decay of plants. The emissions from these sources are estimated on a countywide level. 
In addition to the various source classifications, there are also various types of emission inventories. The first is the actual base year inventory. This inventory is the base year emissions that correspond to the meteorological data used, which for this modeling effort is data from 2002. These emissions are used for evaluating the air quality model performance. 
The second type of inventory is the typical base year inventory. This inventory is similar to the actual base year inventory, except that for sources whose emissions change significantly from year to year, a more typical emission value is used. In this modeling effort, typical emissions were developed for the electric generating units (EGUs) and the wildland fire emissions. The air quality modeling runs using the typical base year inventory provide results which are then used to calculate relative reduction factors for future years. These relative reduction factors for future years are then used to demonstrate reasonable progress toward visibility goals.


KY SIP Narrative, Page 20:

Table 4.1.6-1 is a summary of the 2002 baseline emission inventory for Kentucky. The complete inventory and discussion of the methodology is contained in Appendix D. The emissions summaries for other VISTAS states can also be found in Appendix D.

Table 4.1.6-1 2002 Emissions Inventory Summary for Kentucky (tons per year).


KY SIP Narrative, Page 43:

The inventories for 2009 and 2018 account for post-2002 emission reductions from promulgated and proposed federal, state, local, and site-specific control programs as of July 1, 2004. In general, emissions inventories were developed for 2009 and 2018 using current control information in Kentucky. 
For EGUs, IPM results were adjusted based on state and local air agencies knowledge of planned emission controls at specific EGUs. These updates are documented in the MACTEC emissions inventory report "Documentation of the 2002 Base Year and 2009 and 2018 Projection Year Emission Inventories for VISTAS" dated February 2007 (Appendix D). 
For non-EGUs, VISTAS used recently updated growth and control data consistent with the data used in the USEPA's CAIR analyses (Clean Air Interstate Rule Emissions Inventory Technical Support Document, March 2005) supplemented by state and local air agencies data and updated forecasts from the Department of Energy (DOE). 
Area source controls were estimated using known state level Stage I controls on gasoline dispensing facilities and open burning estimates, as well as controls used to project emissions for the USEPA's Heavy Duty Diesel rulemaking and for the CAIR rulemaking. 
Mobile source controls included local controls underlying the 2002 baseline inventory (vehicle emission inspection, Stage II vapor recovery, anti-tampering, etc.) with changes based on specific State input. The future year inventories were developed by running the MOBILE6.2 model for each year modeled. The future year emissions for the off-road mobile sources included in the USEPA NONROAD model were estimated by running the model for each future year. For the other off-road mobile source categories control data and projections for 1996, 2010, 2015, and 2020 were obtained from the USEPA's CAIR Technical Support Document, and straight-line projections were used to estimate 2009 and 2018 levels.
KY SIP Narrative, Page 45:


KY SIP Narrative, Page 90:

The requirements listed in 51.308(g) include the following: 
   1. Description of the status of implementation; 
   2. Summary of emission reductions achieved thus far, including especially the status of implementation of the CAIR compliance plans for EGUs compared to the control assumed in the modeling; 
   3. Assessment of changes in visibility conditions at each Class I area (current vs. baseline), expressed as 5-year averages of annual values for 20 percent best and worst days; 
   4. Analysis of emission changes over the 5-year period, identified by source or activity; 
   5. Analysis of any significant changes in or out of the State which have impeded progress; 
   6. Assessment of the sufficiency of the implementation plan to meet Reasonable progress goals (RPGs); and 
   7. Review and any modifications to our visibility monitoring plan. 

All requirements listed in 51.308(g) shall be addressed in the SIP revision for reasonable progress. In particular, the KYDAQ recognizes that the 2018 projections of EGU controls from the IPM runs represent one solution to how the CAIR requirements will be met. By the time of the first periodic report, the KYDAQ anticipates that the actual compliance strategy for the various utility companies will be much more defined. An assessment of those actual compliance plans will be done for the first periodic report. 

EPA Findings: Kentucky has adequately addressed this provision.
IV. Conclusion

Except for those aspects of the regional haze submittal that rely on CAIR, Kentucky's regional haze SIP, submitted on June 25, 2008, and as amended on May 28, 2010, meets all of the requirements of section 110 of the CAA and meets the regional haze requirements, as set forth in sections 169A and 169B of the Act and in 40 CFR 51.300-308.  The SIP meets these requirements by, among other things, requiring the establishment of reasonable progress goals for each of its Class I areas, determining BART, and submitting a long-term strategy that addresses regional haze visibility impairment for its Class I areas and for those Class I areas in other states that may be impacted by emissions from within Kentucky.  Thus, EPA is proposing a limited approval and limited disapproval of the Kentucky regional haze SIP in the associated Federal Register action.
.




 

                                       

