                                       
                          Technical Support Document 
                                       
               Missouri Regional Haze State Implementation Plan 
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                 Prepared by:
                                       
                                  Michael Jay
                            Environmental Scientist
                                       
                                       
                                       
                 Unites States Environmental Protection Agency
                                   Region 7


                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                 January 2012 
                                       
              Missouri Regional Haze Technical Support Document 
                               Table of Contents

I. Background
   
II. Summary of EPA's Action

III. Evaluation of Regional Haze Submittal
   A.  Identification of Class I Areas
   B. Reasonable Progress Goal Development
      1. Calculations of Natural and Baseline Visibility Conditions
      2. Uniform Rate of Progress
      3. Consultation
      4. Analysis of Statutory Factors
            a. Independent Missouri Analysis
            b. CENRAP Cost Analysis
            c. Four Factor Analysis by Minnesota Pollution Control Agency
            d. Cost and Time Necessary for Compliance Under CAIR
   C. Long-Term Strategy
      1. Technical Basis for Strategy
            a. State Submittal 
      2. Identification of Sources and Factors to be Considered 
            a. State Submittal
      3. Consultation
            a. State Submittal
   D. Best Available Retrofit Technology (BART)
 BART Eligible Sources
 State Submittal
 Sources Subject to BART
 State Submittal
 BART Determination and Statutory Factors
             Identification of Available Retrofit SO2 Control Technology
             Rank and Evaluation of Technical Feasibility for SO2 Control Options
             Evaluation of Visibility Impact of Feasible SO2 Control
             Identification of Available Retrofit NOx Control Technology 
             Rank and Evaluation of Technical Feasibility for NOx Control Options
             Evaluation of Visibility Impact of Feasible NOx Control
   A. 2002 and 2018 Emissions Inventory
      1.	State Submittal
   B. Monitoring Strategy and Other Implementation Requirements
1.	State Submittal 	
   C. Planning and Conclusion 

IV.	Conclusion



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)).  In 1999, EPA and affected states/tribes agreed to create five RPOs to facilitate interstate coordination on regional haze plan submittals and Tribal Implementation Plans (TIPs). Figure 1 below shows a map of all five RPOs. The RPO for the central United States is the Central Regional Air Planning Association (CENRAP).   The State of Missouri is a member of the Central Regional Air Planning Association (CENRAP).  Members of CENRAP include the following states: Arkansas, Iowa, Kansas, Louisiana, Minnesota, Missouri, Nebraska, Oklahoma, and Texas. The Missouri Department of Natural Resources chose to participate in CENRAP to develop the technical analyses needed to fulfill the requirements of the RHR.  

Figure 1- Geographical Area of Regional Planning Organizations 

II.   SUMMARY OF EPA'S ACTION

EPA is proposing a limited approval of a revision to the Missouri State Implementation Plan (SIP) submitted by the State of Missouri in August 2009 that addresses regional haze for the first implementation period.  This revision addresses the requirements of the Clean Air Act (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 this SIP revision to implement the regional haze requirements for Missouri on the basis that the revision, as a whole, strengthens the Missouri SIP.  

On December 30, 2011, EPA proposed a limited disapproval of this same SIP revision because of the deficiencies in the State's August 2009 regional haze SIP submittal arising from the remand by the U.S. Court of Appeals for the District of Columbia (D.C. Circuit) to EPA of the Clean Air Interstate Rule (CAIR).  76 Fed. Reg. 82219 (Dec. 30, 2011).  In that proposed rulemaking EPA is also proposing a Federal Implementation Plan (FIP) to replace reliance on the CAIR requirements in Missouri's SIP with reliance on the Cross-State Air Pollution Rule (Transport Rule) as an alternative to BART.  

III.  	EVALUATION OF REGIONAL HAZE SUBMITTAL
	
 	A.	Identification Of Class I Areas (40 CFR 51.308 (D)(1)(I))
		SIP Pages 12-13
      
The Plan must identify each mandatory Class I area located within the state and in each mandatory Class I  area located outside the state that may be affected by emissions from within the state.  Missouri identified the Class I areas within the state as follows:

Hercules Glades Wilderness Area 
Situated in southwest Missouri, Taney County, Hercules Glades Wilderness Area (Hercules Glades) is managed by the United States Department of Agriculture (USDA) Forest Service as part of the Mark Twain National Forest. The area includes 12,315 acres located in some of the most rugged hills of the Missouri Ozarks. 

Mingo National Wildlife Refuge
The Mingo National Wildlife Refuge (Mingo) is managed by the U.S. Fish and Wildlife Service. The refuge is situated in southeast Missouri, along the Mississippi Flyway. Only a portion of the refuge is a Class I area (7,730 acres of a total 21,676 acres). 

Additionally, Missouri noted that emissions from Missouri may also contribute to visibility impairment in other states' Class I areas, such as the Caney Creek and Upper Buffalo Wilderness Areas in Arkansas (Upper Buffalo).  

EPA Findings: Missouri has appropriately identified the Class I areas. 

      B.	 Reasonable Progress Goals (40 CFR § 51.308(d)(1)(B))
      	
		1.	Calculations of Natural and Baseline Visibility Conditions
            SIP Pages 25-29, 80-86 and Chapters I and III of  Technical Supplement

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.

For each mandatory Class I area located within the state, the state must also determine the average degree of visibility impairment (expressed in deciviews) for the most impaired and least impaired days (i.e. the 20 percent most impaired days and 20 percent least impaired days) for each year from 2000 to 2004.  Baseline conditions are the average of these annual values.  For each mandatory Class I Federal area located within the state, the plan needs to calculate the number of deciviews (dv) by which baseline conditions exceed natural visibility conditions for the most impaired and least impaired days. 

For natural visibility conditions for the best 20% days,  Missouri has deferred to the EPA default values listed in Table One of EPA's 2003 guidance document "Guidance for Estimating Natural Visibility Conditions Under the Regional Haze Rule."  These values were determined to be 3.59 dv for both Mingo and Hercules Glades.  These default values are listed in Table 12 of Appendix B of the SIP.  The guidance document can be found at the following web address: http://www.epa.gov/ttncaaa1/t1/memoranda/rh_envcurhr_gd.pdf.  

For natural visibility for the 20% worst days, Missouri is relying upon analysis provided by CENRAP.  Using the new IMPROVE equation as part of this analysis, Missouri has determined that natural visibility conditions for the Mingo Class I area is best represented by 12.40 dv for the 20 percent worst days. The Hercules Glades Wilderness Class I area is best represented by 11.30 dv for the 20 percent worst days. Table 1 below shows these conditions.  A more detailed description of the state's analysis is provided in section 5.3 and in Appendix F of the state's SIP. 

For baseline visibility for both the 20% worst and best days, Missouri has relied upon the CENRAP analysis.  These conditions were based on the monitoring results from the IMPROVE monitoring network results in Missouri's two Class I areas for the years 2000-2004, and were based on the new IMPROVE algorithm with appropriate justification provided in section 5.2 and Appendix G of the state's submittal.  Mingo Class I area has an established baseline visibility of 13.76 dv for the cleanest 20 percent of the sample days and 28.02 dv for the 20 percent worst visibility days.  The Hercules Glades Class I area has an established baseline visibility of 12.84 dv for the cleanest 20 percent of the sample days and 26.75 dv for the 20 percent worst visibility days. 
            
                                 Class I Area
           Natural Background Conditions for Missouri Class I Areas
     Baseline Visibility Conditions for Missouri Class I Areas, 2000-2004
                                       
                              20% Worst Days (dv)
                              20% Best Days (dv) 
                        Average for 20% Worst Days (dv)
                        Average for 20% Best Days (dv)
Mingo 
                                     12.40
                                     3.59
                                     28.02
                                     13.76
Hercules Glades 
                                     11.30
                                     3.59
                                     26.75
                                     12.84
	Table 1- Natural Background and Baseline Visibility Conditions for Missouri Class I Areas  

EPA Findings: Missouri has appropriately calculated the baseline area for the Class I areas within the State consistent with EPA's June 1, 2007 "Guidance for Setting Reasonable Progress Goals Under the Regional Haze Program" and has met the requirements of 40 CFR 51.308(d)(2).  

		2.	Uniform Rate of Progress
               SIP Pages 48-49, 85

For each mandatory Class I 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 achieve 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 achieve natural visibility conditions by 2064. The state must show that the reasonable progress goals (RPGs) 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 planning period. The RPGs 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.  

The Figures below show the Uniform Rate of Progress (URP) and modeled 2018 visibility projections. The 2018 projections were below the URP for both Missouri Class I areas. Missouri has determined that the modeled rate of visibility improvement by 2018 shown in Table 2 below is reasonable and hereby adopts it as the RPG for the listed Class I areas. Table 2 provides a Uniform Rate of Progress and Reasonable Progress for Class I areas in Missouri.

                                       
Class I Area
Baseline Conditions, 20% worst days
2018 URP
2018 Modeled Predictions
(goal)
2064 Natural Conditions
DV Improvement needed by 2018 assuming RPG
Progress annually to 2018 assuming RPG
Mingo
28.02
24.37
23.71
12.40
4.31
0.308
Hercules Glades
26.75
23.14
23.06
11.30
3.69
0.264
        Table 2- Uniform Rate of Progress and Reasonable Progress Goals for Class I areas in Missouri


      Figure 2- Uniform Rate of Reasonable Progress Glide Path for Hercules-Glades Worst 20% Days 



            
      Figure 3- Uniform Rate of Reasonable Progress Glide Path for Hercules-Glades Best 20 Days 


      Figure 4- Uniform Rate of Reasonable Progress Glide Path for Mingo Worst 20% Days 







      Figure 5- Uniform Rate of Reasonable Progress Glide Path for Mingo Best 20% Days 

      3.	Consultation
      		SIP Pages 85-86 and Chapter I of Technical Supplement

In developing each reasonable progress goal, the State must consult with those states which may reasonably be anticipated to cause or contribute to visibility impairment in the mandatory Class I area.  Missouri has consulted with FLMs and other states/tribes, which are reasonably anticipated to cause or contribute to visibility in Mingo and Hercules Glades.  The State includes a description of the consultation process in its SIP submittal (Appendix E, United States Central Class I Areas Consultation Plan, Missouri Department of Natural Resources, 2007).  In addition, meeting minutes are provided (Appendix U).

EPA Findings: Missouri has appropriately addressed the consultation requirements.

		4.	Analysis of Statutory Factors (40 CFR 51.308(d)(1)(A))
			SIP Pages 80-85 and Chapter I of Technical Supplement

In establishing a reasonable progress goals for any Class I area within the State, the State must consider the statutory factors of CAA §169A(g)(1), which are specified in 40 CFR 51.308(d)(1)(A) and are as follows:

   1)  Consider cost of compliance,
   2)  The time necessary for compliance,
   3)  The energy and non-air quality environmental impacts of compliance, and 
   4)  The remaining useful life of any potentially affected sources.

In addition, the State must demonstrate how these factors were taken into consideration in selecting the RPG.  Missouri demonstrates that these four factors were applied in determining control strategy options for all source categories including point sources, area sources, on-road mobile sources, off-road mobile sources, which are also included in the state's Long-Term Strategy analysis described in section III. C. of this document.  In that section, Missouri identifies the control measures necessary to achieve the RPGs.  In addition to these four factors, all other Clean Air Act related programs were evaluated to determine what effect these other programs have had or will have on existing and future sources, and if any other control strategies would be reasonable in terms of the four factors described above.  For most sources, the state determined that Clean Air Act programs such as NSR permitting, NSPS standards, MACT standards, on-road and off-road engine standards, federal emission trading programs, fuel standards, and various state rules were found to be reasonable, and for these sources no other reasonable measures were deemed appropriate based on the four factors.  

In its plan, Missouri has committed to continue analyzing this Regional Haze SIP to address any deficiencies discovered.  In addition, if other reasonable control strategies are identified for these sources that contribute to visibility impairment, beyond those implemented through this plan, the state has committed to incorporate such strategies into future SIP revisions to be considered along with the five-year progress reports.

To demonstrate that the State has met the RPG requirement described further below, Missouri primarily relies upon the following: 1) an independent analysis completed by Missouri of its current permitting process, current rule requirements for existing sources, and a statewide point sources emissions analysis; 2) a cost analysis by CENRAP; 3) a published report by the Minnesota Pollution Control Agency; and 4) a description of the cost-effectiveness and visibility impacts from the Clean Air Interstate Rule on Missouri's Class I areas.  

                        a.	Independent Missouri Analysis

In accordance with the Clean Air Act, Missouri has a fully implemented construction permitting program for new major sources and significant modifications of existing sources.  The construction permitting program is referred to as a New Source Review (NSR) permitting program.  In Missouri, Nonattainment NSR permitting occurs in the St. Louis Nonattainment Area.  The construction permitting program in an attainment area or unclassifiable area is referred to as a Prevention of Significant Deterioration (PSD) permitting program.  All areas of the state other than the St. Louis Nonattainment Area are either attainment or unclassifiable, and in those areas the PSD program applies.  One of the major components of the PSD program is the implementation of Best Available Control Technology (BACT) for new major sources or for significant modifications of existing major sources.  EPA has approved Missouri's PSD program.

When Missouri performs a BACT analyses for new sources, the State takes into account the same four factors that are required for developing control strategies under a Regional Haze State Implementation Plan.  In nearly all cases, a BACT analysis would require more stringent controls than a RACT or RACM analysis.  As all new stationary emission sources are required to obtain a construction permit prior to commencing construction, the State's construction permit rule ensures that no significant degradation to visibility in Class I areas will occur as a result of new sources coming online.  If a construction permit application models that degradation would occur to the visibility in a Class I area, then control requirements would be added to the construction permit prior to its issuance.  It is these control measures, established through modeling, that ensure no degradation will occur. 

For existing sources, the State has rules from current SIP requirements that cover many emission sources including but not limited to charcoal kilns, cement kilns, large stationary IC engines, small boilers, and other non-EGU categories.  To demonstrate the extent of these requirements, Missouri provides a list of their non-EGU sources that emit greater than 50 tons of NOx, SO2, or PM10 and their corresponding applicable rule requirements (see Section 1.2 of Technical Supplement, January 30, 2012).  In an effort to determine if additional control requirements would be necessary for these existing point sources the State conducted further analysis.  Out of these sources, the State removed facilities that had already undergone a review for BART analysis, or that were screened from the BART analysis because they modeled a negligible impact on visibility, or that had recently gone through RACT review.  Because the BART analysis is based, in part, on an assessment of many of the same factors that must be addressed in establishing the RPG, EPA believes Missouri's approach of removing those facilities from further analysis is appropriate.  In addition, Missouri removed from consideration  non-EGU sources of SO2 and NOx that were recently analyzed as part of the development of the attainment demonstration for the 2006 PM2.5 NAAQS to address Reasonable Available Control Techniques (RACT) and Reasonably Available Control Measures (RACM) for the St. Louis area.  Missouri explained and EPA agrees that, in the development of their RACT/RACM analysis, the statutory factors are inherently taken into consideration.  

Next, the total of the emissions for each pollutant from the remaining list of refined sources (see tables 2-4 of Technical Supplement) were totaled and compared against emissions from all Missouri point emissions for NOx, SO2, and PM10, showing that these emissions constitute 7.3%, 2.8%, and 19.8% of the pollutant totals, respectively.  Further, when compared against the total statewide anthropogenic emissions in 2002 these percentages are reduced to 2.8%, 2.4%, and 5.7%, respectively.  Noting that ammonium sulfate and ammonium nitrate make up the majority of PM speciation mass in Missouri's Class I areas, and assuming that these sources of NOx and SO2 emissions (precursors of ammonium sulfate and ammonium nitrate) are rounded up to 3%, and assuming that these emissions only impact the monitors in the Missouri Class I areas, the emissions from all of these sources combined would only account for an impact of approximately 0.47 deciview on the State's Class I areas.  The State acknowledged that without modeling these specific sources, a precise impact on the Class I areas cannot be determined, but EPA agrees that this approach, with the additional conservative assumption that these emissions would only impact Missouri's Class I areas, is acceptable.  As a result, Missouri reached the conclusion that eliminating these emissions would likely not be noticeably beneficial in either of Missouri's Class I areas and additional research and analysis of new control requirements from non-EGUs is not necessary.  It's important to note that Missouri did not include EGUs as part of the analysis because EGU emissions reductions are relied upon as part of its Long Term Strategy discussed in section III. C. of this document.

To further support this conclusion, Missouri evaluated the refined list of sources to determine if it is reasonable to assume a visibility impact from any of these sources for either of Missouri's Class I areas using the Q/d approach.  Any source with a Q/d value over 10 dv was researched to determine if reasonable controls were being implemented at applicable sources.  For those five identified sources with a Q/d value over 10, as further described in the State's January 30, 2012, technical supplement, Missouri determined that additional controls were not warranted for one of the following reasons: 1) recent permit revisions limit the pollutant of concern; 2) implementation of a consent agreement requiring the shutdown of emissions units coupled with operation limits on remaining units; 3) a recent BACT analysis was undertaken; or 4) cost effective controls were not available and the units are nearing the end of their useful life.

EPA finds that the Q/d approach undertaken by the State is appropriate and follows the guidance provided in the Federal Land Managers' Air Quality Related Values Work Group Phase I Report-Revised (2010). This document  states that "Experience with the FLAG 2000 recommendations in dealing with many new source review applications led the Agencies to believe that an initial screen that would exempt a source from air quality related values (AQRV) impact review based on its annual emissions (Q) and distance (D) from a Class I area may be appropriate in most situations." The following is an excerpt from the FLM report which describes the screening criteria:

As part of its Regional Haze Regulation, the EPA has introduced a screening criteria in its BART guidelines based on a source's annual emission strength and distance from a Class I area. The EPA stated that it would be reasonable to conclude that the following sources would not be considered to cause or contribute to visibility impairment:

--those located more than 50 km from any Class I area that emit less than 500 tons per year of NOx or SO2 (or combined NOx and SO2), and 
--those located more than 100 km from any Class I area that emit less than 1,000 tons per year of NOx or SO2 (or combined NOx and SO2).

In both cases, the annual emissions over distance factor equates to 10.

The Agencies have concluded that a similar approach has merit with respect to new source impacts at Class I areas, for air pollution sources with relatively steady emissions throughout each year. However, the Agencies are modifying the size criteria to also include Particulate Matter less than 10 microns in size (PM10) and sulfuric acid mist (H2SO4) emissions because those pollutants also impair visibility and contribute to other resource impacts. In addition, rather than the two-step BART test, the Agencies are using a fixed Q/D factor of 10 as a screening criteria for sources locating/located greater than 50 km from a Class I area. Furthermore, the Agencies are expanding the screening criteria to include all AQRVs, not just visibility. Therefore, the Agencies will consider a source locating greater than 50 km from a Class I area to have negligible impacts with respect to Class I AQRVs if its total SO2, NOx, PM10, and H2SO4 annual emissions (in tons per year, based on 24-hour maximum allowable emissions), divided by the distance (in km) from the Class I area (Q/D) is 10 or less. The Agencies would not request any further Class I AQRV impact analyses from such sources.

The report is accessible at the following web address:
http://www.nature.nps.gov/air/pubs/pdf/flag/FLAG_2010.pdf

                  b.	CENRAP Cost Analysis

In considering  the four factors, Missouri has also relied upon a cost analysis provided by CENRAP that was performed by Alpine Geophysics that included analysis for the Mingo and Hercules Glades areas of Missouri.  Alpine primarily looked at controls on EGUs; industrial, commercial and institutional (ICI) boilers; internal combustion engines; and cement kilns. Most of the Missouri facilities identified in the analysis were EGUs already participating in federal CAIR rule. The analysis also provided recommendations on additional controls for non-EGUs, which were considered but not adopted due to the following reasons:

   * Proposed controls are not cost effective  -  over $2,000 per ton of SO2 or NOX
   * Emissions below the threshold limit of 100 tons
   * Sources passed the BART screening analysis
   * Sources already installed controls required by the NOX SIP Call

                        c.	Four Factor Analysis by Minnesota Pollution Control Agency

In addition to the CENRAP analyses, the MRPO and the Minnesota Pollution Control
Agency published a report on the four-factor analysis (referred to as the "4-factor report"). The
report looked at the four factors in a nine-state area (Minnesota, Wisconsin, Michigan, Indiana, Illinois, Missouri, Iowa, North Dakota, and South Dakota.). The 4-factor report primarily reviewed controls on EGUs; ICI boilers; reciprocating engines and turbines, and mobile sources. Missouri has determined based on the cost of compliance and remaining useful life, that controls are not reasonably available for non-EGU sources.  Missouri specifically concludes from the report that additional controls from ICI boilers, reciprocating engines, combustion turbines and other point sources are notnot warranted based on cost of controls and visibility improvement .  Missouri determined that for EGUs, emission reductions predicted to result from CAIR would be sufficient for ensuring reasonable progress during the first implementation period (between baseline and 2018).  

                  d.	Cost and Time Necessary for Compliance Under CAIR

In its SIP submittal, Missouri demonstrates that the 2018 visibility goals for Mingo and Hercules Glades have been largely achieved through EGU emission reductions.  Missouri demonstrates that the statutory factor of cost and time necessary for compliance is considered in its CAIR analysis.  The State cites that reductions will be energy and environmental neutral, but are clearly effective in improving haze levels. A discussion in the CAIR rule highlighted below (70 FR 25197), taken directly from the State's plan, addresses the State's analysis of reasonable progress factors of cost and time necessary for compliance.

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 cannot 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[Notice of Proposed Rulemaking] 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.

EPA Findings: Missouri has properly established goals that provide for reasonable progress towards achieving natural visibility conditions in Missouri's Class I areas, and through multiple analyses has properly taken into account the four statutory factors consistent with EPA's June 2007 guidance document Guidance for Setting Reasonable Progress Goals Under the Regional Haze Program. 

      C.   Long Term Strategy (40 CFR  51.308(d)(3))
	
      1. 	Technical Basis for Long Term Strategy (40 CFR  51.308(d)(3)(iii))
            SIP Pages 18-21, 42-50, 87-101
      
The plan must document the technical basis, including modeling, monitoring, and emissions information, on which the State is relying to achieve its reasonable progress goals.   The State may meet this requirement by relying on technical analyses developed by the regional planning organization and approved by all State participants.

Missouri's modeling assessment provided by the CENRAP regional planning organization addresses this requirement.  The analyses provides the following: 1) a determination of the extent of emissions reductions needed from individual states; 2) a determination of emissions needed to meet the progress goal for the Class I area; 3) analyses to support conclusion that the Long-Term Strategy provides for reasonable progress; and 4) analyses to calculate the resulting degree of visibility improvement that would be achieved at each Class I area. 

The CENRAP applied the Comprehensive Air quality Model with extensions (CAMx) and Community Multiscale Air Quality (CMAQ) models in the modeling simulation.  CAMx is a computer modeling system for the integrated assessment of photochemical and particulate air pollution.  CAMx incorporates all of the technical attributes demanded of state-of-the-art photochemical photochemical grid models, including two-way grid nesting, a subgrid-scale Plume-in-Grid module to treat early dispersion of chemistry of point source NOx plumes, and a fast chemistry solver.  The CMAQ model is an eulerian model that simulates the atmospheric surface processes affecting the transport, transformation and deposition of air pollutants and their precursors.  An eulerian model computes the numerical solution of partial differential equations of plumes on a fixed grid.  The use of these models to determine impacts from emissions within the State on visibility impairment is approved by EPA.  Missouri documented and EPA has reviewed the selection of the episodes, modeling domain, emissions inventories, emissions modeling, and meteorological inputs, and model performance evaluation.  More detailed information on methodologies is provided in Appendix of F of the state's submittal. 

The following paragraphs are taken directly from the State's submittal, section 8.4 of page 45, and describes the 2018 modeling scenario that incorporates the ongoing emissions controls Missouri is relying upon to attain their RPGs (see section III.B(2) above).  The ongoing emissions controls, further described in III. C(2) of this technical support document, account for the anticipated net effect on visibility due to projected changes in point, area, and mobile source emissions over the period addressed by the long-term strategy, meeting the requirements of 40 CFR  51.308(d)(3)(v)(G).

                  a.	State Submittal
                        SIP Page 45

The 2018 Base G modeling run reflects emissions growth and "on the books" controls, which are state and federal controls that will be implemented between the 2002 base year and the 2018
future year. The 2018 emissions for EGUs were based on simulations of the IPM that took into
account the effects of the CAIR trading program. In addition, reductions anticipated from BART
controls for EGUs in Oklahoma, Arkansas, Kansas, and Nebraska were included. Emissions for
onroad and offroad mobile sources were based on activity growth and emissions factors from the
EPA MOBILE6 and NONROAD models, respectively, which reflected emissions reductions
from the Tier 2 and Tier 4 mobile source rules. Area sources and non-EGU point sources were
grown to 2018 levels.  The two important regional haze metrics are the average visibility for the worst 20 percent and best 20 percent days from the 2000-2004 five-year baseline period. The results from the 2002 and 2018 CMAQ and CAMx simulations were used in a relative sense to scale observed PM concentrations from the 2000-2004 baseline to 2018 levels from which 2018 visibility estimates were obtained. The CENRAP 2018 visibility conditions were calculated following EPA default visibility projection procedures and are labeled "Method 1 Prediction" in Figures 8.1 and 8.2.  The steps involved in the visibility calculations are described below:

1. For each Class I area and each monitored day, daily visibility based on IMPROVE
data and the new IMPROVE equation was ranked for the five-year baseline period
(2000-2004) to identify the worst 20 percent and best 20 percent visibility days for
each year in the baseline period.

2. The CMAQ air quality model was used to simulate the base year (for CENRAP the
2002 annual period was simulated) and a future-year (2018). The resulting
information was used to develop Class I area-specific relative reduction factors
(RRFs) for each of the six components of light extinction in the IMPROVE equation
(SO4, NO3, EC, OMC, Soil and CM).

3. The RRFs were multiplied by the measured 24-hour PM concentration for each day
from the worst and best 20 percent days in each year from the five-year baseline
period to obtain projected future-year 24-hour PM concentrations for the worst and
best 20 percent days.

4. The future-year (2018) daily extinction was computed using the new IMPROVE
equation and the projected PM concentrations for each of the worst and best 20
percent days in the five-year baseline from step 3.

5. For each of the worst and best 20 percent days within each year of the five-year
baseline, the future-year daily extinction was converted to deciview. The daily
deciview values were averaged within each of the five years separately to obtain five
years (or as many years with valid data in the 2000-2004 baseline) of average
deciview visibility for the worst and best 20 percent days.

6. The five years of deciview visibility were averaged to obtain the 2018 estimated visibility. The 2018 visibility projections for the worst 20 percent days and best 20 percent days are compared against a 2018 point on the Uniform Rate of Progress (URP) glidepath or the "2018 URP point." The 2018 URP point is obtained by constructing a linear visibility glidepath in deciviews from the observed 2000-2004 Baseline for the worst 20 percent days to the 2064Natural Conditions.

 The 2018 URP point is where the linear glidepath crosses the year 2018.
Figures 8.1 and 8.2 present the 2018 visibility projections for Hercules Glades and Mingo. As
seen in these figures, the 2018 visibility projections at both the Hercules Glades and Mingo Class I areas meet the 2018 point on the URP glidepath for the worst visibility days and exhibit no degradation on the best visibility days. For the worst 20 percent days, the 2018 projection for
Hercules Glades is 23.06 dv, compared to the URP point of 23.14 dv. The 2018 projection for
Mingo is 23.71 dv, as compared to the URP point of 24.37 dv.  

EPA Findings:  The technical analyses and modeling to assess uniform rate of progress and to support the LTS were developed consistent with EPA's interim and final modeling guidance.  EPA accepts the CENRAP 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 CENRAP model performance procedures and results, and that the CMAx and CMAQ models are determined to be appropriate tools for the regional haze assessments for the Missouri LTS and Regional Haze SIP.  
	
            2.  	Identification of Sources and Factors to be Considered (40 CFR  51.308(d)(3)(iv-v)
            SIP Pages 18-21, 42-50, 87-101

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: 1) emission reductions due to ongoing air pollution control programs, including measures to address reasonably attributable visibility impairment; 2) measures to mitigate the impacts of construction activities; 3) emissions limitations and schedules for compliance to achieve the reasonable progress goal; 4) source retirement and replacement schedules; 5) 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; 6) enforceability of emission limitations and control measures; and 7) the anticipated net effect on visibility due to projected changes in point, area, and mobile source emissions over the period.  

The State's technical analysis, further described above in III.C(1), identifies all anthropogenic sources of visibility impairment considered by the State in developing its long-term strategy.  In this analysis, the State considered major and minor stationary sources, mobile sources, and area sources, meeting the requirements of 40 CFR 51.308(d)(3)(iv).  

The following, taken directly from Missouri's Regional Haze Plan,  demonstrates the State has appropriately taken into consideration the minimum factors, described above, in developing its long term strategy, meeting the requirements of 40 CFR 51.308(d)(3)(v):

                  a.	State Submittal
                        SIP Page 88

40 CFR 51.308(d)(3)(v) requires Missouri to consider several factors in developing its long-term
strategy. The ongoing air pollution control programs described in section 11.4.1. below were
used in the modeling demonstration to show that Missouri Class I Areas meet the 2018 RPG.
Additional controls beyond the RPG modeling demonstration that may be considered for use in
the long term strategy at a future date are described in section 11.4.2 below.

      11.4.1 Ongoing air pollution control programs  -  "on the books" controls

40 CFR 51.308(d)(3)(v)(A) requires states to consider emission reductions from ongoing
pollution control programs. The NOX and SO2 emissions reductions resulting from these
ongoing programs will help improve air quality throughout the state of Missouri.  Missouri used the following "on the books" control programs in the modeling demonstration to meet the RPG requirements as presented in Table 10.1 of this Regional Haze Plan. The substantial improvements in visibility impairment at the Mingo site for the worst 20 percent days from 2002 (141 Mm-1) to 2018 (96 Mm-1) is primarily due to reductions in SO4 (35 Mm-1 improvement) from elevated point sources (Figures E-7a through E-7d of the TSD). Elevated point sources also contribute over half to the total extinction for the worst 20 percent days at Hercules Glades in 2002 (Figures E-6a and E-6b of the TSD). Going from 2002 to 2018 the contributions due to elevated point sources (essentially CAIR and NOX SIP call), on-road mobile
and non-road mobile are reduced substantially, over 25 Mm-1.

      11.4.1.1 Clean Air Interstate Rule (CAIR)

On March 10, 2005, EPA signed the CAIR, following three years modeling study and cost
analysis on SO2 and NOX controls.  As required by CAIR, Missouri developed draft rules through the workgroup process.  The rules were presented for public hearing at the December 7, 2006, MACC Meeting and they were adopted at the February 1, 2007, MACC Meeting. The rules establish a cap and trade system for NOX and SO2 emissions, and Missouri sources will be included in the national program. The state rules are 10 CSR 10-6.362 Clean Air Interstate Rule Annual NOX Trading Program and 10 CSR 10-6.366 Clean Air Interstate Rule SOX Trading Program. The state rules include schedules for compliance, sources affected by the rule and emissions limitations.  Table 11.1 summarizes the NOX emissions cap for each unit for each calendar year between 2009-2014 and 2015 and beyond. Table 11.2 summarizes the SO2 emissions cap for each unit for each calendar year between 2010-2014 and 2015 and beyond. These rules can be found in Appendix V.

The long-term strategy also includes any CAIR controls that are being undertaken in other
impacting states that were identified during the Central Class I areas consultation process.

      11.4.1.2 BART

Twenty-six potential BART sources have been identified. Twenty-five have been dropped
through the screening and refined analyses. The remaining source (Holcim  -  Clarksville)
entered into a consent agreement with the Missouri Department of Natural Resources to limit
emissions of SO2 and NOX. No other sources were found to be subject to BART and, therefore,
implementation of an emissions trading program, other emission controls or other alternative
measures in place of BART are not necessary. Detailed analyses can be found in Chapter 9 of the SIP.  Missouri will include BART controls proposed by the eight other impacting states in its long-term strategy. Arkansas, Oklahoma and Texas have already had their BART rules proposed. Missouri is working with these states to document the emissions reduction and control measures required from their sources. Since not all of the BART determinations are completed for other states, the five-year review will be the mechanism used to adjust the Reasonable Progress Goal based on the other states' final BART determinations.

      11.4.1.3 Other federal ongoing air pollution control programs

      Tier 2

Tier 2 standards are federal emission standards for passenger cars, light trucks and larger
passenger vehicles. The program is designed to focus on reducing the emissions most
responsible for the ozone and PM impact from these vehicles  -  NOX and non-methane organic
gases, consisting primarily of hydrocarbons and contributing to VOCs. The Tier 2 standards will
reduce new vehicle NOX levels to an average of 0.07 grams per mile. For new passenger cars
and light duty trucks, these standards were phased in starting in 2004, and the standards were
fully phased in by 2007. For heavy trucks and similar vehicles, the Tier 2 standards will be
phased in beginning in 2008, with full compliance in 2009.  During the phase-in period from 2004-2007, all passenger cars and light trucks not certified to the primary Tier 2 standards had to meet an interim average standard of 0.30 g/mi NOX. During the period 2004-2008, heavy trucks and similar vehicles not certified to the final Tier 2 standards will phase in to an interim program with an average standard of 0.20 g/mi NOX, with those not covered by the phase-in meeting a per-vehicle standard (i.e., an emissions "cap") of 0.60 g/mi NOX trucks and 0.09 g/mi for similar vehicles.

      Tier 4

EPA's Clean Air Nonroad Diesel Rule (Tier 4) requires stringent pollution controls on diesel
engines used in industries such as construction, agriculture and mining, and it will slash sulfur
content of diesel fuel. This rule is the latest in a series of actions that are designed to reduce
emissions from nearly every type of diesel vehicle and equipment. This nonroad diesel program
combines cleaner engine technologies with cleaner fuel  -  similar to the on-highway diesel
program. The standards will cut emissions from nonroad diesel engines by over 90 percent.
Nonroad diesel equipment, as described in this rule, currently accounts for 47 percent of diesel
PM and 25 percent of NOX from mobile sources nationwide.  Sulfur levels will also be reduced in nonroad diesel fuel by 99 percent from current levels (from approximately 3,000 parts per million (ppm) now to 15 ppm in 2010). The lower sulfur fuel will also reduce PM from engines in existing nonroad equipment. It makes it possible for engine manufacturers to use advanced clean technologies, similar to catalytic technologies used in passenger cars. The new engine standards take effect, based on engine horsepower, starting in 2008.

      11.4.1.4 NOX SIP Call

The NOX  SIP call was designed to assist downwind ozone areas in attaining the one-hour and 8-
hour ozone NAAQS by providing upwind NOX emission control. This rulemaking was
developed through the EPA's interpretation of the Ozone Transport Assessment Group
recommendations and subsequent modeling and cost analysis of NOX controls to reduce ozone
transport. The final NOX  SIP call was published in the Federal Register on October 27, 1998.
Missouri's initial rule in response to the NOX  SIP Call, 10 CSR 10-6.350 Emission Limitations
and Emissions Trading of Oxides of Nitrogen, was adopted by the MACC on April 24, 2003.
The rule established an emission limitation of 0.25 lbs NOX /MMBtu heat input for electric
generating units in the eastern one-third of the state and a lower limit of 0.18 lb/MMBtu heat
input for Labadie, Rush Island, and Meramec power plants. EGUs in the western two-third of
the state were limited to an emission rate of 0.35 lbs NOX /MMBtu of heat input. Cyclone
boilers (Sibley and Asbury power plants) that burn tire-derived fuels are allowed to meet 0.68 lbs NOX /MMBtu heat input. The compliance date was May 1, 2004.

On April 21, 2004, the EPA finalized the second phase of NOX  SIP call. Phase II of the SIP call
excluded the portion known as the "coarse grid" (the western 2/3 of Missouri) from the NOX  SIP
Call, defined the area of the eastern 1/3 of Missouri to include the same counties as established in 10 CSR 10-6.350, with the one exception of not including Phelps County, and revised the cap for NOX emissions from the previous statewide budget of 114,532 tons of NOX per ozone season to a partial state budget of 61,406 tons of NOX  per ozone season in the eastern 1/3 of Missouri. The budget assumed control levels of 0.15 lbs/MMBtu for electric generating units, 82 percent
emissions reductions for large natural gas-fired stationary internal combustion engines, 90
percent emissions reductions for diesel and dual fuel stationary internal combustion engines, 60
percent emissions reductions for non-utility boilers and turbines, and 30 percent emissions
reductions for cement manufacturing plants. Small cogeneration units were excluded from the
NOX  SIP Call. Small cogeneration units are units that supply one-third or less of their potential
electrical output capacity, or 25 megawatts or less, to any utility power distribution system for
sale.  The department's Air Pollution Control Program developed 10 CSR 10-6.360 Control of NOX Emissions from Electric Generating Units and Non-Electric Generating Boilers, 10 CSR 10-6.380 Control of NOX Emissions from Portland Cement Kilns, and 10 CSR 10-6.390 Control of NOX Emission from Large Stationary Internal Combustion Engines. This set of three rules
constitutes Missouri's response to EPA's NOX  SIP Call. These rules were presented at public
hearing on April 28, 2005 and were adopted at the May 26, 2005 MACC meeting. The state rules
include schedules for compliance, sources affected by the rule and emissions limitations. Table
11.3 summarizes the NOX allowances for each unit during the ozone season. 


      11.4.2 Additional controls beyond RPG modeling demonstration

Ongoing air pollution control programs, as described in Section 11.4.1, are sufficient to meet the
2018 Uniform Rate of Progress for the Mingo and Hercules Glades Class 1 areas. These
ongoing programs such as CAIR, BACT, or BART have been demonstrated to be very cost effective in reducing the visibility in Missouri's Class I areas.  Additional controls not included in the modeling demonstration in the plan may be considered during the five-year review. A number of control strategies include SO2/NOX Reasonably Available Control Technology (RACT) in the St Louis PM2.5 plan, Illinois Multi-Pollutant Strategy, regional SO2 and NOX control strategy proposed by Alpine Geophysics (Alpine) for CENRAP.

      11.4.2.1 SO2 and NOX RACT in St Louis

Missouri is in the process of preparing an implementation plan to address the St. Louis PM2.5
nonattainment problem. In addition to the development of an attainment demonstration, the
PM2.5 implementation rule requires states to develop all RACT and Reasonably Achievable
Control Measures (RACM). All non-EGU SO2 and NOX sources were identified in the St. Louis
PM2.5 nonattainment area that had actual emissions exceeding 25 tons per year. These include
large boilers, stationary internal combustion engines, two glass melting furnaces, a biosolids
incinerator, a cement kiln and a lead smelter. The emission reductions associated with these
sources will be determined and included in the PM2.5 plan.

      11.4.2.2 Illinois Multi-Pollutant Regulation

In 2006, a multi-pollutant standard (MPS) rule was approved by the Illinois Pollution Control
Board and the Joint Committee on Administrative Rules. This multi-pollutant rule will result in
measurable reduction in mercury, SO2, and NOX emissions. The rule targets the three largest
coal-fired power plant companies in Illinois: Midwest Generation, Ameren and Dynegy. These
three companies represent 88 percent of Illinois' 17,007 Megawatts of electric generating
capacity from coal-fired plants. By implementation of this rule, the Illinois Environmental
Protection Agency estimates the total emissions reduction from all three power companies is
233,600 tons per year of SO2 and 61,434 tons per year of NOX. This is a drastic improvement
compared to emissions reduction achieved by the CAIR.

      11.4.2.3 Regional Controls proposed by Alpine Geophysics (Alpine)

In February 2006, Alpine was contracted by CENRAP to assist in developing control strategies
for CENRAP Class I areas. Based on the available cost information and the Area of Influence
(AOI) analyses, Alpine proposed a methodology for constructing control strategy for both EGUs
and non-EGUs. Control technologies for different industrial source categories were identified.
Regional "CAIR-like" EGU controls, and sub-regional (AOI region)  -  Industrial, Commercial,
and Institutional boilers and natural gas compressors controls  -  were recommended by Alpine.
The final report from Alpine can be found in Appendix W.

      11.4.3 Measures to mitigate the impacts of construction activities

40 CFR 51.308(d)(3)(v)(B) requires Missouri to consider measures to mitigate the impacts of
construction activities. Under the NAAQS, any nonattainment area in Missouri is required to
consider construction emissions as part of the general conformity rule. Missouri meets this
commitment through rule 10 CSR 10-6.300 Conformity of General Federal Actions to State
Implementation Plans, which can be found in Appendix V [of the State SIP]. This rule sets forth policy, criteria and procedures for demonstrating and assuring conformity to applicable implementation plans.

      11.4.4 Source retirement and replacement schedules

40 CFR 51.308(d)(3)(v)(D) requires Missouri to consider source retirement and replacement
schedules in developing RPGs. Retirement and replacement will be managed in conformance
with existing SIP/TIP requirements pertaining to PSD and New Source Review (NSR).

      11.4.5 Smoke Management Plan

40 CFR 51.308(d)(3)(v)(E) requires Missouri to consider smoke management techniques for the
purposes of agricultural and forestry management in developing RPGs.  The purpose of the Smoke Management Plan (SMP) adopted by Missouri is to identify the responsibilities of the Missouri Department of Natural Resources, FLMs, and state land managers to coordinate procedures that mitigate the impacts of prescribed fire and wildland fire used for resource benefits on public health, safety and visibility. This plan is designed to meet the policies of the EPA's Interim Air Quality Policy on Wildland and Prescribed Fires (April 1998) and addresses smoke management through various procedures and requirements in place at various agencies throughout the state.  The department does not intend to submit the SMP for inclusion in the Missouri SIP, but a copy of the Missouri SMP is provided in Appendix X for reference. A letter certifying that the SMP meets the basic requirements will be provided to EPA.  The purpose of a SMP is to mitigate the nuisance and public safety hazards (e.g., on roadways and at airports) posed by smoke intrusions into populated areas; to prevent deterioration of air quality and NAAQS violations; and to address visibility impacts in mandatory federal Class I areas. Some strong indications that an area needs a SMP are: (1) citizens increasingly complain of smoke intrusions; (2) the trend of monitored air quality values is increasing (approaching the daily or annual NAAQS for PM2.5 or PM10) because of significant contributions from fires managed for resource benefits; (3) fires cause or significantly contribute to monitored air quality that is already greater than 85 percent of the daily or annual NAAQS for PM2.5 or PM10; or (4) fires in the area significantly contribute to visibility impairment in mandatory federal Class I areas. None of these four indicators currently shows a problem in Missouri. However, the Missouri SMP should provide additional protection to the federal Class I areas.

      11.4.6 Enforceability of Emission Limitations and Control Measures

40 CFR 51.308(d)(3)(v)(F) requires Missouri to ensure that emission limitations and control
measures used to meet RPGs are enforceable.  Missouri has ensured that all emission limitations and control measures used to meet RPGs are enforceable by Missouri law through section 643 of the Revised Statutes of Missouri. In addition, rules developed for CAIR and the NOX SIP call have placed emission limits on both EGU and non-EGU units. These rules can be found in Appendix V.

      11.4.7 Anticipated net effect on visibility resulting from projected changes to emissions

40 CFR 51.308(d)(3)(v)(G) requires Missouri to address the net effect on visibility resulting
from changes projected in point, area and mobile source emissions by 2018.  The emission inventory for Missouri projects changes to point, area and mobile source inventories by the end of the first implementation period resulting from population growth; industrial, energy and natural resources development; land management; and air pollution control. A summary of these changes is given in Tables 7.1 and 7.2 for each of the pollutants addressed in the regional haze plan inventory.  The net effect on visibility in Missouri Class I areas resulting from these emission differences is discussed in the CENRAP Technical Support Document (Appendix F).

EPA Findings: Missouri has appropriately addressed  40 CFR 51.308(d)(3)(iv-v).  The State's analysis shows that all anthropogenic sources of visibility impairment were considered by the State and were properly included in the CENRAP modeling used to assess establishment of RPGs for Missouri's Class I areas (see III.C(1) above).  In addition, EPA finds that the State has properly considered the minimum required factors in developing its long-term strategy.

	3.	Consultation (40 CFR  51.308(d)(3)(i-ii))
      	SIP Pages 18-24, 87-88

As part of the long-term strategy requirements of the rule, provision 40 CFR 51.308(d)(3)(i) specifically describes that, where the State has emissions that are reasonably anticipated to contribute to visibility impairment in any Class I area located in another state or states, the State must consult with other state(s) in order to develop coordinated emissions management strategies.  The State must consult with any other state having emissions that are regionally anticipated to contribute to visibility impairment in any mandatory Class I  area within the State.   Further, provision 40 CFR  51.308(d)(3)(ii) states that where other states cause or contribute to impairment in a mandatory Class I area, the State must demonstrate that it has included in its implementation plan all measures necessary to obtain its share of the emissions reductions needed to meet the progress goal for the area.  If the State has participated in a regional planning process, the State must ensure it has included all measures needed to achieve its apportionment of emission reduction obligations agreed upon through that process.

                  a.	State Submittal
                        SIP Pages 18, 87
                  
Missouri has consulted with other states/tribes in CENRAP, Visibility Improvement State and
Tribal Association of the Southeast (VISTAS), the Midwest Regional Planning Organization
(MRPO), FLMs and EPA Regions 5, 6 and 7 on development of coordinated strategies for
Central Class I areas, including Mingo, Hercules Glades, Upper Buffalo, and Caney Creek.
Technical analyses, such as Area of Influence (AOI) and source apportionment, were developed
as part of consultation planning to determine contributing states (Appendix E).  Missouri provided the Regional Haze Plan to the FLMs for review on August 23, 2007 and notified the FLMs that a public hearing would be held on this plan at a later date. The FLMs provided early comments on the draft plan and a conference call between Missouri, FLMs, and EPA Region 7 was conducted on September 25, 2007 to discuss the comments. Missouri considered all comments the FLMs provided on the early draft of the plan. Regional modeling and other findings were used to develop RPGs for the Arkansas and Missouri Class I areas based on the existing and proposed controls through both state and federal requirements. It was also determined that these RPGs will meet the established URP goals by 2018. The consultation process determined which states significantly impacted the Arkansas and Missouri Class I areas.
Missouri is reasonably anticipated to contribute to the following Class I areas:

1) Mingo National Wildlife Refuge, Missouri
2) Hercules Glades Wilderness, Missouri
3) Upper Buffalo Wilderness Area, Arkansas
4) Caney Creek Wilderness Area, Arkansas

The state's coordination with FLMs on long-term strategy development is described in Chapter
11 of the SIP. The consultation was completed based on a determination that reasonable progress was achieved by contributing states.

      4.2 OTHER STATE CONSULTATIONS

The consultation processes for the Wichita Mountains (WIMO) Class I area in Oklahoma was
completed prior to the March 2008 submittal of this plan. The Oklahoma Department of
Environmental Quality indicated their belief that Missouri sources impact WIMO. However, in
response to the Oklahoma consultation letter, Missouri replied with a letter recommending that
the rationale for determining contributing states deserves further examination. A more inclusive
methodology for the Central Class I areas with four different metrics (Particulate Matter Source
Apportionment Technology (PSAT), PMF, AOI, and emission rate divided by five times the
distance - Q/d) was used in a combined manner with three out of four positive results required
before concluding that a particular state is contributing. The distance between WIMO and
western Missouri's Class I area is approximately 200-250 miles farther west. Because of this
distance, it is counter-intuitive to assume that planned emission controls on Missouri sources
would be significant enough. It seems likely that Missouri would not be included as significant
based on this level/type of PSAT analysis, and emissions/source distance ratio. 

It is also not clear that additional controls in Missouri would be reasonable to reduce the visibility in WIMO. Based on the PSAT analysis presented, over half the elevated point-source impacts to WIMO are due to sources in Oklahoma, Texas, and Louisiana and most of the area source impacts are due to Oklahoma and Texas sources. Point and area are the two largest emission sectors. Controls appear likely to be more efficient in those states, on a cost-per-ton basis, than additional controls in Missouri. 

Consultation processes for the Minnesota Class I areas was also conducted prior to the March 2008 submittal of this plan. The Minnesota Pollution Control Agency indicated that it believes that Missouri impacts the Boundary Waters, but not Voyageurs. 

Minnesota identified Missouri as a contributing state based on LADCO 2002-2003 Trajectory analysis or LADCO 2018 PSAT modeling analysis at over a 5 percent total contribution to haze at either of their Class I areas. The criteria are met marginally at 5.2 percent for 2018 PSAT for the Boundary Waters area only.

Analysis conducted as part of the Causes of Haze II Study shows emissions for the northern
Class I area at Voyageurs National Park indicating the high impact of Minnesota sources, with
only small impact by out of state sources. Causes of Haze II, 2005; Sullivan, Hafner, Brown, MacDonald, Raffuse, Penfold, Roberts, Sonoma Technology.  Voyageurs and the Boundary Waters are very close in proximity, and the overall analysis was intended to apply regionally. The Emission Impact Potential mappings below underscore the impact of Minnesota sources on the area, and how controls on a relative few will provide much greater result than controls on sources outside Minnesota. Comparisons with Hercules Glades EIP show the difference in areas with significant external sources (Hercules) to areas with significant internal sources (Voyageurs, Boundary Waters). The conclusion reached in the Causes of Haze II [Study] is that for Voyageurs, and by geographic proximity, Boundary Waters, important emission source regions are internal (Minnesota and to a lesser extent North Dakota) on the 20 percent worst days. Area of Influence analysis for CENRAP states confirms that Level I Sulfate for both areas barely enters the northwest 20 miles of Missouri, not indicating strong source influence.

The most recent CENRAP PSAT analysis, as Figure 4.3, shows most Minnesota anthropogenic
sources with very high impacts on Boundary Waters, slightly more than 15 inverse megameters
for 2002. For 2018 modeling, it remains at almost 14 inverse megameters. Of other states, only
Wisconsin elevated point impacts are larger than 2 inverse megameters, and Missouri impacts
are 1.6. Based on the AOI and PSAT analyses, it is not reasonable to control the Missouri
sources at the same level as Minnesota sources to achieve a very small impact at the Boundary
Waters Class I area.  Missouri provided a written response to Minnesota regarding the Northern Class I Areas consultation.  Since other states are still involved in their consultation process for their respective plans, Missouri will continue to participate in their consultation processes, as necessary.

40 CFR 51.308(d)(3)(ii) requires Missouri to demonstrate that its implementation plan includes
all measures necessary to obtain its fair share of emission reductions needed to meet RPGs.

Missouri relied on technical analyses developed by CENRAP and additional weight of evidence
analysis developed as part of consultation planning to determine contributing states (Appendix
E). Nine states, including Missouri, Arkansas, Kentucky, Illinois, Indiana, Ohio, Oklahoma,
Tennessee and Texas, were identified as contributing to visibility in Mingo and/or Hercules
Glades Class I areas. The modeling demonstration has shown that the emission reductions from
these contributing states are sufficient to achieve RPGs in Missouri's Class I areas. 

Current visibility is estimated from monitored components of PM2.5 and coarse mass. Models are used in a relative sense to estimate how current concentrations respond to emission reduction measures. Data analysis is used to identify source categories and regions. Current concentrations of particulate matter components are adjusted by the relative modeled response to estimate concentrations at the end the first implementation period in 2018. Future visibility is
estimated from estimated component concentrations of PM2.5 and PM10 at the end of the first
implementation period. The difference between present visibility and future estimated visibility
is compared with the RPGs to determine if the goal is met. The CENRAP technical analyses on
visibility conditions and RPGs projections can be found in Appendix F. All applicable measures
reflected in the modeling demonstration and weight of evidence analysis have been incorporated
in the state's long-term strategy. Section 11.4 provides information on these control measures.

      11.2  CONSULTATION
      
Missouri will continue to coordinate and consult with the consultation stakeholders during the development of future progress reports and plan revisions, as well as during the implementation of programs having the potential to contribute to visibility impairment in the mandatory Class I areas. Face to face meetings will be held if deemed necessary.  Otherwise, consultation will be in the form of conference calls and/or letter or email correspondence.  The Central Class I areas consultation will be initiated through the Central Class I areas contacts listed in the consultation plan (Appendix E).

Missouri also participated in the consultation processes for Arkansas, Oklahoma and Minnesota; and will continue to participate in other consultation processes in response to any other states that request our participation.

      11.3  SHARE OF EMISSION REDUCTIONS

40 CFR 51.308(d)(3)(ii) requires Missouri to demonstrate that its implementation plan includes all measures necessary to obtain its fair share of emission reductions needed to meet RPGs.

Missouri relied on technical analyses developed by CENRAP and additional weight of evidence analysis developed as part of consultation planning to determine contributing states (Appendix
E).  Nine states, including Missouri, Arkansas, Kentucky, Illinois, Indiana, Ohio, Oklahoma,
Tennessee and Texas, were identified as contributing to visibility in Mingo and/or Hercules
Glades Class I areas.  The modeling demonstration has shown that the emission reductions from these contributing states are sufficient to achieve RPGs in Missouri's Class I areas.

Current visibility is estimated from monitored components of PM2.5 and coarse mass.  Models
are used in a relative sense to estimate how current concentrations respond to emission reduction measures.  Data analysis is used to identify source categories and regions.  Current
concentrations of particulate matter components are adjusted by the relative modeled response to estimate concentrations at the end the first implementation period in 2018.  Future visibility is estimated from estimated component concentrations of PM2.5 and PM10 at the end of the first implementation period. The difference between present visibility and future estimated visibility is compared with the RPGs to determine if the goal is met.  The CENRAP technical analyses on visibility conditions and RPGs projections can be found in Appendix F.  All applicable measures reflected in the modeling demonstration and weight of evidence analysis have been incorporated in the state's long-term strategy.  Section 11.4 provides information on these control measures.

EPA Findings: Missouri has appropriately addressed  40 CFR 51.308(d)(3)(i-ii).  The State's plan provides documentation of the consultation process with states within the CENRAP as well as with states in other RPOs.  Missouri identified Class I areas reasonably anticipated to be impacted by emissions from the State and has concluded the consultation process was completed based on a determination that these Class I areas will achieve their respective reasonable progress goals.  Additionally, EPA finds that the State properly consulted with Oklahoma and Minnesota and relied upon multiple technical analyses to determine that no additional emissions reductions were needed by Missouri for Class I areas in those States. 

	D.  	Best Available Retrofit Technology

Section 169A of the CAA directs States to assess certain large, older emission sources for additional controls in order to address visibility impacts.  States are directed to conduct BART determinations for such sources in specific source categories that contribute to visibility impairment in Class I areas.  The 1999 Regional Haze Rule includes the BART requirement, and directs States to include BART in their regional haze SIPs.  On July 6, 2005, EPA published a revised final rule, including Appendix Y to 40 CFR Part 51, the Guidelines for BART Determinations Under the Regional Haze Rule (hereinafter referred to as the "BART Guidelines") that provides direction to states on determining which of these sources should be subject to BART, and how to determine BART for each source.

A BART-eligible source is one which has the potential to emit 250 tons or more of a visibility-impairing air pollutant, was put in place between August 7, 1962 and August 7, 1977, and whose operations fall within one or more of 26 specifically listed source categories. Under the CAA, BART is required for any BART-eligible source that a State determines ``emits any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in any such area.''  

For those sources subject to BART, Section 169A(g)(7) of the CAA requires that States must consider the following factors in making BART determinations: (1) the costs of compliance, (2) the energy and non-air quality environmental impacts of compliance, (3) any existing pollution control technology in use at the source, (4) the remaining useful life of the source, and (5) the degree of improvement in visibility which may reasonably be anticipated to result from the use of such technology.  

      1.	BART Eligible Sources

The implementation plan should include a list of all BART-eligible sources within the State.  The Guidelines for BART Determinations under the Regional Haze Rules (40 CFR 51, Appendix Y) contains criteria for a State to follow in identifying sources.

Using criteria from Guidelines for BART Determinations under the Regional Haze Rules (40 CFR 51, Appendix Y) Missouri identified those sources within the State that were BART-eligible.  Missouri's SIP Narrative contains this discussion in pages 51-54.
                  
                  
                  
                  
                  a.	State Submittal 
                  	SIP Pages 51-54

The State of Missouri identified potentially BART-eligible source by reviewing the emission inventory database and extracting data for facilities within the 26 categories identified. A survey was conducted for the facilities within this group asking for large source identification and the timing of the installation and operation of those sources. The sources listed in Table 9.1 have been identified as sources that meet the criteria for inclusion as BART-eligible sources. Beyond the three primary visibility-impairing pollutants (SO2, NOX, and PM10), the sources were also asked to identify ammonia and VOC emissions. The survey and resultant tabular response information are contained in Appendix J and Appendix I,  [of the SIP] respectively.


      Table 9.1- Facilities with BART-eligible Units in the State of Missouri
      
In addition to the original SIP submittal the State provided a supplemental technical analysis on January 30[th], 2012, that included additional analysis of potential BART sources.  Section 2 of this document discusses the additional analysis performed.  Missouri re-examined their original list of sources and in the process found six facilities needing additional review.  The following is taken directly from the State's supplemental submittal describing the process they undertook.

First a search was done on the type of operating permit the source has been most recently issued.  Basic and intermediate sources do not have a PTE above 100 TPY for any pollutant in their permit.  Intermediate sources agree to take limits in their permit to be below 100 TPY for all criteria pollutant they emit so they are not required to have a Part 70 (aka Title V) permit.  Therefore, all sources with a Basic or Intermediate operating permit are immediately eliminated from BART consideration because their total plant wide PTE for any BART pollutant is well below 250 TPY.  Columns were added to the spreadsheet to show which sources have basic (BAS) or intermediate (INT) permits.  Intermediate permit numbers were listed in a separate column when available.

After this initial review, there were approximately 30 sources that have Part 70 (Title V) permits and required further examination.  The majority of these sources have relatively recent operating permits.  The Air Program reviewed these sources to determine BART applicability based on dates of the equipment as well as calculated PTE for the plant and/or BART eligible points/units. Notes of explanation were added to the "comment" column indicating if plant wide PTE showed that the pollutants of concern were each below 250 TPY.   In addition, the dates for all of the units were verified to confirm exemption, a number of emission points/units were determined to not be emitters of the pollutants of concern, and finally only the BART-eligible units at sources were calculated for PTE.  The most recent operating permit for each Title V permitted facility is listed in a column of the spreadsheet in Appendix A.  If there is not a recently completed operating permit available, there has been sufficient work done by Air permitting staff on draft permits for these sources to allow for the calculation of PTEs.

EPA Findings:  The State's original SIP submittal and supplemental submittal includes a list of all BART-eligible sources within the State and includes supporting documentation describing the analysis performed.  Missouri thoroughly reviewed their permit database and documented in their supplemental submittal the various reason(s) why specific sources were excluded as BART-eligible. Appendix A of the State's supplemental submittal includes a detailed description of a sources and units BART applicability including references to permits where applicable. Missouri has appropriately identified BART-eligible sources.
	
      2.	Sources Subject to BART

A state has various options in addressing their BART-eligible sources including exempting them from BART.  A BART-eligible source in the State may be exempted from a BART demonstration by using an acceptable approach demonstrating that the source or group of sources do not violate the established visibility impairment threshold. The  exemption threshold for determining if a BART-eligible source contributes to regional haze cannot exceed 0.5 deciviews.  Those sources not exempt must perform a full BART analysis.

                        a.	State submittal

Missouri chose to use multiple methods to exclude sources from a full BART demonstration.  These methods are described below.  The following, taken directly from Missouri's Regional Haze Plan, demonstrates the State's approach for analyzing the BART eligible sources either as a group or individually.  

Upon completion of the survey, the BART-eligible sources were divided into four distinct groups: (1) electric generating units participating in the CAIR trading program, (2) sources that have final new source review (construction) permits requiring a BART-eligible unit "shutdown" or no current operating permit at the facility, (3) sources that have gone through a subsequent construction permitting exercise for units that would have been BART-eligible based on original installation date, and (4) all other units that underwent a screen-modeling evaluation to determine the visibility impact on the applicable Class I area(s).

The first and fourth groups were evaluated initially using a screen-modeling technique for visibility impact. The first group (CAIR EGUs) was modeled collectively using all BART eligible sources for only the PM impacts on the applicable Class I areas (NOTE: After EPA comment, two additional CAIR EGU facilities were discovered to be BART-eligible and additional analysis was performed). 

For the first group of sources, the CAIR EGU's identified in Table 9.1 on page 52f the SIP, Missouri's screening modeling for PM only indicated maximum modeled impacts for the collective group at less than 0.5 deciviews.  These results are contained in Table 9.3 of the SIP submittal (Table 9 of this TSD, below) and include Hercules Glades, Mingo and Upper Buffalo Class I areas.  The other visibility impairing pollutants, NOx and SO2, are being addressed by  CAIR.  Modeling results can be found in Complete CALPUFF/CALPOST Modeling Files in Appendix K Parts 1-6 of the SIP.

The second group, sources without an operating permit or that were shutting down, were handled in the following manner in the SIP submittal:

The second group included some or all of the BART-eligible sources at four installations. River Cement, Doe Run  -  Glover, Continental Cement, and Mississippi Lime all had units that were part of a voluntary shutdown or being removed due a specific construction permit condition. To be clear, the Doe Run  -  Glover facility does not have a current operating permit for any of the BART-eligible units at the facility. The department sent a notice to the company on November 13, 2007, that detailed the termination of the operating permit for the BART-eligible units after a thirty-day period allowed for the company to provide additional information. This period expired on December 13, 2007, and the permit was closed out. Therefore, the pyro-process units at Glover cannot restart without a new construction permit requiring Best Available Control Technology (BACT) evaluation for each. These BART eligible units were not included in the modeling analyses and are shown in Table 9.2. During the public comment period, Mississippi Lime Company provided a comment that changed the units that were subject to the BART air quality screening evaluation. Mississippi Lime Company has provided a permit modification that includes the continued use of Peerless Rotary Kiln #4 (EP69) at the facility. The unit was originally subject to shutdown provisions in the applicable permit, but the updated information required a new evaluation of the facilities' BART-eligible units.  The other unit in the shutdown provision (PRK #3  -  EP68) has been dismantled. 


Table 9- Units Removed from BART Consideration Due to Shutdown or federal Permit Requiring Shutdown

The third source category contained a single source, Doe Run  -  Buick, that had undergone recent permit modifications.  This source was reviewed for BART applicability by reviewing the recent BACT analysis the source was required to perform as part of a permit action and making a BART determination based on this analysis.  Missouri's SIP on page 56 states as follows: 

Missouri reviewed the SO2 BACT finding for the blast furnace and found that three different technologies were considered: wet scrubbing, dry scrubbing, and de-sulfurization of feed materials. Dry scrubbing was eliminated as technically infeasible due to excess hazardous waste generation. Wet scrubbing of the exhaust gas in a packed tower scrubber was selected as the best available technical option for significant control. However, the capital cost of this type of unit was estimated to be $24,100,000 with an annualized system cost of $11,000,000. The cost per ton SO2 reduced was $3,537, which demonstrated that this alternative was cost-ineffective. The continued use of chemical de-sulfurization of the battery paste input to the smelting process at the Buick facility was determined to be BACT and resulted in the SO2 emission limit in the 2005 permit. The BART finding is equivalent to the BACT finding in this case due to the high cost of the wet scrubbing technology and the on-going benefit of the desulfurization process. In addition to this cost finding and to better understand the visibility impact of the BART-eligible unit, the department conducted a sensitivity analysis of the Doe Run Buick facility on the nearby Mingo Class I area. The results of this sensitivity are included below in the refined screening discussion under Section 9.3.7  -  Doe Run Buick.

In addition to the SIP analysis for this source, Missouri submitted an additional supplemental modeling analysis on January 30, 2012.  Missouri performed both screening and refined modeling for this source.  The results of the refined modeling indicate this source will not have a visibility impact over the 0.5 deciview threshold at either of the two Missouri Class I areas.  Results of this analysis can be found in section 2.3 of the supplemental technical analysis.

For the fourth group of sources, CALPUFF modeling was performed on individual sources to exempt the source based on visibility impacts below the 0.5 deciview threshold.  This modeling included both screening and refined modeling and also relied on  Missouri's supplemental technical analysis for several sources where emissions were mistakenly omitted in the original analysis.  The screening analysis used a conservative approach taking the maximum modeled visibility impacts while the refined modeling was required if the screening modeling showed a potential visibility impact.  The refined modeling allowed for the use of a 98 percentile impact while the screening modeling took the maximum impact.  Missouri's SIP on page 57 states:
 
The State of Missouri utilized two different methods for evaluation of visibility impacts: (1) Method 2  -  modeled relative humidity factors are calculated for each hour/day of the modeling period and (2) Method 6  -  an average relative humidity factor is applied for each Class 1 area being evaluated. Based on the analyses, it was determined that Method 2 provides more conservative results for visibility calculation. Since only Method 6 was required by the BART rulemaking, the use of Method 2 gives added confidence to the findings regarding sources that did not trigger refined modeling. Some or all of the following Class 1 areas were evaluated based on source location: Mingo, Hercules Glades, Upper Buffalo, and Mammoth Cave (Kentucky). The screening evaluation criterion was a maximum deciview impact of greater than 0.5 deciview to require a refined analysis. Six sources were identified during the source-specific screening analyses and these sources were notified to provide refined CALPUFF modeling analyses and/or the department conducted the refined screening analyses. In accordance with the guidelines, a contribution threshold of 0.5 deciview (98th percentile) was used for determining which sources were subject to BART using the refined modeling approach.

As stated above, Missouri chose a screening evaluation criterion of a maximum deciview impact greater than 0.5 deciview to require a refined analysis. Six sources were identified during the source-specific screening analyses and these sources were notified to provide refined CALPUFF modeling analyses and/or the State of Missouri conducted the refined screening analyses.  The six sources are identified in bold in Table 10 below.


      Table 10- CALPUFF/CALPOST Screening Results

In addition to the maximum impact metric shown above in the screening analyses, Missouri evaluated the number of days with visibility impacts over 0.5 deciview threshold to decide which Class I area would be necessary to include in the refined analyses. The results shown in Table 11 illustrate the number of days over the threshold for each source/Class I area combination. Again, these results were utilized to identify which Class I areas were more likely to be impacted and which sources would need further refined analyses.




      Table 11- Number of Days over the 0.5 Deciview Thresholds

Based on the screening analyses, Missouri required the University of Missouri-Columbia (UMC) to submit refined analysis for Mingo, Hercules Glades, and Upper Buffalo; and Noranda to submit refined modeling for Mingo. Refined modeling was also utilized for BASF Corporation, Independence Power and Light  -  Blue Valley, Columbia Municipal Power Plant, Marshall Municipal, and Holcim - Clarksville. Each of these BART-eligible sources was evaluated for Mingo, Hercules Glade, and Upper Buffalo Class I areas.

Based on this review process, the source parameter, emissions, and many other CALPUFF /CALPOST issues were addressed to allow for the development of the refined analysis. The following are brief descriptions of these analyses: 

   * University of Missouri-Columbia (UMC)

The refined analysis presented demonstrates that Boiler 10 at the University of Missouri  -  Columbia does not exceed the 98th percentile visibility impact threshold of 0.5 deciview for CALPOST Method 2 methodology, which is more conservative than the required Method 6 methodology. When comparing the screening analysis to the refined analysis, it was apparent that the use of the PM10 speciation profile by the source did not impact the visibility change.  The use of the speciaition profile did not impact visibility changes because  SO2 emissions from this boiler contribute over 90 percent of the visibility impact at all the relevant Class I areas evaluated when the impacts are above 0.5 deciviews.  Further, the SO2 emissions calculation from this source was based on a potential maximum emission rate (maximum hourly design rate * emission factor). The refined analysis showed the SO2 impact did not exceed the standard.  

   * Noranda

Missouri's BART analysis of Noranda did not evaluate the maximum allowable rates for all pollutants. However, most of the units at Noranda were analyzed using  the maximum allowable rates. However, there were five units used  maximum 24-hour actual SO2 emissions in the analysis: (1) EP-59, Monitor  -  Potline 1; (2) EP-60, Monitor  -  Potline 2; (3) EP-61, Stack  -  Potline 1&2; (4) EP-98, Carbon Bake Furnace 1; and (5) EP-99, Carbon Bake Furnace 2. 

Missouri utilized all the emission calculations in both the screening and the refined meteorological evaluations for Noranda's impacts, which can be found in Appendix M.
Missouri identified one issue with respect to the refined analyses completed by Noranda's contractor.  The CALMET model required the use the location of the surface, upper air, and precipitation stations to develop the appropriate 3-D meteorological fields for input into CALPUFF. During the review, Missouri discovered that the location of the precipitation stations utilized in the Noranda project were based on incorrect latitude and longitude data procured from another source. Therefore, at this time, the results from the emission calculations presented in Tables 12 and 13 reflect the previous Noranda submittal and the corrected submittal. Comparing the results shows that  both sets of calculations are nearly identical and reflect that Noranda's BART-eligible units do not cause or contribute to a visibility problem at Mingo.

The results demonstrate that the BART-eligible units at Noranda do not exceed the 98th percentile visibility impact threshold of 0.5 deciviews even if the more conservative CALPOST Method 2 methodology is used.


      Table 12- Noranda Refined Analysis Results (Original Submittal)

     Table 13- Noranda Refined Analysis Results (Revised CALMET)
 
   * Independence Power and Light  -  Blue Valley

The emission rates used for the BART-eligible unit (Boiler #3) at Independence Power and Light were generated by using the maximum boiler heat input (540 MMBTU/hr) and the minimum heat content for the coal used over the last five years (10,100 BTU/lb in 2004).  The University of Missouri  -  Columbia originated the refined meteorological dataset for this analysis.The results of the refined analysis for Mingo, Hercules, and Upper Buffalo are presented in Table 14 including the 8[th] highest (98 percentile) values. Also, as with the MDNR analysis of the University of Missouri  -  Columbia's boiler, the same PM10 direct conversion to positive matrix factorization PMF calculation methodology was utilized to determine visibility impacts in CALPOST.

The results demonstrate that the source at Independence Power and Light does not exceed the 98[th] percentile visibility impact threshold of 0.5 deciviews even if the more conservative CALPOST Method 2 methodology was used.


Table 14- Independence Power and Light Refined Analysis Results

   * Marshall Municipal Utilities

The emission rates used for the BART-eligible unit (Coal-fired Boiler  -  EP05) at Marshall were generated by using the maximum boiler heat input (235 MMBTU/hr) and the minimum heat content for the coal used over the last five years (10,653 BTU/lb in 2005).   The University of Missouri  -  Columbia originated  the refined meteorological dataset for this analysis. The results of the refined analysis for Mingo, Hercules, and Upper Buffalo are presented in Table 15, including the 8th highest (98 percentile) values. Also, as with the MDNR analysis of the University of Missouri  -  Columbia's boiler, the same PM10 direct conversion to PMF calculation methodology was utilized to determine visibility impacts in CALPOST. 

The results demonstrate that the source at Marshall Municipal Utilities does not exceed the 98[th] percentile visibility impact threshold of 0.5 deciview even if the more conservative CALPOST Method 2 methodology was used.


   Table 15- Marshall Municipal Utilities Refined Analysis Results

   * Columbia Municipal Power Plant

The emission rates used for the BART-eligible unit (Boiler #7 -  EP02) at Columbia were generated by using the maximum boiler heat input (371 MMBTU/hr) and the minimum heat content for the coal used over the last five years (13,304 BTU/lb in 2004).  The University of Missouri  -  Columbia originated the refined meteorological dataset for this analysis. The results of the refined analysis for Mingo, Hercules, and Upper Buffalo are presented in Table 16 including the 8th-highest (98 percentile) values. The same PM10 direct conversion to PMF calculation methodology was utilized to determine visibility impacts in CALPOST as with the other boilers reviewed. 

The results demonstrate that the source at Columbia Municipal Power Plant does not exceed the 98[th] percentile visibility impact threshold of 0.5 deciviews even if the more conservative CALPOST Method 2 methodology was used.


   Table 16- Columbia Municipal Utilities Refined Analysis Results

   * BASF Corporation  -  Palmyra

The emission rates used for the BART-eligible units (124,000 Gallon Nitric Acid Tank  -  PR08, D Incinerator  -  TC01, A Incinerator  -  PR53, B Incinerator  -  PR54, and 2 Natural Gas Fired Boilers  -  UTIL07) at BASF were generated by utilizing the potential throughput for each unit and the maximum emission factor over the last five years.  The results of the refined analysis for Mingo, Hercules, and Upper Buffalo are presented in Table 17 including the 8th-highest (98 percentile) values. The same PM10 direct conversion to PMF calculation methodology was utilized to determine visibility impacts in CALPOST. The results demonstrate that the BART-eligible units at BASF Corporation-Palmyra Plant do not exceed the maximum visibility impact threshold of 0.5 deciviews even if the more conservative CALPOST Method 2 methodology was used.

  Table 17- BASF Corporation Refined Analysis Results

   * Doe Run  -  Buick

For Doe Run  -  Buick the initial SIP submittal had modeling that exempting this source from further BART analysis as it was determined that impacts were less than the 0.5 deciview threshold that Missouri is using.  However, the supplemental analysis that MO submitted on January 30, 2012 explained that the blast furnace units at the source were not included in the original analysis because of a recent BACT determination for the units.  Missouri included an additional modeling analysis in Section 2.3 of the January 30, 2012, supplemental technical analysis that included emissions from these BART units. The emission rates used in the original visibility analysis for the BART-eligible unit (Blast Furnace -Main Stack EP08) at Doe Run Buick was generated using the 2005 construction permit limits for this source. Specifically, the facility-wide SO2 limit was utilized, while the unit specific limits for NOx and PM10 were used. The emission rates are as follows:

      SO2 -- 18,630.1 lb SO2/day
      NOX  -  20.6 lb NOX /day
      PM10  -  184.8 lb PM10/day

The University of Missouri  -  Columbia originated the refined meteorological dataset for this original analysis and supplementary analysis. The results of the refined analysis from the January 30, 2012 supplementary technical analysis for Mingo and Hercules Glade are presented in Table 18. Also, as with the MDNR's analysis of the University of Missouri  -  Columbia's boiler, the same PM10 direct conversion to PMF calculation methodology was utilized to determine visibility impacts in CALPOST. These results demonstrate that the current set of emissions at Doe Run  -  Buick will not have an adverse impact in Mingo or Hercules Glade.

                                 Class I Area
                                     Year
                          8th Highest (98%) M6 Impact
                          8th Highest (98%) M2 Impact
                                   Hercules
                                     2001
                                     0.245
                                     0.227
                                   Hercules
                                     2002
                                     0.274
                                     0.294
                                   Hercules
                                     2003
                                     0.351
                                     0.407
                                     Mingo
                                     2001
                                     0.428
                                     0.449
                                     Mingo
                                     2002
                                     0.295
                                     0.382
                                     Mingo
                                     2003
                                     0.305
                                     0.446
   Table 18- Doe Run Buick Refined Analysis Results from Section 2.3 of the supplemental submittal.

   * Holcim  -  Clarksville

Before the August 2009 regional haze state implementation plan submittal, there were two different modeling analyses conducted by Missouri for Holcim. The first evaluation utilized SO2 and NOX emissions from the calculation documented below (tons clinker per hour X lb pollutant / ton clinker). The other analysis utilized a maximum 24-hour actual emission rate for SO2 and NOX from 2004-07 based on Holcim's continuous emission monitoring system. The emission rates used in this analysis for the largest BART-eligible units at Holcim Clarksville (EP14  -  Main Stack) were generated by using the maximum clinker throughput for the kiln (175 tons/hour) and the maximum emission factors over the last five years. The highest annual emission factor for the last five years for each pollutant of interest is: SO2  -  22.97 lb SO2/ton clinker, 13.89 lb NOX/ton clinker, and 0.22 lb PM10/ton clinker. The calculated emissions were:

SO2 -- 175 tons clinker/hour * 22.97 lb SO2/ ton clinker * 24 hours/day = 96,474.0 lb SO2/day
NOX  -  175 tons clinker/hour * 13.89 lb NOX /ton clinker * 24 hours/day = 58,338.0 lb NOX /day
PM10  -  175 tons clinker/hour * 0.22 lb PM10/ton coal * 24 hours/day = 924.0 lb PM10/day

The remaining emission points (mostly handling of materials) had only PM10 emissions and the emissions were calculated based on maximum hourly design rate and emission factor for the particular operation. These emission points are further documented in Appendix I of the SIP.  The University of Missouri  -  Columbia originated the refined meteorological dataset for this analysis. 

The results of the refined analysis for Mingo, Hercules, and Upper Buffalo are presented in Table 19 including the 8th-highest (98 percentile) values. The same PM10 direct conversion to PMF calculation methodology was utilized to determine visibility impacts in CALPOST as with the other boilers reviewed.


    Table 19- Holcim  -  Clarksville Refined Analysis Results

The refined modeling impacts from Holcim  -  Clarksville exceed the 0.5 deciview threshold for all three Class 1 areas for both the Method 2 and the Method 6 CALPOST methodology.  Missouri determined this source would be subject to BART.



EPA Findings:
Missouri has appropriately identified the BART-eligible sources and those units subject to a further BART controls analysis.  Missouri successfully developed and provided the technical analyses to support the BART assessment provision for the non-CAIR units making a demonstration that certain units can be exempted from further BART controls based on CALPUFF modeling results.  For the CAIR units the State demonstrated that the PM emissions from the entire group would have impacts less than 0.5 deciview, therefore, these sources were exempted from further BART analysis for PM controls. 

The CALPUFF modeling system was appropriately applied by Missouri in the assessment of visibility impairment in Class I areas for specific sources subject to the BART Rule.  The BART Rule indicates 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 NNSR/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 Missouri to use the CALPUFF modeling system for its BART exemption modeling. Missouri established a 0.5 deciview threshold for making its BART exemption determination based on guidance contained in the Guidelines for BART Determinations. After going through the analysis described above MO found one source, Holcim  -  Clarksville that could not be exempted from further BART analysis.  This source is addressed below in the sources subject to BART section.

References used in the EPA review: 

   1. 40 CFR Part 5, Appendix W: Guideline on Air Quality Model
   2. Appendix Y to 40 CFR Part 51, the Guidelines for BART Determinations Under the Regional Haze Rule
   
      3.	BART Determination and Statutory Factors
         SIP Starting at Page 73 and Appendix N
      
For each BART-eligible unit in the State that is 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) is required. To address the requirements for BART, the State must take into consideration all of the  "Five Factors" from 40 CFR 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. 

 As indicated above Missouri found only one unit Holcim  -  Clarksville that is required to perform a source-specific BART determination.  The remaining BART-eligible units were EGU's included in CAIR or were sources exempted from further review based on their visibility impacts.  The State's submittal regarding BART for Holcim-Clarksville states as follows:

Therefore, MDNR contacted Holcim to pursue controls of SO2 and NOX for BART. The PM10 impact on visibility was less than 1 percent and did not constitute enough impact to pursue control. Holcim initially responded (prior to the March 2008 submittal) that they were in the process of retrofitting the kiln under permit #082007-019 that required them to install a mid-kiln firing system for NOX control. As a point of reference, MDNR determined that mid-kiln firing was equivalent to a 30 percent reduction from 1990 cement kiln NOX emissions based on EPA information provided during the development of the NOX  SIP call.

After the March 2008 SIP submittal, Holcim provided additional information detailing the maximum 24-hour SO2 and NOX emissions using continuous emission monitoring (CEM) data.  Holcim installed the CEMs in 2004, began collection of hourly emission data, and has maintained the necessary quality assurance measures to ensure accuracy of this data. The use of maximum 24-hour emission rates is the preferred methodology for determination of visibility impacts from BART sources (if available). MDNR utilized the maximum 24-hour emission rates provided by Holcim with the necessary filling of missing hourly data. The PM10 emission rates used in the new analysis were identical to the previous screening and refined modeling analyses. The resultant emissions obtained from the CEM data were:

SO2  -  4,900.95 lb SO2/hr = 117,622.7 lb SO2/day on 12/2/2004
NOX  -  3049.39 lb NOX/hr = 73,185.3 lb NOX /day on 11/24/2007

These emissions were modeled using the refined analysis method. The Method 6 results are presented in Table 20.


     Table 20- Holcim  -  Clarksville Post Submittal Refined Analysis Results

Because the analysis found that the unit was not able to be excluded from further BART analysis, Missouri requested the source submit a full BART analysis.  Holcim's BART submittals are included in Appendices N-P of the State SIP and relevant portions are excerpted herein.  :
 
Subsequently, Holcim performed an analysis to determine BART for each visibility affecting pollutant (VAP) for the kiln. The VAPs emitted by the kiln include NOx, SO2, and particulate matter with a mass mean diameter smaller than ten microns (PM10) of various forms (filterable coarse particulate matter [PMc], filterable fine particle matter [PMf], elemental carbon [EC], inorganic condensable particulate matter [IOR CPM] as sulfates [SO4], and organic condensable particulate matter [OR CPM] also referred to as secondary organic aerosols [SOA]). 

            a.	Identification of Available Retrofit SO2 Control Technologies
	
Step 1 of the BART determination identified all available retrofit SO2 control technologies. A list of control technologies was obtained by reviewing the U.S. EPA's Clean Air Technology Center, publicly-available air permits, applications, and technical literature published by the U.S. EPA, state agencies, and Regional Planning Organizations (RPOs).
The available retrofit SO2 control technologies are summarized in Table 24.


         Table 24- Available SO2 Control Technologies at Holcim

Step 2 of the BART determination is to eliminate technically infeasible SO2 control technologies that were identified in Step 1. Missouri's SIP states as follows.  

 Fuel Substitution 

Holcim uses a mixture of coal, petroleum coke, alternative fuels (synfuel), and oil as the primary fuels for the kiln. For example, the 2007 actual fuel usage breakdown on an energy input basis was 3.4 percent coal, 84 percent petroleum coke, 11.8 percent synfuel, and 0.8 percent oil (the fuel usages are also on an as received basis). The sulfur content of the petroleum coke is approximately 5.72 percent and the sulfur content of the coal is approximately 3.45 percent.

The design of the long wet kiln system is such that some of the SO2 resulting from fuel combustion may be emitted and the rest is absorbed in clinker or CKD. Therefore, if Holcim reduces sulfur in the fuel input to the kiln; a corresponding reduction in SO2 emissions from the kiln would be expected. Fuel sulfur content could be reduced by burning a coal with a lower sulfur content of 0.7 percent, in lieu of the current coal/coke, which would result in a lower overall fuel sulfur content.

Determining the specific reduction in SO2 emissions from a reduction in fuel sulfur is complicated as the reactions in the kiln system are complex. To calculate an SO2 control effectiveness, based on switching to a lower sulfur fuel, a high number of assumptions must be made with a very low confidence in the accuracy. 

Based on the 2007 data, if all of the current coal and coke is replaced with low sulfur coal, the sulfur input from fuel is calculated to be reduced by approximately 85 percent. The low sulfur coal has a lower heat content (26 Gj/Metric ton) and higher moisture content than the coke currently being used (33 Gj/Metric ton), so a higher volume of low sulfur coal is needed than coke reduced. These reductions are the maximum reductions in tons of SO2 that can be expected, especially if overall emissions of SO2 increase from other sources (such as raw materials).
The actual reduction will vary significantly, especially on a short term basis. Holcim estimates that the net reduction in SO2 would be in the range of 40 percent to 50 percent. Holcim considers this technology to be technically feasible and will consider it further.

 Raw Material Substitution / Selective Mining 

In a long wet kiln, not only the pyritic sulfur, but total sulfur in the raw materials will have an impact on SO2 emissions. Part of the pyritic sulfur reacts with oxygen and forms SO2 at the relatively lower temperature zone of the kiln. The rest of the sulfur, such as sulfates and sulfur compounds, enters the kiln at higher temperature zones, where more SO2 is volatized. Some of this SO2 will pass the length of the kiln without being absorbed and will thus be emitted to the stack.

Using raw materials with lower sulfur content can reduce the potential for SO2 emissions from a wet kiln system. The limestone, shale, and other raw materials used at the Clarksville plant contain pyrites and total sulfur in varying concentrations. If zones or layers in the on-site quarry could be identified and mined selectively such that lower sulfur content materials are used, the emission rate of SO2 could be reduced. Holcim has conducted a complete quarry investigation, and based on the quarry scheduling optimization (QSO) model and computation, after 1-2 years of mining, the plant will be facing an area that has a higher total sulfur content than that currently being used.

Thus, there will be higher total sulfur content in the raw material in the next 3-10 years. The total sulfur in the raw materials is expected to increase by an additional ~30% without selective quarrying, and by ~20% if some form of selective quarrying can be identified. Similarly, no significant reduction of pyritic sulfur from selective mining is anticipated.
Based on the lack of available data to predict if selective mining would be feasible, Holcim has concluded that selective mining is not considered a technically feasible SO2 control technology for the kiln.

 Dry Lime Injection / Dry Lime Scrubbing  

Dry Lime Injection, or Dry Lime Scrubbing (DLS), consists of injecting hydrated lime, Ca(OH)2, into the flue gas. The Ca(OH)2 reacts with SO2 in the flue gas stream to create fine particles of CaSO3 or CaSO4. The particles are collected in the particulate matter control device (PMCD) serving the kiln. The current PMCD was not sized/designed to handle the additional particulate matter loading that would result from this technology. Consequently, adding DLS could cause PM emissions and opacity to increase above permitted levels requiring Holcim to replace the existing PMCD (an ESP) with a new PMCD (a Baghouse).

Holcim is aware of only one other long term application of this technology on a wet kiln, which is on a smaller wet kiln in Belgium. Consequently, very little data exists to directly quantify the feasibility or benefit (emission reduction) of such a system. Regardless, Holcim considers the technology to be technically feasible.

 Wet Lime Scrubbing
         
Wet lime scrubbing (WLS) is a name for a traditional tailpipe wet scrubber. This process involves passing the flue gas from the main PMCD through a sprayed aqueous suspension of Ca(OH)2 or CaCO3 (limestone) that is contained in an appropriate scrubbing device. In the case of the Clarksville plant, the basic underlying economics would dictate the use of ground limestone as the scrubbing reagent. Use of the cement kiln dust as a scrubbing reagent was not considered as a viable option for Clarksville due to its high chlorine content and a large amount of inert. In WLS, the aqueous suspension of scrubbing reagent is not taken to dryness as it is in DLS. The SO2 reacts with the scrubbing reagent to form CaSO3.H20 or gypsum that is collected and retained as aqueous sludge. The sludge is either dewatered and disposed of or used as synthetic gypsum.

The scrubbing efficiency of WLS can vary from an estimated 80 percent to 95 percent of the SO2 in the flue gas treated by the scrubber5. Further, WLS is a high maintenance process with high rates of downtime expected from build up or plug up of mist-eliminators or spray nozzles and the severe wear and corrosion of components. Holcim has found that high levels of hydrocarbons (THC) in the gas stream have caused significant corrosivity and foam build-up at their Dundee plant. Further, it significantly influences the system availability and the efficiency. The THC levels at the Clarksville plant may also lead to build up and plugging, and thus availability (uptime) of the WLS of 95 percent is assumed.

Despite these identified drawbacks, WLS is considered a technically feasible BART option.

            b.	Rank and Evaluation of Technically Feasible SO2 Control 		Options 

The third step in the BART analysis is to rank the technically feasible options according to effectiveness. Table 25 provides the effectiveness of each technology in the form of an annual average efficiency.

      Table 25- Ranking of Technically Feasible Kiln SO2 Control Technologies

                  c.	Evaluation of Visibility Impacts for Feasible SO2 Controls

Step four of the BART analysis procedure is the impact analysis. The BART determination guidelines list the four factors to be considered in the impact analysis:

:: Cost of compliance
:: Energy impacts
:: Non-air quality impacts; and
:: The remaining useful life of the source

Holcim has conducted an impact analysis for the remaining SO2 control options. Missouri's impact analysis from the State SIP states as follows. 

 Wet Lime Scrubbing

      Cost of Compliance

Holcim utilized a recent WLS vendor bid as the basis for the economic analysis to determine the annualized cost for WLS. Holcim divided the annualized cost of WLS by the annual tons of SO2 reduced to determine the cost effectiveness for WLS. The "annual tons reduced" were determined by subtracting the estimated controlled annual emissions from the existing annual emissions. The existing annual emissions should be considered both on a projected actual and a potential to emit (PTE) basis. The projected actual (PA) annual SO2 emissions provided to the MDNR in the recent Mid Kiln Firing (MKF) permit application was 11,481 tons/year. The PTE listed in the MKF permit is 13,298 tons/year. 

The estimated controlled annual emissions were calculated by applying 80 percent to 95 percent control efficiency and a 95 percent control device uptime, to the projected actual annual and PTE emissions. Table 26 provides a summary of the cost effectiveness analysis related to WLS. The detailed cost analysis table is provided in Appendix P.


    Table 26- Cost Analysis Summary for Wet Lime Scrubbing

The significant increases in cost per ton of clinker produced from using WLS, as shown in Table 26, would likely eliminate any profit margin currently realized by the plant. Thus, it would not be economically feasible to operate the plant with WLS.

      Energy Impacts

A wet scrubber requires an additional fan of considerable horsepower to move the flue gas through the scrubber. The exhaust gas reheat requirement will utilize approximately 1,000,000 MMcf/year of natural gas, which will itself lead to an increase in NOx emissions of 97 tons/yr.

      Non Air-Quality Impacts

Without reheating, a frequent steam plume and/or detached plume can be expected at the discharge of the wet scrubber that would result in visual impairment in the area. The WLS technology could generate over 50,000 tons per year of waste (sludge) that will require disposal in a landfill.

      Remaining Useful Life

The remaining useful life of the kiln does not impact the annualized cost of WLS because the useful life is anticipated to be at least as long as the capital cost recovery period, which is 15 years.
 Fuel Substitution 
      
      Cost of Compliance

The increased cost of using low sulfur coal includes the relative increase in fuel cost as well as the cost of a new coal mill system. Low sulfur coal is harder than the current coal/coke utilized and has lower heat content; consequently, a higher volume of coal grinding will be needed than the current mill can achieve. The increased grinding requirement would also have an additional energy requirement. A bid for a new coal mill, classifier, and mill motors, was obtained from GEBR. Pfeiffer USA Inc. Table 27 provides a summary of the cost effectiveness analysis related to Fuel Substitution. The detailed cost analysis table is provided in Appendix P.

The cost effectiveness analysis does not include the cost to construct any new storage or handling facilities for the low sulfur coal that may be required. The control cost factors were obtained from the EPA's Control Cost Manual, 6th Edition. Some of the factors have been scaled, as indicated, based on the construction being a retrofit, rather than greenfield, and company knowledge of the actual cost of similar size/type projects.


    Table 27- Cost Analysis Summary for Fuel Substitution 

The 3.0 percent sulfur coal option (used 23 percent control as the average) is cost effective and will be considered further.

The low sulfur CO coal option (40 to 50 percent control) is not cost effective and results in significant increases in cost per ton of clinker produced that would likely eliminate any profit margin currently realized by the plant. Thus, it would not be economically feasible to operate the plant with the low sulfur CO coal scenario.

In addition, for this analysis Holcim did consider the blending of solid fuel but considered it also to not be cost effective. In order to properly blend solid fuels for a precise distribution to control sulfur and chlorine inputs and not disrupt production due to fuel quality swings, an additional solid fuel equipment handling system would have to be installed. At a minimum, for each additional solid fuel used a system consisting of feed chutes, transfer conveyors, a storage silo, and a weight belt would be required. It was conservatively estimated that the cost would be similar to the cost of a new mill and thus increase the cost well above the 3.0% sulfur coal option.

      Energy Impacts

The low sulfur CO coal will require additional energy for grinding. The actual increase is difficult to estimate. The regional coal option also relies that a reliable source of tire derived fuel (TDF) and Hazardous Waste Derived Fuel (HWDF) is available to supplement 23% of the thermal requirements for clinker production

      Non Air-Quality Impacts

None.

      Remaining Useful Life

The remaining useful life of the kiln does not impact the annualized cost because the useful life is anticipated to be at least as long as the capital cost recovery period, which is 15 years. However, the existing coal mill would need to be replaced at a significant expense.

 Dry Lime Scrubbing

      Cost of Compliance

The increased cost of DLS includes the cost of hydrated lime as well as the injection system and replacing the existing ESP with a new baghouse. As the DLS injection system would likely be a custom application, Holcim's engineering department has estimated that the DLS injection system equipment cost would be approximately $1,000,000. Holcim obtained a bid for retrofitting the existing ESP with a Baghouse from GE Energy. A detailed analysis has not yet been completed to determine if the retrofit will be capable of handling the additional dust loading from the DLS system.

Table 28 provides a summary of the cost effectiveness analysis related to DLS. The detailed cost analysis table is provided in Appendix A. The control cost factors were obtained from the EPA's Control Cost Manual, 6th Edition. Some of the factors have been scaled, as indicated, based on the construction being a retrofit, rather than greenfield, and company knowledge of the actual cost of similar size/type projects. 


    Table 28- Cost Analysis Summary for Dry Lime Scrubbing 

The significant increases in cost per ton of clinker produced from using DLS, as shown in
Table 28, would likely eliminate any profit margin currently realized by the plant. Thus, it would not be economically feasible to operate the plant with DLS.

      Energy Impacts

Additional electricity is needed for the pump used to inject the lime into the kiln gas.

      Non Air-Quality Impacts

Utilizing DLS could also increase the amount of [cement kiln dust] sent to the landfill.

      Remaining Useful Life

The remaining useful life of the kiln does not impact the annualized cost because the useful life is anticipated to be at least as long as the capital cost recovery period, which is 15 years.
A final impact analysis was conducted to assess the visibility improvement for the existing emission rate when compared to the emission rate of WLS, Fuel Substitution, and DLS. The existing emission rates, and emission rates associated with controls, were modeled using CALPUFF. The existing emission rate is the same rate that was modeled for the BART applicability analysis. The SO2 emission rate associated with WLS, Fuel Substitution, and DLS is the existing emission rates less the average anticipated control of 87.5 percent, 45 percent, and 25 percent respectively. The emission rates are summarized in Table 29.

           Table 29- Summary of Emission Rates Modeling in SO2 Control Visibility Impact 

Comparisons of the existing visibility impacts and the visibility impacts based on WLS, Fuel Substitution, and DLS are provided in Table 30. The visibility improvement associated with the controls are also shown in Table 30; this value was calculated as the difference between the existing visibility impairment and the visibility impairment for the controlled emission rates as measured by the 98th percentile modeled visibility impact.


      Table 30- Summary of Modeled Impacts from SO2 Control Impact Analysis

In order to determine BART for SO2, Holcim reviewed each control option's availability, as well as its cost of compliance, energy impacts, and non-air quality impacts, as well as the remaining useful life of the kiln. Table 31 summarizes the cost effectiveness for the controls based on the tons of SO2 reduced and the visibility improvement in deciviews.

           Table 31- Summary of Cost Effectiveness of SO2 Control Options

Based on the five step analysis outlined by EPA, this analysis demonstrates that the cost of compliance associated with WLS, Fuel Substitution with low sulfur CO coal, and DLS is extremely high on a $/ton of SO2 removed basis, and especially on a $/DV of improvement basis, and is not economically feasible, especially if the cost per ton of clinker produced is taken into account. Fuel Substitution with a regional 3.0 percent sulfur coal, resulting in a 23 percent reduction in SO2 emissions, is economically feasible and is being proposed as BART. Holcim proposes that BART for SO2 for the Holcim Clarksville Kiln is fuel substitution or equivalent technologies that will achieve a 23 percent reduction in the maximum daily SO2 emission rate of 117,345 lbs.
Consistent with BART applicability which is determined on a daily basis, Holcim proposes to comply with an enforceable limit for SO2 of 90,356 lbs/day which is a 23% reduction in the maximum daily SO2 emission rate of 117,345 lbs.

      d.	Identification of Available Retrofit NOx Control Technologies

In Portland cement kilns, the NOx that is generated is primarily classified into one of two categories, i.e., thermal NOx or fuel NOx. Thermal NOx occurs as a result of the high-temperature oxidation of molecular nitrogen present in the combustion air. Fuel NOx is created by the oxidation of nitrogenous compounds present in the fuel. It is also possible for nitrogenous compounds to be present in the raw material feed and become oxidized to form additional NOx referred to as feed NOx. Due to the high flame temperature in the burning zone of the rotary kiln (3400[o] F), NOx emissions from the kiln tend to be mainly comprised of thermal NOx. Although NOx emissions from cement kilns include both nitrogen oxide (NO) and nitrogen dioxide (NO2), typically, less than 10% of the total NOx in the flue gas is NO2.

The Holcim kiln is the only BART source which emits NOx, thus a NOx BART evaluation was performed only for the kiln. The maximum actual 24-hour kiln NOx emission rate that was modeled for the BART applicability determination is summarized in Table 32. The NOx 24-hour maximum actual emission rate was determined from analyzer data for November 24, 2007.


       Table 32- Existing Actual Maximum 24-Hour NOx Emission Rate

Step 1 of the BART determination is the identification of all available retrofit NOX control technologies. A list of control technologies was obtained by reviewing the U.S. EPA's Clean Air
Technology Center, control equipment vendor information, publicly-available air permits, applications, and technical literature published by the U.S. EPA and the RPOs. The available retrofit NOX control technologies are summarized in Table 33.


              Table 33- Possible NOx Control Technologies

Step 2 of the BART determination is to eliminate technically infeasible NOX control technologies that were identified in Step 1.

 Low-NOx Burner in the Rotary Kiln

Low- NOx burners (LNBs) reduce the amount of NOx formed at the flame. The principle of all LNBs is the same: stepwise or staged combustion and localized exhaust gas recirculation (i.e. at the flame). As applied to the rotary cement kiln, the low-NOx burner creates primary and secondary combustion zones at the end of the main burner pipe to reduce the amount of NOx initially formed at the flame. In the high-temperature primary zone, combustion is initiated in a fuel-rich environment in the presence of a less than stoichiometric oxygen concentration. The oxygen-deficient condition at the primary combustion site minimizes thermal and fuel NOx formation and produces free radicals that chemically reduce some of the NOx that is being generated in the flame.

In the secondary zone, combustion is completed in an oxygen-rich environment. The temperature in the secondary combustion zone is much lower than in the first; therefore, lower NOx formation is achieved as combustion is completed. CO that has been generated in the primary combustion zone as an artifact of the sub-stoichiometric combustion is fully oxidized in the secondary combustion zone. Low-NOx burners are considered to be a technically feasible option for NOX control. Holcim already operates a LNB, therefore the technology will not be considered further.

 Flue Gas Recirculation

Flue gas recirculation involves the use of oxygen-deficient flue gas from some point in the process as a substitute for primary air in the main burner pipe in the rotary kiln. Flue gas recirculation (FGR) lowers the peak flame temperature and develops localized reducing conditions in the burning zone through a significant reduction of the oxygen content of the primary combustion "air." The intended effect of the lower flame temperature and reducing conditions in the flame is to decrease both thermal and fuel NOx formation in the rotary kiln.
While FGR is a practiced control technology in the electric utility industry, Holcim is not aware of any attempt to apply FGR to a cement kiln because of the unique process requirements of the industry, i.e., a hot flame is required to complete the chemical reactions that form clinker minerals from the raw materials. The process of producing clinker in a cement kiln requires the heating of raw materials to about 2700°F for a brief but appropriate time to allow the desired chemical reactions that form the clinker minerals to occur. A short, high-temperature flame of about 3400°F is necessary to meet this process requirement. The long/lazy flame that would be produced by FGR would result in the production of lower or unacceptable quality clinker because of the resulting undesirable mineralogy. Clinkering reactions must take place in an oxidizing atmosphere in the burning zone to generate clinker that can be used to produce acceptable cement. FGR would tend to produce localized or general reducing conditions that also could detrimentally affect clinker quality. Due to these important limitations on the application of FGR and the lack of a successful demonstration on cement kiln in the United States, FGR is not a technically feasible control option for NOx control at this time.

 Cement Kiln Dust Insufflations
 
Cement kiln dust (CKD) is a residual byproduct that can be produced by any of the four basic types of cement kiln systems. CKD is most often treated as a waste even though there are some beneficial uses. However, as a means of recycling usable CKD to the cement pyroprocess, CKD sometimes is injected or insufflated into the burning zone of the rotary kiln in or near the main flame. The presence of these cold solids within or in close proximity to the flame has the effect of cooling the flame and/or the burning zone thereby reducing the formation of thermal NOx. The insufflations process is somewhat counterintuitive because a basic requirement of a cement kiln is a very hot flame to heat the clinkering raw materials to about 2700[o] F in as short a time as possible. The Clarksville plant already uses this technology and it is already included in the baseline. 

 Mid  - Kiln Firing of Solid Fuel With Mixing Air Fan 

Secondary combustion is defined as follows: a portion of the fuel is fired in a location other than the burning zone. This reduces thermal NOx generation because the temperature in the secondary combustion zone is less than 2100 °F. Mid-kiln firing (MKF) of solid fuels is an example of secondary combustion and includes fuels such as used tires, oil filter fluff, plastics, spent activated carbon and carbon black, asphalt shingles, diaper manufacturing waste, and other combustible solids. MKF improves clinker process energy efficiency, allows for greater operational flexibility with respect to fuel types, and is currently listed as a NOx control technology in 10 CSR 10-6.380 Control of NOx Emissions from Portland Cement Kilns.
By adding fuel mid-kiln, MKF changes both the flame temperature and flame length.  These changes should reduce thermal NOx formation by burning part of the fuel at a lower temperature and by creating reducing conditions at the mid-kiln fuel injection point which may destroy some of the NOx formed upstream in the kiln burning zone.

Clarksville has the largest long kiln in the world. The kiln has a 7 meter diameter and a very high thermal capacity. Using whole tires to replace 10% of total fuel consumption will require four whole tires being fed to the mid-kiln door per kiln revolution, 12% fuel replacement would require 5 tires per revolution. The maximum tire feed rate per revolution that Holcim is aware of, on similar applications, is three tires per revolution.

Holcim is concerned that the greater the number of tires fed per revolution, on a continuous basis, the greater the potential for process upsets from unstable feeding. Holcim has found that kilns being fed even one to three tires per revolution can have problems with stable, uniform feeding. In addition, if too many tires burn at the bottom of the kiln, a high local temperature could result, and then disturb the normal operation of the kiln and potentially increase NOx. Further, due to the large kiln diameter, the reducing zone created by burning tires may only impact a small cross section of the entire cross section of the kiln, thus having less of an overall reduction in NOx than anticipated.

In an effort to better understand these uncertainties, Holcim hired CINAR Company, the expert in this field, to conduct modeling of the system and to predict NOx reduction. Their study predicted that a 15% NOx reduction would occur at 10% replacement (replace 10% of the current fuel with tires) and 27% NOx reduction for 15% replacement.

To determine the thermal substitution rate (TSR) of tires that the Clarksville plant is capable of utilizing, three additional factors were considered:
1. Tire availability. The local market only has sustainable resources of 10-12% TSR;
2. Tire feeding limit: 12% TSR equates to five tires being fed per revolution.
3. The thermal stability of operation in a large kiln. 15% TSR is predicted to be the maximum for short term periods, whereas 10-12% TSR is predicted to be achievable on a long term basis.

Based on the lack of experience using MKF of tires on kilns the size of Clarksville, Holcim is relying on the computer modeling (regardless of the general uncertainty that exists with computer models) to estimate the NOx reduction. Holcim anticipates that MKF of tires may achieve up to 20% percent NOx reduction at a TSR of 12 percent on a long term basis. MKF is considered to be a technically feasible option for NOx control. Further, Holcim has already received a construction permit that would allow the installation of MKF, whereas other technologies would require a new construction permit application process, the result of which is unknown.

 Selective Noncatalytic Reduction

In the relatively narrow temperature window of 1600o to 1995°F, ammonia (NH3) reacts with NOx without the need for a catalyst to form water and molecular nitrogen in accordance with the following simplified reactions.

4NO + 4NH3 + O2 --> 4N2 + 6H2O
2NO2 + 4NH3 + O2 --> 3N2 + 6H2O

As applied to NOx control from cement kilns and other combustion sources, this technology is called selective noncatalytic reduction (SNCR). Above this temperature range, the NH3 is oxidized to NOx thereby increasing NOx emissions. Below this temperature range, the reaction rate is too slow for completion and unreacted NH3 may be emitted from the pyroprocess. This temperature window generally is available at some location within the rotary kiln. The NH3 could be delivered to the kiln shell through the use of anhydrous NH3, or an aqueous solution of NH3 (ammonium hydroxide) or urea.

Based on the concerns with NH3 slip at high molar ratios, and the uncertainty regarding the level of effectiveness of the reagent in Clarksville's large diameter kiln, Holcim anticipates that at a molar ratio of about 1.0, an average annual control efficiency of 20 percent could likely be achieved without excessive NH3 slip. However, a pilot study would need to be conducted to verify this. Regardless, SNCR is considered to be a technically feasible option for NOX control.

 Selective Catalytic Reduction

Selective Catalytic Reduction (SCR) is an add-on control technology for the control of emissions of the oxides of nitrogen (NOx) from a combustion process. SCR has been successfully employed in the electric power industry. The basic SCR system consists of a system of catalyst grids placed in series with each other within a vessel that is located in a part of the process where the normal flue gas temperature is in the required range. An ammonia-containing reagent is injected at a controlled rate upstream of the catalyst grids that are designed to ensure relatively even flue gas distribution within the grids, to provide good mixing of the reagent and flue gas, and to result in minimum ammonia (NH3)slip. The NH3 reacts with NOx compounds (i.e., NO and NO2) on the surface of the catalyst in equal molar amounts (i.e., one molecule of NH3 reacts with one molecule of NOx). Common reagents include aqueous NH3, anhydrous NH3 and urea [(NH2)2CO]. In the presence of the catalyst, the injected ammonia is converted by OH- radicals to ammonia radicals (i.e., NH2 -), which, in turn, react with NOx to form N2 and H2O. The SCR catalyst enables the necessary reactions to occur at lower temperatures than those required for Selective Non-Catalytic Reduction (SNCR). While catalysts can be effective over a larger range of temperatures, the optimal temperature range for SCR is 570 - 750º F.

Because the reaction rate of NH3 and NOx is temperature dependent, the temperature of the flue gas stream to be controlled is the most important consideration in applying SCR technology to any combustion source. The optimum temperature range for SCR application is about 300º C (570º F) to 450º C (840º F). This range of normal process temperature would occur within the kiln of a long wet kiln, rather than in the exhaust gas between the wet kiln and the PMCD inlet.
SCR has not been applied to any wet cement plant in the world and is not considered an available technology.

A technology is considered "available" if it can be obtained by the applicant through commercial channels or is otherwise available within the common sense meaning of the term. An available technology is "applicable" if it can reasonably be installed and operated on the source type under consideration. A technology that is available and applicable is technically feasible. Availability in this context is further explained using the following process commonly used for bringing a control technology concept to reality as a commercial product: concept stage; research and patenting; bench scale or laboratory testing; pilot scale testing; licensing and commercial demonstration; and Commercial sales.

A control technique is considered available, within the context presented above, if it has reached the licensing and commercial sales stage of development. A source would not be required to experience extended time delays or resource penalties to allow research to be conducted on a new technique. Neither is it expected that an applicant would be required to experience extended trials to learn how to apply a technology on a totally new and dissimilar source type. Consequently, technologies in the pilot scale testing stages of development would not be considered available. An exception would be if the technology were proposed and permitted under the qualifications of an innovative control device consistent with the provisions of 40 CFR 52.21(v) or, where appropriate, the applicable SIP [in which case it would be considered available].

Therefore, SCR is eliminated from further consideration as BART for NOx control at the Clarksville plant.

            e.	Rank / Evaluation of Technically Feasible NOx Control 				Options 
	
The third step in the BART analysis is to rank the technically feasible options according to effectiveness. Table 34 presents potential NOx technically feasible control technologies by effectiveness.


           Table 34- Ranking of Technically Kiln NOx Control Technologies

Step four for the BART analysis is the impact analysis. The BART determination guidelines list four factors to be considered in the impact analysis:

:: Cost of compliance
:: Energy impacts
:: Non-air quality impacts; and
:: The remaining useful life of the source

Missouri considered the following in its impact analysis.  
 MKF

      Cost of Compliance

Holcim anticipates that MKF and SNCR have relatively the same level of effectiveness.
Because SNCR would require a pilot study to prove or verify the effectiveness of NOx reduction and the potential associated opacity issues due to ammonia slip (see Appendix B), Holcim is accepting the use of MKF as BART. Because MKF, the most stringent control option available is BART, the cost of compliance was not required to be evaluated. Regardless, a cost analysis has been conducted for informational purposes. The detailed cost analysis table is provided in Appendix P. The control cost factors were obtained from the EPA's Control Cost Manual, 6th Edition.

      Energy Impacts and Non Air-Quality Impacts
      
There are no known adverse energy or non-air impacts from MKF. MKF of tires has the benefit of eliminating tires from landfills and illegal dumping. It also reduces CO2 emissions (a Green House Gas) and reduces fossil fuel use.

      Remaining Useful Life

The remaining useful life of the kiln does not impact the annualized costs of MKF because the useful life is anticipated to be at least as long as the capital cost recovery period, which would be 15 years.

         		f.	Evaluation of Visibility Impact of Feasible NOx Controls
         
The final impact analysis was conducted to assess the visibility improvement for existing emission rates when compared to the emission rate with MKF. The existing emission rates and emission rates associated with MKF were modeled using CALPUFF. The existing emission rates are the same rates that were modeled for the BART applicability analysis. The NOx emission rate associated with MKF was the existing emission rate less an average reduction of 20 percent. The emission rate is summarized in Table 35.

     Table 35- Summary of Emission Rates Modeling in NOx Control Visibility Impact 

Comparisons of the 98th percentile existing visibility impacts and the visibility impacts based on MKF are provided in Table 36. The visibility improvement associated with MKF are also shown in Table 35; this was calculated as the difference between the existing visibility impairment and the visibility impairment for the remaining control options as measured by the 98th percentile modeled visibility impact.

         Table 36- NOx Control Visibility Impact Analysis

As seen in Tables 36, the MKF option results in a visibility improvement of up to 11.1 percent in the Hercules Glades Class I area.

Based on the five-step analysis outlined by EPA, MKF was identified as the highest ranking feasible add-on control technology. Economic, energy and environmental impacts were assessed for this technology and the visibility improvements were evaluated against existing conditions. The visibility impact analysis demonstrates that the utilization of MKF to achieve a 2,440 lb/hr NOx emission rate results in up to an 11.1 percent visibility improvement. Neither non-air quality nor energy impacts associated with this control technology eliminate it in favor of retaining the existing rates as BART.

Holcim has determined that BART for the Holcim Clarksville Kiln is the installation and operation of a Mid Kiln Firing (MKF) system or equivalent that will achieve a 20 percent reduction in the maximum daily NOX emission rate of 73,185 lbs. Consistent with BART applicability, which is determined on a daily basis, Holcim proposes to comply with an enforceable limit for NOx of 58,548 lbs/day which is a 20% reduction in the maximum daily NOX emission rate of 73,185 lbs.

Based on the lack of site specific, or significant industry data, for the use of this technology on wet cement kilns, it is possible that Holcim will further evaluate the MKF system and determine that MKF results in limited or no additional benefit. In the future, an alternative technology or methodology may become feasible and could be implemented as needed. Holcim will continue to utilize the NOx controls that are already in place, including LNB, insufflations, and the use of alternative fuels as available.

In summary, based on the five-step analysis, Holcim proposes the following as BART:

:: PM10  -  Holcim has determined that the existing electrostatic precipitator constitutes BART.
This control device is effective for controlling PM10 from a wet kiln.

:: NOX  -  Holcim has determined that BART for the Holcim Clarksville Kiln is the installation and operation of a Mid Kiln Firing (MKF) system or equivalent technologies that will achieve a 20 percent reduction in the maximum daily NOX emission rate of 73,185 lbs that was used for visibility impact modeling.

:: SO2  -  Holcim proposes that BART for the Holcim Clarksville Kiln is fuel substitution or equivalent technologies that will achieve a 23 percent reduction in the maximum daily SO2 emission rate of 117,345 lbs that was used for visibility impact modeling.

The proposed BART control strategies will result in reductions of the visibility impacts attributable to the Clarksville plant. A summary of the visibility improvement at Class I areas based on the existing emission rates and proposed BART emission rates is provided in Table 37.

         Table 37- Visibility Impairment Improvement

Holcim proposed to implement BART on or before 5 years after EPA approval of Missouri's Regional Haze Rule State Implementation Plan as per 40 CFR 51 Appendix Y. Because the actual implementation date is at least 5 years from the date of this document, Holcim requested that it not be bound to the technologies reviewed during this analysis, but to the daily limits stated in the report. Changes in technology may offer opportunities to obtain even better reductions with lower cost impact. Consistent with BART applicability which is determined on a daily basis, Holcim proposed to comply with an enforceable limit for SO2 of 90,356 lbs/day which is a 23% reduction in the maximum daily SO2 emission rate of 117,345 lbs. and a NOx limit of 58,548 lbs/day which is a 20% reduction in the maximum daily NOX emission rate of 73,185 lbs.

Missouri comprehensively reviewed the source's BART analysis and determined that the mid-kiln firing of tires (using 12 percent total heat input substitution) and a switch from petroleum coke as the primary kiln fuel to 3 percent sulfur coal (along with the tire derived fuel for NOx control) would be BART for this source.  Missouri and the source did not initially agree on the emission limits proposed by Holcim.  Missouri required the source to pursue more aggressive emission limits than originally recommended based on the cost analysis of feasible controls. The cost analysis Missouri relied upon for their BART determination is found in Appendix Q.  Pages 77 and 78 of the SIP state as follows. 

The SO2 finding was based primarily on the economic evaluation of several different control scenarios for the kiln including both wet and dry scrubbing of the kiln exhaust gas. The department found the overall net cost per ton of cement produced (~$15-20/ton) was excessive and would have seriously compromised the Clarksville kiln's ability to compete in the cement production market. The fuel switch provided a significant amount of emission reductions at a cost of approximately $3/ton cement produced. The overall cost per ton SO2 reduced was calculated as $1,148/ton removed. The NOX finding was based on information provided by Cinar (kiln design contractor for Holcim) regarding NOX emission rates when burning tirederived fuel and the department's serious concern over the use of selective non-catalytic reduction on the world's largest long wet cement kiln. The increased certainty of the mid-kiln firing of tires, along with the use of low-NOX burners and cement kiln dust insufflation, provide a firm basis for the finding of NOX BART at Clarksville. In addition, this reduction of emissions was to have very little additional cost for the plant due to the planned installation of the mid-kiln firing project. The department findings result in a 20 percent reduction of NOX and a 27 percent reduction of SO2 from the maximum 30-day average emissions using the CEM data. As provided in the presumptive BART finding for utilities, the emission limits are expressed as 30- day rolling averages of 42,287 pounds NOX/day and 58,787 pounds SO2/day. The calculation of these averages will include all hours when the kiln is not operating and those emissions for both pollutants will be entered as zero pounds. 

The consent decree, included in Appendix S of the State's submittal, requires the Holcim  -  Clarksville Plant kiln system (Emission Point ID EP-14 main kiln stack) to meet the following rates and work practices, within 4 years after the EPA approves the State's RH SIP or expeditiously as practicable:

      1) NOx  -  42,287 lb/day using a 30 day rolling average
      2) SO2  -  58,787 lb/day using a 30 day rolling average
      3) the facility must monitor using existing CEMS  

The State also performed a visibility analysis on the BART rates and found significant reductions in the visibility impacts.  The SIP on page 77 states:

Refined visibility modeling was conducted with the BART limits prescribed by the department using the previous refined modeling analysis as a baseline to determine the extent of visibility improvement for each of the three impacted Class I areas. The results are provided in Table 9.18.



These results demonstrate less than the 0.5 deciview threshold for the 98th percentile visibility impact at the Hercules Glade and Upper Buffalo Wilderness Areas. Further, great improvement at the more proximate Mingo National Wildlife Refuge was also demonstrated. The highest 98[th] percentile impact was in 2001 and this impact was reduced 47 percent from the original refined modeling estimate.

EPA Findings:   Missouri has appropriately addressed this provision. The State appropriately documented the technical analysis to assess the need and implementation of BART controls as applicable for the Holcim emission units for PM, SO2 and NOx  emissions.  This involved identification of existing controls on applicable emission units, potential control options for applicable emission precursors, and the application of the Five Factors from 40 CFR 51.308(e)(1)(ii)(A) to determine the feasibility of controls that constituted BART.  
The State required Holcim to evaluate and submit three separate BART analysis for the source.  

For each BART-eligible source in the State subject to BART, the implementation plan should include emission limitations representing BART and schedules for compliance with BART.   Each BART-eligible source in the State subject to BART, shall be required to install and operate BART as expeditiously as practicable, but in no event later than 5 years after approval of the implementation plan revision, shall maintain the control equipment required by this subpart and shall establish procedures to ensure such equipment is properly operated and maintained. 

The State and source have also entered into a BART Final Consent Agreement on April 19, 2009, with federally enforceable controls that represent BART, included in Appendix S of the State SIP.  The Consent Agreement includes emission limits on the BART source for NOx and SO2 in Section 19 of the agreement.  Section 19 also requires the controls be installed as expeditiously as practicable but no later than four years after approval of Missouri's regional haze plan. Section 20 provides a description of the compliance demonstration requiring CEMS. 

Missouri has entered into a Consent Agreement with the only BART-subject source in the State other than the CAIR sources.  The Consent Agreement contains federally enforceable BART limitations along with a schedule for compliance before the required 5 year timeframe. The Consent Agreement also ensures continued compliance and maintenance of the equipment to be demonstrated with Continous Emissions Monitoring Systems (CEMS).

The State documented an improvement in visibility at affected Class I areas.  While post-BART control modeled impacts at Mingo are still slightly above 0.5 deciview the impairment has significantly improved with the proposed BART controls. After reviewing the methodology and analyses presented in the SIP and Appendices, EPA finds Missouri's BART implementation for Holcim to be acceptable.  

      E.  	2002 and 2018 Emissions Inventory (40 CFR 51.308(d)(4)(v)) (See also EPA 	TSD Section III.C.1.)
		SIP Pages 34-41

The plan 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 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.

            1.	State Submittal
            SIP Page 34

7.1 2002 AND 2018 EMISSIONS INVENTORY SUMMARY

As specified in the EPA guidance document, Emissions Inventory Guidance for Implementation
of Ozone and Particulate Matter National Ambient Air Quality Standards (NAAQS) and Regional Haze Regulations (August 2005), the regional haze emissions inventory includes carbon monoxide (CO), nitrogen oxides (NOx), sulfur dioxide (SO2), volatile organic compounds (VOCs), fine particulate (PM2.5), coarse particulate (PM10), and ammonia (NH3). Missouri used the CENRAP Base G emissions inventory for both the baseline year of 2002 and future year of 2018. Tables 7.1 and 7.2 summarize the Missouri 2002 and 2018 inventories, respectively. Tables H.1-8 in Appendix H include the complete 2002 and 2018 emissions inventory for Missouri.
                                       
             Table 7.1: 2002 Missouri Emissions Inventory Summary

Source Sector
 NOX (TPY*)
  SO2 (TPY)
 PM10 (TPY)
 PM2.5 (TPY)
  CO (TPY)
 VOC (TPY)
 NH3 (TPY)
Point EGU**
  145,437.9
  272,128.1
    4,093.2
    2,523.2
   11,357.0
    1,796.4
      19.2
Point NEGU***
   36,143.8
   97,117.0
   15,092.2
    7,045.3
  107,756.3
   38,473.6
    6,233.9
Area
   31,337.8
   48,510.9
   29,975.9
   26,385.8
  135,292.9
  204,940.2
    2,276.7
Offroad Mobile
   99,305.6
    9,350.5
   13,063.5
   11,985.3
  754,272.8
  141,183.3
      73.9
Onroad Mobile
  189,852.3
    5,353.5
    4,486.6
    3,297.4
 1,585,277.1
   97,245.6
    5,993.5
Fire
    3,539.6
     936.2
   12,407.2
   10,642.3
  151,389.6
   12,867.9
    1,447.2
Ag and Soil Ammonia
                                                                            0.0
                                                                            0.0
                                                                            0.0
                                                                            0.0
                                                                            0.0
                                                                            0.0
  152,904.1
Fugitive Dust
                                                                            0.0
                                                                            0.0
   95,240.0
   19,006.9
                                                                            0.0
                                                                            0.0
                                                                            0.0
Road Dust
                                                                            0.0
                                                                            0.0
  367,390.3
   55,011.6
                                                                            0.0
                                                                            0.0
                                                                            0.0
Biogenics
   22,518.6
                                                                            0.0
                                                                            0.0
                                                                            0.0
  134,123.4
 1,428,260.0
                                                                            0.0
Totals
  528,135.5
  433,396.3
  541,748.9
  135,897.8
 2,879,469.2
 1,924,767.1
  168,948.5


                    
                    
                    
                    Table 7.2:  2018 Missouri Emissions Inventory Summary
Source Sector
  NOX (TPY)
  SO2 (TPY)
 PM10 (TPY)
 PM2.5 (TPY)
  CO (TPY)
 VOC (TPY)
 NH3 (TPY)
Point EGU
   84,619.8
  289,330.1
   18,958.2
   17,036.6
   15,752.7
    2,080.5
     874.4
Point NEGU
   49,290.8
   66,731.1
   23,598.8
   10,171.7
  184,350.9
   54,908.6
    8,600.2
Area
   35,212.8
   49,726.1
   29,193.0
   25,528.5
  120,114.9
  265,737.4
    4,411.8
Offroad Mobile
   59,624.9
     565.2
    8,371.3
    7,675.0
  739,932.9
   72,794.1
      84.8
Onroad Mobile
   50,860.9
     797.4
    1,415.5
    1,415.5
  895,481.6
   39,672.3
    8,316.0
Fire
    3,539.6
     936.2
   12,407.2
   10,642.3
  151,389.6
   12,867.9
    1,447.2
Ag and Soil Ammonia
                                                                            0.0
                                                                            0.0
                                                                            0.0
                                                                            0.0
                                                                            0.0
                                                                            0.0
  182,451.5
Fugitive Dust
                                                                            0.0
                                                                            0.0
  106,045.3
   21,147.2
                                                                            0.0
                                                                            0.0
                                                                            0.0
Road Dust
                                                                            0.0
                                                                            0.0
  313,576.4
   46,957.9
                                                                            0.0
                                                                            0.0
                                                                            0.0
Biogenics
   22,518.6
                                                                            0.0
                                                                            0.0
                                                                            0.0
  134,123.4
 1,428,260.0
                                                                            0.0
TOTALS
  305,667.4
  408,086.1
  513,565.8
  140,574.6
 2,241,146.0
 1,876,320.7
  206,185.9

* Tons Per Year
** Electric Generating Unit
*** Non-Electric Generating Unit

7.2 OVERVIEW OF EMISSIONS INVENTORY DEVELOPMENT

      7.2.1 Point Sources

The 2002 point source inventory is based on information reported by facilities on Emission Inventory Questionnaires (EIQs). The 2002 EIQ data collection process was conducted by the department's Air Pollution Control Program and the local air pollution agencies of St. Louis County and the City of St. Louis. As the coordinating agency for point source inventory development, the department's Air Pollution Control Program performed the overall quality assurance procedures and submitted the data to EPA's 2002 National Emissions Inventory (NEI) to meet the requirements of the CERR.

Following submission of the Missouri point source inventory to the 2002 NEI, additional quality
assurance, and revision of the data was completed through the CENRAP process. E. H. Pechan
& Associates (Pechan), through a contract with CENRAP, obtained the Missouri point source
inventory and worked with the department's Air Pollution Control Program to make corrections
where needed. In particular, an error that resulted in the double counting of emissions from a
number of emission units was corrected. The problem affected VOC emissions only. For
example, for the Chrysler-North facility (291890231), emission unit number 20949, which
emitted a total of 112 tons/year (about 0.3 tons/day) VOC in 2002, was associated with stack
numbers 44387 and 44388. Instead of being proportioned between the two stacks, the total
amount of 112 tons/year was linked to each stack, which doubled the emissions. In all, this
problem resulted in overstating VOC emissions in the St. Louis nonattainment area by a total of
751 tons/year (roughly 2 tons/day). Other revisions included corrections to facility coordinates
and stack parameters. Pechan also converted the point source inventory to the Sparse Matrix
Operator Kernel Emissions/Inventory Data Analyzer (SMOKE/IDA) format. Pechan's work is
described in detail in the two documents included in Appendix H: The Consolidation of
Emissions Inventories (April 28, 2005) and Refinement of CENRAP's 2002 Emissions Inventories (August 31, 2005).

The 2018 point source emissions inventory was prepared by CENRAP. For non-EGUs, the 2002
emissions were projected to 2018 by applying growth and control factors using the SMOKE
model. The growth and control factors were prepared by Pechan and are documented in the
following report in Appendix H.4: Development of Growth and Control Inputs for CENRAP
2018 Emissions Draft Technical Support Document (May 2005). The control factors for non-
EGU point sources account for Maximum Achievable Control Technology (MACT) standards
and the NOX SIP Call for industrial boilers. In addition, the newly permitted Holcim cement kiln
in Ste. Genevieve County was added to the 2018 non-EGU point inventory.

The Integrated Planning Model (IPM) version 2.1.9 model output for 2018 was used for 2018
EGU point source emissions. The SMOKE IDA formatted version of the 2018 Integrated
Planning Model (IPM) 2.1.9 file was prepared by Pechan for CENRAP. See the Pechan report,
Refinement of CENRAP's 2002 Emissions Inventories (August 31, 2005), in Appendix H.3 for
more information. The proprietary IPM model has been used by the EPA to simulate electrical
power generation and electrical power distribution scenarios based upon "least-cost"
assumptions for future years and, simultaneously, generate estimates of pollutant emissions
associated with these scenarios. The IPM run was conducted by ICF under contract to the RPOs.
This run corresponds with the "VISTASII_PC_1f" modeling run. This run specifically
addressed the emission reductions to be realized through implementation of CAIR assuming all
states participate in the EPA's trading program, Acid Rain Program (Title IV  -  Phases I and II),
NOX SIP Call, and state and local regulations, while incorporating unit-level updates provided by power company stakeholders.

The University of California-Riverside (UCR) ran the SMOKE model for 2018 point source
emissions. The edited IPM file for EGUs was processed in SMOKE without adjustments. The
growth and control factors for non-EGUs were applied using the SMOKE model. The technical
support document in Appendix F describes UCR's work on the 2018 point source inventory.

      7.2.2 Area Sources

The 2002 area source inventory includes emissions estimates prepared by the department's Air
Pollution Control Program and CENRAP, with remaining gaps filled in with data from the NEI.
Table H.9 in Appendix H.1 lists the source of the emissions estimates for each SCC in the base
year area source inventory. For the categories developed by the department's Air Pollution

Control Program, the data and methods used are described in the document Missouri Statewide
Estimates for the 2002 National Emissions Inventory (NEI): Area Sources (January 8, 2007) in
Appendix H.5. The data and methods used to develop the prescribed burning, agricultural dust,
and soil agricultural ammonia inventories for CENRAP can be found in the following reports
prepared by Sonoma Technology in Appendix H: Research and Development of Planned
Burning Emission Inventories for the Central States Regional Air Planning Association (July 30,
2004), Emission Inventory Development for Mobile Sources and Agricultural Dust Sources for
the Central States (October 28, 2004), and Research and Development of Ammonia Emission
Inventories for the Central States Regional Air Planning Association (October 30, 2003).
Documentation of EPA's methods for the NEI may be found on EPA's Clearinghouse for
Inventories and Emission Factors (CHIEF) website at
http://www.epa.gov/ttn/chief/net/2002inventory.html.

In a contract with CENRAP, Pechan consolidated the area source data from the various sources,
conducted additional quality assurance, and worked with the department's Air Pollution Control
Program to make revisions where needed. In particular, corrections were made to doublecounting error of industrial surface coating VOC emissions. Pechan also converted the area source inventory to the SMOKE/IDA format. Pechan's work is described in detail in two
documents included in Appendix H: The Consolidation of Emissions Inventories (April 28,
2005) and Refinement of CENRAP's 2002 Emissions Inventories (August 31, 2005).  

To prepare the area inventories for modeling, UCR made several modifications to the IDA files by removing selected sources either to model them as separate source categories or to omit them from simulations completely. Fugitive and road dust sources were extracted from all stationary-area inventories and adjusted by transport factors following Methodology to Estimate the
Transportable Fraction (TF) of Fugitive Dust Emissions for Regional and Urban Scale Air Quality Analyses (Pace 2005).

The 2018 area source emissions inventory was based on data provided by CENRAP states. Area
source growth and control factors were prepared by Pechan and are documented in the following report in Appendix H.5: Development of Growth and Control Inputs for CENRAP 2018 Emissions Draft Technical Support Document (May 2005). The control factors reflect New
Source Performance Standards (NSPS) for residential wood combustion and Stage II vapor
recovery controls, including onboard vapor recovery.

UCR ran the SMOKE model for the 2018 area source emissions. The growth and control factors
for area sources were applied within SMOKE. The technical support document in Appendix F
describes UCR's work on the 2018 area source inventory. Windblown dust from nonagricultural
land use categories and fire emissions were held constant from 2002 to 2018.

      7.2.3 Offroad Mobile Sources

The 2002 offroad mobile source includes emissions estimates prepared by the department's Air
Pollution Control Program and CENRAP, with remaining gaps filled in with EPA NEI data.
Table H.10 in Appendix H.1 lists the source of the emissions estimates for each SCC in the base
year offroad mobile inventory. The majority of the offroad mobile inventory was developed by
Sonoma Technology under a contract with CENRAP. The methods and data used by Sonoma
are described in the report Emissions Inventory Development for Mobile Sources and
Agricultural Dust Sources for the Central States (October 28, 2004) in Appendix H.7.
Information on the NONROAD model is at http://www.epa.gov/otaq/nonrdmdl.htm.

Pechan, under a contract with CENRAP, consolidated the offroad mobile source inventories
from the various data sources, quality-assured the data, worked with the department's Air
Pollution Control Program to make corrections where needed, and created SMOKE/IDAformatted files. In particular, Pechan made corrections to the fuel oxygenate content used in the NONROAD model. Pechan's work is described in detail in the two documents included in Appendix H: The Consolidation of Emissions Inventories (April 28, 2005) and Refinement of CENRAP's 2002 Emissions Inventories (August 31, 2005).

The 2018 offroad mobile inventory was based on inputs from CENRAP states. Growth and
control factors for locomotives, aircraft, and commercial marine vessels were prepared by
Pechan. The control factors accounted for federal standards for commercial marine vessels and
locomotives. For the remaining offroad mobile categories, Pechan ran the EPA's
NONROAD2004 model for 2018.  EPA's NONROAD2004 model accounts for growth in
equipment populations and incorporates the effects of most final federal standards, including the
Tier 4 diesel engine standards and the exhaust emission standards for large spark-ignition
engines, diesel marine, and land-based recreational engines. Pechan's methods are described in
greater detail in the following report in Appendix H.4: Development of Growth and Control
Inputs for CENRAP 2018 Emissions Draft Technical Support Document (May 2005).

UCR and applied the growth and control factors to non-NONROAD categories using the
SMOKE model. In addition, UCR processed NONROAD-model categories in SMOKE without
adjustments. The technical support document in Appendix F describes UCR's work on the 2018
offroad inventory

      7.2.4 Onroad Mobile Sources

The department's Air Pollution Control Program and CENRAP, with contractor support,
developed the 2002 and 2018 onroad mobile source emissions inventories. Sonoma Technology
provided 2002 VMT data and MOBILE6 input files for all counties in the CENRAP region.
MOBILE6 input files were provided only for the months of January and July for 2002. The
methods and data used by Sonoma are described in the report Emissions Inventory Development
for Mobile Sources and Agricultural Dust Sources for the Central States (October 28, 2004) in
Appendix H.7. UCR prepared MOBILE6 input files for the remaining months of 2002 and
processed the 2002 mobile emissions using the MOBILE6 model within the SMOKE
framework.

Pechan prepared the VMT and MOBILE6 inputs for the 2018 onroad mobile source emissions
inventory. The VMT growth factors and MOBILE6 input files were provided in SMOKE
format. The MOBILE6 input files incorporated state/local control program information,
including Reformulated Gasoline and the inspection and maintenance program in the St. Louis
nonattainment area and low Reid vapor pressure (RVP) gasoline in the Kansas City maintenance area. For each county or group of counties modeled, two SMOKE-formatted MOBILE6 files were prepared: one representing July conditions and one representing January conditions. Pechan's methods are described in greater detail in the following report in Appendix H.4: Development of Growth and Control Inputs for CENRAP 2018 Emissions Draft Technical
Support Document (May 2005). 

UCR prepared MOBILE6 input files for the remaining months of 2018 and processed the 2018 onroad mobile emissions by running the MOBILE6 model within the SMOKE framework. The SMOKE model applies the VMT growth factors. The MOBILE6 model accounts for federal motor vehicle controls, including light-duty motor vehicle engine standards and low-sulfur gasoline, and the federal heavy-duty diesel engine standards and low-sulfur diesel. The technical support document in Appendix F describes UCR's onroad mobile emissions inventory processing.

      7.2.5 Biogenic Emissions

UCR generated biogenic emissions by running the BEIS3 model within the SMOKE framework.
BEIS3 is a system integrated into SMOKE for deriving emissions estimates of biogenic gasphase
pollutants from land use information, emissions factors for different plant species, and
hourly, gridded meteorology data. Biogenic emissions were held constant from 2002 to 2018.
The technical support document in Appendix F describes the development of the biogenic
emissions inventory.

7.3 PERIODIC UPDATES OF EMISSIONS INVENTORIES

Recognizing the importance of maintaining current, valid emissions information, the
department's Air Pollution Control Program commits to periodically updating the Missouri
statewide emissions inventories. The point source inventories will be updated on an annual basis, and the area, onroad mobile, and offroad mobile inventories will be updated every three years.  The three-year updates will begin with the inventory for calendar year 2008, and follow with 2011, 2014, and so on, consistent with EPA's emissions inventory reporting requirements.

In addition to completing regular updates of Missouri's emissions inventory, the Air Pollution
Control Program commits to periodically reviewing emissions information for other states and
future-year emissions projections and making adjustments where needed. This effort will consist
of reviewing and updating any technical data and assumptions regarding emissions growth rates, implementation of emissions controls, and geographic distribution of emissions. The periodic reviews will be coordinated with other states and consultation partners and will be conducted in conjunction with the five-year progress reports discussed in section 2.6.

EPA Findings:  Missouri has appropriately addressed this provision.

	F.	Monitoring Strategy and Other Implementation Plan Requirements (40 CFR 			51.308(d)(4))
		SIP Pages 30-33

The State's plan must include a monitoring strategy for measuring, characterizing, and reporting of regional haze visibility impairment that is representative of all mandatory Class I areas within the State and/or summarize monitoring strategy of States with affected mandatory Class I areas. Compliance with this requirement may be met through participation in the IMPROVE network.

      1.	State Submittal
            SIP Page 30

6.2 CURRENT MONITORING STRATEGY

Upon the creation of CENRAP, the newly formed Monitoring Workgroup identified large
visibility data voids in Southern Arkansas, Iowa, Kansas, Southern Minnesota, Nebraska, and
Oklahoma. Only five IMPROVE sites were located in the CENRAP region. Between 2000 and
2003, five more IMPROVE sites and 15 IMPROVE protocol sites were installed. In Missouri,
IMPROVE Sites are located at Hercules Glades and Mingo (Figures 6.1 and 6.2). An
IMPROVE protocol sampler is located at the site near El Dorado Springs (Figure 6.3). Missouri
commits to meet the requirements under 40 CFR 51.308(d)(4)(iv) to report to EPA visibility data
for each of Missouri's Class I areas annually.

The filter samples from the IMPROVE modules are sent for analysis to the Crocker Nuclear
Laboratory of the University of California in Davis and the data is posted to the IMPROVE
website and the VIEWS website.4 Details regarding the monitors (location, date of installation,
etc., and monitoring data) are found at the VIEWS website. This fulfills Missouri's reporting
requirement of visibility data (electronic) under subsection (iv).

6.3 FUTURE MONITORING STRATEGY

In order to assess progress in reducing visibility impairment in Class I areas, the existing
IMPROVE and IMPROVE Protocol sites will be maintained contingent upon continued national
funding to measure, characterize and report regional haze visibility impairment to satisfy
requirements of subsection (i). If EPA elects to revise funding for this network, Missouri will
evaluate the IMPROVE protocol site at El Dorado Springs. Any changes appropriate to
continued monitoring of Regional Haze for the Missouri Class I areas will be evaluated during
the Missouri five-year review. The five-year review will include the following aspects:

* QA IMPROVE data from Mingo and Hercules Glades.
* Calculate current visibility conditions for most impaired and least impaired days.
* Calculate differences between current conditions and baseline conditions.
* Determine whether RPGs are being met.
Missouri will also evaluate technology changes and the need for new monitors as appropriate.

6.4 SPECIAL MONITORING STUDIES

Special monitoring in the CENRAP region for ammonia was conducted from November 1, 2003
through June 28, 2006. In all, approximately 7,200 individual ammonia and associated
measurements were attempted in the course of this project. One of the primary outcomes of this
sampling was the disclosure that high concentrations of ammonia are occurring in the northern
and central CENRAP regions with a considerable regularity. It seems likely that these are due to
the agricultural sources that have been documented as emitters of ammonia, including animal
raising and fertilizer application.5
 
EPA Findings: Missouri has appropriately addressed this provision.

      G.	Planning and Conclusion (40 CFR 51.308(f-h))
		SIP Pages 101-103

40 CFR 51.308(f) requires a state/tribe to revise its regional haze implementation plan and submit a plan revision to EPA by July 31, 2018 and every ten years thereafter. In accordance with the requirements listed in 40 CFR  51.308(f) of the federal Regional Haze Rule, Missouri in its SIP has committed to revising and submitting this regional haze implementation plan by July 31, 2018 and every ten years thereafter.  40 CFR 51.308(g) 

The Missouri Department of Natural Resources is responsible for developing and submitting the required SIP revisions and periodic reports. In addition, 40 CFR 51.308(g) requires periodic reports evaluating progress towards the RPGs established for each mandatory Class I area. In accordance with the requirements listed in 40 CFR  51.308(g) of the federal rule for regional haze, Missouri has committed to submitting a report on reasonable progress to EPA every five years following the initial submittal of the plan. The reasonable progress report will evaluate the progress made towards the RPG for each mandatory Class I area located within Missouri and in each mandatory Class I area located outside Missouri that may be affected by emissions from within Missouri. The report will be in the form of a SIP revision.

To establish the criteria in evaluating progress, Missouri claims that all requirements listed in 40 CFR 51.308(g) shall be addressed in the plan revision for reasonable progress. These criteria are as follows:

   * Assessment of visibility conditions and changes for each Class I area in
      Missouri;  
   * Implementation status of control measures included in plan and a summary of emissions reductions achieved from measures;
   * Analysis of emission reductions by pollutant, identified by source or activity;
   * Assessment of any significant changes in anthropogenic emissions;
   * Assessment of whether current plan is sufficient to meet RPGs; and
   * Review of Missouri's visibility monitoring strategy and any necessary strategy.

Using the modeling and monitoring as described in the SIP and appendices, Missouri has made a determination that the plan will continue to meet the goal of showing progress towards reducing visibility. Using the consultation process described in the plan, Missouri demonstrates that the visibility goals have been adequately addressed. In addition, through participation in other states' consultation processes, Missouri demonstrates that contribution to other states visibility goals has also been adequately addressed.

For the 2013 five-year review process, Missouri intends to conduct a five factor analysis (four factors plus visibility impact) to address reasonable progress goals set for the Mingo and Hercules Glades Class 1 Areas. The findings of the five-year progress report will determine which action is appropriate and necessary. Depending on the findings of the five-year progress report, Missouri has committed to one of the actions listed in 40 CFR 51.308(h):

   * If Missouri determines that the existing plan requires no further substantive revision in order to achieve established goals, the state will provide the EPA with a negative declaration that further revision of the plan is not needed at this time.
   * If Missouri determines that the existing plan may be inadequate to ensure reasonable progress due to emissions from other states that participated in the regional planning process, Missouri will provide notification to the EPA and the other states that participated in regional planning. Missouri will collaborate with the other states through the regional planning process to address the plan's deficiencies.
   * Where Missouri determines that the current plan may be inadequate to ensure reasonable progress due to emissions from another country, the state shall provide notification, along with available information, to the EPA.
   * Where Missouri determines that the existing plan is inadequate to ensure reasonable progress due to emissions within the state, Missouri shall revise its plan to address the plan's deficiencies within one year.

Modeling and monitoring information will be used and the consultation process described in the plan will be followed in carrying out the scheduled incremental administrative and technical actions required in the plan. Any resulting plan revisions could include a revision to goals, contingency measures, the monitoring strategy, and any other parts of the plan as deemed necessary.

EPA Findings: Missouri has appropriately addressed these provisions.

IV.	CONCLUSION

Except for those aspects of the Regional Haze submittal that rely on EPA's CAIR Rule, Missouri's Regional Haze SIP, submitted on August 5, 2009, including the technical supplement submitted on January 30, 2012, meets all of the requirements of Section 110 of the CAA and Regional Haze requirements, as set forth in Sections 169A and 169B of the Act and in 40 CFR 51.300-308, including by requiring the establishment of reasonable progress goals for each of its Class I areas, determining Best Available Retrofit Technology (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 Missouri.








                                       
                                       
                                       
	

