Part III

As Enacted ( Federal Register)

Cited authorities 24

Cited in

[Federal Register: July 6, 2005 (Volume 70, Number 128)]

[Rules and Regulations]

[Page 39103-39172]

From the Federal Register Online via GPO Access [wais.access.gpo.gov]

[DOCID:fr06jy05-22]

[[Page 39103]]

Part III

Environmental Protection Agency

40 CFR Part 51

Regional Haze Regulations and Guidelines for Best Available Retrofit Technology (BART) Determinations; Final Rule

[[Page 39104]]

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 51

[FRL-7925-9]

RIN 2060-AJ31

Regional Haze Regulations and Guidelines for Best Available Retrofit Technology (BART) Determinations

AGENCY: Environmental Protection Agency (EPA).

ACTION: Final rule.

SUMMARY: On July 1, 1999, EPA promulgated regulations to address regional haze (64 FR 35714). These regulations were challenged, and on May 24, 2002, the U.S. Court of Appeals for the District of Columbia Circuit issued a ruling vacating the regional haze rule in part and sustaining it in part. American Corn Growers Ass'n v. EPA, 291 F.3d 1 (D.C. Cir. 2002). Today's rule addresses the court's ruling in that case.

In addition, prior to the court's decision, EPA had proposed guidelines for implementation of the Best Available Retrofit Technology (BART) requirements under the regional haze rule, (66 FR 38108, July 20, 2001). The proposed guidelines were intended to clarify the requirements of the regional haze rule's BART provisions. We proposed to add the guidelines and also proposed to add regulatory text requiring that these guidelines be used for addressing BART determinations under the regional haze rule. In addition, we proposed one revision to guidelines issued in 1980 for facilities contributing to ``reasonably attributable'' visibility impairment.

In the American Corn Growers case, the court vacated and remanded the BART provisions of the regional haze rule. In response to the court's ruling, on May 5, 2004 we proposed new BART provisions and reproposed the BART guidelines. The American Corn Growers court also remanded to the Agency its decision to extend the deadline for the submittal of regional haze plans. Subsequently, Congress amended the deadlines for regional haze plans (Consolidated Appropriations Act for Fiscal Year 2004, Public Law 108-199, January 23, 2004). The May 5, 2004 proposed rule also contained an amendment to the regional haze rule to conform to the new statutory deadlines.

We received numerous comments on both the July 20, 2001 proposal and the May 5, 2004 reproposal. Today's final rule reflects our review of the public comments.

DATES: The regulatory amendments announced herein take effect on September 6, 2005.

ADDRESSES: Docket. All documents in the docket are listed in the EDOCKET index at http://www.epa.gov/edocket. Although listed in the

index, some information is not publicly available, i.e., CBI or other information whose disclosure is restricted by statute. Certain other material, such as copyrighted material, is not placed on the Internet and will be publicly available only in hard copy form. Publicly available docket materials are available either electronically in EDOCKET or in hard copy at the OAR Docket, EPA/DC, EPA West, Room B102, 1301 Constitution Ave., NW., Washington, DC. The Public Reading Room is open from 8:30 a.m. to 4:30 p.m., Monday through Friday, excluding legal holidays. The telephone number for the Public Reading Room is (202) 566-1744, and the telephone number for the OAR Docket is (202) 566-1742.

FOR FURTHER INFORMATION CONTACT: Kathy Kaufman at (919) 541-0102 or by e-mail at Kaufman.Kathy@epa.gov or Todd Hawes at 919-541-5591 or by e- mail Hawes.Todd@epa.gov.

SUPPLEMENTARY INFORMATION:

Regulated Entities. This final rule will affect the following: State and local permitting authorities and Indian Tribes containing major stationary sources of pollution affecting visibility in federally protected scenic areas.

This list is not intended to be exhaustive, but rather provides a guide for readers regarding entities likely to be regulated by this action. This list gives examples of the types of entities EPA is now aware could potentially be regulated by this action. Other types of entities not listed could also be affected. To determine whether your facility, company, business, organization, etc., is regulated by this action, you should examine the applicability criteria in Part II of this preamble. If you have any questions regarding the applicability of this action to a particular entity, consult the people listed in the preceding section.

Outline. The contents of today's preamble are listed in the following outline.

Overview of Today's Proposed Actions II. Background
1. Regional Haze Rule
2. Partial Remand of the Regional Haze Rule in American Corn Growers
3. Changes in Response to American Corn Growers
4. Center for Energy and Economic Development v. EPA
5. Relationship Between BART and the Clean Air Interstate Rule (CAIR)
6. Overview of the BART Process III. Detailed Discussion of the BART Guidelines
7. Introduction
8. Scope of the Rule--Whether to Require States to Follow the Guidelines for All BART Sources
9. How to Identify BART-Eligible Sources
10. How to Determine Which BART-Eligible Sources are Subject to BART
11. The BART Determination Process IV. Effect of This Rule on State Options for Using Alternative Strategies In Lieu of Source-by-Source BART V. Statutory and Executive Order Reviews
12. Executive Order 12866: Regulatory Planning and Review
13. Paperwork Reduction Act
14. Regulatory Flexibility Act
15. Unfunded Mandates Reform Act
16. Executive Order 13132: Federalism
17. Executive Order 13175: Consultation and Coordination with Indian Tribal Governments
18. Executive Order 13045: Protection of Children from Environmental Health Risks and Safety Risks
19. Executive Order 13211: Actions That Significantly Affect Energy Supply, Distribution, or Use.
National Technology Transfer Advancement Act
1. Executive Order 12898: Federal Actions to Address Environmental Justice in Minority Populations and Low-Income Populations
Overview of Today's Actions

Today's rulemaking provides the following changes to the regional haze regulations:

(1) Revised regulatory text in response to the American Corn Growers court's remand, to require that the BART determination include an analysis of the degree of visibility improvement resulting from the use of control technology at each source subject to BART,

(2) Revised regulatory text in 40 CFR 51.308(b) and deletion of 40 CFR 51.308(c) Options for regional planning in response to Congressional legislation amending the deadlines for submittal of regional haze implementation plans. This provision had provided for an alternative process for States to submit regional haze implementation plans in attainment areas,

(3) BART guidelines, contained in a new Appendix Y to 40 CFR part 51,

(4) New and revised regulatory text, to be added to 40 CFR 51.308(e), regarding the use of Appendix Y in establishing BART emission limits, and

(5) Revised regulatory language at 40 CFR 51.302 to clarify the relationship between New Source Performance Standards (NSPS) and BART for reasonably attributable visibility impairment.

How This Preamble Is Structured. Section II provides background on the

[[Page 39105]]

Clean Air Act (CAA) BART requirements as codified in the regional haze rule, on the D.C. Circuit Court decision which remanded parts of the rule, and on the April 2004 reproposal responding to the remand. Section III discusses specific issues in the BART guidelines in more detail, including background on each issue, major comments we received on the July 2001 proposal and May 2004 reproposal, and our responses to those comments. Section IV provides a discussion of how this rulemaking complies with the requirements of Statutory and Executive Order Reviews.
Background
1. The Regional Haze Rule
  
  In 1999, we published a final rule to address a type of visibility impairment known as regional haze (64 FR 35714, July 1, 1999). The regional haze rule requires States to submit implementation plans (SIPs) to address regional haze visibility impairment in 156 Federally- protected parks and wilderness areas. These 156 scenic areas are called ``mandatory Class I Federal areas'' in the Clean Air Act (CAA)\1\ but are referred to simply as ``Class I areas'' in today's rulemaking. The 1999 rule was issued to fulfill a long-standing EPA commitment to address regional haze under the authority and requirements of sections 169A and 169B of the CAA.
  
  \1\ See, e.g. CAA Section 169A(a)(1).
  
  As required by the CAA, we included in the final regional haze rule a requirement for BART for certain large stationary sources that were put in place between 1962 and 1977. We discussed these requirements in detail in the preamble to the final rule (64 FR at 35737-35743). The regulatory requirements for BART were codified at 40 CFR 51.308(e) and in definitions that appear in 40 CFR 51.301.
  
  The CAA, in sections 169A(b)(2)(A) and in 169A(g)(7), uses the term ``major stationary source'' to describe those sources that are the focus of the BART requirement. To avoid confusion with other CAA requirements which also use the term ``major stationary source'' to refer to a somewhat different population of sources, the regional haze rule uses the term ``BART-eligible source'' to describe these sources. The BART-eligible sources are those sources which have the potential to emit 250 tons or more of a visibility-impairing air pollutant, were 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 which 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.'' Accordingly, for stationary sources meeting these criteria, States must address the BART requirement when they develop their regional haze SIPs.
  
  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 nonair 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.
  
  These statutory factors for BART were codified at 40 CFR 51.308(e)(1)(ii).
  
  In the preamble to the regional haze rule, we committed to issuing further guidelines to clarify the requirements of the BART provision. The purpose of this rulemaking is to fulfill this commitment by providing guidelines to assist States as they identify which of their BART-eligible sources should undergo a BART analysis (i.e., which are ``sources subject to BART'') and select controls in light of the statutory factors listed above (``the BART determination'').
2. Partial Remand of the Regional Haze Rule in American Corn Growers v. EPA
  
  In response to challenges to the regional haze rule by various petitioners, the D.C. Circuit in American Corn Growers \2\ issued a ruling striking down the regional haze rule in part and upholding it in part. This section discusses the court's opinion in that case as background for the discussion of specific changes to the regional haze rule and the BART guidelines presented in the next two sections, respectively.
  
  \2\ American Corn Growers et al. v. EPA, 291 F.3d 1 (2002).
  
  We explained in the preamble to the 1999 regional haze rule that the BART requirements in section 169A(b)(2)(A) of the CAA demonstrate Congress' intent to focus attention directly on the problem of pollution from a specific set of existing sources (64 FR 35737). The CAA requires that any of these existing sources ``which, as determined by the State, emits any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility [in a Class I area],'' shall install the best available retrofit technology for controlling emissions.\3\ In determining BART, the CAA requires the State to consider several factors that are set forth in section 169(g)(2) of the CAA, including the degree of improvement in visibility which may reasonably result from the use of such technology.
  
  \3\ CAA sections 169A(b)(2) and (g)(7).
  
  The regional haze rule addresses visibility impairment resulting from emissions from a multitude of sources located across a wide geographic area. Because the problem of regional haze is caused in large part by the long-range transport of emissions from multiple sources, and for certain technical and other reasons explained in that rulemaking, we had adopted an approach that required States to look at the contribution of all BART sources to the problem of regional haze in determining both applicability and the appropriate level of control. Specifically, we had concluded that if a source potentially subject to BART is located within an upwind area from which pollutants may be transported downwind to a Class I area, that source ``may reasonably be anticipated to cause or contribute'' to visibility impairment in the Class I area. Similarly, we had also concluded that in weighing the factors set forth in the statute for determining BART, the States should consider the collective impact of BART sources on visibility. In particular, in considering the degree of visibility improvement that could reasonably be anticipated to result from the use of such technology, we stated that the State should consider the degree of improvement in visibility that would result from the cumulative impact of applying controls to all sources subject to BART. We had concluded that the States should use this analysis to determine the appropriate BART emission limitations for specific sources.\4\
  
  \4\ See 66 FR at 35737-35743 for a discussion of the rationale for the BART requirements in the 1999 regional haze rule.
  
  In American Corn Growers v. EPA, industry petitioners challenged EPA's interpretation of both these aspects of the BART determination process and raised other challenges to the rule. The court in American Corn Growers concluded that the BART provisions in the 1999 regional haze rule were inconsistent with the provisions in the CAA ``giving the states broad authority over BART determinations.'' 291 F.3d at 8. Specifically, with respect to the test for determining whether a source is subject to BART, the court held that the
  
  [[Page 39106]]
  
  method that EPA had prescribed for determining which eligible sources are subject to BART illegally constrained the authority Congress had conferred on the States. Id. The court did not decide whether the general collective contribution approach to determining BART applicability was necessarily inconsistent with the CAA. Id. at 9. Rather, the court stated that ``[i]f the [regional haze rule] contained some kind of a mechanism by which a state could exempt a BART-eligible source on the basis of an individualized contribution determination, then perhaps the plain meaning of the Act would not be violated. But the [regional haze rule] contains no such mechanism.'' Id. at 12.
  
  The court in American Corn Growers also found that our interpretation of the CAA requiring the States to consider the degree of improvement in visibility that would result from the cumulative impact of applying controls in determining BART was inconsistent with the language of the Act. 291 F.3d at 8. Based on its review of the statute, the court concluded that the five statutory factors in section 169A(g)(2) ``were meant to be considered together by the states.'' Id. at 6.
3. Changes in Response to American Corn Growers
  
  Today's rule responds to the American Corn Growers court's decision on the BART provisions by including changes to the regional haze rule at 40 CFR 51.308, and by finalizing changes to the BART guidelines. This section outlines the changes to the regional haze rule due to the court's remand. It also explains the minor change we are making to the section of the regulation governing the use of the 1980 BART guidelines when conducting BART analyses for certain power plants for reasonably attributable (i.e., localized) visibility impairment. 1. Determination of Which Sources Are Subject to BART
  
  Today's action addresses the American Corn Growers court's vacature of the requirement in the regional haze rule requiring States to assess visibility impacts on a cumulative basis in determining which sources are subject to BART. Because this requirement was found only in the preamble to the 1999 regional haze rule (see 291 F.3d at 6, citing 64 FR 35741), no changes to the regulations are required. Instead, this issue is addressed in the BART guidelines, which provide States with appropriate techniques and methods for determining which BART-eligible sources ``may reasonably be anticipated to cause or contribute to any impairment of visibility in any mandatory Class I Federal area.'' These processes, to address the holding of American Corn Growers by eliminating the previous constraint on State discretion, are explained in further detail in sections II.D. and III below. 2. Consideration of Anticipated Visibility Improvements in BART Determinations
  
  Pursuant to the remand in American Corn Growers, we are amending the regional haze rule to require the States to consider the degree of visibility improvement resulting from a source's installation and operation of retrofit technology, along with the other statutory factors set out in CAA section 169A(g)(2), when making a BART determination. This has been accomplished by listing the visibility improvement factor with the other statutory BART determination factors in 40 CFR 51.308(e)(1)(A), so that States will be required to consider all five factors, including visibility impacts, on an individual source basis when making each individual source BART determination.
4. Center for Energy and Economic Development v. EPA
  
  After the May 2004 reproposal of the BART guidelines, the D.C. Circuit decided another case where BART provisions were at issue, Center for Energy and Economic Development v. EPA, 398 F.3d 653, 2005 (``CEED''). In this case, the court granted a petition challenging provisions of the regional haze rule governing the optional emissions trading program for certain western States and Tribes (the ``WRAP Annex Rule'').
  
  The court in CEED affirmed our interpretation of CAA section 169A(b)(2) as allowing for non-BART alternatives where those alternatives are demonstrated to make greater progress than BART. (CEED, slip. op. at 13). The court, however, took issue with provisions of the regional haze rule governing the methodology of that demonstration. Specifically, 40 CFR 51.308(e)(2) requires that visibility improvements under source-specific BART--the benchmark for comparison to the alternative program--be estimated based on the application of BART controls to all sources subject to BART. (This section was incorporated into the WRAP Annex rule by reference at 40 CFR 51.309(f)). The court held that we could not require this type of group BART approach--vacated in American Corn Growers in a source- specific BART context--even in a program in which State participation was wholly optional.
  
  The BART guidelines as proposed in May 2004 contained a section offering guidance to States choosing to address their BART-eligible sources under the alternative strategy provided for in 40 CFR 51.308(e)(2). This guidance included criteria for demonstrating that the alternative program achieves greater progress towards eliminating visibility impairment than would BART.
  
  In light of the D.C. Circuit's decision in CEED, we have not included the portion of the proposed BART guidelines addressing alternative programs in today's rulemaking. We remain committed to providing States with the flexibility to address BART through alternative means, and we note again that our authority to do so was upheld in CEED. Therefore, we intend to revise the provisions of the regional haze rule governing such alternatives and provide any additional guidance needed in a subsequent rulemaking conducted as expeditiously as practicable.
5. Relationship Between BART and the Clean Air Interstate Rule (CAIR)
  
  On March 10, 2005, EPA issued the Clean Air Interstate Rule (CAIR), requiring reductions in emissions of sulfur dioxide (SO2) and nitrogen oxides (NOX) in 28 eastern States and the District of Columbia. When fully implemented, CAIR will reduce SO2emissions in these states by over 70 percent and NOXemissions by over 60 percent from 2003 levels. The CAIR imposes specified emissions reduction requirements on each affected State, and establishes an EPA-administered cap and trade program for EGUs in which States may participate as a means to meet these requirements. The relationship between BART and the Clean Air Interstate Rule (CAIR) is discussed in section IV. below.
6. Overview of the BART Process
  
  The process of establishing BART emission limitations can be logically broken down into three steps: First, States identify those sources which meet the definition of ``BART-eligible source'' set forth in 40 CFR 51.301. Second, States determine whether such sources ``emit[] any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility [in a Class I area.]'' A source which fits this description is ``subject to BART.'' Third, for each source subject to BART, States then identify the appropriate type and the level of control for reducing emissions.
  
  [[Page 39107]]
  
  Identifying BART-eligible sources. The CAA defines BART-eligible sources as those sources which fall within one of 26 specific source categories, were built during the 15-year window of time from 1962 to 1977, and have potential emissions greater than 250 tons per year. The remand did not address the step of identifying BART-eligible sources, which is conceptually the simplest of the three steps.
  
  Sources reasonably anticipated to cause or contribute to visibility impairment (sources subject to BART). As we noted in the preamble to the 1999 regional haze rule, defining the individual contributions of specific sources of the problem of regional haze can be time-consuming and expensive. Moreover, Congress established a very low threshold in the CAA for determining whether a source is subject to BART. We are accordingly finalizing several approaches for States for making the determination of whether a source ``emits any pollutants which may reasonably be anticipated to cause or contribute to any visibility impairment.'' Certain of these approaches would allow States to avoid undertaking unnecessary and costly studies of an individual source's contribution to haze by allowing States to adopt more streamlined processes for determining whether, or which, BART-eligible sources are subject to BART.
  
  In 1999, we adopted an applicability test that looked to the collective contribution of emissions from an area. In particular, we stated that if ``a State should find that a BART-eligible source is `reasonably anticipated to cause or contribute' to regional haze if it can be shown that the source emits pollutants within a geographic area from which pollutants can be emitted and transported downwind to a Class I area.'' \5\ States certainly have the discretion to consider that all BART-eligible sources within the State are ``reasonably anticipated to cause or contribute'' to some degree of visibility impairment in a Class I area.
  
  \5\ 64 FR 335740, July 1, 1999. The regional haze rule discusses at length why we believe that States should draw this conclusion. 64 FR at 35739-35740.
  
  This is consistent with the American Corn Growers court's decision. As previously noted, the court's concern with our original approach governing BART applicability determinations was that it would have ``tie[d] the states' hands and force[d] them to require BART controls at sources without any empirical evidence of the particular source's contribution to visibility impairment.'' 291 F.3d at 8. By the same rationale, we believe it would be an impermissible constraint of State authority for the EPA to force States to conduct individualized analyses in order to determine that a BART-eligible source ``emits any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in any [Class I] area.'' \6\ American Corn Growers did not decide whether consideration of visibility impact on a cumulative basis would be invalid in all circumstances. 291 F.3d at 9. Given the court's emphasis on the importance of the role of the States in making BART determinations, we believe that a State's decision to use a cumulative analysis at the eligibility stage is consistent with the CAA and the findings of the D.C. Circuit.
  
  \6\ CAA section 169A(b)(2)(A).
  
  We believe a State may conclude that all BART-eligible sources within the State are subject to BART.\7\ Any potential for inequity towards sources could be addressed at the BART determination stage, which contains an individualized consideration of a source's contribution in establishing BART emission limits.
  
  \7\ See 64 FR at 35714, 35721; see also Supporting Information for Proposed Applicability of Regional Haze Regulations, Memorandum by Rich Damberg to Docket A-95-38, U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, July 29, 1997.
  
  States also have the option of performing an analysis to show that the full group of BART-eligible sources in a State cumulatively may not be reasonably anticipated to cause or contribute to any visibility impairment in Class I areas. We anticipate that in most, if not all States, the BART-eligible sources are likely to cause or contribute to some visibility impairment in Class I areas. However, it is possible that using a cumulative approach, a State could show that its BART sources do not pose a problem.
  
  Finally, States may consider the individualized contribution of a BART-eligible source to determine whether a specific source is subject to BART. Specifically, States may choose to undertake an analysis of each BART-eligible source in the State in considering whether each such source meets the test set forth in the CAA of ``emit[ting] any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in any [Class I] area.'' Alternatively, States may choose to presume that all BART-eligible sources within the State meet this applicability test, but provide sources with the ability to demonstrate on a case by case basis that this is not the case. Either approach appears consistent with the D.C. Circuit's statement that a collective contribution approach may be appropriate so long as the States are allowed to exempt sources on the basis of an individualized contribution determination. 291 F.3d at 8.
  
  Today's guidelines include different options States can use to assess whether source should be subject to BART. States need to determine whether to make BART determinations for all of their BART- eligible sources, or to consider exempting some of them from BART because they may not reasonably be anticipated to cause or contribute to any visibility impairment in a Class I area. For assessing the impact of BART-eligible sources on nearby Class I areas, we are including a process whereby the States would use an air quality model able to estimate a single source's contribution to visibility impairment and a different process whereby States could exempt groups of sources with common characteristics based on representative model plant analyses. Finally, States may use cumulative modeling to show that no sources in a State are subject to BART.
  
  The BART determination. The State must determine the appropriate level of BART control for each source subject to BART. Section 169A(g)(7) of the CAA requires States to consider the following factors in making BART determinations: (1) The costs of compliance, (2) the energy and nonair 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. The remand did not address the first four steps of the BART determination. The remand did address the final step, mandating that we must permit States to take into account the degree of improvement in visibility that would result from imposition of BART on each individual source when deciding on particular controls.
  
  The first four factors are somewhat similar to the engineering analysis in the original BART guidelines proposed in 2001 and reproposed in 2004. The BART guidelines also contains a detailed discussion of available and cost-effective controls for reducing SO2and NOXemissions from large coal-fired electric generating units (EGUs).
  
  For assessing the fifth factor, the degree of improvement in visibility from various BART control options, the States may run CALPUFF or another appropriate dispersion model to predict visibility impacts. Scenarios would be
  
  [[Page 39108]]
  
  run for the pre-controlled and post-controlled emission rates for each of the BART control options under review. The maximum 24-hour emission rates would be modeled for a period of three or five years of meteorological data. States have the flexibility to develop their own methods to evaluate model results.
Detailed Discussion of the Final BART Guidelines
1. Introduction
  
  In this section of the preamble, we discuss changes or clarifications to the reproposed BART guidelines. Where relevant, we also respond to comments received during the comment period on the 2001 proposal. For each provision of the guidelines that we are changing or clarifying, we provide discussion of, as appropriate:
  
  --Background information, --How the provision was addressed in the May 2004 reproposal (and in the 2001 proposal, if different from the reproposal), --A summary of comments received on the provision, either from the May 2004 reproposal, from the July 2001 proposal, or from both, and --The changes or clarifications that we are finalizing and the reasons for these changes or clarifications.
2. Scope of the Rule--Whether To Require States To Follow the Guidelines for All BART Sources
  
  Background. Section 169A(b)(1) of the CAA requires EPA to issue regulations to provide guidelines to States on the implementation of the visibility program. In addition, the last sentence of section 169A(b) states:
  
  In the case of a fossil-fuel fired generating powerplant having a capacity in excess of 750 megawatts, the emission limitations required under this paragraph shall be determined pursuant to guidelines, promulgated by the Administrator under paragraph (1).
  
  This statutory requirement clearly requires us to promulgate BART guidelines that the States must follow in establishing BART emission limitations for power plants with a total capacity exceeding the 750 megawatt cutoff. The statute is less clear regarding the import of the guidelines for sources other than 750 megawatt power plants.
  
  Proposed rules. Both the 2001 proposal and the 2004 reproposal included a requirement for States to follow the procedures set out in the guidelines in determining BART for sources in all of the 26 listed BART categories. The 2001 proposal requested comment on whether the regional haze rule should: (1) Require the use of the guidelines only for 750 megawatt utilities, with the guidelines applying as guidance for the remaining categories, or (2) require the use of the guidelines for all of the affected source categories.
  
  Comments. We received comments on this issue in both 2001 and 2004. Comments varied widely on whether we can or should require the use of the guidelines for all of the affected source categories.
  
  Comments from State, local and tribal air quality agencies generally supported our proposal to require the use of the guidelines for all of the source categories. These comments cited a need for national consistency in the application of the BART requirement across the source categories, and from State to State. One State agency commenter questioned our legal authority to require the use of the guidelines for all source categories; and several State agency commenters, while supporting the proposal, requested that we provide clarification of the legal authority for requiring the States to use the guidelines in establishing BART emission limitations for all categories.
  
  Comments from the utility industry, from various manufacturing trade groups, and from individual companies were critical of the proposal to require States to follow the guidelines generally. Many commenters also argued that EPA lacked the authority to issue guidelines for any industrial category other than 750 megawatt powerplants, whether the use of such guidelines were mandatory or not. Other commenters stated that the language in the CAA clearly restricts the scope of mandatory guidelines to larger powerplants. The commenters cited the legislative history of the 1977 Clean Air Act amendments in support of this position, and frequently claimed that requiring the guidelines for all 26 categories of sources would deprive States of flexibility in implementing the program.
  
  Comments from environmental organizations and the general public supported the approach in the proposed rule and stated that EPA is obligated to establish regional haze BART guidelines by rulemaking for all 26 categories of stationary sources. Environmental organization comments noted that while Congress expressed a particular concern for 750 MW powerplants, this added emphasis on one sector does not change requirements in the Act for all BART-eligible sources. Accordingly, these commenters believed that we should not construe a special emphasis on powerplants as a restriction on our authority to require use of the guidelines for all categories.
  
  Final rule. The CAA and the relevant legislative history make clear that EPA has the authority and obligation to publish mandatory guidelines for powerplants exceeding 750 megawatts. As previously noted, Congress in section 169A(b) of the CAA expressly provided that emission limitations for powerplants larger than 750 megawatts ``shall be determined pursuant to guidelines promulgated by the Administrator.'' (Emphasis added). This unambiguous language leaves little room to dispute that the guidelines EPA is required to promulgate must be used by States when making BART determinations for this class of sources.
  
  Having carefully considered the comments and further reviewed the CAA and the legislative history, we have concluded that it would not be appropriate for EPA to require States to use the guidelines in making BART determinations for other categories of sources. The better reading of the Act indicates that Congress intended the guidelines to be mandatory only with respect to 750 megawatt powerplants. Thus, while we acknowledge the State agency comments and the policy reasons support consistency across States, we are not requiring States to use the BART guideline for these other categories. In response to State concerns about equitable application of the BART requirement to source owners with similar sources in different States, we do encourage States to follow the guidelines for all source categories but are not requiring States to do so. States should view the guidelines as helpful guidance for these other categories.
  
  We disagree with comments that the CAA and the legislative history prohibit us from issuing guidance for other source categories. As the guidelines make clear, States are not required to follow the approach in the guidelines for sources other than 750 megawatt powerplants. As such, although we believe that the guidelines provide useful advice in implementing the BART provisions of the regional haze rule, we do not believe that they hamper State discretion in making BART determinations.
3. How To Identify BART-Eligible Sources
  
  Section II of the BART guidelines contains a step-by-step process for identifying stiationary sources that are ``BART-eligible'' under the definitions in the regional haze rule. The four basic steps are:
  
  Step 1: Identify the emission units in the BART categories.
  
  Step 2: Identify the start-up dates of those emission units.
  
  [[Page 39109]]
  
  Step 3: Compare the potential emissions from units identified in Steps 1 and 2 to the 250 ton/year cutoff.
  
  Step 4: Identify the emission units and pollutants that constitute the BART-eligible source.
  
  In this section of the preamble, we discuss some of the comments we received on the steps in this process, and any changes we are making in light of those comments. Step 1: Identify the Emission Units in the BART Categories
  
  The BART guidelines list the 26 source categories that the CAA uses to describe the types of stationary sources that are BART-eligible. Both proposals clarified the descriptions of particular source categories.
  
  Comments. The final rule addresses comments on the following source categories. Some comments discussed below were submitted in response to the 2001 propoosal and were not addressed in the reproposal; other comments were submitted in response to the reproposal in 2004.
  
  (1) ``Charcoal production facilities.'' We received comments in 2001 from two industry trade groups requesting that the final guidelines explicitly exclude ``low-emission'' charcoal production facilities from BART. These comments cited a 1975 study considered by Congress in development of the BART category list in the 1977 CAA amendments. This 1975 study noted that some charcoal production facilities have much higher emissions factors (i.e., 352 pounds of PM per ton of charcoal produced versus 20 to 25 pounds of PM per ton of charcoal produced). Accordingly, the comments asserted that the intent of Congress in the 1977 CAA amendments was to provide incentives for higher-emitting facilities to reduce their emissions, rather than to make the entire category BART-eligible.
  
  (2) ``Chemical process plants.'' In 2001 a trade group representing the pharmaceutical industry requested that we determine in the guidelines that the term ``chemical process plants'' does not include pharmaceutical plants.
  
  (3) ``Primary aluminum ore reduction.'' Comments from the aluminum industry in 2001 noted that not all emissions units at these facilities are necessarily involved in ``primary ore reduction.'' Thus, the comments recommended that we clarify that contiguous sources that are not related to primary aluminum ore reduction, such as fabricating facilities and ingot operations, are not BART-eligible. Further, the comments recommended that we use definitions in the NSPS for primary aluminum plants to describe the BART-eligible emissions units.
  
  (4) ``Fossil-fuel fired steam electric plants of more than 250 million Btu/hour heat input.'' The 2004 reproposal contained the clarification, requested by commenters, that this source category refers only to those fossil-fuel fired steam electric plants that generate electricity for sale. One commenter objected to this clarification on the basis that emissions from co-generators would be excluded; many other commenters supported the clarification. Another commenter requested that we also clarify that this category includes only those steam electric plants that burn greater than 50 percent fossil fuel, in order to be consistent with the definition of fossil- fuel boilers proposed in the guidelines. Other commenters requested that we clarify whether the definition includes units which are located at a steam electric plant, but which themselves are not in any of the 26 BART source categories, such as simple cycle turbines, emergency diesel engines, and reciprocating internal combustion engines (RICE).
  
  Several commenters opined that the category should exclude combined cycle units with heat recovery steam generators that lack auxiliary firing, arguing that these units should count as simple cycle turbines. These commenters pointed to other EPA regulatory programs that treat combined cycle units with supplemental firing differently from combined cycle units without supplemental firing. They argued that we should only consider a combined cycle unit to be a ``steam electric plant'' if it has supplemental firing.
  
  (5) ``Fossil-fuel boilers of more than 250 million Btu/hour heat input.'' The 2004 reproposal clarified that this category should be read as including only those boilers individually greater than 250 million Btu/hour heat input. We received many comments on this interpretation, both in favor and opposed. Those favoring this interpretation (generally industry commenters) cited the implementation burden that including smaller boilers would pose, the high cost- effectiveness of controlling smaller boilers, and the relatively smaller impact on regional haze that smaller boilers would pose. They also noted that this interpretation is most consistent with definitions in the NOXSIP call and new source performance standards (NSPS).
  
  Commenters opposing this interpretation (environmental groups, one state, and one regional planning organization) noted that regarding all boilers, irrespective of size, as BART-eligible so long as the aggregate heat input exceeds 250 million Btu/hour is more consistent with the definition of stationary source under the Prevention of Significant Deterioration (PSD) program. These commenters noted that under the CAA, BART and PSD are complementary programs aimed at regulating the same source categories; either one or the other applies depending upon when the source was constructed.
  
  The 2004 reproposal also clarified that if a boiler smaller than 250 million Btu/hour heat input is an integral part of an industrial process in a BART source category other than electric utilities, then the boiler should be considered part of the BART-eligible source in that category. Under these circumstances, the boiler, as part of the BART-eligible source, should be considered for emission control. Some commenters opposed this interpretation, asserting that it would result in an ``arbitrary and capricious'' inconsistency, in that some smaller boilers would be BART-eligible, and others would not. These commenters also noted that these boilers could be included in regional haze SIPs as necessary for making ``reasonable progress'' toward CAA visibility goals, even if they are not considered to be BART-eligible.
  
  Final rule. After considering the comments, we have made the following determinations on the definitions of the following source categories:
  
  (1) ``Charcoal production facilities.'' We believe that in using the term ``charcoal production facilities'' Congress intended to encompass all types of charcoal production facilities. We do not agree with comments that any inferences can necessarily be made regarding the presence of different PM emission factors for different types of charcoal production facilities in the 1975 report. For example, if Congress only intended to regulate a subset of the charcoal production industry, then we believe Congress could have easily indicated this in the source category title, as was done for ``kraft pulp mills'' and for ``coal cleaning plants (thermal dryers).'' We also note that it is more likely that plants in the charcoal production industry with lower emission factors have emissions that are less than the 250 tons per year cutoff for BART eligibility.
  
  (2) ``Chemical process plants.'' We believe that there is a clear precedent to include pharmaceutical manufacturing operations as ``chemical process plants.'' In the standard industrial classification (SIC) system, pharmaceutical operations are generally
  
  [[Page 39110]]
  
  in SIC codes 2833 and 2834, which are a subset of 2-digit category 28 ``Chemical and Allied products.'' Similarly, in the new North American Industrial Classification Codes (NAICS), pharmaceutical manufacturing is codes 32541 and 32542, which is a subset of the ``chemical manufacturing subsector'' which is code 325. Accordingly, in the PSD program, pharmaceutical plants have been treated as ``chemical process plants.'' The commenter is correct in noting that EPA has consistently distinguished between chemical manufacturing and pharmaceutical manufacturing. Examples where different standards or guidelines are established included control technique guideline (CTG) documents, NSPS standards under section 111 of the CAA, and, most recently, maximum achievable control technology (MACT) standards under section 112 of the CAA. We do not agree that these differentiations for emissions standards necessarily require differentiation for purposes of determining BART eligibility. Therefore we believe pharmaceuticals should not be excluded from BART. However, we expect that because of the MACT standards, there is a very low probability that BART determinations will lead to further control requirements from chemical production processes at pharmaceutical plants.
  
  (3) ``Primary aluminum ore reduction.'' We agree with commenters that BART-eligible units in this source category should be defined consistently with the NSPS definition for primary aluminum ore reduction. Therefore we have added a clarification to that effect in the final BART guidelines. We note that this definition is also consistent with the definition at 40 CFR 63.840, which establishes applicability for this source category for the MACT program.
  
  (4) ``Fossil-fuel fired steam electric plants of more than 250 million Btu/hour heat input.'' We have retained the clarification that this source category refers only to those fossil-fuel fired steam electric plants that generate electricity for sale. We believe that this clarification helps to distinguish those plants that are electric utilities from plants in other industrial categories. We also believe that while large co-generators would be excluded from the fossil-fuel fired steam electric plant source category, most large co-generators will be BART-eligible under the fossil-fuel fired boilers source category.
  
  We do not believe it makes sense for this category to include only those steam electric plants that burn greater than 50 percent fossil fuel. We do not believe that a boiler should be excluded from BART review simply because it is located at a plant which burns less than 50 percent fossil fuel. Emissions from any such boiler could be a significant contributor to regional haze, and as such, we believe that each fossil-fuel fired boiler merits a BART review.
  
  We do wish to clarify that units which are located at a steam electric plant, but which themselves are not in any of the 26 BART source categories, should not be considered to be BART-eligible units. We believe that Congress intended that BART review be focused on units in the source categories it delineated. This interepretation is most consistent with the definition of BART-eligible source as we have explained it elsewhere in this preamble in reference to whether entire plants are included if only some units at the plant meet the statutory criteria.
  
  Finally, we believe that all combined cycle units are included in the definition of fossil fuel fired steam electric plant, regardless of whether the combined cycle unit's heat recovery steam generator lacks auxilliary firing. Commenters are correct that some EPA programs have treated combined cycle units with supplemental firing differently from combined cycle units without supplemental firing. However, while some EPA programs do not consider a unit to be a combined cycle unit unless it contains supplemental firing, the definition at issue here is the definition of fossil-fuel fired steam electric plant, not fossil-fuel fired unit. The CAA defines both ``stationary source'' (for visibility purposes) and ``major emitting facility'' (for PSD purposes) to include ``fossil fuel fired steam electric plants.'' In previous guidance for PSD, we have explained that combined cycle gas turbines do fall within the category of ``fossil-fuel fired steam electric plants.'' \8\
  
  \8\ See http://www.epa.gov/Region7/programs/artrd/air/nsr/nsrmemos/turbines.pdf .
  
  (5) ``Fossil-fuel boilers of more than 250 million Btu/hour heat input.'' We have decided to retain the interpretation that this category should be read as including only those boilers individually greater than 250 million Btu/hour heat input. We agree with commenters who noted that including smaller boilers would pose considerable implementation burden. As noted in the 2004 reproposal notice, we do not believe that this interpretation is likely to have a substantial impact. Because smaller boilers are generally less cost-effective to control, we believe that BART review would be unlikely to result in a significant amount of control on these boilers.
  
  We are also retaining the clarification that if a boiler smaller than 250 million Btu/hour heat input is an integral part of an industrial process in a BART source category other than electric utilities, then the boiler should be considered part of the BART- eligible source in that category. (By ``integral to the process'', we mean that the process uses any by-product of the boiler, or vice-versa. We have added this clarification to the definition in the BART guidelines.) We believe that if a State is already considering a BART- eligible industrial process for control, and a boiler is integrated into that process, it makes common sense not to prematurely rule out control options any of the emissions from that process as a whole. (Note that a boiler which is not integral, but is simply attached to a plant, should not be included.) For example, Kraft pulp mills may have boilers that are not serving the energy infrastructure of the plant but typically are serving a process directly by using the waste liquor from the process. Including such a boiler in consideration of control options for the process adds minimal additional burden while leaving maximum discretion to the State in determining BART for the process as a whole.
  
  We are also clarifying today that we have determined that this category should include all individual boilers of greater than 250 million Btu/hour heat input burning any amount of fossil fuel, as opposed to only those boilers that burn greater than 50 percent fossil fuel. We believe that it is quite possible that boilers of this size could contribute to regional haze in a Class I area even if they burn less than 50 percent fossil fuel. Therefore we believe that each fossil fuel-fired boiler merits a BART review. Step 2: Identify the Start-up Dates of Those Emission Units
  
  Background. BART applies only to a major stationary source which ``was in existence on August 7, 1977 but which has not been in operation for more than fifteen years as of such date.'' The visibility regulations define ``in existence'' and ``in operation'' in 40 CFR 51.301. Under these regulations, promulgated in 1980, ``in existence'' means
  
  that the owner or operator has obtained all necessary preconstruction approvals or permits * * * and either has (1) begun, or caused to begin, a continuous program of physical on-site construction of the facility or (2) entered into binding agreements or contractual obligations.
  
  [[Page 39111]]
  
  The term ``in operation'' means engaged in activity related to the primary design function of the source.
  
  Step 2 also addresses the treatment of ``reconstructions'' and ``modifications.'' Under the definition of BART-eligible facility, sources which were in operation before 1962 but reconstructed during the 1962 to 1977 time period are treated as new sources as of the time of reconstruction.\9\ The same policies and procedures for identifying reconstructed ``affected facilities'' under the NSPS are used to determine whether a source has been reconstructed for purposes of the BART requirements. ``Modifications'' under the CAA refers to physical change or change in the method of operation at a source which has led to an increase in emissions. In the proposed BART guidelines, we stated that the best interpretation of the visibility provisions is that a modification to a source does not change an emission's unit construction date for purposes of BART applicability. We requested comment on an alternative interpretation that we believed would be more difficult to implement. Under this approach, sources built before 1962 but modified during the 1962 to 1977 time frame would be considered ``new'' at the time of modification.
  
  \9\ However, sources reconstructed after 1977, which reconstruction had gone through NSR/PSD permitting, are not BART- eligible.
  
  Comments. We received comments in 2001 and 2004 on the discussion in the guideline of the term ``in existence.'' These comments were critical of our statement in the guidelines that sources which had ``commenced construction,'' that is, those which had entered into binding contracts, would be considered to be in existence, even if actual operations did not begin until after the August 7, 1977 cutoff date. These commenters asserted that Congress did not intend to treat a source as ``existing'' in 1977 if it was not yet built.
  
  Other commenters interpreted the proposed guidelines as expanding the definition of BART-eligible sources by requiring States to find that all emission units at a facility are BART-eligible if one part of the facility was built within the 1962-1977 time period. Other comments did not suggest that we had already expanded the definition in the proposed guidelines, but did suggest that we should expand the definition in that way in the final guidelines. Some commenters noted that there was a degree of confusion in the regulated community on whether the proposed guidelines were requiring BART for all units at a power plant, including those that were in operation before August 7, 1962, if these units are co-located with one or more units that were put in place within the 1962-1977 time period. These commenters requested that we clarify that such pre-1962 units would not be BART- eligible.
  
  Some commenters asserted that our proposed approach is unworkable, because the approach requires States to identify all emissions units put in place between the 1962 and 1977. Some of these commenters asserted that Congress intended that BART would apply only if entire plants satisfy the statutory criteria. These comments suggested that BART should apply only if an entire plant that is one of the 26 listed source category types had been placed in operation at a discrete point within the 15 year time period for BART eligibility. These commenters asserted that our proposed guidelines, which involved the identification and aggregation of individual emission units within the 1962-1977 time period, were inconsistent with Congress' intent. Other comments suggested that EPA could improve implementation of the program by covering discrete projects rather than individual emissions units. A few commenters suggested that for purposes of identifying such discrete projects, we consider using the term ``process or production unit'' that we used in hazardous air pollutant regulations under CAA section 112(g).
  
  One commenter requested that the guidelines clarify that emissions from ``linked'' emission units should not be considered in determining BART eligibility. That is, even if changes in emissions from one unit could affect the emissions from a ``linked'' unit that was not put in place within the 1962-1977 time period, that would not affect whether the ``linked'' unit was BART-eligible. Another commenter suggested that the approach set forth in the guidelines for identifying BART-eligible sources is inappropriate because the particular set of units identified as BART-eligible will not necessarily ``provide a reasonable and logical platform for the installation of controls.''
  
  Other commenters stated that facilities that had been modified after 1977 should not be included in the pool of sources subject to BART. Such facilities, it was argued, already meet the BART requirements because of the controls installed to meet the requirements of PSD, NSR, or the NSPS.
  
  Final rule. We disagree with the comments recommending that we interpret the term ``in existence'' to refer to sources that are in actual operation. The discussion of this term in Step 2 is based on the regulatory definition which has been in place since 1980. The guidelines reiterate this definition and provide examples of its application. Interpreting the term ``in existence'' as suggested by commenters would not be consistent with the plain language of the regulations.
  
  In the 2001 and 2004 proposed guidelines, we noted that ``the term `in existence' means the same thing as the term `commence construction' as that term is used in the PSD regulations.'' Commenters were critical of this statement, claiming that EPA was unlawfully reinterpreting section 169A in the guidelines. The statement in Step 2 of guidelines, however, is not a reinterpretation of the term ``in existence,'' but merely a statement noting that the definitions used in the visibility regulations and the PSD regulations are essentially identical.
  
  To the extent that commenters are claiming that the existing regulatory definition of ``in existence'' is unlawful, EPA's interpretation of this term in promulgating the 1980 regulations was a reasonable one. First, it is worth noting that the regulations adopting this interpretation of the term ``in existence'' were in effect in 1990 and implicitly endorsed by Congress in its 1990 amendments to the CAA.\10\ Moreover, the definition at issue accurately reflects Congress' intent that the BART provision apply to sources which had been ``grandfathered'' from the new source review permit requirements in parts C and D of title I of the CAA. For all the above reasons, we are neither revising the regional haze regulations to change the definition of ``in existence,'' nor adopting a strained interpretation of the regulation in the guidelines.
  
  \10\ See CAA section 193.
  
  We agree with commenters that the definition of ``BART-eligible source'' does not require States to find that all emission units at a facility are subject to the requirement of the BART provisions if only one part of the facility was built within the 1962-1977 time period. We received comments on this issue in 2001 and clarified in 2004 that the BART guidelines do not direct States to find that all boilers at a facility are BART-eligible if one or more boilers at the facility were put in place during the relevant time period. Under Step 2 of the process for identifying BART-eligible sources set out in the guidelines, States are required to identify only those boilers that were put in place between 1962 and 1977. As explained in the preamble to the 2004 reproposed guidelines, only these boilers are potentially subject to BART.
  
  [[Page 39112]]
  
  We do not agree with those commenters claiming that Congress clearly intended to apply BART only if an ``entire plant'' was put into place between 1962 and 1977. Most of the BART source categories are broad descriptions types of industrial facilities such as ``kraft pulp mills,'' ``petroleum refineries'' or ``primary copper smelters.'' For such source categories, the implication of commenters' argument would that if any portion of the plant was in operation before August 7, 1962, then Congress intended to exempt the entire plant from BART. Such an interpretation is problematic and inequitable. For example, under this approach BART would not apply if a company chose to expand its production by building a second production line at an existing line in 1965, but would apply if the same company chose to build the same equipment at a greenfield site. Under the approach set forth in the guidelines, such a production line would be treated similarly under either set of facts. We do not believe that either the plain language of the statute or the relevant legislative history indicate that Congress intended for major-emitting sources of visibility-impairing pollutants to be exempted from the BART requirements because a plant contains some emission units that began operation before 1962.
  
  Also, we disagree with the comment that modifications after 1977 should change an emissions' unit date of construction for purposes of BART applicability. The commenter's suggestion that such sources already meet BART requirements may be accurate, but does not provide a basis for exempting the source from review. As we note in the guideline, the review process will take into account the controls already in place and the State may find that these controls are consistent with BART.
  
  We agree with the comments related to ``linked'' emission units. The comment appears to address whether emissions from the ``linked'' units are considered in determining BART eligibility. In the guidelines, we are focusing on only the emissions units that were put in place during the 1962 to 1977 dates and the emissions from those units. We agree that even if changes in emissions from one unit could affect the emissions from a ``linked'' unit that was not put in place within the 1962-1977 time period, this would not affect whether the ``linked'' unit was BART-eligible.
  
  We disagree with commenters that the approach set forth in the guidelines for identifying BART-eligible sources is inappropriate because the particular set of units identified as BART-eligible will not necessarily ``provide a reasonable and logical platform for the installation of controls.'' We do not agree that this factor is relevant to the identification of those emissions units which meet the definition of BART-eligible source. Such factors are important in the States' consideration of control strategies and options but do not clearly relate to the first step of identifying those sources which fall within one of 26 source categories, were built during the 15 year window of time from 1962 to 1977, and have potential emissions of greater than 250 tons per year. We do thus agree generally with the commenter's recommendation of allowing States to consider the particular history and control potential of units in determining BART, but do not agree that it is relevant to the predicate question of identifying the BART-eligible source.
  
  Finally, the approach to identifying a ``BART-eligible source'' in the guidelines is based on the definitions in the regional haze rule of the relevant terms. For 750 MW power plants, States are required to apply the definitions as set forth in the guidelines; for other sources, States may adopt a different approach to the task of identifying BART-eligible sources, so long as that approach is consistent with the Act and the implementing regulations. In other words, while the guidelines adopt an approach for large power plants which involves the aggregation of all emissions units put into place between 1962 and 1977, States have the flexibility to consider other reasonable approaches to the question of identifying BART-eligible sources for other source categories.
  
  For 750 MW power plants, many of the issues identified by commenters with the approach of looking at a facility on an emission unit by emission unit basis do not exist. Unlike many types of industrial processes, power plants consist generally of a discrete number of very large emission units. For other types of facilities such as kraft pulp mills or chemical process plants which may have many small emission units that have undergone numerous changes, the guidelines do not limit the ability of the States to approach the question of identifying BART-eligible sources in ways which make sense for the particular sources given their design and history. Step 3: Compare the Potential Emissions to the 250 Ton/Yr Cutoff.
  
  Background. Step 3 of the guidelines addresses the question of whether the units identified in Steps 1 and 2 have emissions in excess of the threshold for major sources set forth in section 169A(g)(7) of the CAA. The guidelines pose the following questions to help the States in determining whether the relevant emissions units have the potential to emit in excess of the 250 tons per year threshold of any single visibility-impairing pollutant:
  
  (1) What pollutants should I address?
  
  The 2001 proposed guidelines included the following list of visibility-impairing pollutants: SO2, NOX, particulate matter, volatile organic compounds (VOCs), and ammonia. We proposed in 2001 and again in 2004 that States use PM10as the indicator for particulate matter. As explained in the guidelines, there is no need to have separate 250 ton thresholds for PM10and PM2.5because emissions of PM10include the components of PM2.5as a subset. In addition, because of various uncertainties associated with regulating VOCs and ammonia, we requested comment in 2004 on the level of discretion States should exercise in making BART determinations for VOCs and took ammonia off the list of visibility-impairing pollutants.
  
  In both proposals, we clarified that the 250 tons per year cutoff applies to emissions on a pollutant by pollutant basis. In other words, a source is subject to BART only if it emits at least 250 tons per year of an individual visibility-impairing pollutant.
  
  (2) What does the term ``potential'' emissions mean?
  
  The proposed guidelines in 2001 and the reproposed guidelines in 2004 excerpt the definition of ``potential to emit'' from the regulations at 40 CFR 51.301. As the definition makes clear, the potential to emit of a source is calculated based on its capacity to emit a pollutant taking into account its physical and operational design. Under this definition, federally enforceable emission limits may be taken into account in calculating a source's potential emissions; however, emission limitations which are enforceable only by State and local agencies, but not by EPA and citizens in Federal court, cannot be used to limit a source's potential to emit for purposes of the regional haze program.
  
  (3) What is a ``stationary source?'
  
  As explained above, States are required to make a BART determination only for ``stationary sources'' of a certain size that fall within one of 26 types of industrial categories listed in the statute and that were built within a certain time frame. The regional haze rule contains definitions that are relevant to the determination of the emissions units that comprise a ``stationary source.'' First, the regulations at 40 CFR 51.301
  
  [[Page 39113]]
  
  define ``stationary source'' as ``any building, structure, facility, or installation which emits or may emit any air pollutant.'' Second, the terms ``building, structure, or facility'' are defined in part based on grouping pollutant-emitting activities by industrial category:
  
  Building, structure, or facility means all of the pollutant- emitting activities which belong to the same industrial grouping, are located on one or more contiguous or adjacent properties, and are under the control of the same person (or persons under common control). Pollutant-emitting activities must be considered as part of the same industrial grouping if they belong to the same Major Group (i.e., which have the same two-digit code) as described in the Standard Industrial Classification Manual, 1972 as amended by the 1977 Supplement (U.S. Government Printing Office stock numbers 4101- 0066 and 003-005-00176-0 respectively).
  
  In the 2001 proposed guideline, we noted that support facilities, i.e. facilities used to convey, store, or otherwise assist in the production of the principal product, are considered to fall within the same industrial grouping as the primary facility. To clarify this, in 2004 we proposed to add language to the guideline noting that emission units at a plant, even if they are a ``support facility'' for purposes of other programs, would not be subject to BART unless they were within one of the 26 listed source categories and were built within the 1962 to 1977 time frame. Discussion of ``What Pollutants Should I Address?''
  
  Comments. PM10as an indicator. Some comments questioned the use of PM10(which includes both coarse and fine particulate matter) as the indicator for particulate matter. Commenters noted that the coarse fraction, that is particulate matter between 10 and 2.5 micrograms in diameter, fundamentally differs compared to the fine mass in how it interacts with light. Commenters suggested that only the fine mass (PM2.5) component of particulate matter is likely to contribute to visibility impairment. Accordingly, these commenters recommended that the 250 ton cutoff for particulate matter should be based upon emissions of PM2.5.
  
  Ammonia. Many commenters addressed the exclusion of ammonia from the list of visibility-impairing pollutants. A number of commenters, primarily from industry but also from one state and one regional planning organization, supported the exclusion of ammonia. These commenters generally cited the complexity and variability of ammonia's role in the formation of PM2.5in the atmosphere, the relative greater benefits of controlling NOXand SO2, the uncertainties in the inventory of ammonia emissions, and the inherent complexities of gauging the contribution of potential ammonia reductions to improving visibility in Class I areas. In addition, commenters noted that few, if any, point sources emit ammonia in amounts that exceed the 250 ton per year threshold.
  
  Other commenters, including a number of environmental groups and several states, regional planning organizations, and industry commenters, argued that ammonia should be included in the list of visibility-impairing pollutants in the guidelines. In support of this view, commenters cited evidence that ammonia is a known precursor to PM2.5. One commenter noted that improvements are being made to ammonia inventories and to the understanding of ammonia's role in the formation of haze. Other commenters pointed to a National Park Service (NPS) analysis of monitoring data that indicates that visibility-impairment due to nitrate aerosol formation (to which ammonia contributes) is of significant concern \11\ and to a 2003 direction to policy-makers from the North American Research Strategy for Tropospheric Ozone (NARSTO) \12\ indicating that consideration of control strategies needs to include ammonia in combination with other precursors to particle formation. Many commenters also argued that EPA should encourage or allow the States to consider ammonia in their visibility protection plans, and noted that ammonia reductions could be a cost-effective way to improve visibility under certain conditions.
  
  \11\ See http://wrapair.org/forums/ioc/meetings/030728/index.html (specifically presentation by John Vimont, National Park
  
  Service).
  
  \12\ NARSTO, Particulate Matter Assessment for Policy Makers: A NARSTO Assessment. P. McMurry, M. Shepherd, and J. Vickery, eds. Cambridge University Press, Cambridge, England (2004).
  
  Volatile Organic Compounds (VOCs). Several commenters responded to our request for comments on whether States should treat VOCs in urban areas differently from VOCs in rural areas. Environmental groups and a few States argued that the current state of scientific knowledge does not support a differentiation between urban and rural sources of VOCs. One environmental commenter cited evidence that organic aerosols are a major constituent of visibility-reducing aerosols and that VOCs are important precursors to the formation of secondary organic aerosols. One commenter also stated that VOCs may play a particularly significant role in particle formation in those rural areas with significant nearby sources of NOX. Commenters also cited evidence that the contribution of VOC to particle formation likely varies widely in different areas of the country, and argued that States should retain flexibility to address local VOC sources if they determine that those sources are contributors of concern.
  
  Several industry commenters stated that more focus should be placed on controlling VOCs in urban rather than rural areas. A few commenters from industry argued that VOCs in rural areas have not been shown to be a significant contributor to particle formation, and should be excluded from the list of pollutants to be addressed in the BART process. One argued that VOCs should be excluded from BART entirely based upon uncertainties in the current state of knowledge, and a few argued that VOCs from both power plants and rural sources should be excluded from BART, based on low emissions and the cost of controls. One regional planning organization requested that EPA clarify the definitions of ``urban'' and ``rural'' areas.
  
  Final rule. PM10as an indicator. While it is always necessary to assess PM2.5impacts, we agree with commenters who stated that the coarse fraction is less efficient at light scattering than fine particles, there is ample evidence that the coarse fraction does contribute to visibility impairment.\13\ For example, standard methods for calculating reconstructed light extinction routinely include a calculation for the contribution to light extinction from the coarse fraction, an implicit recognition that these particles contribute measurably to visibility impairment.\14\ We do recognize that coarse PM is likely to contribute more to regional haze in arid areas than humid areas. We believe that, as the Grand Canyon Visibility Transport Commission (GCTVC) recognized,\15\ States in the arid West in particular should take the coarse fraction of particulate matter into account in determining whether a source meets the threshold for BART applicability.
  
  \13\ See Fine particles: Overview of Atmospheric Chemistry, Sources of Emissions, and Ambient Monitoring Data, Memorandum to Docket OAR 2002-0076, April 1, 2005.
  
  \14\ These methods are described at the following Web site: http://vista.cira.colostate.edu/improve/Tools/ReconBext/reconBext.htm .
  
  \15\ Grand Canyon Visibility Transport Commission, Recommendations for Improving Western Vistas, Report to the U.S. EPA, June 10, 1996.
  
  Because long-range transport of fine particles is of particular concern in the formation of regional haze, we also
  
  [[Page 39114]]
  
  believe that it is very important to estimate the PM2.5 fraction of direct particulate emissions as correctly as possible. In addition, we believe that air quality modeling results will be more meaningful provide a more accurate prediction of a source's impact on visibility if the inputs account for the relative particle size of directly emitted particulate matter (e.g. PM10vs. PM2.5).
  
  States should consider whether their current test methods for measuring particulate matter emissions from stationary sources account for the condensible fraction of particulate matter and consider revising any such stationary source test methods to account for the condensible fraction of particulate emissions. See the source testing technical support document (TSD) in the docket for this rule, which discusses test methods for particulate matter in more detail.\16\
  
  \16\ Fine particles: Overview of Source Testing Approaches, Memorandum to Docket OAR 2002-0076, April 1, 2005.
  
  Ammonia. In regard to ammonia, we believe there is sufficient uncertainty about emission inventories and about the potential efficacy of control measures from location to location such that the most appropriate approach for States to take is a case-by-case approach. There are scientific data illustrating that ammonia in the atmosphere can be a precursor to the formation of particles such as ammonium sulfate and ammonium nitrate; \17\ however, it is less clear whether a reduction in ammonia emissions in a given location would result in a reduction in particles in the atmosphere and a concomitant improvement in visibility. In other words, the question of whether ammonia contribute to visibility impairment in a specific instance can be a difficult one.
  
  \17\ See Fine particles: Overview of Atmospheric Chemistry, Sources of Emissions, and Ambient Monitoring Data, Memorandum to Docket OAR 2002-0076, April 1, 2005.
  
  It may be that States will not be faced often with the question of addressing ammonia in making BART determinations. As noted above, States are required to make BART determinations only for stationary sources that fall within certain industrial categories. The types of sources subject to the BART provisions are not typically significant emitters of ammonia. Because of this, it is unlikely that including ammonia on the list of visibility-impairing pollutants in the BART guidelines would have much impact on the States' determinations of whether a source is BART-eligible. Thus, while ammonia can contribute to visibility impairment, we believe the decision whether to consider ammonia as a visibility-impairing pollutant in a specific case where a potential BART source actually emits more than 250 tons per year of ammonia is best left to the State.
  
  VOCs. Organic compounds can be categorized according to their varying degrees of volatility: highly reactive, volatile compounds with six or fewer carbon atoms which indirectly contribute to PM formation through the formation of oxidizing compounds such as the hydroxyl radical and ozone; semivolatile compounds with between seven and 24 carbon atoms which can exist in particle form and can readily be oxidized to form other low volatility compounds; and high molecular weight organic compounds--those with 25 carbon atoms or more and low vapor pressure--which are emitted directly as primary organic particles and exist primarily in the condensed phase at ambient temperatures. The latter organic compounds are considered to be primary PM2.5 emissions and not VOCs for BART purposes.
  
  Current scientific and technical information shows that carbonaceous material is a significant fraction of total PM2.5mass in most areas and that certain aromatic VOC emissions such as toluene, xylene, and trimethyl-benzene are precursors to the formation of secondary organic aerosol.\18\ However, while progress has been made in understanding the role of VOCs in the formation of organic PM, this relationship remains complex, and issues such as the relative importance of biogenic versus anthropogenic emissions remain unresolved.
  
  \18\ Ibid.
  
  Therefore we believe that the best approach for States to follow in considering whether VOC emissions are precursors to PM2.5 formation is a case-by-case approach. States should consider, in particular, whether a source's VOC emissions are those higher-carbon VOCs that are more likely to form secondary organic aerosols. In addition, given the variable contribution of a given amount of VOC emissions to PM2.5formation, States may also wish to exercise discretion in considering only relatively larger VOC sources to be BART-eligible.
  
  After careful consideration of the comments, we agree with commenters who assert that EPA should not suggest a general distinction between the relative contributions of urban and rural VOC emissions to particle formation. The state of knowledge in this area is complex and rapidly evolving. Monitoring data in the East \19\ suggest that there may be a greater contribution to particle formation in urban areas from VOCs as compared to rural areas, but we recognize that further research is needed to better determine the extent of the contribution of specific VOC compounds to organic PM mass. We do not agree, however, with commenters who make the blanket assertion that rural VOCs are not a significant contributor to particle formation, as it is possible that in specific areas, such as where NOXemissions are high, rural anthropogenic VOCs could potentially play a significant role.
  
  \19\ Ibid.
  
  Discussion of the Term ``Potential'' Emissions
  
  Comments. A number of commenters were critical of the restriction in the regional haze rule that allows States to credit federally enforceable limitations on emissions but not limitations that are enforceable only by States and local agencies. These commenters believed that this restriction had been rejected by the D.C. Circuit for a number of other EPA regulations and noted that EPA has developed policies that currently credit state-enforceable limits. The comments recommended that EPA issue guidance consistent with what commenters claimed were current policies for other regulations. In addition, we received comments arguing that in determining whether a source is a major stationary source, the States should consider a source's actual-- rather than potential--emissions. These commenters stated that using a source's potential emissions overstates a source's actual emissions and impacts on visibility.
  
  Final rule. CAA section 169A(g)(7) defines a ``major stationary source'' as a source with the potential to emit 250 tons or more any pollutant. Based inter alia on that statutory definition, EPA's implementing regulations define BART-eligible sources as those with the potential to emit 250 tons or more of any air pollutant. As these definitions clearly require consideration of a source's potential emissions, the guidelines state that a State should determine whether a source's potential emissions exceed the 250 ton threshold in determining whether the source is BART-eligible.
  
  As explained in the 2001 and 2004 proposed guidelines, the regional haze regulations define ``potential to emit.'' The guidelines repeat that regulatory definition and provide an example illustrating its application. EPA did not propose to change the definition in 2001 or 2004, but merely highlighted the
  
  [[Page 39115]]
  
  current definition in 40 CFR 51.301. Although we noted in the 2001 proposed guidelines that we expected to undertake a rulemaking to determine whether only federally enforceable limitations should be taken into account in the regional haze program definition, we have not yet begun the process for such a rulemaking. However, we consider the comments criticizing EPA's definition of ``potential to emit'' as a request for reconsideration of the visibility regulations and will take these requests into account in determining any future rulemaking efforts to address the general definition of ``potential to emit.'' For the time being, we believe that States may consider federally enforceable limits or emissions limitations in State permits, which are enforceable under State law, in determining a source's ``potential to emit.'' Discussion of What Emissions Units Should Be Considered Part of a ``Stationary Source''
  
  Comments. A number of comments in 2001 expressed concern with our statement that a ``support facility'' should be grouped with a primary facility in determining which emissions units belong to the same industrial grouping. These comments generally coincided with comments discussed above that EPA should determine BART on a plantwide basis, rather than by aggregating emissions units. Commenters on the 2004 reproposal noted with approval the clarification that ``support facilities'' should only be considered BART-eligible if these units themselves were both constructed within the 1962-1977 time frame and fell within one of the listed source categories.
  
  Two commenters felt that we should more clearly define the BART- eligible source, either by identifying emission units within source categories, or by somehow accounting for the specific set of emission units, within the fenceline, to which controls would logically apply.
  
  Final rule. The guidelines continue to note that the definition of ``building, structure or facility'' in the regional haze rule is based upon aggregating emissions units within the same industrial grouping. This discussion in the guidelines is consistent with the language in the definition of ``building, structure or facility'' in the regional haze rule which contains a specific reference to the 2-digit SIC classifications. The BART guidelines refer to this definition and explain how 2-digit SIC codes are used in determining the scope of BART for a given plantsite. (In the rare situation where industrial groupings in separate 2-digit SIC codes exist at a single plant site, then there would be more than one separate ``stationary source'' present. In that situation, each ``stationary source'' should be looked at individually for purposes of determining BART-eligibility.)
  
  We agree that more clarity is needed to account for situations where a specific set of units constitute the logical set to which BART controls would apply. The CAA requires BART at certain major stationary sources. Accordingly we believe it could be appropriate, at the BART determination step, for States to allow sources to ``average'' emissions across a set of BART-eligible emission units within a fenceline, so long as the amount of emission reductions from each pollutant being controlled for BART would be at least equal to those reductions that would be obtained by simply controlling each unit. We have added language to the guidelines to this effect. Step 4: Identify the Emission Units and Pollutants That Constitute the BART-Eligible Source
  
  Background. The final step in identifying a ``BART-eligible source'' is to use the information from the previous three steps to identify the universe of equipment that makes up the BART-eligible source. The 2001 and 2004 proposed BART guidelines stated that if the emissions from the list of emissions units at a stationary source exceed a potential to emit of 250 tons per year for any individual visibility-impairing pollutant, then that collection of emissions units is a BART-eligible source. The guidelines also stated that a BART analysis would be required for each visibility-impairing pollutant emitted from this collection of emissions units.
  
  In the 2004 reproposed BART guidelines, we noted that we believed that section 169A(b)(2)(A) of the CAA requires a State to undertake a BART analysis for ``any'' visibility-impairing pollutant emitted by a BART-eligible source, regardless of the amount emitted. We proposed, however, to provide the States with the flexibility to identify de minimis levels for pollutants at BART-eligible sources, but limited that flexibility so that any such de minimis levels could not be higher than those used in the PSD program: 40 tons per year for SO2, NOX, and VOC, and 15 tons per year from PM10. We requested comment on this provision and on the use of de minimis values. Discussion of Whether To Include All Emitted Visibility-Impairing Pollutants in the BART Analysis
  
  Comments. A number of commenters supported the concept of including all pollutants in the BART analysis once an individual pollutant triggers the BART review. Other commenters, although supportive of the concept generally, recommended that we should add the pollutants together before the comparison with the threshold.
  
  A number of commenters disagreed with EPA's conclusion that the CAA requires States to make a BART determination for any visibility- impairing air pollutant emitted by a BART eligible source. These commenters stated that undertaking a BART analysis for all pollutants emitted by a major stationary source is an unnecessary administrative burden with minimal environmental benefit. Commenters argued that Congress intended for BART to apply only to those pollutants for which a source is major. Commenters accordingly recommended that the 250 ton per year threshold apply to each pollutant emitted by a source and that BART apply only to those pollutants which meet this threshold. A number of these commenters argued alternatively that only those pollutants from a source demonstrated, individually, to cause or contribute to visibility impairment are required to go through a BART determination.
  
  Final rule. We disagree with the comment that emissions of different visibility-impairing pollutants must be added together to determine whether a source exceeds the 250 ton per year threshold. The CAA, in section 169A(g)(7), defines a ``major stationary source'' as one with the potential to emit 250 tons or more of ``any pollutant.''
  
  We disagree with comments that the BART analysis is required only for those pollutants that individually exceed the 250 ton per year threshold. Section 169A(b)(2)(A) specifically requires States to submit SIPs that include a requirement that a major stationary source
  
  which, as determined by the State * * * emits any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in any [Class I area], shall procure, install, and operate * * * the best available retrofit technology, as determined by the State * * * for controlling emissions from such source for the purpose of eliminating or reducing any such impairment.
  
  The regional haze regulations similarly require that the States submit a SIP that contains
  
  [[Page 39116]]
  
  A determination of BART for each BART-eligible source in the State that emits any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in any mandatory Class I Federal area.
  
  40 CFR 51.308(e)(1)(ii). Nothing in these statutory or regulatory requirement suggests that the BART analysis is limited to those pollutants for which a source is considered major. At best, these provisions can be read as requiring a BART determination only for those emissions from a specific source which do, in fact, cause or contribute to visibility impairment in a particular Class I area, or which could reasonably be anticipated to do so. Commenters, however, have not presented any evidence that as a general matter emissions of less than 250 tons per year of PM2.5, SO2, or other visibility-impairing pollutants from potential BART sources do not ``cause or contribute to any impairment of visibility'' in any of the Class I areas covered by the regional haze rule. As there is no such evidence currently before us, there is no basis to conclude that the States are required to make BART determinations only for those pollutants emitted in excess of 250 tons per year.
  
  At the same time, we agree with certain commenters that the CAA does not require a BART determination for any visibility impairing pollutant emitted by a source, regardless of the amount. After reviewing the language of the Act and the comments received, we have concluded that our interpretation of the relevant language in section 169A(b)(2)(A) of the Act in the 2004 proposed guidelines is not necessarily the best reading of the BART provisions. Section 169A(b)(2)(A) of the Act can be read to require the States to make a determination as to the appropriate level of BART controls, if any, for emissions of any visibility impairing pollutant from a source. Given the overall context of this provision, however, and that the purpose of the BART provision is to eliminate or reduce visibility impairment, it is reasonable to read the statute as requiring a BART determination only for those emissions from a source which are first determined to contribute to visibility impairment in a Class I area.
  
  The interpretation of the requirements of the regional haze program reflected in the discussion above does not necessitate costly and time- consuming analyses. Consistent with the CAA and the implementing regulations, States can adopt a more streamlined approach to making BART determinations where appropriate. Although BART determinations are based on the totality of circumstances in a given situation, such as the distance of the source from a Class I area, the type and amount of pollutant at issue, and the availability and cost of controls, it is clear that in some situations, one or more factors will clearly suggest an outcome. Thus, for example, a State need not undertake an exhaustive analysis of a source's impact on visibility resulting from relatively minor emissions of a pollutant where it is clear that controls would be costly and any improvements in visibility resulting from reductions in emissions of that pollutant would be negligible. In a scenario, for example, where a source emits thousands of tons of SO2but less than one hundred tons of NOX, the State could easily conclude that requiring expensive controls to reduce NOX would not be appropriate. In another situation, however, inexpensive NOXcontrols might be available and a State might reasonably conclude that NOXcontrols were justified as a means to improve visibility despite the fact that the source emits less than one hundred tons of the pollutant. Moreover, as discussed below, we are revising the regional haze regulations to allow the States to exempt de minimis emissions of SO2, NOX, and PM2.5from the BART determination process which should help to address the concerns of certain commenters associated with the burden of a broad BART analysis.
  
  De minimis levels
  
  Comments. Many commenters agreed that we should establish de minimis levels for individual pollutants in order to allow States and sources to avoid BART determinations for pollutants emitted in relatively trivial amounts. Many commenters suggested that States would be unlikely to impose emission limits for pollutants emitted at the proposed de minimis levels because it would not be cost-effective to do so and such emission reductions could not be expected to produce any perceptible improvements in visibility. Several commenters agreed that the pollutant coverage requirements for BART eligibility should be consistent with those for the PSD program, but others argued that BART should be required only for pollutants emitted in amounts greater than 250 tons per year. Commenters also noted that the guidelines were not clear as to whether the de minimis provision would apply on a plant- wide or unit by unit basis. A few commenters also noted that the final guidelines should clarify where in the BART determination process de minimis levels may be used.
  
  Other commenters opposed the use of de minimis exemptions. These commenters argued that it would be unreasonable to rule categorically that a certain level of emissions had a trivial impact on visibility without assessing the impacts of these emissions in particular circumstances. These commenters argued that States should consider the emissions of all visibility-impairing pollutants in a BART determination regardless and that, consequently, there should be no de minimis levels.
  
  Final rule. As proposed in 2004, we believe that it is reasonable to give States the flexibility to establish de minimis levels so as to allow them to exempt from the BART determination process pollutants emitted at very low levels from BART-eligible sources. As explained by the D.C. Circuit, ``categorical exemptions from the requirements of a statute may be permissible `as an exercise of agency power, inherent in most statutory schemes, to overlook circumstances that in context may fairly be considered de minimis.' '' \20\ The ability to create de minimis exemptions from a statute is a tool to be used in implementing the legislative design.\21\
  
  \20\ EDF et al. v. EPA, 82 F.3d 451, 466 (D.C. Cir. 1996) citing Alabama Power v. Costle, 636 F.2d 323 (D.C. Cir. 1979).
  
  \21\ Id.
  
  The intent of Congress in requiring controls on emissions from certain major stationary sources was to eliminate or reduce any anticipated contribution to visibility impairment from these sources. This, as section 169A(b)(2)(A) states, is the ``purpose'' of BART. In making a determination as to the appropriate level of controls, however, the States are required to take into account not only the visibility benefits resulting from imposing controls on these sources but also the costs of complying with the BART provision. The BART provision is accordingly designed to ensure that the States take into consideration all emissions of certain stationary sources in making a BART determination, but also to provide States with the flexibility to include the costs and benefits of controlling these sources in the calculus of determining the appropriate level of BART.
  
  We believe it would be permissible for States to create de minimis levels at a low level. If a State were to undertake a BART analysis for emissions of less than 40 tons of SO2or NOXor 15 tons of PM10from a source, it is unlikely to result in anything but a trivial improvement in visibility. This is
  
  [[Page 39117]]
  
  because reducing emissions at these levels would have little effect on regional emissions loadings or visibility impairment. We believe most States would be unlikely to find that the costs of controlling a few tons of emissions were justified. Because the overall benefits to visibility of requiring BART determinations for emissions of less than the de minimis levels would be trivial, we are amending the regional haze rule to make clear that the States have this flexibility.
  
  The de minimis levels discussed today apply on a plant-wide basis. Applying de minimis levels on a unit by unit basis as suggested by certain commenters could exempt hundreds of tons of emissions of a visibility-impairing pollutant from BART analysis. In at least some of the twenty-six source categories covered by the BART provisions, a single control device can be used to control emissions from multiple units. Thus, it is possible that while emissions from each unit are relatively trivial, the costs of controlling emissions from multiple units might be cost-effective in light of the BART-eligible source's total emissions of the pollutant at issue. States should consider the control options in such situations and determine the appropriate approach for the specific source.
  
  We are revising the regional haze rule to provide States with the ability to establish de minimis levels up to the levels proposed in 2004. We believe States may, if they choose, exclude from the BART determination process potential emissions from a source of less than forty tons per year of SO2or NOX, or 15 tons per year for PM10. (Note also that for sources that are BART- eligible for one pollutant, we also believe that States could allow those sources to model the visibility impacts of pollutants at levels between de minimis and 250 tons in order to show that the impact is negligible and should be disregarded. See section D below). In the guidelines, we include this as part of the BART determination in section IV of the guidelines. (We note that these emission levels represent the maximum allowable de minimis thresholds--States retain their discretion to set the thresholds at lesser amounts of each pollutant, or to not provide any pre-determined de mininis levels.) We believe that this approach is the clearest method for exempting trivial emissions from the BART determination process. Alternatively, States may find it useful to exclude de minimis emissions in identifying whether a source is subject to BART in section III of the guidelines. Either approach is consistent with the regulation issued in this rule.
4. How To Determine Which BART-eligible Sources Are ``Subject to BART''
  
  Cause or Contribute
  
  Background. Under section 169A(b)(2)(A) of the Act, each State must review its BART eligible sources and determine whether they emit ``any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in [a Class I] area.'' If a source meets this threshold, the State must then determine what is BART for that source.
  
  Proposed rule. In the reproposed guidelines, we identified three options for States to use in determining which BART-eligible sources meet the test set forth in section 169A(b)(2)(A) of the CAA. To determine whether a BART-eligible source is ``reasonably anticipated to cause or contribute to visibility impairment,'' the first proposed option was that a State could choose to consider the collective contribution of emissions from all BART-eligible sources and conclude that all BART-eligible sources within the State are ``reasonably anticipated to cause or contribute'' to some degree of visibility impairment in a Class I area. The preamble to the 1999 regional haze rule explains at length why we believe that looking to the collective contribution of many sources over a broad area is a reasonable approach, and we explained in the 2004 reproposed guideline that we believed that a State's decision to use a cumulative analysis at this stage of the BART determination process would be consistent with the CAA and the findings of the D.C. Circuit in American Corn Growers.
  
  The second proposed option was to allow a State to demonstrate, using a cumulative approach, that none of its BART-eligible sources contribute to visibility impairment. Specifically, we proposed to provide States with the option of performing an analysis to show that the full group of BART-eligible sources in a State cumulatively do not cause or contribute to visibility impairment in any Class I areas.
  
  As a third option, we proposed that a State may choose to determine which sources are subject to BART based on an analysis of each BART- eligible source's individual contribution. We labeled this option as an ``Individualized Source Exemption Process,'' and proposed that States use an air quality model to determine an individual source's contribution to visibility impairment, calculated on a 24 hour basis, using allowable emissions, and compared to an established threshold.
  
  Comments. Several commenters expressed the view that EPA was misinterpreting the American Corn Growers case to allow the States to apply a collective contribution test in determining whether BART- eligible sources are subject to BART. These commenters took the position that, because this approach does not allow for a source to show that it does not individually cause or contribute to visibility impairment, it is incompatible with the language of section 169A(b)(2)(A)of the Act. They argued that EPA should modify the provisions in the proposed rule to ensure that an individual source is afforded the opportunity to conduct an analysis to demonstrate that its emissions do not impair visibility in any Class I area. Conversely, several commenters indicated that the option to determine that all potential BART sources contribute to regional haze should be the starting point of determining BART eligibility.
  
  Many industry commenters and some States supported the second proposed option which would allow a State to demonstrate through an analysis of the collective contribution of all its BART-eligible sources that none of these sources contribute to visibility impairment. Several of these commenters added, however, that if this cumulative analysis were to show a contribution, then, consistent with the decision in American Corn Growers, the State must allow each individual source to demonstrate that its own emissions do not, by themselves, contribute to the problem of visibility impairment. One commenter requested clarification on what visibility threshold a State should use in determining that no sources are reasonably anticipated to cause or contribute to any impairment in a Class I area.
  
  A number of commenters supported the third option for determining BART applicability based on an analysis of source-specific effects on visibility. However, many of the commenters stated that the CAA requires that the States either conduct such an analysis in determining those sources subject to BART, or allow an individual source to make a showing that it does not cause or contribute to visibility impairment. In addition, although supportive of the general notion of allowing for an exemption process for BART-eligible sources, several commenters stated that the third option contained burdensome modeling requirements, and that States need a more flexible, straightforward,
  
  [[Page 39118]]
  
  and less costly method to make the ``cause or contribute'' determination.
  
  Several environmental groups commented that the proposed options potentially go too far in allowing sources to be exempted from the BART requirements. These commenters asserted that EPA should clarify that States may not allow a BART-eligible source to avoid the BART requirements without an affirmative demonstration by the State, or by the source, showing that the source does not emit any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in a Class I area. Absent such a demonstration, they argue, a State may not choose to waive the requirement to conduct a BART review of the source.
  
  Final rule. The final BART guidelines adopt the general approach contained in the reproposal, providing the States with several options for identifying the sources subject to BART. The final BART guidelines describe the options contained in the reproposal as well as one new option. The discussion of options in the final guidelines are structured somewhat differently from the reproposal, and the options are explained in greater detail. The guidelines reaffirm that a State may choose to consider all BART-eligible sources to be subject to BART, and to make BART determinations for all its BART-eligible sources.\22\ For States that choose to consider exempting some or all of their BART- eligible sources from review, the guidelines then discuss three options that States may use to determine whether its sources are ``reasonably anticipated to cause or contribute'' to visibility impairment at a Class I area. Options 1 and 3 are similar to options in the 2004 reproposal; under option 1, States may use an individual source attribution approach, while option 3 provides the States with an approach for demonstrating that no sources in a State should be subject to BART. Option 2 is new; it is an approach for using model plants to exempt individual sources with common characteristics.
  
  \22\ States choosing this approach should use the data being developed by the regional planning organizations, or on their own, as part of the regional haze SIP development process to make the showing that the State contributes to visibility impairment in one or more Class I areas.
  
  Threshold for visibility impact. One of the first steps in determining whether sources cause or contribute to visibility impairment for purposes of BART is to establish a threshold (quantified in units called ``deciviews'') against which to measure the visibility impact of one or more sources. We believe that a single source that is responsible for a 1.0 deciview change or more should be considered to ``cause'' visibility impairment; a source that causes less than a 1.0 deciview change may still contribute to visibility impairment and thus be subject to BART.
  
  The guidelines note that because of varying circumstances affecting different Class I areas, the appropriate threshold for determining whether a source ``contributes to any visibility impairment'' for the purposes of BART may reasonably differ across States. Although the appropriate threshold may vary, the Guidelines state that the contribution threshold used for BART applicability should not be higher than 0.5 deciviews. We discuss threshold issues in greater detail in the subsection immediately following this one, entitled Metric for Visibility Degradation. Pollutants
  
  The guidelines direct that States should look at SO2, NOX, and direct particulate matter (PM) emissions in determining whether sources cause or contribute to visibility impairment, including both PM10and PM2.5. Consistent with the approach for identifying BART-eligible sources, States do not need to consider less than de minimis emissions of these pollutants from a source.
  
  States may use their best judgement to determine whether VOC or ammonia emissions are likely to have an impact on visibility in an area. In addition, they may use PM10or PM2.5as an indicator for PM2.5in determining whether a source is subject to BART. In determining whether a source contributes to visibility impairment, however, States should distinguish between the fine and coarse particle components of direct particulate emissions. Although both fine and coarse particulate matter contribute to visibility impairment, the long-range transport of fine particles is of particular concern in the formation of regional haze. Air quality modeling results used in the BART determination will provide a more accurate prediction of a source's impact on visibility if the inputs into the model account for the relative particle size of any directly emitted particulate matter (i.e. PM10vs. PM2.5).
  
  We believe that PM10is likely to contribute more to regional haze in arid areas than humid areas. As the Grand Canyon Visibility Transport Commission (GCTVC) recognized,\23\ States in the arid West, in particular, will need to take the coarse fraction of particulate matter into account in determining whether a source meets the threshold for BART applicability.
  
  \23\ Grand Canyon Visibility Transport Commission, Recommendations for Improving Western Vistas, Report to the U.S. EPA, June 10, 1996.
  
  Option 1. We agree with commenters supporting the use of an individual source analysis in determining if a BART-eligible source causes or contributes to visibility impairment. Consistent with American Corn Growers, this option provides a method for a State to evaluate the visibility impact from an individual source and show that the source is not reasonably anticipated to cause or contribute to visibility degradation in a Class I area and thus may be exempt from BART. (Note also that an individual source analysis is used to inform the BART determination). In general, a dispersion model is used to assess the visibility impact from a single source, and that impact is compared to a threshold which is determined by the State. The threshold (quantified in deciviews) is the numerical metric that is used to define ``cause or contribute''; if a source's impact is below the threshold, a State may exempt the source from BART; otherwise the source would be subject to BART.
  
  We discuss specific issues on the individualized source attribution process, including changes since proposal and issues raised by commenters, in the subsections immediately following this one: Metric for visibility degradation; Use of CALPUFF for visibility modeling; The use of natural conditions in determining visibility impacts for reasonable progress and comparison to threshold values; Modeling protocol; and Alternatives for determining visibility impacts from individual sources.
  
  Option 2. In the final guideline, we describe a modified approach, using model plants based on representative sources sharing certain characteristics, that the States may use to simplify the BART determination process, either to exempt (individually or as a group) those small sources that are not reasonably anticipated to cause or contribute to visibility impairment, or to identify those large sources that clearly should be subject to BART review. States could use the CALPUFF model, for example, to estimate levels of visibility impairment associated with different combinations of emissions and distances to the nearest Class I area. In carrying out this approach, the State could then reflect groupings of specific types of sources with important common characteristics, such as emissions, stack heights and plume characteristics, and develop ``composite model plants.'' Based on CALPUFF
  
  [[Page 39119]]
  
  analyses of these model plants, a State may find that certain types of sources are clearly reasonably anticipated to cause or contribute to visibility impairment. Conversely, representative plant analyses may show that certain types of sources are not reasonably anticipated to cause or contribute to visibility impairment. Based on the modeling results, a State could exempt from BART all sources that emit less than a certain amount per year and that are located a certain distance from the nearest Class I area.
  
  Our analyses of visibility impacts from model plants provide a useful example of the type of analyses that might be used to exempt categories of sources from BART.\24\ Based on our model plant analysis, EPA believes that a State could reasonably choose to exempt sources that emit less than 500 tons per year of NOXor SO2(or combined NOXand SO2), as long as they are located more than 50 kilometers from any Class I area; and sources that emit less than 1000 tons per year of NOXor SO2(or combined NOXand SO2) that are located more than 100 kilometers from any Class I area.
  
  \24\ Supplement to CALPUFF Analysis in Support of the June 2005 Changes to the Regional Haze Rule, U.S. Environmental Protection Agency, June 15, 2005, Docket No. OAR-2002-0076.
  
  In our analysis, we developed two model plants (a EGU and a non- EGU), with representative plume and stack characteristics, for use in considering the visibility impact from emission sources of different sizes and compositions at distances of 50, 100 and 200 kilometers from two hypothetical Class I areas (one in the East and one in the West). Because the plume and stack characteristics of these model plants were developed considering the broad range of sources within the EGU and non-EGU categories, they do not necessarily represent any specific plant. However, the results of these analyses may be instructive in the development of an exemption process for groups of BART-eligible sources, without modeling each of these sources individually.
  
  States may want to conduct their own model plant analysis that take into account local, regional, and other relevant factors (such as meteorology, sulfur dioxide, nitrogen dioxide, and ammonia). If so, you may want to consult your EPA Regional Office to ensure that any relevant technical issues are resolved before you conduct your modeling.
  
  In preparing our hypothetical examples, we have made a number of assumptions and exercised certain modeling choices; some of these have a tendency to lend conservatism to the results, overstating the likely impacts, while others may understate the modeling results. On balance, when all of these factors are considered, we believe that our examples reflect realistic treatments of the situations being modeled.\25\ A summary of the more significant elements and their implications is provided below.
  
  \25\ CALPUFF Analysis in Support of the June 2005 Changes to the Regional Haze Rule, U.S. Environmental Protection Agency, June 15, 2005, Docket No. OAR-2002-0076.
  
  Features of the modeling examples which may understate visibility impacts
  
  An annual emission rate was used for the example modeling (e.g. 10,000 TPY divided by 365 days divided by 24 hours). ``Real world'' sources have variable emission rates, and in any 24 hour period may be operating well above the annual rate.
  
  The monthly average relative humidity was used, rather than the daily average humidity, and would contribute to lowering the peak values in daily model averages.
  
  A 24-hour average was calculated from modeled hourly visibility impacts, reducing the impact of any one particular hour that could be higher due to a number of meteorological effects.
  
  Features of the modeling examples which may overstate visibility impacts
  
  We located receptors using a grid of concentric circles for distances of 50, 100 and 200 km. A receptor was placed every 10 degrees around each circle, and highest impacts were reported regardless of direction from the source. In actuality, receptors would be located only in the Class I area, or in only one direction from the source.
  
  We used simplified chemistry (i.e. for conversion of SO2and NOXto fine particles) and disperson techniques which tend to overstate model impacts.
  
  Special care should be used to ensure that the criteria used in the modeling are appropriate for a given State. Our modeling may not be appropriate for every region of the country, due to the unique characteristics of different Class I areas and varying meteorological and geographical conditions in different regions. In addition, States may want to design their own model plants taking into account the types of sources at issue in their region.
  
  Option 3. Under the BART guidelines, a State may consider exempting all its BART-eligible sources from BART by conducting analyses that show that all of the emissions from BART-eligible sources in their State, taken together, are not reasonably anticipated to cause or contribute visibility impairment. To make such a showing, a State could use CALPUFF or another appropriate dispersion model to evaluate the impacts of individual sources on downwind Class I areas, aggregating those impacts to determine the collective contribution from all-BART eligible sources in the State. A State with a sufficiently large number of BART-eligible sources could also make such a showing using a photochemical grid model.\26\
  
  \26\ For regional haze applications, regional scale modeling typically involves use of a photochemical grid model that is capable of simulating aerosol chemistry, transport, and deposition of airborne pollutants, including particulate matter and ozone. Regional scale air quality models are generally applied for geographic scales ranging from a multi-state to the continental scale. Because of the design and intended applications of grid models, they may not be appropriate for BART assessments, so States should consult with the appropriate EPA Regional Office prior to carrying out any such modeling.
  
  We agree with commenters who pointed out that the option of allowing a State to demonstrate that the full group of BART-eligible sources in the State do not contribute to visibility impairment would, by default, satisfy an individual source contribution assessment. Commenters have not shown any reason to believe that if the sum total of emissions from the BART-eligible sources in a State do not ``cause or contribute'' to visibility impairment in any Class I area, that emissions from one such source will meet the threshold for BART applicability. A State following this approach accordingly need not undertake an affirmative demonstration based on a source by source analysis of visibility impacts to find that its sources are not subject to BART. Metric for Visibility Degradation
  
  Background. The 2004 reproposed guidelines contained a proposed threshold for the States to use in determining whether an individual source could be considered to cause visibility impairment in a Class I area. We proposed a 0.5 deciview change relative to natural background conditions,\27\ as a numerical threshold for making this determination.\28\
  
  \27\ Guidance for Estimating Natural Visibility Conditions Under the Regional Haze Rule, ( U.S. Environmental Protection Agency, September 2003. http://www.epa.gov/ttncaaa1/t1/memoranda/rh_envcurhr_gd.pdf. Natural background conditions, expressed in
  
  deciviews, are defined for each Class I area. EPA has issued guidance for estimating natural background conditions which has estimates of default conditions as well as measures to develop refined estimates of natural conditions.
  
  \28\ In the proposal we noted that a 0.5 deciview change in visibility is linked to ``perceptibility,'' or a just noticeable change in most landscapes. National Acid Precipitation Assessment Program (NAPAP), Acid Deposition: State of Science and Technology Report 24, Visibility: Existing and Historical Conditions--Causes and Effects (Washington, DC, 1991) Appendix D at 24-D2 (``changes in light extinction of 5 percent will evoke a just noticeable change in most landscapes''). Converting a 5 percent change in light extinction to a change in deciviews yields a change of approximately 0.5 deciviews.
  
  [[Page 39120]]
  
  We proposed the CALPUFF model as the preferred approach for predicting whether a single source caused visibility impairment if the modeled results showed impacts from the source that exceeded the threshold on any given day during a five-year period. We also proposed that if a source had an estimated impact on visibility of less than 0.5 deciviews, a State could choose to exempt the source from further BART analysis.
  
  Comments. We received numerous comments supporting the proposed threshold. A number of commenters stated that the 0.5 deciview threshold is appropriate given the low triggering threshold for applicability established by Congress, and that the literature supports it as the minimum level of perceptibility. Some commenters cited published documentation supporting their assertions that a minimum change in deciviews necessary for perceptibility is 0.5 deciviews.\29\
  
  \29\ Ibid.
  
  Other commenters criticized the threshold as too low. They stated that a change of 0.5 deciviews is inconsistent with language in the regional haze rule pointing to 1.0 deciview as the appropriate perceptibility threshold, and they cited more recent literature justifying perceptibility as greater than a change of 1 deciview.\30\
  
  \30\ Henry, R.C., Just-Noticeable Differences in Atmospheric Haze, Journal of the Air & Waste Management Association, 52:1238- 1243, October 2002.
  
  One commenter said that we should allow States and regional planning organizations (RPOs) the flexibility to determine appropriate visibility-impact thresholds in light of current knowledge about a range of perceptibility thresholds. Another commenter said that we should explain our basis for establishing a threshold of a one-time impact of greater than 0.5 deciviews, in light of the overall goal of the regional haze program. Yet another commenter said that the proposal would ``change the regulatory role of the deciview metric by converting it into a regulatory 0.5 deciview standard (versus a `goal') for defining how States must exercise their authority and discretion in determining whether an individual source `causes or contributes' to visibility impairment in a Class I area.''
  
  Several commenters said that the 0.5 deciview threshold is too high. A recurring comment was that the statutory BART applicability test from CAA Section 169A(b)(2)(A) contains two separate elements: ``causation'' of any visibility impairment and ``contribution'' to any such impairment. Commenters pointed out that by setting a threshold of 0.5 deciviews, we had combined ``cause or contribute'' into a single test of causality, thus effectively eliminating the ``contribution'' element of the BART applicability test. The commenters asserted that a single BART-eligible source can ``contribute'' to visibility impairment with impacts much lower than 0.5 deciviews. They argued that we must set the minimum threshold for individual source contribution to visibility impairment at the lowest level detectable by modeling or other appropriate analysis, and that this minimum individual contribution level must in any event be set at no greater than a 0.1 deciview change relative to natural conditions, which is a clearly measurable level. One commenter suggested that a cause or contribute threshold be set at some percentage of the ``just noticeable'' change of 0.5 deciviews.
  
  Another commenter said that in a case where multiple sources each have a visibility impact of less than a 0.5 deciview change, but together result in a change of more than 0.5 deciview, each of these sources contributes to the resulting visibility impairment. This commenter asserted that BART guidelines that result in exemptions for these ``contributing'' sources would subvert the goals of the regional haze program.
  
  Similarly, several commenters suggested that if any combination of BART eligible sources causes visibility impairment in a Class I area of more than 0.5 deciviews (by CALPUFF modeling for any 24-hour period, for example), that State should determine that each individual source is subject to BART. Thus, the commenter added, the court's concern about the lack of ``empirical evidence of a source's contribution to visibility impairment'' would be addressed.
  
  Two commenters said that our requirement to use the maximum 24-hour value over the 5-year period of meteorological data in the modeling, as proposed, is too stringent, unreasonable, inappropriate, and departs from the previous methodologies for the regional haze program. Additionally they said that the threshold is restrictive because the single highest 24-hour modeled impact over a three- or five-year period may be influenced by short-term weather conditions, like high humidity, and the BART applicability determination should not be made based on a one-time occurrence.
  
  One commenter said that whatever the final threshold for a single- source impact for BART sources, EPA should clarify that the purpose of this modeling assessment is to evaluate a source's anticipated contribution to uniform regional haze over the Class I area. EPA should state that the assumption of a uniform haze contribution based on CALPUFF modeling eliminates the need to assess issues related to the size of the Class I area, views within a Class I area, and weather impact interactions. Finally, one commenter said that thresholds should be established separately for the eastern and western regions of the United States, as natural visibility conditions are established separately for eastern and western regions in the guidance.
  
  Final Rule. Today's guidelines advise States to use a deciview metric in defining ``cause or contribute,'' as explained further below. The fact that the deciview is also used to track progress toward the goal of natural visibility does not in any way indicate that we are ``converting'' a ``goal'' into a requirement.\31\ Use of the same metric in the ``cause or contribute'' context as used for establishing reasonable progress goals, tracking changes in visibility conditions, and defining baseline, current, and natural conditions simply provides for a consistent approach to quantifying visibility impairment.
  
  \31\ Moreover, the fact that the ultimate purpose of the visibility provisions is expressed as a ``goal'' does not mean that all aspects of the program are merely aspirational. CAA section 169A(a)(4) requires EPA to establish regulations to ensure that reasonable progress is made toward the national visibility goal, and 169A(b)(2) provides that EPA must require SIPs to contain emission limits, schedules of compliance, and other measures as may be necessary to make reasonable progress towards meeting the goal.
  
  In response to commenters who said we conflated the ``cause or contribute'' test, we are clarifying that for purposes of determining which sources are subject to BART, States should consider a 1.0 deciview change or more from an individual source to ``cause'' visibility impairment, and a change of 0.5 deciviews to ``contribute'' to impairment.\32\
  
  \32\ If ``causing'' visibility impairment means causing a humanly perceptible change in visibility in virtually all situations (i.e. a 1.0 deciview change), then ``contributing'' to visibility impairment must mean having some lesser impact on the conditions affecting visibility that need not rise to the level of human perception.
  
  In a regulatory context, we believe that a State's decision as to an
  
  [[Page 39121]]
  
  appropriate threshold for contribution could depend upon the number of sources affecting a class I area. To illustrate, if there were only one emissions source affecting visibility in a class I area, that source could have a deciview impact only slightly below the perceptibility threshold without contributing to noticeable impairment. However, if there were 100 sources each changing visibility by 0.1 deciviews, the total impact would be a 10-deciview change in visibility. In this hypothetical example, all 100 sources would be contributing, in equal amounts, to substantial visibility impairment.
  
  Because circumstances will vary in different locations, we believe that States should have discretion to set an appropriate threshold depending on the facts of the situation. We believe, however, that it would be difficult for a State to justify a threshold higher than 0.5 deciviews. In particular, 0.5 deciviews represents one half of the 1.0 deciview level that we are equating with a single source ``causing'' visibility degradation. Typically, there are multiple sources that affect visibility in class I areas, so a source causing a 0.5 deciview change can be expected to be contributing to noticeable visibility impairment.
  
  In determining whether the maximum threshold of 0.5 deciviews or a lower threshold is appropriate for purposes of BART, we believe that States should consider the number of emissions sources affecting the class I area and the magnitude of the individual sources' impacts.\33\ In general, a larger number of sources causing impacts in a class I area may warrant a lower contribution threshold. In selecting a threshold, States may want to take into account the fact that individual sources have varying amounts of impact on visibility in class I areas. Depending on the facts regarding the number of sources affecting a class I area and their modeled impacts, the State could set a threshold that captures those sources responsible for most of the total visibility impacts, while still excluding other sources with very small impacts.\34\
  
  \33\ All states are working together in regional planning organizations, and we expect that states will have modeling information that identifies sources affecting visibility in individual class I areas, and the magnitude of their impacts.
  
  \34\ Under our guidelines, the contribution threshold should be used to determine whether an individual source is reasonably anticipated to contribute to visibility impairment. You should not aggregate the visibility effects of multiple sources and compare their collective effects against your contribution threshold because this would inappropriately create a ``contribution to contribution'' test.
  
  We also note that under this guidance, States would have discretion in setting the threshold for ``contributes to'' based on modeled impacts of sources. Consistent with American Corn Growers, we are not requiring States to find sources subject to BART regardless of their impact on Class I areas. We are suggesting that, in establishing a threshold for assessing contribution for BART, it may be logical to draw a line between ``contribution'' and ``non-contribution'' based on the number and magnitude of the various sources affecting the Class I areas at issue. Such an approach gives States the ability to assess the empirical evidence showing contribution and to design an appropriate regulatory regime in light of the nature of the problem. We note that for 750 MW power plants, such a line drawing exercise is likely to be unnecessary, as such sources will in most or all cases have impacts far exceeding 1.0 deciviews.
  
  Finally, we disagree that separate threshold levels should be established based on geography because a unit change in visibility expressed in deciviews, perceived or measured, is the same regardless of geography. As explained in the 1999 regional haze rule, the deciview can be used to express changes in visibility impairment in a way that corresponds to human perception in a linear manner. As a result, using the deciview as the metric for measuring visibility means, for example, that a one deciview change in a highly impaired environment would be perceived as roughly the same degree of change as one deciview in a relatively clear environment, and geography is not a factor. Interpretation of CALPUFF Results
  
  The standard CALPUFF modeling run provides day-by-day estimates of a source's visibility effects over a five-year period. In the proposed BART guideline, we indicated that if the maximum daily visibility value at any receptor over the five years modeled is greater than the ``cause or contribute'' threshold, then the State should conclude that the source is subject to BART. A number of commenters took issue with our proposal to use the 24-hour maximum modeled visibility impact over five years of meteorological data. Several of them pointed out, for example, that the maximum modeled 24-hour impact may be an outlier unduly influenced by weather. We agree that the maximum modeled effect in a five-year period could be the result of unusual meteorology. We also recognize that, although CALPUFF is the best currently available tool for analyzing the visibility effects of individual sources, it is a model that includes certain assumptions and uncertainties. Thus, we agree with commenters that a State should not necessarily rely on the maximum modeled impact in determining whether a source may reasonably be anticipated to contribute to visibility impairment in a Class I area.
  
  The final guideline states that it would be reasonable for States to compare the 98th percentile of CALPUFF modeling results against the ``contribution'' threshold established by the State for purposes of determining BART applicability. Some stakeholders have argued for the 90th percentile value, or even lower, contending that EPA should not use extreme cases to make BART applicability decisions. EPA agrees that, in most cases, important public policy decisions should not be based on the extreme tails of a distribution. We have concluded, however, that the 98th percentile is appropriate in this case.
  
  The use of 90th percentile value would effectively allow visibility effects that are predicted to occur at the level of the threshold (or higher) on 36 or 37 days a year. We do not believe that such an approach would be consistent with the language of the statute. Second, we note that the 98th percentile value would only be used to determine whether a particular BART-eligible source would be subject to further review by the State. In determining what, if any, emission controls should be required, the State will have the opportunity to consider the frequency, duration, and intensity of a source's predicted effect on visibility.
  
  On the other hand, there are other features of our recommended modeling approach that are likely to overstate the actual visibility effects of an individual source. Most important, the simplified chemistry in the model tends to magnify the actual visibility effects of that source. Because of these features and the uncertainties associated with the model, we believe it is appropriate to use the 98th percentile--a more robust approach that does not give undue weight to the extreme tail of the distribution. The use of the 98th percentile of modeled visibility values would appear to exclude roughly 7 days per year from consideration. In our judgment, this approach will effectively capture the sources that contribute to visibility impairment in a Class I area, while minimizing the likelihood that the highest modeled visibility impacts might be caused by unusual meteorology or conservative assumptions in the model.
  
  [[Page 39122]]
  
  Use of CALPUFF for Visibility Modeling
  
  Background. In providing the States with the option of making a determination as to which sources are subject to BART based on a consideration of each source's individual contribution to visibility impairment, we proposed that States use an air quality model such as CALPUFF. We also proposed that States use a CALPUFF or other EPA approved model in the BART analysis itself. The CALPUFF system, as explained in the 2004 reproposed guideline, consists of a diagnostic meteorological model, a gaussian puff dispersion model with algorithms for chemical transformation and complex terrain, and a post processor for calculating concentration fields and visibility impacts.
  
  The regional haze rule addresses visibility impairment caused by emissions of fine particles and their precursors. As fine particle precursors, such as SO2or NOX, are dispersed, they react in the atmosphere with other pollutants to form visibility- impairing pollutants. In fact, Congress implicitly recognized in 1977 the role of chemical transformation in creating visibility impairment, when it stated that the ``visibility problem is caused primarily by emissions of SO2, [NOX], and particulate matter.'' \35\ In most cases, to predict the impacts of a source's specific contribution to visibility impairment, a State will need a tool that takes into account not only the transport and diffusion of directly emitted PM2.5but also one that can address chemical transformation.
  
  \35\ H.R. Rep. No. 95-294 at 204 (1077).
  
  Because the air quality model CALPUFF is currently the best application available to predict the impacts of a single source on visibility in a Class I area, we proposed that a CALPUFF assessment be used as the preferred approach first, for determining whether an individual source is subject to BART, and second, in the BART determination process. The CALPUFF assessment is specific to each source, taking into account the individual source's emission characteristics, location, and the particular meteorological, topographical, and climatological conditions of the area in which the source is located, any of which may have an impact on the transport of PM2.5and its precursors. CALPUFF can be used to estimate not only the effects of directly emitted PM2.5emissions from a source, but also to predict the visibility impacts from the transport and chemical transformation of fine particle precursors.
  
  The CALPUFF model is generally intended for use on scales from 50 km to several hundred kilometers from a source. As a general matter, States will typically need to assess the impacts of potential BART sources on Class I areas located more than 50 km from the source.\36\ However, in situations where the State is assessing visibility impacts for source-receptor distances less than 50 km, we proposed that States use their discretion in determining visibility impacts, giving consideration to both CALPUFF and other EPA-approved methods. As an example, we suggested that States could use an appropriate local-scale plume impact model, such as PLUVUEII,\37\ to determine whether a source's emissions are below a level that would be reasonably anticipated to cause or contribute to visibility impairment in any Class I area.
  
  \36\ To determine whether a BART-eligible source ``may reasonably be anticipated to cause or contribute to any visibility impairment in any Class I area,'' it may not always be sufficient for the State to predict the impacts of a BART-eligible source only on the nearest Class I area (or on the nearest receptor in the nearest Class I area). The particular meteorological and topographical conditions, for example, could mean that a source's greatest impacts occurred at a Class I area other than the nearest one.
  
  \37\ PLUVUEII is a model used for estimating visual range reduction and atmospheric discoloration caused by plumes resulting from the emissions of particles, nitrogen oxides, and sulfur oxides from a single source. The model predicts the transport, dispersion, chemical reactions, optical effects and surface deposition of point or area source emissions. It is available at http://www.epa.gov/scram001/tt22.htm#pluvue .
  
  Comments. A number of States, environmental groups, and some industry commenters strongly supported the use of CALPUFF as proposed. Many commenters supported the use of CALPUFF but indicated that States must have the flexibility to use additional tools for their individual source analyses. Some suggested options for the ``cause or contribute'' determination were the use of photochemical grid models, or more simplified, non-modeling approaches. Commenters claimed that States must have the option to incorporate advances in science and technologies into models or other applications that may produce more accurate simulations of meteorology, chemistry, and visibility impairment. Other industry groups and States argued that CALPUFF has significant limitations, especially simulating complex atmospheric chemistry, and that EPA's recommendation of CALPUFF as the preferred approach is therefore inappropriate.
  
  Another issue raised by commenters was the use of CALPUFF for estimating secondary particulate matter formation. Commenters recognized that CALPUFF was incorporated into the ``Guideline on Air Quality Models'' at 40 CFR part 51, appendix W in April 2003 as the preferred model for Prevention of Significant Deterioration (PSD) increment and National Ambient Air Quality Standards (NAAQS) compliance assessments of long range transport of primary emissions of SO2and PM2.5. However, commenters stated that CALPUFF has not been incorporated into the Guideline on Air Quality Models for predicting the secondary formation of PM. The commenters remarked that EPA guidance indicates that photochemical grid models be used to simulate secondary PM formation and concluded on this basis that the application of CALPUFF as we proposed is in conflict with our guidance.
  
  Final rule. We believe that CALPUFF is an appropriate application for States to use for the particular purposes of this rule, to determine if an individual source is reasonably anticipated to cause or contribute to impairment of visibility in Class I areas, and to predict the degree of visibility improvement which could reasonably be anticipated to result from the use of retrofit technology at an individual source. We encourage States to use it for these purposes.\38\
  
  \38\ The model code and its documentation are available at no cost for download from http://www.epa.gov/scram001/tt22.htm#calpuff.
  
  CALPUFF is the best modeling application available for predicting a single source's contribution to visibility impairment. It is the only EPA-approved model for use in estimating single source pollutant concentrations resulting from the long range transport of primary pollutants. In addition, it can also be used for some purposes, such as the visibility assessments addressed in today's rule, to account for the chemical transformation of SO2and NOX. As explained above, simulating the effect of precursor pollutant emissions on PM2.5concentrations requires air quality modeling that not only addresses transport and diffusion, but also chemical transformations. CALPUFF incorporates algorithms for predicting both. At a minimum, CALPUFF can be used to estimate the relative impacts of BART-eligible sources. We are confident that CALPUFF distinguishes, comparatively, the relative contributions from sources such that the differences in source configurations, sizes, emission rates, and visibility impacts are well-reflected in the model results. States can make judgements
  
  [[Page 39123]]
  
  concerning the conservativeness or overestimation, if any, of the results. In fact, although we focused on the use of CALPUFF for primary pollutants in revising the Guideline of Air Quality Modeling, section
  
  7.2.1.e. of the Guideline states:
  1. CALPUFF (Section A.3) may be applied when assessment is needed of reasonably attributable haze impairment or atmospheric deposition due to one or a small group of sources. This situation may involve more sources and larger modeling domains than that to which VISCREEN ideally may be applied. The procedures and analyses should be determined in consultation with the appropriate reviewing authority (paragraph 3.0(b) and the affected FLM(s).
  We believe that our proposed use of CALPUFF is thus fully in keeping with the Guideline on Air Quality Models, especially in light of the low triggering threshold for determining whether a source is reasonably anticipated to cause or contribute to visibility impairment in a Class I area, and the fact that the modeling results are used as only one of five statutory criteria evaluated to determine BART emission limits.
  
  Even so, as commenters point out, CALPUFF has not yet been fully evaluated for secondary pollutant formation. For the specific purposes of the regional haze rule's BART provisions, however, we have concluded that CALPUFF is sufficiently reliable to inform the decision making process.
  
  EPA revised the Guideline on Air Quality Models in 2003, in part, to add CALPUFF to the list of approved models for particular uses. At that time, we considered comments that CALPUFF should be approved for use in predicting the impact of secondary emissions on particulate matter concentrations. As we stated in the revision, CALPUFF represents a substantial improvement in methods for assessing long-range transport of air pollutants. However, as explained in the response to comments for that rulemaking, the modeling results in the context of a PSD review may be used as the sole determining factor in denying a source a permit to construct.\39\ Although its use in simulating long-range transport is beneficial, given the significance of the modeling results in assessing increment consumption due to a single source's impacts, we made a determination that it would not be appropriate in the rulemaking revising Appendix W to approve CALPUFF for use in modeling secondary emissions.
  
  \39\ Under CAA section 165(a), a major emitting facility may not be constructed unless the owner or operator of the facility demonstrates that the emissions from the facility will not cause or contribute air pollution in excess of an increment or NAAQS.
  
  In contrast to the significance of the modeling results in the PSD context, the use of CALPUFF in the context of the regional haze rule is not determinative of a source's ability to construct or operate. A State may use CALPUFF to determine whether a source can reasonably be anticipated to cause or contribute to visibility impairment and so should be subject to additional review to determine if the source should be subject to control.
  
  Based on our analysis of the power plants covered by the guidelines, we believe that all but a handful of these plants have impacts of greater than 1.0 deciview on one or more Class I areas.\40\ In fact, we anticipate that most of these plants are predicted to have much higher maximum impacts.\41\ Because of the scale of the predicted impacts from these sources, CALPUFF is an appropriate or a reasonable application to determine whether such a facility can reasonably be anticipated to cause or contribute to any impairment of visibility. In other words, to find that a source with a predicted maximum impact greater than 2 or 3 deciviews meets the contribution threshold adopted by the States does not require the degree of certainty in the results of the model that might be required for other regulatory purposes.
  
  \40\ CALPUFF Analysis in Support of the Regional Haze Rule, U.S. Environmental Protection Agency, April 15, 2005, Docket No. OAR- 2002-0076.
  
  \41\ Ibid.
  
  In the unlikely case that a State were to find that a 750 MW power plant's predicted contribution to visibility impairment is within a very narrow range between exemption from or being subject to BART, the State can work with EPA and the FLM to evaluate the CALPUFF results in combination with information derived from other appropriate techniques for estimating visibility impacts to inform the BART applicability determination. Similarly for other types of BART eligible sources, States can work with the EPA and FLM to determine appropriate methods for assessing a single source's impacts on visibility.
  
  As discussed in section E. below we also recommend that the States use CALPUFF as a screening application in estimating the degree of visibility improvement that may reasonably be expected from controlling a single source in order to inform the BART determination. As we noted in 2004, this estimate of visibility improvement does not by itself dictate the level of control a State would impose on a source; ``the degree of improvement in visibility which may reasonably be anticipated to result from the use of [BART]'' is only one of five criteria that the State must consider together in making a BART determination. The State makes a BART determination based on the estimates available for each criterion, and as the CAA does not specify how the State should take these factors into account, the States are free to determine the weight and significance to be assigned to each factor. CALPUFF accordingly is an appropriate application for use in combination with an analysis of the other statutory factors, to inform decisions related to BART.
  
  We understand the concerns of commenters that the chemistry modules of the CALPUFF model are less advanced than some of the more recent atmospheric chemistry simulations. To date, no other modeling applications with updated chemistry have been approved by EPA to estimate single source pollutant concentrations from long range transport. In its next review of the Guideline on Air Quality Models, EPA will evaluate these and other newer approaches and determine whether they are sufficiently documented, technically valid, and reliable to approve for general use. In the meantime, as the Guideline makes clear, States are free to make their own judgements about which of these or other alternative approaches are valid and appropriate for their intended applications.
  
  Theoretically, the CALPUFF chemistry simulations, in total, may lead to model predictions that are generally overestimated at distances downwind of 200 km. Again, States can make judgements concerning the conservativeness or overestimation, if any, of the results.
  
  The use of other models and techniques to estimate if a source causes or contributes to visibility impairment may be considered by the State, and the BART guidelines preserve a State's ability to use other models. Regional scale photochemical grid models may have merit, but such models have been designed to assess cumulative impacts, not impacts from individual sources. Such models are very resource intensive and time consuming relative to CALPUFF, but States may consider their use for SIP development in the future as they are adapted and demonstrated to be appropriate for single source applications. However, to date, regional models have not been evaluated for single source applications. Their use may be more appropriate in the cumulative modeling options discussed
  
  [[Page 39124]]
  
  above.\42\ In evaluating visibility improvement as one of the five factors to consider in setting BART controls, other models, used in combination with CALPUFF may be helpful in providing a relative sense of the source's visibility impact and can aid in informing the BART decision. A discussion of the use of alternative models is given in the Guideline on Air Quality in appendix W, section 3.2.
  
  \42\ For regional haze applications, regional scale modeling typically involves use of a photochemical grid model that is capable of simulating aerosol chemistry, transport, and deposition of airborne pollutants, including particulate matter and ozone. Regional scale air quality models are generally applied for geographic scales ranging from a multi-state to the continental scale. Because of the design and intended applications of grid models, they may not be appropriate for BART assessments, so States should consult with the appropriate EPA Regional Office prior to carrying out any such modeling.
  
  The Use of Natural Conditions in Determining Visibility Impacts for Reasonable Progress and Comparison to Threshold Values
  
  Background. As set out in section 169A(a) of the CAA and stated in the 1999 regional haze rule, a return to natural visibility conditions, or the visibility conditions that would be experienced in the absence of human-caused impairment, is the ultimate goal of the regional haze program. To measure progress toward this goal, the regional haze rule requires that a comparison with natural conditions for the 20 percent best and worst days to calculate ``reasonable progress'' determinations. Default values for natural visibility conditions are provided in EPA guidance.\43\ In the 2004 reproposal of the BART guidelines, we proposed that changes in visibility, expressed in deciviews, should be determined by comparing the impact from a single source to natural visibility conditions. That impact should then be compared to a threshold impact, also expressed in deciviews, to assess if a BART-eligible source should be subject to a BART review.
  
  \43\ Guidance for Estimating Natural Visibility Conditions Under the Regional Haze Rule, U.S. Environmental Protection Agency, September 2003. http://www.epa.gov/ttncaaa1/t1/memoranda/rh_envcurhr_gd.pdf. Natural background conditions, expressed in
  
  deciviews, are defined for each Class I area. EPA has issued guidance for estimating natural background conditions which has estimates of default conditions as well as measures to develop refined estimates of natural conditions.
  
  Comments. Opposing commenters said that a return to natural conditions is unattainable as it would require the elimination of every manmade source, and that changes should be compared against currently existing conditions. They added that true ``natural conditions'' cannot be verified, do not account for manmade emissions from other countries, and are not a realistic target for improvement. Further, they argued that natural conditions are a ``goal'' representing a benchmark that is relevant to the States' determination, under the regional haze program, of the level of ``reasonable progress'' to achieve; however they stated that there is no legal requirement (and there could not be a legal requirement) that the natural conditions goal ultimately must be achieved. Several commenters added that current visibility conditions make more sense as a baseline because sources that are subject to BART today will likely not be in operation in the 2064 time frame. A commenter added that using current visibility conditions for the analysis will give a more realistic, real-world prediction of whether controlling the source pursuant to BART will actually improve visibility. The commenter said that Congress did not intend for sources to have to consider retrofitting controls under the BART provision if those sources currently are not impacting real-world visibility. Other utility groups stated that in addition to international emissions, the estimated natural visibility conditions failed to account for natural phenomena such as sea salt, wildfires, and natural organics. One commenter noted that natural visibility estimates will be revised and refined over time and it would be unwise to compare impacts and improvements to a moving baseline.
  
  On the other hand, numerous commenters supported the use of natural visibility conditions as a baseline for measuring visibility improvements. Several environmental groups said that any increase in the baseline beyond natural visibility conditions will unlawfully distort and weaken the BART requirement by effectively raising the applicability threshold in less protected, highly polluted areas, which would be illogical. Further, they pointed out that these BART-eligible sources clearly are contributing to the very manmade visibility impairment that the Act is explicitly designed to remedy by a return to natural conditions. They added that measuring natural conditions as opposed to some other baseline condition is a more appropriate approach, given that the planning goal is to achieve natural visibility by the end of the program. They also added that a baseline other than natural conditions would never assure ``reasonable progress''.
  
  Finally, two commenters asked for clarification on the values for natural conditions to be used for estimating changes in visibility. The commenters appeared to assume that we intended for the comparison to be done for natural visibility conditions on the 20 percent best days.
  
  Final Rule. We disagree with commenters saying that the use of natural conditions as the baseline for making visibility impact determinations is inappropriate. The visibility goal of the CAA is both the remedying of existing impairment, and prevention of future impairment. The court, in American Corn Growers, upheld our interpretation of that goal as the return to natural visibility conditions.\44\ Long-term regional haze strategies are developed to make ``reasonable progress'' towards the CAA goal, and States must demonstrate reasonable progress in their regional haze State implementation plans (SIPs). Since the BART program is one component of that demonstration, visibility changes due to BART are appropriately measured against the target of natural conditions.
  
  \44\ See also our explanation of the CAA goal provided in the regional haze rule at 64 FR at 35720-35722. We note that the court in American Corn Growers also observed, ``the natural visibility goal is not a mandate, it is a goal.'' 291 F.3d at 27.
  
  In establishing the goal of natural conditions, Congress made BART applicable to sources which ``may be reasonably anticipated to cause or contribute to any impairment of visibility at any Class I area''. Using existing conditions as the baseline for single source visibility impact determinations would create the following paradox: the dirtier the existing air, the less likely it would be that any control is required. This is true because of the nonlinear nature of visibility impairment. In other words, as a Class I area becomes more polluted, any individual source's contribution to changes in impairment becomes geometrically less. Therefore the more polluted the Class I area would become, the less control would seem to be needed from an individual source. We agree that this kind of calculation would essentially raise the ``cause or contribute'' applicability threshold to a level that would never allow enough emission control to significantly improve visibility. Such a reading would render the visibility provisions meaningless, as EPA and the States would be prevented from assuring ``reasonable progress'' and fulfilling the statutorily-defined goals of the visibility program. Conversely, measuring improvement against clean conditions would ensure reasonable progress toward those clean conditions.
  
  [[Page 39125]]
  
  With regard to BART-eligible sources not being in operation for the duration of the program, a State, in making BART determinations, is explicitly directed by the CAA to account for the remaining useful life of a source. Thus, States may factor into their reasonable progress estimates those shut-downs that are required and effected in permit or SIP provisions. In addition, as provided for under our guidance,\45\ proper accounting for international emissions and natural phenomena is in the 5 year SIP progress report, not in the setting of natural visibility estimates. Finally, these final BART guidelines use the natural visibility baseline for the 20 percent best visibility days for comparison to the ``cause or contribute'' applicability thresholds. We believe this estimated baseline is likely to be reasonably conservative and consistent with the goal of natural conditions.
  
  \45\ Guidance for Estimating Natural Visibility Conditions Under the Regional Haze Rule, U.S. Environmental Protection Agency, September 2003. http://www.epa.gov/ttncaaa1/t1/memoranda/rh_envcurhr_gd.pdf .
  
  Modeling Protocol
  
  Background. The 2004 guidelines proposed that a written modeling protocol be submitted for assessing visibility impacts from sources at distances greater than 200 km from a Class I area. The proposal indicated that the protocol should include a description of the methods and procedures to follow, for approval by the appropriate reviewing authority; critical items to include in the protocol are meteorological and terrain data, source-specific information (stack height, temperature, exit velocity, elevation, and allowable emission rate of applicable pollutants), and receptor data from appropriate Class I areas.
  
  Comments. All of the comments supported the development of a written modeling protocol. Industry, Federal, and State commenters said a modeling protocol should be required of all States and stakeholders who are performing the BART modeling analysis. Commenters said the protocol should allow all interested parties an opportunity to understand the modeling approach and how the results will be used, and that the State should provide opportunity for comments on the procedures prior to the publication of the final results.
  
  Many utility groups commented that the protocol should provide States with flexibility and that the choice of models should be at the States' (or RPOs') discretion. Some commenters stressed that it is important that states and sources retain the flexibility to decide how to set up and run the selected model, while others asked for specific guidance on the setup of CALPUFF or other approved models, including on specific parameters (e.g. how to adjust for cases where sources are greater than 200 km from a Class I area).
  
  Regarding the approval of a modeling protocol, some commenters said that the protocol should be approved by EPA. Others stated, however, that we should have only an advisory role in development of the protocol. They said that States are in a better position to determine which modeling input values best reflect conditions in their States.
  
  Several commenters representing environmental groups said we should develop a CALPUFF protocol that must be followed and should include, among other items, meteorological data (i.e., where available 5 years of data should be used), emissions reported for the same meteorological years, documented source parameters, model physical parameters, and assumed background concentrations for ozone and ammonia (based on nearby reliable observations and/or regional modeling results). They added that a protocol developed by EPA would help to produce consistent BART determinations across various sources and geographic areas for both shorter and longer distances. FLMs stated that this is also an appropriate time to create regional modeling platforms for CALPUFF, which would allow States and sources to run the model more expeditiously and more consistently. They recommended that we consider a multi-agency process to reach agreement on an appropriate modeling protocol prior to allowing BART applicability and control determinations to be based on model results. FLMs added that it would be helpful to establish a national procedure for this process, including a methodology for establishing natural background conditions, background ammonia concentrations, and determining sulfuric acid emission rates. Such a process, they said, could reasonably be engaged in prior to deadlines for state implementation plans, and would not delay implementation of the BART guidelines. The FLMs noted that consistent, nationally applicable guidance is essential, and that once it is developed, virtually no deviations should be allowed. Finally, they added that the CALPUFF modeling exercises should follow the Interagency Workgroup on Air Quality Modeling (IWAQM) Phase 2 Summary Report and Recommendations for Modeling Long Range Transport Impacts,\46\ but that we, in consultation with the FLMs and States, should also publish additional guidance to address more recent issues such as particle speciation, emission rate averaging times, and ``natural obscuration.'' Another State commenter said that The Guideline on Air Quality Models (CFR Part 51, Appendix W) should be included along with the IWAQM Report as a reference for CALPUFF setup. One RPO commented that we should provide data, perhaps using example facilities, to demonstrate the effect of the process so that States can get a better feeling for which sources are likely to fall below the 0.5 deciview threshold. This would help States understand the net effect of all of the parameters chosen in the exemption process.
  
  \46\ Interagency Workgroup on Air Quality Modeling (IWAQM) Phase 2 Summary Report and Recommendations for Modeling Long Range Transport Impacts, U.S. Environmental Protection Agency, EPA-454/R- 98-019, December 1998.
  
  Commenters also said that we should continuously revise modeling protocols by providing a modeling clearinghouse to States, and further, that we should consider new models for use, such as the Community Multiscale Air Quality (CMAQ) model.
  
  There were specific comments requesting guidance for calculating visibility impacts and other general modeling concerns. One technical comment was that the guidelines should specify that the IMPROVE monitor is the receptor by which modeled visibility impacts should be evaluated with the CALPUFF model. Another commenter suggested using recent scientific evidence to update the light extinction coefficients used by CALPUFF to calculate visibility changes. These commenters also stated that CALPUFF might be improved by capping the relative humidity to lower values than are currently used.
  
  Additional commenters representing utility organizations discussed how to identify Class I areas that should be modeled. They said that the guidelines should require sources to model only the nearest Class I area (or possibly the two closest), and one commenter said that we should provide a reasonable methodology to minimize the effort needed to address impacts from BART-eligible sources on multiple Class I areas.
  
  Final Rule. We agree that States should adopt modeling protocols for all modeling demonstrations, regardless of the distance from the BART-eligible source and the Class I area impacted. We are therefore dropping the 200 km and greater distance requirement from the guidelines. As noted in the 2004 re-
  
  [[Page 39126]]
  
  proposal, we believe that potential uncertainties in model performance may be greater at distances greater than 200 km for a source. A modeling protocol may reduce the need for additional analyses. We favor coordination among States, EPA regions, RPOs, and other federal agencies to agree on a modeling protocol(s) which would provide consistent application.
  
  In developing a modeling protocol, we also encourage States to use the framework provided for model setup in EPA's IWAQM. CALPUFF model users may find default settings in that document which may be appropriate for their modeling situations and add an element of consistency to model applications. The Guideline on Air Quality Models (CFR Part 51, Appendix W) also provides useful guidance.
  
  We do, however, understand and agree that States have flexibility developing a modeling protocol. Moreover, the diversity of the nation's topography and climate, and variations in source configurations and operating characteristics, dictate against a strict modeling ``cookbook''. A State may need to address site-specific circumstances at individual sources potentially affecting a specific Class I area. For example, in a particular area a State may have available emissions data, that is more representative of the modeling domain, which may supplement the model defaults. States may want to consult with the appropriate EPA regional office and Federal Land Managers in adjusting the model input parameters. The modeling input recommendations in the IWAQM report are designed for visibility impact applications, and those defaults allow for tailoring for a given application (e.g. puff splitting). The model developers Web site \47\ also has a series of frequently asked questions with answers to assist users in tailoring model applications.
  
  \47\ http://www.src.com/calpuff/calpuff1.htm.
  
  We agree that we have only an advisory role in development of the protocol as the States better understand the BART-eligible source configurations and the geophysical and meteorological data affecting their particular Class I area(s).
  
  In the protocol development process, we support the idea of designing example runs, as we have done in our example analysis for EGUs,\48\ so that States may get a better understanding of what visibility impacts might be expected from a particular type of source or sources. Once a protocol has been finalized, a State may be able to use example runs as a proxy in making BART determinations which could potentially eliminate the need for case-by-case review for every BART- eligible source. A common sense approach should be taken, particularly where an analysis may add a significant resource burden to a State. For example, if there are multiple Class I areas in relatively close proximity to a BART-eligible source, a State may model a full field of receptors at the closest Class I area. Then a few strategic receptors may be added at the other Class I areas (perhaps at the closest point to the source, a receptor at the highest and lowest elevation in the Class I area, a receptor at the IMPROVE monitor, and a few receptors that are expected to be at the approximate plume release height). If the highest modeled impacts are observed at the nearest Class I area, a State may choose not to analyze the other Class I areas any further and additional analyses might be unwarranted.
  
  \48\ CALPUFF Analysis in Support of the June 2005 Changes to the Regional Haze Rule,U.S. Environmental Protection Agency, June 15, 2005, Docket No. OAR-2002-0076.
  
  As models are revised and advances in science are incorporated into the models, we can make certain that revisions to protocols are made accordingly. We will work closely with States and FLMs, as should States; we expect that States will also work closely with FLMs throughout the protocol development process. We expect a similar protocol development process for other models that may be used, once those models are developed to predict and track single source impacts and demonstrate acceptable model performance. States should contact the appropriate FLM and EPA regional office for the latest guidance and modeling updates. Alternatives for Determining Visibility Impacts From Individual Sources
  
  Background. In the 2004 reproposal, we requested comment on the following alternatives to CALPUFF modeling for determining whether individual sources cause or contribute to visibility impairment: look- up tables developed from screening-level air quality modeling; running CALPUFF in a simpler screening mode than the preferred approach; a source ranking methodology; and an emissions divided by distance (Q/D) method. Except for the simplified CALPUFF approach, all alternatives were based on developing a relationship between source emissions and the source's distance to a Class I area. Each of these approaches was intended to reduce the resource burden on States.
  
  Comments. Some commenters supported the use of alternative approaches, while others suggested that the alternatives could be used either in conjunction, or in hierarchical fashion, with modeling approaches. Many commenters were opposed to their use. The opposing comments were consistent in stating that the alternatives were inappropriate because they did not account for important factors such as terrain, local meteorological data, prevailing wind directions (which influence pollutant transport), and differences in stack release parameters. Commenters added that there is no direct connection between emissions, distance, and visibility impairment, and that the methods treat SO2and NOXequally for impairment estimates.
  
  Final Rule. We disagree that the alternatives are necessarily inappropriate, but we share most of the concerns articulated by the opposing commenters. We believe that alternatives should not be used to exempt a source from BART review without more rigorous evaluations and sensitivity tests showing that the results are at least as conservative as the CALPUFF model. We know of at least one study showing that, for one location and for one year, there is no guarantee that the simplified CALPUFF technique is as conservative as the preferred approach \49\. While we are not adopting in the guideline any specific alternative to modeling for power plants greater than 750MW, a State may develop its own alternative approach for the other source categories to determine if a source would be subject to BART, provided that the alternative demonstrates a sufficient basis to determine clearly that the source causes or contributes to visibility impairment, or that more refined analysis is warranted. Use of an alternative approach could be a conservative non-modeling method for easing a State's resource burden. We believe conservatism is needed because of the purpose of the test: i.e. solely to determine if a closer look at the source is warranted.
  
  \49\ Analysis of the CALMET/CALPUFF Modeling System in a Screening Mode, U.S. Environmental Protection Agency, November 1998, Docket No. OAR-2002-0076.
5. The BART Determination Process
  
  Background. CAA section 169A(g)(7) directs States to consider five factors in making BART determinations. The regional haze rule codified these factors in 40 CFR 51.308(e)(1)(ii)(B), which directs States to identify the ``best system of continuous emissions control
  
  [[Page 39127]]
  
  technology'' taking into account ``the technology available, the costs of compliance, the energy and nonair quality environmental impacts of compliance, any pollution control equipment in use at the source, and the remaining useful life of the source.'' Section IV. of the BART guidelines provides a step-by-step guide to conducting a BART determination which takes these factors into account.
  
  This section of the preamble addresses a number of issues relative to the process for conducting a BART determination contained in Section IV of the BART guidelines. 1. What Is Meant by ``Technical Feasibility of the Control Options'' in Step 2 of the BART Determination?
  
  Comments. We received several comments on this discussion, both on the 2001 proposal and on the 2004 reproposal. One commenter recommended that the concept of available technology for regional haze should be expanded to include those in the pilot scale testing phase, because these guidelines will precede the installation of controls by about 10 years. Other commenters believed that the discussion of technical feasibility introduced terms and concepts that were not clear, for example, what is meant by ``commercial demonstration.'' One commenter raised issues with deeming technologies used in foreign countries ``available'' unless their performance has been demonstrated in the United States. A few commenters expressed concern with the provision in the guidelines that new technologies should be considered up to the time of a State's public comment period on the BART determination. The commenter believed that this could create an endless review loop for States if new technologies continually became available.
  
  Final rule. In the final guidelines, we have largely retained the language that was in the proposed guidelines. Because the guidelines call for consideration of technologies that become available by the time of the State's public comment process on the BART determination, technologies should be considered that become available well after we finalize the BART guidelines. We also note, for clarity, that the Guidelines state that technologies need to be both licensed and commercially available (i.e. commercially demonstrated and sold). 2. How Should the Costs of Control Be Estimated in Step 4 of the BART Determination?
  
  Comments. This section of the guidelines remained unchanged between the 2001 proposal and the 2004 reproposal. Comments varied, ranging from questioning the reliance on EPA's OAQPS Control Cost Manual Fifth Edition, February 1996, EPA 453/B-96-001 (hereafter called the ``Control Cost Manual'') to requesting that we not include the concept of incremental cost effectiveness in the guidelines. A commenter expressed concerns that incremental cost effectiveness calculations, the cost of implementing each succeeding control option, is too dependent on the number of interim options included in the analysis. Moreover, the commenter believed that incremental cost calculations increase the complexity of the analysis, and they also increase the possibility for inconsistent cost results.
  
  Final rule. We have finalized this section of the guidelines with some changes to how it was proposed. States have flexibility in how they caculate costs. We believe that the Control Cost Manual provides a good reference tool for cost calculations, but if there are elements or sources that are not addressed by the Control Cost Manual or there are additional cost methods that could be used, we believe that these could serve as useful supplemental information.
  
  In addition, the guidelines continue to include both average and incremental costs. We continue to believe that both average and incremental costs provide information useful for making control determinations. However, we believe that these techniques should not be misused. For example, a source may be faced with a choice between two available control devices, control A and control B, where control B achieves slightly greater emission reductions. The average cost (total annual cost/total annual emission reductions) for each may be deemed to be reasonable. However, the incremental cost (total annual costA-B/total annual emission reductionsA-B) of the additional emission reductions to be achieved by control B may be very great. In such an instance, it may be inappropriate to choose control B, based on its high incremental costs, even though its average cost may be considered reasonable.
  
  Finally, it is important to note that, while BART determinations are focused at individual sources, it is likely that in response to SIP requirements, States will be making BART determinations for many units in a subject source category all at the same time. In doing so, States are likely to compare costs across each source category as well as looking at costs for individual units in order to respond to SIP requirements in an efficient manner (from the State's perspective). 3. How Should ``Remaining Useful Life'' Be Considered in Step 4 of the BART Determination?
  
  Comments. We received a number of comments on the issue of remaining useful life, both on the 2001 proposal and on the 2004 reproposal. One commenter asserted that remaining useful life should not be considered in the cost analysis and that if a source is in operation at the time of a State's SIP submittal, it must have plans to install controls. Other commenters believed that, to the extent that assertions regarding a plant's remaining useful life influences the BART decision, there must be an enforceable requirement for the plant to shut down by that date. Other comments questioned whether Congress intended enforceable restrictions in order to take into account the remaining useful life and whether EPA had the authority under the CAA to require plant shutdowns.
  
  A number of comments were received regarding our request for comments on how to provide flexibility for situations where market conditions change. Some comments interpreted this provision as a loophole that would allow sources to continue operation for a number of years without BART. Another comment supported the concept of allowing a source to later change its mind, so long as BART is installed.
  
  Final rule. We have retained the approach in the proposed guidelines, including the provision for flexibility for sources to continue operating, with BART in place, should conditions change. We believe that the CAA mandates consideration of the remaining useful life as a separate factor, and that it is appropriate to consider in the analysis the effects of remaining useful life on costs. We believe that, because the source would not be allowed to operate after the 5- year point without such controls, the option for providing flexibility would not create a loophole for sources. Moreover, any source operating after this point without BART controls in place would be subject to enforcement actions for violating the BART limit. For any source that does not agree to shut down before the 5-year point, the State should identify a specific BART emission limit that would apply after this point in time. 4. How Should ``Visibility Impacts'' Be Considered in Step 5 of the BART Determination?
  
  Background. The fifth statutory factor addresses the degree of improvement in
  
  [[Page 39128]]
  
  visibility which may reasonably be anticipated to result from the use of the ``best control technology'' for sources subject to BART. The 2004 reproposal focuses on the use of single source emissions modeling to evaluate the BART control options. As part of the BART determination, we proposed that a State or individual source would run CALPUFF, or another EPA-approved model, to estimate, in deciviews, a BART source's visibility impact at a Class I area. The source would run the model once using its allowable emission rates, and then again at the various post-control emissions rates being evaluated for the BART determination. The 24-hour model results would then be tabulated for the pre- and post-control scenarios, for the average of the 20 percent worst modeled days at each receptor, over the time period of meteorology modeled. The difference in the averages for each receptor is the expected degree of improvement in visibility. Alternatively, the proposal requested comment on the option of using the hourly modeled impacts from CALPUFF at each receptor and determining the improvement in visibility based on the number of hours above the 0.5 deciview threshold for both the pre- and post-control model runs. We also requested comment on combinations of the proposed and alternative options and on the use of the simpler screening version of CALPUFF to do the analysis.
  
  Comments. Several environmental groups said that issues relating to the determination of visibility improvement for evaluating BART controls are in many ways the same as for determining which BART- eligible sources are subject to BART. Thus, the commenter pointed out, the issues concerning the BART applicability test, discussed in section D., are all equally applicable here, including comments on: using the 0.5 deciview threshold on an aggregate basis for determining visibility impairment and potential exemption for BART-eligible sources, use of a natural visibility baseline versus current visibility, using a substantially lower deciview threshold than 0.5 deciviews to determine the contribution to visibility impairment by an individual source, and demonstration of those thresholds by means of appropriate modeling rather than other less reliable and more subjective techniques.
  
  An industry commenter claimed that the American Corn Growers case emphasized the fact that the CAA clearly provides that BART determinations should balance the visibility benefits of controls comprehensively against their burdens; the commenter noted that this is not mentioned in our proposal; the commenter said that although the proposal would allow States to run the CALPUFF model, it fails to specify how they might consider the results.
  
  One State commenter opposed the use of visibility modeling for the purpose of informing the choice of control option, stating that it is unnecessary, confusing and without adequate standards or guidance for implementation. The State added that the analysis of control options in the BART process should yield the greatest, most cost-effective control efficiency for NOXand SO2at or above our presumptive levels of control. Moreover, it said that analysis of the degree of visibility improvement may result in very small increments of visibility improvements within Class I areas from an individual source, thus tilting the selection to the lower control efficiency option. The State added that we should remove this criterion from the analysis to ensure that the best cost effective controls will result. Another State agency said that modeling impacts should not be considered in BART determinations because they are not considered when determining BACT for the PSD program.
  
  A variety of commenters pointed out several areas where the guidelines should be improved or clarified in regard to the degree of visibility improvement determination:
  
  We should clarify that the analysis is pollutant-specific (e.g., the modeling evaluation of a BART control option for SO2reduction should not be combined with the modeling evaluation of a BART control option for NOX.)
  
  We should clarify that only the closest Class I area must be modeled.
  
  We should describe CALPUFF as one possible model to use, rather than as the only model that may be used.
  
  States and sources should have the flexibility to perform multiple modeling runs based on different levels of available control.
  
  Predicted visibility improvements that are imperceptible should be given no weight in determining the level of control that constitutes BART.
  
  States should be allowed to establish a factor for the required degree of visibility improvement.
  
  Several industry and utility commenters expressed concern about using allowable emission rates to predict visibility impacts for BART control options; they argued that actual emission rates should be considered instead. Three commenters stated that we must make clear that States should use emission rates that will be permissible at the time BART controls take effect, not current emissions rates.
  
  Additional comments from utilities, industry, and one State opposed the approach wherein the results from the 20 percent worst modeled days (pre- and post-control) were used to evaluate the visibility improvements expected from the various control options. Some believed this was too stringent, while others said it was not stringent enough. Two utilities added that the criteria should use the 20 percent worst days based on monitored data, not modeled data. An environmental group stated that sources should not be limited to just the worst days, but the improvements should be based upon controls reducing visibility impairment on any day. The commenter added that this rationale ignores the middle 60 percent of days in which visibility may worsen, because sources may increase emissions on these days as a trade-off for cutting emissions on the worst days. The commenter further argued that there are no data to support our assertion that improvement on the worst days means improvement on other days. They noted that default ``natural condition'' deciview values for Class I areas in our natural conditions guidance exist only for the average of the 20 percent best and worst days. The commenter added that we used the average default natural conditions (for the 20 percent best days) for the visibility impairment analysis, but there are no default ``maximum 24-hour'' values in the guidance.
  
  Nine commenters supported implementation of visibility improvement thresholds, which were not proposed in 2004. A State commenter said it is unclear how the modeled net visibility improvement would be specifically utilized in the BART analysis, and requested a target level of improvement or a de minimis level by which to measure improvement. Two industry commenters suggested alternatives to the 24- hour value. One said that setting a threshold for comparison, as in the BART-applicability test, is more appropriate than the overall comparison of the 20 percent worst case days, and that the threshold for comparison should be on at least a daily average (or longer), not an hourly average, due to the possibility of short-term spikes based on certain meteorological conditions.
  
  These commenters also said that a comparison of the number of days above or below a certain threshold is preferable since below a certain threshold, the impacts of visibility are not perceptible; unlike concentration
  
  [[Page 39129]]
  
  levels of certain pollutants (i.e., ozone) which do not have a threshold below which there are no effects, there are concentration levels of particulate below which there is no visibility impact. They also asserted that comparing the number of days would allow for a more complete picture of how controls would potentially improve visibility. As noted previously, a small number of unusual meteorological conditions can produce significant spikes on a single day or days. Since the overall goal of the regional haze rule is long-term visibility improvement, they said that a comparison of the total number of days exceeding a threshold over multiple years will provide a better overall indicator of visibility improvement. One commenter suggested that if we retain the maximum 24-hour value for the visibility impairment analysis, we should at least allow the use of only 1 year, rather than 5 years, of meteorological data. That would simplify the modeling and would lessen the chance that one day with atypical, extreme conditions would dictate the result.
  
  One FLM supported our proposed method to determine visibility improvement associated with installation of BART. However, with regard to the use of hourly data instead of 24 hour data for the degree of visibility improvement assessment, another FLM said that while hourly model data are, by their nature, less reliable in predicting actual conditions, a measure that reports the total number of hours above a given threshold would still be a useful measure of the long-term effect of BART control. They said we should require States to report a combination of measures of the visibility improvement expected from BART. Such measures would be the change in the 20 percent worst days as well as a metric that examines the amount of time during a year that the source's visibility impact would exceed a threshold with and without BART.
  
  Another utility commenter added that, if a BART control option would result in no perceptible improvement in visibility at a Class I area, then it is not a cost-effective option. This commenter said that based on Pitchford and Malm (1994) \50\ and Henry (2002) \51\ a 2 deciview threshold of perception would be appropriate, with a 1 deciview threshold providing a margin of safety. Another commenter said that we should clarify that visibility improvement differences among BART control options should be considered insignificant if the differences are less than the perceptibility threshold level, which should be set in excess of 1 deciview. Other commenters said the minimum threshold should be 1 deciview.
  
  \50\ Pitchford, M. and Malm, W., ``Development and Applications of a Standard Visual Index,'' Atmospheric Environment, V. 28, no. 5, March 1994.
  
  \51\ Henry, R.C. ``Just-Noticeable Differences in Atmospheric Haze'', Journal of the Air & Waste Management Association, 52:1238- 1243, October 2002.
  
  Final Rule. We disagree with the comment that modeling should not be part of a BART review because it is not considered for BACT. CAA section 169A(g)(2) clearly requires an evaluation of the expected degree of improvement in visibility from BART controls. All five statutory factors, including cost-effectiveness and expected visibility improvement, should be reflected in the level of BART control that the State implements. We believe that modeling, which provides model concentration estimates that are readily converted to deciviews, is the most efficient way to determine expected visibility improvement.
  
  For the purposes of determining visibility improvement, States may evaluate visibility changes on a pollutant-specific basis. If expected improvement is shown from the various control choices, the State can weigh the results with the other four BART determination factors when establishing BART for a particular source. For example, a State can use the CALPUFF model to predict visibility impacts from an EGU in examining the option to control NOXand SO2with SCR technology and a scrubber, respectively. A comparison of visibility impacts might then be made with a modeling scenario whereby NOXis controlled by combustion controls. If expected visibility improvements are significantly different under one control scenario than under another, then a State may use that information, along with information on the other BART factors, to inform its BART determination.
  
  Even though the visibility improvement from an individual source may not be perceptible, it should still be considered in setting BART because the contribution to haze may be significant relative to other source contributions in the Class I area. Thus, we disagree that the degree of improvement should be contingent upon perceptibility. Failing to consider less-than-perceptible contributions to visibility impairment would ignore the CAA's intent to have BART requirements apply to sources that contribute to, as well as cause, such impairment.
  
  Although we are not requiring States to use allowable emission rates to predict the anticipated future visibility impacts of BART controls, we disagree that daily average actual emission rates should be used to make this assessment. Emissions from a source can vary widely on a day to day basis; during peak operating days, the 24-hour actual emission rate could be more than double the daily average. On the other hand, in the long term, estimating visibility impacts based on allowable emission rates for every hour of the year may unduly inflate the maximum 24 hour modeled impairment estimate from a BART- eligible source. The emissions estimates used in the models are intended to reflect steady-state operating conditions during periods of high capacity utilization. We do not generally recommend that emissions reflecting periods of start-up, shutdown, and malfunction be used, as such emission rates could produce higher than normal effects than would be typical of most facilities. Where States have information on a source's daily emissions, an emission rate based on the maximum actual emissions over a 24 hour period for the most recent five years may be a more appropriate gauge of a source's potential impact as it would ensure that peak emission conditions are reflected, but would likely not overestimate a source's potential impact on any given day. We have accordingly included this change to the final guidelines. We recommend that the State use the highest 24-hour average actual emission rate, for the most recent three or five year period of meteorological data, to characterize the maximum potential benefit.
  
  Because each Class I area is unique, we believe States should have flexibility to assess visibility improvements due to BART controls by one or more methods, or by a combination of methods, and we agree with the commenters suggestions to do so. We believe the maximum 24-hour modeled impact can be an appropriate measure in determining the degree of visibility improvement expected from BART reductions (or for BART applicability). We have pointed out, however, that States should have flexibility when evaluating the fifth statutory factor. A State is encouraged to account for the magnitude, frequency, and duration of the contributions to visibility impairment caused by the source based on the natural variability of meteorology. These are important elements to consider as they would provide useful information on both the short term peak impact and long term average assessments which are critical in making the visibility assessment.
  
  We agree with the suggestion that the use of a comparison threshold, as is
  
  [[Page 39130]]
  
  done for determining if BART-eligible sources should be subject to a BART determination, is an appropriate way to evaluate visibility improvement. However, we believe the States have flexibility in setting absolute thresholds, target levels of improvement, or de minimis levels since the deciview improvement must be weighed among the five factors, and States are free to determine the weight and significance to be assigned to each factor. For example, a 0.3, 0.5, or even 1.0 deciview improvement may merit stronger weighting in one case versus another, so one ``bright line'' may not be appropriate.
  
  In addition, comparison thresholds can be used in a number of ways in evaluating visibility improvement (e.g. the number of days or hours that the threshold was exceeded, a single threshold for determining whether a change in impacts is significant, a threshold representing an x percent change in improvement, etc.). In our example modeling analysis of a hypothetical source,\52\ we used three different 24-hour thresholds (1.0, 0.5, and 0.1 deciviews) and examined the number of days that those thresholds were exceeded for a source with a 90 percent change, for example, in SO2emissions (i.e. 10,000 TPY and 1,000 TPY). The number of days that the thresholds were exceeded in the 10,000 TPY case was substantial, and the visibility improvement due to the reduction in emissions was dramatic (i.e. the number of days exceeding the thresholds dropped considerably).\53\
  
  \52\ CALPUFF Analysis in Support of the June 2005 Changes to the Regional Haze Rule, U.S. Environmental Protection Agency, June 15, 2005, Docket No. OAR-2002-0076.
  
  \53\ Ibid.
  
  Other ways that visibility improvement may be assessed to inform the control decisions would be to examine distributions of the daily impacts, determine if the time of year is important (e.g. high impacts are occurring during tourist season), consideration of the cost- effectiveness of visibility improvements (i.e. the cost per change in deciview), using the measures of deciview improvement identified by the State, or simply compare the worst case days for the pre- and post- control runs. States may develop other methods as well. 5. In What Sequence Should Alternatives Be Assessed in Step 5 of the BART Determination?
  
  Background. Both the 2001 proposal and the 2004 reproposal requested comments on two options for evaluating the ranked options. Under the first option, States would use a sequential process for conducting the impacts analysis, beginning with a complete evaluation of the most stringent control option. If a State determines that the most stringent alternative in the ranking does not impose unreasonable costs of compliance, taking into account both average and incremental costs, the analysis begins with a presumption that this level is selected. Under this option, States would then proceed to consider whether energy and non-air quality environmental impacts would justify selection of an alternative control option. If there are no outstanding issues regarding energy and non-air quality environmental impacts, the analysis is ended and the most stringent alternative is identified as the ``best system of continuous emission reduction.'' If a State determines that the most stringent alternative is unacceptable due to such impacts, this approach would require them to document the rationale for this finding for the public record. Then, the next most- effective alternative in the listing becomes the new control candidate and is similarly evaluated. This process would continue until the State identifies a technology which does not pose unacceptable costs of compliance, energy and/or non-air quality environmental impacts.
  
  We also requested comment on an alternative decision-making approach that would not begin with an evaluation of the most stringent control option. For example, States could choose to begin the BART determination process by evaluating the least stringent technically feasible control option or by evaluating an intermediate control option drawn from the range of technically feasible control alternatives. Under this approach, States would then consider the additional emissions reductions, costs, and other effects (if any) of successively more stringent control options. Under such an approach, States would still be required to (1) display all of the options and identify the average and incremental costs of each option; (2) consider the energy and non-air quality environmental impacts of each option; and (3) provide a justification for adopting the technology selected as the ``best'' level of control, including an explanation of its decision to reject the other control technologies identified in the BART determination.
  
  In selecting a ``best'' alternative, the proposed guidelines included a discussion on whether the affordability of controls should be considered. As a general matter, for plants that are essentially uncontrolled at present and emit at much greater levels per unit of production than other plants in the category, we believe it is likely that additional control will be cost-effective. The proposed guidelines noted, however, that we recognize there may be unusual circumstances that justify taking into consideration the conditions of the plant and the economic effects of requiring the use of a given control technology. These effects would include effects on product prices, the market share, and profitability of the source. We did not intend, for example, that the most stringent alternative must always be selected if that level would cause a plant to shut down, while a slightly lesser degree of control would not have this effect.
  
  Comments. We received comments supporting both of the approaches for evaluating ranked control alternatives. Many commenters, including commenters from State agencies, were supportive of the first approach. Comments from State air quality agencies were strongly supportive of this approach. These commenters believed that this approach is consistent with past approaches by States for considering control options for case-by-case determinations, is well understood by all parties, and thus easier to implement. The first approach also was strongly supported in comments from environmental organizations and private citizens. Some comments noted that the plain terminology ``best'' suggests that there must be a sound reason for not using the most stringent control level.
  
  Many comments from industrial trade organizations were critical of the first approach and believed that any requirement to use this approach would reduce State discretion because this approach, in the judgment of the commenters, would amount to use of the most stringent alternative as a default. Some of these comments asserted that the approach in option 1 would shift the BART analysis away from a cost- benefit approach mandated by the CAA towards a BACT-like technology analysis. Other commenters believed that EPA should recognize that BART, as a control requirement for retrofitting existing sources, should differ from BACT or other controls for new equipment. A number of comments, in supporting the second approach, believed that this approach provides greater consideration of the incremental cost of each succeeding option.
  
  Final rule. In the final guidelines, we have decided that States should retain the discretion to evaluate control options in whatever order they choose, so long as the State explains its analysis of the CAA factors. We agree with
  
  [[Page 39131]]
  
  commenters who asserted that the method for assessing BART controls for existing sources should consider all of the statutory factors. 6. What Should Be the Presumptive Limits for SO2and NOXfor Utility Boilers?
  
  Background. In the 2004 reproposal, we proposed that States, as a general matter, should require EGUs greater than 250 MW in size at power plants larger than 750 MW to control 95 percent of their SO2emissions, or control to within an SO2 emission range of 0.1 to 0.15 lb/mmBtu. We also proposed to establish a rebuttable presumption that States should impose these BART SO2limits on all EGUs greater than 250 MW, regardless of the size of the power plant at which they are located.
  
  For NOX, we proposed that sources currently using controls such as SCRs to reduce NOXemissions during part of the year should be required to operate those controls year-round. For power plants without post-combustion controls, we proposed to establish a presumptive emissions limit of 0.20 lbs/mmbtu for EGUs greater than 250 MW in size. We requested comment on the rate of NOX emissions that can be achieved with combustion modifications on specific types of boilers. Many commenters responded both in favor and in opposition to these proposed BART presumptive limits.
  
  Comments. A number of utility groups said the presumptive SO2emissions control approach inappropriately ignores the need for a visibility impact evaluation which is required in step 5 of the proposed case-by-case BART engineering analysis. They said that setting presumptive limits infringes on a state's authority to establish BART on a case-by-case basis considering not only visibility improvement, but the other statutory factors as well. The commenters said that visibility is both Class I area and source specific, which is the reason Congress gave the States the lead role and discretion in the BART program to determine which sources need to install or upgrade controls. Through the use of presumptions and default values, however, our prescriptive process, as proposed, would make the installation of maximum controls more likely without regard to visibility benefits. Instead, they argued, we should give the states maximum flexibility to use the five statutory factors in their BART determinations. Commenters said sources must be allowed to assess the visibility improvements of a variety of control options.
  
  Several utilities raised concern that sources with existing controls should not be required to meet the presumptive limits without the chance to evaluate the degree of visibility improvement expected from the additional emission reduction requirements. They said that if a source can demonstrate a reduction in visibility impairment below the specified threshold (whether that threshold is our currently proposed 0.5 deciview or an alternative level) with less stringent controls, then neither we nor States should impose, by default, more stringent reduction requirements.
  
  Commenters from industry, utilities, and States said that we had not indicated what previously-controlled sources must do to comply with BART, while we had determined what controls are necessary for uncontrolled sources. They were concerned that the guidelines would lead States to require previously-controlled sources to remove the controls and replace them with even newer controls at great cost and very little, if any, improvement in emission levels and visibility in Class I areas. Commenters added that States should be able to use their discretion to determine whether additional controls are needed.
  
  Some commenters were concerned that the proposed rule would require some plants to install SCR to meet the NOXcontrol level proposed, as the potential retrofit of SCR technology for the BART determination may be supported by the degree of visibility improvement expected. They said that the guidelines indicate that if a State finds that a source's visibility contribution warrants the installation of SCR, then SCR may be imposed. The commenter added, however, that the guidelines also need to provide for instances where the visibility condition warrants a lesser control level than what would be achieved by advanced combustion control; the commenter claimed there was reference to this concept in the preamble but not the guidelines.
  
  Final rule. In these guidelines, we are finalizing specific presumptive limits for SO2and NOXfor certain EGUs based on fuel type, unit size, cost effectiveness, and the presence or absence of pre-existing controls. The presumptive limits finalized in today's rulemaking reflect highly cost-effective technologies as well as provide enough flexibility for States to take particular circumstances into account.
  
  The presumptive limits apply to EGUs at power plants with a total generating capacity in excess of 750 MW. As explained in greater detail below, for these sources we are establishing a BART presumptive emission limit for coal-fired EGUs greater than 200 MW in size without existing SO2control. These EGUs should achieve either 95 percent SO2removal, or an emission rate of 0.15 lb SO2/mmBtu, unless a State determines that an alternative control level is justified based on a careful consideration of the statutory factors. For NOX, we are establishing a set of BART presumptive emission limits for coal-fired EGUs greater than 200 MW in size based upon boiler size and coal type, and based upon whether selective catalytic reduction (SCR) or selective noncatalytic reduction (SNCR) are already employed at the source. See section d. below for a table listing those specific limits. Based on our analysis of emissions from power plants, we believe that applying these highly cost-effective controls at the large power plants covered by the guidelines would result in significant improvements in visibility and help to ensure reasonable progress toward the national visibility goal.
  
  States, as a general matter, must require owners and operators of greater than 750 MW power plants to meet these BART emission limits. We are establishing these requirements based on the consideration of certain factors discussed below. Although we believe that these requirements are extremely likely to be appropriate for all greater than 750 MW power plants subject to BART, a State may establish different requirements if the State can demonstrate that an alternative determination is justified based on a consideration of the five statutory factors.
  
  In addition, while States are not required to follow these guidelines for EGUs located at power plants with a generating capacity of less than 750 MW, based on our analysis detailed below, we believe that States will find these same presumptive controls to be highly-cost effective, and to result in a significant degree of visibility improvement, for most EGUs greater than 200 MW, regardless of the size of the plant at which they are located. A State is free to reach a different conclusion if the State believes that an alternative determination is justified based on a consideration of the five statutory factors. Nevertheless, our analysis indicates that these controls are likely to be among the most cost-effective controls available for any source subject to BART, and that they are likely to result in a significant degree of visibility improvement.
  
  The rest of this section discusses these presumptive limits for SO2and NOXfor EGUs and the additional
  
  [[Page 39132]]
  
  visibility impact and cost-effectiveness analyses we have performed since proposal of the guidelines in 2004.
  1. Visibility Analysis for SO2and NOX Emissions From EGUs. In the 2004 reproposal, our preliminary CALPUFF modeling \54\ suggested that controlling a single 250 MW EGU at a 90 percent level would improve visibility substantially from that source. Based on the expected degree of improvement in visibility and the use of highly effective control technologies that are available for sources of this capacity and greater, we concluded that the specific control levels in the proposal were appropriate. Even at that level of control however, our analysis indicated that emissions from the source might still cause a perceptible impact on visibility.
    
    \54\ Summary of Technical Analyses for the Proposed Rule, Mark Evangelista, U.S. Environmental Protection Agency, April 12, 2004, Docket No. OAR-2002-0076.
    
    Following comments that we had ignored the need to consider the degree of improvement in visibility which could reasonably be anticipated from the use of the presumptive control technologies, we undertook a more comprehensive modeling analysis of the anticipated visibility impacts of controlling large EGUs. Based on this modeling analysis, we anticipate that a majority of the currently uncontrolled EGUs at power plants covered by the guideline are predicted to have 24- hour maximum impacts of greater than a change of 2 or 3 deciviews.\55\ Our modeling examples included scenarios that were representative of typical EGUs, but, in our first hypothetical run 1, we conservatively assumed SO2emissions of 10,000 tons per year (TPY) and NOXemissions of approximately 3,500 TPY.\56\ Such levels of emissions are well below those that may be expected of an uncontrolled 200 MW EGU. The number of days during any year that such sources are predicted to have visibility impacts of greater than 0.5 deciviews or even 1.0 deciview were 29 days and 12 days on average, respectively, at 50 km from a hypothetical Class I area in the East; if the 98th percentile were considered, there would be five days above a 1.0 deciview change.
    
    \55\ CALPUFF Analysis in Support of the the June 2005 Changes to the Regional Haze Rule, U.S. Environmental Protection Agency, June 15, 2005, Docket No. OAR-2002-0076.
    
    \56\ Ibid.
    
    The modeled emission rates in the example were conservative; for much larger EGUs with capacities of 750 MW or more, and emission rates much higher than those which were modeled, visibility degradation is expected to be far worse. Clearly there is a substantial degree of visibility improvement which is likely from emission reductions at these sources.
    
    Although we are confident that the EGUs for which we are establishing presumptive limits each have a significant impact on visibility at one or more Class I areas, a State retains the option and flexibility to conduct its own analysis or allow a source to demonstrate that it should not be subject to BART (based on its visibility effects).
  2. BART Presumptive Limits for SO2From Coal-Fired Units. For currently uncontrolled coal-fired EGUs greater than 200 MW in size located at power plants greater than 750 MW, we are establishing a presumptive BART limits of 95 percent SO2 removal, or an emission rate of 0.15 lb SO2/mmBtu. We are not establishing a presumptive limit for EGUs with existing post- combustion SO2controls or for EGUs that burn oil.
    
    In 2004, we proposed presumptive limits for SO2of 95 percent control or a comparable performance level of 0.1 to 0.15 lbs per million BTU as controls that would be achievable and cost- effective. We requested comment on the removal effectiveness of flue gas desulfurization (``FGD'' or ``scrubber'' controls) for various coal types and sulfur content combinations. Having considered the comments received, we have determined that there is ample data to support the determination that the BART presumptive limits outlined in today's action are readily achievable by new wet or semi-dry FGD systems across a wide range of coal types and sulfur contents based on proven scrubber technologies currently operational in the electric industry.\57\
    
    \57\ Technical Support Document for BART SO2 Limits for Electric Generating Units, Memorandum to Docket OAR 2002-0076, April 1, 2005.
    
    We agree with the commenters who stated that our dual recommendation provided equity across sources burning coals of varying sulfur content. We believe the presumptive limits provide enough flexibility that absent unique circumstances, any BART-eligible coal- fired EGU will be able to achieve one of the limits with a new FGD system. We expect that BART-eligible EGUs burning medium to high sulfur coal will be able to achieve a removal efficiency of 95 percent in a cost effective manner by utilizing various wet FGD technologies, and that those EGUs burning lower sulfur coals could meet the emission limit of 0.15lb/mmBtu in a cost effective manner by utilizing dry FGD technologies. As described below, EPA's unit specific economic modeling showed that the majority of BART eligible units greater than 200 MW can meet the presumptive BART limit at a cost of $400 to $2000 per ton of SO2removed.
    
    Some commenters expressed concerns that the proposed limits were too stringent in particular for: (1) EGUs less than 750 MW in size, (2) EGUs burning low sulfur coals, and (3) EGUs burning lignite coals. However, numerous examples exist of smaller EGUs and EGUs burning low sulfur or lignite coals achieving these SO2limits at reasonable cost.\58\ We recognize that semi-dry FGD systems are most commonly utilized on units burning lower sulfur coals and are not typically designed for removal efficiencies of 95 percent or greater. However, we believe that most of these EGUs can readily achieve the presumptive emission rate limit of 0.15 lb SO2/mmBtu. An analysis of EPA's RACT/BACT/LEAR Clearinghouse Dry FGD cost effectiveness data ranged from $393 to $2132 per ton SO2 removed, with an average cost effectiveness of $792 per ton.\59\
    
    \58\ Ibid.
    
    \59\ Summary of BART Source Analyses, Memorandum from Bill Balcke and Doran Stegura, Perrin Quarles Associates, Inc., to Chad Whiteman, EPA March 24, 2003. See 2001 emissions data in BART AR file, attached.
    
    We received a few comments expressing the belief that the presumptive limits should be more stringent, given that BART emission limits will not be fully implemented until 2013 or 2014. We recognize that while some scrubber units currently achieve reductions greater than 95 percent, not all units can do so. The individual units that currently achieve greater than 95 percent control efficiencies do not necessarily represent the wide range of unit types across the universe of BART-eligible sources. An analysis of the Department of Energy's U.S. FGD Installation Database supports our belief that 95 percent removal efficiencies would be obtainable by all types of EGUs burning medium and high sulfur coal by 2014, including BART-eligible EGUs. In addition, we note that the presumption does not limit the States' ability to consider whether a different level of control is appropriate in a particular case. If, upon examination of an individual EGU, a State determines that a different emission limit is appropriate based upon its analysis of the five factors, then the State may apply a more or less stringent limit.
    
    Our analysis of presumptive BART limits accounted for variations in existing SO2controls. We accordingly considered (1) coal- fired EGUs without
    
    [[Page 39133]]
    
    existing SO2controls, and (2) coal-fired EGUs with existing SO2controls. This analysis consisted of the following key elements: (1) Identification of all potentially BART-eligible EGUs, and (2) technical analyses and industry research to determine applicable and appropriate SO2control options, (3) economic analysis to determine cost effectiveness for each potentially BART-eligible EGU, and (4) evaluation of historical emissions and forecast emission reductions for each potentially BART-eligible EGU.\60\
    
    \60\ Ibid.
    
    We identified 491 potentially BART-eligible coal-fired units based on the following criteria: (1) The unit was put in place between August 7, 1962 and August 7, 1977, and (2) the unit had the potential to emit more than 250 tons annually of SO2. Our assessment of potential controls included various industry case studies, technical papers, public comments, BACT analyses, and historical Acid Rain emissions data. Our analysis is described in detail in the TSD.\61\
    
    \61\ Ibid.
    
    We calculated cost effectiveness and projected SO2 emission reductions on a per unit basis based on removal efficiencies of 90 percent for dry FGD systems, in particular spray dry lime systems, and 95 percent for wet FGD systems, in particular limestone forced oxidation systems. Based on our analysis, the average cost effectiveness for controlling all BART-eligible EGUs greater than 200 MW without existing SO2controls was estimated to $919 per ton of SO2removed. Moreover, the range of costs effectiveness numbers demonstrates that the majority of these units can meet the presumptive limits at a cost of $400 to $2000 per ton of SO2removed.
    
    Figure 1
    
    Calculated
    
    Percent of Percent of BART average cost
    
    estimated Unit capacity (MW)
    
    Tons (K) of SO2 eligible coal- effectiveness for removable BART SO2 emitted in 2001 fired unit's 2001 MW grouping ($/ emissions from emissions ton SO2 removed) coal-fired units*
    
    300 MW...........................
    
    5712
    
    85.2 .................
    
    84.8 All Units.........................
    
    6707
    
    100
    
    984
    
    100 BART Units (>200MW)...............
    
    6246
    
    92.2
    
    919
    
    91.9
    
    In establishing presumptive BART limits, we were cognizant of the fact that upgrading an existing scrubber system is typically considered more cost effective than constructing a new scrubber system. However, due to the diverse and complex nature of upgrading existing FGD systems (scrubber type, reagents, online year, absorber characteristics, current operating procedures, etc.), there is no single solution or standard appropriate for all EGUs. As a result, we are not including specific numerical presumptive limits for EGUs with pre-existing scrubbers. However, for scrubbers currently achieving removal efficiencies of at least 50 percent, we recommend States evaluate a range of scrubber upgrade options available for improving the SO2removal performance of existing units. There are numerous scrubber enhancements available to upgrade the average removal efficiencies of all types of existing scrubber systems, and the guidelines contains a discussion of the options that States should evaluate in making BART determinations for EGUs with existing scrubbers.
    
    The guidelines do not require EGUs with existing FGD systems to remove these controls and replace them with new controls, but the guidelines do state that coal fired EGUs with existing SO2 controls achieving removal efficiencies of less than 50 percent should consider constructing a new FGD system to meet the presumptive limits of 95 percent removal or 0.15 lb/mmBtu in addition to evaluating the suite of upgrade options. For these EGUs, the suite of available ``upgrades'' may not be sufficient to remove significant SO2 emissions in a cost effective manner, and States may determine that these EGUs should be retrofitted with new FGD systems.
  3. BART Limits for SO2 From Oil-Fired Units. We are not establishing a presumptive BART limit for SO2from oil-fired EGUs. The guidelines state that the most appropriate control option for oil-fired EGUs, regardless of capacity, is to set limits on the sulfur content of the fuel oil burned in the unit.
    
    Commenters suggested EPA evaluate two primary control options for BART oil-burning units: (1) Sulfur content fuel oil limitations, and (2) flue gas desulfurization systems. We have been unable to find any FGD application in the U.S. electric industry on an oil-fired unit. As a result, our analysis for oil-fired units focused on benchmarking previously imposed fuel oil restrictions on the electric industry and (2) a regional economic analysis of switching from high sulfur to low sulfur fuel oil.
    
    Our study of currently imposed fuel oil restrictions on the electric industry suggested that all BART-eligible EGUs currently have some sort of imposed sulfur content or emission rate limitation. Of the 74 BART-eligible oil-burning EGUs, 32 currently have sulfur fuel oil restrictions of less than 1 percent, and 67 have some sort of sulfur content limitation. In addition, our economic analysis suggests that switching to low sulfur fuel oil is a cost effective method in reducing SO2emission from oil fired units.
    
    As approximately 43 percent of the BART eligible oil units currently have a sulfur content limitation that is either equivalent to, or more stringent than, one percent sulfur by weight, the guidelines require States to consider a one percent or lower sulfur by weight fuel oil restriction on all BART eligible EGUs as part of their BART analysis, and recommends that States establish appropriate and sustainable sulfur content fuel oil restrictions, taking into
    
    [[Page 39134]]
    
    account fuel oil availability. States should accordingly evaluate a one percent sulfur content limitation as a starting point of their BART determination for oil-fired EGUs subject to BART.
  4. BART Presumptive Limits for NOX From Coal-fired Units. In the 2004 reproposal, in discussing NOXcontrols on EGUs, we explained that there are two somewhat distinct approaches to reducing emissions of NOXat existing sources. One is to use combustion controls (including careful control of combustion air and low-NOXburners). The other approach is removal technology applied to the flue gas stream (such as SCRs and SNCRs).
    
    For EGUs currently using controls such as SCRs or SNCRs to reduce NOXduring part of the year, we are establishing a presumption that use of these same controls year-round is BART. (Some commenters supported year-round operation of these controls. One commenter suggested the cost of year-round operation of SCRs would be significant. However, our analysis showed year-round operation of existing SCRs compared to operation during the 5-month ozone season only to be highly cost effective (average cost-effectiveness of $170 per ton).) Although only a few BART-eligible sources currently have SNCRs installed, we note that States may wish to consider SCR as an alternative to annual operation of SNCR in light of the relatively high operating costs associated with SNCR.
    
    For sources without post-combustion controls (i.e., SCRs and SNCRs), we are establishing a presumption as to the appropriate BART limits for coal-fired units based on boiler design and coal type. These presumptions apply to EGUs greater than 200 MW at power plants with a generating capacity greater than 750 MW and are based on control strategies that are generally cost-effective for all such units.
    
    In 2004 we noted that, unlike the methods for controlling SO2(which fall within a fairly narrow range of cost effectiveness and control efficiencies), the removal efficiencies and costs associated with the control techniques for NOXvary considerably, depending on the design of the boiler and the type of coal used. In response to comments on the proposal, we have performed additional analyses of all individual BART-eligible coal-fired units \62\ and our analyses indicated that both cost effectiveness and post- control rates for NOXdo depend largely on boiler design and type of coal burned. Based on these analyses, we believe that States should carefully consider the specific NOXrate limits for different categories of coal-fired utility units, differentiated by boiler design and type of coal burned, set forth below as likely BART limits.
    
    \62\ See Technical Support Document for BART NOX Limits for Electric Generating Units and Technical Support Document for BART NOXLimits for Electric Generating Units Excel Spreadsheet, Memorandum to Docket OAR 2002-0076, April 15, 2005.
    
    In today's action, EPA is setting presumptive NOXlimits for EGUs larger than 750 MW. EPA's analysis indicates that the large majority of the units can meet these presumptive limits at relatively low costs. Because of differences in individual boilers, however, there may be situations where the use of such controls would not be technically feasible and/or cost-effective. For example, certain boilers may lack adequate space between the burners and before the furnace exit to allow for the installation of over-fire air controls. Our presumption accordingly may not be appropriate for all sources. As noted, the NOXlimits set forth here today are presumptions only; in making a BART determination, States have the ability to consider the specific characteristics of the source at issue and to find that the presumptive limits would not be appropriate for that source.
    
    The table below indicates the types of boilers installed at the 491 BART-eligible coal-fired EGUs. Dry-bottom wall-fired boiler units and tangentially-fired boiler units make up a large majority of the total BART-eligible EGUs.
    
    Table 1.--Population of BART-Eligible Coal-Fired EGUs
    
    Number
    
    Number
    
    Number
    
    Boiler type
    
    Units > 200 MW All units Units > 200 MW at 750 MW plants
    
    Cyclone...................................................
    
    56
    
    35
    
    19 Cell Burner...............................................
    
    35
    
    35
    
    29 Dry Bottom--Wall fired....................................
    
    188
    
    121
    
    77 Dry Bottom Turbo-fired....................................
    
    14
    
    10
    
    4 Stoker....................................................
    
    5
    
    0
    
    0 Tangentially-fired........................................
    
    186
    
    164
    
    112 Wet Bottom................................................
    
    6
    
    5
    
    5 Other.....................................................
    
    1
    
    0
    
    0
    
    Total BART-eligible coal-fired EGUs...................
    
    491
    
    370
    
    246
    
    For all types of boilers other than cyclone units, the limits in Table 2 are based on the use of current combustion control technology. Current combustion control technology is generally, but not always, more cost-effective than post-combustion controls such as SCRs. For cyclone boilers, SCRs were found to be more cost-effective than current combustion control technology;\63\ thus the NOXlimits for cyclone units are set based on using SCRs. SNCRs are generally not cost-effective except in very limited applications and therefore were not included in EPA's analysis. The types of current combustion control technology options assumed include low NOXburners, over- fire air, and coal reburning.
    
    \63\ The current combustion control technology EPA analyzed for cyclone units is coal reburning.
    
    We are establishing presumptive NOXlimits in the guidelines that we have determined are cost-effective for most units for the different categories of units below, based on our analysis of the expected costs and performance of controls on BART-eligible units greater than 200 MW. We assumed that coal-fired EGUs would have space available to install separated over-fire air. Based on the large number of units of various boiler designs that have installed separated over- fire air, we believe this assumption to be reasonable. It is
    
    [[Page 39135]]
    
    possible, however, that some EGUs may not have adequate space available. In such cases, other NOXcombustion control technologies could be considered such as Rotating Opposed Fire Air (``ROFA''). The limits provided were chosen at levels that approximately 75 percent of the units could achieve with current combustion control technology. The costs of such controls in most cases range from just over $100 to $1000 per ton. Based on our analysis, however, we concluded that approximately 25 percent of the units could not meet these limits with current combustion control technology. However, our analysis indicates that all but a very few of these units could meet the presumptive limits using advanced combustion controls such as rotating opposed fire air (``ROFA''), which has already been demonstrated on a variety of coal-fired units. Based on the data before us, the costs of such controls in most cases are less than $1500 per ton.
    
    \64\ No Cell burners, dry-turbo-fired units, nor wet-bottom units burning lignite were identified as BART-eligible, thus no presumptive limit was determined. Similarly, no wet-bottom units burning sub-bituminous were identified as BART-eligible.
    
    \65\ These limits reflect the design and technological assumptions discussed in the technical support document for NOXlimits for these guidelines, e.g., EPA assumed space would be available for over-fire air. See Technical Support Document for BART NOXLimits for Electric Generating Units and Technical Support Document for BART NOX Limits for Electric Generating Units Excel Spreadsheet, Memorandum to Docket OAR 2002- 0076, April 15, 2005.
    
    Table 2.--Presumptive NOX Emission Limits for BART-Eligible Coal-Fired Units \64\
    
    NOX presumptive Unit type
    
    Coal type
    
    limit (lb/mmbtu) \65\
    
    Dry-bottom wall-fired......... Bituminous............
    
    0.39 Sub-bituminous........
    
    0.23 Lignite...............
    
    0.29 Tangential-fired.............. Bituminous............
    
    0.28 Sub-bituminous........
    
    0.15 Lignite...............
    
    0.17 Cell Burners.................. Bituminous............
    
    0.40 Sub-bituminous........
    
    0.45 Dry-turbo-fired............... Bituminous............
    
    0.32 Sub-bituminous........
    
    0.23 Wet-bottom tangential-fired... Bituminous............
    
    0.62
    
    Table 3.--Average Cost-Effectiveness of NOX Controls for BART-Eligible Coal-Fired Units
    
    Number units National average Unit type
    
    Coal type
    
    nation-wide
    
    ($/ton)
    
    Dry-bottom wall-fired.................... Bituminous.......................
    
    114
    
    1229 Sub-bituminous...................
    
    66
    
    576 Lignite..........................
    
    3
    
    1296 Tangential-fired......................... Bituminous.......................
    
    105
    
    567 Sub-bituminous...................
    
    72
    
    281 Lignite..........................
    
    9
    
    614 Cell Burners............................. Bituminous.......................
    
    32
    
    1287 Sub-bituminous...................
    
    3
    
    1021 Dry-turbo-fired.......................... Bituminous.......................
    
    7
    
    775 Sub-bituminous...................
    
    7
    
    599 Wet-bottom............................... Bituminous.......................
    
    6
    
    378 Cyclones (with SCR)...................... All..............................
    
    56
    
    900
    
    The advanced combustion control technology we used in our analysis, ROFA, is recently available and has been demonstrated on a variety of unit types. It can achieve significantly lower NOXemission rates than conventional over-fire air and has been installed on a variety of coal-fired units including T-fired and wall-fired units. We expect that not only will sources have gained experience with and improved the performance of the ROFA technology by the time units are required to comply with any BART requirements, but that more refinements in combustion control technologies will likely have been developed by that time. As a result, we believe our analysis and conclusions regarding NOXlimits are conservative.\66\ For those units that cannot meet the presumptive limits using current combustion control technology, States should carefully consider the use of advanced combustion controls such as ROFA in their BART determination.
    
    \66\ See Technical Support Document for BART NOX Limits for Electric Generating Units and Technical Support Document for BART NOX Limits for Electric Generating Units Excel Spreadsheet, Memorandum to Docket OAR 2002-0076, April 15, 2005.
    
    A detailed discussion of our analysis is in the docket.\67\ For data on emissions and existing control technology in use at the BART- eligible EGUs, we used EPA's Clean Air Markets Division database.\68\
    
    \67\ Id.
    
    \68\ Reporting requirements for the Acid Rain Program and NOXSIP Call affected sources, see 40 CFR 75 subpart G (parts 7562-64), and EPA Clean Air Markets Division Web site, data and maps page (http://www.epa.gov/airmarkets).
6. Selective Catalytic Reduction (``SCR'') and Cyclone Units
  
  We also analyzed the installation of SCRs at BART-eligible EGUs, applying SCR to each unit and fuel type. The cost-effectiveness was generally higher than for current combustion control technology except for one unit type, cyclone units. Because of the relatively high NOXemission rates of cyclone units, SCR is more cost- effective. Our analysis indicated that the cost-effectiveness of applying SCR on coal-fired cyclone units is typically less than $1500 a ton, and that the average cost-
  
  [[Page 39136]]
  
  effectiveness is $900 per ton.\69\ As a result, we are establishing a presumptive NOXlimit for cyclone units based on the use of SCR. For other units, we are not establishing presumptive limits based on the installation of SCR. Although States may in specific cases find that the use of SCR is appropriate, we have not determined that SCR is generally cost-effective for BART across unit types.
  
  \69\ See Technical Support Document for BART NOX Limits for Electric Generating Units and Technical Support Document for BART NOX Limits for Electric Generating Units Excel Spreadsheet, Memorandum to Docket OAR 2002-0076, April 15, 2005.
  
  Oil and Gas-Fired Units
  
  For oil-fired and gas-fired units, we believe that installation of current combustion control technology is highly cost-effective and should be considered in determining BART for these sources. We performed an analysis of BART-eligible oil and gas-fired units similar to the analysis done for coal-fired units. Our analysis indicated that a number of units can make significant reductions in NOX emissions which are cost-effective through the application of current combustion control technology.\70\ However, for a number of units, the use of combustion controls does not appear to be cost-effective. As a result, we determined that it would be inappropriate to establish a general presumption regarding likely BART limits. As a result, the guidelines only indicate that States should consider the installation of current combustion control technology on oil and gas-fired units.
  
  \70\ Id.
How Does Today's Rule Affect States Options for Using Alternative Strategies in Lieu of Source-by-Source BART?

Background

Over the past several years, there have been a number of rule makings and court decisions on the subject of BART and BART-alternative programs. In order to understand today's actions, it is useful to again review the regulatory and litigation history, with a specific focus on BART-alternative issues.

As noted in part I of this preamble, the 1999 regional haze rule included provisions for BART, codified at 40 CFR 51.308(e), and in definitions that appear in 40 CFR 51.301. Among these provisions was section 308(e)(2), allowing States to implement cap and trade programs, or other alternative programs, in lieu of BART. Section 308(e)(2) provided that trading program alternatives must be demonstrated to achieve greater reasonable progress than BART, and provided the general parameters for making this demonstration. Of particular relevance, section 308(e)(2) directed States, in the course of estimating emissions reductions anticipated from source-by-source BART, to determine what comprises BART based on the four non-visibility factors, and then estimate visibility improvements based on the application of BART to all sources subject to BART. In other words, section 308(e)(2) indicated that states should use what has since been termed a ``group BART'' approach to estimating the source-by-source BART benchmark, for comparison to the alternative program. Section (e)(2) did not prescribe the specific criteria to be used to compare the progress estimated from source-by-source BART to that anticipated from the trading program. The preamble discussion indicated that the comparison should be based on both emission reductions and visibility improvement, but did not provide further specificity. See 64 FR at 35741-35743.

Specific criteria for making the comparison to programs was proposed in the BART Guidelines (40 CFR 51 App. Y) in 2001. These criteria--sometimes referred to as the ``better-than-BART test'' consist of the following. First, if the geographic distribution of emissions reductions from the two programs is expected to be similar, the comparison can be made based on emissions alone. Second, if the distribution of emissions reductions is anticipated to be significantly different, then a two-pronged visibility improvement test is employed. The first prong is that the alternative program must not result in a degradation of visibility at any Class I area. The second prong is that the alternative program must result in greater visibility improvement overall, based on an average across all affected Class I areas. See 66 FR 38133.

In 2002, the D.C. Circuit decided American Corn Growers. The court in that decision invalidated ``the BART provisions'' on the basis that EPA had improperly constrained State authority by requiring them to bifurcate visibility from the other statutory factors when making BART determinations, and by specifying that visibility impairment should be considered on a group basis when determining whether a BART eligible source is subject to BART. 291 F.3d 1, 8.

Because EPA's policy of allowing alternative programs to BART was not at issue in American Corn Growers, the decision contained no discussion of how such alternative programs would be compared to BART-- neither the step of estimating emissions from source-by-source BART, nor the criteria for the actual comparison (i.e., the test). Therefore, EPA interpreted the court's vacature of the BART provisions to apply to the source-by-source BART regulations under 40 CFR 51.308(e)(1). Accordingly, in our May 2004 reproposal of the BART guidelines, we did not propose any changes in section 308(e)(2), and we retained the section on trading programs in the guidelines (Appendix Y) as that section was proposed in 2001.

In June 2004, in the Supplemental Notice of Proposed Rulemaking (SNPR) for the Clean Air Interstate Rule (CAIR), we proposed to conclude that the CAIR will achieve greater reasonable progress than would BART for SO2and NOXat BART-eligible EGUs in CAIR affected States and therefore may be treated as a program in lieu of BART for those sources. In doing so, we discussed regional haze rule section 308(e)(2) as precedent for the policy of allowing trading programs to substitute for BART.\71\ However, noting that the CAIR trading program affected only one category of BART-eligible sources (EGUs), rather than all BART-eligible categories as envisioned for State-developed BART-alternative programs under section 308(e)(2), we proposed adding a 308(e)(3) applicable only to CAIR. This section would provide that states that comply with the CAIR by subjecting EGUs to the EPA administered cap and trade program may consider BART satisfied for NOXand SO2from BART-eligible EGUs. In the CAIR SNPR and supporting documentation,\72\ we provided analyses demonstrating that CAIR would achieve greater emission reductions than BART, and would make greater reasonable progress according to the two- pronged visibility test previously proposed in the BART guidelines.

\71\ Section 308(e)(2) was based, in turn, on the precedent set by our interpretation of CAA 169A(b)(2) in a single BART-source context--see 64 FR 35739, citing Central Arizona Water Conservation District, 990 F.2d 1531 (1993).

\72\ ``Supplemental Air Quality Modeling Technical Support Document (TSD) for the Clean Air Interstate Rule (CAIR), May, 2004.'' http://www.epa.gov/cair/pdfs/saqmtsd.pdf.

In February 2005, in CEED v. EPA, the D.C. Circuit invalidated a BART-alternative program developed by the Western Regional Air Partnership (WRAP), which was also based on a requirement of group-BART analysis in setting source-by-source benchmark. It is important to note that the two-pronged better-than-BART test was not

[[Page 39137]]

at issue in CEED, as neither the States nor EPA had employed that test in determining that the WRAP's program achieved greater progress than BART. The issue on which the court based its decision was not how the two programs were compared, but how States were required to estimate reductions from source-by-source BART in order to make the comparison. The implications of this case to today's action are discussed in more detail below.

Finally, on March 10, 2005 we promulgated the final CAIR. In the final CAIR, we presented refined and updated analyses continuing to show that CAIR makes greater progress than BART. We concluded at that time that we should defer a final ``better than BART'' determinations until (1) the source-by-source BART guidelines for EGU were promulgated, and (2) the criteria for comparing alternatives to BART were also finalized. We are taking both of those actions today, and, as explained below, are therefore also making our final determination that CAIR achieves greater progress than BART and may be used by States as a BART substitute.

Final Criteria for Comparing Visibility Progress of an Alternative Program to BART

Proposed Rule. As noted, the criteria for determining if an alternative measure achieves greater reasonable progress than BART (also known as the ``better than BART'' test or the two- pronged visibility test) were first proposed in the 2001 BART guideline proposal and reproposed in the identical form in the 2004 BART guidelines reproposal. The test appeared as an element of the guideline's overview of the steps involved in developing a trading program consistent with regional haze rule section 308(e)(2).

Specifically, the guidelines provided that States could first look at the geographic distribution of emissions under the trading program. ``If [the] distribution of emissions is not substantially different than under BART, and greater emissions reductions are achieved, then the trading program would presumptively achieve ``greater reasonable progress.'' (69 FR at 25231). If the distribution of emissions is expected to be different, then States are directed to conduct an air quality modeling study. The guidelines then provide that

``[t]he modeling study would demonstrate ``greater reasonable progress'' if both of the following two criteria are met:

--Visibility does not decline in any Class I area, and --Overall improvement in visibility, determined by comparing the average differences over all affected Class I areas

Comments Received

Several commenters stated that the trading criteria contained in the proposed BART guidelines were, along with other parts of the guidelines, beyond EPA's authority to impose under the CAA.

Several State commenters asked for clarification of what should be considered a significantly different geographic distribution of emission reductions, for purposes of proceeding to the two-pronged visibility test.

One comment, submitted by environmental groups in response to our preliminary application of the two-pronged test to the CAIR in the CAIR rulemaking, goes to the permissibility of that test in general and is therefore relevant to the finalization of the test. Specifically, these commenters stated that because section 169A(b)(2)(A) requires BART for an eligible source which may reasonably be anticipated to cause or contribute to any impairment of visibility in any Class I area, EPA is without basis in law or regulation to base a better-than-BART determination on an analysis that uses averaging of visibility improvement across different Class I areas.

Final Action. We are amending the regional haze rule to incorporate the two- prong visibility test as it was previously proposed in the BART guideline proposals. Specifically, we are adding the test to the rule provisions at section 51.308(e)(3).

The EPA has the authority to prescribe this methodology under its general rulemaking authority provided by CAA section 301(a), and under CAA sections 169A(4) and 169(e). The latter provisions require EPA to promulgate regulations to assure reasonable progress towards the national visibility goal and to assure compliance with the requirements of section 169A, which include the requirements for BART under section 169A(b)(2)(A), and to promulgate such measures as may be necessary to carry out these regulations. The EPA has determined that source-by- source BART need not be required when it is not necessary to meet reasonable progress because greater progress can be achieved by an alternative means. The D.C. Circuit in CEED upheld this interpretation of the BART provisions' relationship to the broader reasonable progress requirements of the Act. 398 F.3d at 660. In order to assure that such alternative programs meet the reasonable progress goals of the CAA, EPA has the authority, and perhaps a duty, to promulgate regulations governing how that determination is made.

Moreover, these requirements for making the ultimate comparison between an alternative program and BART do not affect in any way how states make BART determinations or how they determine which sources are subject to BART. It is in those areas where the Act and legislative history indicate that Congress evinced a special concern with insuring that States would be the decision makers. Nothing in American Corn Growers or CEED suggests that those cases rendered EPA's rulemaking authority under section 169A(a)(4) completely inoperable in any BART context.

With respect to the use of average overall improvement, we explained in the CAIR NFR preamble that we disagree with comments that CAA section 169A(b)(2)'s requirement of BART for sources reasonably anticipated to contribute to impairment at any Class I area means that an alternative to the BART program must be shown to create improvement at each and every Class I Area. Even if a BART alternative is deemed to satisfy BART for regional haze purposes, based on average overall improvement as opposed to improvement at each and every Class I Area, CAA section 169A(b)(2)'s trigger for BART based on impairment at any Class I area remains in effect, because a source may become subject to BART based on ``reasonably attributable visibility impairment'' at any area. See 40 CFR 51.302. In addition, within a regional haze context, not every measure taken is required to achieve a visibility improvement at every class I area. BART is one component of long term strategies to make reasonable progress, but it is not the only component. The requirement that the alternative achieves greater progress based on the average improvement at all Class I areas assures that, by definition, the alternative will achieve greater progress overall. Though there may be cases where BART could produce greater improvement at one or more class I areas, the no-degradation prong assures that the alternative will not result in worsened conditions anywhere than would otherwise exist, and the possibility of BART for reasonably attributable visibility protects against any potential ``hot spots.'' Taken together, the EPA believes these factors make a compelling case that the proposed test properly defines ``greater reasonable progress.'' The EPA anticipates that regional haze implementation plans will also contain measures addressing other sources as

[[Page 39138]]

necessary to make progress at every mandatory Federal Class I area.

We are therefore finalizing the test criteria in the same form in which they were proposed as part of the BART guidelines. We also recognize that the test criteria leave some terms and conditions undefined, and we believe States and Tribes should retain the discretion to reasonably interpret and apply these terms as appropriate to the context of the particular program at issue.

First, in the proposed test we did not specify the time period which should serve as the starting point for comparison under the first prong. That is, we did not specify whether potential degradation should be determined in relation to visibility conditions existing at the time of the proposed program, or in relation to base case visibility projections for the time of program implementation. While either option is, we believe, reasonable, in this rulemaking we have used the future projected base case, for the following reasons.

The underlying purpose of both prongs of the test is to assess whether visibility conditions at Class I areas would be better with the alternative program in place than they would without it. The first prong ensures that the program does not cause a decline in visibility at any particular Class I area. It addresses the possibility that the alternative program might allow local increases in emissions which could result in localized degradation. The second prong assesses whether the alternative program produces greater visibility improvement in the aggregate than would source specific BART.

In both cases, the logical reference point is visibility conditions as they are expected to be at the time of program implementation but in the absence of the program. This insures that the visibility improvements or degradations determined are due to the programs being compared--source-specific BART and the cap-and-trade alternative--and not to other extrinsic factors. For example, if large increases in wild land fires are expected, due to accumulation of fuel from past forest management practices, a degradation of visibility from current conditions may be expected. It would be irrational to disapprove an alternative program because of a modeled degradation from current conditions, where that degradation is actually anticipated because of smoke from such fires--sources which are not subject to the CAA BART provisions. By comparing the alternative to future projected baseline conditions, such extrinsic variables are accounted for. We are thus able to ascertain (to the extent possible where future projections are concerned) whether visibility under the alternative would decline at any Class I area, all other things being equal.

Therefore, in applying the test to the CAIR, we used the future (2015) projected baseline. We believe, however, that States should have discretion in determining the most appropriate baseline for this prong of the test, as long as the State's method is reasonable.

Second, although the proposed test indicated that dispersion modeling should be used to determine visibility differences for the worst and best 20 percent of days, the guideline did not specify the relationship between the worst and best days and the two prongs of the test. We believe that each prong of the test should ideally be based on an examination of both the worst and best 20 percent of days. Thus, under the first prong, visibility must not decline at any one Class I area on either the best 20 percent or the worst 20 percent days \73\ as a result of implementing the alternative program; and, under the second prong both the best and worst days should be considered in determining whether the alternative program produces greater average improvement.

\73\ The regional haze rule requires States to establish reasonable progress goals for each Class I area that provide for improvement in visibility for the most impaired days and ensure no degradation in visibility for the most impaired days. The reasonable progress test in the regional haze rule remains as a separate test from better than BART. The SIPs must contain measures to achieve the reasonable progress goal; such measures could include not only stationary source programs such as BART but also programs to address emissions from other types of sources. The no degradation (on the 20 percent best days) component of the reasonable progress test must still be applied to the final future year emissions control strategy. This does not directly impact the conclusions of the better than BART test.

Third, the proposed guidelines did not define ``affected'' Class I areas for purposes of the comparison. In applying the test to the CAIR, we considered all federal mandatory Class I areas in the contiguous 48 States for which data was available. The principal Class I areas affected by the CAIR are those in the eastern U.S., therefore we calculated average improvement separately for the eastern areas, but also considered affects at all Class I areas nationally. We believe this was appropriate for a federally mandated program of the scope and magnitude of the CAIR. However, this may not be necessary for every BART-alternative program developed by States in the future, especially if proposed programs are limited to smaller geographic areas or are limited to source categories having significantly less widespread impacts than EGUs. In such circumstances, it may be reasonable for the States and Tribes involved to develop criteria for ``affected'' Class I areas. For example, the affected region could be considered to be the States and Tribes involved in the trading program as well as immediately adjacent States, or Class I areas within adjacent States that are within some defined distance of participating States.

With respect to comments on the degree of difference in the geographic distribution of emissions necessary to trigger application of the two prong test, we believe it is not necessary for EPA to define that in the rule. For our CAIR analysis, we explained in the SNPR that the fact that CAIR would produce greater emissions reductions than BART in most States, but less reductions than BART in a few States, was sufficient reason to employ the two pronged visibility test, 69 FR 32704. For other programs developed by States, a State would have the ability to make a reasonable decision as to whether there was a sufficient basis to make the demonstration that an alternative program would be better than BART based on modeling of the emissions distributions alone, or whether the State should proceed with the two- pronged visibility test. The State's discretion is subject as always to the condition that it must be reasonably exercised, and must be supported by adequate documentation of the analyses.

Finally, on a related issue, we note that in a separate rule making to follow soon after today's action, we will be soliciting comments on whether there might be other means of demonstrating that an alternative program makes greater reasonable progress than BART, in addition to the two-pronged visibility test we are finalizing in today's action. Such other means might take into account additional policy considerations, as well as the relative degree of visibility improvement of the two programs.
1. Final Determination That CAIR Makes Greater Reasonable Progress Than BART
  
  Proposal. As noted in the background section above, in both the CAIR SNPR, and NFR, we discussed the proposed approach of allowing States to treat CAIR as an in-lieu-of BART program for EGUs in CAIR- affected States. In both actions, we presented analyses based on emission projections and air quality modeling showing that CAIR will achieve greater reasonable progress
  
  [[Page 39139]]
  
  towards the national visibility goal than would BART for affected EGUs. These analyses were conducted according to the criteria for making such ``better than BART'' determinations which had been proposed in the BART guidelines, and which have now been finalized in the regional haze rule at 40 CFR 51.308(e)(3), as discussed above in section IV.B. Below, we briefly recap these prior analyses. See 69 FR 32684, 32702-32707 and 70 FR 25162, 25299-25304 and associated Technical Support Documents \74\ for full details.
  
  \74\ Supplemental Air Quality Modeling Technical Support Document (TSD) for the Clean Air Interstate Rule (CAIR), May, 2004. http://www.epa.gov/cair/pdfs/saqmtsd.pdf; Demonstration that CAIR
  
  Satisfies the `Better-than-BART' Test as proposed in the Guidelines for Making BART Determinations, EPA Docket Number OAR-2003-0054- YYYY, March 2005. http://www.epa.gov/cair/pdfs/finaltech04.pdf.
  
  Scenarios Examined
  
  The CAIR is applicable to 28 States and the District of Columbia and requires levels of SO2and NOXemissions reductions based on those achievable on a highly cost effective basis from EGUs. BART, on the other hand, is applicable nationwide and covers 25 additional industrial categories, as well as EGUs, of a certain vintage. In our comparison, we sought to determine whether the CAIR cap and trade program for EGUs will achieve greater reasonable progress than would BART for EGUs only. Therefore, the relevant scenarios to examine were (1) SO2and NOXemissions from all EGUs nationwide after the application of BART controls to all BART- eligible EGUs (``nationwide BART''), and (2) SO2and NOXemissions from all EGUs nationwide after the emissions reductions attributable to CAIR in the CAIR region and application of BART controls to all BART-eligible EGUS outside the CAIR region (``CAIR + BART''). The latter scenario reflects the fact that source-by-source BART would remain a federal requirement outside the CAIR region, unless and until it is replaced by some other state or federally required program. Thus, in order to more accurately project CAIR emissions, it is necessary to impose BART controls outside the CAIR region, to account for potential load and emission shifting among EGUs.
  
  In addition to these two scenarios, a third was used--the future base case in the absence of either program. This third scenario was used to ensure that CAIR would not cause degradation from otherwise existing conditions. See section IV.B above for a discussion of why the future baseline is an appropriate comparison point for the first prong of the ``better than BART'' test.
  
  At the SNPR stage, a ``CAIR + BART'' scenario was not available, as the only projections available at that time had been developed for other purposes. Thus, the ``CAIR'' scenario used then, which was based on the Clear Skies proposal, was imperfect for purposes of this analysis in that it assumed SO2reductions on a nationwide basis (rather than in the CAIR region only) and assumed NOX reductions requirements in a slightly different geographic region than covered by the proposed CAIR.
  
  For the CAIR NFR, we redid the emissions projections for both the Nationwide BART and CAIR + BART in the West scenarios. For the former, we increased the number of BART-eligible units included by lowering the assumed threshold for BART applicability from 250 MW capacity for both NOXand SO2to 100 MW for SO2and 25 MW for NOX, and by reviewing the list of potentially BART- eligible EGUs. For the latter scenario, we produced emissions projections based on application of CAIR-level emission reductions in the States proposed for inclusion in the CAIR in the SNPR.
  
  Emission Projections. For the analyses in both the SNPR and NFR, we used the Integrated Planning Model (IPM) to estimate emissions expected from the scenarios described above. Tables 1 and 2 present the results from the SNPR and NFR, respectively.
  
  Table 1.--EGU SO2 and NOX Emissions--as Projected in CAIR SNPR [In thousands of tons per year]
  
  Additional 2015 Modeled reduction from 2015 Base case 2015 ``CAIR'' nationwide e
  
  ``CAIR'' EGU emissions
  
  Bart
  
  (nationwide BART minus ``CAIR'')
  
  Nationwide SO2..........................
  
  9,081
  
  5,260
  
  7,012
  
  1,752 Nationwide NOX..........................
  
  3,950
  
  2,248
  
  2,781
  
  533
  
  Table 2.--EGU SO2 and NOX Emissions--as Projected in CAIR NFR [In thousands of tons per year]
  
  Additional reduction from 2015 Base case 2015 CAIR + BART 2015 Nationwide CAIR + BART EGU emissions
  
  BART
  
  (nationwide BART minus CAIR+BART)
  
  Nationwide SO2..........................
  
  9,084
  
  4,735
  
  7,162
  
  2,427 Nationwide NOX..........................
  
  3,721
  
  1,816
  
  2,454
  
  638
  
  As can be seen in the numbers in the right-most column, CAIR produced far superior emission reductions to nationwide BART, and the superiority of CAIR over BART increased between the SNPR and NFR projections, when the scenarios were corrected to more accurately reflect the anticipated reality in 2015.
  
  Air Quality Modeling Results. The proposed ``better-than-BART'' test provided that if the distribution of emission reductions is substantially the same under the alternative program as under BART, then the demonstration can be made simply by comparing emission reductions. If, however, the distribution is significantly different,
  
  [[Page 39140]]
  
  then visibility modeling is required in order to apply the two pronged test previously described. As noted above, CAIR emission reductions were vastly greater than those under BART. However, because there were some differences in the geographic distribution of reductions on a state-by-state basis, in order to be conservative we conducted air quality modeling and evaluated CAIR under the two pronged test.
  
  Specifically, using the above emissions projections, we completed numerous air quality modeling runs and postprocessing calculations to determine the impacts of emissions and emissions control strategies on visibility in Class I areas. We quantified the impacts of the CAIR and BART controls on visibility impairment by comparing the results of the future-year (2015) base case model runs with the results of the CAIR + BART and nationwide BART control strategy model runs. We quantified visibility impacts on the 20 percent best and 20 percent worst visibility days.
  
  For the SNPR modeling, we used the Regional Modeling System for Aerosols and Deposition (REMSAD) model to calculate these visibility impacts. This modeling used base year meteorology from 1996. Complete year ambient monitoring data, which is necessary to model future improvements in visibility, was available for 1996 from Inter-agency Monitoring of Protected Visual Environments (IMPROVE) monitors located at 44 Class I areas--13 within the CAIR region and 31 outside of it.
  
  For the NFR modeling, we used the Community Multiscale Air Quality (CMAQ) model. The base year meteorology used in the CMAQ modeling was 2001. This later base year enabled us to look at more Class I areas, because there were more IMPROVE monitors which had complete year data for 2001 than there had been in 1996. Specifically, 81 of the 110 IMPROVE sites have complete ambient air quality data for 2001. Moreover, because in some cases a given IMPROVE monitor is designated as representing more than one Class I area, these 81 sites are representative of 116 Class I areas. Twenty nine of the 116 are in the East (east of 100 degrees longitude) and 87 are in the West.
  
  Using the modeling results, we then applied the two prong better than BART test which had been defined in the proposed BART rule. As explained above, under the first prong, visibility must not decline at any Class I area, as determined by comparing the predicted visibility impacts at each affected Class I area under the (CAIR) trading program with future base case visibility conditions. Under the second prong, overall visibility, as measured by the average improvement at all affected Class I areas, must be better under the trading program than under source-specific BART. The future year air quality modeling results were used to make this demonstration.
  
  The visibility impacts of the CAIR + BART scenario were compared to base case 2015 visibility conditions (without CAIR or BART) to determine whether the CAIR resulted in a degradation of visibility at any Class I area. We also compared these visibility impacts with the visibility impacts of nationwide BART implementation, to assess whether the proposed CAIR would result in greater average visibility improvement than nationwide BART.
  
  The CAIR passed the first prong by not causing a degradation of visibility at any Class I area, either in the West or nationally. This was true in both the SNPR and NFR modeling. The visibility projections for each Class I area are presented in the respective TSD's.\75\
  
  \75\ See Footnote [74], Supra.
  
  The overall results are presented in tables 3 and 4 below, representing the SNPR and NFR modeling respectively.
  
  Table 3.--Average Visibility Improvement in 2015 vs. 2015 Base Case (Deciviews) as Modeled Using REMSAD in CAIR SNPR
  
  ``CAIR'' Scenario
  
  Nationwide BART Class I areas
  
  --------------------------------------------------------------- East \76\ National
  
  East
  
  National
  
  20 percent Worst Days...........................
  
  2.0
  
  0.7
  
  1.0
  
  0.4 20 percent Best Days............................
  
  0.7
  
  0.2
  
  0.4
  
  0.1
  
  Table 4.--Average Visibility Improvement in 2015 vs. 2015 Base Case (Deciviews) as Modeled Using CMAQ in CAIR NFR
  
  CAIR + BART in West
  
  Nationwide BART Class I Areas
  
  --------------------------------------------------------------- East \76\ National
  
  East
  
  National
  
  20 percent Worst Days...........................
  
  1.6
  
  0.5
  
  0.7
  
  0.2 20 percent Best Days............................
  
  0.4
  
  0.1
  
  0.2
  
  0.1
  
  As can be see from the tables, although the models produced different absolute values, in both cases CAIR produced significantly greater visibility improvement than nationwide BART. For example, looking at the 20 percent worst days at Eastern Class I areas (the areas most influenced by the CAIR, since it is an eastern program), in both cases the visibility improvements from CAIR were at least twice as great as under nationwide BART (i.e., in the SNPR, 2.0 deciviews compared to 1.0 deciviews improvement, and in the NFR, 1.6 deciviews compared to 0.7 deciviews improvement).
  
  \76\ Eastern Class I areas are those in the CAIR affected states, except areas in west Texas which are considered western and therefore included in the national average, plus those in New England
  
  This historical overview is given in the interest of providing a more complete understanding of the analyses presented at various stages in the CAIR rule making progress. In the end, however, it is the analyses presented in the CAIR NFR on which we are basing our determination that CAIR makes greater reasonable progress towards the national visibility goals than does nationwide BART. Therefore, these NFR results are examined more closely in the ``Final Action'' section below, in light of additional emissions projections we
  
  [[Page 39141]]
  
  have conducted to insure that changes to the CAIR and BART rules made subsequent to the CAIR NFR do not affect that determination. Comments Received and EPA's Responses
  
  Although many comments were received regarding our proposal to determine that CAIR makes greater reasonable progress than BART, nearly all of them related either to the terms of the test itself, or to policy and legal implications of allowing CAIR required reductions to substitute for source-by-source BART. These are addressed in sections B (above ) and D (below) respectively. One commenter asserted, with respect to modeling presented in the SNPR, that the improvement of CAIR compared to source-specific BART is so slight it may be potentially within the margin of error, and therefore insufficient for the better than BART demonstration, or for assuring that no hot spots will occur.
  
  The EPA disagrees that the difference between CAIR and BART in the SNPR visibility projections was not significant. The visibility results presented in the NFR continue to show that the CAIR cap and trade program with BART in the non-CAIR region provides considerably more visibility improvement compared to nationwide BART (for EGUs only). The NFR modeling results show that the average visibility improvement from CAIR on the 20 percent worst days at 29 Eastern Class I areas is 1.6 deciviews (dv) compared to only a 0.7 dv improvement from nationwide BART controls. In the ``better than BART'' TSD we have provided modeling results for 116 individual Class I areas. The modeling shows that CAIR will not create any ``hot spots.'' On the 20 percent worst days, all of the Eastern Class I areas show more visibility improvement under CAIR+BART than under BART alone. In many of the Western Class I areas, nationwide BART and CAIR + BART in the West provide about the same visibility benefits. (This is to be expected, since the CAIR is only applicable in the East.) While the visibility benefits are similar in the West (outside of the CAIR region), they are clearly not similar in the East, where the CAIR is predicted to achieve twice as much visibility improvement compared to BART.
  
  Final action. The CAIR vs. BART comparison presented in the CAIR NFR was developed while both rules were under development and therefore subject to change. Since the emissions projections and air quality modeling presented in the CAIR NFR was completed, several changes were, in fact, made to the CAIR region. In addition, since that time our assumptions regarding the likely maximum BART emission reductions from EGUs also changed. Therefore, we recalculated the emission projections to see if the rule changes could possibly affect the determination that CAIR will achieve greater reasonable progress than BART.
  
  Most significantly, the final CAIR included Arkansas, Delaware, and New Jersey only for purposes of significant contribution to ozone non- attainment by summertime NOXemissions, whereas our modeling had been based on the assumption that these States would be included for contribution to PM2.5non-attainment by SO2 and NOXemissions. The new emission projections are based on the application of CAIR only for ozone in these States.
  
  With respect to the nationwide BART, for SO2the NFR projections assumed the application of a 90 percent control or 0.10 lbs/mmBtu at uncontrolled EGUs greater than 100 MW. In the new projections, the control assumptions were changed to 95 percent or 0.15 lbs/mmbtu, to reflect the presumptive control levels in the final BART guidelines. For NOX, the NFR projections were based on an assumed emission rate of 0.2 lbs/mmBTU at all BART eligible EGUs nationwide. The new projections are based on the assumption of combustion controls on all BART eligible units except cyclones which have SCR, and the operation of all existing SCR and SNCRs annually, instead of just in the ozone season. Finally for both pollutants, the threshold for application of controls was increased to 200 MW, to better reflect the presumptions included in the final BART guidelines.
  
  We used IPM to project 2015 emissions given these new parameters. The results are presented in Table 5 below, which also includes the CAIR NFR projections (as reported in Table 2) for the reader's convenience.
  
  Table 5.--EGU SO2 and NOX Emissions--as Projected in CAIR NFR and as Projected in Subsequent Update (In thousands of tons per year)
  
  Additional reduction from 2015 CAIR + BART 2015 Nationwide CAIR + BART BART
  
  (nationwide BART minus CAIR+BART)
  
  CAIR NFR:
  
  Nationwide SO2........................................
  
  4,735
  
  7,162
  
  2,427
  
  Nationwide NOX........................................
  
  1,816
  
  2,454
  
  638 Updated Projections:
  
  Nationwide SO2........................................
  
  5,042
  
  7,953
  
  2,911
  
  Nationwide NOX........................................
  
  2,000
  
  2,738
  
  738
  
  The updated emissions estimates for both the BART and CAIR with BART in the West scenarios are slightly higher than the NFR emissions estimates, but the difference between the CAIR + BART and nationwide BART scenarios are even larger compared to the NFR determination. For SO2, the updated CAIR + BART achieves about 2.9 million tons more reductions than updated nationwide BART in 2015. For NOX, the updated CAIR + BART policy is projected to result in about 738,000 tons more emissions reductions than the updated BART nationwide policy in 2015. The difference between the updated CAIR + BART and nationwide BART scenarios is now larger by 484,000 tons of SO2reduction (i.e., 2,911,000 - 2,427,000) and 100,000 tons of NOXreduction (i.e. 738,000 - 638,000). Implications of New Emission Projections for the Two-Pronged Test
  
  The first prong of the better than BART test specifies that no degradation of visibility can occur at any Class I area. In order to be sure that Class I areas do not experience a degradation in
  
  [[Page 39142]]
  
  visibility, we examined the updated State by State emissions estimates. Compared to the 2015 base case, in the updated CAIR + BART case, there are no individual Statewide increases in either SO2or NOX(except for a very small 1,000 ton increase in NOXin Connecticut and 2,000 ton increase in SO2 in New Jersey).\77\ That is consistent with the NFR CAIR + BART case in which no degradation was found. Consequently we have determined that no degradation would occur under the updated CAIR + BART emissions scenario.
  
  \77\ The 1,000 ton per year increase in NOXin Connecticut represents approx. 0.003 percent of the total EGU NOXin the 2015 base case and the 2,000 ton per year increase in SO2in New Jersey represents approx. 0.0005 percent of the total EGU SO2. Since the impacts on visibility from EGU SO2and NOXare generally regional in nature, we would expect this small increase to have little or no impact on visibility in any Class I area.
  
  The second prong of the better than BART test specifies a greater average visibility improvement from the CAIR trading program (CAIR + BART). The average visibility improvement from the NFR CAIR + BART case was much greater (on the 20 percent worst visibility days) than the nationwide BART case. In the scenario we modeled for the NFR, the larger visibility improvement from CAIR + BART was achieved by reducing SO2emissions by an additional 2.4 million tons per year compared to nationwide BART and NOXemissions by an additional 638,000 tons per year compared to natiowide BART.
  
  In the updated scenario, the emissions difference between the CAIR + BART and nationwide BART cases are even larger (2.9 million tons of SO2and 738,000 tons of NOX).\78\ The distribution of emission reductions changed somewhat in the new projections--that is, some States saw a larger difference between CAIR and BART, while in other States the difference was smaller. The largest change was in Kentucky, where the new projections showed that emission reductions from CAIR were even greater than those from BART by an additional 200,000 tons per year. Among States where the change between projections went the other direction--that is, showing that BART reductions were closer to CAIR reductions than previously projected-- the greatest changes were in Alabama and Pennsylvania, where the difference between the programs decreased by 46,000 and 45,000 tons, respectively.
  
  \78\ The difference between the updated CAIR + BART and nationwide BART scenarios is larger than the difference between the modeled CAIR + BART and nationwide BART scenarios. The ``difference of the differences'' is 485,000 tons of SO2and 100,000 tons of NOX.
  
  Perhaps more importantly, in the new projections, there are fewer States in which BART reductions are greater than CAIR reductions. In the NFR projections, there were 12 States \79\ where nationwide BART SO2reductions were greater than CAIR + BART reductions.\80\ In those 12 States, BART emissions achieved approx. 686,000 more tons of SO2reduction compared to CAIR + BART. In the rest of the States, CAIR + BART achieved an additional 3.1 million tons per year of SO2reduction compared to BART. All told, the modeling showed that visibility improvement was greater under the CAIR than under BART on an overall average basis, both at eastern Class I areas and at all Class I areas nationally. In the new projections, CAIR + BART achieved an additional 3.4 million tons per year of SO2 reduction compared to BART in 39 of the 48 States. In the remaining 9 States \81\ BART achieved approx. 472,000 more tons of SO2 reduction compared to CAIR + BART in the west.\82\
  
  \79\ California, Delaware, Florida, Georgia, Iowa, Louisiana, Michigan, Mississippi, Missouri, North Carolina, Texas, and Wisconsin.
  
  \80\ There were also four States where BART NOX emissions reductions were slightly higher than CAIR + BART (a total of 4,000 tons per year). Those States are Connecticut, Delaware, New Jersey, and Oklahoma.
  
  \81\ Alabama, Louisiana, Michigan, Mississippi, Missouri, New Jersey, North Carolina, Texas, Wisconsin.
  
  \82\ We performed a similar analysis using projections including the Clean Air Mercury Rule, CAMR, which was promulgated after the CAIR NFR. The CAMR emission projections show slight additional emission reductions of SO2and NOXas compared to the projections CAIR + BART without CAMR, and are nearly identical in terms of geographic distribution. Therefore CAIR + BART + CAMR, like CAIR + BART, passes the two-pronged test for demonstrating greater reasonable progress than BART. This is discussed in more detail in the TSD accompanying today's action.
  
  Due to the fact that the new projections show that the difference between CAIR and BART reductions is even greater than previously estimated, and the visibility improvements due to CAIR + BART were previously modeled to be much larger than BART, we can state with a high degree of confidence that the updated CAIR + BART scenario passes the second prong of the better than BART test.
2. Revision to Regional Haze Rule To Allow CAIR States To Treat CAIR as a BART-Substitute for EGUs
  
  In the SNPR, we proposed that States which adopt the CAIR cap and trade program for SO2and NOXwould be allowed to treat the participation of EGUs in this program as a substitute for the application of BART controls for these pollutants at affected EGUs. To implement this, we proposed an amendment to the Regional Haze Rule which would add a subpart 40 CFR 51.308(e)to read as follows:
  
  A State that opts to participate in the Clean Air Interstate Rule cap-and-trade program under part 96 AAA-EEE need not require affected BART-eligible EGUs to install, operate, and maintain BART. A State that chooses this option may also include provisions for a geographic enhancement to the program to address the requirement under Sec. 51.302(c) related to BART for reasonably attributable impairment from the pollutants covered by the CAIR cap and trade program.\83\
  
  \83\ A geographic enhancement is a method, procedure, or process to allow a broad regional strategy, such as the CAIR cap & trade program, to accommodate BART for reasonably attributable impairment. For example, it could consist of a methodology for adjusting allowance allocations at a source which is required to install BART controls.
  
  We proposed that this would be codified at 40 CFR 51.308(e)(3); however, that section now incorporates the ``better than BART'' test as discussed above. In today's action, as described below we are finalizing this provision of the rule, where it will be codified as section 308(e)(4).
  
  The EPA's authority to treat emissions reductions required by the CAIR as satisfying BART was not affected by CEED. As noted, the D.C. Circuit in CEED upheld the proposition that EPA can approve implementation plans which rely on alternative strategies to BART, as long as greater reasonable progress is achieved. CEED, 398 F.3d at 660. Moreover, the CAIR program is not infected in any way with the ``group BART'' methodology held invalid by the court. That is because CAIR emission reductions levels were not based on the invalid ``group-BART'' approach or any other assumptions regarding BART, but were developed for other reasons. Specifically, the CAIR was developed to assist with attainment of the NAAQS for PM2.5and ozone. Had EPA not performed the comparison of CAIR to BART for visibility progress purposes, the CAIR emission reduction requirements would remain unchanged. Therefore, EPA is not imposing an invalid BART requirement on States, but rather allowing States, at their option, to utilize the CAIR cap and trade program as a means to satisfy BART for affected EGUs.
  
  We received numerous comments on this proposal, which are summarized along with our responses in the CAIR NFR preamble at 70 FR 25300-25302 and in the Response to Comment document. To summarize our responses to some of the most important comments:
  
  [[Page 39143]]
  
  (1) We note that we are not constraining the discretion of States to determine which sources are subject to BART and to make BART determinations. CAIR-affected States are not required to accept our determination that CAIR may substitute for BART. Under the amended rule, States simply have the option of accepting this determination.
  
  (2) The EPA does not believe that anything in the CAA or relevant case law prohibits a State from considering emissions reductions required to meet other CAA requirements when determining whether source by source BART controls are necessary to make reasonable progress. Whatever the origin of the emission reduction requirement, the relevant question for BART purposes is whether the alternative program makes greater reasonable progress. As discussed above, EPA has determined that CAIR does so with respect to SO2and NOX from EGUs in the CAIR region.
  
  Moreover, the fact that BART and CAIR originate from different provisions of the CAA does not mean that CAIR and BART emissions reductions would be additive if BART-eligible EGUs in the CAIR program were required to install and operate BART. Such source specific control requirements would simply result in a redistribution of emission reductions, as other EGUs could buy the excess allowances generated by the installation of controls at BART units. The net result would be the same level of emission reductions, but at a higher total cost, because the ability of the market to find the most cost effective emission reductions would be constrained.
  
  (3) Although regional haze rule section 308(e)(2) is not directly applicable, as the CAIR is covered by the special provision newly codified at section 308(e)(4), this determination is consistent with the policy contained in section 308(e)(2) requiring in-lieu of BART programs be based on emissions reductions ``surplus to reductions resulting from measures adopted to meet requirements as of the baseline date of the SIP.'' The baseline date for regional haze SIPs is 2002;\84\ therefore CAIR reductions are surplus to requirements as of that year.
  
  \84\ See Memorandum from Lydia Wegman and Peter Tsirigotis, 2002 Base Year Emission Inventory SIP Planning: 8-hr Ozone, PM2.5, and Regional Haze Programs, November 8, 2002. http://www.epa.gov/ttn/oarpg/t1/ memoranda/2002bye--gm.pdf.
  
  (4) We agree with commenters that it was premature to make a final determination whether CAIR makes greater reasonable progress than BART in the final CAIR because at that time the BART guidelines and the criteria for making such determinations had not been finalized. In today's action, both those rule makings are complete and therefore such a determination is ripe.
  
  (5) We disagree with commenters who thought that CAIR should be considered ``better than BART'' regardless of whether a State participates in the cap and trade program. Our demonstration that CAIR makes greater reasonable progress than BART is based only on an examination of emissions reductions from EGUs under both programs. The CAIR emissions projections and modeling assumes that EGU emissions will be capped at the levels specified in the CAIR. Therefore, States that choose to meet their CAIR emission reduction requirements in a manner other than through the participation of EGUs in the CAIR cap and trade program would have to develop an appropriate demonstration that the measures they employ make greater reasonable progress than would BART for any affected source categories, if the State wanted its CAIR- required reductions to substitute for source-by-source BART.
  
  (6) We disagree with commenters who asserted that CAIR should satisfy BART for States that are subject to CAIR only for ozone season NOX. We explained in the final CAIR preamble that a State subject to CAIR for NOXpurposes only would have to make a supplementary demonstration that BART has been satisfied for SO2, as well as for NOXon an annual basis. We wish to clarify here that a State which is only subject to CAIR for NOX, but which also chooses to participate in the CAIR trading program for both SO2and NOX, may consider BART to be satisfied for both SO2and NOXfrom EGUs. Because we modeled these States as controlling for both SO2and NOXin the CAIR NFR, our better than BART demonstration presented in that action would be valid in that scenario. Conversely, if such States choose to participate only in the ozone season NOXtrading program, the updated projections presented in today's action demonstrate that BART would be satisfied for NOX, but such states would still need to address BART for SO2emissions from EGUs.
  
  (7) We noted in the final CAIR preamble that although we believe it is unlikely that a State or FLM will find it necessary to certify reasonably attributable visibility impairment at any Class I area, as a legal matter that possibility exists. That is, the determination that CAIR makes greater reasonable progress than BART is made in the context of BART for regional haze under CAA 169B, and does not preclude a finding of reasonably attributable impairment under CAA 169A. The CAIR cap and trade program does not include geographic enhancements to accommodate the situation where BART is required based on reasonable attribution at a source which participates in the trading program, but States retain the discretion to include such enhancements in their SIPs.
  
  (8) Our determination that CAIR makes greater reasonable progress than BART for EGUs is not a determination that CAIR satisfies all reasonable progress requirements in CAIR affected States. Each State, whether in the CAIR region or not, is required to set reasonable progress goals for each Class I area within the State as required in regional haze rule section 308(d)(1), and to develop long term strategies, considering all anthropogenic sources of visibility impairing pollutants, as required by section 308(d)(3).
  
  In setting the reasonable progress goals, the State is to consider the amount of visibility improvement needed to achieve a uniform rate of progress towards natural background conditions in the year 2064. (This uniform rate of progress is sometimes referred to as the default glide-path). The State is also to consider the statutory reasonable progress factors contained in CAA section 169A(g)(1).\85\
  
  \85\ Similar to the BART factors, the reasonable progress factors are: the cost of compliance, the time necessary for compliance, the energy and nonair quality environmental impacts of compliance, and the remaining useful life of any existing sources subject to such requirements.
  
  In doing so, we anticipate that States will take into account the degree to which CAIR emissions reductions are projected to bring visibility conditions at its Class I areas in line with the default glide path. In some States, the improvements expected from CAIR, combined with the application of the reasonable progress factors to other source sectors, may result in a determination that few additional emissions reductions are reasonable for the first long term strategy period. Nonetheless, each State is required to set its reasonable progress goals as provided by the regional haze rule and cannot assume that CAIR will satisfy all of its visibility-related obligations.
Statutory and Executive Order Reviews
1. Executive Order 12866: Regulatory Planning and Review
  
  Under Executive Order 12866 (58 FR 51735, October 4, 1993), EPA must determine whether the regulatory action is ``significant'' and, therefore, subject to Office of Management and Budget
  
  [[Page 39144]]
  
  (OMB) review and the requirements of the Executive Order. The Order defines ``significant regulatory action'' as one that is likely to result in a rule that may:
  
  (1) Have an annual effect on the economy of $100 million or more or adversely affect in a material way the economy, a sector of the economy, productivity, competition, jobs, the environment, public health or safety, or State, local, or Tribal governments or communities;
  
  (2) Create a serious inconsistency or otherwise interfere with an action taken or planned by another agency;
  
  (3) Materially alter the budgetary impacts of entitlements, grants, user fees, or loan programs or the rights and obligations of recipients thereof; or
  
  (4) Raise novel legal or policy issues arising out of legal mandates, the President's priorities, or the principles set forth in the Executive Order.
  
  Pursuant to the terms of Executive Order 12866, it has been determined that this rule is a ``significant regulatory action,'' thus EPA has submitted this rule to OMB for review. The drafts of the rules submitted to OMB, the documents accompanying such drafts, written comments thereon, written responses by EPA, and identification of the changes made in response to OMB suggestions or recommendations are available for public inspection at EPA's Air and Radiation Docket and Information Center (Docket Number OAR-2002-0076). The EPA has prepared the document entitled ``Regulatory Impact Analysis of the Final Clean Visibility Interstate Rule or Guidelines for Best Available Retrofit Technology Determinations Under the Regional Haze Regulations'' (RIA) to address the requirements of this executive order. 1. What Economic Analyses Were Conducted for the Rulemaking?
  
  The analyses conducted for this final rule provide several important analyses of impacts on public welfare. These include an analysis of the social benefits, social costs, and net benefits of three possible regulatory scenarios that States may follow to implement the BART rule and guidelines. The economic analyses also address issues involving requirements of the Paperwork Reduction Act (PRA), potential small business impacts, unfunded mandates (including impacts for Tribal governments), environmental justice, children's health, energy impacts, and other statutory and executive order requirements. 2. What Are the Benefits and Costs of This Rule?
  
  The benefit-cost analysis shows that substantial net economic benefits to society are likely to be achieved due to reductions in emissions resulting from this rule. The results detailed below show that this rule would be beneficial to society, with annual net benefits (benefits less costs) ranging from approximately $1.9 to $12.0 billion in 2015. These alternative net benefits estimates reflect differing assumptions about State actions taken to implement BART and about the social discount rate used to estimate the annual value of the benefits and costs of the rule. All amounts are reflected in 1999 dollars. The range of benefits and costs reported for the BART represent estimates of EPA's assessment of State actions that will likely be taken to comply with the BART rule and guidelines. a. Control Scenarios
  
  Today's rule sets forth presumptive requirements for States to require EGUs to reduce SO2and NOXemissions for units greater than 200 megawatts (MW) in capacity at plants greater than 750 MW in capacity that significantly contribute to visibility impairment in Federal Class I areas (national parks). The analysis conducted in the RIA presents alternative control scenarios of possible additional controls for EGUs located at plants less than 750 MW in capacity. The EPA also calculated the amount of SO2and NOXemissions reductions for several illustrative scenarios that reflect alternative State actions regulating industries with non- EGU sources. The analyses conducted include three regulatory alternative scenarios that States may choose to follow to comply with BART. The alternatives include three scenarios of increasing stringency--Scenario 1, Scenario 2, and Scenario 3. A brief discussion of the these alternatives for the EGUs and all other sources follows. More details of the alternative control scenarios and associated control costs are discussed in the RIA. i. Electric Generating Units
  
  In the revised BART guidelines, we have included presumptive control levels for SO2and NOXemissions from coal-fired electric generating units greater than 200 megawatts (MW) in capacity at plants greater than 750 MW in capacity. Given the similarities of these units to other BART-eligible coal-fired units greater than 200 MW at plants 750 MW or less, EPA's guidance suggests that States control such units at similar levels for BART. The guidelines would require 750 MW power plants to meet specific control levels of either 95 percent control or controls of 0.15 lbs/MMBtu, for each EGU greater than 200 MW, unless the State determines that an alternative control level is justified based on a careful consideration of the statutory factors.\86\ Thus, for example, if the source convincingly demonstrates unique circumstances affecting its ability to cost-effectively reduce its emissions, the State may take that into account in determining whether the presumptive levels of control are appropriate for the facility. For an EGU greater than 200 MW in size, but located at a power plant smaller than 750 MW in size, States may also find that such controls are cost-effective when taking into consideration the costs of compliance in the BART analysis in applying the five factor test for the BART determination. In our analysis we have assumed that no additional controls will occur where units have existing scrubbers and that no controls will occur for oil-fired units. While these levels may represent current control capabilities, we expect that scrubber technology will continue to improve and control costs will continue to decline.
  
  \86\ These levels are commonly achievable by flue gas desulfurization controls (``scrubbers'').
  
  For NOX, for those large EGUs that have already installed selective catalytic reduction (SCR) or selective non- catalytic reduction (SNCR) during the ozone season, States should require the same controls for BART. However, those controls should be required to operate year-round for BART. For sources currently using SCR or SNCR for part of the year, states should presume that the use of those same controls year-round is highly cost-effective. For other sources, the guidelines establish presumptive emission levels that vary depending largely upon boiler type and fuel burned. For coal-fired cyclone units with a size greater than 200 MW, our analysis assumes these units will install SCR. For all other coal-fired units, our analysis assumed these units will install current combustion control technology. In addition, we assume no additional controls for oil and/ or gas-fired steam units.
  
  We present alternative regulatory scenarios. Scenario 2 represents our application of the presumptive limits described above to all BART eligibility EGUs greater than 200 MW. For Scenario 1, we assume that only 200 MW BART-eligible EGUs located at facilities above 750 MW capacity will comply with the SO2requirements and NOXcontrols. In this scenario, no
  
  [[Page 39145]]
  
  facilities less than 750 MW capacity are assumed to install BART controls. For Scenario 1, we assume that units with existing SCRs will operate those SCR units year round annually. In contrast in Scenario 3, we analyzed SO2controls equivalent to 95 percent reductions or 0.1 lbs per MMBtu on all previously uncontrolled units. NOXcontrols for this most stringent scenario presume SCRs will be installed on all units greater than 100 MW capacity and combustion controls will be installed on units greater than 25 MW but less than 100 MW capacity. The EPA analyzed the costs of each BART scenario using the Integrated Planning Model (IPM). The EPA has used this model extensively in past rulemakings to analyze the impacts of regulations on the power sector.
  
  The analysis presented assumes that BART-eligible EGUs affected by the Clean Air Interstate Rule (70 FR 25162) have met the requirements of this rule. Thus, no additional controls for EGUs beyond CAIR are anticipated or modeled for the 28 State plus District of Columbia CAIR region. In addition, we are assuming no additional SO2 controls for sources located in States of Arizona, Utah, Oregon, Wyoming, and New Mexico or Tribal lands located in these States due to agreements made with the Western Regional Air Partnership (WRAP). ii. Sources Other Than Electric Generating Units
  
  As previously discussed there are 25 source categories potentially subject to BART in addition to EGUs (referred to as non-EGU source categories) as defined by the CAA. The EPA evaluated a set of SO2and NOXemission control technologies available for these source categories and estimated the associated costs of control using AirControlNET. The control scenarios evaluated reflect control measure cost caps of up to $1,000 per ton (Scenario 1), $4,000 per ton (Scenario 2), and $10,000 per ton (Scenario 3). The EPA also conducted a cost analysis for control costs of up to $2,000 per ton and $3,000 per ton, and the results of this analysis are presented in the RIA. The analysis consists of applying SO2and NOXcontrols to each non-EGU source category up to the specified cost per ton ``cap'' in each scenario. These cost per ton caps are specified in average cost terms. As control stringency is increased, the marginal costs are also estimated for each non-EGU source category. The scenarios examined are based on the costs of technologies such as scrubbers for SO2control, and varying types of technologies for NOXcontrol. Scrubbers are the most common type of SO2control for most non-EGU sources for each scenario, while combustion controls such as low NOX burners (LNB) and post-combustion controls such as selective noncatalytic reduction (SNCR) and selective catalytic reduction (SCR) are commonly applicable to most of the non-EGU source categories. Combustion controls are commonly applied as part of Scenario 1, while SNCR and SCR are more commonly applied either by themselves or in combination with combustion controls as part of Scenarios 2 and 3. Analyses are not available for 8 of the 25 non-EGU source categories, because there are no available control measures for these sources or there are no sources in these categories included in the non-EGU emissions data utilized in these analyses. All of these results are estimated using a nationwide database of BART-eligible non-EGU sources that is based on information collected from Regional Planning Organizations (RPOs) in the fall of 2004. b. Baseline and Year of Analysis
  
  The final rule sets forth the guidelines for States and Tribes for meeting the BART requirements under the CAA and the Regional Haze Rule. The Agency considered all promulgated CAA requirements and known State actions in the baseline used to develop the estimates of benefits and costs for this rule including the recently promulgated Clean Air Interstate Rule (70 FR 25162) and the proposal to include New Jersey and Delaware in the final CAIR region for fine particulate matter (70 FR 25408). However, EPA did not include within the baseline the actions States may take to implement the ozone and PM2.5NAAQS standards nor the recently promulgated Clean Air Mercury Rule. No additional SO2controls were assumed for any EGUs within the five WRAP States of Utah, Arizona, Wyoming, Oregon or New Mexico that have existing agreements to achieve reduction goals.
  
  In the analysis, the controls and reductions are assumed to be required in 2015, a date that is generally consistent with the expected timing of the rule. States must submit SIPs relevant to the BART requirements in January 2008. After approval of the SIP, there is a 5 year compliance date. Thus, controls are likely to be installed and in operation by the end of 2013 or the beginning of 2014 to comply with the rule. In addition, EPA had existing inventories, modeling, and base case runs for 2015 to use for the analysis. The year 2015 is used in this analysis. All estimates presented in this report represent annualized estimates of the benefits and costs of BART in 2015 rather than the net present value of a stream of benefits and costs in these particular years of analysis. c. Cost Analysis and Economic Impacts
  
  For the affected region, the projected annual private incremental costs of BART to the power industry (EGU source category) range from $253 to $896 million in 2015 depending upon the scenario evaluated. These costs represent the private compliance cost to the electric generating industry of reducing NOXand SO2 emissions that EPA believes States may require to comply with BART.
  
  In estimating the net benefits of regulation, the appropriate cost measure is ``social costs.'' Social costs represent the welfare costs of the rule to society. These costs do not consider transfer payments (such as taxes) that are simply redistributions of wealth. The social costs of this rule for the EGU sector only are estimated to range from approximately $119 to $567 million in 2015 assuming a 3 percent discount rate. These EGU sector costs become $141 to $688 million in 2015 assuming a 7 percent discount rate.
  
  Overall, the impacts of the BART are modest, particularly in light of the large benefits we expect. Retail electricity prices are projected to increase roughly 0.1 percent with BART in the 2015 timeframe under Scenario 2. Coal-fired generation, as well as coal production and natural gas-fired generation are projected to remain essentially unchanged as a result of this rule. It is also not expected that BART will change the composition of new generation built to meet growth in electricity demand. BART is also not expected to impact coal or natural gas prices.
  
  For today's rule, EPA analyzed the costs for the EGU source category using the Integrated Planning Model (IPM). The IPM is a dynamic linear programming model that can be used to examine the economic impacts of air pollution control policies for SO2 and NOXthroughout the contiguous U.S. for the entire power system. Documentation for IPM can be found in the docket for this rulemaking or at http://www.epa.gov/airmarkets/epa-ipm.
  
  The EPA also conducted an analysis of State actions in requiring emission controls for BART eligible sources in the non-EGU source categories. For the nation, the projected annual private incremental costs range from $150 million to $2.24 billion for industries with affected non-EGU sources. This cost range results from different assumptions about possible actions
  
  [[Page 39146]]
  
  States may take to comply with BART and alternative discount rates of 3 and 7 percent. The non-EGU private incremental control cost estimates are assumed to approximate the social costs of the rule for the non-EGU sector. The EPA analyzed the costs to non-EGUs sources using AirControlNET. The AirControlNET is a software tool that can be used to estimate the private costs and emission reductions of air pollution control policies for SO2, NOX, and other criteria pollutants throughout the contiguous U.S. for all manufacturing industries and many other industries. Documentation for AirControlNET can be found in the docket for this rulemaking or at http://www.epa.gov/ttn/ecas/AirControlNET.htm .
  
  In summary, the EPA estimates that the annual social costs of this rule for the EGU and non-EGU source categories range from approximately $0.3 to $2.9 billion annually, based on alternative scenarios of State actions in response to the BART rule and guidelines assuming 3 or 7 percent discount rates. Estimates are reflected in 1999 dollars. d. Human Health Benefit Analysis
  
  Our analysis of the health and welfare benefits associated with this rule are presented in this section. Briefly, the analysis projects major benefits from implementation of the rule in 2015. As described below, thousands of deaths and other serious health effects would be prevented. We are able to monetize annual benefits ranging from approximately $2.2 to $14.3 billion in 2015. This range reflects different assumptions about States actions in response to the BART rule and the applicable discount rate (3 percent or 7 percent).
  
  Table IV-1 presents the primary estimates of reduced incidence of PM- and visibility-related health effects for 2015 for the regulatory control strategy the EPA expects States may follow to comply with BART. In 2015 for Scenario 2, we estimate that PM-related annual benefits include approximately 1,600 fewer premature fatalities, 890 fewer cases of chronic bronchitis, 2,200 fewer non-fatal heart attacks, 2300 fewer hospitalizations (for respiratory and cardiovascular disease combined-- admissions and emergency room visits) and result in significant reductions in days of restricted activity due to respiratory illness (with an estimate of one million fewer cases) and approximately 170,000 fewer work-loss days. We also estimate substantial health improvements for children from reduced upper and lower respiratory illness, acute bronchitis, and asthma attacks.
  
  Ozone health-related benefits are expected to occur during the summer ozone season (usually ranging from May to September in the Eastern U.S.). Since we did not conduct ozone modeling for this rulemaking, we are unable to quantify or monetize the ozone related benefits that will likely result from BART.
  
  Table IV-2 presents the estimated monetary value of reductions in the incidence of health and welfare effects. Annual PM-related health benefits and visibility benefits are estimated to range from approximately $2.2 to $14.3 billion annually. This range of estimates reflects different scenarios about States actions in response to the BART rule and the applicable discount rate (3 percent or 7 percent). Estimated annual visibility benefits in southeastern and southwestern Class I areas range from approximately $80 million to $420 million annually in 2015. All monetized estimates are stated in 1999$. These estimates account for growth in real gross domestic product (GDP) per capita between the present and 2015. As the table indicates, total benefits are driven primarily by the reduction in premature fatalities each year. Reductions in premature mortality account for over 90 percent of total benefits.
  
  Table IV-3 presents the total monetized net benefits for 2015. This table also indicates with a ``B'' those additional health and environmental benefits of the rule that we were unable to quantify or monetize. These effects are additive to the estimate of total benefits. A listing of the benefit categories that could not be quantified or monetized in our benefit estimates are provided in Table IV-4. We are not able to estimate the magnitude of these unquantified and unmonetized benefits. While EPA believes there is considerable value to the public for the PM-related benefit categories that could not be monetized, we believe these benefits may be small relative to those categories we were able to quantify and monetize. In contrast, EPA believes the monetary value of the ozone-related premature mortality benefits could be substantial, but we were unable to estimate the benefits for this rulemaking. e. Quantified and Monetized Welfare Benefits
  
  Only a subset of the expected visibility benefits--those for Class I areas in the southeastern and southwestern U.S. are included in the monetary benefits estimates we project for this rule. We believe the benefits associated with these non-health benefit categories are likely significant. For example, we are able to quantify significant visibility improvements in Class I areas in the Northeast and Midwest, but are unable at present to place a monetary value on these improvements. Similarly, we anticipate improvement in visibility in residential areas where people live, work and recreate in the nation for which we are currently unable to monetize benefits. For the Class I areas in the southeastern and southwestern U.S., we estimate annual benefits ranging from $80 to $420 million beginning in 2015 for visibility improvements. The value of visibility benefits in areas where we were unable to monetize benefits could also be substantial.
  
  Table IV-1.--Clean Air Visibility Rule: Estimated Reduction in Incidence of Adverse Health Effects in 2015 \a,b\
  
  Incidence reduction Health Effect
  
  ----------------------------------------------------- Scenario 1
  
  Scenario 2
  
  Scenario 3
  
  PM-Related Endpoints:
  
  Premature mortality \c\ Adult, age 30 and over............................
  
  400
  
  1,600
  
  2,300 Infant, age 18)
  
  120
  
  450
  
  640 \e\..................................................
  
  Emergency room visits for asthma (age 18 years and
  
  370
  
  1,300
  
  1,800 younger).............................................
  
  Acute bronchitis (children, age 8-12).................
  
  550
  
  2,100
  
  3,000
  
  [[Page 39147]]
  
  Lower respiratory symptoms (children, age 7-14).......
  
  6,600
  
  25,000
  
  36,000
  
  Upper respiratory symptoms (asthmatic children, age 9-
  
  5,000
  
  19,000
  
  27,000 18)..................................................
  
  Asthma exacerbation (asthmatic children, age 6-18)....
  
  8,100
  
  31,000
  
  44,000
  
  Work loss days (adults, age 18-65)....................
  
  44,000
  
  170,000
  
  240,000
  
  Minor restricted-activity days (MRADs) (adult age, 18-
  
  260,000
  
  1,000,000
  
  1,400,000 65)..................................................
  
  \a\ Incidences are rounded to two significant digits. These estimates represent benefits from BART nationwide. The modeling used to derive these incidence estimates assumes the final CAIR program in the baseline including the CAIR promulgated rule and the proposal to include SO2 and annual NOX controls for New Jersey and Delaware. Modeling used to develop these estimates assumes annual SO2 and NOX controls for Arkansas for CAIR resulting in a slight understatement of the reported benefits and costs for BART. The recently promulgated CAMR has not been considered in the baseline for BART. \b\ Ozone benefits are expected for BART, but are not estimated for this analysis. \c\ Adult premature mortality based upon studies by Pope et al., 2002. Infant premature mortality is based upon studies by Woodruff, Grillo, and Schoendorf, 1997. \d\ Respiratory hospital admissions for PM include admissions for chronic obstructive pulmonary disease (COPD), pneumonia, and asthma. \e\ Cardiovascular hospital admissions for PM include total cardiovascular and subcategories for ischemic heart disease, dysrhythmias, and heart failure.
  
  Table IV-2. Estimated Monetary Value of Reductions in Incidence of Health and Welfare Effects for the Clean Air Visibility Rule in 2015 [In millions of 1999$] a,b
  
  Scenario 1
  
  Scenario 2
  
  Scenario 3
  
  Health Effects:
  
  Premature mortality c,d Adult >30 years 3 percent discount rate....................
  
  $2,330
  
  $9,180
  
  $13,000 7 percent discount rate....................
  
  1,960
  
  7,730
  
  10,900 Infant Premature mortality b. Chronic respiratory damage. Premature aging of the lungs. Nonasthma respiratory emergency room visits. Increased exposure to Uvb. Hospital Admissions : respiratory. Emergency room visits for asthma. Minor restricted activity days. School loss days. Asthma attacks. Cardiovascular emergency room visits. Acute respiratory symptoms. Ozone--Welfare.................... Yields for: --Commercial forests, --Fruits and vegetables, and --Commercial and noncommercial crops. Damage to urban ornamental plants. Recreational demand from damaged forest aesthetics. Ecosystem functions. Increased exposure to UVb. PM--Health c...................... Premature mortality: short- term exposuresd. Low birth weight. Pulmonary function. Chronic respiratory diseases other than chronic bronchitis. Nonasthma respiratory emergency room visits. Exposure to UVb (+/-) e. PM--Welfare....................... Visibility in many Class I areas. Residential and recreational visibility in non- Class I areas. Soiling and materials damage. Ecosystem functions. Exposure to UVb (+/-)e. Nitrogen and Sulfate Deposition-- Commercial forests due to Welfare.
  
  acidic sulfate and nitrate deposition. Commercial freshwater fishing due to acidic deposition. Recreation in terrestrial ecosystems due to acidic deposition. Existence values for currently healthy ecosystems.
  
  [[Page 39149]]
  
  Commercial fishing, agriculture, and forests due to nitrogen deposition. Recreation in estuarine ecosystems due to nitrogen deposition. Ecosystem functions. Passive fertilization due to nitrogen deposition. Mercury Health g.................. Incidence of neurological disorders. Incidence of learning disabilities. Incidence of developmental delays. Potential reproductive effectsf. Potential cardiovascular effectsf, including: --Altered blood pressure regulation f --Increased heart rate variability f --Incidence of myocardial infarction f Mercury Deposition Welfare g...... Impacts on birds and mammals (e.g., reproductive effects). Impacts to commercial, subsistence, and recreational fishing.
  
  a In addition to primary economic endpoints, there are a number of biological responses that have been associated with ozone health effects including increased airway responsiveness to stimuli, inflammation in the lung, acute inflammation and respiratory cell damage, and increased susceptibility to respiratory infection. The public health impact of these biological responses may be partly represented by our quantified endpoints. b Premature mortality associated with ozone is not currently included in the primary analysis. Recent evidence suggests that short-term exposures to ozone may have a significant effect on daily mortality rates, independent of exposure to PM. EPA is currently conducting a series of meta-analyses of the ozone mortality epidemiology literature. EPA will consider including ozone mortality in primary benefits analyses once a peer-reviewed methodology is available. c In addition to primary economic endpoints, there are a number of biological responses that have been associated with PM health effects including morphological changes and altered host defense mechanisms. The public health impact of these biological responses may be partly represented by our quantified endpoints. d While some of the effects of short term exposures are likely to be captured in the estimates, there may be premature mortality due to short term exposure to PM not captured in the cohort study upon which the primary analysis is based. e May result in benefits or disbenefits. See discussion in Section 5.3.4 for more details. f These are potential effects as the literature is insufficient. g Mercury emission reductions are not anticipated for BART for the EGU source category due to the cap-and-trade program promulgated for the Clean Air Mercury Rule (March 2005); however, the geographic location of mercury reductions may change as a result of this rule. EPA believes any such effects for these sources would be minimal. Mercury reductions are expected for the non-EGU source categories. The mercury reduction for BART from the non-EGU source categories is expected to be small in comparison to reductions resulting from the recently promulgated Clean Air Interstate Rule and the Clean Air Mercury Rule (March 2005).
  1. How Do the Benefits Compare to the Costs of This Final Rule?
    
    In estimating the net benefits of regulation, the appropriate cost measure is ``social costs.'' Social costs represent the welfare costs of the rule to society. The social costs of this rule for the EGU and non-EGU sector sources are estimated to range from approximately $0.3 to $2.9 billion in 2015. This range depends upon the control scenario assumed and applicable discount rates of 3 percent and 7 percent. The net benefits (social benefits minus social costs) of the rule range from approximately $1.9 + B billion or $12.0 + B billion depending upon the scenario evaluated and the applicable discount rate (3 and 7 percent) annually in 2015. Implementation of the rule is expected to provide society with a substantial net gain in social welfare based on economic efficiency criteria.
    
    There is uncertainty surrounding the actions States are likely to take to comply with the BART guidelines. States will determine BART- eligible sources based upon CAA criteria, determine those BART-eligible sources reasonably anticipated to cause or contribute to visibility impairment in Class I areas and then apply a 5 factor test for BART determinations. The range of estimated benefits, costs, and resulting net benefits for BART reflects the uncertainty concerning States responses to BART and represents EPA's best estimates of the benefit- cost outcomes of alternative compliance scenarios.
    
    The annualized cost of BART, as quantified here, is EPA's best assessment of the cost of actions States are likely to take to comply with the rule. The EGU portion of these costs are generated from rigorous economic modeling of changes in the power sector due to the BART rule and guidelines. This type of analysis using IPM has undergone peer review and been upheld in Federal courts. The direct cost includes, but is not limited to, capital investments in pollution controls, operating expenses of the pollution controls, investments in new generating sources, and additional fuel expenditures. The EPA believes that these costs reflect, as closely as possible, the additional costs of the BART rule and guidelines to industry. However, there may exist certain costs that EPA has not quantified in these estimates. These costs may include costs of transitioning to the BART, such as the costs associated with the retirement of smaller or less efficient EGUs, employment shifts as workers are retrained at the same company or re-employed elsewhere in the economy. Costs may be understated since an optimization model was employed that assumes cost minimization, and the regulated community may not react in the same manner to comply with the rule. Although EPA has not quantified these potential additional costs, the Agency believes that they are small compared to the quantified costs of the program on the power sector. The annualized cost estimates presented are the best and most accurate based upon available information.
    
    The non-EGU portion of these costs are generated from extensive cost modeling based on applying illustrative regulatory scenarios to the non-EGU source categories. These costs represent potential impacts to non-EGU sources from State-imposed BART requirements. The direct cost includes, but is not limited to, capital investments in pollution controls, operating and maintenance expenses of the pollution controls, and additional fuel expenditures. The EPA believes that these costs reflect, as closely as possible, the potential additional costs of the BART rule and guidelines to industries with non-EGU sources. However, there
    
    [[Page 39150]]
    
    may exist certain costs that EPA has not quantified in these estimates. These costs may include costs of transitioning to the BART rule and guidelines, such as the costs associated with the retirement of smaller or less efficient non-EGUs, employment shifts as workers are retrained at the same company or re-employed elsewhere in the economy, and costs associated with applying both SO2and NOX controls at one facility at the same time. Costs may be understated since the non-EGU cost modeling presumed a least-cost approach, and the potentially regulated community may not react in the same manner to comply with the rules. Although EPA has not quantified these costs, the Agency believes that they are small compared to the quantified costs of the program on industries with potentially affected non-EGU sources. The annualized cost estimates presented are the best and most accurate based upon available information. In a separate analysis, EPA estimates the indirect costs and impacts of higher electricity prices and costs applicable to the non-EGU sectors on the entire economy [see Regulatory Impact Analysis for the Final Clean Visibility Rule, Appendix A (June 2005)].
    
    The costs presented here are EPA's best estimate of the direct private costs of the BART rule and guidelines. For purposes of benefit- cost analysis of this rule, EPA has also estimated the additional costs of BART using alternate discount rates for calculating the social costs, parallel to the range of discount rates used in the estimates of the benefits of BART (3 percent and 7 percent). Using these alternate discount rates, the social costs of BART range from $0.3 to $2.9 billion in 2015. (Note the portion of these annual costs associated with non-EGU sources represents incremental private cost estimates that are used as a proxy for the social costs of the rule.)
    
    Every benefit-cost analysis examining the potential effects of a change in environmental protection requirements is limited to some extent by data gaps, limitations in model capabilities (such as geographic coverage), and uncertainties in the underlying scientific and economic studies used to configure the benefit and cost models. Gaps in the scientific literature often result in the inability to estimate quantitative changes in health and environmental effects. Gaps in the economics literature often result in the inability to assign economic values even to those health and environmental outcomes that can be quantified. While uncertainties in the underlying scientific and economics literatures (that may result in overestimation or underestimation of benefits) are discussed in detail in the economic analyses and its supporting documents and references, the key uncertainties which have a bearing on the results of the benefit-cost analysis of this rule include the following:
    
    Uncertainty concerning actions States will undertake to comply with BART;
    
    EPA's inability to quantify potentially significant benefit categories;
    
    Uncertainties in population growth and baseline incidence rates;
    
    Uncertainties in projection of emissions inventories and air quality into the future;
    
    Uncertainty in the estimated relationships of health and welfare effects to changes in pollutant concentrations including the shape of the C-R function, the size of the effect estimates, and the relative toxicity of the many components of the PM mixture;
    
    Uncertainties in exposure estimation; and
    
    Uncertainties associated with the effect of potential future actions to limit emissions.
    
    Despite these uncertainties, we believe the benefit-cost analysis provides a reasonable indication of the expected economic benefits of the rulemaking in future years under a set of reasonable assumptions.
    
    In valuing reductions in premature fatalities associated with PM, we used a value of $5.5 million per statistical life. This represents a central value consistent with a range of values from $1 to $10 million suggested by recent meta-analyses of the wage-risk value of statistical life (VSL) literature.\87\
    
    \87\ Mrozek, J.R. and L.O. Taylor, What determines the value of a life? A Meta Analysis, Journal of Policy Analysis and Management 21 (2), pp. 253-270.
    
    The benefits estimates generated for this rule are subject to a number of assumptions and uncertainties, that are discussed throughout the Regulatory Impact Analysis document [Regulatory Impact Analysis for the Final Clean Air Visibility Rule (April 2005)]. As Table IV-2 indicates, total benefits are driven primarily by the reduction in premature fatalities each year. Elaborating on the previous uncertainty discussion, some key assumptions underlying the primary estimate for the premature mortality category include the following:
    
    (1) EPA assumes inhalation of fine particles is causally associated with premature death at concentrations near those experienced by most Americans on a daily basis. Plausible biological mechanisms for this effect have been hypothesized for the endpoints included in the primary analysis and the weight of the available epidemiological evidence supports an assumption of causality.
    
    (2) EPA assumes all fine particles, regardless of their chemical composition, are equally potent in causing premature mortality. This is an important assumption, because the proportion of certain components in the PM mixture produced via precursors emitted from EGUs may differ significantly from direct PM released from automotive engines and other industrial sources, but no clear scientific grounds exist for supporting differential effects estimates by particle type.
    
    (3) EPA assumes the C-R function for fine particles is approximately linear within the range of ambient concentrations under consideration. In the PM Criteria Document, EPA recognizes that for individuals and specific health responses there are likely threshold levels, but there remains little evidence of thresholds for PM-related effects in populations.\88\ Where potential threshold levels have been suggested, they are at fairly low levels with increasing uncertainty about effects at lower ends of the PM2.5concentration ranges. Thus, EPA estimates include health benefits from reducing the fine particles in areas with varied concentrations of PM, including both regions that are in attainment with fine particle standard and those that do not meet the standard.
    
    \88\ U.S. EPA. (2004). Air Quality Criteria for Particulate Matter. Research Triangle Park, NC: National Center for Environmental Assessment-RTP Office; Report No. EPA/600/P-99/002aD.
    
    The EPA recognizes the difficulties, assumptions, and inherent uncertainties in the overall enterprise. The analyses upon which the BART rule and guidelines are based were selected from the peer-reviewed scientific literature. We used up-to-date assessment tools, and we believe the results are highly useful in assessing this rule.
    
    There are a number of health and environmental effects that we were unable to quantify or monetize. A complete benefit-cost analysis of BART requires consideration of all benefits and costs expected to result from the rule, not just those benefits and costs which could be expressed here in dollar terms. A listing of the benefit categories that were not quantified or monetized in our estimate are provided in Table IV-4. These effects are denoted by ``B'' in Table IV-3 above, and are additive to the estimates of benefits.
    
    [[Page 39151]]
  2. What Are the Unquantified and Unmonetized Benefits of BART Emissions Reductions?
    
    Important benefits beyond the human health and welfare benefits resulting from reductions in ambient levels of PM2.5and ozone are expected to occur from this rule. These other benefits occur both directly from NOXand SO2emissions reductions, and indirectly through reductions in co-pollutants such as mercury. These benefits are listed in Table IV-4. Some of the more important examples include: Reductions in NOXand SO2emissions required by BART will reduce acidification and, in the case of NOX, eutrophication of water bodies. Reduced nitrate contamination of drinking water is another possible benefit of the rule. This final rule will also reduce acid and particulate deposition that cause damages to cultural monuments, as well as, soiling and other materials damage.
    
    To illustrate the important nature of benefit categories we are currently unable to monetize, we discuss two categories of public welfare and environmental impacts related to reductions in emissions required by BART: reduced acid deposition and reduced eutrophication of water bodies. a. What Are the Benefits of Reduced Deposition of Sulfur and Nitrogen to Aquatic, Forest, and Coastal Ecosystems?
    
    Atmospheric deposition of sulfur and nitrogen, more commonly known as acid rain, occurs when emissions of SO2and NOXreact in the atmosphere (with water, oxygen, and oxidants) to form various acidic compounds. These acidic compounds fall to earth in either a wet form (rain, snow, and fog) or a dry form (gases and particles). Prevailing winds can transport acidic compounds hundreds of miles, across State borders. Acidic compounds (including small particles such as sulfates and nitrates) cause many negative environmental effects, including acidification of lakes and streams, harm to sensitive forests, and harm to sensitive coastal ecosystems. i. Acid Deposition and Acidification of Lakes and Streams
    
    The extent of adverse effects of acid deposition on freshwater and forest ecosystems depends largely upon the ecosystem's ability to neutralize the acid. The neutralizing ability [key indicator is termed Acid Neutralizing Capacity (ANC)] depends largely on the watershed's physical characteristics: geology, soils, and size. Waters that are sensitive to acidification tend to be located in small watersheds that have few alkaline minerals and shallow soils. Conversely, watersheds that contain alkaline minerals, such as limestone, tend to have waters with a high ANC. Areas especially sensitive to acidification include portions of the Northeast (particularly, the Adirondack and Catskill Mountains, portions of New England, and streams in the mid-Appalachian highlands) and southeastern streams. ii. Acid Deposition and Forest Ecosystem Impacts
    
    Current understanding of the effects of acid deposition on forest ecosystems focuses on the effects of ecological processes affecting plant uptake, retention, and cycling of nutrients within forest ecosystems. Recent studies indicate that acid deposition is at least partially responsible for decreases in base cations (calcium, magnesium, potassium, and others) from soils in the northeastern and southeastern United States. Losses of calcium from forest soils and forested watersheds have now been documented as a sensitive early indicator of soil response to acid deposition for a wide range of forest soils in the United States.
    
    In red spruce stands, a clear link exists between acid deposition, calcium supply, and sensitivity to abiotic stress. Red spruce uptake and retention of calcium is impacted by acid deposition in two main ways: leaching of important stores of calcium from needles and decreased root uptake of calcium due to calcium depletion from the soil and aluminum mobilization. These changes increase the sensitivity of red spruce to winter injuries under normal winter conditions in the Northeast, result in the loss of needles, slow tree growth, and impair the overall health and productivity of forest ecosystems in many areas of the eastern United States. In addition, recent studies of sugar maple decline in the Northeast demonstrate a link between low base cation availability, high levels of aluminum and manganese in the soil, and increased levels of tree mortality due to native defoliating insects.
    
    Although sulfate is the primary cause of base cation leaching, nitrate is a significant contributor in watersheds that are nearly nitrogen saturated. Base cation depletion is a cause for concern because of the role these ions play in surface water acid neutralization and their importance as essential nutrients for tree growth (calcium, magnesium and potassium).
    
    This regulatory action will decrease acid deposition in the transport region and is likely to have positive effects on the health and productivity of forest systems in the region. iii. Coastal Ecosystems
    
    Since 1990, a large amount of research has been conducted on the impact of nitrogen deposition to coastal waters. Nitrogen is often the limiting nutrient in coastal ecosystems. Increasing the levels of nitrogen in coastal waters can cause significant changes to those ecosystems. In recent decades, human activities have accelerated nitrogen nutrient inputs, causing excessive growth of algae and leading to degraded water quality and associated impairments of estuarine and coastal resources.
    
    Atmospheric deposition of nitrogen is a significant source of nitrogen to many estuaries. The amount of nitrogen entering estuaries due to atmospheric deposition varies widely, depending on the size and location of the estuarine watershed and other sources of nitrogen in the watershed. There are a few estuaries where atmospheric deposition of nitrogen contributes well over 40 percent of the total nitrogen load; however, in most estuaries for which estimates exist, the contribution from atmospheric deposition ranges from 15-30 percent. The area of the country with the highest air deposition rates (30 percent deposition rates) includes many estuaries along the northeast seaboard from Massachusetts to the Chesapeake Bay and along the central Gulf of Mexico coast.
    
    In 1999, National Oceanic and Atmospheric Administration (NOAA) published the results of a 5-year national assessment of the severity and extent of estuarine eutrophication. An estuary is defined as the inland arm of the sea that meets the mouth of a river. The 138 estuaries characterized in the study represent more than 90 percent of total estuarine water surface area and the total number of U.S. estuaries. The study found that estuaries with moderate to high eutrophication represented 65 percent of the estuarine surface area.
    
    Eutrophication is of particular concern in coastal areas with poor or stratified circulation patterns, such as the Chesapeake Bay, Long Island Sound, and the Gulf of Mexico. In such areas, the ``overproduced'' algae tends to sink to the bottom and decay, using all or most of the available oxygen and thereby reducing or eliminating populations of bottom-feeder fish and shellfish, distorting the normal population balance between different aquatic organisms, and in extreme cases, causing dramatic fish kills. Severe and persistent eutrophication often directly impacts human activities. For example,
    
    [[Page 39152]]
    
    fishery resource losses can be caused directly by fish kills associated with low dissolved oxygen and toxic blooms. Declines in tourism occur when low dissolved oxygen causes Noxious smells and floating mats of algal blooms create unfavorable aesthetic conditions. Risks to human health increase when the toxins from algal blooms accumulate in edible fish and shellfish, and when toxins become airborne, causing respiratory problems due to inhalation. According to the NOAA report, more than half of the nation's estuaries have moderate to high expressions of at least one of these symptoms'an indication that eutrophication is well developed in more than half of U.S. estuaries.
    
    This rule is anticipated to reduce nitrogen deposition in the nation. Thus, reductions in the levels of nitrogen deposition will have a positive impact upon current eutrophic conditions in estuaries and coastal areas in the country. 5. Are There Health or Welfare Disbenefits of the BART That Have Not Been Quantified?
    
    In contrast to the additional benefits of the rule discussed above, it is also possible that this rule will result in disbenefits in some areas of the region. Current levels of nitrogen deposition in these areas may provide passive fertilization for forest and terrestrial ecosystems where nutrients are a limiting factor and for some croplands.
    
    The effects of ozone and PM on radiative transfer in the atmosphere can also lead to effects of uncertain magnitude and direction on the penetration of ultraviolet light and climate. Ground level ozone makes up a small percentage of total atmospheric ozone (including the stratospheric layer) that attenuates penetration of ultraviolet--b (UVb) radiation to the ground. The EPA's past evaluation of the information indicates that potential disbenefits would be small, variable, and with too many uncertainties to attempt quantification of relatively small changes in average ozone levels over the course of a year (EPA, 2005a). The EPA's most recent provisional assessment of the currently available information indicates that potential but unquantifiable benefits may also arise from ozone-related attenuation of UVb radiation (EPA, 2005b). Sulfate and nitrate particles also scatter UVb, which can decrease exposure of horizontal surfaces to UVb, but increase exposure of vertical surfaces. In this case as well, both the magnitude and direction of the effect of reductions in sulfate and nitrate particles are too uncertain to quantify (EPA, 2004). Ozone is a greenhouse gas, and sulfates and nitrates can reduce the amount of solar radiation reaching the earth, but EPA believes that we are unable to quantify any net climate-related disbenefit or benefit associated with the combined ozone and PM reductions in this rule.
2. Paperwork Reduction Act
  
  Today's rule clarifies, but does not modify the information collection requirements for BART. Therefore, this action does not impose any new information collection burden. However, the OMB has previously approved the information collection requirements contained in the existing regulations [40 CFR Part 51] under the provisions of the Paperwork Reduction Act, 44 U.S.C. 3501 et seq. and has assigned OMB control number 2060-0421, EPA ICR number 1813.04. A copy of the OMB approved Information Collection Request (ICR) may be obtained from Susan Auby, Collection Strategies Division; U.S. Environmental Protection Agency (2822T); 1200 Pennsylvania Ave., NW, Washington, DC 20460 or by calling (202) 566-1672.
  
  Burden means the total time, effort, or financial resources expended by persons to generate, maintain, retain, or disclose or provide information to or for a Federal agency. This includes the time needed to review instructions; develop, acquire, install, and utilize technology and systems for the purposes of collecting, validating, and verifying information, processing and maintaining information, and disclosing and providing information; adjust the existing ways to comply with any previously applicable instructions and requirements; train personnel to be able to respond to a collection of information; search data sources; complete and review the collection of information; and transmit or otherwise disclose the information.
  
  An agency may not conduct or sponsor, and a person is not required to respond to a collection of information unless it displays a currently valid OMB control number. The OMB control numbers for EPA's regulations in 40 CFR are listed in 40 CFR part 9.
3. Regulatory Flexibility Act
  
  EPA has determined that it is not necessary to prepare a regulatory flexibility analysis in connection with this final rule.
  
  For purposes of assessing the impacts of today's rule on small entities, small entity is defined as: (1) A small business as defined by the Small Business Administrations' regulations at 13 CFR 121.201; (2) a small governmental jurisdiction that is a government of a city, county, town, school district or special district with a population of less than 50,000; and (3) a small organization that is any not-for- profit enterprise which is independently owned and operated and is not dominant in its field.
  
  Table IV-5 lists potentially impacted BART industry source categories and the current applicable small business criteria established by the Small Business Administration.
  
  Table IV-5. Potentially Affected BART Source Categories and Small Business Size Standards
  
  NAICS \a\
  
  Description
  
  Size standard \b\
  
  221112 \c,d\.......................... Fossil fuel-fired electric electric output 2requirements between an emission rate and a removal rate.
  
  For a State that chooses to require case-by-case BART, today's rule would establish presumptive levels of controls for SO2and NOXfor certain EGUs that the State finds are subject to BART. Based on its consideration of various factors set forth in the regulations; however, a State may conclude that a different level of control is appropriate. The States will accordingly exercise substantial intervening discretion in implementing the final rule. Additionally, we have assessed that the compliance dates for the rule will provide adequate time for EGUs to install the required emission controls.
National Technology Transfer and Advancement Act

Section 12(d) of the National Technology Transfer Advancement Act of 1995 (NTTAA), Public Law 104-113, section 12(d)(15 U.S.C. 272 note) directs EPA to use voluntary consensus standards (VCS) in its regulatory activities unless to do so would be inconsistent with applicable law or otherwise impractical. Voluntary consensus standards are technical standards (e.g., materials specifications, test methods, sampling procedures, and business practices) that are developed or adopted by VCS bodies. The NTTAA directs EPA to provide Congress, through OMB, explanations when the EPA decides not to use VCS.

This action does not involve technical standards; thus, EPA did not consider the use of any VCS.
1. Executive Order 12898: Federal Actions To Address Environmental Justice in Minority Populations and Low-Income Populations
  
  Executive Order 12898, ``Federal Actions to Address Environmental Justice in Minority Populations and Low-Income Populations,'' requires federal agencies to consider the impact of programs, policies, and activities on minority populations and low-income populations. According to EPA guidance,\90\ agencies are to assess whether minority or low-income populations face risks or a rate of exposure to hazards that are significant and that ``appreciably exceed or is likely to appreciably exceed the risk or rate to the general population or to the appropriate comparison group.'' (EPA, 1998)
  
  \90\ U.S. Environmental Protection Agency, 1998. Guidance for Incorporating Environmental Justice Concerns in EPA's NEPA Compliance Analyses. Office of Federal Activities, Washington, D.C., April, 1998.
  
  In accordance with Executive Order 12898, the Agency has considered whether this rule may have disproportionate negative impacts on minority or low income populations. Negative impacts to these sub- populations that appreciably exceed similar impacts to the general population are not expected because the Agency expects this rule to lead to reductions in air pollution emissions and exposures generally.
2. Congressional Review Act
  
  The Congressional Review Act, 5 U.S.C. 801 et seq., as added by the Small Business Regulatory Enforcement Fairness Act of 1996, generally provides that before a rule may take effect, the agency promulgating the rule must submit a rule report, which includes a copy of the rule, to each House of the Congress and to the Comptroller General of the United States. The EPA will submit a report containing this rule and other required information to the U.S. Senate, the U.S. House of Representatives, and the Comptroller General of the United States prior to publication of the rule in the Federal Register. A major rule cannot take effect until 60 days after it is published in the Federal Register. This action is a ``major rule'' as defined by 5 U.S.C. 804(2).
  
  List of Subjects in 40 CFR Part 51
  
  Environmental protection, Air pollution control, Administrative practice and procedure, Intergovernmental relations, Lead, Nitrogen dioxide, Ozone, Particulate matter, Reporting and recordkeeping
  
  [[Page 39156]]
  
  requirements, Sulfur oxides, Volatile organic compounds.
  
  Dated: June 15, 2005. Stephen L. Johnson, Administrator.
  
  0 For the reasons set forth in the preamble, part 51 of chapter I of title 40 of the Code of Federal Regulations is amended as follows:
  
  PART 51--REQUIREMENTS FOR PREPARATION, ADOPTION, AND SUBMITTAL OF IMPLEMENTATION PLANS
  
  0 1. The authority citation for part 51 continues to read as follows:
  
  Authority: 23 U.S.C. 101; 42 U.S.C. 7410-7671q.
  
  0 2. Section 51.302 is amended by revising paragraph (c)(4)(iii) to read as follows:
  
  Sec. 51.302 Implementation control strategies for reasonably attributable visibility impairment.
  
  * * * * *
  
  (c) * * *
  
  (4) * * *
  
  (iii) BART must be determined for fossil-fuel fired generating plants having a total generating capacity in excess of 750 megawatts pursuant to ``Guidelines for Determining Best Available Retrofit Technology for Coal-fired Power Plants and Other Existing Stationary Facilities'' (1980), which is incorporated by reference, exclusive of appendix E to the Guidelines, except that options more stringent than NSPS must be considered. Establishing a BART emission limitation equivalent to the NSPS level of control is not a sufficient basis to avoid the analysis of control options required by the guidelines. This document is EPA publication No. 450/3-80-009b and has been approved for incorporation by reference by the Director of the Federal Register in accordance with 5 U.S.C. 552(a) and 1 CFR part 51. It is for sale from the U.S. Department of Commerce, National Technical Information Service, 5285 Port Royal Road, Springfield, Virginia 22161. It is also available for inspection from the National Archives and Records Administration (NARA). For information on the availability of this material at NARA, call 202-741-6030, or go to: http://www.archives.gov/federal_register/index.html .
  
  * * * * *
  
  0 3. Section 51.308 is amended by revising paragraph (b), removing and reserving paragraph (c), revising paragraphs (e)(1)(ii), (e)(3), and (e)(4), and adding paragaphs (e)(5) and (6) to read as follows:
  
  Sec. 51.308 Regional haze program requirements.
  
  * * * * *
  
  (b) When are the first implementation plans due under the regional haze program? Except as provided in Sec. 51.309(c), each State identified in Sec. 51.300(b)(3) must submit, for the entire State, an implementation plan for regional haze meeting the requirements of paragraphs (d) and (e) of this section no later than December 17, 2007.
  
  (c) [Reserved] * * * * *
  
  (e) * * *
  
  (1) * * *
  
  (ii) A determination of BART for each BART-eligible source in the State that emits any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in any mandatory Class I Federal area. All such sources are subject to BART.
  
  (A) The determination of BART must be based on an analysis of the best system of continuous emission control technology available and associated emission reductions achievable for each BART-eligible source that is subject to BART within the State. In this analysis, the State must take into consideration the technology available, the costs of compliance, the energy and nonair quality environmental impacts of compliance, any pollution control equipment in use at the source, the remaining useful life of the source, and the degree of improvement in visibility which may reasonably be anticipated to result from the use of such technology.
  
  (B) The determination of BART for fossil-fuel fired power plants having a total generating capacity greater than 750 megawatts must be made pursuant to the guidelines in appendix Y of this part (Guidelines for BART Determinations Under the Regional Haze Rule).
  
  (C) Exception. A State is not required to make a determination of BART for SO2or for NOXif a BART-eligible source has the potential to emit less than 40 tons per year of such pollutant(s), or for PM10if a BART-eligible source emits less than 15 tons per year of such pollutant. * * * * *
  
  (3) A State which opts under 40 CFR 51.308(e)(2) to implement an emissions trading program or other alternative measure rather than to require sources subject to BART to install, operate, and maintain BART may satisfy the final step of the demonstration required by that section as follows: If the distribution of emissions is not substantially different than under BART, and the alternative measure results in greater emission reductions, then the alternative measure may be deemed to achieve greater reasonable progress. If the distribution of emissions is significantly different, the State must conduct dispersion modeling to determine differences in visibility between BART and the trading program for each impacted Class I area, for the worst and best 20 percent of days. The modeling would demonstrate ``greater reasonable progress'' if both of the following two criteria are met:
  
  (i) Visibility does not decline in any Class I area, and
  
  (ii) There is an overall improvement in visibility, determined by comparing the average differences between BART and the alternative over all affected Class I areas.
  
  (4) A State that opts to participate in the Clean Air Interstate Rule cap-and-trade and trade program under part 96 AAA-EEE need not require affected BART-eligible EGU's to install, operate, and maintain BART. A State that chooses this option may also include provisions for a geographic enhancement to the program to address the requirement under Sec. 51.302(c) related to BART for reasonably attributable impairment from the pollutants covered by the CAIR cap-and-trade program.
  
  (5) After a State has met the requirements for BART or implemented emissions trading program or other alternative measure that achieves more reasonable progress than the installation and operation of BART, BART-eligible sources will be subject to the requirements of paragraph (d) of this section in the same manner as other sources.
  
  (6) Any BART-eligible facility subject to the requirement under paragraph (e) of this section to install, operate, and maintain BART may apply to the Administrator for an exemption from that requirement. An application for an exemption will be subject to the requirements of Sec. 51.303(a)(2)-(h). * * * * *
  
  0 4. Appendix Y to Part 51 is added to read as follows:
  
  Appendix Y to Part 51--Guidelines for BART Determinations Under the Regional Haze Rule
  
  Table of Contents
Introduction and Overview
1. What is the purpose of the guidelines?
2. What does the CAA require generally for improving visibility?
3. What is the BART requirement in the CAA?
4. What types of visibility problems does EPA address in its regulations?
  
  [[Page 39157]]
5. What are the BART requirements in EPA's regional haze regulations?
6. What is included in the guidelines?
7. Who is the target audience for the guidelines?
8. Do EPA regulations require the use of these guidelines? II. How to Identify BART-eligible Sources
9. What are the steps in identifying BART-eligible sources?
  1. Step 1: Identify emission units in the BART categories
  2. Step 2: Identify the start-up dates of the emission units
  3. Step 3: Compare the potential emissions to the 250 ton/yr cutoff
  4. Final step: Identify the emission units and pollutants that constitute the BART-eligible source. III. How to Identify Sources ``Subject to BART'' IV. The BART Determination: Analysis of BART Options
10. What factors must I address in the BART Analysis?
11. What is the scope of the BART review?
12. How does a BART review relate to maximum achievable control technology (MACT) standards under CAA section 112?
13. What are the five basic steps of a case-by-case BART analysis?
  1. Step 1: How do I identify all available retrofit emission control techniques?
  2. Step 2: How do I determine whether the options identified in Step 1 are technically feasible?
  3. Step 3: How do I evaluate technically feasible alternatives?
  4. Step 4: For a BART review, what impacts am I expected to calculate and report? What methods does EPA recommend for the impacts analyses?
    1. Impact analysis part 1: how do I estimate the costs of control?
    2. What do we mean by cost effectiveness?
    3. How do I calculate average cost effectiveness?
    4. How do I calculate baseline emissions?
    5. How do I calculate incremental cost effectiveness?
    6. What other information should I provide in the cost impacts analysis?
    7. What other things are important to consider in the cost impacts analysis?
    8. Impact analysis part 2: How should I analyze and report energy impacts?
    9. Impact analysis part 3: How do I analyze ``non-air quality environmental impacts?''
    10. Impact analysis part 4: What are examples of non-air quality environmental impacts?
    11. How do I take into account a project's ``remaining useful life'' in calculating control costs?
  5. Step 5: How should I determine visibility impacts in the BART determination?
14. How do I select the ``best'' alternative, using the results of Steps 1 through 5?
  1. Summary of the impacts analysis
  2. Selecting a ``best'' alternative
  3. In selecting a ``best'' alternative, should I consider the affordability of controls?
  4. SO2limits for utility boilers
  5. NOXlimits for utility boilers V. Enforceable Limits/Compliance Date
Introduction and Overview
1. What is the purpose of the guidelines?
  
  The Clean Air Act (CAA), in sections 169A and 169B, contains requirements for the protection of visibility in 156 scenic areas across the United States. To meet the CAA's requirements, we published regulations to protect against a particular type of visibility impairment known as ``regional haze.'' The regional haze rule is found in this part at 40 CFR 51.300 through 51.309. These regulations require, in 40 CFR 51.308(e), that certain types of existing stationary sources of air pollutants install best available retrofit technology (BART). The guidelines are designed to help States and others (1) identify those sources that must comply with the BART requirement, and (2) determine the level of control technology that represents BART for each source.
2. What does the CAA require generally for improving visibility?
  
  Section 169A of the CAA, added to the CAA by the 1977 amendments, requires States to protect and improve visibility in certain scenic areas of national importance. The scenic areas protected by section 169A are ``the mandatory Class I Federal Areas * * * where visibility is an important value.'' In these guidelines, we refer to these as ``Class I areas.'' There are 156 Class I areas, including 47 national parks (under the jurisdiction of the Department of Interior--National Park Service), 108 wilderness areas (under the jurisdiction of the Department of the Interior--Fish and Wildlife Service or the Department of Agriculture--U.S. Forest Service), and one International Park (under the jurisdiction of the Roosevelt-Campobello International Commission). The Federal Agency with jurisdiction over a particular Class I area is referred to in the CAA as the Federal Land Manager. A complete list of the Class I areas is contained in 40 CFR 81.401 through 81.437, and you can find a map of the Class I areas at the following Internet site: http://www.epa.gov/ttn/oarpg/t1/fr_notices/classimp.gif .
  
  The CAA establishes a national goal of eliminating man-made visibility impairment from all Class I areas. As part of the plan for achieving this goal, the visibility protection provisions in the CAA mandate that EPA issue regulations requiring that States adopt measures in their State implementation plans (SIPs), including long- term strategies, to provide for reasonable progress towards this national goal. The CAA also requires States to coordinate with the Federal Land Managers as they develop their strategies for addressing visibility.
3. What is the BART requirement in the CAA?
  1. Under section 169A(b)(2)(A) of the CAA, States must require certain existing stationary sources to install BART. The BART provision applies to ``major stationary sources'' from 26 identified source categories which have the potential to emit 250 tons per year or more of any air pollutant. The CAA requires only sources which were put in place during a specific 15-year time interval to be subject to BART. The BART provision applies to sources that existed as of the date of the 1977 CAA amendments (that is, August 7, 1977) but which had not been in operation for more than 15 years (that is, not in operation as of August 7, 1962).
  2. The CAA requires BART review when any source meeting the above description ``emits any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility'' in any Class I area. In identifying a level of control as BART, States are required by section 169A(g) of the CAA to consider:
    
    (a) The costs of compliance,
    
    (b) The energy and non-air quality environmental impacts of compliance,
    
    (c) Any existing pollution control technology in use at the source,
    
    (d) The remaining useful life of the source, and
    
    (e) The degree of visibility improvement which may reasonably be anticipated from the use of BART.
  3. The CAA further requires States to make BART emission limitations part of their SIPs. As with any SIP revision, States must provide an opportunity for public comment on the BART determinations, and EPA's action on any SIP revision will be subject to judicial review.
4. What types of visibility problems does EPA address in its regulations?
  1. We addressed the problem of visibility in two phases. In 1980, we published regulations addressing what we termed ``reasonably attributable'' visibility impairment. Reasonably attributable visibility impairment is the result of emissions from one or a few sources that are generally located in close proximity to a specific Class I area. The regulations addressing reasonably attributable visibility impairment are published in 40 CFR 51.300 through 51.307.
  2. On July 1, 1999, we amended these regulations to address the second, more common, type of visibility impairment known as ``regional haze.'' Regional haze is the result of the collective contribution of many sources over a broad region. The regional haze rule slightly modified 40 CFR 51.300 through 51.307, including the addition of a few definitions in Sec. 51.301, and added new Sec. Sec. 51.308 and 51.309.
5. What are the BART requirements in EPA's regional haze regulations?
  1. In the July 1, 1999 rulemaking, we added a BART requirement for regional haze. We amended the BART requirements in 2005. You will find the BART requirements in 40 CFR 51.308(e). Definitions of terms used in 40 CFR 51.308(e)(1) are found in 40 CFR 51.301.
  2. As we discuss in detail in these guidelines, the regional haze rule codifies and clarifies the BART provisions in the CAA. The rule requires that States identify and list ``BART-eligible sources,'' that is, that States identify and list those sources that fall within the 26 source categories, were put in place during the 15-year window of time from 1962 to 1977, and have potential
    
    [[Page 39158]]
    
    emissions greater than 250 tons per year. Once the State has identified the BART-eligible sources, the next step is to identify those BART-eligible sources that may ``emit any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility.'' Under the rule, a source which fits this description is ``subject to BART.'' For each source subject to BART, 40 CFR 51.308(e)(1)(ii)(A) requires that States identify the level of control representing BART after considering the factors set out in CAA section 169A(g), as follows:
    
    --States must identify the best system of continuous emission control technology for each source subject to BART taking into account the technology available, the costs of compliance, the energy and non-air quality environmental impacts of compliance, any pollution control equipment in use at the source, the remaining useful life of the source, and the degree of visibility improvement that may be expected from available control technology.
  3. After a State has identified the level of control representing BART (if any), it must establish an emission limit representing BART and must ensure compliance with that requirement no later than 5 years after EPA approves the SIP. States may establish design, equipment, work practice or other operational standards when limitations on measurement technologies make emission standards infeasible.
6. What is included in the guidelines?
  1. The guidelines provide a process for making BART determinations that States can use in implementing the regional haze BART requirements on a source-by-source basis, as provided in 40 CFR 51.308(e)(1). States must follow the guidelines in making BART determinations on a source-by-source basis for 750 megawatt (MW) power plants but are not required to use the process in the guidelines when making BART determinations for other types of sources.
  2. The BART analysis process, and the contents of these guidelines, are as follows:
  (a) Identification of all BART-eligible sources. Section II of these guidelines outlines a step-by-step process for identifying BART-eligible sources.
  
  (b) Identification of sources subject to BART. As noted above, sources ``subject to BART'' are those BART-eligible sources which ``emit a pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in any Class I area.'' We discuss considerations for identifying sources subject to BART in section III of the guidance.
  
  (c) The BART determination process. For each source subject to BART, the next step is to conduct an analysis of emissions control alternatives. This step includes the identification of available, technically feasible retrofit technologies, and for each technology identified, an analysis of the cost of compliance, the energy and non-air quality environmental impacts, and the degree of visibility improvement in affected Class I areas resulting from the use of the control technology. As part of the BART analysis, the State should also take into account the remaining useful life of the source and any existing control technology present at the source. For each source, the State will determine a ``best system of continuous emission reduction'' based upon its evaluation of these factors. Procedures for the BART determination step are described in section IV of these guidelines.
  
  (d) Emissions limits. States must establish emission limits, including a deadline for compliance, consistent with the BART determination process for each source subject to BART. Considerations related to these limits are discussed in section V of these guidelines.
7. Who is the target audience for the guidelines?
  1. The guidelines are written primarily for the benefit of State, local and Tribal agencies, and describe a process for making the BART determinations and establishing the emission limitations that must be included in their SIPs or Tribal implementation plans (TIPs). Throughout the guidelines, which are written in a question and answer format, we ask questions ``How do I * * *? '' and answer with phrases ``you should * * *, you must * * * '' The ``you'' means a State, local or Tribal agency conducting the analysis. We have used this format to make the guidelines simpler to understand, but we recognize that States have the authority to require source owners to assume part of the analytical burden, and that there will be differences in how the supporting information is collected and documented. We also recognize that data collection, analysis, and rule development may be performed by Regional Planning Organizations, for adoption within each SIP or TIP.
  2. The preamble to the 1999 regional haze rule discussed at length the issue of Tribal implementation of the requirements to submit a plan to address visibility. As explained there, requirements related to visibility are among the programs for which Tribes may be determined eligible and receive authorization to implement under the ``Tribal Authority Rule'' (``TAR'') (40 CFR 49.1 through 49.11). Tribes are not subject to the deadlines for submitting visibility implementation plans and may use a modular approach to CAA implementation. We believe there are very few BART- eligible sources located on Tribal lands. Where such sources exist, the affected Tribe may apply for delegation of implementation authority for this rule, following the process set forth in the TAR.
8. Do EPA regulations require the use of these guidelines?
  
  Section 169A(b) requires us to issue guidelines for States to follow in establishing BART emission limitations for fossil-fuel fired power plants having a capacity in excess of 750 megawatts. This document fulfills that requirement, which is codified in 40 CFR 51.308(e)(1)(ii)(B). The guidelines establish an approach to implementing the requirements of the BART provisions of the regional haze rule; we believe that these procedures and the discussion of the requirements of the regional haze rule and the CAA should be useful to the States. For sources other than 750 MW power plants, however, States retain the discretion to adopt approaches that differ from the guidelines.
How to Identify BART-Eligible Sources

This section provides guidelines on how to identify BART- eligible sources. A BART-eligible source is an existing stationary source in any of 26 listed categories which meets criteria for startup dates and potential emissions.
1. What are the steps in identifying BART-eligible sources?
Figure 1 shows the steps for identifying whether the source is a ``BART-eligible source:''

Step 1: Identify the emission units in the BART categories,

Step 2: Identify the start-up dates of those emission units, and

Step 3: Compare the potential emissions to the 250 ton/yr cutoff.

Figure 1. How to determine whether a source is BART-eligible:

Step 1: Identify emission units in the BART categories

Does the plant contain emissions units in one or more of the 26 source categories? [rtarr2] No [rtarr2] Stop [rtarr2] Yes [rtarr2] Proceed to Step 2

Step 2: Identify the start-up dates of these emission units

Do any of these emissions units meet the following two tests?

In existence on August 7, 1977 AND

Began operation after August 7, 1962 [rtarr2] No [rtarr2] Stop [rtarr2] Yes [rtarr2] Proceed to Step 3

Step 3: Compare the potential emissions from these emission units to the 250 ton/yr cutoff

Identify the ``stationary source'' that includes the emission units you identified in Step 2.

Add the current potential emissions from all the emission units identified in Steps 1 and 2 that are included within the ``stationary source'' boundary.

Are the potential emissions from these units 250 tons per year or more for any visibility-impairing pollutant? [rtarr2] No [rtarr2] Stop [rtarr2] Yes [rtarr2] These emissions units comprise the ``BART-eligible source.''
1. Step 1: Identify Emission Units in the BART Categories
2. The BART requirement only applies to sources in specific categories listed in the CAA. The BART requirement does not apply to sources in other source categories, regardless of their emissions. The listed categories are:
  
  (1) Fossil-fuel fired steam electric plants of more than 250 million British thermal units (BTU) per hour heat input,
  
  (2) Coal cleaning plants (thermal dryers),
  
  (3) Kraft pulp mills,
  
  (4) Portland cement plants,
  
  (5) Primary zinc smelters,
  
  (6) Iron and steel mill plants,
  
  (7) Primary aluminum ore reduction plants,
  
  [[Page 39159]]
  
  (8) Primary copper smelters,
  
  (9) Municipal incinerators capable of charging more than 250 tons of refuse per day,
  
  (10) Hydrofluoric, sulfuric, and nitric acid plants,
  
  (11) Petroleum refineries,
  
  (12) Lime plants,
  
  (13) Phosphate rock processing plants,
  
  (14) Coke oven batteries,
  
  (15) Sulfur recovery plants,
  
  (16) Carbon black plants (furnace process),
  
  (17) Primary lead smelters,
  
  (18) Fuel conversion plants,
  
  (19) Sintering plants,
  
  (20) Secondary metal production facilities,
  
  (21) Chemical process plants,
  
  (22) Fossil-fuel boilers of more than 250 million BTUs per hour heat input,
  
  (23) Petroleum storage and transfer facilities with a capacity exceeding 300,000 barrels,
  
  (24) Taconite ore processing facilities,
  
  (25) Glass fiber processing plants, and
  
  (26) Charcoal production facilities.
3. Some plants may have emission units from more than one category, and some emitting equipment may fit into more than one category. Examples of this situation are sulfur recovery plants at petroleum refineries, coke oven batteries and sintering plants at steel mills, and chemical process plants at refineries. For Step 1, you identify all of the emissions units at the plant that fit into one or more of the listed categories. You do not identify emission units in other categories.
  
  Example: A mine is collocated with an electric steam generating plant and a coal cleaning plant. You would identify emission units associated with the electric steam generating plant and the coal cleaning plant, because they are listed categories, but not the mine, because coal mining is not a listed category.
4. The category titles are generally clear in describing the types of equipment to be listed. Most of the category titles are very broad descriptions that encompass all emission units associated with a plant site (for example, ``petroleum refining'' and ``kraft pulp mills''). This same list of categories appears in the PSD regulations. States and source owners need not revisit any interpretations of the list made previously for purposes of the PSD program. We provide the following clarifications for a few of the category titles:
  
  (1) ``Steam electric plants of more than 250 million BTU/hr heat input.'' Because the category refers to ``plants,'' we interpret this category title to mean that boiler capacities should be aggregated to determine whether the 250 million BTU/hr threshold is reached. This definition includes only those plants that generate electricity for sale. Plants that cogenerate steam and electricity also fall within the definition of ``steam electric plants''. Similarly, combined cycle turbines are also considered ``steam electric plants'' because such facilities incorporate heat recovery steam generators. Simple cycle turbines, in contrast, are not ``steam electric plants'' because these turbines typically do not generate steam.
  
  Example: A stationary source includes a steam electric plant with three 100 million BTU/hr boilers. Because the aggregate capacity exceeds 250 million BTU/hr for the ``plant,'' these boilers would be identified in Step 2.
  
  (2) ``Fossil-fuel boilers of more than 250 million BTU/hr heat input.'' We interpret this category title to cover only those boilers that are individually greater than 250 million BTU/hr. However, an individual boiler smaller than 250 million BTU/hr should be subject to BART if it is an integral part of a process description at a plant that is in a different BART category--for example, a boiler at a Kraft pulp mill that, in addition to providing steam or mechanical power, uses the waste liquor from the process as a fuel. In general, if the process uses any by-product of the boiler and the boiler's function is to serve the process, then the boiler is integral to the process and should be considered to be part of the process description.
  
  Also, you should consider a multi-fuel boiler to be a ``fossil- fuel boiler'' if it burns any amount of fossil fuel. You may take federally and State enforceable operational limits into account in determining whether a multi-fuel boiler's fossil fuel capacity exceeds 250 million Btu/hr.
  
  (3) ``Petroleum storage and transfer facilities with a capacity exceeding 300,000 barrels.'' The 300,000 barrel cutoff refers to total facility-wide tank capacity for tanks that were put in place within the 1962-1977 time period, and includes gasoline and other petroleum-derived liquids.
  
  (4) ``Phosphate rock processing plants.'' This category descriptor is broad, and includes all types of phosphate rock processing facilities, including elemental phosphorous plants as well as fertilizer production plants.
  
  (5) ``Charcoal production facilities.'' We interpret this category to include charcoal briquet manufacturing and activated carbon production.
  
  (6) ``Chemical process plants.'' and pharmaceutical manufacturing. Consistent with past policy, we interpret the category ``chemical process plants'' to include those facilities within the 2-digit Standard Industrial Classification (SIC) code 28. Accordingly, we interpret the term ``chemical process plants'' to include pharmaceutical manufacturing facilities.
  
  (7) ``Secondary metal production.'' We interpret this category to include nonferrous metal facilities included within SIC code 3341, and secondary ferrous metal facilities that we also consider to be included within the category ``iron and steel mill plants.''
  
  (8) ``Primary aluminum ore reduction.'' We interpret this category to include those facilities covered by 40 CFR 60.190, the new source performance standard (NSPS) for primary aluminum ore reduction plants. This definition is also consistent with the definition at 40 CFR 63.840.
5. Step 2: Identify the Start-Up Dates of the Emission Units
6. Emissions units listed under Step 1 are BART-eligible only if they were ``in existence'' on August 7, 1977 but were not ``in operation'' before August 7, 1962.
  
  What does ``in existence on August 7, 1977'' mean?
7. The regional haze rule defines ``in existence'' to mean that:
  
  ``the owner or operator has obtained all necessary preconstruction approvals or permits required by Federal, State, or local air pollution emissions and air quality laws or regulations and either has (1) begun, or caused to begin, a continuous program of physical on-site construction of the facility or (2) entered into binding agreements or contractual obligations, which cannot be canceled or modified without substantial loss to the owner or operator, to undertake a program of construction of the facility to be completed in a reasonable time.'' 40 CFR 51.301.
  
  As this definition is essentially identical to the definition of ``commence construction'' as that term is used in the PSD regulations, the two terms mean the same thing. See 40 CFR 51.165(a)(1)(xvi) and 40 CFR 52.21(b)(9). Under this definition, an emissions unit could be ``in existence'' even if it did not begin operating until several years after 1977.
  
  Example: The owner of a source obtained all necessary permits in early 1977 and entered into binding construction agreements in June 1977. Actual on-site construction began in late 1978, and construction was completed in mid-1979. The source began operating in September 1979. The emissions unit was ``in existence'' as of August 7, 1977.
  
  Major stationary sources which commenced construction AFTER August 7, 1977 (i.e., major stationary sources which were not ``in existence'' on August 7, 1977) were subject to new source review (NSR) under the PSD program. Thus, the August 7, 1977 ``in existence'' test is essentially the same thing as the identification of emissions units that were grandfathered from the NSR review requirements of the 1977 CAA amendments.
8. Sources are not BART-eligible if the only change at the plant during the relevant time period was the addition of pollution controls. For example, if the only change at a copper smelter during the 1962 through 1977 time period was the addition of acid plants for the reduction of SO2emissions, these emission controls would not by themselves trigger a BART review.
  
  What does ``in operation before August 7, 1962'' mean?
  
  An emissions unit that meets the August 7, 1977 ``in existence'' test is not BART-eligible if it was in operation before August 7, 1962. ``In operation'' is defined as ``engaged in activity related to the primary design function of the source.'' This means that a source must have begun actual operations by August 7, 1962 to satisfy this test.
  
  Example: The owner or operator entered into binding agreements in 1960. Actual on-site construction began in 1961, and construction was complete in mid-1962. The source began operating in September 1962. The emissions unit was not ``in operation'' before August 7, 1962 and is therefore subject to BART.
  
  What is a ``reconstructed source?'
9. Under a number of CAA programs, an existing source which is completely or
  
  [[Page 39160]]
  
  substantially rebuilt is treated as a new source. Such ``reconstructed'' sources are treated as new sources as of the time of the reconstruction. Consistent with this overall approach to reconstructions, the definition of BART-eligible facility (reflected in detail in the definition of ``existing stationary facility'') includes consideration of sources that were in operation before August 7, 1962, but were reconstructed during the August 7, 1962 to August 7, 1977 time period.
10. Under the regional haze regulations at 40 CFR 51.301, a reconstruction has taken place if ``the fixed capital cost of the new component exceeds 50 percent of the fixed capital cost of a comparable entirely new source.'' The rule also states that ``[a]ny final decision as to whether reconstruction has occurred must be made in accordance with the provisions of Sec. Sec. 60.15 (f)(1) through (3) of this title.'' ``[T]he provisions of Sec. Sec. 60.15(f)(1) through (3)'' refers to the general provisions for New Source Performance Standards (NSPS). Thus, the same policies and procedures for identifying reconstructed ``affected facilities'' under the NSPS program must also be used to identify reconstructed ``stationary sources'' for purposes of the BART requirement.
11. You should identify reconstructions on an emissions unit basis, rather than on a plantwide basis. That is, you need to identify only the reconstructed emission units meeting the 50 percent cost criterion. You should include reconstructed emission units in the list of emission units you identified in Step 1. You need consider as possible reconstructions only those emissions units with the potential to emit more than 250 tons per year of any visibility-impairing pollutant.
12. The ``in operation'' and ``in existence'' tests apply to reconstructed sources. If an emissions unit was reconstructed and began actual operation before August 7, 1962, it is not BART- eligible. Similarly, any emissions unit for which a reconstruction ``commenced'' after August 7, 1977, is not BART-eligible.
  
  How are modifications treated under the BART provision?
13. The NSPS program and the major source NSR program both contain the concept of modifications. In general, the term ``modification'' refers to any physical change or change in the method of operation of an emissions unit that results in an increase in emissions.
14. The BART provision in the regional haze rule contains no explicit treatment of modifications or how modified emissions units, previously subject to the requirement to install best available control technology (BACT), lowest achievable emission rate (LAER) controls, and/or NSPS are treated under the rule. As the BART requirements in the CAA do not appear to provide any exemption for sources which have been modified since 1977, the best interpretation of the CAA visibility provisions is that a subsequent modification does not change a unit's construction date for the purpose of BART applicability. Accordingly, if an emissions unit began operation before 1962, it is not BART-eligible if it was modified between 1962 and 1977, so long as the modification is not also a ``reconstruction.'' On the other hand, an emissions unit which began operation within the 1962-1977 time window, but was modified after August 7, 1977, is BART-eligible. We note, however, that if such a modification was a major modification that resulted in the installation of controls, the State will take this into account during the review process and may find that the level of controls already in place are consistent with BART.
15. Step 3: Compare the Potential Emissions to the 250 Ton/Yr Cutoff
  
  The result of Steps 1 and 2 will be a list of emissions units at a given plant site, including reconstructed emissions units, that are within one or more of the BART categories and that were placed into operation within the 1962-1977 time window. The third step is to determine whether the total emissions represent a current potential to emit that is greater than 250 tons per year of any single visibility impairing pollutant. Fugitive emissions, to the extent quantifiable, must be counted. In most cases, you will add the potential emissions from all emission units on the list resulting from Steps 1 and 2. In a few cases, you may need to determine whether the plant contains more than one ``stationary source'' as the regional haze rule defines that term, and as we explain further below.
  
  What pollutants should I address?
  
  Visibility-impairing pollutants include the following:
  
  (1) Sulfur dioxide (SO2),
  
  (2) Nitrogen oxides (NOX), and
  
  (3) Particulate matter.
  
  You may use PM10as an indicator for particulate matter in this intial step. [Note that we do not recommend use of total suspended particulates (TSP) as in indicator for particulate matter.] As emissions of PM10include the components of PM2.5as a subset, there is no need to have separate 250 ton thresholds for PM10and PM2.5; 250 tons of PM10represents at most 250 tons of PM2.5, and at most 250 tons of any individual particulate species such as elemental carbon, crustal material, etc.
  
  However, if you determine that a source of particulate matter is BART-eligible, it will be important to distinguish between the fine and coarse particle components of direct particulate emissions in the remainder of the BART analysis, including for the purpose of modeling the source's impact on visibility. This is because although both fine and coarse particulate matter contribute to visibility impairment, the long-range transport of fine particles is of particular concern in the formation of regional haze. Thus, for example, air quality modeling results used in the BART determination will provide a more accurate prediction of a source's impact on visibility if the inputs into the model account for the relative particle size of any directly emitted particulate matter (i.e. PM10vs. PM2.5).
  
  You should exercise judgment in deciding whether the following pollutants impair visibility in an area:
  
  (4) Volatile organic compounds (VOC), and
  
  (5) Ammonia and ammonia compounds.
  
  You should use your best judgment in deciding whether VOC or ammonia emissions from a source are likely to have an impact on visibility in an area. Certain types of VOC emissions, for example, are more likely to form secondary organic aerosols than others.\1\ Similarly, controlling ammonia emissions in some areas may not have a significant impact on visibility. You need not provide a formal showing of an individual decision that a source of VOC or ammonia emissions is not subject to BART review. Because air quality modeling may not be feasible for individual sources of VOC or ammonia, you should also exercise your judgement in assessing the degree of visibility impacts due to emissions of VOC and emissions of ammonia or ammonia compounds. You should fully document the basis for judging that a VOC or ammonia source merits BART review, including your assessment of the source's contribution to visibility impairment.
  
  \1\ Fine particles: Overview of Atmospheric Chemistry, Sources of Emissions, and Ambient Monitoring Data, Memorandum to Docket OAR 2002-006, April 1, 2005.
  
  What does the term ``potential'' emissions mean?
  
  The regional haze rule defines potential to emit as follows:
  
  ``Potential to emit'' means the maximum capacity of a stationary source to emit a pollutant under its physical and operational design. Any physical or operational limitation on the capacity of the source to emit a pollutant including air pollution control equipment and restrictions on hours of operation or on the type or amount of material combusted, stored, or processed, shall be treated as part of its design if the limitation or the effect it would have on emissions is federally enforceable. Secondary emissions do not count in determining the potential to emit of a stationary source.
  
  The definition of ``potential to emit'' means that a source which actually emits less than 250 tons per year of a visibility-impairing pollutant is BART-eligible if its emissions would exceed 250 tons per year when operating at its maximum capacity given its physical and operational design (and considering all federally enforceable and State enforceable permit limits.)
  
  Example: A source, while operating at one-fourth of its capacity, emits 75 tons per year of SO2. If it were operating at 100 percent of its maximum capacity, the source would emit 300 tons per year. Because under the above definition such a source would have ``potential'' emissions that exceed 250 tons per year, the source (if in a listed category and built during the 1962- 1977 time window) would be BART-eligible.
  
  How do I identify whether a plant has more than one ``stationary source?''
16. The regional haze rule, in 40 CFR 51.301, defines a stationary source as a ``building, structure, facility or installation which emits or may emit any air pollutant.'' \2\
  
  [[Page 39161]]
  
  The rule further defines ``building, structure or facility'' as:
  
  \2\ Note: Most of these terms and definitions are the same for regional haze and the 1980 visibility regulations. For the regional haze rule we use the term ``BART-eligible source'' rather than ``existing stationary facility'' to clarify that only a limited subset of existing stationary sources are subject to BART.
  
  all of the pollutant-emitting activities which belong to the same industrial grouping, are located on one or more contiguous or adjacent properties, and are under the control of the same person (or persons under common control). Pollutant-emitting activities must be considered as part of the same industrial grouping if they belong to the same Major Group (i.e., which have the same two-digit code) as described in the Standard Industrial Classification Manual, 1972 as amended by the 1977 Supplement (U.S. Government Printing
  
  Office stock numbers 4101-0066 and 003-005-00176-0, respectively).
17. In applying this definition, it is necessary to determine which facilities are located on ``contiguous or adjacent properties.'' Within this contiguous and adjacent area, it is also necessary to group those emission units that are under ``common control.'' We note that these plant boundary issues and ``common control'' issues are very similar to those already addressed in implementation of the title V operating permits program and in NSR.
18. For emission units within the ``contiguous or adjacent'' boundary and under common control, you must group emission units that are within the same industrial grouping (that is, associated with the same 2-digit SIC code) in order to define the stationary source.\3\ For most plants on the BART source category list, there will only be one 2-digit SIC that applies to the entire plant. For example, all emission units associated with kraft pulp mills are within SIC code 26, and chemical process plants will generally include emission units that are all within SIC code 28. The ``2- digit SIC test'' applies in the same way as the test is applied in the major source NSR programs.\4\
  
  \3\ We recognize that we are in a transition period from the use of the SIC system to a new system called the North American Industry Classification System (NAICS). For purposes of identifying BART- eligible sources, you may use either 2-digit SICS or the equivalent in the NAICS system.
  
  \4\ Note: The concept of support facility used for the NSR program applies here as well. Support facilities, that is facilities that convey, store or otherwise assist in the production of the principal product, must be grouped with primary facilities even when the facilities fall wihin separate SIC codes. For purposes of BART reviews, however, such support facilities (a) must be within one of the 26 listed source categories and (b) must have been in existence as of August 7, 1977, and (c) must not have been in operation as of August 7, 1962.
19. For purposes of the regional haze rule, you must group emissions from all emission units put in place within the 1962-1977 time period that are within the 2-digit SIC code, even if those emission units are in different categories on the BART category list.
  
  Examples: A chemical plant which started operations within the 1962 to 1977 time period manufactures hydrochloric acid (within the category title ``Hydrochloric, sulfuric, and nitric acid plants'') and various organic chemicals (within the category title ``chemical process plants''). All of the emission units are within SIC code 28 and, therefore, all the emission units are considered in determining BART eligibility of the plant. You sum the emissions over all of these emission units to see whether there are more than 250 tons per year of potential emissions.
  
  A steel mill which started operations within the 1962 to 1977 time period includes a sintering plant, a coke oven battery, and various other emission units. All of the emission units are within SIC code 33. You sum the emissions over all of these emission units to see whether there are more than 250 tons per year of potential emissions.
20. Final Step: Identify the Emissions Units and Pollutants That Constitute the BART-Eligible Source
  
  If the emissions from the list of emissions units at a stationary source exceed a potential to emit of 250 tons per year for any visibility-impairing pollutant, then that collection of emissions units is a BART-eligible source.
  
  Example: A stationary source comprises the following two emissions units, with the following potential emissions: Emissions unit A
  
  200 tons/yr SO2
  
  150 tons/yr NOX
  
  25 tons/yr PM Emissions unit B
  
  100 tons/yr SO2
  
  75 tons/yr NOX
  
  10 tons/yr PM
  
  For this example, potential emissions of SO2are 300 tons/yr, which exceeds the 250 tons/yr threshold. Accordingly, the entire ``stationary source'', that is, emissions units A and B, may be subject to a BART review for SO2, NOX, and PM, even though the potential emissions of PM and NOXat each emissions unit are less than 250 tons/yr each.
  
  Example: The total potential emissions, obtained by adding the potential emissions of all emission units in a listed category at a plant site, are as follows:
  
  200 tons/yr SO2 150 tons/yr NOX 25 tons/yr PM
  
  Even though total emissions exceed 250 tons/yr, no individual regulated pollutant exceeds 250 tons/yr and this source is not BART- eligible.
  
  Can States establish de minimis levels of emissions for pollutants at BART-eligible sources?
  
  In order to simplify BART determinations, States may choose to identify de minimis levels of pollutants at BART-eligible sources (but are not required to do so). De minimis values should be identified with the purpose of excluding only those emissions so minimal that they are unlikely to contribute to regional haze. Any de minimis values that you adopt must not be higher than the PSD applicability levels: 40 tons/yr for SO2and NOXand 15 tons/yr for PM10. These de minimis levels may only be applied on a plant-wide basis.
How to Identify Sources ``Subject to BART''

Once you have compiled your list of BART-eligible sources, you need to determine whether (1) to make BART determinations for all of them or (2) to consider exempting some of them from BART because they may not reasonably be anticipated to cause or contribute to any visibility impairment in a Class I area. If you decide to make BART determinations for all the BART-eligible sources on your list, you should work with your regional planning organization (RPO) to show that, collectively, they cause or contribute to visibility impairment in at least one Class I area. You should then make individual BART determinations by applying the five statutory factors discussed in Section IV below.

On the other hand, you also may choose to perform an initial examination to determine whether a particular BART-eligible source or group of sources causes or contributes to visibility impairment in nearby Class I areas. If your analysis, or information submitted by the source, shows that an individual source or group of sources (or certain pollutants from those sources) is not reasonably anticipated to cause or contribute to any visibility impairment in a Class I area, then you do not need to make BART determinations for that source or group of sources (or for certain pollutants from those sources). In such a case, the source is not ``subject to BART'' and you do not need to apply the five statutory factors to make a BART determination. This section of the Guideline discusses several approaches that you can use to exempt sources from the BART determination process.
1. What Steps Do I Follow To Determine Whether a Source or Group of Sources Cause or Contribute to Visibility Impairment for Purposes of BART?
1. How Do I Establish a Threshold?
  
  One of the first steps in determining whether sources cause or contribute to visibility impairment for purposes of BART is to establish a threshold (measured in deciviews) against which to measure the visibility impact of one or more sources. A single source that is responsible for a 1.0 deciview change or more should be considered to ``cause'' visibility impairment; a source that causes less than a 1.0 deciview change may still contribute to visibility impairment and thus be subject to BART.
  
  Because of varying circumstances affecting different Class I areas, the appropriate threshold for determining whether a source ``contributes to any visibility impairment'' for the purposes of BART may reasonably differ across States. As a general matter, any threshold that you use for determining whether a source ``contributes'' to visibility impairment should not be higher than 0.5 deciviews.
  
  In setting a threshold for ``contribution,'' you should consider the number of emissions sources affecting the Class I areas at issue and the magnitude of the individual sources' impacts.\5\ In general, a larger number of sources causing impacts in a Class I area may warrant a lower contribution threshold. States remain free to use a threshold lower than 0.5 deciviews if they conclude that the
  
  [[Page 39162]]
  
  location of a large number of BART-eligible sources within the State and in proximity to a Class I area justify this approach.\6\
  
  \5\ We expect that regional planning organizations will have modeling information that identifies sources affecting visibility in individual class I areas.
  
  \6\ Note that the contribution threshold should be used to determine whether an individual source is reasonably anticipated to contribute to visibility impairment. You should not aggregate the visibility effects of multiple sources and compare their collective effects against your contribution threshold because this would inappropriately create a ``contribute to contribution'' test.
2. What Pollutants Do I Need to Consider?
  
  You must look at SO2, NOX, and direct particulate matter (PM) emissions in determining whether sources cause or contribute to visibility impairment, including both PM10and PM2.5. Consistent with the approach for identifying your BART-eligible sources, you do not need to consider less than de minimis emissions of these pollutants from a source.
  
  As explained in section II, you must use your best judgement to determine whether VOC or ammonia emissions are likely to have an impact on visibility in an area. In addition, although as explained in Section II, you may use PM10an indicator for particulate matter in determining whether a source is BART-eligible, in determining whether a source contributes to visibility impairment, you should distinguish between the fine and coarse particle components of direct particulate emissions. Although both fine and coarse particulate matter contribute to visibility impairment, the long-range transport of fine particles is of particular concern in the formation of regional haze. Air quality modeling results used in the BART determination will provide a more accurate prediction of a source's impact on visibility if the inputs into the model account for the relative particle size of any directly emitted particulate matter (i.e. PM10vs. PM2.5).
3. What Kind of Modeling Should I Use To Determine Which Sources and Pollutants Need Not Be Subject to BART?
  
  This section presents several options for determining that certain sources need not be subject to BART. These options rely on different modeling and/or emissions analysis approaches. They are provided for your guidance. You may also use other reasonable approaches for analyzing the visibility impacts of an individual source or group of sources.
  
  Option 1: Individual Source Attribution Approach (Dispersion Modeling)
  
  You can use dispersion modeling to determine that an individual source cannot reasonably be anticipated to cause or contribute to visibility impairment in a Class I area and thus is not subject to BART. Under this option, you can analyze an individual source's impact on visibility as a result of its emissions of SO2, NOXand direct PM emissions. Dispersion modeling cannot currently be used to estimate the predicted impacts on visibility from an individual source's emissions of VOC or ammonia. You may use a more qualitative assessment to determine on a case-by-case basis which sources of VOC or ammonia emissions may be likely to impair visibility and should therefore be subject to BART review, as explained in section II.A.3. above.
  
  You can use CALPUFF \7\ or other appropriate model to predict the visibility impacts from a single source at a Class I area. CALPUFF is the best regulatory modeling application currently available for predicting a single source's contribution to visibility impairment and is currently the only EPA-approved model for use in estimating single source pollutant concentrations resulting from the long range transport of primary pollutants.\8\ It can also be used for some other purposes, such as the visibility assessments addressed in today's rule, to account for the chemical transformation of SO2and NOX.
  
  \7\ The model code and its documentation are available at no cost for download from http://www.epa.gov/scram001/tt22.htm#calpuff.
  
  \8\ The Guideline on Air Quality Models, 40 CFR part 51, appendix W, addresses the regulatory application of air quality models for assessing criteria pollutants under the CAA, and describes further the procedures for using the CALPUFF model, as well as for obtaining approval for the use of other, nonguideline models.
  
  There are several steps for making an individual source attribution using a dispersion model:
4. Develop a modeling protocol. Some critical items to include in the protocol are the meteorological and terrain data that will be used, as well as the source-specific information (stack height, temperature, exit velocity, elevation, and emission rates of applicable pollutants) and receptor data from appropriate Class I areas. We recommend following EPA's Interagency Workgroup on Air Quality Modeling (IWAQM) Phase 2 Summary Report and Recommendations for Modeling Long Range Transport Impacts \9\ for parameter settings and meteorological data inputs. You may use other settings from those in IWAQM, but you should identify these settings and explain your selection of these settings.
  
  \9\ Interagency Workgroup on Air Quality Modeling (IWAQM) Phase 2 Summary Report and Recommendations for Modeling Long Range Transport Impacts, U.S. Environmental Protection Agency, EPA-454/R- 98-019, December 1998.
  
  One important element of the protocol is in establishing the receptors that will be used in the model. The receptors that you use should be located in the nearest Class I area with sufficient density to identify the likely visibility effects of the source. For other Class I areas in relatively close proximity to a BART-eligible source, you may model a few strategic receptors to determine whether effects at those areas may be greater than at the nearest Class I area. For example, you might chose to locate receptors at these areas at the closest point to the source, at the highest and lowest elevation in the Class I area, at the IMPROVE monitor, and at the approximate expected plume release height. If the highest modeled effects are observed at the nearest Class I area, you may choose not to analyze the other Class I areas any further as additional analyses might be unwarranted.
  
  You should bear in mind that some receptors within the relevant Class I area may be less than 50 km from the source while other receptors within that same Class I area may be greater than 50 km from the same source. As indicated by the Guideline on Air Quality Models, 40 CFR part 51, appendix W, this situation may call for the use of two different modeling approaches for the same Class I area and source, depending upon the State's chosen method for modeling sources less than 50 km. In situations where you are assessing visibility impacts for source-receptor distances less than 50 km, you should use expert modeling judgment in determining visibility impacts, giving consideration to both CALPUFF and other appropriate methods.
  
  In developing your modeling protocol, you may want to consult with EPA and your regional planning organization (RPO). Up-front consultation will ensure that key technical issues are addressed before you conduct your modeling.
5. With the accepted protocol and compare the predicted visibility impacts with your threshold for ``contribution.'' You should calculate daily visibility values for each receptor as the change in deciviews compared against natural visibility conditions. You can use EPA's ``Guidance for Estimating Natural Visibility Conditions Under the Regional Haze Rule,'' EPA-454/B-03-005 (September 2003) in making this calculation. To determine whether a source may reasonably be anticipated to cause or contribute to visibility impairment at Class I area, you then compare the impacts predicted by the model against the threshold that you have selected.
  
  The emissions estimates used in the models are intended to reflect steady-state operating conditions during periods of high capacity utilization. We do not generally recommend that emissions reflecting periods of start-up, shutdown, and malfunction be used, as such emission rates could produce higher than normal effects than would be typical of most facilities. We recommend that States use the 24 hour average actual emission rate from the highest emitting day of the meteorological period modeled, unless this rate reflects periods start-up, shutdown, or malfunction. In addition, the monthly average relative humidity is used, rather than the daily average humidity--an approach that effectively lowers the peak values in daily model averages.
  
  For these reasons, if you use the modeling approach we recommend, you should compare your ``contribution'' threshold against the 98th percentile of values. If the 98th percentile value from your modeling is less than your contribution threshold, then you may conclude that the source does not contribute to visibility impairment and is not subject to BART.
  
  Option 2: Use of Model Plants To Exempt Individual Sources With Common Characteristics
  
  Under this option, analyses of model plants could be used to exempt certain BART-eligible sources that share specific characteristics. It may be most useful to use this type of analysis to identify the types of small sources that do not cause or contribute to visibility impairment for purposes of BART, and thus should not be subject to a BART review. Different Class I areas may have different characteristics, however, so
  
  [[Page 39163]]
  
  you should use care to ensure that the criteria you develop are appropriate for the applicable cases.
  
  In carrying out this approach, you could use modeling analyses of representative plants to reflect groupings of specific sources with important common characteristics. Based on these analyses, you may find that certain types of sources are clearly anticipated to cause or contribute to visibility impairment. You could then choose to categorically require those types of sources to undergo a BART determination. Conversely, you may find based on representative plant analyses that certain types of sources are not reasonably anticipated to cause or contribute to visibility impairment. To do this, you may conduct your own modeling to establish emission levels and distances from Class I areas on which you can rely to exempt sources with those characteristics. For example, based on your modeling you might choose to exempt all NOX-only sources that emit less than a certain amount per year and are located a certain distance from a Class I area. You could then choose to categorically exempt such sources from the BART determination process.
  
  Our analyses of visibility impacts from model plants provide a useful example of the type of analyses that can be used to exempt categories of sources from BART.\10\ In our analyses, we developed model plants (EGUs and non-EGUs), with representative plume and stack characteristics, for use in considering the visibility impact from emission sources of different sizes and compositions at distances of 50, 100 and 200 kilometers from two hypothetical Class I areas (one in the East and one in the West). As the plume and stack characteristics of these model plants were developed considering the broad range of sources within the EGU and non-EGU categories, they do not necessarily represent any specific plant. However, the results of these analyses are instructive in the development of an exemption process for any Class I area.
  
  \10\ CALPUFF Analysis in Support of the June 2005 Changes to the Regional Haze Rule, U.S. Environmental Protection Agency, June 15, 2005, Docket No. OAR-2002-0076.
  
  In preparing our analyses, we have made a number of assumptions and exercised certain modeling choices; some of these have a tendency to lend conservatism to the results, overstating the likely effects, while others may understate the likely effects. On balance, when all of these factors are considered, we believe that our examples reflect realistic treatments of the situations being modeled. Based on our analyses, we believe that a State that has established 0.5 deciviews as a contribution threshold could reasonably exempt from the BART review process sources that emit less than 500 tons per year of NOXor SO2(or combined NOXand SO2), as long as these sources are located more than 50 kilometers from any Class I area; and sources that emit less than 1000 tons per year of NOX or SO2(or combined NOXand SO2) that are located more than 100 kilometers from any Class I area. You do, however, have the option of showing other thresholds might also be appropriate given your specific circumstances.
  
  Option 3: Cumulative Modeling To Show That No Sources in a State Are Subject to BART
  
  You may also submit to EPA a demonstration based on an analysis of overall visibility impacts that emissions from BART-eligible sources in your State, considered together, are not reasonably anticipated to cause or contribute to any visibility impairment in a Class I area, and thus no source should be subject to BART. You may do this on a pollutant by pollutant basis or for all visibility- impairing pollutants to determine if emissions from these sources contribute to visibility impairment.
  
  For example, emissions of SO2from your BART-eligible sources may clearly cause or contribute to visibility impairment while direct emissions of PM2.5from these sources may not contribute to impairment. If you can make such a demonstration, then you may reasonably conclude that none of your BART-eligible sources are subject to BART for a particular pollutant or pollutants. As noted above, your demonstration should take into account the interactions among pollutants and their resulting impacts on visibility before making any pollutant-specific determinations.
  
  Analyses may be conducted using several alternative modeling approaches. First, you may use the CALPUFF or other appropriate model as described in Option 1 to evaluate the impacts of individual sources on downwind Class I areas, aggregating those impacts to determine the collective contribution of all BART-eligible sources to visibility impairment. You may also use a photochemical grid model. As a general matter, the larger the number of sources being modeled, the more appropriate it may be to use a photochemical grid model. However, because such models are significantly less sensitive than dispersion models to the contributions of one or a few sources, as well as to the interactions among sources that are widely distributed geographically, if you wish to use a grid model, you should consult with the appropriate EPA Regional Office to develop an appropriate modeling protocol.
The BART Determination: Analysis of BART Options

This section describes the process for the analysis of control options for sources subject to BART.
1. What factors must I address in the BART review?
  
  The visibility regulations define BART as follows:
  
  Best Available Retrofit Technology (BART) means an emission limitation based on the degree of reduction achievable through the application of the best system of continuous emission reduction for each pollutant which is emitted by . . . [a BART-eligible source]. The emission limitation must be established, on a case-by-case basis, taking into consideration the technology available, the costs of compliance, the energy and non-air quality environmental impacts of compliance, any pollution control equipment in use or in existence at the source, the remaining useful life of the source, and the degree of improvement in visibility which may reasonably be anticipated to result from the use of such technology.
  
  The BART analysis identifies the best system of continuous emission reduction taking into account:
  
  (1) The available retrofit control options,
  
  (2) Any pollution control equipment in use at the source (which affects the availability of options and their impacts),
  
  (3) The costs of compliance with control options,
  
  (4) The remaining useful life of the facility,
  
  (5) The energy and non-air quality environmental impacts of control options
  
  (6) The visibility impacts analysis.
2. What is the scope of the BART review?
  
  Once you determine that a source is subject to BART for a particular pollutant, then for each affected emission unit, you must establish BART for that pollutant. The BART determination must address air pollution control measures for each emissions unit or pollutant emitting activity subject to review.
  
  Example: Plantwide emissions from emission units within the listed categories that began operation within the ``time window'' for BART \11\ are 300 tons/yr of NOX, 200 tons/yr of SO2, and 150 tons/yr of primary particulate. Emissions unit A emits 200 tons/yr of NOX, 100 tons/yr of SO2, and 100 tons/yr of primary particulate. Other emission units, units B through H, which began operating in 1966, contribute lesser amounts of each pollutant. For this example, a BART review is required for NOX, SO2, and primary particulate, and control options must be analyzed for units B through H as well as unit A.
  
  \11\ That is, emission units that were in existence on August 7, 1977 and which began actual operation on or after August 7, 1962.
3. How does a BART review relate to Maximum Achievable Control Technology (MACT) Standards under CAA section 112, or to other emission limitations required under the CAA?
  
  For VOC and PM sources subject to MACT standards, States may streamline the analysis by including a discussion of the MACT controls and whether any major new technologies have been developed subsequent to the MACT standards. We believe that there are many VOC and PM sources that are well controlled because they are regulated by the MACT standards, which EPA developed under CAA section 112. For a few MACT standards, this may also be true for SO2. Any source subject to MACT standards must meet a level that is as stringent as the best-controlled 12 percent of sources in the industry. Examples of these hazardous air pollutant sources which effectively control VOC and PM emissions include (among others) secondary lead facilities, organic chemical plants subject to the hazardous organic NESHAP (HON), pharmaceutical production facilities, and equipment leaks and wastewater operations at petroleum refineries. We believe that, in many cases, it will be unlikely that States will identify emission controls more stringent than the MACT standards without
  
  [[Page 39164]]
  
  identifying control options that would cost many thousands of dollars per ton. Unless there are new technologies subsequent to the MACT standards which would lead to cost-effective increases in the level of control, you may rely on the MACT standards for purposes of BART.
  
  We believe that the same rationale also holds true for emissions standards developed for municipal waste incinerators under CAA section 111(d), and for many NSR/PSD determinations and NSR/PSD settlement agreements. However, we do not believe that technology determinations from the 1970s or early 1980s, including new source performance standards (NSPS), should be considered to represent best control for existing sources, as best control levels for recent plant retrofits are more stringent than these older levels.
  
  Where you are relying on these standards to represent a BART level of control, you should provide the public with a discussion of whether any new technologies have subsequently become available.
4. What Are the Five Basic Steps of a Case-by-Case BART Analysis?
  
  The five steps are:
  
  STEP 1--Identify All \12\ Available Retrofit Control Technologies,
  
  \12\ In identifying ``all'' options, you must identify the most stringent option and a reasonable set of options for analysis that reflects a comprehensive list of available technologies. It is not necessary to list all permutations of available control levels that exist for a given technology--the list is complete if it includes the maximum level of control each technology is capable of achieving.
  
  STEP 2-- Eliminate Technically Infeasible Options,
  
  STEP 3-- Evaluate Control Effectiveness of Remaining Control Technologies,
  
  STEP 4-- Evaluate Impacts and Document the Results, and
  
  STEP 5--Evaluate Visibility Impacts.
  1. STEP 1: How do I identify all available retrofit emission control techniques?
  2. Available retrofit control options are those air pollution control technologies with a practical potential for application to the emissions unit and the regulated pollutant under evaluation. Air pollution control technologies can include a wide variety of available methods, systems, and techniques for control of the affected pollutant. Technologies required as BACT or LAER are available for BART purposes and must be included as control alternatives. The control alternatives can include not only existing controls for the source category in question but also take into account technology transfer of controls that have been applied to similar source categories and gas streams. Technologies which have not yet been applied to (or permitted for) full scale operations need not be considered as available; we do not expect the source owner to purchase or construct a process or control device that has not already been demonstrated in practice.
  3. Where a NSPS exists for a source category (which is the case for most of the categories affected by BART), you should include a level of control equivalent to the NSPS as one of the control options.\13\ The NSPS standards are codified in 40 CFR part 60. We note that there are situations where NSPS standards do not require the most stringent level of available control for all sources within a category. For example, post-combustion NOXcontrols (the most stringent controls for stationary gas turbines) are not required under subpart GG of the NSPS for Stationary Gas Turbines. However, such controls must still be considered available technologies for the BART selection process.
    
    \13\ In EPA's 1980 BART guidelines for reasonably attributable visibility impairment, we concluded that NSPS standards generally, at that time, represented the best level sources could install as BART. In the 20 year period since this guidance was developed, there have been advances in SO2control technologies as well as technologies for the control of other pollutants, confirmed by a number of recent retrofits at Western power plants. Accordingly, EPA no longer concludes that the NSPS level of controls automatically represents ``the best these sources can install.'' Analysis of the BART factors could result in the selection of a NSPS level of control, but you should reach this conclusion only after considering the full range of control options.
  4. Potentially applicable retrofit control alternatives can be categorized in three ways.
    
    Pollution prevention: use of inherently lower-emitting processes/practices, including the use of control techniques (e.g. low-NOXburners) and work practices that prevent emissions and result in lower ``production-specific'' emissions (note that it is not our intent to direct States to switch fuel forms, e.g. from coal to gas),
    
    Use of (and where already in place, improvement in the performance of) add-on controls, such as scrubbers, fabric filters, thermal oxidizers and other devices that control and reduce emissions after they are produced, and
    
    Combinations of inherently lower-emitting processes and add-on controls.
  5. In the course of the BART review, one or more of the available control options may be eliminated from consideration because they are demonstrated to be technically infeasible or to have unacceptable energy, cost, or non-air quality environmental impacts on a case-by-case (or site-specific) basis. However, at the outset, you should initially identify all control options with potential application to the emissions unit under review.
  6. We do not consider BART as a requirement to redesign the source when considering available control alternatives. For example, where the source subject to BART is a coal-fired electric generator, we do not require the BART analysis to consider building a natural gas-fired electric turbine although the turbine may be inherently less polluting on a per unit basis.
  7. For emission units subject to a BART review, there will often be control measures or devices already in place. For such emission units, it is important to include control options that involve improvements to existing controls and not to limit the control options only to those measures that involve a complete replacement of control devices.
    
    Example: For a power plant with an existing wet scrubber, the current control efficiency is 66 percent. Part of the reason for the relatively low control efficiency is that 22 percent of the gas stream bypasses the scrubber. A BART review identifies options for improving the performance of the wet scrubber by redesigning the internal components of the scrubber and by eliminating or reducing the percentage of the gas stream that bypasses the scrubber. Four control options are identified: (1) 78 percent control based upon improved scrubber performance while maintaining the 22 percent bypass, (2) 83 percent control based upon improved scrubber performance while reducing the bypass to 15 percent, (3) 93 percent control based upon improving the scrubber performance while eliminating the bypass entirely, (this option results in a ``wet stack'' operation in which the gas leaving the stack is saturated with water) and (4) 93 percent as in option 3, with the addition of an indirect reheat system to reheat the stack gas above the saturation temperature. You must consider each of these four options in a BART analysis for this source.
  8. You are expected to identify potentially applicable retrofit control technologies that represent the full range of demonstrated alternatives. Examples of general information sources to consider include:
    
    The EPA's Clean Air Technology Center, which includes the RACT/BACT/LAER Clearinghouse (RBLC);
    
    State and Local Best Available Control Technology Guidelines--many agencies have online information--for example South Coast Air Quality Management District, Bay Area Air Quality Management District, and Texas Natural Resources Conservation Commission;
    
    Control technology vendors;
    
    Federal/State/Local NSR permits and associated inspection/performance test reports;
    
    Environmental consultants;
    
    Technical journals, reports and newsletters, air pollution control seminars; and
    
    The EPA's NSR bulletin board--http://www.epa.gov/ttn/nsr ;
    
    Department of Energy's Clean Coal Program--technical reports;
    
    The NOXControl Technology ``Cost Tool''-- Clean Air Markets Division Web page--http://www.epa.gov/airmarkets/arp/nox/controltech.html ;
    
    Performance of selective catalytic reduction on coal- fired steam generating units--final report. OAR/ARD, June 1997 (also available at http://www.epa.gov/airmarkets/arp/nox/controltech.html );
    
    Cost estimates for selected applications of NOXcontrol technologies on stationary combustion boilers. OAR/ARD June 1997. (Docket for NOXSIP Call, A- 96-56, item II-A-03);
    
    Investigation of performance and cost of NOX controls as applied to group 2 boilers. OAR/ARD, August 1996. (Docket for Phase II NOXrule, A-95-28, item IV-A-4);
    
    Controlling SO2Emissions: A Review of Technologies. EPA-600/R-00-093, USEPA/ORD/NRMRL, October 2000; and
    
    The OAQPS Control Cost Manual.
    
    You are expected to compile appropriate information from these information sources.
  9. There may be situations where a specific set of units within a fenceline constitutes the
    
    [[Page 39165]]
    
    logical set to which controls would apply and that set of units may or may not all be BART-eligible. (For example, some units in that set may not have been constructed between 1962 and 1977.)
  10. If you find that a BART source has controls already in place which are the most stringent controls available (note that this means that all possible improvements to any control devices have been made), then it is not necessary to comprehensively complete each following step of the BART analysis in this section. As long these most stringent controls available are made federally enforceable for the purpose of implementing BART for that source, you may skip the remaining analyses in this section, including the visibility analysis in step 5. Likewise, if a source commits to a BART determination that consists of the most stringent controls available, then there is no need to complete the remaining analyses in this section.
  11. STEP 2: How do I determine whether the options identified in Step 1 are technically feasible?
    
    In Step 2, you evaluate the technical feasibility of the control options you identified in Step 1. You should document a demonstration of technical infeasibility and should explain, based on physical, chemical, or engineering principles, why technical difficulties would preclude the successful use of the control option on the emissions unit under review. You may then eliminate such technically infeasible control options from further consideration in the BART analysis.
    
    In general, what do we mean by technical feasibility?
    
    Control technologies are technically feasible if either (1) they have been installed and operated successfully for the type of source under review under similar conditions, or (2) the technology could be applied to the source under review. Two key concepts are important in determining whether a technology could be applied: ``availability'' and ``applicability.'' As explained in more detail below, a technology is considered ``available'' if the source owner may obtain it through commercial channels, or it 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.
    
    What do we mean by ``available'' technology?
  12. The typical stages for bringing a control technology concept to reality as a commercial product are:
    
    Concept stage;
    
    Research and patenting;
    
    Bench scale or laboratory testing;
    
    Pilot scale testing;
    
    Licensing and commercial demonstration; and
    
    Commercial sales.
  13. A control technique is considered available, within the context presented above, if it has reached the stage of licensing and commercial availability. Similarly, we do not expect a source owner to conduct extended trials to learn how to apply a technology on a totally new and dissimilar source type. Consequently, you would not consider technologies in the pilot scale testing stages of development as ``available'' for purposes of BART review.
  14. Commercial availability by itself, however, is not necessarily a sufficient basis for concluding a technology to be applicable and therefore technically feasible. Technical feasibility, as determined in Step 2, also means a control option may reasonably be deployed on or ``applicable'' to the source type under consideration.
    
    Because a new technology may become available at various points in time during the BART analysis process, we believe that guidelines are needed on when a technology must be considered. For example, a technology may become available during the public comment period on the State's rule development process. Likewise, it is possible that new technologies may become available after the close of the State's public comment period and before submittal of the SIP to EPA, or during EPA's review process on the SIP submittal. In order to provide certainty in the process, all technologies should be considered if available before the close of the State's public comment period. You need not consider technologies that become available after this date. As part of your analysis, you should consider any technologies brought to your attention in public comments. If you disagree with public comments asserting that the technology is available, you should provide an explanation for the public record as to the basis for your conclusion.
    
    What do we mean by ``applicable'' technology?
    
    You need to exercise technical judgment in determining whether a control alternative is applicable to the source type under consideration. In general, a commercially available control option will be presumed applicable if it has been used on the same or a similar source type. Absent a showing of this type, you evaluate technical feasibility by examining the physical and chemical characteristics of the pollutant-bearing gas stream, and comparing them to the gas stream characteristics of the source types to which the technology had been applied previously. Deployment of the control technology on a new or existing source with similar gas stream characteristics is generally a sufficient basis for concluding the technology is technically feasible barring a demonstration to the contrary as described below.
    
    What type of demonstration is required if I conclude that an option is not technically feasible?
  15. Where you conclude that a control option identified in Step 1 is technically infeasible, you should demonstrate that the option is either commercially unavailable, or that specific circumstances preclude its application to a particular emission unit. Generally, such a demonstration involves an evaluation of the characteristics of the pollutant-bearing gas stream and the capabilities of the technology. Alternatively, a demonstration of technical infeasibility may involve a showing that there are unresolvable technical difficulties with applying the control to the source (e.g., size of the unit, location of the proposed site, operating problems related to specific circumstances of the source, space constraints, reliability, and adverse side effects on the rest of the facility). Where the resolution of technical difficulties is merely a matter of increased cost, you should consider the technology to be technically feasible. The cost of a control alternative is considered later in the process.
  16. The determination of technical feasibility is sometimes influenced by recent air quality permits. In some cases, an air quality permit may require a certain level of control, but the level of control in a permit is not expected to be achieved in practice (e.g., a source has received a permit but the project was canceled, or every operating source at that permitted level has been physically unable to achieve compliance with the limit). Where this is the case, you should provide supporting documentation showing why such limits are not technically feasible, and, therefore, why the level of control (but not necessarily the technology) may be eliminated from further consideration. However, if there is a permit requiring the application of a certain technology or emission limit to be achieved for such technology, this usually is sufficient justification for you to assume the technical feasibility of that technology or emission limit.
  17. Physical modifications needed to resolve technical obstacles do not, in and of themselves, provide a justification for eliminating the control technique on the basis of technical infeasibility. However, you may consider the cost of such modifications in estimating costs. This, in turn, may form the basis for eliminating a control technology (see later discussion).
  18. Vendor guarantees may provide an indication of commercial availability and the technical feasibility of a control technique and could contribute to a determination of technical feasibility or technical infeasibility, depending on circumstances. However, we do not consider a vendor guarantee alone to be sufficient justification that a control option will work. Conversely, lack of a vendor guarantee by itself does not present sufficient justification that a control option or an emissions limit is technically infeasible. Generally, you should make decisions about technical feasibility based on chemical, and engineering analyses (as discussed above), in conjunction with information about vendor guarantees.
  19. A possible outcome of the BART procedures discussed in these guidelines is the evaluation of multiple control technology alternatives which result in essentially equivalent emissions. It is not our intent to encourage evaluation of unnecessarily large numbers of control alternatives for every emissions unit. Consequently, you should use judgment in deciding on those alternatives for which you will conduct the detailed impacts analysis (Step 4 below). For example, if two or more control techniques result in control levels that are essentially identical, considering the uncertainties of emissions factors and other parameters
    
    [[Page 39166]]
    
    pertinent to estimating performance, you may evaluate only the less costly of these options. You should narrow the scope of the BART analysis in this way only if there is a negligible difference in emissions and energy and non-air quality environmental impacts between control alternatives.
  20. STEP 3: How do I evaluate technically feasible alternatives?
    
    Step 3 involves evaluating the control effectiveness of all the technically feasible control alternatives identified in Step 2 for the pollutant and emissions unit under review.
    
    Two key issues in this process include:
    
    (1) Making sure that you express the degree of control using a metric that ensures an ``apples to apples'' comparison of emissions performance levels among options, and
    
    (2) Giving appropriate treatment and consideration of control techniques that can operate over a wide range of emission performance levels.
    
    What are the appropriate metrics for comparison?
    
    This issue is especially important when you compare inherently lower-polluting processes to one another or to add-on controls. In such cases, it is generally most effective to express emissions performance as an average steady state emissions level per unit of product produced or processed.
    
    Examples of common metrics:
    
    Pounds of SO2emissions per million Btu heat input, and
    
    Pounds of NOXemissions per ton of cement produced.
    
    How do I evaluate control techniques with a wide range of emission performance levels?
  21. Many control techniques, including both add-on controls and inherently lower polluting processes, can perform at a wide range of levels. Scrubbers and high and low efficiency electrostatic precipitators (ESPs) are two of the many examples of such control techniques that can perform at a wide range of levels. It is not our intent to require analysis of each possible level of efficiency for a control technique as such an analysis would result in a large number of options. It is important, however, that in analyzing the technology you take into account the most stringent emission control level that the technology is capable of achieving. You should consider recent regulatory decisions and performance data (e.g., manufacturer's data, engineering estimates and the experience of other sources) when identifying an emissions performance level or levels to evaluate.
  22. In assessing the capability of the control alternative, latitude exists to consider special circumstances pertinent to the specific source under review, or regarding the prior application of the control alternative. However, you should explain the basis for choosing the alternate level (or range) of control in the BART analysis. Without a showing of differences between the source and other sources that have achieved more stringent emissions limits, you should conclude that the level being achieved by those other sources is representative of the achievable level for the source being analyzed.
  23. You may encounter cases where you may wish to evaluate other levels of control in addition to the most stringent level for a given device. While you must consider the most stringent level as one of the control options, you may consider less stringent levels of control as additional options. This would be useful, particularly in cases where the selection of additional options would have widely varying costs and other impacts.
  24. Finally, we note that for retrofitting existing sources in addressing BART, you should consider ways to improve the performance of existing control devices, particularly when a control device is not achieving the level of control that other similar sources are achieving in practice with the same device. For example, you should consider requiring those sources with electrostatic precipitators (ESPs) performing below currently achievable levels to improve their performance.
  25. STEP 4: For a BART review, what impacts am I expected to calculate and report? What methods does EPA recommend for the impacts analysis?
    
    After you identify the available and technically feasible control technology options, you are expected to conduct the following analyses when you make a BART determination:
    
    Impact analysis part 1: Costs of compliance,
    
    Impact analysis part 2: Energy impacts, and
    
    Impact analysis part 3: Non-air quality environmental impacts.
    
    Impact analysis part 4: Remaining useful life.
    
    In this section, we describe how to conduct each of these three analyses. You are responsible for presenting an evaluation of each impact along with appropriate supporting information. You should discuss and, where possible, quantify both beneficial and adverse impacts. In general, the analysis should focus on the direct impact of the control alternative.
    1. Impact analysis part 1: how do I estimate the costs of control?
  26. To conduct a cost analysis, you:
    
    (1) Identify the emissions units being controlled,
    
    (2) Identify design parameters for emission controls, and
    
    (3) Develop cost estimates based upon those design parameters.
  27. It is important to identify clearly the emission units being controlled, that is, to specify a well-defined area or process segment within the plant. In some cases, multiple emission units can be controlled jointly. However, in other cases, it may be appropriate in the cost analysis to consider whether multiple units will be required to install separate and/or different control devices. The analysis should provide a clear summary list of equipment and the associated control costs. Inadequate documentation of the equipment whose emissions are being controlled is a potential cause for confusion in comparison of costs of the same controls applied to similar sources.
  28. You then specify the control system design parameters. Potential sources of these design parameters include equipment vendors, background information documents used to support NSPS development, control technique guidelines documents, cost manuals developed by EPA, control data in trade publications, and engineering and performance test data. The following are a few examples of design parameters for two example control measures:
    
    Control device
    
    Examples of design parameters
    
    Wet Scrubbers......................... Type of sorbent used (lime, limestone, etc.). Gas pressure drop. Liquid/gas ratio. Selective Catalytic Reduction......... Ammonia to NOX molar ratio. Pressure drop. Catalyst life.
  29. The value selected for the design parameter should ensure that the control option will achieve the level of emission control being evaluated. You should include in your analysis documentation of your assumptions regarding design parameters. Examples of supporting references would include the EPA OAQPS Control Cost Manual (see below) and background information documents used for NSPS and hazardous pollutant emission standards. If the design parameters you specified differ from typical designs, you should document the difference by supplying performance test data for the control technology in question applied to the same source or a similar source.
  30. Once the control technology alternatives and achievable emissions performance levels have been identified, you then develop estimates of capital and annual costs. The basis for equipment cost estimates also should be documented, either with data supplied by an equipment vendor (i.e., budget estimates or bids) or by a referenced source (such as the OAQPS Control Cost Manual, Fifth Edition, February 1996, EPA 453/B-96-001).\14\ In order to maintain and improve consistency, cost estimates should be based on the OAQPS Control Cost Manual, where possible.\15\ The Control Cost Manual addresses most control technologies in sufficient detail for a BART analysis. The cost analysis should also take into account any site- specific design or other conditions identified above that affect the cost of a particular BART technology option.
    
    \14\ The OAQPS Control Cost Manual is updated periodically. While this citation refers to the latest version at the time this guidance was written, you should use the version that is current as of when you conduct your impact analysis. This document is available at the following Web site: http://www.epa.gov/ttn/catc/dir1/cs1ch2.pdf .
    
    \15\ You should include documentation for any additional information you used for the cost calculations, including any information supplied by vendors that affects your assumptions regarding purchased equipment costs, equipment life, replacement of major components, and any other element of the calculation that differs from the Control Cost Manual.
    
    [[Page 39167]]
    1. What do we mean by cost effectiveness?
      
      Cost effectiveness, in general, is a criterion used to assess the potential for achieving an objective in the most economical way. For purposes of air pollutant analysis, ``effectiveness'' is measured in terms of tons of pollutant emissions removed, and ``cost'' is measured in terms of annualized control costs. We recommend two types of cost-effectiveness calculations--average cost effectiveness, and incremental cost effectiveness.
    2. How do I calculate average cost effectiveness?
      
      Average cost effectiveness means the total annualized costs of control divided by annual emissions reductions (the difference between baseline annual emissions and the estimate of emissions after controls), using the following formula:
      
      Average cost effectiveness (dollars per ton removed) = Control option annualized cost \16\
      
      \16\ Whenever you calculate or report annual costs, you should indicate the year for which the costs are estimated. For example, if you use the year 2000 as the basis for cost comparisons, you would report that an annualized cost of $20 million would be: $20 million (year 2000 dollars).
      
      Baseline annual emissions--Annual emissions with Control option
      
      Because you calculate costs in (annualized) dollars per year ($/ yr) and because you calculate emissions rates in tons per year (tons/yr), the result is an average cost-effectiveness number in (annualized) dollars per ton ($/ton) of pollutant removed.
    3. How do I calculate baseline emissions?
  31. The baseline emissions rate should represent a realistic depiction of anticipated annual emissions for the source. In general, for the existing sources subject to BART, you will estimate the anticipated annual emissions based upon actual emissions from a baseline period.
  32. When you project that future operating parameters (e.g., limited hours of operation or capacity utilization, type of fuel, raw materials or product mix or type) will differ from past practice, and if this projection has a deciding effect in the BART determination, then you must make these parameters or assumptions into enforceable limitations. In the absence of enforceable limitations, you calculate baseline emissions based upon continuation of past practice.
  33. For example, the baseline emissions calculation for an emergency standby generator may consider the fact that the source owner would not operate more than past practice of 2 weeks a year. On the other hand, baseline emissions associated with a base-loaded turbine should be based on its past practice which would indicate a large number of hours of operation. This produces a significantly higher level of baseline emissions than in the case of the emergency/standby unit and results in more cost-effective controls. As a consequence of the dissimilar baseline emissions, BART for the two cases could be very different.
    1. How do I calculate incremental cost effectiveness?
  34. In addition to the average cost effectiveness of a control option, you should also calculate incremental cost effectiveness. You should consider the incremental cost effectiveness in combination with the average cost effectiveness when considering whether to eliminate a control option. The incremental cost effectiveness calculation compares the costs and performance level of a control option to those of the next most stringent option, as shown in the following formula (with respect to cost per emissions reduction):
    
    Incremental Cost Effectiveness (dollars per incremental ton removed) = (Total annualized costs of control option) - (Total annualized costs of next control option) / (Control option annual emissions) - (Next control option annual emissions)
    
    Example 1: Assume that Option F on Figure 2 has total annualized costs of $1 million to reduce 2000 tons of a pollutant, and that Option D on Figure 2 has total annualized costs of $500,000 to reduce 1000 tons of the same pollutant. The incremental cost effectiveness of Option F relative to Option D is ($1 million - $500,000) divided by (2000 tons - 1000 tons), or $500,000 divided by 1000 tons, which is $500/ton.
    
    Example 2: Assume that two control options exist: Option 1 and Option 2. Option 1 achieves a 1,000 ton/yr reduction at an annualized cost of $1,900,000. This represents an average cost of ($1,900,000/1,000 tons) = $1,900/ton. Option 2 achieves a 980 tons/ yr reduction at an annualized cost of $1,500,000. This represents an average cost of ($1,500,000/980 tons) = $1,531/ton. The incremental cost effectiveness of Option 1 relative to Option 2 is ($1,900,000 - $1,500,000) divided by (1,000 tons - 980 tons). The adoption of Option 1 instead of Option 2 results in an incremental emission reduction of 20 tons per year at an additional cost of $400,000 per year. The incremental cost of Option 1, then, is $20,000 per ton - 11 times the average cost of $1,900 per ton. While $1,900 per ton may still be deemed reasonable, it is useful to consider both the average and incremental cost in making an overall cost-effectiveness finding. Of course, there may be other differences between these options, such as, energy or water use, or non-air environmental effects, which also should be considered in selecting a BART technology.
  35. You should exercise care in deriving incremental costs of candidate control options. Incremental cost-effectiveness comparisons should focus on annualized cost and emission reduction differences between ``dominant'' alternatives. To identify dominant alternatives, you generate a graphical plot of total annualized costs for total emissions reductions for all control alternatives identified in the BART analysis, and by identifying a ``least-cost envelope'' as shown in Figure 2. (A ``least-cost envelope'' represents the set of options that should be dominant in the choice of a specific option.)
    
    [[Page 39168]]
    
    [GRAPHIC] [TIFF OMITTED] TR06JY05.000
    
    Example: Eight technically feasible control options for analysis are listed. These are represented as A through H in Figure 2. The dominant set of control options, B, D, F, G, and H, represent the least-cost envelope, as we depict by the cost curve connecting them. Points A, C and E are inferior options, and you should not use them in calculating incremental cost effectiveness. Points A, C and E represent inferior controls because B will buy more emissions reductions for less money than A; and similarly, D and F will buy more reductions for less money than C and E, respectively.
  36. In calculating incremental costs, you:
    
    (1) Array the control options in ascending order of annualized total costs,
    
    (2) Develop a graph of the most reasonable smooth curve of the control options, as shown in Figure 2. This is to show the ``least- cost envelope'' discussed above; and
    
    (3) Calculate the incremental cost effectiveness for each dominant option, which is the difference in total annual costs between that option and the next most stringent option, divided by the difference in emissions, after controls have been applied, between those two control options. For example, using Figure 2, you would calculate incremental cost effectiveness for the difference between options B and D, options D and F, options F and G, and options G and H.
  37. A comparison of incremental costs can also be useful in evaluating the viability of a specific control option over a range of efficiencies. For example, depending on the capital and operational cost of a control device, total and incremental cost may vary significantly (either increasing or decreasing) over the operational range of a control device. Also, the greater the number of possible control options that exist, the more weight should be given to the incremental costs vs. average costs. It should be noted that average and incremental cost effectiveness are identical when only one candidate control option is known to exist.
  38. You should exercise caution not to misuse these techniques. For example, you may be faced with a choice between two available control devices at a source, control A and control B, where control B achieves slightly greater emission reductions. The average cost (total annual cost/total annual emission reductions) for each may be deemed to be reasonable. However, the incremental cost (total annual costA - B/total annual emission reductionsA - B) of the additional emission reductions to be achieved by control B may be very great. In such an instance, it may be inappropriate to choose control B, based on its high incremental costs, even though its average cost may be considered reasonable.
  39. In addition, when you evaluate the average or incremental cost effectiveness of a control alternative, you should make reasonable and supportable assumptions regarding control efficiencies. An unrealistically low assessment of the emission reduction potential of a certain technology could result in inflated cost-effectiveness figures.
    1. What other information should I provide in the cost impacts analysis?
      
      You should provide documentation of any unusual circumstances that exist for the source that would lead to cost-effectiveness estimates that would exceed that for recent retrofits. This is especially important in cases where recent retrofits have cost- effectiveness values that are within what has been considered a reasonable range, but your analysis concludes that costs for the source being analyzed are not considered reasonable. (A reasonable range would be a range that is consistent with the range of cost effectiveness values used in other similar permit decisions over a period of time.)
      
      Example: In an arid region, large amounts of water are needed for a scrubbing system. Acquiring water from a distant location could greatly increase the cost per ton of emissions reduced of wet scrubbing as a control option.
    2. What other things are important to consider in the cost impacts analysis?
      
      In the cost analysis, you should take care not to focus on incomplete results or partial calculations. For example, large capital costs for a control option alone would not preclude selection of a control measure if large emissions reductions are projected. In such a case, low or reasonable cost effectiveness numbers may validate the option as an appropriate BART alternative irrespective of the large capital costs. Similarly, projects with relatively low capital costs may not be cost effective if there are few emissions reduced.
    3. Impact analysis part 2: How should I analyze and report energy impacts?
  40. You should examine the energy requirements of the control technology and determine whether the use of that technology results in energy penalties or benefits. A source owner may, for example, benefit from the combustion of a concentrated gas stream rich in volatile organic compounds; on the other hand, more often extra fuel or electricity is required to power a control
    
    [[Page 39169]]
    
    device or incinerate a dilute gas stream. If such benefits or penalties exist, they should be quantified to the extent practicable. Because energy penalties or benefits can usually be quantified in terms of additional cost or income to the source, the energy impacts analysis can, in most cases, simply be factored into the cost impacts analysis. The fact of energy use in and of itself does not disqualify a technology.
  41. Your energy impact analysis should consider only direct energy consumption and not indirect energy impacts. For example, you could estimate the direct energy impacts of the control alternative in units of energy consumption at the source (e.g., BTU, kWh, barrels of oil, tons of coal). The energy requirements of the control options should be shown in terms of total (and in certain cases, also incremental) energy costs per ton of pollutant removed. You can then convert these units into dollar costs and, where appropriate, factor these costs into the control cost analysis.
  42. You generally do not consider indirect energy impacts (such as energy to produce raw materials for construction of control equipment). However, if you determine, either independently or based on a showing by the source owner, that the indirect energy impact is unusual or significant and that the impact can be well quantified, you may consider the indirect impact.
  43. The energy impact analysis may also address concerns over the use of locally scarce fuels. The designation of a scarce fuel may vary from region to region. However, in general, a scarce fuel is one which is in short supply locally and can be better used for alternative purposes, or one which may not be reasonably available to the source either at the present time or in the near future.
  44. Finally, the energy impacts analysis may consider whether there are relative differences between alternatives regarding the use of locally or regionally available coal, and whether a given alternative would result in significant economic disruption or unemployment. For example, where two options are equally cost effective and achieve equivalent or similar emissions reductions, one option may be preferred if the other alternative results in significant disruption or unemployment.
    1. Impact analysis part 3: How do I analyze ``non-air quality environmental impacts?''
  45. In the non-air quality related environmental impacts portion of the BART analysis, you address environmental impacts other than air quality due to emissions of the pollutant in question. Such environmental impacts include solid or hazardous waste generation and discharges of polluted water from a control device.
  46. You should identify any significant or unusual environmental impacts associated with a control alternative that have the potential to affect the selection or elimination of a control alternative. Some control technologies may have potentially significant secondary environmental impacts. Scrubber effluent, for example, may affect water quality and land use. Alternatively, water availability may affect the feasibility and costs of wet scrubbers. Other examples of secondary environmental impacts could include hazardous waste discharges, such as spent catalysts or contaminated carbon. Generally, these types of environmental concerns become important when sensitive site-specific receptors exist or when the incremental emissions reductions potential of the more stringent control is only marginally greater than the next most-effective option. However, the fact that a control device creates liquid and solid waste that must be disposed of does not necessarily argue against selection of that technology as BART, particularly if the control device has been applied to similar facilities elsewhere and the solid or liquid waste is similar to those other applications. On the other hand, where you or the source owner can show that unusual circumstances at the proposed facility create greater problems than experienced elsewhere, this may provide a basis for the elimination of that control alternative as BART.
  47. The procedure for conducting an analysis of non-air quality environmental impacts should be made based on a consideration of site-specific circumstances. If you propose to adopt the most stringent alternative, then it is not necessary to perform this analysis of environmental impacts for the entire list of technologies you ranked in Step 3. In general, the analysis need only address those control alternatives with any significant or unusual environmental impacts that have the potential to affect the selection of a control alternative, or elimination of a more stringent control alternative. Thus, any important relative environmental impacts (both positive and negative) of alternatives can be compared with each other.
  48. In general, the analysis of impacts starts with the identification and quantification of the solid, liquid, and gaseous discharges from the control device or devices under review. Initially, you should perform a qualitative or semi-quantitative screening to narrow the analysis to discharges with potential for causing adverse environmental effects. Next, you should assess the mass and composition of any such discharges and quantify them to the extent possible, based on readily available information. You should also assemble pertinent information about the public or environmental consequences of releasing these materials.
    1. Impact analysis part 4: What are examples of non-air quality environmental impacts?
      
      The following are examples of how to conduct non-air quality environmental impacts:
      
      (1) Water Impact
      
      You should identify the relative quantities of water used and water pollutants produced and discharged as a result of the use of each alternative emission control system. Where possible, you should assess the effect on ground water and such local surface water quality parameters as ph, turbidity, dissolved oxygen, salinity, toxic chemical levels, temperature, and any other important considerations. The analysis could consider whether applicable water quality standards will be met and the availability and effectiveness of various techniques to reduce potential adverse effects.
      
      (2) Solid Waste Disposal Impact
      
      You could also compare the quality and quantity of solid waste (e.g., sludges, solids) that must be stored and disposed of or recycled as a result of the application of each alternative emission control system. You should consider the composition and various other characteristics of the solid waste (such as permeability, water retention, rewatering of dried material, compression strength, leachability of dissolved ions, bulk density, ability to support vegetation growth and hazardous characteristics) which are significant with regard to potential surface water pollution or transport into and contamination of subsurface waters or aquifers.
      
      (3) Irreversible or Irretrievable Commitment of Resources
      
      You may consider the extent to which the alternative emission control systems may involve a trade-off between short-term environmental gains at the expense of long-term environmental losses and the extent to which the alternative systems may result in irreversible or irretrievable commitment of resources (for example, use of scarce water resources).
      
      (4) Other Adverse Environmental Impacts
      
      You may consider significant differences in noise levels, radiant heat, or dissipated static electrical energy of pollution control alternatives. Other examples of non-air quality environmental impacts would include hazardous waste discharges such as spent catalysts or contaminated carbon.
    2. How do I take into account a project's ``remaining useful life'' in calculating control costs?
  49. You may decide to treat the requirement to consider the source's ``remaining useful life'' of the source for BART determinations as one element of the overall cost analysis. The ``remaining useful life'' of a source, if it represents a relatively short time period, may affect the annualized costs of retrofit controls. For example, the methods for calculating annualized costs in EPA's OAQPS Control Cost Manual require the use of a specified time period for amortization that varies based upon the type of control. If the remaining useful life will clearly exceed this time period, the remaining useful life has essentially no effect on control costs and on the BART determination process. Where the remaining useful life is less than the time period for amortizing costs, you should use this shorter time period in your cost calculations.
  50. For purposes of these guidelines, the remaining useful life is the difference between:
    
    (1) The date that controls will be put in place (capital and other construction costs incurred before controls are put in place can be rolled into the first year, as suggested in EPA's OAQPS Control Cost Manual); you are conducting the BART analysis; and
    
    (2) The date the facility permanently stops operations. Where this affects the BART determination, this date should be assured by a federally- or State-enforceable restriction preventing further operation.
  51. We recognize that there may be situations where a source operator intends to shut down a source by a given date, but wishes to retain the flexibility to continue
    
    [[Page 39170]]
    
    operating beyond that date in the event, for example, that market conditions change. Where this is the case, your BART analysis may account for this, but it must maintain consistency with the statutory requirement to install BART within 5 years. Where the source chooses not to accept a federally enforceable condition requiring the source to shut down by a given date, it is necessary to determine whether a reduced time period for the remaining useful life changes the level of controls that would have been required as BART.
    
    If the reduced time period does change the level of BART controls, you may identify, and include as part of the BART emission limitation, the more stringent level of control that would be required as BART if there were no assumption that reduced the remaining useful life. You may incorporate into the BART emission limit this more stringent level, which would serve as a contingency should the source continue operating more than 5 years after the date EPA approves the relevant SIP. The source would not be allowed to operate after the 5-year mark without such controls. If a source does operate after the 5-year mark without BART in place, the source is considered to be in violation of the BART emissions limit for each day of operation.
  52. Step 5: How should I determine visibility impacts in the BART determination?
    
    The following is an approach you may use to determine visibility impacts (the degree of visibility improvement for each source subject to BART) for the BART determination. Once you have determined that your source or sources are subject to BART, you must conduct a visibility improvement determination for the source(s) as part of the BART determination. When making this determination, we believe you have flexibility in setting absolute thresholds, target levels of improvement, or de minimis levels since the deciview improvement must be weighed among the five factors, and you are free to determine the weight and significance to be assigned to each factor. For example, a 0.3 deciview improvement may merit a stronger weighting in one case versus another, so one ``bright line'' may not be appropriate. [Note that if sources have elected to apply the most stringent controls available, consistent with the discussion in section E. step 1. below, you need not conduct, or require the source to conduct, an air quality modeling analysis for the purpose of determining its visibility impacts.]
    
    Use CALPUFF,\17\ or other appropriate dispersion model to determine the visibility improvement expected at a Class I area from the potential BART control technology applied to the source. Modeling should be conducted for SO2, NOX, and direct PM emissions (PM2.5and/or PM10). If the source is making the visibility determination, you should review and approve or disapprove of the source's analysis before making the expected improvement determination. There are several steps for determining the visibility impacts from an individual source using a dispersion model:
    
    \17\ The model code and its documentation are available at no cost for download from http://www.epa.gov/scram001/tt22.htm#calpuff.
    
    Develop a modeling protocol.
    
    Some critical items to include in a modeling protocol are meteorological and terrain data, as well as source-specific information (stack height, temperature, exit velocity, elevation, and allowable and actual emission rates of applicable pollutants), and receptor data from appropriate Class I areas. We recommend following EPA's Interagency Workgroup on Air Quality Modeling (IWAQM) Phase 2 Summary Report and Recommendations for Modeling Long Range Transport Impacts \18\ for parameter settings and meteorological data inputs; the use of other settings from those in IWAQM should be identified and explained in the protocol.
    
    \18\ Interagency Workgroup on Air Quality Modeling (IWAQM) Phase 2 Summary Report and Recommendations for Modeling Long Range Transport Impacts, U.S. Environmental Protection Agency, EPA-454/R- 98-019, December 1998.
    
    One important element of the protocol is in establishing the receptors that will be used in the model. The receptors that you use should be located in the nearest Class I area with sufficient density to identify the likely visibility effects of the source. For other Class I areas in relatively close proximity to a BART-eligible source, you may model a few strategic receptors to determine whether effects at those areas may be greater than at the nearest Class I area. For example, you might chose to locate receptors at these areas at the closest point to the source, at the highest and lowest elevation in the Class I area, at the IMPROVE monitor, and at the approximate expected plume release height. If the highest modeled effects are observed at the nearest Class I area, you may choose not to analyze the other Class I areas any further as additional analyses might be unwarranted.
    
    You should bear in mind that some receptors within the relevant Class I area may be less than 50 km from the source while other receptors within that same Class I area may be greater than 50 km from the same source. As indicated by the Guideline on Air Quality Models, this situation may call for the use of two different modeling approaches for the same Class I area and source, depending upon the State's chosen method for modeling sources less than 50 km. In situations where you are assessing visibility impacts for source- receptor distances less than 50 km, you should use expert modeling judgment in determining visibility impacts, giving consideration to both CALPUFF and other EPA-approved methods.
    
    In developing your modeling protocol, you may want to consult with EPA and your regional planning organization (RPO). Up-front consultation will ensure that key technical issues are addressed before you conduct your modeling.
    
    For each source, run the model, at pre-control and post-control emission rates according to the accepted methodology in the protocol.
    
    Use the 24-hour average actual emission rate from the highest emitting day of the meteorological period modeled (for the pre- control scenario). Calculate the model results for each receptor as the change in deciviews compared against natural visibility conditions. Post-control emission rates are calculated as a percentage of pre-control emission rates. For example, if the 24-hr pre-control emission rate is 100 lb/hr of SO2, then the post control rate is 5 lb/hr if the control efficiency being evaluated is 95 percent.
    
    Make the net visibility improvement determination.
    
    Assess the visibility improvement based on the modeled change in visibility impacts for the pre-control and post-control emission scenarios. You have flexibility to assess visibility improvements due to BART controls by one or more methods. You may consider the frequency, magnitude, and duration components of impairment. Suggestions for making the determination are:
    
    Use of a comparison threshold, as is done for determining if BART-eligible sources should be subject to a BART determination. Comparison thresholds can be used in a number of ways in evaluating visibility improvement (e.g. the number of days or hours that the threshold was exceeded, a single threshold for determining whether a change in impacts is significant, or a threshold representing an x percent change in improvement).
    
    Compare the 98th percent days for the pre- and post- control runs.
    
    Note that each of the modeling options may be supplemented with source apportionment data or source apportionment modeling.
5. How do I select the ``best'' alternative, using the results of Steps 1 through 5?
  1. Summary of the Impacts Analysis
    
    From the alternatives you evaluated in Step 3, we recommend you develop a chart (or charts) displaying for each of the alternatives:
    
    (1) Expected emission rate (tons per year, pounds per hour);
    
    (2) Emissions performance level (e.g., percent pollutant removed, emissions per unit product, lb/MMBtu, ppm);
    
    (3) Expected emissions reductions (tons per year);
    
    (4) Costs of compliance--total annualized costs ($), cost effectiveness ($/ton), and incremental cost effectiveness ($/ton), and/or any other cost-effectiveness measures (such as $/deciview);
    
    (5) Energy impacts;
    
    (6) Non-air quality environmental impacts; and
    
    (7) Modeled visibility impacts.
  2. Selecting a ``best'' alternative
  3. You have discretion to determine the order in which you should evaluate control options for BART. Whatever the order in which you choose to evaluate options, you should always (1) display the options evaluated; (2) identify the average and incremental costs of each option; (3) consider the energy and non-air quality environmental impacts of each option; (4) consider the remaining useful life; and (5) consider the modeled visibility impacts. You should provide a justification for adopting the technology that you select as the ``best'' level of control, including an explanation of the
    
    [[Page 39171]]
    
    CAA factors that led you to choose that option over other control levels.
  4. In the case where you are conducting a BART determination for two regulated pollutants on the same source, if the result is two different BART technologies that do not work well together, you could then substitute a different technology or combination of technologies.
  5. In selecting a ``best'' alternative, should I consider the affordability of controls?
  6. Even if the control technology is cost effective, there may be cases where the installation of controls would affect the viability of continued plant operations.
  7. There may be unusual circumstances that justify taking into consideration the conditions of the plant and the economic effects of requiring the use of a given control technology. These effects would include effects on product prices, the market share, and profitability of the source. Where there are such unusual circumstances that are judged to affect plant operations, you may take into consideration the conditions of the plant and the economic effects of requiring the use of a control technology. Where these effects are judged to have a severe impact on plant operations you may consider them in the selection process, but you may wish to provide an economic analysis that demonstrates, in sufficient detail for public review, the specific economic effects, parameters, and reasoning. (We recognize that this review process must preserve the confidentiality of sensitive business information). Any analysis may also consider whether other competing plants in the same industry have been required to install BART controls if this information is available.
  8. Sulfur dioxide limits for utility boilers
    
    You must require 750 MW power plants to meet specific control levels for SO2of either 95 percent control or 0.15 lbs/ MMBtu, for each EGU greater than 200 MW that is currently uncontrolled unless you determine that an alternative control level is justified based on a careful consideration of the statutory factors. Thus, for example, if the source demonstrates circumstances affecting its ability to cost-effectively reduce its emissions, you should take that into account in determining whether the presumptive levels of control are appropriate for that facility. For a currently uncontrolled EGU greater than 200 MW in size, but located at a power plant smaller than 750 MW in size, such controls are generally cost- effective and could be used in your BART determination considering the five factors specified in CAA section 169A(g)(2). While these levels may represent current control capabilities, we expect that scrubber technology will continue to improve and control costs continue to decline. You should be sure to consider the level of control that is currently best achievable at the time that you are conducting your BART analysis.
    
    For coal-fired EGUs with existing post-combustion SO2 controls achieving less than 50 percent removal efficiencies, we recommend that you evaluate constructing a new FGD system to meet the same emission limits as above (95 percent removal or 0.15 lb/ mmBtu), in addition to the evaluation of scrubber upgrades discussed below. For oil-fired units, regardless of size, you should evaluate limiting the sulfur content of the fuel oil burned to 1 percent or less by weight.
    
    For those BART-eligible EGUs with pre-existing post-combustion SO2controls achieving removal efficiencies of at least 50 percent, your BART determination should consider cost effective scrubber upgrades designed to improve the system's overall SO2removal efficiency. There are numerous scrubber enhancements available to upgrade the average removal efficiencies of all types of existing scrubber systems. We recommend that as you evaluate the definition of ``upgrade,'' you evaluate options that not only improve the design removal efficiency of the scrubber vessel itself, but also consider upgrades that can improve the overall SO2removal efficiency of the scrubber system. Increasing a scrubber system's reliability, and conversely decreasing its downtime, by way of optimizing operation procedures, improving maintenance practices, adjusting scrubber chemistry, and increasing auxiliary equipment redundancy, are all ways to improve average SO2removal efficiencies.
    
    We recommend that as you evaluate the performance of existing wet scrubber systems, you consider some of the following upgrades, in no particular order, as potential scrubber upgrades that have been proven in the industry as cost effective means to increase overall SO2removal of wet systems:
    
    (a) Elimination of Bypass Reheat;
    
    (b) Installation of Liquid Distribution Rings;
    
    (c) Installation of Perforated Trays;
    
    (d) Use of Organic Acid Additives;
    
    (e) Improve or Upgrade Scrubber Auxiliary System Equipment;
    
    (f) Redesign Spray Header or Nozzle Configuration.
    
    We recommend that as you evaluate upgrade options for dry scrubber systems, you should consider the following cost effective upgrades, in no particular order:
    
    (a) Use of Performance Additives;
    
    (b) Use of more Reactive Sorbent;
    
    (c) Increase the Pulverization Level of Sorbent;
    
    (d) Engineering redesign of atomizer or slurry injection system.
    
    You should evaluate scrubber upgrade options based on the 5 step BART analysis process.
  9. Nitrogen oxide limits for utility boilers
    
    You should establish specific numerical limits for NOXcontrol for each BART determination. For power plants with a generating capacity in excess of 750 MW currently using selective catalytic reduction (SCR) or selective non-catalytic reduction (SNCR) for part of the year, you should presume that use of those same controls year-round is BART. For other sources currently using SCR or SNCR to reduce NOXemissions during part of the year, you should carefully consider requiring the use of these controls year-round as the additional costs of operating the equipment throughout the year would be relatively modest.
    
    For coal-fired EGUs greater than 200 MW located at greater than 750 MW power plants and operating without post-combustion controls (i.e. SCR or SNCR), we have provided presumptive NOX limits, differentiated by boiler design and type of coal burned. You may determine that an alternative control level is appropriate based on a careful consideration of the statutory factors. For coal-fired EGUs greater than 200 MW located at power plants 750 MW or less in size and operating without post-combustion controls, you should likewise presume that these same levels are cost-effective. You should require such utility boilers to meet the following NOXemission limits, unless you determine that an alternative control level is justified based on consideration of the statutory factors. The following NOXemission rates were determined based on a number of assumptions, including that the EGU boiler has enough volume to allow for installation and effective operation of separated overfire air ports. For boilers where these assumptions are incorrect, these emission limits may not be cost- effective.
    
    [[Page 39172]]
    
    Table 1.--Presumptive NOX Emission Limits for BART-Eligible Coal-Fired Units.\19\
    
    NOX presumptive Unit type
    
    Coal type
    
    limit (lb/ mmbtu) \20\
    
    Dry-bottom wall-fired......... Bituminous............
    
    0.39 Sub-bituminous........
    
    0.23 Lignite...............
    
    0.29 Tangential-fired.............. Bituminous............
    
    0.28 Sub-bituminous........
    
    0.15 Lignite...............
    
    0.17 Cell Burners.................. Bituminous............
    
    0.40 Sub-bituminous........
    
    0.45 Dry-turbo-fired............... Bituminous............
    
    0.32 Sub-bituminous........
    
    0.23 Wet-bottom tangential-fired... Bituminous............
    
    0.62
    
    Most EGUs can meet these presumptive NOXlimits through the use of current combustion control technology, i.e. the careful control of combustion air and low-NOXburners. For units that cannot meet these limits using such technologies, you should consider whether advanced combustion control technologies such as rotating opposed fire air should be used to meet these limits.
    
    \19\ No Cell burners, dry-turbo-fired units, nor wet-bottom tangential-fired units burning lignite were identified as BART- eligible, thus no presumptive limit was determined. Similarly, no wet-bottom tangential-fired units burning sub-bituminous were identified as BART-eligible.
    
    \20\ These limits reflect the design and technological assumptions discussed in the technical support document for NOXlimits for these guidelines. See Technical Support Document for BART NOX Limits for Electric Generating Units and Technical Support Document for BART NOX Limits for Electric Generating Units Excel Spreadsheet, Memorandum to Docket OAR 2002- 0076, April 15, 2005.
    
    Because of the relatively high NOXemission rates of cyclone units, SCR is more cost-effective than the use of current combustion control technology for these units. The use of SCRs at cyclone units burning bituminous coal, sub-bituminous coal, and lignite should enable the units to cost-effectively meet NOXrates of 0.10 lbs/mmbtu. As a result, we are establishing a presumptive NOXlimit of 0.10 lbs/mmbtu based on the use of SCR for coal-fired cyclone units greater than 200 MW located at 750 MW power plants. As with the other presumptive limits established in this guideline, you may determine that an alternative level of control is appropriate based on your consideration of the relevant statutory factors. For other cyclone units, you should review the use of SCR and consider whether these post-combustion controls should be required as BART.
    
    For oil-fired and gas-fired EGUs larger than 200MW, we believe that installation of current combustion control technology to control NOXis generally highly cost-effective and should be considered in your determination of BART for these sources. Many such units can make significant reductions in NOX emissions which are highly cost-effective through the application of current combustion control technology.\21\
    
    \21\ See Technical Support Document for BART NOX Limits for Electric Generating Units and Technical Support Document for BART NOX Limits for Electric Generating Units Excel Spreadsheet, Memorandum to Docket OAR 2002-0076, April 15, 2005.
Enforceable Limits/Compliance Date

To complete the BART process, you must establish enforceable emission limits that reflect the BART requirements and require compliance within a given period of time. In particular, you must establish an enforceable emission limit for each subject emission unit at the source and for each pollutant subject to review that is emitted from the source. In addition, you must require compliance with the BART emission limitations no later than 5 years after EPA approves your regional haze SIP. If technological or economic limitations in the application of a measurement methodology to a particular emission unit make a conventional emissions limit infeasible, you may instead prescribe a design, equipment, work practice, operation standard, or combination of these types of standards. You should consider allowing sources to ``average'' emissions across any set of BART-eligible emission units within a fenceline, so long as the emission reductions from each pollutant being controlled for BART would be equal to those reductions that would be obtained by simply controlling each of the BART-eligible units that constitute BART-eligible source.

You should ensure that any BART requirements are written in a way that clearly specifies the individual emission unit(s) subject to BART regulation. Because the BART requirements themselves are ``applicable'' requirements of the CAA, they must be included as title V permit conditions according to the procedures established in 40 CFR part 70 or 40 CFR part 71.

Section 302(k) of the CAA requires emissions limits such as BART to be met on a continuous basis. Although this provision does not necessarily require the use of continuous emissions monitoring (CEMs), it is important that sources employ techniques that ensure compliance on a continuous basis. Monitoring requirements generally applicable to sources, including those that are subject to BART, are governed by other regulations. See, e.g., 40 CFR part 64 (compliance assurance monitoring); 40 CFR 70.6(a)(3) (periodic monitoring); 40 CFR 70.6(c)(1) (sufficiency monitoring). Note also that while we do not believe that CEMs would necessarily be required for all BART sources, the vast majority of electric generating units potentially subject to BART already employ CEM technology for other programs, such as the acid rain program. In addition, emissions limits must be enforceable as a practical matter (contain appropriate averaging times, compliance verification procedures and recordkeeping requirements). In light of the above, the permit must:

Be sufficient to show compliance or noncompliance (i.e., through monitoring times of operation, fuel input, or other indices of operating conditions and practices); and

Specify a reasonable averaging time consistent with established reference methods, contain reference methods for determining compliance, and provide for adequate reporting and recordkeeping so that air quality agency personnel can determine the compliance status of the source; and

For EGUS, specify an averaging time of a 30-day rolling average, and contain a definition of ``boiler operating day'' that is consistent with the definition in the proposed revisions to the NSPS for utility boilers in 40 CFR Part 60, subpart Da.\22\ You should consider a boiler operating day to be any 24-hour period between 12:00 midnight and the following midnight during which any fuel is combusted at any time at the steam generating unit. This would allow 30-day rolling average emission rates to be calculated consistently across sources.

\22\ 70 FR 9705, February 28, 2005.

[FR Doc. 05-12526 Filed 7-5-05; 8:45 am]

BILLING CODE 6560-50-P

Subscribers can access the reported version of this case.

You can sign up for a trial and make the most of our service including these benefits.

Request your trial

Why Sign-up to vLex?

Over 100 Countries

Search over 120 million documents from over 100 countries including primary and secondary collections of legislation, case law, regulations, practical law, news, forms and contracts, books, journals, and more.
Thousands of Data Sources

Updated daily, vLex brings together legal information from over 750 publishing partners, providing access to over 2,500 legal and news sources from the world’s leading publishers.
Find What You Need, Quickly

Advanced A.I. technology developed exclusively by vLex editorially enriches legal information to make it accessible, with instant translation into 14 languages for enhanced discoverability and comparative research.
Over 2 million registered users

Founded over 20 years ago, vLex provides a first-class and comprehensive service for lawyers, law firms, government departments, and law schools around the world.

Subscribers are able to see a list of all the cited cases and legislation of a document.

You can sign up for a trial and make the most of our service including these benefits.

Request your trial

Why Sign-up to vLex?

Over 100 Countries

Search over 120 million documents from over 100 countries including primary and secondary collections of legislation, case law, regulations, practical law, news, forms and contracts, books, journals, and more.
Thousands of Data Sources

Updated daily, vLex brings together legal information from over 750 publishing partners, providing access to over 2,500 legal and news sources from the world’s leading publishers.
Find What You Need, Quickly

Advanced A.I. technology developed exclusively by vLex editorially enriches legal information to make it accessible, with instant translation into 14 languages for enhanced discoverability and comparative research.
Over 2 million registered users

Founded over 20 years ago, vLex provides a first-class and comprehensive service for lawyers, law firms, government departments, and law schools around the world.

Subscribers are able to see a list of all the documents that have cited the case.

You can sign up for a trial and make the most of our service including these benefits.

Request your trial

Why Sign-up to vLex?

Over 100 Countries

Search over 120 million documents from over 100 countries including primary and secondary collections of legislation, case law, regulations, practical law, news, forms and contracts, books, journals, and more.
Thousands of Data Sources

Updated daily, vLex brings together legal information from over 750 publishing partners, providing access to over 2,500 legal and news sources from the world’s leading publishers.
Find What You Need, Quickly

Advanced A.I. technology developed exclusively by vLex editorially enriches legal information to make it accessible, with instant translation into 14 languages for enhanced discoverability and comparative research.
Over 2 million registered users

Founded over 20 years ago, vLex provides a first-class and comprehensive service for lawyers, law firms, government departments, and law schools around the world.

Subscribers are able to see the revised versions of legislation with amendments.

You can sign up for a trial and make the most of our service including these benefits.

Request your trial

Why Sign-up to vLex?

Over 100 Countries

Search over 120 million documents from over 100 countries including primary and secondary collections of legislation, case law, regulations, practical law, news, forms and contracts, books, journals, and more.
Thousands of Data Sources

Updated daily, vLex brings together legal information from over 750 publishing partners, providing access to over 2,500 legal and news sources from the world’s leading publishers.
Find What You Need, Quickly

Advanced A.I. technology developed exclusively by vLex editorially enriches legal information to make it accessible, with instant translation into 14 languages for enhanced discoverability and comparative research.
Over 2 million registered users

Founded over 20 years ago, vLex provides a first-class and comprehensive service for lawyers, law firms, government departments, and law schools around the world.

Subscribers are able to see any amendments made to the case.

You can sign up for a trial and make the most of our service including these benefits.

Request your trial

Why Sign-up to vLex?

Over 100 Countries

Search over 120 million documents from over 100 countries including primary and secondary collections of legislation, case law, regulations, practical law, news, forms and contracts, books, journals, and more.
Thousands of Data Sources

Updated daily, vLex brings together legal information from over 750 publishing partners, providing access to over 2,500 legal and news sources from the world’s leading publishers.
Find What You Need, Quickly

Advanced A.I. technology developed exclusively by vLex editorially enriches legal information to make it accessible, with instant translation into 14 languages for enhanced discoverability and comparative research.
Over 2 million registered users

Founded over 20 years ago, vLex provides a first-class and comprehensive service for lawyers, law firms, government departments, and law schools around the world.

Subscribers are able to see a visualisation of a case and its relationships to other cases. An alternative to lists of cases, the Precedent Map makes it easier to establish which ones may be of most relevance to your research and prioritise further reading. You also get a useful overview of how the case was received.

Request your trial

Why Sign-up to vLex?

Over 100 Countries

Search over 120 million documents from over 100 countries including primary and secondary collections of legislation, case law, regulations, practical law, news, forms and contracts, books, journals, and more.
Thousands of Data Sources

Updated daily, vLex brings together legal information from over 750 publishing partners, providing access to over 2,500 legal and news sources from the world’s leading publishers.
Find What You Need, Quickly

Advanced A.I. technology developed exclusively by vLex editorially enriches legal information to make it accessible, with instant translation into 14 languages for enhanced discoverability and comparative research.
Over 2 million registered users

Founded over 20 years ago, vLex provides a first-class and comprehensive service for lawyers, law firms, government departments, and law schools around the world.

Subscribers are able to see the list of results connected to your document through the topics and citations Vincent found.

You can sign up for a trial and make the most of our service including these benefits.

Request your trial

Why Sign-up to vLex?

Over 100 Countries

Search over 120 million documents from over 100 countries including primary and secondary collections of legislation, case law, regulations, practical law, news, forms and contracts, books, journals, and more.
Thousands of Data Sources

Updated daily, vLex brings together legal information from over 750 publishing partners, providing access to over 2,500 legal and news sources from the world’s leading publishers.
Find What You Need, Quickly

Advanced A.I. technology developed exclusively by vLex editorially enriches legal information to make it accessible, with instant translation into 14 languages for enhanced discoverability and comparative research.
Over 2 million registered users

Founded over 20 years ago, vLex provides a first-class and comprehensive service for lawyers, law firms, government departments, and law schools around the world.

Part III

You can sign up for a trial and make the most of our service including these benefits.

Why Sign-up to vLex?

Over 100 Countries

Thousands of Data Sources

Find What You Need, Quickly

Over 2 million registered users

You can sign up for a trial and make the most of our service including these benefits.

Why Sign-up to vLex?

Over 100 Countries

Thousands of Data Sources

Find What You Need, Quickly

Over 2 million registered users

You can sign up for a trial and make the most of our service including these benefits.

Why Sign-up to vLex?

Over 100 Countries

Thousands of Data Sources

Find What You Need, Quickly

Over 2 million registered users

You can sign up for a trial and make the most of our service including these benefits.

Why Sign-up to vLex?

Over 100 Countries

Thousands of Data Sources

Find What You Need, Quickly

Over 2 million registered users

You can sign up for a trial and make the most of our service including these benefits.

Why Sign-up to vLex?

Over 100 Countries

Thousands of Data Sources

Find What You Need, Quickly

Over 2 million registered users

Why Sign-up to vLex?

Over 100 Countries

Thousands of Data Sources

Find What You Need, Quickly

Over 2 million registered users

You can sign up for a trial and make the most of our service including these benefits.

Why Sign-up to vLex?

Over 100 Countries

Thousands of Data Sources

Find What You Need, Quickly

Over 2 million registered users