In the Supreme Court of the United StatesIn the Supreme Court of the United StatesIn the Supreme Court of the United StatesIn the Supreme Court of the United StatesIn the Supreme Court of the United States

STATE OF OKLAHOMA; OKLAHOMAINDUSTRIAL ENERGY CONSUMERS;

OKLAHOMA GAS AND ELECTRIC COMPANY,Petitioners,

v.

UNITED STATES ENVIRONMENTALPROTECTION AGENCY; SIERRA CLUB,

Respondents.

On Petition for Writ of Certiorari to theUnited States Court of Appeals for the Tenth Circuit

PETITION FOR WRIT OF CERTIORARI

E. Scott PruittOklahoma Attorney GeneralPatrick R. WyrickSolicitor General Counsel of RecordP. Clayton EubanksDeputy Solicitor GeneralOKLAHOMA ATTORNEYGENERAL’S OFFICE313 NE 21st StreetOklahoma City, OK 73105(405) 521-3921patrick.wyrick@oag.ok.gov

Attorneys for PetitionerState of Oklahoma

Becker Gallagher · Cincinnati, OH · Washington, D.C. · 800.890.5001

NO.

Michael GravesThomas P. SchroedterHALL ESTILL,Attorneys at Law320 South Boston Ave.Suite 200Tulsa, OK 74103-3706(918) 594-0443

Attorneys for PetitionerOklahoma IndustrialEnergy Consumers

(additional counsel listed on inside cover)

Brian J. MurrayCharles T. WehlandDennis MurashkoJONES DAY77 West Wacker DriveChicago, IL 60601(312) 782-3939

Michael L. RiceJONES DAY717 Texas, Suite 3300Houston, TX 77002(832) 239-3640

Attorneys for PetitionerOklahoma Gas And ElectricCompany

i

QUESTION PRESENTED

The Regional Haze Program of the Clean Air Actallocates to the States the task of fashioning and thenimplementing plans to improve the aesthetic quality ofair over certain federal lands. The question presentedis whether, despite that allocation of powers to theStates, the United States Environmental ProtectionAgency may nonetheless conduct a de novo review ofthe State of Oklahoma’s plan, in conflict with both thelimited authority granted to the agency under the Actand decisions of this and other courts that haverecognized the primary role given to the States inimplementing the Clean Air Act.

ii

PARTIES TO THE PROCEEDINGS

Petitioners State of Oklahoma, Oklahoma Gas andElectric Company, and Oklahoma Industrial EnergyConsumers were petitioners in the court below.Respondents are the United States EnvironmentalProtection Agency and the Sierra Club, and wererespondent and intervenor-respondent, respectively, inthe court below.

iii

RULE 29.6 STATEMENT

Oklahoma Gas and Electric Company is a wholly-owned subsidiary of OGE Energy Corp. No publiclyheld corporation owns 10% or more of the stock of OGEEnergy Corp. Petitioner Oklahoma Industrial EnergyConsumers is a non-partisan, unincorporatedassociation of large consumers of energy with facilitieslocated in the State of Oklahoma.

iv

TABLE OF CONTENTS

QUESTION PRESENTED . . . . . . . . . . . . . . . . . . . . i

PARTIES TO THE PROCEEDINGS . . . . . . . . . . . . ii

RULE 29.6 STATEMENT . . . . . . . . . . . . . . . . . . . . iii

TABLE OF AUTHORITIES . . . . . . . . . . . . . . . . . . vi

PETITION FOR A WRIT OF CERTIORARI . . . . . . 1

OPINIONS BELOW . . . . . . . . . . . . . . . . . . . . . . . . . 1

JURISDICTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

STATUTES AND REGULATIONS INVOLVED . . . 1

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

STATEMENT OF THE CASE . . . . . . . . . . . . . . . . . 5

REASONS FOR GRANTING THE WRIT . . . . . . . 16

I. The Tenth Circuit’s decision conflicts withdecisions of this Court and other federalcourts of appeal on the allocation of federal-state authority. . . . . . . . . . . . . . . . . . . . . . . . 16

II. The conflict over federal-state authority is arecurring problem of national importance. . 24

CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

v

APPENDIX

Appendix A Denial of Petitions for Review of FinalDecision Issued by the United StatesEnvironmental Protection Agency andJudgment in the United States Courtof Appeals for the Tenth Circuit(July 19, 2013) . . . . . . . . . . . . . . . App. 1

Appendix B Final Rule, Environmental ProtectionAgency, 40 CFR Part 52(December 28, 2011) . . . . . . . . . App. 56

Appendix C Order Denying Rehearing in theUnited States Court of Appeals for theTenth Circuit(October 31, 2013) . . . . . . . . . . App. 209

Appendix D Statutes and Regulations . . . . App. 21142 U.S.C.A. § 7491 . . . . . . . . . . App. 21140 C.F.R. § 51.308 . . . . . . . . . . App. 218

Appendix E Excerpts from Regional HazeImplementation Plan Revision, Stateof Oklahoma, Department ofEnvironmental Quality (February 2, 2010) . . . . . . . . . . App. 244

Appendix F Order in the United States Court ofAppeals for the Tenth Circuit(June 22, 2012) . . . . . . . . . . . . . App. 246

vi

TABLE OF AUTHORITIES

CASES

Alaska Dep’t of Envtl. Conservation v. EPA, 540 U.S. 461 (2004) . . . . . . . . . . . . . . . . 19, 20, 26

Am. Corn Growers Ass’n v. EPA, 291 F.3d 1 (D.C. Cir. 2002) . . . . . . . . . . . 8, 16, 17

Arizona v. EPA, No. 13-70366 (9th Cir., filed Jan. 31, 2013) . . . 25

Arizona Pub. Serv. Co. v. EPA, 562 F.3d 1116 (10th Cir. 2009) . . . . . . . . . . . . . 21

Chevron U.S.A., Inc. v. Natural Resources DefenseCouncil, Inc., 467 U.S. 837 (1984) . . . . . . . . . . . . . . . . . . . . . . 16

Cliffs Natural Res., Inc. v. EPA, No. 13-1758 (8th Cir., filed Apr. 4, 2013) . . . . . 25

EME Homer City Generation, L.P. v. Envtl. Prot.Agency, 696 F.3d 7 (D.C. Cir. 2012) . . . . . . . . . . . . . . . . . 8

Louisiana Dep’t of Env. Quality v. EPA, No. 12-60672 (5th Cir., filed Sept. 4, 2012) . . . 25

Martinez, et al. v. EPA, No. 11-9567 (10th Cir., filed Oct. 21, 2011) . . . 25

Michigan v. EPA, No. 13-2130 (8th Cir., filed May 22, 2013) . . . . 25

Nebraska v. EPA, No. 12-3084 (8th Cir., filed Sept. 4, 2012) . . . . 25

vii

New York v. United States, 505 U.S. 144 (1992) . . . . . . . . . . . . . . . . . . . 26, 27

N. Dakota v. EPA, 730 F.3d 750 (8th Cir. 2013) . . . . . . . . . . . . 25, 26

PPL Montana, LLC v. EPA, No. 12-73757 (9th Cir., filed Nov. 16, 2012) . . . 25

Train v. Nat. Res. Def. Council, Inc., 421 U.S. 60 (1975) . . . . . . . . . . . . . . . . . . . 2, 8, 18

Union Elec. Co. v. EPA, 427 U.S. 246 (1976) . . . . . . . . . . . . . . . . . . . . 8, 18

Utah v. EPA, No. 13-9535 (10th Cir., filed Mar. 21, 2013) . . . 25

STATUTES, REGULATIONS, AND LEGISLATION

28 U.S.C. § 1254(1) . . . . . . . . . . . . . . . . . . . . . . . . . . 1

40 C.F.R. pt. 51 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 C.F.R. § 51.308 . . . . . . . . . . . . . . . . . . . . . . . . 440 C.F.R. § 51.308(d)(1)(i)(B) . . . . . . . . . . . . . . . . 740 C.F.R. § 51.308(e)(1)(ii)(A) . . . . . . . . . . . 3, 6, 8

70 Fed. Reg. 39,104 (July 6, 2005) . . . . . . . . . 2, 7

40 C.F.R. pt. 5276 Fed. Reg. 81,728 (Dec. 28, 2011) . . . . . . 12, 13

42 U.S.C. § 7401(a)(3) . . . . . . . . . . . . . . . . . . . . . . . . 2

42 U.S.C. § 7407 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

42 U.S.C. § 7410 . . . . . . . . . . . . . . . . . . . . . . . . . . 1, 8

42 U.S.C. § 7491 . . . . . . . . . . . . . . . . . . . . . . . . . . 1, 3

viii

42 U.S.C. § 7491(a)(1) . . . . . . . . . . . . . . . . . . . . . . 2, 5

42 U.S.C. § 7491(b)(1) . . . . . . . . . . . . . . . . . . . . . . . . 7

42 U.S.C. § 7491(b)(2) . . . . . . . . . . . . . . . . . . . . . . . . 7

42 U.S.C. § 7491(b)(2)(A) . . . . . . . . . . . . . . . . . . . . . 6

42 U.S.C. § 7491(g)(2) . . . . . . . . . . . . . . . . . . . . 3, 5, 6

70 Fed. Reg. 39,106/1-2 . . . . . . . . . . . . . . . . . . . . . . . 8

70 Fed. Reg. 39,107/3 . . . . . . . . . . . . . . . . . . . . . . . . 8

70 Fed. Reg. 39,127 . . . . . . . . . . . . . . . . . . . . . . . . . 13

70 Fed. Reg. 39,132 . . . . . . . . . . . . . . . . . . . . . . . . . 10

70 Fed. Reg. 39,137 (July 6, 2005) . . . . . . . . . . . . . . 2

76 Fed. Reg. 52,388 (Aug. 22, 2011) (New Mexico) . 25

76 Fed. Reg. 81,728 (Dec. 28, 2011) (Oklahoma) . . 25

77 Fed. Reg. 14,604 (Mar. 12, 2012) (Arkansas) . . 25

77 Fed. Reg. 20,894 (Apr. 6, 2012) (North Dakota) 25

77 Fed. Reg. 33,022 (June 4, 2012) (Wyoming) . . . 25

77 Fed. Reg. 39,425 (July 3, 2012) (Louisiana) . . . 25

77 Fed. Reg. 40,150 (July 6, 2012) (Nebraska) . . . 25

77 Fed. Reg. 50,936 (Aug. 23, 2012) (Nevada) . . . . 25

77 Fed. Reg. 57,864 (Sept. 18, 2012) (Montana) . . 25

77 Fed. Reg. 71,533 (Dec. 3, 2012) (Michigan) . . . . 25

77 Fed. Reg. 72,512 (Dec. 5, 2012) (Arizona) . . . . . 25

ix

77 Fed. Reg. 74,355 (Dec. 14, 2012) (Utah) . . . . . . 25

78 Fed. Reg. 8,706 (Feb. 6, 2013) (Minnesota) . . . . 25

Alaska National Interest Lands Conservation Act,16 U.S.C. §§ 3101 et seq. . . . . . . . . . . . . . . . . . . 26

Clean Air Act, 42 U.S.C. § XXX . . . . . . . . . . . passim§ 7410 (sec. 110) . . . . . . . . . . . . . . . . . . . . . . . . . . 1

§ 110(a)(2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8§ 7491 (sec. 169A) . . . . . . . . . . . . . . . . . . . . . . . . 1§ 7492 (sec. 169B) . . . . . . . . . . . . . . . . . . . . . . . . 1

§ 169A(b)(2)(A) . . . . . . . . . . . . . . . . . . . . . . . 17§ 169A(g)(2) . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Clean Water Act, 33 U.S.C. §§ 1251 et seq. . . . . . . 26

Occupational Safety and Health Act, 29 U.S.C.§§ 651 et seq. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Resource Conservation and Recovery Act, 42 U.S.C.§§ 6901 et seq. . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

OTHER AUTHORITIES

http://www.epa.gov/airquality/visibility/what.html. 2

Oklahoma Regional Haze State ImplementationPlan, available at http://www.deq.state.ok.us/AQDnew/rulesandplanning/Regional_Haze/SIP/index.htm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1

PETITION FOR A WRIT OF CERTIORARI__________

Petitioners respectfully pray for a writ of certiorarito review the judgment of the United States Court ofAppeals for the Tenth Circuit in this case.

OPINIONS BELOW

A divided panel of the Tenth Circuit filed its opinionon July 19, 2013. App. 1. That opinion is reported at723 F.3d 1201.

JURISDICTION

The judgment of the court of appeals was entered onJuly 19, 2013. Petitions for panel and en bancrehearing were denied on October 31, 2013. App. 209. This petition for certiorari is filed within ninety days ofthe denial of the petitions for rehearing. Thejurisdiction of this Court is invoked under 28 U.S.C.§ 1254(1).

STATUTES AND REGULATIONS INVOLVED

This case involves a challenge to a final rule thatthe United States Environmental Protection Agency(“EPA”) promulgated under sections 110, 169A and169B of the Clean Air Act, 42 U.S.C. §§ 7410, 7491,7492. EPA published the final rule on December 28,2011, at 76 Fed. Reg. 81,728. Also involved areregulations that EPA promulgated to effectuate therelevant sections of the Clean Air Act. Thoseregulations are found at 40 C.F.R. pt. 51, subpt. P.

The full text of pertinent statutory and regulatoryprovisions are set forth in the appendix to this petition.App. 211.

2

INTRODUCTION

This case raises recurring issues of nationalimportance concerning the ability of the States toexercise their statutory authority under the Clean AirAct’s “Regional Haze Program”—a program that affectsforty-five States and territories. The Regional HazeProgram was added to the Clean Air Act in 1977, andaims to mitigate and ultimately prevent any“impairment of visibility in mandatory Class I Federalareas” due to “manmade air pollution.” 42 U.S.C.§ 7491(a)(1).1

The Clean Air Act recognizes that “air pollutionprevention . . . and air pollution control at its source isthe primary responsibility of States and localgovernments.” 42 U.S.C. § 7401(a)(3); see also id.§ 7407(a). Even in the Clean Air Act, wherecooperative federalism is a dominant theme, theRegional Haze Program is unique in the amount ofpower reserved to the States. See, e.g., Train v. NaturalResources Defense Council, Inc., 421 U.S. 60, 79 (1975);70 Fed. Reg. 39,104, 39,137 (July 6, 2005) (“the Act andlegislative history indicate that Congress evinced aspecial concern with insuring that States would be thedecision makers.”). Congress was especially concernedwith maximizing state authority in this context

1 According to EPA, “[h]aze is caused when sunlight encounterstiny pollution particles in the air. Some light is absorbed byparticles. Other light is scattered away before it reaches anobserver. More pollutants mean more absorption and scattering oflight, which reduce the clarity and color of what we see. Sometypes of particles such as sulfates, scatter more light, particularlyduring humid conditions.” See http://www.epa.gov/airquality/visibility/what.html.

3

because the Regional Haze Program’s goals andstandards are purely aesthetic, unrelated to publichealth and safety.

Congress thus vested the States—not EPA—withthe authority to develop and implement “StateImplementation Plans” under the Regional HazeProgram that include “reasonable progress” measuresand “best available retrofit technology” (“BART”)determinations (i.e., determinations as to whattechnology might best control emissions from certainqualifying sources, like electricity generating plants). See 42 U.S.C. § 7491. Congress further mandated thatStates, not EPA, decide what constitutes BART foreligible facilities. See 42 U.S.C. § 7491(g)(2). And theClean Air Act does not require the State to reach anyparticular result in doing so; rather, it only requiresthat the State balance five statutory factors, and reacha decision of its own based on that balancing. 42 U.S.C.§ 7491(g)(2); 40 C.F.R. § 51.308(e)(1)(ii)(A).

Oklahoma’s Regional Haze State ImplementationPlan accordingly balanced the five BART factors, anddetermined that the BART for reducing sulfur dioxide(SO2) emissions from its six qualifying sources (i.e.,electricity-generating power plants) was to requirethose facilities to use only low sulfur coal, which burnssignificantly cleaner than cheaper high sulfur coal, andemits about fifty percent less SO2. Oklahoma submittedits Plan to EPA with this BART determination for SO2.

Under the guise of reviewing Oklahoma’s BARTdetermination for compliance with the statutoryrequirement that Oklahoma balance the five prescribedfactors, EPA conducted a de novo review of thosedeterminations and rejected Oklahoma’s Plan. EPA

4

then substituted a Federal Implementation Plan inplace of Oklahoma’s Plan that required the powerplants to reduce SO2 emissions to virtually zero.Without this Court’s intervention, Petitioner OklahomaGas and Electric Company (“OG&E”) must eitherconvert its power plants to natural gas long beforenecessary or install “scrubbers” onto them, at anestimated cost of $1.2 billion. Worse still, EPA admitsthat either option will result in visibility improvementsthat are barely perceptible to the human eye.

A divided panel of the Tenth Circuit held that notonly did EPA have the authority to review and rejectOklahoma’s Plan, but that EPA’s reasoning in doing sowas entitled to highly deferential “arbitrary andcapricious” review. That decision is unquestionablywrong, and demands this Court’s urgent review. Inoverriding Oklahoma’s BART determination in thismanner, EPA usurped authority that the Clean Air Actclearly delegates to the States, upsetting the balance ofpower that Congress carefully sought to create in theClean Air Act and its Regional Haze Program. And inconflict with decisions of other courts recognizing theproper allocation of authority under the Clean Air Act,the Tenth Circuit sanctioned that result, reflexivelydeferring not to the States, as the Regional HazeProgram required, but to EPA.

The Tenth Circuit’s decision threatens every State’sability to exercise the statutory authority vested inthem by Congress to make BART determinations underthe Regional Haze Program. And the threat runsdeeper. The same question of understanding,respecting, and implementing the shared authoritybetween the States and the federal government arises

5

under a broad range of other federal statutes with asimilar cooperative-federalism approach.

The Court should resolve the question presentedwithout delay. Waiting for further percolation is not apractical option, as the multi-billion dollar federalplans that are being foisted upon the States cannotpractically be undone once implementation has begun,and at least nine other States have had their StateImplementation Plans rejected and replaced withFederal Implementation Plans. In Oklahoma alone,EPA’s actions will cost OG&E $1.2 billion dollars withno discernable return other than a marked increase inwhat Oklahoma ratepayers will pay for theirelectricity. Accordingly, this Court’s immediate reviewis urgently needed to preserve the delicate balance ofpower that Congress established in the Regional HazeProgram (and other federal statutes that reflect similardivisions of authority) and to settle this importantissue that will recur time and again as Regional HazeState Implementation Plans are reviewed throughoutthe nation.

STATEMENT OF THE CASE

Statutory and Regulatory Background. Inunequivocal terms, Congress intended that the Stateswould implement the Regional Haze Program’saesthetic goal of “remedying . . . impairment ofvisibility in mandatory class I Federal areas.” 42 U.S.C.§ 7491(a)(1), (g)(2). As such, the Clean Air Actmandates that a State submit a Plan to EPA laying outthe State’s plan for achieving that goal. The CleanAir Act requires that, with regard to certain sourcesthat contribute to visibility impairments, StateImplementation Plans must include:

6

except as otherwise provided . . . a requirementthat each major stationary source which is inexistence on August 7, 1977, but which has notbeen in operation for more than fifteen years asof such date, and which, as determined by theState (or the Administrator in the case of a[Federal Implementation Plan]) emits any airpollutant which may reasonably be anticipatedto cause or contribute to any impairment ofvisibility in any such area, shall procure, install,and operate, as expeditiously as practicable (andmaintain thereafter) the best available retrofittechnology, as determined by the State (or theAdministrator in the case of a [FederalImplementation Plan]) for controlling emissionsfrom such source for the purpose of eliminatingor reducing any such impairment.

§ 7491(b)(2)(A) (emphases added). The twice-deployedphrase “as determined by the State” is unambiguous.

In other words, the State—not EPA—must:1) determine which of the eligible major stationarysources in the State contribute to visibility impairment;and then 2) determine BART for controlling theemissions causing that impairment at that source. Id.When determining BART, the State must balance fivefactors for each qualifying source: (i) the costs ofcompliance; (ii) the energy and non-air qualityenvironmental impacts of compliance; (iii) any existingpollution control technology in use at the source;(iv) the remaining useful life of the source; and (v) thedegree of improvement in visibility that may beexpected as a result of such technology. 42 U.S.C.§ 7491(g)(2); 40 C.F.R. § 51.308(e)(1)(ii)(A).

7

The EPA’s role, in turn, is limited to ensuring thateach state plan “contain[s] such emission limits,schedules of compliance and other measures as may benecessary to make reasonable progress toward meetingthe national goal.” § 7491(b)(2). To carry out thislimited role, EPA is tasked with creating guidelines forthe States “on appropriate techniques and methods forimplementing this section.” § 7491(b)(1). To this end,the Clean Air Act advises EPA that StateImplementation Plans must contain “such emissionlimits, schedules of compliance and other measures asmay be necessary to make reasonable progress towardmeeting” the national visibility goal. § 7491(b)(2).2 EPA has thus promulgated “Regional Haze Regulationsand Guidelines for Best Available Retrofit Technology(BART) Determinations; Final Rule.” (“BARTGuidelines”). 70 Fed. Reg. 39,104 (July 6, 2005)(codified at 40 C.F.R. pt. 51).

Pursuant to the Clean Air Act, a State must applythe BART Guidelines only when a State makes a BARTdetermination for a powerplant of at least 750MW. Forsources under 750MW, the Guidelines becomediscretionary. As the BART determinations at issue inOklahoma were for sources greater than 750MW, Oklahoma was obligated to apply the BARTGuidelines.

2 In its regulations, EPA established that the goal of naturalvisibility conditions be attained by the year 2064. 40 C.F.R.§51.308(d)(1)(i)(B). BART is just the first of the reasonableprogress control measures to be employed over the course of thesixty-year Visibility Program.

8

In the Guidelines, EPA acknowledges that it is theStates that identify which BART sources “mayreasonably be anticipated to cause or contribute to anyimpairment of visibility in any mandatory Class IFederal area,” 70 Fed. Reg. 39,106/1-2, and it is theStates who “must determine the appropriate level ofBART control for each source subject to BART.” 70Fed. Reg. 39,107/3. In short, as emphasized by theD.C. Circuit, the Clean Air Act “give[s] the States broadauthority over BART determinations” and how a stateweighs the BART factors. Am. Corn Growers Ass’n v.EPA, 291 F.3d 1, 8 (D.C. Cir. 2002).

The bottom line is EPA may only reject a State’sdetermination when it finds that the State’sdetermination does not accomplish the goals of theRegional Haze Program. See 42 U.S.C. § 7410; 40C.F.R. 51.308(e)(1)(ii)(A). The Clean Air Act “gives the[EPA] no authority to question the wisdom of a State’schoices of emission limitations” if such choices are “partof a plan which satisfies the standards of § 110(a)(2).” Train, 421 U.S. 60, 79; see also Union Elec. Co. v. EPA,427 U.S. 246, 250 (1976) (the Clean Air Act providesthat EPA “shall approve the proposed plan if it hasbeen adopted after public notice and hearing” and if itmeets the “specified criteria” set forth in Clean Air Act§ 110(a)(2)). In short, the division of authority betweenEPA and the States “is strict,” and establishes a“federalism bar.” EME Homer City Generation, L.P. v.EPA, 696 F.3d 7, 29 (D.C. Cir. 2012). This “statutoryfederalism bar prohibits EPA from using the [StateImplementation Plan] process to force States to adoptspecific control measures.” Id.

9

Factual Background. Oklahoma contains a singlearea subject to the Regional Haze Program: the WichitaMountains National Wildlife Refuge, which makes upa portion of a small mountain range in sparselypopulated far southwestern Oklahoma. As required bythe Regional Haze Program, Oklahoma identified sixmajor stationary sources as contributing to visibilityimpairment at the Wildlife Refuge—two units atOG&E’s Muskogee Generating Station, two units at itsSooner Generating Station, and two units owned andoperated by Public Service Company of Oklahoma.3 Theunits are located in northeastern Oklahoma. Theclosest is 145 miles from the Wildlife Refuge, while thefarthest is 201 miles away.

Petitioner OG&E is Oklahoma’s largest electricityprovider and serves approximately 785,000 customersover 30,000 square miles in Oklahoma and westernArkansas. Another Petitioner, Oklahoma IndustrialEnergy Consumers, represents many of Oklahoma’slargest consumers of electricity—mainly industrialconsumers engaged in energy price-sensitive industriessuch as pulp and paper, cement, refining, glass,industrial gases, plastic, film and food processing, andwho employ thousands of Oklahoma citizens. BothOG&E and the Oklahoma Industrial EnergyConsumers participated in Oklahoma’s StateImplementation Plan process.

Oklahoma’s State Implementation Plan. WhenOklahoma began the process of determining BART forits six qualifying sources, before it were both a 2008

3 The Public Service Company of Oklahoma reached a negotiatedsettlement with EPA. Its two units are not at issue.

10

cost analysis for the OG&E Units—which both EPAand the Oklahoma Department of EnvironmentalQuality had stated was prepared in conformity withEPA Air Pollution Control Cost Manual—and a 2009cost analysis prepared at EPA’s and Oklahoma’srequest that was more site-specific than the 2008 costestimate.

Oklahoma’s “on the ground” analyses demonstratedthat the installation of scrubbers on each of the fourOG&E Units would cost more than $1.2 billion, orbetween approximately $7,000 and $10,000 per ton ofSO2 removed, which is between three and one-half andfive times the upper limit of EPA’s expected costs forthis technology. See 70 Fed. Reg. at 39,132 (estimatingan average cost of $919 per ton and a cost range of$400 to $2,000 per ton of SO2 removed). Additionally,because OG&E had voluntarily begun using low sulfurcoal some years prior, the effectiveness of scrubbers toreduce actual SO2 emissions was greatly reduced.

The State unequivocally concluded that scrubberswere not cost effective for the OG&E Units. Not onlywere the scrubbers too expensive in light of theminimal visibility benefits that would result from theiruse, but their high costs would compel OG&E to extendthe life of the coal-fired units to allow it to recoup theenormous capital costs. A broad spectrum of otherparties, such as environmental advocates like theOklahoma Chapter of the Sierra Club, supported theState’s conclusion at the time.

Oklahoma concluded that making the continueduse of low sulfur coal mandatory constituted BART forSO2 emissions from the OG&E Units. Oklahomadetermined that this requirement would result in an

11

annual average SO2 emission rate of 0.55 lb/mmBtu,less than half the average annual emission rate of1.176 lb/mmBtu that EPA projected if cheaper highsulfur coal was used.

On February 17, 2010, Oklahoma submitted to EPAits State Implementation Plan containing these BARTdeterminations. Oklahoma explained in its Plan that:

[Oklahoma] conducted a thorough case-by-casefive-factor BART analysis for each of the BART-subject units. [Oklahoma] determined that[scrubbers are] not cost-effective for SO2 controlfor any of the six coal-fired . . . electric unitsreviewed. . . . This determination is based on thecapital cost of add-on controls, the costeffectiveness both in dollars per ton and dollarsper deciview of add-on controls, the long termviability of coal with respect to otherenvironmental programs, and nationalcommitments. . . . Revised cost estimates wereprovided by the affected facilities that are basedon vendor quotes and go well beyond the defaultmethodology recommended by EPA guidance.The cost estimates are credible, detailed, andspecific for the individual facilities. The finalestimate for [scrubbers] for the six coal-firedunits was on average 153% greater than thehigh end costs assumed by [Oklahoma] in theDraft [State Implementation Plan]. These costsput the projects well above costs reported forother BART determinations, and above thelevels [Oklahoma] considered reasonable for costeffectiveness both in terms of dollars per ton of

12

pollutant removed and dollars per deciview (e.g.,$10,000,000/dv) of improved visibility.

(Oklahoma Regional Haze State Implementation Plan),App. 245 (available at: http://www.deq.state.ok.us/AQDnew/rulesandplanning/Regional_Haze/SIP/index.htm) (emphasis added).

EPA’s Rejection of Oklahoma’s Plan. OnDecember 28, 2011, EPA published a final rule withrespect to the Oklahoma Plan, disapproving the State’sSO2 BART determinations for the six Oklahoma unitsbased on EPA’s own balancing of the five statutoryfactors. See Partial Approval of Oklahoma StateImplementation Plan and Promulgation of FederalImplementation Plan, 76 Fed. Reg. 81,728 (Dec. 28,2011) (“Final Rule”); App. 56. Instead of acceptingOklahoma’s approach, EPA implemented a markedlydifferent approach through a Federal ImplementationPlan that imposed a 30-day average SO2 emission limitof 0.06 lbs/MMBtu for each of the four OG&E Units.App. 70. If OG&E wishes to continue to operate thefour affected coal units, the limit imposed by EPA inthe Final Rule would require the installation of ascrubber at each unit within five years.

To justify rejecting Oklahoma’s Plan, EPA hired itsown analyst who expressly and remarkably:

1. assumed that OG&E was burning high-sulfur coal, even though that had not beenthe case for years, and despite the fact thatEPA’s own BART Guidelines mandated theuse of actual historic baseline emissions,App. 139;

13

2. assumed that OG&E could use smaller,cheaper scrubbers, despite the fact thatOG&E showed that those smaller scrubberswould act as governors on its units andprevent the units from producing enoughelectricity to meet peak demand, id.;

3. concluded that the useful life of the scrubberswas thirty years—rather than the twenty-year useful life used by Oklahoma—despitethe fact that EPA had itself used twentyyears in prior cost analyses, App. 159; and

4. rejected Oklahoma’s cost estimates fordeviating from EPA’s Control Cost Manual,even though EPA had previouslyacknowledged that “States have flexibility inhow they calculate costs,” 70 Fed. Reg. at39127, and the State had real-world, site-specific vendor quotes to support thoseestimates. App. 135-36. Ironically, EPA thenturned around and itself deviated from theControl Cost Manual without any site-specific cost support in estimating muchlower installation costs.

In short, EPA’s analyst dramatically overstated thecost-effectiveness of the scrubbers. EPA, in turn, usedthe analyst’s conclusions as a basis for rejectingOklahoma’s Plan. Worse still, EPA’s final rule for thefirst time employed the “overnight method” forcalculating costs (i.e., assuming that an entire plantcould be constructed in a single day) and the days ofvisibility improvement metric for conducting visibilityanalysis (a cumulative analysis that fails to performthe required analysis for each source), depriving

14

petitioners of the opportunity to comment on those newmethodologies.

Petitioners filed requests for reconsideration withEPA in February 2012, but no action has been taken onthose requests.

Proceedings Below. On February 24, 2012,Petitioners filed petitions for review challenging EPA’spartial disapproval of the Oklahoma Plan andsimultaneous promulgation of EPA’s FederalImplementation Plan as arbitrary and capricious,contrary to law, and in violation of the AdministrativeProcedure Act’s notice-and-comment requirements.4 OnJune 22, 2012, the Tenth Circuit issued an order tostay the Federal Implementation Plan pending thehearing by the merits panel. App. 246.

A divided panel of the Tenth Circuit subsequentlydenied the petitions for review. Reasoning that “all the[Clean Air Act] did was shift the initial responsibilityfor making BART determinations from EPA to theState,” App. 16 (emphasis added), the majorityconcluded that not only was Oklahoma entitled to nodeference in its initial BART determination, but thatEPA was entitled to deference in “reviewing” andrejecting Oklahoma’s Plan. App. 19-20. The majoritythus reviewed EPA’s rejection of Oklahoma’s Plan tosee if it was “arbitrary, capricious, an abuse ofdiscretion, or otherwise not in accordance with law.” Id.In applying that review, the majority noted that the

4 The Tenth Circuit consolidated for review the separate petitionsfiled by the State of Oklahoma and Oklahoma Industrial EnergyConsumers (No. 12-9526) and OG&E (No. 12-9527).

15

deference it afforded to EPA was “especially strong”because the challenged decisions involved technicalmatters within EPA’s area of expertise, and, therefore,“[l]eft to evaluate the arguments of the parties’ experts,we must give deference to the EPA.” App. 32-33.

Even affording such deference to EPA, the majoritythought it a “close case,” but in its view, it “ultimately”could not adopt Oklahoma’s analyses “given that theEPA was aware of, and provided explanationscontradicting, petitioners’ comments.” App. 33.

Turning to EPA’s Federal Implementation Plan, themajority applied “the same arbitrary and capriciousstandard…used to evaluate the EPA’s rejection of[Oklahoma’s Plan],” App. 28-29, and concluded that theFederal Implementation Plan was neither arbitrary norcapricious. App. 28-45.

Judge Kelly dissented, expressing the view thatwhile “[u]sually the court grants deference to the EPA’stechnical determinations…[t]he EPA deserves no suchdeference, however, where it does not support aconclusion contradicting Oklahoma’s first, reasonable,detailed technical conclusion.” App. 52. Judge Kellyfurther explained that while “the EPA has at leastsome authority to review BART determinations withina state’s [Plan], it has no authority to conditionapproval of a [State Implementation Plan] basedsimply on a preference for a particular controlmeasure.” App. 53. Judge Kelly concluded that“Oklahoma considered the cost and resulting benefit ofsuch a large investment in scrubbers, and itsconclusion was not unreasonable.” Id. Judge Kellyconcluded that EPA acted “arbitrarily and capriciously”by exaggerating the effectiveness of the scrubbers in

16

order to make them seem cost effective. EPA,“[k]nowing these calculations violated [its own]manual,” developed an alternative way to attempt tojustify the scrubbers: it simply changed “the size of thescrubbers to smaller, less expensive ones,” but did sowithout providing “any evidence that a significantlysmaller scrubber was sufficient to meet OG&E’sneeds.” App. 51-52. Consequently, Judge Kelly wouldhave found EPA’s actions unlawful and would not havedeferred to EPA’s technical judgments and experts.App. 52. Judge Kelly also concluded that EPA failed toprovide record evidence to support why its owncontrary BART determinations were justified.

REASONS FOR GRANTING THE WRIT

I. The Tenth Circuit’s decision conflicts withdecisions of this Court and other federalcourts of appeal on the allocation of federal-state authority.

In reflexively affording EPA Chevron deference, thepanel below departed from other circuits, which haveresoundingly recognized that States, not EPA, areentitled to deference in formulating plans under theClean Air Act. Chevron U.S.A., Inc. v. NaturalResources Defense Council, Inc., 467 U.S. 837 (1984).The decision below radically departed from the clearly-designed cooperative-federalism mechanism of theClean Air Act (specifically, the Regional HazeProgram), converting it into federal supremacy whereEPA is permitted to replace a State’s determinationwith its own.

1. The decision below squarely conflicts with theD.C. Circuit’s decision in American Corn Growers

17

Association v. EPA, 291 F.3d 1 (D.C. Cir. 2002) (“CornGrowers”), in which the D.C. Circuit invalidated EPA’sattempt to mandate the manner in which States mustconsider one of the five BART factors and stressed thatEPA’s actions were “inconsistent with the Act’sprovisions giving the states broad authority over BARTdeterminations.” Id. at 8. By dictating that the Statesmake BART determinations in a particular manner,EPA had impermissibly “constrain[ed] authorityCongress conferred on the states.” Id. at 9. The courtemphasized that the “states . . . play the lead role indesigning and implementing regional haze programs,”id. at 8 (citing Clean Air Act §§ 169A(b)(2)(A);169A(g)(2)), and that the phrase “as determined by theState” is unique to the Regional Haze Program. Indeed,no similar language appears in other air programsregarding, for example, best available controltechnology (“BACT”) or National Ambient Air QualityStandards. As such, the D.C. Circuit had no troubleconcluding that Congress had unequivocally left BARTdeterminations to the States. Id. at 7–8.

Regardless, the panel below concluded that CornGrowers did “not alter [its] conclusion.” App. 14. Thepanel held that while Corn Growers recognized that theClean Air Act “shift[ed] the power to determine BARTfrom the EPA to the states,” Congress intended only toprevent “the EPA from directly making t[he] BARTdecisions, ” and that EPA retained the ability toindirectly make those decisions through its “authorityto ensure that…BART decisions comply with thestatute.” Id.

But that is a distinction without a difference. Byreplacing Oklahoma’s careful work in the State

18

Implementation Plan with EPA’s own, de novoapproach, driven by its hired consultant’s post-StateImplementation Plan work, EPA did make BARTdecisions directly. The fact that EPA operated underthe cloak of reviewing Oklahoma’s Plan is beside thepoint if, at the end of the day, the result is the same.

2. The decision below also cannot be reconciledwith this Court’s nearly four decades of recognizing theClean Air Act’s “division of responsibilities” betweenthe States and the federal government. Train v.Natural Resources Defense Council, Inc., 421 U.S. 60,79 (1975). In Train, the Court observed that EPA “isplainly charged by the Act with the responsibility forsetting the national ambient air standards.” But “[j]ustas plainly,” EPA “is relegated by the Act to a secondaryrole in the process of determining and enforcing thespecific, source-by-source emission limitations whichare necessary if the national standards it has set are tobe met.” Id. (emphasis added). As the Court explained,“[t]he Act gives the [EPA] no authority to question thewisdom of a State’s choices of emission limitations ifthey are part of a plan which satisfies the [Act’s]standards.” Id. “[S]o long as the ultimate effect of aState’s choice of emission limitations is compliancewith the national standards for ambient air, the Stateis at liberty to adopt whatever mix of emissionlimitations it deems best suited to its particularsituation.” Id.; see also Union Elec. Co. v. EPA, 427U.S. at 269 (“Congress plainly left with the States, solong as the national standards were met, the power todetermine which sources would be burdened byregulation and to what extent”).

19

3. Likewise, the Court in Alaska Department ofEnvironmental Conservation v. EPA, 540 U.S. 461(2004), examined EPA’s attempt to override Alaska’s“best available control technology” (“BACT”)determination under the National Ambient Air QualityStandards program—a program in which the Clean AirAct gives EPA an even greater supervisory role than inRegional Haze cases. Despite the express authorityconferred by the Clean Air Act on EPA to reject aState’s BACT determination, this Court held thatEPA’s role was limited to reviewing whether the State’sBACT determination was reasonably moored to theClean Air Act and faithful to the statute’s definition ofBACT. Id. at 484. Unwilling to accord its normaldeference to EPA’s actions, this Court found that EPA’soversight role was limited to determining whether theState’s determination “is not based on a reasonedanalysis” and is “arbitrary.” Id. at 490-91. Even EPAagreed that it must accord appropriate deference to aState’s determination and that it lacked authority to“second guess” a state’s decision. Id.

Thus, in reviewing EPA’s purely supervisory role,this Court held that “the production and persuasionburdens remain with EPA and the underlying questiona reviewing court resolves remains the same: Whetherthe state agency’s BACT determination wasreasonable, in light of the statutory guides and thestate administrative record.” Id. at 494.

This Court’s reasoning in Alaska Department shouldhave applied with even greater force to EPA’s review ofBART determinations. BACT is a continually-evolving,health-driven emission level applicable to newconstruction or modification. BART, on the other hand,

20

is a one-time, cost-benefit-based, visibility standard forsources constructed prior to and unmodified since 1977.The BACT provisions impose obligations reflected bysuch strong, normative terms as “maximum” and“achievable” that are not found in the Clean Air Act’sdefinition of BART. Id. at 484–91. There is also noprovision in BACT that is comparable to the RegionalHaze Program’s unequivocal mandate that BART is “asdetermined by the State.”

In evaluating EPA’s rejection of Oklahoma’s BARTdetermination, the panel below largely ignored AlaskaDepartment, and disregarded its guidance. The panelsanctioned EPA’s second-guessing of the cost estimatesused in the State’s BART determination, contendingthat they failed to comply with EPA guidelines. EPAerroneously argued—and the panel majorityagreed—that Oklahoma failed to follow a particularcosting methodology, that OG&E’s costing assumptionswere flawed, and that even OG&E’s detailed costestimates, provided at EPA’s request, were inadequate.Oklahoma raised numerous objections and counter-arguments to these conclusions during the FederalImplementation Plan rulemaking process, but thepanel found that Oklahoma failed to show that EPA’sapproach was arbitrary and capricious. This findingrequires Oklahoma to disprove the validity of EPA’sconclusions, but Alaska Department mandatesotherwise. It is EPA that bears the burden of showingthat Oklahoma’s costing methods, and ultimately itsBART determination, were unreasonable, and thepanel erred in holding EPA to a lesser standard.

4. Contrary to the approach taken by the majority,for regional haze, the State, not EPA, is the “authorized

21

agency” entitled to deference under Arizona PublicService Co. v. EPA, 562 F.3d 1116, 1123 (10th Cir.2009). By according EPA the deference that is reservedby the Clean Air Act to the State, the panel majority’sdecision undermined the State’s exercise of itsauthority under the Clean Air Act. Repeatedlythroughout the course of its review of EPA’s decision,the panel deferred to EPA’s preferences as long asEPA’s hired consultant provided some explanation forEPA’s conclusions. But the test should not have beenwhether EPA’s approach could be justified. It shouldhave been whether EPA had a basis to say that theState’s approach violated some mandatory requirementin the Regional Haze Regulations or was itselfarbitrary. The panel’s decision leaves States unable todetermine with certainty what approach to regionalhaze is acceptable because the majority gives EPA thefreedom to rely on any one of multiple possibleinterpretations of baseline emissions, of therequirements of EPA’s Control Cost Manual, or even ofthe engineering necessary to identify technicallyfeasible controls. By painting its review of EPA’s actionin rejecting the Oklahoma Plan with the same broad,deferential brush that it viewed EPA’s adoption of itsown Federal Implementation Plan, the panelimproperly disturbed the State’s authority to determineBART, contrary to the Clean Air Act and the long lineof decisions described above.

Because the majority was overly deferential in itsreview of EPA’s action, it failed to conduct ameaningful examination of the explanations underlyingEPA’s cost analysis. The panel should have requiredEPA to show why the State’s rule was unreasonablebefore EPA could reject Oklahoma’s Plan, and the

22

panel should have done so giving deference toOklahoma’s determinations, not EPA’s. If the panelhad given EPA’s rejection of Oklahoma’s Plan that levelof review, it would have found that EPA’s explanationswere frequently based on assumptions unsupported bythe record, contrary to basic engineering or economicrealities, or based on materials or analysis that EPAdid not provide to Oklahoma as part of the stateadministrative record for its consideration during theState’s lengthy process for making its BARTdetermination. This aspect of the decision isparticularly important because the majority recognizedthat even under its deferential standard of review, itwas a “close case.” The panel’s wholesale deferenceallowed EPA to:

1. ignore technical design requirements for thescrubbers needed to maintain the existingfunctionality of the OG&E Units and ignoredEPA’s own guidelines requiring the use ofpast actual emissions to measure the effect ofthe addition of scrubbers.

2. deviate from the twenty-year useful life ofscrubbers used by Oklahoma even though atwenty-year life has been used in other costanalyses and has been acknowledged by EPAas being consistent with its Control CostManual. That error alone resulted inunderstating the scrubber’s annual capitalcosts by thirty percent.

3. reject Oklahoma’s 2009 cost estimates thatdeviated from the Control Cost Manual eventhough the State had site-specific vendorquotes to support those costs. Meanwhile,

23

EPA deviated from the Control Cost Manualwithout any site-specific cost support.

4. support the Federal Implementation Plan byaggregating visibility improvements frommultiple units, even though OG&Especifically objected to that methodology inthe administrative proceeding for beinginconsistent with the Regional Hazeregulations.

EPA could not justify its rulemaking and achieve itsdesired result of requiring scrubbers on the OG&EUnits without these errors because OG&E voluntarilyadopted the use of low sulfur coal many years ago, thusminimizing any adverse impact on visibility. TheOklahoma Plan would have ensured reasonable furtherprogress toward maintaining that limited impact bymaking that voluntary choice mandatory. EPA,however, was unsatisfied with anything less than theinstallation of scrubbers, which put it in the awkwardposition of having to justify huge costs from which onlymarginal visibility benefits will flow. It was thus littlewonder EPA’s rulemaking was not a model of expertagency work.

The only question for EPA on review of the State’sdetermination should have been whether it representeda reasonable application of EPA guidelines based onthe record that existed when the State made itsdecision, and in conducting this review, EPA shouldhave given the State’s determinations the same level ofdeference that it expects when its decisions arereviewed. Rather than review the State’s determinationfor proper and reasonable exercise of its discretion,EPA hired a consultant to second-guess Oklahoma’s

24

choices, App.9, created projections of scrubber costsusing its discretionary choices and assumptions afterthe state administrative record was closed, id., andsubstituted its judgments for the site-specific analysisconducted by Oklahoma. EPA did not give Oklahomathis information to consider in making its BARTdetermination. The Clean Air Act does not authorizeEPA to approach its review of the State’s BARTdeterminations in that way, and the majority’s decisionundermines the authority given to the State.

On this basis alone, this Court’s review iswarranted and urgently needed.

II. The conflict over federal-state authority is arecurring problem of national importance.

This Court’s review is urgently needed in light ofthe important and recurring nature of the questionpresented—not only under the Regional Haze Programbut also a broad range of other federal statutesexemplifying various allocations of authority betweenthe States and federal government.

1. In this, just the first phase of the long-termRegional Haze Program,5 EPA has disapproved stateBART determinations or taken similar action in twelveStates and has a pending disapproval in another

5 Given the stringency of EPA’s other regulations applicable tofacilities in the eastern States, EPA has not for the most partrequired those States to make BART determinations for electricitygenerating sources.

25

State.6 While the Tenth Circuit’s decision was the firstof multiple expected judicial decisions reviewing EPAdisapprovals of state BART determinations under theRegional Haze Program in circuits around the country,it will certainly not be the last. At least nine otherproceedings are now pending, involving seven otherstate plans, including two more in the Tenth Circuit.7

And more may be yet to come.

Just as it rushed to do in North Dakota’s case beforethe Eighth Circuit, EPA will certainly waste no time inusing the Tenth Circuit’s decision before court aftercourt in these BART cases. And if those courts rely onthe Tenth Circuit’s decision like the Eighth Circuit did,

6 77 Fed. Reg. 72,512 (Dec. 5, 2012) (Arizona); 77 Fed. Reg. 14,604(Mar. 12, 2012) (Arkansas); 77 Fed. Reg. 39,425 (July 3, 2012)(Louisiana); 77 Fed. Reg. 71,533 (Dec. 3, 2012) (Michigan); 78 Fed.Reg. 8,706 (Feb. 6, 2013) (Minnesota); 77 Fed. Reg. 57,864 (Sept.18, 2012) (Montana); 77 Fed. Reg. 40,150 (July 6, 2012)(Nebraska); 77 Fed. Reg. 50,936 (Aug. 23, 2012) (Nevada); 76 Fed.Reg. 52,388 (Aug. 22, 2011) (New Mexico); 77 Fed. Reg. 20,894(Apr. 6, 2012) (North Dakota); 76 Fed. Reg. 81,728 (Dec. 28, 2011)(Oklahoma); 77 Fed. Reg. 74,355 (Dec. 14, 2012) (Utah); 77 Fed.Reg. 33,022 (June 4, 2012) (Wyoming) (proposed).

7 Arizona v. EPA, No. 13-70366 (9th Cir., filed Jan. 31, 2013);Louisiana Dep’t of Env. Quality v. EPA, No. 12-60672 (5th Cir.,filed Sept. 4, 2012); Michigan v. EPA, No. 13-2130 (8th Cir., filedMay 22, 2013); Cliffs Natural Res., Inc. v. EPA, No. 13-1758 (8thCir., filed Apr. 4, 2013) (Michigan and Minnesota); PPL Montana,LLC v. EPA, No. 12-73757 (9th Cir., filed Nov. 16, 2012); Nebraskav. EPA, No. 12-3084 (8th Cir., filed Sept. 4, 2012); Martinez, et al.v. EPA, No. 11-9567 (10th Cir., filed Oct. 21, 2011) (New Mexico);North Dakota v. EPA, 730 F.3d 750 (8th Cir. 2013); Oklahoma v.EPA (the instant case); Utah v. EPA, No. 13-9535 (10th Cir., filedMar. 21, 2013).

26

see North Dakota v. EPA, 730 F.3d 750, 761 (8th Cir.2013), the ripple effect will magnify the harms causedby the Tenth Circuit’s decision. Billion-dollar plan afterbillion-dollar plan will be forced on the States. Andonce construction begins in order to implement thoseplans, it cannot practically be unwound.

Additionally, the decision not only harms Oklahomanow, it also limits the technical tools Oklahoma hasavailable to it in developing future Regional Haze StateImplementation Plans for the remaining forty-six yearsof the Regional Haze Program. Oklahoma’s next StateImplementation Plan is due in 2018, and the BARTdeterminations made now—whether by Oklahoma orEPA—will directly affect the choices and decisionsmade by Oklahoma for the next half-a-century. EPA’simposition of a Federal Implementation Plan in thisfirst planning period of the Regional Haze programunlawfully ties Oklahoma’s hands as to what it can doin the future—a direct repudiation of Congress’smandate that the States lead the design andimplementation of the Regional Haze Program.

2. Much like it did in Alaska Department, EPA herehas yet again called into question numerous otherstatutes that embody the principle of cooperativefederalism. See, e.g., New York v. United States, 505U.S. 144, 167-168 (1992) (identifying “numerous federalstatutory schemes” of this nature, including the CleanWater Act, 33 U.S.C. §§ 1251 et seq., the OccupationalSafety and Health Act, 29 U.S.C. §§ 651 et seq., theResource Conservation and Recovery Act, 42 U.S.C.§§ 6901 et seq., and the Alaska National Interest LandsConservation Act, 16 U.S.C. §§ 3101 et seq.). Much likethe Clean Air Act, these statutes are based on shared

27

federal-state responsibility, whereby the federalgovernment sets standards and the States—if they optto undertake the responsibility—are given broadflexibility in implementing those standards.Cooperative-federalism promotes federalism because“state governments remain responsive to the localelectorate’s preferences; state officials remainaccountable to the people.” New York v. United States,505 U.S. at 167-168. The Tenth Circuit’s decision, bytransferring from the State to EPA core discretionaryauthority under the leading cooperative-federalismstatutory regimes, threatens to undermine the balanceof power struck by Congress and accepted by the Stateswhen they assumed the responsibilities offered underthe Act. For this reason too, the decision merits review.

CONCLUSION

For these reasons, the petition for a writ ofcertiorari should be granted, and the judgment belowreversed.

28

Respectfully submitted,

E. Scott PruittOklahoma Attorney GeneralPatrick R. WyrickSolicitor General Counsel of RecordP. Clayton EubanksDeputy Solicitor GeneralOklahoma AttorneyGeneral’s Office313 NE 21st StreetOklahoma City, OK 73105(405) 521-3921fc.docket@oag.ok.govscott.pruitt@oag.ok.govpatrick.wyrick@oag.ok.govclayton.eubanks@oag.ok.gov

Attorneys for Petitioner State of Oklahoma

Michael GravesThomas P. SchroedterHALL ESTILL, Attorneys at Law320 South Boston Ave., Suite 200Tulsa, OK 74103-3706(918) 594-0443mgraves@hallestill.comtschroedter@hallestill.com

Attorneys for PetitionerOklahoma Industrial EnergyConsumers

29

Brian J. MurrayCharles T. WehlandDennis MurashkoJONES DAY77 West Wacker DriveChicago, IL 60601(312) 782-3939bjmurray@jonesday.comctwehland@jonesday.comdmurashko@jonesday.com

Michael L. RiceJONES DAY717 Texas, Suite 3300Houston, TX 77002(832) 239-3640mlrice@jonesday.com

Attorneys for Petitioner Oklahoma Gas And ElectricCompany

APPENDIX

i

APPENDIX

TABLE OF CONTENTS

Appendix A Denial of Petitions for Review of FinalDecision Issued by the United StatesEnvironmental Protection Agency andJudgment in the United States Courtof Appeals for the Tenth Circuit(July 19, 2013) . . . . . . . . . . . . . . . App. 1

Appendix B Final Rule, Environmental ProtectionAgency, 40 CFR Part 52(December 28, 2011) . . . . . . . . . App. 56

Appendix C Order Denying Rehearing in theUnited States Court of Appeals for theTenth Circuit(October 31, 2013) . . . . . . . . . . App. 209

Appendix D Statutes and Regulations . . . . App. 21142 U.S.C.A. § 7491 . . . . . . . . . . App. 21140 C.F.R. § 51.308 . . . . . . . . . . App. 218

Appendix E Excerpts from Regional HazeImplementation Plan Revision, Stateof Oklahoma, Department ofEnvironmental Quality (February 2, 2010) . . . . . . . . . . App. 244

Appendix F Order in the United States Court ofAppeals for the Tenth Circuit(June 22, 2012) . . . . . . . . . . . . . App. 246

App. 1

APPENDIX A

PUBLISH

UNITED STATES COURT OF APPEALSTENTH CIRCUIT

[Filed July 19, 2013]

No. 12-9526_______________________________________STATE OF OKLAHOMA; OKLAHOMA )INDUSTRIAL ENERGY CONSUMERS, )an unincorporated association, )

Petitioners, ))

v. ))

UNITED STATES ENVIRONMENTAL )PROTECTION AGENCY, )

Respondent. )--------------------------------------------------------- )SIERRA CLUB, )

Intervenor-Respondent, ))

and ))

PACIFICORP; AMERICAN COALITION )FOR CLEAN COAL ELECTRICITY; )NATIONAL PARKS )CONSERVATION ASSOCIATION, )

Amici Curiae. )_______________________________________)

App. 2

No. 12-9527_______________________________________OKLAHOMA GAS & ELECTRIC ) COMPANY, )

Petitioner, ))

v. ))

UNITED STATES ENVIRONMENTAL )PROTECTION AGENCY, )

Respondent. )--------------------------------------------------------- )SIERRA CLUB, )

Intervenor-Respondent, ))

and ))

PACIFICORP; AMERICAN COALITION )FOR CLEAN COAL ELECTRICITY; )NATIONAL PARKS CONSERVATION )ASSOCIATION, )

Amici Curiae. )_______________________________________)

PETITION FOR REVIEW OF FINAL DECISIONISSUED BY THE UNITED STATES

ENVIRONMENTAL PROTECTION AGENCYEPA-R06-OAR-2010-0190

E. Scott Pruitt, Oklahoma Attorney General, (P.Clayton Eubanks, Assistant Attorney General; MichaelGraves and Thomas P. Schroedter of Hall Estill, Tulsa,Oklahoma, with him on the briefs), Oklahoma City,Oklahoma, for Petitioners State of Oklahoma andOklahoma Industrial Energy Consumers.

App. 3

Brian J. Murray of Jones Day, Chicago, Illinois,(Thomas E. Fennell of Jones Day, Dallas, Texas;Michael L. Rice of Jones Day, Houston, Texas; CharlesT. Wehland of Jones Day, Chicago, Illinois, on thebriefs), for Petitioner Oklahoma City Gas & ElectricCompany.

Stephanie J. Talbert, United States Department ofJustice, Environment & Natural Resources Division,Environmental Defense Section, Washington, D.C.(Ignacia S. Moreno, Assistant Attorney General; M. LeaAnderson and Barbara Nann, Of Counsel, UnitedStates Environmental Protection Agency, with her onthe brief), for Respondent.

Andrea Issod, (Elena Saxonhouse and Sanjay Narayanwith her on the brief), San Francisco, California, for theIntervenor, Sierra Club.

Michael G. Jenkins, Assistant General Counsel,PacifiCorp Energy, Salt Lake City, Utah, and E. BlaineRawson, Ray Quinney & Nebeker, P.C., Salt Lake City,Utah, filed an amicus curiae brief on behalf ofPacifiCorp, Amicus Curiae.

Paul M. Seby and Marian C. Larsen of Moye WhiteLLP, Denver, Colorado, filed an amicus curiae brief onbehalf of the American Coalition for Clean CoalElectricity, Amicus Curiae.

Stephanie Kodish, Knoxville, Tennessee, filed anamicus curiae brief on behalf of the National ParksConservation Association, Amicus Curiae.

Before BRISCOE, Chief Judge, KELLY andLUCERO, Circuit Judges.

App. 4

BRISCOE, Chief Judge.

In these consolidated petitions for review,petitioners1 challenge a final rule promulgated by theUnited States Environmental Protection Agency underthe Clean Air Act. The petitioners argue that the EPAimpermissibly rejected Oklahoma’s plan to limit theemissions of sulfur dioxide at Oklahoma Gas andElectric Company power plants and replaced it with itsown more stringent regulations, which petitionerscontend usurped the state’s authority and will requiresizable expenditures on unnecessary technology. Weconclude that the EPA has authority to review thestate’s plan and that it lawfully exercised thatauthority in rejecting it and promulgating its own.Exercising our jurisdiction under 42 U.S.C.§ 7607(b)(1), we deny the petitions for review.

I

A. Statutory Background

The Clean Air Act “uses a cooperative-federalismapproach to regulate air quality.” U.S. Magnesium,LLC v. EPA, 690 F.3d 1157, 1159 (10th Cir. 2012).Under the Clean Air Act (CAA), the EnvironmentalProtection Agency (EPA) must create and reviewnational ambient air quality standards for certainpollutants. See 42 U.S.C. §§ 7408, 7409. States thenhave the responsibility to adopt state implementationplans (SIPs), “which provide[] for implementation,

1 The petitioners include the state of Oklahoma, the OklahomaIndustrial Energy Consumers interest group, and the OklahomaGas and Electric Company.

App. 5

maintenance, and enforcement” of those primary andsecondary air quality standards. § 7410(a)(1).

States, however, exercise this authority with federaloversight. The EPA reviews all SIPs to ensure that theplans comply with the statute. The EPA may notapprove any plan that “would interfere with anyapplicable requirement” of this chapter of the UnitedStates Code. § 7410(l). The EPA has a duty to create itsown federal implementation plan (FIP) if either: 1) it“finds that a State has failed to make a requiredsubmission or finds that the plan or plan revisionsubmitted by the State does not satisfy the minimumcriteria established under subsection (k)(1)(A) of thissection”; or 2) it “disapproves a State implementationplan submission in whole or in part.” § 7410(c)(1). Theduty to promulgate a FIP exists “unless the Statecorrects the deficiency, and the Administrator approvesthe plan or plan revision, before the Administratorpromulgates such Federal implementation plan.” Id.

At issue in this case are the portions of the CAAthat seek to protect visibility at certain national parksand wildlife areas. The CAA requires that the EPApromulgate regulations “to assure . . . reasonableprogress toward” preventing any future and“remedying . . . any existing, impairment of visibility inmandatory class I Federal areas which impairmentresults from manmade air pollution.” § 7491(a)(1),(a)(4). It also requires that the EPA ensure that eachstate plan “contain[s] such emission limits, schedulesof compliance and other measures as may be necessaryto make reasonable progress toward meeting thenational goal.” § 7491(b)(2).

App. 6

Relevant in this case are the CAA’s mandatesregarding sources that contribute to visibilityimpairments. SIPs must include:

except as otherwise provided . . . a requirementthat each major stationary source which is inexistence on August 7, 1977, but which has notbeen in operation for more than fifteen years asof such date, and which, as determined by theState (or the Administrator in the case of a[FIP]) emits any air pollutant which mayreasonably be anticipated to cause or contributeto any impairment of visibility in any such area,shall procure, install, and operate, asexpeditiously as practicable (and maintainthereafter) the best available retrofit technology,as determined by the State (or theAdministrator in the case of a [FIP]) forcontrolling emissions from such source for thepurpose of eliminating or reducing any suchimpairment.

§ 7491(b)(2)(A). To simplify, a state—or the EPA, whenpromulgating a FIP—must: 1) determine which of theeligible major stationary sources in their statecontributes to visibility impairment; and then2) determine the “best available retrofit technology” forcontrolling the emissions causing that impairment atthat source. Id. When determining “best availableretrofit technology” (BART):

the State (or the Administrator in determiningemission limitations which reflect suchtechnology) shall take into consideration [1] thecosts of compliance, [2] the energy and nonairquality environmental impacts of compliance,

App. 7

[3] any existing pollution control technology inuse at the source, [4] the remaining useful life ofthe source, and [5] the degree of improvement invisibility which may reasonably be anticipated toresult from the use of such technology.

§ 7491(g)(2).

The CAA requires that the EPA create guidelinesfor the states “on appropriate techniques and methodsfor implementing this section.” § 7491(b)(1). For “afossil-fuel fired generating powerplant having a totalgenerating capacity in excess of 750 megawatts, theemission limitations required under this paragraphshall be determined pursuant” to the regulationspromulgated by the EPA. § 7491(b). The EPA haspromulgated these BART guidelines at 40 C.F.R.§ 51.308(e).

B. Procedural Background

In 2005, the EPA issued an updated version of itsRegional Haze Rule that required states to submit SIPrevisions by December 17, 2007. See Regional HazeProgram Requirements, 40 C.F.R. § 51.308(b). OnJanuary 15, 2009, the EPA took final action in findingthat Oklahoma—along with 31 other states, theDistrict of Columbia, and the U.S. VirginIslands—failed to submit a SIP that addressed any ofthe Regional Haze elements by this deadline. SeeFinding of Failure To Submit State ImplementationPlans Required by the 1999 Regional Haze Rule, 74Fed. Reg. 2392-01 (Jan. 15, 2009). This triggered theEPA’s duty to promulgate a federal implementationplan within two years. See 42 U.S.C. § 7410(c)(1).

App. 8

Before the EPA promulgated a FIP, however,Oklahoma submitted its SIP. See Oklahoma RegionalHaze State Implementation Plan, Joint Appendix (JA)at 55 (Feb. 17, 2010). At issue in this petition are theSIP’s BART determinations with respect to two unitsat Oklahoma Gas & Electricity’s (OG&E’s) MuskogeeGenerating Station and two units at its SoonerGenerating Station. The Oklahoma SIP set a sulfurdioxide (SO2) emissions limits of 0.65 lb/mmBtu(thirty-day average) and 0.55 lb/mmBtu (annualaverage) for each of these four units. See OG&EMuskogee Generating Station BART Review, JA at 187(Jan. 15, 2010); OG&E Sooner Generating StationBART Review, JA at 221 (Jan. 15, 2010). The BART foreach of these units included OG&E’s continued use oflow-sulfur coal. The SIP considered, but rejected, anemissions limit that would require the installation ofso-called scrubbers to remove SO2. See MuskogeeBART Review, JA at 213; Sooner BART Review, JA at247. “The cost for [dry scrubbers] is too high, thebenefit too low and these costs, if borne, further extendthe life expectancy of coal as the primary fuel in theSooner [and Muskogee] facilit[ies] for at least 20 yearsand beyond,” according to OG&E’s BART analyses. Seeid.

On March 22, 2011, the EPA proposed a rule thatwould partially approve and partially disapproveOklahoma’s SIP. Proposed Rule, 76 Fed. Reg.16,168-01, 16,169 (Mar. 22, 2011). The SO2 emissionlimitations for OG&E’s four units were among theparts of the SIP that the EPA proposed disapproving.The EPA said that Oklahoma failed to follow thepromulgated regulations in determining BART. Id. at16,182. Specifically, the EPA said that Oklahoma “did

App. 9

not properly ‘take into consideration the costs ofcompliance’ when it relied on cost estimates thatgreatly overestimated the costs of dry and wetscrubbing to conclude these controls were not costeffective.” Id. (quoting 40 C.F.R. § 51.308(e)(1)(ii)(A)).

“Given that scrubbers are typically considered to behighly cost-effective controls for power plants such asthose at issue, [the EPA] retained a consultant toindependently assess the suitability and costs ofinstalling these controls.” Id. The EPA found thescrubbers to be substantially more cost effective thanOklahoma did. Id. at 16,183. For example, Oklahomaestimated the cost of the scrubbers to be $7,147 per tonof SO2 removed at one of the Sooner Generating Stationunits. Id. The EPA projected scrubbers at that sameunit would cost $1,291 per ton of SO2 removed. Id.

In addition to proposing the partial disapproval ofthe SIP, the EPA proposed creating its own federalimplementation plan in the same action. Id. at 16,168.The EPA proposed an SO2 emissions limit of0.06 lb/mmBtu (thirty-day average). Id. at 16,193-94.Based on this limit, the EPA believed the use of dryscrubbers would be cost effective. Id. at 16,183. Afternotice and comment, the EPA published the final ruleenacting these emissions limits. See Final Rule, 76 Fed.Reg. 81,728-01 (Dec. 28, 2011).

On February 24, 2012, the state of Oklahoma andthe Oklahoma Industrial Energy Consumers filed inthis court a petition seeking review of the final rule(Case No. 12-9526). OG&E filed its petition for reviewthe same day (Case No. 12-9527). We later issued anorder granting a motion to consolidate these petitions.

App. 10

The petitioners also took steps to stay theapplication of the rule. The same day they filedpetitions for review, the petitioners filed with the EPAa motion for reconsideration and a request for anadministrative stay.2 The petitioners also filed a motionin this court seeking a stay pending a hearing on themerits. A two-judge panel of this court granted thepetitioners’ motion to stay the portion of the rulerequiring the reduction of SO2 emissions at these fourOG&E units. Oklahoma v. EPA, Nos. 12-9526 and12-9527, at 1-2 (10th Cir. June 22, 2012). Meanwhile,appellate briefing progressed.

The petitioners raise a number of objections to thefinal rule, arguing that the EPA has usurped thestate’s authority in an effort to force OG&E to spendmore than one-billion dollars to install unnecessarytechnology in the next five years. First, they argue thatthe EPA exceeded its statutory authority bydisapproving Oklahoma’s BART determination.

2 Under the CAA, the filing of a petition for reconsideration doesnot affect the finality of an EPA action for the purposes of judicialreview. See 42 U.S.C. § 7607(b)(1) (“The filing of a petition forreconsideration by the Administrator of any otherwise final rule oraction shall not affect the finality of such rule or action forpurposes of judicial review nor extend the time within which apetition for judicial review of such rule or action under this sectionmay be filed, and shall not postpone the effectiveness of such ruleor action.”); see also Natural Res. Def. Council v. Abraham, 355F.3d 179, 203 n.11 (2d Cir. 2004). While the Third Circuit has heldthat a pending petition for reconsideration deprived it ofjurisdiction under the CAA, see W. Penn Power Co. v. EPA, 860F.2d 581, 587-88 (3d Cir. 1988), it reached this result before theCAA was amended to prevent petitions for reconsideration fromaffecting finality. Clean Air Act, Amendments, Pub. L. No.101-549, § 706, 104 Stat. 2399 (1990).

App. 11

Second, they argue that, even if the EPA had thisauthority, the EPA acted arbitrarily and capriciouslyby disapproving Oklahoma’s SIP. Third, they arguethat the EPA acted arbitrarily and capriciously inpromulgating its FIP. Fourth, the petitioners arguethat the EPA failed to provide them adequate notice ofaspects of the final rule. Finally, the petitioners arguethe EPA violated the CAA by promulgating the FIP inthe same action in which it partially disapproved of theSIP and after the two-year deadline to promulgate aFIP had expired.

II

The petitioners argue that the EPA exceeded itsstatutory authority by rejecting Oklahoma’s BARTdeterminations and replacing them with its own. Thepetitioners say that the EPA’s action tramples on thediscretion that Congress afforded states to make thesedecisions. The CAA’s cooperative-federalism policysupports this view, the petitioners say. Morespecifically, the petitioners point to the statute’slegislative history and its language—mandating BART“as determined by the State.” In the petitioners’ view,this all indicates that the statute unambiguouslyprescribes a limited role for the EPA as regards BARTdeterminations.

In interpreting the CAA, we must follow theguidance set forth in Chevron, U.S.A., Inc. v. NaturalRes. Def. Council, Inc., 467 U.S. 837 (1984). “If thestatute is clear, we apply its plain meaning and theinquiry ends.” Ariz. Pub. Serv. Co. v. EPA, 562 F.3d1116, 1123 (10th Cir. 2009) (quotation omitted). “If thestatute is silent or ambiguous about the question atissue . . . we defer to the authorized agency and apply

App. 12

the agency’s construction so long as it is a reasonableinterpretation of the statute.” Id. (quotation omitted).“[A]dministrative implementation of a particularstatutory provision qualifies for Chevron deferencewhen it appears that Congress delegated authority tothe agency generally to make rules carrying the forceof law, and that the agency interpretation claimingdeference was promulgated in the exercise of thatauthority.” United States v. Mead Corp., 533 U.S. 218,226-27 (2001).

We agree with the EPA that the statute providesthe agency with the power to review Oklahoma’s BARTdetermination for these four units. The EPA may notapprove any plan revision “if the revision wouldinterfere with any applicable requirement concerningattainment and reasonable further progress . . . or anyother applicable requirement of this chapter.” 42 U.S.C.§ 7410( l). And under § 7410(a)(2)(J) SIPs must “meetthe applicable requirements of . . . part C of thissubchapter”—which includes the provisions of the CAArelated to visibility. See §§ 7491, 7492.

The visibility statute itself requires the EPA topromulgate regulations that “require each applicableimplementation plan . . . to contain such emissionlimits, schedules of compliance and other measures asmay be necessary to make reasonable progress towardmeeting the national goal.” § 7491(b)(2). That includesa requirement that the state make BARTdeterminations. And while it is undoubtedly true thatthe statute gives states discretion in balancing the fiveBART factors, it also mandates that the state adhere tocertain requirements when conducting a BARTanalysis. The state plan must include a BART

App. 13

determination for any eligible plant that “mayreasonably be anticipated to cause or contribute to anyimpairment of visibility in any such area.” § 7491(b)(2).In addition, § 7491(b) requires that the BARTdetermination for units at power plants like those atissue here—having a total generating capacity ofgreater than 750 megawatts—“shall be determinedpursuant” to the EPA regulations. See 42 U.S.C.§ 7491(b) (“In the case of a fossil-fuel fired generatingpowerplant having a total generating capacity in excessof 750 megawatts, the emission limitations requiredunder this paragraph shall be determined pursuant toguidelines, promulgated by the Administrator underparagraph (1).”); see also EPA Br. at 7.

As required by the statute, the EPA haspromulgated regulations providing guidelines formaking BART determinations. Like the statute, theregulations require that BART determinations at largepower plants follow these guidelines. 40 C.F.R.§ 51.308(e)(1)(ii)(B) (“The determination of BART forfossil-fuel fired power plants having a total generatingcapacity greater than 750 megawatts must be madepursuant to the guidelines appendix Y of this part(Guidelines for BART Determinations Under theRegional Haze Rule).”). The EPA rejected Oklahoma’sSIP because the BART determinations failed to complywith these guidelines. See EPA Br. at 22 (“Specifically,EPA concluded that Oklahoma failed to reasonablyconsider the ‘cost of compliance’ factor by calculatingcosts as required by the BART guidelines, which led toan ‘unreasoned and unjustified’ BART determination.”).Given that the statute mandates that the EPA mustensure SIPs comply with the statute, we fail to see howthe EPA would be without the authority to review

App. 14

BART determinations for compliance with theguidelines.

The D.C. Circuit’s opinion in American CornGrowers Ass’n v. EPA, 291 F.3d 1 (D.C. Cir. 2002), doesnot alter this conclusion. At issue in Corn Growers wasa provision of the Regional Haze Rule that requiredstates to make BART decisions based in part on thegeographical location of a source, as opposed to itsactual emissions. Id. at 4-5. The rule requiredBART-eligible sources be subject to BART “even absentempirical evidence of that source’s individualcontribution to visibility impairment in a Class I areaso long as the source is located within a region thatmay contribute to visibility impairment.” Id. at 5.When making the BART determination, the stateneeded to “analyze the degree of visibility improvementthat would be achieved . . . as a result of the emissionreductions achievable from all sources subject to BARTlocated within the region that contributes to visibilityimpairment.” Id. at 6 (quotation and emphasisomitted). The D.C. Circuit held the EPA’s approachwas “inconsistent” with the CAA. Id. at 7-8.

The D.C. Circuit cited two ways in which the rulewas inconsistent with the statute. First, the EPA’sapproach “distort[ed] the judgment Congress directedthe states to make for each BART-eligible source” bytreating one of the five BART factors differently thanthe others. Id. at 6. The rule, for instance, prevented astate from “consider[ing] the degree to which newequipment at a particular source would help cure thehaze in some distant national park.” Id. at 7. “UnderEPA’s take on the statute, it is therefore entirelypossible that a source may be forced to spend millions

App. 15

of dollars for new technology that will have noappreciable effect on the haze in any Class I area.” Id.

Second, the D.C. Circuit said that the ruleimpermissibly “constrain[ed] authority Congressconferred on the states.” Id. at 9. The court said thatthe statute and the legislative history suggested thatthe states had broad authority to weigh the statutoryfactors and make BART determinations. Id. at 8. TheD.C. Circuit noted that the Conference Report on the1977 amendments to the CAA specifically referenced“an agreement to reject the House bill’s provisionsgiving EPA the power to determine whether a sourcecontributes to visibility impairment and, if so, whatBART controls should be applied to that source.” Id.The agreement instead added the language delegatingthis authority to the state. Id. “The Conference Reportthus confirms that Congress intended the states todecide which sources impair visibility and what BARTcontrols should apply to those sources.” Id. The HazeRule, though, “ties the states’ hands and forces them torequire BART controls at sources without anyempirical evidence of the particular source’scontribution to visibility impairment in a Class I area.”Id.

Here, though, the statute and the legislative historysupport our conclusion that the EPA may reject BARTdeterminations that do not comply with the guidelines.True, the modification of the original House bill reflectsan intent to shift the power to determine BART fromthe EPA to the states. But, as above, it still placedstatutory limits on those state decisions. While thelegislative history may evidence an intent to preventthe EPA from directly making those BART decisions, it

App. 16

does not necessarily evidence an intent to deprive theEPA of any authority to ensure that these BARTdecisions comply with the statute. In the present case,the EPA did not reject the petitioners’ BARTdetermination because it disagreed with the way itbalanced the five factors. It rejected the BARTdetermination because it failed to follow theguidelines—as required by the statute—in calculatingone of those factors.

All the conference agreement referenced by the D.C.Circuit did was shift the initial responsibility formaking BART determinations from the EPA to thestate. But that does not differ from other parts of theCAA—states have the ability to create SIPs, but theyare subject to EPA review. In addition, the ConferenceReport emphasized that the BART determinations forlarge power plants must comply with EPA guidelines:

The agreement clarifies that the State,rather than the Administrator, identifies thesource that impairs visibility in the Federalclass I areas identified and thereby fall withinthe requirements of this section. . . . .

In establishing emission limitations for anysource which impairs visibility, the State shalldetermine what constitutes ‘best availableretrofit technology’ (as defined in this section) inestablishing emission limitations on a source-by-source basis to be included in the Stateimplementation plan so as to carry out therequirements of this section. The regulationsand Federal guidelines required by the Housepassed bill for determining this technology are

App. 17

eliminated for all sources other than fossil fuelelectric generating plants with a total generatingcapacity in excess of 750 megawatts.

H.R. Rep. No. 95-564, at 155 (1977) (Conf. Rep.)(emphasis added).

The Senate discussion about the Conference Reportalso highlighted the role that the guidelines play inBART determinations for large power plants:

[Senator] McClure. And while those existingsources are limited to the 28 major sourcescontained in the Senate bill’s definition of majoremitting facilities, exempting any such sourcewhich has the maximum potential to emit lessthan 250 tons per year, Federal guidelines applyonly to fossil-fuel fired generating plants inexcess of 750 megawatts?

[Senator] Muskie. That is correct.

[Senator] McClure. Under the conferenceagreement, does the State retain sole authorityfor identification of sources for the purpose ofvisibility issues under this section?

[Senator] Muskie. Yes; the State, not theAdministrator, identifies a source that mayimpair visibility and thereby falls within therequirement of section 128.

[Senator] McClure. And does this also hold truefor determination of “Best Available RetrofitTechnology”?

[Senator] Muskie. Yes; here again it is the Statewhich determines what constitutes “Best

App. 18

Available Retrofit Technology,” as defined insection 128. The Federal guidelines apply only tothe large powerplants we have described.

123 Cong. Rec. S26,854 (daily ed. Aug. 4, 1977)(emphasis added). The last sentence—omitted bypetitioners in their brief—makes clear that the statuterequires that the BART determination here complywith the guidelines. See Pet. Opening Br. at 15. Andbecause the EPA monitors SIPs for compliance with thestatute, it must monitor BART determinations forcompliance with the guidelines. To be sure, theguidelines themselves might somehow conflict with thestatute. But the petitioners have not argued that anyconflict exists here.3 We therefore hold that the EPA

3 In its amicus brief, the American Coalition for Clean CoalElectricity asserts that some conflict between the guidelines andstatute may exist because:

EPA can provide the States with guidelines only “onappropriate techniques and methods,” including“(A) methods for identifying, characterizing, determining,quantifying, and measuring visibility impairment inFederal areas referred to in paragraph (1), (B) modelingtechniques (or other methods) for determining the extentto which manmade air pollution may reasonably beanticipated to cause or contribute to such impairment, andmethods for preventing and remedying such manmade airpollution and resulting visibility impairment.” Seecross-reference from [42 U.S.C. § 7491(b)] (last paragraph)to [§ 7941(b)(1) to § 7491(a)(3)]. Thus, EPA’s role is toprovide procedural and technical guidance to the States inmaking BART determinations.

See Am. Coalition for Clean Coal’s Br. at 12 (emphasis added).However, the statute does not limit the guidelines so restrictively.First, the guidelines must be made as part of regulations that

App. 19

had the authority to review Oklahoma’s BARTdetermination with respect to these two power plants.

III

Having held that the EPA possesses the authorityto review these BART decisions, we must nowdetermine whether the EPA lawfully exercised thatauthority when it rejected Oklahoma’s SIP. Petitionersargue that the EPA took arbitrary and capriciousaction in rejecting two sets of cost estimates they usedin determining BART. The EPA, on the other hand,argues that it properly rejected these estimates—and,thus, the SIP that relied on them—for failure to complywith its guidelines.

We follow the standards of the AdministrativeProcedure Act (APA) in reviewing the EPA’s actionsunder the CAA. See Magnesium, 690 F.3d at 1164.Under the APA, we must hold unlawful any agencyaction that is “arbitrary, capricious, an abuse ofdiscretion, or otherwise not in accordance with law.” 5U.S.C. § 706(2). “Under the arbitrary or capriciousstandard, we must determine whether the agencyconsidered the relevant data and rationally explained

ensure “reasonable progress toward meeting the national goals”specified in the statute. § 7491(a)(4), (b)(1). Second, thoseguidelines must merely “tak[e] into account” recommendationsfrom a report to Congress on the methods and techniquesreferenced in the statute, 42 U.S.C. § 7491(b)(1), which includes“methods for preventing and remedying such manmade airpollution and resulting visibility impairment.” 42 U.S.C.§ 7491(a)(3)(C). Moreover, the amicus brief fails to explain why theEPA could provide these regulations providing procedural andtechnical guidance, but yet lacks the authority to ensure statescomplied with them.

App. 20

its decision.” Ariz. Pub. Serv. Co., 562 F.3d at 1122.“Agency action is arbitrary or capricious if the agencyhas relied on factors which Congress has not intendedit to consider, entirely failed to consider an importantaspect of the problem, offered an explanation for itsdecision that runs counter to the evidence before theagency, or is so implausible that it could not beascribed to a difference in view or the product of agencyexpertise.” Id. at 1123. (quotation omitted). “Even whenan agency explains its decision with less than idealclarity, a reviewing court will not upset the decision onthat account if the agency’s path may reasonably bediscerned.” Alaska Dep’t of Envtl. Conservation, 540U.S. at 497 (quotation omitted). In addition, we notethat “[w]hen an agency interprets its own regulation,the Court, as a general rule, defers to it unless thatinterpretation is plainly erroneous or inconsistent withthe regulation.” Decker v. Nw. Envtl. Def. Ctr., 133 S.Ct. 1326, 1337 (2013) (quotation omitted).

A. 2008 Cost Estimates

The petitioners argue that the EPA arbitrarilyrejected a set of cost estimates that OG&E submittedto the EPA in 2008 (2008 Cost Estimates). Thepetitioners claim that the “EPA acknowledged that‘OG&E did utilize the “EPA Air Pollution Control CostManual” when constructing its [May 2008] costestimates.’” Pet. Opening Br. at 20 (quoting EPARegion 6 Comments on ODEQ’s BART EngineeringAnalyses, JA at 1132 (Nov. 4, 2008)). The 2008 CostEstimates were “more than ten times EPA’s statedaverage costs per ton for this technology, and nearlyfive times as much as the upper limit of EPA’s expectedcost range.” Id. at 21. The EPA should have addressed

App. 21

these numbers, petitioners say, because they prove thescrubbers were not cost effective.

This argument is without merit. The EPA neverstated that the 2008 Cost Estimates complied with theControl Cost Manual.4 In context, the EPA simplyacknowledged that OG&E purported to have used themanual in constructing these estimates. Indeed, itrequested that OG&E note any deviations from the costmanual—as required by the guidelines5—after pointingout that OG&E’s estimates “seem[ed] high compared towhat EPA has seen in other BART analysis.” EPA

4 The entire EPA comment included the following:

Regarding its cost estimates, OG&E’s estimates seem highcompared to what EPA has seen in other BART analyses.OG&E cites increased equipment costs, in part due to the“sellers market” that resulted from the CAIR program.Since the CAIR has been vacated, OG&E should solicitrevised bids from pollution control equipment vendors.Region 6 is aware of similarly sized and configuredfacilities that estimate much lower costs for theinstallation of wet or dry FGD systems. Region 6 notes thatOG&E did utilize the “EPA Air Pollution Control CostManual” when constructing its cost estimates. However,OG&E should also note any areas in which where it hasdeviated from that guidance.

EPA Region 6 Comments, JA at 1132 (emphases added).

5 The guidelines state that “cost estimates should be based on theOAQPS Control Cost Manual, where possible.” 40 C.F.R. pt. 51app. Y(IV)(D)(4)(a). However, it also instructs that parties “shouldinclude documentation for any additional information regardingpurchased equipment costs, equipment life, replacement of majorcomponents, and any other element of the calculation that differsfrom the Control Cost Manual.” 40 C.F.R. pt. 51 app. Y(IV)(D)(4)(a)n.15.

App. 22

Region 6 Comments, JA at 1132; see also 40 C.F.R. pt.51 app. Y(IV)(D)(4)(a) n.15.6

And in any event, the EPA did not “ignore[]” the2008 Cost Estimates. Rather, the EPA explained that“[t]hese 2008 costs are not valid under the overnightcosting method” required by the manual. Response toTechnical Comments, JA at 1236. The 2008 CostEstimates “contain[ed] . . . fundamental methodologicalflaws, such as including escalation and Allowance forFunds Used During Construction (AFUDC).” Id. “Thecost of scrubbers would not be substantially higherthan those reported for other similar projects if OG&Ehad used the costing method and basis, i.e., overnightcosts in current dollars, prescribed by the Control Cost

6 Petitioners argue in their reply brief that the EPA did notexplicitly raise this response in its comments and therefore cannotuse this reasoning to justify its decision on appellate review. Pet.Reply Br. at 10 n.1 (quoting Motor Vehicles Mfrs. Ass’n v. StateFarm Mut. Auto. Ins. Co., 463 U.S. 29, 50 (1983) (“[A]n agency’saction must be upheld, if at all, on the basis articulated by theagency itself.”)). But an agency “need not address every comment,[although] it must respond in a reasoned manner to those thatraise significant problems.” Covad Commc’ns Co. v. FCC, 450 F.3d528, 550 (D.C. Cir. 2006) (quotation omitted). “The failure torespond to comments is significant only insofar as it demonstratesthat the agency’s decision was not based on a consideration of therelevant factors.” Id. (citations and quotation omitted). “In makingthe ‘keystone’ inquiry whether the [agency] engaged in reasoneddecisionmaking, the reviewing court is to consider the largeradministrative record.” Mt. Diablo Hosp. v. Shalala, 3 F.3d 1226,1234 (9th Cir. 1993) (quotation omitted).

Here, the letter itself—which is part of the record—quiteclearly explains the EPA’s concerns about the 2008 CostEstimates. And the petitioners did, in fact, submit new costestimates.

App. 23

Manual,” the EPA said. Id. The EPA therefore had areasonable basis for rejecting the 2008 Cost Estimatesas not complying with the guidelines.

B. The 2009 Cost Estimates

The petitioners also argue that the EPA actedarbitrarily in rejecting the cost estimates submitted in2009 (2009 Cost Estimates). These cost estimatesincluded more site-specific data than the ones OG&Einitially submitted. The petitioners say that the EPAshould defer to Oklahoma’s determination that the“site-specific cost information submitted by OG&E in2009 was credible, detailed, and specific for theindividual facilities, going well beyond the defaultmethodology recommended by the EPA guidance.” Pet.Opening Br. at 23 (quotation omitted). This argumentappears to be premised, in part, on our accepting therepresentation that the EPA conceded the 2008 CostEstimates complied with the guidelines. Thepetitioners argue in their reply brief that for “EPA toacknowledge receipt of cost estimates that it agreedwere in compliance with the CCM, ask for site-specificcost estimates that go beyond the CCM, and then rejectthose site-specific estimates for not conforming to theCCM exemplifies arbitrary and capricious results-oriented decisionmaking.” Pet. Reply Br. at 11.

We disagree with the petitioners’ characterizationof the EPA’s actions. The EPA did not reject the mereuse of any site-specific costs as not complying with theguidelines. It rejected this particular use of site-specificnumbers because it “recognized that how OG&Especified those vendor quotes and [its] subsequent useof them in its cost analysis was flawed.” Response toTechnical Comments, JA at 1308. The guidelines

App. 24

require that states provide support for any site-specificcosts that depart from the generic numbers in theControl Cost Manual. See 40 C.F.R. pt. 51 app.Y(IV)(D)(4)(a) n.15 (“You should include documentationfor any additional information you used for the costcalculations, including any information supplied byvendors that affects your assumptions regardingpurchased equipment costs, equipment life,replacement of major components, and any otherelement of the calculation that differs from the ControlCost Manual.”). OG&E never delivered to the EPAthese detailed vendor estimates, preventing the EPAfrom conducting an adequate review to ensure thesedepartures from the manual—and Oklahoma’sapproval of them—were justified. See 76 Fed. Reg. at81,745 (“[M]uch of the documentation OG&E andothers cite to support deviations from the Control CostManual was not provided to us. Thus, we were unableto analyze their contents and determine whether thesedeviations were appropriate.”); see, e.g., Response toTechnical Comments, JA at 1239 (“For instance,although OG&E provided two spreadsheets that listedtheir cost line items, these spreadsheets, each over 600lines in length (and including line items such asseeding and fertilizing the grass at the plant sites),were stripped of all cell calculations, preventing anymeaningful review.”). And, as we discuss below inevaluating the EPA’s action in promulgating its FIP,many of OG&E’s costing assumptions wereunjustified.7

7 We recognize that the EPA has less discretion when it takesactions to reject a SIP than it does when it promulgates a FIP.However, we believe that the EPA had reason to make the

App. 25

Moreover, the EPA’s consultant noted that, even ifthe departures from the generic numbers had beenproperly documented, the resulting costs were notanalyzed in compliance with the manual. See RevisedBART Cost-Effectiveness Analysis, JA at 1517 (October2010) (“[I]t is possible to follow the generic costingmethod in the Cost Manual, relying on vendor quotesand other information to estimate scrubber capital andO&M costs. [OG&E’s consultants] used vendorestimates for equipment costs but did not follow thegeneric costing method.”). The guidelines say thatstates should follow the manual’s methodology so thatprojects can be more easily compared. The EPA saidthat OG&E should have used the “overnight” costingmethodology. Instead, “OG&E and others incorrectlyassume that BART cost effectiveness should be basedon the ‘all-in’ cost method, which includes all of thecosts of a financial transaction, including interest,commissions, and any other fees from a financialtransaction up to the date that the project goes intooperation, as of the assumed commercial operatingdates of the scrubbers, 2014 and 2015.” 76 Fed. Reg. at81,744.

For their part, the petitioners argue generally thatthe Control Cost Manual does not require the use of the“overnight” method used by the EPA. However, they donot point to any specific parts of the guidelines or theControl Cost Manual that contradict the EPA’sapproach. Instead, the petitioners argue that the EPA

adjustments described in Section IV, Part B, even under the higherstandard we would apply when evaluating its actions in rejectinga SIP. OG&E has yet to provide any justification for providingestimates that departed from the guidelines.

App. 26

itself conceded that its methodology excluded onlyinflation. Pet. Opening Br. at 31. (“EPA claimed thatthe CCM required compliance with a ‘constant dollar’approach . . . . The constant dollar approach allowscomparability in the BART context by removing theeffects of inflation from cost estimates.”). Petitionersbelieve this means that it properly included the othercosts, such as the Allowance for Funds Used DuringConstruction, in its analyses.

Unlike the petitioners, we do not read the EPA’sstatement as proof that the EPA believed theconstant-dollar method requires removing onlyinflation. The EPA’s consultant referenced themanual’s adherence to the “constant-dollar” method asthe reason for excluding inflation in the estimates. SeeRevised BART Cost-Effectiveness Analysis, JA at 1517(“The cost metric estimated in the Manual is real orconstant-dollar costs in that the effect of inflation hasbeen removed.”). But the consultant never said that theControl Cost Manual excluded only inflation. In fact,the consultant explained throughout her report thatthe Control Cost Manual also required excluding manyof the other costs on which OG&E had relied. See JA at1519 (“Cost items such as escalation of costs, bond cost,and AFUDC are not part of the Cost Manualmethodology for estimating costs.”); see, e.g., JA at1520 (“[Allowance for Funds Used DuringConstruction] are not part of the constant dollarapproach found in the EPA Control Cost Manual andshould not be included in the BART cost-effectivenessanalysis.”). Viewing these comments as a whole, we donot believe the EPA or its consultant ever concededinflation was the only cost that needed to be eliminatedfrom OG&E’s estimates.

App. 27

The petitioners also refer to two affidavits fromtheir experts that detail what they believe is the propercosting methodology. Even if we permitted thesearguments to be incorporated by reference, we cannotconsider these affidavits because they are outside of theadministrative record. 42 U.S.C. § 7607(d)(7)(A).

Aware of this hurdle, petitioners criticize the EPA’sprocedure, arguing that they had no opportunity toobject to the EPA’s use of the so-called “overnight”costing method because the EPA used it for the firsttime in the final rule. To be sure, the EPA used theterm “overnight” method for the first time in the finalrule. However, the EPA excluded the same costs in thefinal rule that it did throughout the entire process—the“overnight” method was simply the shorthand it usedin the Final Rule to describe the exclusion of thesecosts. In fact, the petitioners’ own comments to theEPA belie the argument they have made to us, sincethey challenged the exclusion of these costs in theadministrative proceeding. See, e.g., Ex. B to OG&E’sComments: May 2011 BART Cost Analysis Report, JAat 1156 (May 20, 2011) (“EPA’s consultant incorrectlyargues that an AFUDC is not part of the constantdollar approach found in the EPA Control Cost Manualand should not be included in the BARTcost-effectiveness analyses.” (quotation omitted)). Wesee no reason to excuse petitioners’ failure to raisethese substantive arguments in their brief.

Additionally, we do not have jurisdiction to considerany procedural error that might have occurred as aresult of the EPA allegedly using the “overnight”method for the first time in the Final Rule. Under theCAA, “[o]nly an objection to a rule or procedure which

App. 28

was raised with reasonable specificity during theperiod for public comment . . . may be raised duringjudicial review.” 42 U.S.C. § 7607(d)(7)(B). The factthat the petitioners could not have raised theirobjection about the use of the “overnight” method untilafter the final rule was published does not excuse themfrom the requirement they first raise the issue with theEPA. “Rather, the CAA requires a petitioner to firstraise its objection to the agency th[r]ough a petition forreconsideration.” Appalachian Power Co. v. EPA, 249F.3d 1032, 1065 (D.C. Cir. 2001). Petitioners here,though, filed their petition for reconsideration the sameday they filed this petition for review. We thereforelack jurisdiction to rule on this procedural objection.See 42 U.S.C. § 7607(d)(7)(B) (“If the Administratorrefuses to convene [a reconsideration proceeding], suchperson may seek review of such refusal in the UnitedStates court of appeals for the appropriate circuit . . . .”)(emphasis added).

IV

Petitioners also challenge decisions made by theEPA in promulgating the FIP. First, the petitionersargue that the EPA itself failed to follow the guidelinesbecause it did not base its cost-effectiveness analysis onthe historical emissions baseline. Second, petitionersargue that the EPA based its analysis on incorrecttechnological assumptions about the size of thescrubber that needs to be built. Third, the petitionerscriticize many of the adjustments the EPA made to itscost estimates. Finally, the petitioners assert that“[s]crubbers at the OG&E units would not have asignificant impact on visibility.” Id. at 36. We reviewthese challenges under the same arbitrary and

App. 29

capricious standard we used to evaluate the EPA’srejection of the SIP. However, we do so whilerecognizing this requires a slightly differentperspective: evaluating the EPA’s own choices underthe guidelines, as opposed to evaluating its choice toreject the Oklahoma SIP under the guidelines.

A. Baseline Emissions/Technical Feasibility

Petitioners argue the EPA acted arbitrarily when itpromulgated its FIP because it ignored the units’ pastrates of SO2 emissions in conducting its analysis.Under the guidelines, “[a]verage cost effectivenessmeans the total annualized costs of control divided byannual emissions reductions (the difference betweenbaseline annual emissions and the estimate ofemissions after controls).” 40 C.F.R. pt. 51 app.Y(IV)(D)(4)(c). “The baseline emissions rate shouldrepresent a realistic depiction of anticipated annualemissions for the source.” Id. at Y(IV)(D)(4)(d). “Ingeneral, for the existing sources subject to BART, youwill estimate the anticipated annual emissions basedupon actual emissions from a baseline period.” Id.“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) willdiffer from past practice, and if this projection has adeciding effect in the BART determination, then youmust make these parameters or assumptions intoenforceable limitations.” Id.

OG&E has been voluntarily using low-sulfur coal atthese power plants. It therefore estimated thatscrubbers would remove around 14,000 tons of SO2 peryear at each of the power plants. Comments of OG&Eon Proposed EPA Rule, JA at 1106. The EPA, however,

App. 30

assumed that OG&E would begin to use high-sulfurcoal if it installed the proposed scrubbers. The EPA,therefore, estimated that scrubbers would remove43,428 tons of SO2 per year at one of the power plants,and 46,458 tons of SO2 per year at the other. Id. Byassuming OG&E would remove larger amounts of SO2,while using cheaper, high-sulfur coal, the resultingconclusion was that the scrubbers appeared more costeffective.

This links with petitioners’ second objection: thatthe EPA impermissibly based its analysis on theconstruction of smaller, less expensive—but allegedlytechnically infeasible—scrubbers. Petitioners claimthat EPA wrongly assumed that OG&E had the optionof building a smaller scrubber based on the sulfurcontent of the coal it would burn. Scrubber size doesnot depend on the sulfur content of coal, the petitionerssay. Rather, scrubber size must “reflect the maximumpotential heat input from the facility, and that numberis essentially the same whether a facility burns high orlow sulfur coal.” Pet. Opening Br. at 28. A smallerscrubber would be impossible to build, or wouldsignificantly diminish the units’ electricity production,petitioners say.

The EPA says it made its calculations in response tothe flawed assumptions made by the petitioners. In theEPA’s view, the petitioners’ analysis assumed that theywould be building a much more powerful—and, moreimportant, costly—scrubber system than was actuallyneeded. The petitioners still assumed, though, thatthey would use low-sulfur coal in the future, makingthe scrubbers seem less cost effective than theyactually would be, the EPA said.

App. 31

The EPA’s consultant, Dr. Phyllis Fox, analyzed twooptions to account for these alleged flaws in OG&E’sanalysis. In Option 1, Fox evaluated the costeffectiveness of the high powered scrubbers using anahistorical baseline. Revised BART Cost-EffectivenessAnalysis, JA at 1513. That is, the analysis assumedthat OG&E—while currently using low-sulfurcoal—would begin using cheaper, higher sulfur coal ifit actually built these larger scrubbers. Id. at 1280.This increased the scrubbers’ cost effectiveness becauseit anticipated the removal of greater amounts of SO2.Id. at 1513-14.

In Option 2, the EPA evaluated the costeffectiveness of a lower powered scrubber using thehistorical baseline emissions. Id. at 1514. In order todesign this less powerful scrubber, the EPA used amodel that OG&E’s consultants had created for theEPA in a different action.8 JA at 1283. This optionincreased the cost effectiveness of the scrubbers—assuming the continued use of low-sulfur coal—becauseit assumed the scrubbers would be less expensive.

The evaluation of the petitioners’ argumentessentially hinges on the technical feasibility of asmaller unit. We agree with the petitioners that theguidelines typically require the use of the historicalemissions baseline. And we agree that if the EPA wereto, without justification, increase the historicalemissions baseline, it would encourage the use of

8 The EPA used this model in the final rule. The EPA’s consultanthad used a less precise model for basic engineering in the proposal,which the EPA conceded was an “oversimplification.” Response toTechnical Comments, JA at 1283.

App. 32

high-sulfur coal simply for the purpose of removinggreater amounts of SO2.

But what is framed as an argument over thebaseline is really an argument over the size of theproposed scrubbers. The cost effectiveness of scrubberswould routinely be understated if proposals includedcosting for scrubbers much larger than needed. “Whileit may be prudent to overdesign for many reasons, thecost of overdesign should not be attributed to BART,especially when the emission reductions do notconsider the overdesign,” the EPA said. Response toTechnical Comments. Id. at 1283. If OG&E wantedlarger scrubbers, then the EPA needed some way toisolate the BART-related costs from the non-BART-related costs. The EPA took this into account byadjusting the baseline emissions under the assumptionthat, if OG&E built the scrubbers it proposed, it wouldpresumably shift to using cheaper forms of coal.

Left to evaluate the arguments of the parties’experts, we must give deference to the EPA. See SanJuan Citizens Alliance v. Stiles, 654 F.3d 1038, 1045(10th Cir. 2011) (“The deference we give agency actionis especially strong where the challenged decisionsinvolve technical or scientific matters within theagency’s area of expertise.” (quotation omitted)).9 While

9 The petitioners argue that we should not afford the EPAdeference because the EPA’s consultant did not speak to OG&Edirectly and did not visit the site. Pet. Br. at 19 n.9. They cite noauthority for this proposition. Further, the EPA said in its finalrule that it “met with OG&E and its consultant concerning thedevelopment of our proposal and had extensive communicationsclarifying particular technical points,” which it conveyed to itsconsultant to incorporate in her report. 76 Fed. Reg. at 81,728.

App. 33

the petitioners criticize some of the engineeringassumptions made by the EPA, they do not explain whythe EPA was not justified in relying on OG&E’s ownconsultant’s model, or why the EPA’s detailedresponses in its technical support document wereinsufficient in addressing its concerns. For example,the petitioners state that “the smaller scrubberenvisioned by EPA [option 2] might work for some daysof operation at the OG&E Units, but it would precludeOG&E from producing electricity at higher levels whenneeded.” Pet. Opening Br. at 29 n.18. But the EPAassumed in its analysis that the plants would operateat a 100% capacity factor. Technical SupportDocument, JA at 1348.

To be sure, our dissenting colleague raises a numberof valid concerns about the EPA’s actions, and weacknowledge that this is a close case. But, ultimately,we cannot adopt OG&E’s analysis given that the EPAwas aware of, and provided explanations contradicting,petitioners’ comments.10 For instance, the petitionerscontend that the EPA relied too heavily on the sulfurcontent of the fuel in downsizing the scrubbers. “Ascrubber must be sized to reflect the maximumpotential heat input from the facility, and that numberis essentially the same whether a facility burns high or

10 The dissent argues that “[t]he EPA deserves no deference [ontechnical determinations], however, where it does not support aconclusion contradicting Oklahoma’s first, reasonable, detailedtechnical analysis.” Dissent at 3. However, the dissent does notdisagree with our conclusion that the EPA had sufficient reasonsfor rejecting cost estimates—rife with errors—submitted byOG&E. It is not clear how the dissent reached the conclusion that,despite these errors, other portions of Oklahoma’s analysis were“reasonable.”

App. 34

low sulfur coal,” the petitioners say. Pet. Opening Br.at 28. The EPA “agree[d] that the sulfur content of thefuel, taken by itself, will not significantly affect the size(or cost) of the gas path portions of the FGD system.”Response to Technical Comments, JA at 1283.However, the EPA noted that the “design and sizing ofa scrubber is generally divided into two major systems:(1) flue gas path and (2) reactant handling system.” Id.at 1284. The EPA pointed out that OG&E’s ownconsultant had said in a previous case that “gross unitsize in MW . . . and sulfur content of the fuel are themajor variables” needed “in order to predict futureretrofit costs.” Id. at 1285. Ultimately, the EPAconcluded that “the use of the lower sulfur coal alonewould reduce the capital cost of the scrubber by about$7 million or 3%.” Id. at 1284. Given that we mustdefer to the EPA’s technical judgments, we cannot saythe EPA acted arbitrarily on the basis of the recordbefore us and the petitioners’ arguments in their brief.

B. 2009 Cost Estimates

The petitioners also argue that the EPA improperlyadjusted the 2009 Cost Estimates in promulgating theFIP. Petitioners make three arguments. First, thepetitioners argue that the EPA made unreasonableadjustments to the site-specific numbers put forth bypetitioners. Second, the petitioners argue that the EPAimproperly relied on the “overnight” cost methodinstead of the constant-dollar approach. Finally, thepetitioners argue the EPA selectively manipulatedmany of the other input variables in order to forceOG&E to install scrubbers.

App. 35

1) Double Counting and Discounts

The petitioners claim that in performing theanalysis that formed the basis of the FIP, the EPAconsultant arbitrarily discounted some of the costs thatthe petitioners had used in their analysis. Reviewingthe record, we do not believe these adjustments werearbitrary or capricious. The EPA explained the basisfor each of the adjustments it made. For example, theEPA discounted vendor quotes to reflect the likelyefficiencies from building multiple identical units.Revised BART Cost-Effectiveness Analysis, JA at1530-31. It applied a 5% discount after a “search of theliterature revealed a range of 4% to 10% savings fromoptimized equipment.” Response to TechnicalComments, JA at 1229.

Further, the EPA provided an extensive response tothe petitioners’ comments on this point. The petitionersargued during the notice-and-comment period that theEPA should not have assumed that the quotes did notalready contain a multiple-unit discount because thevendors knew they were constructing multiple units. Inresponse, the EPA explained why it believed thesequotes did not factor in efficiency discounts. Id. at 1228.OG&E had specifically asked for single-unit estimates;the prices OG&E used in its report were “exactlydouble the per-unit prices in the vendor quotes.” Id. at1229. There was “no evidence that any of these vendorquotes considered multiple unit discounts,” eventhough “[b]ased on common industry practice, [theEPA] expected to see vendor discounts for multipleunits in OG&E’s site specific cost estimates.” Id. at1228. In addition, one of the vendors specifically “notedthat the cost of the equipment and the design could be

App. 36

optimized to provide more economical operation.” Id. at1229.

Likewise, the EPA explained many of the otheradjustments it made to the costs submitted by OG&E.For instance, OG&E’s proposal estimated contingencycosts at 14% of the total project capital cost plusescalation, without providing any details on how thisnumber was calculated. Revised BART Cost-Effectiveness Analysis, JA at 1520. The consultantflagged this number as high, noting that the ControlCost Manual only permits the consideration of alimited number of contingency costs. “A contingencyfactor in a cost effectiveness analysis ‘should bereserved (and applied to) only those items that couldincur a reasonable but unanticipated increase but arenot directly related to the demolition, fabrication, andinstallation of the system,’” the consultant said. Id. at1521 (quoting Control Cost Manual, Chapter 2, CostEstimation: Concepts and Methodology, Sec. 2.5.4,p. 2-30). In the absence of any justification for OG&E’sestimate, the consultant adjusted contingency costs tothe standard 3% of purchased equipment costs. Id.

Similarly, the consultant adequately explained thedecision to apply a discount to the “owner’s costs”estimated by OG&E. OG&E estimated “owner’s cost” at5% of total capital expended, much higher than the 2%its consultant had used in other BART analyses. Id. at1527. The EPA’s consultant acknowledged that ownersincur some costs related to construction. Id. But theconsultant explained that many of the costs that OG&Eincluded in its estimate of “owner’s costs”—such as siteoversight—had already been included under theindirect capital costs of engineering/procurement and

App. 37

construction. Id. In light of this, the consultantremoved owner’s costs from the analysis. Id. (“Owner’scosts are not separately included in BART costeffectiveness analyses and have been more than doublecounted here.”).

We do not believe these types of adjustments werearbitrary or capricious. Moreover, even after the EPAmade these adjustments, at least some of these site-specific costs were higher than the generic numbers theEPA could otherwise have used. See, e.g., id. at 1528(“This change reduces the engineering andprocurement costs by $12,733,100 at Sooner and by$12,944,277 at Muskogee. These values are still highcompared to estimates based on the Cost Manualmethod of 10% purchased equipment costs.” (footnotesomitted)). In fact, the EPA consultant concluded thatthe EPA’s cost estimates were likely higher than actualcosts would be. Id. at 1510 (“Actual costs could be evenlower as I was unable to correct all of the overestimatesthat I identified due to lack of support and underlyingcalculations.”).

2) Constant-Dollar Basis

Petitioners also argue that the EPA departed fromthe Control Cost Manual in removing costs aside frominflation from its estimates. As above, we see no meritto this argument. The petitioners have not made apersuasive case in their briefs that either the EPA’s

App. 38

methodology conflicts with the manual11 or that theEPA conceded it should only have excluded inflation.

3) Selective Manipulation

Finally, the petitioners claim the EPA selectivelymanipulated its data to meet its desired result. Thepetitioners argue the EPA manipulated the data inthree ways. First, petitioners argue that the EPAarbitrarily accepted some site-specific numbers and notothers. Second, petitioners again reiterate complaintsabout the EPA’s adjustments to its estimates. Third,petitioners argue that the EPA assumed too long of auseful life for the scrubbers.

a) Site-specific costs

First, the petitioners argue that the EPA chose onlyto adopt site-specific cost estimates that were higherthan the generic estimates used in the manual. Wecannot agree. As the EPA explained, it used thesesite-specific numbers when there was, in fact, accuratedocumentation—supporting a departure from themanual. Response to Technical Comments, JA at 1273(“We used the Control Cost Methodology (overnightcosts, no inflation, no AFUDC, no income taxes, etc.)and site-specific values when they were valuable andcorrect.”). Petitioners want us to view it as suspiciousthat the EPA accepted the site-specific estimates wherethey were lower than the generic numbers in theguidelines. But one could find it equally suspicious that

11 We are not necessarily endorsing the EPA’s approach to costingmethodology. We note only that the petitioners chose on appeal toabandon the arguments that they made in the administrativehearing that the manual did not support the EPA’s methodology.

App. 39

the petitioners did not provide documentation for itsestimates that were much greater than the generic costnumbers in the manual. Given that many of thesite-specific numbers provided to the EPA were notproperly documented in accordance with the guidelines,we do not find it arbitrary or capricious that the EPArejected them.

b) Double counting

The petitioners also argue that the EPA relied onflawed assumptions when it reduced some of the costestimates for double counting or overestimation. Again,though, the consultant and the EPA thoroughlydocumented why they made these adjustments. We seeno basis for reversing the EPA’s decision on thisground.

c) Useful life of scrubbers

Finally, the petitioners argue that the EPA assumedtoo long of a useful life for the scrubbers. The longerthe useful life of the scrubbers, the more cost effectivethey will appear to be. The EPA assumed a useful lifeof thirty years in its analysis. The petitioners arguethat the EPA departed from the Control Cost Manual’sstandard useful life of twenty years for large pollutioncontrol systems.

But the EPA adequately explained why it chose toassume a useful life of thirty years. As the petitionersacknowledge, the Control Cost Manual does notreference any specific useful life for scrubbers. See Pet.Opening Br. at 34. Therefore, the EPA considered,among other factors: 1) the fact that scrubbers installedbetween 1975 and 1985 are still in use; 2) thestandards from cost estimates handbooks and

App. 40

published papers, and 3) the fact that the EPA hasassumed a 30-year lifetime for scrubbers since “at least1981.” See JA at 1263. The petitioners do not explainwhy this explanation was inadequate.12

Instead, the petitioners argue that the EPA’sdecision was unjustified because the EPA assumed thatthe petitioners would use high-sulfur coal in analyzingOption 1, but then, in its response to the comments,said that the scrubbers would operate in the mild,low-sulfur environment in assuming a useful life ofthirty years. This argument is premised on aninaccurate characterization of the EPA’s comments. In

12 The petitioners characterize the EPA as departing from themanual’s standard useful life of twenty years based solely on anunsupported determination “that significant advances have beenmade in the material of construction and baghouse design sincethis publication.” Pet. Opening Br. at 34. But the EPA was merelyexplaining that its thirty-year useful life for dry scrubbers wasreasonable even in the context of the other control systemsactually mentioned in the manual:

OG&E also asserts that the Control Cost Manual assumesa 20 year useful life for other large air pollution systems,citing 20 years for a fabric filter baghouse and 20 years forSCR. The cite of 20 years for the baghouse actually says:“For fabric filters, the system lifetime varies from 5 to 40years, with 20 years being typical,” citing to a 1980 report.Significant advances have been made in the material ofconstruction and baghouse design since this publication.Further, the 1981 EPA/TVA report cited above, “TechnicalReview of Dry FGD Sytems and Economic Evaluation ofSpray Dryer FGD System” assumes a 30 year economicand tax life for an SDA/FF, which includes the fabric filter.Regardless, our 30 year estimate for a scrubber, whichincludes a baghouse, is well within the reported range.

Response to Technical Comments, JA at 1264 (footnotes omitted).

App. 41

context, the EPA said that even scrubbers inhigh-sulfur environments have useful lives of at leastthirty years:

The subject application, the use of a scrubber toremove SO2 from low sulfur coal, is a mildenvironment for a scrubber, compared to highsulfur applications, which have alreadydemonstrated 30 year lifetimes. The corrosionpotential and bag plugging issues in a low sulfurapplication are much lower than in a comparablehigh sulfur application.

Response to Technical Comments, JA at 1264(emphasis added).

Finally, the petitioners argue that the EPA hasassumed a shorter useful life for scrubbers attached toother projects. However, the petitioners did not raisethis particular argument regarding the useful life ofthe scrubbers during the administrative review period.Under the CAA, we may review “[o]nly an objection toa rule or procedure which was raised with reasonablespecificity during the period for public comment.” 42U.S.C. § 7607(d)(7)(B); see, e.g., Am. Farm BureauFed’n v. EPA, 559 F.3d 512, 538 (D.C. Cir. 2009) (percuriam) (“In the comments submitted in response tothe EPA’s proposed revocation of the annual standard,the environmental petitioners argued only that therecord evidence demonstrated adverse effects fromlong-term coarse PM exposure; they did not raise theircurrent argument that an annual standard is necessaryto prevent adverse effects from short-term exposure.”).The EPA raised this jurisdictional bar in its brief, EPAResponse Br. at 43 n.9; the petitioners offered noresponse in their reply. Therefore, even if this

App. 42

argument had merit, we decline to consider it in thispetition for review.

C. Scrubbers Would Not Have a Significant Impacton Visibility

Finally, the petitioners assert that the emissionslimits proposed in the FIP would not have a significantimpact on visibility in the region. The petitionersargue: 1) that the EPA should have used thedollar-per-deciview method in evaluating the benefitsto visibility in installing the scrubbers; and 2) that theEPA impermissibly aggregated the visibilityimprovements the scrubbers would create acrossfacilities. Neither of these claims has merit.

Oklahoma first suggests EPA should not haverejected the visibility analysis it conducted in the SIP,which used the dollar-per-deciview method. Thisargument is misguided. The EPA rejected the SIPbecause of the flawed cost estimates. Whenpromulgating its own implementation plan, it did notneed to use the same metric as Oklahoma. Theguidelines merely permit the BART-determiningauthority to use dollar per deciview as an optionalmethod of evaluating cost effectiveness. See 40 C.F.R.pt. 51 app. Y(IV)(E)(1).13

13 We note, however, that in both its final rule and in its brief theEPA asserts that the guidelines require the use of thedollar-per-ton metric in evaluating cost effectiveness. Theguidelines themselves are a bit unclear. In the section on costeffectiveness, the guidelines mention only the dollar-per-tonmetric. 40 C.F.R. pt. 51 app. Y(IV)(D)(4)(c). However, theguidelines later state that in evaluating alternatives, “werecommend you develop a chart (or charts) displaying for each of

App. 43

And in the final rule, the EPA explained why it didnot use the dollar-per-deciview metric used byOklahoma. “Generally speaking, while the metric canbe useful if thoughtfully applied, we view the use of the$/deciview metric as suggesting a level of precision inthe calculation of visibility impacts that is not justifiedin many cases.” 76 Fed. Reg. at 81,747. The EPA hasnever mandated the use of this metric, and has notdeveloped “thresholds of acceptable costs per deciviewimprovement.” Id. While the federal land managershave developed threshholds, these threshholds wereapparently developed without input from the EPA andwithout notice-and-comment review. EPA Br. at 54n.13. In light of this, we do not find it arbitrary orcapricious that the EPA chose not to use thedollar-per-deciview metric in evaluating BART optionsin creating the FIP. We therefore also conclude thatany argument by the petitioners that the dollar-per-deciview measurement proves the scrubbers are notcost effective lacks merit. See Pet. Reply Br. at 16.

The petitioners next argue that the EPAimpermissibly aggregated the visibility improvementsacross facilities to make the scrubbers look moreeffective. The petitioners say that this is evidenced by

the alternatives” that includes, among other factors, the cost ofcompliance defined as “compliance—total annualized costs ($), costeffectiveness ($/ton), and incremental cost effectiveness ($/ton),and/or any other cost-effectiveness measures (such as$/deciview).” Id. app. Y(IV)(E)(1) (emphasis added).

We do not rule here on whether the EPA would be justified inrejecting a SIP because it relied on the dollar-per-deciview metric.We hold only that it was reasonable for the EPA to use thedollar-per-ton metric despite Oklahoma’s earlier analysis.

App. 44

the EPA’s reference to the total improvement invisibility (2.89 deciviews) that would result fromplacing scrubbers on all four of these units. Viewedseparately, the improvements from the addition ofscrubbers at each unit would not be worth the cost ofthe scrubbers, the petitioners say.

While we agree that the EPA referenced theaggregated visibility improvement across the Class Iareas, we do not agree that it used that number toguide its analysis. Instead, it evaluated theimprovements facility-by-facility. The Response to theTechnical Comments includes data on the visibilityimprovement at a number of different areas, brokendown by facility. See JA at 1495-98. Further, the EPAsaid its modeling “indicates that visibilityimprovements anticipated from the installation of dryscrubbers at each facility will result in reducingmodeled impacts . . . from each facility at all nearbyClass I areas to levels below 0.5 dv, with improvementsgreater than 1.0 dv at some Class I areas.” See 76 Fed.Reg. at 81,739 (emphasis added).

The petitioners also argue that the EPA erredbecause it considered the visibility improvementfacility-by-facility instead of unit-by-unit. Even if thisargument had merit, we have no jurisdiction toconsider it. The petitioners never raised this objectionduring the rulemaking process. See 42 U.S.C.§ 7607(d)(7)(B) (“Only an objection to a rule orprocedure which was raised with reasonable specificityduring the period for public comment (including anypublic hearing) may be raised during judicial review.”).In fact, the petitioners actually commented that theEPA needed to make the visibility determinations on a

App. 45

facility-by-facility basis. See JA at 1108 (“Whetherreviewing a SIP or proposing its own FIP, EPA, likeindividual states, is required to consider the visibilityimprovement associated with scrubbers on afacility-by-facility basis.”).

Likewise, the petitioners argue that the EPA didnot provide sufficient notice of the approach it used inits final rule. The petitioners say the EPA used a newmetric—days of visibility improvement—that it hadnever used before. Even if this argument has merit, wecannot consider it on appeal. Again, we may onlyaddress issues that were raised during the rulemakingprocess. As discussed above, it does not matter that thepetitioners could not have raised their objection beforethe promulgation of the final rule. “Rather, the CAArequires a petitioner to first raise its objection to theagency through a petition for reconsideration.”Appalachian Power Co., 249 F.3d at 1066. We thereforedecline to consider the petitioners’ argument here. See42 U.S.C. § 7607(d)(7)(B) (“If the Administrator refusesto convene [a reconsideration proceeding], such personmay seek review of such refusal in the United Statescourt of appeals for the appropriate circuit . . . .”(emphasis added)).

V

In addition to these arguments concerning theEPA’s substantive analysis, the petitioners raise anumber of challenges to the procedures the EPA usedin promulgating the rule. First, they argue that theEPA may not promulgate a FIP in the same action inwhich the agency disapproves a SIP. Second, thepetitioners argue that the EPA lost the authority topromulgate a FIP because the agency failed to act

App. 46

within two years after its duty to promulgate a FIP wasfirst triggered.

The CAA creates a high bar for any petitionerchallenging an EPA action on procedural grounds. Thepetitioner must prove: 1) that the failure to observe theprocedure was “arbitrary and capricious”; 2) that theobjection was “raised with reasonable specificity duringthe period for public comment”; and 3) that the errorswere “so serious and related to matters of such centralrelevance to the rule that there is a substantiallikelihood that the rule would have been significantlychanged if such errors had not been made.” 42 U.S.C.§ 7607(d)(9)(D), (d)(7), (d)(8). “The essential message ofso rigorous a standard is that Congress was concernedthat EPA’s rulemaking not be casually overturned forprocedural reasons, and we of course must respect thatjudgment.” Sierra Club v. Costle, 657 F.2d 298, 391(D.C. Cir. 1981).

The petitioners first argue that the EPA violatedrequired procedures by promulgating its FIP in thesame action in which it disapproved the SIP. Thepetitioners argue the statute requires the EPA firsttake action on the SIP because it says, according topetitioners, that the “EPA shall propose a FIP ‘unlessthe State corrects the deficiency,’ thereby reflectingCongress’s intention for States to have the power todesign their own SIP and have an opportunity tocorrect a SIP before a FIP is issued.” Pet. Opening Br.at 40 (quoting § 7410(c)). It also makes a policyargument that permitting the EPA to disapprove SIPsin the same action in which it promulgates a FIP will“blur [the] distinction” between the EPA’s role inreviewing SIPs and promulgating FIPs. Id. at 41.

App. 47

We do not agree that the EPA’s actions violated theprocedural requirements of the CAA. The petitioners’parsing of the statute relies on a truncated quotation.Under 42 U.S.C. § 7410(c)(1), the EPA must create aFIP after either the state has failed to make therequired SIP submission or the EPA has disapprovedpart of the state’s SIP. This duty continues to exist“unless the State corrects the deficiency, and theAdministrator approves the plan or plan revision,before the Administrator promulgates such Federalimplementation plan.” § 7410(c)(1) (emphasis added).Once the EPA issued findings that Oklahoma failed tosubmit the required SIP under the Regional Haze Rule,the EPA had an obligation to promulgate a FIP. Thestatute itself makes clear that the mere filing of a SIPby Oklahoma does not relieve the EPA of its duty. Andthe petitioners do not point to any language thatrequires the EPA to delay its promulgation of a FIPuntil it rules on a proposed SIP. As the EPA points out,such a rule would essentially nullify any time limitsthe EPA placed on states. States could forestall thepromulgation of a FIP by submitting one inadequateSIP after another.

In any case, even if we agreed that the EPA shouldnot have promulgated the FIP in the same action as itrejected the SIP, it is not clear the petitioners wouldmeet the high bar for overturning an EPA action onprocedural grounds. It may be poor policy to try todistinguish between the SIP and FIP in a single action.But the petitioners make no attempt to show theprocedural error was “so serious and related to mattersof such central relevance to the rule that there is asubstantial likelihood that the rule would have beensignificantly changed if such errors had not been

App. 48

made.” § 7607(d)(9)(D), (d)(8). Indeed, after the EPAraised this heightened standard of review of proceduralactions in its brief, the petitioners were silent as to anyissue regarding this procedural deficiency in their replybrief.

The petitioners also assert the EPA violated thestatute because § 7410(c)(1) says that the“Administrator shall promulgate a Federalimplementation plan at any time within 2 years after”the EPA’s duty to promulgate a FIP is triggered. Thepetitioners argue the EPA lost authority to promulgateits FIP because more than two years had passed sincethe EPA made its initial finding that Oklahoma failedto submit a SIP. Although the statute undoubtedlyrequires that the EPA promulgate a FIP within twoyears, it does not stand to reason that it loses its abilityto do so after this two-year period expires. Rather, theappropriate remedy when the EPA violates the statuteis an order compelling agency action.

The Supreme Court rejected an argument similar tothe petitioners’ argument in Brock v. Pierce County,476 U.S. 253 (1986). At issue in Brock was a lawrequiring that the Secretary of Labor “issue a finaldetermination as to the misuse of [ComprehensiveEmployment and Training Act] funds by a grantrecipient within 120 days after receiving a complaintalleging such misuse.” Brock, Id. at 254-55 (quotationomitted). The petitioners in that case argued that theSecretary lost his power to recover those funds if he didnot make a final determination within 120 days. Id. at255.

In the absence of a more clear statutory directive,the Court refused to accept this argument. Rather,

App. 49

when “there are less drastic remedies available forfailure to meet a statutory deadline”—such as a motionto compel agency action—“courts should not assumethat Congress intended the agency to lose its power toact.” Id. at 260. The Court “would be most reluctant toconclude that every failure of an agency to observe aprocedural requirement voids subsequent agencyaction, especially when important public rights are atstake.” Id.

The petitioners do not explain why the principles ofBrock would not also control this case. The provisionhere is “clearly intended to spur [the EPA] to action,not to limit the scope of [its] authority.” Id. at 265. Inthe absence of any other indication from Congress, theappropriate remedy is simply a suit to compel agencyaction, not to eliminate the EPA’s authority to file aFIP. See Mont. Sulphur & Chem. Co. v. EPA, 666 F.3d1174, 1190-91 (9th Cir. 2012).

VI

In conclusion, we hold that the EPA had theauthority to review Oklahoma’s BART determinations.Moreover, it exercised that authority properly.Accordingly, we DENY the petition for review of theEPA’s final rule. The stay pending hearing by themerits panel is hereby lifted.

KELLY, Circuit Judge, concurring in part anddissenting in part.

Although I agree with much of the court’s analysis,I dissent with respect to whether certain EPA actionswere arbitrary and capricious. See Ct. Op. Pt. IV(A)

App. 50

(analyzing the EPA’s calculations of baseline emissionsand its determination regarding the technicalfeasibility of the smaller scrubbers on which it basedits cost/benefit analysis). Therefore, I would grant thepetition for review.

As an initial matter, the court states that “[w]ereview these challenges [to the EPA’s FIP] under thesame arbitrary and capricious standard we used toevaluate the EPA’s rejection of the SIP.” Ct. Op. Pt. IV.The court notes, however, that “we do so whilerecognizing this requires a slightly differentperspective: evaluating the EPA’s own choices underthe guidelines, as opposed to evaluating its choice toreject the Oklahoma SIP under the guidelines.” Id.; seealso id. at Pt. III(B), n.7 (“We recognize that the EPAhas less discretion when it takes actions to reject a SIPthan it does when it promulgates a FIP.”). That mayusually be the case, but here the EPA rejectedOklahoma’s SIP and promulgated its own FIP in thesame rulemaking action. Many of the same reasons forrejecting the SIP were used to justify the FIP.Therefore, to the extent it makes a difference, I am notconvinced we owe any more deference to the EPA inevaluating these challenges than we would if thisaction were solely a rejection of a state SIP.

The EPA rejected Oklahoma’s cost estimates forscrubbers and provided two options of its own. Theseoptions arbitrarily and capriciously (1) assumed OG&Ewould burn coal they are not burning and have noplans to burn and (2) used scrubbers that do not fit andare not technically feasible.

For purposes of the cost benefit analysis, theControl Cost Manual (CCM) requires that anticipated

App. 51

annual emissions be calculated based upon past actualemissions. 40 C.F.R. pt. 51, App. Y(IV)(D)(4)(d). TheEPA ignored the historical emissions baseline inOption 1. OG&E has been using low-sulfur coal and itscontinued use would have resulted in scrubbersremoving about 14,000 tons of SO2 per year at eachaffected power plant. JA 1106. The EPA, however,assumed that OG&E would begin using high-sulfurcoal which would result in the removal of betweenroughly 43,000 to 46,000 tons of SO2 per year at eachaffected power plant. Id. at 1106, 1513–14. Petitionerscommented that any assumption by the EPA of achange in coal was improper and unsupportable. See id.at 1088, 1096–99 (OG&E EPA comments); see also id.at 300–02, 313 (OG&E state comments). Knowing thesecalculations violated the manual, the EPA developedOption 2.

Option 2 was based on OG&E’s continued use oflow-sulfur coal—the correct emissions baseline asrequired by the CCM. However, the EPA changed thesize of the scrubbers to smaller, less expensive ones. Id.at 1514–16. Petitioners argued extensively that thesesmaller scrubbers were technically infeasible. See, e.g.,id. at 1099–1102. In particular, the comments point outthe size of a scrubber is not dependent on the type ofcoal used but on gas flow and the maximum potentialheat input. Id. at 1101–02, 1144 (OG&E EPAcomments); see also id. at 384–430 (OG&E statecomments).

EPA admitted in response that the type of coalalone does not affect scrubber size and its estimate fora smaller scrubber was a result of oversimplification.See id. at 1283–84. Therefore, the only question is

App. 52

whether the EPA provided support for the technicalfeasibility of the smaller scrubbers it ultimatelyrequired. The EPA, however, relies almost exclusivelyon a cost model by OG&E’s consultant Sargent &Lundy in a different action and its statements that theamount of SO2 removed can in fact affect scrubber sizeat least somewhat. See id. at 1283, 1348. This reportdoes not describe why it matters or how it would affectthe size of the scrubbers in this case.

The EPA rejected Oklahoma’s evidentiary supportwith no clear evidence of its own to support its contraryconclusion. Usually the court grants deference to theEPA’s technical determinations. See San Juan CitizensAlliance v. Stiles, 654 F.3d 1038, 1045 (10th Cir. 2011).The EPA deserves no such deference, however, whereit does not support a conclusion contradictingOklahoma’s first, reasonable, detailed technicalconclusion. See JA 384–430; Lockheed Martin Corp. v.Admin. Review Bd., U.S. Dep’t of Labor, No. 11-9524,2013 WL 2398691, at *3 (10th Cir. June 4, 2013).Therefore I would conclude that the failure of theagency to provide any evidence that a significantlysmaller scrubber was sufficient to meet OG&E’s needsis arbitrary and capricious.

Finally, it is worth noting that the EPA’s regionalhaze program is distinct in the amount of power givento the states. See, e.g., Train v. Nat. Res. Def. Council,Inc., 421 U.S. 60, 79 (1975); 70 Fed. Reg. 39,104, 39,137(July 6, 2005) (“[H]ow states make BARTdeterminations or how they determine which sourcesare subject to BART” are among the issues “where theAct and legislative history indicate that Congressevinced a special concern with insuring that States

App. 53

would be the decision makers.”). There are a number ofreasons for this approach, not the least of which is thatits goals and standards are purely aesthetic ratherthan directly related to health and safety. The EPA’srule here requires OG&E to make a $1.2 billion dollarinvestment over the next five years that will, evenunder EPA’s estimate, result in no appreciable changein visibility. Moreover, there is no evidence thisinvestment will have any effect whatsoever on airquality. It surely will, however, result in adversechanges to what Oklahoma ratepayers will pay forelectricity.

Although the EPA has at least some authority toreview BART determinations within a state’s SIP, ithas no authority to condition approval of a SIP basedsimply on a preference for a particular control measure.Texas v. EPA, 690 F.3d 670, 684 (5th Cir. 2012); seeEME Homer City Generation, L.P. v. EPA, 696 F.3d 7,29 (D.C. Cir. 2012) (reviewing a different rule andconcluding that the CAA “prohibits EPA from using theSIP process to force States to adopt specific controlmeasures”). Oklahoma considered the cost andresulting benefit of such a large investment inscrubbers, and its conclusion was not unreasonable.

Assuming the EPA has authority to reviewOklahoma’s SIP in the manner it did, its rejection ofthe SIP and promulgation of its own FIP was arbitraryand capricious. Accordingly, I respectfully dissent onthis issue.

App. 54

UNITED STATES COURT OF APPEALSFOR THE TENTH CIRCUIT

[Filed July 19, 2013]

No. 12-9526_______________________________________STATE OF OKLAHOMA; OKLAHOMA )INDUSTRIAL ENERGY CONSUMERS, )an unincorporated association, )

Petitioners, ))

v. ))

UNITED STATES ENVIRONMENTAL )PROTECTION AGENCY, )

Respondent. )--------------------------------------------------------- )SIERRA CLUB, )

Intervenor - Respondent, ))

and ))

PACIFICORP; AMERICAN COALITION )FOR CLEAN COAL ELECTRICITY; )NATIONAL PARKS )CONSERVATION ASSOCIATION, )

Amici Curiae. )_______________________________________)

No. 12-9527_______________________________________OKLAHOMA GAS & ELECTRIC ) COMPANY, )

Petitioner, ))

App. 55

v. ))

UNITED STATES ENVIRONMENTAL )PROTECTION AGENCY, )

Respondent. )--------------------------------------------------------- )SIERRA CLUB, )

Intervenor - Respondent, ))

and ))

PACIFICORP; AMERICAN COALITION )FOR CLEAN COAL ELECTRICITY; )NATIONAL PARKS CONSERVATION )ASSOCIATION, )

Amici Curiae. )_______________________________________)

JUDGMENT

Before BRISCOE, Chief Judge, KELLY andLUCERO, Circuit Judges.

This petition for review originated from the UnitedStates Environmental Protection Agency and wasargued by counsel.

It is the judgment of this Court that the action ofthe United States Environmental Protection Agency isaffirmed.

Entered for the Court

/s/ Elisabeth A. ShumakerELISABETH A. SHUMAKER, Clerk

App. 56

APPENDIX B

81728 Federal Register / Vol. 76, No. 249 /Wednesday, December 28, 2011 / Rules and

Regulations

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 52

[EPA–R06–OAR–2010–0190; FRL–9608–4]

Approval and Promulgation of ImplementationPlans; Oklahoma; Federal Implementation Planfor Interstate Transport of Pollution AffectingVisibility and Best Available Retrofit TechnologyDeterminations

AGENCY: Environmental Protection Agency (EPA).

ACTION: Final rule.

SUMMARY: EPA is partially approving and partiallydisapproving a revision to the Oklahoma StateImplementation Plan (SIP) submitted by the State ofOklahoma through the Oklahoma Department ofEnvironmental Quality on February 19, 2010, intendedto address the regional haze requirements of the CleanAir Act (CAA). In addition, EPA is partially approvingand partially disapproving a portion of a revision to theOklahoma SIP submitted by the State of Oklahoma onMay 10, 2007 and supplemented on December 10, 2007to address the requirements of CAA section110(a)(2)(D)(i)(II) as it applies to visibility for the 19978-hour ozone and 1997 fine particulate matter National

App. 57

Ambient Air Quality Standards. This CAA requirementis intended to prevent emissions from one state frominterfering with the visibility programs in anotherstate. EPA is approving certain core elements of theSIP including Oklahoma’s: determination of baselineand natural visibility conditions; coordinating regionalhaze and reasonably attributable visibility impairment;monitoring strategy and other implementationrequirements; coordination with states and FederalLand Managers; and a number of NOX, SO2, and PMBART determinations. EPA is finding that Oklahoma’sregional haze SIP did not address the sulfur dioxideBest Available Retrofit Technology requirements for sixunits in Oklahoma in accordance with the RegionalHaze requirements, or the requirement to preventinterference with other states’ visibility programs. EPAis promulgating a Federal Implementation Plan toaddress these deficiencies by requiring emissions to bereduced at these six units. This action is being takenunder section 110 and part C of the CAA.

DATES: This final rule is effective on: January 27,2012.

ADDRESSES: EPA has established a docket for thisaction under Docket ID No. EPA–R06–OAR–2010–0190. All documents in the docket arelisted in the Federal eRulemaking portal index athttp://www.regulations.gov and are available eitherelectronically at http://www.regulations.gov or in hardcopy at EPA Region 6, 1445 Ross Ave., Dallas, TX,75202–2733. To inspect the hard copy materials, pleaseschedule an appointment during normal business hourswith the contact listed in the FOR FURTHER

App. 58

INFORMATION CONTACT section. A reasonable feemay be charged for copies.

FOR FURTHER INFORMATION CONTACT: JoeKordzi, EPA Region 6, (214) 665–7186,kordzi.joe@epa.gov.

SUPPLEMENTARY INFORMATION:Throughout this document wherever “we,” “us,” “our,”or “the Agency” is used, we mean the EPA.

Overview

The CAA requires that states develop andimplement SIPs to reduce the pollution that causesvisibility impairment over a wide geographic area,known as Regional Haze (RH). CAA sections 110(a) and169A. Oklahoma submitted a RH plan to us onFebruary 19, 2010. On March 22, 2011, we proposed topartially approve and partially disapprove certainelements of Oklahoma’s SIP. 76 FR 16168. Today, weare taking final action by partially approving andpartially disapproving the elements of Oklahoma’s RHSIP addressed in our proposed rule. As discussed in theproposal for this rule, the CAA requires us topromulgate a Federal Implementation Plan (FIP) if astate fails to make a required SIP submittal or we findthat the state’s submittal is incomplete orunapprovable. CAA section 110(c)(1). Therefore, we arepromulgating a FIP to address the deficiencies inOklahoma’s RH plan.

One important element of the RH requirements ofthe CAA is that the Best Available Retrofit Technology(BART) must be selected and implemented for certainsources. The process of establishing BART emissionlimitations can be logically broken down into three

App. 59

steps. First, states identify those sources which meetthe definition of “BART-eligible source” set forth in 40CFR 51.301. Second, states determine whether suchsources “emit any air pollutant which may reasonablybe anticipated to cause or contribute to any impairmentof visibility in any such area” (a source which fits thisdescription is “subject to BART”). Third, for each sourcesubject to BART, states then identify the appropriatetype and the level of control for reducing emissions,” byconducting a five-step analysis: Step 1: Identify AllAvailable Retrofit Control Technologies, Step2: Eliminate Technically Infeasible Options, Step3: Evaluate Control Effectiveness of Remaining ControlTechnologies, Step 4: Evaluate Impacts and Documentthe Results, and Step 5: Evaluate Visibility Impacts.

We agree with Oklahoma’s identification of sourcesthat are BART eligible and subject to BART. Inaddition, we are approving a number of BARTdeterminations from Oklahoma’s RH SIP. We are notable to approve Oklahoma’s sulfur dioxide (SO2) BARTdeterminations for the OG&E’s Sooner Units 1 and 2,the OG&E Muskogee Units 4 and 5, and the AEP/PSONortheastern Units 3 and 4. In reviewing the SO2

BART determinations for these six units,1 we noted thestate’s cost estimates for SO2 scrubbers were high incomparison to other similar units, and we thereforeseparately assessed the costs of installation of controls

1 When we say “six BART sources,” or “six units,” we mean Units4 and 5 of the Oklahoma Gas and Electric Muskogee plant inMuskogee County; Units 1 and 2 of the Oklahoma Gas and ElectricSooner plant in Noble County; and Units 3 and 4 of the AmericanElectric Power/Public Service Company of Oklahoma Northeasternplant in Rogers County.

App. 60

for these units using well established costingmethodologies for BART determinations. As a result ofthis review, we proposed disapproval of the Oklahoma’s SO2 BART determinations for these six units becausethe Oklahoma’s costing methodology was not inaccordance with RH requirements. Consistent with thedisparity in cost estimations we identified in ourproposed disapproval, our revised cost estimateindicates that dry scrubber control technology is about½ to ¾ less expensive than was calculated byOklahoma. We have therefore determined it isappropriate to finalize our proposed disapproval of theOklahoma’s SO2 BART determinations for the sixunits, because we conclude that the flaws in the state’scost estimations were significant, and that the statetherefore lacked adequate record support and areasoned basis for its determinations regarding the costeffectiveness of controls as needed for the final steps ofthe BART analysis and as required by the RH Rule(RHR). We are also disapproving the state’s submittedLong Term Strategy because it relies on these BARTlimits which we are disapproving. We will of courseconsider, and would prefer, approving a SIP if the statesubmits a revised plan for these units that we canapprove.

We are approving the remaining sections of the RHSIP submission. This includes certain core elements ofthe SIP including Oklahoma’s (1) determination ofbaseline and natural visibility conditions,(2) coordinating regional haze and reasonablyattributable visibility impairment, (3) monitoringstrategy and other implementation requirements,(4) coordination with states and Federal Land

App. 61

Managers, and (5) the following BART determinationsfrom Oklahoma’s RH SIP:

• The SO2, nitrogen oxides (NOX), and particulatematter (PM) BART determinations for the OklahomaGas and Electric (OG&E) Seminole Units 1, 2, and 3.

• The NOX and PM BART determinations forOG&E’s Sooner Units 1 and 2.

• The NOX and PM BART determinations for theOG&E Muskogee Units 4 and 5.

• The SO2, NOX, and PM BART determinations forthe American Electric Power/Public Service Companyof Oklahoma (AEP/PSO) Comanche Units 1 and 2.

• The SO2, NOX, and PM BART determinations forthe AEP/PSO Northeastern Unit 2.

• The NOX and PM BART determination for theAEP/PSO Northeastern Units 3 and 4.

• The SO2, NOX, and PM BART determination forthe AEP/PSO Southwestern Unit 3.

In addition to the Regional Haze Requirements,CAA section 110(a)(2)(D)(i)(II) requires that theOklahoma SIP ensure that emissions from sourceswithin Oklahoma do not interfere with measuresrequired in the SIP of any other state under part C ofthe CAA to protect visibility. This requirement iscommonly referred to as the visibility prong of“interstate transport,” which is also called the “goodneighbor” provision of the CAA. Oklahoma submitteda SIP to meet the requirements of interstate transportfor the 1997 8-hour ozone National Ambient AirQuality Standards (NAAQS) and the fine particulate

App. 62

matter (PM2.5) NAAQS on May 10, 2007, andsupplemented it on December 10, 2007. In the May 10,2007, submittal, Oklahoma stated that it intended forits RH submittal to satisfy the requirements of thevisibility prong. We proposed to partially approve andpartially disapprove this submission as it relied uponthe Regional Haze SIP that we were proposing topartially approve and partially disapprove. Inevaluating whether Oklahoma’s SIP ensures thatemissions from sources within Oklahoma do notinterfere with the visibility programs of other states,we found that the regional modeling conducted by theCentral Regional Air Programs (CENRAP),participated in by Oklahoma, included reductions atthe six units that were not required by the OklahomaSIP. Since this modeling was used by other states andOklahoma in establishing their Reasonable ProgressGoals, we find that the Oklahoma SIP does not ensurethat emissions from sources within Oklahoma do notinterfere with measures required in the SIP of anyother state under Part C of the CAA to protectvisibility.

To address the deficiencies identified in ourdisapproval of these SO2 BART determinations and thedisapproval of the SIP submission as it pertains to thevisibility prong of interstate transport, we arefinalizing a FIP to control emissions from the six units.Our FIP requires that these six units reduce emissionsof SO2 to improve the scenic views at four nationalparks and wilderness areas: the Caney Creek andUpper Buffalo Wilderness Areas in Arkansas, theWichita Mountains National Wildlife Refuge inOklahoma, and the Hercules Glades Wilderness Areain Missouri. Improved air quality also results in public

App. 63

health benefits. This FIP can be replaced by a futurestate plan that meets the applicable CAArequirements.

All six units are coal-fired electricity generatingunits. Our FIP requires the six units to reduce their SO2 pollution to an emission rate of 0.06 pounds permillion BTU, calculated on the basis of a rolling 30boiler operating day average. This can be accomplishedby retrofitting the six units with dry flue gasdesulfurization technology, commonly known as “SO2

scrubbers.” In addition, any technology that can meetthis SO2 emission limit may be implemented at the sixsubject units. For example, EPA believes that theselimits can also be met by wet scrubbing technology orswitching to natural gas.

We held a 60 day public comment period on thisaction, and an open house and a public hearing in bothTulsa and Oklahoma City. Many public commentersdisagreed with aspects of our cost analysis for SO2

BART for the six affected units. After careful review ofinformation provided during the public commentperiod, we revised our calculation of the total projectcost for the four OG&E units from our proposed rangeof approximately $312,423,000 to $605,685,000, to ourfinal range of approximately $589,237,000 to$607,461,000. We made no changes to the cost basis forthe two AEP/PSO units from our proposal. As such, theassociated cost investment for AEP/PSO is$274,100,000. Even with these changes to our costanalysis we conclude that we cannot approve the SIP’s SO2 emission limits and instead must adopt theproposed emission limits for the six units. However, inconsideration of comments about the time needed to

App. 64

comply with our FIP, we have extended the time forcompliance with the SO2 emission limit from theproposed three years to five years.

This investment will reduce the visibility impactsdue to these facilities by over 60 to 80% at each one ofthe four national parks and wilderness areas in thearea, and promote local tourism by decreasing thenumber of days when pollution impairs scenic views.Although today’s action is taken to address visibilityimpairments, we believe it will also reduce publichealth impacts by decreasing SO2 pollution byapproximately 95%.

This action is being taken under section 110 andpart C of the CAA.

Table of Contents

I. Summary of Our Proposal

A. Regional Haze

B. Interstate Transport of Pollutants andVisibility Protection

II. Final Decision

A. Regional Haze

B. Interstate Transport of Pollutants andVisibility Protection

C. Compliance Timeframe

III. Analysis of Major Issues Raised by Commenters

A. Comments Generally Favoring Our Proposal

B. Comments Generally Against Our Proposal

App. 65

C. Comments on Legal Issues

1. General Legal Comments

2. Comments Asking EPA To Consider AllRules

3. Comments on Interstate Transport

D. Comments on Modeling

E. Summary of Responses to Comments on the SO2 BART Cost Calculation

1. Control Cost Manual Methodology

2. Revised Cost Calculations for the OG&EUnits

3. Cost Calculations for the AEP/PSO Units

4. Conclusion

F. Summary of Responses to VisibilityImprovement Analysis Comments

G. Summary of Responses to CommentsReceived on the SO2 BART Emission Limit

H. Summary of Responses to CommentsReceived on the SO2 BART ComplianceTimeframe

I. Comments Supporting Conversion to NaturalGas and/or Renewable Energy Sources

J. Comments Arguing Our Proposal WouldHurt the Economy and/or Raise ElectricityRates

App. 66

K. Comments Arguing Our Proposal WouldHelp the Economy

L. Comments on Health and EcosystemBenefits and Other Pollutants

M. Miscellaneous Comments

IV. Statutory and Executive Order Reviews

I. Summary of Our Proposal

On March 22, 2011, we published the proposal onwhich we are now taking final action. 76 FR 16168. Weproposed to partially approve and partially disapproveOklahoma’s RH SIP revision submitted on February19, 2010. We also proposed to partially approve andpartially disapprove a portion of a SIP revision wereceived from the State of Oklahoma on May 10, 2007,as supplemented on December 10, 2007, for thepurpose of addressing the “good neighbor” provisions ofthe CAA section 110(a)(2)(D)(i)(II) with respect tovisibility for the 1997 8-hour ozone NAAQS and thePM2.5 NAAQS.

A. Regional Haze

We proposed to approve Oklahoma’s determinationthat Units 4 and 5 of the OG&E Muskogee plant, Units1 and 2 of the OG&E Sooner plant, and Units 3 and 4of the AEP/PSO Northeastern plant are subject toBART under 40 CFR 51.308(e). However, we proposedto disapprove the SO2 BART determinations for Units4 and 5 of the OG&E Muskogee plant; Units 1 and 2 ofthe OG&E Sooner plant; and Units 3 and 4 of theAEP/PSO Northeastern plant because they do notcomply with our regulations under 40 CFR 51.308(e).

App. 67

We also proposed to disapprove the long term strategy(LTS) under section 51.308(d)(3) because Oklahomahas not shown that the strategy is adequate to achievethe reasonable progress goals set by Oklahoma and byother nearby states. The visibility modeling Oklahomaused to support its SIP revision submittal assumed SO2 reductions from the six sources identified abovethat Oklahoma did not secure when making its BARTdeterminations for these sources. The OklahomaDepartment of Environmental Quality (ODEQ)participated in the Central Regional Air PlanningAssociation (CENRAP) visibility modeling developmentthat assumed certain SO2 reductions from these sixBART sources. ODEQ also consulted with other stateswith the understanding that these reductions would besecured. We proposed a FIP to address these defects inBART and the LTS.

We proposed a FIP that included SO2 BARTemission limits on these sources. We proposed that SO2 BART for Units 4 and 5 of the OG&E Muskogeeplant, Units 1 and 2 of the OG&E Sooner plant, andUnits 3 and 4 of the AEP/PSO Northeastern plant is an SO2 emission limit of 0.06 lbs/MMBtu that appliesindividually to each of these units on a rolling 30 daycalendar average. Additionally, we proposedmonitoring, recordkeeping, and reporting requirementsto ensure compliance with these emission limitations.We proposed that compliance with the emission limitsbe within three years of the effective date of our finalrule. We solicited comments on alternative timeframes,of from two years up to five years from the effectivedate of our final rule. We also proposed that, shouldOG&E and/or AEP/PSO elect to reconfigure the aboveunits to burn natural gas as a means of satisfying their

App. 68

BART obligations under section 51.308(e), conversionshould be completed within the same time frame. Wesolicited comments as to, considering the engineeringand/or management challenges of such a fuel switch,whether the full five years allowed under section51.308(e)(1)(iv) following our final approval would beappropriate.

We proposed to disapprove section VI.E of theOklahoma RH SIP entitled, “Greater ReasonableProgress Alternative Determination.” We also proposedto disapprove the separate executed agreementsbetween ODEQ and OG&E, and ODEQ and AEP/PSOentitled “OG&E Regional Haze Agreement, Case No.10–024,” and “PSO Regional Haze Agreement, Case No.10–025,” housed within Appendix 6–5 of the RH SIP.We proposed that these portions of the submittal areseverable from the BART determinations and the LTS.These alternative determinations are not fundamentalrequirements of a RH program, so disapproval of themdoes not create a regulatory gap in the SIP. Therefore,no FIP is required.

We proposed no action on whether Oklahoma hassatisfied the reasonable progress requirements ofEPA’s regional haze SIP requirements found at section51.308(d)(1).

We also proposed to approve the remaining sectionsof the RH SIP submission.

B. Interstate Transport of Pollutants and VisibilityProtection

We proposed to partially approve and partiallydisapprove a portion of a SIP revision we received fromthe State of Oklahoma on May 10, 2007, as

App. 69

supplemented on December 10, 2007, for the purpose ofaddressing the “good neighbor” provisions of the CAAsection 110(a)(2)(D)(i) with respect to visibility for the1997 8-hour ozone NAAQS and the PM2.5 NAAQS.This proposal addressed the requirement of section110(a)(2)(D)(i)(II) that emissions from Oklahomasources do not interfere with measures required in theSIP of any other state under part C of the CAA toprotect visibility.

Having proposed to disapprove these provisions ofthe Oklahoma SIP, we proposed a FIP to address therequirements of section 110(a)(2)(D)(i)(II) with respectto visibility to ensure that emissions from sources inOklahoma do not interfere with the visibility programsof other states. We proposed to find that the controlsproposed under the proposed FIP, in combination withthe controls required by the portion of the OklahomaRH submittal that we proposed to approve, will serveto prevent sources in Oklahoma from emittingpollutants in amounts that will interfere with efforts toprotect visibility in other states.

II. Final Decision

A. Regional Haze

We are partially approving, partially disapproving,and taking no action on various portions of Oklahoma’sRH SIP revision submitted on February 19, 2010. Weare finalizing a FIP to address the defects in thoseportions of this SIP that are mandatory requirementsthat we are disapproving.

We are disapproving the SO2 BART determinationsfor Units 4 and 5 of the Oklahoma OG&E Muskogeeplant; Units 1 and 2 of the OG&E Sooner plant; and

App. 70

Units 3 and 4 of the AEP/PSO Northeastern plant. Weare disapproving the LTS under section 51.308(d)(3).

We are finalizing a FIP that specifically imposes SO2 BART emission limits on these sources. We findthat SO2 BART for Units 4 and 5 of the OG&EMuskogee plant, Units 1 and 2 of the OG&E Soonerplant, and Units 3 and 4 of the AEP/PSO Northeasternplant is an SO2 emission limit of 0.06 lbs/MMBtu thatapplies individually to each of these units. As wediscuss elsewhere in this action and in a supplementalresponse to comments document (Supplemental RTC),2

we find there is ample support for this decision.However, in response to a comment we received, we arechanging our proposed averaging period for theseemission limits from a straight rolling 30 day calendaraverage to one calculated on the basis of a boileroperating day (BOD). We also received a commentrequesting that we revise our proposed unit-by-unit SO2 limit, and replace it with a plant wide average SO2

limit. As we note in our response to this comment,although we are open to combining the BOD and plantwide averaging techniques, this presents a significant

2 The full title of the Supplemental RTC document is the “Responseto Technical Comments for Sections E through H of the FederalRegister Notice for the Oklahoma Regional Haze and VisibilityTransport FIP,” and it is available in the docket for thisrulemaking. This document is referred to as the “SupplementalRTC” throughout this rulemaking. We received many lengthy, andhighly technical, comments concerning our SO2 BART costanalysis, the visibility improvement analysis, the emission limit,and the compliance timeframe. While this notice generallyaddresses all of the issues commenters raised, the SupplementalRTC is intended to address comments on these four categories ingreater detail.

App. 71

technical challenge in having a verifiable, workable,and enforceable algorithm for calculating such anaverage. Due to our obligation to ensure theenforceability of the emission limits we are imposing inour FIP and the technical challenges of meeting thatobligation through a plant wide limit, we are notincluding a plant wide average SO2 limit in our finalFIP. We leave it to Oklahoma to take up this matter ina future SIP revision, should it decide to do so. We areconfident that this issue can be addressed prior to theinstallation of the emission controls required to satisfyour FIP.

We are promulgating monitoring, recordkeeping,and reporting requirements to ensure compliance withthese emission limitations.

We are disapproving section VI.E of the OklahomaRH SIP entitled, “Greater Reasonable ProgressAlternative Determination.” We are also disapprovingthe separate executed agreements between ODEQ andOG&E, and ODEQ and AEP/PSO entitled “OG&ERegional Haze Agreement, Case No. 10–024,” and “PSORegional Haze Agreement, Case No. 10–025,” housedwithin Appendix 6–5 of the RH SIP. We find that theseportions of the submittal are severable from the BARTdeterminations and the LTS. These alternativedeterminations are not fundamental requirements of aRH program, so disapproval of them does not create agap in the SIP. For these reasons, no FIP is required.

We are taking no action on whether Oklahoma hassatisfied the reasonable progress requirements ofEPA’s RH SIP requirements found at section51.308(d)(1).

App. 72

We are approving the remaining sections of the RHSIP submission. This includes certain core elements ofthe SIP including Oklahoma’s (1) determination ofbaseline and natural visibility conditions,(2) coordinating regional haze and reasonablyattributable visibility impairment, (3) monitoringstrategy and other implementation requirements,(4) coordination with states and Federal LandManagers, and (5) the following BART determinationsfrom Oklahoma’s RH SIP:

• The SO2, nitrogen oxides (NOX), and particulatematter (PM) BART determinations for the OklahomaGas and Electric (OG&E) Seminole Units 1, 2, and 3.

• The NOX and PM BART determinations forOG&E’s Sooner Units 1 and 2.

• The NOX and PM BART determinations for theOG&E Muskogee Units 4 and 5.

• The SO2, NOX, and PM BART determinations forthe American Electric Power/Public Service Companyof Oklahoma (AEP/PSO) Comanche Units 1 and 2.

• The SO2, NOX, and PM BART determinations forthe AEP/PSO Northeastern Unit 2.

• The NOX and PM BART determination for theAEP/PSO Northeastern Units 3 and 4.

• The SO2, NOX, and PM BART determination forthe AEP/PSO Southwestern Unit 3.

App. 73

B. Interstate Transport of Pollutants and VisibilityProtection

We are partially approving and partiallydisapproving a portion of a SIP revision we receivedfrom the State of Oklahoma on May 10, 2007, assupplemented on December 10, 2007, for the purpose ofaddressing the “good neighbor” provisions of the CAAsection 110(a)(2)(D)(i) with respect to visibility for the1997 8-hour ozone NAAQS and the PM2.5 NAAQS.

We are finalizing a FIP to address the requirementsof section 110(a)(2)(D)(i)(II) with respect to visibility toensure that emissions from sources in Oklahoma do notinterfere with the visibility programs of other states.We find that the controls under this FIP, incombination with the controls required by the portionof the Oklahoma RH submittal that we are approving,will serve to prevent sources in Oklahoma fromemitting pollutants in amounts that will interfere withefforts to protect visibility in other states.

C. Compliance Timeframe

In response to comments we received, we find thatcompliance with the emission limits of our FIP must bewithin five years of the effective date of this rule. Thiscompliance timeframe includes the election toreconfigure the six units to burn natural gas.

III. Analysis of Major Issues Raised byCommenters

We received both written comments and oralcomments at the Public Hearings in Oklahoma Cityand Tulsa. We also received comments by the Internetand the mail. The comments are summarized and

App. 74

discussed below. The full text received from thesecommenters is included in the docket associated withthis action.

A. Comments Generally Favoring Our Proposal

Comment: We received many letters in support ofour rulemaking from members representing variousorganizations that were similar in content and format,and are represented by two types of positive commentletters in the docket for this rulemaking. Each of thesecomment letters supports our proposed decision for thesix coal units identified above. More than 500 of theseletters specifically urge us to require emissionsreductions from these six units in our final decision.

We received two letters from Federal LandManagers in support of this rulemaking. Thesecomments include support for our proposed disapprovalof the Long Term Strategy under Section 51.308(d)(3)and our proposed disapproval of the GreaterReasonable Progress Alternative Determination(section 51.308), as well as support for our proposedFIP requiring an emissions limitation of 0.06 lb of SO2/MMBtu for each of the six units identified above.These comments also include agreement that EPA’sproposed controls are cost-effective, reasonable andattainable, and that they constitute BART. Theseletters also included support for requiring compliancewith the proposed emission limitations within threeyears from the effective date of the final rule, but couldaccept compliance within five years.

At the Public Hearing in Oklahoma City, positivecomments were received from representatives of anatural gas producer and from public citizens. Some

App. 75

comments included support for our proposeddisapproval of the Oklahoma SIP submittal, as well asfor finalizing our proposed FIP. Included with thesecomments was the belief expressed that not controllingthese sources will not make electricity cheap. Anotheridea presented at this hearing was that, whereas cheapelectricity does not make an economy healthy,renewable energy does. Data for eight states waspresented, including Washington State in which 75percent of the electricity comes from renewableresources. Other comments were that clean air is abasic necessity of life and not a luxury, and that cleanair is not something that should be traded or bargainedaway in the name of profit. Further, these commentsincluded encouragement for the shortest possibletimeline for compliance.

Comments were also received in support of ourproposal at the Public Hearing in Tulsa. Onecommenter noted that in the background for theproposed FIP, we accepted almost all of themethodologies and conclusions put forth by the ODEQ,with the exception of BART for SO2 removal. Anothercommenter mentioned that the concept of being a goodneighbor and reducing air pollution is a criticalcomponent of the CAA.

Response: We acknowledge these commenters fortheir support of this action. We also note that severalof the specific emissions and timeframe limitationssupported by these commenters in the proposal havebeen modified in this final action based on all of theinformation received during the comment period.Please see the docket associated with this action foradditional detail. Additionally, some of the specific

App. 76

issues that these commenters raised are addressedelsewhere in this notice.

B. Comments Generally Against Our Proposal

We received written comments, as well as oralcomments at the Public Hearings in Oklahoma Cityand Tulsa, that generally did not support our proposedrulemaking. Most of these commenters expressedconcerns about the economic impact of this rulemaking.Due to the specific nature of these comments, weaddress them more fully in the remainder of this noticeand in the Supplemental RTC. The full text of thesecomments is included in the docket associated with thisaction.

We also received one unspecific negative commentfrom an individual, which did not includedocumentation, rationale, or data for us to respond tobeyond our responses provided elsewhere in this notice.

C. Comments on Legal Issues

1. General Legal Comments

Comment: We received several comment lettersquestioning whether we have CAA authority todisapprove Oklahoma’s BART determination anddetermine BART through a FIP. These commentersincluded the Oklahoma Attorney General, OG&E,several industry trade organizations, and AEP/ PSO.We also received a comment letter signed by multipleattorneys general from throughout the United States.3

3 The signatories of this May 2011 comment letter were theattorney generals of Oklahoma, Alabama, Kentucky, Maine, theN. Mariana Islands, South Carolina, Texas, and Utah.

App. 77

The commenters generally contend that our proposalwould “usurp” or encroach on the state’s authority andthat EPA lacks the authority to substitute its ownjudgment or policy preferences for the state’sdeterminations. The Oklahoma Attorney Generalcomments that our role is “simply one of support” andthat state determinations are entitled to “specialdeference.” Similarly, one commenter states that wecannot “second-guess” the state and redo a BARTanalysis with no deference to the state’s findings. Thatcommenter also states that we have not articulated anystandard under which we may judge the validity of astate’s BART determination.

Response: Congress crafted the CAA to provide forstates to take the lead in developing implementationplans, but balanced that decision by requiring EPA toreview the plans to determine whether a SIP meets therequirements of the CAA. EPA’s review of SIPs is notlimited to a ministerial type of “rubber-stamping” of astate’s decisions. EPA must consider not only whetherthe state considered the appropriate factors but actedreasonably in doing so. In undertaking such a review,EPA does not “usurp” the state’s authority but ensuresthat such authority is reasonably exercised. EPA hasthe authority to issue a FIP either when EPA has madea finding that the state has failed to timely submit aSIP or where EPA has found a SIP deficient. Here,EPA has authority and we have chosen to approve asmuch of the Oklahoma SIP as possible and to adopt aFIP only to fill the remaining gap. Our action today isconsistent with the statute. In finalizing our proposeddeterminations, we are approving the state’sdeterminations in identifying BART eligible sourcesand largely approving the state’s BART determinations

App. 78

for thirteen different emission units subject to BART.We are, however, disapproving the state’s SO2 BARTdeterminations for six of those units. As explained inthe proposal, the state’s SO2 BART determinations forthe six OG&E and AEP/PSO units are not approvablebecause ODEQ “did not properly follow therequirements of section 51.308(e)(1)(ii)(A).” 76 FR16168, at 16182. Specifically, ODEQ did not properly“take into consideration the costs of compliance,” whenit relied on cost estimates that greatly overestimatedthe costs of controls. We have determined that thefaults in ODEQ’s cost methodology were significantenough that they resulted in BART determinations forSO2 that were both unreasoned and unjustified.Accordingly, those determinations that relied onsignificantly flawed cost estimations are notapprovable.

In the absence of approvable BART determinationsin the SIP for SO2 for BART eligible sources inOklahoma, we are obliged to promulgate a FIP tosatisfy the CAA requirements. Likewise, in the absenceof an approvable SIP that addresses the requirementthat emissions from Oklahoma sources do not interferewith measures required in the SIP of any other state toprotect visibility, we are obliged to promulgate a FIP toaddress the defect. This authority and responsibilityexists under CAA section 110(c)(1). We also arerequired by the terms of a consent decree withWildEarth Guardians, lodged with the U.S. DistrictCourt for the Northern District of California to ensurethat Oklahoma’s CAA requirements for110(a)(2)(D)(i)(II) are finalized by December 13, 2011.Because we have found the state’s SIP submissions donot adequately satisfy either requirement in full and

App. 79

because we have previously found that Oklahomafailed to timely submit these SIP submissions, we havenot only the authority but a duty to promulgate a FIPthat meets those requirements. Our action in large partapproves the RH SIP submitted by Oklahoma; thedisapproval of the SO2 BART determinations andimposition of the FIP is not intended to encroach onstate authority. This action is only intended to ensurethat CAA requirements are satisfied using ourauthority under the CAA. We note that Oklahoma maysubmit a new SIP revision addressing the issue of SO2

controls for these six units, in which case we will assessit against Clean Air Act and Regional Haze Rulerequirements as a possible replacement for the FIP.

Comment: Multiple commenters have cited tovarious CAA statutory provisions to support theircontention that the State of Oklahoma has authority or“primary authority,” where EPA has no authority orlesser authority. On this point, commenters have citedCAA Sections 169A(b)(2)(A) and 169A(g)(2).Specifically, Section 169A(b)(2)(A) reads in part thatregulations to protect visibility shall require theinstallation and operation of BART “as determined bythe State (or the Administrator in the case of a planpromulgated under section 7410(c) of the this title).”Section 169A(g)(2) begins, “in determining [BART] theState (or the Administrator in determining emissionslimitations which reflect such technology) shall” takeinto consideration several requisite statutory factors.The commenters place special emphasis on thereferences to the “the State” in these provisions andcontend that the plain language of the statute providesthat states, and not EPA, have authority to determineBART.

App. 80

Response: We agree that states have authority todetermine BART, but we disagree with commenters’assertions that EPA has no authority or lesserauthority to determine BART when promulgating aFIP. As the parenthetical in section 169A(b)(2)(A)indicates, the Administrator has the authority todetermine BART “in the case of a plan promulgatedunder section 7510(c).” In other words, theAdministrator has explicit authority to determineBART when promulgating a FIP. In our proposal, westated that we must consider the same factors as stateswhen proposing a FIP to address BART. 76 FR 16168,at 16187. Our BART determination follows the factorsprescribed by CAA Section 169A(g)(2). We disagreethat the language of the CAA limits our authority todetermine BART in the case of a FIP.

Comment: Commenters who have argued that theplain language of the CAA requires that states are theprimary or only BART determining authorities havealso cited our preamble language from past FederalRegister publications that they believe reinforces theircontention. For example, several commenters cited 70FR 39104, at 39107, which reads in part, “the Statemust determine the appropriate level of BART controlfor each source subject to BART.” Commenters havealso cited the preamble to our proposal, where wewrote, “States are free to determine the weight andsignificance to be assigned to each factor” when makingBART determinations. 76 FR 16168, at 16174. Finally,some commenters have stated the preamble of the RHRsupports their contentions when it states: “In somecases, the State may determine that a source hasalready installed sufficiently stringent emissioncontrols for compliance with other programs (e.g., the

App. 81

acid rain program) such that no additional controlswould be needed for compliance with the BARTrequirement.” 64 FR 35714, at 35740.

Response: We agree that states are assignedstatutory and regulatory authority to determine BARTand that many past EPA statements have confirmedstate authority in this regard. Although the states havethe freedom to determine the weight and significanceof the statutory factors, they have an overridingobligation to come to a reasoned determination. Asdetailed in our proposal and the supporting TechnicalSupport Document (TSD), the state’s SO2 BARTdeterminations for the six OG&E and AEP/PSO unitswere premised on flawed cost assumptions. Since theseSO2 BART determinations of the state are notapprovable, we are obliged to step into the shoes of thestate and arrive at our BART determinations.

Comment: Commenters have also cited other CAAprovisions. One commenter states that 169A(b) onlyallows for EPA to issue guidelines with technical andprocedural guidance for determining BART, not toissue rules that dictate the outcome (except for fossil-fueled power plants with capacity that exceeds 750MW). That commenter also contends that our lack ofauthority relative to the states is shown through CAASection 169A(f), which provides that the meeting of thenational visibility goal is not a “nondiscretionary duty”of the Administrator. AEP/PSO comments that theprovisions of CAA Section 169B shows that states havespecial authority to act together through visibilitytransport commissions. The Oklahoma AttorneyGeneral cites CAA Section 101(a)(3), which provides

App. 82

that air pollution control at its source “is the primaryresponsibility of States and local governments.”

Response: States shoulder significantresponsibilities in CAA implementation and ineffectuating the requirements of the RHR. EPA has theresponsibility of ensuring that state plans, includingRH SIPs, conform to CAA requirements. None of theCAA provisions cited by commenters change ourconclusion that we have authority to issue a FIP tosatisfy BART requirements given that Oklahoma’s RHSIP is not fully approvable. We cannot approve a RHSIP that fails to address BART with a reasonedconsideration of the costs of compliance. Our inabilityto approve the state’s BART determinations for SO2

means we must follow through on our non-discretionary duty to promulgate a FIP. Under theCAA, we were required to do this by January 2011, twoyears after EPA found that Oklahoma failed to submita RH SIP. 74 FR 2392. The language of CAA Section169A(f), which concerns the meeting of the nationalgoal, is not related to the review of a state’s BARTdeterminations or our determinations on theiradequacy or the timing of our action.

Comment: Many commenters expressed the viewthat their statutory arguments are reinforced bylegislative history of the 1977 CAA amendments.Several commenters refer to statements of SenatorEdmund Muskie regarding the conference agreementon the provisions for visibility protection in thoseamendments. Senator Muskie had stated that underthe conference agreement the state, “not theAdministrator,” identifies BART eligible sources anddetermines BART. 123 Cong. Rec. 26854 (August 4,

App. 83

1977). Commenters have also noted that Am. CornGrowers Ass’n v. EPA, 291 F.3d 1 (D.C. Cir. 2002) usedlegislative history, including the Conference Report onthe 1977 amendments, when the Court had invalidatedpast regulatory provisions regarding BART forconstraining state authority. The Court stated that theConference report confirmed that Congress “intendedthe states to decide which sources impair visibility andwhat BART controls apply to those sources.”

Response: We agree that the CAA places therequirements for determining BART for BART-eligiblesources on states. As discussed above, the CAA alsorequires the Administrator to determine BART in theabsence of an approvable determination from the state.Because we have determined that Oklahoma’s BARTdeterminations for SO2 for the six OG&E and AEP/PSOunits do not conform with section 51.308(e) and are notapprovable, we are authorized and at this timerequired to promulgate a FIP.

Comment: Several commenters have asserted ourproposal is inconsistent with the decision of the DCCircuit in Am. Corn Growers Ass’n v. EPA, 291 F.3d 1(D.C. Cir. 2002). They contend that language in thedecision affirms their views regarding state authorityand EPA’s lack of authority in regulating the problemof regional haze. In particular, the American CornGrowers decision had described states as playing “thelead role” in designing and implementing regional hazeprograms, Id. at 3, and described the CAA as “givingthe states broad authority over BART determinations.”Id. at 8.

Response: We disagree that our proposal isinconsistent with the American Corn Growers decision.

App. 84

We have determined that Oklahoma utilized flawedcost assessments and incorrectly estimated thevisibility impacts of controls. We have determinedthese issues resulted in non-approvable SO2 BARTdeterminations for the six OG&E and AEP/PSO units.We recognize the state’s broad authority over BARTdeterminations, and recognize the state’s authority toattribute weight and significance to the statutoryfactors in making BART determinations. As a separatematter, however, a state’s BART determination mustbe reasoned and based on an adequate record.Although we have largely approved the state’s RH SIP,we cannot agree that CAA requirements are satisfiedwith respect to these SO2 BART determinations.

Comment: One commenter contends that stateshave broader authority for regional haze, because it isnot a human health-based regulation. Anothercommenter similarly suggests that states are the“appropriate decision makers” because regional haze isabout haze, not health.

Response: We do not agree that the CAA or RHRprescribes a different degree of authority to statesbased on the program having the goal of improvingvisibility as opposed to preventing adverse humanhealth effects. Among other things, the CAA requiresstates to submit plans that satisfy NAAQS standardsset to protect both public health and welfare. Nothingin the terms of the CAA or its implementation historydirects that SIP submittals addressing visibility aresubject to a different standard of evaluation than SIPsubmittals that directly address public health issuesassociated with air pollutants. The distinction is notpertinent to state authority to develop RH SIPs and

App. 85

does not diminish our responsibility and authority torequire that they conform to the RHR.

Comment: Several commenters have more generallyasserted that we lack authority to disapprove the RHSIP, because of past cases where we have lackedauthority in particular SIP disapproval actions. Thesecommenters have cited, in particular to Florida Power& Light Co. v. Costle, 650 F.2d 579, 581 (5th Cir. 1981)(EPA must approve a SIP that “meets statutorycriteria”), Train v. NRDC, 421 U.S. 60, 79 (1975), andCommonwealth of Vir. v. EPA, 108 F.3d 1397 (D.C. Cir.1997). Under these cases, the commenters assert thatwe cannot question the wisdom of a state’s choices orrequire particular control measures if plan provisionssatisfy CAA standards.

Response: States are required by the CAA toaddress the BART requirements in their SIP. Ourdisapproval of the SO2 BART determinations in theOklahoma RH SIP is authorized under the CAAbecause the state’s SO2 BART determinations for thesix OG&E and AEP/PSO units do not satisfy thestatutory criteria. The state’s analysis of the costeffectiveness of controls was flawed due to reasonsdiscussed elsewhere in this notice. While states haveauthority to exercise different choices in determiningBART, the determinations must be reasonablysupported. Oklahoma’s errors in taking intoconsideration the costs of compliance were significantenough that we cannot conclude the state determinedBART according to CAA standards. The cases cited bythe commenters stress important limits on EPAauthority in reviewing SIP submissions, but ourdisapproval of these SO2 BART determinations for the

App. 86

six units has an appropriate basis in our CAAauthority.

Comment: A citizen commenter asserts that ourproposal is indicative of “raw unconstitutional power.”

Response: The commenter has cited no specificprovisions of the Constitution. In any case, we regardneither the RHR, which has previously been subject toreview by the D.C. Circuit, nor our underlyingstatutory authority for this action to beunconstitutional. We are acting under statutoryresponsibilities established in the 1977 and 1990amendments to the CAA. As is the case for anyexecutive agency under the authority of the President,the Constitution has charged us with theimplementation and enforcement of laws written byCongress. The administration of the CAA andimplementation of the RHR is accordingly notunconstitutional.

Comment: AEP/PSO and another commenter havecommented that our proposed action improperlycombines matters under Oklahoma’s RH SIP withunrelated matters addressed in the 2007 InterstateTransport SIP. Both commenters have stated that ourdisapproval of the Interstate Transport SIP would beinconsistent with our guidance in 2006. They contendour 2006 guidance had suggested conclusions regardingwhether emissions from any one state could interferewith measures of neighboring states to protect visibilitycould only be reached when a neighboring state’s RHSIP had been approved. These commenters believeOklahoma’s Interstate Transport SIP obligations underCAA Section 110(a)(2)(D)(i)(II) can be approvedbecause there were no EPA-approved regional haze

App. 87

SIPs at the time of submittal or when we reviewed theOklahoma submission.

Response: We disagree with contention of thecommenters that RH SIP requirements and thevisibility requirements of section 110(a)(2)(D)(i)(II) areunrelated. We are addressing them simultaneouslybecause the purposes and requirements of theinterstate transport provisions of the CAA with respectto visibility and the RH program are intertwined.Section 110(a)(2)(D)(i)(II) does not explicitly definewhat is required in SIPs to prevent the prohibitedimpact on visibility in other states. However, becausethe RH program requires measures that must beincluded in SIPs specifically to protect visibility, EPA’s2006 Guidance4 recommended that RH SIPsubmissions meeting the requirements of the visibilityprogram could satisfy the requirements of CAA section110(a)(2)(D)(i)(II) with respect to visibility.Subsequently, in instances in which some states didnot make the RH SIP submission, in whole or in part,or did not make an approvable RH SIP submission, weevaluated whether those states could comply withsection 110(a)(2)(D)(i)(II) by other means. Thus, wehave elsewhere determined that states may also beable to satisfy the requirements of CAA section110(a)(2)(D)(i)(II) with something less than anapproved RH SIP, see, for example, our determinations

4 See, “Guidance for State Implementation Plan (SIP) Submissionsto Meet Current Outstanding Obligations Under Section110(a)(2)(D)(i) for the 8-Hour Ozone and PM2.5 National AmbientAir Quality Standards,” from William T. Harnett, Director AirQuality Policy Division, OAQPS, to Regional Air Division Director,Regions I–X, dated August 15, 2006 (the “2006 Guidance”).

App. 88

regarding Colorado (76 FR 22036) and Idaho (76 FR36329). In other words, an approved RH SIP is not theonly possible means to satisfy the requirements of CAAsection 110(a)(2)(D)(i)(II) with respect to visibility;however, such a SIP could be sufficient. Given thisreasoning, we do not agree with commenters’contentions that our action improperly combines twounrelated programs.

Regarding our guidance on submissions in Augustof 2006, we explicitly stated that “at this point in time,”it was not possible to assess whether emissions fromsources in the state would interfere with measures inthe SIPs of other states. As subsequent events havedemonstrated, we were mistaken as to the assumptionthat all states would submit RH SIPs in December of2007, as required by the RHR, and mistaken as to theassumption that all such submissions would meetapplicable RH program requirements and therefore beapproved shortly thereafter. Thus the premise of the2006 Guidance that it would be appropriate to awaitsubmission and approval of such RH SIPs beforeevaluating SIPs for compliance with section110(a)(2)(D)(i)(II) was in error. Our 2006 Guidance wasclearly intended to make recommendations that wererelevant at that point in time, and subsequent eventshave rendered it inappropriate in this specific action.We must therefore act upon Oklahoma’s submission inlight of the actual facts, and in light of the statutoryrequirements of section 110(a)(2)(D)(i). In order toevaluate whether the state’s SIP currently in factcontains provisions sufficient to prevent the prohibitedimpacts on the required programs of other states, weare obligated to consider the current circumstances andinvestigate the level of controls at Oklahoma sources

App. 89

and whether those controls are or are not sufficient toprevent such impacts.

We reject the argument that Oklahoma’s submittalshould be approvable because surrounding states haveyet to submit RH SIPs that have been approved. Theargument fails to address what would happen if adownwind state were never to submit the required RHSIP, or were never to submit a RH SIP that wasapprovable. On its face, the commenter’s argument issimply inconsistent with the objectives of the statute toprotect visibility programs in other states if a statenever submits an approvable RH SIP. Second, thisapproach is flatly inconsistent with the timingrequirements of section 110(a)(1) which specifies thatSIP submissions to address section 110(a)(2)(D)(i),including the visibility prong of that section, must bemade within three years after the promulgation of anew or revised NAAQS. We acknowledge that therehave been delays with both RH SIP submissions bystates and our actions on those RH SIP submissions,but that fact does not support a reading of the statutethat overrides the timing requirements of the statute.At this point in time, states are required to havesubmitted regional haze plans to EPA that establishreasonable progress goals for Class I areas. Thisrequirement applies whether or not states have in factsubmitted such plans. We believe that there are meansavailable now to evaluate whether a state’s section110(a)(2)(d)(i)(II) SIP submission meets the substantiverequirement that it contain provisions to prohibitinterference with the visibility programs of otherstates, and therefore that further delay, until all RHSIPs are submitted and fully approved, is unwarranted

App. 90

and inconsistent with the key objective to protectvisibility.

As detailed in our proposal, we believe based on theinformation currently before us that animplementation plan that provides for emissionsreductions consistent with the assumptions used in themodeling of other CENRAP states will ensure thatemissions from Oklahoma sources do not interfere withthe measures designed to protect visibility in otherstates. 76 FR 16168, at 16193. The Oklahoma SO2

BART determinations for the six OG&E and AEP/PSOunits did not require these sources to meet the level ofcontrol assumed in the CENRAP modeling. As wediscuss elsewhere in our response to comments,Oklahoma engaged in a regional planning process. Thisregional planning process included a forum in whichstate representatives built emission inventories thatassumed that specific pollution sources would becontrolled to specific levels. This included assumptionsthat the six OG&E and AEP/PSO units would becontrolled to presumptive BART emission levels forSO2. Visibility modeling projections subsequentlyassumed those emission reductions, and other statesrelied on those reductions as part of their reasonableprogress demonstrations. Accordingly and consistentwith our proposal, we are partially disapproving theOklahoma SIP revision submitted to address therequirements of CAA section 110(a)(2)(D)(i)(II). TheFIP remedies the inadequacy in the Oklahoma SIP byrequiring controls for the six units that at least achievethe level of control assumed in the CENRAP modeling.

Comment: AEP/PSO and another commenter haveasserted that the promulgation of revised NAAQS for

App. 91

ozone and PM2.5 in 1997 did not trigger any additionalSIP obligations with respect to section110(a)(2)(D)(i)(II). A commenter believes that theserevised NAAQS are not meaningfully related tovisibility requirements in Title I Part C, of the CAA.The commenters ask EPA to determine that noobligation to address Part C visibility components of aSIP arose from those NAAQS revisions.

Response: Reduced visibility is an effect of airpollution, and the emissions of PM2.5 and ozone and itsprecursors can contribute to visibility impairment. SIPplanning for the control of these pollutants on thepromulgation of a new NAAQS will therefore implicatecontrol measures and issues relating to visibility. CAAsection 110(a)(1) therefore requires implementationplans submitted in the wake of a newly promulgatedNAAQS to address whether the state has adequateprovisions to prevent interference with the efforts ofother states to protect visibility. The obligation toaddress Part C visibility components expressly followsfrom the language of 110(a) concerning when plansmust be submitted and what each implementation planmust contain.

Comment: OG&E contends that EPA’s proposal todisapprove the state’s BART determination is faulty,because the agency relied “without critical review” onwhat the commenter describes as the “opinion” of acontracted consultant. The commenter contends EPA’sour consultant is unqualified to evaluate costs ofinstalling and operating scrubbers at the OG&E Units,because our consultant “has no experience designingscrubbers or estimating their costs.” Additionally,OG&E states our consultant lacked relevant knowledge

App. 92

about the OG&E Units and the facilities at which theseunits are located, and did not attempt to communicatewith OG&E or its contractor about the particulardesign parameters, engineering specifications, or otherintricacies associated with the OG&E units. Thecommenter believes the consultant’s report containsopinions that “lack adequate foundation.” On this basis,OG&E states that EPA cannot lawfully rely on theconsultant’s report.

Response: As an initial matter, we do not agree thatour regulatory actions are subject to evidentiary rulesregarding expert testimony, as this comment suggests.Our consultant’s detailed report was incorporated astechnical support for our regulatory determinationsand is not properly characterized as an opinion. Thecontention that we accepted the consultant’s reportwithout critical review is false. As was stated in ourproposal, only after we thoroughly reviewed andevaluated the report was it made a part of our TSD. 76FR 16168, at 16182–16183. Furthermore, we met withOG&E and its consultant concerning the developmentof our proposal and had extensive communicationsclarifying particular technical points. This informationwas coordinated with our consultant and wasincorporated into her report. Thus, we worked closelywith our consultant in the development of her report.

Comment: A commenter states that EPA’s proposedBART determination would violate Executive Order13132, Federalism.

Response: We do not agree that our proposal or thisfinal action violates Executive Order 13132. EPA istaking actions specified under the CAA in partiallyapproving and partially disapproving the Oklahoma

App. 93

RH SIP. The CAA also specifies the responsibility ofEPA to issue a FIP when states have not met theirrequirements under the CAA. EPA is promulgating thisFIP to fill the regulatory gap created by the partialdisapproval. Under the FIP, the state retains itsauthority to submit future RH SIPs consistent withCAA and RHR requirements; we do not discount thepossibility of a future, approvable RH SIP submissionthat results in the modification or withdrawal of theFIP. This rulemaking does not change the distributionof power between the states and EPA. Consistent withthis, in the Executive Orders section of thisrulemaking, we have determined that Executive Order13132 does not apply to this action.

Comment: A commenter states that EPA cannotpropose a FIP until after it has taken final action todisapprove a state implementation plan. Thecommenter cites to part of CAA section 110(c)(1) whichstates that the Administrator shall promulgate a FIP“at any time within 2 years after” the Administrator“disapproves a State implementation plan submission.”The commenter states that EPA should withdraw theproposed FIP, take final action only on the SIP, andonly then propose a FIP, if one is necessary.

Response: We have the authority to promulgate aFIP concurrently with a disapproval action. This timingfor FIP promulgation is authorized under CAA section110(c)(1). As has been noted in past FIP promulgationactions, the language of CAA section 110(c)(1), by itsterms, establishes a two-year period within which wemust promulgate the FIP, and provides no furtherconstraints on timing. See, e.g., 76 FR 25178, at 25202.Oklahoma failed to submit its regional haze SIP to us

App. 94

by December 2007, as required by Congress. Two yearslater, Oklahoma had still not submitted its regionalhaze SIP. When we made a finding in 2009 thatOklahoma had failed to submit its regional haze SIP,(see 74 FR 2392), that created an obligation for us topromulgate a FIP by January 2011. We are exercisingour discretion to promulgate the FIP concurrently withour disapproval action because of the applicablestatutory deadlines requiring us at this time topromulgate RH BART determinations to the extentOklahoma’s BART determinations are not approvable.

Comment: OG&E expresses the view that we haveimproperly combined a proposed disapproval of theOklahoma SIP with our own BART determination. Thecommenter contends that the fact we would reach adifferent BART determination is not “itself sufficientgrounds to disapprove the SIP.” The commenterbelieves EPA desired to have scrubbers installed on theOG&E units and is only proposing to substitute its ownBART determination “to mask the fact that it lacks anymeritorious grounds to disapprove ODEQ’s BARTdetermination.”

Response: Our grounds for disapproving ODEQ’s SO2 BART determination were articulated in ourproposal, and we have not claimed that having arrivedat a different SO2 BART determination constitutes abasis for disapproval. Instead, as was clear in ourproposal, we were obliged to develop an SO2 BARTdetermination because Oklahoma’s SO2 BARTdetermination was flawed and not approvable. The factthat Oklahoma’s SO2 BART determination was notapprovable caused us to develop a BART determination

App. 95

that adheres to the requirements of section51.308(e)(1)(ii)(A).

Comment: OG&E comments that we cannot justifyour disapproval based on aggregate visibilityimprovements. The commenter asserts that when wereview a SIP or propose a FIP, the agency is requiredto consider the visibility improvement associated withscrubbers on a facility-by-facility basis. The commenterpoints to a portion of our proposal where we stated thatmodeling demonstrates a “2.89 deciview improvementin visibility,” 76 FR 16168, at 16186, and notes thestatement is based on combining impacts fromscrubbers at multiple units. The commenter assertsthis approach violates the individual facility approachdictated by CAA as outlined in the American CornGrowers case and violates the RHR and the guidelinesthat responded to that case outcome. In particular, thecommenter cites to the preamble language at 70 FR39104, at 39106 which describes how the RHR wasamended “to require the States to consider the degreeof visibility improvement resulting from a source’sinstallation and operation of retrofit technology, alongwith the other statutory factors.” The commenterattributes significance to EPA’s phrasing, which hadstated in part, “* * * States will be required to considerall five factors, including visibility impacts, on anindividual source basis when making each individualsource BART determination.”

Another commenter also contends we based our SO2

BART proposal for the six OG&E and AEP/PSO unitson a visibility estimate of an 8.20 dv cumulativeimprovement over multiple Class I areas. Further, thiscommenter contends we have claimed this visibility

App. 96

improvement will result from emission reductions at allthree facilities combined, which the commentercharacterizes as a form of aggregation that isimpermissible, as BART must be determined on asource-by-source basis. The commenter also stated thatanalysis should be focused on the visibility impacts atthe most impacted area, not all areas. The commenterclaims our rules indicate that it is appropriate to modelimpacts at the nearest Class I area as well as impactsat other nearby Class I areas. However, in the case ofthe latter category of areas, merely for the purpose of“determin[ing] whether effects at those [other] areasmay be greater than at the nearest Class I area.” 70 FR39104, at 39170. Further, continues the commenter,the rules state that “[i]f the highest modeled effects areobserved at the nearest Class I area, you may choosenot to analyze the other Class I areas any further* * *.” Id. Based on this, the commenter states thatthat the BART rules contemplate a visibilityimprovement analysis that only is focused on visibilityimpacts in the most impacted area, not all areas.

Response: We proposed disapproval of theOklahoma SO2 BART determination for the six OG&Eand AEP/PSO units in part because we disagreed withODEQ’s cost analysis, and our own visibility modelingindicated SO2 controls would result in significantvisibility improvement. In so doing, we adhered to therequirements of section 51.308(e). Oklahoma’s SO2

BART determinations for the six units were based onflawed costing methodologies. Our determinationsregarding visibility improvement are not inconsistentwith the CAA or the court’s interpretation in AmericanCorn Growers of the individual facility approach thatmust be utilized when making BART determinations.

App. 97

Although we noted in the proposal the combinedvisibility improvement at four Class I areas due to theinstallation of SO2 controls at the six OG&E andAEP/PSO units, our FIP is not based on an analysis ofvisibility improvements that are aggregated acrossmultiple facilities. Rather, we assessed the visibilityimprovement of each facility separately.

Our visibility modeling shows that the six OG&Eand AEP/PSO units “causes or contributes” to visibilityimpairment—as the phrase is defined in the RHR5—atfour Class I areas. As Table 1 indicates, the number ofdays per year each Class I area is impacted at this levelby each facility’s emissions are expected to decreasedrastically at each Class I area as the result ofinstallation of SO2 BART emission controls at the sixunits. Clearly, the visibility benefits from SO2 BARTemission reductions will be spread among all affectedClass I areas, not only the most affected area, andshould be considered in evaluation of benefits fromproposed reductions. The portion of the BARTGuidelines (40 CFR 51 Appendix Y, IV.D.5) that thecommenter referenced states: “If the highest modeledeffects are observed at the nearest Class I area, youmay choose not to analyze the other Class I areas anyfurther as additional analyses might be unwarranted.”This section of the BART Guidelines addresses how todetermine visibility impacts as part of the BARTdetermination and is intended to make clear that ifcertain controls would be justified based on the impactsat the nearest Class I area, the state is not required to

5 States should consider a 1.0 deciview change or more from anindividual source to “cause” visibility impairment, and a change of0.5 deciviews to “contribute” to impairment. 70 FR 39120.

App. 98

undertake an exhaustive analysis of impacts acrossmultiple Class I areas. Several paragraphs later in theBART Guidelines is the following: “You have flexibilityto assess visibility improvements due to BART controlsby one or more methods. You may consider thefrequency, magnitude, and duration components ofimpairment,” emphasizing the flexibility in method andmetrics that exists in assessing the net visibilityimprovement.

Comment: OG&E comments that we had improperlyanalyzed the “contingent BART determination thatapplies if EPA rejects ODEQ’s determination that lowsulfur coal is BART and all appeals are exhausted.”The commenter says the contingent BARTdetermination should not have been analyzed as aBART alternative under 40 CFR 51.308, because it is“not a BART alternative.” If the contingentdetermination were to be effectuated, the commenterasserts that scrubbers would then constitute BARTitself, not an alternative to BART scrutinized underseparate rules. The commenter also asserts that thecontingent BART finding would be consistent with thestatutory timeframe for installation of BART (viz., “inno event later than five years” under CAA section169A(g)(4)), because the contingent BART findingwould not be triggered until the appellate process hadconcluded and because a final appellate ruling might bemade before 2013, which could result in a time forcompliance that is shorter than five years.

Response: The RHR does not afford the option ofsubmitting contingent BART determinations thatwould apply and become effective when EPAdisapproves and successfully defends its disapproval of

App. 99

a state’s BART determination. This item in the RH SIPcould not be evaluated as a BART determination,because it is not on its face a BART finding. Thiscomponent of the RH SIP submission inherentlyspeculates on the actions and outcomes of review byEPA and the courts, and is contrary to the SIPplanning and review expected under the RHR and theCAA, more generally. Accordingly, we properlyevaluated these provisions as an alternative to BARTand determined that the contingent BARTdetermination was not approvable under 40 CFR51.308. We disagree that it could be reviewed underany other provision and found to be consistent with theRHR.

Comment: OG&E comments that we had improperlyanalyzed the “2026 compliance option” as failing tomeet the standards of a BART alternative. In thecommenter’s view, the 2026 compliance is not a BARTalternative but only a measure “to implement a long-term strategy in the name of reasonable progress.”OG&E asserts that ODEQ has authority for this under51.308(d)(3), and that implementation of thecompliance option could reduce emissions more thanwould be possible with dry scrubbers, and that ourevaluation of the 2026 compliance option loses sight ofthe long-term national goal.

Response: We disagree that the contingent SIPprovision can be recognized as implementing a long-term strategy. As discussed in our response regardingthe “contingent BART determination,” this componentof the RH SIP is not on its face reviewable as a BARTdetermination and fails to satisfy the requirements ofSection 51.308. The contingent SIP is predicated on

App. 100

speculative actions and outcomes of review by EPA andcourts, and does not comport with established SIPplanning and approval processes under the CAA.

Comment: A commenter expressed concern thatEPA has ignored the regional haze plan supported byODEQ and local utilities, and states, “EPA hasassumed the State’s role under the Clean Air Act andhas simply chosen not to exercise its discretion toapprove the Greater Reasonable Progress AlternativeDetermination.” Another commenter also submitted acomment requesting that EPA use the Oklahoma RHSIP as a guideline in the decision making process.Another commenter from the office of Oklahoma’sAttorney General states that we “should defer to thestate plan,” because Oklahoma is in a superior positionto make decisions regarding energy policy.

Response: We note that our action today largelyapproves the regional haze plan submitted byOklahoma. We are, however, finalizing disapprovals ofthe state’s SO2 BART determinations and the “GreaterReasonable Progress Alternative Determination”referenced by the commenter. We have determined thatneither of these components of the RH SIP submissionconforms to CAA and RHR requirements. BecauseOklahoma’s SO2 BART determinations are not beingapproved, we have promulgated a FIP that determinesSO2 BART for the six OG&E and AEP/PSO units in amanner consistent with RHR requirements. We agreethat this action, as with any FIP, may be said toassume a planning role ordinarily belonging to thestate. Even with the finalization of the FIP, the statenevertheless retains its authority to submit future RHSIPs consistent with CAA and RHR requirements; we

App. 101

do not discount the possibility of a future, approvableRH SIP submission that results in the modification orwithdrawal of the FIP. In the meantime, sources mustcomply with the requirements of the FIP and theapproved components of Oklahoma’s RH SIP.

2. Comments Asking EPA To Consider All Rules

Comment: OG&E comments that installation ofscrubbers will consume a significant amount ofadditional power that would need to be generated byburning additional fuel. The commenter suggests thatincreased GHG emissions from the additional fuelcombustion could trigger the requirement to obtain aprevention of significant deterioration (PSD) permit forgreenhouse gas emissions (GHGs). The commenterasserts that a PSD permit application process “can take18–24 months” and, if the process is necessary, it mightbe impossible to accommodate any PSD permitapplication process in a three-year compliance period.The commenter further contends the permittingprocess will impose costs and the terms of the PSDpermit might impose costs if changes to the method ofoperation or additional control technologies arerequired. The commenter says we failed to account forthese costs in our cost evaluation.

Response: We agree that the installation of SO2 dryscrubbers at the six OG&E and AEP/PSO units couldconceivably increase the emissions of other regulatednew source review pollutants, including GHGs, to thepoint where PSD review is triggered. Any PSD permitthat is necessary would have to be obtained fromODEQ, which is the permitting authority in Oklahoma.Whether or not PSD permitting is required would bebased on design-specific considerations and

App. 102

applicability determinations that will vary with eachunit. OG&E has not provided underlying data or factsto substantiate first, that PSD permitting could not beavoided through controls designed to consume lesspower, and second that a PSD permit, if needed, wouldimpose additional or collateral costs that wouldmaterially change our cost evaluation. We also disagreewith the assertion that PSD permitting will require18–24 months; Oklahoma’s SIP for PSD permitting,consistent with CAA section 165(c), establishes a oneyear objective for granting or denying PSD permitapplications. As we discuss elsewhere in this notice andin our Supplemental RTC, we find that compliancewith SO2 BART for the six units is extended to fiveyears, which should provide ample opportunity tosatisfy PSD permitting requirements, if any.

Comment: A commenter states that the proposedthree-year compliance period is not justified. Thecommenter contends that we should consider otherregulations that we are formulating for the powersector that will affect the six units covered by the FIP.The commenter mentions the Clean Air TransportRule, the proposed Air Toxics rule, the projected NSPS,and rules for GHGs, coal combustion waste, andimplementation of 316(b) of the Clean Water Act. Thecommenter states the compliance period is inadequatebecause utilities would not have sufficient time todevelop a plan that addresses all of the regulations weare considering, including BART, because those rulesmay affect how they choose to comply with any givenBART limitations. The commenter also thinks weshould be required to analyze whether the compliancetimeframe is appropriate by examining whether theother regulations will cause delays because of

App. 103

simultaneous demands for materials, equipment,supplies, and labor.

In related comments, OG&E and anothercommenter state that other regulatory developmentsthat impact coal burning power plants in the periodsince Oklahoma submitted its SIP should be consideredin our BART analysis, including the utility MACTproposal, the cooling water intake proposal, and thecoal ash disposal proposal. OG&E further citesadditional possible regulations through revision of theNAAQS, and the clean air transport proposal. OG&Estates the control requirements and costs of these otherrules should be considered in establishing theremaining useful life of the OG&E units for the BARTanalysis. OG&E is concerned that depending on theoutcome of these rulemaking processes, some or all ofthe units in question may not continue to beeconomically viable. The Governor of Oklahoma alsosubmitted a comment requesting EPA to consider theimpact that subsequent rulemakings may have on theissue of regional haze.

Response: We agree that multiple regulatory actionsare pending that will affect the power sector and agreethat regulatory development should be coordinatedwhen possible. We also recognize the importance oflong-term and coordinated planning on the part ofowners of industrial sources that are subject to BART.The visibility requirements of the CAA were put inplace in 1977 and 1990, and our implementingregulations adopted in 1999, and the regional hazerequirement for installation and operation of BART, inparticular, must be carried out expeditiously. We haveno basis and no supporting evidence from the

App. 104

commenter or any other source to conclude thatsignificant market constraints for materials,equipment, supplies and labor would arise to make athree-year compliance period unachievable, but we dorecognize the importance of planning within anycompliance period. As we discuss elsewhere in thisnotice and in the Supplemental RTC, we have extendedthe compliance timeframe from the three years weproposed. Compliance with the SO2 BART emissionlimits in our FIP must be within five years of theeffective date of our final rule, which is the maximumtime permitted by statute.

With regard to the BART analysis, the BARTguidelines do allow for consideration of the remaininguseful life of facilities when considering the costs ofpotential BART controls. Such a claim would have to besecured by an enforceable requirement. Neither OG&Enor AEP/PSO claimed any such restrictions on theoperation of these six units. Consequently, we assumeda remaining useful life of 30 years in our BARTanalysis. If OG&E and/or AEP/ PSO decide the units inquestion have a shorter useful life such that installingscrubbers is no longer cost effective, and are willing toaccept an enforceable requirement to that effect, arevised BART analysis could be submitted by theplant(s) in question and our FIP could be re-analyzedaccordingly. Similarly, we could also review a revisedSIP submitted by ODEQ.

The RHR follows from statutory requirements of theCAA that are separate and independent from theregulatory requirements mandated by othercomponents of the CAA and by other federal statutoryschemes cited by the commenters. Even assuming the

App. 105

cited regulations were finalized and costs of theseregulations were non-speculative, they have no bearingon the cost effectiveness analysis used to determineBART. Whether or not SO2 BART is cost effective inconjunction with possibly unrelated environmentalcontrols that may be separately required by otherstatutes such as the Clean Water Act is not part of thestatutory formulation that Congress prescribed toaddress regional haze.

3. Comments on Interstate Transport

Comment: We received two comments emphasizingthat regional haze is a problem that is not alwayscontained by state boundaries. One of the commentersstates that a “regional approach is critical” and notesthat CAA Section 169B(c)(1) authorizes theestablishment of visibility transport regions. Thecommenter states that visibility issues for the WichitaMountains Wilderness Area (WMWA) make it a“candidate for consideration of the establishment of atransport region.” The commenter believes that aregional examination or study of all the issues willallow development of the long range strategies and leadto cost-effective management of all pollution sourcesthat impair visibility in the region’s Class I areas.

Response: We agree that pollutants from one ormore states can significantly contribute to visibilityimpairment in the Class I areas of different states.CAA section 110(a)(2)(D)(i)(II) explicitly provides thatstates must have SIPs with adequate provisions toprevent interference with the efforts of other states toprotect visibility. Our FIP action ensures that sourcesin Oklahoma meet the RH requirements for BART andthe visibility requirements of section 110(a)(2)(D)(i)(II).

App. 106

We also agree that a regional approach to addressingvisibility transport is important, which is why EPAfunded Regional Planning Organizations (RPOs), suchas the Central Regional Air Planning Organization(CENRAP), in which Oklahoma participated. Statessuch as Oklahoma engaged in the RPO process foryears in order to co-develop strategies for mitigatingregional haze. At this time, we do not believe thatdelaying or setting aside these strategies in order tofurther study regional haze through the formation of atransport region is appropriate. However, we note theAdministrator has statutory discretion to establish atransport region in the future and may do so on theAdministrator’s own motion or on consideration of a“petition from the Governors of at least two affectedStates.” CAA Section 169B(c)(1).

D. Comments on Modeling

Comment: AEP/PSO stated that visibilityimprovements expected by installing controls under ourFIP are nearly identical to the improvements from theactions included in the ODEQ SIP submission, and thatthe FIP controls will not provide a noticeableimprovement in visibility. The commenter concludesthat the actions included in the ODEQ SIP submissionare just as effective in reducing visibility impairmentas the FIP. We received additional comments thatinstallation of controls proposed in the FIP wouldresult in imperceptible or nearly imperceptibleimprovements in visibility. Information is provided inthe comments that claims to support the statementthat there is “virtually no distinguishable” differencebetween the controlled and uncontrolled cases.

App. 107

Response: We performed visibility modeling as partof the SO2 BART determination analysis. A change ofapproximately one deciview (dv) is generally regardedas a perceptible change in visibility. 70 FR 39104, at39118. “For purposes of determining which sources aresubject to BART, states should consider a 1.0 deciviewchange or more from an individual source to ‘cause’visibility impairment, and a change of 0.5 deciviews to‘contribute’ to impairment.”6 70 FR 39104, at 39120.Our modeling indicates that visibility improvementsanticipated from the installation of dry scrubbers ateach facility will result in reducing modeled impacts(maximum of 98th percentile daily maximum dv) fromeach facility at all nearby Class I areas to levels below0.5 dv, with improvements greater than 1.0 dv at someClass I areas. We also evaluated the amount ofimprovement in the number of days that each facilitywould either cause or contribute to visibilityimpairment. As detailed in Table 1 below, thereductions resulting from our FIP would almostcompletely eliminate days when any of the threefacilities’ BART units have a perceptible impact(greater than 1.0 dv). These reductions would alsosignificantly decrease the number of days that have a0.5 deciview impact (or greater).

6 “If ‘causing’ visibility impairment means causing a humanlyperceptible change in visibility in virtually all situations (i.e. a 1.0deciview change), then ‘contributing’ to visibility impairment mustmean having some lesser impact on the conditions affectingvisibility that need not rise to the level of human perception.” 70FR 39104, at 39120.

App. 108

TA

BL

E 1

—A

VE

RA

GE

NU

MB

ER

OF

DA

YS

PE

R Y

EA

R E

AC

H F

AC

ILIT

Y’S

VIS

IBIL

ITY

IM

PA

CT

SE

XC

EE

D 1

.0 A

ND

0.5

DE

CIV

IEW

S

Cla

ss I

area

Dis

tan

ceto

un

it(k

m)

Ave

rage

# o

f da

ys/y

r >

1.0

dvA

vera

ge #

of

days

/yr

> 0.

5 dv

Bas

elin

eL

NB

LN

B &

DF

GD

Bas

elin

eL

NB

LN

B &

DF

GD

Soo

ner

Un

its

1 &

2

Can

eyC

reek

345

31

014

50

Her

cule

s-G

lade

s36

32

00

93

0

Upp

erB

uff

alo

327

21

011

50

Wic

hit

aM

oun

tain

s23

418

101

3825

3

App. 109

TO

TA

LA

vera

ge #

of day

s/ye

ar

2512

172

383

Mu

skog

ee U

nit

s 4

& 5

Can

eyC

reek

180

177

046

283

Her

cule

s-G

lade

s23

07

50

2214

1

Upp

erB

uff

alo

164

158

034

252

Wic

hit

aM

oun

tain

s32

412

70

2620

2

App. 110

TO

TA

LA

vera

ge #

of day

s/ye

ar

5127

012

886

8

Nor

thea

ster

n U

nit

s 3

& 4

Can

eyC

reek

263

106

030

171

Her

cule

s-G

lade

s24

46

40

1711

0

Upp

erB

uff

alo

211

84

021

121

Wic

hit

aM

oun

tain

s32

311

70

2416

2

App. 111

TO

TA

LA

vera

ge #

of day

s/ye

ar

3521

093

554

App. 112

In addition, in a situation where the installation ofBART may not result in a perceptible improvement invisibility, the visibility benefit may still be significant,as explained by the preamble of the RHR: “Failing toconsider less-than-perceptible contributions to visibilityimpairment would ignore the CAA’s intent to haveBART requirements apply to sources that contribute to,as well as cause, such impairment.” 70 FR 39104, at39129. Given that sources are subject to BART basedon a contribution threshold of no greater than 0.5deciviews, it would be inconsistent to automaticallyrule out additional controls where the improvement invisibility may be less than 1.0 deciview or even 0.5deciviews. A perceptible visibility improvement is nota requirement of the BART determination becausevisibility improvements that are not perceptible maystill be determined to be significant. We considered thereduction in visibility impairment at WichitaMountains, Caney Creek, Upper Buffalo, and Hercules-Glades to be significant. Installation of dry scrubbersat each facility will result in significant visibilityimprovements, reducing the number of days withimpaired visibility due to each of these sources at allimpacted Class I areas (Table 1).

Comment: AEP/PSO stated that we should acceptthe visibility analysis results provided in ODEQ’s SIPfor determining BART for SO2 because the results ofboth our and ODEQ’s visibility modeling are notsignificantly different.

Response: We disagree that ODEQ’s modeling wassufficient for evaluating the visibility impacts to informour BART determination. Given that the emissionrates that we proposed as SO2 BART differed from

App. 113

those assumed in ODEQ’s BART visibility modeling, itwas necessary to perform our own CALPUFF visibilitymodeling. In doing so, we followed EPA/FLM guidanceand practices to assess the anticipated visibilityimprovements from the use of dry and wet scrubberswith emission rates of 0.06 and 0.04 lb of SO2/MMBtu,respectively. ODEQ, in contrast, used emission rates of0.10 and 0.08 lb of SO2/MMBtu for dry and wetscrubbers, respectively, in its modeling. As a result,ODEQ underestimated the visibility improvementsassociated with the use of dry and wet scrubbers.Furthermore, ODEQ’s BART visibility analyses reliedon pollutant-specific modeling to evaluate the visibilitybenefits from the use of available SO2 emissioncontrols. As discussed in the TSD that accompanied theproposed action and elsewhere in our response tocomments, due to the complexity of atmosphericchemistry and chemical transformation amongpollutants, we modeled all visibility impairingpollutants together to fully assess the visibilityimprovement anticipated from the use of controls. Asdetailed in the TSD, we also had updated emissionestimates for sulfuric acid emissions based on thelatest information, and corrected PM speciation thatwas included in our modeling. We therefore disagreewith the commenter and have explained why weneeded to do our own BART CALPUFF visibilityanalysis. We modeled the emission rates determined tobe achievable by the available and technologicallyfeasible controls in accordance with the appropriateprocedures, utilizing current practices and modelversions that were acceptable to us at the time theywere conducted in the latter half of 2010, and we areconfident in using our results as one of the five factorsin making a BART determination.

App. 114

Comment: A commenter stated that in our visibilityanalysis, we updated the PM speciation analysis forboth Sooner and Muskogee to use National ParkService (NPS) speciation profiles for dry bottom boilersrather than wet bottom boilers calculated in ODEQ’sSIP submission and used updated coal properties. Thecommenter concludes that the difference betweenODEQ’s PM speciation and EPA’s should not impactthe BART analysis because primary PM speciesemitted directly from the stack generally have littleoverall impact on visibility impairment, and PMspecific controls are not being considered for BART. Inaddition, the commenter states that we used differentestimates for sulfuric acid emissions used to representemissions of sulfate particles. The commenter statesthat this sulfate emission rate is not likely to be asignificant factor in the overall visibility impairmentand therefore the differences between ODEQ’smodeling and EPA’s modeling is not significant.Because the results are not significantly differentbetween EPA’s and ODEQ’s visibility modeling, thecommenter asserts that we have no basis for notaccepting the visibility modeling provided in the SIP.

Response: As discussed in the TSD, it was necessaryfor us to perform CALPUFF visibility modeling toassess the anticipated visibility improvements from theuse of dry and wet scrubbers at the achievable SO2

emission rates of 0.06 and 0.04 lbs/MMBtu,respectively. Because revised modeling was necessaryto support our proposed BART determination, weperformed modeling following EPA/FLM guidance andpractices, and corrected errors noted during our reviewof ODEQ’s modeling. Our modeling included revisedPM speciation to correct errors in PM speciation that

App. 115

was included in ODEQ’s modeling. As detailed in theTSD, ODEQ used incorrect coal properties andemission factors in calculating the PM speciation usedin their modeling. In addition, we estimated sulfuricacid emissions using the best current informationavailable from the Electric Power Research Institute(EPRI)7 and the correct coal properties. ODEQestimates of sulfuric acid emissions for Sooner andMuskogee failed to account for removal in the existingair heater or ESP. ODEQ’s estimates of sulfuric acidemissions from the Northeastern units were based onan assumption of 3 ppm sulfur content conversion inthe flue gas. Furthermore, sulfuric acid emissionestimates used in ODEQ’s PM pollutant-specificmodeling were based on the erroneous PM speciationdiscussed above.

We agree with the commenter that primary PM andsulfuric acid emissions from the sources modeled maynot significantly impact visibility. However, inperforming our own modeling analysis to support ourBART determination, we saw no reason to not makecorrections and estimate emissions based on acceptedmethodology using the best current information,correct emission factors and coal properties. Becauseemissions of PM and sulfuric acid vary between wetand dry scrubbers and do have some impact onvisibility conditions, we utilized the best estimates forthe emissions of these species to fully account for thedifference in visibility impacts between the base caseand the two control cases modeled.

7 “Estimating Total Sulfuric Acid Emissions from Stationary PowerPlants: Version 2010a. EPRI, Palo Alto, CA: 2010. 1020636.”

App. 116

Comment: AEP/PSO asserted that we incorrectlyrejected the ODEQ visibility improvement evaluationbecause ODEQ applied various controls usingpollutant-specific baseline and control model runs, asopposed to using all visibility impairing pollutants inthe calculation of the baseline and control model runs.The commenter states that our BART guidelines arenot specific as to how to evaluate visibilityimprovement for the application of BART controls. Thecommenter asserts that the pollutant specificCALPUFF modeling approach is a reasonable butsimplistic method to look at the improvement invisibility impairment attributable to NOX, SO2, or PMand is consistent with our guidance contained in aBART Q&A document that states that the controltechnology visibility analysis can be conducted forsingle units and individual pollutants.

Response: The referenced BART Q&A document8

states that it may be appropriate to conduct a unit byunit, pollutant by pollutant analysis, depending on thetypes of units and control measures underconsideration. As discussed in the TSD, due to thenonlinear nature and complexity of atmosphericchemistry and chemical transformation amongpollutants, all relevant pollutants should be modeledtogether to predict the total visibility impact at each

8 “Q&A’s for Source by Source BART rule,” dated July 6, 2005. Thisdocument is not available on EPA’s Web site and is a draftdocument reflecting the preliminary views of EPA staff on anumber of questions submitted by stakeholders.

App. 117

Class I area receptor.9 The referenced Q&A documentprovides clarification and guidance on performingvisibility analyses for BART. The emissions of NOX andSO2, should be modeled together to determine thevisibility impacts, and in evaluation of controls andcombinations of controls in determining BART for asource. As seen in our modeling results for wet and dryscrubbers included in our proposal and TSD, thechemical interaction between pollutants andbackground species can lead to situations where thereduction of emissions of a pollutant can actually leadto an increase in visibility impairment. Therefore, tofully assess the visibility benefit anticipated from theuse of controls, all pollutants should be modeledtogether. As discussed elsewhere in this response tocomments, it was necessary for us to performCALPUFF visibility modeling to assess the anticipatedvisibility improvements from the use of dry and wetscrubbers at the achievable SO2 emission rates of 0.06and 0.04 lb/MMBtu, respectively. Because revisedmodeling was necessary to support our proposed BARTdetermination, we performed modeling followingEPA/FLM guidance and practices, including modelingall visibility impairing pollutants together to fullyassess the total visibility benefit anticipated fromemission reductions.

Comment: AEP/PSO stated that when we calculatedvisibility improvement during our BART analysis, weused the monthly average humidity adjustment factors

9 “Regional Haze Regulations and Guidelines for Best AvailableRetrofit Technology (BART) Determinations,” from Joseph Paisie,Geographic Strategies Group, OAQPS, to Kay Prince, BranchChief, EPA Region 4, dated July 19, 2006.

App. 118

provided in Table A–2 of our 2003 Guidance documentfor the assessment of natural background visibility,whereas, ODEQ used Table A–3 in its visibilitycalculations. The commenter states that there is noguidance that requires the use of humidity factors fromTable A–2 as opposed to Table A–3. In addition, thecommenter states that the use of humidity factors fromTable A–2 instead of A–3 should not make a significantdifference in the overall visibility impairment and doesnot provide a basis for our rejection of the visibilitymodeling provided in the SIP submittal.

Response: EPA guidance for estimating naturalvisibility conditions under the RHR provides monthlysite-specific relative humidity factors for use incalculating visibility impairment.10 Table A–2 of theguidance contains the “recommended” values based onthe representative IMPROVE site location. Table A–3provides data based on the centroid of the area as“supplemental information.” Relative humidity factorsare used with the original IMPROVE equation tocalculate extinction from measured or predictedpollutant concentrations. The factors used by ODEQare not the recommended values and are given in theguidance document only as supplemental information.Furthermore, EPA guidance for tracking progressunder the RHR contains that same information alsolabeled Table A–2 and A–3 and is consistent with the

10 See, “Guidance for Estimating Natural Visibility ConditionsUnder the Regional Haze Rule,” EPA–454/B–03–005, September2003.

App. 119

above guidance material.11 This guidance states thatthe site specific values provided in Table A–2 for eachmandatory federal Class I area are recommended to beused for all visibility and tracking progress calculationsfor that Class I area. Table A–3 is supplemental dataprovided for informational purposes. We used therecommended values from Table A–2 of these guidancedocuments to calculate visibility using the originalIMPROVE equation.

As discussed elsewhere in this response tocomments, we find that our CALPUFF visibilitymodeling was necessary to assess the anticipatedvisibility improvements from the use of dry and wetscrubbers at the achievable emission rates that weredetermined during our analysis of the available controltechnology. We performed our CALPUFF visibilitymodeling following EPA/FLM guidance and practices.As detailed in the following response to comment, weused the revised IMPROVE equation to estimatevisibility impacts. The revised IMPROVE equationutilizes a separate set of relative humidity adjustmentfactors available from the Federal Land Managers’ AirQuality Related Values Work Group (FLAG) Phase IReport.12 We also evaluated modeling results using theoriginal IMPROVE equation to quantify the sensitivity

11 “Guidance for Tracking Progress Under the Regional Haze Rule,”EPA–454/B–03–004, September 2003.

12 “Federal Land Managers’ Air Quality Related Values WorkGroup (FLAG) Phase I Report—Revised (2010) Natural ResourceReport NPS/ NRPC/NRR—2010/232,” National Park Service, U.S.Department of the Interior, available at http://www.nature.nps.goV/air/Pubs/pdf/flag/ FLAG_2010.pdf.

App. 120

of our results to the choice in visibility impairmentalgorithm. In applying the original IMPROVE equationfor this sensitivity analysis, we utilized therecommended relative humidity factors provided in theguidance.

Comment: AEP/PSO stated that ODEQ used themost up-to-date version of the visibility model availableand utilized the original IMPROVE equation that wasapproved for use at the time the SIP was prepared. Thecommenter stated that when we performed ourmodeling we used the revised IMPROVE equation. Thecommenter states that the use of this differentequation is the largest variable causing the ODEQmodeling results to be different from our modelingresults. The commenter concludes that because ODEQused the most up-to-date version of the equation at thetime the SIP was prepared, the subsequent release ofnew methods should not be the basis for overriding theresults provided in the SIP.

Response: The original IMPROVE equation and therevised IMPROVE equation refer to two differentversions of algorithms used to estimate visibilityimpairment from pollutant concentrations. The revisedequation is a more recently available, refined version ofthe original equation and is now considered by EPAand FLM representatives to be the better approach toestimating visibility impairment. Compared to theoriginal IMPROVE equation, this revised IMPROVEequation has less bias, accounts for more pollutants,incorporates more recent data, and is based on

App. 121

considerations of relevance for the calculations neededfor assessing progress under the RHR.13

As discussed elsewhere in this response tocomments, it was necessary for us to performCALPUFF visibility modeling to assess the anticipatedvisibility improvements from the use of dry and wetscrubbers at the achievable SO2 emission rates of 0.06and 0.04 lb/MMBtu, respectively for Step 5 of theBART analysis. As part of our BART analysis, weperformed CALPUFF modeling to assess the impacts ofthe SO2 BART proposed controls on the sources at issueon visibility impairment. Because the revisedIMPROVE equation is the preferred method foranalyses being conducted at this time,14 we estimated

13 Revised IMPROVE algorithm for Estimating Light Extinctionfrom Particle Speciation Data, IMPROVE, January 2006(http://vista.cira. colostate.edu/improve/Publications/GrayLit/gray_literature.htm); Hand, J.L., Douglas, S.G., 2006, Review ofthe IMPROVE Equation for Estimating Ambient Light ExtinctionCoefficients—Final Report (http://vista.cira.colostate.edu/improve/Publications/GrayLit/016_IMPROVEEeqReview/IMPROVEeqReview.htm).

14 U.S. EPA. Additional Regional Haze Questions. U.S.Environmental Protections Agency. August 3, 2006, available athttp://www.wrapair.org/forums/iwg/documents/Q_and_A_for_Regional_Haze_8–03–06.pdf#search=%22%22 New%20IMPROVE%20equation%22%22; WRAP presentation, “Update on IMPROVELight Extinction Equation and Natural Conditions Estimates” TomMoore, May 23, 2006; U.S. Forest Service, National Park Service,and U.S. Fish and Wildlife Service. 2010. Federal land managers’air quality related values work group (FLAG): phase Ireport—revised (2010). Natural Resource ReportNPS/NRPC/NRR—2010/232. National Park Service, Denver,Colorado.

App. 122

the CALPUFF visibility impacts using this peerreviewed algorithm. We also evaluated modelingresults using the original IMPROVE equation toquantify the sensitivity of our results to the choice invisibility impairment algorithm. Visibility benefitsestimated using the original IMPROVE equation werelarger than those estimated with the revisedIMPROVE equation at all four Class I areas includedin the modeling. We note that, using either equation,visibility benefits were projected for the installation ofscrubbers and support the conclusion that dryscrubbers are the appropriate BART control for eachfacility.

Comment: AEP/PSO states that we incorrectlycompared baseline visibility impairment with visibilityimprovement for controlled cases. The commenterstates that both the Oklahoma SIP and the proposedFIP compared an inherently higher 24-hour average forthe baseline with an inherently lower 30-day averagefor the controlled case. The commenter states that thesame averaging period should be used so decisions arenot biased toward greater SO2 emission reductions. Thecommenter also states that our analysis is consistentwith many other BART analyses and determinationsprepared by EPA, states and industry, but inconsistentwith the proposed BART determination for the FourCorners Power Plant in New Mexico and BARTguidance from the State of Colorado.

Response: The approach that we have taken forestimating the visibility impacts of wet and dryscrubbing is appropriate based on the approach set outin the BART Guidelines. The BART guidelines statethat in estimating visibility impacts:

App. 123

Use the 24-hour average actual emission ratefrom the highest emitting day of themeteorological period modeled (for the pre-control scenario). Calculate the model results foreach receptor as the change in deciviewscompared against natural visibility conditions.Post-control emission rates are calculated as apercentage of pre-control emission rates. Forexample, if the 24-hr pre-control emission rate is100 lb/hr of SO2, then the post control rate is5 lb/hr if the control efficiency being evaluated is95 percent.

The BART guidelines also state:

The emissions estimates used in the modelsare intended to reflect steady-state operatingconditions during periods of high capacityutilization. We do not generally recommend thatemissions reflecting periods of start-up,shutdown, and malfunction be used, as suchemission rates could produce higher than normaleffects than would be typical of most facilities.

The BART guidelines provide a consistent approachto assess the visibility improvement due to theinstallation of controls allowing comparison betweenBART assessments. Setting the baseline using thehighest emitting day during the period being assessedprovides a consistent approach for sources to assesstheir baseline impacts and gives an assessment of themaximum impact the source will have on visibility.ODEQ, EPA and AEP agreed on how to model thebaseline emissions, including the baseline emissionrates, in a previous modeling protocol and subsequentmodeling reports. ODEQ’s RH SIP, and EPA’s proposed

App. 124

FIP incorporated this same baseline emission rateapproach that is consistent with previous agreementsand analyses that AEP had conducted.

In modeling the post-control emission rates, weconsidered the reasonably anticipated control efficiencyof the available control technology taking into accountthat the BART modeling should reflect steady-stateoperating conditions and should not generally reflectperiods of start-up, shutdown and malfunction. Asdiscussed previously in our TSD and elsewhere in thisnotice and the Supplemental RTC, control efficienciesreasonably achievable by dry scrubbing and wetscrubbing were determined to be 95% and 98%respectively. We also note that OG&E directed itsvendors to provide bids on a dry SO2 scrubber systemthat was designed to remove 95% of the SO2. The twoAEP sources were modeled with baseline SO2 emissionrates of 5230.8 and 5034.6 lb/hr for Units #3 and #4respectively. These rates for the two AEP sources weremodeled using the firing rate of each unit with baselineSO2 emission rates of 0.9 lb/MMBtu which, asdiscussed above, are the same rates, previouslyprovided by AEP and utilized by ODEQ in theOklahoma RH SIP for the baseline emission rates.Applying the expected 95% reduction in emission ratesfor a dry scrubber, in accordance with the examplegiven in the BART guidelines, would result in anemission rate of 0.045 lb/MMBtu. This value is lowerthan our proposed BART SO2 emission limit of0.06 lb/MMBtu. The 0.06 lb/MMBtu emission limit wechose was based on a thorough review of achievableemission rates of current Dry Flue Gas Desulfurization(DFGD) scrubbers and the example method for theBART guidelines that yields 0.045 lb/MMBtu is not

App. 125

appropriate in this case for estimating future emissionrate for modeling. We chose to model the future SO2

emission rate of 0.06 lb/MMBtu rather than0.045 lb/MMBtu because this is consistent with ourproposed BART emission limit and is a reasonableestimate of future emissions in order to estimate thefuture visibility improvement from baseline levels. Ourapproach of modeling the proposed emission limit isconsistent with the approach taken by ODEQ in theirSIP and in our action on the BART FIP for the State ofNew Mexico and is not as conservative as using theemission rate based on percentage reduction asoutlined in the BART guideline.

As discussed elsewhere, the BART determination isbased on consideration of five factors, including thedegree of improvement in visibility which mayreasonably be anticipated to result from the use of suchtechnology. The visibility modeling is intended to givea reasonable best estimate of the visibility impactsfrom an evaluation of emission reductions. Thevisibility analysis is only one of the factors in a BARTdetermination. In this final action, we are setting a SO2

limit of 0.06 lb/MMBtu to be calculated on a 30-dayrolling average Boiler Operating Day. We modeled the0.06 lb/MMBtu in our proposal, which equates to a 93percent reduction in emissions, because we havedetermined this emission rate to be achievable. Thispercentage reduction is less than would be expectedfrom the installation of a DFGD that has beenoptimally designed (refer to Figure 7 and 8 of theSupplemental RTC and the associated responses tocomments).

App. 126

We recognize that sources complying with a 30 dayaverage may at times operate above the 30 day averageemission limit but they will have to balance those timesby operating below the limit at other times. Thisvariability is difficult to assess, though a prudentsource will strive to remain below the 30-day emissionlimit as much as possible. In some instances, it may beappropriate to model a slightly higher emission ratewhen limiting the emissions using a 30-day average toaccount for potential variability, when the amount ofvariability is well understood. In this case, we believeusing the 30 day average emission limit is a reasonableapproach to project future emissions that wouldreasonably be anticipated in accordance with BARTguidelines because we have no reason to think thevariability in the future case will be large enough toimpact our evaluation of the five factors.

We did not believe it was appropriate to assessvariability based on past history of emissions at thefacilities because there is inherently more variability inhistoric data when facilities are not specificallycontrolling to achieve low SO2 emissions and thefacility emissions instead can vary due to the range oftypes of coal purchased. As the limits are reduced to alevel in the range that was proposed in our action, theamount of variability that would exist is expected todecrease, as the source must demonstrate complianceon a 30-day BOD compliance level with a much tighterlimit than it had previously. We have seen this inevaluation of some sources in comparing their pre-control emission variability with their post-controlemission variability.

App. 127

As discussed in a later response to comment, wenote the TS Power Plant near Dunphy, Nevada, whichhas a similar permitted SO2 emission limit to ourBART FIP, maintained a 30-day BOD emission ratebelow 0.06 lb/MMBtu for an approximately 20-monthperiod of time in 2010–2011. This plant burns a similarPowder River Basin (PRB) coal as the six AEP/PSO andOG&E units. In addition, the Wygen II facility, locatedoutside Gillette, Wyoming, and the Weston 4 facility,near Wausua, Wisconsin, also burn coal similar to theOG&E and AEP/PSO’s units and have been able tomaintain 30-day BOD SO2 emission rates below0.06 lb/MMBtu for significant periods of time duringthe years of 2009–2011. CEM data for the TS Plant(Figure 7 of the Supplemental RTC) shows limitedvariability in 24-hr emissions. We note that this dataincludes periods of start-up, shutdown, andmalfunction that would normally be considered whenevaluating the emission rate to be modeled to representsteady-state operating conditions for BART modeling.In evaluation of other facilities we did find where theyhad operated for months at a significantly loweremission rate than 0.06 lb/MMBtu, with limitedvariability under steady-state conditions.

The commenter pointed to other actions andguidance concerning emission rate estimates andindicated that we were not consistent with thoseapproaches. The commenter pointed to the EPA Region9 proposal for the Four Corners power plant, whichused the percent reduction approach and the 24-hourmaximum actual baseline emission rate to estimate afuture controlled emission rate. We note that weevaluated this technique (see discussion earlier in thisresponse) that is outlined in the BART guideline as one

App. 128

acceptable technique and it resulted in a value(0.045 lb/MMBtu) that was not reasonable compared tothe 30-day emission limit (0.06 lb/MMBtu) that weproposed and determined to be technically feasible. Thecommenter also pointed to guidance that Colorado hasdeveloped for their BART sources that indicates amaximum 24-hour future controlled emission rateshould be used in conjunction with using the maximumactual 24-hour baseline emission rate.

The BART guidelines state:

Make the net visibility improvementdetermination.

Assess the visibility improvement based onthe modeled change in visibility impacts for thepre-control and post-control emission scenarios.

You have flexibility to assess visibilityimprovements due to BART controls by one ormore methods. You may consider the frequency,magnitude, and duration components ofimpairment.

The BART guidelines allow for some flexibility inhow to assess visibility improvements due to BARTcontrols. As we discuss elsewhere in this response, weconsider issues related to frequency, magnitude andduration of emission levels that may occur incomparison to our proposed 0.06 lb/MMBtu 30-daylimit and the potential for impacting the visibilityprojections. We concluded that the amount of times thevariability of emissions would exceed 0.06 lb/MMBtu ona maximum daily process would not be expected to beof sufficient magnitude to have a large impact on ourvisibility improvement estimates. We agree that the

App. 129

BART guidelines allow for some flexibility in howvisibility improvement determinations are conducted.We considered processes similar to Colorado’sapproach, including the methodology given as anexample in the BART guidelines, but determined wedid not have sufficient information to accuratelyestimate the future maximum 24-hour emission rateand furthermore concluded that existing modelingindicated that small changes would not significantlyimpact our visibility improvement estimates. Overall,the BART guidelines give some flexibility to how thevisibility improvements can be calculated and theapproach that we have used is reasonable based on theinformation available and is not inconsistent with theBART guidelines.

We conducted modeling for future emission rates of0.04 and 0.06 lb/MMBtu of SO2 in our proposal. Wenote that at these low SO2 emission rates, the mostimpacted days were more nitrate driven days becausethe SO2 rates were low. Therefore, a slight increase inemission rates on the order of 10% or so for a maximum24-hour emission rate would not be expected to resultin much change in visibility estimates. We do note thatother modeling conducted by the source’s consultantsand the state indicates that a significant increase inthe controlled SO2 emission rate would decrease thevisibility impairment improvements from installationof controls and result in much lower relative visibilityimprovement. As further discussed elsewhere in thisresponse we find our future emission rate to be areasonable assessment of the visibility improvementdue to the setting of a 0.06 lb/MMBtu on a 30-day BODlimit.

App. 130

In summary, we find our approach to modeling thebaseline and control case emissions was a reasonableestimate of reduction in impairment and notinconsistent with the BART guideline. We recognizethat it is possible that the facility will operate atslightly higher emission rates at times, but it is alsotrue that to remain in compliance over a 30-day rollingaverage, it will also have to operate at lower emissionrates than 0.06 lbs/MMBtu. Furthermore, we haveshown that other facilities have demonstrated that it isfeasible to operate below 0.06 lbs/MMBtu for extendedperiods of time. Finally, we have noted that even ifemissions are slightly higher than 0.06 lbs/MMBtu, attimes, it would not be expected to increase the visibilityimpairment significantly because at these lowconcentrations, visibility impairment due to AEP/PSOsources is primarily due to nitrates. We find theapproach for estimating improvements in visibility dueto our proposed emission level that we have used isappropriate based on the information available and isnot inconsistent with the BART guidelines. For thesereasons, we believe the proposal was based on areasonable assessment of visibility improvements forconsideration as one of the five factors of the BARTdecision.

Comment: A commenter submitted a review of ourmodeling results for controlling SO2 emissions, notinga 2.89 deciview improvement in visibility at theWichita Mountains and a cumulative improvement invisibility total of 8.20 deciviews. The commenterbelieves our CALPUFF modeling is appropriate andconcurs with our emission calculations and speciation.They do, however, note several “possibly incorrect inputvalues” regarding base elevations of several units and

App. 131

the stack gas exit velocity of one unit. The commenterexpressed the view that corrected values would notsubstantially change results and conclusions. Thecommenter also contends that EPA’s proposed SO2

BART may benefit Oklahoma and the facilities,because the commenter believes that based on resultsof their dispersion modeling, the units are currentlycontributing to violations of the one-hour SO2 NAAQS.

Response: We agree with the commenter that ourmodeling calculations and speciations are appropriate.We further agree with the commenter’s noted visibilityimprovement resulting from the SO2 controls that weare requiring in the FIP. It is true that states will berequired to submit plans demonstrating attainment ormaintenance of the new one-hour SO2 NAAQS.However, this is not a consideration for our action,which is directed solely to ensuring the state has metthe BART requirements of the RHR and therequirements of CAA section 110(a)(2)(D)(i)(II). Withrespect to the noted “possibly incorrect input values,”we agree that correcting these values would notsubstantially change our results and conclusions.

E. Summary of Responses to Comments on the SO2

BART Cost Calculation

We received many comments on issues concerningour cost calculations for our proposed SO2 BARTdeterminations on the six OG&E and AEP/PSO units.The full text received from these commenters isincluded in the docket associated with this action.Additionally, our summary and response for thesecomments is provided in the “Response to TechnicalComments for Sections E through H of the FederalRegister Notice for the Oklahoma Regional Haze and

App. 132

Visibility Transport FIP,” (or Supplemental RTC), andit is available in the docket. Although we summarizethem here, please see the Supplemental RTC for a fullaccounting of the issues and how they influenced ourfinal decision. We deviate in sections E., F., G., and H.,from the comment-response format of the rest of thenotice, as many of the comments summarized hereinwere drawn from multiple, lengthy, and highlytechnical comments.

The significant aspects of our approach to costestimations in consideration of all comments aresummarized in this section. Overall, our finalrulemaking retains the basis for the cost effectivenessevaluation and cost estimates we employed in ourproposal. However, as discussed in more detail below,we are changing several factors in the cost calculationsfor the four OG&E units as a result of the commentswe received. We are making no changes to the costcalculations for the two AEP/PSO units.

1. Control Cost Manual Methodology

The Control Cost Manual must be followed to theextent possible when calculating the cost of BARTcontrols.15 This is necessary to ensure that a consistent

15 Very limited situations exist under which an analyst can departfrom the Control Cost Manual methodology under the RH rule.“The basis for equipment cost estimates also should bedocumented, either with data supplied by an equipment vendor(i.e., budget estimates or bids) or by a referenced source (such asthe OAQPS Control Cost Manual, Fifth Edition, February 1996,EPA 453/B–96–001). In order to maintain and improveconsistency, cost estimates should be based on the OAQPS ControlCost Manual, where possible. The Control Cost Manual addresses

App. 133

methodology is used when comparing cost effectivenessdeterminations. The Control Cost Manual allows site-specific conditions to be incorporated in certaincircumstances. Site-specific conditions can includevendor quotes, space constraints, a design feature thatcould complicate installing a control, or unusualcircumstances that introduce a cost not contemplatedby the Control Cost Manual. OG&E incorporated manyof these into its cost evaluation. However, the RHRspecifically requires that the analyst document anysuch site-specific conditions.16 Thus, the RHR placesthe burden on the analyst to make this demonstration,and on EPA to approve it, disapprove it, or document itwhen promulgating a FIP. Nevertheless, with theexceptions noted herein and in our Supplemental RTC,we approved many of those site-specific costmodifications.

The Control Cost Manual uses the overnightmethod of cost estimation, widely used in the utilityindustry.17 The U.S. Energy InformationAdministration (EIA) defines “overnight cost” as “anestimate of the cost at which a plant could be

most control technologies in sufficient detail for a BART analysis.”70 FR 39104, at 39166.

16 A cost determination can deviate from the Control Cost Manualmethodology if you “include documentation for any additionalinformation you used for the cost calculations, including anyinformation supplied by vendors that affects your assumptionsregarding purchased equipment costs, equipment life, replacementof major components, and any other element of the calculation thatdiffers from the Control Cost Manual.” Id.

17 See Control Cost Manual, Section 2.3 to 2.4.

App. 134

constructed assuming that the entire process fromplanning through completion could be accomplished ina single day. This concept is useful to avoid any impactof financing issues and assumptions on estimatedcosts.”18 EIA presents all of its projected plant costs interms of overnight costs. The overnight cost is thepresent value cost that would have to be paid as a lumpsum up front to completely pay for a constructionproject.19 The overnight method is appropriate forBART determinations because it allows differentpollution control equipment to be compared in ameaningful manner. Because “different controls havedifferent expected useful lives and will result indifferent cash flows, the first step in comparingalternatives is to normalize their returns using theprinciple of the time value of money * * * . The processthrough which future cash flows are translated intocurrent dollars is called present value analysis. Whenthe cash flows involve income and expenses, it is alsocommonly referred to as net present value analysis. Ineither case, the calculation is the same: Adjust thevalue of future money to values based on the samepoint in time (generally year zero of the project),employing an appropriate interest (discount) rate andthen add them together.”20 This is the overnightmethod, in which costs are calculated based on current

18 EIA, “Updated Capital Cost Estimates for Electricity GenerationPlants,” November 2010, footnote. 2, available at: http://www.eia.gov/oiaf/beck_plantcosts/?src=email.

19 Steven Stoft, Power Economics: Designing Markets forElectricity, 2002.

20 Id., page 2–18.

App. 135

dollars. Therefore, consistent with our proposal, wefind that the overnight method is appropriate forcalculating costs for all six units.

OG&E and others incorrectly assume that BARTcost effectiveness should be based on the “all-in” costmethod, which includes all of the costs of a financialtransaction, including interest, commissions, and anyother fees from a financial transaction up to the datethat the project goes into operation, as of the assumedcommercial operating dates of the scrubbers, 2014 and2015. This is an entirely different method than thatprescribed in the Control Cost Manual. OG&E andothers conclude that dry scrubbers are not cost effectivefor the six units, based on all-in costs reported in 2014to 2015 dollars, compared to costs estimated at othersimilar facilities based on overnight costs and 2009 andearlier dollars. This comparison is an invalid becauseOG&E’s 2014 and 2015 all-in costs are much higherthan the corresponding overnight costs, as prescribedby the Control Cost Manual. This makes the estimatedcost of scrubbers at the six units appear to be higherthan scrubbers required at other similar facilitiescosted using the overnight method. Many of thecorrections we make to ODEQ’s cost estimates for thesix OG&E and AEP/ PSO units are due to the fact thatODEQ did not follow this provision of the Control CostManual in its SIP submittal. Please refer to ourSupplemental RTC in the docket for more informationabout how the overnight costing methodology isemployed by the Control Cost Manual.

2. Revised Cost Calculations for the OG&E Units

OG&E’s cost estimates deviate from the ControlCost Manual, which is based on the overnight cost

App. 136

approach. In its cost estimates, OG&E has improperlyincluded allowances for excessive contingenciesallowances for funds during construction (AFUDC),double counted certain expenses, and improperly reliedon the Electric Power Research Institute (EPRI) costmodel, CUECost. These deviations from the ControlCost Manual, occurring because of the reliance uponthe all-in cost methodology, artificially increase the costof scrubbing at Sooner and Muskogee, compared to thecost at other similar facilities using the overnight costmethodology.

OG&E’s cost estimates relied on vendor quotes andsite specific estimates for certain additional costs. Wesupport the use of vendor quotes and site specificestimates but only as used within the parameters ofthe overnight cost methodology. The Guidelines, citedin this comment, are clear that “[y]ou should includedocumentation for any additional information you usedfor the cost calculations, including any informationsupplied by vendors that affects your assumptionsregarding purchased equipment costs, equipment life,replacement of major components, and any otherelement of the calculation that differs from the ControlCost Manual.”21 However, much of the documentationOG&E and others cite to support deviations from theControl Cost Manual was not provided to us. Thus, wewere unable to analyze their contents and determinewhether these deviations were appropriate. Also,although OG&E provided two spreadsheets that listedits cost line items, these spreadsheets, each over 600lines in length, were stripped of all formulas for cell

21 70 FR 39104, at 39166, footnote 15.

App. 137

calculations, preventing any meaningful review,despite our request for that material.

Capital Recovery Factor

We are changing one input to the cost calculationsfor the four OG&E units based on a comment wereceived from OG&E concerning the Capital RecoveryFactor (CRF). OG&E states that, while the ControlCost Manual includes a default rate of 7% for the socialdiscount interest rate, we should use a site-specificsocial discount interest rate for the four OG&E units.This rate includes several site-specific variables,including income tax. The commenter states that theCRF includes not only recovery of principal but also areturn on the principal, with the rate of return equal tothe discount rate. OG&E states that for an investorowned utility, such as itself, which is financed by a mixof debt and equity, the discount rate is equal to theweighted average of the equity return and debt return.

We agree that a site-specific social discount interestrate is appropriate based on the documentationprovided by the commenter. However, we disagree thatsuch a rate can include income tax. The Control CostManual states “this Manual methodology does notconsider income taxes.” Control Cost Manual, page 2–9.The site-specific social discount interest rate, excludingincome tax, is 6.01%, which is less than the default rateof 7%. Thus, we have revised our cost effectivenessanalysis in Exhibits 1 and 2 for Options 1 and 2, to usethe levelized interest rate of 6.01%, as reported byOG&E, adjusted to remove income taxes. This rate isconsistent with OG&E’s real average cost of capital andfalls within the range of 3% to 7% recommended byOMB for regulatory cost analyses. This correction

App. 138

moderately improved the cost effectiveness, thuslowering the calculation of $/ton SO2 removed. Fordetailed information on our calculation, please see theSupplemental RTC.

Construction Management

In our proposal, we revised the cost estimate toremove what we took to be double counting of theBalance of Plant (BOP) construction managementcosts. OG&E explained in a comment that crew wagerates do not include contractor general andadministrative (G&A) costs and that constructionmanagement is the cost of third-party constructionmanagement, different from the BOP profits contractorand different from the owner. Based on thisexplanation, we have restored the constructionmanagement costs in our revised Options 1 and 2 costestimates in Exhibits 1 and 2. This correction slightlydiminished the cost effectiveness, thus raising thecalculation of $/ton SO2 removed.

Scrubber Design and Emission Baseline Mismatch

We retain both our Option 1 and Option 2 costeffectiveness approaches to the mismatch between thedesign of OG&E’s SO2 scrubbers and the coal theycurrently burn. OG&E specified to its vendors thatthey provide cost estimates for SO2 scrubber systemsdesigned to treat the exhaust gases from a coal thatcontains much higher amounts of sulfur than coals thatwere typically burned in the baseline period(2004–2006). However, in calculating the costeffectiveness, OG&E used its historical baselineemissions, which resulted from the burning of thoselower sulfur coals. Thus, OG&E costed scrubbers that

App. 139

were overdesigned based on the coal that was, and is,typically burned. This resulted in two errors that bothcombined to make the control technology appear lesscost effective.

First, the BART Guidelines require that wecalculate cost effectiveness on the basis of annualizedcost divided by tons of pollutant removed from theemissions baseline ($/ton). Therefore, use of a baselinethat is lower than would result from burning thehigher sulfur coal the scrubber was designed to treat,lowers the denominator in the $/ton equation, andskews the cost effectiveness calculation to appear lesscost effective. We account for this mismatch in Option1 by raising the baseline to match the higher sulfurcoal the scrubber system was designed to treat.

Second, although we have adjusted our calculationin response to OG&E’s comments, we conclude that theover designed scrubber system was more expensivethan necessary to treat the coal OG&E historicallyburned and continues to burn. We account for thismismatch in Option 2 by slightly decreasing the capitalcosts to reflect a scrubber designed to treat the exhaustgases from the coal OG&E has historically burned,while retaining the historical emission baseline.

We find that, whether OG&E chooses to burn itscurrent coal, or burn a coal that its scrubber systemwas designed to treat, the resulting cost effectivenesslies in the range defined by Options 1 and 2 (below). We find that both options are cost effective in light ofthe five-step BART analysis.

App. 140

Cost Adjustment of Scrubber in Option 2

As we describe above, in calculating costeffectiveness under Option 2 in our proposal, we alsoanalyzed the cost of a dry scrubber for the OG&E units,assuming the scrubber would be re-sized to scrub thecoal being currently burned. We did this using a costscaling equation based on the differences between thesulfur content of the coal OG&E typically burns versusthe coal their scrubber system was designed to treat.OG&E responded in a comment to us that the exhaustgas flow rate, rather than the sulfur content, is theprimary variable that affects scrubber sizing. Thus, theuse of a higher sulfur coal would not significantly affectthe size, and hence the cost of a scrubber. Based on theinformation OG&E supplied, we re-adjusted the cost ofOption 2 based on certain design algorithms in the dryscrubber absorber (SDA) cost model developed byOG&E’s contractor, Sargent & Lundy for EPA.22 Theresults of this analysis indicate that the use of thelower sulfur coal alone would reduce the capital cost ofthe scrubber by about $7 million or 3%.

Other Issues Concerning Site-Specific Costs

In addition to those comments that resulted in amodification to our cost basis, two others meritparticular emphasis. These comments led us toinvestigate two other line item costs to determinewhether we underestimated the costs of the scrubbersfor the four OG&E units by not using site-specific

22 Sargent & Lundy, IPM Model—Revisions to Cost andPerformance for APC Technologies, SDA FGD Cost DevelopmentMethodology, Final, August 2010, Table 1.

App. 141

values. We determined that, even if we made changesto the cost calculations to account for these site-specificcost line items, the cost of controls would be even morecost-effective than our proposed range. These line itemscosts are: (1) Auxiliary power; and (2) capacity factorfor Option 2. These issues were uncovered during thecourse of preparing our response to comments, but didnot directly follow from information provided by thecomments. Thus, we did not further modify our costbasis, but discuss these issues as they serve to furtherillustrate why we believe our cost basis likelyoverestimates the costs of control and that ourconclusions that dry scrubbers for the six OG&E andAEP/PSO units are cost effective and are reasonable.

a. Auxiliary Power

We received a comment that EPA incorrectlylowered OG&E’s auxiliary power costs for theDFGD/FF control systems on the premise that the unitcost of electricity used in the cost estimate was higherthan the cost to OG&E to produce electricity. Auxiliarypower is the sum of the demand by the scrubber,baghouse, and booster fans (the latter required toovercome the increase in backpressure from addingthese controls) and is accounted for in a BART costeffectiveness analysis. OG&E used average year-roundmarket retail rates of $85.93/MWh (2015 dollars) forSooner and $83.83/MWh (2014 dollars) for Muskogee asthe best long-run measure of auxiliary power costs. Thecost of auxiliary power affects the cost effectivenesscalculation in both Option 1 and Option 2.

We have concluded that our proposed cost of$50/MWh is an appropriate estimate of the cost ofauxiliary power for the four OG&E units. We arrived at

App. 142

this number because OG&E’s summary of auxiliarypower costs indicates the range used for other similarfacilities is $30/MWh to $50/MWh.23 We took the mostconservative view based on this report and adopted thehighest value in this range. However, even if we wereto take OG&E’s view that a site-specific auxiliarypower cost is more appropriate, we disagree that wecould use the market-value of power for purposes of theBART determination because the utility would not paymarket price. We estimate that the actual site-specificcost of auxiliary power for the four OG&E units is nomore than $36/MWh. However, because we arrived atthis figure due to independent research that we do notview as being a logical outgrowth of the comment wereceived, we have not revised our cost effectivenessanalysis to use $36/MWh. Instead, we retain the$50/MWh figure we proposed. We view this example asfurther evidence that OG&E’s scrubber costs areartificially inflated, and that the cost of controls underboth options in our FIP is reasonable.

b. Capacity Factor in Option 2

ODEQ calculated future annual emissions assuminga 90% capacity factor. In comparison, during the yearsthat established the emission baseline (2004–2006), theunits operated only 78.5% of the time, on average.Thus, ODEQ’s calculation of emission reductions fromscrubbers compares uncontrolled 2004–2006 baselineemissions, when the units operated at 78.5% of

23 December 28, 2009 S&L FollowUp Report, Attach. C, pdf 109(Gerald Gentleman—$45.65/ MWh;White Bluff—$47/MWh;Boardman/Northeastern/Naughton—$50/MWh; NebraskaCity—$30/MWh).

App. 143

capacity, to controlled emissions when burning ahigher sulfur coal, with the units operating at 90%capacity. This mismatch results in two errors inestimating the cost of Option 2: The future emissionswere overestimated, but certain operating costs wereunderestimated. Correcting these errors in the costcalculations would make Option 2 even more costeffective than our proposed calculations, as theresulting decrease in the operating costs would offsetthe increase in the capacity factor in the $/toncalculation. However, because we arrived at theseerrors due to independent research that we do not viewas being a logical outgrowth of the comment wereceived, we have not revised our cost effectivenessanalysis in Option 2. We view this example as furtherevidence that OG&E’s scrubber costs are artificiallyinflated, and that the cost of controls under bothoptions in our FIP is reasonable.

We made no additional changes to our costevaluation as a result of the comments we received. Assummary of our final $/ton cost effectivenesscalculations are provided below:

Proposal(Sooner/Muskogee)

Final(Sooner/Muskogee)

Option 1 $1,291/$1,317 $1,239/$1,276

Option 2 $2,048/$2,366 $2,747/$3,032

App. 144

3. Cost Calculations for the AEP/PSO Units

We received a number of comments from AEP/PSOconcerning our SO2 BART cost estimate for the twoNortheastern units. Some of these comments objectedto our incorporation of OG&E’s site specific informationin AEP/PSO’s scrubber cost estimate. Other commentsobjected to specific line item costs in our cost estimatesfor both wet and dry scrubbers. We proposed the costeffectiveness of dry scrubbing to be $1,544/ton, and thecost effectiveness of wet scrubbers to be approximately9% more. As we note in more detail in our separateSupplemental RTC, the ODEQ SO2 BART evaluation ofAEP/PSO Northeastern units 3 and 4 does not provideany support for its assumption that the cost of dryscrubbers is $555/kW to $582/kW, figures we considerto be high in comparison to other BART scrubberdeterminations. However, the Northeastern units arevery similar to the Sooner and Muskogee units, forwhich vendor quotes were available for dry scrubbers.We used these vendor quotes to support our costanalysis for the Northeastern units. After havingreviewed all comments concerning our SO2 BART costestimates for the AEP/ PSO units, we have determinedthat no changes were warranted to our proposed costestimates. Thus, absent any supporting informationfrom AEP/PSO for any of the capital costs it presents,we find our BART SO2 cost evaluation to be wellfounded, representative of the AEP/PSO units inquestion, and based on the best information availableto us.

4. Conclusion

We find that under Option 1, the costs to complywith the FIP will be $1,239/ton for Units 1 and 2 of the

App. 145

OG&E Sooner plant and $1,276/ton for Units 4 and 5 ofthe OG&E Muskogee plant. Under Option 2, the cost tocomply with the FIP will be $2,747/ton for Units 1 and2 of the OG&E Sooner plant and $3,032/ton for Units4 and 5 of the OG&E Muskogee plant. For Units 3 and4 of the AEP/PSO Northeastern plant, we find that thecosts to comply with the FIP remain at $1,544/ton, aswe proposed. We find these ranges to be cost effectivefor these six units under the five-step analysis forBART under the RHR. As previously stated, ourcomplete, technical responses to comments received onthe issue of costs are in the Supplemental RTC in thedocket.

F. Summary of Responses to Visibility ImprovementAnalysis Comments

We received comments on Step 5 of BART: Degreeof improvement in visibility which may reasonably beanticipated to result from the use of scrubbertechnology. Commenters contested our determinationthat OG&E and AEP/PSO’s facilities significantlycontribute to visibility impairment. We explain that wefind that dry scrubbers are cost effective for the sixOG&E and AEP/PSO units, in light of the visibilityimprovement these controls are predicted to achieve.Commenters also disputed our determination not touse the $/deciview metric in the Step 5 BART analysiswhen this approach was used by ODEQ. OG&Eprovided a $/deciview analysis for its units andcomparable BART determination performed by us. Inour analysis for our BART FIP for OG&E andAEP/PSO, we did not evaluate $/deciview. We explainthat the BART Guidelines list the $/deciview metric asan optional cost effectiveness measure that can be

App. 146

employed along with the required $/ton metric for usein a BART evaluation. The metric can be useful incomparing control strategies or as additionalinformation in the BART determination process;however, due to the complexity of the technical issuessurrounding regional haze, we have neverrecommended the use of this metric as a cutpoint inmaking BART determinations. We note that to use the$/deciview metric as the main determining factor wouldmost likely require the development of thresholds ofacceptable costs per deciview of improvement for BARTdeterminations for both single and multiple Class Ianalyses. We have not developed such thresholds foruse in BART determination made by us. As OG&Eacknowledges, EPA did not use this metric as part ofits proposed BART determinations for either the FourCorners Power Plant FIP in AZ, or the San JuanGenerating Station FIP in NM. Generally speaking,while the metric can be useful if thoughtfully applied,we view the use of the $/deciview metric as suggestinga level of precision in the calculation of visibilityimpacts that is not justified in many cases. While wedid not use a $/deciview metric, we did, however,consider the visibility benefits and costs of controltogether, as noted above by weighing the costs in lightof the predicted visibility improvement.

G. Summary of Responses to Comments Received on theSO2 BART Emission Limit

We received comments stating we did notadequately support our SO2 BART emission limit of0.06 lbs/MMBtu for the six OG&E and AEP/PSO units.In analyzing the control technology, the RHR mandatesthat we take into account the most stringent emission

App. 147

control level that the technology is capable of achieving.70 FR 39104, at 39166. In accordance with the RHR,when identifying an emissions performance level toevaluate under BART, consideration of recentregulatory decisions and performance data (e.g.manufacturer’s data, engineering estimates, and theexperience of other sources) is required. Id. Indetermining our SO2 BART emission limit of0.06 lbs/MMBtu, we drew on a number of sources ofinformation. These include industry reports, vendorquotes, the engineering analysis contained in the TSD,and the historical emissions data for other similar coalfired power plants. As we state in the TSD and affirm,a dry scrubber at Sooner or Muskogee, designed ascosted, could meet an SO2 emission limit of0.06 lb/MMBtu based on 30-day BOD average, whenburning coal containing 0.51 to 1.18 lb/MMBtu SO2. Weconclude the same is true for the AEP/PSONortheastern units because they have historicallyburned coal with a sulfur content within this range.24

Among other objections, OG&E states we cannotrely on the SO2 emission performance of new facilitiesas an indicator of the performance potential of retrofitscrubbers. OG&E presents data on what it states arethe best performing scrubber installations in theUnited States, and contends that the lowest emissionrate achieved by a retrofit on an annual basis is0.088 lbs/MMBtu. We explain that a scrubber,regardless of type, is not influenced by whether the fluegas comes from a new boiler or an old boiler located inan existing plant. The scrubber merely reacts to

24 TSD, Appendix C, page 43.

App. 148

physical and chemical characteristics of the gas stream.Therefore, although we use other sources ofinformation to justify our SO2 BART emission limit, wefind that considering emission data from new scrubberinstallations to support our decision is appropriate. Inso doing, we analyzed the historical emissions data ofseveral units that we discuss above in response toanother comment, which OG&E included in itscomment. We reviewed the performance of three unitsthat are of similar size and burn similar coal. One unit,TS Power Plant, has an emission limit that requiresemissions to be significantly controlled and has beenable to maintain its emissions below 0.06 lbs/MMBtuon a 30 day BOD basis continuously. We also reviewedthe performance of two other units that demonstratethe ability to maintain emissions below the0.06 lbs/MMBtu limit for long periods of time. We notethat these units do not have as constraining emissionlimits so they do not have to control their emissions asclosely. This and other sources of information weoutline above and in our Supplemental RTC cause usto conclude our proposed SO2 BART emission limit of0.06 lbs/MMBtu, calculated on the basis of a 30 dayBOD, for the six OG&E and AEP/PSO units istechnically feasible and therefore the correct SO2 limitfor BART.

OG&E also states that we should include in ourproposed SO2 BART emission limit a compliancemargin. OG&E suggests that a SO2 emission of 0.10 isrequired to provide a “reasonable margin for operatingfluctuations and compliance.” We reply that we aremodifying the compliance averaging period from a 30calendar period to a 30 day Boiler Operating Day(BOD) period. As the BART Guidelines direct, “[y]ou

App. 149

should consider a boiler operating day to be any 24-hour period between 12:00 midnight and the followingmidnight during which any fuel is combusted at anytime at the steam generating unit.”25 To calculate a 30day rolling average based on boiler operating day, theaverage of the last 30 “boiler operating days” is used. Inother words, days are skipped when the unit is down,as for maintenance. This, in effect, provides a marginby eliminating spikes that occur at the beginning andend of outages, and is consistent with the BARTGuidelines.

In our separate Supplemental RTC, we also discussseveral other objections OG&E raises in its comments.These include objections to our reliance on a NationalLime Association scrubber performance chart, OG&E’scontention that our proposed SO2 BART emission ismore representative of a LAER limit, and the technicalcapability of dry scrubbing. After addressing theseissues, we find that our proposed SO2 BART emissionfor the six OG&E and AEP/ PSO units remains at rateof 0.06 lbs/MMBtu.

H. Summary of Responses to Comments Received on theSO2 BART Compliance Timeframe

We proposed that compliance with our SO2 BARTemission limits be within three years of the effectivedate of our final rule. We solicited comments onalternative timeframes, from as few as two (2) years toup to five (5) years from the effective date of our finalrule. We received comments that retrofitting ofscrubbers is now routine in the United States and that

25 70 FR 39104, at 39172.

App. 150

approximately 290 coal-fired units totaling about116,000 MW nationwide have been retrofit withscrubbers since 1990. The commenter cites to manyexamples of SO2 scrubbers being installed at coal-firedpower plants within a three year timeframe. OG&Eand others state that our proposed three year schedulefocuses on actual construction timelines, but fails toacknowledge or allow sufficient time for theengineering, design, and permit processes that must becompleted prior to the commencement of construction.They state a compliance schedule of from 52–54months would be required.

Although we do not specify what technology the sixOG&E and AEP/PSO units must use to satisfy the SO2

BART emission limit, we expect that either dry or wetSO2 scrubbers will be used, or that the SO2 limit will bemet by switching one or more of the units to naturalgas. We agree that SO2 scrubbers have been installedat other facilities with construction timeframes of threeyears or less. However, we also agree with OG&E andAEP/PSO that there may be issues such as PSDpermitting, and the construction/expansion of a landfillthat may not be reflected in the example compliancetimes reported by the commenter. Therefore, we findthat compliance with the emission limits be within fiveyears of the effective date of our final rule.

I. Comments Supporting Conversion to Natural Gasand/or Renewable Energy Sources

Comment: Several parties submitted commentsnoting that switching to natural gas-fired electricity isfeasible and demonstrated in practice. One of thecommenters points out that, of the three subject sites,two have existing major natural gas supplies (OG&E

App. 151

Muskogee and AEP/PSO Northeastern) and that fuelswitching will require construction of new or expandednatural gas supply and electric interconnectionfacilities. The commenter states that expanding alongexisting gas supply lines would cost less and take lesstime than constructing a new line. The commentershave stressed that natural gas produces comparativelylow emissions of many pollutants, including haze-causing pollutants, air toxics, and greenhouse gases.Commenters also noted use of natural gas as a fuelsource would eliminate the need to manage coalcombustion waste and scrubber waste. Severalcommenters who support the switch from coalcombustion to natural gas combustion cited theavailability and abundance of natural gas as a naturalresource, particularly in Oklahoma.

Response: We agree that switching of existing coal-fired power generating units to natural gas, eitherthrough conversion of existing boilers or installation ofnew power generating units, is technically feasible anddemonstrated in practice. As stated in our proposal, theowners of the units subject to the FIP may elect toreconfigure the units to burn natural gas as means ofsatisfying their BART obligations under section51.308(e). Switching to natural gas would be anacceptable method of complying with the limitsproposed in the FIP, because natural gas combustioninherently results in much lower SO2 emissions. Weagree that natural gas may result in lower emissions ofother pollutants and offer other environmentaladvantages. The owners of each subject unit may takethese advantages, as well as the availability andpricing information, into consideration as they evaluate

App. 152

this option for complying with SO2 BART emissionlimits.

Comment: Eight commenters responded to ourrequest for comments on the compliance deadline forthe six BART-subject units and whether it would beappropriate to extend that deadline for those utilitiesthat elected to switch from coal to natural gas in orderto comply with the BART emission limits. Several ofthese commenters note that switching to natural gascan be accomplished in less than three years if utilitiesenter into long-term power purchase agreements withexisting natural gas-fired power generators but utilitiesthat choose to construct new gas-fired units or convertexisting units will likely require more time. Theyindicate that the requirements to engage in competitivebidding, complete engineering designs, preparebudgets, obtain necessary permits, and equipmentinstallation will likely require up to five years tocomplete. One of these commenters points out thatOG&E has already studied fuel-switching at thesystem and plant levels and that the typical lead timeof construction of new natural gas-fired combined cyclecombustion turbines is four years.

Numerous commenters express their support forextending the compliance deadline to five years forunits that will be converted to, or replaced with,natural gas-fired power generating units. Thesecommenters cite the broad collateral benefits andoverall superiority of switching to a cleaner fuel sourceover installing additional controls on the existing unitsand continuing to burn coal.

Multiple other commenters, however, expressed theopinion that the utilities have had ample time already

App. 153

to transition away from coal to cleaner or renewablepower generation and that the affected utilities shouldphase out the BART-subject coal-fired units as quicklyas possible. These commenters feel that the proposedcompliance deadline of three years is adequate.

ODEQ submitted comments supporting a fourteenand one-half month extension (to four years and twoand one-half months total) on the installation ofscrubbers and a seven and one-half year extension (toten and one-half years total) for switching to naturalgas.

Response: We thank the commenters for theirresponses to our request for comments on the proposedcompliance deadline. As we have discussed elsewherein our response to comments we find that a compliancedeadline of five years is appropriate for the six OG&Eand AEP/PSO units to comply with our FIP SO2

emission limit. After reviewing the informationprovided by the commenters, we find that the samecompliance deadline of five years is appropriate for anyof the six OG&E and AEP/PSO units that elect tocomply with the FIP SO2 emission limit by convertingan existing unit to natural gas or replacing it with anew, natural gas-fired unit.

Comment: Several commenters providedinformation concerning underutilized electricalgeneration capacity through natural gas combustion inOklahoma. One commenter further suggested that fuelswitching could be achieved by imposition of annualemissions caps on the BART-subject, coal-fired units.According to the commenter, such a scheme wouldprovide the affected utilities with the flexibility to shiftpower generation to existing gas-fired generating units

App. 154

or purchase power from merchant generators. Thecommenter states that there is an exception provisionin the RH regulations at 40 CFR 51.308(e)(2) thatallows for imposition of operating limits on BART-eligible units in lieu of conventional BART reductionsif the regulating authority implements an emissiontrading program.

Another commenter noted that switching to naturalgas-fired generation, either through conversion ofexisting units or replacement with new units, wouldresult in power plants better suited to integrate withvariable wind power generation.

Response: Section 51.308(e)(2) allows Oklahoma toimplement an emissions trading program or otheralternative measure in lieu of BART. Among otherrequirements, such an alternative to BART mustachieve greater reasonable progress than would beachieved through the installation and operation ofBART. However, Oklahoma did not include such aprogram as part of its RH SIP, and we cannot requireOklahoma to establish an emission trading programthat would support annual emission caps or operationallimits on the six BART-subject units. We also note thatas a practical matter, there is no longer adequate timeto develop and implement such an emissions tradingprogram and meet our consent decree deadline withWildEarth Guardians of December 13, 2011 if weattempted to develop and implement such an emissiontrading program as part of our action.26 Whether or notexisting natural gas-fired power generation capacity in

26 See, WildEarth Guardians v. Jackson, Case No. 4:09-cv-02453-CW (N. Dist. Cal.).

App. 155

Oklahoma and other parts of the Southwest Power Poolis underutilized has no direct bearing on our SO2 BARTdeterminations.

Comment: We received multiple comments fromnumerous parties concerning the economics ofswitching from coal-fired to natural gas-fired powergeneration. These comments focused on a wide range ofeconomic issues, including cost-benefit analysis of oneBART compliance alternative over another, future riskto ratepayers due to future maintenance andcompliance costs, economic impact of increasingreliance on renewable energy sources, and ancillarybenefits to the economy of switching from coal tonatural gas or renewable energy sources.

Many of the comments we received pertain to theadditional economic burden of addressing coalcombustion and scrubber waste that would continue tobe generated by the six BART-subject coal-fired unitsif the utilities elect to comply with the BARTrequirements of the proposed FIP by installingscrubber units, rather than fuel switching. Onecommenter provided an economic analysis indicatingthat containment of the coal ash and scrubber wastewould cost $180 million in capital investment and$2–$5 million annually for disposal of residuals if theutilities can sell the fly ash, or up to $9 millionannually if the fly ash cannot be sold. The commenterfurther asserts that scrubbing all six of the BART-subject coal-fired units could generate up to 600,000tons per year of flue gas desulfurization wastebyproducts, the disposal of which could cost anadditional $22 million annually. Two commenters haveasserted that the power generation capacity of the six

App. 156

OG&E and AEP/PSO units can be replaced with theconstruction of new, modern natural gas-firedcombined cycle turbines for less money than would berequired to install scrubbers on the coal-fired units tomeet BART emission limits.

Other comments focused on the likely imposition offuture, additional environmental regulatory compliancecosts associated with continued firing of coal, such asrequirements for new baghouses to control emissions ofparticulate matter and metals, construction ofimproved and expanded containment of coalcombustion residuals, and carbon emission reductionsor sequestration. These commenters noted thatattempting to further extend the lives of the six OG&Eand AEP/PSO units is a bad investment when suchadditional controls for other pollutants are foreseeable,and that switching to natural gas power generationwould reduce the risk to ratepayers of the eventual costincreases associated with these additional regulatoryrequirements.

Several commenters noted that the six OG&E andAEP/PSO units are approaching the end of their usefullives and that switching to natural gas and renewableenergy sources will decrease the risk to ratepayers ofincreased maintenance costs due to the advanced ageof the units.

Other commenters, some of whom identifiedthemselves as ratepayers at the affected utilities,indicated that they would be willing to pay an increasein power rates in exchange for power that wasgenerated by cleaner fuels or renewable energy sources.These commenters cited the overall health andenvironmental benefits that would result from a

App. 157

transition away from coal-fired power and expressedtheir belief that such benefits would outweigh anypotential increase in electricity rates.

Finally, two commenters suggested that switchingto natural gas and/or renewable energy sources wouldhave collateral economic benefits by creating new jobsand providing general economic stimulus in the region.

Response: We affirm that each of the sources subjectto BART under the FIP can acceptably meet theemission limits in the FIP by switching to natural gas.As the companies evaluate how to satisfy their BARTobligations, we encourage them to consider switchingfrom coal to natural gas at the six affected units as thismay offer numerous, significant long-term financialand environmental benefits over the option ofcontinued use of coal with additional controls. As wasstated in our proposal, we do not wish to dissuadecompanies from exercising this option. As we discusselsewhere in our response to comments andSupplemental RTC, we find that a compliance deadlineof five years is appropriate for any of the six OG&E andAEP/PSO units that elect to comply with the FIP SO2

emission limit by converting an existing unit to naturalgas or replacing it with a new, natural gas-fired unit.

Comment: Several commenters expressed concernover the potential rate increases that might result froma switch to natural gas or some form of renewableenergy sources and the impact of those rate increaseson households with low or fixed incomes.

Response: The companies owning each of the sourcessubject to BART are only required to satisfy the SO2

BART emission limits at those sources. Our action only

App. 158

contemplates the reconfiguration of existing units. Wehave determined that reconfiguration would be costeffective with application of dry and wet scrubbingtechnology. Though the SO2 BART emission limits mayalso be met with reconfiguration of the units to burnnatural gas, the companies themselves are free todetermine whether this option best responds to futurecustomer needs and preferences, including anypotential impact on rates. As we state elsewhere in thisresponse to comments and the Supplemental RTC,although we based our BART determination of the useof SO2 dry scrubbers, the owners of the six units inquestion are free to consider any technology to meettheir SO2 BART obligations, including switching tonatural gas. We acknowledge the potential benefitsthat the commenters suggest of switching the units inquestion to burn natural gas. Renewable energytechnology is not a retrofit option for the sourcessubject to BART and is accordingly outside the scope ofour action.

Comment: Several commenters have expressed theview that it does not make good economic sense toinvest heavily in new control equipment in order tomeet BART on units that are so close to retirement.Some of these commenters point out that it makesmore sense to invest in new natural gas-fired unitsinstead of converting the existing boilers to burnnatural gas, given the size of the investments beingconsidered and the advanced age of the existing coal-fired units.

Several of the comments focused on the long-termeconomic benefits of construction of new natural gas-

App. 159

fired units over conversion of the existing boilers at thesix coal-fired units to meet the BART emission limits.

Response The BART guidelines do allow forconsideration of the remaining useful life of facilitieswhen considering the costs of potential BART controls.Such a claim would have to be secured by anenforceable requirement. Neither OG&E nor AEP/PSOclaimed any such restriction on the operation of thesesix units and Oklahoma did not submit any enforceabledocument for action by us. Consequently, we assumeda remaining useful life of 30 years in our BARTanalysis.

If OG&E and/or AEP/PSO decide the units inquestion have a shorter useful life such that installingscrubbers is no longer cost effective, and are willing toaccept an enforceable requirement to that effect, arevised BART analysis could be submitted by theplant(s) in question and our FIP could be re-analyzedaccordingly. Similarly, we could also review a revisedSIP submitted by ODEQ.

Comment: Numerous commenters expressed broadsupport for transitioning away from coal and otherfossil fuels to sources of energy that are completelyrenewable, such as wind and solar-generated power.These commenters recommend that the BART-subjectunits should be replaced with wind-powered unitswhere possible and that natural gas should be used forpower generation during periods of low wind yield. Oneof the commenters notes that Oklahoma and otherparts of the Southwest Power Pool (SPP) haveenormous potential for wind farm development andthat as of July 2010 the SPP transmissioninterconnection queue had 111 wind generation

App. 160

projects totaling over 20,000 MW and an additional7,470 MW of incremental wind development.Comments received on this subject also noted that windpower can be developed at relatively low costs and thatthe money the utilities currently spend on theimportation of coal and handling the byproducts of itscombustion would be better spent on construction ofadditional wind generating capacity.

Response: Renewable energy technology is not aretrofit option for the sources subject to BART and istherefore outside the scope of our SO2 BARTdetermination. We do generally acknowledge thatmany kinds of renewable energy do not produce haze-causing pollutants, and transitioning to those sourcesof energy could lead to visibility improvements.

Comment: We received opinions and data from fourcommenters expressing support for increased energyefficiency efforts as a technique for lowering powerdemand and therefore reducing the combustion of fossilfuels and its impact on the environment. One of thesecommenters noted that the affected utilities havebegun some energy efficiency programs and that withincreased effort they should be able to realize thesuccesses of other programs elsewhere in the countrythat have seen cumulative reductions in annual powerconsumption of 5–8 percent since 2004. The commenternotes that OG&E, in particular, should be able toreduce power demand by up to 1,200 GWh/year and2,100 GWh/year after five and ten years, respectively,at an annual reduction goal of one percent, or as muchas 1,800 GWh/year and 3,100 GWh/year after five andten years, respectively, at an annual reduction goal ofone and a half percent.

App. 161

Response: While not specifically within the scope byour SO2 BART determination or our approval of otheraspects of the state’s RH SIP, we acknowledge thatefficiency programs that reduce reliance on sources ofhaze-causing pollutants may promote visibilityimprovements.

Comment: OG&E states that if it is required todecide whether to install scrubbers or retire andreplace electric generating units with natural gas onroughly the same time frame, the economic analysissuggests that rate increases to customers will be lowerwith scrubbers. Installation of scrubbers is projected tocost more than $1.5 billion. OG&E is concerned thatwith this type of capital investment, it would be lockedeconomically into maximizing the use of its coal-firedunits for the foreseeable future. OG&E states theagreement outlined by ODEQ in the SIP (and rejectedby EPA) would reduce “the cumulative SO2 emissionsfrom Sooner Units 1 and 2 and Muskogee Units 4 and5 [to] approximately fifty-seven percent (57%) less thanwould be achieved through the installation andoperation of Dry FGD with SDA at all four (4) units.”OG&E states it should have the flexibility to takeadvantage of evolving technologies and to utilize theselocal clean energy sources at its plants in the future,while achieving the same (or better) reduction inimpact on visibility. OG&E states EPA’s failure toconsider these issues in the proposal is short-sighted,and arbitrary, capricious and contrary to applicablelaw.

Response: We find the approximately $1.2 billioncost claimed by OG&E in its BART analysis (referencedabove as $1.5 billion) for the installation of SO2 dry

App. 162

scrubbers is in error. As discussed elsewhere in ourresponse to comments and Supplemental RTC, basedon our Option 1 and Option 2 analyses, we find thetotal project costs to range between $290,418,007 to$299,400,007 for Sooner Units 1 and 2, and from$298,818,917 to 289,791,940 for Muskogee. Further, aswe also discuss in our proposal, although we based ourSO2 BART determination on the basis of dry SO2

scrubbers, OG&E is free to employ other technologiesto meet this limit, including switching to natural gas,as long as that switch is completed in the same BARTtimeframe. We discuss the BART compliance deadlinein the response to another comment.

Comment: A commenter stated we failed to consider“the costs of compliance” of converting the six coal-firedgenerating units to natural gas. Without anyexplanation, contends OIEC, we proposed that thesegenerating units could be converted to natural gas “asa means of satisfying their BART obligations.* * *” 76FR 16168, at 16194. The commenter states we failed toconsider the costs of compliance of conversion tonatural gas, as required by the CAA section 169A(g)(2),and the BART Guidelines, Part 51, Appendix.Y(IV)(D)(4)(a). The commenter states the FIP shouldtherefore be withdrawn.

Response: The commenter’s reference to ourproposal27 is fully reproduced as follows:

Should OG&E and/or AEP/PSO elect toreconfigure the above units to burn natural gas,as a means of satisfying their BART obligations

27 76 FR 16168, at 16194.

App. 163

under section 51.308(e), that conversion shouldbe completed by the same timeframe. We invitecomments as to, considering the engineeringand/or management challenges of such a fuelswitch, whether the full 5 years allowed undersection 308(e)(1)(iv) following the effective dateof our final rule would be appropriate.

Under the RHR,28 we cannot, and did not, evaluatethe costs associated with switching the six OG&E andAEP/PSO units over to natural gas for BART. However,after conducting the BART analysis and adopting ofemissions limits, alternatives to installing controltechnologies may achieve the same emission limits. Weare open to alternative mechanisms to achieve theBART emissions limits we adopted. As stated in ourproposal, we merely afforded OG&E and/or AEP/PSOthe opportunity to switch to natural gas as a means ofsatisfying BART. We also indicated we were willing toconsider comments to extend the BART compliancetimeframe to the full amount of time allowed under theRHR to accommodate that conversion. Although webased our BART determination of the use of SO2

scrubbers, the six units in question are free to considerany technology or alternative mechanism to meet their SO2 BART obligations.

J. Comments Arguing Our Proposal Would Hurt theEconomy and/or Raise Electricity Rates

Comment: Several commenters expressed concernabout adverse effects of electrical bill increases, stating

28 70 FR 39104, at 39164: “note that it is not our intent to directStates to switch fuel forms, e.g. from coal to gas.”

App. 164

that analyses prepared by the state’s utilities, businessgroups and the Oklahoma Corporation Commissionestimate our proposal could increase utility bills inOklahoma significantly, with some estimates as high as30 percent. Some commenters stated that the rateincrease would result in decreased business investmentin Oklahoma; while others stated that it will hurtexisting businesses, local governments, and familiesalready struggling from the recession. Severalcommenters noted that the rate increase will have adisproportionate adverse impact on senior citizens andthe disadvantaged, especially individuals living onfixed incomes. Commenters urged us to consider thecost implications of our proposal as we balance thegoals of the CAA with the economic impact onconsumers, communities, and businesses. Specifically,one commenter stated that installation of scrubbertechnologies on aging coal-fired facilities may not bethe most cost-effective or environmental approach.Several commenters ask EPA to consider all of thealternatives available, including switching to naturalgas over a longer timeframe. One commenter furtherstated that EPA’s proposal is not cost effective and doesnot significantly improve visibility. Commenters urgedEPA to adopt the Oklahoma State plan. A commenterthat supported the proposal stated that while the FIPcould cause rates to increase somewhat, Oklahoma hasthe eighth lowest average electricity rates in thecountry, rates are higher in neighboring states, and thedifference in rates may result from the fact that otherstates have emission controls on a higher percentage oftheir coal plants.

Response: The federal regulations implementing theCAA’s BART provisions require that we evaluate

App. 165

(1) cost of compliance, (2) the energy and non-airquality environmental impacts of compliance, (3) anyexisting pollution control technology in use at thesource, (4) remaining useful life of source, and(5) degree of improvement in visibility which mayreasonably be anticipated to result from the use of suchtechnology. 40 CFR 51.308(e)(1)(ii)(A). After a carefulcost review, we have determined that benefits invisibility from implementing our proposal outweigh theincrease in costs for the facilities. As discussed in ourproposal, we disagree with OG&E’s and AEP/PSO’scost estimate for installing scrubbers on the six unitsaddressed by our FIP. After careful review ofinformation provided during the public commentperiod, we revised our calculation of the total projectcost for the four OG&E units from our proposed rangeof approximately $312,423,000 to $605,685,000, to ourfinal range of approximately $589,237,000 to$607,461,000. We made no changes to the cost basis forthe two AEP/PSO units from our proposal. As such, theassociated cost investment for AEP/PSO is$274,100,000. In light of the visibility benefits wepredict will occur, we consider this to be cost effective.We take our duty to estimate the cost of controls veryseriously, and make every attempt to make athoughtful and well informed determination. We notethat our cost estimate, being about half that of OG&E’swill result in significantly less costs being passed on torate payers. We also note that our FIP allows for any ofthe six units to switch to natural gas within five yearsof this final action instead of installing the controltechnology.

App. 166

K. Comments Arguing Our Proposal Would Help theEconomy

Comments: We also received comments that theproposed FIP would help the economy in a variety ofways. One commenter stated that environmentalregulations like the RHR improve the economy andcreate jobs; and industry always finds a way to managethe cost of implementation. One commenter states thatcleaner air will boost Oklahoma’s productivity and jobcreation.

Response: Although, we did not consider thepotential positive benefits to local economics in makingour decision today, we do acknowledge that improvedvisibility may have a positive impact on tourism. Also,installing the controls required by the BARTdetermination on the six units will take three years orlonger to complete. These projects will require well-paid, skilled labor that can potentially be drawn fromthe local area, which would seem to benefit theeconomy.

Finally, as we have noted elsewhere in our responseto comments, although our action concerns visibilityimpairment, this action may also result in significantimprovements in human health. Improved humanhealth will reduce the healthcare costs and reduce thenumber of missed school and work days in thecommunity.

L. Comments on Health and Ecosystem Benefits andOther Pollutants

Comments: Several commenters state thatpollutants that cause visibility impairment also harm

App. 167

public health. Specifically, commenters assert thefollowing:

RH pollutants include NOX, SO2, PM,ammonia, and sulfuric acid. NOX is a precursorto ground level ozone, which is associated withrespiratory diseases, asthma attacks, anddecreased lung function. NOX also reacts withammonia, moisture, and other compounds toform particulates that can cause and worsenrespiratory disease, aggravate heart disease,and lead to premature death. Similarly, SO2

increases asthma symptoms, leads to increasedhospital visits, and can form particulates thataggravate respiratory and heart diseases andcause premature death. Both NOX and SO2 causeacid rain. PM can penetrate into the lungs andcause health problems, such as prematuremortality, lung disease, aggravated asthma,chronic bronchitis, and heart attacks.

Commenters cite to EPA’s estimates that in 2015,full implementation of the RHR nationally will prevent1,600 premature deaths, 2,200 non-fatal heart attacks,960 hospital admissions, and over 1 million lost schooland work days. The RHR will result in health benefitsvalued at $8.4 to $9.8 billion annually. More than100,000 children and 365,000 adults are diagnosedwith asthma in Oklahoma, and hospitalizations inOklahoma due to asthma cost roughly $57.9 million in2007 alone. Commenters also cite to a Clean Air TaskForce finding that the six units at issue in the proposedrule annually cause approximately 118 deaths, 181heart attacks, 2,037 asthma attacks, 86 hospital

App. 168

admissions, 74 cases of chronic bronchitis, and 129emergency room visits.

Some commenters also relay personal stories of thehealth impacts on themselves and their families fromthe emissions at issue. One commenter is disappointedthat the air quality in Oklahoma is so poor that theODEQ often warns active adults to avoid prolongedoutdoor exposure. She notes that ozone action daysprevent children from playing outside in the summer.Several children have been hospitalized due to asthmaand other illnesses that the commenters attribute tothe emissions at issue. One commenter contends thatmany people who are impacted by this rulemaking arenot aware of the rulemaking process, or their rightsunder that process. Commenters further state that it isEPA’s responsibility to protect the air quality andprevent these negative health effects.

Several commenters also assert that NOX and SO2

emissions from coal plants harms crops like pecans,barley, and oats, which puts the livelihoods of localfarmers at risk, impacts the health of those whoconsume the contaminated food, and increases the costof food.

Some commenters want this rulemaking to addresshealth issues. One commenter states that, while theRHR was designed to provide redress for visibilityimpairment, the BART Guidelines expressly providefor the consideration of non-air quality environmentalimpacts in step four of the five-step BART process. Thisconsideration includes the environmental impact onhuman health.

App. 169

One commenter states that the power plants havehad plenty of time to change operations to comply, butthey have failed to do so. Several commenters assertthat Oklahoma is unable to properly manage water andair pollution because special interest groups trumpscience. Another commenter states that coal pollutionis devastating tourism and wildlife in Oklahoma. Onecommenter states that cleaner air will improve thehealth of its citizens. Some commenters assert thatcustomers are subsidizing the cost of electricity withtheir health, lives, and livelihoods. One commenterstated that the increase in electricity costs is offset byreducing the healthcare costs to the community to treatillnesses and deaths caused by air pollution from theplants. Another commenter points out that powerplants are also built near the most vulnerable andunderserved populations in the state, based on theargument that the plants will bring needed jobs. Onecommenter concludes that it is unfair and unethical tohold citizens hostage to the idea that they must choosebetween electricity and good health. Severalcommenters feel that it is appropriate for industry tobear the burden of the cost, rather than pass it on tocitizens of the state in the form of healthcare costs.These commenters are amenable to paying higherelectricity rates in exchange for healthier air andwater. Several commenters request that EPA imposethe strongest possible regulation of emissions andenforcement of the CAA.

Another commenter notes that President Nixoncreated EPA to protect the environment and the CAAwas passed to protect air quality in our national parksand wilderness areas. President Reagan’s acid rainprogram cost less than industry or EPA estimated; and

App. 170

hopefully, installing scrubbers on these coal plants willalso cost less than estimated. Further, the CAA allowsEPA to limit sulfur oxides, nitrogen dioxides, organiccompounds, and particulates to ensure the quality ofthe air in the region. Several commenters state thatcoal pollutes throughout the process during extraction,burning, and disposal. One commenter states that thetrue cost of coal is the cost of its transportation,remediation of coal pollution, and lost tourism and badpublic relations in states where coal production occursthrough mountaintop removal. Many commentersrecommend that Oklahoma convert to more efficientsources of energy such as natural gas, wind, and solarpower.

One commenter asserts that he suffered from severechildhood asthma caused by allergies before the coal-fired power plants were built. He states that affordableelectricity from the plants allows him to keep hiswindows closed, thereby preventing allergens fromentering his home.

Response: We appreciate the commenters’ concernsregarding the negative health impacts of emissionsfrom the six units at issue. We agree that the sameNOX emissions that cause visibility impairment alsocontribute to the formation of ground-level ozone,which has been linked with respiratory problems,aggravated asthma, and even permanent lung damage.We also agree that SO2 emissions that cause visibilityimpairment also contribute to increased asthmasymptoms, lead to increased hospital visits, and canform particulates that aggravate respiratory and heartdiseases and cause premature death; and that bothNOX and SO2 cause acid rain. We agree that the same

App. 171

PM emissions that cause visibility impairment can beinhaled deep into lungs, which can cause respiratoryproblems, decreased lung function, aggravated asthma,bronchitis, and premature death. We agree that thesepollutants can have negative impacts on plants andecosystems, damaging plants, trees, and othervegetation, and reducing forest growth and crop yields,which could have a negative effect on species diversityin ecosystems. Therefore, although our action concernsvisibility impairment, we note the potential forsignificant improvements in human health and theecosystem.

The CAA states that the non-air qualityenvironmental impacts of compliance are aconsideration in determining BART. See CAA Section169A(g)(2). The BART Guidelines allow for theconsideration of non-air quality environmental impactsunder 40 CFR 51, Appendix Y(IV)(D)(j). See also, 70 FR39104, at 39169. However, this BART factor generallyis considered in order to determine if a control optionthat is otherwise technically feasible should beeliminated due to adverse environmental impacts. Suchimpacts could include solid or hazardous wastegeneration and discharges of polluted water as a resultof the control device. Although we may note potentialhealth benefits from the reduction of air pollutants dueto the installation of a BART control, we do notconsider them as part of the BART determination.While we received many comments concerning healthimpacts from the ongoing operations of BART-eligiblesources, we received no comments asserting that dryand wet scrubbers should be differentiated oreliminated as compliance options based on non-airquality environmental impacts.

App. 172

Although we appreciate the commenters’encouragement that we adopt even stricter standards,after considering all the comments we received, as wehave stated elsewhere in this notice, we believe thatthe standards proposed in our proposal establish BARTand will prevent visibility impairment from the sixunits.

Issues that the commenters raise about the effect ofEPA’s action on the cost of electricity are addressedelsewhere in this notice. Additionally, comments thatrecommend that the six units switch to natural gas orother sources of renewable energy are addressedelsewhere in this notice.

Comments: Several commenters note that coal-plantemissions contain other toxins including mercury, lead,cadmium, chromium, dioxins, formaldehyde, arsenic,radioactive isotopes, oxide, and radon gas. Anothercommenter is concerned that the toxicity of thepollutants in regional haze is higher in close proximityto the source of emissions.

Specifically, several commenters state that poorreclamation of coal ash from AEP’s Shady Point powerplant causes negative health impacts in Bokoshe,Oklahoma. These commenters are concerned about thehealth effects of fly ash because they state it containsarsenic, mercury, lead, cadmium, and other toxins.They describe the project as consisting of transportingcoal ash from the plant to an abandoned lead mine inBokoshe. Commenters claim that the result is a fiftyfoot wall of toxic coal ash at the reclamation site inBokoshe. Commenters state that pollution from thereclamation project has damaged property and people’shealth. They state that fugitive emissions from the

App. 173

trucks and the reclamation site run off into the groundwater, polluting drinking water supplies. Onecommenter also states that fly ash has been used inOklahoma as repair material for county roads.Commenters state that sixteen to twenty familiesliving nearby have cancer, children have asthma, andcalves in the area are stillborn. One commenter statesthat EPA’s proposal to put scrubbers on the units atissue will help address asthma, but these scrubberswill cause emissions of toxic fly ash.

Several commenters are concerned that themercury, chromium, and arsenic from the coal-firedpower plants are contaminating food, primarily fish.One commenter contends that these chemicals arecarcinogenic and bioaccumulate. As a result, they state,some fish in Oklahoma have high levels of toxicmaterials and cannot be consumed. Commenters notethat mercury contamination is so extreme that largerfish species are unsafe for pregnant women to eat. Onecommenter states that mercury is a neurotoxin thatnegatively affects a child’s ability to talk, walk, read,and learn. Several commenters point out that ODEQhas issued advisories that prohibit eating fish fromcertain lakes because the mercury content isdangerously high. One commenter further states thatsixteen out of fifty of the lakes in Oklahoma haveelevated levels of mercury.

Response: Although we appreciate the commenters’concerns regarding the potential negative healthimpacts from toxic emissions from the six units atissue, we note that we are not quantifying any toxicemissions that may be emitted, and such emissions arenot considered to be visibility impairing pollutants.

App. 174

Therefore, consideration of the toxic emissions isoutside the scope of this rulemaking under the RHR.However, please note that other provisions of the CAA,as well as other environmental statutes andregulations address toxic emissions, such as the onesnoted here. EPA implements such programs to protecthuman health and the environment from the negativeimpacts of these pollutants, and Oklahoma’s SIP isrequired to include provisions consistent with theseFederal requirements to the extent that they areapplicable.

Comment: One commenter mentions the impacts ofthe transport of emissions from existing and plannedcoal plants in Texas, stating that sixty percent ofmercury pollution in Oklahoma comes from Texas. Herequests that EPA accelerates mercury testing inOklahoma’s land and lakes.

Response: While we understand the commenter’sconcern with the impacts of transport emission fromTexas on water bodies in Oklahoma, mercury testing ofwater bodies is outside the scope of our action. Mercuryis not considered a visibility impairing pollutant; it isan air toxic regulated under CAA requirements thatare distinct from the RHR and CAA section110(a)(2)(D)(i)(II).

Comments: Several commenters discuss the impactof coal power on climate change. One commenter alsonotes that we should regulate CO2 because ninety-seven percent of scientists agree that it is causingclimate change. He contends that coal fired powerplants are contributing to climate change, stating thatthe CO2 level has risen from 280 ppm during the pre-industrial age to 380 ppm today. He cites the IPCC and

App. 175

others who state that the CO2 level should not exceed350 ppm. He also discusses the increasingtemperatures and potential for sea level rise in thenear future. The commenter states that we need toaddress climate change now.

Response: While we understand the commenters’concerns with respect to climate change, considerationof climate change is outside the scope of our action onthe RHR. While CO2 is a greenhouse gas (GHG), it isnot considered a visibility impairing pollutant.However, EPA implements regulations that addressGHGs in order to protect the public and theenvironment from the negative impacts of climatechange. Additionally, Oklahoma’s SIP is required toinclude provisions consistent with those Federalrequirements.

M. Miscellaneous Comments

Comment: OG&E states that we found a defect inOklahoma’s Long Term Strategy (LTS) becauseCENRAP modeling assumed the presumptive SO2

BART limit (0.15 lb/mmBtu) for OG&E’s Sooner andMuskogee facilities, which was not secured byOklahoma in its SIP. OG&E states we reasoned thatthe proposed FIP was necessary to cure these defects.OG&E asserts we may not pre-determine the BARTSO2 emissions limit based on assumptions made duringregional modeling, but the emissions limit should bedetermined based on the five statutory factors asapplied to an individual facility.

Further, OG&E states our reasoning with respect tothe Oklahoma LTS is in error. When setting reasonableprogress goals for their own Class I areas, OG&E

App. 176

states, the states are authorized to consider the samefive statutory factors that are used in determiningBART, including the costs of additional controls. OG&Estates that Oklahoma did not specify additional SO2

controls for the Sooner and Muskogee units as part ofOklahoma’s LTS for the Wichita Mountains. OG&Enotes that for Class I areas in other states, a statemust ensure that it has included in its LTS allmeasures needed to achieve its apportionment ofemission reduction obligations agreed upon through theregional planning process. 40 CFR 51.308(d)(2)(ii).OG&E states that ODEQ found that its LTS requiredno further controls for Oklahoma sources becauseemissions from Oklahoma were found (through theregional planning process) to impair visibility at allrelevant Class I areas other than Wichita Mountainsonly insignificantly. Thus, OG&E reasons, theOklahoma LTS is consistent with the agreementsreached during regional planning. OG&E states wefailed to justify, or explain, our basis for assuming thatthe regional planning process would have come to adifferent conclusion concerning Oklahoma’s impact onother states’ Class I areas if a different SO2 emissionrate had been assumed for the Sooner and Muskogeeunits in question.

Response: We disagree with OG&E’s assertion thatOklahoma’s decision not to require controls for the sixOG&E and AEP/PSO units is consistent with the RHrequirements for the LTS, section 51.308(d)(3)(ii),which requires:

Where other States cause or contribute toimpairment in a mandatory Class I Federalarea, the State must demonstrate that it has

App. 177

included in its implementation plan allmeasures necessary to obtain its share of theemission reductions needed to meet the progressgoal for the area. If the State has participated ina regional planning process, the State mustensure it has included all measures needed toachieve its apportionment of emission reductionobligations agreed upon through that process.

Oklahoma did engage in a regional planningprocess. This regional planning process included aforum in which state representatives built emissioninventories that assumed that specific pollution sourceswould be controlled to specific levels. This includedassumptions that the six OG&E and AEP/PSO unitswould be controlled to presumptive BART emissionlevels for SO2. Visibility modeling projectionssubsequently assumed those emission reductions.However, Oklahoma, in its subsequent RH SIP, did notinclude these promised reductions on which the otherstates are presently relying.

We note the CENRAP RPO process was open andrepresentatives from industry occasionally attendedCENRAP meetings and had an opportunity to engagein this process. ODEQ engaged in consultations under51.308(d)(3)(i), which requires that where the State hasemissions that are reasonably anticipated to contributeto visibility impairment in any mandatory Class IFederal area located in another State or States, theState must consult with the other State(s) in order todevelop coordinated emission management strategies.The State must consult with any other State havingemissions that are reasonably anticipated to contribute

App. 178

to visibility impairment in any mandatory Class IFederal area within the State.

All states that engaged in these consultations wereinvolved in the discussions leading up to, and theactual construction of the emission inventories and themodeling strategy. These LTS consultations thereforeassumed OG&E’s Sooner and Muskogee sources wouldbe controlled to the presumptive limit levels and madedecisions regarding whether additional controls toaddress LTS were needed on that basis. Thus, we aredisapproving Oklahoma’s LTS.

Furthermore, and notwithstanding the above LTSdiscussion, we disagree with OG&E’s assertion that ourBART analysis of the six OG&E and AEP/PSO units isdue to the CENRAP modeling. As we discussed in ourproposal, we arrived at our proposed BARTdetermination for the six units in question afterperforming the BART analysis required under theRHR.

Comment: AEP/PSO commented that we shouldclarify that new monitoring systems proposed undersection 52.1923(e) do not need to be installed for bothUnit 3 and Unit 4 of the Northeastern plant if the samefuel is used for both units. Instead, they reason, stackemissions should be apportioned to the units based onunit to stack load ratios. AEP/PSO claims theequipment necessary to report emissions for each unitindividually will add approximately $250,000 to thecost to comply, and provides no better data onemissions to the atmosphere.

Response: We are affirming that we are in factrequiring that the monitoring described in section

App. 179

52.1923(e) must be installed separately for each ofUnits 3 and 4 of the AEP/PSO Northeastern plant eventhough the same fuel is used for both units. We do notfind that it is proper to calculate the emissions of eachunit based on its load ratio, as individual SO2 scrubberswill likely have slightly different performancecharacteristics and we need to ensure that both units’scrubbers are working properly by monitoring theemissions unit by unit.

Comment: AEP/PSO believes there is a conflictbetween the language in section 52.1923(d) and (e).Section 52.1923(d) states that if a valid SO2 pounds perhour or heat input is not available for any hour for aunit, that heat input and SO2 pounds per hour shall notbe used in the calculation of the 30-day rolling averagefor SO2.

Section 52.1923(e) states that when valid SO2

pounds per hour, or SO2 pounds per million Btuemission data are not obtained because of continuousmonitoring system breakdowns, repairs, calibrationchecks, or zero and span adjustments, emission datamust be obtained by using other monitoring systemsapproved by the EPA to provide emission data for aminimum of 18 hours in each 24 hour period and atleast 22 out of 30 successive boiler operating days.

Response: We do not see a conflict between thelanguage in sections 52.1923(d) and (e). Paragraph (d)refers to short term, discrete data acquisition problemsand paragraph (e) refers to more serious problems thatmay arise due to fundamental underlying problemswith the monitoring system.

App. 180

Comment: One commenter called for an integratedand comprehensive strategy for EGUs to meet CAArequirements, noting that EGU emissions are subjectto the RHR, the PM2.5 NAAQS, and the NationalEmissions Standards for Hazardous Air Pollutants.The commenter stated that to effectively addressimpacts to human health and RH caused by EGUemissions, the FIP or SIP should require (1) SCR tocontrol NOX, (2) wet scrubbers to control SO2, and(3) wet electrostatic precipitators to controlcondensable particulate matter and acid mists. Thecommenter also asked us to reconsider our proposal toaccept ODEQ’s NOX BART determination, because(1) according to our proposal additional NOX reductionswould achieve significant improvement in visibilityover baseline, (2) Nitrate particulates from EGUs areprimarily responsible for the majority of visibilityimpairment during winter days, and (3) the full benefitof wet scrubber controls may not be achieved unlessBART controls on NOX is also required. Concerning SO2, the commenter expressed concern that theproposal would “approve” a dry scrubber system, alongwith an older electrostatic precipitator at the OG&ESooner facility that would achieve poor control of PM2.5 emissions. The commenter added that the proposedrule does not provided adequate information to allowthe public to understand and compare control measuresor to comprehend the extent of underperformance of PM2.5 controls.

Another commenter requested additional controlsand monitoring for ammonia and sulfuric acid.Specifically the commenter (1) requested that we setemission limits for ammonia and sulfuric acid mist,similar to those proposed for the San Juan Generating

App. 181

Station in New Mexico (76 FR 491), (2) stated theirsupport for requiring continuous emissions monitors tomonitor ammonia, and (3) urged us to require stacktesting for sulfuric acid on a more frequent basis thanannual monitoring.

Response: The purpose of our plan is to address theCAA BART requirements. Our evaluation found that:

• The NOX controls adopted by the state meet theCAA BART requirements;

• The SO2 BART controls we proposed in our FIP,in addition to the state adopted NOX controls, wouldlead to significant improvement in visibility and meetthe CAA BART requirements;

• Additional NOX controls would not be costeffective; and

• Additional pollutant controls are not needed tomeet the CAA BART requirements.

Regarding the request for ammonia and sulfuricacid mist emission limits and monitoring, we didpropose ammonia and sulfuric acid limits andmonitoring, as part of our New Mexico RH FIP for theSan Juan Generating Station. 76 FR 491. We did thisbecause we were concerned about the potential forammonia slip, as a result of the operation of SelectiveCatalytic Reduction (SCR), and the potential for thegrowth in sulfuric acid emissions if they were notlimited in an enforceable manner. As explained in ourresponse to comments in that action, we ultimatelydetermined that neither an ammonia limit, nor

App. 182

ammonia monitoring was warranted.29 We did,however, limit sulfuric acid emissions, verified byannual stack testing due to the potential for visibilityimpairment from increased sulfuric acid emissionsassociated with operation of SCR. These issues are notapplicable here, as our BART FIP is concerned with thereduction of SO2, which is not controlled by SCR, andour visibility modeling does not indicate the need tocontrol or monitor sulfuric acid or ammonia emissions.

Comment: One commenter stated that bymandating scrubbers on coal plants that we are tryingto phase out does not make sense. Another commenterasked why switching to low sulfur coal is notconsidered a viable alternative instead of mandatinginstallation of expensive wet gas scrubbers. A thirdcommenter stated that the EPA continues to bog downelectricity producers with burdensome paperwork andlegal uncertainty and that the EPA RHR is a perfectexample of the EPA’s lack of economic reality.

Response: We are not attempting to phase out theOklahoma coal plants that are subject to our FIP. Thepurpose of our FIP is to control SO2 emissions from sixOklahoma EGUs that contribute to RH in order to meetthe CAA BART requirements. To that end we aresetting emissions limits for SO2. We are not requiringcertain control technologies or fuel sources. Asdiscussed earlier, we used the CAA’s BART evaluationcriteria for our plan and found that it is reasonable andrealistic. The paperwork required will ensurecompliance with the BART FIP.

29 76 FR 52388, at 52407.

App. 183

Comment: One commenter expressed his view thatcitizens should ask EPA to set and enforce regulationsfor haze because the state regulations were inadequate.Another commenter stated that we should reject lowerstandards suggested by others.

Response: We agree with the commenter thatOklahoma’s RH SIP was inadequate in its control ofSO2 from the six OG&E and AEP/PSO units. We findthat our FIP will require the proper amount of SO2

control in order to comply with the RHR.

Comment: A request was submitted that we hold apublic hearing on our proposal in Tulsa, Oklahoma.

Response: Originally we scheduled one publichearing in Oklahoma City. In response to the requestwe added a second hearing in Tulsa on April 14, 2011.The transcripts of both public hearings are available inthe docket.

Comment: One commenter asked us to work withODEQ and the electrical power providers to develop acost effective plan.

Response: We find that the SO2 controls required byour FIP are, for the reasons discussed elsewhere in ourresponse to comments and Supplemental RTC, costeffective. We are, however, willing to work with ODEQand others to develop a SIP that could replace our FIP.Such a SIP will need to meet the CAA and EPA’s RHregulations and be consistent with EPA’s guidance.

Comment: One commenter supported our proposal’s(1) determination that Oklahoma’s SO2 BART limits donot meet the RH regulations, (2) analysis of thevisibility improvement resulting from BART controls,

App. 184

(3) determination that low NOX burners areappropriate as BART, and (4) determination thatexisting electrostatic precipitators and a0.1 lbs/MMBtu emissions limit is appropriate as BARTfor particulate matter.

Response: We appreciate the comments.

Comment: Comments were received expressingconcern over other sources of air pollution, such aslandfills, coal-fired power plants, the Tar Creeksuperfund site and sources in Texas.

Response: While we understand the commenter’sconcern with the impacts of other sources of pollution,the scope of this action is limited to assessing whethercertain elements of the Oklahoma RH SIP meet the RHrequirements of the CAA, including BART, andaddressing any deficiencies identified. We note alsothat other state and federal statutes and regulationsaddress other sources of air pollution, such as thosereferenced by the commenters, to protect human healthand the environment from the negative impacts ofthese pollutants.

Comment: Two commenters provided questions atthe Oklahoma City public hearing. Several questionsrelate to Class 1 areas, such as: designation of Class 1areas; location of Class 1 areas in relation to the sixunits and other coal-fired units; frequency, degree, andseason of visibility impact in Class 1 areas; andtourism at the Class 1 areas. Other questions concerncost of compliance by the six units, such as: annual andtotal cost; cost and benefit analysis of comparing thecost of compliance to “visitor impact days”; economicimpacts to the region; and EPA’s authority to

App. 185

implement the FIP. Finally, some questions concernthe Wichita Wildlife Refuge specifically andcontemplate sources of haze impacting that Class 1area, other than the six units.

Response: In general, answers to these questionsare: (1) Found in our proposal or in supportingdocuments for our proposal, (2) furnished in responseto other comments, or (3) not a necessary or relevantconsideration for our action. For responses to thesecomments, please see the “Addendum Responding toQuestions Received” available in the electronic docketfor this rulemaking.

Comment: We received comments not related to theproposal. These included comments on:

• Enforcement by EPA and ODEQ;

• A RH educational plan;

• Emissions from the LaFarge cement company;and

• Eliminating coal as a source of energy.

Response: While these and other comments may beimportant topics for discussion, we are not addressingthese topics as they are outside the scope of ourrulemaking.

IV. Statutory and Executive Order Reviews

A. Executive Order 12866: Regulatory Planning andReview and Executive Order 13563: ImprovingRegulation and Regulatory Review

This action finalizes a source-specific FIP for sixunits at coal-fired power plants in Oklahoma (OG&E

App. 186

Sooner Plant Units 1 and 2, OG&E Muskogee PlantUnits 4 and 5, and AEP/PSO Northeastern Plant Units3 and 4). This type of action is exempt from ExecutiveOrders 12866 (58 FR 51735, October 4, 1993) and13563 (76 FR 3821, January 21, 2011).

B. Paperwork Reduction Act

This action does not impose an informationcollection burden under the provisions of thePaperwork Reduction Act, 44 U.S.C. 3501 et seq.Burden is defined at 5 CFR 1320.3(b). Under thePaperwork Reduction Act, a “collection of information”is defined as a requirement for “answers to * * *identical reporting or recordkeeping requirementsimposed on ten or more persons * * * .” 44 U.S.C.3502(3)(A). Because the FIP only applies to six units atthree power plants (OG&E Sooner Plant, OG&EMuskogee Plant, and AEP/PSO Northeastern Plant)the Paperwork Reduction Act does not apply. See 5CFR 1320(c).

C. Regulatory Flexibility Act

The Regulatory Flexibility Act (RFA) generallyrequires an agency to prepare a regulatory flexibilityanalysis of any rule subject to notice and commentrulemaking requirements under the AdministrativeProcedure Act or any other statute unless the agencycertifies that the rule will not have a significanteconomic impact on a substantial number of smallentities. Small entities include small businesses, smallorganizations, and small governmental jurisdictions.

For purposes of assessing the impacts of today’s ruleon small entities, small entity is defined as: (1) A smallbusiness as defined by the Small Business

App. 187

Administration’s (SBA) regulations at 13 CFR 121.201;(2) a small governmental jurisdiction that is agovernment of a city, county, town, school district orspecial district with a population of less than 50,000;and (3) a small organization that is any not-for-profitenterprise which is independently owned and operatedand is not dominant in its field.

After considering the economic impacts of thisaction on small entities, I certify that this action willnot have a significant economic impact on a substantialnumber of small entities. The FIP for the OG&ESooner Plant, the Muskogee Plant, and the AEP/PSONortheastern Plant being finalized today does notimpose any new requirements on small entities. SeeMid-Tex Electric Cooperative, Inc. v. FERC, 773 F.2d327 (D.C. Cir. 1985).

D. Unfunded Mandates Reform Act

Title II of the Unfunded Mandates Reform Act of1995 (UMRA), 2 U.S.C. 1531–1538, requires Federalagencies, unless otherwise prohibited by law, to assessthe effects of their regulatory actions on state, local,and tribal governments and the private sector. Thisrule does not contain a Federal mandate that mayresult in expenditures of $100 million or more, adjustedfor inflation, for state, local, and tribal governments, inthe aggregate, or the private sector in any one year.Our cost estimate indicates that the total annual costof compliance with this rule is below this threshold.Thus, this rule is not subject to the requirements ofsections 202 or 205 of UMRA.

This rule is also not subject to the requirements ofsection 203 of UMRA because it contains no regulatory

App. 188

requirements that might significantly or uniquelyaffect small governments. This rule contains regulatoryrequirements that apply only to six units at coal-firedpower plants in Oklahoma (OG&E Sooner Plant Units1 and 2, OG&E Muskogee Plant Units 4 and 5, andAEP/PSO Northeastern Plant Units 3 and 4).

E. Executive Order 13132: Federalism

This action does not have federalism implications.It will not have substantial direct effects on the states,on the relationship between the national governmentand the states, or on the distribution of power andresponsibilities among the various levels ofgovernment, as specified in Executive Order 13132.This action merely prescribes EPA’s action to addressthe state not fully meeting its obligation to prohibitemissions from interfering with other states measuresto protect visibility. Thus, Executive Order 13132 doesnot apply to this action. In the spirit of Executive Order13132, and consistent with EPA policy to promotecommunications between EPA and state and localgovernments, EPA specifically solicited comment onthe proposed rule from state and local officials.

F. Executive Order 13175: Consultation andCoordination With Indian Tribal Governments

This final action does not have tribal implications,as specified in Executive Order 13175 (65 FR 67249,November 6, 2000), because the action EPA is takingneither imposes substantial direct compliance costs ontribal governments, nor preempts tribal law. Therefore,the requirements of section 5(b) and 5(c) of theExecutive Order do not apply to this rule. Consistentwith EPA policy, EPA nonetheless provided outreach to

App. 189

Oklahoma Tribes on several occasions in March andApril 2011, and offered consultation regarding thisaction. EPA did not receive any requests forconsultation on this rule.

G. Executive Order 13045: Protection of Children FromEnvironmental Health Risks and Safety Risks

EPA interprets EO 13045 (62 FR 19885, April 23,1997) as applying only to those regulatory actions thatconcern health or safety risks, such that the analysisrequired under section 5–501 of the EO has thepotential to influence the regulation. This action is notsubject to EO 13045 because it implements specificstandards established by Congress in statutes.

H. Executive Order 13211: Actions ConcerningRegulations That Significantly Affect Energy Supply,Distribution, or Use

This action is not subject to Executive Order 13211(66 FR 28355 (May 22, 2001)), because it is not asignificant regulatory action under Executive Order12866.

I. National Technology Transfer and Advancement Act

Section 12(d) of the National Technology Transferand Advancement Act of 1995 (“NTTAA”), Public Law104–113, 12(d) (15 U.S.C. 272 note) directs EPA to usevoluntary consensus standards in its regulatoryactivities unless to do so would be inconsistent withapplicable law or otherwise impractical. Voluntaryconsensus standards are technical standards (e.g.,materials specifications, test methods, samplingprocedures, and business practices) that are developedor adopted by voluntary consensus standards bodies.

App. 190

NTTAA directs EPA to provide Congress, throughOMB, explanations when the Agency decides not to useavailable and applicable voluntary consensusstandards. This rule would require the affected units atthe OG&E Sooner Plant, the Muskogee Plant, and theAEP/PSO Northeastern Plant to meet the applicablemonitoring requirements of 40 CFR part 75. Part 75already incorporates a number of voluntary consensusstandards. Consistent with the Agency’s PerformanceBased Measurement System (PBMS), Part 75 sets forthperformance criteria that allow the use of alternativemethods to the ones set forth in Part 75. The PBMSapproach is intended to be more flexible and costeffective for the regulated community; it is alsointended to encourage innovation in analyticaltechnology and improved data quality. At this time,EPA is not recommending any revisions to Part 75;however, EPA periodically revises the test proceduresset forth in Part 75. When EPA revises the testprocedures set forth in Part 75 in the future, EPA willaddress the use of any new voluntary consensusstandards that are equivalent. Currently, even if a testprocedure is not set forth in Part 75, EPA is notprecluding the use of any method, whether itconstitutes a voluntary consensus standard or not, aslong as it meets the performance criteria specified;however, any alternative methods must be approvedthrough the petition process under 40 CFR 75.66 beforethey are used.

App. 191

J. Executive Order 12898: Federal Actions To AddressEnvironmental Justice in Minority Populations andLow-Income Populations

Executive Order 12898 (59 FR 7629, February 16,1994), establishes federal executive policy onenvironmental justice. Its main provision directsfederal agencies, to the greatest extent practicable andpermitted by law, to make environmental justice partof their mission by identifying and addressing, asappropriate, disproportionately high and adversehuman health or environmental effects of theirprograms, policies, and activities on minoritypopulations and low-income populations in the UnitedStates.

EPA has determined that this rule will not havedisproportionately high and adverse human health orenvironmental effects on minority or low-incomepopulations because it increases the level ofenvironmental protection for all affected populationswithout having any disproportionately high andadverse human health or environmental effects on anypopulation, including any minority or low-incomepopulation. Our FIP limits emissions of SO2 from sixunits at coal-fired power plants in Oklahoma (OG&ESooner Plant Units 1 and 2, OG&E Muskogee PlantUnits 4 and 5, and AEP/PSO Northeastern Plant Units3 and 4). In addition to our FIP, we also approve SIPelements that also limit the emission of otherpollutants, including PM and NOX.

K. Congressional Review Act

The Congressional Review Act, 5 U.S.C. 801 et seq.,as added by the Small Business Regulatory

App. 192

Enforcement Fairness Act of 1996, generally providesthat before a rule may take effect, the agencypromulgating the rule must submit a rule report, whichincludes a copy of the rule, to each House of theCongress and to the Comptroller General of the UnitedStates. EPA will submit a report containing this actionand other required information to the U.S. Senate, theU.S. House of Representatives, and the ComptrollerGeneral of the United States prior to publication of therule in the Federal Register. A major rule cannottake effect until 60 days after it is published in theFederal Register. This action is not a “major rule” asdefined by 5 U.S.C. 804(2). This rule will be effective onJanuary 27, 2012.

L. Judicial Review

Under section 307(b)(1) of the CAA, petitions forjudicial review of this action must be filed in the UnitedStates Court of Appeals for the appropriate circuit byFebruary 27, 2012. Pursuant to CAA section307(d)(1)(B), this action is subject to the requirementsof CAA section 307(d) as it promulgates a FIP underCAA section 110(c). Filing a petition for reconsiderationby the Administrator of this final rule does not affectthe finality of this action for the purposes of judicialreview nor does it extend the time within which apetition for judicial review may be filed, and shall notpostpone the effectiveness of such rule or action. Thisaction may not be challenged later in proceedings toenforce its requirements. See CAA section 307(b)(2).

List of Subjects in 40 CFR Part 52

Air pollution control, Environmental protection,Best available retrofit technology, Incorporation by

App. 193

reference, Intergovernmental relations, Interstatetransport of pollution, Nitrogen dioxide, Ozone,Particulate matter, Regional haze, Reporting andrecordkeeping requirements, Sulfur dioxide, Visibility.

Dated: December 13, 2011.

Lisa P. Jackson, Administrator.

For the reasons set out in the preamble, title 40,chapter I, of the Code of Federal Regulations isamended as follows:

PART 52—[AMENDED]

# 1. The authority citation for part 52 continues to readas follows:

Authority: 42 U.S.C. 7401 et seq.

Subpart LL—[Amended]

# 2. Section 52.1920 is amended as follows: # a. The table in paragraph (c) is amended by adding insequential order under “Subchapter 8. Permits for Part70 Sources” a new heading for part 11 and a new entryfor “(252:100:8–70 to 252:100:8–77)”. # b. The first table in paragraph (e) is amended byadding at the end a new entry for “Interstate transportfor the 1997 ozone and PM2.5 NAAQS(Noninterference with measures required to preventsignificant deterioration of air quality or to protectvisibility in any other State)”, immediately followed byan entry for “Regional haze SIP”. “ # c. The second table in paragraph (e) entitled “EPAApproved Statutes in the Oklahoma SIP” is amendedby removing the entry for “Interstate transport for the

App. 194

1997 ozone and PM2.5 NAAQS.”

The amendments read as follows:

§ 52.1920 Identification of plan. * * * * *

(c) * * *

App. 195

EP

A A

PP

RO

VE

D O

KL

AH

OM

A R

EG

UL

AT

ION

S

Sta

te c

itat

ion

Tit

le/

subj

ect

Sta

teef

fect

ive

date

EP

A a

ppro

val d

ate

Exp

lan

atio

n

**

**

*

PA

RT

11.

Vis

ibil

ity

Pro

tect

ion

Sta

nda

rds

(252

:100

:8–7

0to 25

2:10

0:8–

77)

Vis

ibil

ity

Pro

tect

ion

Sta

nda

rds.

6/15

/200

712

/28/

11 [

Inse

rtF

R p

age

nu

mbe

rw

her

e do

cum

ent

begi

ns]

App. 196

(e)

* *

*

EP

A A

PP

RO

VE

D N

ON

-RE

GU

LA

TO

RY

PR

OV

ISIO

NS

AN

D Q

UA

SI-

RE

GU

LA

TO

RY

ME

AS

UR

ES

IN T

HE

OK

LA

HO

MA

SIP

Nam

e of

SIP

prov

isio

nA

ppli

cabl

ege

ogra

phic

or

non

-at

tain

men

tar

ea

Sta

tesu

bmit

tal/

effe

ctiv

e da

te

EP

A a

ppro

val

date

Exp

lan

atio

n

**

**

*

Inte

rsta

tetr

ansp

ort

for

the

1997

ozo

ne

and

PM

2.5

NA

AQ

S(N

onin

terf

eren

cew

ith

mea

sure

sre

quir

ed t

opr

even

t

Sta

tew

ide

5/1/

2007

11/2

6/20

10, 7

5F

R 7

2701

12/2

8/11

[In

sert

cit

atio

nof

pu

blic

atio

n].

Non

inte

rfer

ence

wit

h m

easu

res

requ

ired

to

prev

ent

sign

ific

ant

dete

rior

atio

n o

fai

r qu

alit

y in

an

yot

her

Sta

te

App. 197

prev

ent

sign

ific

ant

dete

rior

atio

n o

fai

r qu

alit

y or

to

prot

ect

visi

bili

ty in

an

yot

her

Sta

te).

oth

er S

tate

appr

oved

11

/26/

2010

.N

onin

terf

eren

cew

ith

mea

sure

sre

quir

ed t

opr

otec

t vi

sibi

lity

in a

ny

oth

erS

tate

par

tial

lyap

prov

ed12

/28/

11.

Reg

ion

al h

aze

SIP

:S

tate

wid

e2/

17/2

010

12/2

8/11

[In

sert

cit

atio

nof

pu

blic

atio

n].

Cor

ere

quir

emen

ts o

f40

CF

R 5

1.30

8

App. 198

(a)

Det

erm

inat

ion

of b

asel

ine

and

nat

ura

lvi

sibi

lity

con

diti

ons.

(b)

Coo

rdin

atin

gre

gion

al h

aze

and

reas

onab

lyat

trib

uta

ble

visi

bili

tyim

pair

men

t.

App. 199

c) M

onit

orin

gst

rate

gy a

nd

oth

erim

plem

enta

tion

requ

irem

ents

.

(d)

Coo

rdin

atio

nw

ith

Sta

tes

and

Fed

eral

Lan

dM

anag

ers.

(e)

BA

RT

dete

rmin

atio

ns

exce

pt f

or t

he

foll

owin

g S

O2

BA

RT

dete

rmin

atio

ns:

Un

its

4 an

d 5

ofth

e O

klah

oma

Gas

an

dE

lect

ric

(OG

&E

)

App. 200

Mu

skog

eepl

ant;

Un

its

1an

d 2

of t

he

OG

&E

Soo

ner

plan

t; a

nd

Un

its

3 an

d 4

of t

he

Am

eric

anE

lect

ric

Pow

er/P

ubl

icS

ervi

ceC

ompa

ny

ofO

klah

oma

(AE

P/P

SO

)N

orth

east

ern

plan

t.

App. 201

# 3. Section 52.1923 is added to read as follows:

§ 52.1923 Best Available Retrofit Requirements(BART) for SO2 and Interstate pollutanttransport provisions; What are the FIPrequirements for Units 4 and 5 of the OklahomaGas and Electric Muskogee plant; Units 1 and 2 ofthe Oklahoma Gas and Electric Sooner plant; andUnits 3 and 4 of the American ElectricPower/Public Service Company of OklahomaNortheastern plant affecting visibility?

(a) Applicability. The provisions of this sectionshall apply to each owner or operator, or successiveowners or operators, of the coal burning equipmentdesignated as: Units 4 or 5 of the Oklahoma Gas andElectric Muskogee plant; Units 1 or 2 of the OklahomaGas and Electric Sooner plant; and Units 3 or 4 of theAmerican Electric Power/Public Service Company ofOklahoma Northeastern plant.

(b) Compliance Dates. Compliance with therequirements of this section is required within fiveyears of the effective date of this rule unless otherwiseindicated by compliance dates contained in specificprovisions.

(c) Definitions. All terms used in this part but notdefined herein shall have the meaning given them inthe CAA and in parts 51 and 60 of this title. For thepurposes of this section:

24-hour period means the period of time between12:01 a.m. and 12 midnight.

Air pollution control equipment includes selectivecatalytic control units, baghouses, particulate or

App. 202

gaseous scrubbers, and any other apparatus utilized tocontrol emissions of regulated air contaminants thatwould be emitted to the atmosphere.

Boiler-operating-day means any 24-hour periodbetween 12:00 midnight and the following midnightduring which any fuel is combusted at any time at thesteam generating unit.

Daily average means the arithmetic average of thehourly values measured in a 24-hour period.

Heat input means heat derived from combustion offuel in a unit and does not include the heat input frompreheated combustion air, recirculated flue gases, orexhaust gases from other sources. Heat input shall becalculated in accordance with 40 CFR part 75.

Owner or Operator means any person who owns,leases, operates, controls, or supervises any of the coalburning equipment designated as:

Unit 4 of the Oklahoma Gas and Electric Muskogeeplant; or

Unit 5 of the Oklahoma Gas and Electric Muskogeeplant; or

Unit 1 of the Oklahoma Gas and Electric Soonerplant; or

Unit 2 of the Oklahoma Gas and Electric Soonerplant; or

Unit 3 of the American Electric Power/PublicService Company of Oklahoma Northeasternplant; or

App. 203

Unit 4 of the American Electric Power/PublicService Company of Oklahoma Northeasternplant.

Regional Administrator means the RegionalAdministrator of EPA Region 6 or his/her authorizedrepresentative.

Unit means one of the coal fired boilers coveredunder Paragraph (a), above.

(d) Emissions Limitations.

SO2 emission limit. The individual sulfur dioxideemission limit for a unit shall be 0.06 pounds permillion British thermal units (lb/MMBtu) as averagedover a rolling 30 boiler-operating-day period. For eachunit, SO2 emissions for each calendar day shall bedetermined by summing the hourly emissionsmeasured in pounds of SO2. For each unit, heat inputfor each boiler-operating-day shall be determined byadding together all hourly heat inputs, in millions ofBTU. Each boiler-operating-day the thirty-day rollingaverage for a unit shall be determined by addingtogether the pounds of SO2 from that day and thepreceding 29 boiler-operating-days and dividing thetotal pounds of SO2 by the sum of the heat input duringthe same 30 boiler-operating-day period. The resultshall be the 30 boiler-operating-day rolling average interms of lb/MMBtu emissions of SO2. If a valid SO2

pounds per hour or heat input is not available for anyhour for a unit, that heat input and SO2 pounds perhour shall not be used in the calculation of the 30boiler-operating-day rolling average for SO2.

App. 204

(e) Testing and monitoring.

(1) No later than the compliance date of thisregulation, the owner or operator shall install,calibrate, maintain and operate Continuous EmissionsMonitoring Systems (CEMS) for SO2 on Units 4 and 5of the Oklahoma Gas and Electric Muskogee plant;Units 1 and 2 of the Oklahoma Gas and Electric Soonerplant; and Units 3 and 4 of the American ElectricPower/Public Service Company of OklahomaNortheastern plant in accordance with 40 CFR 60.8and 60.13(e), (f), and (h), and Appendix B of Part 60.The owner or operator shall comply with the qualityassurance procedures for CEMS found in 40 CFR part75. Compliance with the emission limits for SO2 shallbe determined by using data from a CEMS.

(2) Continuous emissions monitoring shall applyduring all periods of operation of the coal burningequipment, including periods of startup, shutdown, andmalfunction, except for CEMS breakdowns, repairs,calibration checks, and zero and span adjustments.Continuous monitoring systems for measuring SO2 anddiluent gas shall complete a minimum of one cycle ofoperation (sampling, analyzing, and data recording) foreach successive 15-minute period. Hourly averagesshall be computed using at least one data point in eachfifteen minute quadrant of an hour. Notwithstandingthis requirement, an hourly average may be computedfrom at least two data points separated by a minimumof 15 minutes (where the unit operates for more thanone quadrant in an hour) if data are unavailable as aresult of performance of calibration, quality assurance,preventive maintenance activities, or backups of datafrom data acquisition and handling system, and

App. 205

recertification events. When valid SO2 pounds per hour,or SO2 pounds per million Btu emission data are notobtained because of continuous monitoring systembreakdowns, repairs, calibration checks, or zero andspan adjustments, emission data must be obtained byusing other monitoring systems approved by the EPAto provide emission data for a minimum of 18 hours ineach 24 hour period and at least 22 out of 30 successiveboiler operating days.

(f) Reporting and Recordkeeping Requirements.Unless otherwise stated all requests, reports,submittals, notifications, and other communications tothe Regional Administrator required by this sectionshall be submitted, unless instructed otherwise, to theDirector, Multimedia Planning and PermittingDivision, U.S. Environmental Protection Agency,Region 6, to the attention of Mail Code: 6PD, at 1445Ross Avenue, Suite 1200, Dallas, Texas 75202–2733.For each unit subject to the emissions limitation in thissection and upon completion of the installation ofCEMS as required in this section, the owner oroperator shall comply with the following requirements:

(1) For each emissions limit in this section, complywith the notification, reporting, and recordkeepingrequirements for CEMS compliance monitoring in 40CFR 60.7(c) and (d).

(2) For each day, provide the total SO2 emitted thatday by each emission unit. For any hours on any unitwhere data for hourly pounds or heat input is missing,identify the unit number and monitoring device thatdid not produce valid data that caused the missinghour.

App. 206

(g) Equipment Operations. At all times, includingperiods of startup, shutdown, and malfunction, theowner or operator shall, to the extent practicable,maintain and operate the unit including associated airpollution control equipment in a manner consistentwith good air pollution control practices for minimizingemissions. Determination of whether acceptableoperating and maintenance procedures are being usedwill be based on information available to the RegionalAdministrator which may include, but is not limited to,monitoring results, review of operating andmaintenance procedures, and inspection of the unit.

(h) Enforcement.

(1) Notwithstanding any other provision in thisimplementation plan, any credible evidence orinformation relevant as to whether the unit would havebeen in compliance with applicable requirements if theappropriate performance or compliance test had beenperformed, can be used to establish whether or not theowner or operator has violated or is in violation of anystandard or applicable emission limit in the plan.

(2) Emissions in excess of the level of the applicableemission limit or requirement that occur due to amalfunction shall constitute a violation of theapplicable emission limit.

# 4. Section 52.1928 is added to read as follows:

§ 52.1928 Visibility protection.

(a) The following portions of the Oklahoma RegionalHaze (RH) State Implementation Plan submitted onFebruary 19, 2010 are disapproved:

App. 207

(1) The SO2 BART determinations for Units 4 and5 of the Oklahoma Gas and Electric (OG&E) Muskogeeplant; Units 1 and 2 of the OG&E Sooner plant; andUnits 3 and 4 of the American Electric Power/PublicService Company of Oklahoma (AEP/PSO)Northeastern plant;

(2) The long-term strategy for regional haze;

(3) “Greater Reasonable Progress AlternativeDetermination” (section VI.E), and

(4) Separate executed agreements between ODEQand OG&E, and ODEQ and AEP/PSO entitled “OG&ERegional Haze Agreement, Case No. 10–024, and “PSORegional Haze Agreement, Case No. 10–025,” housedwithin Appendix 6–5 of the RH SIP.

(b) The portion of the State Implementation Planpertaining to adequate provisions to prohibit emissionsfrom interfering with measures required in anotherstate to protect visibility, submitted on May 10, 2007and supplemented on December 10, 2007 isdisapproved.

(c) The SO2 BART requirements for Units 4 and 5of the Oklahoma Gas and Electric (OG&E) Muskogeeplant; Units 1 and 2 of the OG&E Sooner plant; andUnits 3 and 4 of the American Electric Power/PublicService Company of Oklahoma (AEP/PSO)Northeastern plant, the deficiencies in the long-termstrategy for regional haze, and the requirement for aplan to contain adequate provisions to prohibitemissions from interfering with measures required inanother state to protect visibility are satisfied by § 52.1923.

App. 208

[FR Doc. 2011–32572 Filed 12–27–11; 8:45 am]

BILLING CODE 6560–50–P

App. 209

APPENDIX C

UNITED STATES COURT OF APPEALSFOR THE TENTH CIRCUIT

[Filed October 31, 2013]

No. 12-9526_______________________________________STATE OF OKLAHOMA, et al., )

Petitioners, ))

v. ))

UNITED STATES ENVIRONMENTAL )PROTECTION AGENCY, )

Respondent. )--------------------------------------------------------- )SIERRA CLUB, )

Intervenor - Respondent, ))

and ))

PACIFICORP, et al., )Amici Curiae. )

_______________________________________)

No. 12-9527_______________________________________OKLAHOMA GAS & ELECTRIC ) COMPANY, )

Petitioner, ))

v. )

App. 210

)UNITED STATES ENVIRONMENTAL )PROTECTION AGENCY, )

Respondent. )--------------------------------------------------------- )SIERRA CLUB, )

Intervenor - Respondent, ))

and ))

PACIFICORP, et al., )Amici Curiae. )

_______________________________________)

ORDER

Before BRISCOE, Chief Judge, KELLY, andLUCERO, Circuit Judges.

Petitioners’ petition for panel rehearing in 12-9526is denied. Petitioner’s petition for panel rehearing in12-9527 is also denied. Judge Kelly would grant panelrehearing in both cases, consistent with his concurringand dissenting opinion.

The petitions for rehearing en banc weretransmitted to all of the judges of the court who are inregular active service. As no member of the panel andno judge in regular active service on the courtrequested that the court be polled, those petitions arealso denied.

Entered for the Court

/s/ Elisabeth A. ShumakerELISABETH A. SHUMAKER, Clerk

App. 211

APPENDIX D

42 U.S.C.A. § 7491

§ 7491. Visibility protection for Federal class I areas

Currentness

(a) Impairment of visibility; list of areas; study andreport

(1) Congress hereby declares as a national goalthe prevention of any future, and the remedying ofany existing, impairment of visibility in mandatoryclass I Federal areas which impairment resultsfrom manmade air pollution.

(2) Not later than six months after August 7,1977, the Secretary of the Interior in consultationwith other Federal land managers shall review allmandatory class I Federal areas and identify thosewhere visibility is an important value of the area.From time to time the Secretary of the Interior mayrevise such identifications. Not later than one yearafter August 7, 1977, the Administrator shall, afterconsultation with the Secretary of the Interior,promulgate a list of mandatory class I Federal areasin which he determines visibility is an importantvalue.

(3) Not later than eighteen months after August7, 1977, the Administrator shall complete a studyand report to Congress on available methods forimplementing the national goal set forth in

App. 212

paragraph (1). Such report shall includerecommendations for--

(A) methods for identifying, characterizing,determining, quantifying, and measuringvisibility impairment in Federal areas referredto in paragraph (1), and

(B) modeling techniques (or other methods)for determining the extent to which manmadeair pollution may reasonably be anticipated tocause or contribute to such impairment, and

(C) methods for preventing and remedyingsuch manmade air pollution and resultingvisibility impairment.

Such report shall also identify the classes or categoriesof sources and the types of air pollutants which, aloneor in conjunction with other sources or pollutants, mayreasonably be anticipated to cause or contributesignificantly to impairment of visibility.

(4) Not later than twenty-four months afterAugust 7, 1977, and after notice and public hearing,the Administrator shall promulgate regulations toassure (A) reasonable progress toward meeting thenational goal specified in paragraph (1), and (B)compliance with the requirements of this section.

(b) Regulations

Regulations under subsection (a)(4) of this sectionshall--

(1) provide guidelines to the States, taking intoaccount the recommendations under subsection(a)(3) of this section on appropriate techniques and

App. 213

methods for implementing this section (as providedin subparagraphs (A) through (C) of such subsection(a)(3)), and

(2) require each applicable implementation planfor a State in which any area listed by theAdministrator under subsection (a)(2) of this sectionis located (or for a State the emissions from whichmay reasonably be anticipated to cause orcontribute to any impairment of visibility in anysuch area) to contain such emission limits,schedules of compliance and other measures as maybe necessary to make reasonable progress towardmeeting the national goal specified in subsection (a)of this section, including--

(A) except as otherwise provided pursuant tosubsection (c) of this section, a requirement thateach major stationary source which is inexistence on August 7, 1977, but which has notbeen in operation for more than fifteen years asof such date, and which, as determined by theState (or the Administrator in the case of a planpromulgated under section 7410(c) of this title)emits any air pollutant which may reasonably beanticipated to cause or contribute to anyimpairment of visibility in any such area, shallprocure, install, and operate, as expeditiously aspracticable (and maintain thereafter) the bestavailable retrofit technology, as determined bythe State (or the Administrator in the case of aplan promulgated under section 7410(c) of thistitle) for controlling emissions from such sourcefor the purpose of eliminating or reducing anysuch impairment, and

App. 214

(B) a long-term (ten to fifteen years) strategyfor making reasonable progress toward meetingthe national goal specified in subsection (a) ofthis section.

In the case of a fossil-fuel fired generatingpowerplant having a total generating capacity inexcess of 750 megawatts, the emissionlimitations required under this paragraph shallbe determined pursuant to guidelines,promulgated by the Administrator underparagraph (1).

(c) Exemptions

(1) The Administrator may, by rule, after noticeand opportunity for public hearing, exempt anymajor stationary source from the requirement ofsubsection (b)(2)(A) of this section, upon hisdetermination that such source does not or will not,by itself or in combination with other sources, emitany air pollutant which may reasonably beanticipated to cause or contribute to a significantimpairment of visibility in any mandatory class IFederal area.

(2) Paragraph (1) of this subsection shall not beapplicable to any fossil-fuel fired powerplant withtotal design capacity of 750 megawatts or more,unless the owner or operator of any such plantdemonstrates to the satisfaction of theAdministrator that such powerplant is located atsuch distance from all areas listed by theAdministrator under subsection (a)(2) of this sectionthat such powerplant does not or will not, by itselfor in combination with other sources, emit any air

App. 215

pollutant which may reasonably be anticipated tocause or contribute to significant impairment ofvisibility in any such area.

(3) An exemption under this subsection shall beeffective only upon concurrence by the appropriateFederal land manager or managers with theAdministrator’s determination under thissubsection.

(d) Consultations with appropriate Federal landmanagers

Before holding the public hearing on the proposedrevision of an applicable implementation plan to meetthe requirements of this section, the State (or theAdministrator, in the case of a plan promulgated undersection 7410(c) of this title) shall consult in person withthe appropriate Federal land manager or managersand shall include a summary of the conclusions andrecommendations of the Federal land managers in thenotice to the public.

(e) Buffer zones

In promulgating regulations under this section, theAdministrator shall not require the use of anyautomatic or uniform buffer zone or zones.

(f) Nondiscretionary duty

For purposes of section 7604(a)(2) of this title, themeeting of the national goal specified in subsection(a)(1) of this section by any specific date or dates shallnot be considered a “nondiscretionary duty” of theAdministrator.

App. 216

(g) Definitions

For the purpose of this section--

(1) in determining reasonable progress thereshall be taken into consideration the costs ofcompliance, the time necessary for compliance, andthe energy and nonair quality environmentalimpacts of compliance, and the remaining useful lifeof any existing source subject to such requirements;

(2) in determining best available retrofittechnology the State (or the Administrator indetermining emission limitations which reflect suchtechnology) shall take into consideration the costs ofcompliance, the energy and nonair qualityenvironmental impacts of compliance, any existingpollution control technology in use at the source, theremaining useful life of the source, and the degreeof improvement in visibility which may reasonablybe anticipated to result from the use of suchtechnology;

(3) the term “manmade air pollution” means airpollution which results directly or indirectly fromhuman activities;

(4) the term “as expeditiously as practicable”means as expeditiously as practicable but in noevent later than five years after the date of approvalof a plan revision under this section (or the date ofpromulgation of such a plan revision in the case ofaction by the Administrator under section 7410(c) ofthis title for purposes of this section);

App. 217

(5) the term “mandatory class I Federal areas”means Federal areas which may not be designatedas other than class I under this part;

(6) the terms “visibility impairment” and“impairment of visibility” shall include reduction invisual range and atmospheric discoloration; and

(7) the term “major stationary source” means thefollowing types of stationary sources with thepotential to emit 250 tons or more of any pollutant:fossil-fuel fired steam electric plants of more than250 million British thermal units per hour heatinput, coal cleaning plants (thermal dryers), kraftpulp mills, Portland Cement plants, primary zincsmelters, iron and steel mill plants, primaryaluminum ore reduction plants, primary coppersmelters, municipal incinerators capable ofcharging more than 250 tons of refuse per day,hydrofluoric, sulfuric, and nitric acid plants,petroleum refineries, lime plants, phosphate rockprocessing plants, coke oven batteries, sulfurrecovery plants, carbon black plants (furnaceprocess), primary lead smelters, fuel conversionplants, sintering plants, secondary metal productionfacilities, chemical process plants, fossil-fuel boilersof more than 250 million British thermal units perhour heat input, petroleum storage and transferfacilities with a capacity exceeding 300,000 barrels,taconite ore processing facilities, glass fiberprocessing plants, charcoal production facilities.

App. 218

40 C.F.R. § 51.308

§ 51.308 Regional haze program requirements.

Effective: August 6, 2012Currentness

(a) What is the purpose of this section? This sectionestablishes requirements for implementation plans,plan revisions, and periodic progress reviews to addressregional haze.

(b) When are the first implementation plans dueunder the regional haze program? Except as providedin § 51.309(c), each State identified in § 51.300(b)(3)must submit, for the entire State, an implementationplan for regional haze meeting the requirements ofparagraphs (d) and (e) of this section no later thanDecember 17, 2007.

(c) [Reserved]

(d) What are the core requirements for theimplementation plan for regional haze? The State mustaddress regional haze in each mandatory Class IFederal area located within the State and in eachmandatory Class I Federal area located outside theState which may be affected by emissions from withinthe State. To meet the core requirements for regionalhaze for these areas, the State must submit animplementation plan containing the following planelements and supporting documentation for allrequired analyses:

(1) Reasonable progress goals. For eachmandatory Class I Federal area located within theState, the State must establish goals (expressed in

App. 219

deciviews) that provide for reasonable progresstowards achieving natural visibility conditions. Thereasonable progress goals must provide for animprovement in visibility for the most impaireddays over the period of the implementation plan andensure no degradation in visibility for the leastimpaired days over the same period.

(i) In establishing a reasonable progress goal forany mandatory Class I Federal area within theState, the State must:

(A) Consider the costs of compliance, the timenecessary for compliance, the energy and non-airquality environmental impacts of compliance,and the remaining useful life of any potentiallyaffected sources, and include a demonstrationshowing how these factors were taken intoconsideration in selecting the goal.

(B) Analyze and determine the rate ofprogress needed to attain natural visibilityconditions by the year 2064. To calculate thisrate of progress, the State must comparebaseline visibility conditions to natural visibilityconditions in the mandatory Federal Class Iarea and determine the uniform rate of visibilityimprovement (measured in deciviews) thatwould need to be maintained during eachimplementation period in order to attain naturalvisibility conditions by 2064. In establishing thereasonable progress goal, the State mustconsider the uniform rate of improvement invisibility and the emission reduction measuresneeded to achieve it for the period covered by theimplementation plan.

App. 220

(ii) For the period of the implementation plan, ifthe State establishes a reasonable progress goalthat provides for a slower rate of improvement invisibility than the rate that would be needed toattain natural conditions by 2064, the State mustdemonstrate, based on the factors in paragraph(d)(1)(i)(A) of this section, that the rate of progressfor the implementation plan to attain naturalconditions by 2064 is not reasonable; and that theprogress goal adopted by the State is reasonable.The State must provide to the public for review aspart of its implementation plan an assessment ofthe number of years it would take to attain naturalconditions if visibility improvement continues at therate of progress selected by the State as reasonable.

(iii) In determining whether the State’s goal forvisibility improvement provides for reasonableprogress towards natural visibility conditions, theAdministrator will evaluate the demonstrationsdeveloped by the State pursuant to paragraphs(d)(1)(i) and (d)(1)(ii) of this section.

(iv) In developing each reasonable progress goal,the State must consult with those States which mayreasonably be anticipated to cause or contribute tovisibility impairment in the mandatory Class IFederal area. In any situation in which the Statecannot agree with another such State or group ofStates that a goal provides for reasonable progress,the State must describe in its submittal the actionstaken to resolve the disagreement. In reviewing theState’s implementation plan submittal, theAdministrator will take this information intoaccount in determining whether the State’s goal for

App. 221

visibility improvement provides for reasonableprogress towards natural visibility conditions.

(v) The reasonable progress goals established bythe State are not directly enforceable but will beconsidered by the Administrator in evaluating theadequacy of the measures in the implementationplan to achieve the progress goal adopted by theState.

(vi) The State may not adopt a reasonableprogress goal that represents less visibilityimprovement than is expected to result fromimplementation of other requirements of the CAAduring the applicable planning period.

(2) Calculations of baseline and natural visibilityconditions. For each mandatory Class I Federalarea located within the State, the State mustdetermine the following visibility conditions(expressed in deciviews):

(i) Baseline visibility conditions for the mostimpaired and least impaired days. The period forestablishing baseline visibility conditions is 2000 to2004. Baseline visibility conditions must becalculated, using available monitoring data, byestablishing the average degree of visibilityimpairment for the most and least impaired daysfor each calendar year from 2000 to 2004. Thebaseline visibility conditions are the average ofthese annual values. For mandatory Class I Federalareas without onsite monitoring data for 2000-2004,the State must establish baseline values using themost representative available monitoring data for

App. 222

2000-2004, in consultation with the Administratoror his or her designee;

(ii) For an implementation plan that issubmitted by 2003, the period for establishingbaseline visibility conditions for the period of thefirst long-term strategy is the most recent 5—yearperiod for which visibility monitoring data areavailable for the mandatory Class I Federal areasaddressed by the plan. For mandatory Class IFederal areas without onsite monitoring data, theState must establish baseline values using the mostrepresentative available monitoring data, inconsultation with the Administrator or his or herdesignee;

(iii) Natural visibility conditions for the mostimpaired and least impaired days. Natural visibilityconditions must be calculated by estimating thedegree of visibility impairment existing undernatural conditions for the most impaired and leastimpaired days, based on available monitoringinformation and appropriate data analysistechniques; and

(iv)(A) For the first implementation plan addressingthe requirements of paragraphs (d) and (e) of thissection, the number of deciviews by which baselineconditions exceed natural visibility conditions forthe most impaired and least impaired days; or

(B) For all future implementation planrevisions, the number of deciviews by whichcurrent conditions, as calculated underparagraph (f)(1) of this section, exceed natural

App. 223

visibility conditions for the most impaired andleast impaired days.

(3) Long-term strategy for regional haze. EachState listed in § 51.300(b)(3) must submit along-term strategy that addresses regional hazevisibility impairment for each mandatory Class IFederal area within the State and for eachmandatory Class I Federal area located outside theState which may be affected by emissions from theState. The long-term strategy must includeenforceable emissions limitations, complianceschedules, and other measures as necessary toachieve the reasonable progress goals establishedby States having mandatory Class I Federal areas.In establishing its long-term strategy for regionalhaze, the State must meet the followingrequirements:

(i) Where the State has emissions that arereasonably anticipated to contribute to visibilityimpairment in any mandatory Class I Federal arealocated in another State or States, the State mustconsult with the other State(s) in order to developcoordinated emission management strategies. TheState must consult with any other State havingemissions that are reasonably anticipated tocontribute to visibility impairment in anymandatory Class I Federal area within the State.

(ii) Where other States cause or contribute toimpairment in a mandatory Class I Federal area,the State must demonstrate that it has included inits implementation plan all measures necessary toobtain its share of the emission reductions neededto meet the progress goal for the area. If the State

App. 224

has participated in a regional planning process, theState must ensure it has included all measuresneeded to achieve its apportionment of emissionreduction obligations agreed upon through thatprocess.

(iii) The State must document the technical basis,including modeling, monitoring and emissionsinformation, on which the State is relying todetermine its apportionment of emission reductionobligations necessary for achieving reasonableprogress in each mandatory Class I Federal area itaffects. The State may meet this requirement byrelying on technical analyses developed by theregional planning organization and approved by allState participants. The State must identify thebaseline emissions inventory on which its strategiesare based. The baseline emissions inventory year ispresumed to be the most recent year of theconsolidate periodic emissions inventory.

(iv) The State must identify all anthropogenicsources of visibility impairment considered by theState in developing its long-term strategy. TheState should consider major and minor stationarysources, mobile sources, and area sources.

(v) The State must consider, at a minimum, thefollowing factors in developing its long-termstrategy:

(A) Emission reductions due to ongoing airpollution control programs, including measuresto address reasonably attributable visibilityimpairment;

App. 225

(B) Measures to mitigate the impacts ofconstruction activities;

(C) Emissions limitations and schedules forcompliance to achieve the reasonable progressgoal;

(D) Source retirement and replacementschedules;

(E) Smoke management techniques foragricultural and forestry management purposesincluding plans as currently exist within theState for these purposes;

(F) Enforceability of emissions limitationsand control measures; and

(G) The anticipated net effect on visibility dueto projected changes in point, area, and mobilesource emissions over the period addressed bythe long-term strategy.

(4) Monitoring strategy and otherimplementation plan requirements. The State mustsubmit with the implementation plan a monitoringstrategy for measuring, characterizing, andreporting of regional haze visibility impairment thatis representative of all mandatory Class I Federalareas within the State. This monitoring strategymust be coordinated with the monitoring strategyrequired in § 51.305 for reasonably attributablevisibility impairment. Compliance with thisrequirement may be met through participation inthe Interagency Monitoring of Protected VisualEnvironments network. The implementation planmust also provide for the following:

App. 226

(i) The establishment of any additionalmonitoring sites or equipment needed to assesswhether reasonable progress goals to addressregional haze for all mandatory Class I Federalareas within the State are being achieved.

(ii) Procedures by which monitoring data andother information are used in determining thecontribution of emissions from within the State toregional haze visibility impairment at mandatoryClass I Federal areas both within and outside theState.

(iii) For a State with no mandatory Class IFederal areas, procedures by which monitoring dataand other information are used in determining thecontribution of emissions from within the State toregional haze visibility impairment at mandatoryClass I Federal areas in other States.

(iv) The implementation plan must provide forthe reporting of all visibility monitoring data to theAdministrator at least annually for each mandatoryClass I Federal area in the State. To the extentpossible, the State should report visibilitymonitoring data electronically.

(v) A statewide inventory of emissions ofpollutants that are reasonably anticipated to causeor contribute to visibility impairment in anymandatory Class I Federal area. The inventorymust include emissions for a baseline year,emissions for the most recent year for which dataare available, and estimates of future projectedemissions. The State must also include acommitment to update the inventory periodically.

App. 227

(vi) Other elements, including reporting,recordkeeping, and other measures, necessary toassess and report on visibility.

(e) Best Available Retrofit Technology (BART)requirements for regional haze visibility impairment.The State must submit an implementation plancontaining emission limitations representing BARTand schedules for compliance with BART for eachBART—eligible source that may reasonably beanticipated to cause or contribute to any impairment ofvisibility in any mandatory Class I Federal area, unlessthe State demonstrates that an emissions tradingprogram or other alternative will achieve greaterreasonable progress toward natural visibilityconditions.

(1) To address the requirements for BART, theState must submit an implementation plancontaining the following plan elements and includedocumentation for all required analyses:

(i) A list of all BART—eligible sources withinthe State.

(ii) A determination of BART for eachBART—eligible source in the State that emits anyair pollutant which may reasonably be anticipatedto cause or contribute to any impairment ofvisibility in any mandatory Class I Federal area. Allsuch sources are subject to BART.

(A) The determination of BART must bebased on an analysis of the best system ofcontinuous emission control technology availableand associated emission reductions achievablefor each BART—eligible source that is subject to

App. 228

BART within the State. In this analysis, theState must take into consideration thetechnology available, the costs of compliance, theenergy and nonair quality environmentalimpacts of compliance, any pollution controlequipment in use at the source, the remaininguseful life of the source, and the degree ofimprovement in visibility which may reasonablybe anticipated to result from the use of suchtechnology.

(B) The determination of BART for fossil-fuelfired power plants having a total generatingcapacity greater than 750 megawatts must bemade pursuant to the guidelines in appendix Yof this part (Guidelines for BARTDeterminations Under the Regional Haze Rule).

(C) Exception. A State is not required to makea determination of BART for SO2 or for NOX if aBART—eligible source has the potential to emitless than 40 tons per year of such pollutant(s), orfor PM10 if a BART—eligible source has thepotential to emit less than 15 tons per year ofsuch pollutant.

(iii) If the State determines in establishing BARTthat technological or economic limitations on theapplicability of measurement methodology to aparticular source would make the imposition of anemission standard infeasible, it may insteadprescribe a design, equipment, work practice, orother operational standard, or combination thereof,to require the application of BART. Such standard,to the degree possible, is to set forth the emissionreduction to be achieved by implementation of such

App. 229

design, equipment, work practice or operation, andmust provide for compliance by means whichachieve equivalent results.

(iv) A requirement that each source subject toBART be required to install and operate BART asexpeditiously as practicable, but in no event laterthan 5 years after approval of the implementationplan revision.

(v) A requirement that each source subject toBART maintain the control equipment required bythis subpart and establish procedures to ensuresuch equipment is properly operated andmaintained.

(2) A State may opt to implement or requireparticipation in an emissions trading program orother alternative measure rather than to requiresources subject to BART to install, operate, andmaintain BART. Such an emissions tradingprogram or other alternative measure must achievegreater reasonable progress than would be achievedthrough the installation and operation of BART. Forall such emission trading programs or otheralternative measures, the State must submit animplementation plan containing the following planelements and include documentation for allrequired analyses:

(i) A demonstration that the emissions tradingprogram or other alternative measure will achievegreater reasonable progress than would haveresulted from the installation and operation ofBART at all sources subject to BART in the State

App. 230

and covered by the alternative program. Thisdemonstration must be based on the following:

(A) A list of all BART—eligible sources withinthe State.

(B) A list of all BART—eligible sources andall BART source categories covered by thealternative program. The State is not required toinclude every BART source category or everyBART—eligible source within a BART sourcecategory in an alternative program, but eachBART—eligible source in the State must besubject to the requirements of the alternativeprogram, have a federally enforceable emissionlimitation determined by the State and approvedby EPA as meeting BART in accordance withsection 302(c) or paragraph (e)(1) of this section,or otherwise addressed under paragraphs (e)(1)or (e)(4)of this section.

(C) An analysis of the best system ofcontinuous emission control technology availableand associated emission reductions achievablefor each source within the State subject to BARTand covered by the alternative program. Thisanalysis must be conducted by making adetermination of BART for each source subjectto BART and covered by the alternative programas provided for in paragraph (e)(1) of thissection, unless the emissions trading program orother alternative measure has been designed tomeet a requirement other than BART (such asthe core requirement to have a long-termstrategy to achieve the reasonable progress goalsestablished by States). In this case, the State

App. 231

may determine the best system of continuous emission control technology and associatedemission reductions for similar types of sourceswithin a source category based on bothsource-specific and category-wide information,as appropriate.

(D) An analysis of the projected emissionsreductions achievable through the tradingprogram or other alternative measure.

(E) A determination under paragraph (e)(3) ofthis section or otherwise based on the clearweight of evidence that the trading program orother alternative measure achieves greaterreasonable progress than would be achievedthrough the installation and operation of BARTat the covered sources.

(ii) [Reserved]

(iii) A requirement that all necessary emissionreductions take place during the period of the firstlong-term strategy for regional haze. To meet thisrequirement, the State must provide a detaileddescription of the emissions trading program orother alternative measure, including schedules forimplementation, the emission reductions requiredby the program, all necessary administrative andtechnical procedures for implementing the program,rules for accounting and monitoring emissions, andprocedures for enforcement.

(iv) A demonstration that the emission reductionsresulting from the emissions trading program orother alternative measure will be surplus to thosereductions resulting from measures adopted to meet

App. 232

requirements of the CAA as of the baseline date ofthe SIP.

(v) At the State’s option, a provision that theemissions trading program or other alternativemeasure may include a geographic enhancement tothe program to address the requirement under§ 51.302(c) related to BART for reasonablyattributable impairment from the pollutantscovered under the emissions trading program orother alternative measure.

(vi) For plans that include an emissions tradingprogram that establishes a cap on total annualemissions of SO2 or NOX from sources subject to theprogram, requires the owners and operators ofsources to hold allowances or authorizations to emitequal to emissions, and allows the owners andoperators of sources and other entities to purchase,sell, and transfer allowances, the following elementsare required concerning the emissions covered bythe cap:

(A) Applicability provisions defining thesources subject to the program. The State mustdemonstrate that the applicability provisions(including the size criteria for including sourcesin the program) are designed to prevent anysignificant potential shifting within the State ofproduction and emissions from sources in theprogram to sources outside the program. In thecase of a program covering sources in multipleStates, the States must demonstrate that theapplicability provisions in each State coveressentially the same size facilities and, if sourcecategories are specified, cover the same source

App. 233

categories and prevent any significant, potentialshifting within such States of production andemissions to sources outside the program.

(B) Allowance provisions ensuring that thetotal value of allowances (in tons) issued eachyear under the program will not exceed theemissions cap (in tons) on total annual emissionsfrom the sources in the program.

(C) Monitoring provisions providing forconsistent and accurate measurements ofemissions from sources in the program to ensurethat each allowance actually represents thesame specified tonnage of emissions and thatemissions are measured with similar accuracy atall sources in the program. The monitoringprovisions must require that boilers, combustionturbines, and cement kilns in the programallowed to sell or transfer allowances mustcomply with the requirements of part 75 of thischapter. The monitoring provisions must requirethat other sources in the program allowed to sellor transfer allowances must provide emissionsinformation with the same precision, reliability,accessibility, and timeliness as informationprovided under part 75 of this chapter.

(D) Recordkeeping provisions that ensure theenforceability of the emissions monitoringprovisions and other program requirements. Therecordkeeping provisions must require thatboilers, combustion turbines, and cement kilnsin the program allowed to sell or transferallowances must comply with the recordkeepingprovisions of part 75 of this chapter. The

App. 234

recordkeeping provisions must require thatother sources in the program allowed to sell ortransfer allowances must comply withrecordkeeping requirements that, as comparedwith the recordkeeping provisions under part 75of this chapter, are of comparable stringency andrequire recording of comparable types ofinformation and retention of the records forcomparable periods of time.

(E) Reporting provisions requiring timelyreporting of monitoring data with sufficientfrequency to ensure the enforceability of theemissions monitoring provisions and otherprogram requirements and the ability to auditthe program. The reporting provisions mustrequire that boilers, combustion turbines, andcement kilns in the program allowed to sell ortransfer allowances must comply with thereporting provisions of part 75 of this chapter,except that, if the Administrator is not thetracking system administrator for the program,emissions may be reported to the trackingsystem administrator, rather than to theAdministrator. The reporting provisions mustrequire that other sources in the programallowed to sell or transfer allowances mustcomply with reporting requirements that, ascompared with the reporting provisions underpart 75 of this chapter, are of comparablestringency and require reporting of comparabletypes of information and require comparabletimeliness and frequency of reporting.

App. 235

(F) Tracking system provisions which providefor a tracking system that is publicly availablein a secure, centralized database to track in aconsistent manner all allowances and emissionsin the program.

(G) Authorized account representativeprovisions ensuring that the owners andoperators of a source designate one individualwho is authorized to represent the owners andoperators in all matters pertaining to thetrading program.

(H) Allowance transfer provisions providingprocedures that allow timely transfer andrecording of a l lowances, minimizeadministrative barriers to the operation of theallowance market, and ensure that suchprocedures apply uniformly to all sources andother potential participants in the allowancemarket.

(I) Compliance provisions prohibiting asource from emitting a total tonnage of apollutant that exceeds the tonnage value of itsallowance holdings, including the methods andprocedures for determining whether emissionsexceed allowance holdings. Such method andprocedures shall apply consistently from sourceto source.

(J) Penalty provisions providing formandatory allowance deductions for excessemissions that apply consistently from source tosource. The tonnage value of the allowances

App. 236

deducted shall equal at least three times thetonnage of the excess emissions.

(K) For a trading program that allowsbanking of allowances, provisions clarifying anyrestrictions on the use of these bankedallowances.

(L) Program assessment provisions providingfor periodic program evaluation to assesswhether the program is accomplishing its goalsand whether modifications to the program areneeded to enhance performance of the program.

(3) A State which opts under 40 CFR 51.308(e)(2)to implement an emissions trading program orother alternative measure rather than to requiresources subject to BART to install, operate, andmaintain BART may satisfy the final step of thedemonstration required by that section as follows:If the distribution of emissions is not substantiallydifferent than under BART, and the alternativemeasure results in greater emission reductions,then the alternative measure may be deemed toachieve greater reasonable progress. If thedistribution of emissions is significantly different,the State must conduct dispersion modeling todetermine differences in visibility between BARTand the trading program for each impacted Class Iarea, for the worst and best 20 percent of days. Themodeling would demonstrate “greater reasonableprogress” if both of the following two criteria aremet:

(i) Visibility does not decline in any Class I area,and

App. 237

(ii) There is an overall improvement in visibility,determined by comparing the average differencesbetween BART and the alternative over all affectedClass I areas.

(4) A State subject to a trading programestablished in accordance with § 52.38 or § 52.39under a Transport Rule Federal ImplementationPlan need not require BART—eligible fossilfuel-fired steam electric plants in the State toinstall, operate, and maintain BART for thepollutant covered by such trading program in theState. A State that chooses to meet the emissionreduction requirements of the Transport Rule bysubmitting a SIP revision that establishes a tradingprogram and is approved as meeting therequirements of § 52.38 or § 52.39 also need notrequire BART—eligible fossil fuel-fired steamelectric plants in the State to install, operate, andmaintain BART for the pollutant covered by suchtrading program in the State. A State may adoptprovisions, consistent with the requirementsapplicable to the State for a trading programestablished in accordance with § 52.38 or § 52.39under the Transport Rule Federal ImplementationPlan or established under a SIP revision that isapproved as meeting the requirements of § 52.38 or§ 52.39, for a geographic enhancement to theprogram to address the requirement under§ 51.302(c) related to BART for reasonablyattributable impairment from the pollutant coveredby such trading program in that State.

(5) After a State has met the requirements forBART or implemented emissions trading program

App. 238

or other alternative measure that achieves morereasonable progress than the installation andoperation of BART, BART—eligible sources will besubject to the requirements of paragraph (d) of thissection in the same manner as other sources.

(6) Any BART—eligible facility subject to therequirement under paragraph (e) of this section toinstall, operate, and maintain BART may apply tothe Administrator for an exemption from thatrequirement. An application for an exemption willbe subject to the requirements of § 51.303(a)(2)-(h).

(f) Requirements for comprehensive periodicrevisions of implementation plans for regional haze.Each State identified in § 51.300(b)(3) must revise andsubmit its regional haze implementation plan revisionto EPA by July 31, 2018 and every ten years thereafter.In each plan revision, the State must evaluate andreassess all of the elements required in paragraph (d)of this section, taking into account improvements inmonitoring data collection and analysis techniques,control technologies, and other relevant factors. Inevaluating and reassessing these elements, the Statemust address the following:

(1) Current visibility conditions for the mostimpaired and least impaired days, and actualprogress made towards natural conditions duringthe previous implementation period. The period forcalculating current visibility conditions is the mostrecent five year period preceding the required dateof the implementation plan submittal for whichdata are available. Current visibility conditionsmust be calculated based on the annual averagelevel of visibility impairment for the most and least

App. 239

impaired days for each of these five years. Currentvisibility conditions are the average of these annualvalues.

(2) The effectiveness of the long-term strategyfor achieving reasonable progress goals over theprior implementation period(s); and

(3) Affirmation of, or revision to, the reasonableprogress goal in accordance with the procedures setforth in paragraph (d)(1) of this section. If the Stateestablished a reasonable progress goal for the priorperiod which provided a slower rate of progressthan that needed to attain natural conditions by theyear 2064, the State must evaluate and determinethe reasonableness, based on the factors inparagraph (d)(1)(i)(A) of this section, of additionalmeasures that could be adopted to achieve thedegree of visibility improvement projected by theanalysis contained in the first implementation plandescribed in paragraph (d)(1)(i)(B) of this section.

(g) Requirements for periodic reports describingprogress towards the reasonable progress goals. EachState identified in § 51.300(b)(3) must submit a reportto the Administrator every 5 years evaluating progresstowards the reasonable progress goal for eachmandatory Class I Federal area located within theState and in each mandatory Class I Federal arealocated outside the State which may be affected byemissions from within the State. The first progressreport is due 5 years from submittal of the initialimplementation plan addressing paragraphs (d) and (e)of this section. The progress reports must be in theform of implementation plan revisions that complywith the procedural requirements of § 51.102 and

App. 240

§ 51.103. Periodic progress reports must contain at aminimum the following elements:

(1) A description of the status of implementationof all measures included in the implementation planfor achieving reasonable progress goals formandatory Class I Federal areas both within andoutside the State.

(2) A summary of the emissions reductionsachieved throughout the State throughimplementation of the measures described inparagraph (g)(1) of this section.

(3) For each mandatory Class I Federal areawithin the State, the State must assess thefollowing visibility conditions and changes, withvalues for most impaired and least impaired daysexpressed in terms of 5—year averages of theseannual values.

(i) The current visibility conditions for the mostimpaired and least impaired days;

(ii) The difference between current visibilityconditions for the most impaired and least impaireddays and baseline visibility conditions;

(iii) The change in visibility impairment for themost impaired and least impaired days over thepast 5 years;

(4) An analysis tracking the change over the past5 years in emissions of pollutants contributing tovisibility impairment from all sources and activitieswithin the State. Emissions changes should beidentified by type of source or activity. The analysis

App. 241

must be based on the most recent updatedemissions inventory, with estimates projectedforward as necessary and appropriate, to accountfor emissions changes during the applicable 5—yearperiod.

(5) An assessment of any significant changes inanthropogenic emissions within or outside the Statethat have occurred over the past 5 years that havelimited or impeded progress in reducing pollutantemissions and improving visibility.

(6) An assessment of whether the currentimplementation plan elements and strategies aresufficient to enable the State, or other States withmandatory Federal Class I areas affected byemissions from the State, to meet all establishedreasonable progress goals.

(7) A review of the State’s visibility monitoringstrategy and any modifications to the strategy asnecessary.

(h) Determination of the adequacy of existingimplementation plan. At the same time the State isrequired to submit any 5—year progress report to EPAin accordance with paragraph (g) of this section, theState must also take one of the following actions basedupon the information presented in the progress report:

(1) If the State determines that the existingimplementation plan requires no furthersubstantive revision at this time in order to achieveestablished goals for visibility improvement andemissions reductions, the State must provide to theAdministrator a negative declaration that further

App. 242

revision of the existing implementation plan is notneeded at this time.

(2) If the State determines that theimplementation plan is or may be inadequate toensure reasonable progress due to emissions fromsources in another State(s) which participated in aregional planning process, the State must providenotification to the Administrator and to the otherState(s) which participated in the regional planningprocess with the States. The State must alsocollaborate with the other State(s) through theregional planning process for the purpose ofdeveloping additional strategies to address theplan’s deficiencies.

(3) Where the State determines that theimplementation plan is or may be inadequate toensure reasonable progress due to emissions fromsources in another country, the State shall providenotification, along with available information, to theAdministrator.

(4) Where the State determines that theimplementation plan is or may be inadequate toensure reasonable progress due to emissions fromsources within the State, the State shall revise itsimplementation plan to address the plan’sdeficiencies within one year.

(i) What are the requirements for State and FederalLand Manager coordination?

(1) By November 29, 1999, the State mustidentify in writing to the Federal Land Managersthe title of the official to which the Federal LandManager of any mandatory Class I Federal area can

App. 243

submit any recommendations on theimplementation of this subpart including, but notlimited to:

(i) Identification of impairment of visibility inany mandatory Class I Federal area(s); and

(ii) Identification of elements for inclusion in thevisibility monitoring strategy required by § 51.305and this section.

App. 244

APPENDIX E

Regional Haze Implementation Plan Revision

State of Oklahoma

Department of Environmental Quality

February 2, 2010

* * *

[p.81]

* * *

As outlined in the previous section and described indetail in Appendix 6-4, DEQ conducted a thoroughcase-by-case five-factor BART analysis for each of theBART-subject units. DEQ determined that Dry-FlueGas Desulfurization with Spray Dryer Absorber (“DryFGD with SDA”) is not cost-effective for S02 control forany of the six coal-fired steam electric units reviewed,i.e., OG&E Sooner Units 1 and 2, OG&E MuskogeeUnits 4 and 5, and PSO Northeastern Units 3 and 4.This determination is based on the capital cost of add-on controls, the cost effectiveness both in dollars perton and dollars per deciview of add-on controls, and thelong term viability of coal with respect to otherenvironmental programs, and national commitments.In addition to information provided prior to the publichearing, DEQ considered public comments, andadditional information provided by the affectedfacilities in response to questions raised by thecommentors and DEQ staff. Revised cost estimates

App. 245

were provided by the affected facilities that are basedon vendor quotes and go well beyond the defaultmethodology recommended by EPA guidance. The costestimates are credible, detailed, and specific for theindividual facilities. The final estimate for Dry FGDwith SDA for the six coal-fired units was on average153% greater than the high end costs assumed by DEQin the Draft SIP. These costs put the projects wellabove costs reported for other BART determinations,and above the levels DEQ considered reasonable forcost effectiveness both in terms of dollars per ton ofpollutant removed and dollars per deciview (e.g.,$10,000,000/dv) of improved visibilty. Tables VI-14 andVI-15 give data on these measures of cost-effectiveness.

[p.82]

* * *

DEQ has determined that the cost for DFGD is toohigh and the benefit too low. These costs would furtherextend the life expectancy of coal as the primary fuel inthe Sooner facility for at least 20 years and beyond.Consequently, DEQ has determined BART for the sixcoal-fired steam electric units to be the use (orcontinued use) of low sulfur coal. Additionalexplanation of DEQ’s rationale and conclusions isincluded in the BART Determinations in Appendix 6-4.

* * *

App. 246

APPENDIX F

UNITED STATES COURT OF APPEALSFOR THE TENTH CIRCUIT

[Filed June 22, 2012]

Nos. 12-9526 & 12-9527

(No. EPA-R06-OAR-2010-0190)_______________________________________STATE OF OKLAHOMA; OKLAHOMA )INDUSTRIAL ENERGY CONSUMERS, )an unincorporated association; )OKLAHOMA GAS & ELECTRIC )COMPANY, )

Petitioners, ))

v. ))

UNITED STATES ENVIRONMENTAL )PROTECTION AGENCY, )

Respondent. )--------------------------------------------------------- )SIERRA CLUB, )

Intervenor-Respondent. )_______________________________________)

ORDER

Before KELLY and HOLMES, Circuit Judges.

Petitioners, the State of Oklahoma, OklahomaIndustrial Energy Consumers, and the Oklahoma Gas& Electric Company, seek a stay pending review of that

App. 247

portion of the Environmental Protection Agency’s finalrule requiring the reduction of sulfur dioxide emissionsat four electric generating units. We conclude that thestay factors have been met in this case, and wetherefore GRANT the motion for stay pending hearingby the merits panel.

Entered for the Court

/s/ Elisabeth A. ShumakerELISABETH A. SHUMAKER, Clerk