Vol. 77 Wednesday, No. 234 December 5, 2012

Part III

Environmental Protection Agency

40 CFR Part 52 Approval, Disapproval and Promulgation of Air Quality Implementation Plans; ; Regional Haze State and Federal Implementation Plans; Final Rule

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00001 Fmt 4717 Sfmt 4717 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72512 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

ENVIRONMENTAL PROTECTION http://www.regulations.gov or in hard (14) The initials CEMS mean or refer AGENCY copy at EPA Region 9, 75 Hawthorne to continuous emission monitoring Street, San Francisco, California. Please system. 40 CFR Part 52 note that while many of the documents (15) The term Cholla refers to Cholla in the docket are listed at http:// Power Plant. [EPA–R09–OAR–2012–0021, FRL–9754–3] www.regulations.gov, some information (16) The term Class I area refers to a 1 Approval, Disapproval and may not be specifically listed in the mandatory Class I Federal area. Promulgation of Air Quality index to the docket and may be publicly (17) The term coal-fired BART units Implementation Plans; Arizona; available only at the hard copy location refers to Apache Generating Station Regional Haze State and Federal (e.g., copyrighted material, large maps, Units 2 and 3; Cholla Power Plant Units Implementation Plans multi-volume reports or otherwise 2, 3 and 4 and Coronado Generating voluminous materials), and some may Station Units 1 and 2. AGENCY: Environmental Protection not be available at either locations (e.g., (18) The initials COFA mean or refer Agency (EPA). confidential business information). To to close-coupled overfire air. (19) The term Coronado refers to ACTION: Final rule. inspect the hard copy materials, please schedule an appointment during normal Coronado Generating Station. SUMMARY: EPA is taking final action to business hours with the contact listed (20) The initials CY mean or refer to approve in part and disapprove in part directly below. Calendar Year. a portion of Arizona’s State (21) The initials EGU mean or refer to FOR FURTHER INFORMATION CONTACT: Electric Generating Unit. Implementation Plan (SIP) submittal for Thomas Webb, U.S. EPA, Region 9, its regional haze program and to (22) The initials ESPs mean or refer to Planning Office, Air Division, Air-2, 75 electrostatic precipitators. promulgate a Federal Implementation Hawthorne Street, San Francisco, CA Plan (FIP) for the disapproved elements (23) The words EPA, we, us or our 94105. Thomas Webb can be reached at mean or refer to the United States of the SIP. The State and Federal plans telephone number (415) 947–4139 and are to implement the regional haze Environmental Protection Agency. via electronic mail at (24) The initials FGD mean or refer to program in Arizona for the first [email protected]. flue gas desulfurization. planning period through 2018. This SUPPLEMENTARY INFORMATION: (25) The initials FGR mean or refer to final rule addresses only the portion of flue gas recirculation. the SIP related to Arizona’s Throughout this document, wherever ‘‘we,’’ ‘‘us,’’ or ‘‘our,’’ is used, we mean (26) The initials FIP mean or refer to determination of Best Available Retrofit Federal Implementation Plan. Technology (BART) to control emissions the United States Environmental Protection Agency (EPA). (27) The initials FLMs mean or refer from eight units at three electric to Federal Land Managers. generating stations: Apache Generating Definitions (28) The initials FR mean or refer to Station, Cholla Power Plant and For the purpose of this document, we the Federal Register. Coronado Generating Station. Consistent are giving meaning to certain words or (29) The initials GEP mean or refer to with our proposal, EPA approves in this initials as follows: Good Engineering Practice. final rule the State’s determination that (1) The words or initials CAA or Act (30) The initials IMPROVE mean or the three sources are subject to BART, mean or refer to the Clean Air Act, refer to Interagency Monitoring of and approves the State’s emissions unless the context indicates otherwise. Protected Visual Environments limits for sulfur dioxide (SO2) and (2) The initials ACC refer to the monitoring network. particulate matter less than or equal to Arizona Corporation Commission. (31) The initials IWAQM mean or 10 micrometers (PM10) at all the units, (3) The initials ACCCE mean or refer refer to Interagency Workgroup on Air but disapproves Arizona’s BART to American Coalition for Clean Coal Quality Modeling. emissions limits for nitrogen oxides Electricity. (32) The initials IPM mean or refer to (NOX) at the coal-fired units of the three (4) The initials ADEQ mean or refer to Integrated Planning Model. power plants. We also are promulgating the Arizona Department of (33) The initials LNB mean or refer to a FIP that contains new emissions limits Environmental Quality. low-NOX burners. for NOX at these coal-fired units and (5) The initials AEPCO mean or refer (34) The initials LTS mean or refer to compliance schedules for to Arizona Electric Power Cooperative. Long-Term Strategy. implementation of BART as well as (6) The initials AFUDC mean or refer (35) The initials MMBtu mean or refer requirements for equipment to allowance for funds used during to Million British thermal units. maintenance, monitoring, recordkeeping construction. (36) The initials MW mean or refer to and reporting for all units and all (7) The term Apache refers to Apache megawatts. pollutants at the three sources. In Generating Station. (37) The initials MWh mean or refer today’s action, we are revising some (8) The initials APS mean or refer to to megawatt hours. elements of the proposed FIP in Company. (38) The initials NEI mean or refer to response to comments and additional (9) The words Arizona and State National Emission Inventory. information that we received. mean the State of Arizona. (39) The initials NH3 mean or refer to (10) The initials BART mean or refer ammonia. DATES: Effective date: This rule is to Best Available Retrofit Technology. (40) The initials NOX mean or refer to effective January 4, 2013. (11) The term BART units refers to nitrogen oxides. Compliance dates: The owners/ Apache Generating Station Units 1, 2 (41) The initials NP mean or refer to operators of each unit subject to this and 3; Cholla Power Plant Units 2, 3 and National Park. final rule shall comply by the dates 4 and Coronado Generating Station specified in the regulatory text. Units 1 and 2. 1 Although states and tribes may designate as ADDRESSES: EPA has established docket (12) The initials CBI mean or refer to Class I additional areas which they consider to have visibility as an important value, the requirements of number EPA–R09–OAR–2012–0021 for Confidential Business Information. the visibility program set forth in section 169A of this action. Generally, documents in the (13) The initials CCM mean or refer to the CAA apply only to ‘‘mandatory Class I Federal docket are available electronically at EPA’s Cost Control Manual. areas.’’

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00002 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72513

(42) The initials NPRM mean or refer (72) The initials WEP mean or refer to emissions limits for NOX at all of the to Notice of Proposed Rulemaking. Weighted Emissions Potential. coal-fired BART units (i.e., all of the (43) The initials O&M mean or refer (73) The initials WFGD mean or refer BART units except for Apache Unit 1). to operation and maintenance. to wet flue gas desulfurization. We also proposed to disapprove the (44) The initials OC mean or refer to (74) The initials WRAP mean or refer compliance schedules and requirements organic carbon. to the Western Regional Air Partnership. for equipment maintenance and (45) The initials OFA mean or refer to Table of Contents operation, including monitoring, over fire air. recordkeeping and reporting (46) The initials PM mean or refer to I. Background requirements for BART at all of the particulate matter. A. Summary of Our Proposed Action BART units, since these were not (47) The initials PM mean or refer to B. Legal Basis for Our Final Action 10 II. Overview of Final Action included in the SIP submittal. particulate matter with an aerodynamic III. Final BART Determinations Proposed FIP: The proposed FIP diameter of less than 10 micrometers IV. EPA’s Responses to Comments contained BART emissions limits for (coarse particulate matter). A. General Comments on ADEQ’s NOX at all of the coal-fired BART units, (48) The initials PM2.5 mean or refer Approach to BART as well as compliance deadlines and to fine particulate matter with an B. Comments on ADEQ’s Individual BART requirements for equipment aerodynamic diameter of less than 2.5 Analyses and Determinations maintenance and operation, including C. General Comments on EPA’s BART FIP micrometers. monitoring, recordkeeping and (49) The initials PNG mean or refer to Analyses and Determinations D. Source-Specific Comments on EPA’s reporting, to ensure the enforceability of pipeline natural gas. BART Analyses and Determinations the BART limits for all of the BART (50) The initials ppm mean or refer to E. Comments on Enforceability units. Because our proposed FIP parts per million. Requirements in EPA’s BART FIP emission limits would likely result in (51) The initials PSD mean or refer to F. Comments on Legal Issues changes in stack conditions from those Prevention of Significant Deterioration. G. Other Comments anticipated in the SIP, we invited (52) The initials RACT mean or refer V. Summary of Final Action comment on whether an alternative test VI. Statutory and Executive Order Reviews to Reasonably Available Control method to the one required in the SIP Technology. I. Background is acceptable for PM10. In addition, we (53) The initials RAVI mean or refer A. Summary of Our Proposed Action specifically sought comment on whether to Reasonably Attributable Visibility we should require lower SO2 emissions Impairment. Our notice of proposed rulemaking limits or removal efficiency (54) The initials RATA mean or refer (NPRM) was signed on July 2, 2012, and requirements for any of the coal-fired to relative accuracy test audit. was published in the Federal Register BART units. Finally, in the regulatory 2 (55) The initials RHR mean or refer to on July 20, 2012. In that notice, we text in our NPRM, we proposed to the Regional Haze Rule, originally proposed to approve in part and incorporate by reference into the FIP promulgated in 1999 and codified at 40 disapprove in part a portion of two provisions of the Arizona CFR 51.301–309. Arizona’s Regional Haze SIP (submitted Administrative Code, R18–2–310 and (56) The initials RMB refer to RMB on February 28, 2011) and proposed a R18–2–310.01, which we characterized Consulting & Research, Inc. FIP to address the deficiencies in the as establishing an affirmative defense (57) The initials RMC mean or refer to disapproved portions of the SIP. The for excess emissions due to Regional Modeling Center. proposed rule addressed the BART malfunctions.3 (58) The initials RP mean or refer to requirements for eight units at three Reasonable Progress. electric generating stations: Arizona B. Legal Basis for Our Final Action (59) The initials RPG or RPGs mean or Electric Power Company’s (AEPCO) Our action is based on an evaluation refer to Reasonable Progress Goal(s). Apache Generating Station (Apache) of Arizona’s Regional Haze SIP (60) The initials RPOs mean or refer Units 1, 2 and 3; Arizona Public submitted on February 28, 2011, to meet to regional planning organizations. Service’s (APS) Cholla Power Plant the requirements of Section 308 of the (61) The initials SCR mean or refer to (Cholla) Units 2, 3 and 4; and Salt River RHR. We evaluated the SIP against the Selective Catalytic Reduction. Project’s (SRP) Coronado Generating requirements of the RHR and Clean Air (62) The initials SIP mean or refer to Station (Coronado) Units 1 and 2. We Act (CAA) sections 169A and 169B. We State Implementation Plan. did not propose action on any other part also applied the general SIP (63) The initials SNCR mean or refer of Arizona’s SIP related to the remaining requirements in CAA section 110. Our to Selective Non-catalytic Reduction. requirements of the Regional Haze Rule authority for action on Arizona’s (64) The initials SO mean or refer to 2 (RHR). In summary, we proposed the Regional Haze SIP is based on CAA sulfur dioxide. following: section 110(k). Our authority to (65) The initials SOFA mean or refer Proposed Approval: We proposed to promulgate a FIP is based on CAA to separated over fire air. approve Arizona’s determination that section 110(c). (66) The initials SRP mean or refer to the following sources and units are Agricultural subject to BART: Apache Units 1, 2 and II. Overview of Final Action Improvement and Power District. 3; Cholla Units 2, 3 and 4; and Coronado EPA is taking final action to approve (67) The initials TCI mean or refer to Units 1 and 2 (collectively ‘‘BART in part and disapprove in part a portion total capital investment. units’’). We proposed to approve of Arizona’s SIP for Regional Haze, and (68) The initials tpy mean tons per Arizona’s BART emissions limits for to promulgate a FIP for the disapproved year. SO2 and PM10 at all three sources and elements of the SIP. This final rule only (69) The initials TSD mean or refer to units and the emissions limit for NOX at addresses the BART requirements for Technical Support Document. Apache Unit 1. the eight BART units identified above. (70) The initials VOC mean or refer to Proposed Disapproval: We proposed volatile organic compounds. to disapprove Arizona’s BART 3 Those provisions also include an affirmative (71) The initials WA mean or refer to defense for excess emissions due to startups and Wilderness Area. 2 77 FR 42834. shutdowns, which we did not intend to incorporate.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00003 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72514 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

Most notably, and with the exception of proposal, and have concluded that some We have revised certain elements of Apache Unit 1, the FIP includes NOX changes are warranted based on public our proposed FIP based on public emission limits for all the units that are comments and additional information comments and additional information as achievable with SCR. At this time, EPA we received in response to questions follows: is not taking action on the State’s other raised in the proposal. • Apache Units 2 and 3: The final BART determinations or any other parts Final Approval: EPA is approving emissions limit for NOX is 0.070 pounds of the SIP regarding the remaining Arizona’s determination that the per million British thermal units (lb/ requirements of the RHR. following sources and units are subject MMBtu) determined as an average of the EPA takes very seriously a decision to to BART: Apache Units 1, 2 and 3; two units, based on a rolling 30-boiler- disapprove any state plan. To approve a Cholla Units 2, 3 and 4; and Coronado operating-day average. Compared to the state plan, EPA must be able to find that Units 1 and 2 (collectively ‘‘BART proposed emissions limit of 0.050 lb/ the state plan is consistent with the units’’). We are approving the emissions MMBtu on each unit, this higher limit requirements of the CAA and EPA’s limits for NOX, PM10 and SO2 at Apache and the addition of a two-unit average regulations. Further, EPA’s oversight Unit 1 as proposed. We are approving provides an extra margin of compliance role requires us to ensure fair the State’s emissions limits for PM10 and to account for periods of startup and implementation of CAA requirements SO2 for all the units. shutdown as well as additional by states across the country, even while Final Disapproval: Based on our operational flexibility for Apache given acknowledging that individual evaluation described in the proposal AEPCO’s status as a small entity. When decisions from source to source or state and in this document, we are either one of the two units is not to state may not have identical disapproving the State’s BART operating, its emissions from its own outcomes. In this instance, for the emissions limits for NOX at all the preceding thirty boiler-operating-days reasons described in our proposal and BART units except for Apache Unit 1, will continue to be included in the two- in this document, we find that the for which the SIP’s BART determination unit average. The final compliance date State’s NOX BART determinations for consists of fuel switching to pipeline for this NOX limit remains five years the coal-fired units are not consistent natural gas (PNG). We also are from the date of publication of this final with the applicable statutory and disapproving the compliance schedules rule. For SO2 and PM10 we are regulatory requirements. Furthermore, and requirements for equipment extending the compliance deadline to the Arizona Regional Haze SIP does not maintenance and operation, including four years from publication of this final include the necessary compliance monitoring, recordkeeping and rule in order to provide AEPCO with schedules and requirements for reporting requirements for BART at all sufficient time to implement upgrades equipment maintenance and operation, the BART units since these were not to the existing scrubbers and including monitoring, recordkeeping included in the Arizona Regional Haze electrostatic precipitators (ESPs) at these and reporting requirements for BART. 4 units. SIP. • As a result, EPA believes this final Final Federal Implementation Plan: Cholla Units 2, 3 and 4: The final disapproval is the only path that is We are promulgating a FIP that includes emissions limit for NOX is 0.055 lb/ consistent with the Act at this time. emissions limitations representing MMBtu determined as an average of the We encourage the State to submit a three units, based on a rolling 30-boiler- BART for NOX at all the coal-fired revised SIP to replace all portions of our BART units. The FIP also includes operating-day average. Compared to the FIP, and are ready to work with the compliance schedules and requirements proposed emissions limit of 0.050 lb/ State to develop a revised plan. The for equipment maintenance, monitoring, MMBtu on each unit, the higher limit CAA requires states to prevent any testing, recordkeeping and reporting for and three-unit average provide an extra future and remedy any existing man- margin of compliance to account for all the BART units. For PM10 at all made impairment of visibility in 156 units, we allow the use of Method 5 as periods of startup and shutdown. When national parks and wilderness areas an alternative to Method 201A/202. In any of the three units is not operating, designated as Class I areas. Arizona has addition, the FIP includes a removal its emissions from its own preceding a wealth of such areas. The three power thirty boiler-operating-days will efficiency requirement for SO2 on plants affect visibility at 18 national Cholla Units 2, 3 and 4, which will continue to be included in the three- parks and wilderness areas, including ensure that the scrubbers on these units unit average. As proposed, the final the Grand Canyon, Mesa Verde and the are properly operated and maintained. compliance date to install and operate Petrified Forest. The State and EPA Finally, we are incorporating into the controls is five years from the date of

must work together to ensure that plans FIP an affirmative defense provision for publication of this final rule. For SO2, we are adding a removal efficiency are in place to make progress toward excess emissions due to malfunctions.5 natural visibility conditions at these requirement of 95 percent for the national treasures. 4 For each BART source, the SIP must include a scrubbers on Cholla Units 2, 3 and 4, in order to ensure that these scrubbers are III. Final BART Determinations requirement to install and operate control equipment as expeditiously as practicable (40 CFR properly operated and maintained, This section is a summary of EPA’s 51.308(e)(1)(iv)); a requirement to maintain control consistent with 40 CFR 51.308(e)(1)(v). final action on the BART determinations equipment (40 CFR 51.308(e)(1)(v)); and procedures We are retaining the other compliance to ensure control equipment is properly operated for the BART units at Apache, Cholla and maintained, including requirements for deadlines as proposed, except for Cholla and Coronado electric generating monitoring, recordkeeping and reporting (40 CFR Unit 2, where we are extending the stations. Please refer to Table 1 that 51.308(e)(1)(v)). compares this final rule to the proposal 5 In the regulatory text in our NPRM, we proposed promulgating today include an adequate margin of that was published on July 20, 2012. to incorporate by reference into the FIP two compliance to account for periods of startup and provisions of the Arizona Administrative Code, shutdown. Accordingly, as indicated by the title of Where EPA has modified our proposal R18–2–310 and R18–2–310.01, which we this provision in our proposed regulatory text to respond to comments or additional characterized as establishing an affirmative defense (‘‘Affirmative Defense for Malfunctions’’), we are information, we explain our analysis in for excess emissions due to malfunctions. However, only incorporating into the FIP the malfunction- the next section titled ‘‘EPA’s Responses those provisions also include an affirmative defense related provisions of these rules and not the startup for excess emissions due to startups and and shutdown provisions. Our final regulatory text to Comments.’’ We have fully shutdowns, which we did not intend to incorporate. clarifies this distinction and also incorporates the considered all comments on our As explained below, the emission limits that we are definition of malfunction.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00004 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72515

compliance deadline to April 1, 2016, MMBtu determined as an average of the When either one of the two units is not for both SO2 and PM10 in order to two units, based on a rolling 30-boiler- operating, its emissions from its own provide APS with sufficient time to operating-day average. Compared to the preceding thirty boiler-operating-days install a new wet flue gas proposed emissions limits of 0.050 on will continue to be included in the two- desulfurization (FGD) system and fabric Unit 1 and 0.080 on Unit 2, this new unit average. The final compliance date filter on this unit. limit based on a two-unit average for the two units is five years from the • Coronado Units 1 and 2: The final provides an extra margin of compliance date of publication of this final rule. emissions limit for NOX is 0.065 lb/ to account for startup and shutdown.

TABLE 1—SUMMARY OF CHANGES FROM PROPOSAL TO FINAL RULE: EMISSIONS LIMITS (LB/MMBTU) AND COMPLIANCE DATES IN SIP AND FIP

NOX PM10 SO2 Source Proposal Final Proposal Final Proposal Final

Apache Unit 1 .... 0.056, Five years 0.056, Five years 0.0075, 180 days 0.0075, 180 days 0.00064, 180 days 0.00064, 180 days. Apache Unit 2 .... 0.050, Five years 0.070 (across two 0.03, 180 days ..... 0.03, Four years 0.15, 180 days ..... 0.15, Four years. units) Apache Unit 3 .... 0.050, Five years Five years 0.03, 180 days ..... 0.03, Four years 0.15, 180 days ..... 0.15, Four years. Cholla Unit 2 ...... 0.050, Five years 0.055 (across 0.015, Jan 1, 2015 0.015, Apr 1, 2016 0.15, 180 days ..... Add 95 percent ef- three units) ficiency Apr 1, 2016. Cholla Unit 3 ...... 0.050, Five years Five years 0.015, 180 days ... 0.015, 180 days 0.15, 180 days ..... Add 95 percent ef- ficiency 1 year. Cholla Unit 4 ...... 0.050, Five years 0.015, 180 days ... 0.015, 180 days 0.15, 180 days ..... Add 95 percent ef- ficiency 1 year. Coronado Unit 1 0.050, Five years 0.065 (across two 0.03, 180 days ..... 0.03, 180 days 0.08, 180 days ..... 0.08, 180 days. units) Coronado Unit 2 0.080, June 1, Five years 0.03, 180 days ..... 0.03, 180 days 0.08, 180 days ..... 0.08, 180 days. 2014.

IV. EPA’s Responses to Comments Arizona Department of Environmental added that EPA apparently approved, We are responding to comments on Quality (ADEQ), and a consortium of without any scrutiny, ADEQ’s our proposed rule published on July 20, conservation organizations (National determination that Cholla Unit 1 is not 2012.6 We held an initial public hearing Parks Conservation Association, Sierra BART-eligible. The commenter (Earthjustice) in Phoenix, Arizona, on July 31, 2012. Club, Physicians for Social requested that EPA properly analyze the In response to concerns that more time Responsibility—Arizona Chapter, Dine’ BART-eligibility of Cholla Unit 1. was needed to analyze the proposal and Citizens Against Ruining Our Specifically, the commenter requested develop comments, we added two Environment, Grand Canyon Trust, and San Juan Citizens Alliance) represented that EPA identify which ‘‘aspects of the additional public hearings in Holbrook process by which ADEQ identified its and in Benson, Arizona, on August 14 by Earthjustice. All of the comments we received along with attached technical eligible-for-BART and subject-to-BART and 15, respectively, and extended the sources’’ it disagrees with, the basis of public comment deadline to September reports and analyses are available for 7 review in the docket. each disagreement, and whether any 18, 2012. The three public hearings such disagreement implicates Cholla were attended by hundreds of citizens, A. General Comments on ADEQ’s Unit 1. In addition, the commenter local and state government officials, Approach to BART stated that EPA’s independent analysis workers and officials from the power of this issue must be supported by the plants, and representatives from 1. ADEQ’s Identification of BART Sources following information, which is needed environmental organizations. Testimony to verify the actual date that Cholla Unit Comment: One commenter and comments from the three public 1 began operating: hearings are organized in the docket by (Earthjustice) stated that EPA must • The document entitled ‘‘Operating location and available for viewing at provide further factual support for its Notes for May 1962’’ referenced in www.epa.gov/region9/air/actions/ determination that Cholla Unit 1 is not ADEQ’s SIP; arizona.html and http:// BART-eligible. The commenter • All available 1962 operating records www.regulations.gov. indicated that the record lacks the for Cholla Unit 1; We also received a number of written requisite support for this conclusion. • All initial CAA construction and comments, including extensive Recounting the history of ADEQ’s operating permits issued to Cholla Unit comments from stakeholders and finding that Unit 1 is not BART-eligible, 1; government agencies who offered policy the commenter noted that APS claimed, • All emissions data from the year and technical analyses addressing the and ADEQ concurred, that Unit 1 is not 1962 for Cholla Unit 1; details of our proposed rule. These BART-eligible based on a 50-year-old • Notes of the meeting between stakeholders included AEPCO, APS, document entitled ‘‘Operating Notes For ADEQ and APS in August 2007 or any SRP, PacifiCorp, Arizona Utilities Group May 1962’’ which allegedly shows that other time ADEQ and APS discussed the (AUG), National Park Service (NPS), Unit 1 began operations on May 1, 1962, BART-eligibility of Cholla Unit 1; and and was thus placed into operation just • Any other documentation that 6 77 FR 42834. months before the August 7, 1962, either supports or contradicts whether 7 77 FR 45326 (July 31, 2012). BART-eligibility cut-off. The commenter Cholla Unit 1 was placed into

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00005 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72516 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

commercial operation before August 7, eligibility when we take action on the control that is BART on a case-by-case 1962. remainder of the Arizona Regional Haze basis. Response: We did not specifically SIP. propose to take action on ADEQ’s a. ADEQ’s Approach to Costs of determination that Cholla Unit 1 is not 2. ADEQ’s BART Control Analyses Compliance BART-eligible and our statement that Comment: One commenter Comment: One commenter (NPS) ‘‘we do not agree with all aspects of the (PacifiCorp) states that EPA improperly agreed with EPA’s conclusions that the process by which ADEQ identified its focuses on only two factors, costs and costs of control were not calculated by eligible-for-BART and subject-to-BART visibility improvement, in rejecting the ADEQ in accordance with the BART sources’’ was not intended to apply to ADEQ’s entire NO BART Guidelines and that costs were included X for items not allowed by EPA Control this unit. Nonetheless, we agree with determination. The commenter states Cost Manual (CCM or the Manual) (e.g., the commenter that it is appropriate to that EPA inappropriately places more owner’s costs, surcharge, escalation, and give some consideration to this issue in weight on these factors. the context of today’s rulemaking Allowance for Funds Utilized During action, which covers ADEQ’s BART Response: EPA disagrees with the Construction—AFUDC), which inflates determinations for the other three units comment that we inappropriately the total cost of compliance and the cost at Cholla. focused on costs and visibility per ton of pollutant reduced. According Contrary to the commenter’s improvement in our decision to to the commenter, a review of industry assertion, the WRAP did not find Cholla disapprove ADEQ’s NOX BART data (detailed in Appendix A of the Unit 1 subject to BART. The WRAP determinations. As outlined in our commenter’s submission) indicates that document cited by the commenter proposal, we considered ADEQ’s the total capital investment (TCI) for merely indicates that ADEQ notified evaluation of the energy and non-air SCR retrofits is typically about $200/ APS on July 13, 2007 that Cholla Units quality environmental impacts of kW, while the TCI estimates for Apache 1–4 were ‘‘Potentially Subject to compliance of the control technologies, and Cholla equaled or exceeded $250/ BART.’’ 8 The WRAP’s ‘‘Arizona BART any existing pollution control kW. Eligibility TSD’’ further explains that: technology in use at each of the sources, The commenter (NPS) noted that the and the remaining useful life of each [Cholla] Unit 1 is listed as potentially date BART Guidelines recommend use of the eligible as information shows that the source, to be generally reasonable and Manual if vendor data are not available. emissions unit was in service only 2 months consistent with the RHR and the BART The commenter conducted detailed cost prior to the cut-off date. Recommend Guidelines.14 However, we also found analyses of SCR using an approach that requesting additional supporting that the costs of control were not the commenter believes is similar to that documentation for final determination.9 calculated in accordance with the BART used by EPA in its evaluation of SCR on ADEQ received this additional Guidelines, and that the visibility the Colstrip power plant—using the cost documentation from APS in August impacts were not appropriately methodologies of the Manual and 2007 in the form of a document dated evaluated and considered. These relying on EPA’s Integrated Planning May 23, 1962, and entitled ‘‘Operating findings formed part of the basis for our Model (IPM) to reflect the most recent Notes For May 1962.’’ 10 This document disapproval of ADEQ’s NOX BART cost levels. The commenter observed indicates that, ‘‘[o]n Tuesday, May 1, determinations. that most of the ADEQ SCR cost 1962, unit [#1 was] placed into Comment: Several commenters estimates were based on TCI costs that commercial operation.’’ 11 After objected to EPA’s use of non-specific were relatively high ratios of the reviewing this documentation, ADEQ and undefined parameter levels for both reported direct capital costs (DCC). The concurred that Unit 1 was not BART the ‘‘cost-effectiveness’’ and ‘‘sufficient commenter indicated that according to eligible.12 visibility improvement’’ parameters in the Manual, the ratio of TCI to DCC is Following the close of the public rejecting ADEQ’s SIP. One commenter 141 percent, while ADEQ’s estimates comment period, we requested and were as follows: (Pacificorp) further noted that states • received from APS a copy of the cannot meet EPA’s specific targets At Apache, TCI is 179 percent of ‘‘Operating Notes For May 1962’’ along unless and until those targets are clearly DCC for both units and included $6 with additional information concerning defined. million in costs for each unit not the operation of Cholla Unit 1.13 We typically allowed by EPA. Response: The RHR and the BART • have placed these materials in the At Cholla, TCI is 258 percent of Guidelines do not require the DCC for all three units and included $11 docket and, based on our initial review, development of specific thresholds, but we believe this documentation is million in costs for Units 2 and 3 (each) rather require evaluation of each BART and $15 million for Unit 4 that are not sufficient to confirm ADEQ’s determination on a case-by-case basis determination that this unit is not typically allowed by EPA. for each source.15 We have not • BART-eligible. However, because this At Coronado, data were not established a specific cost threshold that question was not addressed in our sufficient to calculate these values. makes a particular control option BART proposed rulemaking, we are not taking The commenter asserted that this based on just a dollars per ton number, final action on it at this time. We intend supports EPA’s concern that control and there is not a specific target, either to address Cholla Unit 1’s BART costs submitted by the utilities either in terms of cost-effectiveness or included costs not typically allowed by visibility improvement, for ADEQ to 8 Exhibit 17 to Earthjustice Comments, WRAP EPA or were inadequately documented. BART Clearinghouse (Oct. 24, 2008). meet. All five factors must be evaluated Response: We appreciate the 9 ‘‘Supporting Documentation on Emissions Unit and weighed to determine the level of information provided by the National Bart Eligibility Analysis’’, section 5.1.2. Park Service and are in agreement that 10 Arizona Regional Haze SIP at page 155. 14 See 77 FR 42841. ADEQ’s cost estimates of SCR are 11 Id. 15 See, e.g., BART Guidelines, 40 CFR Part 51, overestimated. As indicated in our 12 Id. Appendix Y, section IV.D.5 (‘‘a 0.3 deciview 13 Email from Sue Kidd, APS, to Colleen improvement may merit a stronger weighting in one proposal, our cost estimates for SCR McKaughan, EPA (October 10, 2012, 9:17 a.m.) and case versus another, so one ‘‘bright line’’ may not generally produced lower values than attachments. be appropriate.’’) those in the Arizona Regional Haze SIP,

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00006 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72517

and at a level that we consider cost- relevant information for multiple NOX BART for each of the Cholla units. For effective. Although we recognize that control options, but does not provide Apache and Coronado, ADEQ NPS’s estimates produce even lower evidence that this information has been determined that the current control values than those from our proposal, we analyzed in any way. In the case of technology (hot-side ESPs) was BART have not updated our own cost Apache and Coronado, the Arizona and eliminated the most stringent estimates to reflect NPS’s comments Regional Haze SIP does not analyze this control technology (fabric filters). We since we already consider SCR to be cost information in even a qualitative note that PM emissions from EGUs cost-effective. We do note that in order manner. In the case of Cholla, the terms typically contribute only a small to address the comments from the ‘‘least expensive’’ and ‘‘most expensive’’ percentage of the modeled visibility utilities, we have performed are used, but only in the context of impact from EGUs, and that controlling supplemental cost analyses for each providing a reference for visibility their emissions results in very small facility based on the costs provided by impacts, and not in the context of an visibility benefit. For example, the utilities, and in doing so have evaluation of costs. This does not CALPUFF visibility modeling accounted for those costs not allowed by constitute ‘‘consideration,’’ as it performed by WRAP indicates that for CCM methodology. involves little more than ensuring the Apache, the maximum baseline PM10 Comment: Two commenters (ADEQ presence of cost values, with no visibility impact at the most affected and AUG) stated that EPA did not and judgment, analysis, or interpretation of Class I area (Chiricahua NM) is 0.04 cannot show that ADEQ failed to their meaning. dv.17 Assuming that a more stringent consider relevant cost information in Second, we disagree with the control technology could achieve 100 making its BART determination, the commenter’s characterization of our percent PM control and eliminate this State fully complied with its CAA disapproval as based on a ‘‘preference’’ entire visibility impact, a more stringent obligations, and EPA’s rationale is for a different format or form of PM10 BART determination would insufficient to reject ADEQ’s cost explanation for ADEQ’s BART therefore achieve, at most, a visibility determinations. AUG asserted that: determinations. As discussed in the benefit of 0.04 dv. Although ADEQ did previous paragraph, ADEQ has not not document its analysis or weighing of Arizona has expressly stated that it has considered each of the BART factors. EPA discussed its BART determination the five factors in arriving at the PM10 plainly cannot—and does not—demonstrate rationale, particularly with regard to BART determinations for Apache or that Arizona failed to take the costs of costs of compliance, in any format. Coronado, additional analysis would compliance with BART emission limits into While ADEQ’s RH SIP does include cost not have the potential to result in consideration. The state is required to do no information, it provides no explanation selection of a more stringent control more than that, and EPA cannot lawfully regarding how, or even if, this cost technology in light of the small disapprove the state’s determinations on the information was used in arriving at its potential for visibility benefit. basis that the Agency would prefer a different NO BART determinations. Although With regard to SO , ADEQ selected form of, or format for, explanation of those X 2 determinations. we agree that the RHR does provide the most stringent control technology states significant discretion in their (wet FGD) for all units at Apache, The commenters further stated that the consideration of the BART factors, Cholla, and Coronado. Although ADEQ other reason EPA rejected ADEQ’s cost AUG’s comment presupposes that these did not ‘‘take into account the most determinations is that EPA believed that costs were considered. The Arizona stringent emission control level that the ADEQ relied on inadequately Regional Haze SIP does not indicate that technology is capable of achieving,’’ documented costs. The commenters they were considered. correcting for this flaw would not have contended that there is nothing in the Comment: ADEQ noted that the same the potential to result in the selection of CAA or BART rules that requires a state principles were used for the PM10 and a more stringent control technology, to present any particular level of cost SO2 BART evaluations as were used for since wet FGD, which is the most documentation or that limits a state’s the NOX BART evaluation, yet EPA stringent control technology, was discretion in its consideration of the accepted the approach for only PM10 already selected as BART. Further cost factor in making a BART and SO2. discussion of our evaluation of ADEQ’s determination. Response: We disagree that we BART analyses for PM10 and SO2 is Response: We disagree with this accepted ADEQ’s approach for PM10 and provided below. comment. First, while Arizona may SO2. Although we did not disapprove Comment: The commenters stated have ‘‘expressly stated’’ that it ADEQ’s PM10 and SO2 BART that one of EPA’s reasons for rejecting considered each of the BART factors, it determinations, the absence of a ADEQ’s cost determinations is because must do more than ‘‘state’’ that it disapproval of these determinations the costs are inconsistent with the CCM. considered a BART factor, but must also should not be construed to represent The commenters noted that use of the provide some type of analysis acceptance of the approach by which outdated Manual is not required by the demonstrating that it considered the they were developed. We acknowledge CAA or the BART rules and provide BART factors.16 Although ADEQ has that ADEQ took a similar approach in its references in which EPA has stated that presented information relevant to each analyses for PM10 and SO2 as for NOX, the Manual is only one tool that can be of the BART factors, it has not provided and that these analyses exhibit the same used but that other cost data should also an explanation regarding how this deficiencies we have noted elsewhere be considered. information was used to develop its for the NOX BART determinations. Response: We partially agree with this BART determinations. Specifically in However, we did not disapprove the comment. We acknowledge that our the case of cost calculations, the PM10 and SO2 determinations because BART guidelines state, ‘‘In order to Arizona Regional Haze SIP includes we find that the shortcomings in these maintain and improve consistency, cost analyses did not result in unreasonable estimates should be based on the [CCM], 16 See, e.g., BART Guidelines, 40 CFR Part 51, BART determinations and therefore where possible’’ and that ‘‘[w]e believe Appendix Y, section IV.E.2. (‘‘You should provide were generally ‘‘harmless errors.’’ a justification for adopting the technology that you that the [CCM] provides a good select as the ‘best’ level of control, including an With regard to PM10, we note that explanation of the CAA factors that led you to ADEQ determined the most stringent 17 See Docket Item No. B–12, ‘‘Summary of WRAP choose that option over other control levels.’’) control technology (fabric filters) was RMC BART Modeling for Arizona.’’

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00007 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72518 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

reference tool for cost calculations, but equitable BART determinations across c. ADEQ’s Approach to Degree of if there are elements or sources that are states and across sources. Visibility Improvement not addressed by the Control Cost Comment: One commenter (SRP) Comment: Several commenters Manual or there are additional cost stated that ADEQ appropriately (American Coalition for Clean Coal methods that could be used, we believe considered the ‘‘dollars-per-deciview’’ Electricity (ACCCE), AEPCO, APS, that these could serve as useful cost-effectiveness of different control AUG, , PacifiCorp, SRP) 18 supplemental information.’’ The options, which is reasonable and asserted that EPA improperly dismissed Manual contains two types of entirely within the broad discretion ADEQ’s visibility impacts analyses. The information: (1) Study level cost afforded to the states under the CAA. commenters cited the BART Guidelines estimates of capital and operation and SRP stated that because BART is a (70 FR 39170, July 6, 2005) to assert that maintenance (O&M) costs for certain component of the CAA’s visibility there is no prescribed method for states specific types of pollution control program, it is more crucial to evaluate to consider and weigh visibility impacts equipment, such as SCR, and (2) a control costs in relation to the visibility and, thus, EPA has no legal grounds for broader costing methodology, known as improvements that may be expected disapproving a SIP based on the method the overnight method. We agree that the using a dollars per deciview ($/dv) the State has chosen to consider language of the BART Guidelines does metric. visibility impacts or improvements. The not require strict adherence to the study Response: The BART Guidelines commenters added that whatever EPA’s level equations and cost methods used require that cost-effectiveness be preference, it has no discretion to to estimate capital and O&M costs. calculated in terms of annualized substitute its method or its conclusion We consider the use of the broader dollars per ton of pollutant removed, or for those of the State. According to the costing methodology used by the CCM, $/ton, but also list the $/deciview ratio commenters, it is clear that the BART the overnight method, as crucial to our as an additional cost-effectiveness rules envision—or, at a minimum, ability to assess the reasonableness of measure that can be employed along allow—a visibility improvement the costs of compliance. Evaluation of with $/ton for use in a BART analysis that is focused on visibility the cost of compliance factor requires an evaluation.21 However, the $/dv metric impacts in the most impacted area. evaluation of the cost-effectiveness is only useful to the extent that it Regarding ADEQ’s BART associated with the various control reflects appropriately calculated costs determination at Coronado in particular, options considered for the facility. A and visibility benefits. As explained one commenter (SRP) noted that ADEQ proper evaluation of cost-effectiveness elsewhere in this document, we have evaluated a visibility index derived allows for a reasoned comparison not determined that ADEQ did not evaluate from an average of modeled visibility only of different control options for a costs and visibility benefits in a manner improvements at the nine Class I areas given facility, but also of the relative consistent with the RHR and the BART closest to Coronado. The commenter costs of controls for similar facilities. If Guidelines. Therefore, while ADEQ asserted that this approach was well the cost-effectiveness of a control certainly had the discretion to take $/dv within the State’s discretion to assess technology for a particular facility is into consideration as part of its BART visibility under the BART rules. outside the range for other similar analyses, the values that it relied upon Another commenter (AUG) argued this facilities, the control technology may be in doing so were not reasonable. consideration of an average visibility rejected as not cost-effective.19 In order impacts index is an even more thorough for this type of comparison to be b. ADEQ’s Approach to Energy and type of evaluation than that required by meaningful, the cost estimates for these Non-Air Quality Environmental Impacts the BART rules. One commenter (AEPCO) added that facilities must be performed in a Comment: One commenter (SRP) EPA’s proposal to disapprove ADEQ’s consistent manner. Without an ‘apples- stated that EPA inappropriately NO BART determinations was largely to-apples’ comparison of costs, it is downplayed the energy and non-air X based on its concern with ADEQ’s impossible to draw rational conclusions quality factor in its review of ADEQ’s reliance on the Western Regional Air about the reasonableness of the costs of BART analysis. Another commenter compliance for particular control Partnership (WRAP) modeling. (ADEQ) noted that because fly ash By contrast, another commenter options. Use of the CCM methodology is ammonia residues have the potential to intended to allow a fair comparison of asserted that since the facilities’ contaminate ground and surface waters, modeling results indicated that controls pollution control costs between similar ADEQ included potential environmental applications for regulatory purposes. would contribute to visibility impacts and the economics of disposing improvements in multiple Class I areas, This is why the BART guidelines the fly ash in its BART analysis. specify the use of the CCM where ADEQ should consider these benefits possible 20 and why it is reasonable for Response: We do not agree that we rather than looking at the benefits in us to insist that the CCM methodology inappropriately downplayed the energy only a single Class I area. The be observed in the cost estimate process. and non-air quality environmental commenter believes that overlooking However, we note that the overnight impacts factor in our review of ADEQ’s significant visibility benefits in this way method has been used for decades for BART analyses. ADEQ provided only considerably understates the overall regulatory control technology cost brief consideration of this factor in its benefit of controls to improved analyses, and that its use ensures BART analyses and did not explain how visibility. The commenter contended it weighed this factor against the other that the procedure followed by ADEQ is statutory factors. Because ADEQ’s 18 BART Guidelines, 40 CFR Part 51, Appendix Y, not a sufficient basis for making BART section IV.D.4.a. analysis of this factor was limited in determinations for sources with 19 See Id. section IV.D.4.f (‘‘A reasonable range [of scope, our evaluation of this factor in substantial benefits across many Class I cost-effectiveness values] would be a range that is reviewing the SIP was similarly limited. areas. consistent with the range of cost-effectiveness We discuss our analysis of this factor in Response: EPA’s proposed values used in other similar permit decisions over a period of time.’’) our FIP action below. disapproval of ADEQ’s NOx BART 20 BART Guidelines, 40 CFR Part 51, Appendix Y, determinations was not based on any section IV.D.4. 21 BART Guidelines sections IV.D.4.c and IV.E. concern with the WRAP modeling

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00008 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72519

protocol, upon which ADEQ relied for in determining the appropriate level of BART for each BART-eligible source in its BART analyses. On the contrary, we controls, EPA does not agree that these the State that emits any air pollutant found that the modeling procedures benefits may be ignored.26 which may reasonably be anticipated to relied upon by ADEQ were ‘‘in accord While there may be no single cause or contribute to any impairment with EPA guidance.’’ 22 However, we prescribed method to consider and of visibility in any mandatory Class I noted that ADEQ’s use of the results of weigh visibility impacts, the BART Federal area.’’ 27 The BART Guidelines modeling in making BART decisions Guidelines do require that certain explain that, ‘‘[i]f the emissions from the was ‘‘problematic in several visibility impacts be included in the list of emissions units at a stationary respects.’’ 23 In other words, our concern considering and weighing. EPA source exceed a potential to emit of 250 with the visibility analysis was not with disagrees that state flexibility extends to tons per year for any visibility-impairing the technical adequacy of the modeling categorically excluding consideration of pollutant, then that collection of itself, but rather with how ADEQ visibility improvements occurring at emissions units is a BART-eligible interpreted the results of this modeling. multiple Class I areas. Considering source.’’ 28 Therefore, it is that In its BART analyses for Apache and benefits at multiple areas does not collection of units for which one must Cholla, ADEQ considered visibility necessarily require use of the make a BART determination. The improvements only at the single Class I ‘‘cumulative’’ improvement approach Guidelines state ‘‘you must conduct a area with the greatest modeled impact (i.e., the direct sum of improvements at visibility improvement determination from a facility. This neglects all the areas), but does require that for the source(s) as part of the BART improvements that would occur at other improvements at those areas be taken determination. * * *’’ 29 This requires nearby Class I areas, and in general is into account in some way. For example, consideration of the visibility not adequate for assessing the overall one could simply list visibility improvement from BART applied to the visibility benefit from candidate BART improvements at the various areas, and facility as a whole. controls. As noted by commenters, the qualitatively weigh the number of areas The RHR and the Guidelines do not BART Guidelines provide that, ‘‘[i]f the and the magnitudes of the preclude consideration of visibility highest modeled impacts are observed at improvements. However, ADEQ did not improvement from controls on the nearest Class I area, [a State] may do this for any of the sources covered individual units, but that would be in choose not to analyze the other Class I by this action. addition to considering the areas any further and additional With respect to ADEQ’s consideration improvement from the whole facility. analyses might be unwarranted.’’ 24 of visibility improvements for The BART Guidelines clearly allow for Commenters argued that this language Coronado, EPA agrees that average the consideration of technical feasibility shows that Arizona’s exclusive focus on visibility index used by ADEQ could be and cost-effectiveness on a unit-by-unit improvements at a single Class I area is acceptable in itself as part of assessing basis where appropriate, but those allowed under the BART Guidelines. multiple area impacts and considerations fall under different However, this language is not intended improvements; indeed it is a variant of factors than the assessment of the degree as an invitation for states to ignore the cumulative improvement approach. of visibility improvement, and do not significant visibility improvements at However, without any consideration of remove the obligation to consider multiple Class I areas. Rather, it is particular area improvements, the visibility improvement from BART intended to provide a way of averaging process causes especially applied to the facility as a whole. In streamlining a complex and difficult large benefits at some individual areas sum, while the State has some flexibility modeling exercise where ‘‘an analysis to be diluted or lost, effectively in choosing a specific procedure to may add a significant resource burden discounting some of the more important consider these cumulative area and to a State.’’ 25 For example, when the effects of the controls. In addition, the multiple unit benefits, when such visibility benefits at the most impacted approach is counter to ADEQ’s benefits are significant, it is not Class I area alone are sufficient to justify emphasis elsewhere in the SIP on the reasonable to ignore them altogether as the selection of the most stringent importance of considering the visibility ADEQ did. control technology as BART, then improvement at the single area having Comment: One commenter (NPS) analysis of additional areas would be the largest impact from a given facility. agrees with EPA that the ammonia unnecessary and the state could Finally, ADEQ provided no discussion background concentration assumed by conserve resources by not modeling the of how the results of the visibility index ADEQ for Cholla and Coronado may be impacts on those additional areas. Here, were weighed against the other BART too low, ranging from 1 part per billion by contrast, ADEQ did not perform its factors. (ppb) down to 0.2 ppb. According to the In addition, ADEQ considered own modeling at all, but instead relied commenter, EPA guidance recommends visibility improvements from controls on modeling performed by contractors the use of a 1 ppb ammonia background on only a single emitting unit at a time, for the facilities. This modeling for areas in the west, absent compelling despite the fact that each of the three indicated that the installation of more evidence to the contrary. sources has multiple BART-eligible stringent controls (i.e., SNCR or SCR) Other commenters (APS and AUG) units. This neglects the full would result in visibility benefits at state that the Interagency Workgroup on improvement that would result from multiple Class I areas, yet ADEQ chose Air Quality Modeling (IWAQM) controls on the facility, with the to consider the benefits only at the most recommended value of 1ppb is outdated potential for dismissing emitting unit and should not be used now that better impacted area. Where, as here, the benefits that are individually small, but benefits of controls have been modeled data have been gathered and since the that collectively could have a significant CALPUFF model was updated to allow for a number of surrounding areas and visibility benefit. The RHR requires RH consideration of these benefits is useful for monthly, rather than yearly, average SIPs to include a ‘‘determination of ammonia concentrations. APS also 22 77 FR 42841. noted that EPA Region 9 has explicitly 26 See, e.g., 76 FR 52388, 52430 (San Juan 23 Id. Generating Station); 77 FR 51620, 51631–51632 27 24 BART Guidelines, 40 CFR Part 51, Appendix Y, (Four Corners Power Plant); and 77 FR 51915, 40 CFR 51.308(e)(1)(ii). section IV.D.5. 51922–51923 (Roseton and Danskammer Generating 28 40 CFR Part 51, Appendix Y, section II.A.4. 25 See 70 FR 39126. Stations). 29 Id. section IV.D.5.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00009 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72520 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

approved the use of the same monthly- recommended by the commenter.32 APS, the amount of measured varying background ammonia Since this method accounts only for ammonium is comparable to and at concentrations, which were based on ammonium, and not remaining free times much greater than the amount of actual field measurements, in running gaseous ammonia, the total ammonia ammonia. the CALPUFF model for two other sites originally available to form visibility- New ammonia monitoring data were located close to Cholla and that were impairing compounds may actually be collected by SRP at several sites used by ADEQ in its analysis. These higher. Because of uncertainty in the between NGS and the two nearest Class values range from 1 ppb during the ‘‘back-calculation’’ method, and I areas, Capitol Reef National Park and summertime to 0.2 ppb during cold criticism of it, EPA relied on it in the Grand Canyon National Park, from winter months. EPA has also stated in FCPP FIP only as corroboration for the December 2009 through April 2010. The response to comments on the Montana IWAQM default of 1 ppb.33 monitoring report,35 cited by regional haze FIP (77 FR 57864, Nevertheless, it supports the idea that commenter APS, describes a September 18, 2012) that ‘‘it is winter ammonia levels in the Class I surprisingly high spatial variability in preferable to use ambient ammonia areas affected by emissions from sources ammonia concentrations. The two measurements when such data are in Arizona are likely substantially monitors in the Cameron area south of available rather than using default higher than 0.2 ppb. NGS (and east-southeast of the Grand background ammonia concentrations.’’ EPA agrees with commenters that it Canyon) showed consistent Another commenter (Navajo Nation) would be preferable to use actual concentration differences despite being agrees that EPA should use actual, monitoring data to determine less than five miles from each other; this recorded data wherever possible, background ammonia concentrations. may be due to relatively localized However, much of the existing data especially ammonia background values. ammonia sources. These sites also cited by the commenters is from other AUG concludes that EPA has no basis showed consistently lower states, and so is unlikely to be for rejecting the use of refined measurements than the Halls Crossing representative for evaluating visibility background ammonia concentration site, north of NGS (and southwest of impacts at Arizona’s Class I areas. values in disapproving the SIP. Capitol Reef). The range in Further, the data comprises only 30 concentrations was comparable to the Response: The IWAQM Guidance is ammonia itself, and not ammonium; or range seen between the AECOM values the only guidance available for choosing if it does include ammonium, that is not at the low end, and EPA’s back- ammonia background concentrations. cited by the commenters. Visibility- calculated values at the high end. Because of the paucity of monitoring impairing ammonium sulfate and Unfortunately, because of the variability data and the uncertainty in other ammonium nitrate are formed from and its unknown causes, the data ammonia estimation methods, EPA ammonia, SO2, and NOX. Therefore the collected did not lead to a clear picture concludes that it is appropriate to use ammonium represents part of the pool of appropriate and representative the default 1 ppb from the IWAQM of ammonia that could be available to ammonia background concentrations to Guidance. interact with the SO2 and NOX from a use with CALPUFF. As stated by the commenter, EPA did facility and contribute to visibility In any case, as mentioned above, originally accept monthly varying impacts, and should be accounted for in some nearby monitored data reported in ammonia values of 0.2 to 1.0 ppb for estimating ammonia background Sather’s paper show considerably higher BART analyses performed by AECOM concentrations. In several of the ammonia than recommended by some 34 for APS for the Four Corners Power research papers cited by commenter commenters, so it is not clear that Plant (FCPP), and by SRP for the Navajo values lower than 1 ppb should be used. 32 See, e.g., Proposed Rule: Source Specific Generating Station (NGS). However, Federal Implementation Plan for Implementing Best EPA concludes that there is not a shortly after that, the USDA Forest Available Retrofit Technology for Four Corners compelling case for using ammonia Service brought to EPA’s attention Power Plant: Navajo Nation Technical Support background concentrations other than ammonia monitored in the Four Corners Document, pages 59–61, 65–66, 68–73. the 1 ppb found in the only area showing concentrations up to 3 33 Id. at page 68. authoritative guidance document 34 31 RoMANS—Rocky Mountain Atmospheric ppb, described in a journal paper by Nitrogen & Sulfur Study, William C. Malm and available on this topic and supported by Mark Sather and others. EPA and the Jeffrey L. Collett. National Park Service, CSU–CIRA, the FLMs. Forest Service also estimated ammonia Fort Collins, CO. ISSN 0737–5352–84. October Comment: Two commenters (APS and concentrations by ‘‘back calculating’’ 2009. http://www.nature.nps.gov/air/Studies/ AUG) noted that the RHR and BART romans.cfm. Table 3.9 on p.3–38 shows ammonium the amount of ammonia needed to form comparable to or about half of ammonia, depending Guidelines are silent regarding whether the ammonium nitrate and ammonium on measurement method. It also shows that the visibility improvements should be sulfate collected at Arizona and New spring time mean and maximum ammonium are modeled on a unit-by-unit basis or a Mexico sites in the IMPROVE about 0.22 and 0.57 mg/m3, respectively, or 0.38 and plant-wide basis, and there is no legal 0.78 ppb; and the mean and maximum ammonia are monitoring network. This yielded about 0.38 and 1.0 mg/m3 or 0.51 and 1.4 ppb. The requirement that units be modeled concentrations ranging from 0.4 to 1.3 sum of these means and maxima is 0.81 and 2.2 aggregately. Given that visibility ppb, with winter values considerably ppb, respectively. Figure 4.26 on p.4–26 shows benefits are approximately additive, the higher than the AECOM 0.2 ppb daily sums of ammonium and ammonia, with commenters contend that it is values of 2.5–5 mg/m3 or 3.6–7.2 ppb occurring frequently. These are substantially higher than unreasonable for EPA to conclude that 30 Interagency Workgroup On Air Quality values cited by the commenters. ‘‘NH3 Monitoring Modeling (IWAQM) Phase 2 Summary Report And in the Upper Green River Basin, Wyoming’’, by John Kreidenweis & Jeffrey L. Collett Jr. Journal of the Air Recommendations For Modeling Long Range V. Molenar, H. James Newell, Jeffrey Collett, et al. & Waste Management Assoc. vol.54, issue 5, 2004, Transport Impacts (EPA–454/R–98–019), EPA Extended Abstract #70, A&WMA Specialty pages 585–592. DOI:10.1080/ OAQPS, December 1998, http://www.epa.gov/ Conference ‘‘Aerosol & Atmospheric Optics: Visual 10473289.2004.10470927, Table 1 p. 587 shows scram001/7thconf/calpuff/phase2.pdf. Air Quality and Radiation’’, Moab, Utah, 28 April— ammonium about four times as high as ammonia. 31 Mark E. Sather et al., ‘‘Baseline ambient 2 May 2008, p.3 Figure 1 and p.4 Figure 3 show 35 ‘‘Measurements of Ambient Background gaseous ammonia concentrations in the Four ammonium comparable to ammonia in summer and Ammonia on the Colorado Plateau and Visibility Corners area and eastern Oklahoma, USA’’. Journal far greater in winter. ‘‘Aerosol Ion Characteristics Modeling Implications’’, Salt River Project, Dr. Ivar of Environmental Monitoring, 2008, 10, 1319–1325, During the Big Bend Regional Aerosol and Visibility Tombach, Consultant, and Robert Paine, AECOM DOI: 10.1039/b807984f. Observational Study,’’ Taehyoung Lee, Sonia M. Environment, September 2010.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00010 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72521

ADEQ’s BART analyses failed to 3 does not comply with the RHR’s these units.37 Therefore, we proposed to consider any significant visibility effect requirements. As discussed further approve ADEQ’s determination that merely because ADEQ modeled the below, we performed a supplemental BART for PM10 at Apache Units 2 and units separately. In addition, AUG notes analysis using the version of AEPCO’s 3 is upgrades to the existing ESPs and that it is necessary to determine the cost estimate that adheres to our a PM10 emissions limit of 0.03 lb/ effects of emissions from units assumptions regarding costs that are MMBtu. individually so that projected visibility allowed by the CCM (i.e., capital costs One commenter asserted that this impacts can be considered in light of for the installation of SCR with LNB and limit is too lenient, since other coal- costs and other impacts associated with OFA of $164.9 million), and we also fired units are achieving lower limits, BART-candidate controls for that considered the fact that AEPCO is a based on test data submitted by various particular unit, and modeling units small entity under the Regulatory utilities to EPA as part of an Information together could obscure these Flexibility Act.36 Collection Request (ICR) for the Mercury comparisons. and Air Toxics (MATS) Rule.38 EPA Response: Considering the visibility b. ADEQ’s BART Analysis and disagrees with this comment. The benefits of multiple units together does Determination for PM10 MATS Rule establishes an emission not preclude a state from also Comment: One commenter (NPS) standard of 0.030 lb/MMBtu filterable considering individual unit benefits, as agreed with ADEQ and EPA that BART PM (as a surrogate for toxic non- well as individual unit costs. EPA does for PM10 at Apache Units 2 and 3 is mercury metals) as representing not agree that modeling the units upgrades to the existing electrostatic Maximum Achievable Control together obscures these other precipitators (ESPs) and a PM10 Technology (MACT) for coal-fired comparisons. Rather, the benefit of emissions limit of 0.03 lb/MMBtu. The EGUs.39 This standard derives from the controls for an entire BART-eligible commenter noted that ADEQ stated that average emission limitation achieved by source is a factor that should be PM10 emissions would be measured by the best performing 12 percent of considered along with those other conducting EPA Method 201/202 tests. existing coal-fired EGUs (taking into comparisons. In any case, whether In contrast, a second commenter account the variability in the testing considered unit by unit or all units (Earthjustice) disagreed with EPA’s results for these facilities), based upon together, visibility improvement has no proposal to approve ADEQ’s PM10 to the same test data referred to by the effect on the assessment of cost- BART determination for Apache Units 2 commenter.40 The BART Guidelines effectiveness as measured by dollars per and 3. The commenter contended that provide that, ‘‘unless there are new ton of reductions. EPA proposed to approve the BART technologies subsequent to the MACT determination despite acknowledging standards which would lead to cost- B. Comments on ADEQ’s Individual that ADEQ did not conduct a full BART effective increases in the level of BART Analyses and Determinations analysis for PM10 because it control, you may rely on the MACT 1. ADEQ’s BART Analyses and overestimated costs and failed to standards for purposes of BART.’’ 41 Determinations for Apache Unit 1 consider upgrades to the existing ESPs. Therefore, we are approving ADEQ’s determination that a PM limit of 0.03 Comment: One commenter (NPS) However, the commenter believes that 10 lower emission rates are achievable and, lb/MMBtu represents BART for these concurred with ADEQ’s and EPA’s units. proposals for BART at Apache Unit 1. as a result, that EPA should disapprove Response: We acknowledge NPS’s ADEQ’s BART determination, conduct a c. ADEQ’s BART Analysis and full five-factor BART analysis and set a concurrence. Determination for SO2 lower emission limit as BART for PM10. 2. ADEQ’s BART Analyses and According to the commenter, the Sahu Comment: One commenter Determinations for Apache Units 2 and report demonstrates that nearly 150 (Earthjustice) disagreed with EPA’s 3 EGUs across the nation with a variety of proposal to approve ADEQ’s SO2 BART PM controls achieve emission rates determination at Apache Units 2 and 3. a. ADEQ’s BART Analysis and The commenter states the approval is Determination for NOX lower than 0.03 lb/MMBtu. The commenter asserted that neither ADEQ contrary to the RHR because ADEQ’s Comment: One commenter nor EPA provided any explanation why BART determination is not supported (Earthjustice) commended EPA’s Apache Units 2 and 3 could not by a valid five-factor analysis. The commenter states that EPA cannot decision to disapprove ADEQ’s NOX similarly meet a lower emission limit. BART determination for Apache Units 2 Response: As we noted in our speculate that it would reach the same and 3. The commenter stated that EPA proposal, ADEQ’s BART analysis did conclusion as ADEQ, and it must correctly concluded that ADEQ’s BART not demonstrate that all potential undertake an independent full five- determination for NOX inflated the costs upgrades to the existing ESPs at Apache 37 of more-stringent NOX controls by Units 2 and 3 were fully evaluated or 77 FR 42847. 38 Information Collection Request For National including costs not allowed by EPA Cost that the costs were calculated in Control Manual, provided little Emission Standards For Hazardous Air Pollutants compliance with the Control Cost (NESHAP) for Coal- And Oil-Fired Electric Utility reasoning about the visibility benefits of Manual. However, we concluded that Steam Generating Units (OMB Control No. additional NOX controls, and did not this was a harmless error because of the 2060-0631). See http://www.epa.gov/ttn/atw/utility/ utilitypg.html for detailed information obtained weigh the visibility impacts at all relatively small visibility improvement nearby Class I areas. The commenter through this ICR. associated with PM10 reductions from 39 77 FR 9304, 9450, 9458 (February 16, 2012) asserted that because ADEQ’s BART (codified at 40 CFR 60.42Da(a), 60.50Da(b)(1)). analysis does not comply with the 36 As explained in our proposal, a firm primarily 40 See Memorandum from Jeffrey Cole (RTI RHR’s requirements, EPA must engaged in the generation, transmission, and/or International) to Bill Maxwell (EPA) regarding disapprove ADEQ’s BART distribution of electric energy for sale is small if, ‘‘National Emission Standards for Hazardous Air determinations for Apache Units 2 and including affiliates, the total electric output for the Pollutants (NESHAP) Maximum Achievable Control preceding fiscal year did not exceed 4 million Technology (MACT) Floor Analysis for Coal- and 3. megawatt hours. 77 FR 42867. AEPCO sold under Oil-fired Electric Utility Steam Generating Units for Response: We agree that ADEQ’s 3 million megawatt hours in 2011 and is therefore Final Rule’’ (Dec. 16, 2011). BART analysis for Apache Units 2 and a small entity. 41 40 CFR Part 51, Appendix Y, Section IV.C.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00011 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72522 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

factor BART analysis. The commenter requirement would not be appropriate pollution control practices for argues that an SO2 limit of 0.04 lb/ for these units. minimizing emissions’’ at all times. We MMBtu is achievable and cost-effective While new wet scrubbers are capable expect that this requirement will help to for Apache Units 2 and 3 according to of achieving 95 percent or better ensure that the scrubbers on Apache 42 the Sahu report. The commenter further removal of SO2, the Apache scrubbers Units 2 and 3 are properly maintained asserts that, based on this report, were manufactured in the 1970s and and operated under all conditions. scrubber upgrades can achieve SO2 designed to meet a limit of 0.8 lb/ removal efficiencies of 98 percent and MMBtu (i.e., a control efficiency of up 3. ADEQ’s BART Analyses and should have been investigated. to 70 percent).43 For such existing Determinations for Cholla Units 2, 3 and Another commenter (NPS) noted that scrubbers achieving greater than 50 4 that AEPCO’s BART reports indicate percent control, the BART Guidelines Comment: One commenter (APS) that uncontrolled SO2 emissions are (which are not mandatory for these remarked that EPA stated that APS’s 0.69 lb/MMBtu, and that the ADEQ units) do not provide a presumptive contractor did not provide supporting BART proposal would reduce SO2 limit or removal efficiency, but information for its capital cost estimate, emissions by 78 percent down to 0.15 recommend consideration of cost- such as detailed equipment lists. The lb/MMBtu. Based on the SO2 control effective scrubber upgrades designed to commenter argues that detailed data submitted by the commenter, the improve the system’s overall SO2 equipment lists are typically not commenter asserted that other BART removal efficiency.44 In August 2009, necessary for the level of accuracy upgrades are achieving higher removal AEPCO provided information to ADEQ needed for the process selection phase efficiencies and/or lower SO2 limits. concerning potential scrubber upgrades of a project and noted that its contractor 45 The commenter believes that the at Apache Units 2 and 3. AEPCO used vendor quotes for the major pieces existing scrubbers can be upgraded to noted that it was in the process of of equipment and factors for achieve better removal efficiency and upgrading its limestone grinding system construction, balance of plant, lower emission rates than the 78 percent and described other potential upgrades, electrical, owner’s costs, surcharges, and 0.15 lb/MMBtu proposed by EPA. such as improving operation of the AFUDC and contingency. The commenter cited various examples scrubber bypass damper system, Response: We do not agree with this of upgraded scrubbers achieving limits upgrading the mist eliminator wash comment. The BART Guidelines of less than 0.15 lb/MMBtu or removal system, adding another sieve tray, and provide that: rates of greater than 90 percent. modifying the flue gas inlet. The By contrast, ADEQ and AEPCO enclosed ‘‘Wet FGD Implementation You should include documentation for any expressed opposition to both a lower Plan’’ indicated that AEPCO intended to additional information you used for the cost limit and a removal efficiency proceed with upgrading the limestone calculations, including any information supplied by vendors that affects your requirement. ADEQ asserted that ‘‘the grinding system, improving operation of limits included in the state SIP assumptions regarding purchased equipment the scrubber bypass damper system, and costs, equipment life, replacement of major submittal are acceptable as BART’’ and upgrading the mist eliminator wash components, and any other element of the ‘‘imposing dual-limitations will be system, but that ‘‘[t]he remaining wet calculation that differs from the [CCM].47 unnecessary and burdensome for the FGD options were not selected on the facility.’’ AEPCO commented that ADEQ basis of low probability of successfully Thus, detailed cost documentation is necessary to the extent that cost permit conditions, which require SO2 making a significant difference in absorption systems to be operated and scrubber performance and/or high assumptions differ from the CCM. In this case, several of ADEQ’s and APS’s maintained at all times in a manner cost.’’ 46 consistent with good air pollution Based on this information, we cost assumptions for control costs at control practices for minimizing conclude that no further cost-effective Cholla differed from the CCM, but no emissions, is sufficient, and an scrubber upgrades are likely to be such documentation was provided as additional control efficiency limit is not feasible for this facility and we are part of the Arizona Regional Haze SIP. necessary. An efficiency limit would therefore deferring to ADEQ’s a. BART Analysis and Determination for also require modification to the determination that 0.15 lb/MMBtu NOX monitors to include the capability to represents BART for these units. Given measure scrubber inlet SO2 in addition the age of these scrubbers, we find that Comment: One commenter to stack emissions, which would require an additional removal efficiency (Earthjustice) commended EPA’s additional capital and O&M requirement would be unnecessarily decision to disapprove ADEQ’s NOX expenditures. burdensome. This approach is BART determination for Cholla Units 2, Response: We proposed to approve consistent with our consideration of 3 and 4. The commenter stated that EPA ADEQ’s determination that BART for AEPCO’s status as a small entity in our correctly concluded that ADEQ’s BART SO2 at Apache Units 2 and 3 is upgrades FIP determination. We note that our determination for NOX inflated the costs to the existing scrubbers with an final FIP includes a requirement to of more-stringent NOX controls by associated emission limit of 0.15 lb/ maintain and operate air pollution including costs not allowed by the MMBtu (30-day rolling average). control equipment at all units in ‘‘a Manual, and substantially However, we also solicited comment on manner consistent with good air underestimated the visibility benefits of whether an efficiency requirement additional NOX controls. The should be part of the BART 42 See BART Guidelines, 40 CFR Part 51, commenter asserted that because requirement, since Apache has the Appendix Y, section IV.E.4. ADEQ’s BART analysis does not comply ability to use coal from various sources 43 See Apache Title V Permit Technical Support with the RHR’s requirements, EPA must that have varying sulfur content. After Document (2007), Table 9; Title V Permit (2007), disapprove ADEQ’s BART reviewing the comments received on Attachment B, section II.E.1.a. determinations for Cholla Units 2, 3 and 44 See BART Guidelines, 40 CFR Part 51, our proposal, we have concluded that Appendix Y, section IV.E.4. 4. the emission limit set by ADEQ 45 Letter from Michelle Freeark, AEPCO, to Trevor appropriately reflects BART for SO2 at Baggiore, ADEQ (July 8, 2009). 47 40 CFR Part 51, Appendix Y, section IV.4.a., these units and that a removal efficiency 46 Id. note 15.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00012 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72523

Response: As explained in our 0.015 lb/MMBtu represents what can be analysis required by the RHR, although proposal and elsewhere in this continuously achieved with a properly EPA identified nothing in the docket to document, we agree that ADEQ’s BART operated baghouse on these units. The support its claim that a full BART analyses and determinations for NOX at fabric filters (i.e., baghouses) at Cholla analysis would have yielded the same Cholla Units 2, 3 and 4 do not comply will all be new since they are scheduled result. The commenter states that EPA with the requirements of the CAA and to be installed between 2008 and 2016. cannot speculate that it would reach the RHR. We are therefore disapproving Recent PSD BACT limits for coal-fired same conclusion as ADEQ, and it must these determinations. EGUs with new baghouses have undertake an independent full five- typically ranged from 0.01 to 0.015 lb/ b. BART Analysis and Determination for factor BART analysis. MMBtu using Method 5. The commenter further stated that PM10 As to the commenter’s position that ADEQ’s SO2 BART analysis for Cholla Comment: One commenter (NPS) bag material selection would influence Units 2, 3 and 4 is also flawed because agreed with EPA’s proposal to approve the level of PM that could be achieved, ADEQ failed to analyze controls and ADEQ’s BART determination for Cholla EPA notes that there are a number of upgrades that would result in emission Units 2, 3 and 4 of an emission limit of factors that influence a utility’s rates lower than the BART Guidelines’ 0.015 lb/MMBtu for PM10 based on the selection of proper bag material such as presumptive BART limits. According to use of fabric filters, the most stringent bag life, compatibility with exhaust gas the commenter, EPA has recognized control technology available. In stream and control of other pollutants multiple times that the presumptive contrast, a second commenter such as mercury (Hg) or sulfuric acid BART limits are merely the starting (Earthjustice) disagreed with EPA’s mist (H2SO4). In addition, it should be point for the BART determination, not proposal to approve ADEQ’s PM BART noted that the latest revision to the EGU the ending point. Moreover, the determination for Cholla Units 2, 3 and NSPS requires modified units to meet a commenter asserted that the 4. The commenter contended that EPA 48 PM limit of 0.015 lb/MMBtu. Also, as presumptive limits are often outdated proposed to approve the BART noted above, the recent EGU MATS rule with the result that appropriate determination despite acknowledging sets a PM emissions standard of 0.03 lb/ consideration of the five statutory BART that ADEQ did not conduct a full BART MMBtu, and the BART Guidelines factors can result in far lower emission analysis for PM because fabric filters are provide that, ‘‘unless there are new rates than presumptive BART. The the most stringent PM control technologies subsequent to the MACT commenter cited statements by EPA technology available and ADEQ’s 0.015 standards which would lead to cost- Region 6 (76 FR 64186, 64203, October lb/MMBtu emission limit is effective increases in the level of ‘‘consistent’’ with other EGUs 17, 2011, regarding proposed actions on control, you may rely on the MACT Arkansas’ RH SIP) and EPA Region 9 (77 employing fabric filters (citing 77 FR standards for purposes of BART.’’ 49 42849). However, the commenter FR 51633 regarding the final RH FIP for Therefore, we are finalizing our the Four Corners Power Plant). believes that lower emission rates are proposed approval of ADEQ’s BART achievable with fabric filters and, as a Earthjustice also presented determination for PM10 at Cholla Units documentation that the commenter result, that EPA should disapprove 2, 3 and 4. ADEQ’s BART determination, conduct a believes to show that lower SO2 full five-factor BART analysis and set a c. ADEQ’s BART Analysis and emission limits are achievable and cost- lower emission limit as BART for PM10. Determination for SO2 effective at Cholla Units 2, 3 and 4. According to the commenter, the BART Comment: Citing various examples of According to the commenter, a report Guidelines’ exemption from a full five- submitted with the comments (the lower SO2 limits at other coal-fired 50 factor analysis for the most stringent units, one commenter argued that the ‘‘Stamper report’’) shows that a control technology is not applicable in existing scrubbers at Cholla can be proper BART determination for Cholla

this case because improvements to the upgraded to achieve lower emission would have found that 98 percent SO2 fabric filters are possible and a lower rates than the 0.15 lb/MMBtu proposed control efficiency achieving a 0.04 lb/ MMBtu emission limit is BART for the emission rate is thus achievable. by EPA. Based on the SO2 control data The latter commenter (Earthjustice) submitted by the commenter, the units, and that even with the less- stated that had EPA conducted the PM10 commenter asserted that other BART stringent 95 percent SO2 control BART analysis required by the RHR, it upgrades are achieving higher removal efficiency that is the basis of ADEQ’s would show that an emission rate lower BART determinations, ADEQ should efficiencies and/or lower SO2 limits. than 0.015 lb/MMBtu is BART for Another commenter (Earthjustice) have required an SO2 emission limit of Cholla. According to the commenter, an disagreed with EPA’s proposal to 0.10 lb/MMBtu because 0.15 lb/MMBtu limit does not reflect 95 percent SO2 expert report accompanying the approve ADEQ’s SO2 BART commenter’s submission (the ‘‘Sahu determination for Cholla Units 2, 3 and removal. report’’) demonstrates that upgrades to 4. The commenter states the approval is Another commenter (APS) noted that the fabric filters can achieve a lower contrary to the RHR because ADEQ’s the SO2 content of the coal source for emission limit and, moreover, that BART determination is not supported the Cholla plant is up to 3.0 lbs/MMBtu, nearly 100 EGUs across the nation with by a valid five-factor analysis, which the and the maximum rate of removal that a variety of PM controls achieve commenter believes had flaws in its cost will be continuously achievable after emission rates lower than 0.015 lb/ and visibility improvement analyses. the plant upgrades its scrubbers is 95 MMBtu. The commenter asserted that The commenter alleged that EPA percent. Therefore, the commenter neither ADEQ nor EPA provided any proposed to approve the SO2 BART asserts that 0.15 lb/MMBtu is the explanation why Cholla Units 2, 3 and determinations based on unsupported 4 could not similarly meet a lower speculation that the outcome would be 50 Attachment 1 to Earthjustice Comments, emission limit. the same if EPA performed the BART Technical Support Document to Comments of Response: We are finalizing our Conservation Organizations, Proposed Arizona Regional Haze Partial SIP Approval and Partial FIP approval of ADEQ’s PM10 BART 48 77 FR 9450 (February 16, 2012) (codified at 40 SO2 and NOX BART Determinations for Cholla determination at Cholla Units 2, 3 and CFR 60.42Da). Units 2, 3 and 4 (September 17, 2012), prepared by 4. We find that an emission limit of 49 40 CFR Part 51, Appendix Y, Section IV.C. Victoria Stamper.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00013 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72524 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

maximum achievable SO2 emissions As noted by APS, this limit would costs of compliance for the various limit. require a removal efficiency of 95 control options. Response: A number of commenters percent when these units are burning Similarly, another commenter indicated that lower emission levels are this ‘‘worst-case’’ (highest-sulfur) coal (Earthjustice) supported EPA’s being achieved at other sources with (i.e., 3.0 lb/MMBtu). Therefore, we are disapproval of ADEQ’s NOX BART wet FGDs and western coal. However, finalizing our approval of ADEQ’s BART determination for Coronado Units 1 and none of these examples are based on limit of 0.15 lb/MMBtu of SO2 for these 2. For the reasons discussed by the coal with as high a potential SO2 level units. commenter above for Cholla Units 2, 3 as the coal that is currently burned at However, we remain concerned that and 4, the commenter agreed with what Cholla. APS historically burned coal this worst case coal is not representative the commenter said was EPA’s from the McKinley mine located on the of the typical coal that APS will receive conclusion that all of ADEQ’s BART Navajo Reservation at the Cholla units. from the El Segundo and Lee Ranch determinations are fatally flawed in Following the closure of this mine, APS mines. APS’s current contract for this numerous respects (e.g., inflated costs obtained coal from various sources until coal indicates that the minimum sulfur and underestimated visibility benefits). the company signed a long-term content is equivalent to 1.88 lb/MMBtu Specific to Coronado, the commenter contract for coal from the El Segundo of SO2 for the El Segundo coal and 1.64 agreed that ADEQ failed to provide and Lee Ranch mines in New Mexico.51 lb/MMBtu of SO2 for the Lee Ranch detailed and verifiable cost information The sulfur content of coal from these Coal.57 When burning this lower-sulfur and to properly consider the costs of two mines is substantially higher than coal, the units would only need to compliance for each control option in Powder River Basin (PRB) coal and also achieve 90 to 92 percent control in order its BART analysis (citing 77 FR 42851). much higher than coal from the former to meet the BART limit of 0.15 lb/ In addition, the commenter indicated source, the McKinley mine.52 The MMBtu of SO2. While APS has stated that ADEQ failed to properly evaluate current coal contract for these units that the scrubbers on Cholla Units 2, 3 the visibility benefits of more-stringent indicates that the typical sulfur content and 4 will be able to continuously NOX controls at Coronado, used a novel of this coal is equivalent to 2.4 lb/ achieve a removal efficiency of 95 and unapproved metric to measure percent, the Arizona Regional Haze SIP visibility benefits, failed to consider MMBtu SO2 and can be as high as 3.0 lb/MMBtu.53 Given that the transition to does not include a requirement or cumulative visibility benefits across all this coal has already occurred and that procedures to ensure that the scrubbers affected Class I areas, and used incorrect company has entered into a contract to are operated and maintained to achieve background ammonia concentrations in continue purchasing this coal until this level of control. Therefore, in order its modeling. The commenter added that 2024, we consider emissions based on to ensure that these scrubbers are ADEQ also failed to explain how it this coal supply to ‘‘represent a realistic properly operated and maintained, evaluated the five statutory BART depiction of anticipated annual consistent with 40 CFR 51.308(e)(1)(v), factors and selected BART based on the emissions for the source.’’ 54 The RHR we are finalizing a removal efficiency factors. The commenter asserted that and the BART Guidelines do not require requirement for SO2 of 95 percent on a because ADEQ’s BART analysis does states to restrict or alter a facility’s 30-day rolling basis for Cholla Units 2, not comply with the RHR’s selection of the coal supply in order to 3 and 4. This requirement is explained requirements, EPA properly meet a specific limit. further under ‘‘Comments on disapproved ADEQ’s NOX BART APS’s comments on the proposal Enforceability Requirements in EPA’s determinations for Coronado. indicate that the company intends to BART FIP.’’ Response: We agree that ADEQ’s BART analysis for NOX at Coronado upgrade the existing SO2 controls at 4. ADEQ’s BART Analyses and Units 1 and 2 did not comply with the Unit 2 to a new wet flue gas Determinations for Coronado Units 1 desulfurization (FGD) system, identical requirements of the CAA and RHR. and 2 Comment: One commenter (SRP) to those already installed on Units 3 and a. ADEQ’s BART Analysis and stated that EPA must accept ADEQ’s 4.55 APS further explained that: Determination for NOX BART determination for NOX because it The coal source for [Cholla] is El Segundo Comment: One commenter (NPS) was a complete and thorough five-factor and Lee Ranch coal with an SO content of 2 agreed with EPA that ADEQ’s BART analysis conducted in accordance with up to 3.0 lbs/mmBtu.The maximum rate of the BART Guidelines and resulted in a removal that will be continuously achievable selection of LNB with OFA for after the scrubber upgrades * * * are Coronado is not adequately supported reasonable and appropriate performed is 95 percent. If compliance is for the following reasons: determination of NOX BART for determined on a 30-day rolling average basis, • ADEQ did not consider the typical Coronado. the maximum SO2 emission limit achievable visibility metrics of benefit at the area Response: We do not agree with this at Cholla on a continuous basis is, therefore, with maximum impact, nor benefits comment. As explained in the NPRM 0.15 lb/mmBtu. 56 summed over the areas. and elsewhere in this document, Given this information, EPA finds that • Using the default 1 ppb ammonia ADEQ’s BART determinations for NOX the ADEQ BART limit of 0.15 lb/MMBtu background concentration would have did not comply with the requirements of the RHR or the BART Guidelines. represents BART for SO2 at these units. increased estimated impacts and control benefits. Therefore, we are finalizing our 51 See ‘‘Additional APS Cholla BART response’’, • There is no weighing of the disapproval of these NOX BART Appendix B. visibility benefits and visibility cost- determinations, including the 52 See, e.g., ‘‘APS Cholla Unit 2 BART report’’, effectiveness for the various candidate determinations at Coronado Units 1 and Table 2–2. controls and the various Class I areas. 2. 53 See ‘‘Additional APS Cholla BART response’’, • ADEQ does not indicate whether it Appendix B, Section 6.2. b. ADEQ’s BART Analysis and 54 considered any cost thresholds to be BART Guidelines, 40 CFR Part 51, Appendix Y, Determination for PM10 section IV.D.4.d. reasonable or expensive in analyzing the 55 ‘‘Comments of Arizona Public Service Comment: One commenter (NPS) Company’’, page 27. 57 See ‘‘Additional APS Cholla BART response’’, agreed with EPA’s proposal to approve 56 Id. page 63. Appendix B, Section 6.2. ADEQ’s PM10 BART determination for

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00014 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72525

Coronado Units 1 and 2, noting that that than 0.03 lb/MMBtu. The commenter EPA’s request for comment on whether emissions of PM10 from Coronado Units asserted that neither ADEQ nor EPA a lower emission limit may be 1 and 2 are currently controlled by hot- provided any explanation why achievable when the units are burning side ESPs and that, in terms of the Coronado Units 1 and 2 could not a lower-sulfur coal, the commenter consent decree, SRP is required to similarly meet a lower emission limit. responded that it is inappropriate for optimize its ESPs to achieve a PM10 Response: EPA acknowledges that EPA to establish a BART limit that emission rate of 0.030 lb/MMBtu. ADEQ did not perform a rigorous five- would be premised on any restriction of Another commenter (SRP) stated that factor BART analysis for PM10 at SRP’s fuel supply. According to the EPA’s approval of the Arizona BART Coronado. However, a full five-factor commenter, this type of restriction determination for PM is reasonable 10 analysis would be very unlikely to would increase unit operating costs and and appropriate, believing it to be result in a change of control technology reduce operational flexibility, and EPA consistent with the CAA and supported for PM Modeling of visibility impacts 10. provides no technical record to support by the technical record in this from direct PM emissions has shown 10 consideration of this emissions rulemaking. The commenter does not very small impairment for EGU PM10 believe any upgrades to the ESPs are emissions in comparison to visibility reduction option. warranted as part of the BART impairment resulting from SO2 and NOX Another commenter (Earthjustice) determination, as SRP has in place a emissions. The existing hotside ESPs at disagreed with EPA’s proposal to plan to optimize performance of the Coronado Units 1 and 2 control PM10 by approve ADEQ’s SO2 BART existing equipment. The commenter greater than 98 percent. In addition, SRP determination. The commenter states noted that as part of the consent decree is required under a Consent Decree to the approval is contrary to the RHR between SRP and EPA for Coronado, optimize the performance of these ESPs because ADEQ’s BART analyses are not SRP is required to operate the ESPs ‘‘at and to meet a PM limit of 0.030 lb/ supported by a valid five-factor analysis. all times when the Unit it serves is in MMBtu as measured by Method 5.58 The The commenter states that EPA cannot operation to maximize PM emission consent decree also requires Coronado speculate that it would reach the same reductions, provided that such to install and conduct performance conclusion as ADEQ, and it must operation of the ESP is consistent with specification testing of a particulate the technological limitations, undertake an independent full five- matter CEMS (PMCEMS). factor BART analysis, which the manufacturers’ specifications, and good Installing the best control, a baghouse, commenter believes would result in a engineering and maintenance practices would result in a cost exceeding SO2 BART limit of 0.04 lb/MMBtu based for the ESP,’’ and this requirement also $100,000/ton of additional PM removed. on a 30-day rolling average. Earthjustice is reflected in Coronado’s current Title From a cost and visibility improvement V operating permit. standpoint, it is not justifiable to require further asserted that, according to the The commenter also noted that the replacement of controls that can achieve Sahu report, WFGD can achieve SO2 PM10 limit in the recently promulgated a reasonably low emission level on a removal efficiencies of 98 percent and MATS Rule will be more stringent than continuous basis. As noted previously, the use of low-sulfur coals, which can the PM10 limit proposed as BART. The 0.030 lb/MMBtu is the limit for further reduce SO2 emissions, also commenter indicated that it makes filterable PM in the recently issued EGU should have been investigated. sense for BACT to be more stringent MATS rule. Therefore, we are finalizing Response: EPA does not agree that we than BART, and it likewise is our approval of ADEQ’s BART should disapprove the ADEQ BART appropriate for the MATS requirements determination for PM at these units. to impose more stringent compliance 10 determination and set an emission limit obligations on utilities than a BART c. ADEQ’s BART Analysis and as low as 0.04 lb/MMBtu for SO2. EPA determination since MATS is intended Determination for SO2 does acknowledge that while burning some coals, such as from PRB, these to protect the public health from Comment: One commenter (NPS) limits can be achieved at new units hazardous air pollutants, while BART is noted that the consent decree between (though only achieved continuously aimed at aesthetic concerns that EPA and SRP requires installation of Congress intended the states to address wet flue gas desulfurization (WFGD) over longer than 30-day averages), but very gradually. systems on both Coronado units to EPA does not find that this limit would In contrast, a third commenter be consistently achievable at Coronado. achieve a 30-day rolling average SO2 (Earthjustice) disagreed with EPA’s removal efficiency of at least 95 percent Coronado receives its coal supply by rail proposal to approve ADEQ’s PM10 or a 30-day rolling average SO2 line and has access to various sources of BART determination for Coronado Units emissions rate of no greater than 0.080 coal including PRB, Colorado and New 1 and 2. The commenter contended that lb/MMBtu. The commenter added that Mexico coals. As mentioned previously, EPA proposed to approve the BART EPA proposed to approve ADEQ’s BART the RHR and the BART Guidelines do determination despite acknowledging SO emission limit of 0.08 lb/MMBtu not require emission limits to be set at that ADEQ did not conduct a full BART 2 (30-day rolling average) for Coronado a level that would restrict the flexibility analysis for PM because EGUs with 10 Units 1 and 2, which the commenter of EGUs to use available coals with ESPs elsewhere have BART limits of indicated would be consistent with the varying sulfur content. 0.03 lb/MMBtu. However, the more stringent limits on WFGD commenter believes that lower emission The consent decree between EPA and upgrades that the commenter has seen. SRP described in our proposal requires rates are achievable and, as a result, that One commenter (SRP) stated that installation of wet flue gas EPA should disapprove ADEQ’s BART EPA’s approval of ADEQ’s BART determination, conduct a full five-factor desulfurization (WFGD) systems (i.e., determination for SO is reasonable and BART analysis and set a lower emission 2 appropriate, believing it to be supported new scrubbers) at both units at limit as BART for PM 10. According to by the technical record. In response to Coronado by January 1, 2013. These the commenter, the Sahu report scrubbers are required to achieve either demonstrates that nearly 150 EGUs 58 Consent Decree in United States v. Salt River 0.080 lb/MMBtu of SO2 or 95 percent across the nation with a variety of PM Project, CV 08–1479–PHX–JAT (entered Dec. 19, reduction of SO2 across the FGD, both controls achieve emission rates lower 2008).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00015 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72526 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

over a rolling 30-day basis.59 ADEQ has emissions based upon actual emissions from anticipated future emissions, we selected 0.08 lb/MMBtu as the BART a baseline period.62 updated the baseline period for each emission limit for these units. We find This provision is consistent with the unit to ensure that it reflected these that this is an appropriate limit for these statutory requirement that each BART changes.67 units and are finalizing our approval of determination take into consideration With respect to Coronado Unit 2, we this determination. ‘‘any existing pollution control also took into account the federally- We also note that the recently technology in use at the source.’’ 63 enforceable emissions limits set by a promulgated EGU MATS rule, which While the Guidelines do not specify Consent Decree between the United uses an SO2 limit as an acceptable particular dates for this ‘‘baseline States and SRP, which was entered in surrogate for limiting the emissions of period’’ for BART analyses, in order to 2008.68 Again, this is consistent with hazardous acid gases, has set the limit ‘‘represent a realistic depiction of the BART Guidelines, which provide at 0.20 lb/MMBtu of SO2 for existing anticipated annual emissions for the that: 60 EGUs like Coronado Units 1 and 2. source’’ the baseline can account for When you project that future operating controls already installed on the source, C. General Comments on EPA’s BART parameters (e.g., limited hours of operation or, where appropriate, controls which or capacity utilization, type of fuel, raw FIP Analyses and Determinations are required to be installed in the near materials or product mix or type) will differ 1. Selection of Baseline Period future. from past practice, and if this projection has a deciding effect in the BART determination, Comment: Several commenters In many instances, the 2000–2004 time frame was used as a baseline then you must make these parameters or expressed disagreement with our assumptions into enforceable limitations. In general approach to the selection of period for BART determinations because this time frame reflected the absence of enforceable limitations, you baseline periods. One commenter (NPS) calculate baseline emissions based upon stated a general preference for the use of existing controls in use at BART sources continuation of past practice.69 a baseline period that represents pre- at the time BART analyses were performed, following the issuance of the Consistent with this provision, for control emissions, as advised in the Coronado we used the consent decree- BART Guidelines, to estimate baseline final BART Guidelines in 2005. In Arizona’s case, the initial BART mandated NOX emission limit of 0.08 emissions for the purpose of calculating lb/MMBtu in order to ensure that the the average cost-effectiveness of the analyses were performed in 2007, using baseline periods that varied by source: baseline emissions rate would represent complete control system (e.g., a realistic depiction of anticipated combustion controls plus SCR). The 2002–2007 for Apache; 2001–2003 for Cholla; and 2001–2003 for Coronado.64 annual emissions for Unit 2. commenter believes that this avoids any We note that such an ‘‘updated biasing of the calculations by sources These periods appear to reflect controls in existence at the time that these BART baseline’’ might not be appropriate in all that install combustion controls during analyses were performed. Our proposed instances. For instance, if it appeared the BART evaluation process. NPCA disapproval of certain aspects of that controls had been installed early in asserted that the ‘‘proper’’ baseline for Arizona’s BART determinations was not order to avoid a more stringent BART BART determinations is 2001–2004. based on any flaw in the choice of determination, it would presumably not ADEQ asserted EPA violated the RHR baseline period. be appropriate to use a baseline provision in 51.308(d)(2)(i), which However, having proposed to representing these new controls. We specifies the period for establishing disapprove Arizona’s BART find no evidence of such intent here. baseline visibility conditions as 2000– determinations for NO on other Rather, with respect to Coronado, the 2004, by using the period between 2008 X grounds, we were obligated to conduct installation of new NOX and SO2 and 2011 as a baseline period for EPA’s our own five-factor BART analyses for controls was required by a consent BART analyses. NO for these sources. At the time we decree. With respect to Cholla, the Response: We disagree that our use of X conducted our analysis in 2011 and installation of newly installed NOX and updated baseline periods for BART 2012, several of these units had been SO2 controls coincided with increases determinations is inappropriate or retrofitted with additional NOX controls in potential emissions of these inconsistent with the CAA or the RHR. that were not in place between 2000 and pollutants resulting from a change in While the RHR specifies 2000–2004 as 2004. In particular, Cholla had installed coal supply.70 Therefore, the more the baseline for purposes of measuring LNB on Units 2, 3 and 4 in 2008 to recent baseline is likely to be more reasonable progress at Class I areas 2009, and Coronado had installed LNB representative of future operating during the first implementation at Unit 1 in 2009.65 In addition, during conditions at these units. 61 period, neither the RHR nor the BART this time period, Cholla completed its Contrary to the assertions of some Guidelines require that this particular transition to a different coal with much commenters, use of updated baselines timeframe be used as the baseline for 66 higher potential NOX emissions. Thus, did not unfairly penalize those sources BART determinations at individual in order to take into account existing that reduced their NOX emissions in sources. Rather, the Guidelines provide controls and to ensure that the baseline advance of a final BART determination. that, for purposes of calculating the period accurately represented Rather, the updated baseline effectively costs of compliance: lowered the baseline visibility impacts The baseline emissions rate should 62 BART Guidelines, 40 CFR Part 51, Appendix Y, from these sources by reducing the represent a realistic depiction of anticipated section IV.D.4.d.1 baseline emissions. As a result, the annual emissions for the source. In general, 63 CAA 169A(g)(2), 42 U.S.C. 7491(g)(2); see also projected benefits of additional controls for the existing sources subject to BART, you 40 CFR 52.308(e)(ii)(A). 64 See, e.g., SIP Appendix D at 4; Apache Unit 2 will estimate the anticipated annual 67 BART analysis at 2–2; Cholla. 77 FR 42861. 65 68 77 FR 42859, 42861. Although no new NOX Consent Decree in United States v. Salt River 59 Consent Decree in United States v. Salt River controls were installed at Apache during this Project, CV 08–1479–PHX–JAT (entered December Project, CV 08–1479–PHX–JAT (entered Dec. 19, timeframe, we determined that more recent 19, 2008). 2008). emissions data (2008–2011 rather than 2005–2007) 69 BART Guidelines, 40 CFR Part 51, Appendix Y, 60 77 FR 9490 (February 16, 2012), codified in were more likely to represent future emissions. 77 section IV.D.4.d.2. Table 2 to Subpart UUUUU of 40 CFR Part 63. FR 42856. 70 See Docket Item B–09, ‘‘Additional APS Cholla 61 See 40 CFR 51.308(d)(2)(i). 66 77 FR 42856, 42859, 42861. BART response’’, Appendix B, Section 6.2.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00016 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72527

were less than if we had used the the plan will take into consideration all limit of 0.065 lb/MMbtu. For each of the original baseline. This approach is emissions reductions achieved during three sources, we have established the consistent with the RHR and the BART the first implementation period, compliance determination method such Guidelines because it accurately reflects consistent with the requirements of the that when one unit is not operating, the controls in place at the time we CAA and the RHR. emissions from its own preceding thirty boiler-operating-days will continue to performed our BART analysis. 2. Control Efficiencies and Emission be included in the 30-day rolling Nonetheless, in order to address Reductions for Alternative Controls commenters’ concerns about the effect average. In the case of Coronado, for of the updated baselines on our Comment: One commenter (NPS) example, during periods when only one proposed determinations, we have also concurred with EPA’s reliance on an unit operates, this method allows the taken into account the original baseline SCR level of performance of 0.05 lb/ one operating unit to average out short- periods considered by ADEQ, as part of MMBtu. The commenter noted that this term emission spikes by using the most the supplemental cost analyses level is consistent with EPA’s recent thirty boiler-operating-day value described below. determination for the San Juan from the non-operating unit. Otherwise, Finally, we note that the use of a more Generating Station in New Mexico and averaging across units would not be recent baseline for purposes of our EPA’s assumptions for the Colstrip and possible during such periods, since the BART analyses does not alter the Corette power plants in Montana. emissions value from the non-operating baseline used for purposes of measuring Response: We acknowledge the unit would be zero since it is not reasonable progress. As noted by several commenter’s concurrence. As described operating. commenters, the RHR specifies that, for further below, information received in Comment: One commenter purposes of setting RPGs and measuring comments on our proposal continues to (Earthjustice), based on a report progress: support the use of an SCR level of submitted with the comments (the performance of 0.05 lb/MMBtu on an ‘‘Sahu report’’), stated that SCR can The period for establishing baseline annual average basis. Accordingly, we achieve greater NOX reductions and visibility conditions is 2000 to 2004. Baseline have retained the use of 0.05 lb/MMBtu visibility conditions must be calculated, visibility benefits at less cost than EPA’s using available monitoring data, by in our cost calculations (which are calculations. According to the establishing the average degree of visibility based on annual emissions). However, commenter, while SCR systems are impairment for the most and least impaired in setting emission limits on a 30-day capable of achieving 90 percent or days for each calendar year from 2000 to rolling average basis, it is necessary to greater removal, EPA’s proposed NOX 2004. The baseline visibility conditions are account for startup and shutdown emission limit of 0.05 lb/MMBtu 71 the average of these annual values. events, which raise the average emission represents control levels of less than 90 In its Regional Haze SIP, Arizona used rates over this shorter period of time. percent at each of the Apache, Cholla IMPROVE monitoring data from 2000– Therefore, we have revised our and Coronado units. Accordingly, the 2004 to calculate baseline visibility for proposed emission limits for SCR at commenter believes that EPA should the best and worst visibility days for each of the sources. As explained below, have analyzed SCR with an emission each Class I Area.72 Since these baseline we have also taken into account other limit of 0.04 lb/MMBtu because this visibility conditions are calculated site-specific factors in revising the level is achievable at 90 percent based on monitored conditions at Class emissions limits. In the case of Apache removal. I areas, they reflect actual emissions that Units 2 and 3, we have performed a The commenter (Earthjustice), based occurred during the 2000–2004 time supplemental analysis using AEPCO’s on a separate report submitted with the frame, rather than any subsequently cost estimates that are allowed by the comments (the ‘‘Stamper report’’), stated implemented controls. CCM (capital costs for the installation of that SCR systems are capable of In developing its long-term strategy, a SCR with LNB and OFA of $164.9 achieving 90 percent or greater removal state must consider inter alia million). We also considered comments, and EGUs elsewhere are subject to NOX ‘‘[e]missions limitations and schedules the size of the Apache facility, AEPCO’s emission limits as low as 0.03 lb/ for compliance to achieve the classification as a small entity, the MMBtu. The commenter cited several reasonable progress goal’’ and the economic effects of requiring the use of Prevention of Significant Deterioration ‘‘anticipated net effect on visibility due SCR on Apache Units 2 and 3, and (PSD) permit limits based on BACT to projected changes in point, area, and AEPCO’s arguments regarding an SCR determinations, including a 0.03 lb/ mobile source emissions over the next emissions limit of 0.07 lb/MMBtu. As MMbtu limit at Plant Washington, 10 years.’’ 73 This would include any discussed below in this preamble, we issued by Georgia Environmental reductions in emissions from BART have concluded that in this case it is Protection Division, and 0.035 lb/ sources that are implemented prior to a appropriate to revise the 30-day rolling MMBtu for Desert Rock, issued by EPA final BART determination, as well as average SCR limit to 0.070 lb/MMBtu, Region 9. Accordingly, the commenter any reductions resulting from such a with a ‘‘bubble’’ across Apache Units 2 believes that EPA should have analyzed determination. Thus, in setting its RPGs and 3. In the case of Cholla, we have SCR with an emission limit of 0.04 lb/ for 2018, a state may receive ‘‘credit’’ for taken into account the need to MMBtu because this level is achievable any reductions achieved during the first accommodate startup and shutdown at 90 percent control for each of the implementation period, regardless of events in the 30-day rolling average and units. whether or not those reductions are have revised the limit to 0.055 lb/ Response: We agree with the reflected in the ‘‘baseline’’ emissions for MMBtu, with a bubble across Units 2, 3 information provided by the a particular BART source. and 4. Finally, in the case of Coronado, commenters that SCR technology has EPA has not yet proposed action on we have taken into account both the the potential to achieve 90 percent and Arizona’s RPGs or long-term strategy. need to accommodate startup and greater rates of removal, as well as Our ultimate action on these elements of shutdown events, as well as the existing achieve emission rates of less than 0.05 consent decree, which sets an emission lb/MMBtu. However, we disagree with 71 40 CFR 51.308(d)(2)(i). limit of 0.080 lb/MMBtu for Unit 2. the commenter’s assertion that emission 72 AZ Regional Haze SIP at page 39. Based on these considerations, we have limits associated with BART must meet 73 40 CFR 51.308(d)(3)(v)(G). set a two-unit 30-day rolling average the lowest emission rate achieved with

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00017 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72528 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

that technology at any coal-fired power averaging time, including startup and in this context, constrain) the plant. The RHR provides that: shutdown periods). determination of what the most • The determination of BART must be based These units are based on new appropriate BART limit representative on an analysis of the best system of construction, which can be designed to of a given technology is for a given continuous emission control technology optimize NOX reduction in other aspects facility. The emission limit set to reflect available and associated emission reductions of combustion (i.e., pulverizer design, BART must be determined based on a achievable for each BART-eligible source that boiler height, etc.). consideration and weighing of the is subject to BART * * * 74 • There is inadequate data available statutory BART factors. Although there Additionally, the BART Guidelines state to confirm the long-term achievability of are some similarities between the top- that: ‘‘[i]n assessing the capability of the the limits because the units have not down BACT determination process and control alternative, latitude exists to begun operation or only recently the five-step BART determination consider special circumstances became operational. process, we note that a BACT pertinent to the specific source under Other commenters note that, as part of determination is based almost review, or regarding the prior the Cross State Air Pollution Rule exclusively on cost-effectiveness, and application of the control alternative’’ 75 (CSAPR), EPA concluded that a NOX does not, for example, take visibility and that ‘‘[t]o complete the BART limit below 0.06 lb/MMBtu is not improvement at Class I areas into process, you must establish enforceable achievable through retrofit of SCR on account.78 emission limits that reflect the BART coal-fired electric generating units.77 One of the commenters noted that in requirements * * *’’.76 The five-factor AEPCO and APS also note that based on IPM modeling performed in support of BART analysis described in the data from the RACT/BACT/LAER the CSAPR rulemaking, we used an SCR Guidelines is a case-by-case analysis Clearinghouse, new coal-fired EGUs emission rate of 0.06 lb/MMBtu for that considers site-specific factors in with SCR are only required to achieve certain retrofit coal-fired EGUs, stating assessing the best technology for 0.05 lb/MMBtu averaged over 12 that this was the most stringent continuous emission controls. After a months, and it is not appropriate to emission rate assumed achievable for technology is determined as BART, the assume that a retrofit coal-fired unit can retrofit units. It is important to note that BART Guidelines require establishment achieve this limit averaged over 30 days. IPM is a tool that operates using a large of an emission limit that reflects the SRP notes that the proposed limit for number of variables with values BART requirements, but does not Coronado Unit 1 is more stringent than determined based upon a wide variety specify that the emission limit must the recently promulgated NSPS for of assumptions. These assumptions, and represent the maximum level of control electric utility steam generating units the values upon which they are based, achieved by the technology selected as constructed after May 3, 2011 (40 CFR will necessarily change based upon the BART. While the BART Guidelines and part 60, subpart Da), which establishes needs and context of the project or the RHR do not preclude selection of the a limit of 0.70 lb/MWh (0.077 lb/ rulemaking for which IPM is used. It is maximum level of control achieved by MMBtu) for new units, and 1.1 lb/MWh therefore not appropriate to a given technology as BART, the (0.11 lb/MMbtu) for modified units. automatically consider a particular emission limit set to reflect BART must APS also provided a report, originally assumption or variable value (in this be determined based on a consideration prepared by RMB Consulting & case, SCR emission rate) used in one and weighting of the five statutory Research, Inc. (RMB) for comment on application of IPM to represent a BART factors. Therefore, limits set as the Regional Haze FIP for San Juan uniform standard or constraint against BACT during PSD review (e.g., Desert Generating Station, suggesting that the which all other uses of IPM should be Rock), or emission rates achieved from Subpart Da limits represent the most compared. the operation of individual facilities stringent level of available control. The In the case of the CSAPR rulemaking under an emissions trading program RMB report states that EPA’s Guidelines cited by the commenter, IPM was used (e.g., Clean Air Act Interstate Rule indicate that state regulatory agencies to set state-wide budgets for NOX based (CAIR)) may provide important should consider NSPS limits in the on assumptions that would be information, but should not be BART evaluation except in cases where minimally achievable to a broad array of construed to automatically represent the the NSPS might be considered outdated covered sources. The emission data and most appropriate BART limit for a given (e.g., ‘‘technology determinations from constraints fed into IPM therefore technology. the 1970s or early 1980s’’), which is not represented sector-wide modeling Comment: Several commenters (APS, the case for the recently promulgated assumptions, which is a much different AEPCO, SRP, AUG, Pacificorp) note that NSPS Subpart Da. use and context than a BART the proposed NOX emission rate, as Response: We do not agree that our determination, which must ‘‘take into based on SCR technology, is more consideration of a NOX emission limit of account the most stringent emission stringent than many other EPA actions. 0.050 lb/MMBtu was inappropriate. We control level’’ in order to establish a In its comments, SRP provided a note that, in its submitted comments, source-specific emission limit. As a contractor’s report that found that the Earthjustice identified several recently result, we disagree that the 0.06 lb/ proposed limit is inconsistent with issued permits that establish emission MMBtu assumption used in the CSAPR BACT determinations that EPA has limits for SCR that are more stringent rulemaking should be construed to approved for new coal-fired units in the than our proposal. While limits set as following ways: BACT during PSD review may provide • 78 We note that a Class I area impact analysis Although there have been several important information about the must be performed by certain PSD projects as part units permitted with similar emissions capabilities of various control of the permit application process. However, the limits, none of these limits are directly technologies, they should not be visibility results are not used in the BACT equivalent (same numeric limit and determination, which is typically determined prior construed to automatically represent (or to performing the visibility modeling, and are not used to determine the appropriate level of control 74 40 CFR 51.308(e)(1)(ii)(A). 77 Citing 76 FR 1109, 1115, January 7, 2011; EPA, except in those cases where the visibility impact is 75 BART Guidelines, 40 CFR Part 51, Appendix Y, Transport Rule Engineering Feasibility Response to sufficiently high to warrant mitigation measures section IV.D.3. Comments, Docket ID No. EPA–HQ–OAR–2009– that end up involving additional emission 76 Id. section V. 0491–4529, at 13, July 6, 2011. reductions.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00018 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72529

represent the most stringent emission commenters asserted that EPA cannot facility. For each source, we considered control level for SCR. ignore presumptive BART limits combustion controls as a potential Similarly, we also disagree that the because, as part of the BART option for BART.86 recently promulgated NSPS Subpart Da Guidelines, they are binding regulatory We also disagree with commenters’ represents the most stringent emission presumptions that should only be assertions that our selection of non- control level for SCR. First, we deviated from based on a careful presumptive BART technology as BART acknowledge that while the BART consideration of the BART factors (70 is flawed or presumptively incorrect. In Guidelines state that ‘‘EPA no longer FR 39171). the BART Guidelines EPA explained concludes that the NSPS level of EPA’s Proposed Rule, however, does that: controls automatically represents ‘the not reflect any such consideration. 79 For coal-fired EGUs greater than 200 MW best these sources can install’ ’’ this Indeed, EPA’s Proposed Rule never even located at greater than 750 MW power plants was written in the context of older mentions the presumptive limits except and operating without post-combustion NSPS subparts with technology to note that Arizona considered them. controls (i.e. SCR or SNCR), we have evaluations that could potentially be (77 FR 42847). The nature of and basis provided presumptive NOX limits, outdated and not representative of for EPA-established presumptive NOX differentiated by boiler design and type of current pollution control technology BART limits for the relevant units at coal burned. You may determine that an performance. We also acknowledge that, Apache, Cholla, and Coronado show alternative control level is appropriate based while the BART Guidelines provide for that EPA’s determination in its on a careful consideration of the statutory ‘‘situations where NSPS standards do proposed FIP that SCR—a much more factors. For coal-fired EGUs greater than 200 MW located at power plants 750 MW or less not require the most stringent level of costly, post-combustion technology— in size and operating without post- available control for all sources within represents BART at these facilities is, at combustion controls, you should likewise a category’’ and cite NSPS Subpart GG least, presumptively incorrect. Because presume that these same levels are cost- (stationary gas turbines) as a subpart EPA failed to consider the presumptive effective. You should require such utility that does not consider post-combustion limits in developing its proposed FIP’s boilers to meet the following NOX emission 80 controls, the recently promulgated BART limits for NOX, the Proposed Rule limits, unless you determine that an NSPS Subpart Da does consider post- is flawed and must be withdrawn. alternative control level is justified based on combustion controls such as SCR.81 The commenters also note that the consideration of the statutory factors.87 Despite this language, however, we RHR also established presumptive Therefore, the presumptive emission disagree with the commenter’s assertion BART emission limits for NOX limits in the BART Guidelines are that NSPS Subpart Da represents the emissions from fossil fuel-fired units rebuttable, and the five statutory factors most stringent emission control level for through notice-and-comment enumerated in the BART Guidelines SCR, or that an NSPS Subpart, even a rulemaking. The presumptive NOX provide the mechanism for establishing recently promulgated one, should be emissions limits for coal-fired EGUs different requirements. Specifically, as treated as a ‘‘floor’’ for establishing vary according to individual source explained in the preamble to the BART BART emission limits. While the BART characteristics, including fuel firing Guidelines: Guidelines provide that, ‘‘you may rely configuration (tangential/wall-fired, on MACT standards for purposes of opposed wall-fired, cyclone) and type of If, upon examination of an individual EGU, 82 a State determines that a different emission BART,’’ they do not indicate that the fuel burned (bituminous, sub- limit is appropriate based upon its analysis same is true for the NSPS standards. An bituminous, lignite, etc.). Commenters of the five factors, then the State may apply NSPS standard must establish an also argued that, because EPA shifted a more or less stringent limit.88 emission rate that is appropriate for all the baseline for BART, it did not the units within its category,83 which in include combustion controls, such as Thus, the establishment of presumptive the case of Subpart Da includes a variety LNB, in its analysis, and only BART limits, and the corresponding of boiler types, coal types, and baseline considered higher cost post-combustion technology upon which those limits are emission rates that may not be controls (SNCR and SCR). based, does not preclude states or EPA representative of the Apache, Cholla, Response: We disagree with the from setting limits that differ from those and Coronado units. Specifically in the commenters’ assertions that we ignored presumptions. The five-factor analysis we performed for these facilities case of the RMB report, which was the presumptive BART NOX limits. prepared for the San Juan Generating Because Apache, Cholla and Coronado demonstrates that, taking into Station, the assertion that the Subpart all have access to and have historically consideration the expected remaining Da standards represent the most burned both bituminous and sub- useful life and the existing controls stringent level of available control is bituminous coal,84 there is no single present at the facilities, SCR is cost- effective, results in the most visibility undermined by the report’s findings that presumptive NOX limit that applies to emission modeling indicates that the any of these units.85 Therefore, rather improvement of all feasible control San Juan units could achieve NOX than rely upon the numerical values of technologies, and that these factors are not outweighed by SCR’s potential emission rates in the range of 0.047 to the presumptive NOX limits listed in the 0.068 lb/MMBtu, which are emission BART Guidelines, we have considered energy and non-air quality rates lower than the Subpart Da the technological basis for presumptive environmental impacts. As a result, regardless of the appropriateness of SCR standards. NOX BART limits, such as the use of Comment: Multiple commenters combustion control technology, boiler (AUG, APS, SRP) stated that EPA must type, and coal type, as part of the five- 86 At Apache Units 2 and 3, we considered consider presumptive BART limits. The combustion controls (LNB plus OFA) as one of the factor analysis we performed for each control scenarios. At Cholla and Coronado, combustion controls were considered as part of the 79 BART Guidelines, 40 CFR Part 51, Appendix Y, 84 See, e.g., Final Report, Apache Unit 2 BART baseline emission rate and were a potential BART section IV.D.1, n. 13. Analysis, Table 3–1 (December 2007); Cholla Unit option in the event that the five-factor analysis 80 Id. section IV.D.1. 2 BART Report, page ES–2; SRP Comments on indicated that no additional controls beyond the 81 40 CFR Part 60, Subpart Da. Proposed Rule (September 2012), RMB Technical baseline were justified. 82 Id. section IV.C. Memorandum, page 3. 87 BART Guidelines, 40 CFR Part 51, Appendix Y, 83 Or subcategories, which Subpart Da does not 85 See BART Guidelines, 40 CFR Part 51, section IV.E.5. establish except for ‘‘new’’ and ‘‘modified’’ units. Appendix Y, Table 1. 88 70 FR 39132.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00019 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72530 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

as a control technology for most units manually placing them into a use of such information. The on a national scale, our five-factor spreadsheet that calculated the capital commenters also note that the RHR analyses establish that NOX BART limits and O&M costs associated with explicitly provides that the cost analysis more stringent than the presumptive pollution control options. While we should take into account any site- limits are appropriate for these units. relied upon the results of these specific information that affect the costs spreadsheet calculations, we did not of a particular BART technology option, 3. Costs of Compliance then use those results to run IPM, as the and the Corn Growers court explained Comment: Several commenters stated type of information generated by an that BART determinations must be that EPA inappropriately conducted its actual IPM model run (e.g., generation made on a source-specific basis. cost analysis using generalized data and dispatch decisions, capacity decisions) Response: While we agree that BART a regional model, whereas the CAA is not relevant to our action. We determinations must be made on a requires a BART determination to be documented our use of the equations source-specific basis, we do not agree based, in part, on a site-specific cost from IPM’s air pollution control that site-specific information is required evaluation. One commenter (Navajo technology cost component by placing for all aspects of a BART analysis. Nation) stated that EPA should justify the raw cost calculation spreadsheet in Nonetheless, in order to address its use of the IPM and explain why it the docket for our proposal.89 This commenters’ concerns that our proposal did not use or request line item costs spreadsheet contained the IPM was based on cost information that was from the facilities to make its analysis equations, corresponding variable insufficiently site-specific and that the more site-specific. This commenter also values, selected notes regarding costs of the SCR with LNB and OFA stated that EPA’s reliance on the IPM is assumptions and variable ranges as well control option, in particular, are not misplaced because the model integrates as selected tables from IPM Base Case representative of actual installation health-based regulations and not the v4.10 documentation. Since we did not costs at these facilities, we have RHR. perform an actual IPM model run, the performed a supplemental cost analysis. Another commenter (SRP) added that spreadsheet and contractor’s report in The supplemental cost analyses for each the proposed rule and the TSD say the docket for our proposal sufficiently facility are described in Section IV.D of almost nothing about how IPM was used document our use of the cost this document, and incorporate much of to calculate costs, instead directing the methodologies from the IPM air the cost information provided by the public to an EPA contractor report for pollution control cost component. facilities in their comments. In more information. The commenter We disagree with commenters’ performing this supplemental cost asserted that no contractor report in the characterization of the cost development analysis, we have adopted a ‘‘hybrid’’ docket for the rulemaking supplies methodology contained in IPM as approach that relies on cost estimates additional detail on precisely how IPM generalized or outdated. As noted in the provided by the facilities for certain line was used. The commenter believes that documentation for IPM’s cost items, but still retains the use of the this failing renders EPA’s proposed rule development methodology for SCR, the CCM methodology as described in the inconsistent with the CAA’s public cost estimate methodology is based following response. notice requirements. upon two databases of actual SCR Comment: Several commenters stated Response: As described in our projects.90 These databases include that EPA’s cost estimating techniques proposal, the IPM is a multi-regional 2004 and 2006 industry cost estimates are flawed and its reliance on the linear programming model of the U.S. prepared for the Midwestern Ozone outdated EPA CCM led to electric power sector. IPM relies upon a Group (MOG), and a proprietary in- underestimates of costs. Several of these very large number of data inputs and house database maintained by commenters noted that EPA claimed provides forecasts of least-cost capacity engineering firm Sargent & Lundy (S&L). that owner’s costs, surcharges and expansion, electricity dispatch, and The MOG information was cross- Allowance for Funds Used During emission control strategies for meeting referenced with actual 2009 projects, Construction (AFUDC) are not allowed energy demand and environmental, and escalated accordingly. S&L then by EPA’s CCM and refute that these transmission, dispatch, and reliability used the information in these databases costs are not allowed by the Manual. constraints. EPA has used IPM to to develop the equations described in The commenters state that while the evaluate the cost and emissions impacts the cost component taking into account Manual does not have specific line of proposed policies, such as the recent the pre-control NOX emission level, items for owner’s costs and surcharges, Mercury and Air Toxics Standard degree of reduction, coal type, facility it discusses some of the items that roll (MATS) to limit pollutant emissions size, and numerous other unit-specific up into these categories. APS, for from the electric power sector. factors. While a costly engineering example, states that: We wish to clarify that, for our evaluation that included site visits proposed action on Arizona’s Regional would potentially produce a more Owner’s costs are home office and plant support costs that are charged directly to Haze SIP, we did not actually run IPM. refined cost estimate that could be Rather, we used one component of IPM, specific projects. These would include costs considered more site-specific than our related to project management, engineering, specifically, the component that own, we disagree that our approach has construction support, start-up, training, etc. develops the costs of air pollution produced cost estimates that are either Surcharges are home office costs associated control technologies. Broadly speaking, ‘‘generic’’ or ‘‘generalized.’’ with a project that may not be charged IPM relies upon numerous components Comment: Several commenters directly to that project. These costs would be and sub-components to specify contended that where specific related to overhead loads, procurement, constraints and variable values that feed knowledge is available, the CCM is accounting, finance, etc.91 into the model algorithms used during oriented to allow and provide for the APS also notes that there is a line item an actual IPM model run. The air for AFUDC in the Manual but provides pollution control cost development 89 Document ID: EPA–R09–OAR–2012–0021– that it is assumed to be zero percent, but component is just one of these 0008, File name: G–15_MODELING_ that in its experience AFUDC is a real FILES_EGU_BART_Costs_Apache_ numerous components. We relied upon Cholla_Coronado_FINAL2. cost and is never zero percent. In the cost information and equations 90 http://www.epa.gov/airmarkets/progsregs/epa- contained in this component by ipm/docs/v410/Appendix52A.pdf. 91 APS comments, page 12.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00020 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72531

addition, the commenters state neither interest and return on equity that would costs associated with fly ash ammonia the CAA nor the BART Guidelines accrue over the construction period and removal in selecting BART. Further, require the Manual to be used to adds them to the rate base when additional problems during disposal of determine the costs of compliance. construction is completed and the assets fly ash may cause environmental Response: With regard to owner’s are used. Although it is included in damage and should not be discounted. costs and surcharges, we agree with capital costs, AFUDC primarily Response: EPA disagrees with this commenters’ assertions that the CCM represents a tool for utilities to capture comment. First, we note that ammonia does discuss some of the items that roll their cost of borrowing and return on adsorption in the fly ash is expected to up into these line items as they have equity during construction periods. be minimal from SCR because excess described in comments. For the control AFUDC is not allowed as a capitalized ammonia would likely react with option of SCR with LNB and OFA, for cost associated with a pollution control sulfuric acid to form particulate example, the CCM does provide for device under CCM’s overnight costing ammonium sulfate or ammonium ‘‘Engineering and Home Office Fees’’ 92 methodology, and is specifically bisulfate, which would not pose the that could potentially include some of disallowed for SCRs (i.e., set to zero) in same odor problem in fly ash reuse as the home office and plant support costs the CCM.93 Therefore, in reviewing adsorbed ammonia. Second, the described in comments. These types of other BART determinations, EPA has facilities’ own BART analyses did not costs are often included in estimates consistently excluded AFUDC.94 include costs of fly ash disposal or under some type of engineering/ Comment: The ACCCE notes that the ammonia removal in the cost estimates procurement/project services line item. Manual specifically states that it does for SCR, which indicates that they do In the case of the cost estimates not directly address the controls needed not consider these potential costs to be provided by the utilities (both those to control air pollution at EGUs, citing significant. Finally, we note that the submitted to ADEQ as part of the the following quote from the Control Arizona Department of Transportation original BART analysis, and those Cost Manual: has designated fly ash from each of the 96 submitted to us in comments on our * * * this Manual does not directly three sources as approved material. As proposal), we note that their cost address the controls needed to control air explained in our proposed rulemaking estimates are not organized to list line pollution at electrical generating units and the accompanying TSD, the item(s) that clearly correspond to (EGUs) because of the differences in presence of ammonia does not impact ‘‘Engineering and Home Office Fees,’’ accounting for utility sources. Electrical the integrity of the use of fly ash in and do not provide information utilities generally employ the EPRI Technical concrete.97 Therefore, we have no indicating where these costs may be Assistance Guidance (TAG) as the basis for their cost estimation processes. information that suggests that included. As a result, while owner’s installation of SCR would result in a costs and surcharge are not line items Response: We disagree with the change to the facilities’ current fly ash included in the CCM, in this instance, commenter’s assertion that the CCM disposal and re-use practices. as a conservative assumption, we have does not address control costs needed to Comment: One commenter (SRP) included the portion of owner’s costs/ control air pollution at EGUs. The quote stated that EPA downplayed the energy surcharge in the total cost, up to the cited by the commenter contains a and non-air quality factor its revised value specified for ‘‘Engineering and footnote that reads as follows: BART determination in the proposed Home Office Fees’’ indicated by the This does not mean that this Manual is an FIP, presenting the narrow conclusion CCM. inappropriate resource for utilities. In fact, that potential energy and non-air quality We disagree with commenters’ many power plant permit applications use impacts do not warrant elimination of assertions that AFUDC is a cost that the Manual to develop their costs. However, any of the otherwise feasible control should be incorporated into our cost comparisons between utilities and across the options for NO at any of the sources. industry generally employ a process called X analysis, as it is inconsistent with CCM The commenter asserted that this methodology. The utility industry uses ‘‘levelized costing’’ that is different from the methodology used here.95 narrow consideration of this factor is a method known as ‘‘levelized costing’’ not tenable because this factor must be The quote is merely a factual to conduct its internal comparisons, weighed and considered in conjunction observation that electric utilities, in which is different from the methods with the other BART factors in the their planning and cost estimating for specified by the CCM. Utilities use overall assessment of what control their own purposes, use a different ‘‘levelized costing’’ to allow them to option constitutes BART for a particular accounting method than required by the recover project costs over a period of source. The commenter believes that CCM. The footnote clarifies that the several years and, as a result, realize a EPA’s approach minimizes the role of CCM is appropriate for utilities for reasonable return on their investment. this factor in a BART analysis, which is regulatory purposes. The CCM uses an approach sometimes beyond EPA’s authority.98 referred to as overnight costing, which 4. Energy and Non-Air Environmental Response: EPA does not agree with treats the costs of a project as if the Impacts this comment. The RHR and the BART project were completed ‘‘overnight’’, Guidelines allow the reviewing with no construction period and no Comment: One commenter (ADEQ) stated that EPA should consider the authority (State, Tribe, or EPA) the interest accrual. Since assets under discretion to determine how to weigh construction do not provide service to 93 CCM (Tables 1.4 and 2.5 show AFUDC value and in what order to evaluate the current customers, utilities cannot as zero). charge the interest and allowed return 94 See, e.g., 77 FR 20894, 20916–17 (Apr. 6, 2012) 96 Approved Materials Source List, Fly Ash, on equity associated with these assets to (explaining in support of the North Dakota Regional Natural Pozzolan, and Lime, Revised July 10, 2012, customers while under construction. Haze FIP, ‘‘we maintain that following the available at http://www.azdot.gov/Highways/ Under the ‘‘levelized costing’’ overnight method ensures equitable BART Materials/. determinations * * *.’’); 76 FR 52388, 52399–400 97 See 77 FR 42853–4284, TSD at 38. methodology, AFUDC capitalizes the (August 22, 2011) (explaining in the New Mexico 98 Citing Corn Growers, 291 F.3d at 6–7 (finding Regional Haze FIP that the Manual does not allow that EPA’s original 1999 regional haze rules had 92 As described in Table 2–5 of the CCM, AFUDC). improperly divorced consideration of the BART Engineering and Home Office Fees represent 10 95 EPA Air Pollution Cost Control Manual, Sixth visibility benefits factor from the other BART percent of purchased equipment costs. Edition page 1–3. factors).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00021 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72532 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

statutory factors (cost of compliance, the provides a supplemental measure that would be to simply list visibility energy and non-air quality combines information on the number of improvements at the various areas, and environmental impacts of compliance, areas and on individual area qualitatively weigh the number of areas any existing pollution control improvement. and the magnitudes of the technology in use at the source, the In contrast, several commenters improvements. The cumulative sum is remaining useful life of the source, and (ADEQ, AEPCO, APS and AUG) simply an easily understood and the degree of improvement in visibility disagreed that EPA’s new visibility objective way of weighing cumulative which may reasonably be anticipated to metric, ‘‘cumulative visibility visibility improvement, as part of the result from the use of such technology), improvement,’’ is an appropriate metric, overall BART decision. as long as the reviewing authority asserting that this metric incorrectly Comment: One commenter performed justifies its selection of the ‘‘best’’ level inflates the estimated visibility NO2 modeling by scaling tropospheric of control and explains the CAA factors improvements of various control column NO2 derived from satellite that led the reviewing authority to options and should not be used. The measurements, as portrayed in imagery choose that option over other control commenters further stated that this from the Institute of Environmental levels.99 In this case, having metric does not appear anywhere in the Physics, University of Bremen, disapproved the state’s BART CAA, RHR or BART Guidelines, and Germany. The commenter states that determinations for NOX at several units, that these rules and guidelines SCR would reduce NO2 closer to ‘‘all of the rights and duties that would specifically give discretion to states to background levels. otherwise fall to the State accrue instead determine how to take into account Response: While the facilities to EPA.’’ 100 This includes a significant visibility impacts in a BART evaluation. considered for BART control are not the degree of discretion in deciding how to In addition, the RHR (at 70 CFR 39170) only NOX sources in the area, the weigh the five factors, so long as that supports identifying the single Class I commenter’s scaling of the weighing is accompanied by reasoned area that would have the greatest concentrations in the satellite images explanation for adopting the technology visibility effects from emission controls according to the reductions expected selected as BART, based on the five and does not support adding from SCR can give a rough idea of its factors, and in accordance with the improvements from multiple Class I NO2 benefit. However, to assess BART Guidelines. EPA has provided a areas in determining visibility effects. visibility impacts, the model used must detailed explanation of our BART The commenters affirmed that EPA account for the formation of visibility- evaluation process and five-factor should use a change in deciview at the impairing ammonium nitrate particles. analyses in our proposal, TSD and Class I area with the highest impact as Under the BART Guidelines, CALPUFF elsewhere in this document. We have its visibility metric, consistent with is the recommended model that weighed the potential energy and EPA’s RHR and the method used by incorporates this nitrate chemistry. non-air environmental quality impacts other EPA regions and states. Alternative models could potentially be of the various control options along The commenters further stated that to used if they had the ability to handle with the other statutory factors in our be relevant to the environmental effect this and other chemical transformations BART analyses and have concluded that that the regional haze program and had undergone a rigorous impacts do not warrant elimination of addresses, the metric by which visibility performance evaluation. any of the otherwise feasible control improvement is determined for Comment: One commenter (NPS) 101 options for NOX at any of the sources. purposes of assessing BART for a commended EPA for the thoroughness particular facility must reflect actual of its visibility modeling analyses and 5. Remaining Useful Life of the Source human perception of visibility. The the methodologies used. The commenter Comment: One commenter (APS) did commenters added that the cumulative noted that EPA used CALPUFF methods not dispute EPA’s assumption of a impact approach used by EPA has no tie 6 and 8 and modeled against annual twenty-year useful life of the emission to human perception and can only average and 20 percent best natural control equipment in its annualized cost distort a BART analysis. The background conditions. The commenter calculations. commenters believe that this approach also pointed out that EPA modeled all Response: EPA agrees with the arbitrarily magnifies the benefit that pollutants while varying NOX emissions commenter that this is an appropriate might be associated with emission to evaluate the effects of changing this assumption for these sources. limitations at a single source. one pollutant. Response: EPA agrees with NPS on Response: EPA acknowledges the 6. Degree of Improvement in Visibility the need to consider visibility comment. It was our intention to Comment: One commenter (NPS) improvements at all the nearby Class I estimate visibility impacts accurately agreed with EPA that a more complete areas as part of a comprehensive and transparently so that one could assessment of visibility improvement for assessment of the degree of visibility more easily compare results to earlier candidate BART controls would include improvement due to BART controls. applications of CALPUFF and clearly consideration of the number of areas EPA disagrees with some other understand the effect of old versus affected and the degree of improvement commenters that cumulative revised IMROVE equations (methods 6 expected at all Class I areas rather than improvement over multiple areas is an and 8) as well as alternative natural focusing on a single area. The inappropriate metric, or that examining background conditions. We modeled all commenter commended EPA for its a single Class I area is sufficient. The pollutants together in order to account reliance on deciview improvement and cumulative improvement metric (i.e., for chemical interactions among the the number of areas showing the simple sum of impacts or various pollutants and also the improvement, plus its consideration of improvements over all the Class I areas) nonlinear dependence of deciviews cumulative improvement, which is not intended to correspond to a single upon extinction. human’s perception at a given time and Comment: One commenter (APS) 99 See BART Guidelines, 40 CFR Part 51, place. The approach is simply one way stated that EPA’s proposal noted that it appendix Y, section IV.E.2. 100 Central Arizona Water Conservation Dist. v. of assessing improvements at multiple is appropriate to use Method 6a, 6b, 8a EPA, 990 F.2d 1531, 1541 (9th Cir. 1993). areas, for consideration along with other or 8b in CALPOST within the CALPUFF 101 See 77 FR 42853–4284, TSD at 38. visibility metrics. Another approach model, yet EPA inappropriately rejected

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00022 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72533

ADEQ’s use of Method 6a in its own benchmark for average cost- assessment of visibility impacts. The analysis and instead used Method 8b, effectiveness, we note that the BART commenters made the following which yielded higher predicted Guidelines do not require the arguments to support their contention visibility improvements in Class I areas. development of a specific threshold. that EPA’s modeling overestimates the Response: EPA did not reject ADEQ’s The BART Guidelines, however, require visibility benefits associated with BART use of visibility method 6a, which that cost-effectiveness be calculated in control options. First, EPA used an remains a viable method for past terms of annualized dollars per ton of outdated version of the CALPUFF visibility modeling work under an pollutant removed, or $/ton.103 We model (version 5.8) that over-predicts agreed upon protocol. Method 6a considered cost of controls by visibility benefits. Based on citations comprises CALPOST Method 6, the old discussing the total capital costs, annual provided by the commenters, CALPUFF IMPROVE equation for translating costs, $/ton, and incremental $/ton, and version 6.42 has been shown to provide pollutant concentration into visibility considered the degree of visibility better agreement with observed levels of impacts, and annual average (the ‘‘a’’) improvement by discussing the nitrates. The commenters provided natural background concentrations. individual and cumulative deciview modeling results using CALPUFF However, for new visibility modeling, improvement resulting from the various version 6.42 for EPA’s consideration. such as EPA performed for the FIP, control technology options, as well as Second, EPA’s outdated use of constant method 8b is preferable. Method 8b the percent change in improvement. Our ammonia background concentration of comprises CALPOST Method 8, the consideration of other metrics in 1.0 ppb over-predicts visibility benefits revised IMPROVE equation, and best 20 addition to $/dv in no way relegates and fails to account for known monthly percent of days (the ‘‘b’’) natural visibility improvement to a secondary or seasonal variations. EPA backgrounds. The revised IMPROVE role. Finally, we note that the FLMs’ inappropriately rejected ADEQ’s use of equation has superior performance for recommended ‘‘benchmarks’’ for dollars variable background concentrations, assessing visibility, and is per deciview are for average dollars per which was well within the state’s recommended by the Federal Land deciview not incremental dollars per discretion. Several of these commenters Managers for regional haze assessments deciview.104 Neither the BART also noted that a case study 107 by performed for New Source Review Guidelines nor the FLMs recommend Terhorst and Berkman based on the 102 permitting. EPA believes that using consideration of incremental dollars per 2005 closure of the Mohave Generating the best 20 percent of days as a basis for deciview. Station found virtually no evidence that background concentrations is desirable Comment: One commenter (NPS) closure resulted in improved visibility since visibility impacts due to emissions cautioned against any implication in at the Grand Canyon. In addition, SRP from facilities are most noticeable on EPA’s analyses that visibility stated that EPA must consider visibility the best days, that is, most visible to improvement must exceed 0.5 dv to be benefits from NOX controls within the visitors of Class I areas. EPA assessed significant. The commenter believes that context of nitrate contributions to the results of both methods (and also the such an approach would be contrary to regional haze. Studies of visibility ‘‘6b’’ and ‘‘8a’’ combinations), but the BART Guidelines. impairment on the Colorado plateau primarily relied on 8b as the most Response: EPA agrees that the 0.5 dv show that nitrate aerosols contribute appropriate method in the BART threshold for ‘‘contribute to visibility only two to five percent to haze. context. impairment’’ is only for the initial Response: EPA disagrees with the Comment: One commenter (APS) Subject-to-BART screening test and it is commenters that any new CALPUFF objected to EPA shifting the CAA’s a maximum even for that purpose, version should be used for the BART mandate to compare costs and benefits according to the BART Guidelines.105 determination. EPA relied on version under the BART program to an Smaller improvements from controls 5.8 of CALPUFF because it is EPA- assessment of ‘‘cost-effectiveness’’ ($/ approved version in accordance with ton) without specifying the threshold should be considered in BART determinations, since they can be the Guideline on Air Quality Models level of what is cost-effective. APS also (‘‘GAQM’’, 40 CFR 51, Appendix W, noted that in the absence of a specific beneficial in considering effects from controls on multiple sources.106 We section 6.2.1.e). EPA updated the threshold for cost-effectiveness, the specific version to be used for regulatory FLMs have referred to a benchmark of have used the 0.5 dv level simply as one purposes on June 29, 2007, including $20 million per deciview as the upper point of comparison, a ‘‘benchmark’’ or minor revisions as of that date. The limit. The commenter also presented ‘‘yardstick,’’ to gauge the magnitude of approved CALPUFF modeling system data showing the incremental costs of impacts under various control scenarios. includes CALPUFF version 5.8, level going from LNB/OFA to SNCR or SCR Comment: Several commenters (APS, 070623, and CALMET version 5.8 level to be over $20 million per deciview for AUG, Navajo Nation, PacifiCorp and 070623. CALPUFF version 5.8 has been Cholla. SRP) asserted that EPA’s proposed NOX Response: The commenter is correct BART determination rests on a flawed thoroughly tested and evaluated, and that the BART Guidelines list the $/ has been shown to perform consistently deciview ratio as an additional cost- 103 BART Guidelines section IV.D.4.c. with the initial 2003 version in the 104 effectiveness metric that can be See, e.g. National Park Service Comments on analytical situations for which Best Available Retrofit Technology for Apache, CALPUFF has been approved. Any employed along with $/ton for use in a Cholla, and Coronado Power Plants in Arizona BART evaluation, and we have included (September 17, 2012) at 6. other version, and especially one with this information in our proposal. While 105 BART Guidelines, 40 CFR Part 51, Appendix such fundamental differences in its the FLMs have indicated that they Y, section III.A.1 (‘‘As a general matter, any handling of chemistry, would be threshold that you use for determining whether a consider $20 million/dv to be a considered an ‘‘alternative model’’, source ‘‘contributes’’ to visibility impairment subject to the provisions of GAQM should not be higher than 0.5 deciviews.’’) 102 Federal Land Managers’ Air Quality Related 106 See, e.g. 70 FR 39129 (‘‘Even though the section 3.2.2(b), requiring full model Values Work Group (FLAG) Phase I Report— visibility improvement from an individual source Revised (2010), U.S. Forest Service, National Park may not be perceptible, it should still be considered 107 Terhorst, Jonathan and Berkman, Mark, ‘‘Effect Service, U.S. Fish and Wildlife Service, October in setting BART because the contribution to haze of Coal-fired Power Generation on Visibility in a 2010. http://www.nature.nps.gov/air/Pubs/pdf/flag/ may be significant relative to other source Nearby National Park’’, Atmospheric Environment FLAG_2010.pdf. contributions in the Class I area.’’) 44, 2524, 2530 (Apr. 2010).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00023 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72534 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

documentation, peer-review, and this comprehensive evaluation process from wildfires, garbage burning along performance evaluation. No such and remains EPA-approved version, and the Mexico/US border, and dust storms. information for the later CALPUFF is thus the appropriate version for EPA’s Response: We do not agree that we are versions that meet the requirements of BART determinations of these facilities. front-loading emission reductions or section 3.2.2(b) has been submitted to or The ammonia issue has already been that we have lost sight of the ‘‘end approved by EPA. Experience has addressed above. EPA believes that goal.’’ While the goal of the regional shown that when the full evaluation there is no compelling alternative to the haze program is to achieve natural procedure is not followed, errors that use of the default 1 ppb background visibility conditions in all mandatory are not immediately apparent can be concentration. Class I Federal areas by 2064, the introduced along with new model The Terhorst & Berkman study cited requirement for states to implement features. For example, changes by the commenter is worthy of BART applies only during the first introduced to CALMET to improve consideration as the Regional Haze planning period ending in 2018.109 simulation of over-water convective program evolves, but one study does not Where a State has not met the RHR mixing heights caused their periodic invalidate CALPUFF, which has had requirements related to BART, EPA is collapse to zero, even over land, so that multiple performance evaluations and obligated to disapprove that portion of CALPUFF concentration estimates were has gone through public comment and the State’s submittal. And, as explained no longer reliable.108 rulemaking. It also does not remove the elsewhere in this document, because the The change from CALPUFF version legal requirement to perform BART FIP clock has already expired for the 5.8 to CALPUFF 6.4 is not a simple determinations for eligible facilities. Arizona Regional Haze plan, we are model update to address minor issues, While nitrate appears to be a smaller required to promulgate a FIP for any but a significant change in the model contributor to visibility impairment disapproved portion of the SIP. Our science that requires its own rulemaking than some other compounds, section action fulfills part of this duty. with public notice and comment before 169A of the Clean Air Act requires We agree that there are various other it can be relied on for regulatory BART determinations on BART-eligible factors that contribute to haze at purposes. EGUs regardless of ambient visibility Arizona’s Class I areas. However, these Furthermore, it should be noted that conditions. Application of BART is one other factors are not relevant to the the US Forest Service and EPA review BART requirements, which govern of CALPUFF version 6.4 results for a means by which we can ensure the continuation of downward emission and today’s action. Under the RHR, causes of limited set of BART applications haze other than BART sources are showed that differences in its results visibility impairment trends. Modeling shows maximum visibility impacts of addressed under separate requirements from those of version 5.8 are driven by for reasonable progress and a long-term two input assumptions not associated 1.2 to 4.5 deciviews depending on the facility, which are not negligible strategy. We will address the remaining with the chemistry changes in 6.4. Use requirements of the RHR for the first of the so-called ‘‘full’’ ammonia limiting contributions to visibility impairment. Even if an individual pollutant or implementation period in Arizona, method and finer horizontal grid including requirements for reasonable resolution are the primary drivers in the source category appears small to some commenters, the many segments of the progress toward the 2064 goal, in a predicted differences in modeled separate rulemaking action. visibility impacts between the model emissions inventory taken together do versions. These input assumptions have cause visibility impairment, and each D. Source-Specific Comments on EPA’s been previously reviewed by EPA and must be addressed in order to make BART Analyses and Determinations progress towards the national goal of the FLMs and have been rejected based 1. EPA’s BART Analysis and on lack of documentation, inadequate remedying visibility impairment from Determination for NO at Apache Units peer review, and lack of technical man-made pollution. EPA identifies X 2 and 3 justification and validation. stationary sources as an important Introducing a new regulatory model is category to evaluate under the Regional a. Control Efficiencies Haze program, including a BART a long process. EPA intends to conduct Comment: Various commenters analysis. a comprehensive evaluation of the latest (ADEQ, AEPCO and AUG) asserted that Comment: Several commenters argued CALPUFF version along with other EPA’s proposed BART determination for that the proposed FIP is inconsistent ‘‘chemistry’’ air quality models, Apache Units 2 and 3 was premised on with the goal of the RHR, which is to including a full statistical performance the assumption that SCR can achieve an make progress toward natural visibility evaluation, verification of its scientific emission limit of 0.050 lb/MMBtu conditions by the year 2064. Another basis, and determination of whether the continuously on a 30-day rolling commenter added that Arizona’s energy underlying science has been average, including periods of startup, providers have already invested time incorporated into the modeling system shutdown and equipment malfunctions, and money (hundreds of millions of correctly. To accommodate such a but that this limit has not been shown dollars) in order to reach the long-term model, there would have to be an to be feasible. They argued that EPA had goal of achieving natural background evaluation of the effect on the regulatory failed to support either its proposed visibility by 2064, and that the framework for its use, including in New BART determination or its reliance on accelerated timeline proposed by the Source Review permitting, and also this limit in its BART analysis. In rule would result in astronomical costs. changes to the Guideline on Air Quality addition, AEPCO and AUG stated that Another commenter stated that EPA is Models and other modeling guidance, in EPA inappropriately relied on vendor front-loading as many emission consultation with the FLMs. CALPUFF information to support an emission rate reductions as possible in the first five version 5.8 has already gone through of 0.050 lb/MMBtu using SCR. AEPCO years of this program, while ignoring also noted that it considered this 108 ‘‘CALPUFF Regulatory Update’’, Roger W. other causes of visibility impairment, support anecdotal and stated that it Brode, Presentation at Regional/State/Local such as fires, in its FIP. Other Modelers Workshop, June 10–12, 2008; http:// www.cleanairinfo.com/ commenters suggested that Arizona’s 109 See 40 CFR 51.308(f) (future Regional Haze regionalstatelocalmodelingworkshop/archive/2008/ haze is produced by a number of plans must address reasonable progress and long- agenda.htm. environmental factors, like pollution term strategy, but not BART).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00024 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72535

cannot form the basis for a BART MMBtu on a rolling 30-day average and type to which Apache has access, we determination, as BACT rules expressly could only achieve in the range of 0.053 also note that AEPCO meets the provide that EPA does ‘‘not consider a to 0.072 lb/MMBtu.111 We have definition of ‘‘small entity’’ as vendor guarantee alone to be sufficient reviewed the analyses provided by SRP established for electric utility justification that a control option will and note that while the results of SRP’s companies by the U.S. Small Business work.’’ AEPCO requested that if EPA analysis indicate that Coronado could Administration.114 We considered retains the SCR limits, that they be set meet a 0.050 lb/MMBtu limit on an AEPCO’s small entity status 115 and how at 0.07 lb/MMBtu due to the annual average basis,112 we agree that to provide AEPCO with operational infeasibility of complying with a lower the Coronado units cannot achieve an flexibility consistent with application of limit at the Apache station. Also, due to SCR emission rate of 0.050 lb/MMBtu the five-factor BART analysis. Based on the load-following and cycling nature of on rolling 30-day average. As a result, these considerations, we have decided the units and the need to accommodate we conclude that 0.050 lb/MMBtu is to raise the rolling 30-day average startups and shutdowns, AEPCO appropriate as annual average design emission limit from the proposed level requested that any lower limits be set as value, but not as 30-day rolling average of 0.050 lb/MMBtu to 0.070 lb/MMBtu. an annual average limit. emission limit at the Coronado units. A rolling 30-day average of 0.070 lb/ Response: We partially agree with this While we acknowledge that Apache 2 MMBtu represents an upward revision comment. In our proposal, our analysis and 3 are not identical to the Coronado of 40 percent from an annual average was based on an SCR annual average units, we do note the following design value of 0.050 lb/MMBtu and design value of 0.050 lb/MMBtu, which similarities: corresponds to the upper end of the was subsequently proposed as a rolling • Both the Apache and Coronado range of lb/MMBtu values considered 30-day average emission limit. We units are of the same boiler type (Riley achievable by SRP’s analysis. We disagree that our use of 0.050 lb/MMBtu turbo). consider this magnitude of upward as an annual average design value is • Both the Apache and Coronado revision appropriate to accommodate merely anecdotally supported or based units were constructed and placed into emissions from startup and shutdown on vendor literature/guarantees alone. operation at approximately the same events, as well to provide AEPCO a As discussed in our proposal, the ability time. Construction commenced on the sufficient measure of operational of SCR to achieve control efficiencies in Apache units in 1976, and they were flexibility as a small entity. In addition, the range of 80 to 90 percent is well placed into operation in 1979. The in response to comments requesting that established. Although the information Coronado units were placed into emission limits be established across operation in 1979 and 1980. units,116 consistent with the BART included in our proposal did include • vendor estimates, it also included Both the Apache and Coronado Guidelines,117 we have decided to set summaries of SCR control efficiencies units have access to, and could the emission limit as a ‘‘bubble’’ limit that were achieved in practice. We have potentially use, a bituminous and sub- across Apache Units 2 and 3. We are 113 further supplemented the record to bituminous coal blend. therefore finalizing a 30-day rolling • Although the historical operating include more recent examples average BART emission limit of 0.070 profiles of the Apache and Coronado illustrating that SCR, as a technology, is lb/MMBtu for Apache Units 2 and 3 as units are not identical, both the Apache capable of achieving control efficiencies a ‘‘bubble’’ across these two units. and Coronado units are cycling units in the range of 80 to 90 percent. For the Comment: One commenter (AEPCO) that exhibit a greater number of startup Apache units, an annual average requested that if EPA establishes an and shutdown events than baseload emission rate of 0.050 lb/MMBtu units. 114 represents 87 to 89 percent control. As noted in our NPRM (77 FR 42867). Based on these similarities, we 115 See EPA’s Action Development Process, Final While these values represent the upper similarly conclude that the Apache Guidance for EPA Rulewriters: Regulatory range of SCR control and are more units cannot achieve an SCR emission Flexibility Act as amended by the Small Business Regulatory Enforcement Fairness Act, November stringent than the control efficiencies rate of 0.050 lb/MMBtu on a rolling 30- used in the BART analyses prepared by 2006, at 3. This EPA guidance document states that day average, but that use of 0.050 lb/ prior to the enactment of the Small Business 110 AEPCO, we reaffirm that these values MMBtu as an annual average design Regulatory Enforcement Fairness Act, EPA are appropriate, given that they are still value is appropriate. We agree that exceeded the requirements of the Regulatory within the range of what is achievable Flexibility Act (RFA) by preparing a regulatory when establishing a rolling 30-day flexibility analysis for every rule that would have with SCR and that the Apache units are BART emission limit that is based upon any impact on any number of small entities. In view among the highest baseline NOX an annual average design value, it is of the changes made by SBREFA, however, EPA emission rate units considered in our appropriate to provide a compliance decided to implement the RFA as written—a proposal. We agree with the commenter regulatory flexibility analysis as specified by the margin for periods of startup and RFA is not required simply because the rule has that, when establishing a 30-day rolling shutdown. In addition to considering some impact on some number of small entities: average BART emission limit that would the boiler type, age of the units, and coal ‘‘Instead, such analysis will be required only in apply at all times, it is appropriate to cases where we will not certify that the rule will not have significant economic impact on a accommodate emissions associated with 111 As discussed in further detail in the responses substantial number of small entities’’, but ‘‘It startup and shutdown events in on Coronado, this range of values corresponds to an remains EPA policy that program offices should developing the emission limit. SRP SCR unit designed to operate during all periods of assess the direct adverse impact of every rule on raised similar concerns in comments on normal operation and loading conditions. small entities and minimize any adverse impact to 112 As discussed in further detail in the responses the extent feasible, regardless of the magnitude of Coronado 1 and 2. As discussed in more on Coronado, this is specifically in regards to the impact or the number of small entities affected.’’ detail in our responses on Coronado, Coronado Unit 1. 116 Although AEPCO did not specifically request SRP submitted information suggesting 113 The Apache units have access to a number of this, this comment was made in comments that the Coronado units cannot achieve bituminous and sub-bituminous coal blends. See, submitted by Arizona Utility Group on behalf of all e.g., Final Report, Apache Unit 2 BART Analysis, of the utilities. As a result, we are also establishing an SCR emission rate of 0.050 lb/ Table 3–1 (December 2007). While the Coronado bubble limits for the Apache units. units currently burn 100 percent sub-bituminous 117 BART Guidelines, 40 CFR Part 51, Appendix 110 See Docket Items B–03 and B–04, Appendix A. Powder River Basin coal, they have historically Y, section V (‘‘You should consider allowing AEPCO’s calculations are based on 83–85 percent burned a mixture of PRB with bituminous coal. See sources to ‘‘average’’ emissions across any set of SCR control efficiency, and 24-hour average SRP Comments on Proposed Rule (September BART-eligible emission units within a fenceline emission rates of 0.07 lb/MMBtu. 2012), RMB Technical Memorandum, page 3. * * *).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00025 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72536 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

SNCR limit, that the limits for Apache not provide any information in its for the Coronado units in evaluating Units 2 and 3 be set at 0.23 lb/MMBtu. comments documenting how or to what SNCR performance and an appropriate The commenter notes that while there extent these issues justify a 0.23 lb/ SNCR emission limit for the Apache are some differences in past utilization, MMBtu emission limit (rolling 30-day units. As noted in our responses to the units are functionally identical and average). We note that AEPCO’s original comments on Coronado, SRP submitted that, based on the best information BART analysis also identified an SNCR a conceptual design estimate for SNCR available, a limit of 0.23 lb/MMBtu is emission estimate of 0.23 lb/MMBtu, for Coronado 1 that included a vendor likely the best consistently achievable but did not discuss the extent to which estimate of 25 percent control efficiency limit given the load-following, unit- startup, shutdown, and malfunction from LNB emission rates. As noted in cycling and startup and shutdown events are accounted for in this our responses for Coronado, while this issues that must be addressed as part of emission rate. is less stringent than the 30 percent unit operation. We note, however, that SRP also Response: Although AEPCO stated in provided information in its comments SNCR control efficiency used by our comments that ‘‘based on the best regarding SNCR performance at contractor, we consider it a reasonable information available, a limit of 0.23 lb/ Coronado Unit 1. Again, because of the estimate. Based upon 25 percent control MMBtu is likely the best achievable similarities between the Apache units efficiency, annual average emission limit’’ and cited unit cycling and and the Coronado units, we consider it rates for the SNCR with LNB and OFA startup/shutdown issues, AEPCO did useful to examine information provided option are presented in Table 2.

TABLE 2—APACHE: SNCR EMISSION RATE ESTIMATE [Annual average]

Control Average Control technology efficiency Apache 2 Apache 3 (lb/MMBtu) (lb/MMBtu) across units (percent) 1 (lb/MMBtu)

OFA ...... 0.37 0.44 0.40 LNB+OFA ...... 30 0.26 0.31 0.28 SNCR+LNB+OFA ...... 25 0.19 0.23 0.21 1 This represents the incremental control efficiency from the previous control option, not the overall control efficiency from the baseline case of OFA.

If we were to establish a BART For the purposes of our cost $2,275/ton to $2,908/ton, which EPA emission limit corresponding to the use calculations or visibility modeling, considers cost-effective. According to of SNCR technology, we would use the however, we have retained the use of Earthjustice, when the cost-effectiveness annual average SNCR emission rates our original SNCR emission rates. A less of SCR is calculated using more accurate presented in Table 2 as our basis, rather stringent SNCR emission rate would, by costs and proper baselines, the result is than our original estimates based on 30 itself, primarily serve to make the next a cost-effective SCR investment that percent SNCR control efficiency. As most stringent control option, SCR, reduces NOX at a cost of $2,640/ton at noted in a separate response, when appear to remove a greater amount of Unit 2 and $2,275/ton at Unit 3. using an annual average design emissions. This in turn would make the Response: Based upon a review of the emission rate to establish a rolling 30- SCR control option appear more commenters’ calculations, we recognize day limit that will apply during periods incrementally cost-effective (i.e., by that there are certain aspects of cost of startup, shutdown, and malfunction, removing a greater amount of emissions, calculations that would result in lower we consider it appropriate to provide relative to SNCR, for the same cost). As $/ton values under different some type of measure that provides a discussed in our proposal and in other assumptions. As noted in our proposal, compliance margin for such events. responses to comments, we already we already consider the SCR with LNB First, we would set the SNCR emission consider SCR to be cost-effective, and it and OFA control option to be cost- limit as a ‘‘bubble’’ limit across Apache is not determinative to our decision to effective at $/ton values that are 2 and 3. As seen in Table 2, the annual find that SCR is ‘‘even more’’ somewhat higher than those calculated average SNCR emission rate, averaged incrementally cost-effective. by the commenters. As a result, we across both units, is 0.21 lb/MMBtu. A b. Costs of Compliance decline to modify our estimates of cost- 0.23 lb/MMBtu emission limit, as effectiveness to reflect these comments, requested by AEPCO, established on a Comment: Two commenters (NPS and as it is not in any way determinative to rolling 30-day average represents an Earthjustice) conducted their own our decision to find that SCR is ‘‘even approximate 10 percent increase from analyses of the cost and cost- more’’ cost-effective or that the the 0.21 lb/MMBtu annual average effectiveness of SCR with LNB and OFA incremental cost-effectiveness value emission rate. We would consider this for reducing emissions of NOX at between SCR and SNCR is ‘‘even more’’ magnitude of upward revision Apache Units 2 and 3. NPS used the incrementally cost-effective. appropriate to accommodate startup, cost methodologies of the CCM, relied Comment: One commenter (AEPCO) shutdown, and malfunction events as on the IPM to reflect the most recent stated EPA underestimated the site- well as the unit cycling nature of the SCR cost levels, and submitted the specific costs for installing SCR at Apache units. As a result, if established, detailed calculations as Appendix B to Apache, due principally to EPA’s we would consider the BART emission its comments. The commenter’s analysis substitution of general data used in the limit corresponding to the SNCR with yielded cost-effectiveness values of IPM model for the site-specific data LNB and OFA option to be 0.23 lb/ $2,392/ton to $3,144/ton. The used by ADEQ. The commenter stated MMBtu, established as a bubble across commenter noted that EPA’s analysis that EPA needs to reevaluate its both units. yielded cost-effectiveness values of numbers in light of AEPCO’s site-

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00026 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72537

specific analysis. For operation and EPA did not include contractor indirect each unit, while the annualized costs of maintenance costs, AEPCO estimates costs and contingency with the capital, LNB and OFA estimated by ADEQ are total costs of $1,760,000, which is engineering and construction costs, nor only about $533,000 per unit. In slightly lower than EPA’s estimate of did EPA include any owner’s costs or addition, the commenter notes that the $1,822,463, with the main difference allowance for funds during marginal improvement in visibility with due to EPA’s higher allowance for construction, including interest during SCR over LNB and OFA would be less maintenance. For the base unit costs, construction. AEPCO does not believe than 1 deciview. EPA used a 25 percent reduction factor EPA should disallow these costs. Response: We disagree with for ‘‘low dust’’ for Unit 3. AEPCO’s AEPCO’s estimates with these costs are commenters’ assertions that we vendors do not believe there will be any $85,666,000, compared with EPA’s substantial reduction in cost based on estimate of $33,279,000. underestimated the costs of SCR, or that ‘‘low dust,’’ and estimates that installed The commenter stated that based on the cost of SCR is disproportional to its costs will be approximately $39,094,000 AEPCO’s estimated installed costs of benefits. In developing our proposed compared to EPA’s estimate of SCR, the cost burden is disproportional action for Apache Units 2 and 3, we $33,279,000 for this unit. AEPCO to the benefits. Adding the costs of SCR examined the cost estimates for the SCR estimates that the bare module cost will to EPA’s estimate for LNB and OFA, the with LNB and OFA control option be near $48,119,000, rather than the annualized cost is $3,508 per ton and contained in AEPCO’s original BART $25,599,000 that EPA estimates, because $13.9 million per deciview. analysis.118 By comparison, the SCR EPA only included costs for induced Another commenter (ACCCE) stated with LNB and OFA cost estimates we draft (ID) fan upgrades and did not that EPA’s proposal to require SCR at developed for our proposed action 119 account for the additional costs of Apache Units 2 and 3 must be do not differ significantly. A upgrading existing or running new abandoned due to the high costs of SCR. comparison of capital cost, total annual electrical service to support the The commenter notes that according to cost, and cost-effectiveness for these two additional electrical loads required by EPA’s estimates, costs of SCR with LNB estimates are summarized in Tables 3 SCR. The commenter also stated that and OFA would be about $6 million for and 4.

TABLE 3—APACHE UNIT 2: COST COMPARISON OF SCR WITH LNB AND OFA

Total annual Emissions Average cost- Capital cost cost removed effectiveness ($) ($/yr) (tpy) ($/ton)

EPA estimate ...... $44,779,657 $5,869,299 2,019 $2,908 AEPCO original estimate ...... 48,740,300 6,102,740 3,250 1,878

TABLE 4—APACHE UNIT 3: COST COMPARISON OF SCR WITH LNB AND OFA

Total annual Emissions Average cost- Capital cost cost removed effectiveness ($) ($/yr) (tpy) ($/ton)

EPA estimate ...... $43,812,028 $6,103,078 2,683 $2,275 AEPCO original estimate ...... 48,740,300 6,062,302 2,778 2,182

We note that while we used a different McDonnell.120 AEPCO provided two compared to a 25 percent cost reduction cost estimation methodology than sets of revisions: one in which it used in our estimate); AEPCO, our estimates of capital cost retained our assumptions regarding • Use of higher capacity factor (0.85 and total annual cost are very similar to costs not included in the CCM, such as for both units, compared to 0.62 and the company’s original estimates and AFUDC and owner’s costs, and another 0.71); differ, for example, by only 8 percent set in which it included those costs. In • Lower SCR NOX removal efficiency and 4 percent (respectively) at Apache both cases, these analyses also (based on an SCR emission rate of 0.07 Unit 2. More importantly, we note that contained revisions in order to reflect lb/MMBtu, compared to 0.05 lb/ AEPCO’s original estimates for Apache capital costs and O&M costs that AEPCO MMBtu); Units 2 and 3 actually show lower $/ton considered more representative and • Inclusion of an additional 15 values than our own, meaning that appropriate for the Apache units. These percent engineering, procurement, AEPCO’s original estimate indicates that revisions included the following: contracting fee (not included in our cost SCR with LNB and OFA is cost- • Higher bare module SCR costs, estimate); and • effective. involving the inclusion and upward And certain other different revision of specific constituent cost assumptions regarding O&M costs that In submitted comments, AEPCO result in similar total O&M costs. provided multiple analyses comparing items (e.g., concrete and piling, ductwork); AEPCO then included our estimate of our SCR (stand alone) cost estimate with LNB and OFA costs with its SCR revised estimates prepared by • Use of lower cost reduction for the (standalone) costs to arrive at its overall engineering firm Burns and low-dust SCR design as reflected in bare module cost (10 percent cost reduction, cost estimate for the SCR with LNB and OFA control option. As discussed 118 Docket Item No. B–01, Arizona Regional Haze 120 The analysis was included in Attachment 1 to elsewhere in this preamble, we have SIP, Appendix D, page 49. AEPCO’s Comments on the page titled ‘‘SCR Capital decided to finalize a 30-day rolling 119 See 77 FR 42856, Table 16. Cost Comparison.’’ average BART emission limit of 0.070

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00027 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72538 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

lb/MMBtu for Apache Units 2 and 3, items noted above, such as bare module 5, this results in revised SCR with LNB and a ‘‘bubble’’ across these two units SCR costs, we are willing to defer to and OFA cost-effectiveness values of to provide AEPCO an adequate margin AEPCO’s judgment on these issues in $3,450/ton and $2,973/ton for Apache 2 for compliance. Although this 30-day order to address AEPCO’s concerns that and 3, respectively, that are still within limit accommodates the possibility of our cost estimate was insufficiently site- a range that we consider cost-effective multiple startups in a given 30-day specific. As a supplemental cost when considered in conjunction with period, we expect such spikes to be estimate, we have used the version of the visibility improvement associated smoothed out over the course of a year, AEPCO’s cost estimate that adheres to with SCR. so that the annual average remains our assumptions regarding costs that are closer to 0.05 lb/MMBtu. For the other allowed by the CCM. As shown in Table

TABLE 5—APACHE 2 AND 3: COST ESTIMATE OF SUPPLEMENTAL SCR WITH LNB AND OFA

Parameter Apache 2 Apache 3 Notes

SCR Capital Cost ($) ...... 71,938,250 71,938,250 1 LNB+OFA Capital Cost ($) ...... 10,543,189 10,543,189 2 SCR+LNB+OFA Capital Cost ($) ...... 82,481,439 82,481,439 ...... Interest Rate (percent) ...... 7.0 7.0 ...... Equipment Lifetime (years) ...... 20 20 ...... Capital Recovery Factor ...... 0.094 0.094 2 Annualized Capital Cost ($/yr) ...... 7,785,664 7,785,664 ...... Fixed O&M ($/yr) ...... 466,000 466,000 1 Variable O&M ($/yr) ...... 1,294,600 1,294,600 1 Total Annual O&M ($/yr) ...... 1,760,600 1,760,600 ......

Total Annual Cost ($/yr) ...... 9,546,264 9,546,264 ...... Heat Rate (MMBtu/hr) ...... 2,316 2,223 2 Baseline Emission Rate (annual average lb/MMBtu) ...... 0.371 0.438 ...... SCR Emission Rate (annual average lb/MMBtu)) ...... 0.050 0.050 2 SCR Control Efficiency (percent) ...... 87 89 ...... Annual Capacity Factor ...... 0.85 0.85 1 Baseline Emissions (tpy) ...... 3,198 3,625 ...... SCR Emissions (tpy) ...... 431 414 ...... Emissions Removed (tpy) ...... 2,767 3,211 ...... Annual Cost ($/yr) ...... 9,546,264 9,546,264 ...... Emissions Removed (tpy) ...... 2,767 3,211 ......

Average Cost-Effectiveness ($/ton) ...... 3,450 2,973 ......

Comment: One commenter (AEPCO) Comment: One commenter (AEPCO) limitations in obtaining funding for stated that according to EPA’s estimates stated that the Appendix Y BART capital improvements. As a single of SNCR costs, the incremental costs of Guidelines (40 CFR 51, App. Y, section generating station, with multiple units SNCR with LNB and OFA compared to IV.E.3.2) provide that the State and EPA subject to BART requirements, the LNB and OFA are $3.3 million with a must consider the economic effects of cooperative is unable to spread costs maximum incremental improvement of BART determinations. AEPCO estimates over unaffected units, other facilities or 0.47 dv at Chiricahua Wilderness Area. that to install and operate SCR with a large system of units and ratepayers. The commenter stated that this LNB and OFA, rates would need to rise Also, as a cooperative, AEPCO is owned improvement in deciviews is by more than 17.5 percent. Further, the by its members and cannot sell stock or insignificant compared with cost. units could have to shut down if the other equities to raise funding, and must cost of power from those units is out of seek long-term financing from the Rural Response: As described above, EPA is line with the cost of power in the open Utilities Service, which has a limited not limited to considering incremental market. Moreover, due to contract budget and is being asked to fund efforts costs and benefits in comparing BART expirations, AEPCO has no certainty for other cooperatives and rural utilities alternatives. The visibility benefits of that even its existing 147,643 meters to meet CAIR, CSAPR, other SIP SNCR at Chiricahua are a full 1 will be available to defray costs. AEPCO initiatives, and the upcoming EGU deciview with an annual cost of $6.6 asserted that these factors are exactly MACT. In addition, the terms of million and a cost-effectiveness of the types of circumstances that were AEPCO’s mortgage agreement would $2,056 $/ton averaged over the two designed to be acknowledged in the necessitate a rate increase of more than emitting units. In this case, even the BART Guidelines. 16 percent to accommodate SCR, and it incremental cost-effectiveness of $2,837 One commenter (AEPCO) stated that is not certain whether the Arizona $/ton is well within the range that we EPA failed to follow the requirements of Corporation Commission (ACC) would consider cost-effective. The incremental CAA section 51.308 and Appendix Y in grant such a rate increase or what the visibility benefit of 0.47 dv is also its cost analysis by failing to review the long term impact would be on AEPCO’s substantial, and additional benefits affordability of the final cost on AEPCO working and patronage capital. would occur at multiple Class I areas. as a single facility cooperative, but AEPCO also stated that the operating Considered as a contribution to rather examined only the cost per ton and financing costs are unreasonable for visibility impairment, EPA disagrees and the cost per deciview. EPA should the Apache plant. EPA estimates the that this improvement from SNCR is also consider the implications of SCR system alone will have operating insignificant. AEPCO’s cooperative status and its and maintenance costs of $3.3 million,

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00028 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72539

which is 35 percent of AEPCO’s total for the SCR system, and the current that this is reasonable, as average annual net revenue of $9.5 million for 2010 and equity capital of $94 million in 2011 increases have been up to 3 times as more than the net revenue of $1.9 would cover the entire cost of high as this increase, and this rate will million for 2011. AEPCO estimates that installation. The report also shows that likely be offset by a settlement award of it will need to increase rates by $22.5 AEPCO will receive refunds from a $63 million. The commenter also noted million a year over the O&M costs just settlement with two railroads totaling that while the incomes of its customer to finance SCR with LNB and OFA on $63 million. The commenter further base are relatively low, the cost of living Units 2 and 3. This combined cost is 14 refuted that AEPCO may not be able to in the area is also lower than the times AEPCO’s net revenues in 2011 borrow sufficient funds for SCR. The national average. The commenter and 2.8 times 2010 net revenues. This commenter stated that RUS loan funds further noted that utilities in similarly cost does not include other are not raised or subsidized by economically disadvantaged areas have expenditures that will be required for taxpayers, and the RUS does not successfully installed modern pollution Units 1, 2 and 3 for BART. With only anticipate any shortage in funding. In controls costing significantly more than 147,643 metered customers and with addition, the commenter claimed that the cost of SCR at Apache. the National Rural Utility Cooperative many of these customers in low income Response: It is not EPA’s intention to Finance Corporation (NRUCFC) is areas, rate increases for these customers endanger the economic viability of financed by private investors, and are not trivial. The commenter also Apache Generating Station or to place AEPCO should not have any difficulty stated that SNCR also is not affordable an undue burden on AEPCO’s borrowing from the NRUCFC, if due to the operating costs. AEPCO customers. EPA has considered the estimates SNCR with LNB and OFA necessary. Another commenter (ACCCE) stated comments on these issues very operating costs to be $6.8 million, carefully. Regarding the legal basis for which is three times AEPCO’s net that the large costs of SCR may adversely impact AEPCO and its our decision, neither the CAA nor the revenue 2011 and over two-thirds of net RHR requires states or EPA to consider revenues in 2010. customers due to AEPCO’s small size, the low income profiles of AEPCO’s the affordability of controls or ratepayer Another commenter (Earthjustice) service area, and AEPCO’s ability to impacts as part of a BART analysis. stated that SCR costs will not threaten obtain financing. The commenter urges Rather, the CAA and RHR require AEPCO’s continued viability or have a EPA to give full consideration to consideration of ‘‘the costs of severe impact on its operations, which AEPCO’s comments submitted June 29, compliance, the energy and non-air are the only two affordability conditions 2012, on these issues. quality environmental impacts of allowed to be considered under the Commenters from AEPCO’s member compliance, any existing pollution BART Guidelines (Appendix Y, Section cooperatives stressed the unique control technology in use at the source, IV.E.3.). The commenter noted that economic and engineering challenges the remaining useful life of the source, guidance and case law on Reasonably they face—low population density, the and the degree of improvement in Available Control Technology (RACT) demands of servicing vast remote areas visibility which may reasonably be and BACT determinations, which make with rugged topography, and anticipated to result from the use of clear that affordability issues are given transmission grid capacity limitations such technology.’’ 121 relatively little weight, are instructive that make it difficult to import power. The BART Guidelines do allow for for BART determinations due to the They noted that the majority of their (but do not require) the consideration of similar analysis. For RACT and BACT, power comes from the Apache ‘‘affordability’’ as part of the ‘‘costs of the commenter explained that Congress Generating Station, so the cost impact of compliance’’ under certain intended that all sources in a source SCR installation would be especially circumstances, noting that: category bear similar costs for pollution acute, resulting in rate increases ranging reduction and that sources should not from an estimated 15 percent to 30 1. Even if the control technology is cost effective, there may be cases where the be able to avoid cost-effective controls percent. The commenters pointed out installation of controls would affect the due to poor financial position, as this that their customer base has average would reward inefficient or poorly- viability of continued plant operations. incomes well below the national and 2. There may be unusual circumstances managed sources. The commenter cited Arizona averages, and would be that justify taking into consideration the two cases regarding RACT and BACT especially hard hit by large rate conditions of the plant and the economic economic feasibility (Michigan v. increases; many customers struggle to effects of requiring the use of a given control Thomas, 805 F.2d 176, 180 (6th Cir. pay their power bills as it is. The technology. These effects would include 1986), Nat’l Steel Corp., Great Lakes commenters stated that AEPCO and the effects on product prices, the market share, Steel Div. v. Gorsuch, 700 F.2d 314, 324 associated cooperatives cannot finance and profitability of the source. Where there (6th Cir. 1983)). The commenter also or absorb the costs of SCR at the Apache are such unusual circumstances that are noted that detailed economic data is Generating Station. The commenters judged to affect plant operations, you may required for sources to raise indicated that closure of the large, load- take into consideration the conditions of the affordability issues under RACT and following coal-fired units would plant and the economic effects of requiring BACT, and the detailed economic threaten the reliability of the electrical the use of a control technology. Where these analysis called for in the BART system, particularly with the limited effects are judged to have a severe impact on plant operations you may consider them in Guidelines should be similarly robust capacity of the local grid to import where EPA considers affordability the selection process, but you may wish to power from other areas. provide an economic analysis that issues for ‘‘unusual circumstances.’’ The Another commenter (Earthjustice) demonstrates, in sufficient detail for public commenter also stated that Apache’s cited a report by Paul Chernick at review, the specific economic effects, continued viability is not threatened, Resource Insight Inc., which estimates parameters, and reasoning * * * Any based on a report by Paul Chernick at that any rate increases at Apache would analysis may also consider whether other Resource Insight Inc., which shows that be limited to a 2 percent to 5 percent competing plants in the same industry have AEPCO’s average operating margin over increase at most, resulting in an average the last four years would cover 185 extra cost of $3.28 per month on 121 CAA section 169A(g)(2), 42 U.S.C. 7491(g)(2); percent of the annual debt repayment customer bills. The commenter stated 40 CFR 51.308(e)(1)(ii)(A).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00029 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72540 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

been required to install BART controls if this AEPCO’s annual sales. We noted in the magnitude lower than the SCR costs information is available.122 NPRM that the projected costs of SCR described elsewhere in this document. We interpret the question of with LNB and OFA are approximately Therefore, even if we were to take them affordability as a specific question of $12 million per year, and that this into account, they would not whether the viability of continued plant exceeds AEPCO’s net margins of $9.5 substantially affect our analyses. operations will be affected by the million in 2010 and $1.9 million in Regarding the comment that the cost pollution control technology in 2011,125 although the report by Paul of SCR with LNB and OFA at Apache question. Although one commenter Chernick at Resource Insight Inc., could be covered with funds from asserted that the costs of SCR with LNB submitted by Earthjustice, notes that AEPCO’s operating margins or legal and OFA could cause a shutdown of AEPCO’s margin in 2008 was $17.4 settlements, while Apache Generating Apache Units 2 and 3 if it causes power million. Station does have annual operating costs from those units to be out of line In addition to conducting this initial margins that vary according to various with the cost of power on the open economic impact assessment, we conditions, it is not necessarily true that market, the commenter did not provide requested information from AEPCO on AEPCO can cover the costs of pollution evidence or analysis that supports this the economics of operating Apache control equipment exclusively from assertion. We agree that the terms of Generating Station and what impact the these funds, or from the settlement AEPCO’s mortgage require AEPCO to installation of SCR may have on the agreement mentioned in the comment. have sufficient revenue to meet the economics of operating Apache Because AEPCO is a member-owned financial metrics of Times Interest Generating Station. We received a utility, operating margins and other Earned Ratio and Debt Service Coverage description of plant conditions and surplus funds may be earmarked to be ratio. But AEPCO is eligible to finance potential economic effects before the returned to its member cooperatives on additional debt related to air pollution NPRM was published,126 and received a rotating basis. While some of these controls, and it has not shown that such additional information during the funds may be available for capital financing is unavailable to it. Securing comment period. We noted in the expenditures such as pollution controls, a rate increase from ACC may be time NPRM that if our analysis of this we have assumed for the purpose of our consuming, and thus supports our information indicated that installation analysis that financing will be necessary decision to grant AEPCO five years for of SCR would have a severe impact on to achieve the pollution reductions installation of such controls. However, the economics of operating Apache required by our action. the information provided to us does not Generating Station, we would For electric utilities, EPA has not show that installation of SCR would incorporate such considerations in our customarily analyzed or considered affect the viability of continued plant selection of BART. ratepayer impacts in BART determinations.129 operations. AEPCO is not being treated The BART cost figures provided in Nevertheless, we also differently from other competing plants this final action do not include other analyzed ratepayer impacts in an effort in its industry: many other electric expenditures that will be required for to assess the potential effects of our utilities, including other rural electric Apache Units 1, 2 and 3 to meet the action on AEPCO as a small entity. EPA cooperatives, are also being required to BART emission limits included in requested an electricity rate analysis install BART controls. Arizona’s Regional Haze SIP. Under the through our contractor, EC/R Inc., to Nonetheless, we performed additional CAA, EPA is not permitted to consider assist us in evaluating the possible analysis to understand better the economic feasibility when taking action electricity rate increases discussed in impacts of the proposed pollution the comments above. Our contractor on a SIP.127 To the extent these costs are controls on AEPCO as a small entity. As noted that AEPCO’s analysis appears to relevant to our FIP action, we note that we explained in our proposal, the U.S. place the entire burden of the AEPCO did not provide any cost Small Business Administration (SBA) incremental capital and O&M costs on estimates for the required upgrades to defines an electric utility company as its Member Co-ops and their retail the existing ESPs and scrubbers at small if, including its affiliates, it is customers. However, the analysis Apache Units 2 and 3 and estimated primarily engaged in the generation, should account for a share of the SCR that the total first year annualized cost transmission and/or distribution of cost going to off-system sales volumes of the required controls at Apache Unit electric energy for sale and its total and not only allocated to member rates. 1 (LNB and FGR) would be $0.552 electric output for the preceding fiscal The contractor’s Incremental Cost million.128 These costs are two orders of year did not exceed 4 million megawatt Model calculated an increment in hours (MWh).123 In 2011, AEPCO 125 See Docket Item H–1 Arizona Electric Power revenue requirements for AEPCO’s member cooperatives sold 2,453,272 Cooperative, Inc. Annual Report Electric for Year member cooperatives of 12.7 percent 124 MWh of electricity. As explained in Ending December 31, 2011 submitted to Arizona under the scenario that spreads the the proposal, we conducted an initial Corporation Commission Utilities Division, incremental SCR cost across all kWh assessment of the potential adverse available at http://www.azcc.gov/Divisions/Utilities/ Annualpercent20Reports/2011/Electric/ produced at Apache, both Member Co- impacts on AEPCO of requiring SCR Arizona_Electric_Power_Cooperative_Inc.pdf. ops and off-system or non-Member with LNB and OFA. Using publicly 126 Docket Item C–16, Letter from Michelle sales. Under the alternative scenario available information, EPA estimated Freeark (AEPCO) to Deborah Jordan (EPA), that the incremental cost for SCR is that the annualized cost of requiring AEPCO’s Comments on BART for Apache covered exclusively by member Generating Station, June 29, 2012. SCR in Units 1 and 2 would likely be cooperatives, the incremental revenue in the range of 3 percent of AEPCO’s 127 Union Electric Co., v. EPA, 427 U.S. 246, 255– 66 (1976); 42 U.S.C. 7410(a) (2). assets and between 6 and 7 percent of 128 Arizona Regional Haze SIP, Appendix D, calculations and thus the cost-effectiveness of the Table 10.3; see also Comments of Arizona Electric various control options considered. See 77 FR 4284. 122 BART Guidelines, 40 CFR Part 51, Appendix Power Cooperative, Inc., Proposed Disapproval of 129 Exceptions include EPA’s Regional Haze FIP Y, section IV.E.3. AZ RH SIP and EPA’s Proposed RH BART FIP for Hawaii, where we analyzed potential rate 123 77 FR 42866–42867; see also 13 CFR 121.201, (September 18, 2012) page 9. In our proposal, we impacts due to the unique energy situation in footnote 1. noted that these control cost calculations include Hawaii, 77 FR 61478, 61488, and EPA’s BART FIP 124 Annual Report for year ending December 31, costs that are disallowed by EPA’s Control Cost for Four Corners Power Plant, where we examined 2011, from AEPCO to Arizona Corporation Manual, such as owner’s costs and AFUDC. Both of potential rate impacts as part of tribal consultation, Commission. these elements have the effect of inflating cost 77 FR 51620, 51625–51626.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00030 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72541

requirement was 15.4 percent.130 As estimate was based on the capital cost mentioned above. The potential rate explained in the preceding responses, estimates originally published in our increases for residential users in 2019, this analysis is based on a capital cost NPRM. the first full year of incremental capital for the installation of SCR with LNB and AEPCO sells electricity through its expenditures for pollution controls OFA of $164.9 million, which matches member cooperatives, and not directly installed in 2017 (and the year with the the costs claimed by AEPCO in their to residential and business customers, largest incremental cost impact), range comment letter minus certain charges but EC/R also analyzed the impact of an from 4.5 percent, or $5.75 per month excluded by EPA CCM. This difference increase in the cost of electricity over 2011 rates, to 10.6 percent, or in the estimated capital cost for SCR generation on the monthly bills of $10.75 per month over 2011 rates.131 also accounts for much of the electricity users serviced by AEPCO’s EC/R noted that the assumptions it discrepancy between AEPCO’s and Member Co-ops. Table 6 indicates the made in constructing its model may Earthjustice’s estimates of electricity incremental retail costs of electricity to cause the impact to rates to be rate increases, since Earthjustice’s end users under the two scenarios conservatively overstated.

TABLE 6—INCREMENTAL RETAIL COSTS DUE TO SCR [As 2019 costs would impact 2011 retail rates]

Residential class only Combined residential, commercial & industrial Scenario Range of Percent Average $ per Average $ per Percent Average $ per Average $ per outcomes Increase year per month per Increase year per month per (percent) customer customer (percent) customer customer

A: Members Pay all Low ...... 5.4 $83 $6.92 5.8 $125 $10.42 SCR Costs. High ...... 10.6 129 10.75 12.0 220 18.33 B: Members Pay Portion Low ...... 4.5 69 5.75 4.8 103 8.58 of SCR Costs. High ...... 8.8 107 8.92 9.9 182 15.17

While these projected rate increases household income levels in the areas analyses conducted by EPA and the are not trivial, they are comparable to served by AEPCO’s Class A member commenters attempted to project the average historical rate increases for cooperatives to average household revenue requirements and possible rate AEPCO, Arizona, and U.S. incomes in the United States. In 2011 increases that would be required if SCR ratepayers.132 They are also projected to the median income for U.S. households with LNB and OFA are required at occur seven years in the future. Again, was $50,502. Using the supplemental Apache, BART and other environmental in discussing the limitations of this information provided by AEPCO, we regulatory requirements form only one retail rate analysis, EC/R noted that the calculated that the median income for part of the complex business conditions results of the retail rate assessment AEPCO’s Member Co-ops’ ratepayers under which utility rate decisions take should be considered conservative by was $49,303. In addition, we aggregated place, especially over extended time design. the data on median household income periods. It is the responsibility of utility Regarding the comment that utilities by zip code into four incomes ranges. in similarly economically disadvantaged companies to work with the appropriate Seventy-one percent of the median areas have successfully installed regulatory agencies to implement any household incomes by zip code were in modern pollution controls costing necessary rate changes in a manageable the $40,000 and above income ranges significantly more than the cost of SCR fashion. and twenty-nine percent were in the at Apache, we note that none of the Accordingly, because neither these median household income range of installed controls listed in Earthjustice’s projected rate increases nor any comment letter were installed under the $20,000 to $39,999. We found that the household incomes in AEPCO’s Member submitted information or analysis RHR. Accordingly, EPA cannot rely on indicate that a requirement to install them as precedents for the Apache Co-ops’ service area are in the same range as average U.S. household SCR with LNB and OFA will affect the Generating Station BART analysis. viability of Apache Generating Station, Regarding the comment on the income, so an increase in AEPCO’s EPA is finalizing its determination that economic vulnerability of AEPCO’s electricity rates should not cause greater this level of control represents BART. ratepayer population, EPA reviewed the hardship than a similar increase 133 However, we are also taking into supplemental information on per capita elsewhere in the country. EPA’s and median household incomes. responsibility under the CAA and the account AEPCO’s status as a small Because electric utility bills are likely RHR is to implement BART at Apache entity as part of our determination. In paid at the household and not Generating Station. As discussed particular, in its comments on our individual, or per capita, level, we elsewhere in this document, the five- proposal, AEPCO requested that ‘‘EPA believe that median household income factor analysis indicates SCR with LNB set the final BART limits in terms of lb/ is an appropriate metric for assessment. and OFA represents BART for NOX at MMBtu only and not as a specified We used census data to compare Apache Units 2 and 3. While the technology’’ to provide AEPCO with

130 Apache Plant: Report on SCR Incremental Cost 132 Energy Information Administration (EIA) State 133 Arizona Regional Haze SIP, BART Assessment. Prepared by Energy Strategies, LLC for Historical Tables for 2011, Released: October 1, Determination for Apache Generating Station, EC/R, Inc. (November 2012). 2012. Average Price by State by Provider, 1990– Supplemental Economic Analysis. Memorandum 131 Id. 2011. http://www.eia.gov/electricity/data/state/ from Larry Sorrels and Robin Langdon, EPA Office avgprice_annual.xls, last accessed November 5, of Air Quality Planning and Standards (November 2012. 5, 2012).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00031 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72542 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

‘‘maximum flexibility.’’ 134 AEPCO also AEPCO’s ability to use multiple fuels. on any population, including any requested that if EPA decided to finalize However, the BART emissions limit we minority or low-income population emission limits consistent with SCR that are establishing for Apache Units 2 and from our final action. Disadvantaged the limits be set at 0.07 lb/MMBtu.135 3 will still allow AEPCO a choice of populations also will be able to enjoy Given the unusual status of AEPCO as using multiple fuels across the units at the visibility improvements in Class I a small entity and a rural electric the Apache facility. areas anticipated from the emissions cooperative, we believe that it is b. Visibility Improvement reductions required by this final consistent with EPA policy to minimize rulemaking. adverse impact to this small entity to Comment: One commenter (NPS) EPA took several steps to ensure the extent that such action is feasible agreed with EPA’s analysis of the transparency and meaningful and consistent with our BART analysis. visibility impacts of the alternative NOX participation in the rule development To allow this small entity the maximum control options for Apache Units 2 and process for this BART FIP. In response flexibility that is consistent with our 3 at the various impacted Class I areas, to numerous requests, we extended the analysis of the five factors, we have as presented in EPA’s TSD, including public comment period on our proposal determined that it is appropriate to set EPA’s conclusions that ‘‘the and increased the number of public the BART limit as a 30-day rolling improvements from SCR are hearings in Arizona from one to three. average 0.070 lb/MMBtu limit, with a substantially greater than for the other In addition, all three hearings had five year compliance deadline. As candidate controls’’ and that ‘‘the Spanish language interpretation services AEPCO noted, this approach may allow modeled degree of visibility and the hearing on August 14 in minor changes in configuration of the improvement supports SCR as BART for Holbrook, Arizona, also offered optimal system to allow AEPCO’s Apache.’’ The commenter also indicated interpretation in Dine´. compliance at somewhat lower cost. that it compiled BART analyses data We disagree that Executive Order This 30-day rolling average 0.070 lb/ from across the United States, which 12898 requires EPA to consider the MMBtu limit is also applied as a revealed that the average cost per economic effects of our proposed action ‘‘bubble’’ across Units 2 and 3. This deciview proposed by either a state or on disadvantaged populations. As EPA’s approach allows for short term emission a BART source is $14 to $18 million. Environmental Appeals Board (EAB) spikes from startups and provides this The commenter pointed out that for all has explained: small entity with additional operational of the NOX control options at the Apache plant, including SCR, both the Executive Order 12898 instructs federal flexibility within the constraints of the agencies to address, as appropriate, BART emissions limit. $/max deciview and the $/cumulative ‘‘disproportionately high and adverse human Comment: One commenter (AEPCO) deciview are well below this range. health or environmental effects of [their] stated that EPA should not consider fuel Response: We acknowledge the programs, policies, and activities on minority switching from the current mix to all commenter’s agreement with our and low-income populations * * *.’’ The natural gas at Apache Unit 1 to be analysis. Our supplemental analysis, Executive Order, thus, speaks to human costless. AEPCO states that if it loses the discussed in more detail above, was health and environmental effects; it does not ability to use multiple fuels, its conducted using a capital cost for the require federal agencies to consider issues negotiating leverage with natural gas installation of SCR with LNB and OFA regarding cost or rate changes.136 suppliers will be greatly reduced, and it of $164.9 million. For the 0.070 limit on Therefore, Executive Order 12898 does will not be able to obtain gas at Apache Units 2 and 3 that we are not require us to consider potential reasonably competitive rates. AEPCO finalizing in this action, this economic effects. Nonetheless, as argued that this cost at Apache Unit 1 supplemental analysis found an average explained elsewhere in this document, should be considered by EPA in its cost per deciview ($/max deciview) of in consideration of AEPCO’s status as a overall evaluation of the affordability of $12.7 million and a cumulative average small entity and consistent with EPA controls at Apache. cost per deciview ($/cumulative policy encouraging consideration of the Response: EPA is approving ADEQ’s deciview) of $3.1 million. potential social and economic impacts emissions limit for Apache Unit 1. As 137 c. Other Comments of EPA actions, we have conducted noted by the commenter, Tables 6 and an analysis of the affordability of 7 of our proposed action (77 FR 42844) Comment: One commenter noted that installing SCR at Apache Units 2 and 3. listed ‘‘fuel switch to PNG’’ as a control EPA is required by the Executive Order This analysis indicates that installation option in the context of the PM10 and on Environmental Justice to consider all of SCR would not affect the viability of SO2 BART analyses, in addition to ‘‘fuel potential economic and environmental continued plant operations at Apache switch to low-sulfur fuel oil.’’ The impacts on minorities and low-income and would result in an average rate annualized costs for both options were populations that its decisions on BART, increase for residential member utility listed as zero in both analyses. The in this case, will have on AEPCO and its customers of (at most) $11 per month in information contained in Tables 6 and customers. The commenter stated that 2019 compared to 2011 rates. 7 does not represent our analysis for over four in ten of AEPCO’s customers Comment: One commenter indicated Apache Unit 1, but reflects the are minorities. In similar remarks, that because AEPCO is a small electric information contained in ADEQ’s PM10 another commenter cautioned EPA that cooperative, EPA is required by the and SO2 BART analyses. ADEQ’s BART such increases would impact at-risk Regulatory Flexibility Act to prepare a analyses for Apache 1 eliminated more populations. regulatory flexibility analysis for this stringent control technologies such as Response: In establishing BART rulemaking. fabric filters and wet FGD, and requirements for the facilities in this determined that a fuel switch to natural final rulemaking, EPA is increasing the 136 In re: Upper Blackstone Water Pollution gas was BART. Natural gas is a level of environmental protection for all Abatement District, Order Denying Review In Part commodity, and its price fluctuates due affected populations by requiring and Remanding In Part, NPDES Appeal Nos. 08–11 to 08–18 & 09–06. (May 28, 2010) slip op at 105. substantial NOX emission reductions. to factors beyond the constraints on (internal citation omitted). Thus, EPA does not expect any 137 See, e.g., Interim Guidance on Considering 134 AEPCO Comments page 18. disproportionately high and adverse Environmental Justice During the Development of 135 Id. human health or environmental effects an Action page 4, footnote 4.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00032 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72543

Response: We agree that AEPCO is remained 2001–2006, LNB and OFA b. Control Efficiencies considered small entity for purposes of would also have been considered in the Comment: In arguing against the the Regulatory Flexibility Act (RFA). analysis. APS noted that EPA concurred achievability of EPA’s proposed limit, However, the RFA does not require a with ADEQ’s BART determination for one commenter (APS) noted that regulatory flexibility analysis when a SO2 and PM10 emissions for these same according to the study that EPA placed rule has an impact on only one small units using a baseline of 2001–2006. In in the docket (IPM Model—Revisions to entity (as opposed to a significant addition, one commenter (Earthjustice) Cost and Performance for APC impact on a substantial number of small asserted the baseline period (2008– Technologies, 2010, Sargent & Lundy), entities). Nonetheless, EPA policy is to 2011) understates NOX emissions the Agency’s minimum emissions limit assess the direct adverse impact of every reductions compared to the baseline of 0.05 lb/MMBtu is specific to Powder rule on small entities and minimize any period of 2001–2004. River Basin coal and the minimum level adverse impact to the extent feasible, In contrast, one commenter (NPS) for bituminous coal is 0.07 lb/MMBtu. regardless of the magnitude of the concurred with EPA’s use of 2011 as the The commenter also stated that because impact or number of small entities baseline period for Cholla units 2, 3 and this is a minimum emissions level, it is affected. Therefore, we gave AEPCO 4 since it represents the first complete probably too aggressive even for a BART additional opportunities to participate calendar year at which it is certain that determination based on bituminous in the rulemaking process. Specifically, the Cholla plant operated using the full coal. The commenter also stated that prior to issuing our proposed rule, we quantity of a higher NOX-emitting coal these rates may be appropriate for new informed AEPCO that our proposed that the plant is committed to purchase units under ideal conditions as BACT action would address BART under its current coal contract. The are not appropriate for BART. requirements for units at AEPCO’s commenter submitted a graph of annual Another commenter (AUG) stated that Apache facility. We also requested NOX emission rates for the units at the EPA’s record in support of the putative information from AEPCO on the Cholla plant, which the commenter achievability of a 0.050 lb/MMBtu economics of operating Apache believes to show the impact of recently emission limit at Apache, Cholla, and Generating Station and what impact the added combustion controls and higher- Coronado is extremely thin and installation of SCR may have on the NOX coal. unpersuasive. AUG states that EPA has economics of operating Apache Response: As explained in a previous not, for instance, demonstrated through Generating Station. We have considered response, we do not agree that use of the the development of an SCR conceptual the comments we received concerning updated baseline for Cholla was design or some other, similar site AEPCO’s status as a small entity and the incorrect or inappropriate. Moreover, specific analysis that SCR can achieve potential economic impact of our updating the baseline did not eliminate this emission rate at any of these proposed action on AEPCO. Our LNB and OFA from consideration as particular facilities, and that EPA must discussion of affordability above BART, since existing controls can affirmatively establish that its selected includes our response to these constitute BART if additional controls BART rate is in fact achievable at these comments and delineates the changes are not warranted based on the five- facilities. we made from our initial proposal in factor analysis. For example, EPA In addition, AUG asserted that EPA’s order to give AEPCO flexibility as a recently approved a determination by proposed limit of 0.050 lb/MMBtu is small entity. We have also taken into Colorado that existing LNB at Comanche inconsistent with the following EPA consideration the potential impact of Units 1 and 2 constituted BART where actions: the reporting, recordkeeping, and other ‘‘the State determined that the added • As part of CSAPR, EPA concluded compliance requirements of this rule, as expense of achieving lower limits that a NOX limit below 0.06 lb/MMBtu set forth in the regulatory text. Because through different controls was not is not achievable through retrofit of SCR AEPCO is an electric utility that is reasonable based on the high cost- on coal-fired electric generating already subject to reporting, effectiveness [$9,900/ton] coupled with units.140 recordkeeping and other compliance the low visibility improvement (under • In EPA’s proposed rule for North requirements under the CAA, AEPCO 0.2 dv) afforded.’’ 138 In this case, by Dakota, EPA based its BART analysis on already has access to the professional contrast, the cost-effectiveness of post a 0.05 lb/MMBtu emission rate, but then skills necessary for the preparation of combustion controls is reasonable and proposed to adopt a 0.07 lb/MMBtu the reports and records necessary for the expected visibility improvements limit because EPA concluded the more compliance with the FIP. are substantial, as explained below. stringent rate would not allow a sufficient margin of compliance (citing 2. Cholla Units 2, 3 and 4 Nonetheless, in order to address the commenter’s concerns that we did not 76 FR 58570, 58610, September 21, a. Selection of Baseline Period properly consider LNB and OFA as a 2011). • Comment: Several commenters potential control option and therefore In its final rule for South Dakota, asserted that EPA incorrectly and precluded a BART determination of EPA set a NOX limit of 0.10 lb/MMBtu inappropriately changed the control LNB and OFA, we have used a baseline for an electric generating plant to allow baseline period in its NOX BART period of 2001–2003, which for an adequate margin of compliance analysis for Cholla. APS and PacifiCorp corresponds to the period used in APS’s (citing 77 FR 24845, 24848, 24849, April 26, 2012). contend that the 2011 NOX emissions original BART analysis. Our • were already controlled by LNB and supplemental cost analysis for Cholla is In Colorado’s recently approved OFA at Cholla Units 2, 3 and 4, which summarized in Table 10.139 regional haze SIP, the NOX BART for penalized APS and PacifiCorp for their Craig Station is an emission rate of 0.27 lb/MMBtu based on SNCR and SCR for voluntary use of these controls. In 138 77 FR 18052, 18066 (March 15, 2012) addition, since LNB and OFA were (Proposed Rule); pre-publication version of Final their units and the NOX BART for already in use, EPA inappropriately Rule, signed September 10, 2012, available at: http://www.epa.gov/region8/air/FinalActionOn 140 Citing 76 FR 1109, 1115, January 7, 2011; EPA, only considered higher cost post- ColoradoRegionalHazePlanSep2012.pdf. Transport Rule Engineering Feasibility Response to combustion controls (SCR and SNCR) in 139 A spreadsheet titled ‘‘Supplemental Cost Comments, Docket ID No. its BART analysis. If the baseline Analysis 2012–11–15.xls’’ is in the docket. EPA-HQ-OAR-2009-0491-4529, at 13, July 6, 2011.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00033 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72544 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

Hayden Station is an emission rate of most stringent emission control level that the In our proposal, we explained that 0.07 lb/MMBtu for one unit and 0.08 lb/ technology is capable of achieving. You SCR, as a technology, can achieve a MMBtu at another unit based on SCR. should consider recent regulatory decisions level of performance between 80 to 90 and performance data (e.g., manufacturer’s Response: We disagree that the SCR percent reduction, even on a retrofit emission rate for the Cholla units data, engineering estimates and the experience of other sources) when basis, and especially when combined should be established at 0.07 lb/MMBtu identifying an emissions performance level with LNB and OFA. Although the per IPM guidance for bituminous coal. or levels to evaluate. commenters indicate that they do not Based on the coal information provided In assessing the capability of the control consider our support for this position 141 in the original Cholla BART analyses, alternative, latitude exists to consider special persuasive, they have not specifically the Lee Ranch/El Segundo Mine coal circumstances pertinent to the specific disputed the claim that SCR can, as a being used at Cholla does exhibit some source under review, or regarding the prior technology, achieve this level of properties that would fall in the range application of the control alternative. performance. We have included of bituminous coal (nitrogen and However, you should explain the basis for moisture content), but also exhibits choosing the alternate level (or range) of additional documents, including vendor properties that fall in the range of sub- control in the BART analysis. Without a experience lists of SCR projects, which showing of differences between the source bituminous coal (fixed carbon, heat indicate that SCR has been capable of and other sources that have achieved more achieving this level of performance.143 value). As a result, we do not agree that stringent emissions limits, you should the Lee Ranch/El Segundo coal can In determining whether special conclude that the level being achieved by circumstances exist at the Cholla units clearly be classified as a bituminous those other sources is representative of the coal. achievable level for the source being that may justify using a different range More broadly, we disagree with analyzed.142 of control, we examined the Clean Air commenters’ assertion that 0.05 lb/ We therefore disagree with commenters’ Markets Database (CAMD) for tangential MMBtu (rolling 30-day average) is an assertion that the BART Guidelines coal-fired units operating with SCR, inappropriate SCR emission limit for the require a SCR conceptual design or either stand alone or in conjunction Cholla units. Although BART other site specific engineering analysis with LNB and OFA, and on a retrofit determinations are performed on a site- in order to demonstrate a level of basis. We identified the 10 best such specific basis, the process for performance. The BART Guidelines performing units, and have listed them establishing the technical feasibility of a indicate that one should take into in Table 7. In addition, we have listed control technology and its associated account the most stringent emission their best-performing annual average emission performance level are control level that the technology is emission rate as well as the percent described in the BART Guidelines as capable of achieving and then document reduction associated with that emission follows: any special circumstances for selecting rate by comparing it to annual average It is important, however, that in analyzing an alternate level or range of control in emission rates from its pre-SCR period the technology you take into account the the BART analysis. of operation.144

TABLE 7—BEST PERFORMING TANGENTIAL COAL-FIRED EGUS WITH RETROFIT SCRS

SCR Emission rate Control State Facility name Unit ID efficiency Control technology (lb/MMBtu) Year (percent)

TX ...... W A Parish ...... WAP7 0.038 2007 73 SCR 1 TX ...... W A Parish ...... WAP8 0.038 2006 77 SCR 1 VA ...... Chesterfield Power Station ...... 6 0.041 2009 89 SCR+LNB+COFA/SOFA NC ...... Marshall ...... 3 0.045 2011 85 SCR+LNB+SOFA TN ...... Kingston ...... 6 0.051 2009 88 SCR+LNB+SOFA TN ...... Kingston ...... 8 0.052 2009 88 SCR+LNB+SOFA TN ...... Kingston ...... 9 0.052 2009 89 SCR TN ...... Kingston ...... 7 0.054 2009 88 SCR+LNB+SOFA MN ...... Boswell Energy Center ...... 3 0.054 2009 86 SCR+LNB+SOFA TX ...... Sandow ...... 4 0.059 2011 83 SCR+LNB+SOFA 1 In the case of the Parish units, we note that their <80 percent control efficiency is the result of low pre-SCR emission rates.

In the case of the Cholla units, which a 24-hour average basis.145 Although the of APS’s own BART analyses for the are also tangential coal-fired EGUs, our commenters have stated that they Cholla units are based upon control estimate of the level of performance of disagree with this level of control efficiencies in a similar range. The the SCR with LNB and OFA control efficiency and the emission rate original BART analyses performed by option corresponds to 80 to 85 percent associated with it, they have not APS and submitted to ADEQ included control efficiency, which is in the low- submitted information for the Cholla visibility modeling indicating that SCR to mid-range of SCR performance. We units documenting special with LNB and OFA can achieve in the used these control efficiencies in our circumstances that would justify a lower range of 83 to 86 percent control cost calculations on an annual average effective range of control efficiency for efficiency for Cholla Units 2, 3 and 4. basis, and in our visibility modeling on SCR. In fact, we note that certain aspects APS calculated these control

141 143 ‘‘Additional APS Cholla BART response’’, Kurtides, Ted ‘‘Lessons Learned from SCR Reduction Control of NOX emissions from Fossil Appendix B. Reactor Retrofit’’, Presented at COAL–GEN (August Fuel-fired Electric Power Plants’’ (May 2009). 142 BART Guidelines, 40 CFR Part 51, Appendix 6–8, 2003); Hitachi SCR/NOX catalyst experience 144 ‘‘Tangentially-fired coal unit SCR retrofit (February 2010); Haldor Topsoe SCR catalyst Y, section IV.D.3. emission data.’’ reference list (October 2009); Institute of Clean Air Companies, ‘‘White Paper—Selective Catalytic 145 See 77 FR 42859, Table 18.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00034 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72545

efficiencies based upon the difference 2003 baseline period and a 24-hour MMBtu. This information is between the highest 24-hour average average SCR emission rate of 0.07 lb/ summarized in Table 8. emission rate observed over a 2001–

TABLE 8—SCR WITH LNB AND OFA CONTROL EFFICIENCY ESTIMATE [APS estimate]

Baseline NOX emissions SCR+LNB+OFA (24-hour average) Emission rate Unit Control (lb/MMBtu) 1 Control tech Period (lb/MMBtu) efficiency 2 (percent)

Cholla 2 ...... 0.503 CCOFA 2001–03 0.07 86 Cholla 3 ...... 0.410 CCOFA 2001–03 0.07 83 Cholla 4 ...... 0.415 CCOFA 2001–03 0.07 83 1 Per Table 2–1 of the original BART analysis for each unit, Docket Items B–06 through B–08. 2 Per Appendix A of the original BART analysis for each unit, Docket Items B–06 through B–08.

APS submitted updated visibility use SCR emission rates consistent with the Cholla units can achieve the values modeling to us as part of comments on these control efficiencies in other in Table 9. These values are consistent our proposal, and with the exception of aspects of its BART analysis, such as on with our own estimates of SCR with Cholla Unit 2, the baseline emissions an annual average basis in cost LNB and OFA performance, and support and associated SCR control efficiencies calculations. If the control efficiencies the use of a 0.05 lb/MMBtu emission do not differ from the original calculated by APS are applied to rate, on an annual average basis, in our analysis.146 We note that APS did not baseline annual average emission rates, cost calculations.147

TABLE 9—SCR WITH LNB AND OFA EMISSION RATE [Per APS Control Efficiency Estimate]

Baseline NOX emissions SCR+LNB+OFA (Annual ave) emission rate Unit Control (lb/MMBtu) Ctrl tech Period efficiency (lb/MMBtu) (percent)

Cholla 2 ...... 0.326 CCOFA 2001–03 86 0.045 Cholla 3 ...... 0.304 CCOFA 2001–03 83 0.052 Cholla 4 ...... 0.296 CCOFA 2001–03 83 0.050

With regard to establishing the BART SIP that established a BART emission to be ‘‘smoothed out.’’ 149 Since the limit emission limit of 0.05 lb/MMBtu on a limit of 0.10 lb/MMBtu (30-day rolling) was established on a shorter averaging rolling 30-day average, the commenters for Big Stone I, based on the use of SCR period than the design basis (from 365 note that in the proposed Regional Haze technology, also citing a need for days to 30 days), there are fewer days FIP for North Dakota, we stated the compliance margin for BART limits that (i.e., data values) with which such following for the Milton R Young must apply at all times including short-term spikes can be ‘‘smoothed Station Unit 1, a coal-fired boiler for startup, shutdown, and malfunction (77 out.’’ In the instances noted by the which we also proposed a NOX BART FR 24849). We agree with the commenter, a less stringent value (from determination based on the use of SCR commenter that it is appropriate to 0.05 to 0.07 for MR Young 1) was technology: accommodate startup and shutdown established for the shorter averaging In proposing a BART emission limit of 0.07 events when establishing a rolling 30- period. lb/MMBtu, we adjusted the annual design day BART emission limit. Since these In order to accommodate emissions rate of 0.05 lb/MMBtu upwards to allow for events, particularly startup, generate from startup and shutdown events, we a sufficient margin of compliance for a 30- elevated levels of emissions, the are finalizing two revisions to our day rolling average limit that would apply at particular day during which such an all times, including startup, shutdown, and proposed emission limit of 0.050 lb/ malfunction.148 event occurs will appear as a short-term MMBtu (rolling 30-day average). First, ‘‘spike.’’ On an annual average basis, we are finalizing the limit as a ‘‘bubble’’ The commenter also notes that we such short-term spikes can be averaged limit across Cholla Units 2, 3 and 4. By approved South Dakota’s Regional Haze with 365 other values that allow them establishing the rolling 30-day limit

146 In the visibility modeling submitted a part of 147 In addition, APS’s comments also included an 148 76 FR 58610. their comments, APS apparently identified a higher SNCR design estimate based upon LNB 149 The precise method by which such short term maximum 24-hour average value from the 2001– performance of 0.22 lb/MMBtu. Achieving an SCR spikes will be ‘smoothed out’ over the period of a 2003 baseline period than the one identified in emission rate of 0.05 lb/MMBtu from this emission year will vary based upon the precise compliance Table 8 for Cholla Unit 2. This results in an rate would represent only 77 percent control determination methodology. The suggestion that it estimated SCR with LNB and OFA control efficiency. This is well within the range of what would be averaged with the other 364 days’ values SCR can achieve, even with a lower inlet NOX is just a generic description of one type of averaging efficiency of 87 percent. emission rate. process.

VerDate Mar<15>2010 20:20 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00035 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72546 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

across all three units, this allows the visibility estimates must be updated to construction of utility-scale air quality spike in emissions associated with a reflect these levels. control systems. startup/shutdown event at one unit to Response: We partially agree with this Response: We disagree with the be smoothed with the emission values comment. In submitted comments, APS commenter’s assertion that we have not from the other operating units. Second, provided a conceptual design estimate provided sufficient information we are also finalizing a less stringent for SNCR which was based upon 25 regarding our cost calculations. In the value in order to establish an emission percent control efficiency (incremental docket for our proposal, we included limit that accommodates the startup and from LNB) and a resulting emission rate the raw cost calculation spreadsheets shutdown events associated with the of 0.17 lb/MMBtu. While this control that contain the cost calculation operating profile of the Cholla units. In efficiency is less than the 30 percent equations, corresponding variable determining what magnitude of revision control efficiency used by our values, selected notes regarding is appropriate, we examined the contractor, we consider it to be a assumptions and variable ranges, as emissions of the Cholla units, as reasonable estimate based upon the well as selected tables from the IPM 154 reported to CAMD, over a 2001–2003 vendor quotes provided by APS.152 Base Case v4.10. In addition, web 150 baseline period. We calculated We disagree with the use of an LNB links were also provided (both in the annual average emission rates and 30- emission rate of 0.22 lb/MMBtu, as the raw cost calculation spreadsheet and in day rolling average emission rates using Cholla units have not demonstrated a our proposal) to the location on the a calculation methodology consistent ability to operate at this publicly available EPA Web site that corresponding to a bubble limit across emission rate under the current coal contains full IPM documentation. We 151 note that both SRP and AEPCO were all three units. Based on this contract for Lee Ranch/El Segundo coal. able to locate this spreadsheet, as both methodology, we determined that the Based upon a review of CAMD emission utilities submitted control cost estimates maximum annual average emission rate data since the installation of LNB, we as part of their comments that revised for these units was approximately 0.32 acknowledge that the Cholla units have, certain variable values and assumptions lb/MMBtu, while the maximum 30-day to varying degrees, operated with LNB in our contractor’s raw calculation rolling average emission rate was at emission rates consistent with APS’s spreadsheet. This information was approximately 0.35 lb/MMBtu. This assertion of 0.22 lb/MMBtu during this initially developed by EPA represents an 8 percent difference period. However, as noted in our contractors 155 and was reviewed by between the highest rates observed on proposal, calendar year 2011 an annual and 30-day rolling average. EPA staff. Following the close of the represented the first year at which the public comment period on our proposed We recognize that this variability Cholla plant operated at the ‘‘full’’ between annual average and 24-hour rulemaking, APS provided additional minimum purchase quantity under its information concerning its own cost average emission rates is based on new contract for Lee Ranch/El Segundo operation of the Cholla units with LNB estimates. We have placed this coal, which is a higher NOX-emitting and OFA, and may not be directly information to the docket and taken it coal than what was previously used. into account as part of this final representative of the variability Since the beginning of 2011 to associated with operation of SCR. We rulemaking, as explained below. September 2012, Cholla Units 3 and 4 Comment: One commenter (APS) are therefore finalizing an emission rate have operated at or below an emission of 0.055 lb/MMBtu as a bubble limit stated that EPA’s cost-effectiveness rate of 0.22 lb/MMBtu for only five to numbers in the proposed FIP are across Cholla Units 2, 3 and 4, which six months of this 21 month period, and represents a 10 percent upward revision incorrect. The commenter stated that Cholla Unit 2 has not operated at or EPA used a capital recovery factor of 9.4 from the annual average design value. below this emission rate in any month When combined with the 3-unit bubble, percent, assuming an interest rate of 7 during this period.153 Therefore, an LNB percent, but APS states that a capital this represents an emission limit that we emission rate of 0.22 lb/MMBtu is not consider appropriate to ensure design recovery factor of 13.4 percent should supported by the actual recent operation be used to account for income and and operation of the emission control of the Cholla units, so it is unlikely to system to provide the best available property taxes and the cost of capital be an appropriate representation of authorized by ACC in the last rate case. retrofit control. anticipated future emissions. Comment: EPA based LNB/SOFA The commenter also stated that EPA analysis uses emissions factors for SCR emission rates on 2011 NOX emissions c. Costs of Compliance rates, which is not an accurate that are not appropriate for the type of Comment: One commenter (APS) assessment of the capability of the coal used, the units, or the averaging stated that, for EPA’s capital costs installed LNB and SOFA. Arizona set period. In addition, APS noted the cost estimate, no back-up material was the BART limit for Cholla Units 2, 3 and values used in the IPM model and EPA’s provided, even when directly requested 4 at 0.22 lb/MMBtu. All three units were CCM may be outdated, which may also by APS. This lack of information makes able to meet this limit in their lead to underestimation of the true it impossible for APS to comment on the acceptance test after LNB and SOFA costs. APS estimates cost-effectiveness validity of EPA’s cost estimates. The were retrofitted, and APS believes they ranging from $7,719/ton to $8,894/ton, commenter also stated that EPA has not can meet it long term. In addition, an with incremental costs ranging from established its contractor or SNCR design study performed by Black subcontractor responsible for the costs 154 and Veatch indicated that an SNCR Document ID: EPA–R09–OAR–2012–0021– estimates as experienced in the 0008, File name: G–15_MODELING_FILES_EGU_ system could obtain a control efficiency _ _ _ _ _ engineering, procurement and BART Costs Apache Cholla Coronado FINAL2 of approximately 25 percent, which 155 Specifically, the initial cost estimates were would correspond to an emission rate of developed by Jim Staudt of Andover Technology 152 0.17 lb/MMBtu. EPA’s cost and Black and Veatch’s report cites lower inlet Partners. While there is no requirement for EPA to NOX concentrations to the SNCR system. A lower establish that its contractors are ‘‘experienced in the inlet NOX emission rate makes it more difficult to engineering, procurement, and construction of 150 ‘‘Cholla CAMD emission data (daily) 2001–03’’ reduce NOX emissions, which makes a lower utility-scale air quality control systems,’’ Dr. Staudt 151 Please consult the regulatory language in our removal efficiency reasonable. has extensive expertise and experience in the field 153 final action for the NOX compliance determination ‘‘Cholla CAMD emission data (monthly) 2010– of air pollution control at power plants. See: methodology associated with the bubble limit. 12.’’ www.andovertechnology.com/staudt.html.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00036 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72547

$8,759/ton to $10,329/ton compared to were incorrect, we have performed a well as AFUDC, which is inconsistent EPA’s estimates of $3,115/ton to $3,473/ supplemental analysis using portions of with CCM methodology. ton, with incremental costs ranging from the updated cost estimates provided by • Use of a 7 percent interest rate: We $3,257/ton to $3,813/ton. APS included APS in its comments. In this have retained the use of a 7 percent costs for surcharges, current AFUDC supplemental analysis, we have interest rate in calculating the capital and fixed charge rates, and emissions generally relied upon APS’s estimates of recovery factor, and disagree with APS’s factors based on the capability of the capital costs and operating costs. While assertion that a 13.4 percent interest rate existing LNB and OFA at the plant, we do not find that these estimates were typical SNCR removal rates, and sufficiently supported with detailed is appropriate. For cost analyses related minimum SCR emissions for site-specific information in all to government regulations, an bituminous coal. instances, we are using them as a appropriate ‘‘social’’ interest (discount) In contrast, one commenter conservative assumption (i.e., an rate should be used. EPA calculated (Earthjustice) stated that SCR at Cholla assumption that would tend to capital recoveries using 3 percent and 7 is more cost-effective than EPA’s overestimate rather underestimate the percent interest rates in determining calculations suggest, in that EPA annualized cost of controls). As cost-effectiveness for the Regulatory overestimated the costs by (1) using an discussed in a previous response, we Impact Analysis (RIA) for the BART unjustifiably high 7 percent interest consider it appropriate to observe the Guidelines.157 158 We consider our use of rate; (2) amortizing costs over a 20-year broader cost methodology used in EPA’s an interest rate of 7 percent to calculate life of the SCR system, rather than a CCM, and have adjusted or eliminated capital recovery to be a conservative more realistic life of 30 years or more; certain cost items not allowed by the approach. and (3) overestimating the costs of the CCM. A line-by-line comparison of • Use of original baseline period: As SCR catalyst, reagent, auxiliary power APS’s cost estimate and our revisions discussed elsewhere in our responses, and property taxes and insurance. In can be found in the docket for this we consider our use of a more recent addition, the commenter asserted that rulemaking action.156 A summary of baseline as consistent with BART EPA baseline period understates NOX cost estimates based on this Guidelines. However, in order to emissions reductions compared to the supplemental analysis is in Table 10, baseline period of 2001–2004. and includes the following: address commenter’s concerns that we According to the commenter, when the • Inclusion of APS’s updated cost did not properly consider LNB and OFA cost-effectiveness of SCR is calculated estimates: We have adopted a ‘hybrid’ as a potential control option and using more accurate costs, proper approach in which we have used APS’s therefore precluded a BART baselines and appropriate emission capital cost and O&M cost estimates, determination of LNB and OFA, we rates, the result is an even more cost- while excluding those cost items not have used a baseline period of 2001– effective SCR investment that reduces allowed by CCM methodology. As 2003, which corresponds to the period NOX at a cost of $1,901/ton at Unit 2, discussed in a previous comment, we used in APS’s original BART analysis. $1,940/ton at Unit 3 and $2,076/ton at have included owner’s costs up to the This represents a time period prior to Unit 4. amount provided for ‘‘Engineering and the installation of LNB, during which Response: Although we do not agree Home Office Fees’’ as described by the the control technology in place on the that our cost-effectiveness estimates CCM. We have excluded surcharge as Cholla units was only OFA.

TABLE 10—CHOLLA CONTROL COST ESTIMATES (PER APS COMMENTS, WITH EPA REVISIONS)

Annualized Annual Total annual Control options Capital cost capital cost O&M cost cost ($) ($/yr) ($/yr) ($/yr)

Cholla 2: LNB+OFA ...... $4,482,254 $423,093 $120,000 $543,093 SNCR w/LNB+OFA ...... 16,617,408 1,568,566 1,254,500 2,823,066 SCR w/LNB+OFA ...... 87,713,386 8,279,523 1,626,683 9,906,206 Cholla 3: LNB+OFA ...... 3,848,807 363,300 120,000 483,300 SNCR w/LNB+OFA ...... 19,238,125 1,815,943 1,254,500 3,070,443 SCR w/LNB+OFA ...... 83,461,195 7,878,146 1,570,766 9,448,912 Cholla 4: LNB+OFA ...... 5,334,618 503,550 170,000 673,550 SNCR w/LNB+OFA ...... 24,885,052 2,348,973 1,737,393 4,086,366 SCR w/LNB+OFA ...... 119,083,832 11,240,671 2,350,182 13,590,853

A summary of emission rates and We note that while APS has provided control cost estimates are based on an emission reductions associated with emission estimates for this baseline annual average ($/year), we have each control option is in Table 11. As period, the values provided, both in the calculated annual emission rates for the noted previously, these emission original BART analysis and in OFA baseline using the annual average estimates are based on a 2001–2003 submitted comments, appear to emission data reported to CAMD over baseline period, during which the represent the highest 24-hour average this 2001–2003 baseline period. Cholla units operated only with OFA. value for modeling purposes. Since Comparing a baseline value on a 24-

156 Docket ID No. EPA–R09–OAR–2012–0021. Available Retrofit Technology (BART) 158 A 7 percent interest rate is recommended by 157 Regulatory Impact Analysis for the Final Clean Determinations Under the Regional Haze Office of Management and Budget, Circular A–4, Air Visibility Rule or the Guidelines for Best Regulations, EPA–0452/R–05–004 (June 2005). Regulatory Analysis, http://www.whitehouse.gov/ omb/circulars-a004-a-4/.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00037 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72548 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

hour average basis (as provided by APS) apples’’ comparison, as some portion of differences between moving from a 24- to a control option value on an annual the emission reduction in such a hour average to an annual average basis. average basis is not an ‘‘apples-to- comparison would be attributable to the

TABLE 11—CHOLLA EMISSION ESTIMATES

Emission Annual Emission rate Emissions Control options factor Heat rate capacity removed (lb/MMBtu) (MMBtu/hr) factor (lb/hr) (tpy) (tpy)

Cholla 2: OFA (only) ...... 0.326 3,022 0.91 985 3,927 ...... LNB+OFA ...... 0.295 3,022 0.91 892 3,554 373 SNCR w/LNB+OFA...... 0.207 3,022 0.91 624 2,488 1,440 SCR w/LNB+OFA ...... 0.050 3,022 0.91 151 602 3,325 Cholla 3: OFA (only) ...... 0.304 3,480 0.86 1058 3,985 ...... LNB+OFA ...... 0.254 3,480 0.86 885 3,335 650 SNCR w/LNB+OFA...... 0.178 3,480 0.86 620 2,334 1,651 SCR w/LNB+OFA ...... 0.050 3,480 0.86 174 655 3,330 Cholla 4: OFA (only) ...... 0.296 4,399 0.93 1302 5,304 ...... LNB+OFA ...... 0.260 4,399 0.93 1144 4,661 643 SNCR w/LNB+OFA...... 0.182 4,399 0.93 801 3,263 2,042 SCR w/LNB+OFA ...... 0.050 4,399 0.93 220 896 4,408

Cost-effectiveness values for each and annual emissions removed listed in control technology are summarized in the previous tables. Table 12, based on the total annual costs

TABLE 12—CHOLLA CONTROL OPTION COST-EFFECTIVENESS

Cost-effectiveness Total annual Emissions ($/ton) Control options cost ($/yr) removed (tpy) Average Increment

Cholla 2: OFA (only) ...... LNB+OFA ...... 543,093 373 1,454 ...... SNCR w/LNB+OFA ...... 2,823,066 1,440 1,961 2,138 SCR w/LNB+OFA ...... 9,906,206 3,325 2,979 3,757 Cholla 3: OFA (only) ...... LNB+OFA ...... 483,300 650 743 ...... SNCR w/LNB+OFA ...... 3,070,443 1,651 1,860 2,586 SCR w/LNB+OFA ...... 9,448,912 3,330 2,838 3,799 Cholla 4: OFA (only) ...... LNB+OFA ...... 673,550 643 1,047 ...... SNCR w/LNB+OFA ...... 4,086,366 2,042 2,001 2,441 SCR w/LNB+OFA ...... 13,590,853 4,408 3,083 4,016

Even based on cost estimates revised two control options, the SNCR- and visibility impacts of the alternative NOX to use APS’s capital and O&M cost SCR-based systems. As discussed in our control options for Cholla Units 2, 3 and estimates, we still consider the cost- proposed action, and in other responses 4 at the various impacted Class I areas, effectiveness values of SCR, on an in this document, we have not as presented in EPA’s TSD. The average ($2,838 to $3,083/ton) and identified any energy or non-air quality commenter also indicated that its incremental ($3,757 to $4,016/ton) impacts that warrant eliminating SCR estimates of the two $/deciview basis, to not be cost-prohibitive. We from consideration for the Cholla units. measures of cost-effectiveness were consider these results supportive of our Combined with the modeled visibility similar to those of EPA. Specifically, the proposed determination that SCR with improvement associated with this commenter’s analysis yielded values of LNB and OFA is cost-effective. We note control option, these cost estimates $19.9 million for the ‘‘$/max deciview’’ that while the LNB and OFA option is continue to support the selection of SCR metric and $3.7 million for ‘‘$/ cumulative deciview.’’ the least expensive (i.e., lowest annual with LNB and OFA as BART for NOX at cost) and is the most cost-effective of the the Cholla units. Response: We acknowledge the comment. control technologies (i.e., has the lowest d. Visibility Improvement $/ton value), it is also the least effective Comment: One commenter (APS) control option. It removes substantially Comment: One commenter (NPS) hired a contractor to perform modeling fewer emissions than either of the other agreed with EPA’s analysis of the with CALPUFF version 5.8 and the

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00038 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72549

updated version of 6.42 to measure the benefits. The BART Guidelines weighted average of the individual sensitivity of various emission control recommend consideration of both stacks’ absolute exit temperatures. EPA scenarios at Cholla Units 2, 3 and 4 average and incremental cost- found that impacts and improvements including two different background effectiveness,162 but do not expressly decreased by some 11 percent when ammonia concentrations. The contractor require or recommend consideration of merged stacks are used. The found that regardless of which model incremental visibility improvement. improvement from SCR at Petrified version or background ammonia value Rather, they provide for consideration of Forest remains over 1.0 dv, with was used, the highest predicted net visibility improvement (i.e., ‘‘the continued substantial benefit at Grand visibility improvement of SNCR or SCR, visibility improvement based on the Canyon. A merged stack for Units 2 and compared to LNB and OFA, is lower modeled change in visibility impacts for 3 was also assumed in additional than the threshold for human the pre-control and post-control modeling EPA performed to address perceptibility of 1.0 deciview. emission scenarios’’ as opposed to the H2SO4 emissions for Cholla, as Moreover, retrofitting SNCR or SCR at change between different control described below. Cholla will not lead to any perceptible scenarios).163 EPA’s Guideline on Air Quality improvement in visibility at any of the Comment: One commenter (APS) Models (40 CFR part 51, Appendix W) 13 Class I areas within 300 km of the noted that Cholla Units 2 and 3 have at section 6.2.2 requires that facilities be Cholla facility. separate flues but share a single stack, modeled using a stack height consistent Response: EPA disagrees with the which EPA failed to recognize in its with GEP, rather than a higher actual ammonia concentration and CALPUFF visibility modeling. The commenter also stack height, in order to prohibit ‘‘stack model version used by the commenter noted that EPA failed to use the height credit’’ from being used in for reasons discussed above. Further, we appropriate Good Engineering Practice developing emission limits.166 By do not agree that the consideration of (GEP) stack height correction required building very tall stacks instead of visibility improvement must directly by EPA’s own rules for modeling. applying emission controls, facilities reflect human perception. The CAA and Because these errors result in visibility could avoid violating the NAAQS the RHR require, as part of each BART impacts in opposite directions, the net locally, but would contribute to higher analysis, consideration of ‘‘the degree of effect is less than 5 percent, based on levels of emissions regionally, and cause improvement in visibility which may modeling that APS has conducted. higher total pollutant levels downwind. reasonably be anticipated to result from Response: If the commenter is correct In short, the requirement to use GEP the use of such technology.’’ 159 The that there were two errors that nearly stack height generally results in regulations do not require that the cancel out, then this would appear to conservative modeling, thereby improvement anticipated to result from have little effect on EPA’s decision. The removing the incentive to build a particular technology at a particular maximum area benefit of SCR was artificially tall stacks to evade controls. source be perceptible by a single human modeled by EPA to be 1.34 dv at Choosing a stack height or taking credit being in order to be relevant as part of Petrified Forest National Park, and 1.06 for a stack height increase is not at issue a BART determination. As EPA dv at Grand Canyon National Park; a 5 in a BART determination. The visibility explained in the preamble to the BART percent reduction in these would still impacts and improvements shown in Guidelines: result in substantial visibility benefits. EPA’s BART modeling are closer to the EPA’s modeling was based on stack actual values if actual stack heights are Even though the visibility improvement parameters provided by APS in a used. Insofar as GEP is relevant, using from an individual source may not be 164 perceptible, it should still be considered in letter that did not mention the shorter GEP heights would tend to setting BART because the contribution to merged stack, although it was increase both pre- and post-control haze may be significant relative to other mentioned in APS’s BART analysis 165 impacts, and to scale up the estimated source contributions in the Class I area. Thus, submitted to ADEQ. Stack parameters visibility improvements. The overall we disagree that the degree of improvement for Unit 4 provided in the commenter’s effect would be to strengthen the case should be contingent upon perceptibility.160 modeling do not match either of those for EPA’s proposed controls. Thus, in our visibility improvement documents (exit velocity of 77.1 feet/ Comment: Based on a report analysis, we have not considered second versus 52 feet/second in APS’s submitted with the comments, one perceptibility as a threshold criterion for letter). In addition, it is unclear how commenter (Earthjustice) stated that had considering improvements in visibility. parameters for the merged stack in the EPA’s BART analysis included lower Rather, we have considered visibility commenter’s modeling were derived emission rates and proper baselines, the improvement in a holistic manner, (except that the area of the merged stack visibility benefits of SCR at Cholla Units taking into account all reasonably used is equal to the sum of the areas of 2, 3 and 4 would be even greater than anticipated improvements in visibility the individual stacks cited in the APS the 7.21 dv cumulative visibility benefit expected to result at all Class I areas letter). Nevertheless EPA acknowledges discussed in the proposed rule. within 300 kilometers of each source. that Units 2 and 3 should have been Response: As explained in the general Improvements smaller than 0.5 dv may modeled together as a single stack. EPA discussion regarding selection of be warranted considering the number of conducted additional modeling to assess baseline periods above, we do not agree Class I areas involved, and the fact that this affect, assuming the same total stack that we used an improper baseline. in the aggregate, small improvements exit area and volume flow rate as for the However, we agree that higher baselines from controls on multiple BART and individual stacks, and a volume- and lower post-control emissions would other sources will contribute to show greater benefits than our modeling visibility progress.161 162 BART Guidelines, 40 CFR Part 51, Appendix showed, and would further support our In addition, EPA is not obligated to Y, Section IV.d.4.b. proposal for SCR. focus on incremental costs and benefits 163 Id. Section IV.D.5. 164 to the exclusion of absolute costs and ‘‘Request for Information Relating to Cholla 166 Guideline on Air Quality Models 6.2.2.a. ‘‘The Power Plant’’, letter from Sue Kidd, Director, use of stack height credit in excess of Good Corporate Environmental Policy and Programs, to Engineering Practice (GEP) stack height or credit 159 CAA section 169A(g)(2), 40 CFR Francisco Don˜ ez, EPA, (February 3, 2012). resulting from any other dispersion technique is 51.308(e)(1)(ii)(A). 165 ‘‘BART Analysis for Cholla Unit 2,’’ Prepared prohibited in the development of emission 160 70 FR 39129. for APS by CH2MHill (January 2008). limitations by 40 CFR 51.118 and 40 CFR 51.164.’’

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00039 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72550 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

Comment: One commenter (APS) projected sulfuric acid emissions after modeling; the maximum area base case stated that EPA incorrectly applied installation and operation of SCR using impact is 3.51 dv at Petrified Forest H2SO4 mitigation factors from an the EPRI methodology is dependent on compared to 4.53 dv previously. But for Electric Power Institute (EPRI) report 167 future decisions made by the facility on some areas the impacts from controls in reaching its conclusion that H2SO4 the type of SCR catalyst and number of declined more than the impacts from production is not a problem with SCR layers used, as well as numerous the base case, leading to the somewhat at Cholla. The commenter stated that assumptions about loss to downstream surprising result that the improvement this factor is actually 90 percent rather components (i.e., air preheaters and due to controls actually increased than 99 percent in the report, but that baghouses), the true values of which are relative to the original modeling. The this factor only applies to sub- currently not yet defined or known for maximum area benefit of SCR in the bituminous coal because of the high Cholla. An increase in sulfuric acid new modeling is 1.55 dv compared to calcium content in the ash of these emissions from the installation of SCR 1.34 dv in the original. The cumulative coals. The commenter stated that testing may trigger major modification PSD area benefit decreased very slightly to at the Four Corners Power Plant (FCPP), permit requirements at a low threshold 7.19 dv compared to 7.21 in the original. which has similar coal ash calcium of seven tons per year.168 Based on this improved estimate of content to that at Cholla, indicates that Preconstruction permitting review may sulfate emission based on the EPRI 15 percent removal by the fabric filters also be triggered from significant method, the case for SCR appears to be would be likely. The commenter stated emissions increases of PM2.5 from SCR strengthened, since the maximum that the H2SO4 emissions created by the installation at Cholla. If one of these visibility improvement is larger than SCR will exceed the NSR significance pollutants triggers PSD, the permitting originally estimated. authority must provide an Additional level, will result in costs associated with e. Other Comments the H2SO4 emissions, and will reduce Impact Analysis under the PSD the improvement in visibility program. The PSD program also requires Comment: One commenter (NPS) anticipated by the retrofitting with SCR. the permitting authority to determine agreed with EPA’s conclusions on Another commenter (ADEQ) also BACT for pollutants that triggered PSD. Cholla that the visibility improvement stated that EPA discounts the impact of For these reasons, Region 9 has associated with the most stringent sulfuric acid mist that will be generated determined that for Cholla, emission option (SCR with LNB and OFA) is by SCR and overestimates the acid mist limits and monitoring requirements for substantial; that SCR with LNB and removal rate. The commenter indicated sulfuric acid are more appropriately OFA is cost-effective on an average basis that testing at another facility shows reviewed in the preconstruction as well as on an incremental basis when compared to the next most stringent H2SO4 removal to be closer to 57 permitting process. percent rather than EPA’s assumed 99 Nevertheless, EPA conducted option (SNCR with LNB and OFA); and percent removal. The commenter noted additional CALPUFF modeling to assess that NOX BART for Cholla Units 2, 3 that if H SO emissions increase above the visibility effect of increased sulfuric and 4 is SCR with LNB and OFA, with 2 4 an associated emission limit for NO on the PSD significance threshold, a PSD acid due to the SCR catalyst. One X each of the units of 0.050 lb/MMBtu, permit and BACT analysis would be scenario used the existing modeling for based on a rolling 30-boiler-operating- required. EPA’s BART analysis fails to Cholla, but added in SCR sulfate calculated by the method in the EPRI day average. consider the costs associated with likely Response: We acknowledge the BACT requirements of low oxidation document. Since the existing modeling used sulfate calculated using PM comment. catalyst, fuel additives or sorbent Comment: One commenter (APS) injection with a polishing baghouse. speciation spreadsheets provided by the National Park Service, this scenario estimated that EPA’s proposed controls Response: EPA’s decision to discount on Cholla Units 2 and 3 will cost $248 the increase of H SO caused by mixes two calculation methods and may 2 4 not be reliable. The sulfate in the million and $103 million, respectively, oxidation from the SCR catalyst was and increase the costs of electricity from actually based on the 90 percent control existing modeling is so large that the additional SCR sulfate from the EPRI those units by over 25 percent. The figure; we erroneously wrote 99 percent commenter stated that given the current (which applies to ammonia reduction method increases total sulfate by only about 5 percent. Visibility benefits only market price for natural gas, the from a wet scrubber). This figure is from proposed BART requirements, expected the 0.10 percent penetration for decreased by about three percent at Petrified Forest, and by an even smaller coal ash regulations, and potential baghouses, the only one available for fraction at other areas. To assess the future carbon legislation could baghouses in the EPRI report. It is not SCR sulfate effect in a more consistent jeopardize the long-term economic clear that results from the testing at manner, EPA calculated sulfate using viability of the entire plant. The FCPP referenced by the commenter may the EPRI method throughout the base commenter also stated that EPA did not be applied directly to Cholla given the case for SCNR, and for SCR. All cases consider the impacts of requiring SCR differences between the facilities. In used a merged stack for Units 2 and 3 on ratepayers’ monthly bills, which addition, the full test results were not and consistent speciation for all units would be about 2 percent to provided, so we cannot rely on the (formerly the speciation for Unit 2 accommodate SCR alone. In addition, commenter’s figures. differed from the others). The sulfate the commenter is concerned about In any case, EPA does not believe that emissions from the EPRI method are potential impacts on the transmission BART is the appropriate context for much lower than from the NPS grid in Arizona, the local economy due addressing this issue. Actual spreadsheets, but SCR increases that to lost jobs, and a reduced diversity in measurements of baseline sulfuric acid amount by a factor of six (even with the APS’s fuel mix if Cholla was to close. emissions have not yet been determined increase the total is still far lower than Response: It is not EPA’s intention to at Cholla. Moreover, the calculation of used in the original modeling). The endanger the economic viability of visibility impacts for all cases are Cholla or to place an undue burden on 167 Estimating Total Sulfuric Acid Emissions from Stationary Power Plants, Version 2010a, 1020636, substantially lower than in the former APS’s customers. Neither the CAA nor Technical Update, Electric Power Research the RHR requires states or EPA to Institute, April 2010). 168 See 40 CFR 52.21(b)(23)(i). consider the affordability of controls,

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00040 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72551

ratepayer impacts or potential job losses review, the specific economic effects, year. Under Title II of UMRA, EPA has as part of a BART analysis. Rather, they parameters, and reasoning.173 determined that this rule does not require consideration of ‘‘the costs of Thus, only under ‘‘unusual contain a Federal mandate that may compliance, the energy and non-air circumstances’’ where a potential result in expenditures that exceed the quality environmental impacts of control option is expected to have a inflation-adjusted UMRA threshold of compliance, any existing pollution ‘‘severe impact on plant operations’’ or $100 million (in 1996 dollars) by State, control technology in use at the source, ‘‘result in significant economic local, or Tribal governments or the the remaining useful life of the source, disruption or unemployment’’ can we private sector in any one year. Even and the degree of improvement in consider economic effects as part of a using the higher cost estimates in our visibility which may reasonably be BART determination. In this case, APS supplemental analysis for the FIP we are anticipated to result from the use of has provided no evidence to support its finalizing today, we estimate that the such technology.’’ 169 assertions that our proposed FIP would total annual costs in the aggregate will APS’s comments appear to be based result in significant rate increases, not exceed $65 million.174 Finally, this in part on a misunderstanding that an jeopardize the plant’s operations, or rule is not subject to Executive Order analysis of ‘‘non-air quality result in any other economic effects. In 13211 (66 FR 28355 (May 22, 2001)), environmental impacts’’ must include the absence of such evidence, APS’s because it is not a significant regulatory economic effects. In fact, the plain assertions regarding plant shutdown, action under Executive Order 12866. language of the statute, as well as the rate increases and job losses are Comment: One commenter (APS) RHR, makes clear that this factor is speculative, and we cannot consider disagreed with EPA’s conclusion that limited to non-air quality environmental them as part of our BART the use of anhydrous ammonia does not impacts.170 The BART Guidelines note determination. pose significant additional safety that examples of such impacts would Comment: One commenter concerns compared to aqueous include ‘‘solid or hazardous waste (PacifiCorp) stated that because the ammonia and urea. The commenter generation and discharges of polluted regional haze actions in Arizona, contends that while anhydrous water from a control device.’’ 171 Wyoming, Colorado and elsewhere will ammonia would be transported by rail, The BART Guidelines do allow for have an impact of $100 million or more safety concerns are not eliminated (but do not require) the consideration of on the company and its customers, EPA because the severity of damage in an ‘‘significant economic disruption or must conduct the regulatory analyses accident can be much greater, if less unemployment’’ as part of ‘‘energy required by the Unfunded Mandates frequent than truck accidents, and impacts.’’ Specifically, the Guidelines Reform Act (UMRA) and Actions constitutes a much higher risk after provide that: Concerning Regulations That delivery. Due to the hazards of moving and storing anhydrous ammonia, the * * * the energy impacts analysis may Significantly Affect Energy Supply, consider * * * whether a given alternative Distribution, or Use (Executive Order Department of Homeland Security and would result in significant economic 13211) before reaching conclusions EPA have additional requirements for disruption or unemployment. For example, regarding BART controls or imposing a anhydrous ammonia that result in where two options are equally cost effective regional haze FIP. additional costs to use it. Urea costs and achieve equivalent or similar emissions Response: The commenter is more than anhydrous ammonia, but it is reductions, one option may be preferred if combining separate regulatory actions. safer and less expensive to use and the other alternative results in significant The commenter is not correct in store. Due to these factors the 172 disruption or unemployment. aggregating the potential private sector commenter stated that SNCR and SCR The Guidelines also allow for mandate of separate rules to evaluate costs should include the use of urea consideration of ‘‘affordability’’ as part whether UMRA applies. UMRA defines rather than anhydrous ammonia. of the ‘‘costs of compliance’’ under the term ‘Federal private sector Response: The BART analyses certain circumstances: mandate’ to mean any provision in submitted by APS indicate that the annualized cost of urea at each of the 1. Even if the control technology is cost regulation that would impose an effective, there may be cases where the enforceable duty upon the private Cholla units would be less than the installation of controls would affect the sector. Under UMRA, the term annualized cost of anhydrous viability of continued plant operations. ‘‘regulation’’ or ‘‘rule’’ means any rule ammonia.175 In addition, the cost 2. There may be unusual circumstances for which the agency publishes a estimates provided by APS in comments that justify taking into consideration the general notice of proposed rulemaking. are based on the use of urea as a reagent. conditions of the plant and the economic The rule being finalized today is limited Accordingly, we have used the cost for effects of requiring the use of a given control to addressing the obligations of three urea in our supplemental cost analysis. technology. These effects would include facilities in Arizona and does not Comment: One commenter (APS) effects on product prices, the market share, noted that Cholla has a long history of and profitability of the source. Where there include other regional haze actions are such unusual circumstances that are occurring in separate rulemakings, such installing pollution control equipment, judged to affect plant operations, you may as for Wyoming and Colorado. take into consideration the conditions of the Under section 202 of UMRA, before 174 Using total annual costs from our plant and the economic effects of requiring promulgating any final rule for which a supplemental analysis, annual aggregate cost equals $64,378,422. This amount consists of: $9,906,206 the use of a control technology. Where these general notice of proposed rulemaking for Cholla Unit 2, $9,448,912 for Cholla Unit 3, and effects are judged to have a severe impact on was published, EPA must prepare a $13,590,853 for Cholla Unit 4 (See Table 10 of this plant operations you may consider them in written statement, including a cost- NFRM); $12,103,941 for Coronado Unit 1 and the selection process, but you may wish to benefit analysis, if that rule includes $235,982 for Coronado Unit 2 (See Tables 15 and 13 of this NFRM); and $9,546,264 for each of provide an economic analysis that any ‘‘Federal mandates’’ that may result demonstrates, in sufficient detail for public Apache Units 2 and 3 (See Table 5 of this NFRM). in expenditures to State, local, and 175 See BART Analysis for Cholla Unit 2, Tribal governments, in the aggregate, or Appendix A, Economic Analysis, Input 169 CAA section 169A(g)(2), 42 U.S.C. 7491(g)(2); to the private sector, of $100 million or Calculations; BART Analysis for Cholla Unit 3, 40 CFR 51.308(e)(1)(ii)(A). Appendix A, Economic Analysis, Input 170 Id. more (adjusted for inflation) in any 1 Calculations; BART Analysis for Cholla Unit 4, 171 BART Guidelines section IV.D.4.h Appendix A, Economic Analysis, Input 172 Id. section IV.E.2. 173 Id. section IV.E.3. Calculations.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00041 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72552 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

has participated in a voluntary for Coronado, and that EPA’s failure to analysis for Coronado is summarized in emissions reduction project, and has consider these controls in its BART Table 15.179 spent over $473 million to reduce assessment makes the proposed rule b. Control Efficiencies emissions. While Unit 1 at Cholla is not invalid. The commenter added that BART-eligible, it is equipped with a emission reductions already achieved at Comment: One commenter (SRP) wet-tray absorber to control SO2, a fabric the facility using LNB with OFA should stated that the SNCR NOX emission rate filter to control particulates, and LNB not be ignored in EPA’s analysis simply evaluated by EPA is incorrect. The with OFA to control NOX emissions. because EPA delayed review of ADEQ’s commenter cited an SNCR conceptual Unit 2 is BART-eligible and has a SIP until 2012. The commenter design estimate prepared by S&L mechanical dust collector for particulate concluded that EPA should give (attached to the submission) asserting control, a wet flooded-disk venturi deference to the baseline emissions that, based on an initial review of SNCR scrubber and absorbers to control SO2, period selected by the State in its SIP implementation at Coronado, the additional particulate controls, and LNB analysis and fully consider LNB with expected NOX reductions would be 25 with OFA to control NOX emissions. OFA as an appropriate basis for BART percent and notes that additional Units 3 and 4 have wet open-spray FGD emission limitations for Coronado. studies would be needed to guarantee absorber to control SO2, fabric filters to Another commenter (NPS) preferred this performance. According to the control particulates, and LNB with OFA the use of a baseline period before the commenter, this estimate also was to control NOX emissions. Unit 2 is installation of LNB with OFA instead of verified by an independent vendor, scheduled to upgrade its SO2 and the post-installation period (May 16, FuelTech, whose assessment was also particulate controls to be identical to 2009 to December 31, 2010) used by attached to the submission. Units 3 and 4 by January 1, 2016. EPA. For Unit 2, the commenter stated The commenter (SRP) assumed that Response: We appreciate that APS has that the federally enforceable limit of EPA evaluated an emission limit that is installed various controls on the Cholla 0.080 lb/MMBtu is a realistic depiction based on a higher reduction efficiency units over the last several years and we of future emissions even though the (i.e., 30 percent) applied to a starting have taken these existing controls into required SCR system has not yet been NOX emission limit of 0.30 lb/MMBtu. account as part of our BART analysis for installed. According to the commenter, given 176 NOX. However, we note that, even Response: As explained in the general Coronado’s current NOX emissions limit with all of these new controls, discussion regarding selection of of 0.320 lb/MMBtu following the emissions from Cholla still cause baseline periods above, we disagree that installation of LNB with OFA on each visibility impairment at nine Class I our use of updated baseline periods for of the units and an SNCR control areas and contribute to impairment at an BART determinations is inappropriate efficiency of 25 percent, the appropriate additional two areas.177 or inconsistent with the CAA or the NOX emission rate to use in the BART Comment: One commenter (APS) RHR. Moreover, updating the baseline analysis would be 0.24 lb/MMBtu, requested that EPA allow the flexibility did not eliminate LNB with OFA from rather than EPA’s assumed value of 0.21 of averaging NOX emissions across all consideration as BART for Coronado lb/MMBtu. The commenter contended the BART-eligible units at the plant. The Unit 1, since existing controls can that this NOX emission rate (i.e., 0.24 lb/ commenter stated that allowing for this constitute BART, if additional controls MMBtu) represents a level that can flexibility would make no difference are not warranted based on the five- likely be achieved on a consistent basis from a visibility improvement factor analysis. For example, EPA based on input from SRP’s vendors who perspective. recently approved a determination by have specific SNCR implementation Response: We agree with the Colorado that existing LNB at Comanche experience. commenter, and have finalized a single Units 1 and 2 constituted BART where Response: We partially agree with this NOX emission limit across Cholla Units ‘‘the State determined that the added comment. Coronado Unit 1 currently 2, 3 and 4. expense of achieving lower limits operates with a federally-enforceable NO emission limit of 0.320 lb/ 3. Comments on Coronado Units 1 and through different controls was not X MMBtu.180 A review of recent emission 2 reasonable based on the high cost- effectiveness [$9,900/ton] coupled with data in CAMD indicates NOX emission a. Selection of Baseline Period the low visibility improvement (under levels below this limit. As noted in our Comment: Two commenters (ADEQ 0.2 dv) afforded.’’ 178 In the case of response to SRP’s comments regarding and SRP) stated that EPA’s selected Coronado, by contrast, the cost- SCR, we agree that when using an baseline emissions period effectiveness of post combustion annual average design emission rate to inappropriately eliminated controls is reasonable and the expected establish a rolling 30-day limit that will consideration of LNB with OFA as a visibility improvements are substantial, apply during periods of startup, viable BART control strategy. SRP as explained below. Nonetheless, in shutdown, and malfunction events, it is asserted that EPA’s decision to include order to address the commenter’s appropriate to include some type of LNB with OFA in its baseline NOX concerns that we did not properly measure that provides a compliance emissions estimate cannot, consistent consider LNB with OFA as a potential margin. with the BART rules, foreclose control option, and therefore precluded In submitted comments, SRP consideration of those controls as BART a BART determination of LNB with provided a conceptual design estimate OFA, we have used a baseline period of for SNCR which was based upon 25 176 77 FR 42854, July 20, 2012 (noting that ‘‘[t]he 2001–2003, which corresponds to the percent control efficiency (incremental baseline emissions used by EPA reflect current fuels period used in SRP’s original BART from LNB) and a resulting emission rate and control technologies in place at the facilities, of 0.24 lb/MMBtu. While this control as well as regulatory requirements the facilities will analysis. Our supplemental cost be required to meet independent of EPA’s BART efficiency is less than the 30 percent determination.’’). 178 77 FR 18052, 18066 (March 15, 2012) 177 See 77 FR 42861, July 20, 2012, Table 20 (Proposed Rule); pre-publication version of Final 179 A spreadsheet titled ‘‘Supplemental Cost (showing baseline impacts from Cholla of over 1 dv Rule, signed September 10, 2012, available at: Analysis 2012–11–15.xls’’ is in the docket. at nine Class I areas, and impacts of over 0.5 dv at http://www.epa.gov/region8/air/FinalActionOn 180 See Coronado Title V Permit, Attachment B, eleven areas). ColoradoRegionalHazePlanSep2012.pdf. section II.E.1.a.ii.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00042 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72553

control efficiency used by our OFA and a corresponding emission Coronado, and the recent PSD consent contractor, we consider it to be a limit of 0.320 lb/MMBtu, making Units decree should, pursuant to the BART reasonable estimate based upon the 1 and 2 currently subject to a 0.320 lb/ Guidelines, be deemed to satisfy BART. vendor quotes provided by SRP.181 MMBtu NOX limit. The commenter Response: We do not agree that we When using a control efficiency of 25 added that Unit 2 will be subject to a improperly ignored the existing consent percent and our baseline period of LNB 0.080 lb/MMBtu NOX emission limit as decree in our proposed BART performance for Coronado Unit 1, we soon as the SCR for that unit is installed determination for NOX at Coronado, estimate an annual average SNCR and operational (i.e., by June 1, 2014), since we specifically took the consent emission rate of 0.22 lb/MMBtu. pursuant to the consent decree, a limit decree into account throughout our NOX For the purposes of our cost that is significantly more stringent than BART analysis.182 We also do not agree calculations and visibility modeling, what the state determined to be BART that the Coronado consent decree however, we have retained the use of for Coronado. represents BACT or BART for NOX. our original SNCR emission rate (0.21 The commenter (SRP) asserted that While the consent decree concerned lb/MMBtu). A less stringent SNCR the consent decree controls are better alleged violations of the PSD provisions emission rate, by itself, would primarily than BART. The commenter pointed out of the CAA, it does not indicate that its make the next most stringent control that once SCR is installed on Unit 2, the provisions represent either BACT or option, SCR, appear to remove a greater facility will be subject to a plant-wide BART. Rather, it specifically provides amount of emissions. This in turn emission limit of 7,300 tons of NOX per that: year under the consent decree which, would make the SCR control option Compliance with the terms of this Consent appear more incrementally cost- according to the commenter, translates Decree does not guarantee compliance with effective by removing a greater amount to an effective emission rate of 0.20 lb/ all applicable federal, state, or local laws or of emissions, relative to SNCR, for the MMBtu for Coronado as a whole, and is regulations. The emission rates and removal same cost. As discussed in our proposal more stringent than the state’s NOX efficiencies set forth herein do not relieve and in response to comments, we BART determination and EPA’s SRP from any obligation to comply with already consider SCR to be cost- presumptive NOX limits. other state and federal requirements under 183 effective. It is not determinative to our The commenter (SRP) also contended the Clean Air Act * * * decision to find that SCR is ‘‘even that EPA’s BART rules support the While the BART Guidelines provide more’’ incrementally cost-effective. conclusion that the existing and that NSR/PSD settlement agreements In the context of establishing a BART currently planned controls are better may represent BART in some instances, emission limit consistent with the use of than NOX BART because those controls they do not establish a presumption that SNCR technology, however, we would and emission rates were agreed to by such settlements represent BART, nor use the annual average SNCR emission SRP and EPA to resolve allegations of do they indicate that a BART analysis is rate of 0.22 lb/MMBtu as our basis, violations of certain requirements of the unnecessary where such a settlement rather than our original estimate based PSD program for both units. According exists.184 In Coronado’s case, we do not on 30 percent SNCR control efficiency. to the commenter, those limits are agree that the consent decree represents intended to reflect compliance with the As noted in a separate response, when BART for NOX for either unit or for the using an annual average design PSD program’s BACT requirements. The facility as a whole. Nonetheless, we are emission rate to establish a rolling 30- commenter noted that BACT, by taking the consent decree into account definition, reflects the maximum degree day limit that would apply during in our BART determination for NOX at periods of startup, shutdown, and of control for new facilities or existing Coronado, as described below. malfunction, we consider it appropriate facilities undergoing a major Comment: In arguing against the to provide some type of measure that modification while BART is to apply to achievability of EPA’s proposed limit, unmodified existing sources. So BACT provides a compliance margin for such two commenters (AUG and SRP) cited a would be expected to be more stringent, events. A 0.24 lb/MMBtu emission report prepared by RMB Consulting & and certainly not less stringent, than limit, as requested by SRP, established Research, Inc. (RMB) for the San Juan BART. The commenter quoted a recent on a rolling 30-day average represents Generating Station in New Mexico, EPA statement about the Four Corners about a 10 percent increase from the which reportedly states that the 0.05 lb/ Power Plant indicating that BART need 0.22 lb/MMBtu annual average emission MMBtu limit imposed on that facility not be equivalent to BACT (citing 77 FR rate. We would consider this magnitude does not represent a consistently 51620, 51636, August 24, 2012). of upward revision appropriate to The commenter (SRP) asserted that achievable level of emissions for the accommodate startup, shutdown, and the BART rules reflect this units at the facility. In addition, SRP malfunction events as well as the unit understanding, providing that PSD contracted with RMB and Sargent and cycling nature of Coronado Unit 1. As settlement agreements generally satisfy Lundy (S&L) to review the ability of the a result, we would consider the BART BART requirements (citing 70 FR Coronado units to achieve the 0.050 lb/ emission limit corresponding to the 39164). According to the commenter, MMBtu emission limit proposed by EPA SNCR with LNB and OFA option to be EPA recently recognized this principle 182 0.24 lb/MMBtu. in its final regional haze rule for North See 77 FR 42849–42850, July 20, 2012, Comment: One commenter (SRP) (summarizing terms of consent decree), 42861– Dakota in which EPA concluded that it 42862 (describing consideration of consent decree stated that EPA improperly ignored the was appropriate to rely on North requirements in baseline for Coronado analyses), Coronado consent decree in its selection Dakota’s BACT determination for the 42863 (noting potential effect of consent decree of the proposed BART controls for NOX. two units at the Milton R. Young Station activities on cost analysis), 42864 (proposing The commenter noted that ADEQ emission limit of 0.080 lb/MMBtu and compliance (0.36 lb/MMBtu and 0.35 lb/MMBtu) to deadline of June 1, 2014 at Coronado Unit 2, determined that NOX BART for satisfy BART because emissions control consistent with the emission limit in the consent Coronado Units 1 and 2 is LNB with technology had not changed appreciably decree). since that BACT determination (citing 183 Consent Decree in United States v. Salt River 181 Project, CV 08–1479–PHX–JAT (entered December Although the report cites lower NOX 77 FR 20897, April 6, 2012). The concentrations, due to the lower inlet NOX emission 19, 2008). rate, removal efficiency is also reduced making it commenter stated that a similar 184 BART Guidelines, 40 CFR Part 51, Appendix more difficult to reduce NOX emissions. situation is present in the case of Y, section IV.C.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00043 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72554 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

using SCR control technology. Their Finally, both AUG and SRP noted that examined both Coronado Units 1 and 2, reports were submitted as attachments the BART Guidelines authorize included the following assumptions: to the commenter’s submission. application of BART emission limits on • Five to six startups (1 cold/ According to the commenter, both a plant-wide basis, rather than a unit-by- remainder warm) per month (which is consultants concluded that a NOX BART unit basis, and that use of plant-wide the maximum observed based on 2001 limit of 0.050 lb/MMBtu is not limits would not affect the expected to 2011 historical performance); achievable at Coronado on a 30-day visibility benefits of controls. Therefore, • Startup emissions based on the rolling average that includes periods of they requested that EPA allow for plant- maximum value observed during that startup, shutdown, and malfunction. wide averaging at Coronado. startup period; The commenter made the following Response: We partially agree with this • Non-startup periods of operation arguments against the achievability of a comment. As noted by the commenters, based on historical load operation, limit of 0.050 lb/MMBtu relying first on the BART Guidelines recommend that which consists of a mixture of low load RMB’s analysis and then on S&L’s states ‘‘consider allowing sources to and high-load cycling operation; analysis. ‘average’ emissions across any set of • Inclusion of a low load temperature RMB’s analysis stated that EPA did BART-eligible emission units within a control system; and not adequately consider the impact of fenceline * * *’’ 185 Given that such a • Maintaining the catalyst guarantee startup and shutdown emissions or the ‘‘bubbling’’ approach would not of 0.04 lb/MMBtu during full load, ability to measure such emissions in its diminish the visibility benefits of steady-state operations over the life of BART determination. RMB examined controls, we have decided to finalize a the catalyst. operating data from 2001 to 2011 in single plant-wide limit across the two The analysis performed by S&L order to identify the number of startup Coronado units. examined only Coronado Unit 2, and events (both ‘‘cold’’ and ‘‘warm’’ starts) In analyzing what emission limit was one element of S&L’s broader and shutdown events associated with would represent BART for NOX on a analysis examining the ability of each unit. RMB’s analysis shows that plant-wide basis, we have taken a Coronado Unit 2 to meet a limit more the average number of startup/shutdown number of factors into consideration. In stringent than the 0.080 lb/MMBtu limit events for Coronado Units 1 and 2 is one our proposal, we used an annual in the consent decree. The analysis per month (each), and that the average design value for SCR of 0.050 performed by S&L was based on the maximum number of startup/shutdown lb/MMBtu at Coronado Unit 1 and following assumptions: events is five per month (Coronado Unit proposed an emission limit for this • One to three startup events per 1) and six per month (Coronado Unit 2). same value on a rolling 30-day average. month; RMB then developed a computer model At Coronado Unit 2, we proposed an • Non-startup periods of operation to estimate the 30-day rolling average emission limit of 0.080 lb/MMBtu, but based entirely on low load cycling the Coronado units could achieve based solicited comment on whether a more scenario (40–100 percent gross load upon the emissions profile of these stringent limit would be feasible and cycling); startup/shutdown events, the maximum cost-effective for Unit 2. SRP submitted • Inclusion of a low load temperature number of startup/shutdown events, comments stating that an emission rate control system; 187 and and an assumption of a NOX emission of 0.05 lb/MMBtu was not achievable by • Maintaining the catalyst guarantee rate of 0.04 lb/MMBtu over the life of either unit, due to the startup/shutdown of 0.04 lb/MMBtu during full load, the catalyst. RMB’s analysis indicates operating profile of the Coronado units. steady-state operations over the life of that the maximum 30-day average the As noted in other responses, BART the catalyst. units could achieve is well above 0.050 limits apply at all times including The results of both of these analyses lb/MMBtu. periods of startup, shutdown, and indicates that the Coronado units could S&L’s analysis focused on the ability malfunction. As a result, we agree with achieve a rolling 30-day emission rate in of Coronado Unit 2, which has been commenters that when establishing a the range of 0.053 to 0.072 lb/MMBtu designed to achieve a 0.08 lb/MMBtu rolling 30-day BART emission limit that based on all the assumptions listed emission rate, to achieve a lower 0.05 is based upon an annual average design above. We acknowledge that different lb/MMBtu emission rate. S&L’s analysis value, it is appropriate to provide a assumptions, such as using fewer considered multiple design changes and compliance margin for periods of startups per month, or using a load examined their potential impact on startup and shutdown. Therefore, we operating profile during non-startup reducing the design emission rate, as have taken into consideration the periods that corresponded to a greater well as the costs and design/ startup/shutdown operating profile of fraction of high-load cycling operations, construction time associated with these the Coronado units. could produce a lower range of emission options. S&L concluded that, at a In submitted comments, SRP included values. However, we find that the minimum, SRP would be required to reports prepared by S&L and RMB assumptions used in both analyses are install a low load temperature control Consulting summarizing an analysis reasonable based on the historic system designed to increase flue gas performed to determine the rolling 30- performance data supplied by SRP and temperatures at the SCR inlet during day emission rates the units could its consultants. Therefore, we have periods of low load cycling to achieve achieve when accounting for startup concluded that a 0.050 lb/MMBtu any additional reduction in average and shutdown events, as well as the emission rate is not achievable on a NOX emissions. S&L’s analysis load cycling operating profile of the 186 rolling 30-day average at either of the concluded that even with a low-load plant. The analyses in the two reports Coronado units.188 Nonetheless, we note temperature control system, Unit 2 were based on slightly different could not consistently achieve the assumptions. RMB’s analysis, which 187 S&L’s analysis also included emission proposed limit when periods of low- modeling of Coronado Unit 2 without the low load load cycling, startup and shutdown are 185 BART Guidelines, 40 CFR Part 51, Appendix temperature control system, which, as discussed in Y, section V. further detail below, is not part of the current SCR taken into account, and could only 186 In addition to the final reports, SRP provided design. achieve within the range of 0.053 to certain supporting spreadsheets upon request. We 188 Nonetheless, we note that the emission 0.072 lb/MMBtu. have placed these spreadsheets in the docket. modeling results (particularly those produced by

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00044 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72555

that the results of these analyses whether the SCR system for Unit 2 system during periods of low load.190 (particularly those produced by the could achieve an emission rate more Without this control system, the RMB report) indicate that Coronado stringent than the 0.080 lb/MMBtu limit Coronado Unit 2 SCR system will not Unit 1 could meet a 0.050 lb/MMBtu in the consent decree for which the SCR operate during periods of low load. limit on an annual average basis. As a was designed, S&L examined a number Under EPA’s visibility regulations, result, we conclude that 0.050 lb/ of different potential measures. One of ‘‘BART means an emission limitation MMBtu is appropriate as an annual these measures was the installation of a based on the degree of reduction average design value, but not as 30-day low load temperature control system, achievable through the application of rolling average emission limit at the which the current SCR design for Unit the best system of continuous emission Coronado units. 2 does not include. reduction * * *’’191 While SCR With respect to Coronado Unit 2, we As described in the S&L report, represents the most stringent technology have also taken into account the fact periods of low load operation generally that Unit 2 is already subject to a consist of operation between loads of for NOX control, an SCR system that is consent decree limit of 0.080 lb/MMBtu 138 MW to 270 MW (operation above designed not to function during a period with a compliance deadline of June 1, 270 MW can be considered ‘‘high’’ of operation that represents a substantial 2014. We consider the SCR system that load). Broadly speaking, the temperature fraction of the unit’s overall operating SRP has designed to meet this limit to in the SCR system will fall below 599 profile cannot be considered constitute ‘‘pollution control equipment degrees F during these periods of low continuous. In examining the in use at the source.’’ Therefore, load operation, which is the minimum installation of a low load temperature consistent with the BART Guidelines, temperature required for effective NOX control system as an upgrade option to we have considered various ways in control. A low load temperature control Coronado Unit 2, we note that the S&L which the performance of the current system increases the temperature at the report estimated the costs for the low SCR design for Unit 2 could be SCR inlet in order to maintain 599 load temperature control system as improved.189 In its analysis examining degrees F, allowing operation of the SCR shown in Table 13.

TABLE 13—S&L’S COST ESTIMATES FOR LOW LOAD TEMPERATURE CONTROL SYSTEM

Annualized Annual O&M Total annual Capital cost 1 Measure capital cost 2 costs costs ($) ($/yr) ($/yr) ($/yr)

Low load temperature control system ...... $2,500,000 $235,982 ...... $235,982 1 Represents the mid-range value of S&L’s estimate of capital costs. 2 Capital costs annualized using a 7 percent interest rate over a 20 year lifetime.

Although it is not clear what annual disagree. Specifically, SRP bases this result, although equipped with an SCR average emission rate can be achieved comment on the visibility improvement system, the maximum 24-hour average by Coronado Unit 2 with installation of associated with a 0.080 lb/MMBtu limit emission rate for Coronado is more a low load temperature control system, and a lower value such as 0.07 or 0.05. accurately represented by an emission the upper range of rolling 30-day Visibility modeling, however, is based rate corresponding to LNB and OFA, emission rates modeled for Coronado on the highest emission rate observed and not SCR. Unit 2 is 0.072 lb/MMBtu. We consider on a 24-hour average, not on a 30-day We consider this distinction crucial. this a conservative estimate (i.e., a high or annual average basis. Since Coronado In our base case modeling runs, the estimate in this case, as the annual Unit 2 is not equipped with a low load maximum 24-hour average emission rate average number will certainly be lower temperature controller and therefore modeled for Coronado Unit 2 was than the 30-day value), and have used cannot operate the SCR during periods represented by a NOX emission rate of this emission rate with the cost of low load operation, emissions from 0.08 lb/MMBtu, corresponding to the information contained in the S&L Coronado Unit 2 during these periods emission limit for SCR in the consent report, to calculate the cost-effectiveness correspond to operation of LNB with decree. However, the highest 24-hour value shown in Table 14. Installation of OFA. A review of Coronado Unit 2’s average emission rate is more accurately a low load temperature controller operating history since June 2011, represented by a 24-hour period of low results in a cost-effectiveness of $1,900/ which is approximately when LNB was load operation, where the SCR system ton, which is in a range that we consider installed, indicates several instances in would not be operating. Based on Acid cost-effective. which it operates at low load for periods Rain Program data reported to CAMD, In addition, SRP stated that it that can exceed a 24-hour calendar day. this corresponds to a NOX emission rate considered the incremental visibility Based on the Acid Rain Program data of 0.23 lb/MMBtu and 13,684 lb/day.193 benefit of an emission limit more provided by SRP and included in By allowing the SCR system to run stringent than 0.080 lb/MMBtu to be CAMD, the longest such period of during all loading periods, the insignificant. In relation to installation continuous low load operation extended installation of a low load temperature of a low load temperature controller, we from May 20 to May 22, 2012.192 As a control system would result in a

the RMB report) indicate that Coronado Unit 1 systems. If SCR were to be installed on Coronado 193 This represents the emission rate on April 1, could meet a 0.050 lb/MMBtu limit on an annual 1, for example, the information SRP has provided 2012, which is the highest emitting day that average basis. As a result, we conclude that the use indicates that such a system would include a low consisted of 24 consecutive hours of low-load of a 0.050 lb/MMBtu as annual average design value load temperature control system. operation, as identified in ‘‘SRP 2 NOX analysis in our proposal was appropriate. 191 40 CFR 51.301. (EPA edits).xls’’ and ‘‘Coronado 2 2011–12Q3 NOX 189 BART Guidelines, 40 CFR Part 51, Appendix 192 We have identified these dates in both sets of Emission Data (daily).xls’’. Y, section IV.D.3. data, per ‘‘SRP 2 NOX analysis (EPA edits).xls’’ and 190 We note that this is not an unusual control ‘‘Coronado 2 2011–12Q3 NOX Emission Data system, and is commonly included in typical SCR (daily).xls’’.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00045 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72556 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

decrease in the maximum 24-hour affected Class I area is 0.52 (Gila visibility improvement substantial, average emission rate from 0.21 lb/ Wilderness). Cumulatively, across all of especially when taking into MMBtu to 0.080 lb/MMBtu. The the affected Class I areas, this results in consideration the cost-effectiveness of visibility improvement associated with visibility improvement of 2.64 installing a low load temperature this emission decrease at the single most deciviews. We consider this degree of control system.

TABLE 14—CORONADO UNIT 2: COST-EFFECTIVENESS

Emission Emission rate 1 Cost- factor Removed Annual cost effectiveness (lb/MMBtu) (lb/hr) (tpy) (tpy) ($/yr) ($/ton)

SCR+LNB+OFA (no low load temp ctrl sys) ...... 0.080 319 1,242 ...... SCR+LNB+OFA (with low load temp ctrl sys) ...... 0.072 287 1,118 124 235,982 1,900 1 Emissions calculated based on 3,984 MMBtu/hr and 0.89 capacity factor, as used in the TSD for our proposal.

In recognition of the work already Comment: While supporting EPA’s provide no source-specific data performed by SRP to meet the consent determination that SCR is BART for explaining why SCR at Unit 1 could not decree emission limit of 0.080 lb/ Coronado Unit 1, one commenter achieve a 0.05 lb/MMBtu NOX emission MMBtu for Unit 2, and to avoid (Earthjustice) stated that lower NOX limit. The commenter asserted that, in interfering with SRP’s ability to meet emission limits are cost-effective and contrast, the Sahu report explains why that requirement by the deadline of June achievable. For Unit 1, the commenter an even lower 0.04 lb/MMBtu emission 1, 2014, we have decided not to require made the following two points based on limit is achievable at Unit 1. a BART emission limit for Coronado 2 a report (the ‘‘Sahu report’’) submitted Accordingly, the commenter believes more stringent than 0.080 lb/MMBtu. with the comments. First, SCR can that EPA should not weaken its BART Instead, we are finalizing a plant-wide achieve even greater NOX reductions at proposal as SRP requested. NOX emission limit for Coronado of less cost than EPA’s calculations. EPA Response: We disagree with the 0.065 lb/MMBtu on a rolling 30-day failed to analyze whether an emission commenter’s assertion that our BART average, which will provide a sufficient limit lower than 0.05 lb/MMBtu is analysis should have examined the compliance margin for startup and achievable and cost-effective with SCR potential for lower ‘‘boiler-out’’ NOX shutdown events. We are also at Unit 1 as required under the BART emission rates.196 The commenter cites structuring the compliance Guidelines. Second, the NOX emissions several examples of other coal-fired determination method so that, when one exiting Coronado Unit 1’s boiler could boilers using PRB coal achieving boiler- of the two units is not operating, its be reduced significantly from the out NOX emission rates in the range of emissions from the preceding thirty current rate of approximately 0.3 lb/ 0.096 to 0.154 lb/MMBtu, and points to boiler-operating-days will continue to MMBtu to a rate of 0.15 to 0.20 lb/ these examples as evidence that the be included in the two-unit average. We MMBtu, which would result in a lower Apache and Coronado units could attain expect that SRP can meet this limit by achievable emission rate. Neither ADEQ lower emission rates through the use of installing a low load temperature nor EPA analyzed the various methods combustion controls. We note that the control system on Unit 2 and an SCR of reducing these NOX emissions. best performing units on this list are system including a low load The commenter (Earthjustice) noted primarily tangential- or wall-fired units, temperature control system on Unit 1. that SRP submitted comments to EPA and that none of the units appear to be We are setting a compliance deadline shortly before EPA issued the proposed Riley turbo-fired boilers. Particularly in for achieving this limit of five years rule arguing that SCR with a 0.05 lb/ the case of the Apache and Coronado from publication of this final rule, MMBtu NOX emission limit is units, which are turbo-fired boilers, we which will ensure that SRP has unachievable at Unit 1 (and Unit 2).195 consider this distinction crucial when adequate time to design and install According to the commenter, SRP determining the appropriate units with these controls without interfering with argued that EPA’s proposal is not which to compare emission the consent decree deadline of June 1, achievable by pointing to BART performance. The Riley-turbo boiler is a 2014 for operation of SCR on Unit 2. proposals in other states that required unique wall-fired boiler design that is Finally, we are including in the SCR with an emission limit less characterized by a venturi-shaped lower regulatory text of the FIP a requirement stringent than 0.05 lb/MMBtu. The section (often described as a ‘‘pinch’’ in that pollution control equipment be commenter countered that these BART the boiler wall) with burners located on designed and capable of operating determinations for other sources in the underside of the pinched wall, tilted properly to minimize emissions during other states do not show that EPA’s slightly downwards.197 It is a relatively all expected operating conditions, BART proposal is unachievable at uncommon design, with only two dozen consistent with the regulatory definition Coronado Unit 1, as BART such units currently in operation.198 of BART as ‘‘an emission limitation determinations are source-specific. The based on the degree of reduction commenter added that SRP’s comments 196 As described by the commenter, the ‘‘boiler- achievable through the application of out’’ NOX emission rate refers to the emission rate the best system of continuous emission ‘‘a requirement established by the State or the after including the effects of combustion controls reduction for each pollutant which is Administrator which limits the quantity, rate, or such as low NOX burners, over-fire air, neural concentration of emissions of air pollutants on a networks, adaptive controls, etc. emitted by an existing stationary continuous basis, including any requirement 197 See ‘‘Design and Operation of Coal-fired 194 facility.’’ relating to the operation or maintenance of a source TURBO furnaces for NOX control’’, Riley Stoker to assure continuous emission reduction * * *’’). Corporation, November 1978. 194 40 CFR 51.301. See also, CAA section 302(k), 195 Citing Docket Item C–16 (SRP Letter to 198 Acid Rain Program data indicates 22 turbo 42 U.S.C. 7602 (defining ‘‘emission limitation’’ as DJordan 06–26–2012). units were in operation in 2011.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00046 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72557

The turbo boiler was developed in the emission rate at full load, steady state emission limit (or a 0.04 lb/MMBtu 1960s and, unlike many other wall-fired conditions, and that SRP’s analysis limit). The commenter noted that EPA boilers, was generally able to meet the indicates Coronado Unit 1 could requested information concerning NOX emission limits contained in the achieve 0.050 lb/MMBtu on an annual whether the amount and management of 1971 New Source Performance average basis, we do not consider this catalyst could be altered to meet a 0.05 Standards for fossil fuel fired steam emission rate achievable as a rolling 30- lb/MMBtu NOX limit at Unit 2, but generators.199 While Babcock Power, day limit based on the number of according to the commenter SRP did not which acquired Riley Stoker, has startup and shutdown events associated provide any such information. As a developed new burner upgrades to meet with its operating profile. result, the commenter urged EPA to more stringent NOX emission standards, Comment: While supporting EPA’s revise its BART determination to require the combustion control designs determination that SCR is BART for SCR with an emission limit lower than available for turbo-fired boilers have not Coronado Unit 2, one commenter 0.08 lb/MMBtu. been through the same number of design (Earthjustice) stated that lower NOX iterations, and are therefore not as emission limits are cost-effective and Response: We disagree with the effective as those for other boiler achievable. For Unit 2, the commenter commenter’s assertion that it is types.200 We therefore do not consider made four major points. First, the NSR appropriate to consider lower ‘‘boiler- it appropriate to compare the ‘‘boiler- consent decree does not exempt out’’ NOX emissions for Coronado Unit out’’ emission rates of the Riley turbo Coronado Unit 2 from a NOX BART 2, for the same reasons we noted in the design with those achieved by determination based on a valid five- previous response for Coronado Unit 1 tangential and more traditional wall- factor BART analysis. Second, contrary on this issue. We also disagree with the fired units. to the argument that the 0.08 lb/MMBtu commenter’s assertion that 0.04 lb/ More specifically, combustion limit on Coronado Unit 2 under the MMBtu is an appropriate SCR emission controls on Coronado 1 (LNB) were consent decree was developed to rate to consider for Coronado Unit 2, installed in 2009, and the commenter address BACT obligations, that emission also for the same reasons we noted in has not indicated any design limit is not BACT, which requires a top- the previous response for Coronado Unit improvements or upgrades that would down analysis that selects the 1 on this issue. achieve improved performance. We note ‘‘maximum degree of reduction.’’ There We agree with the commenter’s that the baseline period for our analysis is no BACT analysis in the consent assertions that the consent decree is not represented the use of combustion decree and no explanation of how the a replacement for a five-factor BART controls (in the form of LNB with OFA) 0.08 lb/MMBtu emission limit was analysis. We also agree that while the and that our emission estimate of LNB selected. In addition, while BACT is based on past actual emission data, as requires case-by-case analysis, the consent decree resolved NSR/PSD reported to CAMD, over the baseline consent decree limit was not specific to obligations such as BACT, a ‘‘top-down’’ period. As part of the supplemental cost Unit 2; it simply required installation of BACT analysis was not performed as analysis we performed, we used a SCR on one of the two units. Third, the part of the consent decree negotiations. baseline period that predated negotiated limit contained in the NSR Based on our review of SRP’s August 24, installation of LNB, and consisted of consent decree cannot replace the 2012 letter and submitted comments, we emission rates corresponding to OFA required five-factor BART analysis for do not consider the SCR system for only.201 Comparing annual average Coronado Unit 2 because BART is more Coronado Unit 2, as currently designed, emission rates during the periods prior stringent than the consent decree’s to constitute BART. As noted in the to and following LNB installation, we emission limit. The Sahu report shows analysis contained in our response to note that Coronado Unit 1 has achieved that an emissions limit lower than 0.08 SRP’s comments, we consider the approximately 25 percent reduction lb/MMBtu is cost-effective and installation of a low-load temperature from installing LNB at an emission rate achievable at Unit 2. Fourth, the NOX controller to be both cost-effective and of approximately 0.30 lb/MMBtu. We emissions exiting Coronado Unit 2’s to result in substantial visibility consider these values reasonable, as it is boiler could be reduced significantly improvement. We are not, however, supported by actual emission data and from the current rate of approximately finalizing a more stringent emission represents a control efficiency similar to 0.33 lb/MMBtu to a rate of 0.15 to 0.20 limit for Coronado Unit 2. Instead, we the 30 percent control efficiency lb/MMBtu, which would result in a are finalizing a requirement that assumed by our contractor. lower achievable emission rate. Neither pollution control equipment be In addition, we disagree with the ADEQ nor EPA analyzed the various designed and capable of operating commenter’s assertion that 0.04 lb/ methods of reducing these NOX properly to minimize emissions during MMBtu is an appropriate SCR emission emissions. SCR with a 0.04 lb/MMBtu all expected operating conditions, limit to consider for Coronado Unit 1. emission limit at Coronado Unit 2 is consistent with the regulatory definition As discussed in the previous response achievable with various control of BART as ‘‘an emission limitation to SRP’s comments, we have examined methods and is even more cost-effective based on the degree of reduction the analysis performed by SRP and than EPA‘s calculations suggest. achievable through the application of determined that a 0.050 lb/MMBtu Because of this, the commenter the best system of continuous emission emission rate is not achievable by requested that EPA revise its BART reduction for each pollutant which is Coronado Unit 1 on a rolling 30-day determination to reflect this lower level. emitted by an existing stationary average. Although we note that SRP’s The commenter (Earthjustice) stated facility.’’ 202 analysis is based on a 0.04 lb/MMBtu that SRP has claimed that a NOX emission limit of 0.05 lb/MMBtu is 202 40 CFR 51.301. See also, CAA section 302(k), 199 ‘‘An Overview of Riley Stoker’s Burner unachievable based on its progress in 42 U.S.C. 7602 (defining ‘‘emission limitation’’ as Development Efforts for NOX Control’’, Riley Stoker constructing the SCR unit required by ‘‘a requirement established by the State or the Corporation, April 7, 1983. Administrator which limits the quantity, rate, or 200 the NSR consent decree, but does not ‘‘Low NOX Combustion System Solutions for concentration of emissions of air pollutants on a Wall Fired,T-Fired, and Turbo Fired Boilers.’’ explain how construction progress to continuous basis, including any requirement Babcock Power, August 28–31, 2006. date would prevent it from calibrating relating to the operation or maintenance of a source 201 Supplemental Cost Analysis 2012–11–15.xls. the SCR to achieve a 0.05 lb/MMBtu to assure continuous emission reduction * * *’’).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00047 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72558 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

c. Costs of Compliance costs. The commenter asserted that EPA on actual costs experienced for Unit 2 Comment: One commenter (NPS) improperly ignored site-specific cost for which SCR has been designed to agreed with EPA that SRP did not estimates for Coronado BART control achieve an emission limit of 0.080 lb/ provide ADEQ with control cost options by substituting its own MMBtu, rather than the 0.050 lb/MMBtu calculations at a level of detail that estimates, and ignored the fact that assumed by EPA for Unit 1. According allowed for a comprehensive review. Arizona has ‘‘the lead role in designing to the commenter, there could be The commenter conducted analysis of and implementing [its] regional haze additional costs for Unit 1 of as much the cost and cost-effectiveness of adding program’’ and ‘‘broad authority over as $117 million for additional catalyst BART determinations’’ (citing Corn and an increased ammonia emission SCR to reduce emissions of NOX at Coronado Unit 1 using the cost Growers, 291 F.3d at 3, 8). The rate, a dry sorbent injection control commenter stated that ADEQ fully system to address increased sulfuric methodologies of the CCM and relying complied with the BART Guidelines acid mist and condensable PM on the IPM to reflect the most recent and was justified in any deviation from emissions, and a fabric filter baghouse SCR cost levels, and submitted the the specific terms of the CCM because and induced draft fans to address detailed calculations as Appendix E to ADEQ engaged in a reasoned, site- increased filterable PM emissions. The its comments. The commenter’s analysis specific cost analysis. The commenter commenter stated that even without yielded a cost-effectiveness value of added that ADEQ has discretion to these additional costs, the site-specific $2,540/ton. The commenter noted that conduct and document its cost cost estimate for an SCR system on Unit EPA’s analysis yielded a cost- assessment at a level that it deems 1 is almost twice the value used by EPA effectiveness value of $2,405/ton, which appropriate, and that the documentation in its BART determination, and for the EPA considers cost-effective. Another that supports ADEQ’s BART SCR system on Unit 2, the actual cost commenter (Earthjustice) also asserted determination is reasonable by any incurred by SRP is likewise almost that SCR at Coronado 1 is cost-effective. objective standard. twice the value used by EPA in its When calculated based on an SCR The commenter (SRP) asserted that BART determination. The commenter emission rate of 0.04 lb/MMBtu, and EPA improperly ignored site-specific concluded that this documentation when accurate cost figures and proper cost estimates for Coronado BART demonstrates the importance of using baselines are used, the commenter control options, instead using the IPM available site-specific cost estimates asserts that SCR would reduce NOX to calculate the capital costs and annual when conducting a BART determination emissions at a cost of just $2,024/ton of operating costs associated with the for Coronado. NO removed. X various NOX control options that EPA Response: We disagree with the NPS commented that it was not able considered. Moreover, the commenter commenter’s assertion that the cost to conduct a cost analysis for Coronado added that no cost estimate derived calculations SRP provided to ADEQ as Unit 2, on which SRP is installing SCR from a model designed to produce part of the original BART analysis, or in to meet an emission limit of 0.080 lb/ generalized information about utilities the subsequent response to ADEQ’s MMBtu under a consent decree with throughout the nation could satisfy the information request, were supported by EPA. However, the commenter used the CAA requirement that BART be sufficient documentation. For example, CCM to evaluate the differences determined based on a site-specific the annual O&M costs associated with between an SCR on this unit at 0.050 lb/ analysis. SRP provided adjusted inputs an SCR system will involve such costs MMBtu versus 0.080 lb/MMBtu. for use in IPM for unit size, gross heat as reagent usage, catalyst replacement According to the commenter, an SCR rate, NOX removal factor, NOX removal costs, and labor costs, among others. meeting the more stringent limit would efficiency, ammonia cost, operating SRP provided no breakdown of annual have essentially the same footprint as labor rate, bare module costs, urea costs O&M costs beyond the total annual the less effective unit, but would require and property taxes and insurance. SRP O&M value. Similarly, SRP’s capital cost an additional layer of catalyst and asserted that when these values are used estimates consist of only a total value, would be seven feet taller. The in the model, the IPM outputs validate accompanied by a capital recovery commenter presented basic design the site-specific costs provided by SRP factor to determine the corresponding parameters for SCR units achieving the (based on detailed SCR and SNCR cost annualized cost. This level of detail two levels of control. comparisons provided in the does not allow us, and could not have Response: We agree with NPS’s comments), although the adjusted IPM allowed ADEQ, to evaluate the assertion that SRP’s cost figures, as results still under-predict the costs reasonableness of SRP’s cost estimates provided in their original BART based on site-specific considerations. for Coronado. As noted in a previous analysis and in the subsequent response The commenter (SRP) stated that its response, we have identified several to ADEQ’s information request, were not site-specific costs for SCR are based on issues with the cost calculations sufficiently documented. While we also the actual cost projections associated performed for the Apache and Cholla agree with the commenters’ assertion with the current SCR installation at Unit units that are inconsistent with the that SCR with LNB and OFA is cost- 2. The commenter stated that SRP has methodology established by EPA’s CCM. effective, we decline to modify our already made substantial progress on SRP’s cost estimates do not provide estimates of cost- effectiveness to reflect the Unit 2 SCR installation with more sufficient detail for us to evaluate if they the cost items noted in these comments, than 40 percent of the project already are consistent with CCM methodology. as it is not in any way determinative to complete, with the engineering design Although we do not agree that our our decision to find that SCR is ‘‘even effort more than 90 percent complete, cost-effectiveness estimates were more’’ cost-effective, or that the and the overall procurement efforts incorrect, we have performed a incremental cost-effectiveness value more than 75 percent complete. As supplemental analysis for Coronado 1 between SCR and SNCR is ‘‘even more’’ such, the commenter believes that the using portions of the updated cost incrementally cost-effective. site-specific costs are appropriate for estimates provided by SRP in its Comment: One commenter (SRP) use in any evaluation of BART controls. comments. Our use of these cost argued that EPA’s cost of compliance In addition, the commenter (SRP) estimates in this supplemental analysis analysis for Coronado is flawed and indicated that its cost estimates for Unit should not be construed to represent an must be replaced with site-specific 1 are conservative since they are based acceptance of SRP’s revision to our IPM

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00048 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72559

assumptions. Rather, this supplemental of its comments. This estimate has did not properly consider LNB and OFA analysis represents a conservative excluded cost items not allowed by the as a potential control option and estimate of costs (i.e., an assumption CCM, such as AFUDC, escalation, and therefore precluded a BART that would tend to overestimate rather owner’s costs, and have been included determination of LNB and OFA, we than underestimate the annualized cost in the supplemental analysis. have used a baseline period of 2001– of controls). A summary of cost • Original baseline period: As 2003, which corresponds to the period estimates based on this supplemental discussed elsewhere in our responses, used in SRP’s original BART analysis. analysis is displayed in Table 15. we consider our use of a more recent This represents a time period prior to • SRP’s revised SNCR cost estimates: baseline as consistent with BART the installation of LNB, during which SRP also submitted a conceptual capital Guidelines. However, in order to the control technology in place on cost estimate for an SNCR system as part address commenter’s concerns that we Coronado 1 was OFA-only.

TABLE 15—CORONADO UNIT 1: CONTROL COST ESTIMATES [Per SRP with EPA revisions]

Annualized Annual O&M Total annual Coronado 1 Capital cost capital cost cost cost control technology ($) ($/yr) ($/yr) ($/yr)

LNB+OFA ...... $6,500,000 $613,554 $0 $613,554 SNCR w/LNB+OFA ...... 14,164,000 1,336,981 5,829,800 7,166,781 SCR w/LNB+OFA ...... 80,633,219 7,611,205 4,492,736 12,103,941

Regarding SRP’s concern that its own average emission limit of 0.050 lb/ considered part of the cost of costs for Coronado Unit 1 are MMBtu will involve greater costs than compliance for the purposes of a BART conservative (i.e., underestimated in a system designed to meet 0.080 lb/ determination. EPA cannot anticipate this context) because they are based on MMBtu, we disagree that the costs for what control technology might be a Coronado Unit 2 design that achieves Coronado Unit 1 are of the magnitude of required in the future for sulfuric acid 0.080 lb/MMBtu instead of 0.050 lb/ those described above for Coronado mist under PSD or minor NSR. The MMBtu, we partially agree. For Unit 2. Based on SRP’s comments, we BART Guidelines do not require the Coronado Unit 2, SRP identified certain note that the SCR reactor box for Unit inclusion of potential future costs that additional costs that would be 2 has been designed for a ‘‘3+1’’ might be associated with permit associated with design changes configuration (i.e., an initial three requirements as part of the cost necessary to meet an emission rate more catalyst layers, with space for a fourth estimates for a BART determination. stringent than the consent decree limit layer to be added later in the system’s of 0.080 lb/MMBtu. The two most Therefore, while we acknowledge that lifetime to maintain the same level of there are costs associated with important changes would be increased effectiveness) and has perhaps already levels of ammonia injection and additional catalyst and increased been fabricated. As a result, ammonia injection, they represent a additional SCR catalyst (in the form of accommodating additional catalyst an additional fourth catalyst layer at the small fraction ($4 million) of the $117 cannot be achieved by increasing the time of initial catalyst fill). The SCR million total identified by SRP. We have volume of the initial three layers, but catalyst is responsible for a certain used certain elements from SRP’s must be achieved by including the amount of SO2 to SO3 conversion, estimates in preparing our supplemental fourth catalyst layer (or some portion of which can then form sulfuric acid cost analysis for Unit 1, but we have not it) during the initial fill. Since each (H2SO4). SRP notes that the additional adjusted SRP’s estimates to reflect these catalyst layer is designed for a certain fourth catalyst layer can be expected to factors since the cost estimates provided amount of SO to SO conversion, result in a collateral increase in sulfuric 2 3 by SRP do not include a level of detail inclusion of an additional layer acid (H2SO4) emissions. A dry sorbent that would allow us to properly make unavoidably results in an increase in the injection (DSI) system may be needed to such adjustments. address this increase in sulfuric acid, overall conversion rate. However, since which itself has the potential to increase an SCR system for Coronado Unit 1 has A summary of emission rates and filterable particulate emissions. not been designed, we consider it emission reductions associated with Addressing this increase in filterable feasible for SRP to specify a design at each control option is in Table 16. As particulate emissions may in turn the outset that accommodates additional noted previously, these emission require installation of a fabric filter volume in the initial catalyst layers, estimates are based on a 2001–2003 baghouse. Of the $117 million in capital thereby achieving a more stringent baseline period, during which the costs identified by SRP, the majority of emission rate without the higher SO2 to Coronado units operated only with these costs ($113 million) are associated SO3 conversion rate associated with a OFA. We have calculated annual with construction of the DSI and fabric fourth catalyst layer. Moreover, even if emission rates for the OFA baseline filter. SRP were required to install a DSI using the annual average emission data While we agree that designing system or DSI and a fabric filter, EPA (lb/MMBtu) reported to CAMD over this Coronado Unit 1 to meet an annual does not agree that these costs should be 2001–2003 baseline period.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00049 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72560 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

TABLE 16—CORONADO 1: ANNUAL EMISSION ESTIMATES

Emission Heat rate Annual Emission rate Emissions Coronado 1 control technology factor (MMBtu/hr) capacity removed (lb/MMBtu) 1 factor (lb/hr) (tpy) (tpy)

OFA (only)...... 0.407 4,316 0.84 1,756 6,462 ...... LNB+OFA ...... 0.303 4,316 0.84 1,308 4,811 1,651 SNCR w/LNB+OFA...... 0.212 4,316 0.84 915 3,368 3,095 SCR w/LNB+OFA ...... 0.050 4,316 0.84 216 794 5,669 1 Annual average basis.

Cost-effectiveness values for each annual costs and annual emissions control technology are summarized in removed listed in the previous tables. Table 17, and are based on the total

TABLE 17—CORONADO 1: CONTROL OPTION COST-EFFECTIVENESS

Total annual Emissions Cost-effectiveness Coronado 1 control technology cost removed ($/ton) ($/yr) (tpy) Avg. Incr.

OFA (only) ...... LNB+OFA ...... $613,554 1,651 $372 SNCR w/LNB+OFA...... 7,166,781 3,095 2,316 4,540 SCR w/LNB+OFA ...... 12,103,941 5,669 2,135 1,918

Based on SRP’s capital and O&M cost commenter added that this failing proper NOX BART assessment, the estimates, we still consider the cost- renders EPA’s proposed rule proposed rule lacks an adequate effectiveness values of SCR, on an inconsistent with the CAA’s public foundation. The commenter stated that average ($2,135/ton) and incremental notice requirements. the high incremental costs of post- ($1,918/ton) basis, to not be cost- Response: We disagree with the combustion NOX control technologies prohibitive. We consider these results commenter’s assertion that we have not when compared to combustion control supportive of our proposed provided sufficient information technologies reinforces the conclusion determination that SCR with LNB and regarding our cost calculations. In the e- that post-combustion control OFA is cost-effective. We note that docket for our proposal, we included technologies cannot be the basis for while the LNB and OFA option is the the raw cost calculation spreadsheets BART for the units at Coronado. least expensive (i.e., lowest annual cost) from our contractor that contain the IPM Response: We disagree with the and is the most cost-effective of the equations, corresponding variable commenter’s assertion that we did not control technologies (i.e., has the lowest values, selected notes regarding perform a sufficient incremental cost $/ton value), it is also the least effective assumptions and variable ranges, as analysis for the Coronado units. In our control option (i.e., removes smallest well as selected tables from the IPM control cost summaries (Table 22 in the Base Case v4.10.203 quantity of NOX). It removes In addition, Web proposed rule and Table 32 in the TSD), substantially fewer emissions than links were provided (both in the raw the column labeled ‘‘incremental cost- either of the other two control options, cost calculation spreadsheet and in our effectiveness’’ represents the $/ton of the SNCR- and SCR-based systems. As proposal) to the location on the publicly the control option when compared to discussed in our proposed action, and available EPA Web site that contains the preceding control option. The in other responses in this document, we full IPM documentation. column labeled ‘‘average cost- Comment: One commenter (SRP) have not identified any energy or non- effectiveness’’, represents the $/ton of stated that EPA failed to follow the air quality impacts that warrant the control option when compared to BART Guidelines by not conducting an eliminating SCR from consideration for the baseline control. In the case of incremental cost analysis for Coronado. Coronado Unit 1. Combined with the Coronado Unit 1, we considered two According to the commenter, the modeled visibility improvement control options beyond the baseline: proposed rule and TSD both provide a associated with the SCR control option, SNCR with LNB and OFA, and SCR single entry for incremental costs for SRP’s cost estimates continue to support with LNB and OFA. The ‘‘single entry each of the Coronado units that reflect the selection of SCR with LNB and OFA for incremental costs’’, as described in the incremental cost of the most as BART for Coronado 1. the comment, represents the stringent NO BART control option Comment: One commenter (SRP) X incremental cost between the SNCR- compared to the baseline. The stated that the proposed rule and the and SCR-based options. An incremental commenter asserted that this is not a TSD say almost nothing about how IPM cost value was not calculated between complete incremental analysis because was used to calculate costs, instead LNB with OFA (which is the option it ignores incremental comparisons directing the public to an EPA preceding the SNCR-based option) and between identified control options. SRP contractor report for more information. SNCR because LNB with OFA contended that in the absence of a The commenter asserted that no represented the baseline control in our contractor report in the docket for the 203 Document ID: EPA–R09–OAR–2012–0021– analysis. The cost-effectiveness of rulemaking supplies additional detail 0008, File name: G–15_MODELING_FILES_EGU moving from LNB with OFA to SNCR on precisely how IPM was used. The _BART_Costs_Apache_Cholla_Coronado_FINAL2. with LNB and OFA is therefore

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00050 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72561

adequately captured in the ‘‘average proposed FIP, one commenter (SRP) requirements to eliminate unnecessary cost-effectiveness’’ column. We do note performed and submitted an assessment and unreasonable compliance burdens. that, in our supplemental cost analysis, of the critical components of a BART • The commenter supports and we have used OFA as the baseline analysis for Coronado, including control appreciates the use of the monitoring control, and have therefore calculated costs and the visibility improvements system certification and quality an incremental cost-effectiveness value associated with the control options. The assurance (QA) procedures in 40 CFR for moving from LNB with OFA to commenter indicated that this analysis Part 75. However, EPA’s proposed SNCR with LNB and OFA. These results shows that even without considering definition of ‘‘valid’’ data is broader are described in a previous comment other energy and non-air quality than 40 CFR Part 75, and EPA also and, as noted in that comment, we environmental impacts associated with should make clear that the ‘‘bias’’ disagree with the commenter’s assertion the implementation of SNCR or SCR, it adjustment procedures in 40 CFR Part that the incremental cost of post- is clear that the visibility benefits 75 do not apply to data used to calculate combustion controls is cost-prohibitive. realized from implementation of post- the 30-day rolling averages. combustion controls are not justified by • The commenter objects to the d. Visibility Improvement the cost. The commenter also submitted proposed additional relative accuracy Comment: One commenter (SRP) the results of this analysis using requirements. Imposing additional asserted that EPA is without basis for CALPUFF version 6.42 in place of relative accuracy test audit (RATA) establishing in the proposed rule a 0.5 version 5.8. The commenter stated that specifications will not increase the deciview comparison threshold as a this analysis provides even stronger accuracy of any monitoring system, but touchstone for analyzing impacts from evidence that selection of post- would increase the difficulty and cost of Coronado BART controls, citing the combustion controls as BART for testing. It also could result in additional BART Guidelines and associated Coronado is inappropriate. missing data if tests must be repeated to preamble. According to the commenter, Response: We disagree with this meet the specifications. To proceed with even if EPA could impose a 0.5 comment. As noted in a separate combined RATA specifications, EPA deciview comparison threshold, it is response, we have performed a also would need to either propose (and only by substituting its own preferred supplemental cost analysis that relies solicit comment on) alternative low- modeling methodology (which the upon many elements of the cost analysis emitter combined specifications that commenters argued is something EPA provided by the commenter. Even with have been demonstrated to be cannot lawfully do) that EPA can project the higher cost estimates provided by consistently achievable, or exempt units that requiring SCR at Unit 1 would the commenter, we consider the costs of meeting any of the applicable 40 CFR barely yield a projected improvement of post-combustion controls such as SNCR Part 75 low-emitter thresholds from more than 0.5 deciview at one area. The and SCR to be cost-effective on a $/ton those specifications. commenter also noted that 0.5 deciview basis. In addition, as noted in a separate • The commenter stated that the is below the level of human response, we disagree with several proposed data availability requirements perceptibility. assumptions used in the commenter’s are unnecessary and too stringent. Response: As explained above, we visibility modeling, such as the use of Source owners and operators already have not used 0.5 dv as a threshold, but an unapproved CALPUFF model have sufficient incentive to obtain valid as one point of comparison such as a version and treatment of ammonia data in order to avoid the increasingly ‘‘benchmark’’ or ‘‘yardstick’’ to gauge background concentrations. We conservative (and ultimately punitive) the magnitude of impacts under various therefore disagree that the visibility missing data substitution procedures control scenarios. benefits modeled by the commenter are that apply under 40 CFR Part 75. representative of the benefits that will Regarding stringency, if a unit has a e. Other Comments accrue with the use of post-combustion significant missing data event during a Comment: The commenter (NPS) controls. The modeling results calendar quarter in which it also has a agreed with EPA’s determination that performed in support of our proposal significant period of unit downtime NOX BART for Coronado Units 1 and 2 indicate substantial visibility benefits, (e.g., as a result of an outage), the is SCR with LNB and OFA. The especially with the SCR control option. percent of operating hours during the commenter noted that EPA proposed on As a result, we do not consider it quarter with valid data could easily be Unit 1 an emission limit for NOX of appropriate to eliminate either of the less than 90 percent. It is in part for this 0.050 lb/MMBtu, based on a rolling 30- post-combustion controls from reason that 40 CFR Part 75 measures boiler-operating-day average, and on consideration as BART. Although SCR data availability over each 8,760- Unit 2 an emission limit of 0.080 lb/ is the most stringent control option, its operating-hour period. EPA should MMBtu, which is consistent with the associated visibility benefits and cost- either eliminate the unnecessary emission limit in the consent decree. effectiveness justify this technology as requirement or provide data to justify its The commenter stated that EPA BART. proposed requirement that take into acknowledged that the emission limit account the differences described above. for Unit 2 established in the consent E. Comments on Enforceability • EPA must modify the quarterly decree was not the result of a BART Requirements in EPA’s BART FIP reporting requirements to be consistent five-factor analysis, and that the consent Comment: One commenter (SRP) with existing requirements. decree does not indicate that SCR at made the following points concerning • EPA must modify the notification 0.080 lb/MMBtu represents BART. The the proposed enforceability requirements in the proposed rule commenter commended EPA for requirements: because they are overly broad and soliciting additional information on the • EPA must modify the monitoring overly prescriptive. First, EPA should feasibility of achieving a more stringent requirements to be consistent with clarify the proposed provision by limit. existing requirements. If EPA proceeds requiring notice only of new controls Response: We acknowledge the to impose additional controls at that will be required to meet the FIP or comment. Coronado beyond those specified in the regional haze SIP. Second, because Comment: In response to EPA’s consent decree and already included in installation of controls is a complex proposed BART determination in the the Coronado permit, it must align these process, and the point at which that

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00051 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72562 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

process is ‘‘complete’’ may not be of the CEMS required by Part 75 and Method 5 or 5e are appropriate where immediately clear, EPA must revise the used for the compliance demonstrations SCR is used. ADEQ suggested Method 5 requirement to use a more objective for this action obtain 90 percent valid or 5e where SCR is used, and states that term and allow sufficient time for data (per Part 75 specifications) for each any collateral increase in acid mist owners and operators to comply. Third, unit over each calendar year. In should be addressed through a because the proposed requirement addition, the rule will require the permitting process. SRP stated that wet duplicates reporting already required for affected units to conduct RATA scrubbers also render Methods 201 and a new add-on NOX emission control evaluations and calculate the quarterly 201A inapplicable, and requested that under 40 CFR Part 75, EPA should rely valid data hours for NOX lb/hour and EPA specify the use of Method 5, 5B, 5I on (and if necessary refer to) the notice heat input. EPA will not finalize the or an approved alternative. required under Part 75. minimum data requirements in the One commenter (NPS) pointed out Response: We partially agree with this proposal, but will require these data to that use of SCR at these units is comment and are adjusting the be calculated (all data for determining expected to result in increased enforceability requirements of the final the relative accuracy in these units are condensable particulate matter in the FIP accordingly. EPA agrees that the available when Part 75 RATAs are form of sulfuric acid mist (H2SO4), Part 75 bias adjustment should not be performed) and reported to both EPA which would have the effect of making applied to the compliance data for the and ADEQ to determine if these data are the emission limit more stringent than BART rules in this action and is making capable of meeting more rigorous QA/ intended by ADEQ, and likely not be changes to the final rule to address this QC requirements in the future. We also achievable in practice. To address EPA’s comment. However, EPA does not agree note that the final rule will add QA/QC request for comment on whether to that only the incentives under the Acid and minimum valid data requirements allow compliance with the PM10 limit to Rain Part 75 rules are sufficient to for the inlet SO2 CEMS that are needed be demonstrated using test methods that assure adequate valid data for this rule. to calculate the SO2 removal efficiencies do not capture condensable particulate Part 75 relies on progressively punitive for the Cholla EGUs. Finally, EPA agrees matter, namely EPA Methods 1 through data substitution procedures to promote that semiannual reporting will be 4 and Method 5 or Method 5e, the good valid data availability for its sufficient for this rule, and the final rule commenter conducted and submitted an program. Our rule does not substitute will reflect this. analysis of H2SO4 emissions. According data, so the incentives of the Part 75 Comment: One commenter (AEPCO) to the commenter, H SO emissions will rules do not exist. Therefore, EPA is requested that EPA clarify that AEPCO 2 4 not be significant, contributing less than requiring that each unit subject to this has longer than 180 days to comply with 10 percent to the PM limit. The rule obtain 90 percent valid data, as the non-SCR limits. The commenter is 10 commenter suggested that the 0.030 lb/ determined under Part 75, for each particularly concerned about the time MMBtu limit proposed by ADEQ for the calendar year. needed for the ESP and scrubber Apache and Coronado units be adjusted It should be noted that the commenter upgrades and believes a five-year period to 0.033 lb/MMBtu to reflect the did not submit any data specific to its for all BART implementation would be increase in total PM attributable to EGUs indicating the difficulty of appropriate. ADEQ also commented that 10 SCR, and that PM emissions would be meeting the proposed valid data the facility will need more than 180 10 measured by conducting EPA Method availability requirements. Also, the days to complete the upgrades needed 201A/202 tests consistent with the other two utility companies affected by to meet the SO2 BART limits, and stated this rule did not make any objection to that a five-year compliance time frame ADEQ’s SIP. the proposed data requirements. from the time the BART limit is In contrast to the previous However, EPA, as a result of this finalized, as specified in RHR Appendix commenters, one commenter comment, has reconsidered the Y, is most appropriate. (Earthjustice) stated that EPA should additional quality assurance and valid Response: EPA agrees that AEPCO approve the test methods in the ADEQ data requirements from the proposal. As would need more than the 180 days in RH SIP (i.e., EPA Methods 201 and 202) indicated by the commenter, the proposed rule. However, we do not and ensure that the BART limit includes measurement and QA requirements for agree that five years is necessary to both filterable and condensable PM NOX lb/hour are not currently required perform the necessary upgrades. The fractions. The commenter asserted that by Part 75. In addition, EPA recognizes final rule will require AEPCO’s two if EPA allows or requires a test method that the calculation of heat input units to meet the SO2 and PM10 limits other than Method 201 and 202, the requires the combination of the flow within four years of the effective date of PM10 BART emission limit would and diluent (O2 or CO2) CEMS along this rule. This time frame will allow effectively be less stringent because it with measurements of temperature and AEPCO to perform the upgrades to the would only apply to filterable PM, and estimation of moisture. In addition in two units during regularly scheduled not total PM. The commenter indicated the final rule, EPA is providing for a maintenance outages. that requiring different test methods multi-unit determination of compliance. Comment: Several commenters would in effect be proposing an even This would compound the valid data (ADEQ, AEPCO, APS, EarthJustice, NPS, less-stringent PM10 BART limit, which concerns of the commenter. EPA SRP) provided feedback on test would require EPA to undertake an requires monitoring data used for methods. AEPCO supported independent BART analysis that compliance determinations to be of maintaining the use of EPA Method demonstrates that the less-stringent known quality as demonstrated through 201A to comply with the proposed emission limit is BART. Consequently, Quality Assurance/Quality Control (QA/ BART PM10 limits. In contrast, ADEQ according to the commenter, if EPA QC) procedures.204 In place of the and APS only supported the use of requires or allows a different test requirement to validate through RATA Methods 201A and 202 if SCR controls method, it must lower the emission testing of the NOX lb/hour measurement are not used. These commenters stated limit to reflect only the filterable PM10 and heat input, EPA will require that all that SCR causes an increase in sulfuric fraction. The commenter added that in acid aerosol mist, which results in an this case, EPA should ensure that 204 See, e.g., 40 CFR 60.13(a) and 40 CFR increase in condensable particulate compliance with the filterable PM10 Appendix F. matter. APS suggested Methods 1–4 and limit is demonstrated with use of CEMS

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00052 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72563

for filterable PM, which is currently EPA is finalizing a decision to allow to do the necessary upgrades for Unit 2. available. either Method 5 or Methods 201A and This unit will require scrubber upgrades Response: ADEQ selected test 202 for demonstrating compliance with that need to be done concurrent with methods 201 and 202 for determining the BART PM10 limits set by ADEQ. the fabric filter installation to compliance with this limit. EPA noted As noted above, the addition of the accommodate the increase in pressure in the proposal that the proposed SCR to these EGUs for NOX control will drop that a new fabric filter will impose. addition of SCR for NOX control would likely increase the condensable PM that ADEQ also stated a compliance date of likely increase the quantity of PM will be measured by Method 202. By April 1, 2016, would be more collected as condensable PM by method offering the option of Method 5 or appropriate than January 1, 2015, for 202 due to an increase in H2SO4 from Methods 201A and 202, the facilities both the PM10 and SO2 limits at Cholla the oxidation of SO2 to SO3. EPA can determine which methods are Unit 2. requested comment on changing the test compatible with their units’ stack Response: EPA agrees with this method from methods 201 and 202 to conditions and will best demonstrate comment and has changed the EPA Method 5 which measures only the the proper operation of their PM compliance date in the final rule to filterable PM. Method 5 measures all controls. Any significant increase in April 1, 2016.209 In addition, as sizes of filterable PM which results in a H2SO4 and the appropriate control of explained above, in order to ensure that higher filterable PM value than Methods this visibility impairing pollutant will the wet FGD (i.e. scrubbers) on all three 201 or Method 201A, which only be addressed through the PSD units at Cholla are properly operated measure filterable PM10. permitting process with a BACT and maintained, consistent with 40 CFR In its comments concerning the determination for H2SO4 control. The 51.308(e)(1)(v), we are finalizing a proposal for Coronado, SRP noted that significance level that triggers removal efficiency requirement for SO2 Method 201A cannot be used in a wet permitting for H2SO4 is an increase of of 95 percent on a 30-day rolling basis exhaust gas stream. We agree with this seven tons per year of this pollutant.208 for Cholla Units 2, 3 and 4. Compliance comment. In promulgating amendments Coronado has already received a PSD with the efficiency requirement will be to Method 201A and Method 202 in permit for H2SO4 that is likely to result determined by SO2 continuous emission 2010, EPA explained that: from the increase in H2SO4 resulting monitoring systems (CEMS) operated at Method 201A cannot be used to measure from the SCR required under the the inlets and outlets of the scrubbers. emissions from stacks that have entrained consent decree. Units 3 and 4 already have SO2 and CO2 moisture droplets (e.g., from a wet scrubber EPA’s AP–42 indicates that CEMS installed after the scrubbers, and approximately one third of the filterable stack) since these stacks may have water Unit 2 has SO2 and CO2 CEMS installed droplets that are larger than the cut size of PM emissions from EGUs are larger than 210 before the scrubbers. Therefore, SO2 the PM10 sizing device. The presence of PM10. This means that the change from and diluent CEMS will need to be moisture would prevent an accurate Method 201 (or 201A) to Method 5 as installed at the inlets to the scrubbers on measurement of total PM10 since any PM10 the compliance method will result in Units 3 and 4. We estimate that the total dissolved in larger water droplets would not this increased measurement of PM. This be collected by the sizing device and would annualized cost for this installation consequently be excluded in determining is offset by the elimination of the (including ongoing operation and total PM10 mass. To measure PM10 in stacks condensable measurement of Method maintenance costs) will be where water droplets are known to exist, 202 and as noted above, the utilities will approximately $51,000 per unit.211 We EPA’s Technical Information Document 09 have the option of using either testing also note that this efficiency (Methods 201 and 201A in Presence of Water approach. requirement will probably result in a Droplets) recommends use of Method 5 of Comment: One commenter (APS) slight increase in operation and appendix A–3 to 40 CFR part 60 (or a requests that EPA change the comparable method) and consideration of the maintenance costs in the form of compliance date for the PM10 limit at 205 total particulate catch as PM10 emissions. additional limestone and scrubber waste Cholla Unit 2 to January 1, 2016, rather disposal expenses. Even considered It is also true that the rarely used than January 1, 2015. The commenter collectively, these additional costs are Method 201 cannot be used in a wet explained that EPA misunderstood the de minimis in comparison to the exhaust stream (also known as a ‘‘wet language of the ADEQ SIP, which refers 206 annualized cost of SCR (i.e., $9,906,206 stack’’). to APS’s commitment to install a fabric to $13,590,853 per unit at Cholla, At this time, the three facilities filter by 2015, to mean installment and according to our supplemental cost subject to this BART rule have a mix of operation by the first of the year, analysis) or the total cost of installing a wet and dry stacks. EPA anticipates that whereas this commitment actually new wet FGD system, which APS has the SO BART limits set by ADEQ will 2 meant by the end of 2015, or December estimated to be $67.0 to $70.9 result in 100 percent of the exhaust gas 31, 2015. The commenter further million.212 In order to allow sufficient undergoing SO2 scrubbing. Neither requested that this date be extended to ADEQ nor EPA is requiring reheat of the April 16, 2016, if the ADEQ approves 209 Although APS requested a deadline of April exhaust gas stream. Therefore, it is APS’s request for a one-year extension 16, 2016, this request was contingent upon ADEQ’s likely that all of the coal-fired units to comply with the Mercury and Air approval of APS’s August 7, 2012 request for a one- covered by this action will have wet Toxics Standards (MATS) before EPA year extension to comply with the MATS. ADEQ’s stacks. So it is doubtful that any finalizes this BART determination. comments indicate that April 1, 2016 is the appropriate deadline for this requirement, so we filterable PM10 method would work as The commenter also requested that have modified the final compliance deadline to 207 the compliance method. Therefore, EPA change the compliance date with April 1, 2016. the 0.15 lb/MMBtu SO2 emissions 210 See Cholla Title V Permit (2012), Table C–3: 205 75 FR 80118, 80121. standard from 180 days after Continuous Emission Monitors. 206 See EPA’s Technical Information Document promulgation to January 1, 2016, or 211 We used EPA’s CEMS Cost Model (available 09, ‘‘Methods 201 and 201A in Presence of Water at http://www.epa.gov/ttn/emc/cem.html) to Droplets’’ (September 9, 1991). April 16, 2016, to allow sufficient time estimate the total annualized cost of adding inlet 207 See, e.g. 75 FR 80126 (‘‘Monitoring the CEMS for SO2 and CO2. See ‘‘CEMS Cost emission of PM10 or PM2.5 from a wet gas stream A consensus method to provide this information Calculation.’’ is a challenging problem that has not been has not emerged.’’) 212 APS Comments, Table 3–8. No annualized addressed successfully despite considerable effort. 208 See 40 CFR 52.21(b)(23)(i). cost was provided.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00053 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72564 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

time for installation of the CEMS, the Comment: One commenter 11-1302, slip op. at 42 (D.C. Cir. Aug. compliance deadline for this removal (Earthjustice) stated that EPA should set 21, 2012) (‘‘CSAPR decision’’); efficiency requirement at these units BART limits for PM2.5 and PM10, rather Luminant Generation Co. v. EPA, 675 will be one year after publication of this than just PM10. The commenter F.3d 917, 921 (5th Cir. 2012); and State final rule for Units 3 and 4. The removal indicated that the BART Guidelines of Texas, et al., v EPA. 690 F.3d 670 (5th efficiency compliance deadline for Unit specify that BART should be evaluated Cir. 2012). 2 will coincide with the compliance and defined for both PM2.5 and PM10 Several commenters stated that EPA date for the lb/MMBtu SO2 emission (citing 40 CFR part 51, Appendix Y, made no finding that Arizona failed to limit for this unit (i.e., April 1, 2016). section II.A.3). satisfy its statutory obligation to Comment: Two commenters requested Response: The BART Guidelines do consider and weigh the BART factors, that EPA implement SCR installation in not require states to set BART limits for and asserted that EPA conceded that the three rather than five years. Earthjustice PM2.5 in addition to limits for PM10. The state had done so in its FIP proposal claimed that the proposed five-year portions of the BART Guidelines cited (citing 77 FR 42851). Some commenters compliance deadline is unreasonable by commenters (i.e. sections II.A.3 and (AEPCO, SRP) stated that EPA proposed and inconsistent with the CAA and RHR III.A.2) pertain to the identification of to disapprove the SIP, in part, because requirements, noting that compliance sources that are BART-eligible and it is not consistent with BART decisions before the ‘‘outside date’’ is required sources that are subject-to-BART, not that other states have made (citing 77 FR whenever earlier compliance is the actual five-factor analysis or 42836), and contended that this finding possible. This commenter contended determination of BART for a given is irrelevant to the approvability of that average SCR installations have source, which is described in section IV ADEQ’s SIP. One commenter (SRP) required 37 to 43 months to implement, of the Guidelines. With respect to the added that ADEQ’s BART and EPA has provided no site-specific five-factor analysis, the Guidelines determinations are entirely legal and factors for these plants to require a provide that, ‘‘[m]odeling should be reasonable and, to the extent that other longer-than-average installation time. conducted for SO2, NOX and direct PM states’ BART determinations may be 213 The commenter notes that ADEQ has an emissions (PM2.5 and/or PM10).’’ The relevant, consistent not only with the ‘‘accelerated permit processing’’ Guidelines thus provide states with the action of other states, but with action program, so that any PSD permits flexibility to consider either PM2.5 or that EPA has approved or proposed to needed to address sulfuric acid mist PM10 emissions or both, as part of their approve for those states (i.e., increases should not require an five-factor analysis. Likewise, the combustion controls as BART for NOX). extension of the compliance deadline to Guidelines do not require that the Two commenters added that EPA five years. The commenter also emission limits reflecting BART should purported to defer to ADEQ’s BART requested that EPA obtain and post to include separate limits for PM2.5 and determinations by indicating that it 214 the docket the outage schedule for these PM10. Thus, we are not required by would prefer to act on a SIP revised to plants, which may provide additional the RHR to set separate BART limits for address the deficiencies perceived by justification for a compliance deadline PM2.5. EPA (citing 77 FR 42839), but the shorter than five years. In contrast, SRP commenters asserted that it is not commented that, if EPA finalizes a F. Comments on Legal Issues deference to invite the State to submit requirement for SCR at Unit 1 ‘‘a Comment: A number of commenters a SIP that conforms to EPA’s policy five-year compliance period is certainly asserted that EPA has acted in a manner choices. The commenters contended warranted.’’ SRP noted that it estimated contrary to the CAA, under which states that in any case, with the court ordered it would require 48 months to install are to play the lead role in designing deadline of November 15, 2012, for EPA SCR at Coronado Unit 2, and that and implementing the regional haze to finalize the proposed FIP, it would be installing SCR on Unit 1 would be even program. These commenters typically impossible for Arizona to prepare and more complicated due to the reduced indicated that EPA is required to defer adopt a revised SIP in time. amount of space following the to the states’ judgment regarding BART Response: We do not agree that our installation on Unit 2. where the state has considered the five partial disapproval of the Arizona Response: We are finalizing a statutory BART factors, and has no Regional Haze SIP is contrary to the compliance deadline of five years from authority to override a state’s BART CAA. As noted by several commenters, final publication of this notice for all determination simply because it States have the lead role in determining SCR-based emission limits. As disagrees with the state’s conclusions. BART for individual sources through explained in our proposal, five years is The commenters often stated that the SIPs. However, EPA also has a crucial a reasonable time frame for SCR design states are empowered by the CAA to role in reviewing SIPs for compliance and installation, particularly where determine how best to weigh each of the with the requirements of the CAA and retrofits of multiple units at a single statutory BART factors and that EPA’s its implementing regulations. Pursuant facility are required. Granting the full only legal role in SIP review is to to CAA section 110, States must submit five years for SCR design and determine whether the state’s plan is SIPs to EPA for review and EPA must installation will allow the facilities to consistent with the CAA. The review SIPs for consistency with the tie in the SCR systems during routinely commenters generally stated the belief Act’s requirements and disapprove any scheduled maintenance outages, which that ADEQ’s BART determinations fully SIP revision that ‘‘would interfere with are typically scheduled for every three complied with the CAA, the Regional any applicable requirement’’ of the years. With respect to Coronado in Haze Rule and the BART Guidelines. Act.215 The CAA also empowers EPA to particular, the five-year compliance The commenters frequently cited call for SIP revisions ‘‘[w]henever [EPA] schedule will allow SRP sufficient time American Corn Growers Ass’n. v. EPA, finds that the applicable to design and install the SCR system on 291 F.3d (D.C. Cir. 2002); EME Homer implementation plan for any area is Unit 1 and to design and install a low- City Generation, L.P. v. EPA, No. substantially inadequate to * * * load temperature controller on Unit 2, comply with any requirement of this which likely must be done in the period 213 BART Guidelines, 40 CFR Part 51, Appendix after the SCR for Unit 2 is placed into Y, section IV.D.5. 215 CAA section 110(a)(1), (k)(3) and (l), 42 U.S.C. operation (June 1, 2014). 214 Id. Section V. 7410(a)(1), (k)(3) and (l).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00054 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72565

chapter,’’ and impose sanctions when Guidelines).222 Thus, the statute from State to State.’’’ 226 By contrast, EPA determines they are ‘‘reasonable provides EPA a key oversight role in Regional Haze SIPs and BART and appropriate for the purpose of reviewing SIPs for compliance with the determinations are subject to detailed ensuring that the requirements [of the RHR and BART requirements. requirements set forth in CAA sections Act] * * * are met.’’ 216 Furthermore, The cases cited by commenters do not 169A, the RHR and the BART the Act mandates that EPA promulgate support an argument that EPA’s role as Guidelines. While in Luminant and a FIP when EPA finds that a State has a reviewer is any less critical in the Texas, the Fifth Circuit found that EPA failed to submit a required SIP to the regional haze context than it is in had failed to tie its disapproval to any Agency, failed to submit a complete SIP, reviewing other SIP components. In requirement of the CAA or EPA’s or where EPA disapproves a SIP.217 American Corn Growers v. EPA, the implementing regulations,227 in this Thus, the CAA provides EPA with a petitioners challenged the original RHR case our disapproval is based on the critical oversight role in ensuring that because, among other things, the RHR SIP’s failure to comply with CAA SIPs meet the requirements of the CAA. treated one of the five statutory factors sections 110(a)(2) and 169A(b)(2)(A), as Nothing in the CAA indicates that differently than the others by requiring implemented through the RHR and the EPA’s role is less important in the States to consider the degree of visibility BART Guidelines. context of the Regional Haze program improvement from imposing BART on a As noted above, CAA section than under other CAA programs. On the group of sources rather than on a 110(a)(2)(J) requires all SIPs to ‘‘meet contrary, CAA section 110(a)(2)(J) source-specific basis.223 The court the applicable requirements’’ of Part C explicitly requires that SIPs ‘‘meet the concluded that such a requirement of Title I of the CAA, including the applicable requirements’’ of Part C of could force States to apply BART requirement that each source found Title I of the CAA including the controls at sources without evidence subject-to-BART, ‘‘procure, install, and requirements for visibility protection set that the individual sources contributed operate, as expeditiously as practicable forth in sections 169A and 169B.218 to visibility impairment at a Class I area, (and maintain thereafter) the best Pursuant to section 169A(b), EPA is which encroached on States’ primary available retrofit technology * * *’’ 228 required to promulgate visibility authority under the regional haze Section 169A(g)(2) further provides that: protection regulations that apply to provisions to determine which In determining best available retrofit ‘‘each applicable implementation plan’’ individual sources are subject to BART technology the State (or the Administrator in (i.e., each SIP or FIP) 219 for each State and what BART controls are appropriate determining emission limitations which containing one or more Class I areas and for each source.224 Therefore, the court reflect such technology) shall take into each State ‘‘emissions from which may vacated the visibility improvement part consideration the costs of compliance, the reasonably be anticipated to cause or of the original RHR as contrary to the energy and nonair quality environmental 225 impacts of compliance, any existing contribute to any impairment of statute. Contrary to some pollution control technology in use at the 220 visibility in any [Class I area].’’ The commenters’ suggestions, however, the source, the remaining useful life of the CAA specifies that these regulations American Corn Growers decision did source, and the degree of improvement in (including the RHR) must require each not address EPA’s authority to reject a visibility which may reasonably be such SIP or FIP to ‘‘contain such State’s BART determinations for failure anticipated to result from the use of such emission limits, schedules of to conform to the CAA, the RHR or the technology.229 compliance and other measures as may BART Guidelines. Similarly, the RHR provides that: be necessary to make reasonable Commenters also cite Luminant progress toward meeting the national Generation v. EPA, 675 F.3d 917, 921 The determination of BART must be based goal,’’ including implementation of (5th Cir. 2012) and Texas v. EPA, 690 on an analysis of the best system of continuous emission control technology BART, as determined by the State (or by F.3d 670 (5th Cir. 2012). Neither of available and associated emission reductions 221 EPA in the case of a FIP). Moreover, these cases involves BART or the CAA’s achievable for each BART-eligible source that the CAA requires that BART for each regional haze provisions at all. Rather, is subject to BART within the State. In this ‘‘fossil-fuel fired generating power plant they involved EPA’s disapprovals of SIP analysis, the State must take into having a total generating capacity in revisions involving Texas’s minor new consideration the technology available, the excess of 750 megawatts’’ must be source review (NSR) program. As noted costs of compliance, the energy and nonair determined pursuant to the guidelines by the Luminant court, ‘‘because ‘the quality environmental impacts of promulgated by EPA (i.e., the BART Act includes no specifics regarding the compliance, any pollution control equipment in use at the source, the remaining useful life structure or functioning of minor NSR of the source, and the degree of improvement 216 See id. 42 U.S.C. 7410(k)(5), (m). programs’ and because the in visibility which may reasonably be 217 See id. section 7410(c)(1). implementing regulations are ‘very anticipated to result from the use of such 218 CAA sections 110(a)(2)(J), 169A and 169B 42 general [,] * * * SIP-approved minor technology.230 U.S.C. 7410(a)(2)(J), 7491 and 7492. NSR programs can vary quite widely 219 Under the CAA, ‘‘applicable implementation ADEQ’s BART determinations for NOX plan’’ is defined as ‘‘the portion (or portions) of the at Apache Units 2 and 3, Cholla Units 222 implementation plan, or most recent revision Id. In this case, Cholla and Coronado each 2, 3 and 4 and Coronado Units 1 and 2 thereof, which has been approved under [CAA have a total generating capacity in excess of 750 section 110], or promulgated under [CAA section megawatts, while Apache has a total plant-wide fall short of these requirements in 110](c) * * * and which implements the relevant generating capacity of 560 megawatts. Thus, the several respects. requirements of [the CAA].’’ CAA section 302(q), 42 BART Guidelines are mandatory for BART First, ADEQ did not analyze the ‘‘best U.S.C. 7602(q). In other words, an ‘‘applicable determinations at Cholla and Coronado and serve system of continuous emission control implementation plan’’ is an EPA-approved SIP or as non-binding guidance with respect to Apache. Tribal Implementation Plan, or an EPA- 223 291 F.3d at 5–9. 226 promulgated FIP. 224 Id. at 7–8. 675 F.3d at 922 (citing 74 FR 51418, 51421 (Oct. 6, 2009). 220 42 U.S.C. 7491(b)(2). In promulgating the RHR, 225 EPA revised the RHR to address the court’s 227 EPA determined that ‘‘all States contain sources decision in American Corn Growers at the same 675 F.3d at 924, 929; 690 F.3d at 679, 682, whose emissions are reasonably anticipated to time as we promulgated the BART Guidelines. 70 686. contribute to regional haze in a Class I area and, FR 39104 (July 6, 2005). The revised RHR and the 228 CAA section 169A(b)(2)(A)., 42 U.S.C. therefore, must submit regional haze SIPs.’’ 64 FR Guidelines were upheld by the D.C. Circuit in 7491(b)(2)(A). 35720; see also 40 CFR 51.300(b)(3). Utility Air Regulatory Group v. EPA, 471 F.3d 1333 229 42 U.S.C. 7491(g)(2). 221 42 U.S.C. 7491(b)(2). (D.C. Cir. 2006). 230 40 CFR 51.308(e)(1)(ii)(A).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00055 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72566 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

technology available and associated visibility impairment at all Class I • A requirement that each source subject emission reductions achievable.’’ Rather areas.232 Thus ADEQ’s BART to BART maintain the control equipment required by this subpart and establish it accepted the source’s own assertions determinations for NOX at Apache Units procedures to ensure such equipment is about what emissions reductions were 2 and 3, Cholla Units 2, 3 and 4 and properly operated and maintained. achievable with various control Coronado Units 1 and 2 do not meet the technologies. For example, in response requirements of CAA section 169A(g)(2) These two requirements are mandatory to comments from the FLMs arguing or 40 CFR 51.308(e)(1)(ii)(A). elements of the RHR and are necessary that SCR could achieve lower rates on In addition, 40 CFR 51.308(e)(1)(ii)(B) to ensure that BART is procured, 30-day-rolling average, ADEQ stated provides that: installed and operated, as expeditiously that: as practicable and maintained The determination of BART for fossil-fuel thereafter, as required under CAA ADEQ’s BART evaluations were based on fired power plants having a total generating section 169A(b)(2)(A). Moreover, CAA site-specific information provided by the capacity greater than 750 megawatts must be section 110(a)(2) requires that emissions applicants. It is the Department’s made pursuant to the guidelines in appendix understanding that such information was Y of this part (Guidelines for BART limits such as BART be ‘‘enforceable’’ based partially on feedback received from Determinations under the Regional Haze and section 302(k) requires emissions vendors and plant personnel who are Rule). limits to be met on a continuous basis. intimately familiar with the specific Arizona’s Regional Haze SIP lacks equipment that is being considered. In that Cholla and Coronado each have a requirements for monitoring, regard, the Department based its BART generating capacity greater than 750 recordkeeping and reporting sufficient computations on the emission rates proposed megawatts. Therefore, the BART to ensure that the BART limits are by the applicant for the different control determinations for these BART sources enforceable and are met on a continuous 231 technology options. must be made pursuant to the BART basis. While it is certainly reasonable to Guidelines. However, ADEQ’s BART Therefore, Arizona’s BART consider site-specific information determinations for these sources did not determinations for Apache, Cholla and provided by the sources as part of a fully comply with the BART Guidelines. Coronado do not meet several BART analysis, it is not reasonable to In particular, as explained more fully requirements of the CAA, the RHR and assume, with no independent analysis, above, contrary to the Guidelines’ the BART Guidelines. Accordingly, we that the sources have appropriately direction that ‘‘cost estimates should be are compelled to partially disapprove identified the emissions reductions based on the OAQPS Control Cost Arizona’s Regional Haze SIP. achievable with the best available Manual, where possible,’’ the control Finally, several commenters cited controls. ADEQ provided no evidence cost calculations supplied by the EME Homer City Generation v. EPA, No. that the sources’ estimates were based utilities and relied upon by ADEQ 11–1302 (D.C. Cir. Aug. 21, 2012). In on legitimate site-specific included line item costs not allowed by EME Homer City Generation, the D.C. considerations or that ADEQ undertook the Control Cost Manual, such as Circuit vacated EPA’s ‘‘Transport Rule’’ any verification of these estimates. owner’s costs, surcharge, and AFUDC. (also known as the ‘‘Cross-State Air Therefore, ADEQ has not demonstrated Thus, ADEQ’s consideration of the ‘‘cost Pollution Rule’’ or ‘‘CSAPR’’), which that its BART determinations were of compliance’’ for these units was not was promulgated by EPA to address ‘‘based on an analysis of the best system consistent with the Guidelines. interstate transport of SO2 and NOX of continuous emission control Furthermore, as explained above, under CAA section 110(a)(2)(D). The technology available and associated ADEQ’s consideration of visibility court found that the Transport Rule emission reductions achievable.’’ benefits was inconsistent with the exceeded EPA’s authority under section Second, ADEQ has not demonstrated Guidelines because the State did not 110(a)(2)(D) because the rule had the that it actually took into consideration consider benefits at multiple Class I potential to require upwind States to the BART factors in making its areas and multiple BART-eligible units reduce emissions by more than their determinations. In particular, while at each source. In addition, ADEQ failed own significant contributions to ADEQ provided information regarding to provide ‘‘a justification for adopting downwind nonattainment and because each of the factors, it gave no the technology [the State selected] as the EPA had not given states an opportunity explanation or rationale for how it ‘best’ level of control, including an to submit SIPs after it quantified their reached a determination based on that explanation of the CAA factors that led obligations for emissions reductions to information. [the State] to choose that option over address transport. Commenters here Finally, ADEQ did not appropriately other control levels.’’ 233 Therefore, point to the D.C. Circuit’s statements consider the ‘‘degree of improvement in ADEQ’s BART determinations for NOX concerning state and federal roles under visibility which may reasonably be at Cholla and Coronado do not comply the CAA and argue that EPA has anticipated’’ from installation of BART with 40 CFR 51.308(e)(1)(ii)(B). exceeded its statutorily mandated role because it did not consider visibility Finally, for all pollutants at all units in proposing to disapprove portions of benefits at all of the affected Class I covered by today’s action, ADEQ’s Arizona’s Regional Haze SIP and areas, nor did it consider the total Regional Haze SIP does not meet the promulgate a FIP. visibility benefit expected to result from requirements of 40 CFR 51.308(e)(1)(iii) While we agree that the general the entire BART-eligible source. and (iv) because it lacks the following principles concerning state and federal Overlooking significant visibility elements: roles under Title I of the CAA apply to benefits at additional areas and from • our action here, we do not agree that our A requirement that each source subject action here is inconsistent with those multiple BART-eligible units to BART be required to install and operate considerably understates the overall BART as expeditiously as practicable, but in principles. In this action, we are benefit of controls to improve visibility no event later than 5 years after approval of fulfilling our statutory duty to review and is contrary to the very purpose of the implementation plan revision. Arizona’s Regional Haze SIP, including BART, i.e., ‘‘eliminating or reducing’’ its BART determinations, for 232 CAA section 169A(b)(2)(A). compliance with the applicable 231 Arizona Regional Haze SIP, Appendix E, 233 BART Guidelines, 40 CFR Part 51, Appendix requirements of the CAA and the RHR, ‘‘Responsiveness Summary’’ at 13. Y, section IV.E.2. and to disapprove any portions of the

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00056 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72567

plan that do not meet those consistency is particularly relevant for the plan revision in part. The plan requirements. Based on our review of BART determinations at fossil-fuel fired revision shall not be treated as meeting the SIP, we proposed to determine that power plants having a capacity in the requirements of this chapter until certain elements of Arizona’s Regional excess of 750 megawatts, which must be the Administrator approves the entire Haze SIP did meet the requirements of made pursuant to the BART plan revision as complying with the the CAA and the RHR, and we proposed Guidelines.237 To the extent a BART applicable requirements of this to approve those elements. However, for determination for such a power plant is chapter.238 the reasons explained in detail in our plainly inconsistent with EPA-approved Some commenters have read this proposal and elsewhere in this determinations for similar sources, it is provision as requiring that EPA act on document, we have concluded that more likely to be inconsistent with the Arizona’s Regional Haze SIP as a whole. Arizona’s BART determinations for NOX RHR and the BART Guidelines and We disagree that this language addresses at several units did not comply with the therefore to warrant greater scrutiny for the question of whether EPA may requirements of the CAA and the RHR. compliance with the applicable consider different elements of a State’s Based on these findings, we are required requirements. plan in separate notice and comment to disapprove these portions of Comment: Several commenters rulemakings. However, even assuming Arizona’s Regional Haze SIP. (ACCCE, ADEQ, APS, SRP) asserted that that this provision of the Clean Air Act In some instances, we expressed our it is contrary to the CAA for EPA to did limit EPA’s ability to act findings of non-compliance with the propose action on only the portions of sequentially on portions of a SIP relevant requirements in terms of ADEQ’s SIP that address the three submission, the requirement to act on a ‘‘disagreement’’ with the state’s power plants that are the subject of the submittal ‘‘as a whole’’ applies only if analysis. These statements were not proposed FIP. One commenter (APS) the submittal meets all of the applicable intended to suggest that our proposed stated that EPA may not ignore all other requirements of the CAA. As explained partial disapproval was simply based on sources of visibility-impairing in our proposal and elsewhere in this policy disagreements with the state. pollutants in the state (nor may it ignore document, we have determined that the Rather we used the term ‘‘disagree’’ as the other categories of visibility- Arizona Regional Haze SIP does not a short hand for our findings that impairing pollutants by focusing only meet all of the applicable requirements specific elements of Arizona’s analyses on nitrates, sulfates and PM) and of the CAA. Specifically, we have did not meet the requirements of the establish BART limitations for the three determined that the submittal as a CAA and the RHR. For example, we affected power plants outside the whole does not meet the requirements noted that, ‘‘[w]e disagree with several context of the long-term strategy and of CAA section 169A(b)(2)(A), as aspects of the NOX BART analysis for larger reasonable progress requirements implemented through the RHR and the Apache Units 2 and 3.’’ 234 We then of the regional haze program. BART Guidelines. Under these went on to list the specific deficiencies Commenters ACCCE, ADEQ and SRP circumstances, we are clearly not in the state’s analysis, and concluded contended that CAA section 110(k)(3) obligated to act on the plan as a whole, that ‘‘we are proposing to disapprove requires EPA either to approve a SIP but are given discretion to act on ADEQ’s BART determination for NOX at submittal ‘‘as a whole’’ or to approve distinct portions of the plan.239 Apache Units 2 and 3, since it does not that SIP submittal in part and While we agree that, as a matter of comply with 40 CFR disapprove it in part in a single policy, it is generally preferable to act 51.308(e)(1)(ii)(A).’’ 235 We made similar rulemaking that addresses in its entirety on plan submissions as a whole, we are findings with respect to ADEQ’s BART ‘‘the plan revision.’’ The commenters currently subject to a court-ordered determination for NOX at Cholla Units indicated that this requirement of the deadline of November 15, 2012 to act on 2, 3 and 4 and Coronado Units 1 and CAA is sensible because it is the plan the BART determinations for Apache 2.236 We have also described in detail, as a whole, with all its elements Generating Station, Cholla Power Plant both in our proposal and in this working together, that must ensure that and Coronado Generating Station.240 document, the other aspects of the the CAA’s regional haze-related goals Although these BART determinations state’s BART determinations that do not are being reached; any other approach to are part of the overall Regional Haze comply with the CAA and the RHR. SIP review and approval would fail to plan for Arizona, they are also severable Finally, some commenters appear to take into account the full array of from that plan, since BART have misunderstood our statement that regulatory choices that Arizona has determinations are made on a source-by- ADEQ’s ‘‘NOX BART determinations for made to address regional haze. source basis and are not dependent the coal-fired units are neither Response: We do not agree that we are upon other elements of the plan.241 consistent with the requirements of the required to act on Arizona’s Regional Act nor with BART decisions that other Haze SIP as a whole. As noted by some 238 42 U.S.C. 7410(k)(3). states have made.’’ As noted by several commenters, our action on Arizona’s 239 See Hall v. EPA, 273 F.3d 1146, 1159 (9th Cir. commenters, the CAA and the RHR Regional Haze SIP is governed by inter 2001) (section 110(k)(3) ‘‘permits the EPA to issue provide states with considerable alia, CAA section 110(k)(3), which ‘partial approvals,’ that is, to approve the States’ SIP discretion in deciding how to weigh the provides that in the case of any revisions in piecemeal fashion’’). 240 EPA agreed to this deadline after concluding statutory factors as a part of a BART submittal on which the Administrator is that litigation would most likely result in a shorter analysis. However, this discretion must required to act under section 110(k)(2), schedule than that to which Plaintiffs had agreed be reasonably exercised in compliance the Administrator shall approve such in negotiation. See Sierra Club v. Johnson, 444 with the applicable requirements. submittal as a whole if it meets all of the F.Supp.2d 46, 58 (D.D.C. 2006) (‘‘this case devolves to a single issue: whether defendant has met the Consistency with other EPA-approved applicable requirements of this chapter. ‘heavy burden’ of demonstrating that it would be BART determinations is one marker of If a portion of the plan revision meets impossible to comply with plaintiff’s proposed reasonableness, as well as compliance all the applicable requirements of this * * *’’). with the requirements of the RHR. Such chapter, the Administrator may approve 241 See 40 CFR 51.308(e)(1)(ii)(A)(‘‘[t]he determination of BART must be based on an the plan revision in part and disapprove analysis of the best system of continuous emission 234 77 FR 42846. control technology available and associated 235 Id. 237 CAA section 169A(b) and 40 CFR emission reductions achievable for each BART- 236 77 FR 42849, 42851. 51.308(e)(1)(ii)(B). Continued

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00057 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72568 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

Therefore, we are taking action on these to the general requirements of section Federal Register.250 No party filed such BART determinations first and we will 308, EPA also adopted specific a petition. act on the remainder of the Arizona provisions that gave a handful of states, At the time of the 2009 Finding, EPA Regional plan in accordance with the including Arizona, the option of anticipated that ADEQ would submit a court-ordered deadlines for that action. submitting a regional haze SIP based on SIP revision covering 309(d)(4) and Comment: One commenter (ADEQ) the recommendations of the Grand 309(g), which would enable EPA to fully asserts that EPA does not have the Canyon Visibility Transport approve ADEQ’s 309 SIP as meeting all authority to adopt a FIP because none of Commission (GCVTC). Under the RHR, of the requirements of the Regional Haze the three triggering events for a FIP a SIP approved by EPA as meeting all Rule, thus ending the FIP clock. under CAA section 110(c)(1) has of the requirements of section 309 However, ADEQ did not submit a 309 occurred. Specifically, the commenter would be ‘‘deemed to comply with the SIP revision to address these two states that: requirements for reasonable progress elements, but instead decided to * * * for EPA to have authority to with respect to the 16 Class I areas [on develop a 308 SIP, which it submitted promulgate a regional haze FIP in Arizona, the Colorado Plateau] for the period to EPA in February 2011. one of three events must have occurred: (1) from approval of the plan through In January 2011, EPA received a a finding of failure to submit a regional haze 2018.’’ 244 notice of intent to sue covering dozens SIP, (2) a finding of failure to satisfy the Arizona made two submittals under of states, including Arizona, stating that minimum criteria for a complete regional section 309 in 2003 and 2004, but never we had not met the statutory deadline haze SIP under section 110(k)(1)(A) or (3) submitted a complete 309 SIP.245 for promulgating Regional Haze FIPs disapproval of a regional haze SIP submitted Rather, on December 24, 2008, ADEQ and/or approving Regional Haze SIPs. by Arizona. None of these three events has This notice was followed by a lawsuit occurred. sent a letter to EPA re-submitting its prior 309 SIP submissions and filed by several advocacy groups 251 With respect to EPA’s January 2009 acknowledging that the submittal did (Plaintiffs) in August 2011. In order finding of failure to submit, the not include provisions to address the to resolve this lawsuit and avoid commenter argues that: requirements of 309(d)(4) or 309(g).246 litigation, EPA entered into a Consent Section 110(c)(1) * * * does not allow These were not minor omissions: Decree with the Plaintiffs, which sets EPA to treat the omission of elements from 309(d)(4) required the submission of deadlines for action for all of the states a SIP submission as a failure to submit a SIP. ‘‘better than BART’’ milestones and a covered by the lawsuit, including Section 110(c)(1) is quite specific. If EPA trading program for SO , as well as Arizona. This decree was entered and believes SIP omissions render a SIP 2 BART requirements for stationary later amended by the Federal District incomplete, the agency may make a finding Court for the District of Columbia over source PM and NOX emissions, and under section 110(k)(1)(A) within the time 252 309(g) required implementation of any the opposition of Arizona. period required by section 110(k)(1)(B) and In opposing the entry of the consent additional measures necessary to start the FIP clock under the second clause decree, Arizona argued that the 2009 of section 110(c)(1)(A). If EPA cannot make demonstrate reasonable progress for the Finding did not give EPA authority to such a finding or, as in this case, fails to do additional Class I areas, in compliance promulgate a Regional Haze FIP for so, the agency may disapprove the SIP, and with the provisions of § 51.308(d)(1) Arizona. The court rejected this start the FIP clock under section 110(c)(1)(B). through (4).247 Thus, as of 2008, ADEQ’s argument, explaining that: By treating the alleged omission of elements Regional Haze SIP, by its own from a SIP as the failure to make a required admission, did not include provisions Arizona contends that the Finding did not submission under the first clause of section constitute a disapproval of the SIPs that had 110(c)(1)(A), EPA is circumventing these addressing BART (or for an alternative to BART) for NO , PM or SO . On previously been submitted because it only procedures. X 2 notes that Arizona did not submit two of January 15, 2009 EPA found that 37 The commenter adds that if EPA did Section 309’s required elements. Ariz. Opp. states, including Arizona, had failed to have the authority to promulgate a [Dkt. # 24] at 6. The Court does not read the make all or part of the required SIP regional haze FIP, it would only have 2009 Finding so narrowly. In the Court’s submissions to address regional haze.248 view, the 2009 Finding reaches a conclusion the authority to address those elements We explained that: that Arizona ‘has failed to make a required of the SIP that EPA identified as having submission or finds that the plan or plan not been submitted, and EPA has never This finding starts the two year clock for revision submitted by the State does not found that Arizona failed to submit a the promulgation by EPA of a FIP. EPA is not satisfy the minimum criteria.’ 42 U.S.C. SIP establishing BART. required to promulgate a FIP if the state 7410(c)(1). Under the CAA, this triggers the Response: We do not agree that we makes the required SIP submittal and EPA EPA’s statutory obligation to promulgate a takes final action to approve the submittal FIP.253 lack authority to issue a FIP addressing within two years of EPA’s finding.249 BART requirements for the three Under the terms of the Consent Decree, sources covered by today’s action. The Under the CAA, any party seeking as amended, EPA is currently subject to commenter’s arguments in this regard judicial review of EPA’s finding of two sets of deadlines for taking action appear to be based on a failure to submit (‘‘2009 Finding’’) was on Arizona’s Regional Haze SIP. misunderstanding of the requirements required to file a petition for review Specifically, the CD requires that: with the appropriate United States of the CAA and the RHR in relation to By the ‘‘Proposed Promulgation Deadlines’’ Arizona’s Regional Haze submittals. Circuit Court of Appeals within 60 days set forth in Table A below EPA shall sign a EPA promulgated the original RHR in of publication of the Finding in the notice(s) of proposed rulemaking in which it 1999.242 As relevant here, section 308 of the RHR requires states to submit SIPs 245 We have included a more detailed history of 250 CAA section 307(b). 42 U.S.C. 7607(b). that establish reasonable progress goals Arizona’s submissions under 309 in the docket for 251 National Parks Conservation Association v. and long-term strategies for achieving this action (Docket No. EPA–R09–OAR–2012– Jackson (D.D.C. Case 1:11–cv–01548). 0021). 252 National Parks Conservation Association v. those goals and provide for 246 Letter from Stephen A. Owens, ADEQ, to Jackson (D.D.C. Case 1:11–cv–01548), Memorandum 243 implementation of BART. In addition Wayne Nastri, EPA (Dec. 14, 2008). Order and Opinion (May 25, 2012) and Minute 247 40 CFR 51.309(d)(4)(i) and (vii), (g)(2). Order (July 2, 2012). eligible source that is subject to BART within the 248 74 FR 2392. 253 See NPCA v. EPA, (D.D.C. Case 1:11–cv– State.’’ 249 Id. at 2393 01548). Dkt # 35, at 3, n. 1.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00058 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72569

proposes approval of a SIP, promulgation of state has not submitted a SIP or when on February 28, 2011, and the SIP was a FIP, partial approval of a SIP and EPA has made a final determination that deemed complete by operation of law promulgation of a partial FIP, or approval of a submitted SIP is not approvable (citing on August 28, 2011, pursuant to CAA a SIP or promulgation of a FIP in the Train v. NRDC, 421 U.S. 60, 79 (1975)). section 110(k)(1)(B).256 This, in turn, alternative, for each State therein, that The commenters believe this principle triggered a deadline of August 28, 2012, collectively meet the regional haze implementation plan requirements that were is confirmed by CAA sections for us to take final action on the SIP, 257 due by December 17, 2007 under EPA’s 307(d)(1)(B), (3) and (6) because EPA pursuant to CAA section 110(k)(1)(B). regional haze regulations. cannot present the relevant factual, We acknowledge that this deadline has By the ‘‘Final Promulgation Deadlines’’ set legal, and policy information and now passed and we intend to act as forth in Table A below, EPA shall sign a rationale necessary to justify a proposed quickly as possible to fulfill our duty to notice(s) of final rulemaking promulgating a or final FIP rule until it has properly act on those portions of the SIP not FIP for each State therein to meet the regional taken final action on any relevant SIP addressed in today’s action. However, haze implementation plan requirements that before it. the fact that we have not acted on the were due by December 17, 2007 under EPA’s One commenter (EEI) also states that entirety of the SIP submittal does not regional haze regulations, except where, by EPA’s assertion that it was compelled to remove or otherwise alter our legal such deadline EPA has for a State therein signed a notice of final rulemaking propose a FIP at the same time that it obligation to promulgate a FIP under unconditionally approving a SIP, or disapproved a portion of the Arizona CAA section 110(c). Our FIP duty does promulgating a partial FIP and unconditional SIP, due to a two-year FIP clock that not terminate until we have actually approval of a portion of a SIP, that started with EPA’s 2009 Finding of approved the submitted SIP. As collectively meet the regional haze Failure to Submit, is inconsistent with explained in our NPRM, TSD and implementation plan requirements that were the CSAPR decision. The commenter elsewhere in this document, we cannot due by December 17, 2007 under EPA’s stated that EPA did not provide approve the State’s BART regional haze regulations. sufficient notice of the problems with determinations for NOX at Apache, Table A, as revised, sets a proposal the SIP to enable Arizona to remedy Cholla and Coronado, nor can we deadline for BART determinations for them, which is precisely the same approve the compliance-related Apache Generating Station, Cholla problem identified by the CSAPR court. requirements that were omitted from the Power Plant and Coronado Generating The commenter adds that EPA must Arizona Regional Haze SIP. Therefore, Station of July 2, 2012 and the final provide the state a realistic opportunity we are obligated to promulgate a FIP to action deadline for these three BART to avoid being pulled into a FIP. Given address these requirements, and we are determinations of November 15, 2012. that EPA has consent decree obligations doing so in today’s action. The deadline for EPA to propose action to finalize BART requirements for the Furthermore, while we agree that the on the remainder of the Arizona EGUs addressed by the proposed SIP by procedural requirements for Regional Haze SIP is December 8, 2012, November 15, 2012, and EPA did not promulgation of a FIP under 110(c) are and the deadline for final action is July propose disapproval of the SIP until set forth in CAA section 307(d),258 we 15, 2013.254 July 20, 2012, a reasonable opportunity do not agree that our action violates that Thus, pursuant to CAA section to develop and receive approval of a provision in any way. Consistent with 110(c)(1) and the court’s orders entering revised SIP was not offered to the state. the requirements of that section, our and amending the Consent Decree, we Response: We do not agree that we are proposal included a summary of the are not only authorized, but are required required to take final action on factual data on which our proposed FIP to issue a FIP for any portion of the Arizona’s Regional Haze SIP before was based, as well as the methodology Arizona SIP that we cannot approve. For promulgating a FIP. Commenters’ used in obtaining the data and in the reasons stated in our proposal and arguments to this effect appear to analyzing the data and the major legal elsewhere in this document, we have conflate the procedural requirements for interpretations and policy determined that we cannot approve the EPA’s issuance of a FIP with procedural considerations underlying the proposed 259 state’s BART determinations for NOX at requirements for action on a SIP. In fact, FIP. In addition, we provided a Apache, Cholla and Coronado, nor can these are two actions are governed by detailed evaluation of Arizona’s BART we approve the compliance-related different provisions of the CAA. analyses for the relevant units, which requirements that were omitted from the As explained in the previous formed the basis for our proposed action Arizona Regional Haze SIP. Therefore, response, EPA’s 2009 finding that on those portions of the Arizona we are obligated to promulgate a FIP to Arizona failed to submit a complete Regional Haze SIP.260 This final address these requirements. Regional Haze SIP triggered a ‘‘FIP rulemaking includes similar information Comment: Several commenters (AUG, clock’’ under CAA section 110(c).255 with respect to the SIP and the FIP, as EEI, PacifiCorp, SRP) stated that EPA This FIP clock could only have been well as ‘‘an explanation of the reasons cannot propose or finalize a NOX BART stopped if Arizona had submitted, and for any major changes in the FIP for these Arizona plants until it has EPA had fully approved a Regional Haze promulgated rule from the proposed taken final action (following notice-and- SIP, before January 15, 2011. Neither of rule’’ and ‘‘a response to each of the comment rulemaking) on ADEQ’s these two things occurred. Therefore, Regional Haze SIP. According to the EPA remains subject to this ‘‘FIP duty.’’ 256 42 U.S.C. 7410(k)(1)(B). commenters, EPA’s authority to propose Our action today fulfills part of that 257 42 U.S.C. 7410(k)(2). duty. 258 See CAA section 307(d)(1)(B), 42 U.S.C. and then take final action to promulgate 7607(d)(1)(B), (‘‘This subsection applies to * * * a FIP comes into existence only when a As several commenters noted, the promulgation or revision of an implementation Arizona submitted a Regional Haze SIP plan by the Administrator under [CAA section 254 On November 13, 2012, the D.C. District Court 110](c)’’] granted a motion by EPA to modify the Consent 255 42 U.S.C. 7410(c). See also Train, 421 U.S. at 259 See CAA section 307(d)(3), 42 U.S.C. Decree to extend the deadlines for promulgation of 64, 79 (explaining that the 1970 CAA Amendments 7607(d)(3). a FIP for any remaining elements of the SIP that are ‘‘sharply increased federal authority and 260 The SIP portion of our action is subject to the disapproved. Under the revised deadlines, EPA will responsibility in the continuing effort to combat air procedural requirements of section 553(b) of propose any necessary FIP elements by March 8, pollution,’’ including giving EPA authority to Administrative Procedure Act (APA), 5 U.S.C. 2013, and finalize such elements by October 15, devise a FIP if the State’s plan fails to satisfy the 553(b), rather than the requirements of CAA 2013. standards of section 7410(a)(2)). subsection 307(d), 42 U.S.C. 7607(d).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00059 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72570 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

significant comments, criticisms, and 2006).267 As explained in our proposal action to approve, in part or in whole, new data submitted in written or oral and elsewhere in this document, the many of these SIPs.272 This stands in presentations during the comment BART Guidelines provide detailed contrast to the situation in EME Homer period.’’ 261 Therefore, our action instructions to states on how to City Generation, where, the court noted complies with the applicable procedural determine which sources are subject to that, ‘‘every Transport Rule State that requirements of the CAA. BART and how to analyze the five submitted a good neighbor SIP for the Finally, we do not agree with statutory factors in order to set 2006 24-hour PM2.5 NAAQS was commenters’ assertions that the D.C. emissions limits representing BART for disapproved.’’ 273 Thus, it is clear that Circuit’s decision in EME Homer City each subject-to-BART source.268 In states had ample opportunity to submit Generation precludes us from 2006, responding to specific questions approvable Regional Haze SIPs before promulgating a partial FIP concurrently from various States and Regional EPA was obligated to promulgate with our partial disapproval of Planning Organizations (RPOs), EPA Regional Haze FIPs under CAA section Arizona’s Regional Haze SIP. In EME issued further guidance to help States 110(c). Homer City Generation, the court found implement the RHR and BART With respect to Arizona’s Regional that EPA had acted improperly in Guidelines.269 Haze SIP in particular, we note that issuing the Transport Rule because we As noted in prior responses, EPA Arizona first made public its proposed simultaneously defined states’ ‘‘good issued a finding of failure to submit for 308 SIP during a comment period neighbor obligations’’ under CAA Regional Haze SIPs on January 15, 2009, beginning on October 28, 2010.274 At section 110(a)(2)(D)(i)(I) and issued FIPs thus triggering a FIP clock under CAA that time, EPA, the National Park to address those obligations.262 The section 110(c).270 By this time, states Service (NPS) (in consultation with the court explained that: had already had more than three years Fish and Wildlife Service) and the U.S. since issuance of the final BART Forest Service all submitted comments * * * the triggers for a FIP are EPA’s finding Guidelines (and more than two years expressing concern about the proposed that the SIP fails to contain a ‘‘required submission’’ or EPA’s disapproving a SIP since the final revisions to the RHR and SIP’s compliance with the CAA, the 275 because of a ‘‘deficiency.’’ But logically, a SIP the issuance of further guidance on the RHR and the BART Guidelines. cannot be deemed to lack a required RHR and BART) to develop their Among other things, EPA noted that the submission or be deemed deficient for failing Regional Haze SIPs. By the time the FIP SIP, ‘‘does not provide a sufficient level to implement the good neighbor obligation clock actually ran out in January 2011, of information and analysis to support until after EPA has defined the State’s good EPA had received Regional Haze SIPs its conclusions.’’ 276 NPS provided neighbor obligation. Once it defines the from nearly every state. EPA has since extensive comments on the proposed obligation, then States may be forced to proposed to approve, in part or in SIP, including detailed evaluations of revise SIPs under Section 110(k)(5) or to whole, the vast majority of these ADEQ’s BART analyses for each of the submit new SIPs under Section 110(a)(1). SIPs.271 We have also has taken final three sources at issue in today’s Only if that revised or new SIP is properly 277 deemed to lack a required submission or is action. In each instance, NPS 267 properly deemed deficient may EPA resort to In response to another D.C. Circuit decision, concluded that ADEQ had not Center for Energy and Economic Development v. conducted a valid BART analysis for a FIP for the State’s good neighbor EPA, 398 F.3d 653 (D.C. Cir. 2005), EPA revised the 263 278 obligation. RHR’s provisions governing alternatives to source- NOX. The Forest Service concurred with the initial comments provided by In essence, the D.C. Circuit found that specific BART determinations on October 13, 2006. These revisions did not alter the requirements for NPS on Arizona’s BART exclusion EPA’s findings of failure to submit and source-specific BART determinations that apply to process and ‘‘strongly disagree[d] with disapprovals of state transport SIPs did Arizona’s BART determinations at issue here. the adequacy of the Arizona reasonable 268 40 CFR Part 51, Appendix Y. While the not trigger FIP obligations under CAA progress analysis.’’ 279 Therefore, ADEQ section 110(c) because these actions Guidelines are only mandatory for fossil fuel-fired electric generating plants with a total generating had the benefit not only of the generally occurred ‘‘before [EPA] told the States capacity in excess of 750 megawatts, States are applicable requirements of the RHR, the what emissions reductions their SIPs encouraged to follow the BART Guidelines in making BART determinations for other types of were supposed to achieve under the 272 sources. Id. section I.H. The Guidelines also set See, e.g., 76 FR 34608 (California); 76 FR good neighbor provision.’’ 264 42557 (Delaware); 76 FR 80754 (Kansas); 77 FR 19 specific presumptive limits for SO2 and NOX for In this case, by contrast, EPA defined these large power plants, but allow states to apply (New Jersey); 77 FR 5191 (District of Columbia); 77 states’ obligations under the RHR and more or less stringent limits based upon source- FR 14604 (Arkansas); 77 FR 17334 (Nevada); 77 FR specific five-factor analyses. 70 FR 39131–39132. 24845 (South Dakota); 77 FR 40150 (Nebraska); 77 the BART Guidelines well in advance of FR 51915 (New York). 269 Memo from Joseph W. Paise Regarding its findings of failure to submit and 273 Regional Haze Regulations and Guidelines for Slip op. at 57. 274 subsequent SIP disapprovals. EPA BART (July 19, 2006); Additional Regional Haze Arizona Regional Haze SIP, Appendix E, promulgated the original RHR on July 1, Questions (Guidance) (Sept. 27 2006). In addition, Public Process. Approximately 60 days prior to the 265 EPA issued final ‘‘Guidance for Setting Reasonable public comment period, ADEQ sent a draft of the 1999. Following the D.C. Circuit’s SIP to the National Park Service and U.S. Forest decision in American Corn Growers, Progress Goals Under the Regional Haze Program’’ on June 1, 2007, but this Guidance is not directly Service. EPA revised the RHR and issued the relevant for individual BART determinations. 275 Id. final BART Guidelines on July 6, 270 74 FR 2392. 276 Id. Letter from Colleen McKaughan, EPA, to 2005.266 The revised RHR and the 271 See, e.g., 76 FR 36450 (Nevada); 77 FR 24794 Eric Massey, ADEQ (Dec. 2, 2010). 277 Guidelines were upheld by the DC (New York); 76 FR 13944 (California); 77 FR 11798 Id. NPS Initial Comments Arizona Draft (Rhode Island); 76 FR 27973 (Delaware); 77 FR Section 308 Regional Haze SIP (Nov. 29, 2010); NPS Circuit in Utility Air Regulatory Group 12770 (Nebraska); 77 FR 18052 (Colorado); 76 FR General BART Comments on ADEQ BART Analyses v. EPA, 471 F.3d 1333 (D.C. Cir. 16168 (Oklahoma); 77 FR 11914 (Vermont); 77 FR (Nov. 29, 2010); NPS Comments AEPCO—Apache 11928 (Wisconsin); 76 FR 52604 (Kansas); 76 FR Generating Station BART Analysis and Determination (Nov. 29, 2010); NPS Comments APS 261 CAA section 307(d)(6)(A) & (B), 42 U.S.C. 64186 (Arkansas); 77 FR 11839 (Maryland); 76 FR Cholla Generating Station BART Analysis and 7607(d)(6)(A) & (B). 58570 (North Dakota); 77 FR 3966 (Illinois); 76 FR 76646 (South Dakota). EPA proposed limited Determination (Nov. 29, 2010); NPS Comments 262 EME Homer City Generation, slip op. at 7. approval and limited disapproval of the Regional SRP’s Coronado Generating Station BART Analysis 263 Id. at 46. Haze SIPs of states covered by the Clean Air and Determination (Nov. 29, 2010); NPS Comments 264 Id. at 47 (emphasis in original). Interstate Rule (CAIR), due to the remand of CAIR on ADEQ BART Exemptions, (Dec. 1, 2010). 265 64 FR 35714. by the D.C. Circuit. See, e.g. 77 FR 3691 (Jan. 25, 278 Id. 266 70 FR 39104. This finding covered 37 states, 2012) (proposing limited approval and limited 279 U.S. Forest Service Specific Comments: the District of Columbia and the Virgin Islands. disapproval of Virginia’s Regional Haze SIP). Arizona Regional Haze SIP (Nov. 29, 2010).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00060 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72571

BART Guidelines and EPA Guidance, Response: We agree that the CAA, the prepared by the Clean Air Task Force, but also specific guidance from EPA and RHR and the BART Guidelines set out the commenter asserted that Cholla, the FLMs pointing out shortcomings in specific requirements that Regional Coronado and Apache collectively cause its Regional Haze SIP. Following receipt Haze SIPs must meet in order to be approximately 41 deaths, 63 heart of these comments, Arizona had the approved by EPA. Our action today attacks and 747 asthma attacks opportunity to revise its SIP to address addresses these requirements as they annually.282 Several other commenters the deficiencies identified by the apply to ADEQ’s BART determinations provided similar comments concerning commenters, but in most instances it for Apache, Cholla and Coronado, but health effects. chose not to do so.280 does not address the requirements as Response: We acknowledge the Finally, while we agree that, in the they apply to the remainder of Arizona’s commenters’ concerns regarding the absence of an expired statutory duty and Regional Haze SIP (e.g., the reasonable adverse health impacts of haze-causing a court-ordered deadline to issue a FIP, progress goals set by the state). EPA will emissions. We agree that the same PM2.5 it would be preferable for us to give propose action on these aspects of the emissions that cause visibility Arizona additional time to revise its SIP shortly and take final action after impairment can cause respiratory Regional Haze SIP prior to promulgation receiving comments. As explained in problems, decreased lung function, of a FIP, we simply do not have this the preceding responses, because of our aggravated asthma, bronchitis, and option under these circumstances. As prior finding of failure to submit, we are premature death. We also agree that the explained in our response to the required to issue a FIP for any portion same NOX emissions that cause previous comment, we are obligated to of the SIP that we cannot approve. Thus, visibility impairment also contribute to issue a FIP to address any gaps left by we are promulgating a FIP for those the formation of ground-level ozone, partial disapprovals of Arizona’s aspects of ADEQ’s BART determinations which has been linked with respiratory Regional Haze SIP. Nonetheless, we for Apache, Cholla and Coronado that problems, aggravated asthma, and even encourage ADEQ to submit a revised SIP we are not approving at this time. permanent lung damage. Finally, we also agree that SO emissions that cause to replace the FIP and will work with G. Other Comments 2 ADEQ to develop such a revised plan to visibility impairment also contribute to meet the requirements of the CAA and 1. Comment on Public Health and increased asthma symptoms, lead to the RHR. Ecosystem Impacts increased hospital visits, and can form particulates that aggravate respiratory Comment: One commenter Comment: A number of commenters and heart diseases and cause premature (Earthjustice) stated that the CAA’s provided comments on the potential death. Thus, to the extent that this FIP Regional Haze program establishes a health effects of our proposal. A number will lead to reductions in these national regulatory floor and requires of other commenters stated that the pollutants, there will be co-benefits for states to develop RH SIPs at least as Regional Haze program’s sole focus is public health. However, for purposes of stringent as this floor (citing 40 CFR the improvement of visibility in Class I this action, we are not authorized to 51.308). According to the commenter, areas, and is not a health-based or consider these benefits and we have not ADEQ’s SIP is legally and technically emissions reduction program. In relation to the Regional Haze program, done so. inadequate because it does not require In our NPRM, while discussing adequate BART emission limits, does any EPA emphasis on health and emissions reduction is inappropriate. Executive Order 13045 (Protection of not achieve ‘‘reasonable progress’’ are Children from Environmental Health required by the RHR and would fail to One commenter (SRP) stated that EPA’s assertion of health benefits is Risks and Safety Risks), we stated that, achieve natural visibility goals by 2064. to the extent the proposed rule will The commenter believes that the unsubstantiated by the proposed rule. A few commenters noted that the air limit emissions of NOX, SO2 and PM10, Arizona RH SIP fails to establish a the rule will have a beneficial effect on program that is at least as stringent as quality in Arizona varies from city to city, and stated that EPA should focus children’s health by reducing air the national floor and that therefore EPA pollution. In this action, while has a legal obligation to disapprove the on the areas with the poorest air quality first, such as Phoenix. discussing Executive Order 13045 SIP and to issue a FIP in its place under (Protection of Children from CAA section 110(c)(1).281 In contrast, one commenter (Earthjustice) stated that the same Environmental Health Risks and Safety pollutants that reduce visibility also Risks), we conclude that this action 280 For example, in response to detailed does not have a disproportionate effect comments from NPS regarding the efficiency and cause significant public health impacts. on children, but again note that to the cost of SCR, ADEQ stated that: The commenter noted that NOX is a ADEQ has determined that the cost computations precursor to ground level ozone, which extent this final action will limit presented by the facilities in support of their BART is associated with respiratory diseases, emissions of NOX, SO2 and PM10, the applications are reasonable. Many of the asthma attacks and decreased lung rule will have a beneficial effect on computations are based on vendor data and site- children’s health by reducing air specific conditions. The Department does not agree function, and that NOX reacts with other that the computations over-estimate the costs of substances to form particulates that can pollution that causes or exacerbates retrofit technologies and under-estimate the cause and worsen respiratory diseases, childhood asthma and other respiratory associated emission decreases and visibility aggravate heart disease, and lead to issues. However, we do not believe it is improvement. necessary or appropriate to quantify the 281 The commenter cited Alaska Dep‘t of Envtl. premature death. The commenter extent of this beneficial effect because Conservation v. EPA, 540 U.S. 461, 470, 484 (2004); indicated that SO2 increases asthma Mont. Sulphur & Chem. Co. v. EPA, 666 F.3d 1174, symptoms, leads to increased hospital we are not relying upon health effects in 1181 (9th Cir. 2012) to support the contention that visits, and can form particulates that the promulgation of this rule. Congress structured the CAA to provide expansive Comment: One commenter EPA oversight to ensure SIPs comply with the CAA. aggravate respiratory and heart diseases The commenter cited 42 U.S.C. 7410(c), (k); EME and cause premature death, and that PM (Earthjustice) stated that the RHR rule Homer City Generation, L.P. v. EPA, No. 11–1302, can penetrate deep into the lungs and llF.3dll, 2012 WL 3570721, at *17 (DC Cir. 282 The commenter cited Clean Air Task Force, Aug. 21, 2012) to support the principle that EPA cause health problems such as Death and Disease From Power Plants, http:// must issue a FIP when it determines that a SIP does aggravated asthma, chronic bronchitis, www.catf.us/fossil/problems/power_plants/existing/ not comply with the CAA. and heart attacks. Based on a report map.php?state=Arizona.

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00061 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72572 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

provides important environmental in electric power rates would do to their have a ‘‘severe impact on plant benefits to plants and animals, soil programs in this time of declining state operations’’ or ‘‘result in significant health and entire ecosystems. The support, and one representative of a economic disruption or commenter noted that NOX and SO2 are local, nonprofit hospital similarly unemployment.’’ None of the the primary causes of acid rain, which voiced the difficulty his facility would commenters have provided any acidifies lakes and streams, can damage have in absorbing large rate increases. evidence that our action today would certain types of trees and soils and One commenter discussed the result in the closure of any of the accelerates the decay of building multiplier effect by which loss of affected units. We discuss many of the materials and paints, including income from any job losses or the potential economic impacts raised as irreplaceable buildings and statues that reduction in disposable income due to concerns here in the context of our are part of our nation’s cultural heritage. increased power bills would ripple analysis of affordability of controls to The commenter added that nitrogen through the local economies and affect AEPCO, above. Finally, we acknowledge deposition, caused by wet and dry local businesses and employment. A that today’s action may have positive deposition of nitrates derived from NOX few commenters discussed the impact economic impacts, as described by emissions, causes well-known adverse on Arizona’s water rates, and advised Earthjustice. However, we have not impacts on ecological systems. The EPA to consider how these rate taken potential economic benefits into commenter also noted that NOX is a increases would affect Arizona’s account in our action. precursor to ozone, which impacts economy. A few commenters asserted 3. Comments From Tribal plants and ecosystems by interfering that the proposed rule is intended to Representatives and Members with plants’ ability to produce food and eliminate coal as a cheap and reliable increasing their susceptibility to disease energy source. Comment: One commenter (Navajo and insects, and also contributes to By contrast, one commenter Nation) stated that comments on our wildfires and bark beetle outbreaks in (Earthjustice) stated that the RHR proposed actions were provided the West by depressing plant water provides substantial economic benefits, pursuant to its government-to- levels and growth. which far outweigh the costs of government relationship with EPA. The Response: We appreciate the pollution control technologies such as commenter stated that this EPA commenter’s concerns regarding the SCR. The commenter noted that EPA rulemaking has adverse implications for negative ecosystem impacts of has valued the RHR’s health benefits at a pending BART FIP for Navajo emissions from the units at issue. We $8.4 to $9.8 billion annually. The Generating Station, which is on Navajo agree that both NOX and SO2 cause acid commenter further asserted that Nation land and burns Navajo coal. The rain and can have negative impacts on requiring power plants to invest in commenter also stated that this rule ecosystems, damaging plants, trees, and pollution controls creates short-term could impact BART decisions for Four other vegetation (including crop yields), construction jobs as well as permanent Corners Power Plant, and San Juan which could have a negative effect on operations and management positions. Generating Station. species diversity in our ecosystems. In addition, the commenter indicated The commenter states that EPA has an However, for purposes of this Regional that the national parks and wilderness obligation to consult with Navajo Nation Haze action, we are not authorized to areas protected by the RHR serve as on a government-to-government basis consider these ecosystem impacts. engines for sustainable local capital, for EPA actions and decisions that may Therefore, while we note the potential with national park visitors contributing affect the Navajo Nation’s interests, and for co-benefits to ecosystem health approximately $30 billion to local reminds EPA that it must defer to tribal resulting from our action today, we have economies and supporting 300,000 jobs government policy decisions, just as it not taken these potential benefits into nationwide. Regarding Arizona would a state, when promulgating a FIP account in this action. specifically, the commenter stated that on tribal lands. over 4.3 million people visited the The commenter further states that 2. Comments on Economic Impacts Grand Canyon in 2010, and this EPA has failed to analyze the Comment: Many commenters, supported over 6,800 jobs and resulted cumulative effects of this rulemaking including state officials, private citizens in over $428 million in visitor spending, and the planned and proposed EPA and representatives of local while tourism at Petrified Forest actions on Navajo Generating Station, governments, schools, and business National Park, Saguaro National Park Four Corners Power Plant, and San Juan groups, expressed concern over and Chiricahua National Monument in Generating Station, including both potential economic effects resulting 2010 supported over 1,100 jobs and visibility improvement and potential from EPA’s proposed BART resulted in over $74 million in visitor regional economic impacts. The determinations, asserting that EPA’s spending. The commenter contended commenter noted that the fossil fuel action would result in rate increases and that studies show that national park economy is vitally important to the Four possibly closures of one or more power visitors highly value clean air, readily Corners region and the Navajo Nation, plants. Some commenters cautioned perceive haze and are willing to cut with many jobs and coal royalties at EPA that rate increases would impact at- short visits to national parks based on stake from loss of the area’s coal fired risk populations, such as seniors on their perception of air quality.283 power plants and their associated fixed incomes. The commenters Response: As explained in our prior mines. The commenter states that EPA emphasized that the three plants have a responses regarding economic issues, must consider these impacts, as well as large financial impact on the the BART Guidelines permit the impacts of utility rate increases, in communities where they are located consideration of economic impacts only this BART decision for NOX. (i.e., they provide jobs and tax revenue) under ‘‘unusual circumstances’’ where a The commenter observed that it is and expressed their concern over the potential control option is expected to possible to go forward without imposing three plants’ economic viability if the a FIP in Arizona, as evidenced by the plants are forced to install SCR to 283 The commenter cited and submitted as Exhibit renewed consideration being given to reduce NO emissions. Several 11 Abt Assocs. Inc., Out of Sight: The Science and the New Mexico regional haze SIP X Economics of Visibility Impairment, at ES–7 (2000), representatives of local school districts available at http://www.abtassociates.com/reports/ under the current stay on the proposed discussed the harm that large increases ES-clear.pdf. FIP for that state. The commenter stated

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00062 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72573

that the Navajo Nation, where two effects of today’s rulemaking together governing the power sector. power plants that are undergoing EPA with rulemaking actions on other BART Nevertheless, as discussed elsewhere in BART determinations are located, determinations as part of our action this document, as part of our shares the concerns of Arizona and New today. As noted above, under the CAA, consideration of the affordability of Mexico regarding the economic impacts the RHR and the BART Guidelines, controls on AEPCO, a small entity, we of requiring SCR. The commenter noted BART determinations are made on a have analyzed the potential rate that the BART decision is not based source-by-basis, taking into account the increases associated with our proposal only on the most effective control five statutory factors. The cumulative for Apache Units 2 and 3. Given the measures, but is to be based on an improvements from the various SIPs, uncertainty inherent in such an analysis of five factors which include FIPs, and BART determinations are analysis, we have used conservative non-air quality impacts such as addressed in analyses under the RHR assumptions in an effort to guard against economic impacts. requirements for Reasonable Progress, understating the potential rate impacts. The commenter also asserted that real Long Term Strategies and future updates Regarding the comment that EPA data should underpin EPA’s decisions, to the SIP, which are separate from should not rely on modeling alone, it is rather than modeling alone. The BART analyses. These cumulative extremely difficult in observational commenter also contended that a public improvements will be influenced by analyses to sufficiently control for all health baseline is needed in order to changes in hundreds or thousands of factors, including emissions from other chart any public health improvements emission sources, so are more sources, to be able to isolate the impacts that result from such emission controls. appropriately addressed through use of of closure of a facility. A model such as Response: EPA appreciates the a grid model, such as CAMx or CMAQ, CALPUFF essentially holds constant a comments provided by the Navajo rather than the CALPUFF model number of factors in order to isolate the Nation on our proposed action pursuant recommended in the BART Guidelines, impacts of a single source. As discussed to its government-to-government which is geared to a far lower number elsewhere in this document, EPA relationship with EPA. As part of of sources, and lacks the detailed affirms that the regulatory version of separate rulemakings, EPA has engaged chemistry of the grid models. CALPUFF is the correct model to use for in consultation with Navajo Nation With regard to the economic concerns these BART determinations. regarding the Four Corners Power raised by the commenter, we are Assessing human exposure and Plant 284 and San Juan Generating required by the CAA and the federal quantifying health benefits are outside Station. EPA is currently engaged in regulations implementing the CAA’s the scope of the requirements of the active consultation with the Navajo BART provisions to evaluate (1) cost of Regional Haze Rule. EPA sets National Nation and other affected tribes on the compliance, (2) the energy and non-air Ambient Air Quality Standards Navajo Generating Station. quality environmental impacts of (NAAQS) to establish levels of air Today’s rule approves Arizona’s SIP compliance, (3) any existing pollution quality that are protective of public (in part) and implements a FIP (in part) control technology in use at the source, health, including the health of sensitive for Apache Units 2 and 3; Cholla Units (4) remaining useful life of source, and populations, for a number of pollutants 2, 3 and 4; and Coronado Units 1 and (5) degree of improvement in visibility including particulate matter. These 2. This action has no retroactive effect which may reasonably be anticipated to ‘‘sensitive’’ populations include on final BART determinations for other result from the use of such technology. asthmatics, children, and the elderly. At facilities. We disagree that this action As explained in our prior responses this time the Navajo Nation is not has a nexus to the BART determination regarding economic issues, the BART identified as out of attainment with any for Navajo Generating Station, because Guidelines permit consideration of of the NAAQS. However, EPA BART analyses, whether performed by economic impacts only under ‘‘unusual recognizes that there are significant the states or EPA, are conducted on a circumstances’’ where a potential concerns about risk and exposure to air source-by-source basis, applying all five control option is expected to have a pollutants on the Navajo Nation and statutory factors to a facility on an ‘‘severe impact on plant operations’’ or EPA will continue discussions with the individual basis. While there are certain ‘‘result in significant economic Navajo Nation and will involve other commonalities among the sources disruption or unemployment.’’ None of federal agencies, as appropriate, to help mentioned by the commenter (e.g., all the commenters have provided any address these concerns. are coal-fired power plants), there are evidence that our action today would Comment: Various other also significant differences that result in the closure of any of the representatives and members of the necessarily affect the case-by-case BART affected units or result in significant Hopi and Navajo Tribes provided oral testimony and/or submitted written analysis. For example, the unit size, unit economic disruption. We also note that comments at one or more of the public age, boiler type, existing controls, type none of the sources affected by today’s hearings. Most tribal community of coal burned and proximity to Class I rulemaking currently purchase coal members supported the proposed FIP areas vary significantly among these from a mine that operates on the Navajo and stated their belief that it will sources. All of these differences have a Nation. improve air quality and human health bearing on at least one of the BART We take our duty to estimate the cost in Arizona. Several commenters factors and thus on the ultimate BART of controls very seriously, and make every attempt to make a thoughtful and recounted their personal experiences determination. Given these various well informed determination. However, with the deterioration of visibility in the distinguishing factors, we do not agree we do not consider a potential increase rural areas in which they live, declining that this rule will affect our BART in electricity rates to be the most water supplies due to water use in determination for Navajo Generating appropriate type of analysis for mining operations, and illnesses that Station. they believe are attributable to air We also do not agree that we are considering the costs of compliance in pollution from the power plants and required to consider the cumulative a BART determination. Projections of electricity rate impacts are inherently mines in the area (e.g., asthma and 284 See document titled: ‘‘Timeline of all tribal fraught with uncertainty due to the bronchitis). A number of commenters consultations on BART.docx’’ in the docket for this numerous variables involved and the pointed out that there are numerous old final rulemaking. complexity of the regulatory regime power plants in and around the Navajo

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00063 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72574 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

Nation, which they believe are causing we do not require the BART analysis to device that has not already been air pollution that contributes to haze consider building a natural gas-fired demonstrated in practice.287 and an increase in the incidence of lung electric turbine although the turbine The Guidelines further provide that: and heart disease and cancer in humans, may be inherently less polluting on a In order to provide certainty in the process, 285 as well as harming native plants and per unit basis.’’ Therefore, we did all technologies should be considered if animals. Some of these commenters not consider such alternatives as part of available before the close of the State’s public advocated for a conversion to renewable our BART analyses. Nonetheless, we comment period. You need not consider energy sources, which they believe will acknowledge that many kinds of technologies that become available after this provide jobs, improve health, and renewable energy do not produce haze- date. reduce emissions that contribute to causing pollutants, and transitioning to The commenter has not provided climate change. One commenter those sources of energy could lead to evidence that this technology has been specifically suggested that EPA promote visibility improvements. demonstrated in practice or that it was alternatives like natural gas and algae The CAA applies equally to all parts available before the close of the State’s ponds as a source of energy. of the United States. In making a public comment period. Therefore, we One commenter indicated that determination in this case, we have have not considered it as a potential reduced haze would improve tourism, applied the applicable provisions of the control option. An additional resulting in increased jobs and tax CAA and the RHR. We have also consideration is that typically 90 receipts. Another tribal commenter considered other applicable percent of the NOX from combustion is stated that before acting, EPA should requirements, including Executive emitted in the form of NO, rather than evaluate the impact on employment and Order 12898,286 which establishes NO2. Since the boiling point of NO is on the Hopi’s revenue from coal if the federal executive policy on 121 K or ¥242 degrees F, much lower FIP causes power plants to close. environmental justice. This Executive than for NO2, and the stack exit One tribal commenter alleged that the Order directs federal agencies, to the temperature is the range of 300–400 K National Academy of Sciences did a greatest extent practicable and or 120–280 degrees F, a large degree of study a number of years ago that permitted by law, to make cooling would be necessary to condense concluded that some areas of the environmental justice part of their the NO, and so the energy costs could country could be designated as mission by identifying and addressing, be substantial. ‘‘national sacrifice areas’’ that would be as appropriate, disproportionately high used for national priorities, irrespective 4. Requests for Extension of Comment of resulting permanent environmental and adverse human health or Period and Additional Hearings damages. According to the commenter, environmental effects of their programs, policies, and activities on minority Comment: A number of commenters many Indian reservations are located in remarked on EPA’s timeline for such areas, such as all of the Navajo and populations and low-income populations in the United States. soliciting public comments, and stated Hopi reservations. The commenter that they believe that the time allowed EPA has determined that our final asserted that the study concluded that was insufficient. One commenter rule will not have disproportionately the well-being of the people in such requested more public hearings, and high and adverse human health or areas can be forfeited so that the rest of another commenter requested a 90-day environmental effects on minority or the country can enjoy cheap energy. extension of the deadline for comments low-income populations because it Response: EPA acknowledges the (starting from July 18, 2012), so that the increases the level of environmental comments. Neither Section 169A of the public has ample time to review, protection for all affected populations CAA nor the BART Guidelines requires analyze, comment, and react to the rule without having any disproportionately that BART analyses include or quantify and in particular EPA’s Technical high and adverse human health or benefits to health or tourism or impact Support Document. The commenter environmental effects on any on employment. EPA does not intend added that an extension would allow population, including any minority or for this action to cause any power plants the ADEQ the opportunity to further low-income population. This rule to close. Although a quantitative collaborate with EPA in revising the requires emissions reductions of NOx analysis of the health and tourism state’s SIP submittal (for the purpose of from three facilities in Arizona. The benefits is beyond the scope of what is nullifying the proposed FIP), and partial approval of the SIP approves required under BART EPA agrees with thereby adhering to the intent of the state law as meeting Federal commenters that emission reductions CAA. achieved to improve visibility will also requirements. Response: As explained above, our improve air quality. Improved air Comment: One commenter suggested proposed rule, which was signed on quality, in turn, affects public health that EPA investigate the technology of July 2, 2012 and published in the and may enhance tourism in the area. cooling steam exhaust through a Federal Register on July 20, 2012,288 EPA notes that even if we had magnetic refrigerator to remove NO2 as provided for a public hearing in quantified the benefits to health and a liquid, since it would condense at the Phoenix, Arizona, on July 31, 2012, and tourism, such an analysis would not relatively high temperature of 294 K or a public comment deadline of August likely have altered the outcome of our 70 degrees F (boiling point). 31, 2012. In response to requests from BART determination. Response: The BART Guidelines various parties for a longer comment Renewable energy technology is not a provide that: period and additional hearings, we retrofit option for the sources subject to extended the public comment period to BART and is therefore outside the scope Technologies which have not yet been applied to (or permitted for) full scale a total of sixty days from publication in of our BART determination. As noted in operations need not be considered as the Federal Register.289 We also the BART Guidelines, ‘‘[w]e do not available; we do not expect the source owner scheduled two more public hearings in consider BART as a requirement to to purchase or construct a process or control redesign the source when considering 287 BART Guidelines, 40 CFR Part 51, Appendix available control alternatives. For 285 BART Guidelines, 40 CFR Part 51, Appendix Y, section IV.D.1. example, where the source subject to Y, section IV.D.1. 288 77 FR 42834. BART is a coal-fired electric generator, 286 59 FR 7629, February 16, 1994. 289 See 77 FR 45326 (July 31, 2012).

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00064 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72575

Southern Arizona (Benson) and in V. Summary of Final Action compliance deadlines and requirements Northern Arizona (Holbrook) on August for equipment maintenance and 14 and 15, 2012, respectively. EPA is taking final action to approve operation, including monitoring, in part and disapprove in part a portion recordkeeping and reporting Comment: One comment letter signed of Arizona’s SIP for Regional Haze and by 728 residents, business owners, requirements for all pollutants at all of to promulgate a FIP for the disapproved the BART units. As a result, we find that citizens and other interested parties elements of the SIP. This final action this final disapproval is the only path urged EPA to extend the comment addresses only the State’s BART that is consistent with the Act at this period on our proposal and provide determinations for the specified units at time. additional hearings near the Cholla the three power plants. We will propose EPA estimates this action will Power Plant. action on the remainder of Arizona’s improve visibility at 18 Class I areas by Response: As noted the preceding Regional Haze SIP in a separate notice. reducing NOX emissions from three response, we extended the comment EPA takes very seriously a decision to power plants by about 22,700 tons per period on our propose rule and we held disapprove portions of a state plan. In year. The total costs associated with additional public hearings, including this instance, we find that the State’s these reductions, according to the one in Holbrook, Arizona, near the NOX BART determinations for the coal- supplemental cost analysis we Cholla Power Plant. fired units are not consistent with the performed based on cost estimates requirements of the Act and the RHR. In provided by the facility owners, are addition, the SIP lacks the necessary summarized in Table 18.

TABLE 18—SUMMARY OF SUPPLEMENTAL COST ANALYSIS

Annualized cap- Total Capital cost ital cost Annual O&M annualized cost Cost- ($) ($/yr) ($/yr) ($/yr) effectiveness

Apache Unit 2 ...... $82,481,439 $7,785,664 $1,760,600 $9,546,264 $3,450 Apache Unit 3 ...... 82,481,439 7,785,664 1,760,600 9,546,264 2,973 Cholla Unit 2 ...... 87,713,386 8,279,523 1,626,683 9,906,206 2,979 Cholla Unit 3 ...... 83,461,195 7,878,146 1,570,766 9,448,912 2,838 Cholla Unit 4 ...... 119,083,832 11,240,671 2,350,182 13,590,853 3,083 Coronado Unit 1 ...... 80,633,219 7,611,205 4,492,736 12,103,941 2,135 Coronado Unit 2 ...... 2,500,000 235,982 ...... 235,982 1,900

VI. Statutory and Executive Order displays a currently valid Office of portion of the Arizona SIP and a Reviews Management and Budget (OMB) control Regional Haze FIP for units at three number. The OMB control numbers for electric generating facilities in Arizona. A. Executive Order 12866: Regulatory Planning and Review our regulations in 40 CFR are listed in Firms primarily engaged in the 40 CFR part 9. generation, transmission, and/or This action finalizes approval of a distribution of electric energy for sale source-specific portion of the Arizona C. Regulatory Flexibility Act are small if, including affiliates, the total SIP and a Regional Haze FIP for units The Regulatory Flexibility Act (RFA) electric output for the preceding fiscal at three facilities in Arizona. This action year did not exceed 4 million megawatt is not a rule of general applicability, and generally requires an agency to prepare a regulatory flexibility analysis of any hours. Only one of the three facilities not a ‘‘significant regulatory action’’ affected by this action is a small entity: under the terms of Executive Order rule subject to notice and comment rulemaking requirements under the AEPCO sold under 3 million megawatt 12866 (58 FR 51735, October 4, 1993). hours in 2011. This type of action is exempt from Administrative Procedure Act or any review under Executive Order (EO) other statute unless the agency certifies Although a regulatory flexibility 12866 (58 FR 51735, October 4, 1993) that the rule will not have a significant analysis as specified by the RFA is not and is therefore not subject to review economic impact on a substantial required when a rule has impact on only under Executive Order 13563 (76 FR number of small entities. Small entities one small entity, EPA estimated the 3821, January 21, 2011). include small businesses, small potential impact to AEPCO of our organizations, and small governmental proposal to require SCR in AEPCO’s B. Paperwork Reduction Act jurisdictions. For purposes of assessing Units 1 and 2. EPA also requested This action does not impose an the impacts of today’s rule on small information from AEPCO on the information collection burden under the entities, small entity is defined as: (1) A economics of operating Apache provisions of the Paperwork Reduction small business as defined by the Small Generating Station and what impact the Act, 44 U.S.C. 3501 et seq. Burden is Business Administration’s (SBA) installation of SCR may have on the defined at 5 CFR 1320.3(b). Because this regulations at 13 CFR 121.201; (2) a economics of operating Apache action will finalize approval of a source- small governmental jurisdiction that is a Generating Station. A summary of the specific portion of the Arizona SIP and government of a city, county, town, comments regarding the impact of this a Regional Haze FIP for units at only school district or special district with a action on AEPCO, and EPA’s response three facilities in Arizona, the population of less than 50,000; and (3) to those concerns, is provided in section Paperwork Reduction Act does not a small organization that is any not-for- I.V. of this preamble. After considering apply. See 5 CFR 1320.3(c). An agency profit enterprise which is independently the economic impacts of this action on may not conduct or sponsor, and a owned and operated and is not small entities, I certify that this action person is not required to respond to a dominant in its field. This action will not have a significant economic collection of information unless it finalizes approval of a source-specific impact on a substantial number of small

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00065 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72576 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

entities. The FIP for the three Arizona contain a Federal mandate that may response to Navajo Nation is provided facilities being issued today does not result in expenditures that exceed the in section I.V. of this preamble. impose new requirements on a inflation-adjusted UMRA threshold of G. Executive Order 13045: Protection of substantial number of small entities $100 million (in 1996 dollars) by State, Children From Environmental Health because one significantly impacted local, or Tribal governments or the Risks and Safety Risks small entity is not a ‘‘substantial’’ private sector in any 1 year. In addition, number. Finalizing approval of a source- this rule does not contain a significant Executive Order 13045: Protection of specific portion of the Arizona Regional Federal intergovernmental mandate as Children from Environmental Health Haze SIP merely approves state law as described by section 203 of UMRA nor Risks and Safety Risks (62 FR 19885, meeting Federal requirements and does it contain any regulatory April 23, 1997), applies to any rule that: imposes no additional requirements requirements that might significantly or (1) Is determined to be economically beyond those imposed by state law. See uniquely affect small governments. significant as defined under Executive Mid-Tex Electric Cooperative, Inc. v. Order 12866; and (2) concerns an FERC, 773 F.2d 327 (D.C. Cir. 1985). E. Executive Order 13132: Federalism environmental health or safety risk that This action does not have federalism we have reason to believe may have a D. Unfunded Mandates Reform Act disproportionate effect on children. EPA (UMRA) implications. It will not have substantial direct effects on the States, on the interprets EO 13045 as applying only to Unfunded Mandates Reform Act of relationship between the national those regulatory actions that concern health or safety risks, such that the 1995 (UMRA), Public Law 104–4, government and the States, or on the establishes requirements for Federal analysis required under section 5–501 of distribution of power and agencies to assess the effects of their the EO has the potential to influence the responsibilities among the various regulatory actions on State, local, and regulation. This action is not subject to levels of government, as specified in Tribal governments and the private EO 13045 because it implements Executive Order 13132, because it sector. Under section 202 of UMRA, specific standards established by addresses the State not fully meeting its EPA generally must prepare a written Congress in statutes. Also, because this obligation to protect visibility statement, including a cost-benefit action only applies to three sources and established in the CAA and this final analysis, for proposed and final rules is not a rule of general applicability, it action will reduce the emissions of NO with ‘‘Federal mandates’’ that may X is not economically significant as from three facilities in Arizona. Thus, result in expenditures to State, local, defined under Executive Order 12866, Executive Order 13132 does not apply and Tribal governments, in the and the rule also does not have a to this action. Although section 6 of aggregate, or to the private sector, of disproportionate effect on children. Executive Order 13132 does not apply $100 million or more (adjusted for However, to the extent this action will to this action, a summary of the inflation) in any one year. Before limit emissions of NO , SO , and PM , concerns raised by State and local X 2 10 promulgating an EPA rule for which a the rule will have a beneficial effect on officials, and EPA’s response to those written statement is needed, section 205 children’s health by reducing air concerns is provided in section I.V. of of UMRA generally requires EPA to pollution that causes or exacerbates this preamble. identify and consider a reasonable childhood asthma and other respiratory number of regulatory alternatives and to F. Executive Order 13175: Consultation issues. adopt the least costly, most cost- and Coordination With Indian Tribal H. Executive Order 13211: Actions effective, or least burdensome Governments alternative that achieves the objectives Concerning Regulations That of the rule. The provisions of section Subject to the Executive Order 13175 Significantly Affect Energy Supply, 205 of UMRA do not apply when they (65 FR 67249, November 9, 2000) EPA Distribution, or Use are inconsistent with applicable law. may not issue a regulation that has tribal This action is not subject to Executive Moreover, section 205 of UMRA allows implications, that imposes substantial Order 13211 (66 FR 28355 (May 22, EPA to adopt an alternative other than direct compliance costs, and that is not 2001)), because it is not a significant the least costly, most cost-effective, or required by statute, unless the Federal regulatory action under Executive Order least burdensome alternative if the government provides the funds 12866. Administrator publishes with the final necessary to pay the direct compliance I. National Technology Transfer and rule an explanation why that alternative costs incurred by tribal governments, or Advancement Act was not adopted. Before EPA establishes EPA consults with tribal officials early any regulatory requirements that may in the process of developing the Section 12(d) of the National significantly or uniquely affect small proposed regulation and develops a Technology Transfer and Advancement governments, including Tribal tribal summary impact statement. We Act of 1995 (NTTAA), Public Law 104– governments, it must have developed believe this rule does not have tribal 113, 12 (10) (15 U.S.C. 272 note) directs under section 203 of UMRA a small implications, as specified in Executive EPA to use voluntary consensus government agency plan. The plan must Order 13175, and will not have standards (VCS) in its regulatory provide for notifying potentially substantial direct effects on tribal activities unless to do so would be affected small governments, enabling governments. Thus, Executive Order inconsistent with applicable law or officials of affected small governments 13175 does not apply to this rule. otherwise impractical. VCS are to have meaningful and timely input in However, in our proposal we requested technical standards (e.g., materials the development of EPA regulatory comment on our proposed rule from specifications, test methods, sampling proposals with significant Federal tribal officials. The Navajo Nation procedures and business practices) that intergovernmental mandates, and Environmental Protection Agency are developed or adopted by the VCS informing, educating, and advising provided comments on our proposed bodies. The NTTAA directs EPA to small governments on compliance with rule, both orally at a public hearing and provide Congress, through annual the regulatory requirements. by letter, which EPA considered in reports to OMB, with explanations Under Title II of UMRA, EPA has developing this final rule. EPA’s when the Agency decides not to use determined that this rule does not summary of these comments and our available and applicable VCS. The

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00066 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72577

rulemaking involves technical parties. 5 U.S.C. 804(3). EPA is not (i) Table 1.1—NOX BART, entry for standards. Therefore, the Agency required to submit a rule report AEPCO [Apache], ST1 [Unit 1] only. conducted a search to identify regarding today’s action under section (ii) Table 1.2—PM10 BART, entries for potentially applicable voluntary 801 because this is a rule of particular AEPCO [Apache], APS Cholla Power consensus standards. However, we applicability and only applies to three Plant and SRP Coronado Generating identified no such standards, and none facilities. Station. were brought to our attention in L. Petitions for Judicial Review (iii) Table 1.3—SO2 BART, entries for comments. Therefore, EPA has decided AEPCO, APS Cholla Power Plant and to use 40 CFR Part 60 Appendix A Under section 307(b)(1) of the Clean SRP Coronado Generating Station. Method 5, 40 CFR Part 51 Appendix M Air Act, petitions for judicial review of ■ 3. Section 52.145 is amended by Methods 201A/202, 40 CFR Part 60 this action must be filed in the United adding paragraphs (e) and (f) to read as Appendix A Method 19, and 40 CFR States Court of Appeals for the follows: Part 75. appropriate circuit by February 4, 2013. Filing a petition for reconsideration by § 52.145 Visibility protection. J. Executive Order 12898: Federal the Administrator of this final rule does Actions To Address Environmental * * * * * not affect the finality of this rule for the Justice in Minority Populations and (e) Approval. On February 28, 2011, purposes of judicial review nor does it Low-Income Populations the Arizona Department of extend the time within which a petition Environmental Quality submitted the Executive Order 12898 (59 FR 7629, for judicial review may be filed, and ‘‘Arizona State Implementation Plan, February 16, 1994), establishes federal shall not postpone the effectiveness of Regional Haze Under Section 308 of the executive policy on environmental such rule or action. This action may not Federal Regional Haze Rule’’ (‘‘Arizona justice. Its main provision directs be challenged later in proceedings to Regional Haze SIP’’). federal agencies, to the greatest extent enforce its requirements. (See CAA (1) With the exception of the NOX practicable and permitted by law, to section 307(b)(2).) make environmental justice part of their BART determinations for Units ST2 and mission by identifying and addressing, List of Subjects in 40 CFR Part 52 ST3 at AEPCO Apache Generating Station; Units 2, 3, and 4 at APS Cholla as appropriate, disproportionately high Environmental protection, Air Power Plant; and Units 1 and 2 at SRP and adverse human health or pollution control, Incorporation by Coronado Generating Station, and the environmental effects of their programs, reference, Intergovernmental relations, BART compliance provisions for all policies, and activities on minority Nitrogen oxides, Sulfur dioxide, BART emissions limits at the eight units populations and low-income Particulate matter, Reporting and at the three power plants, the BART populations in the United States. EPA recordkeeping requirements, Visibility, determinations for AEPCO Apache has determined that this final rule will Volatile organic compounds. not have disproportionately high and Generating Station, APS Cholla Power Dated: November 15, 2012. adverse human health or environmental Plant, and SRP Coronado Generating effects on minority or low-income Lisa P. Jackson, Station in the Arizona Regional Haze populations because it increases the Administrator. SIP meet the applicable requirements of level of environmental protection for all Part 52, chapter I, title 40 of the Code Clean Air Act sections 169A and 169B affected populations without having any of Federal Regulations is amended as and the Regional Haze Rule in 40 CFR disproportionately high and adverse follows: 51.301 through 51.308. human health or environmental effects (f) Source-specific federal on any population, including any PART 52—APPROVAL AND implementation plan for regional haze minority or low-income population. PROMULGATION OF at Apache Generating Station, Cholla This rule requires emissions reductions IMPLEMENTATION PLANS Power Plant, and Coronado Generating Station — (1) Applicability. This of NOX from three facilities in Arizona. ■ 1. The authority citation for Part 52 paragraph (f) applies to each owner/ The partial approval of the SIP merely continues to read as follows: approves state law as meeting Federal operator of the following coal-fired requirements and imposes no additional Authority: 42 U.S.C. 7401 et seq. electricity generating units (EGUs) in requirements beyond those imposed by the state of Arizona: Apache Generating Subpart D—Arizona state law. Station, Units 2 and 3; Cholla Power Plant, Units 2, 3, and 4; and Coronado ■ 2. Section 52.120 is amended by K. Congressional Review Act Generating Station, Units 1 and 2. This adding paragraph (c)(154) to read as The Congressional Review Act, 5 paragraph (f) also applies to each follows: U.S.C. 801 et seq., as added by the Small owner/operator of the following natural Business Regulatory Enforcement § 52.120 Identification of plan. gas-fired EGUs in the state of Arizona: Fairness Act of 1996, generally provides * * * * * Apache Generating Station Unit 1. The that before a rule may take effect, the (c) * * * provisions of this paragraph (f) are agency promulgating the rule must (154) The following plan was severable, and if any provision of this submit a rule report, which includes a submitted February 28, 2011, by the paragraph (f), or the application of any copy of the rule, to each House of the Governor’s designee. provision of this paragraph (f) to any Congress and to the Comptroller General (i) [Reserved] owner/operator or circumstance, is held of the United States. Section 804 (ii) Additional materials. invalid, the application of such exempts from section 801 the following (A) Arizona Department of provision to other owner/operators and types of rules (1) rules of particular Environmental Quality. other circumstances, and the remainder applicability; (2) rules relating to agency (1) Arizona State Implementation of this paragraph (f), shall not be management or personnel; and (3) rules Plan, Regional Haze Under Section 308 affected thereby. of agency organization, procedure, or of the Federal Regional Haze Rule: (2) Definitions. Terms not defined practice that do not substantially affect Appendix D, Arizona BART— below shall have the meaning given to the rights or obligations of non-agency Supplemental Information: them in the Clean Air Act or EPA’s

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00067 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72578 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

regulations implementing the Clean Air Generating Station; and Units 2, 3, and units. Each emission limit shall be Act. For purposes of this paragraph (f): 4 for Cholla Power Plant. based on a rolling 30-boiler-operating- ADEQ means the Arizona Department lb means pound(s). day average, unless otherwise indicated of Environmental Quality. NOX means nitrogen oxides expressed in specific paragraphs. Boiler-operating day means a 24-hour as nitrogen dioxide (NO2). period between 12 midnight and the Owner(s)/operator(s) means any Federal following midnight during which any person(s) who own(s) or who operate(s), Group of coal-fired units emission limitation fuel is combusted at any time in the control(s), or supervise(s) one or more of unit. the units identified in paragraph (f)(1) of Apache Generating Station Coal-fired unit means any of the EGUs this section. Units 2 and 3 ...... 0.070 identified in paragraph (f)(1) of this MMBtu means million British thermal Cholla Power Plant Units 2, section, except for Apache Generating unit(s). 3, and 4 ...... 0.055 Station, Unit 1. Operating hour means any hour that Coronado Generating Station Continuous emission monitoring fossil fuel is fired in the unit. Units 1and 2 ...... 0.065 system or CEMS means the equipment PM10 means filterable total particulate required by 40 CFR Part 75 and this matter less than 10 microns and the (ii) SO2 removal efficiency paragraph (f). condensable material in the impingers requirement. The owners/operators of Emissions limitation or emissions as measured by Methods 201A and 202. Cholla Power Plant Units 2, 3, and 4 limit means any of the Federal Emission Regional Administrator means the shall achieve and maintain a 30-day Limitations required by this paragraph Regional Administrator of EPA Region rolling average SO2 removal efficiency (f) or any of the applicable PM10 and IX or his/her authorized representative. of 95 percent at each unit. SO2 emissions limits for Apache SO2 means sulfur dioxide. (4) Compliance dates. (i) The owners/ Generating Station, Cholla Power Plant, SO4 removal efficiency means the operators of each unit subject to this and Coronado Generating Station quantity of SO2 removed as calculated paragraph (f) shall comply with the NOX submitted to EPA as part of the Arizona by the procedure in paragraph emissions limitations and other NOX- Regional Haze SIP in a letter dated (f)(5)(iii)(B) of this section. related requirements of this paragraph February 28, 2011, and approved into Unit means any of the EGUs identified (f) no later than December 5, 2017. the Arizona State Implementation Plan in paragraph (f)(1) of this section. (ii) The owners/operators of each unit on December 5, 2012. Valid data means data recorded when subject to this paragraph (f) shall Flue Gas Desulfurization System or the CEMS is not out-of-control as comply with the applicable PM10 and FGD means a pollution control device defined by Part 75. SO2 emissions limits submitted to EPA that employs flue gas desulfurization (3) Federal emission limitations.—(i) as part of the Arizona Regional Haze SIP technology, including an absorber NOX emission limitations. The owner/ in a letter dated February 28, 2011, and utilizing lime, fly ash, or limestone operator of each coal-fired unit subject approved into the Arizona State slurry, for the reduction of sulfur to this paragraph (f) shall not emit or Implementation Plan on December 5, dioxide emissions. cause to be emitted NOX in excess of the 2012, as well as the related compliance, Group of coal-fired units mean Units following limitations, in pounds per recordkeeping and reporting of this 1 and 2 for Coronado Generating million British thermal units (lb/ paragraph (f) no later than the following Station; Units 2 and 3 for Apache MMBtu) from any group of coal-fired dates:

Compliance date Unit PM10 SO2

Apache Generating Station, Unit 1 ...... June 3, 2013 ...... June 3, 2013. Apache Generating Station, Unit 2 ...... December 5, 2016 ...... December 5, 2016. Apache Generating Station, Unit 3 ...... December 5, 2016 ...... December 5, 2016. Cholla Power Plant, Unit 2 ...... April 1, 2016 ...... April 1, 2016. Cholla Power Plant, Unit 3 ...... June 3, 2013 ...... June 3, 2013. Cholla Power Plant, Unit 4 ...... June 3, 2013 ...... June 3, 2013. Coronado Generating Station, Unit 1 ...... June 3, 2013 ...... June 3, 2013. Coronado Generating Station, Unit 2 ...... June 3, 2013 ...... June 3, 2013.

(iii) The owners/operators of Cholla (5) Compliance determinations for the requirements found at 40 CFR Part Power Plant Units 2, 3 and 4 shall NOX and SO4—(i) Continuous emission 75, to accurately measure SO2 emissions comply with the SO2 removal efficiency monitoring system. and diluent at the inlet of the sulfur requirement in paragraph (f)(5)(iii)(B) of (A) At all times after the compliance dioxide control device. Apache Unit 1 this section all related compliance, date specified in paragraph (f)(4) of this NOX and diluent CEMs shall be recordkeeping and reporting section, the owner/operator of each operated to meet the requirements of requirements no later than the following coal-fired unit shall maintain, calibrate, Part 75. All valid CEMS hourly data dates: and operate a CEMS, in full compliance shall be used to determine compliance with the requirements found at 40 CFR with the emission limitations for NOX Cholla Power Plant, April 1, 2016. Part 75, to accurately measure SO2, and SO2 in paragraph (f)(3) of this Unit 2. NOX, diluent, and stack gas volumetric section for each unit. When the CEMS Cholla Power Plant, December 5, 2013. flow rate from each unit. In addition, is out-of-control as defined by Part 75, Unit 3. the owner/operator of Cholla Units 2, 3, that CEMs data shall be treated as Cholla Power Plant, December 5, 2013. missing data and not used to calculate Unit 4. and 4 shall calibrate, maintain, and operate a CEMS, in full compliance with the emission average. Each required

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00068 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations 72579

CEMS must obtain valid data for at least four for each group of coal-fired units, (B) The 30-day rolling average SO2 90 percent of the unit operating hours, to calculate the 30-day rolling average removal efficiency for Cholla Units 2, 3, on an annual basis. NOX emission rate for each group of and 4 shall be calculated as follows: (B) The owner/operator of each unit coal-fired units, in pounds of NOX per Step one, sum the total pounds of SO2 shall comply with the quality assurance MMBtu, for each calendar day. Each 30- emitted as measured at the outlet of the procedures for CEMS found in 40 CFR day rolling average NOX emission rate FGD system for the unit during the Part 75. In addition to these Part 75 shall include all emissions and all heat current boiler-operating day and the requirements, relative accuracy test input that occur during all periods previous twenty-nine (29) boiler- audits shall be calculated for both the within any boiler-operating day, operating days as measured at the outlet NOX and SO2 pounds per hour including emissions from startup, of the FGD system for that unit; step measurement and the heat input shutdown, and malfunction. two, sum the total pounds of SO2 measurement. The CEMs monitoring (B) The 30-day rolling average NOX delivered to the inlet of the FGD system data shall not be bias adjusted. The inlet emission rate for Apache Unit 1 shall be for the unit during the current boiler- SO2 and diluent monitors required by calculated in accordance with the operating day and the previous twenty- this rule shall also meet the Quality following procedure: step one, sum the nine (29) boiler-operating days as Assurance/Quality Control (QA/QC) total pounds of NOX emitted from the measured at the inlet to the FGD system requirements of Part 75. The testing and unit during the current boiler-operating for that unit (for each hour, the total evaluation of the inlet monitors and the day and the previous twenty-nine (29) pounds of SO2 delivered to the inlet of calculations of relative accuracy for lb/ boiler-operating days; step two, sum the the FGD system for a unit shall be hr of NOX, SO2 and heat input shall be total heat input to the unit in MMBtu calculated by measuring the ratio of the performed each time the Part 75 CEMS during the current boiler-operating day lb/MMBtu SO2 inlet to the lb/MMBtu undergo relative accuracy testing. In and the previous twenty-nine (29) SO2 outlet and multiplying the outlet addition, relative accuracy test audits boiler-operating days; and step three, pounds of SO2 by that ratio); step three, shall be performed in the units of lb/ divide the total number of pounds of subtract the outlet SO2 emissions MMBtu for the inlet and outlet SO2 calculated in step one from the inlet SO2 NOX emitted during the thirty (30) monitors at Cholla Units 2, 3, and 4. boiler-operating days by the total heat emissions calculated in step two; step Heat input for Apache Unit 1 shall be input during the thirty (30) boiler- four, divide the remainder calculated in measured in accordance with Part 75 operating days. A new 30-day rolling step three by the inlet SO2 emissions fuel gas measurement procedures found calculated in step two; and step five, average NOX emission rate shall be in 40 CFR Part 75, Appendix D. calculated for each new boiler-operating multiply the quotient calculated in step (ii) Compliance determinations for four by 100 to express as a percentage day. Each 30-day rolling average NOX NOX. (A) The 30-day rolling average emission rate shall include all emissions removal efficiency. A new 30-day NO emission rate for each group of X and all heat input that occur during all rolling average SO2 removal efficiency coal-fired units shall be calculated for periods within any boiler-operating day, shall be calculated for each new boiler- each calendar day, even if a unit is not operating day, and shall include all in operation on that calendar day, in including emissions from startup, shutdown, and malfunction. emissions that occur during all periods accordance with the following within each boiler-operating day, (C) If a valid NO pounds per hour or procedure: step one, for each unit, sum X including emissions from startup, heat input is not available for any hour the hourly pounds of NOX emitted shutdown, and malfunction. for a unit, that heat input and NOX during the current boiler-operating day (C) If a valid SO2 pounds per hour at (or most recent boiler-operating day if pounds per hour shall not be used in the the outlet of the FGD system or heat the unit is not in operation), and the calculation of the 30-day rolling input is not available for any hour for average. preceding twenty-nine (29) boiler- a unit, that heat input and SO2 pounds operating days, to calculate the total (iii) Compliance determinations for per hour shall not be used in the pounds of NOX emitted over the most SO2. (A) The 30-day rolling average SO2 calculation of the 30-day rolling recent thirty (30) boiler-operating day emission rate for each coal-fired unit average. period for each coal-fired unit; step two, shall be calculated in accordance with (D) If both a valid inlet and outlet SO2 for each unit, sum the hourly heat input, the following procedure: Step one, sum lb/MMBtu and an outlet value of lb/hr in MMBtu, during the current boiler- the total pounds of SO2 emitted from the of SO2 are not available for any hour, operating day (or most recent boiler- unit during the current boiler-operating that hour shall not be included in the operating day if the unit is not in day and the previous twenty-nine (29) efficiency calculation. operation), and the preceding twenty- boiler-operating days; step two, sum the (6) Compliance determinations for nine (29) boiler-operating days, to total heat input to the unit in MMBtu particulate matter. Compliance with the calculate the total heat input, in during the current boiler-operating day particulate matter emission limitation MMBtu, over the most recent thirty (30) and the previous twenty-nine (29) for each coal-fired unit shall be boiler-operating day period for each boiler-operating day; and step three, determined from annual performance coal-fired unit; step 3, sum together the divide the total number of pounds of stack tests. Within sixty (60) days of the total pounds of NOX emitted from the SO2 emitted during the thirty (30) compliance deadline specified in group of coal-fired units over each unit’s boiler-operating days by the total heat paragraph (f)(4) of this section, and on most recent thirty (30) boiler-operating input during the thirty (30) boiler- at least an annual basis thereafter, the day period (the most recent 30 boiler- operating days. A new 30-day rolling owner/operator of each unit shall operating day periods for different units average SO2 emission rate shall be conduct a stack test on each unit to may be different); step four, sum calculated for each new boiler-operating measure PM10 using EPA Method 5, in together the total heat input from the day. Each 30-day rolling average SO2 40 CFR part 60, Appendix A, or Method group of coal-fired units over each unit’s emission rate shall include all emissions 201A/202 in 40 CFR Part 51, Appendix most recent thirty (30) boiler-operating and all heat input that occur during all M. A test protocol shall be submitted to day period; and step five, divide the periods within any boiler-operating day, EPA and ADEQ a minimum of 30 days total pounds of NOX emitted from step including emissions from startup, prior to the scheduled testing. The three by the total heat input from step shutdown, and malfunction. protocol shall identify which method(s)

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00069 Fmt 4701 Sfmt 4700 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with 72580 Federal Register / Vol. 77, No. 234 / Wednesday, December 5, 2012 / Rules and Regulations

will be used to demonstrate compliance. measurement and hourly heat input appropriate performance or compliance Each test shall consist of three runs, measurement. test had been performed, can be used to with each run at least 120 minutes in (v) Records of all major maintenance establish whether or not the owner or duration and each run collecting a activities conducted on emission units, operator has violated or is in violation minimum sample of 60 dry standard air pollution control equipment, and of any standard or applicable emission cubic feet. Results shall be reported in CEMS. limit in the plan. lb/MMBtu using the calculation in 40 (vi) Any other records required by 40 (10) Equipment operations. At all CFR Part 60 Appendix A Method 19. In CFR Part 75. times, including periods of startup, addition to annual stack tests, the (8) Reporting. All reports and shutdown, and malfunction, the owner owner/operator shall monitor notifications under this paragraph (f) particulate emissions for compliance shall be submitted to the Director of or operator shall, to the extent with the emission limitations in Enforcement Division, U.S. EPA Region practicable, maintain and operate the accordance with the applicable IX, at 75 Hawthorne Street, San unit including associated air pollution Compliance Assurance Monitoring Francisco, CA 94105. control equipment in a manner (CAM) plan developed and approved in (i) The owner/operator shall notify consistent with good air pollution accordance with 40 CFR Part 64. The EPA within two weeks after completion control practices for minimizing averaging time for any other of installation of combustion controls or emissions. Pollution control equipment shall be designed and capable of demonstration of the PM10 compliance Selective Catalytic Reactors on any of or exceedance shall be based on a 6- the units subject to this section. operating properly to minimize hour average. (ii) Within 30 days after the emissions during all expected operating (7) Recordkeeping. The owner or applicable compliance date(s) in conditions. Determination of whether operator of each unit shall maintain the paragraph (f)(4) of this section and acceptable operating and maintenance following records for at least five (5) within 30 days of every second calendar procedures are being used will be based years: quarter thereafter (i.e., semi-annually), on information available to the Regional (i) All CEMS data, including the date, the owner/operator of each unit shall Administrator which may include, but place, and time of sampling or submit a report that lists the daily 30- is not limited to, monitoring results, measurement; parameters sampled or day rolling emission rates for NOX and review of operating and maintenance measured; and results. SO2 for each unit and, for Cholla Units procedures, and inspection of the unit. 2, 3, and 4, the SO removal efficiency, (ii) Daily 30-day rolling emission rates 2 (11) Affirmative defense for for NO and SO and SO removal calculated in accordance with paragraph X 2 2 malfunctions. The following regulations efficiency, when applicable, for each (f)(5) of this section. Included in this are incorporated by reference and made unit, calculated in accordance with report shall be the results of any relative part of this federal implementation plan: paragraph (f)(5) of this section. accuracy test audit performed during (iii) Records of quality assurance and the two preceding calendar quarters. (i) R–18–2–101, paragraph 65; quality control activities for emissions (9) Enforcement. Notwithstanding any (ii) R18–2–310, sections (A), (B), (D) measuring systems including, but not other provision in this implementation and (E) only; and limited to, any records required by 40 plan, any credible evidence or (iii) R18–2–310.01. CFR Part 75. information relevant as to whether the (iv) Records of the relative accuracy unit would have been in compliance [FR Doc. 2012–28565 Filed 12–4–12; 8:45 am] test for hourly NOX and SO2 lb/hr with applicable requirements if the BILLING CODE 6560–50–P

VerDate Mar<15>2010 18:49 Dec 04, 2012 Jkt 229001 PO 00000 Frm 00070 Fmt 4701 Sfmt 9990 E:\FR\FM\05DER3.SGM 05DER3 srobinson on DSK4SPTVN1PROD with