Promulgation of Air Quality Implementation Plans; State of Arkansas; Regional Haze and Interstate Visibility Transport Federal Implementation Plan

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 52

[EPA-R06-OAR-2015-0189; FRL-9924-85-Region 6]

Promulgation of Air Quality Implementation Plans; State of Arkansas; Regional Haze and Interstate Visibility Transport Federal Implementation Plan

AGENCY:

Environmental Protection Agency (EPA).

ACTION:

Proposed rule.

SUMMARY:

The Environmental Protection Agency (EPA) is proposing to promulgate a Federal Implementation Plan (FIP) to address certain regional haze and visibility transport requirements for the State of Arkansas. This FIP would address the requirements of the Regional Haze Rule (RHR) and interstate visibility transport for those portions of Arkansas' State Implementation Plan (SIP) we disapproved in our final action published on March 12, 2012. Specifically, the proposed FIP addresses the requirements for Best Available Retrofit Technology (BART) for those sources for which we did not approve Arkansas' BART determinations, Reasonable Progress Goals (RPGs), reasonable progress controls and a long-term strategy, as well as the interstate visibility transport requirements for pollutants that affect visibility in Class I areas in nearby states. Specific to the reasonable progress controls requirement, we are proposing in the alternative two options for controlling the emissions from the Entergy Independence Plant that is not subject to BART. Under Option 1, we are proposing controls for emissions of SO

2

, and NO

X.

If we take final action on this finding, the source will be subject to controls for both pollutants. Alternatively, under Option 2, we are proposing controls for only emissions of SO

2

for this planning period. In particular, we are soliciting comments on the alternate proposed Options 1 and 2.

DATES:

Comments:

Comments must be received on or before May 16, 2015.

Public Hearing:

We are holding information sessions—for the purpose of providing additional information and informal discussion for our proposal, and public hearings—to accept oral comments into the record, as follows:

Date:

Thursday, April 16, 2015.

Time:

Information Session: 9 a.m.-9:45 a.m. (break from 9:45 a.m.-10 a.m.)

Public hearing: 10 a.m.-11:30 a.m. (break from 11:30 a.m.-1 p.m.)

Information Session: 1 p.m.-1:45 p.m. (break from 1:45 p.m.-2 p.m.)

Public hearing: 2 p.m.-7:30 p.m. (including break from 4 p.m.-4:30 p.m.).

Please see the

ADDRESSES

section for the location of the hearing in North Little Rock, AR

.

ADDRESSES:

Submit your comments, identified by Docket No. EPA-R06-OAR-2015-0189, by one of the following methods:

•

Federal e-Rulemaking Portal: http://www.regulations.gov

. Follow the online instructions for submitting comments.

•

Email: R6AIR_ARHaze@epa.gov

.

•

Mail:

Mr. Guy Donaldson, Chief, Air Planning Section (6PD-L), Environmental Protection Agency, 1445 Ross Avenue, Suite 1200, Dallas, Texas 75202-2733.

•

Hand or Courier Delivery:

Mr. Guy Donaldson, Chief, Air Planning Section (6PD-L), Environmental Protection Agency, 1445 Ross Avenue, Suite 700, Dallas, Texas 75202-2733. Such deliveries are accepted only between the hours of 8 a.m. and 4 p.m. weekdays, and not on legal holidays. Special arrangements should be made for deliveries of boxed information.

•

Fax:

Mr. Guy Donaldson, Chief, Air Planning Section (6PD-L), at fax number 214-665-7263.

Instructions:

Direct your comments to Docket No. EPA-R06-OAR-2015-0189. Our policy is that all comments received will be included in the public docket without change and may be made available online at

www.regulations.gov,

including any personal information provided, unless the comment includes information claimed to be Confidential Business Information (CBI) or other information whose disclosure is restricted by statute. Do not submit information that you consider to be CBI or otherwise protected through

www.regulations.gov

or email. The

www.regulations.gov

Web site is an “anonymous access” system, which means we will not know your identity or contact information unless you provide it in the body of your comment. If you send an email comment directly to us without going through

www.regulations.gov

your email address will be automatically captured and included as part of the comment that is placed in the public docket and made available on the Internet. If you submit an electronic comment, we recommend that you include your name and other contact information in the body of your comment and with any disk or CD-ROM you submit. If we cannot read your comment due to technical difficulties and cannot contact you for clarification, we may not be able to consider your comment. Electronic files should avoid the use of special characters and any form of encryption, and be free of any defects or viruses.

Docket:

All documents in the docket are listed in the

www.regulations.gov

index. Although listed in the index, some information is not publicly available,

e.g.,

CBI or other information whose disclosure is restricted by statute. Certain other material, such as copyrighted material, will be publicly available only in hard copy. SIP materials which are incorporated by reference into 40 Code of Federal Regulations (CFR) part 52 are available for inspection at the following location: Environmental Protection Agency, Region 6, 1445 Ross Avenue, Suite 700, Dallas, TX 75202. Publicly available materials are available either electronically in

www.regulations.gov

or in hard copy at the Region 6 office. The Regional Office hours are Monday through Friday, 8:30 to 4:30, excluding Federal holidays.

Hearing location:

Arkansas Department of Environmental Quality, Commission Room, 1st floor, 5301 Northshore Drive, North Little Rock, AR 72118.

The public hearing will provide interested parties the opportunity to present information and opinions to us concerning our proposal. Interested parties may also submit written comments, as discussed in the proposal. Written statements and supporting information submitted during the comment period will be considered with the same weight as any oral comments and supporting information presented at the public hearing. We will not respond to comments during the public hearings. When we publish our final action, we will provide written responses to all significant oral and written comments received on our proposal. To provide opportunities for questions and discussion, we will hold an information session prior to the public hearing. During the information session, EPA staff will be available to informally answer questions on our proposed action. Any comments made to EPA staff during an information session must still be provided orally during the public hearing, or formally in writing within 30 days after completion of the hearings, in order to be considered in the record. At the public hearings, the hearing officer may limit the time available for each commenter to address the proposal to three minutes or less if the hearing officer determines it to be appropriate. We will not be providing equipment for commenters to

show overhead slides or make computerized slide presentations. Any person may provide written or oral comments and data pertaining to our proposal at the public hearings. Verbatim English language transcripts of the hearing and written statements will be included in the rulemaking docket.

FOR FURTHER INFORMATION CONTACT:

To schedule your inspection, contact Ms. Dayana Medina at (214) 665-7241 or via electronic mail at

medina.dayana@epa.gov

.

SUPPLEMENTARY INFORMATION:

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

Table of Contents

I. Background

II. Overview of Proposed Actions

A. Regional Haze

B. Interstate Transport of Pollutants That Affect Visibility

C. History of State Submittals and Our Actions

D. Our Authority To Promulgate a FIP

III. Our Proposed BART Analyses and Determinations

A. Identification of BART-Eligible Sources and Subject to BART Sources

1. Georgia Pacific-Crossett Mill 6A and 9A Power Boilers

2. AECC Carl E. Bailey Generating Station Unit 1

B. BART Factors

C. BART Determinations and Proposed Federally Enforceable Limits

1. AECC Carl E. Bailey Generating Station

2. AECC John L. McClellan Generating Station

3. AEP Flint Creek Power Plant

4. Entergy White Bluff Plant

5. Entergy Lake Catherine Plant

6. Domtar Ashdown Paper Mill

IV. Our Proposed Reasonable Progress Analysis and Determinations

A. Reasonable Progress Analysis of Point Sources

1. Entergy Independence Plant Units 1 and 2

B. Reasonable Progress Goals

V. Our Proposed Long-Term Strategy

VI. Our Proposal for Interstate Visibility Transport

VII. Summary of Proposed Actions

A. Regional Haze

B. Interstate Visibility Transport

VIII. Statutory and Executive Order Reviews

I. Background

Regional haze is visibility impairment that is produced by a multitude of sources and activities that are located across a broad geographic area and emit fine particulates (PM

2.5

) (

e.g.,

sulfates, nitrates, organic carbon (OC), elemental carbon (EC), and soil dust), and their precursors (

e.g.,

sulfur dioxide (SO

2

), nitrogen oxides (NO

X

), and in some cases, ammonia (NH

3

) and volatile organic compounds (VOC)). Fine particle precursors react in the atmosphere to form PM

2.5

, which impairs visibility by scattering and absorbing light. Visibility impairment reduces the clarity, color, and visible distance that one can see. PM

2.5

can also cause serious health effects and mortality in humans and contributes to environmental effects such as acid deposition and eutrophication.

Data from the existing visibility monitoring network, the “Interagency Monitoring of Protected Visual Environments” (IMPROVE) monitoring network, show that visibility impairment caused by air pollution occurs virtually all the time at most national parks and wilderness areas. The average visual range

1

in many Class I areas (

i.e.,

national parks and memorial parks, wilderness areas, and international parks meeting certain size criteria) in the western United States is 100-150 kilometers, or about one-half to two-thirds of the visual range that would exist without anthropogenic air pollution. In most of the eastern Class I areas of the United States, the average visual range is less than 30 kilometers, or about one-fifth of the visual range that would exist under estimated natural conditions. 64 FR 35715 (July 1, 1999).

1

Visual range is the greatest distance, in kilometers or miles, at which a dark object can be viewed against the sky.

In section 169A of the 1977 Amendments to the Clean Air Act (CAA), Congress created a program for protecting visibility in the nation's national parks and wilderness areas. This section of the CAA establishes as a national goal the prevention of any future, and the remedying of any existing man-made impairment of visibility in 156 national parks and wilderness areas designated as mandatory Class I Federal areas.

2

On December 2, 1980, EPA promulgated regulations to address visibility impairment in Class I areas that is “reasonably attributable” to a single source or small group of sources,

i.e.,

“reasonably attributable visibility impairment.”

3

These regulations represented the first phase in addressing visibility impairment. EPA deferred action on regional haze that emanates from a variety of sources until monitoring, modeling, and scientific knowledge about the relationships between pollutants and visibility impairment were improved.

2

Areas designated as mandatory Class I Federal areas consist of National Parks exceeding 6000 acres, wilderness areas and national memorial parks exceeding 5000 acres, and all international parks that were in existence on August 7, 1977. 42 U.S.C. 7472(a). In accordance with section 169A of the CAA, EPA, in consultation with the Department of Interior, promulgated a list of 156 areas where visibility is identified as an important value. 44 FR 69122 (November 30, 1979). The extent of a mandatory Class I area includes subsequent changes in boundaries, such as park expansions. 42 U.S.C. 7472(a). Although states and tribes may designate as Class I additional areas which they consider to have visibility as an important value, the requirements of the visibility program set forth in section 169A of the CAA apply only to “mandatory Class I Federal areas.” Each mandatory Class I Federal area is the responsibility of a “Federal Land Manager.” 42 U.S.C. 7602(i). When we use the term “Class I area” in this action, we mean a “mandatory Class I Federal area.”

3

45 FR 80084 (December 2, 1980).

Congress added section 169B to the CAA in 1990 to address regional haze issues, and we promulgated regulations addressing regional haze in 1999.

4

The Regional Haze Rule (RHR) revised the existing visibility regulations to integrate into the regulations provisions addressing regional haze impairment and established a comprehensive visibility protection program for Class I areas. The requirements for regional haze, found at 40 CFR 51.308 and 51.309, are included in our visibility protection regulations at 40 CFR 51.300-309. The requirement to submit a regional haze SIP applies to all 50 states, the District of Columbia, and the Virgin Islands. States were required to submit the first implementation plan addressing regional haze visibility impairment no later than December 17, 2007.

5

4

64 FR 35714 (July 1, 1999), codified at 40 CFR part 51, subpart P (Regional Haze Rule).

5

See 40 CFR 51.308(b). EPA's regional haze regulations require subsequent updates to the regional haze SIPs. 40 CFR 51.308(g)-(i).

II. Overview of Proposed Actions

A. Regional Haze

We are proposing to promulgate a FIP as described in this notice and summarized in this section to address those portions of Arkansas' regional haze SIP that we disapproved on March 12, 2012.

6

In our March 12, 2012 final action, we disapproved Arkansas' BART control analyses and determinations for nine units at six facilities and the Reasonable Progress Goals (RPGs) analysis and RPGs set by Arkansas, and partially disapproved the long-term strategy for making reasonable progress. We are proposing this FIP because Arkansas has not provided a revision to its SIP to address the deficiencies identified in our March 12, 2012 partial disapproval. We believe, however, it is preferable for states to take the lead in implementing the Regional Haze requirements as envisioned by the Clean Air Act. We will work with the State of Arkansas if it chooses to develop a SIP

to meet the Regional Haze requirements to replace this proposed FIP.

6

77 FR 14604, March 12, 2012.

The FIP we are proposing includes BART control determinations for sources in Arkansas without previously approved BART determinations and associated compliance schedules and requirements for equipment maintenance, monitoring, testing, recordkeeping, and reporting for all affected sources and units. The BART sources addressed in this FIP cause or contribute to visibility impairment at one or more Class I areas in Arkansas and Missouri. The two Class I areas in Arkansas are the Caney Creek Wilderness Area and the Upper Buffalo Wilderness Area. The two Class I areas in Missouri are the Hercules-Glades Wilderness Area and the Mingo National Wildlife Refuge. In this FIP, we are proposing SO

2

, NO

X

, and PM BART control determinations for nine units at six facilities in Arkansas. We are proposing SO

2

, NO

X

, and PM BART determinations for Unit 1 of the Arkansas Electric Cooperative Corporation (AECC) Carl E. Bailey Generating Station; SO

2

, NO

X

, and PM BART determinations for Unit 1 of the AECC John L. McClellan Generating Station; SO

2

and NO

X

BART determinations for Boiler No. 1 of the American Electric Power (AEP) Flint Creek Power Plant; SO

2

and NO

X

BART determinations for Units 1 and 2 and SO

2

, NO

X

, and PM BART determinations for the Auxiliary Boiler of the Entergy White Bluff Plant; NO

X

BART determination for Unit 4 of the Entergy Lake Catherine Plant; SO

2

and NO

X

BART determinations for Power Boiler No. 1 and SO

2

, NO

X

and PM BART determinations for Power Boiler No. 2 of the Domtar Ashdown Mill. Additionally, for the reasonable progress requirements, we are proposing in the alternative two options for controlling the emissions from the Entergy Independence Plant that is not subject to BART. Under Option 1, under the reasonable progress requirements, we are proposing controls for emissions of SO

2

and NO

X

for Units 1 and 2 of the Entergy Independence Plant. Alternatively, under Option 2, we are proposing controls for only emissions of SO

2

for the first planning period. We solicit comments on this proposed alternative approach. We are also soliciting public comment on any alternative control measures for Entergy White Bluff Units 1 and 2 and Independence Units 1 and 2 that would address the BART and reasonable progress requirements for these four units for this regional haze planning period. The measures in the FIP that we are proposing will reduce emissions that contribute to regional haze in Arkansas' Class I areas and other nearby Class I areas. RPGs are interim visibility goals towards meeting the CAA's national visibility goal of preventing any future, and remedying any existing, impairment of visibility resulting from manmade air pollution in Class I areas. This proposed FIP and the portion of the Arkansas regional haze SIP that we approved on March 12, 2012, together would ensure that progress is made toward natural visibility conditions at these Class I areas. This proposed action and the accompanying documents that are available in the Docket explain the basis for our proposed Arkansas Regional Haze FIP. Please refer to our previous rulemaking on the Arkansas regional haze SIP for additional background regarding the CAA, regional haze, and our RHR.

7

7

77 FR 14604, March 12, 2012.

B. Interstate Transport of Pollutants That Affect Visibility

We propose that a combination of those portions of the Arkansas regional haze SIP that we previously approved and the measures in the FIP will satisfy the visibility requirement of CAA section 110(a)(2)(D)(i)(II) for the 1997 8-hour ozone and 1997 PM

2.5

national ambient air quality standards (NAAQS). CAA section 110(a)(2)(D)(i)(II) requires that states have a SIP, or submit a SIP revision, containing provisions “prohibiting any source or other type of emission activity within the state from emitting any air pollutant in amounts which will . . . interfere with measures required to be included in the applicable implementation plan for any other State under part C [of the CAA] to protect visibility.” Because of the impacts on visibility from the interstate transport of pollutants, we interpret these “good neighbor” provisions of section 110 of the Act as requiring states to include in their SIPs measures to prohibit emissions that would interfere with the reasonable progress goals set to protect Class I areas in other states. For Arkansas, we interpret this to mean that the State must include in its SIP a demonstration that emissions from Arkansas sources and activities will not have the prohibited impacts on other states' existing SIPs. We refer herein to this requirement as the interstate transport visibility requirement. The Arkansas Department of Environmental Quality (ADEQ) submitted a SIP revision to address this requirement on April 2, 2008, and submitted supplemental information on September 27, 2011. The April 2, 2008 submittal stated that Arkansas is relying on the Air Pollution Control and Ecology Commission (APCEC) Regulation 19, Chapter 15, also known as the State BART rulemaking, to satisfy the requirements of section 110(a)(2)(D)(i)(II) with respect to visibility transport. The April 2, 2008 SIP submittal, which was submitted prior to Arkansas' submission of the Arkansas regional haze SIP, also stated that it is not possible to assess whether there is any interference with the measures in the applicable SIP for another state designed to protect visibility for the 1997 8-hour ozone and PM

2.5

NAAQS until Arkansas submits and we approve the Arkansas regional haze SIP. In our final rule published on March 12, 2012, we partially approved and partially disapproved the SIP submittal with respect to the interstate transport visibility requirement under CAA section 110(a)(2)(D)(i)(II), triggering the obligation for us to promulgate a FIP or to fully approve a revised SIP submission from Arkansas to ensure that the requirement is fully addressed.

8

Today's notice describes our proposed FIP, which we propose to find will fully address the deficiencies we identified in our prior partial disapproval action of Arkansas' SIP submittal with respect to the interstate visibility transport requirement under CAA section 110(a)(2)(D)(i)(II) for the 1997 8-hour ozone and 1997 PM

2.5

NAAQS.

8

Id.

C. History of State Submittals and Our Actions

As discussed above, Arkansas submitted a SIP revision on April 2, 2008, to address the interstate transport visibility requirement of CAA section 110(a)(2)(D)(i)(II) for the 1997 8-hour ozone and 1997 PM

2.5

NAAQS. To address the first regional haze implementation period, Arkansas submitted a regional haze SIP on September 23, 2008. On August 3, 2010, Arkansas submitted a SIP revision with non-substantive revisions to the APCEC Regulation 19, Chapter 15, which identified the BART-eligible and subject-to-BART sources in Arkansas and established the BART emission limits that subject-to-BART sources are required to comply with. On September 27, 2011, the State submitted supplemental information on the Arkansas regional haze SIP. We are hereafter referring to these regional haze submittals collectively as the “2008 Arkansas RH SIP.” On March 12, 2012, we partially approved and partially disapproved the 2008 Arkansas RH SIP

and the April 2, 2008 SIP submittal concerning the interstate transport visibility requirements for the 1997 8-hour ozone and 1997 PM

2.5

NAAQS.

9

9

Id.

Our partial disapproval of the 2008 Arkansas RH SIP included a disapproval of the following BART determinations made by Arkansas:

• SO

2

, NO

X

, and PM BART for the AECC Carl E. Bailey Generating Station Unit 1;

• SO

2

, NO

X

, and PM BART for the AECC John L. McClellan Generating Station Unit 1;

• SO

2

and NO

X

BART for the AEP Flint Creek Power Plant No. 1 Boiler;

• SO

2

and NO

X

BART for the bituminous and sub-bituminous coal firing scenarios for the Entergy White Bluff Plant Units 1 and 2;

• SO

2

, NO

X

, and PM BART for the Entergy White Bluff Plant Auxiliary Boiler;

• NO

X

BART for the natural gas firing scenario for the Entergy Lake Catherine Plant Unit 4;

• SO

2

, NO

X

, and PM BART for the fuel oil firing scenario for the Entergy Lake Catherine Plant Unit 4;

• SO

2

and NO

X

BART for the Domtar Ashdown Mill No. 1 Power Boiler; and

• SO

2

, NO

X

, and PM BART for the Domtar Ashdown Mill No. 2 Power Boiler.

In our final action, we also disapproved Arkansas' determinations that the Georgia Pacific-Crossett Mill 6A Boiler is not BART-eligible, and that the 6A and 9A Boilers are not subject to BART. By partially disapproving Arkansas' BART determinations, we also partially disapproved the corresponding provisions of APCEC Regulation 19, Chapter 15. We also disapproved Arkansas' RPGs for its two Class I areas, the Caney Creek Wilderness Area and the Upper Buffalo Wilderness Area, because Arkansas did not meet the requirement under section 169A(g)(1) of the CAA and 40 CFR 51.308(d)(1)(i)(A) to consider the four statutory factors when establishing its RPGs. Additionally, we partially disapproved Arkansas' long-term strategy because it relied on other disapproved portions of the SIP.

D. Our Authority To Promulgate a FIP

Under section 110(c) of the Act, whenever we disapprove a mandatory SIP submission in whole or in part, we are required to promulgate a FIP within 2 years unless we approve a SIP revision correcting the deficiencies before promulgating a FIP. Specifically, CAA section 110(c) provides that the Administrator shall promulgate a FIP within 2 years after the Administrator disapproves a state implementation plan submission “unless the State corrects the deficiency, and the Administrator approves the plan or plan revision, before the Administrator promulgates such Federal implementation plan.” The term “Federal implementation plan” is defined in section 302(y) of the CAA in pertinent part as a plan promulgated by the Administrator to correct an inadequacy in a SIP.

Thus, because we partially disapproved the 2008 Arkansas RH SIP and the SIP submittal addressing the interstate transport visibility requirement, we are required to promulgate a FIP for Arkansas, unless we first approve a SIP revision that corrects the disapproved portions of these SIP submittals. As Arkansas has not as yet submitted a revised SIP following our partial disapproval, we are proposing a FIP to address those portions of the SIP that we disapproved.

III. Our Proposed BART Analyses and Determinations

Following our 2012 disapproval of the 2008 Arkansas RH SIP, Arkansas began the process of generating additional technical information and analysis for the BART determinations. Arkansas gathered technical documentation from the companies whose BART determinations we disapproved. These documents were provided to us and are the basis for our evaluation of BART determinations for the facilities with prior disapproved BART determinations.

A. Identification of BART-Eligible Sources and Subject to BART Sources

States are required to identify all the BART-eligible sources within their boundaries by utilizing the three eligibility criteria in the BART Guidelines (70 FR 39158) and the RHR (40 CFR 51.301): (1) One or more emission units at the facility fit within one of the 26 categories listed in the BART Guidelines; (2) the emission unit(s) began operation on or after August 6, 1962, and the unit was in existence on August 6, 1977; and (3) the potential emissions of any visibility-impairing pollutant from subject units are 250 tons or more per year. Sources that meet these three criteria are considered BART-eligible. Once a list of the BART-eligible sources within a state has been compiled, states must determine whether to make BART determinations for all of them or to consider exempting some of them from BART because they may not reasonably be anticipated to cause or contribute to any visibility impairment in a Class I area. The BART Guidelines present several options that rely on modeling and/or emissions analyses to determine if a source may reasonably be anticipated to cause or contribute to visibility impairment in a Class I area. A source that may not be reasonably anticipated to cause or contribute to any visibility impairment in a Class I area is not “subject to BART,” and for such sources, a state need not apply the five statutory factors to make a BART determination.

1. Georgia Pacific-Crossett Mill 6A and 9A Power Boilers

In our March 12, 2012 final action, we approved Arkansas' identification of BART-eligible sources except for the Georgia-Pacific Crossett Mill 6A Boiler. We also approved Arkansas' determination of which sources are subject to BART, with the exception of its determination that the Georgia-Pacific Crossett Mill 6A and 9A Boilers are not subject to BART. Our basis and analyses for our disapproval of Arkansas' determinations that the 6A Boiler is not BART-eligible and that the 6A and 9A Boilers are not subject to BART is found in our October 17, 2011 proposed rulemaking, March 12, 2012 final rulemaking, and the associated TSDs.

10

10

76 FR 64186 and 77 FR 14604.

A revised Title V permit for the Georgia-Pacific Crossett Mill was issued on August 4, 2011, and again on May 23, 2012. Although no pollution controls were installed, the permitted emission limits for SO

2

and PM

10

for the 6A Boiler and SO

2

, NO

X

, and PM

10

for the 9A Boiler were revised to be more stringent. In a letter dated May 18, 2012,

11

Georgia-Pacific explained to ADEQ that it had conducted additional dispersion modeling in 2011 based on the currently enforceable Title V permit limits for the 6A and 9A Boilers.

12

The results of the 2011 modeling analysis are summarized in the table below. Based on modeling of the current permit limits, the boilers' maximum visibility impact was modeled to be 0.359 dv at Caney Creek (assuming 2002 meteorology). In the letter to ADEQ, Georgia-Pacific stated its belief that the 2011 dispersion modeling analysis and the current Title V permit that enforces the modeled limits are sufficient to

demonstrate no cause or contribution to visibility impairment by the 6A and 9A Boilers, and that the boilers are therefore not subject to BART.

11

May 18, 2012 letter from James W. Cutbirth, Environmental Services Superintendent at Georgia-Pacific Crossett Paper Operations, to Mary Pettyjohn, ADEQ. A copy of this letter can be found in the docket for this proposed rulemaking.

12

See ADEQ Operating Air Permit No. 0597-AOP-R14, issued on May 23, 2012. A copy of the air permit can be found in the docket for this proposed rulemaking.

Table 1—Maximum Modeled Visibility Impacts From 6A and 9A Boilers

[Georgia-Pacific's 2011 Dispersion Modeling Analysis]

Class I area

Maximum Visibility Impact

(dv)

2001

meteorology

2002

meteorology

2003

meteorology

Caney Creek

0.16

0.359

0.296

Upper Buffalo

0.099

0.074

0.099

Hercules-Glades

0.08

0.288

0.125

Mingo

0.123

0.093

0.168

Sipsey

0.171

0.184

0.119

Following discussions with us and ADEQ, Georgia-Pacific provided additional information and documentation to support its contention that the 6A and 9A Boilers are not subject to BART. Georgia-Pacific calculated maximum 24-hour emission rates from the 2001-2003 baseline period using fuel usage data, and then showed that these estimated maximum 24-hour emission rates are below the revised emission rates it used in the 2011 BART screening modeling. In a letter dated April 1, 2013, Georgia-Pacific provided spreadsheets with fuel usage data for the 6A and 9A Boilers for each day during the 2001-2003 baseline period.

13

The 6A Boiler burned only natural gas during the 2001-2003 baseline period, while the 9A Boiler burned both natural gas and bark. Georgia-Pacific used emission factors from AP-42,

Compilation of Air Pollutant Emission Factors,

14

to calculate 24-hour emission rates for SO

2

, NO

X

, and PM

10

(lb/hr) for the 6A and 9A Boilers for each day during the baseline years. The gas and bark usage value for each day was multiplied by the corresponding AP-42 emission factor to calculate the 24-hour emission rate for each day during the baseline period.

15

Georgia-Pacific then determined the maximum 24-hour emission rates for the 6A and 9A Boilers during the baseline period (see table below).

16

13

April 1, 2013 letter from James W. Cutbirth, Environmental Services Superintendent at Georgia-Pacific Crossett Paper Operations, to Mary Pettyjohn, ADEQ. A copy of this letter and all attachments can be found in the docket for this proposed rulemaking.

14

AP-42, Compilation of Air Pollutant Emission Factors,

has been published since 1972 as the primary compilation of EPA's emission factor information. It contains emission factors and process information for more than 200 air pollution source categories. The emission factors have been developed and compiled from source test data, material balance studies, and engineering estimates. The Fifth Edition of AP-42 was published in January 1995. Since then, EPA has published supplements and updates to the fifteen chapters available in Volume I, Stationary Point and Area Sources. The latest emissions factors are available at

http://www.epa.gov/ttnchie1/ap42/

.

15

Please see the TSD for example calculations of the 24-hour emissions rates for the 6A and 9A Boilers. See also the April 1, 2013 letter from James W. Cutbirth, Environmental Services Superintendent at Georgia-Pacific Crossett Paper Operations, to Mary Pettyjohn, ADEQ. The attachments to the April 1, 2013 letter include spreadsheets with the calculated 24-hour emission rates for each day during the 2001-2003 baseline period for the 6A and 9A Boilers. The letter and all attachments are found in the docket for this proposed rulemaking.

16

The maximum 24-hour emission rate for PM

10

for the 9A Boiler is based on the results of stack testing Georgia-Pacific conducted when the boiler was firing bark and gas, since the stack test results yielded a higher emission rate than what Georgia-Pacific calculated using AP-42 emission factors.

Table 2—Georgia-Pacific Crossett Mill 6A and 9A Boiler Maximum 24-Hour Emission Rates From the 2001-2003 Baseline Period

Unit

Maximum 24-Hour Emission Rates (lb/hr)

SO

2

NO

X

PM

10

6A Boiler

0.2

90.7

2.5

9A Boiler

17.9

174.1

72.0

Georgia-Pacific then compared the calculated maximum 24-hour emission rates from the baseline period with the emission rates it modeled in the 2011 BART screening modeling and with the current Title V permit limits (see table below).

17

A comparison of these values shows that the calculated maximum 24-hour emission rates for each pollutant are below the emission rates Georgia-Pacific modeled in the 2011 BART screening modeling, and also below the currently enforceable Title V permit limits.

17

See ADEQ Operating Air Permit No. 0597-AOP-R14, issued on May 23, 2012. A copy of the air permit can be found in the docket for this proposed rulemaking.

Table 3—Georgia-Pacific Crossett Mill—Comparison of Maximum 24-Hour Emission Rates With Modeled Emission Rates and Title V Permit Limits

SO

2

NO

X

PM

10

6A Boiler

Calculated Maximum 24-hr Emission Rate (lb/hr)

0.2

90.7

2.5

Modeled Emission Rate (lb/hr)

0.3

120.0

3.3

Title V permit Limit (lb/hr)

0.3

120.0

3.3

9A Boiler

Calculated Maximum 24-hr Emission Rate (lb/hr)

17.9

174.1

72.0

Modeled Emission Rate (lb/hr)

200.0

218.0

75.8

Title V permit Limit (lb/hr)

199.8

196.0

77.4

Because the 2011 BART screening modeling showed visibility impacts below 0.5 dv from the 6A and 9A Boilers and the recently estimated maximum 24-hour emission rates from the 2001-2003 baseline period are below the modeled emission rates, we propose that it is reasonable to conclude that the boilers had visibility impacts below 0.5 dv during the baseline period. Accordingly, we believe that Georgia‐Pacific's newly provided analysis and documentation, as described above and in our TSD in more detail, is appropriate to demonstrate that the 6A and 9A Boilers are not subject to BART. In comparison to the information available to us when we issued our March 12, 2012 final action on the 2008 Arkansas RH SIP, we believe this newly provided analysis allows for a more accurate assessment of whether or not the 6A and 9A Boilers are subject to BART. Based on this newly provided information, we are proposing to find that while the 6A Boiler is a BART-eligible source, it is not subject to BART. The 9A Boiler is also BART-eligible (as the State determined in the 2008 Arkansas RH SIP), but we are also proposing to find that the 9A Boiler is not subject to BART. Therefore, it is not necessary to perform a BART five factor analysis or to make BART determinations for the Georgia-Pacific Crossett Mill 6A and 9A Boilers.

2. AECC Carl E. Bailey Generating Station Unit 1

In our March 12, 2012 final action on the 2008 Arkansas RH SIP, we noted that the original meteorological databases generated by the Central Regional Air Planning Association (CENRAP) and used by Arkansas to conduct its modeling analyses did not include surface and upper air meteorological observations as EPA guidance recommends. Thus, in its evaluation to determine if a source exceeds the 0.5 dv contribution threshold at potentially affected Class I areas, Arkansas used the maximum value (

i.e.,

1st high value) of modeled visibility impacts instead of the 98th percentile value (

i.e.,

8th high value). The use of the maximum modeled values in the 2008 Arkansas RH SIP was agreed to by us, representatives of the Federal Land Managers, and CENRAP stakeholders. In our March 12, 2012 final action, we also approved Arkansas' determination that the AECC Carl E. Bailey Generating Station (AECC Bailey) Unit 1 is BART-eligible and subject to BART, based on the maximum value of modeled visibility impacts.

Following our March 12, 2012 final action on the 2008 Arkansas RH SIP, AECC hired a consultant to conduct revised modeling of AECC Bailey Unit 1. Unlike the modeling submitted in the 2008 Arkansas RH SIP, the revised modeling shows visibility impacts from Bailey Unit 1 below 0.5 dv, which is the threshold used by Arkansas to determine if a source is subject to BART. However, we already approved Arkansas' determination that the AECC Bailey Unit 1 is subject to BART in our March 12, 2012 final action on the 2008 Arkansas RH SIP.

We do not have the discretion to reopen the issue of whether the source is subject to BART because we already approved the portion of the 2008 Arkansas RH SIP in which Arkansas determined AECC Bailey Unit 1 is subject to BART and Arkansas has not provided us a SIP revision to replace the previous determination.

18

We cannot re-consider our approval of that portion of the 2008 Arkansas RH SIP to have been in error because Arkansas did not submit the revised modeling to us with a request to remove the source from BART and the modeling approach used by Arkansas in that SIP is consistent with our regional haze regulations and was agreed to by us, representatives of the Federal Land Managers, and CENRAP stakeholders prior to submittal of the 2008 Arkansas RH SIP. Therefore, our proposed FIP is not reopening the issue of whether the source is subject to BART, and our final approval of Arkansas' determination that the source is subject to BART remains in place and in the subsection that follows we evaluate AECC Bailey Unit 1 under BART.

18

77 FR 14604, March 12, 2012.

B. BART Factors

The purpose of the BART analysis is to identify and evaluate the best system of continuous emission reduction based on the BART Guidelines.

19

In determining BART, a state, or EPA if promulgating a FIP, must consider the five statutory factors in section 169A of the CAA: (1) The costs of compliance; (2) the energy and nonair quality environmental impacts of compliance; (3) any existing pollution control technology in use at the source; (4) the remaining useful life of the source; and (5) the degree of improvement in visibility which may reasonably be anticipated to result from the use of such technology. See also 40 CFR 51.308(e)(1)(ii)(A). Following the BART Guidelines, the BART analysis is broken down into five steps. Steps 1 through 3 address the availability, technical feasibility and effectiveness of retrofit control options. The consideration of the five statutory factors occurs during steps 4 and 5 of the process.

19

See July 6, 2005 BART Guidelines, 40 CFR 51, Regional Haze Regulations and Guidelines for Best Available Retrofit Technology Determinations.

Step 1—Identify all available retrofit control technologies.

Step 2—Eliminate technically infeasible options.

Step 3—Evaluate control effectiveness of remaining control technologies.

Step 4—Evaluate impacts and document the results.

• Factor 1: Costs of compliance.

• Factor 2: Energy and nonair quality environmental impacts of compliance.

• Factor 3: Existing pollution control technology in use at the source.

• Factor 4: Remaining useful life of the facility.

Step 5—Evaluate Visibility Impacts

• Factor 5: Degree of improvement in visibility which may reasonably be

anticipated to result from the use of retrofit control technology.

C. BART Determinations and Proposed Federally Enforceable Limits

1. AECC Carl E. Bailey Generating Station

The AECC Bailey Unit 1 is a wall-fired boiler with a gross output of 122 megawatts (MW) and a maximum heat input rate of 1,350 million British thermal units per hour (MMBtu/hr). The unit is currently permitted to burn natural gas and fuel oil. The fuel oil burned is currently subject to a sulfur content limit of 2.3% by weight. AECC hired a consultant to perform a BART five factor analysis for Bailey Unit 1.

20

20

See the following BART analyses: “BART Five Factor Analysis, Arkansas Electric Cooperative Corporation Bailey and McClellan Generating Stations,” dated March 2014, Version 4, prepared by Trinity Consultants Inc. in conjunction with Arkansas Electric Cooperative Corporation; and “BART Five Factor Analysis- NO

X

Analysis, Addendum to the July 24, 2012 BART Five Factor Analysis, Arkansas Electric Cooperative Corporation Bailey and McClellan Generating Stations,” dated December 2013, Version 3. A copy of these two BART analyses can be found in the docket for our proposed rulemaking.

The table below summarizes the baseline emission rates modeled for the source. The SO

2

and NO

X

baseline emission rates are the highest actual 24-hour emission rates based on 2001-2003 continuous emission monitoring system (CEMS) data, while the PM baseline emission rates are based on stack testing and AP-42 emission factors.

Table 4—Baseline Emission Rates for AECC Bailey Unit 1

Unit/Fuel scenario

SO

2

(lb/hr)

NO

X

(lb/hr)

Total PM

10

21

(lb/hr)

Inorganic

condensable

(SO

4

)

(lb/hr)

Coarse soil

(PMc)

(lb/hr)

Fine soil

(PMf)

(lb/hr)

Organic

condensable PM

(SOA)

(lb/hr)

Elemental carbon

(EC)

(lb/hr)

Bailey, Unit 1—Natural Gas firing

0.5

443.8

10.2

0.3

0.0

0.0

7.4

2.6

Bailey, Unit 1—Fuel Oil firing

2,375.8

408.8

55.8

4.6

13.7

34.1

0.8

2.7

The NO

X

and

PM baseline emission rates used in AECC's revised modeling for the fuel oil firing scenario were revised from what the State modeled in the 2008 Arkansas RH SIP. The revised NO

X

emission rates for the fuel oil firing scenario are higher than what was modeled in the 2008 Arkansas RH SIP, while the revised PM

10

emission rates for fuel oil firing scenario are lower than what was modeled in the 2008 Arkansas RH SIP. We have some concern with AECC's use of the PM

10

baseline emission rates, which are based on stack testing, because there is no discussion provided on how the stack test results are representative of the maximum 24-hour emissions. However, because the visibility impacts due to PM

10

emissions from Bailey Unit 1 are so small, we believe a closer inspection of the revised PM

10

emission rates and any further updates to these would likely not result in significant changes to the modeled visibility impacts and would not affect our proposed BART decision. As shown in the table below, the percentage of the visibility impairment attributable to PM

10

from Bailey Unit 1 at the Class I area with the highest baseline visibility impacts (Mingo) is 8.10% for the natural gas firing scenario and 1.26% for the fuel oil firing scenario. Most of the visibility impairment is attributable to NO

3

(83.34%) for the natural gas firing scenario and to SO

4

(93.95%) for the fuel oil firing scenario. Therefore, we did not take further steps to adjust the PM

10

emission rates or conduct additional modeling.

21

The National Park Service PM speciation worksheets are typically used to speciate PM

10

into SO

4

, PMc, PMf, SOA, and EC.

AECC's modeling for the baseline emission rates uses the CALPUFF dispersion model to determine the baseline visibility impairment attributable to Bailey Unit 1 at the four Class I areas impacted by emissions from BART sources in Arkansas. These Class I areas are the Caney Creek Wilderness Area, Upper Buffalo Wilderness Area, Hercules-Glades Wilderness Area, and Mingo National Wildlife Refuge. The baseline (

i.e.,

existing) visibility impairment attributable to each unit at each Class I area is summarized in the table below.

Table 5—98th Percentile Baseline Visibility Impairment Attributable to AECC Bailey Unit 1 (2001-2003)

Unit/Fuel scenario

Maximum

(Δdv)

98th percentile

(Δdv)

98th percentile

% SO

4

98th percentile

% NO

3

98th percentile

% PM

10

Bailey Unit 1—Natural Gas firing

Caney Creek

Upper Buffalo

0.219

0.170

0.083

0.072

0.28

0.29

96.36

95.02

3.35

3.43

Hercules-Glades

0.238

0.073

0.22

92.76

3.67

Mingo

0.443

0.102

0.45

83.34

8.10

Bailey Unit 1—Fuel Oil firing

Caney Creek

Upper Buffalo

0.970

0.696

0.330

0.348

87.19

90.73

12.11

8.42

0.57

0.83

Hercules-Glades

0.687

0.368

82.74

14.39

2.08

Mingo

1.592

0.379

93.95

4.68

1.26

a.

Proposed BART Analysis and Determination for SO

2

.

The source does not have existing SO

2

pollution control technology. AECC identified all available control technologies, eliminated options that are not technically feasible, and evaluated the control effectiveness of the remaining control options. Each technically feasible control option was then evaluated in terms of a five factor BART analysis.

AECC's BART evaluation considered both flue gas desulfurization (FGD) and fuel switching as possible controls. AECC found that FGD applications have not been used historically for SO

2

control on fuel oil-fired units in the U.S.

electric industry and therefore considered it a technically infeasible option for control of Bailey Unit 1. Accordingly, AECC did not further consider FGD for SO

2

BART. We concur with AECC's decision to focus the SO

2

BART evaluation on fuel switching. Switching to a fuel with a lower sulfur content is expected to reduce SO

2

emissions in proportion to the reduction in the sulfur content of the fuel, assuming that the fuels have similar heat contents. Bailey Unit 1 burns primarily natural gas, but is also permitted to burn fuel oil. The baseline fuel AECC assumed in the BART analysis is No. 6 fuel oil with 1.81% sulfur content, based on the average sulfur content of the fuel oil from the most recent shipment received by the facility in December 2006. According to the facility, a portion of the fuel oil from this shipment still remains in storage at the facility for future use. AECC evaluated switching to the fuel types shown in the table below.

Table 6—Control Effectiveness of Fuel Switching Options for AECC Bailey Unit 1

Fuel switching options

Estimated SO

2

control

efficiency

%

No. 6 fuel oil, 1% sulfur

45

No. 6 fuel oil, 0.5% sulfur

72

Diesel, 0.05% sulfur

97

Natural gas

99.9

AECC estimated the average cost-effectiveness of switching Bailey Unit 1 to No. 6 fuel oil with 1% sulfur content to be $1,198 per ton of SO

2

removed. Switching from the baseline fuel to No. 6 fuel oil with 0.5% sulfur content was estimated to cost $2,559 per ton of SO

2

removed. The results of AECC's cost analysis are summarized in the table below. For the natural gas switching scenario, AECC found that the current cost of natural gas is actually lower than the cost of the baseline fuel. Therefore, the average cost-effectiveness of switching from the baseline fuel to natural gas is denoted as a negative value (cost savings) in the table

below.

22

The average cost-effectiveness was calculated by dividing the total annual differential cost of switching from the baseline fuel oil to the lower sulfur fuel.

23

The incremental cost-effectiveness calculation compares the costs and performance level of a control option to those of the next most stringent option, as shown in the following formula (with respect to cost per emissions reduction): Incremental Cost Effectiveness (dollars per incremental ton removed) = (Total annualized costs of control option)—(Total annualized costs of next control option)/(Control option annual emissions)—Next control option annual emissions). See BART Guidelines, 40 CFR Part 51, Appendix Y, section IV.D.4.e.

Table 7—AECC Bailey Unit 1: Summary of Costs Associated with Fuel Switching

Fuel

switching

scenario

Average sulfur

content

(%)

Baseline emission rate

( SO

2

tpy)

Controlled emission rate

(SO

2

tpy)

Annual emissions reductions

( SO

2

tpy)

Annual fuel usage

(Mgal/yr)

Fuel cost

($/MMBtu)

Total

annual

differential

cost of fuel

switching

($/yr)

Average cost

effectiveness

22

($/ton)

Incremental

cost

effectiveness

23

($/ton)

Baseline

1.81

37.03

252.86

16.00

No. 6 Fuel Oil—1%

1.00

20.67

16.36

252.86

16.50

19,596

1,198

No. 6 Fuel Oil—0.5%

0.50

10.23

26.80

252.86

17.75

68,587

2,559

4,693

Diesel

0.05

0.99

36.05

287.86

20.95

194,003

5,382

13,558

Natural Gas

0.04

0.01

37.02

38.77

6.19

−384,550

−10,387

−596,446

AECC's evaluation did not identify any energy or non-air quality environmental impacts associated with switching to 1% sulfur No. 6 fuel oil, 0.5% sulfur No. 6 fuel oil, or diesel. The evaluation noted that switching to natural gas may have energy impacts during periods of natural gas curtailment. During periods of natural gas curtailment, natural gas infrastructure maintenance, and other emergencies, the AECC Bailey Generating Station relies on the fuel oil stored at the plant to maintain electrical reliability. AECC's evaluation notes that because of this, it is important to maintain the ability to burn fuel oil at AECC Bailey, even if fuel oil is currently more expensive than natural gas.

With regard to consideration of the remaining useful life of Unit 1, this factor does not impact the SO

2

BART analysis because the emissions control approaches being evaluated for BART do not require capital cost expenditures. Thus, there are no control costs that need to be amortized over the lifetime of the unit.

AECC assessed the visibility improvement associated with fuel switching by comparing the 98th percentile modeled visibility impact of the baseline scenario to the 98th percentile modeled visibility impact of each control scenario. The table below shows a comparison of the baseline visibility impacts and the visibility impacts of the different fuel switching control scenarios that were evaluated, including the cumulative visibility benefits.

Table 8—AECC Bailey Unit 1: Summary of 98th Percentile Visibility Impacts and Improvement due to Fuel Switching

Class I area

Baseline visibility impact

(Δdv)

No. 6 fuel oil—1% sulfur

No. 6 fuel oil—0.5% sulfur

Diesel

Natural gas

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Caney Creek

0.330

0.193

0.137

0.142

0.188

0.084

0.246

0.083

0.247

Upper Buffalo

0.348

0.194

0.154

0.127

0.221

0.069

0.279

0.072

0.276

Hercules-Glades

0.368

0.206

0.162

0.135

0.233

0.069

0.299

0.073

0.295

Mingo

0.379

0.206

0.173

0.170

0.209

0.095

0.284

0.102

0.277

Cumulative Visibility Improvement (Δdv)

0.626

0.851

1.108

1.095

The table above shows that switching to No. 6 fuel oil with 1% sulfur content at Bailey Unit 1 is projected to result in 0.173 dv visibility improvement at Mingo (based on the 98th percentile modeled visibility impacts). The visibility improvement at each of the other three affected Class I areas is projected to be slightly less than that amount, while the cumulative visibility improvement at the four Class I areas is projected to be 0.626 dv. Switching to No. 6 fuel oil with 0.5% sulfur content is projected to result in meaningful visibility improvement. It is projected to result in 0.233 dv visibility improvement at Hercules-Glades. The visibility improvement at each of the other three affected Class I areas is projected to be slightly less than that amount, while the cumulative visibility improvement at the four Class I areas is projected to be 0.851 dv. Switching to diesel or natural gas is also projected to result in meaningful visibility improvement. The visibility improvement at Hercules-Glades is projected to be 0.299 dv for switching to diesel and 0.295 dv for switching to natural gas, and slightly less than that amount at each of the other three affected Class I areas. The cumulative visibility improvement at the four Class I areas is projected to be 1.108 dv for switching to diesel and 1.095 dv for switching to natural gas.

Our Proposed SO

2

BART Determination:

Taking into consideration the five factors, we are proposing to determine that BART for the AECC Bailey Unit 1 is switching to fuels with 0.5% or lower sulfur content by weight. The cost effectiveness of switching to No. 6 fuel oil with 0.5% sulfur content is within the range of what we consider to be cost-effective for BART and it is projected to result in considerable visibility improvement compared to the baseline at the affected Class I areas. Switching to No. 6 fuel oil with 0.5% sulfur content has an estimated average cost-effectiveness of $2,559 per ton of SO

2

removed and is projected to result in visibility improvement ranging from 0.188 to 0.233 dv at each modeled Class I area, and a cumulative visibility improvement of 0.851 dv at the four modeled Class I areas. Switching to natural gas would currently cost less than the baseline fuel and is projected to result in even greater visibility improvement than switching to No. 6 fuel oil with 0.5% sulfur content. However, the BART Guidelines provide that it is not our intent to direct subject-to-BART sources to switch fuel forms, such as from coal or fuel oil to gas (40 CFR part 51, Appendix Y, section IV.D.1). Because natural gas has a sulfur content by weight that is well below 0.5%, the facility may elect to use this type of fuel to comply with BART, but we are not proposing to require a switch to natural gas for Unit 1. Switching to diesel is projected to result in an almost identical level of visibility improvement at each Class I area as switching to natural gas. The incremental visibility improvement of switching to diesel compared to switching to No. 6 fuel oil with a sulfur content of 0.5% is projected to range from 0.058 dv to 0.075 dv at each affected Class I area but the average cost-effectiveness is estimated to be $5,382 per ton of SO

2

removed and the incremental cost-effectiveness compared to switching to No. 6 fuel oil with a sulfur content of 0.5% is estimated to be $13,558 per ton of SO

2

removed, which we do not consider to be very cost-effective in view of the incremental visibility improvement. Because diesel also has a sulfur content by weight that is well below 0.5%, the facility may elect to use this type of fuel to comply with BART, but we are not proposing to require a switch to diesel for Unit 1. We are proposing to determine that SO

2

BART for Bailey Unit 1 is switching to fuels with 0.5% or lower sulfur content by weight. We propose to require that the facility purchase no fuel after the effective date of the rule that does not meet the sulfur content requirement and that 5 years from the effective date of the rule no fuel be burned that does not meet the requirement. We propose that any higher sulfur fuel oil that remains from the facility's 2006 fuel oil shipment cannot be burned past this point. As discussed above, the unit's baseline fuel is No. 6 fuel oil with 1.81% sulfur content, based on the average sulfur content of the fuel oil from the most recent fuel oil shipment received by the facility in 2006. Based on our discussions with the facility, it is our understanding that the unit burns fuel oil primarily during periods of natural gas curtailment and during periodic testing and that the facility still has stockpiles of fuel oil from the most recent shipment. Because the unit burns primarily natural gas and does not ordinarily burn fuel oil on a frequent basis, we believe it is appropriate to allow the facility 5 years to burn its existing supply of No. 6 fuel oil, as the normal course of business dictates and in accordance with any operating restrictions enforced by ADEQ. We believe that a shorter compliance date may result in the facility burning its existing supply of higher sulfur No. 6 fuel oil relatively quickly, resulting in a high amount of SO

2

emissions being emitted by the unit over a short period of time. This is not the intent of our regional haze regulations. We are also proposing regulatory text that includes monitoring, reporting, and recordkeeping requirements associated with this proposed determination.

b.

Proposed BART Analysis and Determination for NO

X

. AECC's BART evaluation examined BART controls for NO

X

for AECC Bailey Unit 1. Bailey Unit 1 does not currently have pollution control equipment for NO

X.

AECC's evaluation identified all available control technologies, eliminated options that are not technically feasible, and evaluated the control effectiveness of the remaining control options. Each technically feasible control option was then evaluated in terms of a five factor BART analysis.

For NO

X

BART, AECC's evaluation considered both combustion and post-combustion controls. The combustion controls evaluated by AECC consisted of flue gas recirculation (FGR), overfire air (OFA), and low NO

X

burners (LNB). The post-combustion controls evaluated consisted of selective catalytic reduction (SCR) and selective non-catalytic reduction (SNCR). AECC found that some boilers may be restricted from installing OFA retrofits due to physical size and space restraints. For purposes of the NO

X

BART evaluation, AECC assumed OFA to be a technically feasible option for Bailey Unit 1, but noted that if OFA was determined to be BART based on the evaluation of the five BART factors, then further analyses would have to be performed to determine if: (1) The dimensions of AECC Bailey's main boilers have sufficient upper furnace volume for OFA mixing and complete combustion and (2) the furnace meets the physical space requirements for OFA ports and air supply ducts. The remaining NO

X

control options were found to be technically feasible.

AECC evaluated three control scenarios: A combination of combustion controls (FGR, OFA, and LNB); the combination of combustion controls and SNCR; and SCR. Based on literature estimates, AECC found that the estimated NO

X

control range for oil and gas wall-fired boilers, such as Bailey Unit 1, is approximately 0.2-0.4 lb/MMBtu using FGR and 0.2-0.3 lb/MMBtu using OFA.

24

When LNB is combined with OFA and FGR, AECC

estimated that a NO

X

controlled emission rate of 0.15—0.20 lb/MMBtu can be achieved at Bailey Unit 1. The NO

X

controlled emission rate of combustion controls combined with SNCR is estimated to be 0.12 lb/MMBtu. The NO

X

control efficiency of SCR is estimated to be 80-90% for gas fired boilers and 70-80% for oil fired boilers, which corresponds to a controlled emission rate of 0.04-0.08 lb/MMBtu for Bailey Unit 1.

24

“Controlling Nitrogen Oxides Under the Clean Air Act: A Menu of Options,” section II, dated July 1994, State and Territorial Air Pollution Program Administrators (STAPPA) and Association of Local Air Pollution Control Officials (ALAPCO).

AECC's cost analysis for NO

X

controls was based on “budgetary” cost estimates it obtained by AECC from the pollution control equipment vendor, Babcock Power Systems. AECC estimated the capital and operating costs of controls based on the vendor's estimates, engineering estimates, and published calculation methods using EPA's Air Pollution Control Cost Manual (EPA Control Cost Manual).

25

We are not aware of any enforceable shutdown date for the AECC Bailey Generating Station, nor did AECC's evaluation indicate any future planned shutdown. This means that the anticipated useful life of the boiler is expected to be at least as long as the capital cost recovery period of controls. Therefore, a 30-year amortization period was assumed in the NO

X

BART analysis as the remaining useful life of Unit 1. The table below summarizes the estimated cost for installation and operation of NO

X

controls for Bailey Unit 1.

25

EPA's “Air Pollution Control Cost Manual,” Sixth edition, January 2002, is located at

www.epa.gov/ttncatc1/products.html#cccinfo.

Table 9—Summary of NO

X

Control Costs for AECC Bailey Unit 1

Control scenario

Baseline emission rate

(NO

X

tpy)

Natural gas controlled emission level

(lb/MMBtu)

26

Fuel oil controlled emission level

(lb/MMBtu)

27

Controlled emission rate

(NO

X

tpy)

Annual emissions reductions

(NO

X

tpy)

Total annual cost

($/yr)

Average cost- effectiveness

($/ton)

Incremental cost-effectiveness($/ton)

Combustion Controls

49.81

0.15

0.15

30.83

18.98

700,477

36,905

Combustion Controls + SNCR

49.81

0.12

0.12

24.79

25.02

1,223,157

48,884

86,536

SCR

28

49.81

0.04

0.08

9.65

40.16

1,555,718

38,738

21,966

AECC

estimated the average cost-effectiveness of installing and operating combustion controls to be $36,905 per ton of NO

X

removed for Bailey Unit 1. The combination of combustion controls and SNCR was estimated to cost $48,884 per ton of NO

X

removed, while SCR was estimated to cost $38,738 per ton of NO

X

removed. In its evaluation, AECC also explained that it expects the cost-effectiveness of NO

X

controls to be lower (

i.e.,

greater dollars per ton removed) in future years due to projected reduced operation of the unit.

26

See the preceding paragraphs for a discussion of the expected controlled emission rates for natural gas vs. fuel oil firing.

27

Id.

AECC did not identify any energy or non-air quality environmental impacts associated with the use of LNB, OFA, or FGR. As for SCR and SNCR, we are not aware of any unusual circumstances at the facility that could create non-air quality environmental impacts associated with the operation of these controls greater than experienced elsewhere and that may therefore provide a basis for their elimination as BART (40 CFR part 51, Appendix Y, section IV.D.4.i.2.). Therefore, we do not believe there are any energy or non-air quality environmental impacts associated with NO

X

controls at AECC Bailey Unit 1 that would affect our proposed BART determination.

AECC assessed the visibility improvement associated with NO

X

controls by modeling the NO

X

emission rates associated with each control option using CALPUFF, and then comparing the visibility impairment associated with the baseline emission rates to the visibility impairment associated with the controlled emission rates as measured by the 98th percentile modeled visibility impact. The tables below show a comparison of the baseline (

i.e.,

existing) visibility impacts and the visibility impacts associated with NO

X

controls.

Table 10—AECC Bailey Unit 1: Summary of the 98th Percentile Visibility Impacts and Improvement Due to NO

X

Controls—Natural Gas Firing

Class I area

Baseline

visibility

impact

(Δdv)

Combustion controls

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Combustion controls + SNCR

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

SCR

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Caney Creek

0.083

0.039

0.044

0.032

0.051

0.014

0.069

Upper Buffalo

0.072

0.034

0.038

0.028

0.044

0.013

0.059

Hercules-Glades

0.073

0.035

0.038

0.029

0.044

0.013

0.06

Mingo

0.102

0.051

0.051

0.043

0.059

0.021

0.081

Cumulative Visibility Improvement (Δdv)

0.171

0.198

0.269

Table 11—AECC Bailey Unit 1: Summary of the 98th Percentile Visibility Impacts and Improvement Due to NO

X

Controls—Fuel Oil Firing

Class I area

Baseline

visibility

impact

(Δdv)

Combustion controls

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Combustion controls + SNCR

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

SCR

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Caney Creek

0.330

0.325

0.005

0.325

0.005

0.323

0.007

Upper Buffalo

0.347

0.332

0.015

0.329

0.018

0.325

0.022

Hercules-Glades

0.367

0.339

0.028

0.333

0.034

0.325

0.042

Mingo

0.378

0.369

0.009

0.367

0.011

0.364

0.014

Cumulative Visibility Improvement (Δdv)

0.057

0.068

0.085

The tables above show that the installation and operation of NO

X

controls is projected to result in a very modest visibility improvement from the baseline. Combustion controls at Bailey Unit 1 are projected to result in visibility improvement of up to 0.051 dv at any single Class I area for the natural gas firing scenario and 0.028 dv for the fuel oil firing scenario (based on the 98th percentile modeled visibility impacts). A combination of combustion controls and SNCR is projected to result in only slight incremental visibility improvement over combustion controls alone. For example, a combination of combustion controls and SNCR at Bailey Unit 1 is projected to result in visibility improvement of up to 0.059 dv at any single Class I area for natural gas firing and 0.034 dv for fuel oil firing, which is an incremental visibility improvement of 0.008 dv for natural gas firing and 0.006 dv for fuel oil firing compared to combustion controls alone. Similarly, the installation and operation of SCR is projected to result in only slight incremental visibility improvement compared to a combination of combustion controls and SNCR.

Our Proposed NO

X

BART Determination:

Taking into consideration the five factors, we are proposing to determine that NO

X

BART for the AECC Bailey Unit 1 is no additional controls, and are proposing that the facility's existing NO

X

emission limit satisfies BART for NO

X

. We are proposing the existing emission limit of 887 lb/hr for NO

X

BART for Bailey Unit 1.

29

As discussed above, the operation of combustion controls at Bailey Unit 1 is projected to result in a maximum visibility improvement of 0.051 dv (Mingo), and a smaller amount of visibility improvement at each of the other affected Class I areas. The installation and operation of combustion controls at Bailey Unit 1 has an average cost-effectiveness of $36,905 per ton of NO

X

removed, which is not within the range of what we consider cost-effective. We believe the relatively small visibility benefit projected from the operation of combustion controls both when combusting fuel oil and natural gas does not justify the estimated cost of those controls. The operation of a combination of combustion controls and SNCR is estimated to cost $48,884 per ton of NO

X

removed, which is also not within the range of what we consider cost-effective. A combination of combustion controls and SNCR is projected to result in only slight incremental visibility benefit compared to combustion controls alone. The operation of SCR is estimated to cost $38,738 per ton of NO

X

removed, which is not cost-effective, and is projected to result in only slight incremental visibility benefit compared to a combination of combustion controls and SNCR. We are proposing to find that NO

X

BART for Bailey Unit 1 is no additional controls and are proposing that the existing NO

X

emission limit of 887 lb/hr is BART for NO

X

and that compliance be demonstrated using the unit's existing CEMS. We are proposing that this emission limitation be complied with for BART purposes from the date of effectiveness of the finalized action. We are also proposing regulatory text that includes monitoring, reporting, and recordkeeping requirements associated with these emission limits.

29

See ADEQ Operating Air Permit No. 0154-AOP-R4, Section IV, Specific Conditions No. 1 and 7.

c.

Proposed BART Analysis and Determination for PM.

PM emissions are inherently low when burning natural gas. Bailey Unit 1 does not currently have pollution control equipment for PM. AECC's BART evaluation considered the following control technologies for PM BART: Dry electrostatic precipitator (ESP), wet ESP, fabric filter, wet scrubber, cyclone (

i.e.,

mechanical collector), and fuel switching. Residual fuel, such as the baseline No. 6 fuel oil burned at Bailey Unit 1, has inherent ash that contributes to emissions of filterable PM. Reductions in filterable PM emissions are directly related to the sulfur content of the fuel.

30

Therefore, switching to No. 6 fuel oil with a lower sulfur content is expected to result in lower filterable PM emissions. AECC's evaluation considered switching to No. 6 fuel oil with 1% sulfur content by weight, No. 6 fuel oil with 0.5% sulfur content by weight, diesel, and natural gas. These are the same lower sulfur fuel types evaluated in the SO

2

BART analysis for the unit.

30

See “AP-42, Compilation of Air Pollutant Emission Factors,” section 1.3.3.1, and Table 1.3-1, available at

http://www.epa.gov/ttnchie1/ap42/.

AECC's evaluation noted that the particulate matter from oil-fired boilers tends to be sticky and small, affecting the collection efficiency of dry ESPs and fabric filters. Dry ESPs operate by placing a charge on the particles through a series of electrodes, and then capturing the charged particles on collection plates, while fabric filters work by filtering the PM in the flue gas through filter bags. The collected particles are periodically removed from the filter bag through a pulse jet or reverse flow mechanism. Because of the sticky nature of particles from oil-fired boilers, dry ESPs and fabric filters are deemed technically infeasible for use at Bailey Unit 1. Wet ESPs, cyclones, wet scrubbers, and fuel switching were identified as technically feasible options for Bailey Unit 1. AECC noted that although cyclones and wet scrubbers are considered technically feasible for use at these boiler types, they are not very efficient at controlling particles in the smaller size fraction, particularly particles smaller than a few microns. However, the majority of the PM emissions from Bailey Unit 1 are greater than a few microns in size.

AECC estimated that switching to a lower sulfur fuel has a PM control efficiency ranging from approximately 44%-99%, depending on the fuel type. The other technically feasible control technologies are estimated to have the following PM control efficiency: Wet ESP—up to 90%, cyclone—85%, and wet scrubber—55%.

AECC evaluated the capital costs, operating costs, and average cost-effectiveness of wet ESPs, cyclones, and wet scrubbers. It also evaluated the average cost-effectiveness of switching to No. 6 fuel oil with 1% sulfur content, No. 6 fuel oil with 0.5% sulfur content, diesel, and natural gas. AECC developed the capital and operating costs of a wet ESP and wet scrubber using the Electric Power Research Institute's (EPRI) Integrated Emissions Control Cost Estimating Workbook (IECCOST) Software. The capital costs of controls (except for fuel switching) were annualized over a 15-year period and then added to the annual operating costs to obtain the total annualized costs. The table below summarizes the average cost-effectiveness of PM controls. The average cost-effectiveness was determined by dividing the annualized cost of controls by the annual PM emissions reductions. The annual emissions reductions were determined by subtracting the estimated controlled annual emission rates from the baseline annual emission rates. AECC estimated the baseline and controlled annual emission rates by conducting a mass balance on the sulfur content of the various fuels evaluated.

We disagree with two aspects of AECC's cost evaluation for PM controls. First, the total annual cost numbers associated with fuel switching should be the same as those used in the SO

2

BART cost analysis for Bailey Unit 1 (see Table 7). In earlier draft versions of AECC's BART analysis, which were provided to us for review, the cost numbers for fuel switching used in the PM and SO

2

BART analyses were identical. In response to comments provided by us, the total annual cost and average cost-effectiveness numbers for fuel switching were revised in the final version of AECC's SO

2

BART analysis. However, it appears that AECC overlooked updating these cost numbers in the final PM BART analysis.

31

In the table below, we have revised the total annual cost of fuel switching for the PM BART analysis to be consistent with the cost estimates from AECC's SO

2

BART analysis, and we have also updated the PM average cost-effectiveness values. The second aspect of AECC's cost evaluation for PM controls that we disagree with is the use of a 15-year capital cost recovery period for calculating the average cost-effectiveness of a wet ESP, wet scrubber, and cyclone. As previously discussed, we are not aware of any enforceable shutdown date for the AECC Bailey Generating Station, nor did AECC's evaluation indicate any future planned shutdown. Therefore, we believe that assuming a 30-year equipment life rather than a 15-year equipment life would be more appropriate for these control technologies. Extending the amortization period from 15 to 30 years has the effect of decreasing the total annual cost of each control option, thereby improving the average cost-effectiveness value of controls (

i.e.,

less dollars per ton removed). However, after considering all five BART factors, we do not believe AECC's assumption of a 15-year amortization period has an impact on our proposed BART decision and therefore we did not revise the amortization period or the average cost-effectiveness calculations for the PM control options. This is discussed in more detail below. The table below summarizes the estimated cost for fuel switching and the installation and operation of PM control equipment for Bailey Unit 1.

31

The final version of AECC's BART analysis for SO

2

and PM, upon which our analysis is largely based, is titled “BART Five Factor Analysis Arkansas Electric Cooperative Corporation Bailey and McClellan Generating Stations, March 2014, Version 4.” A copy of AECC's analysis can be found in the docket for our proposed rulemaking.

Table 12—Summary of Cost of PM Controls for AECC Bailey Unit 1—Baseline Is No. 6 Fuel Oil With 1.81% Sulfur Content by Weight

Control scenario

Baseline

emission rate

(PM tpy)

Control

efficiency

(%)

Controlled

emission rate

(PM tpy)

Annual

emissions

reductions

(PM tpy)

Capital cost

($)

Total

annual cost

($/yr)

Average

cost-

effectiveness

($/ton)

Incremental

cost-

effectiveness

($/ton)

Wet Scrubber

25.63

55.0

11.53

14.09

140,957,713

50,150,862

3,558,286

No. 6 Fuel oil—1% S

25.63

65.7

8.80

16.83

19,596

1,164

−18,296,082

Cyclone

25.63

85.0

3.84

21.78

989,479

1,188,630

54,570

236,168

No. 6 Fuel oil—0.5% S

25.63

89.3

2.75

22.88

68,587

2,997

−1,020,948

Wet ESP

25.63

90.0

2.56

23.06

105,141,431

22,638,340

981,583

125,387,517

Natural Gas

25.63

99.0

0.26

25.37

−384,550

−15,157

−9,966,619

Diesel

25.63

99.5

0.13

25.50

194,003

7,608

4,450,408

The table above shows that the average cost-effectiveness values of all add-on PM control technology options evaluated for AECC Bailey Unit 1 ranged from approximately $55,000 per ton of PM removed to more than $3.5 million per ton of PM removed. The incremental cost-effectiveness of add-on PM control technology options ranged from $236,168 to $125,387,517 per ton of PM removed. Switching to No. 6 fuel oil with either a 1% or 0.5% sulfur content was found to be within the range of what we generally consider cost-effective for BART. Switching to No. 6 fuel oil with 1% sulfur content is estimated to cost $1,164 per ton of PM removed, while switching to No. 6 fuel oil with 0.5% sulfur content is estimated to cost $2,997 per ton of PM removed. As discussed in the SO

2

BART analysis, the current cost of natural gas is actually lower than the cost of the baseline fuel. Therefore, the average cost-effectiveness of switching from the baseline fuel to natural gas is denoted as a negative value in the table above. As discussed above, AECC also explained that it expects the average cost-effectiveness of PM control equipment to be lower (

i.e.,

greater dollars per ton removed) in future years due to projected reduced operation of the unit due to a change in the management of the load control area in which the facility is located.

AECC did not identify any energy or non-air quality environmental impacts associated with fuel switching, but did identify impacts associated with the use of wet ESPs and wet scrubbers due to their electricity usage. Energy use in and of itself does not disqualify a technology (40 CFR part 51, Appendix Y, section IV.D.4.h.1.). In addition, the cost of the electricity needed to operate this equipment has already been factored into the cost of controls. AECC also

noted that both wet ESPs and wet scrubbers generate wastewater streams that must either be treated on-site or sent to a waste water treatment plant, and the wastewater treatment process will generate a filter cake that would likely require landfilling. The BART Guidelines provide that the fact that a control device creates liquid and solid waste that must be disposed of does not necessarily argue against selection of that technology as BART, particularly if the control device has been applied to similar facilities elsewhere and the solid or liquid waste is similar to those other applications. (40 CFR part 51, Appendix Y, section IV.D.4.i.2.). We are not aware of any unusual circumstances at the AECC Bailey Generating Station that could potentially create greater problems than experienced elsewhere related to the treatment of wastewater and any necessary landfilling, nor did AECC's evaluation discuss or mention any such unusual circumstances. Therefore, the need to treat wastewater or landfill any filter cake or other waste in and of itself does not provide a basis for disqualification or elimination of a wet ESP or wet scrubber.

As previously discussed, we are not aware of any enforceable shutdown date for the AECC Bailey Generating Station, nor did AECC's evaluation indicate any future planned shutdown. Therefore, we believe it is appropriate to assume a 30-year amortization period in the PM BART analysis as the remaining useful life of the unit. Assuming a 30-year amortization period, these controls would have a lower estimated total annual cost and would therefore have an improved cost-effectiveness (

i.e.,

less dollars per ton removed) than estimated in AECC's evaluation. However, we did not adjust the amortization period because we do not believe this has an impact on our proposed BART decision. As discussed in the subsection below, the visibility benefit expected from the installation and operation of PM control equipment is too small to justify the cost of these controls. Therefore, we did not revise the amortization period and the average cost-effectiveness calculations for the PM control equipment options.

As switching to lower sulfur fuels has impacts on both SO

2

and PM emissions, AECC's assessment of the visibility improvement associated with fuel switching is addressed in the SO

2

BART analysis for Bailey Unit 1. Table 8 summarizes the visibility improvement associated with controlled emission rates for SO

2

and PM as a result of fuel switching. AECC assessed the visibility improvement associated with wet ESPs, wet scrubbers, and cyclones by modeling the PM emission rates associated with each control option using CALPUFF, and then comparing the visibility impairment associated with the baseline emission rates to the visibility impairment associated with the controlled emission rates as measured by the 98th percentile modeled visibility impact. The controlled PM

10

emission rates associated with wet ESPs, wet scrubbers, and cyclones were calculated by reducing the uncontrolled annual PM

10

emission rates by the pollutant removal efficiency of each control technology. The SO

2

and NO

X

emission rates modeled in the controlled scenarios are the same as those from the baseline scenario, as it is assumed that SO

2

and NO

X

emissions would remain unchanged. The table below shows a comparison of the baseline (

i.e.,

existing) visibility impacts and the visibility impacts associated with PM controls.

Table 13—AECC Bailey Unit 1: Summary of the 98th Percentile Visibility Impacts and Improvement From PM Controls

Class I area

Baseline

visibility

impact

(Δdv)

Wet ESP

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Wet scrubber

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Cyclone

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Caney Creek

0.330

0.327

0.003

0.328

0.002

0.328

0.002

Upper Buffalo

0.347

0.343

0.004

0.345

0.002

0.345

0.002

Hercules-Glades

0.367

0.356

0.011

0.360

0.007

0.361

0.006

Mingo

0.378

0.371

0.007

0.374

0.004

0.374

0.004

Cumulative Visibility Improvement (Δdv)

0.025

0.015

0.014

The table above shows that the operation of a wet ESP, wet scrubber, or cyclone at Bailey Unit 1 is projected to result in minimal visibility improvement at the four affected Class I areas. The modeled visibility improvement from switching to No. 6 fuel oil with 1% sulfur content, No. 6 fuel oil with 0.5% sulfur content, diesel, or natural gas is summarized in Table 8. The modeled visibility improvement shown in Table 8 reflects both SO

2

and PM emissions reductions as a result of switching to fuels with lower sulfur content. However, the majority of the baseline visibility impact at each Class I area when burning the baseline fuel oil is due to SO

2

emissions, while PM

10

emissions contribute only a small portion of the baseline visibility impacts at each Class I area (see Table 5). Accordingly, the majority of the visibility improvement associated with switching to lower sulfur fuels can reasonably be expected to be the result of a reduction in SO

2

emissions.

Our Proposed PM BART Determination:

Taking into consideration the five factors, we propose to determine that PM BART for the AECC Bailey Unit 1 does not require add-on controls. Consistent with our proposed determination for SO

2

BART, we are proposing that PM BART is satisfied by Unit 1 switching to fuels with 0.5% or lower sulfur content by weight. As discussed above, we disagree with AECC's use of a 15-year amortization period in the cost analysis for a wet ESP, wet scrubber, and cyclone. Assuming a 30-year amortization period, these controls would have lower estimated total annual costs and would therefore have an improved cost-effectiveness (

i.e.,

less dollars per ton removed) compared to what was estimated in AECC's evaluation. However, after considering all five BART factors, even if we revised AECC's cost estimates to reflect a 30-year amortization period, resulting in a lower total annual cost and improved cost-effectiveness, we would still not be able to justify the cost of add-on controls in light of the minimal visibility benefit of these controls (see the table above).

We are proposing to determine that PM BART for Bailey Unit 1 is switching to fuels with 0.5% or lower sulfur

content by weight. We propose to require that the facility purchase no fuel after the effective date of the rule that does not meet the sulfur content requirement and that 5 years from the effective date of the rule no fuel be burned that does not meet the requirement. We propose that any higher sulfur fuel oil that remains from the facility's 2006 fuel oil shipment cannot be burned past this point. As previously discussed, the unit's baseline fuel is No. 6 fuel oil with 1.81% sulfur content, based on the average sulfur content of the fuel oil from the most recent shipment received by the facility in 2006. Based on our discussions with the facility, it is our understanding that the unit burns fuel oil primarily during periods of natural gas curtailment and during periodic testing and that the facility still has stockpiles of fuel oil from the most recent fuel oil shipment. Because the unit burns primarily natural gas and does not ordinarily burn fuel oil on a frequent basis, we believe it is appropriate to allow the facility 5 years to burn its existing supply of No. 6 fuel oil, as the normal course of business dictates and in accordance with any operating restrictions enforced by ADEQ. We believe that a shorter compliance date may result in the facility burning its existing supply of higher sulfur No. 6 fuel oil relatively quickly, resulting in a high amount of SO

2

emissions being emitted by the unit over a short period of time. This is not the intent of our regional haze regulations. We are also proposing regulatory text that includes monitoring, reporting, and recordkeeping requirements associated with this proposed determination.

2. AECC John L. McClellan Generating Station

The AECC McClellan Unit 1 is subject to BART. As mentioned previously, we disapproved Arkansas' BART determinations for SO

2

, NO

X

, and PM for McClellan Unit 1 in our March 12, 2012 final action (77 FR 14604). The AECC McClellan Unit 1 is a wall-fired boiler with a gross output of 134 MW and a maximum heat input rate of 1,436 MMBtu/hr. The unit is currently permitted to burn natural gas and fuel oil. The fuel oil burned is currently subject to a sulfur content limit of 2.8% by weight. AECC, through its consultant, performed a five-factor analysis for McClellan Unit 1 (AECC's BART analysis).

32

32

See the following BART analyses: “BART Five Factor Analysis, Arkansas Electric Cooperative Corporation Bailey and McClellan Generating Stations,” dated March 2014, Version 4, prepared by Trinity Consultants Inc. in conjunction with Arkansas Electric Cooperative Corporation; and “BART Five Factor Analysis- NO

X

Analysis, Addendum to the July 24, 2012 BART Five Factor Analysis, Arkansas Electric Cooperative Corporation Bailey and McClellan Generating Stations,” dated December 2013, Version 3. A copy of these two BART analyses can be found in the docket for our proposed rulemaking.

The table below summarizes the baseline emission rates for the source. The SO

2

and NO

X

baseline emission rates are the highest actual 24-hour emission rates based on 2001-2003 CEMS data, while the PM baseline emission rates are based on stack testing and AP-42 emission factors.

Table 14—Baseline Emission Rates for AECC McClellan Unit 1

Unit/fuel scenario

SO

2

(lb/hr)

NO

X

(lb/hr)

Total

PM

10

(lb/hr)

SO

4

(lb/hr)

PMc

(lb/hr)

PMf

(lb/hr)

SOA

(lb/hr)

EC

(lb/hr)

McClellan, Unit 1—Natural Gas

0.6

423.9

10.9

0.3

0.0

0.0

7.9

2.7

McClellan, Unit 1—Fuel Oil

2,747.5

579.8

59.4

5.9

14.2

35.4

1.00

2.8

The NO

X

and PM baseline emission rates AECC modeled for the fuel oil firing scenario were updated from what the State modeled in the 2008 Arkansas RH SIP. The revised NO

X

emission rates for the fuel oil firing scenario are higher than what was modeled in the 2008 Arkansas RH SIP, while the revised PM

10

emission rates for fuel oil firing scenario are lower than what was modeled in the 2008 Arkansas RH SIP. We have some concern with AECC's use of the PM

10

baseline emission rates, which were based on stack testing, because there is no discussion provided on how the stack test results are representative of the maximum 24-hour emissions. However, because the visibility impacts due to PM

10

emissions from McClellan Unit 1 are so small, we believe a closer inspection of the revised PM

10

emission rates and any further updates to these would likely not result in significant changes to the modeled visibility impacts and would not affect our proposed BART decision. As shown in the table below, the percentage of the visibility impairment attributable to PM

10

at the Class I area with the highest visibility impacts (Caney Creek) is 6.63% for the natural gas firing scenario and 0.53% for the fuel oil firing scenario. Most of the visibility impairment is attributable to NO

3

(87.09%) for the natural gas firing scenario and to SO

4

(89.86%) for the fuel oil firing scenario. Therefore, we did not take further steps to adjust the PM

10

emission rates or conduct additional modeling.

AECC modeled the baseline emission rates using the CALPUFF dispersion model to determine the baseline visibility impairment attributable to McClellan Unit 1 at the four Class I areas impacted by emissions from BART sources in Arkansas. These Class I areas are the Caney Creek Wilderness Area, Upper Buffalo Wilderness Area, Hercules-Glades Wilderness Area, and Mingo National Wildlife Refuge. The baseline (

i.e.,

existing) visibility impairment attributable to McClellan Unit 1 at each Class I area is summarized in the table below.

Table 15—98th Percentile Baseline Visibility Impairment Attributable to AECC McClellan Unit 1

[2001-2003]

Unit/fuel scenario

Maximum (Δdv)

98th

Percentile

(Δdv)

98th

Percentile

(% SO

4

)

98th

Percentile

(% NO

3

)

98th

Percentile

(% PM

10

)

98th

Percentile

(% NO

2

)

McClellan Unit 1—Natural Gas:

Caney Creek

0.670

0.125

0.39

87.09

6.63

5.89

Upper Buffalo

0.258

0.052

0.34

91.78

4.82

3.05

Hercules-Glades

0.092

0.040

0.74

86.01

10.18

3.07

Mingo

0.132

0.058

0.33

91.96

5.13

2.58

McClellan Unit 1—Fuel Oil:

Caney Creek

3.007

0.622

89.86

9.62

0.53

0.00

Upper Buffalo

1.323

0.266

98.47

0.95

0.58

0.00

Hercules-Glades

0.662

0.231

78.67

20.16

1.17

0.01

Mingo

0.547

0.228

80.90

17.89

1.20

0.01

a.

Proposed BART Analysis and Determination for SO

2

. AECC's BART evaluation examined BART controls for SO

2

for the AECC McClellan Unit 1. The source does not have existing SO

2

pollution control technology. AECC identified all available control technologies, eliminated options that are not technically feasible, and evaluated the control effectiveness of the remaining control options. Each technically feasible control option was then evaluated in terms of a five factor BART analysis.

The AECC evaluation considered both FGD and fuel switching as possible controls. AECC found that FGD applications have not been used historically for SO

2

control on fuel oil-fired units in the U.S. electric industry and therefore considered it a technically infeasible option for control of McClellan Unit 1. Accordingly, AECC did not further consider FGD for SO

2

BART. We concur with AECC's decision to focus the SO

2

BART evaluation on fuel switching. Switching to a fuel with a lower sulfur content is expected to reduce SO

2

emissions in proportion to the reduction in the sulfur content of the fuel, assuming the fuels have similar heat contents. McClellan Unit 1 burns primarily natural gas, but is also permitted to burn fuel oil. The baseline fuel AECC assumed in the BART analysis is No. 6 fuel oil with 1.38% sulfur content, based on the average sulfur content of the fuel oil from the most recent fuel oil shipment received by the facility in April 2009. A portion of the fuel oil from this shipment still remains in storage at the facility for future use. AECC evaluated switching to the fuel types shown in the table below.

Table 16—Control Effectiveness of Fuel Switching Options for AECC McClellan Unit 1

Fuel switching options

Estimated

SO

2

control

efficiency

(%)

No. 6 fuel oil, 1% sulfur

28

No. 6 fuel oil, 0.5% sulfur

64

Diesel, 0.05% sulfur

96

Natural gas

99.9

AECC estimated the average cost-effectiveness of switching to No. 6 fuel oil with 1% sulfur content to be $2,613 per ton of SO

2

removed for McClellan Unit 1. Switching from the baseline fuel to No. 6 fuel oil with 0.5% sulfur content was estimated to cost $3,823 per ton of SO

2

removed. The results of AECC's cost analysis are summarized in the table below. For the natural gas switching scenario, AECC found that the current cost of natural gas is actually lower than the cost of the baseline fuel. Therefore, the average cost-effectiveness of switching from the baseline fuel to natural gas is denoted as a negative value (cost savings) in the table below.

Table 17—AECC McClellan Unit 1: Summary of Costs Associated With Fuel Switching

Fuel switching

scenario

Average

sulfur content

(%)

Baseline

emission rate

(SO

2

tpy)

Controlled

emission rate

(SO

2

tpy)

Annual

emissions

reductions

(SO

2

tpy)

Annual

fuel usage

(Mgal/yr)

Fuel cost

($/MMBtu)

Total annual

differential

cost of fuel

switching

($/yr)

Average cost

effectiveness

($/ton)

Incremental cost

effectiveness

($/ton)

Baseline

1.38

209.43

1,882.15

16.00

No. 6 Fuel Oil—1%

1.00

153.61

55.81

1,882.15

16.50

145,866

2,613

No. 6 Fuel Oil—0.5%

0.50

75.88

133.55

1,882.15

17.75

510,532

3,823

4,691

Diesel

0.05

7.31

202.11

2,142.73

20.95

1,444,077

7,145

13,616

Natural Gas

0.04

0.07

209.35

288.56

5.97

−2,926,874

−13,980

−603,723

The AECC BART evaluation did not identify any energy or non-air quality environmental impacts associated with switching to 1% sulfur No. 6 fuel oil, 0.5% sulfur No. 6 fuel oil, or diesel. The evaluation noted that switching to natural gas may have energy impacts during periods of natural gas curtailment. During periods of natural gas curtailment, natural gas infrastructure maintenance, and other emergencies, the McClellan Generating Station relies on the fuel oil stored at the plant to maintain electrical reliability. The AECC evaluation notes that because of this, it is important to maintain the ability to burn fuel oil at McClellan, even if fuel oil is currently more expensive to burn than natural gas.

With regard to consideration of the remaining useful life of Unit 1, this factor does not impact the SO

2

BART analysis because the emissions control approaches being evaluated for BART do not require capital cost expenditures. Thus, there are no control costs that need to be amortized over the lifetime of the unit.

AECC assessed the visibility improvement associated with fuel switching by comparing the 98th percentile modeled visibility impact of the baseline scenario (

i.e.,

existing) to the 98th percentile modeled visibility impact of each control scenario. The table below shows a comparison of the baseline visibility impacts and the visibility impacts of the different fuel switching control scenarios that were evaluated, including the cumulative visibility benefits.

Table 18—AECC McClellan Unit 1: Summary of 98th Percentile Visibility Impacts and Improvement Due to Fuel Switching

Class I area

Baseline

visibility

impact

(Δdv)

No. 6 fuel oil—1% sulfur

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

No. 6 fuel oil—0.5% sulfur

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Diesel

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Natural gas

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Caney Creek

0.622

0.537

0.085

0.322

0.3

0.174

0.448

0.125

0.497

Upper Buffalo

0.266

0.231

0.035

0.146

0.12

0.073

0.193

0.052

0.214

Hercules-Glades

0.231

0.202

0.029

0.115

0.116

0.062

0.169

0.040

0.191

Mingo

0.228

0.193

0.035

0.136

0.092

0.080

0.148

0.058

0.17

Cumulative Visibility Improvement (Δdv)

0.184

0.628

0.958

1.072

The table above shows that switching to No. 6 fuel oil with 1% sulfur content at McClellan Unit 1 is projected to result in visibility improvement of 0.085 dv at Caney Creek. The visibility improvement at each of the other three affected Class I areas is projected to be 0.035 dv or less, while the cumulative visibility improvement at the four Class I areas is projected to be 0.184 dv. Switching to No. 6 fuel oil with 0.5% sulfur content is projected to result in considerable visibility improvement. It is projected to result in 0.3 dv visibility improvement at Caney Creek. The visibility improvement at each of the other three affected Class I areas is projected to be 0.12 dv or less, while the cumulative visibility improvement at the four Class I areas is projected to be 0.628 dv. Switching to diesel or natural gas is also projected to result in considerable visibility improvement. The visibility improvement at Caney Creek is projected to be 0.448 dv for switching to diesel and 0.497 dv for switching to natural gas. The cumulative visibility improvement at the four Class I areas is projected to be 0.958 dv for switching to diesel and 1.072 dv for switching to natural gas.

Our Proposed SO

2

BART Determination:

Taking into consideration the five factors, we are proposing to determine that BART for McClellan Unit 1 is switching to fuels with 0.5% or lower sulfur content by weight. The cost of switching to No. 6 fuel oil with 0.5% sulfur content is within the range of what we consider to be cost-effective for BART and it is projected to result in considerable visibility improvement compared to the baseline at the affected Class I areas. Switching to No. 6 fuel oil with 0.5% sulfur content has an estimated average cost-effectiveness of $3,823 per ton of SO

2

removed and is projected to result in visibility improvement ranging from 0.092 to 0.3 dv at each modeled Class I area, and a cumulative visibility improvement of 0.628 dv at the four affected Class I areas. Switching to natural gas currently would cost less than the baseline fuel and is projected to result in even greater visibility improvement than switching to No. 6 fuel oil with 0.5% sulfur content. However, the BART Guidelines provide that it is not our intent to direct subject-to-BART sources to switch fuel forms, such as from coal or fuel oil to gas (40 CFR part 51, Appendix Y, section IV.D.1). Because natural gas has a sulfur content by weight that is well below 0.5%, the facility may elect to use this type of fuel to comply with BART, but we are not proposing to require a switch to natural gas for Unit 1. Switching to diesel is projected to result in considerable visibility improvement. The visibility improvement of switching to diesel is projected to range from 0.148 to 0.448 dv at each modeled Class I area, and the cumulative visibility improvement is 0.958 dv at the four affected Class I areas. The incremental visibility improvement of switching to diesel compared to switching to No. 6 fuel oil with a sulfur content of 0.5% is projected to range from 0.056 dv to 0.148 dv at each affected Class I area. However, the average cost-effectiveness of switching to diesel is estimated to be $7,145 and the incremental cost-effectiveness compared to No. 6 fuel oil with a sulfur content of 0.5% is $13,616 per ton of SO

2

removed, which we do not consider to be cost-effective in view of the incremental visibility improvement. Since diesel also has a sulfur content by weight that is well below 0.5%, the facility may elect to use this fuel type to comply with BART, but we are not proposing to require a switch to diesel for Unit 1. We are proposing to determine that SO

2

BART for McClellan Unit 1 is switching to fuels with 0.5% or lower sulfur content by weight. We propose to require that the facility purchase no fuel after the effective date of the rule that does not meet the sulfur content requirement and that 5 years from the effective date of the rule no fuel be burned that does not meet the requirement. We propose that any higher sulfur fuel oil that remains from the facility's 2009 fuel oil shipment cannot be burned past this point. As discussed above, the unit's baseline fuel is No. 6 fuel oil with 1.38% sulfur content, based on the average sulfur content of the fuel oil from the most recent shipment received by the facility in 2009. Based on our discussions with the facility, it is our understanding that the unit burns fuel oil primarily during periods of natural gas curtailment and during periodic testing and that the facility still has stockpiles of fuel oil from the most recent fuel oil shipment. Because the unit burns primarily natural gas and does not ordinarily burn fuel oil on a frequent basis, we believe it is appropriate to allow the facility 5 years to burn its existing supply of No. 6 fuel oil, as the normal course of business dictates and in accordance with any operating restrictions enforced by ADEQ. We believe that a shorter compliance date may result in the facility burning its existing supply of higher sulfur No. 6 fuel oil relatively quickly, resulting in a high amount of SO

2

emissions being emitted by the unit over a short period of time. This is not the intent of our regional haze regulations. We are also proposing regulatory text that includes monitoring, reporting, and recordkeeping

requirements associated with this proposed determination.

b.

Proposed BART Analysis and Determination for NO

X

. The AECC evaluation examined BART controls for NO

X

for McClellan Unit 1. McClellan Unit 1 does not currently have pollution control equipment for NO

X.

AECC identified all available control technologies, eliminated options that are not technically feasible, and evaluated the control effectiveness of the remaining control options. Each technically feasible control option was then evaluated in terms of a five factor BART analysis.

For NO

X

BART, the AECC evaluation considered both combustion and post-combustion controls. The combustion controls evaluated by AECC consisted of FGR, OFA, and LNB. The post-combustion controls evaluated consisted of SCR and SNCR. AECC found that some boilers may be restricted from installing OFA retrofits due to physical size and space restraints. For purposes of the NO

X

BART evaluation, AECC assumed OFA to be a technically feasible option for McClellan Unit 1, but noted that if OFA was determined to be BART based on the evaluation of the five BART factors, then further analyses would have to be performed to determine if: (1) The dimensions of McClellan's main boilers have sufficient upper furnace volume for OFA mixing and complete combustion and (2) the furnace meets the physical space requirements for OFA ports and air supply ducts. The remaining NO

X

control options were found to be technically feasible.

AECC evaluated three control scenarios: A combination of combustion controls (FGR, OFA, and LNB); the combination of combustion controls and SNCR; and SCR. Based on literature estimates, AECC found that the estimated NO

X

control range for oil and gas wall-fired boilers, such as McClellan Unit 1, is approximately 0.2-0.4 lb/MMBtu using FGR and 0.2-0.3 lb/MMBtu using OFA.

33

When LNB is combined with OFA and FGR, AECC estimated that a NO

X

controlled emission rate of 0.15-0.20 lb/MMBtu can be achieved at McClellan Unit 1. The NO

X

controlled emission rate of combustion controls combined with SNCR is estimated to be 0.10-0.12 lb/MMBtu. The NO

X

control efficiency of SCR is estimated to be 80-90% for gas fired boilers and 70-80% for oil fired boilers, which corresponds to a controlled emission rate of 0.05-0.12 lb/MMBtu for McClellan Unit 1.

33

“Controlling Nitrogen Oxides Under the Clean Air Act: A Menu of Options,” section II, dated July 1994, State and Territorial Air Pollution Program Administrators (STAPPA) and Association of Local Air Pollution Control Officials (ALAPCO).

AECC's cost analysis for NO

X

controls was based on “budgetary” cost estimates it obtained from the pollution control vendor, Babcock Power Systems. AECC estimated the capital and operating costs of controls based on the vendor's estimates, engineering estimates, and published calculation methods using the EPA Control Cost Manual. We are not aware of any enforceable shutdown date for the McClellan Generating Station, nor did AECC's evaluation indicate any future planned shutdown. This means that the anticipated useful life of the boiler is expected to be at least as long as the capital cost recovery period of controls. Therefore, a 30-year amortization period was assumed in the NO

X

BART analysis as the remaining useful life of Unit 1. The table below summarizes the estimated cost for installation and operation of NO

X

controls for McClellan Unit 1.

Table 19—Summary of NO

X

Control Costs for AECC McClellan Unit 1

Control scenario

Baseline

emission rate

(NO

X

tpy)

Natural gas

controlled

emission level

(lb/MMBtu)

Fuel oil

controlled

emission level

(lb/MMBtu)

Controlled

emission rate

(tpy)

Annual

emissions

reductions

(NO

X

tpy)

Total

annual cost

($/yr)

Average cost

effectiveness

($/ton)

Incremental

cost

($/ton)

Combustion Controls

294.04

0.15

0.15

174.89

119.15

746,051

6,261

Combustion Controls + SNCR

294.04

0.12

0.10

136.40

157.64

1,990,988

12,630

32,344

SCR

294.04

0.05

0.12

64.98

229.06

1,732,870

7,565

−3,614

AECC estimated the average cost-effectiveness of installing and operating combustion controls to be $6,261 per ton of NO

X

removed. The combination of combustion controls and SNCR was estimated to cost $12,630 per ton of NO

X

removed, while SCR was estimated to cost $7,565 per ton of NO

X

removed. In its evaluation, AECC also explained that AECC expects the average cost-effectiveness of NO

X

controls to be lower (

i.e.,

greater dollars per ton removed) in future years due to projected reduced operation of the unit.

AECC did not identify any energy or non-air quality environmental impacts associated with the use of LNB, OFA, or FGR. As for SCR and SNCR, we are not aware of any unusual circumstances at the facility that could create non-air quality environmental impacts associated with the operation of these controls greater than experienced elsewhere and that may therefore provide a basis for their elimination as BART (40 CFR part 51, Appendix Y, section IV.D.4.i.2.). Therefore, we do not believe there are any energy or non-air quality environmental impacts associated with NO

X

controls at AECC McClellan Unit 1 that would affect our proposed BART determination.

AECC assessed the visibility improvement associated with NO

X

controls by modeling the NO

X

emission rates associated with each control option using CALPUFF, and then comparing the visibility impairment associated with the baseline emission rates to the visibility impairment associated with the controlled emission rates as measured by the 98th percentile modeled visibility impact. The tables below show a comparison of the baseline (

i.e.,

existing) visibility impacts and the visibility impacts associated with NO

X

controls.

Table 20—AECC McClellan Unit 1: Summary of the 98th Percentile Visibility Impacts and Improvement Due to NO

X

Controls—Natural Gas Firing

[2001-2003]

Class I area

Baseline

visibility impact

(Δdv)

Combustion

controls

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Combustion

controls + SNCR

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

SCR

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Caney Creek

0.125

0.068

0.057

0.056

0.069

0.027

0.098

Upper Buffalo

0.052

0.028

0.024

0.023

0.029

0.012

0.04

Hercules-Glades

0.040

0.021

0.019

0.018

0.022

0.009

0.031

Mingo

0.058

0.031

0.027

0.026

0.032

0.012

0.046

Cumulative Visibility Improvement (Δdv)

0.127

0.152

0.215

Table 21—AECC McClellan Unit 1: Summary of the 98th Percentile Visibility Impacts and Improvement Due to NO

X

Controls—Fuel Oil Firing

[2001-2003]

Class I area

Baseline

visibility impact

(Δdv)

Combustion

controls

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Combustion

controls + SNCR

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

SCR

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Caney Creek

0.621

0.554

0.067

0.542

0.079

0.548

0.073

Upper Buffalo

0.266

0.264

0.002

0.264

0.002

0.264

0.002

Hercules-Glades

0.230

0.209

0.021

0.203

0.027

0.207

0.023

Mingo

0.227

0.203

0.024

0.200

0.027

0.201

0.026

Cumulative Visibility Improvement (Δdv)

0.114

0.135

0.124

The tables above show that the installation and operation of NO

X

controls is projected to result in a very modest visibility improvement from the baseline. Combustion controls at McClellan Unit 1 are projected to result in visibility improvement of up to 0.057 dv at any single Class I area for the natural gas firing scenario and 0.067 dv for the fuel oil firing scenario. A combination of combustion controls and SNCR is projected to result in only slight incremental visibility improvement compared to combustion controls alone. For example, a combination of combustion controls and SNCR at McClellan Unit 1 is projected to result in visibility improvement of up to 0.069 dv at any single Class I area for natural gas firing and 0.079 dv for fuel oil firing, which is an incremental visibility improvement for each fuel firing scenario of 0.012 dv going from combustion controls to combustion controls in combination with SNCR. Similarly, the installation and operation of SCR is projected to result in only slight incremental visibility improvement compared to a combination of combustion controls and SNCR, except for the fuel oil firing scenario. For the fuel oil firing scenario, SCR is projected to result in slightly less than or equal visibility improvement than a combination of combustion controls and SNCR.

Our Proposed NO

X

BART Determination:

Taking into consideration the five factors, we are proposing to determine that NO

X

BART for McClellan Unit 1 is no additional controls, and are proposing that the facility's existing NO

X

emission limits satisfy BART for NO

X

. We are proposing the existing emission limits of 869.1 lb/hr for natural gas firing and 705.8 lb/hr for fuel oil firing for NO

X

BART for McClellan Unit 1.

34

As discussed above, the operation of combustion controls at McClellan Unit 1 is projected to result in a maximum visibility improvement of 0.067 dv (Caney Creek), and a smaller amount of visibility improvement at each of the other Class I areas. The installation and operation of combustion controls at McClellan Unit 1 has an average cost-effectiveness of $6,261 per ton of NO

X

removed, which is not within the range of what we generally consider to be cost-effective. We believe the relatively small visibility benefit projected from the operation of combustion controls both when combusting fuel oil and natural gas does not justify the high estimated cost of those controls. The operation of a combination of combustion controls and SNCR is estimated to cost $12,630 per ton of NO

X

removed, which is not cost-effective. A combination of combustion controls and SNCR is projected to result in only slight incremental visibility benefit compared to combustion controls alone. The operation of SCR is estimated to cost $7,565 per ton of NO

X

removed, which is not generally considered cost-effective, and is projected to result in only slight incremental visibility benefit compared to a combination of combustion controls and SNCR. We are proposing to find that NO

X

BART for McClellan Unit 1 is no additional controls and are proposing that the existing NO

X

emission limits of 869.1 lb/hr for natural gas firing and 705.8 lb/hr for fuel oil firing are BART for NO

X

and that compliance be demonstrated using the unit's existing CEMS. We are proposing that these emissions limitations be complied with for BART purposes from the date of effectiveness of the finalized

action. We are also proposing regulatory text that includes monitoring, reporting, and recordkeeping requirements associated with these emission limits.

34

See ADEQ Operating Air Permit No. 0181-AOP-R5, Section IV, Specific Condition No. 1, 3, and 13.

c.

Proposed BART Analysis and Determination for PM.

McClellan Unit 1 does not currently have pollution control equipment for PM. For PM BART, AECC's evaluation considered the following control technologies: Dry ESP, wet ESP, fabric filter, wet scrubber, cyclone (

i.e.,

mechanical collector), and fuel switching. Residual fuel, such as the baseline No. 6 fuel oil burned at McClellan Unit 1, has inherent ash that contributes to emissions of filterable PM. Reductions in filterable PM emissions are directly related to the sulfur content of the fuel.

35

Therefore, switching to No. 6 fuel oil with a lower sulfur content is expected to result in lower filterable PM emissions. The AECC evaluation considered switching to No. 6 fuel oil with 1% sulfur content by weight, No. 6 fuel oil with 0.5% sulfur content by weight, diesel, and natural gas. These are the same lower sulfur fuel types evaluated in the SO

2

BART analysis for the unit.

35

See “AP-42, Compilation of Air Pollutant Emission Factors,” section 1.3.3.1, and Table 1.3-1, available at

http://www.epa.gov/ttnchie1/ap42/.

The AECC evaluation noted that the particulate matter from oil-fired boilers tends to be sticky and small, affecting the collection efficiency of dry ESPs and fabric filters. Dry ESPs operate by placing a charge on the particles through a series of electrodes, and then capturing the charged particles on collection plates, while fabric filters work by filtering the PM in the flue gas through filter bags. The collected particles are periodically removed from the filter bag through a pulse jet or reverse flow mechanism. Because of the sticky nature of particles from oil-fired boilers, dry ESPs and fabric filters are deemed technically infeasible for use at McClellan Unit 1. Wet ESPs, cyclones, wet scrubbers, and fuel switching were identified as technically feasible options for McClellan Unit 1. AECC noted that although cyclones and wet scrubbers are considered technically feasible for use at these boiler types, they are not very efficient at controlling particles in the smaller size fraction, particularly particles smaller than a few microns. However, the majority of the PM emissions from McClellan Unit 1 are greater than a few microns in size.

AECC estimated that switching to a lower sulfur fuel has a PM control efficiency ranging from approximately 44%-99%, depending on the fuel type. The other technically feasible control technologies are estimated to have the following PM control efficiency: Wet ESP—up to 90%, cyclone—85%, and wet scrubber—55%.

AECC evaluated the capital costs, operating costs, and average cost-effectiveness of wet ESPs, cyclones, and wet scrubbers. AECC also evaluated the average cost-effectiveness of switching to No. 6 fuel oil with 1% sulfur content, No. 6 fuel oil with 0.5% sulfur content, diesel, and natural gas. AECC developed the capital and operating costs of a wet ESP and wet scrubber using the Electric Power Research Institute's (EPRI) Integrated Emissions Control Cost Estimating Workbook (IECCOST) Software. The capital costs of controls (except for fuel switching) were annualized over a 15-year period and then added to the annual operating costs to obtain the total annualized costs. The table below summarizes the average cost-effectiveness of PM controls. The average cost-effectiveness was determined by dividing the annualized cost of controls by the annual PM emissions reductions. The annual emissions reductions were determined by subtracting the estimated controlled annual emission rates from the baseline annual emission rates. AECC estimated the baseline and controlled annual emission rates by conducting a mass balance on the sulfur content of the various fuels evaluated.

We disagree with two aspects of AECC's cost evaluation for PM controls for McClellan Unit 1. First, the total annual cost numbers associated with fuel switching should be the same as those used in the SO

2

BART cost analysis (see Table 17). In earlier draft versions of AECC's analysis, which were provided to us for review, the cost numbers for fuel switching used in the PM and SO

2

BART analyses were identical. In response to comments provided by us, the total annual cost and average cost-effectiveness numbers for fuel switching were revised in the final version of AECC's SO

2

BART analysis. However, it appears that AECC overlooked updating these cost numbers in the final PM BART analysis.

36

In the table below, we have revised the total annual cost of fuel switching for the PM BART analysis to be consistent with the cost estimates from AECC's SO

2

BART analysis, and we have also updated the PM average cost-effectiveness values. The second aspect of AECC's cost evaluation for PM controls that we disagree with is the use of a 15-year capital cost recovery period for calculating the average cost-effectiveness of a wet ESP, wet scrubber, and cyclone. As previously discussed, we are not aware of any enforceable shutdown date for the AECC McClellan Generating Station, nor did AECC's BART evaluation indicate any future planned shutdown. Therefore, we believe that assuming a 30-year equipment life rather than a 15-year equipment life would be more appropriate for these control technologies. Extending the amortization period from 15 to 30 years has the effect of decreasing the total annual cost of each control option, thereby improving the average cost-effectiveness of controls (

i.e.,

less dollars per ton removed). However, after considering all five BART factors, we do not believe AECC's assumption of a 15-year amortization period has an impact on our proposed BART decision and therefore we did not revise the amortization period or the average cost-effectiveness calculations for the PM control equipment options. This is discussed in more detail below. The table below summarizes the estimated cost for fuel switching and the installation and operation of PM control equipment for McClellan Unit 1.

36

The final version of AECC's BART analysis for SO

2

and PM, upon which our analysis is largely based, is titled “BART Five Factor Analysis Arkansas Electric Cooperative Corporation Bailey and McClellan Generating Stations, March 2014, Version 4.” A copy of AECC's analysis can be found in the docket for our proposed rulemaking.

Table 22—Summary of Cost of PM Controls for AECC McClellan Unit 1

Control scenario

Baseline

emission rate

(PM tpy)

Control

efficiency

(%)

Controlled emission rate

(PM tpy)

Annual

emissions

reduction

(PM tpy)

Capital cost

($)

Total annual cost

($/yr)

Average PM cost

effectiveness

($/ton)

Incremental cost

effectiveness

(S/ton)

No. 6 Fuel oil—1% S

136.08

43.6

76.70

59.38

145,866

2,456

Wet Scrubber

136.08

55.0

61.23

74.84

146,303,011

52,056,542

695,549

3,357,741

No. 6 Fuel oil—0.5% S

136.08

82.4

23.94

112.14

510,532

4,553

−1,381,931

Cyclone

136.08

85.0

20.41

115.67

1,432,971

1,721,384

14,882

343,018

Wet ESP

136.08

90.0

13.61

122.47

151,509,333

32,605,907

266,237

4,541,842

Natural Gas

136.08

99.0

1.36

134.72

−2,926,874

−21,725

−2,900,635

Diesel

136.08

99.2

1.10

134.98

1,444,077

10,698

16,811,350

The table above shows that the average cost-effectiveness values of all add-on PM control technology options evaluated for McClellan Unit 1 ranged in cost-effectiveness from approximately $15,000 to $700,000 per ton of PM removed, based on AECC's cost estimates. The incremental cost-effectiveness of add-on PM control technology options ranged from $343,018 to $16,811,350 per ton of PM removed. Switching to No. 6 fuel oil with either a 1% or 0.5% sulfur content was found to be within the range of what we generally consider cost-effective for BART. Switching to No. 6 fuel oil with 1% sulfur content is estimated to cost $2,456 per ton of PM removed, while switching to No. 6 fuel oil with 0.5% sulfur content is estimated to cost $4,553 per ton of PM removed at McClellan Unit 1. As discussed in the SO

2

BART analysis, the current cost of natural gas is actually lower than the cost of the baseline fuel. Therefore, the average cost-effectiveness of switching from the baseline fuel to natural gas is denoted as a negative value in the table above. As discussed above, AECC also explained that it expects the average cost-effectiveness of PM control equipment to be lower (

i.e.,

greater dollars per ton removed) in future years due to projected reduced operation of the units due to a change in the management of the load control area the facilities are located in. Less projected operating time is expected to result in lower annual emissions, which in turn would result in decreased average cost-effectiveness for the add-on PM control technology options.

AECC did not identify any energy or non-air quality environmental impacts associated with fuel switching, but did identify impacts associated with the use of wet ESPs and wet scrubbers due to their electricity usage. Energy use in and of itself does not disqualify a technology (40 CFR part 51, Appendix Y, section IV.D.4.h.1.). In addition, the cost of the electricity needed to operate this equipment has already been factored into the cost of controls. AECC also noted that both wet ESPs and wet scrubbers generate wastewater streams that must either be treated on-site or sent to a waste water treatment plant, and the wastewater treatment process will generate a filter cake that would likely require landfilling. The BART Guidelines provide that the fact that a control device creates liquid and solid waste that must be disposed of does not necessarily argue against selection of that technology as BART, particularly if the control device has been applied to similar facilities elsewhere and the solid or liquid waste is similar to those other applications (40 CFR part 51, Appendix Y, section IV.D.4.i.2.). We are not aware of any unusual circumstances at the AECC McClellan Generating Station that could potentially create greater problems than experienced elsewhere related to the treatment of wastewater and any necessary landfilling, nor did the AECC BART evaluation discuss or mention any such unusual circumstances. Therefore, the need to treat wastewater or landfill any filter cake or other waste in and of itself does not provide a basis for disqualification or elimination of a wet ESP or wet scrubber.

As previously discussed, we are not aware of any enforceable shutdown date for the AECC McClellan Generating Station, nor did the AECC evaluation indicate any future planned shutdown. Therefore, we believe it is appropriate to assume a 30-year amortization period in the PM BART analysis as the remaining useful life of the unit. Assuming a 30-year amortization period, these controls would have a lower estimated total annual cost and would therefore have an improved cost-effectiveness (

i.e.,

less dollars per ton removed) compared to what was estimated in AECC's evaluation. However, we did not adjust the amortization period because we do not believe this has an impact on our proposed BART decision. As discussed in the subsection below, the visibility benefit expected from the installation and operation of PM control equipment is too small to justify the cost of these controls. Therefore, we did not revise the amortization period and the average cost-effectiveness calculations for the PM control equipment options.

As switching to lower sulfur fuels has impacts on both SO

2

and PM emissions, AECC's assessment of the visibility improvement associated with fuel switching is addressed in the SO

2

BART analysis for McClellan Unit 1. Table 18 summarizes the visibility improvement associated with controlled emission rates for SO

2

and PM as a result of fuel switching. AECC assessed the visibility improvement associated with wet ESPs, wet scrubbers, and cyclones by modeling the PM emission rates associated with each control option using CALPUFF, and then comparing the visibility impairment associated with the baseline emission rates to the visibility impairment associated with the controlled emission rates as measured by the 98th percentile modeled visibility impact. The controlled PM

10

emission rates associated with wet ESPs, wet scrubbers, and cyclones were calculated by reducing the uncontrolled annual PM

10

emission rates by the pollutant removal efficiency of each control technology. The SO

2

and NO

X

emission rates modeled in the controlled scenarios are the same as those from the baseline scenario, as it is assumed that SO

2

and NO

X

emissions would remain unchanged. The table below shows a comparison of the baseline (

i.e.,

existing) visibility impacts and the visibility impacts associated with PM controls.

Table 23—AECC McClellan Unit 1: Summary of the 98th Percentile Visibility Impacts and Improvement From PM Controls

Class I area

Baseline

visibility

impact

(Δdv)

Wet ESP

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Wet scrubber

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Cyclone

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

Caney Creek

0.621

0.617

0.004

0.619

0.002

0.619

0.002

Upper Buffalo

0.266

0.263

0.003

0.264

0.002

0.265

0.001

Hercules-Glades

0.230

0.227

0.003

0.228

0.002

0.229

0.001

Mingo

0.227

0.223

0.004

0.224

0.003

0.225

0.002

Cumulative Visibility Improvement (Δdv)

0.014

0.009

0.006

The table above shows that the operation of a wet ESP, wet scrubber, and cyclone at McClellan Unit 1 is projected to result in minimal visibility improvement at the four affected Class I areas. The modeled visibility improvement from switching to No. 6 fuel oil with 1% sulfur content; No. 6 fuel oil with 0.5% sulfur content; diesel; and natural gas are summarized in Table 18. The modeled visibility improvement shown in Table 18 reflects both SO

2

and PM emissions reductions as a result of switching to fuels with lower sulfur content. However, the majority of the baseline visibility impact at each Class I area when burning the baseline fuel oil is due to SO

2

emissions, while PM

10

emissions contribute only a small portion of the baseline visibility impacts at each Class I area (see Table 15). Accordingly, the majority of the visibility improvement associated with switching to lower sulfur fuels can reasonably be expected to be the result of a reduction in SO

2

emissions.

Our Proposed PM BART Determination:

Taking into consideration the five factors, we propose to determine that PM BART for AECC McClellan Unit 1 does not require add-on controls. Consistent with our proposed determination for SO

2

BART, we are proposing that PM BART is satisfied by Unit 1 switching to fuels with 0.5% or lower sulfur content by weight. As discussed above, we disagree with AECC's use of a 15-year amortization period in the cost analysis for a wet ESP, wet scrubber, and cyclone. Assuming a 30-year amortization period, these controls would have a lower estimated total annual cost and would therefore have an improved cost-effectiveness (

i.e.,

less dollars per ton removed) compared to what was estimated in AECC's evaluation. However, after considering all five BART factors, even if we revised AECC's cost estimates to reflect a 30-year amortization period, resulting in a lower total annual cost and improved cost-effectiveness, we would still not be able to justify the cost in light of the minimal visibility benefit of these controls (see the table above).

We are proposing to determine that PM BART for McClellan Unit 1 is switching to fuels with 0.5% or lower sulfur content by weight. We propose to require that the facility purchase no fuel after the effective date of the rule that does not meet the sulfur content requirement and that 5 years from the effective date of the rule no fuel be burned that does not meet the requirement. We propose that any higher sulfur fuel oil that remains from the facility's 2009 fuel oil shipment cannot be burned past this point. As discussed above, the unit's baseline fuel is No. 6 fuel oil with 1.38% sulfur content, based on the average sulfur content of the fuel oil from the most recent shipment received by the facility in 2009. Based on our discussions with the facility, it is our understanding that the unit burns fuel oil primarily during periods of natural gas curtailment and during periodic testing and that the facility still has stockpiles of fuel oil from the most recent fuel oil shipment. Because the unit burns primarily natural gas and does not ordinarily burn fuel oil on a frequent basis, we believe it is appropriate to allow the facility 5 years to burn its existing supply of No. 6 fuel oil, as the normal course of business dictates and in accordance with any operating restrictions enforced by ADEQ. We believe that a shorter compliance date may result in the facility burning its existing supply of higher sulfur No. 6 fuel oil relatively quickly, resulting in a high amount of SO

2

emissions being emitted by the unit over a short period of time. This is not the intent of our regional haze regulations. We are also proposing regulatory text that includes monitoring, reporting, and recordkeeping requirements associated with this proposed determination.

3. AEP Flint Creek Power Plant

The AEP Flint Creek Power Plant Unit 1 is subject to BART. We previously disapproved Arkansas' BART determination for SO

2

and NO

X

for Flint Creek Unit 1 in our March 12, 2012 final action (77 FR 14604). Flint Creek Unit 1 is a dry bottom wall-fired boiler with a nominal generating capacity rating of 558 MW and a nominal design maximum heat input rate of 6,324 MMBtu/hr. The unit burns primarily low-sulfur western coal and is currently equipped with an ESP and low NO

X

burners. AEP hired a consultant to prepare a BART five-factor analysis for the AEP Flint Creek Unit 1 (AEP BART analysis).

37

37

See “BART Five Factor Analysis Flint Creek Power Plant Gentry, Arkansas (AFIN 04-00107),” dated September 2013, Version 4, prepared by Trinity Consultants Inc. in conjunction with American Electric Power Service Corporation for the Southwestern Electric Power Company Flint Creek Power Plant. A copy of this BART analysis is found in the docket for our proposed rulemaking.

The table below summarizes the baseline emission rates for this source. The SO

2

and NO

X

baseline emission rates are the highest actual 24-hour emission rates based on 2001-2003 CEMS data. The emission rates for the PM

10

species reflect the breakdown of the filterable and condensable PM

10

determined from the National Park Service (NPS) “speciation spreadsheet” for

Dry Bottom Boiler Burning Pulverized Coal using only ESP.

38

The sulfate (SO

4

) emission rate was calculated using an EPRI methodology that considers the SO

2

to SO

4

conversion rate and SO

4

reduction

factors for various downstream equipment.

39

38

The NPS Workbook, “PC Dry Bottom ESP Example.xls” updated 03/2006, was obtained from the NPS Web site:

http://www.nature.nps.gov/air/Permits/ect/index.cfm.

Trinity input the following parameters into the workbook for speciation determination: total PM

10

emission rate of 192.5 lb/hr, heat value of 8,500 Btu/lb, sulfur content of 0.31%, ash content of 4.9%.

39

Electric Power Research Institute (EPRI) Estimating Total Sulfuric Acid Emissions from Stationary Power Plants: Version 2010a. EPRI, Palo Alto, CA: 2010.

Table 24—AEP Flint Creek Unit 1: Baseline Maximum 24-Hour Emission Rates

Source

SO

2

(lb/hr)

SO

4

(lb/hr)

NO

X

(lb/hr)

PMc

(lb/hr)

PMf

(lb/hr)

SOA

(lb/hr)

EC

(lb/hr)

Unit 1 (SN-01)

4,728.4

3.1

1,945.0

65.1

50.1

15.1

1.9

AEP modeled the baseline emission rates using the CALPUFF dispersion model to determine the baseline visibility impairment attributable to Flint Creek Unit 1 at the four Class I areas impacted by emissions from BART sources in Arkansas. These Class I areas are the Caney Creek Wilderness Area, Upper Buffalo Wilderness Area, Hercules-Glades Wilderness Area, and Mingo National Wildlife Refuge. The baseline (

i.e.,

2001-2003) visibility impairment attributable to the source at each Class I area is summarized in the table below.

Table 25—Baseline Visibility Impairment Attributable to AEP Flint Creek Unit 1

[2001-2003]

Unit

Caney Creek

Upper Buffalo

Hercules-Glades

Mingo

AEP Flint Creek Unit 1:

Maximum (Δdv)

1.318

2.426

2.103

1.488

98th Percentile (Δdv)

0.963

0.965

0.657

0.631

a.

Proposed BART Analysis and Determination for SO

2

. AEP identified all available control technologies, eliminated options that are not technically feasible, and evaluated the control effectiveness of the remaining control options. Each technically feasible control option was then evaluated in terms of a five factor BART analysis.

The AEP evaluation considered Dry Sorbent Injection (DSI), dry FGD (

i.e.,

dry scrubber), and wet FGD (

i.e.,

wet scrubber) for SO

2

BART. All three options were identified as technically feasible for use at Flint Creek Unit 1. The AEP evaluation noted that depending on residence time, gas stream temperature, and limitations of the particulate control device, DSI control efficiency can range between 40 to 60%.

40

Dry FGD control efficiency generally ranges from 60 to 95%. There are various designs of dry FGD systems, including Spray Dryer Absorber (SDA), Circulating Dry Scrubbing (CDS), and Novel Integrated Desulfurization (NID) technology. According to AEP's evaluation, discussions with vendors indicated that an outlet emission rate of 0.06 lb/MMBtu at Flint Creek Unit 1 would be achievable with NID technology. AEP noted that it has no data to suggest that lower emission levels are sustainably achievable with the NID technology in a retrofit application, and that equipment vendors did not guarantee better performance than this. An emission rate of 0.06 lb/MMBtu represents 92% control from the unit's baseline 30-day average rate of 0.75 lb/MMBtu. AEP's analysis notes that dry FGD using lime as the reagent is capable of achieving 80 to 95% control when used with lower sulfur coals such as those burned at Flint Creek Unit 1. The remainder of AEP's analysis focused on wet FGD and dry FGD (NID). We concur with AEP's decision to focus the remainder of the analysis on the two control options with the highest control efficiency.

40

“Assessment of Control Technology Options for BART-Eligible Sources: Steam Electric Boilers, Industrial Boilers, Cement Plants and Paper and Pulp Facilities” Northeast States for Coordinated Air Use Management (NESCAUM), March 2005.

The estimated capital and operating costs of wet FGD and dry FGD (NID) developed by AEP and used in the cost-effectiveness calculations were based on EPA's Control Cost Manual and supplemented, where available, with vendor and site-specific information obtained by AEP. AEP annualized the capital cost of controls over a 30-year amortization period and then added these to the annual operating costs to obtain the total annualized costs. The average cost-effectiveness was calculated by dividing the total annualized cost of controls by the annual SO

2

emissions reductions. AEP estimated the average cost-effectiveness of a wet FGD system to be $4,919 per ton of SO

2

removed, while the average cost-effectiveness of NID was estimated to be $3,845 per ton of SO

2

removed (see table below).

We disagree with one aspect of AEP's cost analysis.

41

AEP's cost estimates are based on 2016 dollars, which means that they were escalated to a future build date. BART cost analyses should be based on present dollars, and the EPA Control Cost Manual approach explicitly excludes future escalation, as cost comparisons should be made on a current real dollar basis. Escalation of costs from past to the current year of analysis is permitted, as costs are compared based on the time of estimate, but future escalation is not allowed. We expect that de-escalation to 2014 dollars would result in lower cost numbers and overall lower average cost-effectiveness values for all controls evaluated. We believe that wet FGD and NID are both more cost-effective (

i.e.,

less dollars per SO

2

ton removed) than what has been estimated by AEP. However, we did not adjust the cost numbers and the cost-effectiveness values because we do not expect this to change our proposed BART decision. This is discussed in more detail below in the subsection titled “Our Proposed SO

2

BART Determination.”

41

See “BART Five Factor Analysis Flint Creek Power Plant Gentry, Arkansas (AFIN 04-00107),” dated September 2013, Version 4, prepared by Trinity Consultants Inc. in conjunction with American Electric Power Service Corporation for the Southwestern Electric Power Company Flint Creek Power Plant. AEP's SO

2

control cost calculations are found in Appendix A of the BART analysis. An Excel file titled “Consolidated Spreadsheet_2013-09-09” containing spreadsheets with cost information was also provided by AEP Flint Creek in support of the cost analysis. A copy of the BART analysis and the Excel file is found in the docket for our proposed rulemaking.

Table 26—Summary of Cost of SO

2

Controls for AEP Flint Creek Unit 1

Control technology

Baseline

emission rate

(SO

2

tpy)

Controlled emission rate

(SO

2

lb/MMBtu)

Controlled emission rate

(SO

2

tpy)

Annual

emissions

reductions

(SO

2

tpy)

Capital cost

($)

Total annual cost

($/yr)

Average cost effectiveness

($/ton)

Incremental cost-

effectiveness

($/ton)

NID

11,641

0.06

1,120

10,521

281,738,024

40,448,089

3,845

Wet Scrubber

11,641

0.04

747

10,894

374,427,351

53,592,663

4,919

35,240

AEP's evaluation noted that the potential negative energy and non-air quality environmental impacts are greater with wet FGD systems than dry FGD systems. AEP noted that wet FGD requires increased water use and generates large volumes of wastewater and solid waste/sludge that must be treated or stabilized before landfilling, placing additional burden on the wastewater treatment and solid waste management capabilities. We do not expect that water availability would affect the feasibility of wet FGD at Flint Creek Unit 1 because the facility is not located in an exceptionally arid region. Additionally, the BART Guidelines provide that the fact that a control device creates liquid and solid waste that must be disposed of does not necessarily argue against selection of that technology as BART, particularly if the control device has been applied to similar facilities elsewhere (40 CFR part 51, Appendix Y, section IV.D.4.i.2.). In cases where the facility can demonstrate that there are unusual circumstances that would create greater problems than experienced elsewhere, this may provide a basis for the elimination of that control option as BART. But in this case, AEP has not indicated that there are any such unusual circumstances. Another potential negative energy and non-air quality environmental impact associated with wet FGD is the potential for increased power requirements and greater reagent usage compared to dry FGD. The costs associated with increased power requirements and greater reagent usage have already been factored into the cost analysis for wet FGD.

AEP assessed the visibility improvement associated with wet FGD and NID technology by modeling the SO

2

emission rates associated with each control option using CALPUFF, and then comparing the visibility impairment associated with the baseline emission rates to the visibility impairment associated with the controlled emission rates as measured by the 98th percentile modeled visibility impact. The table below compares the baseline (

i.e.,

existing) visibility impacts with the visibility impacts associated with SO

2

controls.

Table 27—AEP Flint Creek Unit 1: Summary of the 98th Percentile Visibility Impacts and Improvement Due to SO

2

Controls

Class I area

Baseline

visibility impact

(Δdv)

NID Technology

Visibility

impact

(Δdv)

Visibility

improvement from baseline

(Δdv)

Wet scrubber

Visibility

impact

(Δdv)

Visibility

improvement from baseline

(Δdv)

Caney Creek

0.963

0.348

0.615

0.334

0.629

Upper Buffalo

0.965

0.501

0.464

0.488

0.477

Hercules-Glades

0.657

0.312

0.345

0.305

0.352

Mingo

0.631

0.217

0.414

0.208

0.423

Cumulative Visibility Improvement (Δdv)

1.838

1.881

The table above shows that the installation and operation of SO

2

controls is projected to result in considerable visibility improvement from the baseline at the four impacted Class I areas. Installation and operation of NID technology is projected to result in visibility improvement of up to 0.615 dv at any single Class I area (based on the 98th percentile modeled visibility impacts), while wet FGD is projected to result in visibility improvement of up to 0.629 dv. Wet FGD is projected to result in very minimal incremental visibility benefit over NID technology, with the projected incremental visibility improvement over NID ranging from 0.007 to 0.014 dv at each Class I area.

Our Proposed SO

2

BART Determination:

Taking into consideration the five factors, we propose to determine that BART for AEP Flint Creek Unit 1 is an emission limit of 0.06 lb/MMBtu on a 30 boiler-operating-day rolling average based on the installation and operation of NID. The operation of NID is projected to result in visibility improvement ranging from 0.352 to 0.629 dv at each affected Class I area (98th percentile basis), and based on AEP's evaluation, is estimated to have an average cost-effectiveness of $3,845 per ton of SO

2

removed. By comparison, AEP estimated wet FGD to have an average cost-effectiveness of $4,919 per ton of SO

2

removed and the incremental cost-effectiveness of wet FGD compared to NID is estimated to be $35,240 per ton of SO

2

removed. As discussed above, we believe that AEP's escalation of the cost of controls to 2016 dollars has likely resulted in the over-estimation of the average cost-effectiveness. Therefore, we believe wet FGD and NID are both more cost-effective (

i.e.,

less dollars per ton of SO

2

removed) than estimated by AEP (see table above). However, we did not adjust the cost numbers and cost-effectiveness calculations because we do not believe that doing so would change our proposed BART determination. We believe that the average cost-effectiveness of both control options was likely over-estimated and the costs associated with wet FGD would continue to be higher than the costs associated with NID if the estimates were adjusted, yet the installation and operation of wet FGD is projected to result in minimal incremental visibility improvement over NID. We are proposing to determine that SO

2

BART

for Flint Creek Unit 1 is an emission limit of 0.06 lb/MMBtu on a 30 boiler-operating-day rolling average based on the installation and operation of NID. We believe that the full compliance time

42

of 5 years is warranted for a new scrubber retrofit and so propose to require compliance with this requirement no later than 5 years from the effective date of the final rule. We are proposing to require that compliance be demonstrated using the unit's existing CEMS. We are also proposing regulatory text that includes monitoring, reporting, and recordkeeping requirements associated with this emission limit.

42

Section 51.308(e)(1)(iv), requires, “each source subject to BART be required to install and operate BART as expeditiously as practicable, but in no event later than 5 years after approval of the implementation plan revision.”

b.

Proposed BART Analysis and Determination for NO

X

.

AEP's BART evaluation examined BART controls for NO

X

for Flint Creek Unit 1 by identifying all available control technologies, eliminating options that are not technically feasible, and evaluating the control effectiveness of the remaining control options. Each technically feasible control option was then evaluated in terms of a five factor BART analysis.

For NO

X

BART, the AEP evaluation considered both combustion and post-combustion controls. The combustion controls considered by AEP consisted of FGR, OFA, and LNB. The post-combustion controls considered consisted of SCR and SNCR. All control options evaluated were found to be technically feasible. AEP estimated that FGR would be able to achieve a controlled emission rate of 0.23-0.29 lb/MMBtu at Unit 1, which is a less stringent emission rate than would be achieved with LNB/OFA. Therefore, FRG was not further considered in the BART evaluation, while LNB/OFA were further considered. AEP evaluated three control scenarios: (1) LNB with OFA (LNB/OFA); (2) the combination of LNB with OFA and SNCR (LNB/OFA + SNCR); and (3) SCR. The baseline NO

X

emission rate assumed by AEP in the analysis is 0.31 lb/MMBtu. AEP estimated that the installation and operation of LNB/OFA at Flint Creek Unit 1 would achieve a NO

X

control level of approximately 0.23 lb/MMBtu on a 30 boiler-operating-day averaging basis. It also estimated that LNB/OFA + SNCR would achieve a NO

X

control level of approximately 0.20 lb/MMBtu, and that SCR would achieve a NO

X

control level of approximately 0.07 lb/MMBtu, also on a 30 boiler-operating-day averaging basis.

AEP estimated the capital costs, operating costs, and average cost-effectiveness of controls based on vendor estimates and published calculation methods. AEP noted that the EPA Control Cost Manual was followed to the extent possible and estimates were supplemented with vendor and site-specific information where available. The cost analysis assumed a 30-year amortization period for LNB/OFA and for SCR, and a 20-year amortization period for SNCR. We discuss the appropriateness of the choice of amortization periods below. The total annual costs were estimated by annualizing the capital cost of controls over either a 30-year or 20-year period and then adding to this value the annual operating cost of controls. AEP determined the annual tons reduced associated with each NO

X

control option by subtracting the estimated controlled annual emission rate from the baseline annual emission rate. The baseline annual emission rate is the average rate as reported by AEP Flint Creek in the 2001-2003 air emission inventories. The average cost-effectiveness of NO

X

controls was calculated by dividing the total annual cost of each control option by the estimated annual NO

X

emissions reductions. The table below summarizes the average-cost effectiveness of NO

X

controls for Flint Creek Unit 1.

Table 28—Summary of NO

X

Control Costs for Flint Creek Unit 1

Control technology

Baseline

emission rate

(NO

X

tpy)

Controlled emission level

(NO

X

lb/MMBtu)

Controlled emission rate

(NO

X

tpy)

Annual

emissions

reductions

(NO

X

tpy)

Capital cost

($)

Total annual cost

($/yr)

Average cost effectiveness

($/ton)

Incremental

cost-

effectiveness

($/ton)

LNB/OFA

5,120

0.23

4,295

826

16,000,000

1,454,621

1,761

LNB/OFA/SNCR

5,120

0.20

3,772

1,348

23,124,235

4,177,782

3,099

5,217

SCR

5,120

0.07

1,251

3,869

121,440,000

13,769,599

3,559

3,805

AEP estimated the average cost-effectiveness of installing and operating LNB/OFA to be $1,761 per ton of NO

X

removed, while the combination of LNB/OFA + SNCR is estimated to cost $3,099 per ton of NO

X

removed, and SCR is estimated to cost $3,559 per ton of NO

X

removed.

AEP did not identify any energy or non-air quality environmental impacts associated with the use of LNB/OFA. As for SCR and SNCR, we are not aware of any unusual circumstances at the facility that could create non-air quality environmental impacts associated with the operation of these controls greater than experienced elsewhere and that may therefore provide a basis for their elimination as BART (40 CFR part 51, Appendix Y, section IV.D.4.i.2.). Therefore, we do not believe there are any energy or non-air quality environmental impacts associated with the operation of NO

X

controls at AEP Flint Creek Unit 1 that would affect our proposed BART determination.

Flint Creek Unit 1 is currently equipped with early generation low NO

X

burners for control of NO

X

emissions. Consideration of the presence of existing pollution control technology at each source is reflected in the BART analysis in two ways: First, in the consideration of available control technologies, and second, in the development of baseline emission rates for use in cost calculations and visibility modeling. The baseline emission rate used in the cost calculations and visibility modeling reflects the operation of these controls. The newer generation low NO

X

burners evaluated by AEP are expected to achieve a higher level of NO

X

control than the currently installed early generation low NO

X

burners.

We are not aware of any enforceable shutdown date for the AEP Flint Creek Power Plant, nor did AEP's evaluation indicate any future planned shutdown. This means that the anticipated useful life of the boiler is expected to be at least as long as the capital cost recovery period of controls. AEP assumed a 30-year amortization period in the evaluation of LNB, OFA, and SCR as the remaining useful life of the unit, and a 20-year amortization period in the evaluation of SNCR. We disagree with AEP's assumption of a 20-year amortization period in the cost analysis of SNCR. Any air pollution controls on the unit are expected to have the same

life as the boiler. Therefore, we believe it is appropriate to assume a 30-year amortization period for SNCR, as was done for SCR and combustion controls. Assuming a 30-year amortization period, SNCR would have a lower estimated total annual cost and would therefore have an improved cost-effectiveness (

i.e.,

less dollars per ton removed) compared to what was estimated in AEP's evaluation. However, we did not adjust the amortization period assumed in AEP's evaluation because we do not believe this has an impact on our proposed BART decision. As discussed in the subsection below, the incremental visibility benefit expected from the installation and operation of SNCR is too small to justify the cost of this control compared to combustion controls alone. Therefore, we did not revise the amortization period and the average cost-effectiveness calculations for SNCR.

AEP assessed the visibility improvement associated with NO

X

controls by modeling the NO

X

emission rates associated with each control option using CALPUFF, and then comparing the visibility impairment associated with the baseline emission rate to the visibility impairment associated with the controlled emission rates as measured by the 98th percentile modeled visibility impact. The table below shows a comparison of the baseline (

i.e.,

existing) visibility impacts and the visibility impacts associated with NO

X

controls.

Table 29—AEP Flint Creek Unit 1: Summary of the 98th Percentile Visibility Impacts and Improvement due to NO

X

Controls

Class I area

Baseline

visibility

impact

(Δdv)

LNB/OFA

Visibility

impact

(Δdv)

Visibility

improvement

from baseline

(Δdv)

LNB/OFA + SNCR

Visibility

impact

(Δdv)

Visibility

Improvement

from Baseline

(Δdv)

SCR

Visibility

Impact

(Δdv)

Visibility

Improvement

from Baseline

(Δdv)

Caney Creek

0.963

0.882*

0.081*

0.849

0.114

0.718

0.245

Upper Buffalo

0.965

0.939

0.026

0.932

0.033

0.895

0.07

Hercules-Glades

0.657

0.633

0.024

0.623

0.034

0.573

0.084

Mingo

0.631

0.617

0.014

0.612

0.019

0.588

0.043

Cumulative Visibility Improvement (Δdv)

0.145

0.2

0.442

* EPA identified a discrepancy in the results presented by AEP and reran the model for the 2003 model year. These values have been adjusted to reflect the results of the EPA model run.

As shown in the table above, the installation and operation of LNB/OFA is projected to result in visibility improvement of up to 0.081 dv at any single Class I area, based on the 98th percentile visibility impairment. The installation and operation of LNB/OFA + SNCR is projected to result in visibility improvement of up to 0.114 dv over the baseline. The installation and operation of SCR is projected to result in visibility improvement of up to 0.245 dv in any single Class I area. The combination of LNB/OFA + SNCR would result in slight incremental visibility benefit over LNB/OFA at Caney Creek and in negligible incremental visibility benefit at the other three affected Class I areas. SCR would result in 0.131 dv incremental visibility benefit over LNB/OFA + SNCR at Caney Creek and less than half as much incremental visibility benefit at the other three affected Class I areas.

Our Proposed NO

X

BART Determination:

Taking into consideration the five factors, we propose to determine that NO

X

BART for Flint Creek Unit 1 is an emission limit of 0.23 lb/MMBtu on a 30 boiler-operating-day rolling average based on the installation and operation of new LNB/OFA. The operation of new LNB/OFA is projected to result in visibility improvement ranging from 0.014 to 0.081 dv at each affected Class I area (98th percentile basis) and is projected to have a cumulative visibility improvement of 0.145 dv across the four affected Class I areas. The operation of LNB/OFA is estimated to have an average cost-effectiveness of $1,761 per ton of NO

X

removed, which we consider to be very cost-effective. By comparison, the operation of LNB/OFA + SNCR is projected to result in small incremental visibility improvement over LNB/OFA, but is estimated to have an average cost-effectiveness of $3,099 per ton of NO

X

removed and an incremental cost-effectiveness of $5,217 per ton of NO

X

removed. We believe that AEP's assumption of a 20-year amortization period for SNCR has likely resulted in lower cost-effectiveness for SNCR. Therefore, we believe LNB/OFA + SNCR is more cost-effective (

i.e.,

less dollars per ton of NO

X

removed) than estimated by AEP (see table above). However, we did not adjust the cost numbers and cost-effectiveness values because we do not believe that doing so would change our proposed BART determination, as the installation and operation of LNB/OFA + SNCR is projected to result in minimal incremental visibility improvement over LNB/OFA alone such that the additional cost of SNCR is not justified.

The operation of SCR is projected to result in visibility improvement ranging from 0.043 to 0.245 dv at each Class I area, and has an average cost-effectiveness of $3,559 per ton of NO

X

removed. The incremental visibility benefit of SCR compared to LNB/OFA + SNCR is projected to be 0.131 dv at Caney Creek and is projected to range from 0.024 to 0.05 dv at the remaining Class I areas. The incremental cost-effectiveness of SCR is estimated to be $3,805 per ton of NO

X

removed. Although we are not adjusting the cost estimate for the reason discussed above, we note that AEP's assumption of a 20-year amortization period for SNCR has the effect of

This text is long and has been trimmed here. Open the source document for the complete record.