Approval and Promulgation of Air Quality Implementation Plans; States of Minnesota and Michigan; Regional Haze State Implementation Plan; Federal Implementation Plan for Regional Haze

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 52

[EPA-R05-OAR-2010-0954; EPA-R05-OAR-2010-0037; FRL-9773-1]

Approval and Promulgation of Air Quality Implementation Plans; States of Minnesota and Michigan; Regional Haze State Implementation Plan; Federal Implementation Plan for Regional Haze

AGENCY:

Environmental Protection Agency (EPA).

ACTION:

Final rule.

SUMMARY:

EPA is finalizing a Federal Implementation Plan (FIP) to implement emission limits that represent Best Available Retrofit Technology (BART) for certain taconite ore processing facilities in Minnesota and Michigan. The Clean Air Act (CAA or the “Act”) and the regional haze rule require implementation plans to contain BART emission limits for sources subject to BART in order to meet the national goal of preventing any future and remedying any existing impairment of visibility in mandatory class I Federal areas arising from manmade air pollution.

DATES:

This final rule is effective on March 8, 2013.

ADDRESSES:

EPA has established a docket for this action under Docket ID Nos. EPA-R05-OAR-2010-0954 and EPA-RO5-OAR-2010-0037. All documents are listed on the

www.regulations.gov

Web site. Although listed in the index, some information is not publicly available, i.e., Confidential Business Information (CBI) or other information whose disclosure is restricted by statute. Certain other material, such as copyrighted material, is not placed on the Internet and will be publicly available only in hard copy form. Publicly available docket materials are available either electronically through

www.regulations.gov

or in hard copy at the Environmental Protection Agency, Region 5, Air and Radiation Division, 77 West Jackson Boulevard, Chicago, Illinois 60604. This facility is open from 8:30 a.m. to 4:30 p.m., Monday through Friday, excluding Federal holidays. We recommend that you telephone Steven Rosenthal, Environmental Engineer, at (312) 886-6052 before visiting the Region 5 office.

FOR FURTHER INFORMATION CONTACT:

Steven Rosenthal, Environmental Engineer, Attainment Planning & Maintenance Section, Air Programs Branch (AR-18J), U.S. Environmental Protection Agency, Region 5, 77 West Jackson Boulevard, Chicago, Illinois 60604, (312) 886-6052,

rosenthal.steven@epa.gov.

SUPPLEMENTARY INFORMATION:

Throughout this document, wherever “we,” “us,” or “our,” is used, we mean the United States Environmental Protection Agency (EPA).

Table of Contents

I. Background Information

A. EPA's Regional Haze Rule and Best Available Retrofit Technology

B. EPA's Legal Authority To Promulgate a FIP

C. Minnesota and Michigan's Regional Haze SIP Submittals

D. EPA's Regional Haze FIP and Related Actions

II. Comments and Responses

III. What action is EPA taking?

A. NO

X

Limits

B. SO

2

Limits

C. CEMS

D. Visibility Benefits and Cost Effectiveness

E. Proposed Disapproval of the States' BART Determinations for Taconite Facilities

IV. Statutory and Executive Order Reviews

I. Background Information

A. EPA's Regional Haze Rule and Best Available Retrofit Technology

The regional haze rule required states to submit State Implementation Plans (SIPs) to implement the rule's requirements by no later than December 17, 2007. Neither Minnesota nor Michigan submitted regional haze SIPs by the required date. The Act requires EPA to promulgate a FIP within two years after EPA finds that a state has failed to make a required SIP submission unless the state corrects the deficiency and EPA subsequently approves the SIP. On January 15, 2009, EPA formally found that both Minnesota and Michigan had failed to timely submit SIPs addressing the regional haze requirements. This finding triggered EPA's duty to either promulgate a regional haze FIP for Minnesota and Michigan or approve subsequently submitted regional haze SIPs.

Minnesota subsequently submitted to EPA a regional haze SIP on December 30, 2009, a draft supplement to the SIP on January 5, 2012, and a final supplement to the SIP on May 8, 2012. Michigan submitted to EPA a regional haze SIP on November 5, 2010. In previous rulemakings, EPA approved in part the states' regional haze SIPs for addressing most regional haze requirements. However, EPA deferred action on the states' BART determinations for taconite facilities in order to further evaluate the sufficiency of those determinations. On August 15, 2012, EPA proposed to disapprove in part the states' regional haze SIPs with regards to their BART determinations for taconite facilities, while simultaneously proposing to promulgate a FIP. In response to comments received related to the sufficiency of EPA's reasoning for proposing disapproval of the Michigan and Minnesota BART determinations, EPA is issuing a separate supplemental notice of proposed rulemaking to solicit additional comments on that issue. Nonetheless, despite the fact that EPA has not finalized its disapproval of the states' BART determinations, EPA has the continuing authority and obligation to promulgate a FIP based on its earlier finding that Minnesota and Michigan had failed to timely submit regional haze SIPs. EPA's duty to promulgate a FIP ends only when it has fully approved a state submission. EPA has determined that the FIP satisfies the requirements of the Act and the regional haze rule.

As described in greater detail in the proposal to this rulemaking (77 FR 49308, August 15, 2012), section 169A of the 1977 Amendments to the CAA created a program for protecting visibility in the nation's national parks and wilderness areas. 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 (45 FR 80084, December 2, 1980). In 1990, Congress added section 169B to the Act to address regional haze issues. Accordingly, EPA promulgated a rule to address regional haze on July 1, 1999 (64 FR 35714), which is codified at 40 CFR part 51, subpart P (“the regional haze rule”). On July 6, 2005, EPA published guidelines to assist states, or EPA when implementing a FIP, in determining which of their sources should be subject to the BART requirements and in determining appropriate emission limits for each applicable source (70 FR 39104), codified at 40 CFR part 51, appendix Y (“BART Guidelines”).

Among other things, section 169A of the Act and 40 CFR 51.308 of the regional haze rule require that states, or EPA when implementing a FIP, assure reasonable progress toward the national goal of achieving natural visibility conditions in Class I areas by submitting an implementation plan that contains emission limits representing BART for certain BART-eligible sources. 40 CFR 51.308(d) and (e). Pursuant to 40 CFR 51.308(e), BART must be determined based upon an analysis of the best

system of continuous emission control technology available and associated emission reductions achievable for each BART-eligible source that is subject to BART. In this analysis, the state, or EPA when implementing a FIP, must take into consideration the technology available, the costs of compliance, the energy and air quality environmental impacts of compliance, any pollution control equipment in use at the source, the remaining useful life of the source, and the degree of visibility improvement reasonably anticipated to result from the use of such technology. CAA section 169A(g)(2); 40 CFR 51.308(e)(1)(ii)(A).

The process of establishing BART emission limits consists of three steps. First, states or EPA identify those sources that meet the definition of “BART-eligible source” set forth at 40 CFR 51.301. Second, states or EPA determine whether such sources “emit any air pollutant which may reasonably be anticipated to cause or contribute to any impairment of visibility in any such area,” and is therefore “subject to BART.” Third, for each source subject to BART, states or EPA then identify the appropriate type and level of control for reducing emissions by conducting a five-step analysis: 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; step 5: evaluate visibility impacts. See BART Guidelines.

The regional haze rule required all states to submit an implementation plan for regional haze meeting the requirements of 40 CFR 51.308(d) and (e) by no later than December 17, 2007. 40 CFR 51.308(b). Neither Minnesota nor Michigan submitted regional haze SIPs to EPA by the required date.

B. EPA's Legal Authority To Promulgate a FIP

Section 110 of the Act requires states to develop implementation plans with enforceable emission limitations and other control measures to meet the applicable requirements of the Act. A state submits its SIPs and SIP revisions to EPA for approval. Congress crafted the Act to provide for states to take the lead in developing SIPs, but balanced that decision by requiring EPA to review the plans to determine whether a SIP meets the requirements of the Act. EPA is required to determine whether the state's submittal meets the requirements of the Act based on information and data available at the time of EPA's review. See

Sierra Club

v.

EPA,

671 F.3d 955 (9th Cir. 2012).

Pursuant to section 110(c)(1)(A) of the Act, if EPA finds that a state has failed to make a required SIP submittal or if EPA finds that a state's required submittal is incomplete, then EPA is required to promulgate a FIP to fill this regulatory gap. Section 110(c)(1)(A) of the Act requires EPA to promulgate a FIP within two years of its finding that a state failed to make a required SIP submission. Further, EPA has a continuing duty to promulgate a FIP even where EPA fails to promulgate a FIP within the required two-year period. EPA's duty to promulgate a FIP continues unless the state corrects the deficiency, and EPA approves the plan or revision before EPA promulgates the FIP.

1

1

It should be noted that in addition to the requirements of section 110(c)(1)(A) of the Act, section 110(c)(1)(B) of the Act requires EPA to promulgate a FIP where EPA has specifically disapproved a state's SIP submittal. Correspondingly, EPA has a continuing duty to promulgate the FIP unless a state corrects the deficiency and EPA approves the plan or revision before EPA promulgates the FIP. Many of the commenters to the proposed FIP assumed that the statutory basis for EPA's authority in promulgating this FIP is section 110(c)(1)(B) of the Act. We acknowledge that the proposed FIP, in identifying potential inadequacies in the Minnesota and Michigan regional haze SIPs, may have given the impression that the authority for promulgating the FIP was a specific determination by EPA that the States' SIPs failed to meet the requirements of the Act and the regional haze rule. However, as clarified in this section, the authority for the promulgation of the FIP arises from section 110(c)(1)(A) of the Act.

In this rulemaking action, EPA has the authority to promulgate a FIP addressing the BART determinations of certain taconite facilities in Minnesota and Michigan based upon the failure of both Minnesota and Michigan to timely submit regional haze SIPs. As discussed above, the regional haze rule required all states to submit a regional haze implementation plan by December 17, 2007. 40 CFR 51.308(b). Neither Minnesota nor Michigan submitted regional haze SIPs to EPA by the required date. Therefore, on January 15, 2009, EPA found that Michigan and Minnesota, as well as certain other states, had failed to submit SIPs addressing the regional haze requirements (74 FR 2392). Based upon that finding, pursuant to section 110(c)(1)(A) of the Act, EPA was under a continuing duty to promulgate FIPs for Minnesota and Michigan to address the regional haze requirements of the Act and the regional haze rule. This FIP is promulgated pursuant to the requirements of section 110(c)(1)(A) of the Act.

C. Minnesota and Michigan's Regional Haze SIP Submittals

Minnesota subsequently submitted to EPA a regional haze SIP on December 30, 2009, a draft supplement to the SIP on January 5, 2012, and a final supplement to the SIP on May 8, 2012. Michigan submitted to EPA a regional haze SIP on November 5, 2010. In general, with regard to the subject-to-BART taconite facilities identified in the respective plans, each State identified Good Combustion Practices (GCP) as the primary control method representing BART for NO

X

.

On January 25, 2012, EPA proposed approval of the Minnesota regional haze plan in which EPA, among other things, proposed to approve BART for the subject-to-BART taconite facilities (77 FR 3681). However, prior to EPA's final action on Minnesota's regional haze plan on June 12, 2012, EPA learned through public comment that Minnesota and Michigan had each failed to thoroughly analyze all feasible BART control technologies for the taconite facilities, and that the SO

2

and NO

X

emission limits set forth in each State's SIP might not reflect BART. Therefore, in light of the uncertainty pertaining to the States' BART determinations for taconite facilities, EPA deferred action on emission limits that Minnesota intended to represent BART for taconite facilities in the final rule approving the Minnesota regional haze SIP (77 FR 34801, June 12, 2012). Correspondingly, EPA proposed approval of certain provisions of the Michigan regional haze SIP, while deferring any action on those provisions of the SIP that addressed the requirement for BART for the one taconite plant in Michigan to which BART applies (77 FR 46912, August 6, 2012). Pursuant to section 110(k)(3) of the Act, EPA may approve a SIP revision in part when only a portion of a SIP revision meets all applicable requirements of the Act.

D. EPA's Regional Haze FIP and Related Actions

EPA proposed a FIP on August 15, 2012 (77 FR 49308) pursuant to section 110(c)(1)(A) of the Act, based on EPA's finding that Minnesota and Michigan failed to timely submit a regional haze SIP, and EPA's continuing duty to promulgate a FIP to address such failure. At the same time, EPA proposed disapproval of the BART determinations for the subject-to-BART taconite facilities made by Minnesota and Michigan for failing to meet the requirements of the Act and the regional haze rule. However, in regards to the proposed disapproval, several

commenters raised concerns that EPA did not provide adequate notice of its rationale for disapproving the States' BART determinations.

Therefore, EPA is taking two separate but related actions. In this rulemaking action, EPA is finalizing the FIP for BART for the subject taconite plants in Michigan and Minnesota. Secondly, in a separate action, EPA is issuing and seeks comment on a supplemental notice of proposed rulemaking elaborating upon the Agency's rationale for proposing partial disapproval of the Minnesota and Michigan SIPs as they pertain to the requirement for BART for taconite plants. The full basis for the partial disapproval is set forth in the separate action.

II. Comments and Responses

On August 15, 2012, EPA published a

Federal Register

Notice entitled “Approval and Promulgation of Implementation Plans; States of Minnesota and Michigan; Regional Haze Federal Implementation Plan” (77 FR 49308). In this notice, the EPA requested comment on EPA's proposed BART determinations and FIP for taconite ore processing facilities located in Minnesota and Michigan. Public comments were accepted at both a public hearing held in St. Paul, Minnesota, on August 29, 2012, and in writing until September 28, 2012.

EPA received comments from Cliffs Natural Resources Inc., ArcelorMittal Minorca Mine Inc., the United States National Park Service (NPS), the Michigan Department of Environmental Quality (MDEQ), the United States Forest Service, the National Parks Conservation Association (NPCA), the Fond du Lac Band of Lake Superior Chippewa, the Leech Lake Band of Ojibwe, the National Tribal Air Association, the Red Cliff Band of Lake Superior Chippewas, U.S. Steel Corporation, and more than 1,000 private citizens.

A. General Comments in Support of the Proposed Rule

Commenter:

National Parks Conservation Association.

Comment:

NPCA supports finalization and implementation of the proposed controls, which will significantly benefit the air quality in the parks, wilderness areas, and communities surrounding these plants.

Commenter:

1,244 private citizens provided similar comments.

Comment:

As a resident of the upper Midwest and a national parks supporter, I want to see natural air quality restored to Voyageurs and Isle Royale National Parks and Boundary Waters Canoe Area Wilderness just as Congress intended. That's why I support EPA's proposal to reduce haze-causing pollution from taconite plants. These large industrial polluters should clean up their air pollution under the Regional Haze Rule.

Reducing haze pollution in our parks will bring healthier air to surrounding communities as well as more visitors who support our local economies. That's why I want EPA to require the most effective methods for reducing air pollution from taconite plants in Michigan and Minnesota. In addition to the emission reductions outlined in EPA's proposed plan, I encourage EPA to evaluate pollution controls that would lead to cleaner air.

Commenter:

Fond du Lac Band of Lake Superior Chippewa.

Comment:

The Band strongly supports the FIP proposed by Region 5, particularly with regard to Region 5's determination that low NO

X

burners are BART for taconite facilities. This option is technically feasible as these burners have already been installed on Minntac's grate-kiln furnaces and are being installed on Essar's straight-grate kiln furnaces. Low NO

X

burners have been shown to be affordable with control costs at roughly $500 per ton, which is well within the range of costs deemed affordable for BART by states and EPA. Low NO

X

burners are a wise choice because they prevent NO

X

from ever being formed. This is a key concept in pollution prevention. Collection and disposal of pollutants can lead to secondary environmental problems, as well as increased energy consumption. The Band contends that installation of these burners is an equity issue. It would be unfair to allow other facilities to operate indefinitely without having to install low NO

X

burners.

Commenter:

Red Cliff Band of Lake Superior Chippewa.

Comment:

Red Cliff supports EPA's proposed requirement for low NO

X

burners for all subject taconite furnaces in Michigan and Minnesota.

Commenter:

Leech Lake Band of Ojibwe.

Comment:

Low NO

X

burners have been installed voluntarily, previous to this action, by two taconite facilities with different furnace systems commonly utilized in the taconite industry. These system installs have shown substantial reductions, up to 60 to 70 percent, can be achieved with a minimal cost of $500 per ton or less. The Band also agrees that taking a preemptive approach by preventing the formation of NO

X

makes sense versus an after-production control technology that is less effective and more costly, both economically and environmentally.

Commenter:

National Tribal Air Association.

Comment:

The Association agrees that using low NO

X

burners as BART for both straight and grate-kilns is a good approach. Not only is the cost to remove NO

X

inexpensive, but these burners can reduce NO

X

by up to 70 percent. Therefore, placing a limit on NO

X

of 1.20 pounds per million British Thermal Units on a 30-day rolling average for facility lines is very reasonable.

Commenter:

National Park Service.

Comment:

NPS agrees with EPA's conclusions that control of emissions from taconite plants in Minnesota and Michigan can be expected to yield significant benefits in reducing visibility impairment in the Class I area in the two states; and that technically feasible controls are available at a reasonable cost for taconite plants that can be expected to provide a visibility benefit that makes those controls warranted.

Commenter:

U.S. Forest Service.

Comment:

The Forest Service supports the proposed FIP to require BART for the taconite plants in Minnesota and Michigan. According to technical analyses by the State of Minnesota and others, the highest contributors to haze in the Boundary Waters from all sources in the U.S. are the taconite industry and power plants. We support the emission controls that the taconite plants would be required to install under the proposed FIP. The FIP demonstrates that these controls are technically feasible and available for the taconite industry to reduce emissions and are already being used by some within the industry. The implementation of the Minnesota regional haze plan is nearly five years past due. Considerable effort and resources have been spent over the past ten years developing the technical information necessary to complete implementation. Much of the technical work was done by states, Tribes, and FLMs working together through multi-state regional planning organizations. The Forest Service has monitored visibility in the Boundary Waters since 1985. The results of this technical work and monitoring support the requirement for BART to reduce impacts to the Boundary Waters.

Response:

EPA acknowledges these commenters' support of the Agency's efforts in developing a FIP for the taconite industry and agrees.

B. Comments Concerning the Adequacy of the Public Comment Period

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs stated that EPA provided inadequate opportunity to

comment on the proposed FIP. Cliffs alleged that 45 days was not a reasonable time period to complete the task of preparing an appropriate response to the proposed FIP given the highly technical concerns surrounding EPA's BART determinations for the taconite industry.

Response:

EPA disagrees that the Agency did not provide an adequate opportunity for public comment. Section 307(h) of the CAA requires EPA to provide “a reasonable period for public participation of at least 30 days” when promulgating a FIP. Here, EPA chose to provide a significantly longer 45-day public comment period in light of the many technical issues surrounding EPA's proposed BART determinations for the taconite industry. EPA believes that 45 days was a reasonable amount of time for Cliffs and others to comment on EPA's proposed FIP.

2

Cliffs' assertion that it should have been granted an extension to conduct a new BART analysis is without merit. Cliffs had several years to conduct a thorough BART analysis, and its failure to timely do so does not bear upon the reasonableness of the length of EPA's comment period. Indeed, the fact that Cliffs was able to prepare an extensive 61-page comment document within the allotted time supports EPA's contention that 45 days was a reasonable period for third parties to comment on the proposed FIP.

2

In fact, Cliffs received a signed copy of the proposed FIP on July 17, 2012, nearly a full month before the formal start of the comment period. Thus, Cliffs had effectively 75 days to prepare its comments.

C. Comments Questioning EPA's Authority To Issue a Federal Implementation Plan

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs stated that EPA has not met the threshold requirements for issuing a FIP. EPA's proposed FIP did not provide a critique of Minnesota and Michigan's BART determinations for the taconite industry or explain why those determinations were inadequate. As a result, Cliffs argued that EPA does not have the legal authority to issue a FIP.

Response:

EPA believes that it has a strong basis for proposing disapproval of Minnesota and Michigan's BART determinations for the taconite industry. Nonetheless, EPA agrees that the proposed rule did not provide a sufficiently detailed critique of the state determinations' inadequacies. As a result, EPA has chosen to issue a supplemental notice of proposed rulemaking providing additional rationale for the Agency's proposed disapproval. Contrary to Cliffs' assertion, however, EPA was not required to make a finding that Minnesota and Michigan's BART determinations were deficient before issuing a FIP. Section 110(c)(1)(A) of the CAA provides that EPA “shall promulgate a [FIP] within 2 years after the Administrator finds that a State has failed to make a required submission * * * unless the State corrects the deficiency, and the Administrator approves the plan or plan revision, before the Administrator promulgates such Federal Implementation plan.”

Pursuant to the regional haze rule, states were required to submit regional haze SIPs no later than December 17, 2007 (64 FR 35714, July 1, 1999). Neither Minnesota nor Michigan made the required submission by this date. Consequently, EPA issued a finding on January 15, 2009 that Minnesota and Michigan, as well as certain other states, had failed to submit SIPs addressing the regional haze requirement (74 FR 2392). This finding triggered EPA's statutory duty to either approve a subsequent state SIP submission or issue a FIP. While it is true that Minnesota and Michigan subsequently submitted regional haze SIPs to EPA, the Agency has not approved either of these plans with respect to the states' BART determinations for the taconite industry. On the contrary, EPA has proposed to disapprove the states' BART determinations for failure to meet the minimum requirements of the CAA. Thus, EPA had both the authority and the continuing obligation to issue a FIP for the taconite industry in Minnesota and Michigan based on the Agency's January 15, 2009 finding of the states' failure to submit.

Commenter:

Michigan Department of Environmental Quality (MDEQ).

Comment:

MDEQ commented that Section 110(c)(1) of the Act authorizes EPA to promulgate a FIP within two years after making a finding that a state's SIP submittal does not satisfy the CAA. However, the CAA does not allow EPA to propose a FIP and simultaneously propose disapproval of the state's SIP.

Response:

MDEQ's interpretation of section 110(c)(1) is incorrect. Once EPA has made a finding of a state's failure to submit, EPA's authority and continuing obligation to issue a FIP does not end until the state has corrected the deficiency and EPA has approved a subsequently submitted SIP. Nowhere in the CAA is there language that limits EPA's authority to simultaneously propose a FIP and propose disapproval of a state's SIP where there has been a prior finding of a failure to submit.

Commenter:

Cliffs Natural Resources, ArcelorMittal Minorca Mine, and Michigan Department of Environmental Quality (MDEQ).

Comment:

Cliffs and MDEQ stated that EPA failed to afford the Minnesota and Michigan SIP proposals the requisite deference. Under the visibility program, states have the primary responsibility for establishing standards, including BART. Thus, Cliffs and MDEQ argued that EPA can disapprove a SIP only where it fails to meet minimum CAA requirements.

Response:

While Congress intended states to take the lead in developing regional haze SIPs, it balanced that decision by requiring EPA to review state plans to determine whether they meet the requirements of the CAA. EPA's review is not limited to a ministerial type of automatic approval of a state's decisions. Rather, EPA must consider not only whether the state considered the appropriate factors, but whether the state acted reasonably in doing so. In undertaking such a review, EPA does not “usurp” the state's authority, but ensures that such authority is reasonably exercised.

Here, EPA firmly maintains that neither state's regional haze SIP met the minimum requirements of the CAA. Among other things, EPA takes issue with the states' assertions that low NO

X

burners are not technically feasible control options for indurating furnaces and that good combustion practices represent BART. Nonetheless, EPA acknowledges that its August 15, 2012 proposed action (77 FR 49308) did not provide a sufficiently detailed analysis of the deficiencies of the states' BART determinations for the taconite industry. Therefore, EPA is publishing a supplemental notice of proposed rulemaking that further addresses the Agency's rational for proposing disapproval of the states' choices regarding taconite BART.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that Minnesota and Michigan engaged in extensive and proper rulemaking efforts. Cliffs then proceeded to provide a detailed history of each state's SIP-development process.

Response:

EPA agrees that Minnesota and Michigan spent considerable time and effort preparing their regional haze SIPs. As stated previously, EPA intends to publish a supplemental notice of proposed rulemaking that further addresses the Agency's rationale for proposing disapproval of Minnesota and Michigan's BART determinations for taconite facilities. EPA reiterates, however, that the Agency had the

authority and continuing obligation to promulgate a FIP for the taconite industry based on the Agency's earlier finding that Minnesota and Michigan had failed to submit regional haze SIPs in a timely manner (74 FR 2392, January 15, 2009).

D. Comments Supporting EPA's Authority To Issue a Federal Implementation Plan

Commenter:

National Parks Conservation Association.

Comment:

At a public hearing on the proposed FIP, representatives for one taconite owner asserted that EPA lacked authority to issue the proposed FIP. The company's assertion has no merit. In fact, EPA has an obligation to develop a FIP under the CAA. The CAA provides states with initial responsibility for identifying sources and determining BART for purposes of regional haze. It is equally clear, however, that EPA retains authority to approve or disapprove the states' determinations and issue a FIP if necessary to correct state plan deficiencies.

EPA is not only well within its authority to promulgate the proposed FIP; it is required to do so because the state plans do not meet the requirements of the CAA. While commenters disagree with some of EPA's proposed BART determinations in the taconite FIP, the record plainly supports EPA's finding that neither the Minnesota nor the Michigan proposal met minimum CAA requirements. The National Park Service, the National Forest Service, and other commenters all submitted detailed technical reviews establishing the many deficiencies in the BART analysis and conclusions of Minnesota and Michigan's plans.

Response:

EPA agrees with the commenter that EPA has the authority and obligation to issue a FIP.

E. Comments Concerning the Use of New Information To Evaluate Minnesota and Michigan's BART Determinations

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA cannot use new information regarding the technical feasibility of low NO

X

burners as a control option for indurating furnaces to undermine the states' BART determinations. Cliffs argued that EPA is seeking to reject Minnesota and Michigan's BART determinations based on information that was not available to either state at the time of their SIP submissions to EPA for approval. To support its position, Cliffs pointed to EPA's BART Guidelines, which state that new technologies need only be considered by a state if they become available before the close of a state's public comment period. Cliffs alleged that low NO

X

burners were not an “available” technology because testing at Minntac and Essar had either not yet commenced or was still ongoing at the time Minnesota and Michigan's periods for public comment had ended.

Commenter:

Michigan Department of Environmental Quality (MDEQ).

Comment:

MDEQ commented that there was not enough information available prior to the close of Michigan's public comment period on June 23, 2010 to indicate that low NO

X

burners had been successfully utilized on indurating furnaces. MDEQ also argued that EPA's proposal to find that low NO

X

burners represent BART for NO

X

at Tilden was impermissibly based on information generated after the close of Michigan's public comment period.

Response:

EPA disagrees with Cliffs and MDEQ's comments for several reasons. First, EPA again reiterates that the Agency had the authority and responsibility to promulgate a FIP for the taconite industry based on the Agency's earlier finding that Minnesota and Michigan had failed to submit regional haze SIPs in a timely manner (74 FR 2392, January 15, 2009). Thus, EPA was entitled to rely on whatever information was available regarding the technical feasibility of low NO

X

burners at the time the Agency performed its BART analysis, including results from the testing at Minntac and Essar.

Nonetheless, even if EPA's authority to promulgate a FIP had been based solely on final disapproval of the states' BART determinations, the information regarding the technical feasibility of low NO

X

burners was not “new” as Cliffs suggests. As the BART Guidelines make clear, technical feasibility encompasses two distinct concepts, “availability” and “applicability.” 40 CFR part 51, appendix Y. A technology is considered “available” if the source owner may obtain it through commercial channels, while it is considered “applicable” if it can reasonably be installed and operated on the source type under consideration. As Cliffs pointed out, only technologies that are “available” at the close of a state's public comment period need be considered as control options by the state.

However, Cliffs' argument that low NO

X

burners were not an “available” technology at the time Minnesota and Michigan's periods for public comment had ended is incorrect. Testing at Minntac and Essar had nothing to do with the “availability” of low NO

X

burners. Rather, the testing at those facilities concerned the “applicability” of low NO

X

burners to the source type in question—indurating furnaces. There can be no dispute that low NO

X

burners were “available” at the time that Minnesota and Michigan developed their regional haze SIPs because this technology has been obtainable through commercial channels as an option for the control of nitrogen oxide emissions for many years. Therefore, Minnesota and Michigan were required to consider low NO

X

burners in their BART analyses, which both states did, albeit dismissively.

Consequently, the sole question presented to the states was one of “applicability”—whether low NO

X

burners could be successfully installed on indurating furnaces. In regards to this question, the BART Guidelines make clear that “a commercially available control option will be presumed applicable if it has been used on the same or a similar source type.” 40 CFR part 51, appendix Y. However, in contrast to the question of “availability,” the Guidelines make no mention of a cut-off date after which states may reject information regarding a technology's “applicability.” Even so, contrary to Cliffs' assertions, both states were aware that low NO

X

burners had been successfully installed on two lines at U.S. Steel's Minntac facility prior to the end of their respective periods for public comment.

3

In a June 23, 2010 letter to the Michigan Department of Natural Resources and Environment regarding the state's draft regional haze SIP, EPA commented that “a low-NO

X

main burner firing solid fuels” had been installed at Minntac and that “work done by other companies had demonstrated that burner designs that lower flame temperature can reduce NO

X

formation in taconite furnaces.”

4

Similarly, in a February 10, 2012 letter to the Minnesota Pollution Control Agency responding to the state's draft regional haze SIP supplement for taconite facilities, EPA explained in detail that “U.S. Steel has demonstrated the development and use of low NO

X

main burners that achieve 70 percent NO

X

reduction on its indurating

lines.”

5

In addition to these comments, both states received comments regarding the technical feasibility of low NO

X

burners from the Forest Service as well. Therefore, both Michigan and Minnesota were aware that low NO

X

burners had been successfully applied to indurating furnaces, and Cliffs' arguments that the results of these studies somehow constitute “new” information are without merit.

3

The comment period for Michigan's regional haze SIP closed on June 23, 2010. The comment period for the Minnesota's regional haze SIP supplement regarding BART at taconite facilities closed on February 3, 2010, but EPA was granted an extension to submit comments. EPA's comments were submitted on February 10, 2010, and were received and considered by MPCA.

4

See Michigan Regional Haze plan: EPA Letter to Michigan Department of Environmental Quality Regarding BART, May 24, 2012 (Docket # EPA-R05-OAR-2010-0954-0008).

5

See MI Haze FIP, EPA 6-23-10 comments to MDEQ on MI Haze submittal (Docket # EPA-R05-OAR-2010-0954-0037).

Finally, even if information regarding the technical feasibility of installing low NO

X

burners to indurating furnaces was not available to Minnesota or Michigan, EPA nonetheless had a duty to consider any new information that subsequently arose when reviewing the states' SIPs. The Ninth Circuit recently held that “if new information indicates to EPA that an existing SIP or SIP awaiting approval is inaccurate or not current, then, viewing air quality and scope of emissions with public interest in mind, EPA should properly evaluate the new information and may not simply ignore it without reasoned explanation of its choice.”

Sierra Club

v.

EPA,

671 F.3d 955, 967 (9th Cir. 2012). Thus, EPA is required, at a minimum, to take new information into account during the SIP approval process and, if necessary, alter its final decision accordingly.

Commenter:

Fond du Lac Band of Lake Superior Chippewa.

Comment:

At the public hearing held in Saint Paul, Minnesota, on August 29, 2012, some commenters voiced the opinion that low NO

X

burners should not be considered as BART because the technology was brought forward after the comment period on the Minnesota regional haze SIP supplement had closed. This is incorrect. Low NO

X

burners were in use and under consideration both before and during the comment period of December 19, 2011 to February 3, 2012. Discussions concerning the installation of low NO

X

burners at Minntac began in 2008, and the burners themselves were installed in 2010, several months before the regional haze SIP supplement was proposed. U.S. Steel's report to MPCA on the performance to-date of their low NO

X

burners at Minntac was submitted in December of 2011, around the time that Minnesota's SIP supplement went on public notice. Additionally, Essar Steel committed to the use of low NO

X

burners in its new plant near Nashwauk in 2010. The record indicates that discussions took place between MPCA and the taconite facilities around the time that the SIP supplement public comment period was open. Because of the timing of U.S. Steel's reports and the fact that no other economically-feasible technology offered more than 15 percent control of NO

X,

those discussions almost certainly included the possibility of requiring low NO

X

burners on taconite furnaces.

In light of the emerging use of low NO

X

burners, the U.S. Steel report, and the discussions indicated in the record, there is no basis for the claim that low NO

X

burners were only brought forward after the comment period. Low NO

X

burner technology was not a surprise and there is no procedural unfairness in the EPA considering it. Furthermore, to the extent that low NO

X

burners can somehow be construed as new information, there is precedent for considering new information while promulgating regulations. For example, on July 20, 2012, EPA informed petitioners that it would reconsider its Mercury and Toxics Standards based on the availability of new technical information.

Commenter:

National Parks Conservation Association.

Comment:

During the public hearing on this matter, a taconite company asserted that EPA's FIP was based on “new information” that is “outside the record” and that the company's “due process” rights were somehow jeopardized by EPA's proposal. As a legal matter, the company's argument has no merit. Likewise, as a practical matter, the company's complaints are unavailing.

Response:

EPA agrees with these commenters that it was appropriate for EPA to rely on whatever information was available regarding the technical feasibility of low NO

X

burners at the time the Agency performed its BART analysis. For a more detailed discussion of this issue, see EPA's previous response to comments from Cliffs, ArcelorMittal, and MDEQ.

F. Comments Concerning EPA's Best Available Retrofit Technology Analysis

Commenter:

National Parks Service.

Comment:

NPS agrees with EPA and with Michigan and Minnesota on the BART-eligibility determinations with respect to the taconite facilities and the states' determination that BART for direct PM is satisfied by the taconite MACT rule.

Response:

EPA acknowledges NPS's support.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs objected to EPA's reference to conversations with industry competitors and their vendors in determining the feasibility of controls for Cliffs' indurating furnaces. Cliffs asserted that EPA ignored information provided by Cliffs and its process engineering firms.

Response:

EPA spent significant time with all affected sources and thoroughly considered all information. EPA acknowledges that it relied heavily upon documented information from Cliffs' competitors in the taconite industry because these companies have experience with low NO

X

burner technology and provided data from actual experience with such technology. It would have been inappropriate for EPA to have ignored substantive information based upon actual experience.

1. Comments Asserting That EPA's BART Analysis Did Not Assess all Available Technologies

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs claimed that EPA's BART determinations were arbitrary because they ignored good combustion practices (GCP) as a BART alternative. Cliffs stated at the MPCA March Citizens Board meeting that “GCP is already required under other federal regulations, including the taconite MACT rule.”

Response:

Cliffs' support of GCP as BART lacks merit because GCP is neither defined by the Minnesota Pollution Control Agency, nor is it typically considered a NO

X

reduction technique. For example, the January 30, 2009 “NO

X

Reduction Analysis” done by Hatch for U.S. Steel's Minntac Iron Ore Pelletizing Operation did not list GCP as a potential NO

X

reduction technology for an indurating furnace.

6

Similarly, the 2008 BACT analysis for JEA—Greenland Energy Center Units 1 and 2 did not list GCP as a potential NO

X

control.

7

In fact, these analyses state that GCP tends to increase NO

X

emissions. This is because measures taken to minimize the formation of NO

X

during combustion inhibit complete combustion, which increases emissions of carbon monoxide. Conversely, GCP aims to reduce carbon monoxide emissions. According to the September 2010 “We Energies Biomass Energy Project Revised Control Technology Review for Carbon Monoxide Emissions for the Biomass-Fired Boiler,” there is an inverse relationship between NO

X

emissions and carbon monoxide emissions, which means that improving

combustion efficiency can increase NO

X

emissions.

8

6

Docket # EPA-R05-OAR-2010-0037-0039.

7

Docket # EPA-R05-OAR-2010-0037-0070.

8

Docket # EPA-R05-OAR-2010-0037-0069.

Concerning the GCP requirement in the taconite MACT rule, GCP for the MACT is not the same as GCP for NO

X

. GCP for MACT is aimed at reducing emissions of products of incomplete combustion (PIC). To minimize PICs, the operating conditions targeted are generally the opposite from those targeted for reducing NO

X

. As explained in the taconite MACT rule (68 FR 61883, October 30, 2003), “The basic method used in reducing NO

X

emissions is a reduction in combustion temperature, which is the opposite strategy needed for minimizing PIC (i.e., increasing combustion temperature).” In conclusion, GCP would be expected to increase NO

X

emissions, not decrease them.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA is required to consider “any existing pollution control technology at the source.” Cliffs argued that EPA failed to adequately consider or consistently apply this threshold factor to the BART determinations in its proposed rule.

Response:

To the extent that Cliffs is referring to its use of GCP as an existing pollution control technology, neither the operational practices that comprise GCP nor their impact on reducing emissions has been documented by Cliffs. As described in detail in the response to the previous comment, EPA does not consider Cliffs' use of GCP to constitute “existing control technology” on these furnaces.

Commenter:

National Parks Conservation Association.

Comment:

NPCA commented that EPA failed to consider fuel-blend alternatives, including greater or exclusive use of natural gas at grate-kiln furnaces, as part of the Agency's BART analysis for SO

2

and NO

X.

Fuel-blend alternatives are a technically feasible control option because indurating furnaces can successfully be operated on alternative fuels, namely fuel blends that consist primarily of natural gas. Contrary to the taconite plant owners' assertions, consideration of alternative fuels is required for BART where changing to cleaner fuel would not necessitate significant changes at any existing facility. There is no legal rationale for excluding this viable pollution control. Additionally, the assertion that alternative fuel costs are uncertain has no merit. There is simply no factual support for price uncertainty being a basis to reject consideration of natural gas as an alternative to coal. Even if significant uncertainty existed, it can be dealt with appropriately in the BART analysis. Finally, the assertion that moving towards a more natural gas-based fuel blend would mean higher NO

X

emissions in exchange for lower SO

2

emissions is a red herring. The existence of such potential secondary impacts is not a reason to discard a BART option prior to analysis. It is a reason to perform the analysis itself.

Response:

Alternative fuels were not considered for the following reasons. The straight-grate furnaces at ArcelorMittal, Hibbing Taconite, and Northshore Mining already burn natural gas. Similarly, U.S. Steel's Keetac and Minntac facilities already burn a fuel mix of natural gas and low-sulfur coal. While fuel-blend alternatives could have been considered for the grate-kiln furnaces at United Taconite and Tilden, EPA proposed to require the most stringent control technology, flue-gas desulfurization (FGD), at these facilities. As the BART Guidelines make clear, where EPA or the states choose the most stringent control option as BART, other control options need not be considered. Therefore, EPA was not required to consider fuel-blend alternatives as part of the Agency's BART analysis. However, EPA notes that subsequent to the proposal, Tilden agreed to convert to natural gas, while United Taconite will be substantially reducing its emissions through the use of natural gas and low-sulfur coal.

Commenter:

National Parks Conservation Association.

Comment:

EPA's NO

X

BART determinations conclude that significant reductions could be achieved cost-effectively by the installation of low NO

X

burners at all taconite kilns. While NPCA concurred with this conclusion, it commented that EPA failed to fully consider the use of regenerative selective catalytic reduction (RSCR). For instance, although RSCR was noted as an available technology in Keetac's BART analysis, EPA's FIP made no note of it. For Tilden, on the other hand, EPA noted this option, but only to point out that the company found it to be infeasible. In fact, this technology appears to be feasible for indurating furnaces. At a minimum, a more thorough evaluation by EPA is necessary. In this case, EPA has not shown that circumstances preclude the application of RSCR to the units in question via evaluation of gas characteristics or demonstration of technical challenges. It has offered no evidence that RSCR is technically infeasible. A fuller evaluation of this technology is warranted as part of a BART determination.

Response:

EPA did evaluate post-combustion NO

X

-control options when it reviewed Minnesota's regional haze plan and agreed with the state's determination that post-combustion control of NO

X

emissions from taconite facilities are not BART. For the proposed and now final rule, EPA evaluated new data on the use of low NO

X

burners at taconite facilities and, after a five-factor BART analysis, determined that low NO

X

burners are BART for these facilities. The BART analyses are fully described in section V of the proposed rule (77 FR 49308). EPA also considered RSCR and related selective catalytic reduction technologies at some of the subject taconite units. EPA concluded in its BART analyses that RSCR and other post-combustion controls do not represent BART for the subject taconite units because, after the installation of low NO

X

burners, the incremental costs of installing further post-combustion controls are unreasonably high. Therefore, this final rule requires that taconite indurating furnaces meet NO

X

emission limits consistent with low NO

X

burner technology.

Commenter:

National Parks Conservation Association.

Comment:

EPA's analysis for SO

2

provides evidence that dry FGD is feasible for taconite facilities, and the Agency requires the use of this technology at the three highest emitting lines (at United Taconite and Tilden). We support these determinations. However, EPA fails to fully analyze the use of dry FGD on the lower-emitting units, instead concluding, without support, that it would not be “economically reasonable.” NPCA asks that EPA analyze whether dry FGD, clearly a feasible technology, could provide cost effective reductions at additional units.

Response:

EPA's BART analysis demonstrated that dry FGD is feasible for the highest emitting lines when those lines are uncontrolled, but determined that the same technology has unreasonably high incremental costs for units with lower uncontrolled emissions. EPA notes, however, that while FGD was originally proposed as BART for the units at United Taconite and Tilden, those facilities have since agreed to operational limits on the types of fuels that may be burned As a result, FGD is no longer being required as BART. Additional discussion of this issue can be found in section III of the preamble.

Commenter:

National Parks Service.

Comment:

It appears that low temperature oxidation is technically and economically feasible for the entire

industry. In addition, tail-end SCR with natural gas reheat has been found technically feasible and borderline economically feasible based on a BACT analysis from several years ago when natural gas prices were much higher. Another form of SCR, RSCR looks promising, but as a new technology, would require trials.

While we would normally prefer to see all of the technically feasible control options evaluated, given the time constraints and the success of the low NO

X

burner technology, it is likely that low NO

X

burners will reduce NO

X

so much that addition of the other technologies would become too expensive for this phase of the regional haze program. We therefore agree that low NO

X

burners at 1.2 lbs NO

X

/MMBTU represent BART for the taconite industry. By setting such a uniform limit, EPA is establishing a “level playing field” that is achievable by all of the taconite plants and will provide substantial (almost 16,000 TPY) NO

X

reductions.

Response:

EPA agrees with the commenter that post-combustion control technologies would likely be expensive for additional pollution reduction. EPA maintains that low NO

X

burners are the appropriate control technology for the indurating furnaces at the taconite facilities. Thus, EPA is finalizing its determination that low NO

X

burners represent BART.

Commenter:

National Parks Service.

Comment:

EPA proposes to determine that BART for SO

2

for straight-grate kilns is existing controls because these furnaces do not burn coal. While true, they burn fuel oil, which can have a high potential for emitting SO

2

depending on the fuel's sulfur content. Although the BART Guidelines do not mandate fuel switching, they encourage evaluation of lower sulfur content fuels. For example, limiting fuel sulfur was an option considered by EPA for oil-fired EGUs in a separate BART rule. We suggest that Minnesota consider use of lower sulfur fuels in future reasonable progress analyses.

Response:

EPA's data indicate that the taconite facilities with straight-grate furnaces use natural gas as the primary fuel with fuel oil as a back-up fuel only. Given the limited use of fuel oil, emission reductions from using lower sulfur fuel would be limited. Nonetheless, EPA agrees with the commenter that the state should consider the impacts of using a lower sulfur fuel in its future reasonable progress analyses and is requiring that the taconite facilities keep records of any future use of fuel oil.

Commenter:

National Parks Service.

Comment:

NPS concurred with EPA's statement that “[the Agency does] not agree that the MPCA and Minntac have adequately documented the infeasibility of all SO

2

controls described.” This observation is especially pertinent with respect to the technical feasibility of spray drying absorption (SDA). According to the taconite industry consultant, SDA is not technically feasible because “the high moisture content of the exhaust would lead to saturation of the baghouse filter cake and plugging of the filters and dust collection system.” On the contrary, SDA requires moisture because a slurry of lime and water is injected into the spray dryer where the slurry reacts with SO

2

to form a dry sulfate power that is then collected in the baghouse. As long as the moisture content of the gas stream is not excessive and the temperature is not too low, SDA becomes a preferred and highly effective SO

2

control option. It is expected that retrofitting the facilities with SDA would eliminate the need for the existing Venturi rod scrubbers used to control PM on most of the taconite furnaces, thus reducing water consumption, gas stream moisture content, and PM emissions due to the higher efficiency of the baghouse.

Response:

In the proposed rule, EPA stated that while the state's documentation for determining the technical feasibility of all SO

2

controls was inadequate, EPA did agree with Michigan's conclusion that additional SO

2

controls, including SDA, were not cost effective and therefore not BART. EPA has not changed its position on this issue in the final rule.

2. Comments Asserting That EPA's Baseline NO

X

Emissions Are Arbitrary

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA's baseline NO

X

assumptions are arbitrary. EPA failed to even consider the actual emissions from each taconite furnace, let alone use them as the starting point for calculating furnace-specific baseline emissions.

Response:

EPA disagrees. In the case-by-case BART analysis for each subject taconite facility, EPA clearly listed the baseline actual annual emissions for each taconite furnace (see, e.g., Table V-B.24 for Hibbing Taconite (77 FR 49321)). In the initial stages of the BART development process, there was a significant lack of emissions data for the taconite facilities, as acknowledged by MPCA. However, additional monitoring and emission reporting from the taconite facilities enabled EPA to determine baseline NO

X

emissions for each facility.

G. Comments Concerning EPA's Analysis of Low NO

X

Burner Technology as NO

X

BART

1. Comments Supporting EPA's Determination That Low NO

X

Burners Are Technically Feasible

Commenter:

National Parks Conservation Association.

Comment:

NPCA commented that EPA's documentation of low NO

X

burners demonstrated that significant, cost-effective emission reductions are afforded by their use on both straight-grate and grate-kiln furnaces. Despite this record, a taconite company raised several concerns about EPA's determination during the public hearing. The concerns were that a given control technology will not transfer between indurating furnaces of the same type, low NO

X

burners will impact processing (product quality, fuel use, etc.), and there has been insufficient time to study various aspects or impacts of this technology.

These concerns are either misplaced or incorrect. As to the first point, low NO

X

burners have been successfully applied to a wide variety of units, including power plants, refineries, chemical companies, and other industrial settings, which burn a wide variety of fuels, including gas and coal. There may be individual differences between the burners at different taconite units. As is always the case, customization to the particular unit will be required. However, the differences among taconite furnaces of the same type are not significant enough to conclude that this clearly robust technology could not be applied to one as well as the others. Indeed, technology transfer would be impossible without such basic assumptions.

As to the impact of low NO

X

burners on operational parameters, EPA's FIP includes information addressing the points of product quality and fuel use. Minntac's experience demonstrates no impact to pellet quality, and after some adjustment, no increase in fuel use.

Finally, far from having had insufficient time to analyze these controls, the taconite facilities have had years in which to contact vendors, do engineering studies and modeling, and perform testing. The regional haze process has been delayed by many years at this point.

Response:

EPA agrees with the commenter that low NO

X

burners can be used to control NO

X

emissions from both straight-grate and grate-kiln indurating furnaces used in the taconite processing industry. EPA also agrees with the commenter that based on data from taconite facilities where low NO

X

burners are either in use or planned product, quality should not be compromised.

Commenter:

National Park Service.

Comment:

NPS commented that it agreed with EPA's proposal that BART for NO

X

for the taconite industry is low NO

X

burners achieving a 70 percent reduction from both straight-grate and grate-kiln furnaces. The proposal for grate-kiln lines is supported by research sponsored by U.S. Steel. The proposal for straight-grate kilns is supported by Essar's testing, which demonstrated a 95 percent reduction in NO

X

emissions for its new kiln.

Response:

As the commenter points out, EPA has determined that low NO

X

burners represent BART. However, EPA is setting an emission limit for each indurating furnace, not a 70 percent control requirement.

2. Comments Asserting That Low NO

X

Burners Are Not Technically Feasibility on Straight-Grate Kilns

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA made an unsupported presumption that the low NO

X

burner technology tested in a

1/4

-scale pilot test by Essar for a new source could be translated to all straight-grate furnaces.

Response:

As indicated in test reports, NO

X

emissions from taconite facilities are generated primarily by the burner. As burner design is the main factor contributing to NO

X

emissions, EPA carefully reviewed results of emission tests of low NO

X

burners for the different taconite furnace types and concluded that low NO

X

burners are technically feasible for straight-grate and grate-kiln furnaces.

Supporting the feasibility of low NO

X

burners on straight-grate kilns is a September 19, 2011 summary of findings presented to the Minnesota Pollution Control Board entitled “Results of Testing at

1/4

-Scale of LE Low NO

X

Burner Prototype for Straight-Grate Pelletizing Furnaces” by Fives North American Combustion, Inc. (Fives) for Essar.

9

After successful bench-scale testing of Fives' low NO

X

LE burners that achieved NO

X

reductions greater than 70 percent in a straight-grate pelletizing furnace, Essar and Fives proceeded with a joint $2 million investment in a test rig to simulate a straight-grate pelletizing furnace. In the

1/4

-scale test rig, the cross-sectional area scaling was very representative of actual furnace geometry, as were the energy inputs and flows. This testing demonstrated an emission rate of 0.25 lbs NO

X

/MMBTU, which is well below the proposed limit of 1.2 lbs NO

X

/MMBTU. Fives concluded that NO

X

emissions in the actual straight-grate furnace should be consistent with those measured in the

1/4

-scale test conditions. The feasibility of low NO

X

burners on straight-grate kilns was also confirmed during a June 20, 2012 call between EPA and a national low NO

X

burner manufacturer, Fives North America. Representatives from the manufacturing company were highly confident of the technical feasibility and application of their technology in straight-grate taconite furnaces. EPA agrees with this assessment.

9

Docket # EPA-R05-OAR-2010-0037-0039.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA failed to conduct an independent, case-by-case feasibility analysis as required by the Step 2 of the BART Guidelines. Minnesota and Michigan previously conducted an extensive case-by-case BART analysis, eliminating low NO

X

burners as technically infeasible for every taconite indurating furnace. Cliffs asserted that EPA has adopted a “one-size-fits-all” approach that is arbitrary and capricious. According to Cliffs, a proper feasibility analysis demonstrates that the technologies selected by EPA are infeasible for Cliffs' indurating furnaces because low NO

X

burners are not technically feasible for straight-grate furnaces and grate-kiln furnaces. Cliffs asserted that the Fives burner designed for Essar cannot be used without source-specific engineering and retrofit design. Cliffs claimed that U.S. Steel spent two years modifying the prototype low NO

X

burners installed on Lines 6 and 7 at its Minntac facility in an attempt to reach desired emission rates while combusting solid fuel, such as coal and biomass, but were ultimately unsuccessful. ArcelorMittal asserted that a proper feasibility analysis demonstrated that the technologies selected by EPA are infeasible for Minorca's indurating furnace. The commenters submitted information describing indurating furnaces, including the different types of furnaces, and explained what they believe are the differences between furnace types.

Response:

EPA believes that its finding that low NO

X

burners are technically feasible for both straight-grate and grate-kiln furnaces is supported by test results on various kiln configurations. Taconite furnaces are all based on one of two technologies. Straight-grate kilns are based on a Dravo-Lurgi design system, while grate-kiln furnaces are based on an Allis-Chalmers design system. EPA understands that each specific taconite furnace has unique operating requirements and specialized equipment. However, all furnaces share the same fundamental design style, either grate-kiln or straight-grate.

In assessing control technologies for a source category, EPA's BART Guidelines state that “control alternatives can include not only existing controls for the source category in question but also take into account technology transfer of controls that have been applied to similar source categories and gas streams.” 40 CFR part 51, appendix Y. The Guidelines go on to explain that “[c]ontrol technologies are technically feasible if either (1) they have been installed and operated successfully for the type of source under review under similar conditions, or (2) the technology could be applied to the source under review.”

Id.

EPA has concluded that there is a clear case for technology transfer of low NO

X

burner technology from grate-kiln furnaces to straight-grate furnaces. First, low NO

X

burner technology has been clearly and successfully demonstrated and applied across various industries for decades. Second, EPA does not consider taconite furnaces to be particularly unique given their similar fundamental designs. In the case of taconite applications, the Fives' testing of a low NO

X

burner prototype on a straight-grate furnace test rig provides reasonable assurance that full-scale applications, given the appropriate time for engineering and shakedown, will be both feasible and effective. In addition, U.S. Steel has already installed and is successfully operating multi-fuel low NO

X

burners on two unique grate-kiln indurating furnaces at their Minntac facility. Prior to the proposed rule, U.S. Steel had already submitted permit applications to install low NO

X

burner technologies on two additional furnaces at Minntac as well. U.S Steel has not indicated any issues with technical feasibility that will prevent the company from applying low NO

X

burners at either its Keetac facility or the remaining furnaces at Minntac. In response to questions from EPA concerning the installation at Minntac, U.S. Steel described the modifications it made allowing for the successful use of low NO

X

burners when burning either coal or natural gas.

10

In EPA's view, this information obtained directly from U.S. Steel rebuts Cliffs' claim that the

prototype low NO

X

burner tests were unsuccessful.

10

Email from U.S. Steel to EPA dated September 19, 2012 (Docket # EPA-R05-OAR-2010-0037-0071).

Nor is EPA persuaded by Cliffs and ArcelorMittal's arguments that the taconite furnaces at their facilities are unique to the extent that low NO

X

burner technology cannot be applied. While EPA understands that a complete engineering analysis will be required to design furnace-specific low NO

X

burners and that a shakedown period will be required to understand and optimize operations, EPA does not believe that the uniqueness of each individual taconite furnace proves technical infeasibility. The compliance times being finalized for each facility in this action account for engineering and shakedown time.

Over the years, the taconite industry has demonstrated that it can re-engineer furnaces to adapt to market changes (such as fuel prices), process changes (to accommodate variation in the type of ore being mined), and new technologies (such as heat recuperation systems).

11

It is clear that depending on the needs and priorities of each company, changes to the furnaces have and can be made.

11

For example, during the Society of Mining Engineers' Annual Meeting in New Orleans, Louisiana from March 2-6, 1986, a presentation was given titled “Design and Performance of the National Steel Pellet Plant High Temperature Heat Recuperation System” (Docket # EPA-R05-OAR-2010-0037-0077). The presentation discussed a high temperature heat recuperation system that was installed at the National Steel Pellet Company facility in Keewatin, Minnesota. The system was similar to those installed at Cliffs' Empire facility and U.S. Steel's Minntac facility.

With respect to ArcelorMittal's comment that a proper feasibility analysis would demonstrate that the technologies selected by EPA are infeasible for Minorca's indurating furnace, EPA relies on a September 27, 2012 report submitted by the commenter and authored by Fives North American titled “Retrofitting Low NO

X

Burners on the ArcelorMittal Minorca Straight-Grate Pelletizing Furnaces.”

12

After review, EPA concludes that this report supports the Agency's conclusion that low NO

X

burners are feasible at the Minorca facility. EPA therefore disagrees with ArcelorMittal's assertion that such technology is infeasible. Fives North American was engaged to perform an engineering study and recommend best options for retrofitting low NO

X

burners at the pelletizing furnace at the Minorca plant in order to achieve NO

X

emission rates below 1.2 lbs NO

X

/MMBtu under expected operating conditions. Fives expressed confidence that the company's experience in manufacturing low NO

X

burners, coupled with the successful results of the

1/4

-scale test at Essar, provided sufficient assurance that the technology could be applied at Minorca while preserving pellet quality and energy efficiency.

12

Docket # EPA-R05-OAR-2010-0037-0037.

3. Comments Concerning EPA's Cost Analysis for Low NO

X

Burners

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA presumed that low NO

X

burners will cost $500/ton at every Cliffs facility despite the fact that none of these facilities identified low NO

X

burners as technically feasible. EPA's $500/ton across-the-board cost estimate for NO

X

control was neither explained in the proposed rule nor supported by the record.

Response:

EPA did not presume that low NO

X

burners will cost $500/ton at every Cliffs facility. EPA's proposed rule stated that “[b]ased on the range of cost-effectiveness values provided, a conservative value of $500/ton will be used as the cost-effectiveness value for low NO

X

burners” (77 FR 49308, 49312). Data made available by U.S. Steel indicate the cost-effectiveness of low NO

X

burners on Minntac's Line 6 indurating furnace was $441/ton of NO

X

reduced with burning a 60 percent coal/40 percent natural gas fuel mix and $221/ton of NO

X

reduced when burning 100 percent natural gas. Barr Engineering and Essar Steel Minnesota have estimated a cost-effectiveness of $370/ton of NO

X

reduced for low NO

X

burner technology on a planned straight-grate natural gas-fired furnace.

13

This furnace is being designed to meet a much more stringent emission limit of 0.25 lbs NO

X

/MMBtu, compared to EPA's proposed limit of 1.2 lbs NO

X

/MMBtu. Thus, EPA's value of $500/ton represents a high-end estimate of expected cost-effectiveness of the selected NO

X

BART controls and is based on itemized costs and annual NO

X

emissions reductions.

13

Essar and Barr Presentation for Society for Mining, Metallurgy and Exploration, Duluth, Minnesota, April 2012 (Docket # EPA-R05-OAR-2010-0037-0039).

4. Comments Concerning the Effectiveness of Low NO

X

Burners

Commenter:

U.S. Steel.

Comment:

U.S. Steel commented that, based upon its experience, the appropriate emission factor when burning solid fuels is 1.5 lbs NO

X

/MMBtu, as opposed to the proposed NO

X

limit of 1.2 lbs NO

X

/MMBtu. U.S. Steel supplemented its comment on October 15, 2012 with data that support a limit of 1.2 lbs NO

X

/MMBtu while burning natural gas and 1.5 lbs NO

X

/MMBtu while burning solid fuels. U.S. Steel proposed that it be subject to the solid fuel limit, unless it utilizes 100 percent natural gas as a fuel for 30 consecutive days. The natural gas limit would then apply and it would remain subject to that limit until such time that solid fuels were utilized.

Response:

Based on a review of the data submitted by U.S. Steel, EPA agrees to revise the NO

X

limits in the final rule to 1.2 lbs NO

X

/MMBtu while an indurating furnace is burning 100 percent natural gas and 1.5 lbs NO

X

/MMBtu when fuels other than natural gas are being used. This revision primarily affects U.S. Steel Keetac, U.S. Steel Minntac, and United Taconite.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA made an unsupported presumption that an emission limit of 1.2 lbs NO

X

/MMBtu is equivalent to a 70 percent NO

X

reduction at every Cliffs facility and that all taconite furnaces emit NO

X

at an uncontrolled baseline rate of 4.0 lbs NO

X

/mmBTU in disregard of furnace variability.

Response:

The NO

X

emission limit that EPA proposed was 1.2 lbs NO

X

/MMBtu on a 30-day rolling average. This emission limit was not based on a percent reduction requirement, but rather was based on a demonstration by U.S. Steel that low NO

X

burners installed at the Minntac facility could achieve this emission limit. This limit is further supported for straight-grate kilns by successful testing of a low NO

X

burner prototype at a 1/4-scale test rig at Essar. It is standard industry practice to perform pilot tests, in which the results of a smaller unit are scaled up to a full production unit. Furthermore, in this case the company was extremely confident that, based upon the results with the 1/4-scale test rig, a limit much lower than 1.2 lbs NO

X

/MMBtu could be achieved on the full production unit. However, based on additional test data of operational low NO

X

burners submitted by U.S. Steel for Lines 6 and 7 at Minntac, EPA is revising its proposed limit of 1.2 lbs NO

X

/MMBtu on a 30-day rolling average. In the final rule, taconite indurating furnaces are subject to a limit of 1.2 lbs NO

X

/MMBtu when only natural gas is burned and 1.5 lbs NO

X

/MMBtu when fuels other than natural gas are used. Both of these limits are based on a 30-day rolling average.

H. Comments Concerning Non-Air Quality Impacts of Low NO

X

Burners

1. Effect on Pellet Quality

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that pellet quality cannot be maintained after the installation of low NO

X

burner technology.

Response:

EPA disagrees. Based on data supplied by U.S. Steel, EPA has concluded that there will be no pellet quality challenges resulting from the installation and operation of low NO

X

burner technology. In an email sent by U.S. Steel to EPA on September 19, 2012, U.S. Steel indicated that pellet quality specifications have not changed since the installation of low NO

X

burners, with zero off-spec shipments to date.

14

There have been no adverse pellet quality issues related to the installation and operation of the low NO

X

burners.

14

Docket # EPA-R05-OAR-2010-0037-0071.

U.S. Steel is required to maintain four pellet quality parameters, after tumble, compressions, reducibility, and low temperature disintegration (LTD), to meet customer specifications. U.S. Steel supplied EPA with data confirming that pellet quality parameters were acceptable after the installation of the Line 6 low NO

X

burner (Table 1). U.S. Steel also included more recent quality parameter data to show that quality continues to remain acceptable. U.S. Steel noted that while compressions have decreased, this has been observed on all process lines, including those without low NO

X

burners, thus indicating an issue with the feed material and not the burners. As also shown, U.S. Steel saw an improvement in reducibility for their pellets after the installation, which U.S. Steel attributes to improved heat distribution in the kiln from the low NO

X

burner.

Table 1—Line 6 Pellet Quality Before and After the Low NO

X

Burner Installation

(Higher values represent better quality)

After tumble

Compression

Reducibility

LTD

Before (11/1/10-4/3/11)

96.0

416

1.15

87.30

After (4/20/11-10/31/11)

96.0

417

1.22

85.58

1/1/12-9/1/12

96.2

410

1.22

86.11

2. Fuel Penalty and Energy Penalty

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that a fuel penalty and an energy penalty will result from the operation of low NO

X

burner technology at its facilities. Additional fans will be required to increase primary air flow through the furnaces because the cooler air that is injected into the burner to control peak flame temperature must be heated. EPA made an unsupported presumption that low NO

X

burners will cause no fuel or energy penalties or other emissions increases at any of the facilities.

Response:

EPA disagrees. The installation and use of low NO

X

burners is not generally considered to result in an energy penalty because one burner is merely being replaced by another. EPA recognizes that there is an increase in electricity needed for the operation of low NO

X

burner fans to assist in movement of air through the system or to heat cooler air that is injected into the burner. These costs can in most cases simply be factored into the cost impacts analysis as they were in this case for both NO

X

and SO

2

controls.

EPA believes that a low NO

X

burner installation that is properly engineered and optimized for a given process will not result in a fuel or energy penalty. EPA's conclusion is based on the U.S. Steel Minntac Line 6 Low NO

X

Main Burner & Facility NO

X

Management Final Report (December 1, 2011),

15

which documented that low NO

X

burners did not cause fuel penalties or other emission impacts. In addition, EPA consulted with a burner manufacturer and reviewed information provided by U.S. Steel regarding potential fuel impacts potentially associated with the operation of low NO

X

burners at a taconite facility.

16

In this correspondence, U.S. Steel Minntac stated that there was a temporary 10.5 percent fuel increase after initial installation of the low NO

X

burner on Line 7. However, during the shakedown period, the fuel increase was alleviated by process optimization (there was a learning curve due to the fact that this was the first low NO

X

burner installed on an iron ore processing line) and balancing the process airflow. The waste gas fan on Line 7 was running at maximum before the burner installation and with the addition of combustion air, the process efficiency decreased and safety issues were created. To alleviate this condition, the waste gas fan airflow capacity was increased in February 2011 on Line 7 to balance the airflow out of the process. In April 2011, the Line 6 low NO

X

burner was installed at U.S. Steel's Minntac facility. After applying what was learned during the shakedown period on Line 7, no increase in process fuel occurred after the installation. U.S. Steel clearly states in its September 19, 2012 email to EPA, “The end result is there is no increase in process fuel due to the installation of the Line 7 low NO

X

burner.”

17

15

Docket # EPA-R05-OAR-2010-0037-0039.

16

U.S. Steel email to EPA dated September 19, 2012 (Docket # EPA-R05-OAR-2010-0037-0071).

17

Docket # EPA-R05-OAR-2010-0037-0071.

In summary, based on available data, EPA believes that with process optimization, proper balancing of process air flows, and proper engineering, Cliffs and ArcelorMittal will be able to achieve similar fuel usage to U.S. Steel and will not incur either a fuel or energy penalty. EPA understands that each company will require a shakedown period similar to that experienced at U.S. Steel and has set the compliance schedules each facility accordingly.

3. Increases in the Emission of Other Pollutants

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that emissions of other pollutants will increase due to the installation of low NO

X

burner technology.

Response:

EPA has determined that there will be no increases in collateral pollutants due to the installation of low NO

X

burner technology. In making this determination, EPA relied on the information supplied by MPCA, U.S. Steel, Coen Company, Inc., and Hatch, as described below.

In a letter dated November 3, 2009 from Coen Company, Inc., Coen stated:

As we have indicated, the kiln burner that we proposed to supply for the above referenced project will not produce more CO as compared with what is being produced by the existing burner. The reason is: carbon monoxide (CO) is formed from lack of fast mixing of NO and oxidant (O

2

) and chemical kinetics of the reaction that is highly dependent on temperature and O

2

concentration. The Coen multi-fuel burner is being designed for higher stoichiometric air (1.00) as compared with the existing burner which has a stoichiometric air of about 0.3 only. Hence, the Coen burner design

promotes a higher amount of premixing of O

2

(oxidant) with fuel to reduce CO production. The flame temperature in both cases is high enough so the oxidation of CO is not kinetically limited. So the new burner design is not kinetically-limited for CO oxidation and the increased premixing in the primary zone will reduce CO emissions.

18

18

Docket # EPA-R05-OAR-2010-0037-0071.

Similarly, in a letter dated November 6, 2009 from Hatch to U.S. Steel, Hatch stated: “Since USS Minntac plans to continue using their current fuels, fuel mixes, and fuel firing rates in conjunction with the low NO

X

burners, with the exception of NO

X

, Hatch does not anticipate any change in the emissions of applicable pollutants. Substantial reduction of NO

X

emissions is also anticipated.”

19

19

Docket # EPA-R05-OAR-2010-0037-0071.

Finally, in a letter to U.S. Steel dated November 20, 2009, Owen Seltz, Engineer, Metallic Mining Section, Industrial Division, MPCA, stated:

Generally, when a reduction in NO

X

emissions from fuel combustion is proposed, the pollutant of concern for potential increase is carbon monoxide (CO). However, due to the design of the proposed low NO

X

main burner, CO emissions are not anticipated to increase. Furthermore, as explained in the manufacturers' letters dated November 3, 2009, and November 6, 2009 and submitted to the MPCA in Minntac's November 12, 2009 letter, due to the design and operation of the proposed burner, CO emissions are expected to decrease.

20

20

Docket # EPA-R05-OAR-2010-0037-0071.

Based on these assurances, EPA is confident that there will be no increases in other pollutants as a result of the installation of low NO

X

burner technology as Cliffs claims.

I. Comments Concerning Sulfur Dioxide (SO

2

) BART Emission Limits

Commenter:

National Parks Conservation Association.

Comment:

NPCA commented that the proposed limits for six of the units (at Northshore, ArcelorMittal, and Hibbing) specifically do not apply when burning fuel oil. This loophole undermines the purpose of a BART analysis and contradicts the CAA requirement for BART to be met on a continuous basis. The final determination must include a limit that encompasses the burning of fuel oil at these facilities.

Response:

Northshore, ArcelorMittal, and Hibbing are straight-grate indurating furnaces and do not burn coal. The primary fuel at these facilities is natural gas. As a result, these facilities have inherently low SO

2

emissions. Fuel oil is used only as a backup fuel. Due to its limited use, there was insufficient test data to set a corresponding SO

2

emission limit for periods when fuel oil is being burned. EPA set the SO

2

emission limits based on available data. For the straight-grate facilities, data was only available for periods in which the furnaces were combusting natural gas. In order to address this issue, EPA has added a regulatory requirement for affected sources to track their use of fuel oil and the resulting SO

2

emissions. This information will be used as the basis for any restrictions that will need to be added, e.g. sulfur content, on the use of fuel oil. These requirements are contained in §§ 52.1183(k)(4) and 52.1235(b)(2)(7).

Commenter:

Michigan Department of Environmental Quality (MDEQ).

Comment:

MDEQ commented that it had based its acceptance of Tilden's BART submittal for SO

2

on the lack of visibility impairment due to SO

2

emissions.

Response:

In the final rule, EPA is no longer requiring add-on controls at the Tilden facility because Tilden has agreed to switch fuels to natural gas within one year of the effective date of this rule.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA should use the methods proposed by Minnesota to set emission limits, citing the limits the state set for the Hibbing Taconite and ArcelorMittal facilities. Cliffs contended that EPA's proposed SO

2

emission limits are unsupported and arbitrary. Cliffs objected to the limits set by EPA because they appear to be based on results from a single stack test, which represents a snapshot in time. Further, the limits ignore a significant amount of available data and bypass the statistical analysis conducted by MPCA.

Response:

In the Agency's review of Minnesota's regional haze SIP supplement, EPA concluded that the limits for taconite facilities proposed by Minnesota do not accurately represent the current level of controls at the facilities. For ArcelorMittal and Hibbing Taconite, Minnesota appears to have set the limit in pounds of SO

2

per long ton (LT) of pellets produced based on a 30-day rolling average, using the Upper Predictive Limit (UPL) approach for normally distributed data. Minnesota did not demonstrate an accurate method to track or record LT of pellets produced. Therefore, the limits proposed by Minnesota are unenforceable. EPA also concluded that the annual testing requirement proposed by the state is insufficient to determine compliance with a limit based on a 30-day rolling average. In this action, EPA is finalizing SO

2

limits for taconite facilities in terms of lbs SO

2

/hr based on a 30-day rolling average, which can be easily and accurately measured using the continuous emissions monitoring system (CEMS) required by this rule.

EPA does agree that the UPL approach is an appropriate method for setting the SO

2

limits. However, the available SO

2

emissions data for the taconite sources generally do not follow a normal, logarithmic, or gamma distribution. For this reason, the UPL should be determined using a nonparametric method, as set forth below. EPA used available stack test and CEMS data from 1990 to the present to recalculate the SO

2

limits for ArcelorMittal and Hibbing, based on the appropriate UPL equation for nonparametric data, in terms of lbs SO

2

/hr on a 30-day rolling average as follows:

UPL

=

x

m

and

m

= (

n

+ 1)*(1 −

a

)

Where:

x

m

= value of the mth data point, when the data is sorted smallest to largest

m

= the rank of the ordered data point, when data is sorted smallest to largest

n

= number of data points

a

= 95th percentile or 0.95

If

m

is not a whole number, a linear interpolation is calculated from the following equation:

ER06FE13.001

Where:

m

i

= the integer portion of

m,

i.e.,

m

truncated at zero decimal places, and

m

d

= the decimal portion of

m

In this final rule, EPA is setting a limit of 38.16 lbs SO

2

/hr for indurating furnace EU026 at ArcelorMittal. This limit must be measured on a 30-day rolling average and does not apply when the subject unit is burning fuel oil. For Hibbing, EPA is finalizing an aggregate limit of 247.8 lbs SO

2

/hr based on a limit of 82.60 lbs SO

2

/hr for each of the three affected lines: EU020, EU021, and

EU022. This limit is also measured on a 30-day rolling average and does not apply when the subject unit is burning fuel oil.

Because of limited stack test data for Hibbing and ArcelorMittal, these sources may, within 20 months of the effective date of this rule, calculate a revised SO

2

limit based on one year of hourly CEMS data, reported in lbs SO

2

/hr, and submit such limit, calculations, and CEMS data to EPA. This limit shall be set in terms of lbs SO

2

/hr, based on the non-parametric UPL equations set forth above, with compliance to be determined on a 30-day rolling average.

Commenter:

Cliffs Natural Resources.

Comment:

Cliffs submitted alternate SO

2

limits for Hibbing based on the UPL equation for normally distributed data.

Response:

While EPA agrees that the UPL approach is the appropriate method for setting the SO

2

limits, EPA disagrees with the alternate SO

2

limits submitted by the Cliffs for the Hibbing facility. Cliffs did not specify whether its suggested limit was daily, instantaneous, or on a 30-day rolling average. But in any case, the limit submitted by Cliffs for Hibbing appears to be calculated using the UPL equation for normally distributed data, a p-value of 0.01 (which would represent a 99.5 percent confidence interval) and m = 1. This is incorrect. According to the UPL method, m represents the number of future runs (i.e., the number of future data points). As the data sets being used in the analyses are one-hour CEMS averages, the value of m should be 720 (30 days times 24 hours) if the limit being set is a 30-day rolling average. Even if compliance is based on the average value of an annual performance test rather than a 30-day rolling average, Minnesota's annual performance testing requires 30 hourly data points, which would result in a value of 30 for

m.

Cliffs also appears to have combined all stacks for each line and averaged all test runs for each set of test data to arrive at one data point for each set of test data, resulting in only 10 data points rather than 720 to calculate the UPL.

In addition, although the raw test data provided SO

2

emissions levels in terms of lbs SO

2

/hr, Cliffs calculated the UPL in terms of lbs SO

2

/LT pellets and then converted the SO

2

limit back into lbs SO

2

/hr by using the maximum design capacity of each line rather than the actual production data collected during testing.

Thus, EPA disagrees with this methodology. The alternate emission limit proposed by Cliffs is significantly higher than the limit that would result from the correct application of the UPL equation for normally distributed data. Further, some of the available data for the Hibbing facility are normally distributed, while other data are not. As noted previously, the available SO

2

emissions data for the taconite industry in general do not follow a normal, logarithmic, or gamma distribution. For this reason, EPA is using the nonparametric UPL equation to calculate the SO

2

emission limits.

Commenter:

Cliffs Natural Resources.

Comment:

Cliffs objected to the proposed 80 percent SO

2

reduction requirement for Northshore, noting that an SO

2

emission limit was also set for the facility. Cliffs contended that EPA failed to cite any justification for the requirement and failed to explain why an 80 percent reduction in SO

2

emissions should be required for Northshore when a similar reduction was not required for any other facility. Cliffs asserted that Northshore's SO

2

emissions have a de minimis impact on visibility, so imposing multiple layers of control requirements would be arbitrary and unnecessary.

Response:

Northshore is subject to BART based on the visibility impacts that were described in the proposal rule. The document entitled “Northshore Mining Company Analysis of Best Available Retrofit Technology (BART)” submitted to MPCA on behalf of Northshore Mining states that “WWESPs are currently in place on the furnace exhausts and are believed to remove 80 to 95 percent of the SO

2

in the exhaust.”

21

Thus, 80 percent is on the low end of the removal efficiency range estimated by Northshore, not an arbitrary number selected by EPA. Further, in a CAA section 114 request to Cliffs and Northshore Mining, EPA requested copies of all stack tests conducted on any emissions unit for any reason, including all test runs, even if a full test series was not completed. In its response to EPA, Cliffs did not provide any SO

2

test data for the subject-to-BART furnaces at Northshore. Without this emissions data, EPA believes using Northshore Mining's prior estimate of 80 to 95 percent control efficiency on its furnace exhausts EPA to impose an 80 percent emissions reduction requirement on stacks SV101, SV102, SV103, SV104, SV105, SV111, SV112, SV113, SV114, and SV115 was appropriate.

21

Docket # EPA-R05-OAR-2010-0037-0034 (p. 24).

Subsequent to the public comment period, Cliffs provided EPA with limited SO

2

emissions data for Northshore and proposed an aggregate limit of 39.0 lbs SO

2

/hr based on a limit of 19.5 lbs SO

2

/hr per line. Cliffs' proposed limit is slightly higher than the limit EPA calculated using the new data and the UPL equation for nonparametric data. However, EPA believes it is reasonable to set an aggregate limit of 39.0 lbs SO

2

/hr, measured on a 30-day rolling average, and to require the source to recalculate this limit when CEMS data are available.

As stated previously, this limit does not apply when the facility is burning fuel oil. In order to address this issue, EPA has added a regulatory requirement for affected sources to track their use of fuel oil and the resulting SO

2

emissions. This information will be used as the basis for any restrictions that will need to be added, e.g. sulfur content, on the use of fuel oil. These requirements are contained in sections 52.1183(k)(4) and 52.1235(b)(2)(7).

In summary, this final rule establishes an aggregate SO

2

emission limit of 39.0 lbs SO

2

/hour, measured on a 30-day rolling average, for Furnace 11 and Furnace 12 at Northshore. Within 20 months of the effective date of this rule, the owner or operator must calculate a revised SO

2

limit based on one year of hourly CEMS emissions data reported in lbs SO

2

/hr and submit such limit, calculations, and data to EPA. This limit shall be set in terms of lbs SO

2

/hr, based on the non-parametric UPL equations previously set forth by EPA, with compliance to be determined on a 30-day rolling average. EPA agrees with the commenter that an 80 percent reduction requirement is no longer needed because it is redundant in light of the final lbs SO

2

/hr emission limit. Consequently, this final rule does not require an additional 80 percent emissions reduction requirement at Northshore.

Commenter:

National Park Service and Cliffs Natural Resources.

Comment:

NPS supported EPA's proposal to require FGD as BART for SO

2

at the United Taconite and Tilden facilities, agreeing with EPA's cost-effectiveness calculations.

Cliffs, on the other hand, disagreed with EPA's cost-effectiveness calculations for the United Taconite and Tilden facilities. Subsequent to the public comment period, Cliffs proposed switching fuels as an alternative to installing FGD scrubbers. Cliffs proposed a combined limit of 529 lbs SO

2

/hr for Lines 1 and 2 at the United Taconite facility, with compliance to be determined on a 30-day rolling average, beginning in 54 months. To meet this limit, the United Taconite furnaces will burn low-sulfur fuels, including increased use of natural gas. For Tilden, Cliffs proposed switching operation to

100 percent natural gas within 12 months with an emissions limit to be set after a year of CEMS data become available.

Response:

Subsequent to the proposed rule, Cliffs has agreed to a federally enforceable aggregate emission limit of 529 lbs SO

2

/hr, based on a 30-day rolling average, at United Taconite, based on the use of low-sulfur fuels. Cliffs has also agreed to convert to the use of 100 percent natural gas at Tilden. Because Tilden will now be restricted to the use of 100 percent natural gas, requiring the installation of SO

2

controls is no longer economically feasible or necessary. Similarly, in light of the reduction in SO

2

emissions that will result from the use of low-sulfur fuels at United Taconite, the cost effectiveness of additional controls has increased to $12,021 per ton for Line 1 and $7,680 per ton for Line 2. Thus, EPA believes that the installation of such controls is no longer economically feasible. In addition to the emission limit proposed by Cliffs, to ensure the use of low-sulfur fuels and SO

2

reductions resulting from the use of low-sulfur fuels at United Taconite, EPA is also requiring that the facility burn either natural gas or a blend of natural gas and coal. EPA is also establishing a limitation on the coal to be used by requiring the coal have a sulfur content no greater than 0.60 percent sulfur by weight based on a monthly block average. The requirement for a sampling and calculation methodology for determining this value is contained within the monitoring plan as required in section 52.1235(e)(8)(x). In summary, EPA is no longer requiring FGD at United Taconite and Tilden as BART for SO

2

in this final rule.

Commenter:

U.S. Steel.

Comment:

U.S. Steel proposed several alternate lbs SO

2

/hr limits for its Minntac facility. These limits were calculated by applying a 99 percent confidence interval utilizing three years of CEMS data. U.S. Steel also proposed alternate limits for producing flux versus acid pellets due to scrubber inefficiencies during acid pellet production.

U.S. Steel proposed, as its first choice, an aggregate limit of 498 lbs SO

2

/hr, on a 30-day rolling average, when all five lines are producing flux pellets; an aggregate limit of 630 lbs SO

2

/hr, on a 30-day rolling average, when Lines 3-5 are producing acid pellets and Lines 6 and 7 are producing flux pellets; and an aggregate limit of 800 lbs SO

2

/hr, on a 30-day rolling average, when all five lines are producing acid pellets. U.S. steel also proposed partially aggregated limits and line-by-line limits for acid pellets and flux pellets. Finally, U.S. Steel proposed that the limit for acid pellets be in effect during acid production and for 30 days thereafter due to the 30-day rolling average.

Response:

EPA compared the limits proposed by U.S. Steel to the limit EPA calculated with the non-parametric UPL method and found them to be comparable. EPA also agrees with the need for a higher limit for acid pellet production.

Therefore, in this final rule, the SO

2

emission limits for U.S. Steel's Minntac facility are 498 lbs SO

2

/hr on Lines 3-7 when all lines are producing flux pellets; 630 lbs SO

2

/hr when Lines 3-5 are producing acid pellets and Lines 6 and 7 are producing flux pellets; and 800 lbs SO

2

/hr on Lines 3-7 when all lines are producing acid pellets. All limits are calculated on a 30-day rolling average.

However, EPA does not agree that the limit for acid pellets should apply during acid production and for 30 days thereafter and thus has not made this change in the final rule. The emission limit for a given 30-day rolling average period will be calculated using a weighted average as follows:

ER06FE13.002

Where:

L

30

= the limit for a given 30-day averaging period

n

f

= the number of days in the 30-day period that the facility is producing flux pellets on Lines 3-7

n

af

= the number of days in the 30-day period that the facility is producing acid pellets on Lines 3-5 and flux pellets on Lines 6 and 7

n

a

= the number of days in the 30-day period that the facility is producing acid pellets on Lines 3-7

Commenter:

U.S. Steel and Cliffs Natural Resources.

Comment:

Cliffs and U.S. Steel commented that it is inappropriate to use a seven percent oxygen correction for emission limits that are not concentration based.

Response:

EPA agrees with the commenters that the use of a seven percent oxygen correction is not necessary when the subject-to-BART facilities elect to comply with an emission limit measured in pounds of pollutant per million British thermal units or pounds of pollutant per hour.

Commenter:

U.S. Steel.

Comment:

U.S. Steel requested that the pH and SO

2

removal efficiency limits for its Keetac facility be deleted because they are redundant with the lbs SO

2

/hr limit.

Response:

EPA agrees with U.S. Steel and has deleted the pH and SO

2

removal efficiency limits from the final rule.

Commenter:

Leech Lake Band of Ojibwe.

Comment:

At United Taconite and Tilden, limiting SO

2

to 5 parts per million by volume or requiring the facilities to meet a 95 percent reduction requirement, on a 30-day rolling average, using dry FGD is achievable and cost-effective.

Commenter:

Red Cliff Band of Lake Superior Chippewas.

Comment:

The Red Cliff Band supported EPA's proposed requirement for additional SO

2

controls in select facilities.

Commenter:

National Tribal Air Association.

Comment:

The Association agreed with using existing SO

2

controls for those taconite facilities where it would be cost prohibitive to convert to a different technology that would only achieve nominal SO

2

reductions. However, in the case of the United Taconite and Tilden facilities, the Association found it relatively inexpensive to use dry FGD to limit SO

2

to 5 parts per million by volume or to meet a 95 percent reduction requirement on a 30-day rolling average.

Commenter:

National Park Service.

Comment:

NPS supported EPA's proposal to require FGD as BART for SO

2

at the United Taconite and Tilden facilities, agreeing with EPA's cost effectiveness calculations and compliance schedule.

Response:

EPA agrees with the commenters that under current operating conditions, FGD is a cost-effective control option for the grate-kiln furnaces at United Taconite and Tilden and represents BART. However, subsequent to the public comment period, Cliffs proposed switching fuels as an alternative to installing FGD scrubbers. Cliffs has since agreed to federally enforceable limits on the types of fuels that may be burned at these facilities. In this final rule, the United Taconite furnaces must burn a combination of natural gas and low-sulfur coal and Tilden will now burn 100 percent natural gas. Given these changes, EPA has determined that requiring the installation of SO

2

controls is no longer economically feasible or necessary.

J. Comments Concerning the Visibility Analysis and Visibility Impacts

Commenter:

Fond du Lac Band of Lake Superior Chippewa.

Comment:

Tables V-C.l0 to V-C.l5 of the FIP demonstrate the changes in visibility that could be expected from the use of low NO

X

burners. While these are only predictions, the expected improvements in visibility in the

Boundary Waters Canoe and Wilderness Area, Voyageurs National Park, and Isle Royale National Park strongly support the use of this technology at the subject-to-BART taconite plants.

Response:

EPA agrees that the BART emission limits have the potential to result in significant improvement in visibility at the affected Class I areas.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA did not conduct a proper analysis of visibility impacts.

Response:

EPA disagrees. EPA's visibility estimates provide ample evidence that the visibility impacts of each subject-to-BART taconite facility are substantial enough to warrant the selected BART controls. EPA's responses to the individual criticisms raised by the commenters on our visibility analysis are discussed in further detail below.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that the proposed rule directly contravened the instructions given to EPA in

American Corn Growers

v.

EPA.

In particular, Cliffs asserted that EPA's method is a “bifurcated” approach to visibility that was rejected in that decision. In Cliffs' view, the

American Corn Growers

court rejected a bifurcated approach in which visibility impacts are treated differently than the other four BART factors. Cliffs noted that for each taconite source, EPA separated its analysis into two distinct sections. Section V.B. of the proposed rule analyzed the first four factors, while EPA separately analyzed visibility improvement in Section V.C. for whichever technology emerged from the four-factor analysis in Section V.B. Thus, in Cliffs' view, the real-world visibility impacts were, at most, a secondary consideration that could not have influenced the evaluation of BART alternatives.

Response:

EPA disagrees. The “bifurcation” referred to in the

American Corn Growers

decision related to EPA's use of a regional, multi-source, group approach to determining the degree of visibility improvement, while analyzing the other four statutory factors on a source-specific basis. The

American Corn Growers

court held that the visibility analysis must not be treated differently and must be a source-specific analysis. Since that decision, EPA and states have consistently conducted the visibility prong of the five-factor analysis on a source-specific basis. In this instance, although EPA presented its visibility analysis in a separate section of the proposed rule, the Agency conducted the analysis on a source-specific basis consistent with the holding in

American Corn Growers.

EPA also disagrees that visibility was a secondary consideration in its analysis. EPA's analysis shows that based on the all of the BART factors, including visibility, the selected controls are warranted. If highly reasonable and cost-effective controls had been available but visibility benefits were slight, EPA would have rejected those controls. Section V.B. of the proposed rule demonstrated that reasonable and cost-effective controls were available. Section V.C. then showed that the visibility benefits to be obtained by requiring controls at each source were significant. Site-specific visibility improvement estimates for each source, derived from regional modeling conducted by the state of Minnesota on a variety of sources in the area, demonstrated that the significant reductions EPA proposed will produce significant visibility improvement in affected Class I areas.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that the approach EPA used in the proposed rule to estimate visibility impacts was arbitrary because it was not site-specific. Rather than extrapolating results from other facilities, EPA should have conducted modeling for the specific sources being regulated. The BART Guidelines instruct EPA to conduct modeling using CALPUFF or other appropriate dispersion models for each source and highlight the importance of source-specific features, such as stack flow rate and release height. EPA's use of “visibility impact ratios” derived from other sources is not consistent with EPA's own guidelines and provides results that are too unreliable for the purpose of a BART visibility analysis. In using the visibility impact ratio approach, EPA is holding itself to a lower standard than it would expect from a state air quality agency conducting a similar BART review.

Response:

EPA's proposed rule acknowledged that there is greater uncertainty associated with the visibility impact ratio approach. Nonetheless, EPA finds this approach to be consistent the BART Guidelines allowance for “appropriate” models and believes the approach provides adequate indication of the visibility benefits of the evaluated controls.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA's visibility analysis was inconsistent with the statutory obligation to consider the degree of visibility improvement that is “reasonably anticipated.” In Cliffs' view, the use of “possible” impacts from an approach extrapolated from other facilities does not satisfy this statutory requirement to consider “reasonably anticipated” visibility impacts. In support of this view, Cliffs noted that EPA extrapolated visibility results from facilities other than taconite plants, such as an electric generating unit. Moreover, Cliffs found EPA's use of sources in the “general area” as unacceptable for the visibility analysis given that wind conditions would affect the taconite facilities differently than the facilities EPA relied upon. Additionally, the commenter noted differences in stack conditions between the taconite facilities being regulated and the sources that EPA relied upon.

Response:

EPA's proposed rule acknowledged the uncertainty associated with the visibility impact ratio approach, but noted that despite the uncertainties, the Agency was confident that the information was adequate to assess potential visibility improvements due to emissions reductions at the specific facilities. While the results obtained from this approach are not expected to be as precise as source-specific CALPUFF modeling, they are based on visibility improvements derived from existing regional scale modeling that was conducted on sources in and around the northern Minnesota area. Given the geographic proximity of the taconite facilities to those that were modeled, EPA believes that the ratio approach provided adequate assurance of the visibility improvements that can be expected from the proposed emission reductions.

The results EPA obtained from its analysis are presented in terms of deciview (DV) change and change in the number of days above the 0.5 DV threshold. In the proposed rule's summary of the impacts at Boundary Waters, Voyageurs, and Isle Royale, these values ranged from 1.3 to 7.1 DVs of improvement with between 17 and 93 fewer days above the 0.5 DV threshold. Therefore, even if the ratio approach was over-estimating visibility improvements by a factor of two or three, the expected benefits would still be significant.

For example, Cliffs submitted CALPUFF modeling that showed the visibility improvements expected from the proposed rule for two of the seven facilities—United Taconite and Tilden. This modeling was only performed at the most impacted of the four affected Class I areas. EPA also notes that these were the only facilities for which new

scrubbers were proposed as BART for SO

2

.

The results of the Cliffs' modeling for United Taconite and Tilden are presented below. The first delta DV value is the subtraction of the two 98th percentile impacts (base minus FIP).

Table 2—United Taconite Predicted Visibility Results for the Most Impacted Areas

Scenario/year

Boundary

Waters days

over 0.5 DV

Boundary

waters 98th

percentile

delta DV

Voyageurs days over 0.5 DV

Voyageurs 98th percentile delta DV

Difference/2002

36

0.594

17

0.454

Difference/2003

39

0.579

14

0.649

Difference/2004

33

0.545

17

0.439

Table 3—Tilden Predicted Visibility Results for the Most Impacted Areas

Scenario/year

Isle Days over 0.5 DV

Isle 98th percentile

delta DV

Seney days over 0.5 DV

Seney 98th percentile delta DV

Difference/2002

2

0.099

3

0.146

Difference/2003

8

0.160

3

0.099

Difference/2004

2

0.112

2

0.125

The baseline emissions associated with the runs above totaled approximately 3,344 tons per year of SO

2

and 3,129 tons per year of NO

X

for United Taconite, and approximately 1,563 tons per year of SO

2

and 2928 tons per year of NO

X

for Tilden. The post-control emissions totaled approximately 233 tons per year of SO

2

and 2,435 tons per year of NO

X

for United Taconite, and approximately 174 tons per year of SO

2

and 2,414 tons per year of NO

X

for Tilden. The United Taconite emissions were based on CEMS data collected under a 100-percent coal-firing scenario, while Tilden emissions were based on stack test information collected under a primarily coal-firing scenario.

EPA believes that Cliffs' modeled baseline emission rates are low based on previous BART modeling and figures from the proposed rule. Expected post-control emissions reductions also appear to be underestimated. The proposed rule identified baseline emissions of approximately 5,330 tons per year for NO

X

and 4,043 tons per year for SO

2

for United Taconite, and approximately 1,153 tons per year of SO

2

and 4,613 tons per year of NO

X

for Tilden. The BART Guidelines recommend that sources use the highest 24-hour average actual emission rate, for the most recent three or five year period of meteorological data, to characterize the maximum potential benefit. By using a low baseline emission rate, Cliffs' modeling underestimates the emissions reductions that will be achieved by the installation of BART controls and the resulting visibility improvements. However, even though the overall SO

2

and NO

X

reductions modeled by Cliffs were over 50 percent lower than the reductions projected in the proposed rule,

22

the results still showed significant visibility improvement at the Boundary Waters. Consequently, EPA believes that Cliffs' modeling provides further evidence that the visibility improvements predicted by the ratio approach are reasonable.

22

This 50 percent discrepancy applies to United Taconite.

Using the CALPUFF model input and meteorological data files submitted by Cliffs, EPA, with substantial assistance from the National Park Service, re-ran the baseline and control-case scenarios for United Taconite and Tilden with data from the proposed rule. For United Taconite, the baseline emissions of SO

2

and NO

X

reflect the emissions presented in the proposed rule. The United Taconite control emissions for NO

X

were based on a 1.2 lbs NO

X

/MMBtu emission limit and heat inputs of 200 MMBtu/hr for line 1 and 260 MMBtu/hr for line 2. The United Taconite control emissions for SO

2

were based on an approximate 94-percent reduction from the base case.

23

For Tilden, the baseline emissions for both visibility pollutants were also based on those contained in the proposed rule. The Tilden control emissions for NO

X

were based on a conversion to 100 percent natural gas, with an 80 percent reduction from the baseline for SO

2

and a 65 percent reduction for NO

X

. The results of this modeling are shown below, but only for the most impacted Class I area.

23

EPA notes that the control emissions for SO

2

at United Taconite differ between those modeled and those that will be achieved based on the final rule because EPA is no longer requiring FGD as a result of the switch to low-sulfur fuels at the facility. This change will result in higher controlled emissions and would be expected to lower the visibility improvements demonstrated in the model slightly.

Table 4—EPA Modeling—United Taconite Predicted Visibility Results for the Most Impacted Area

Scenario/year

Boundary Waters days over 0.5 DV

Boundary Waters 98th percentile delta DV

Difference/2002

80

1.316

Difference/2003

71

1.223

Difference/2004

62

1.358

Table 5—EPA Modeling—Tilden Predicted Visibility Results for the Most Impacted Area

Scenario/year

Seney days over 0.5 DV

Seney 98th percentile delta DV

Difference/2002

0

0.320

Difference/2003

0

0.206

Difference/2004

1

0.165

Again, EPA's CALPUFF modeling shows significant visibility improvement can be expected due to the installation of BART controls at United Taconite and Tilden. EPA believes that these results lend additional support to the accuracy of the visibility analysis that was performed for all facilities in the proposed rule. EPA stands by the results of its ratio approach and believes that it produced reasonable results for the sources examined.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that the proposed rule failed to properly integrate the visibility analysis with cost considerations and that EPA should have identified the costs of controls relative to the visibility improvement using a $/DV metric. In support of this view, Cliffs provided data showing that the costs per DV improvement are very high at two of its facilities: $65 million per DV at United Taconite and $140 million per DV at Tilden. Finally, Cliffs noted that FLMs have cited a threshold of $20 million per DV in correspondence with states, and that any figure beyond this threshold constitutes excessively high costs for the degree of visibility improvement achieved.

Response:

EPA disagrees that a cost per DV analysis was required. The BART Guidelines do not require EPA or the states to conduct such an analysis when evaluating the visibility improvement factor. While the BART Guidelines suggest cost per DV as a possible parameter for consideration, its use is entirely discretionary. There are numerous examples of BART analyses conducted by states and EPA that have not calculated this metric.

Moreover, EPA believes that Cliffs' comment underestimates the visibility impacts from the two facilities that were modeled, leading to erroneous cost per DV figures. As was explained in detail in the response to the previous comment, Cliffs substantially underestimated the baseline emission rates at United Taconite and Tilden, which in turn resulted in emissions reduction estimates that are also too low. The BART Guidelines recommend that the highest 24-hour average actual emission rate, for the most recent three or five-year period of meteorological data, be used to calculate the maximum potential benefit. Overall, the emissions reductions predicted by Cliffs' modeling analysis were less than 50 percent of the emissions reductions projected by EPA for United Taconite.

Finally, Cliffs' reference to the $20 million per DV threshold is misleading. The FLMs recommend that cost per DV be calculated cumulatively to include improvements at all affected Class I areas. Cliffs' analysis, on the other hand, only included visibility improvement at a single Class I area, thereby inflating its total cost per DV figures.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that NO

X

-related visibility improvements should be discounted because nitrate visibility impacts peak in the winter and winter-time visitation at the affected Class I areas is significantly less than during other times of the year. Cliffs noted that the BART Guidelines allow for consideration as to whether impacts occur “during the tourist season.”

Response:

EPA agrees that nitrate impacts are more dominant in the winter. Nonetheless, daily nitrate impacts from April through October are not trivial. EPA also agrees that the BART Guidelines allow states to consider the timing of impacts in addition to other factors related to visibility impairment. However, states are not required to do so, and to our knowledge, neither Michigan nor Minnesota did so in their visibility analyses. EPA is not required to substitute a source's desired exercise of discretion for that of the states. Furthermore, when promulgating a FIP, EPA stands in the shoes of the state. In that capacity, EPA is not required to consider the seasonality of impacts and has chosen not to do so here. Taking into account visitation contradicts the goal of the regional haze rule of improving visibility on the 20 percent best and worst days. Indeed, EPA believes that the experiences of visitors who come to Class I areas during periods other than the peak visitation season are important and should not be discounted.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs cited a number of other flaws in EPA's overall approach to visibility that it believed led to unreliable or overstated impacts from the taconite facilities. First, Cliffs asserted that EPA used natural visibility conditions that were “too clear, excluding conditions such as fires, which had the effect of overstating the impacts of the facilities modeled relative to natural conditions. Second, Cliffs asserted that the chemistry in the current EPA-approved version of CALPUFF, as well as regional photochemical models such as CAMx, overestimates the impact of NO

X

emissions on visibility impairment. Cliffs argued that this is especially true for winter nitrate haze due to the models' static predictions of ammonia background concentrations that should vary seasonally to be in line with monitored observations. As a result, Cliffs concluded that the NO

X

emission reductions that will accompany the installation of BART are being improperly credited with visibility improvements that will not occur in the Minnesota and Michigan Class I areas. Finally, Cliffs cited real-world monitor studies as evidence that large sources that curtailed or shut down operations had little effect on visibility monitors. The first of these studies evaluated the changes in visibility monitoring at the Boundary Waters during periods of low operation at the taconite facilities during 2009. The second study evaluated changes in visibility monitoring at the Grand Canyon after shutdown of the Mohave Power Plant.

Response:

EPA disagrees with these purported flaws in our approach, many of which have been raised in the context of other states' BART determinations. Regarding the issue of natural background conditions, similar issues were addressed in EPA's action on the North Dakota regional haze SIP (77 FR 20909, April 6, 2012). EPA recognizes that variability in natural sources of visibility impairment cause variability in natural haze levels as described in the Agency's “Guidance for Estimating Natural Visibility Conditions under the Regional Haze Rule.”

24

Progress toward natural visibility in Class I areas includes improvement toward natural conditions for the 20 percent worst days and no degradation of visibility on the 20 percent best days. The use of the 20 percent worst days in the calculation of the uniform rate of progress takes into consideration visibility impairment from wild fires, windblown dust, and other natural sources of haze. For the evaluation of visibility impacts for BART sources, however, EPA recommends using the natural visibility baseline for the 20 percent best days for comparison to the “cause or contribute” applicability thresholds. This estimated baseline is reasonably conservative and consistent with the goal of attaining natural visibility conditions. While EPA recognizes that there are natural sources of haze, the use of the 20 percent worst days is inappropriate for the “cause or contribute” applicability thresholds. For example, if visibility impacts were evaluated in comparison to days with very poor natural visibility resulting from nearby wild fires or dust storms, the impacts of BART sources would be significantly reduced relative to these

poor natural visibility conditions and would not be protective of natural visibility on the 20 percent best days.

25

24

Guidance for Estimating Natural Visibility Conditions Under the Regional Haze Rule, U.S. Environmental Protection Agency (September 2003), available at

http://www.epa.gov/ttncaaa1/t1/memoranda/rh_envcurhr_gd.pdf.

“Natural visibility conditions represent the long-term degree of visibility that is estimated to exist in a given mandatory Federal Class I area in the absence of human-caused impairment. It is recognized that natural visibility conditions are not constant, but rather they vary with changing natural processes (e.g., windblown dust, fire, volcanic activity, biogenic emissions). Specific natural events can lead to high short-term concentrations of particulate matter and its precursors. However, for the purpose of this guidance and implementation of the regional haze program, natural visibility conditions represents a long-term average condition analogous to the 5-year average best- and worst-day conditions that are tracked under the regional haze program.” Guidance at 1-1.

25

As part of the settlement of a case brought by the Utility Air Regulatory Group challenging the BART Guidelines, EPA agreed to issue guidance clarifying that states may use either the 20 percent best days or the annual average in estimating natural visibility in the evaluation of a BART source's impacts. This guidance makes clear that states have the flexibility to use either approach in estimating natural background conditions. Here, the states were not required to use the annual average and did not. Similarly, in issuing a FIP, EPA is not required to use the annual average either and chose not to in this case.

In regards to Cliffs' comment on atmospheric chemistry, the approach used by EPA in the proposed rule relied on regional-scale modeling conducted by MPCA where ammonia values varied temporally and spatially. This is in contrast to the approach used in the CALPUFF modeling submitted by Cliffs where a constant 1 ppb monthly average ammonia value was used. While ammonia data is not available for the vicinity of the sources of interest, data is available for sites located in the Class I areas (Fernberg, MN) as well as for sites to the south more representative of northern Wisconsin and southern Minnesota.

26

The available Fernberg ammonia data includes several years of information and has an overall two-week average value of about 0.5 ppb with several two-week periods over 1 ppb. The Perkinstown site located in northern Wisconsin is in an area combining forest, grassland, and agricultural uses and has an overall two-week average ammonia concentration of about 1.5 ppb with several two-week periods over 1.5 ppb. Consequently, the value of 1 ppb used in the modeling submitted by Cliffs is most likely representative of the ammonia concentration in the vicinity of the sources of interest. However, EPA again reiterates that Cliffs' largely baseless criticism of CALPUFF does not apply to the Agency's ratio approach, which relied on regional-scale modeling conducted by the states that included temporal and spatial variations in ammonia concentrations.

26

National Atmospheric Deposition Program,

http://nadp.sws.uiuc.edu/AMoN/sites/data/.

Finally, regarding Cliffs' comment concerning the two monitoring studies, EPA does not find either of the studies to be persuasive with respect to the impacts of taconite sources on visibility. The first study asserts that the 2009 decline in taconite production and a negligible change in visibility are evidence that further controls are not warranted. EPA believes that that it is very difficult to discern any effect from a one-year study and points out that the production decline (as shown in Table 6 below) occurred during the spring and summer, seasons for which Cliffs recognized that nitrate formation is less important.

Table 6—Minnesota 2009 Pellet Production by Month

[Tons]

27

Month

Pellet

production

(tons)

January

2,205,578

February

1,900,003

March

1,620,343

April

958,479

May

181,739

June

340,707

July

849,363

August

1,158,447

September

1,723,336

October

2,008,864

November

2,038,844

December

2,093,403

The second study Cliffs cited, which reviewed visibility monitoring before and after the shutdown of the Mohave Power Plant in Nevada, is a paper by Terhorst and Berkman (Atmospheric Environment, 2010). This paper was subsequently examined and commented on in a paper by White et al. (Atmospheric Environment, January 2012). There, White et al. state: “

[Terhorst and Berkman]'s technical analysis is thoughtfully conceived and executed, but is misleadingly presented as discrediting previous studies and their interpretation by regulators. In reality the Terhorste Berkman analysis validates a consensus on MPP's (Mohave Power Project) visibility impact that was established years before its closure, in a collaborative assessment undertaken jointly by Federal regulators and MPP's owners.” Additionally, EPA has responded to similar comments regarding the Mohave Power Project study and EPA's visibility modeling in our action on the North Dakota regional haze SIP (77 FR 20894, April 6, 2012).

28

27

Source: Minnesota Department of Revenue, Minerals Tax Division, Eveleth, MN.

28

There, EPA stated: “In addition, the study by Terhorst and Berkman does not convince us that use of CALPUFF modeling is inappropriate for this action or that the CALPUFF modeling results should be ignored. A model such as CALPUFF essentially holds constant a number of factors in order to isolate the impacts of a single source. As acknowledged by the study's authors, it is extremely difficult in observational analyses to sufficiently control for all factors, including emissions from other sources, to be able to isolate the impacts of closure of a facility, especially one located over 100 km from the Class I area at issue. In fact, the paper notes that coarse soil mass impacts are an omitted variable in the analytical analysis and that changes in those emissions may have counteracted the visibility improvements expected from the source shutdown” (77 FR 20894, 20910).

Finally, EPA believes that Cliffs, while identifying purported areas where EPA's models exaggerate visibility impacts, overlooks that there are aspects of the models that have been suggested by commenters on our regional haze actions as under-predicting impacts. Some examples include use of 24-hour average emissions impacts. Some examples include use of 24-hour average emissions rather than hourly emissions, use of monthly average relative humidity rather than daily humidity, and use of 98th percentile results to compare to the threshold instead of the highest day.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that certain control options should be rejected because visibility modeling does not indicate the installation of controls will result in a perceptible visibility improvement.

Response:

EPA's disagrees. As explained in the proposed rule, EPA believes that the application of BART will result in perceptible improvements in visibility. Nonetheless, the perceptibility of visibility improvement is not a prerequisite to the selection of a control option as BART. The preamble to the BART Guidelines state, “Even though the visibility improvement from an individual source may not be perceptible, it should still be considered in setting BART because the contribution to haze may be significant relative to other source contributions in the Class I area. Thus, we disagree that the degree of improvement should be contingent upon perceptibility” (70 FR 39104, 39129, July 6, 2005).

Minnesota's regional haze SIP described the importance of the contribution of sources in northeastern Minnesota to visibility impairment in the Boundary Waters and Voyageurs national parks. Accordingly, Minnesota developed a special plan for the northeast region of the state. Minnesota explained:

This area was targeted for controls under the long-term strategy for several reasons. First, the MPCA's analysis of 2002 emissions from the top 18 emitting point sources within Minnesota show that sources from this region make up just

1/3

of the total emissions but provide

2/3

of the total visibility impact. (See Chapter 8, on modeling.) Therefore, they have a much larger impact on the Class I areas than emissions from farther away. In addition, the taconite facilities may be currently uncontrolled or under-controlled for SO

2

or NO

X

, and on the books control strategies are projected to cause fewer

emission decreases in this region than in the remainder of the state.

29

29

Minnesota's 2009 regional haze SIP submittal at 96 (Docket # EPA-R05-OAR-2010-0037-0002).

Thus, Minnesota's assessment supports the determination that taconite facilities contribute to regional haze even if individual impacts are modeled below thresholds for human perceptibility.

K. Comments Concerning Requirements for Continuous Emissions Monitoring

1. Comments in Support of Continuous Emissions Monitoring Requirements

Commenter:

National Parks Conservation Association.

Comment:

EPA's proposed rule includes the use of CEMS as a part of the monitoring, recordkeeping, and reporting necessary to ensure the continuous application of BART. Such monitors provide more accurate data about the emissions from taconite facilities than previous methods. As such, they are essential for tracking emissions and determining their impact on the surrounding communities. Moreover, CEMS can be used as pollution control tools by helping to fine-tune combustion and process controls in a way that periodic stack tests and predictive monitoring cannot. As such, we fully support the required application of CEMS on these sources.

Commenter:

National Park Service.

Comment:

NPS is especially pleased that EPA has proposed testing and CEMS requirements for the subject taconite plants. Our discussions with U.S. Steel, which has led the way in installation and operation of CEMS on indurating furnaces, have led to the mutual agreement that CEMS data is essential for the proper tuning and operation of combustion controls to reduce NO

X

emissions. Minnesota's regional haze SIP discussed the need for requiring CEMS for the taconite industry to monitor NO

X

for a number of reasons. These included setting BART limits, allowing facilities to efficiently manage combustion, resulting in less fuel use and fewer emissions, and tracking progress under the Northeast Minnesota Plan. CEMS data is also essential for assessing the effectiveness of SO

2

controls and should provide an indication of changes in fuel use or sulfur content for use in future regional haze planning.

Commenter:

Leech Lake Band of Ojibwe.

Comment:

The Band agreed with the implementation of CEMS. CEMS will provide the facility and regulators with real time data to ensure that the controls in place are operating at optimum levels, thus saving money for the facility and achieving the required control requirements.

Commenter:

National Tribal Air Association.

Comment:

The Association found requiring CEMS to be a good complement to the NO

X

and SO

2

controls at taconite ore processing facilities. CEMS will provide these facilities with an accurate and timely emissions count of NO

X

and SO

2

and will immediately alert owners to any deviations from such emissions that might require correction.

Response:

EPA acknowledges the comments in support of the proposed CEMS requirements. EPA is finalizing the CEMS requirements as proposed.

2. Comments Questioning EPA's Continuous Emissions Monitoring Requirements

Commenter:

U.S. Forest Service.

Comment:

In the Minnesota regional haze SIP, a statement is made that CEMS “would apply to NO

X

emissions at the facilities burning natural gas and to SO

2

emissions at facilities burning high sulfur fuels.” We do not understand why the NO

X

CEMs are only being required at natural gas-fired furnaces. Those furnaces burning fuels other than natural gas will also investigate NO

X

control strategies and therefore will need the CEMs.

Response:

EPA is requiring CEMS for all seven subject-to-BART taconite facilities. Each taconite facility must monitor its NO

X

and SO

2

emissions with CEMS.

Commenter:

U. S. Steel and Cliffs Natural Resources.

Comment:

Cliffs and U.S. Steel stated that it is inappropriate to require the use of a diluent monitor as part of the monitoring requirements under the proposed rule.

Response:

The need to install a diluent monitor is source-specific and depends on a variety of factors, including the choice of monitors installed. While some subject-to-BART facilities may not need to install a diluent monitor, other facilities may because of their stack characteristics, operating conditions, and monitors chosen. Because the final rule covers a variety of facilities, EPA feels it is appropriate to require a diluent monitor only for those facilities needing such a unit. Therefore, the final rule has been revised to provide the facilities with an option to demonstrate in their monitoring plans whether a diluent monitor is needed or not.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that EPA failed to adequately support its CEMS requirements and that EPA should allow flexibility in the monitoring requirements. More specifically, EPA should require “a comparable method of emission estimation,” such as parametric emissions monitors, for each subject-to-BART source. Cliffs stated that EPA's unsupported generalized statements do not provide adequate justification for the Agency's burdensome monitoring determination.

Response:

EPA clearly states in its technical support for the proposed rule that CEMS are the best method for demonstrating compliance because of the variability in furnace operations and variable fuel usage across the furnaces. The variable fuel feeds and feed material content can impact overall emissions from the process and thereby create the need for continuous monitoring of emissions that impact visibility. Parametric emissions monitor systems are an option for processes that operate at stable, non-variable conditions, but are not appropriate for taconite units. CEMS provide a continuous record of data that can also be used by the facility owner or operator to monitor emissions on a real-time basis. The installation and operation of CEMS and the real-time evaluation of the CEMS data provide several benefits to a facility that can directly lead to practices that reduce emissions during all periods of operation.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that CEMS do not constitute proven monitoring technology with respect to taconite furnaces.

Response:

EPA disagrees. CEMS technology for NO

X

and SO

2

is proven in multiple industries, including the taconite industry. U.S. Steel is successfully using CEMS at its Minntac and Keetac facilities currently, and any problems experienced with the initial installation of CEMS have been resolved. EPA expects facilities will need some time to learn CEMS operation and how it impacts process operations. EPA has incorporated additional time into the final rule before certification is required to allow each facility to learn how CEMS operates.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs asserted that the requirement to install CEMS should be limited to waste stacks. Cliffs proposed to use stack testing data to determine the percent distribution of NO

X

between the hood-exhaust header and the waste-gas header to determine compliance. Cliffs also stated that the use of a single

CEMS per furnace is consistent with Minntac, which currently utilizes NO

X

and SO

2

CEMS only.

Response:

EPA disagrees with Cliffs that the use of CEMS should be limited to waste stacks only. Available information shows that emissions from hood-exhaust stacks can equal about 29 percent of total furnace emissions. Given that nearly one-third of the furnace emissions are from hood exhausts (and can vary), EPA believes that it is appropriate to require CEMS on both waste stacks and hood-exhaust stacks.

Cliffs incorrectly assumed that CEMS are required at each stack venting to the atmosphere. EPA feels it is important to obtain continuous and representative measurements of emissions from subject-to-BART units. If representative measurements of total emissions from subject-to-BART units can be obtained by installing CEMS in a single vent just prior to the common header for the waste-gas stacks and hood-exhaust stacks, then this final rule requires only two CEMS on each stack (one NO

X

CEMS and one SO

2

CEMS) for a total of four CEMS, not ten. The initial monitoring plans required by this final rule will be prepared by the facilities and will provide a means through which an effective monitoring program will be put into place. These plans should include proposals for CEMS types, CEMS numbers, CEMS installation locations, QA/QC procedures, and any other topics and are submitted to EPA for review and approval or disapproval. The installation locations provided in these plans shall be determined based on the requirements of 40 CFR part 60, appendices B and F.

Commenters:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that requiring CEMS on emergency stacks is inappropriate.

Response:

The requirement to install CEMS on emergency stacks depends on the frequency and duration of the use of the emergency stacks during emergency events. If emergency stacks are used on a daily or weekly basis, then emissions from those stacks could have an impact on annual emissions (and visibility) and should be tracked and recorded. If emergency stacks are truly used infrequently for quick releases, then a CEMS may not be necessary. This can be addressed by each facility on a case-by-case basis in its monitoring plan.

Commenters:

Cliffs Natural Resources and ArcelorMittal Minorca Mine

Comment:

Cliffs stated that the subject-to-BART facilities should be exempt from the applicable emission limits during startup, shutdown and malfunction events.

Response:

EPA disagrees. The CAA requires sources to comply with applicable emission limits at all times, including during startup, shutdown and malfunction. See, e.g.,

Sierra Club

v.

EPA,

551 F.3d 1019, 1021 (DC Cir. 2008);

US Magnesium, LLC

v.

EPA,

690 F.3d 1157, 1170 (10th Cir. 2012).

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine and U.S. Steel.

Comment:

Cliffs commented that the requirement to develop and implement a corrective action program for excess emission events should be directed toward the emissions unit and not be part of the CEMS requirement.

Response:

EPA agrees that the corrective action plan for excess emission events should be directed toward the emissions unit. The corrective action plan should establish procedures that operators will follow each time an excess emission event occurs (as identified through the use of real-time CEMS data). These procedures should outline steps to adequately identify causes of excess emissions, actions to be taken to minimize or eliminate those emissions, and evaluate and implement practices to prevent the causes of such excess emissions from reoccurring. The corrective action plan can be an independently developed plan or the procedures can be incorporated into an existing Quality Control Program Plan, corrective action plan, or other existing standard operating plan.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that the dates proposed for installing CEMS are infeasible and suggested an alternative compliance period of 18 months.

Response:

EPA disagrees with the length of time Cliffs asserted is needed to design and install CEMS. EPA recognizes that a certain period of time will be needed if significant upgrades to stacks are necessary. However, the design and installation of a CEMS can be completed far more rapidly than the 18-month period suggested by Cliffs. EPA also believes it is inappropriate to consider the time needed for CEMS installation in a cumulative sequential manner as suggested by Cliffs. Design, engineering requirements, and upgrades to data acquisition systems can be performed at the same time as other activities required by the proposed rule.

EPA also recognizes that once CEMS are installed and operating, there will be a short period of time needed to optimize and become familiar with the system in order to certify the units. EPA believes that the entire process for CEMS installation can be successfully met within the time periods outlined in the proposed rule. However, in response to the comments received, the final rule provides an additional 30 days to certify the CEMS and perform the initial Relative Accuracy Test Audit of the CEMS. The anticipated dates for initial certification and the Relative Accuracy Test Audit must be included in the monitoring plan required by the rule.

Commenter:

U.S. Steel.

Comment:

U.S. Steel commented that it is overly burdensome to require redundant or backup monitoring systems to obtain emissions data during periods of primary CEMS breakdown, repair, calibration check, or zero span adjustment. U.S. Steel proposed to use data gap-filling procedures during those periods when data are not available from the CEMS due to these types of events.

Response:

The purpose of including the requirement to use “other monitoring systems approved by EPA” is to obtain real-time emissions data during periods of primary CEMS breakdown, repair, calibration check, or zero span adjustment. The secondary data can be used to assure data availability and compliance on a continuous basis. However, the requirement for “other monitoring systems” does not mean that a second CEMS system is necessarily needed. Because the duration of these CEMS downtime events is typically short, each subject-to-BART facility can submit proposals for using parametric monitoring or engineering estimates as a surrogate for actual emissions monitoring during these CEMS events. EPA expects that CEMS will be operated at all times, including periods of process unit startup, shutdown and malfunction, except during the events identified above as described at 40 CFR 60.13(e).

EPA also disagrees with the suggestion that gap-filling procedures (i.e., data substitution) should be used for periods of CEMS downtime. Gap-filling procedures are appropriate under 40 CFR part 75 because it is a cap-and-trade program. This final rule is more appropriately related to regulations at 40 CFR part 60, where compliance with an emission limit (rather than annual caps) is required. 40 CFR part 60 prohibits the use of “data substitution” (i.e., gap-filling) because it does not provide accurate emission rates during the CEMS downtime.

Commenter:

U.S. Steel.

Comment:

U.S. Steel commented that it is not appropriate to require initial performance testing of subject-to-BART

facilities or units if the facility is operating a certified CEMS system on the affected units.

Response:

EPA has re-evaluated the need for initial performance testing and agrees with U.S. Steel that it is not necessary to require such testing, for purposes of this rule, at facilities that are or will be operating CEMS when those CEMS will be used to determine compliance. The requirement for initial performance testing has been removed from the final rule. It is important to note that while initial performance testing is being removed for purposes of demonstrating compliance, subject-to-BART units must still be tested as part of the CEMS certification process, although this the certification process will typically not require a 30-day test.

L. Comments Concerning Compliance Schedules

Commenter:

Leech Lake Band of Ojibwe.

Comment:

The Band supported the installation of low NO

X

burners and felt that the 1.5 years allowed for the initial installation, with additional burner installations to follow one year later, is a fair and progressive approach to control NO

X

emissions.

Response:

EPA agrees with the commenter that installing low NO

X

burners is the appropriate approach. In response to additional information submitted by other commenters, however, EPA reviewed the proposed installation schedule has extended it by a number of months in the final rule, as discussed in more detail below.

Commenter:

Cliffs Natural Resources and ArcelorMittal Minorca Mine.

Comment:

Cliffs commented that the BART Guidelines require the states, or EPA when promulgating a FIP, to establish deadlines for compliance with BART emission limits no later than five years from the date of approval or promulgation. EPA's proposed rule, however, contains arbitrary compliance deadlines that are unreasonably short. Cliffs stated that it would take at least four and a half years for it to complete the required engineering, installation, and commissioning of low NO

X

burners for a single furnace. ArcelorMittal stated that “a minimum of 48 months would be necessary to complete this onerous process” for its Minorca Mine facility.

Response:

CAA section 169A(b)(2)(A) requires subject-to-BART sources to install BART and comply with any applicable emission limits “as expeditiously as practicable.” The Act defines this term to mean “as expeditiously as practicable but in no event later than five years after * * * the date of promulgation.” CAA section 169A(g)(4). This language does not indicate that a compliance schedule of five years is to be assumed. Rather, BART must be installed “as expeditiously as practicable,” meaning as soon as the source is capable of installing the controls and meeting the applicable emission limits.

In response to EPA's request for a detailed timeline of the steps required to install low NO

X

burners on a taconite furnace, U.S. Steel provided the following information based on its actual experience with a previous install:

Table 7—U.S. Steel's Estimate Compliance Schedule

Task

Time

Detailed low NO

X

burner engineering

6 months.

Prepare permit applicability determination

2 months.

Procure and manufacture low NO

X

burner

8 months.

Install low NO

X

burner

3 weeks.

Shakedown of low NO

X

burner

6 months.

Total time

22 months, 3 weeks.

In addition to this information, ArcelorMittal included an attachment to its comments of a September 27, 2012 report by Fives North American titled “Retrofitting Low NO

X

Burners on the ArcelorMittal Minorca Straight-Grate Pelletizing Furnace.” In that report, Fives states that to develop an engineering solution that complies with environmental requirements, it is important to allow sufficient time (approximately four to eight months) for engineering analysis and (possibly) testing. The schedule should allow for an additional seven months for fabrication and delivery, followed by an additional two months for installation and commissioning. This amounts to an estimated time of 17 months to achieve compliance.

Based on the timeline provided by U.S. Steel, the vendor estimate from Fives, and concerns from the commenters, EPA is allowing 26 months for a company's first indurating furnace to comply with the final rule. This will allow each source sufficient time to perform an engineering analysis, prepare a permit applicability determination, manufacture and install the low NO

X

burner, and allow for a shakedown period to achieve compliance after the low NO

X

burner has been installed. This is eight months longer than the proposed compliance schedule, allowing for the additional time needed for a shakedown period.

The compliance schedule for additional indurating furnaces is being finalized as proposed. Specifically, a second line has an additional year to comply, for a total of 38 months from the effective date of the rule. A third line has two additional years to comply, for a total of 50 months from the effective date of the rule. This staggered installation schedule will minimize any potential impacts on production. EPA notes that U.S. Steel Minntac is following a shorter schedule consistent with the proposed rule. U.S. Steel Keetac will also follow a modified schedule. For more detail, see the comment below.

Commenter:

U.S. Steel.

Comment:

Due to the lead time associated with acquiring process fans at Keetac, which is estimated to be 52 weeks according to a third-party engineering firm working on the project, and the timing of the major outage schedule, which only occurs once per year, the potential exists to miss the timing window where the two sync up to meet the proposed schedule. Therefore, U.S. Steel requests an additional 12 months to the proposed schedule for installation. In addition, because this will be the first installation of this technology at Keetac, U.S. Steel requests an additional 6 months prior to compliance with the proposed emission limit to allow for a shakedown period to optimize the burner for NO

X

reductions.

Response:

EPA agrees with U.S. Steel that additional time is needed to procure new pre-heat fans and to achieve compliance after installation. In the final rule, Keetac has three years (36 months from the effective date of the

rule) for its single line to comply with the rule.

Commenter:

U.S. Steel.

Comment:

Due to timing of major outage schedules, U.S. Steel requests flexibility in the order of installation for Line 4 and Line 5 at Minntac. U.S. Steel agrees with the overall intent of the proposed schedule, but requests the option to select the order of installation of low NO

X

burners at Lines 4 and 5.

Response:

EPA agrees with the U.S. Steel's request to leave to the discretion of U.S. Steel the order of installation of low NO

X

burners at Lines 4 and 5 at Minntac.

M. Comments Asserting That EPA Must Conduct Government-to-Government Consultation With the Tribes

Commenter:

National Tribal Air Association (NTAA).

Comment:

The Association understands that at the request of the Fond du Lac Band, EPA held a June 28, 2012 conference call with the region's Tribes to discuss the FIP. We appreciate EPA for doing this and highly recommend that the Agency hold similar calls with Tribes for other such actions. However, EPA must also honor its commitment to conduct formal government-to-government consultation in accordance with Executive Order (EO) 13175.

The Association disagrees with EPA's statement that the FIP “does not have tribal implications, as specified in EO 13175. It will not have substantial direct effects on tribal governments. Thus, EO 13175 does not apply to this rule.” The application of BART to taconite ore processing facilities that either are in close proximity to Tribes and their communities or are within Treaty-ceded territory areas maintaining Tribes' usufructary functions is a regulatory action that has Tribal implications. As such, EPA must conduct formal government-to-government consultation with Tribes.

There are also clear purposes for EPA to conduct formal government-to-government consultation with Tribes. First, it provides for more candid conversations between individual Tribes and EPA than would occur otherwise in a group meeting involving other Tribes. Second, each Tribe's circumstances are unique and must be treated as such by EPA. Group meetings would only give short shrift to these circumstances. Third, most cultural resources information is protected from release under statutory exemptions to the Freedom of Information Act. Discussion of such information as part of a group meeting risks its release to the general public and potentially endangers Tribal cultural sites and practices. Finally, the subject matter may be so unique, such as a dispute between an individual Tribe about whose cultural resources might be located within or near a taconite ore processing facility, that government-to-government consultation between the Tribe and EPA could provide the best opportunity for a resolution to the situation versus a group meeting where any number of issues might be discussed in a finite period of time.

Response:

EPA acknowledges that this action may have tribal implications. EPA recognizes that Tribes may have significant interests in regulatory programs even if the potential Tribal impacts are not the types specifically identified in the Executive Order. In this case, EPA initiated consultation with Tribal officials early in the process of developing this regulation to permit them to have meaningful and timely input into its development. While the Tribes ultimately chose not to engage in individual consultation, EPA did communicate with Tribal representatives to ensure that information was made available and that there was sufficient opportunity for questions and discussion. This effort is described in further detail in Section IV of this final rule. EPA appreciates the comments provided by the Tribes and NTAA on this rule, which will benefit Tribes through reduced pollution and improved visibility.

N. Comments Concerning Non-Air Quality Impacts of the Proposed Rule

Commenter:

Cliffs Natural Resources.

Comment:

Cliffs commented that significant environmental impacts will result if the rule is finalized as proposed. Cliffs stated that increased fuel combustion resulting from low NO

X

burner application will result in increased emissions of the products of combustion. Cliffs added that EPA's proposed SO

2

controls also carry ancillary environmental consequences.

Response:

EPA disagrees. The BART Guidelines recognize that environmental concerns become important when sensitive site-specific receptors exist and are impacted by byproducts of the control device. However, the fact that a control device creates liquid and solid waste that must be disposed of does not necessarily argue against that technology as BART. In this case, there are no such sensitive, site-specific issues. To avoid any such issues, EPA rejected the use of wet SO

2

scrubbing at taconite plants in Minnesota because wastewater from wet scrubbing had the potential to interfere with the production of wild rice.

Commenter:

National Tribal Air Association.

Comment:

The Association commented that Tribal traditional practices will benefit by controlling NO

X

and SO

2

emissions from taconite ore processing facilities. Such benefits specifically relate to visibility, health, and acid deposition.

Not only does regional haze, which the FIP addresses, reduce the clarity, color, and visible distance that one can see, it marginalizes Tribal traditional practices that have existed since time immemorial. Many Tribes engage in traditional practices associated with sacred mountains, lakes, or other places

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