
[Federal Register Volume 78, Number 189 (Monday, September 30, 2013)]
[Rules and Regulations]
[Pages 59825-59839]
From the Federal Register Online via the Government Printing Office [www.gpo.gov]
[FR Doc No: 2013-23394]


=======================================================================
-----------------------------------------------------------------------

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 52

[EPA-R05-OAR-2010-0954 and EPA-RO5-OAR-2010-0037; FRL9901-31-Region 5]


Approval and Promulgation of Air Quality Implementation Plans; 
States of Michigan and Minnesota; Regional Haze

AGENCY: Environmental Protection Agency (EPA).

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: In this notice of final rulemaking, EPA is disapproving in 
part the Michigan and Minnesota regional haze State Implementation 
Plans (SIPs) for failure to mandate best available retrofit technology 
(BART) for taconite facilities within these states. This final rule 
supplements a February 6, 2013, action that established Federal 
emission limits representing BART for these facilities.

DATES: This final rule is effective on October 30, 2013.

ADDRESSES: EPA has established dockets for this action under Docket ID 
Numbers EPA-R05-OAR-2010-0954 and EPA-RO5-OAR-2010-0037. All documents 
in the dockets 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 and Maintenance Section, Air Programs 
Branch (AR-18J), Environmental Protection Agency, Region 5, 77 West 
Jackson Boulevard, Chicago, Illinois 60604, (312) 886-6524, 
rosenthal.steven@epa.gov.

SUPPLEMENTARY INFORMATION: Throughout this document whenever ``we,'' 
``us,'' or ``our'' is used, we mean EPA. This supplementary information 
section is arranged as follows:

I. What is the background for this action?
II. What are EPA's responses to the public comments it received?
III. What action is EPA taking?
IV. Statutory and Executive Order Reviews

I. What is the background for this action?

    Minnesota submitted its regional haze SIP on December 30, 2009, a 
draft supplement on January 5, 2012, and a final supplemental 
submission on May 8, 2012. EPA proposed approval of the Minnesota 
regional haze SIP on January 25, 2012 (77 FR 3681). Among other 
actions, the proposed rule proposed to conditionally approve 
Minnesota's regional haze SIP as satisfying the BART requirements of 
the Clean Air Act (CAA or the ``Act'') section 169A(b)(2)(A) and 40 CFR 
51.308(e) for the State's six taconite plants, provided that Minnesota 
submit emission limits representing BART prior to EPA's final action. 
During the comment period on EPA's proposed rule, EPA received comments 
providing evidence that better, cost-effective technology for the 
control of taconite plant emissions was available that Minnesota (and 
Michigan) failed to adequately consider in the SIP revision. Therefore, 
EPA published a final rule approving other aspects of the Minnesota 
regional haze SIP on June 12, 2012 (77 FR 34801), but deferred action 
on BART for Minnesota's taconite facilities.
    Michigan submitted its regional haze SIP on November 5, 2010. EPA 
proposed action on the Michigan regional haze SIP on August 6, 2012 (77 
FR 46912). In this action, EPA proposed to approve several aspects of 
Michigan's regional haze SIP, and proposed to disapprove Michigan's 
BART determinations for a Portland cement plant and a paper mill and 
proposed Federal limits for those two facilities. EPA published final 
action pursuant to this proposal on December 3, 2012 (77 FR 71533). 
However, similar to Minnesota, EPA deferred action on BART for the 
Tilden Mining taconite facility in Michigan.
    On August 15, 2012 (77 FR 49308), EPA published a proposed partial 
disapproval and Federal Implementation Plan (FIP) for BART for taconite 
plants in Minnesota and Michigan. In that action, EPA reviewed relevant 
information regarding the technical feasibility of various options for 
the control of emissions from taconite plants and reviewed other 
information relevant to determining BART for these plants. On February 
6, 2013 (78 FR 8706), EPA published a final rule establishing a FIP to 
implement BART for the taconite facilities in Minnesota and Michigan.
    Also on February 6, 2013 (78 FR 8478), EPA supplemented its 
proposed partial disapproval of the Minnesota and Michigan SIPs for 
failure to require BART for taconite facilities within these states. 
EPA published this notice in response to comments that EPA had not 
adequately explained its rationale for proposing to disapprove the 
States' BART determinations for taconite in its August 15, 2012 
proposed action.

II. What are EPA's responses to the public comments it received?

    In response to its supplemental proposed rulemaking, EPA received 
comments from ArcelorMittal Minorca Mine, Incorporated (ArcelorMittal), 
Cliffs Natural Resources, Inc. (Cliffs), U.S. Representative Richard M. 
Nolan, the Minnesota Pollution Control Agency

[[Page 59826]]

(MPCA), the National Mining Association (NMA), and the Michigan 
Department of Environmental Quality (MDEQ). The following discussion 
provides a summary of the comments and EPA's responses. The comments by 
ArcelorMittal are essentially identical to the comments from Cliffs, 
except that they do not refer to taconite-related issues in Michigan. 
Unless a comment by Cliffs is regarding taconite-related issues in 
Michigan, a comment ascribed to Cliffs is also from ArcelorMittal.

A. Comments by Cliffs and/or ArcelorMittal

    Comment: The CAA gives primary authority for regional haze 
determinations to the states. States are responsible for developing and 
implementing the regional haze program. States are responsible for 
identifying BART-eligible sources, defining BART for each source, 
establishing reasonable progress goals, and developing long-term 
strategies to reduce regional haze in class I Federal areas.
    The D.C. Circuit affirmed the primacy of states in implementing the 
regional haze program in American Corn Growers Ass'n v. EPA, 291 F.3d 1 
(D.C. Cir. 2002). There, the court confirmed the primacy of state 
authority in this area by invalidating EPA's regulations on the grounds 
that they impermissibly constrained state authority. EPA has only a 
limited role in evaluating regional haze SIPs because the CAA calls for 
states to play the lead role in implementing the regional haze program. 
EPA has conceded as much by acknowledging that states retain the 
primary responsibility of developing a viable visibility program and 
states must determine the appropriate level of BART control for each 
source subject to BART.
    Response: Section 110 of the CAA requires states to develop SIPs 
with enforceable emission limitations and other control measures to 
meet the applicable requirements of the Act. A state must then submit 
its SIP 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 SIPs to determine whether a given SIP 
meets all applicable requirements of the Act. See CAA sections 
110(k)(3) and (l).
    The D.C. Circuit's decision in American Corn Growers did not alter 
this balance. The court's decision there was limited to a holding that 
EPA could not require states to evaluate the first four BART factors on 
a source-specific basis, while requiring states to evaluate visibility 
improvement on a group-wide basis. In contrast, the Tenth Circuit 
recently affirmed EPA's authority to evaluate regional haze SIPs for 
compliance with all requirements of the Act, including the visibility 
protection provisions in section 169A and EPA's implementing 
regulations at 40 CFR 51.300-51.309 and 40 CFR pt. 51, app. Y. See 
Oklahoma v. EPA, -- F.3d -- (10th Cir. 2013). As discussed in our 
February 6, 2013, supplemental proposed disapproval, and in the 
response to comments in this final disapproval, EPA has determined 
that, while it agrees with Minnesota and Michigan's identification of 
BART-eligible sources, we find that the states did not satisfy the 
requirements for BART for the taconite facilities.
    Comment: EPA rushed to issue a FIP, apparently leaving EPA without 
enough time to explain to the states how their SIPs were deficient. 
Only now, after finalizing its FIP, does EPA stop to ``take comments'' 
on the basis for its proposed disapproval of the SIPs, which is a sham 
that offers no real opportunity for commenters to influence the outcome 
of the regional haze process already determined by the FIP. EPA should 
stay the FIP until it properly considers public comments on its basis 
for disapproving the SIPs.
    Response: In its comments dated September 28, 2012, addressing 
EPA's August 15, 2012, proposed action, Cliffs commented that ``EPA's 
Proposed Rule does not discuss the validity of the extensive factual 
information and technical analysis underlying Minnesota's and 
Michigan's BART determinations,'' and that ``EPA was forcing the public 
and the States to guess at what EPA believes was wrong with Minnesota's 
and Michigan's SIP submittals.'' In other words, Cliffs commented in 
September 2012 that EPA needed to provide a more extensive explanation 
of its basis for proposing to disapprove Minnesota and Michigan's BART 
determinations for taconite facilities, but now believes that the 
opportunity EPA has provided is rendered meaningless by the 
promulgation of EPA's FIP in February 2012.
    We disagree. As explained in the final rule promulgating the FIP, 
EPA's FIP obligation following a finding of failure to submit remains 
in effect, irrespective of a subsequent state SIP submittal, unless and 
until EPA approves the SIP. See CAA section 110(c). A FIP is a gap-
filling measure only, however. See CAA section 302(y). As a result, a 
FIP promulgated by EPA remains in place only until a state submits a 
SIP correcting the inadequacy and that SIP is approved by EPA. In this 
instance, Cliffs and other commenters had a meaningful opportunity to 
comment on EPA's supplemental proposed disapproval and explain why 
EPA's proposed action was incorrect. Had EPA agreed with Cliffs or 
other commenters and approved the Minnesota and Michigan BART 
determinations for taconite facilities instead, EPA's FIP would have 
been replaced. Therefore, Cliffs' assertions are incorrect. 
Nevertheless, we note that Cliffs' request for a stay is now moot 
because the Eighth Circuit has already granted such a stay.
    Comment: EPA has a limited role in the regional haze process and 
therefore must defer to state determinations that meet minimum 
requirements. EPA's role in the regional haze program is limited to 
approving or disapproving SIPs submitted by the states, and EPA has 
limited discretion to disapprove a SIP, as outlined by CAA section 
110(k).
    The Minnesota SIP was deemed complete on June 30, 2010, and the 
Michigan SIP was deemed complete on May 5, 2011. Once a SIP is deemed 
complete, EPA has 12 months to act on it and ``shall approve such 
submittal as a whole if it meets the applicable requirements . . .'' 
EPA's role is limited to the ministerial function of reviewing SIPs for 
consistency with the Act's requirements.
    Response: The commenters are correct with respect to the schedule 
in CAA section 110(k). However, nothing in this section states, or even 
implies, that EPA must automatically approve a SIP within 12 months 
after a SIP is deemed complete. Further, this section states that EPA 
shall only approve a SIP if it meets all of the applicable requirements 
of the Act. While Congress intended states to take the lead in 
developing regional haze SIPs, it balanced that decision by requiring 
EPA to review the states' SIPs to determine whether they meet the 
applicable requirements of the Act. EPA's review is not limited to the 
ministerial function of rubber-stamping a state's decisions. Rather, in 
reviewing regional haze SIPs in general and BART determinations in 
particular, EPA must consider not only whether the state considered the 
appropriate factors, but also 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.
    Comment: Contrary to EPA's belief, the Ninth Circuit's decision in 
Sierra Club v. EPA, 671 F.3d 955 (9th Cir. 2012), does not authorize 
the Agency to disapprove a complete SIP every time new information 
becomes available. The Ninth Circuit's decision rests heavily on

[[Page 59827]]

CAA section 172(c)(3)'s requirement that nonattainment plans include a 
comprehensive, accurate, current inventory of actual emissions. No such 
provision exists in CAA section 169A, which governs regional haze. 
Rather, EPA is on record instructing states conducting BART 
determinations that ``technologies should be considered if available 
before the close of the State's public comment period'' and that they 
``need not consider technologies that became available after this 
date.'' That is precisely what Minnesota and Michigan did, and EPA 
cannot use a decision from a different jurisdiction based on different 
statutory language to change course now that it prefers a different 
result. Rather, as the D.C. Circuit has held, ``[t]o require states to 
revise completed plans every time'' new information arises ``would lead 
to significant costs and potentially endless delays in the approval 
process.'' Sierra Club v. EPA, 356 F.3d 296, 08 (D.C. Cir. 2004).
    Response: EPA disagrees about the scope of the Ninth Circuit's 
decision, which states that EPA should evaluate any new information 
available and ``may not simply ignore it without reasoned explanation 
or choice.'' Sierra Club, 671 F.3d at 967. The Ninth Circuit does not 
couch this statement narrowly in the context of EPA's review of 
nonattainment plan inventories, but rather presents it broadly as a 
principle of administrative law. Indeed, the Ninth Circuit's full 
holding states: ``But we should not silently rubber stamp agency action 
that is arbitrary and capricious in its reliance on old data without 
meaningful comment on the significance of more current compiled data. 
We hold that EPA's failure to even consider the new data and to provide 
an explanation for its choice rooted in the data presented was 
arbitrary and capricious.'' Id. at 968.
    Irrespective of the significance of Sierra Club v. EPA, however, 
Cliffs mischaracterizes the technical feasibility provisions of the 
BART Guidelines. The statement there that a state need not consider 
technologies that are not commercially available by the end of the 
state's public comment period for its SIP bears no relation to the 
question of whether a technology that has been commercially available 
for decades, such as low NOX burners, is applicable to a 
specific source. Furthermore, even if Cliffs' incorrect reading of the 
BART Guidelines were correct, both Minnesota and Michigan were aware 
that low NOX burners had been successfully applied to 
taconite furnaces before the ends of their respective public comment 
periods. In a June 23, 2010, letter to Michigan regarding the state's 
draft regional haze SIP, EPA commented that ``a low NOX 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 NOX formation in 
taconite furnaces.'' Similarly, in a February 10, 2012, letter to 
Minnesota 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 NOX main burners 
that achieve 70 percent NOX reduction on its indurating 
lines.'' Therefore, both States were well aware that low NOX 
burners had been successfully applied to indurating furnaces. The 
states did not provide reasonable explanations for dismissing this 
information and instead continued to rely on the taconite facilities' 
cursory and unsupported determinations from 2006 that low 
NOX burners were technically infeasible.
    Comment: Even if EPA legitimately determines that a SIP does not 
meet the minimum criteria for approval after giving states appropriate 
deference, EPA's actions remain proscribed by the Act. Section 110(c) 
permits EPA to issue a FIP ``unless the state corrects the deficiency'' 
EPA identified. Had EPA followed the procedure required by the Act and 
addressed the States' SIPs prior to issuing a FIP, any perceived issues 
could have been resolved.
    Response: This comment addresses EPA's final rule promulgating the 
FIP and not EPA's supplemental proposed disapproval and is therefore 
not relevant to this rulemaking. Nevertheless, we point out that Cliffs 
fails to cite the full text of CAA section 110(c)(1), which states that 
EPA ``shall promulgate a [FIP] at any time within 2 years . . . unless 
the State corrects the deficiency, and the Administrator approves the 
plan or plan revision, before the Administrator promulgates such 
[FIP].'' Thus, the plain language of the Act requires SIP approval, not 
merely SIP submission, before EPA's FIP authority is tolled.
    Comment: EPA can disapprove a SIP only where it fails to meet 
minimum CAA requirements. In the case of regional haze, the CAA 
required Minnesota and Michigan to weigh the five statutory factors and 
arrive at reasonable BART technology determinations.
    Minnesota conducted a comprehensive rulemaking process to develop 
its regional haze program, beginning with its analysis of taconite 
sources in 2003. Minnesota began rulemaking efforts shortly after EPA 
promulgated its revised regulations in 2005 and invested ``thousands 
and thousands of hours'' over the next four years collecting and 
analyzing technical data, assessing ground-level operating information 
to make BART determinations that properly weighed the five statutory 
factors from its unique local perspective. After carefully reviewing 
all comments and analyzing all available information, on December 31, 
2009, Minnesota submitted a detailed SIP to EPA that contained a 
determination of the technology that was BART for each taconite furnace 
in the state and for each regional haze pollutant. This SIP was 
supported by more than 1,000 pages of analysis.
    Similarly, Michigan began working to meet its regional haze 
obligations soon after the finalization of EPA's revised regional haze 
regulations and its state guidelines for BART determinations. After 
reviewing all comments and analyzing all available information, on 
November 5, 2010, Michigan submitted a detailed SIP to EPA with 
extensive technical support totaling 1,187 pages that identified BART 
for taconite plants.
    Those submittals demonstrate that both states met their statutory 
regional haze SIP burden, including for BART determinations. EPA must 
give the states' findings the very same deference that EPA so often 
claims it should receive when it holds the primary role in developing a 
substantive standard.
    Response: EPA addressed these very general comments in our February 
6, 2013 supplemental proposed disapproval and addresses these them 
further in our responses to the more specific comments that follow. The 
commenters fail to note that the states' (December 31, 2009 and 
November 5, 2010) SIPs that included thousands of pages lacked, among 
other things, actual NOX emission limits for taconite 
facilities. The States are not entitled to deference in this instance 
because of the numerous gaps and inadequacies in their SIPs, as 
described in the supplemental proposed disapproval and in the responses 
to comments that follow.
    Comment: Minnesota and Michigan properly concluded that low 
NOX burners were not available or technically feasible for 
taconite furnaces at the close of the public comment periods. Pursuant 
to the BART Guidelines, Minnesota and Michigan identified low 
NOX burners as an available control technology at Step 1 of 
the BART analysis. ``Available'' at Step 1 means that the technology 
has a ``practical potential for application to the emissions unit.'' At 
Step 2 of the

[[Page 59828]]

BART analysis, the technologies in Step 1 are evaluated for technical 
feasibility. To be considered technically feasible, technology that has 
not been installed and operated on the source type in question must be 
both ``available'' and ``applicable.'' Availability under Step 2 is 
defined differently than it is under Step 1. Under Step 2, 
``availability'' means commercial availability. A technology is only 
considered commercially available if it is past bench scale and pilot 
testing stages and has reached the licensing and commercial sale 
stages. ``Applicability'' is a technical determination that takes into 
account the technical difficulties that may prevent application of 
available technology to the source in question, such as size and space 
constraints, reliability, and operating problems. The ability to secure 
vendor guarantees is also relevant to the applicability determination.
    EPA attempts to argue that the general existence of low 
NOX burner technology in other, dissimilar applications 
means they are technically feasible for the combustion zones of 
taconite furnaces under Step 2 of the BART analysis. On the contrary, 
the BART Guidelines identify the close of a state's public comment 
period as the cut-off point after which newly ``available'' 
technologies need not be considered by the states. The public comment 
period for Minnesota's BART technology determinations closed on May 16, 
2008, and its supplemental comment period on other aspects of the 
proposed SIP ended September 3, 2009. Michigan's public comment period 
on its non-EGU BART technology determinations closed on June 23, 2010. 
Thus, based on Minnesota and Michigan's reasoned decisions to follow 
the BART Guidelines, only technologies that were ``available'' on May 
16, 2008, and June 23, 2010 (respectively) could be considered.
    EPA also implies that the States did not evaluate low 
NOX burners at all in their BART determinations when in fact 
both States did identify low NOX burners as ``available'' at 
Step 1 for every taconite facility. Minnesota and Michigan acknowledged 
that low NOX burners had been used in other applications 
such as boilers, but properly determined that low NOX 
burners were ``available'' under Step 2 only for the preheat sections 
of the furnaces. None of Cliffs' facilities operate preheat burners, so 
low NOX burners were not technically feasible for any of its 
indurating furnaces.
    Furthermore, EPA's statement that Minnesota possessed information 
suggesting that low NOX burners were ``likely to be a 
successful technology'' for the main burners of taconite furnaces in 
2009 is wholly without support. The record demonstrates that Minnesota 
and Michigan properly determined that low NOX burners were 
not ``available'' for the combustion zones of taconite furnaces by the 
end of the public comment periods. At the time Minnesota was developing 
its SIP, low NOX burners had never been installed in an 
application comparable to a taconite main burner. The Minntac studies 
EPA cites to in support of its claim of commercial availability only 
further support the States' positions. Minntac did not even begin pilot 
testing a new low NOX burner for its grate-kiln furnaces 
until May 2010. That leaves no doubt that low NOX burners in 
the combustion zone were unavailable at the time Minnesota was making 
BART technology determinations because its public comment period closed 
months before pilot testing even began. That testing began four months 
after the close of Michigan's public comment period and continued 
through 2011. Minntac's status reports from May and December 2011 
further confirm that low NOX burners were still in the 
development stage through 2011. Further, Minntac identified a number of 
problems that required modifications to the initial burner and other 
adjustments. Despite all of these adjustments, Minntac never achieved 
the desired emission rates while combusting coal. The Essar low 
NOX burner studies for straight-grate furnaces were even 
further behind in the testing stages than the Minntac studies at the 
time of both SIP submissions. The \1/4\-scale test facility was not 
built until 2011 and final results were not submitted until August 
2011.
    Finally, the information on low NOX burners discussed 
above was available to EPA at the time it proposed approval of 
Minnesota's regional haze SIP in January 2012. EPA cannot now claim 
that it ``did not have the relevant information'' on low NOX 
burners until after it initially proposed approval of Minnesota's 
regional haze SIP.
    Response: Due to the complexity of Cliffs' lengthy comment and the 
interconnectedness of its constituent arguments, it is being addressed 
by a single response. However, each of the four major points raised by 
Cliffs are specifically identified and addressed accordingly.
    EPA agrees with Cliffs that a technology that is both ``available'' 
and ``applicable'' is technically feasible under Step 2 of the case-by-
case BART analysis required under the BART Guidelines. Cliffs is also 
correct that the term ``available'' has somewhat different meanings 
under Step 1 and Step 2. Under Step 1, ``[a]vailable retrofit control 
options are those air pollution technologies with a practical potential 
for application to the emission unit and the regulated pollutant under 
evaluation.'' EPA interprets this use of the term ``available'' broadly 
to include all potential control options, even those that are cutting-
edge or are not currently in use at the source type in question. Under 
Step 2, ``[a] control technique is considered available . . . if it has 
reached the stage of licensing and commercial availability.'' EPA's 
interpretation of this use of the term ``available'' is slightly less 
broad, and includes only those control options that can be obtained 
through ordinary commercial channels.
    However, EPA strongly disagrees with Cliffs attempts to conflate 
the concept of ``availability'' under Step 2, with the separate concept 
of ``applicability.'' While it is true that control technologies that 
are not ``available'' through ordinary commercial channels by the end 
of a state's public comment period need not be considered as BART, the 
same is not true with regards to the question of ``applicability.'' In 
regards to this latter question, states ``need to exercise technical 
judgment in determining whether a control alternative is applicable to 
the source type under consideration.'' Moreover, ``a commercially 
available control option will be presumed applicable if it has been 
used on the same or a similar source type.''
    In the instant context, low NOX burners are an 
``available'' control technology under Step 2 because they can be 
obtained through ordinary commercial channels. Indeed, Fives North 
American and other low NOX burner manufacturers would 
presumably dispute the notion that their products, which have been on 
the market for decades, are not commercially available as Cliffs 
contends. As a result, Minnesota and Michigan were required to exercise 
their technical judgment as to whether low NOX burners were 
``applicable'' to taconite furnaces. In light of the successful 
installation of low NOX burners at Minntac and Essar, which 
both states were aware of prior to the ends of their respective public 
comment periods, Minnesota and Michigan were further required to 
presume the applicability of low NOX burners for taconite 
furnaces because they were in use not just at a similar source type, 
but at the same source type. Since neither Minnesota nor Michigan 
adequately rebutted this presumption or responded

[[Page 59829]]

to comments, but instead relied primarily on cursory technical 
feasibility analyses performed by the taconite companies and their 
contractors in 2006, the states did not comply with the BART Guidelines 
or reasonably ``take into consideration the technology available'' or 
determine the ``best system of continuous emission reduction.'' See 40 
CFR 51.301 and 51.308(e)(1)(ii)(A).
    Contrary to the commenters' assertions, both states were aware that 
low NOX 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.\1\ In a June 23, 2010, letter to the 
Michigan Department of Natural Resources and Environment (now the MDEQ) 
regarding the state's draft regional haze SIP, EPA commented that ``a 
low-NOX 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 
NOX formation in taconite furnaces.'' \2\ 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 NOX main burners 
that achieve 70 percent NOX reduction on its indurating 
lines.'' \3\ In addition to these comments, both states received 
comments regarding the technical feasibility of low NOX 
burners from the Forest Service as well. Therefore, both Michigan and 
Minnesota were aware that low NOX burners had been 
successfully applied to indurating furnaces, and the commenters' 
arguments that the results of these studies somehow constitute ``new'' 
information are without merit.
---------------------------------------------------------------------------

    \1\ 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, 2012, but EPA was granted an extension to submit 
comments. EPA's comments were submitted on February 10, 2012, and 
were received and considered by MPCA.
    \2\ 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).
    \3\ See MN Haze plan, EPA 2-10-12 comments to MPCA in MN May 
8,2012, Suppl. Regional Haze SIP submittal (Docket  EPA-
R05-OAR-2010-0037-0028).
---------------------------------------------------------------------------

    Finally, even if information regarding the technical feasibility of 
installing low NOX 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.
    EPA also disagrees with the commenters' assertions that low 
NOX burners are only commercially available for the preheat 
sections of indurating furnaces. The commenters statement that 
``Minnesota and Michigan acknowledged that low NOX burners 
had been used in other applications such as boilers, but properly 
determined that low NOX burner technology was `available' 
only in the preheat sections of the furnaces'' again confuses the 
concepts of availability and applicability. A control technology cannot 
be commercially available for one application, such as the preheat 
sections of the indurating furnaces, but not commercially available for 
another. Rather, the question is whether the commercially available 
control technology can be applied to the different situations.
    In regards to the installations at Minntac and Essar, a January 30, 
2009, report prepared by Hatch for U.S. Steel strongly recommended that 
U.S. Steel pursue all available technology and potential options 
pertaining to reducing the amount of NOX emissions generated 
by the rotary kiln at the Minntac facility, including the use of a low 
NOX burner. The feasibility of low NOX burners on 
straight-grate kilns is documented in a September 19, 2011 summary of 
findings presented to the Minnesota Pollution Control Board by Fives 
North American Combustion, Inc. (Fives) for Essar. Also, reports on the 
success of U.S. Steel's efforts to use low NOX burners were 
submitted to Minnesota in 2010 and 2011, well before the close of the 
State's comment period on its supplemental regional haze SIP in 
February 2012. These reports, coupled with the comments both Minnesota 
and Michigan received regarding the applicability of low NOX 
burners to taconite facilities, put the States on notice that the 
cursory technical infeasibility determinations in their regional haze 
SIPs were not only inadequate, but inconsistent with a documented 
installation.
    Finally, EPA acknowledges that it was aware that low NOX 
burners were being installed at U.S. Steel's Minntac facility in 2010, 
two years before EPA initially proposed conditional approval of 
Minnesota's BART determinations for taconite facilities. However, EPA 
only became aware of the U.S. Steel test reports from Minntac 
confirming the successful implementation of low NOX burners 
from comments received in response to the January 2012 proposed 
rulemaking. Moreover, commenters urged EPA to take a harder look at the 
technical feasibility of low NOX burners and the adequacy of 
Minnesota's BART determinations for taconite facilities. EPA considered 
the comments and performed additional analysis, which is exactly the 
purpose of the public notice and comment period. Agencies are not 
required to finalize proposed decisions in the face of public comments 
that present compelling evidence that an agency's proposed course of 
action was incorrect.
    Comment: Minnesota and Michigan properly determined that Good 
Combustion Practices (GCP) are BART for taconite furnaces. After 
identifying all technically feasible control options, the states 
performed cost-effectiveness analyses for each furnace and determined 
that no other controls would result in cost-effective NOX 
reductions. GCP will ensure that furnaces are running at their most 
efficient capabilities to complete combustion while consuming as little 
fuel as possible, which will reduce fuel-based NOX emissions 
and minimize thermal NOX by producing only the heat needed 
to make quality pellets. The states performed a proper BART analysis, 
weighing the five statutory factors to arrive at this control option, 
and EPA has no grounds for questioning that judgment.
    EPA cannot credibly attack the legitimacy and enforceability of 
GCP, as EPA itself already requires taconite furnaces to employ GCP as 
part of the Taconite MACT, which requires all sources to ``identify and 
implement a set of site-specific GCP for each type of indurating 
furnace'' that ``correspond to . . . standard operating procedures for 
maintaining the proper and efficient combustion within each indurating 
furnace.'' GCP includes maintaining minimum combustion temperatures and 
maximum CO concentrations in the furnace exhaust gases, and ensuring 
proper burner alignment and fuel-air distribution and mixing. GCP also 
requires routine inspections, preventative maintenance, and performance 
analyses. The requirement to employ and demonstrate compliance

[[Page 59830]]

with GCP is a federally enforceable requirement that has been 
incorporated by reference into the Title V permits for each facility. 
The operation and maintenance plans containing these GCP requirements 
were submitted to the state for each facility to ensure that they 
satisfied the GCP requirements set forth by EPA.
    All of Cliffs' facilities were required to employ GCP as part of 
their Taconite MACT compliance obligations by October 2006, and were 
necessarily employing GCP when they were later required by 
Administrative Order (AO) to conduct NOX testing to 
establish numeric NOX BART emission limits. There is no 
merit to EPA's contention, therefore, that sources failed to use GCP 
while testing under a ``worst-case'' scenario to establish 
NOX BART emission limits. BART limits apply at all times, 
and therefore it is important to establish a limit that sources can 
meet under all operating conditions. As such, the state AOs required 
extended testing to gather over 150 data points that reflected GCP 
under a full range of normal operating conditions. The GCP-based 
NOX limits act as further assurance that sources will 
continue to employ GCP to remain in compliance.
    EPA may not make an about-face on its approval of GCP and the 
emission limits reflecting these controls when nothing has changed 
since its proposed approval. EPA has no basis for changing its position 
and claiming that the new federally enforceable practices and emission 
limits it had already found acceptable are no longer satisfactory. 
Contrary to EPA's claims, the amended state SIPs continue to require 
GCP along with process modifications, and continue to contain emission 
limits (or plans to develop emission limits) based on those controls.
    Even if EPA could demonstrate that additional NOX 
reduction technologies were available during the states' assessment of 
BART for taconite furnaces, NOX BART demonstrations will not 
materially change because availability is just one of the criteria for 
a BART determination. Low NOX burner technologies also fail 
as BART because they will not produce any discernible visibility 
improvement.
    Response: EPA does not agree that Minnesota and Michigan properly 
determined that GCP is BART for taconite furnaces and that it 
identified all technically feasible control options. In its one-size-
fits-all approach to establishing BART, with an essentially identical 
analysis for each taconite facility, Minnesota dismissed low 
NOX burners in the indurating section of the furnace based 
on speculation that they would adversely affect pellet quality. 
However, not only was this position unsupported by corroborating data, 
but U.S. Steel has demonstrated the technical feasibility of low 
NOX burners and documented that they do not adversely affect 
pellet quality.
    EPA also disagrees that GCP underwent a five factor analysis as 
required by the Act and the Regional Haze Rule. In appendix 9.3 of its 
2009 BART Determinations, MPCA states: ``However, the MPCA believes 
that neither ArcelorMittal nor the MPCA has sufficient operating 
parameter data or emissions data to be able to assess whether current 
combustion practices constitute `good' combustion practices. . . .'' 
MPCA basically established an undefined concept, with no specified 
emission reduction potential, as BART. Without identification of a 
specific and quantifiable control requirement, there is no basis for a 
five factor analysis. There is similar language for Minnesota's other 
taconite facilities. Michigan also identified an unspecified GCP, 
without a NOX emission limit, as BART. In addition, 
Minnesota's 2012 regional haze SIP supplement failed to provide any 
indication of what GCP is and what effect it has on emissions.
    The commenters' support of GCP lacks merit for several reasons, 
especially because GCP is not defined by Minnesota or Michigan. Neither 
State's regional haze SIP contained an assessment of combustion 
practices, an analysis of operating parameters in relation to 
emissions, or a definition of operating practices that constitute GCP. 
Such an assessment would be needed to establish GCP, and the lack 
thereof further diminishes GCP as being a meaningful control measure 
for the taconite furnaces. In addition, GCP is not typically considered 
to be a NOX reduction technique. As a relevant example, the 
January 30, 2009, NOX Reduction Analysis performed by Hatch 
for U.S. Steel's Minntac facility fails to list GCP as a potential 
NOX reduction technology for an indurating furnace. As 
another example, the 2008 BACT analysis for JEA--Greenland Energy 
Center Units 1 and 2 also fails to list GCP as a potential 
NOX control. This analysis goes on to state that measures 
taken to minimize the formation of NOX during combustion 
inhibit complete combustion, which increases the emissions of carbon 
monoxide (CO). In other words, GCP, which seeks to promote complete 
combustion rather than inhibit it, would tend to increase 
NOX emissions. The ``September 2010 We Energies Biomass 
Energy Project Revised Control Technology Review for Carbon Monoxide 
Emissions for the Biomass-Fired Boiler'' also discusses the inverse 
relation between NOX emissions and CO emissions, indicating 
how improving combustion efficiency can increase NOX 
emissions. In conclusion, the basic principles of combustion do not 
vary according to the nature of the burner application and GCP is not 
an accepted approach to reduce NOX emissions.
    The commenters state that GCP is already required under other 
Federal regulations, including the Taconite MACT rule. However, GCP for 
the MACT is not the same as GCP for NOX. GCP for the MACT is 
to control products of incomplete combustion (PIC). To minimize PIC, 
the operating conditions targeted are generally the opposite of those 
that would be targeted for reducing NOX. The Taconite MACT 
explains at 68 FR 61883: ``The basic method used in reducing 
NOX emissions is a reduction in combustion temperature, 
which is the opposite strategy needed for minimizing PIC (i.e. 
increasing combustion temperature).'' Therefore, the operation and 
maintenance plans referred to by the commenters to ensure that they 
satisfied the GCP requirements in the Taconite MACT would therefore 
specify conditions that would increase NOX emissions, not 
reduce them.
    Finally, the commenters' statement that ``[l]ow NOX 
burner technologies also fail as BART because they will not produce any 
discernible visibility improvement'' is not germane to this rulemaking. 
Minnesota and Michigan's regional haze SIPs did not assess the 
visibility improvement associated with low NOX burners, or 
conduct a five factor analysis at all, because the States improperly 
rejected the technology as technically infeasible. To the extent that 
Cliffs is attempting to comment once again on EPA's visibility analysis 
for low NOX burners that was conducted as part of the FIP, 
that rulemaking has been finalized.
    Comment: Michigan modeling adequately demonstrates that 
SO2 emissions from Tilden do not cause visibility 
impairment. Michigan conducted source-specific modeling using CALPUFF 
to justify its conclusion that SO2 emissions from Tilden 
Mining do not cause visibility impairment in Class I areas. When 
Cliffs' consultant conducted the proper CAMx modeling for Tilden, the 
results fully supported Michigan's conclusion that SO2 
emissions do not cause visibility impairment. On the 98th percentile 
most impacted day, the visibility improvement at Isle Royale, when 
isolating the sulfate impact, was just 0.14 deciviews. Cliffs' updated 
visibility

[[Page 59831]]

modeling fully supports Michigan's determination that SO2 
reductions at Tilden Line 1 will not create sufficient visibility 
improvement to justify the expense of controls.
    Response: This comment is of limited relevance because the issue 
raised by this comment was not a basis for disapproval. However, 
Cliffs' own modeling shows a combined impact of 0.24 deciviews at Isle 
Royale, Voyageurs, and Boundary Waters. In the discussion of the 
modeling results, Cliffs' report provides visibility impact thresholds 
to provide context for these results. The report states that a 0.10 
deciview difference was defined by other states, such as the 
northeastern states' MANE-VU Regional Planning Organization, as the 
degree of visibility improvement below which additional controls would 
not be justified. Under such a threshold, even the 0.14 deciview 
improvement Cliffs' modeled for Isle Royale (if proven to be accurate) 
would be sufficient to require cost-effective controls.
    Comment: CEMS are not required by the CAA, EPA's regulations 
implementing the Regional Haze program, or the BART Guidelines. The 
states instead have flexibility to choose an appropriate compliance 
demonstration method as long as it is sufficient to show compliance or 
noncompliance, contains a reasonable averaging period consistent with 
established reference methods, and provides adequate recordkeeping and 
reporting for the agency to confirm the source's compliance status. 
Also, consistent with the monitoring flexibility authorized under the 
BART Guidelines, MPCA accepted CEMS data or a ``comparable method of 
emission estimation'' from each BART-affected source for purposes of 
establishing BART emission limits.
    Response: This comment fails to acknowledge or address the primary 
concern identified in EPA's supplemental proposed disapproval, namely 
that absent a CEMS requirement, EPA did not find the emission limits in 
Minnesota's regional haze SIP to be enforceable. Even with a 30-day 
stack testing option in the SIP, EPA believes (as discussed in the 
supplemental proposed disapproval) that the results from this method 
could be challenged at any time as not representative. Minnesota's 
regional haze SIP lacked clarity as to the method to be used to 
determine compliance, while Michigan had no relevant emission limits 
whatsoever addressing BART requirements.
    Comment: EPA does not require CEMS in many of its rules, implicitly 
acknowledging that CEMS are not necessary to demonstrate continuous 
compliance with emission limits. EPA has adopted parametric monitoring 
systems for other regulatory requirements that are designed to ensure 
compliance with health-based emission limits, including the Taconite 
MACT.
    Response: EPA does acknowledge that not all regulations published 
in the Code of Federal Regulations require a CEMS. However, EPA 
stresses that SIPs are approvable only if the emission limits contained 
therein are enforceable, which requires some method(s) to demonstrate 
compliance. EPA continues to believe that Minnesota failed to require 
appropriate methods to demonstrate compliance, while Michigan's SIP 
contained no NOX BART limit at all. Minnesota's limits are 
expressed as 30-day rolling averages and CEMS are needed to determine 
compliance with a 30-day rolling average on a continuing basis, but 
Minnesota in many cases does not require CEMS to provide data for 
evaluating compliance. In the absence of CEMS, Minnesota requires 
``stack testing . . . for 30 hourly data points.'' Even if the average 
of the 30 data points exceeds the emission limit, the data can be 
contested as not necessarily representative of the 720 hours that are 
in a 30-day average. Minnesota has not addressed whether 720 
consecutive hours of stack testing is even practicable, though none of 
the data used to develop emission limits appears to have been collected 
in this manner.
    Comment: Minnesota and Michigan were aware of numerous operating 
difficulties that have been experienced with CEMS usage at taconite 
furnaces. Unlike EPA, the states understood that installing CEMS on a 
taconite furnace is significantly more complex than installing CEMS on 
a boiler. United Taconite found it necessary to seek multiple approvals 
from MPCA to extend its CEMS certification deadline due to CEMS 
maintenance difficulties, and U.S. Steel's Minntac facility, which uses 
CEMS for NOX monitoring, experienced similar problems with 
its original CEMS installation for lines 6 and 7.
    Response: Comments regarding the difficulties of operating CEMS are 
not germane to the question of whether the limits adopted by Minnesota 
can be properly enforced without them. In any case, the initial 
problems faced by U.S. Steel's Minntac facility were resolved shortly 
after they occurred and have not reoccurred since then. EPA continues 
to believe that Minnesota's SIP failed to require appropriate methods 
for assessing compliance with its taconite plant emission limits.
    Comment: Minnesota and Michigan sensibly concluded that requiring 
CEMS would add an unnecessary additional cost to their BART 
determinations that was not warranted. This conclusion is further 
supported by updated cost analyses for CEMS at Cliffs' taconite 
furnaces, which indicate capital costs of $1 million to $1.4 million 
per furnace for CEMS installation, plus hundreds of thousands of 
dollars in additional operating costs for each emission unit.
    Response: Again, this comment appears to be an untimely comment on 
EPA's FIP and is not germane to EPA's proposed finding that Minnesota 
did not provide suitable methods for enforcing its emission limits. 
Cliffs provides a table in attachment B to its comments that lists 
installed capital costs of CEMS. As one example, the installed capital 
cost for Hibbing Line 1 is listed at $1.2 million dollars. The table 
also lists additional costs associated with this line's CEMS, 
including: (1) Annual labor at $311,250 and (2) parts and equipment at 
$97,600 per year, estimating the net present value (NPV) of installed 
CEMS to be $4,430,922. EPA used the number of CEMS that Cliffs 
specified it would need for Hibbing Line 1, as well as some additional 
costs (like scaffolding and platforms for brand new installations), in 
its standard CEMS cost spreadsheet (available at: http://cfpub.epa.gov/oarweb/mkb/contechnique.cfm?ControlID=26). Under this spreadsheet, EPA 
was unable to come close to duplicating the costs claimed by Cliffs. 
EPA is confident that the installation of CEMS at Cliffs' facilities 
will be substantially less than $1.0 to $1.4 million dollars/furnace. 
Even under very extreme circumstances, costs for multiple CEMS would 
barely reach one-third of the costs claimed by Cliffs.
    Comment: Minnesota made all pertinent aspects of the BART 
determinations enforceable through Administrative Orders (AOs). First, 
EPA questions Minnesota's decision to express the SO2 limits 
as lbs SO2/long ton of pellets produced for Northshore and 
Hibbing, claiming that ``pellet production is not routinely measured'' 
and that the AOs do not require recordkeeping of pellet production. 
However, pellet production must be routinely measured for business 
purposes, as finished pellets make up the entire sales business of each 
plant. Production tonnage is measured and cross-checked by a series of 
calibrated conveyer belt scales on a continuous basis.

[[Page 59832]]

    Second, EPA questions the enforceability of NOX emission 
limits for Hibbing because the AO for that facility provides Hibbing an 
opportunity to demonstrate the NOX limits in the AO are not 
feasible. Enforceable NOX limits apply to Hibbing at all 
times.
    Finally, EPA's concerns over the enforceability of the CEMS 
requirement for Hibbing are similarly groundless. The AO requires 
Hibbing to submit a plan to install a CEMS on Line 2 within 60 days of 
the effective date of the AO, and installation and certification of the 
CEMS no later than one year from the due date of the plan.
    Response: EPA is not suggesting that Cliffs fails to measure pellet 
production sufficiently for sales and other business purposes. However, 
there is no indication that Cliffs measures the quantity of finished 
pellets produced each day on each line. Such daily measurements on each 
line would be necessary to establish compliance with a limit measured 
in lbs SO2/long ton of pellets, on a 30-day rolling average, 
as specified in the AOs. Although the AOs contain a general requirement 
for retaining records of operational parameters related to emissions, 
there is no explicit requirement for maintaining daily records of the 
finished production from each line. Such records would be necessary for 
determining compliance with the lbs SO2/long ton limits.
    With respect to the commenters' second point, EPA agrees that 
Minnesota set a NOX limit that will apply to Hibbing's line 
2. However, there are no specific criteria in the Minnesota SIP or the 
AO for Hibbing to ensure that an alternative limit, were it to be 
established, would be set in such a manner so as to satisfy BART. Also, 
the AO did not provide that the alternative limit had to be in the form 
of a SIP revision so as to be Federally enforceable as required by the 
Act.
    Finally, EPA understands that the AO requires Hibbing to submit a 
plan to install a CEMS on Line 2, and possibly on Lines 1 and 3. 
Although there is an explicit requirement for a plan that provides for 
installation of the CEMS, there is not an actual requirement that the 
CEMS be installed.
    Comment: Michigan and Minnesota appropriately determined BART for 
SO2 after conducting a case-by-case evaluation of taconite 
pelletizing furnaces. Minnesota concluded that existing wet 
scrubbers for particulate control used at the Hibbing, Northshore, and 
United Taconite Line 1 furnaces would constitute BART when operated to 
also control SO2 emissions. Minnesota's BART determination 
for United Taconite Line 2 was complicated by an intervening expansion 
project that relied on fuel blending to reduce emissions. Minnesota set 
the numeric BART limits for United Taconite at a level more stringent 
than the level the wet scrubbers alone were expected to consistently 
achieve. Cliffs retained the flexibility under the SIP to meet this 
SO2 limit by installing a polishing scrubber or by adjusting 
the sulfur content in its fuel blend.
    EPA claims that Minnesota's BART determination for United Taconite 
is not approvable because it did not reconsider the cost-effectiveness 
of flue gas desulfurization (FGD) after the expansion project. However, 
EPA cannot reject the SIP on this basis because EPA's own BART 
determination, in the final taconite FIP, also concluded that FGD is 
not cost-effective for United Taconite. Given United Taconite's ability 
to blend existing fuels, United Taconite's anticipated actual baseline 
SO2 emission rate will be low enough to render a reduction 
from FGD not cost-effective. EPA reached the same conclusion that 
Minnesota reached in its SIP that BART for SO2 at United 
taconite was an emission limit, not a control device. The FIP emission 
limit reflects a significant SO2 reduction that can be 
accomplished through fuel blending or polishing controls, which is 
precisely the BART technology determination that EPA claims to object 
to in the SIP. The 0.6-percent sulfur content that EPA adds in the FIP 
does not reduce emissions and does nothing to advance regional haze 
goals. EPA's objection to Michigan's SO2 BART determination 
for Tilden Line 1 must fail on the same basis. Tilden has also 
indicated that it will adjust fuels to ensure that baseline 
SO2 emissions cannot justify FGD as a cost-effective 
control. Therefore, Michigan appropriately set an SO2 limit 
for Tilden that did not reflect expensive add-on controls.
    Response: EPA's FIP did not require FGDs because in November, 2012, 
EPA agreed that FGDs would not be necessary at United Taconite and 
Tilden because Cliffs stated an intention at that time to switch to 
lower sulfur fuels that would result in lower SO2 emissions. 
However, Minnesota and Michigan's BART analyses were based upon the use 
of high sulfur fuels.
    Therefore, this comment is largely misdirected because it is based 
upon EPA's FIP and not on the adequacy of Minnesota and Michigan's BART 
determinations. The commenter's assertion that EPA reached the same 
conclusion that Minnesota reached in its SIP is irrelevant because EPA 
made its determination based upon United Taconite's anticipated use of 
low sulfur fuels (with much lower SO2 emissions) than the 
high sulfur coal in use by United Taconite currently, and upon which 
Minnesota's determination of BART was based.\4\ EPA agreed that FGDs 
are not BART at this anticipated lower emission rate, but does not 
agree that FGDs are not BART when United Taconite is burning high 
sulfur coal. The commenter goes on to object to the 0.6-percent sulfur 
content limit in the FIP. This sulfur content restriction is also not 
relevant to whether or not Minnesota's SIP is approvable because it was 
neither suggested as a control option by the commenter at the time of 
Minnesota's rulemaking, nor considered by Minnesota.
---------------------------------------------------------------------------

    \4\ As indicated in an 11-29-12 email exchange, Cliffs agreed to 
use lower sulfur fuels with the result of reducing its 
SO2 emissions by half.
---------------------------------------------------------------------------

    EPA therefore maintains its position that Minnesota improperly 
rejected the use of FGD as a cost-effective technology for reducing 
SO2 emissions from United Taconite's two lines. Also, as 
discussed in the August 15, 2012, proposed action, EPA believes that 
flue gas scrubbing, particularly in combination with proper fuel 
blending, is considerably more cost-effective than the cost-
effectiveness estimates in Minnesota's regional haze SIP. Furthermore, 
subsequent to Minnesota's initial BART analysis, United Taconite 
switched to using high sulfur fuels on both of its lines, thus making 
FGD a more appropriate control measure to be considered.
    Similarly for Michigan, EPA agreed that if Tilden switches to 100-
percent natural gas,\5\ the use of an FGD would not be cost-effective. 
However, Michigan's BART determination was based upon the use of high 
sulfur coal. As indicated in EPA's August 15, 2012 proposed action, EPA 
determined that an FGD would be more cost-effective than indicated by 
the Michigan regional haze SIP, and an FGD was therefore proposed as 
BART.
---------------------------------------------------------------------------

    \5\ As indicated in an 11-29-12 email exchange, Cliffs agreed to 
switch to natural gas, thereby reducing its SO2 emissions 
by at least 80 percent.
---------------------------------------------------------------------------

    Comment: MPCA's statistical analysis establishing SO2 
limits for the Hibbing facility is correct because the data is normally 
distributed. Barr Engineering provided an analysis showing that the 
data is normally distributed. The approach Minnesota used to establish 
emission limits for each facility was well within the discretion 
afforded to states to identify BART emission limits.
    Response: Non-parametric SO2 emissions data appear to be 
typical across the industry. EPA agrees,

[[Page 59833]]

however, that the available data for the majority of Hibbing's lines 
appear to be normally distributed.\6\ However, as discussed in greater 
detail in the supplemental proposal, MPCA did not correctly apply the 
upper predictive limit (UPL) equation for normally distributed data. 
See 78 FR 8482-8483. If that equation were applied properly, the 
resulting limit for Hibbing would be significantly lower than the limit 
set by MPCA. In addition, the limits set by MPCA were expressed in 
terms of lbs SO2/long ton of pellets produced. As discussed 
in the supplemental proposal, pellet production is not routinely 
measured at the end of an indurating furnace. Further, the AOs do not 
specify methods for determining pellet production by indurating furnace 
and do not specify any requirement to keep records of pellet 
production.
---------------------------------------------------------------------------

    \6\ It should be noted that the Barr Engineering analysis 
combined the data from the different lines. Because even lines of 
the same size can function differently, it would be more appropriate 
to consider the lines separately.
---------------------------------------------------------------------------

    Comment: The SO2 emission limits set by EPA in the FIP 
are identical to the emission limits set by Minnesota for Northshore 
and similar to the emission limits set for Hibbing. Therefore, even if 
EPA has legitimate technical corrections, the resulting changes are not 
substantive and should not result in wholesale rejection of the 
Minnesota regional haze SIP. These issues should have been resolved in 
discussions with Minnesota before EPA issued a FIP.
    Response: EPA disagrees. The SO2 emission limits set by 
Minnesota were expressed in terms of lbs SO2/long ton of 
pellets produced, while the limits set by EPA are expressed in terms of 
lbs SO2/hour. No demonstration has been made that the limits 
are equivalent. Furthermore, the emission limit set by EPA in the FIP 
for Northshore is temporary and must be recalculated after CEMS data 
has been collected.
    Comment: Minnesota and Michigan were not required to reopen the 
BART technology determinations to accommodate EPA's unreasonably 
lengthy SIP review, and EPA may not reject the SIPs on this basis. EPA 
must approve SIPs that satisfy all applicable regulatory requirements 
pursuant to CAA section 110(k)(3). The public comment periods on 
Minnesota and Michigan's BART determinations closed on May 16, 2008, 
and June 23, 2010, respectively. Only technologies that were 
commercially available under Step 2 of the BART analysis at that time 
could be considered by the states in making BART technology 
determinations. EPA may not disapprove a state SIP because the states 
did not include information in their analyses that was not in existence 
at the time the technology determinations closed.
    While the state's determinations were awaiting EPA action, 
Minnesota was engaged in an extensive process of collecting emissions 
data and performing analyses to set emission limits that reflected 
those technology determinations. The states must be able to rely on 
their BART determinations as they proceed to convert them to emission 
limits.
    Response: As discussed previously, and contrary to the commenters' 
assertions, low NOX burners were ``available'' at the end of 
both states' public comment periods and have been for decades. As such, 
both states were required to determine whether low NOX 
burners were applicable to taconite furnaces, a task which they failed 
to do as neither state adequately considered the installation at U.S. 
Steel's Minntac facility or other available information.
    While immaterial, EPA notes that the commenters provide no support 
for their assertion that the States were entitled to rely on their 
prior BART determinations as they ``proceeded to convert them to 
emission limits.'' On the contrary, the Act explicitly requires that 
all BART determinations be in the form of enforceable emission limits. 
See CAA sections 110(a)(2)(A) and 169A(b)(2). Neither the Act nor the 
BART Guidelines contemplate a scenario in which states are permitted to 
select a technology as BART, walling off that decision from further 
public scrutiny and comment, and then wait several years before setting 
emission limits to complete the BART process, ignoring any new 
information brought to their attention in the process.
    Comment: The information that became available after the close of 
the states' public comment periods only further supports Minnesota and 
Michigan's initial BART determinations. Serious concerns continue to 
exist over the feasibility of low NOX burner technology on a 
case-by-case basis, and current cost and modeling information suggest 
that application of this technology would be both more expensive and 
less impactful from a visibility standpoint than EPA presumes. 
Furthermore, the most current cost information on FGD technology 
confirms that FGD technology is not a cost-effective control option for 
United Taconite or Tilden. Minnesota and Michigan made proper BART 
determinations in 2008 and 2010, and none of the information EPA cites 
provides grounds for changing these determinations.
    Response: EPA published the February 6, 2013, supplemental proposed 
disapproval to provide additional information regarding EPA's views on 
Minnesota and Michigan's regional haze SIPs and to solicit additional 
comment regarding the proposal to disapprove the SIPs for failing to 
require BART at the applicable taconite plants. EPA is not soliciting 
further comment on its FIP as the supplemental proposal only addresses 
whether the states' SIPs should be disapproved for failing to provide 
an adequate analysis and require BART for applicable taconite plants. 
The commenters specifically raise the following points: (1) Low 
NOX burner technology is not technically feasible for 
straight-grate furnaces; (2) low NOX burner technology is 
not technically feasible for grate-kiln furnaces; (3) updated cost 
analyses demonstrate that low NOX burner technology is not 
cost-effective in light of limited visibility improvements; and (4) 
updated cost analyses demonstrate that FGD technology is not cost-
effective for Tilden or United Taconite in light of limited visibility 
improvements. These points are not directly relevant to the disapproval 
of Minnesota and Michigan's regional haze SIPs for taconite plants. 
Indeed, given the conclusory nature of Minnesota and Michigan's SIPs 
regarding the feasibility of low NOX burners at these 
facilities, these comments appear to be seeking to provide alternative 
justifications for the States' BART determinations that the plans 
themselves do not rely upon.
    In any case, the commenters' review of the feasibility of low 
NOX burners reflects an overly narrow view of technical 
feasibility. Any installation of control equipment at a facility that 
does not already have that equipment in place requires engineering to 
determine how best to design the equipment to work most effectively 
given the particular features of the particular facility. The 
commenters appear to be arguing that low NOX burners cannot 
be considered technically feasible because, for example, the 
engineering work done to design low NOX burners for the 
Essar facility cannot be directly applied to other facilities. The 
commenters cite selected design features that differ from facility to 
facility, such as the number of windboxes, but the commenters provide 
no reason for EPA to believe that any of these features pose problems 
that could not be solved by appropriate engineering analysis, just as 
has been done at multiple taconite lines and in countless other high 
temperature processes in numerous other industries.

[[Page 59834]]

Comments regarding costs and visibility benefits do not speak to 
whether Minnesota or Michigan appropriately analyzed these costs and 
benefits, and comments regarding time for installation appear to be 
untimely comments on the FIP that are not relevant to this rulemaking 
or the approvability of the States' regional haze SIPs. These issues 
are more appropriately discussed in the FIP, which includes a full and 
appropriate analysis of BART.

B. Comment by MDEQ and Cliffs

    Comment: In spite of the provision in the BART Guidelines that 
states: ``[A]ll technologies should be considered, if available before 
the close of the State's comment period. You need not consider 
technologies that became available after this date,'' MDEQ felt that 
additional review was warranted and included in the SIP a requirement 
that Tilden must conduct further testing to provide the basis for 
NOX emission limits to be incorporated in the air permit for 
the Tilden facility. Tilden has since complied and its Permit to 
Install No. 148-12 contains an enforceable emission limit.
    Response: Michigan has not submitted this permit as part of its 
regional haze SIP. To be approvable, emission limits representing BART 
must be contained in the SIP itself to guarantee Federal 
enforceability. Indeed, the Regional Haze Rule specifically states: 
``The State must submit an implementation plan containing emission 
limitations representing BART.'' 40 CFR 51.308(e) (emphasis added). The 
requirement that BART emission limits be contained in the SIP is 
important because states can unilaterally remove or alter permit limits 
(that are not otherwise contained in an approved SIP or Federal 
standard) without federal approval. Moreover, permits expire. 
Consequently, the existence of a limit that Michigan has not submitted 
as part of its regional haze SIP cannot be considered to remedy 
deficiencies in the SIP.
    EPA additionally notes that the NOX limit in Tilden's 
permit is 2,270 lbs/hr on a 30-day rolling average. According to Table 
3-1 in appendix 9H of Michigan's regional haze SIP, the maximum 24-hour 
actual NOX emissions for Tilden were 26,208 lbs/day. 
Dividing by 24 results in an emission rate of 1,092 lbs/hr, which is 
less than half of the permitted limit. Thus, even if EPA were to agree 
that GCP was BART for this facility, Michigan has not demonstrated that 
a limit more than twice the facility's maximum actual emissions could 
possibly represent GCP.

C. Comments by MDEQ

    Comment: The Michigan regional haze SIP provided extensive 
documentation of a full and appropriate analysis of BART that meets the 
requirements of the CAA.
    Response: As discussed in the supplemental proposal, the Michigan 
regional haze SIP did not meet CAA requirements because of its failure 
to require emission controls that represent BART. The Michigan SIP 
defines BART as GCP, but there is no explanation of what GCP is and no 
NOX limits representing GCP. Michigan relies on a state 
permit that has not been submitted to EPA in the form of a SIP revision 
to argue that there is a limit on Tilden's NOX emissions. 
For the reasons explained in the prior response, this limit is 
insufficient to satisfy the requirements of the CAA and the Regional 
Haze Rule.
    Comment: There was no information prior to the close of Michigan's 
public comment period on June 23, 2010, indicating that low 
NOX burners had been successfully utilized on indurating 
furnaces operating under the same circumstances as Tilden's grate-kiln 
furnace.
    Response: In a June 23, 2010, letter to Michigan regarding its 
regional haze SIP, EPA stated that ``a low NOX main firing 
burner firing solid fuels'' had been installed at Minntac's grate-kiln 
furnace,'' (both Minntac and Tilden have grate-kiln furnaces) and that 
``work done by other companies had demonstrated that burner designs 
that lower flame temperature can reduce NOX formation in 
taconite furnaces.'' Even after being notified of the above, Michigan 
did not conduct an adequate BART review of this technology in its 
regional haze SIP. MDEQ's only response to EPA's comment was that 
``[t]he projects and technologies described above were not selected or 
proposed for BART, but rather were proposed primarily as projects to be 
studied, and were agreed to be completed as part of PSD permitting (not 
BART) or enforcement situations. None of the technologies and projects 
described above have been established for BART at taconite plants in 
Minnesota.'' Such circular logic, that low NOX burners 
should not be evaluated for BART because the States had not selected 
them as BART, is arbitrary and inadequate.
    Comment: The FIP schedule of compliance does not allocate 
sufficient time for the permitting process.
    Response: While not germane to this rulemaking, EPA notes that MDEQ 
has provided no information explaining why Tilden needs more than 26 
months to accommodate permitting requirements.
    Comment: MDEQ disagrees with the contents of EPA's July 2, 2012, 
conversation between EPA, U.S. Steel, and COEN, which is described in 
the August 15, 2012, proposed action. It states: ``There is also no 
increase in combustion related emissions, such as carbon monoxide or 
volatile organic compounds, and there is no reason for SO2 
emissions to increase through use of a low NOX burner.'' 
This conversation failed to recognize that the December 22, 2008 
permitting action for the installation and operation of the auxiliary 
burners and conversion to low NOX burners included a net 
emission increase of 1,607 tons/year of carbon monoxide.
    Response: The 2008 permitting action included an emission estimate 
only. Newer information from November 2009 letters from COEN and Hatch 
document why no increase in CO is expected due to the installation of 
low NOX burners. These assertions are further supported by 
U.S. Steel's draft permit for Lines 4 and 5 at Minntac, which was put 
out for pre-public notice review on May 14, 2013. This permit shows 
only a minimal increase in CO emissions according to PSD calculation 
methodology.

D. Comments by the National Mining Association

    Comment: EPA's proposed disapproval of Minnesota and Michigan's 
regional haze SIPs is contrary to the federalism principles embedded in 
the CAA. Congress purposely created a cooperative federalism scheme in 
the CAA to define the roles of EPA and the states under the regional 
haze program. As the DC Circuit Court of Appeals recently noted in EME 
Homer City Generation, L.P. v. EPA, 696 F.3d 7, 12 (D.D.C. 2012), 
``Under the Clean Air Act, the Federal Government sets air quality 
standards, but States retain the primary responsibility (if the States 
want it) for choosing how to attain those standards within their 
borders. The Act thus leaves it to the individual States to determine, 
in the first instance, the particular restrictions that will be imposed 
on particular emitters within their borders. (If a State refuses to 
participate, the Federal Government regulates the sources directly.)'' 
Procedurally, this requires states to submit SIPs that address regional 
haze and establish BART determinations for sources within their 
borders. The states

[[Page 59835]]

then submit these plans to EPA, who must approve them if they satisfy 
all regulatory requirements. See CAA section 110(k).
    In this instance, both Michigan and Minnesota submitted well-
reasoned SIPs detailing their plans for addressing regional haze 
impacts. The entire taconite industry in the United States resides in 
these two states, and as a result, Minnesota and Michigan have an 
extensive and unparalleled understanding of the taconite mining and 
processing industry. Minnesota and Michigan used this industry 
knowledge and years of work and technical analysis to arrive at BART 
determinations for each BART-eligible taconite furnace. In doing so, 
Minnesota and Michigan carefully considered all available information 
provided through the end of their public comment periods and used this 
information to analyze available control technologies and the 
feasibility of installing these control technologies on each taconite 
furnace. The states then evaluated this information in light of the 
five statutory factors set forth by the CAA.
    Minnesota and Michigan performed these evaluations on a case-by-
case basis, in compliance with the CAA and in accordance with EPA 
guidance, and sought public comment on their determinations. The states 
carefully considered these comments before finalizing their SIPs and 
submitting them to EPA for approval on December 31, 2009 (Minnesota) 
and November 5, 2010 (Michigan).
    EPA is now proposing to disapprove those determinations, not 
because the states erred in their reasoning, but because new 
information was presented in 2012. This information, a report on low-
NOX burner technology used at the Minntac furnace, was not 
available when the states arrived at their BART determinations and it 
had not been peer-reviewed or subject to evaluation by the affected 
stakeholders. EPA's own guidance establishes a cut-off date for state 
technology determinations at the close of the state public comment 
period. The States, therefore, were not required to re-open their BART 
determinations based on this new information, and EPA does not have the 
authority to disapprove a valid and supported SIP based on information 
that was not available to the states at the time of their BART 
determinations.
    EPA claims that it was compelled to consider this new information 
submitted during its 2012 public comment period when deciding whether 
to approve the SIPs, even if the States were not. EPA certainly could 
have considered the new information in many appropriate ways, but it 
chose not to. EPA could have shared the report with the States and then 
deferred to the States' evaluation of the data. Instead, EPA chose to 
ignore the States' assessment that the report was an insufficient 
demonstration that the technology was appropriate for the diverse 
furnace designs in the rest of the industry. EPA could have solicited a 
peer review of the report. Instead, EPA actively ignored the input of 
the furnace design engineers at Metso Engineering, who told the agency 
repeatedly that the burner designs would require 20-50-percent more 
fuel per ton of pellets and could cause pellet quality problems when 
installed on other taconite furnaces. EPA arbitrarily included in the 
record for its proposed FIP only the information that supported low 
NOX burners as BART instead of considering all the 
information made available to the agency and conducting the critical 
technology review that the CAA requires.
    EPA's limited authority under the CAA does not authorize it to 
disapprove a SIP and impose a FIP merely because EPA prefers a 
different BART outcome. EPA's role is not to gather evidence to support 
a predetermined BART decision and actively ignore contrary information. 
When the available information does not provide a clear contrary path, 
EPA must defer to the states' method for weighing the available 
information and to the lawful and appropriate BART decision that arises 
from that method.
    Response: EPA disagrees with NMA's assertion that ``EPA is now 
proposing to disapprove these determinations, not because the states 
erred in their reasoning, but because new information was presented in 
2012.'' In a June 23, 2010, letter to Michigan regarding the state's 
draft regional haze SIP, EPA commented that ``a low NOX 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 NOX formation in 
taconite furnaces.'' Similarly, in a February 10, 2012, letter to 
Minnesota 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 NOX main burners 
that achieve 70 percent NOX reduction on its indurating 
lines.'' Therefore, both states were aware that low NOX 
burners had been successfully applied to indurating furnaces.
    Although NMA states that EPA could have shared the report with the 
States and then deferred to the States' evaluation of the data, 
Minnesota had this information before EPA and made no apparent use of 
it. This information is listed in EPA's February 10, 2012, letter to 
Minnesota, which refers to three reports, from April 13, 2010, to 
December 1, 2011, informing the MPCA of U.S. Steel's success in 
installing low NOX burners on two of its indurating 
furnaces. (EPA also provided a copy of this letter to Michigan.) Even 
if Minnesota did not want to require general use of a proven technology 
on other facilities, there is no conceivable reason why Minnesota 
dismissed low NOX burners as BART at the U.S. Steel Minntac 
lines that were already using them. Although Metso Engineering ``told 
the agency repeatedly that the burner designs would require 20-50% more 
fuel per ton of pellets and could cause pellet quality problems,'' U.S. 
Steel documented that it had neither a fuel penalty nor pellet quality 
problems.
    EPA's action cannot be characterized as disapproving submittals 
that satisfy CAA requirements ``merely because EPA prefers a different 
BART outcome.'' In commenting that EPA must approve state submittals 
that meet minimum CAA requirements, NMA apparently recognizes that EPA 
must disapprove state submittals that fail to meet CAA requirements. By 
dismissing clearly applicable NOX and SO2 
emission control options as infeasible, and by finding a group of 
NOX emission reduction practices (GCP) to be BART without 
defining or conducting the necessary five factor analysis of any 
particular good combustion practice, along with other SIP deficiencies, 
Michigan and Minnesota's submittals fail to satisfy CAA requirements 
regarding BART.
    Finally, in regards to NMA's comment regarding a cut-off date for 
considering new information regarding available technologies, EPA 
provided a thorough response to a similar comment from Cliffs above.
    Comment: EPA may not use a ``sue-and-settle'' approach to 
circumvent CAA requirements and usurp the role of the states. EPA's 
decision to seek comment on proposed deficiencies in the States' 
regional haze SIPs only after finalizing a FIP is contrary to the 
cooperative federalism scheme of the CAA. EPA's well established role 
is to review SIPs, determine whether they meet CAA criteria, and only 
if the state process fails to produce a compliant SIP can EPA issue its 
own FIP. By definition, a FIP may be used only to ``fill all or a 
portion of a gap or otherwise correct all or a portion of an inadequacy 
in a State implementation plan.'' In this instance, EPA has put the

[[Page 59836]]

cart before the horse by finalizing a FIP and then seeking public 
comment on the supposed deficiencies that formed the basis for the FIP 
in the first place. EPA took this strange course of action to meet a 
deadline that the agency agreed to in a consent decree to settle 
litigation brought by the National Parks Conservation Association. EPA 
may not, however, use a self-imposed consent decree deadline to justify 
doing things out of order and in violation of the clear rulemaking 
process set forth in the CAA.
    Prior to issuing its FIP, the only deficiency EPA had identified 
was a failure by Minnesota and Michigan (along with over 30 other 
states) to submit a timely regional haze SIP. Minnesota and Michigan 
rectified this deficiency by submitting their SIPs. These submittals 
triggered EPA's obligations under CAA section 110(k) to review the SIPs 
within one year and work with the states to make any changes necessary 
for federal approval. Instead of meeting its statutory obligation to 
act on the state submissions within one year, EPA entered into a 
consent decree with environmental organizations that set a court-
ordered deadline for action on regional haze. In January 2012, EPA was 
on course to meet that deadline in Minnesota by proposing approval of 
the Minnesota regional haze SIP. However, when EPA decided to change 
course and propose a FIP, EPA had left itself with no time to properly 
identify deficiencies in the SIPs. EPA used the consent decree deadline 
as an excuse to stop working with the states to finalize the SIPs, to 
arbitrarily ignore contrary information, and to deny requests for 
additional time for public comments. EPA's rushed FIP that revealed a 
poor understanding of the taconite industry and significant shortcuts 
in the BART determination process. EPA did not propose disapproval of 
the state SIPs before rushing into a FIP, let alone provide ample 
opportunity for public comment or for the states to rectify any 
perceived deficiencies.
    Had EPA followed proper procedure and discussed this new 
information with the states, EPA would have understood that the new low 
NOX burner trial information would not have changed the 
state BART determinations for taconite furnaces. The taconite industry 
is highly specialized, with each indurating furnace designed to process 
a specific ore type and produce pellets meeting varying specifications. 
An in-depth understanding of each taconite furnace is necessary to 
properly evaluate the applicability of ``new'' technology to these 
sources. Instead of taking the time necessary to understand these 
issues, EPA cited its looming consent decree and rushed through a 
``one-size-fits-all'' FIP that requires all taconite furnaces to expend 
significant resources designing and installing technologies that are 
unproven and could severely impact their ability to manufacture a high 
quality product for global markets. This result would produce severe 
economic consequences for the taconite industry in Minnesota and 
Michigan. EPA may not use a self-imposed consent decree to cut the 
states out of a process Congress intended them to control and inflict 
this type of burden on the taconite industry.
    Response: This comment is generally not relevant to the proposed 
disapproval and is primarily focused on EPA's decision to promulgate a 
FIP, as well as the substance of the FIP. For example, the comments 
objecting to EPA promulgating a FIP before acting on the states' SIPs 
and the comments regarding EPA's alleged ``sue-and-settle'' approach 
appear to be objecting to the timing of EPA's FIP promulgation rather 
than addressing appropriate action on the states' SIPs. NMA believes 
that EPA did not properly consider the ``highly specialized'' nature of 
taconite facilities, but NMA does not identify any particular features 
of any particular facilities that would make more or less control 
feasible. More pertinently, the comment suggests that the one-size-
fits-all nature of the state SIPs (in all cases determining undefined 
``good combustion practices'' to be BART) are an important deficiency.

E. Letter From Congressman Richard M. Nolan

    Congressman Nolan submitted a March 4, 2013, letter to EPA asking 
EPA to approve the Minnesota SIP or amend the FIP to allow the taconite 
facilities sufficient time to comply with NOX BART emission 
limits. The comments submitted by the Congressman focus primarily on 
compliance deadlines and as such are FIP rather than SIP issues. 
Today's action disapproves the Minnesota SIP, however, EPA notes that 
the compliance deadlines in EPA's FIP have already been stayed by the 
Eighth Circuit and EPA is currently reviewing several petitions for 
reconsideration that request additional extensions of the compliance 
deadlines for NOX BART.

F. Comments by the Minnesota Pollution Control Agency

    Comment: The proposed SIP disapproval presents unnecessary 
challenges to the historically strong state/Federal relationship in 
Minnesota. EPA's actions rendered moot a significant investment of time 
and resources by the MPCA.
    Response: EPA also values its strong relationship with MPCA, but 
this concern cannot justify approving a SIP that does not meet CAA 
requirements. EPA appreciates MPCA's efforts and the time it invested 
on the Minnesota regional haze SIP, which EPA approved in full except 
with regard to the BART determinations for taconite facilities. EPA 
encourages MPCA to consider submitting a SIP revision for taconite BART 
that EPA could evaluate for approval and potential replacement of EPA's 
FIP.
    Comment: EPA's BART Guidelines do not require states to re-open 
final BART decisions to consider a technology that becomes available 
after the close of the public comment period on the state's SIP. MPCA 
was clear that its May 2012 regional haze SIP supplement was not re-
opening its 2009 BART determinations for taconite facilities. MPCA had 
valid reasons to exercise this discretion because MPCA knew that the 
pilot project for low NOX burners at Minntac was limited to 
a single type of taconite furnace and that the pilot project clearly 
illustrated that low NOX burners required significant 
testing and operational changes at the two furnaces tested.
    Response: As discussed above in response to a similar comment by 
Cliffs, MPCA is misreading EPA's BART Guidelines. Low NOX 
burners are in wide use across a wide range of industries, many 
involving combustion conditions comparable to those in the taconite 
industry, and this technology has been commercially available since 
long before MPCA even began considering BART for taconite plants. 
Therefore, statements in the BART Guidelines regarding the 
consideration of technologies that become available after the close of 
a state's comment period are not germane here.
    MPCA's comment suggests that MPCA interprets ``available'' to mean 
not just commercially available, but also ``applicable'' at a 
particular facility. This interpretation is inconsistent with the BART 
Guidelines, which clearly differentiate between the concepts of 
``availability'' and ``applicability.'' For a technology like low 
NOX burners that has long been commercially available, the 
BART Guidelines do not provide states with the authority to disregard 
information that indicates that an ``available'' technology has also 
become ``applicable,'' and therefore technically feasible, for use at a 
particular source type.
    Furthermore, MPCA offered multiple comment periods throughout its

[[Page 59837]]

regional haze SIP development process. MPCA wishes to treat its last 
comment period as merely addressing the emission limits for a BART 
technology it had selected previously, without offering the opportunity 
to reconsider whether the selection of that technology was appropriate. 
Thus, even assuming MPCA's interpretation of ``available'' were 
correct, MPCA provides no rationale for interpreting the BART 
Guidelines in such a way so as to foreclose the consideration of 
technologies that become available after the close of one comment 
period, but before the close of another, later comment period. 
Therefore, the BART Guidelines provide no justification for MPCA to 
disregard the information that has come to light over the last several 
years that low NOX burners are ``applicable'' and 
technically feasible for the taconite facilities in the state.
    Installation of a new control technology at a facility commonly 
requires ``significant testing and operational changes.'' Indeed, EPA's 
FIP provided time for exactly this effort. Taken to its logical 
extreme, MPCA's comment suggests that states could not require new 
controls at facilities unless the controls have already been installed 
there. On the contrary, the need for testing and operational changes 
alone cannot justify a finding that controls are technically 
infeasible.
    Comment: EPA should stay implementation of its FIP to resolve 
procedural issues with the SIP actions. EPA's supplemental proposed 
disapproval provides no legal authority, either from the CAA or from 
case law, which allows EPA to adopt a final FIP before EPA formally 
disapproves a state's SIP. EPA's claim that it has a mandate to 
promulgate such a FIP without regard to whether EPA has disapproved the 
State's SIP is unsupported and contrary to case law.
    Response: EPA disagrees. EPA's final rule promulgating the FIP 
clearly explained the Agency's legal authority for taking action. 
Section 110(c) of the CAA requires EPA to promulgate a FIP after 
finding that a state has failed to make a required submission unless 
two events occur before EPA promulgates a FIP: (1) The state corrects 
the deficiency, and (2) EPA approves the state's SIP. We note, however, 
that this comment is moot because the Eighth Circuit has since stayed 
the effective date of EPA's FIP.
    Comment: EPA's June 12, 2012, final action simply stated that EPA 
was deferring action with regard to BART for taconite facilities 
because Minnesota did not select EPA's chosen control technology. By 
publishing the supplemental proposed disapproval on the same day as the 
final FIP, EPA is not providing Minnesota with a meaningful opportunity 
to comment on EPA's proposed action because EPA has already decided on 
its course of action. Under the CAA, EPA's supplemental proposed 
disapproval should afford Minnesota the opportunity to remedy the 
specific issues EPA cites as not adequately meeting the requirements of 
the CAA.
    Response: First, MPCA mischaracterizes EPA's June 12, 2012 final 
action, in which EPA actually stated that it was deferring action to 
evaluate information indicating that BART should be defined as a more 
effective control technology. 77 FR 34801 at 34806. Second, as EPA 
explained in an earlier response to a similar comment from Cliffs, a 
FIP is a gap-filling measure only. See CAA section 302(y). As a result, 
a FIP promulgated by EPA remains in place only until a state submits a 
SIP correcting the inadequacy and that SIP is approved by EPA. In this 
instance, MPCA and other commenters had a meaningful opportunity to 
comment on EPA's supplemental proposed disapproval and explain why 
EPA's proposed action was incorrect. Had EPA agreed with MPCA or other 
commenters and approved the Minnesota and Michigan BART determinations 
for taconite facilities instead, EPA's FIP would have been replaced. 
Therefore, MPCA's assertions are incorrect. Furthermore, we again 
encourage MPCA to consider submitting a SIP revision for taconite BART 
that EPA could evaluate for approval and potential replacement of EPA's 
FIP.
    Comment: EPA's August 15, 2012, proposed action did not propose to 
disapprove Minnesota's regional haze SIP for taconite facilities. The 
only mention of the disapproval of Minnesota's SIP is in the preamble 
where EPA makes an unsubstantiated claim that the disapproval is for a 
failure to make a required submission. MPCA objects to the absence of 
substantive analysis or detail about what required submission Minnesota 
did not make. According to Train v. Natural Resources Defense Council, 
421 U.S. 60 (1975), EPA can only disapprove a SIP if it does not 
adequately meet the requirements of CAA section 110(a)(2).
    Also, no mention of the disapproval of the SIP is made in the 
``Proposed Agency Action'' section of the Federal Register notice, 
meaning that EPA failed to take official agency action on Minnesota's 
regional haze SIP. Finally, EPA's assertion that the August 15, 2012, 
proposed rule served as an ``implicit disapproval'' of Minnesota's 
regional haze SIP represents an acknowledgement that EPA failed to take 
formal agency action. The CAA has no provision for ``implicit 
disapprovals.''
    Response: EPA's August 15, 2012, proposed action noted that the 
BART requirement ``has not been satisfied by Minnesota or Michigan for 
its subject taconite plants,'' and the proposed codification for 
Minnesota stated that ``[t]he requirements of section 169A of the Clean 
Air Act are not met because the regional haze plan submitted by the 
state on December 30, 2009, and on May 8, 2012, does not meet the 
requirements of 40 CFR 51.308(e) with respect to NOX and 
SO2 emissions from [the listed taconite facilities].'' This 
proposed codification represents ``official agency [proposed] action.''
    On the other hand, EPA agreed with comments expressing concern that 
the August 12, 2012 notice did not provide adequate explanation of 
EPA's rationale for proposing to disapprove in part the two States' 
regional haze SIPs. In response to those comments, EPA published a 
supplemental notice on February 6, 2013, at 78 FR 8478, elaborating on 
EPA's rationale for proposing to disapprove in part these SIPs.
    EPA did not claim, in its August 12, 2012 proposed action or 
elsewhere, that the disapproval was for failure to submit a timely 
regional haze SIP. EPA stated that its FIP obligation was based on a 
finding that the states had failed to make the submittal (an obligation 
that remains in the absence of EPA approval of a subsequently submitted 
SIP). Similarly, EPA did not make the assertion regarding ``implicit 
disapproval'' claimed by Minnesota. Instead, EPA stated that its August 
12, 2012 proposed action ``provided a full discussion of why EPA 
proposed to conclude that the BART criteria resulted in more stringent 
control than was required by the States, thus implicitly concluding 
that the state submittals did not require controls representing BART.'' 
Furthermore, the action proposed regulatory text stating that the state 
submittals failed to require BART for the taconite plants.
    In regard to MPCA's comment regarding the DC Circuit's decision in 
Train v. NRDC, that case did not deal with a regional haze SIP, which, 
in addition to satisfying the requirements of CAA section 110(a)(2), 
must also satisfy the requirements of CAA section 169A and the Regional 
Haze Rule.
    Comment: MPCA comments that EPA's claims of authority to promulgate 
a FIP based on a finding of failure to

[[Page 59838]]

submit a SIP fail to recognize that Minnesota ``submitted a SIP within 
the timeline [prescribed] by the 2009 Notice of Deficiency.'' MPCA 
objects that the administrative record omits an April 1, 2010, letter 
finding Minnesota's SIP to be complete.
    Response: This comment regarding FIP authority has been addressed 
in EPA's FIP and is not relevant here. EPA has made the addition to the 
record that MPCA requested.
    Comment: MPCA comments that EPA's supplemental proposed disapproval 
does not remedy EPA's official statements from the June 12, 2012, final 
action in which EPA stated that it would act through a FIP because 
Minnesota did not select EPA's chosen control technology for BART. MPCA 
finds these findings to be contrary to case law, citing Virginia v. 
EPA, 108 F.3d 1397, 1406 (D.C. Cir. 1997), without explanation.
    Response: It is not clear why MPCA commented on statements in the 
June 12, 2012, final action regarding prospective EPA actions, because 
the actual actions as proposed on August 15, 2012, and February 6, 
2013, were available for comment and were more indicative of EPA's 
actions than its prior anticipatory statements. In any case, MPCA is 
misrepresenting EPA's June 12, 2012, statements. Nowhere in this notice 
did EPA ``simply state that Minnesota did not select EPA's chosen 
control technology.'' Instead, EPA noted the ``significant information 
about additional NOX controls [that it received] in comments 
on [the January 25, 2012] proposed rulemaking, [prompting EPA to defer 
action to allow] EPA time to evaluate properly additional potential 
emission controls for the taconite facilities.''
    As explained in the supplemental proposed disapproval, EPA's 
subsequent evaluation led to its view that: (1) Minnesota and Michigan 
did not properly evaluate BART for NOX and SO2 
for taconite plants because they dismissed technically and economically 
feasible control technologies without adequate justification; (2) 
Minnesota and Michigan adopted a ``control technology'' (GCP) for 
NOX that was not defined or properly analyzed; (3) Minnesota 
did not demonstrate that its emission limits in fact represented GCP, 
while Michigan did not include NOX emission limits in its 
SIP at all; and (4) Minnesota failed to make its emission limits 
appropriately enforceable. MPCA notably does not contest EPA's view 
that low NOX burners and FGD are more effective at reducing 
emissions than GCP. More precisely, MPCA does not argue that GCP is 
either equivalent to or better than these technologies. That is, 
Minnesota makes no argument that its chosen technology can equally well 
be considered the best available. Thus, the failure of the states to 
follow EPA's implementing regulations or the BART Guidelines when 
selecting BART for these facilities had the practical consequence of 
the SIPs requiring less than BART. Finally, Virginia v. EPA, to the 
extent it is relevant here, does not instruct EPA to approve SIPs that 
fail to meet CAA requirements.
    Comment: EPA correctly states that Minnesota essentially re-opened 
its 2009 BART determinations for the affected electric generating units 
(EGUs). MPCA finds that an extensive administrative record compiled by 
EPA in support of revised action supported re-opening the EGU BART 
determinations. MPCA finds in contrast that it lacked an adequate 
administrative record to justify a re-opening of the taconite BART 
determinations.
    Response: At issue here is whether MPCA should have more thoroughly 
considered evidence indicating the applicability and effectiveness of 
low NOX burners. This comment suggests that MPCA undertook a 
partial consideration of this evidence in order to evaluate whether a 
more thorough review and ``re-opening'' of its BART determinations was 
necessary. However, as has been elaborated several times in the 
responses above, EPA disagrees that MPCA's partial consideration of 
information regarding the technical feasibility of low NOX 
burners was reasonable or sufficient to satisfy the States' obligations 
under the CAA and Regional Haze Rule to evaluate all technically 
feasible control options when selecting BART.
    Comment: MPCA closes its comments by recommending several 
modifications to the FIP.
    Response: These comments are pertinent to a completed rulemaking 
promulgating the FIP and are not germane to this rulemaking regarding 
disapproval of Minnesota's regional haze SIP.

III. What action is EPA taking?

    EPA is disapproving in part the Michigan and Minnesota regional 
haze SIPs for failure to satisfy BART requirements for NOX 
and SO2 emissions from the subject taconite facilities 
within these states. Specifically, EPA is disapproving in part the 
Michigan and Minnesota regional haze SIPs for failure to comply with 40 
CFR 51.308(e)(1)(ii)(A), which requires BART determinations ``to be 
based on an analysis of the best system of continuous emission control 
technology'' that, among other things, ``take[s] into consideration the 
technology available.'' EPA is also disapproving in part the Michigan 
regional haze SIP for failure to comply with 40 CFR 51.301 and 
51.308(e), which require BART determinations to be in the form of 
enforceable ``emission limitations'' contained in SIPs.
    For NOX emissions, Minnesota's SIP failed to ``take into 
consideration the technology available'' because it summarily dismissed 
a technically feasible control technology, low NOX burners, 
without adequate explanation. Furthermore, by selecting an unidentified 
set of practices as BART instead of low NOX burners, 
Minnesota's SIP failed to require the emission reductions associated 
with ``the best system of continuous emission control technology 
available.'' For SO2 emissions, Minnesota's SIP also failed 
to make BART determinations for certain facilities ``based on an 
analysis of the best system of continuous emission control technology'' 
because Minnesota did not reasonably consider the ``costs of 
compliance'' when it improperly rejected the most stringent control 
option, FGD.
    Similarly, Michigan's SIP also failed to ``take into consideration 
the technology available'' because it too summarily dismissed low 
NOX burners as a technically infeasible control option. 
Also, by selecting an unidentified set of practices as BART instead of 
low NOX burners, Michigan's SIP failed to require the 
emission reductions associated with ``the best system of continuous 
emission control technology available.'' Moreover, unlike Minnesota's 
SIP, Michigan's SIP did not include ``emission limitations representing 
BART'' for its Tilden facility. While Michigan commented that it has 
since issued a permit establishing NOX emission limits for 
the Tilden facility, neither these limits nor any other emission limits 
were included in the SIP as required. Finally, for SO2 
emissions, Michigan's SIP also failed to make BART determinations for 
certain facilities ``based on an analysis of the best system of 
continuous emission control technology'' because Michigan did not 
reasonably consider the ``costs of compliance'' when it improperly 
rejected the most stringent control option, FGD.
    A discussion of how this action relates to the taconite FIP that 
was published on February 6, 2013 is discussed in the February 6, 2013 
supplemental proposed disapproval.

[[Page 59839]]

IV. Statutory and Executive Order Reviews

Executive Order 12866: Regulatory Planning and Review

    Under Executive Order 12866 (58 FR 51735, October 4, 1993), this 
action is not a ``significant regulatory action'' and, therefore, is 
not subject to review by the Office of Management and Budget.

Paperwork Reduction Act

    This rule does not impose an information collection burden under 
the provisions of the Paperwork Reduction Act of 1995 (44 U.S.C. 3501 
et seq.).

Regulatory Flexibility Act

    This action merely disapproves state law as not meeting Federal 
requirements and imposes no additional requirements beyond those 
imposed by state law. Accordingly, the Administrator certifies that 
this rule will not have a significant economic impact on a substantial 
number of small entities under the Regulatory Flexibility Act (5 U.S.C. 
601 et seq.).

Unfunded Mandates Reform Act

    Because this rule disapproves pre-existing requirements under state 
law and does not impose any additional enforceable duty beyond that 
required by state law, it does not contain any unfunded mandate or 
significantly or uniquely affect small governments, as described in the 
Unfunded Mandates Reform Act of 1995 (Pub. L. 104-4).

Executive Order 13132: Federalism

    This action also does not have Federalism implications because it 
does not have substantial direct effects on the states, on the 
relationship between the national government and the states, or on the 
distribution of power and responsibilities among the various levels of 
government, as specified in Executive Order 13132 (64 FR 43255, August 
10, 1999). This action merely disapproves a state rule, and does not 
alter the relationship or the distribution of power and 
responsibilities established in the CAA.

Executive Order 13175: Consultation and Coordination With Indian Tribal 
Governments

    This rule also does not have tribal implications because it will 
not have a substantial direct effect on one or more Indian tribes, on 
the relationship between the Federal Government and Indian tribes, or 
on the distribution of power and responsibilities between the Federal 
Government and Indian tribes, as specified by Executive Order 13175 (59 
FR 22951, November 9, 2000).

Executive Order 13045: Protection of Children From Environmental Health 
and Safety Risks

    This rule also is not subject to Executive Order 13045 ``Protection 
of Children from Environmental Health Risks and Safety Risks'' (62 FR 
19885, April 23, 1997), because it disapproves a state rule.

Executive Order 13211: Actions That Significantly Affect Energy Supply, 
Distribution, or Use

    Because it is not a ``significant regulatory action'' under 
Executive Order 12866 or a ``significant energy action,'' this action 
is also not subject to Executive Order 13211, ``Actions Concerning 
Regulations That Significantly Affect Energy Supply, Distribution, or 
Use'' (66 FR 28355, May 22, 2001).

National Technology Transfer Advancement Act

    In reviewing state submissions, EPA's role is to approve state 
choices, provided that they meet the criteria of the CAA. In this 
context, in the absence of a prior existing requirement for the state 
to use voluntary consensus standards (VCS), EPA has no authority to 
disapprove a state submission for failure to use VCS. It would thus be 
inconsistent with applicable law for EPA, when it reviews a state 
submission, to use VCS in place of a state submission that otherwise 
satisfies the provisions of the CAA. Thus, the requirements of section 
12(d) of the National Technology Transfer and Advancement Act of 1995 
(15 U.S.C. 272 note) do not apply.
    Under section 307(b)(1) of the CAA, petitions for judicial review 
of this action must be filed in the United States Court of Appeals for 
the appropriate circuit by November 29, 2013. Filing a petition for 
reconsideration by the Administrator of this final rule does not affect 
the finality of this action for the purposes of judicial review nor 
does it extend the time within which a petition for judicial review may 
be filed, and shall not postpone the effectiveness of such rule or 
action. This action may not be challenged later in proceedings to 
enforce its requirements. See CAA section 307(b)(2).

List of Subjects in 40 CFR Part 52

    Environmental protection, Air pollution control, Incorporation by 
reference, Intergovernmental relations, Nitrogen dioxide, Particulate 
matter, Reporting and recordkeeping requirements, and Sulfur oxides.

    Dated: September 11, 2013.
Susan Hedman,
Regional Administrator, Region 5.

    40 CFR part 52 is amended as follows:

PART 52--APPROVAL AND PROMULGATION OF IMPLEMENTATION PLANS

0
1. The authority citation for part 52 continues to read as follows:

    Authority:  42 U.S.C. 7401 et seq.


0
2. Section 52.1183 is amended by adding paragraph (o) to read as 
follows:


Sec.  52.1183  Visibility protection.

* * * * *
    (o) The requirements of section 169A of the Clean Air Act are not 
met because the regional haze plan submitted by the state on November 
5, 2010, does not meet the requirements of 40 CFR 51.308(e) with 
respect to NOX and SO2 emissions from Tilden 
Mining Company L.C. of Ishpeming, Michigan. The requirements for this 
facility are satisfied by complying with Sec.  52.1183(k-n)

0
3. Section 52.1236 is amended by adding paragraph (d) to read as 
follows:


Sec.  52.1236  Visibility protection.

* * * * *
    (d) The requirements of section 169A of the Clean Air Act are not 
met because the regional haze plan submitted by the state on December 
30, 2009, and on May 8, 2012, does not meet the requirements of 40 CFR 
51.308(e) with respect to NOX and SO2 emissions 
from United States Steel Corporation, Keetac of Keewatin, Minnesota; 
Hibbing taconite company of Hibbing, Minnesota; United States Steel 
Corporation, Minntac of Mountain Iron, Minnesota; United Taconite, LLC 
of Forbes, Minnesota; ArcelorMittal Minorca Mine, Inc. near Virginia, 
Minnesota; and Northshore Mining Company-Silver Bay of Silver Bay, 
Minnesota. The requirements for these facilities are satisfied by 
complying with the requirements of Sec.  52.1235.

[FR Doc. 2013-23394 Filed 9-27-13; 8:45 am]
BILLING CODE 6560-50-P


