December 7, 2020

Air and Radiation Division
Environmental Protection Agency (EPA), Region 8
1595 Wynkoop Street,
Denver, Colorado 80202–1129

submitted via: Federal Rulemaking Portal: https://www.regulations.gov

Re: Docket ID No. EPA–R08–OAR–2020–0098, Proposed Approval and Promulgation of Implementation Plans; State of Utah; Salt Lake City and Provo, Utah PM2.5 Redesignations to Attainment and Utah State Implementation Plan Revisions

Dear Air and Radiation Division,

Thank you for this opportunity to comment on the EPA proposed approval and promulgation of implementation plans; State of Utah; Salt Lake City and Provo, Utah PM2.5 redesignations to attainment and Utah State Implementation Plan revisions (Docket ID No. EPA–R08–OAR–2020–0098. I make these comments on behalf of the Utah Chapter of Sierra Club, Utah Physicians for a Healthy Environment, HEAL Utah and Western Resource Advocates. As we establish below, many aspects of the rules and plans that EPA is proposing to approve fail to meet the requirements of the Clean Air Act and do not adequately safeguard the public health and environment of Utah. As a result, we ask that EPA not approve those elements of Utah’s rules and plans detailed below.

I. Commenting Organizations

The Utah Chapter of the Sierra Club works to protect Utah’s wild places, wildlife, and waters, as well as the people and communities who depend on them. Our activism and advocacy are based on our strong grassroots networks, citizen-based leadership, and the guidance and skillsets of professional staff support. With over 5,600 members, and growing, we work to protect public lands, promote renewable energy, and support initiatives that promote clean air 2 strategies. We maintain a presence at the Utah Legislature to advocate on the full spectrum of environmental issues and amplify the voices of our members.

Western Resource Advocates is a regional non-profit conservation organization with programs and staff spanning the Intermountain West, including Utah. Our mission is to protect the West’s land, air and water to ensure that vibrant communities exist in balance with nature. To this end,
we work to ensure the West will have clean air and clean water to support healthy communities and vital habitat.

Utah Physicians for a Healthy Environment is dedicated to protecting the health and well-being of the citizens of Utah by promoting science-based health education and interventions that result in progressive, measurable improvements to the environment and our health.

HEAL Utah promotes clean air and renewable energy, and protects public health and the environment from nuclear and other toxic threats. Representing more than 20,000 members, HEAL has a long track record of achieving positive change in the state by mobilizing local communities, promoting science-based public policy and legislation, and strategically working with regulatory agencies.

The organizations’ interest in the present matter is based on the public health crisis that exists as a result of severe and frequent spikes in PM2.5 air pollution that occur in northern Utah. These acute, and often long-lasting episodes of high concentrations of PM2.5 jeopardize the health and
well-being of northern Utah’s residents and visitors. The organizations listed above have each already spent thousands of hours and significant resources on behalf of their members, supporters and staff in efforts to influence the decisions now before EPA.

II. Introduction

Importantly, most of our staff, members and supporters live along Utah’s highly populated Wasatch Front, an area plagued in the winter by considerable spikes in concentrations of fine particular matter (PM2.5). We know firsthand what it is like to live, work and raise families in an area where air pollution threatens our health, our economy and our environment. We understand that because of the high concentrations of PM2.5 air pollution in our communities, our parents risk premature death, our children’s lung function may be permanently stunted, and we are more likely to have heart attacks and strokes. Some of our communities are particularly hard hit by PM2.5 because they are exposed to disproportionately high levels of other pollutants, for example, from highways and oil refineries and are subject to exacerbating inequities such as a lack of access to health care. Additionally, the frequency and severity of ozone pollution along the Wasatch Front has been increasing, meaning our neighborhoods are exposed to unhealthy levels of air pollution in the winter and summer. Because of the importance of your decision making on our health and quality of life, we ask you to follow the letter of the law and require of Utah further emission reductions as demanded by the Clean Air Act.

III. Specific Comments.

A. EPA’s Failure to Extend the Comment Period is Unconscionable and Antidemocratic.

On November 10, 2020, EPA refused to grant the request by Dr. Moench, on behalf of Utah Physicians for a Healthy Environment and other public health and conservation organizations, that the comment period on the agency’s proposed approvals be extended by 30-days. Dr. Moench noted that 30 days was not “enough time for organizations and concerned citizens to respond, especially under the current constraints of the pandemic and upcoming holiday activity.”1

To deny this request for an extension during a pandemic and holiday period effectively deprived the public of a meaningful opportunity to response to the complex, technical and consequential decisions EPA plans to make. This is particularly true given that Utah residents are struggling with care-giving and work-related obligations and emotional and physical distress due to the very high rates COVID-19 infections and hospitalizations in our state. EPA’s refusal to give the public more time to ensure that their government is adequately protecting their health is unconscionable and undermines principles of government accountability and adequate administrative procedures. At precisely the time that EPA should be even more vigilant in carrying out its mission to safeguard public health and listening to those affected by its determinations, the agency has gone the other direction and denied the public a real chance to influence agency decisionmaking.

Moreover, the scope of EPA’s proposed decision is of a significance rarely seen. As you explain, in one single action, EPA is proposing

the redesignation of the Salt Lake City, Utah and Provo, Utah nonattainment areas (NAAs) to attainment for the 2006 24-hour fine particulate matter with an aerodynamic diameter less than or equal to a nominal 2.5 microns (PM2.5) National Ambient Air Quality Standard (NAAQS), and also acting on multiple related State Implementation Plan (SIP) submissions. We are proposing to approve SIP revisions submitted by the State of Utah on January 19, 2017; April 19, 2018; February 4 and 15, 2019; and January 13, May 21, and July 21, 2020. These SIP submissions include revisions to Utah Administrative Code (UAC) Sections R307-110, R307-200, and R307-300 Series; revisions to Utah SIP Sections X.B and E; revisions to Utah SIP Sections IX.H.11, 12, and 13; best available control measures/best available control technologies (BACM/BACT) PM2.5 determinations for Salt Lake City and Provo; maintenance plans for the Salt Lake City and Provo areas for PM2.5; and the request for redesignation under the 2006 24-hour PM2.5 standard. Additionally, the EPA is proposing to approve, through parallel processing, a request to remove startup and shutdown emission limits for Kennecott’s Power Plant in the Utah SIP and the accompanying R307-110-17 revisions.

85 Fed. Reg. 71023, 71023 (Nov. 6, 2020). Clearly, it is unreasonable to grant the public only 30 days to address all these planned actions. Moreover, it is both ironic and telling that EPA has given itself four years to consider just one of the several SIP revisions submitted by Utah – dated January 19, 2017 – but has refused to give the public 60 days to comment on not only this submission but the multiple additional, highly technical and voluminous proposals EPA now has under consideration.

Based on the above, we ask that EPA reissue its proposed decisions and provide, at a minimum, a 60-day public comment period on the proposed decisions encompassed by 85 Fed. Reg. 71023.

B. Because Exposure to PM2.5 Puts Utahns at Significant Risk for Disease and Death, EPA Must Do More to Safeguard Public Health.

A highly reliable analysis of the PM NAAQS was undertaken and reported to EPA on October 22, 2019 by the members of the CASAC Particulate Matter Review Panel that were dismissed without notice by press release on October 10, 2018. After being disbanded, the members formed the Independent Particulate Matter Review Panel (Independent Panel) and commented to EPA on the ongoing review of the PM NAAQS on various occasions, including in an October 22, 2019 submission. In summarizing the findings of that analysis, the Independent Panel found that unequivocally, the current PM NAAQS are not tight enough to protect public health and that leaving them at current levels would result in significant avoidable mortality:

[O]n the basis of more recent evidence, as described below, exposure to ambient PM2.5 at the levels of the current standards is estimated by the EPA to be responsible for tens of thousands of premature deaths in the United States each year.2

After its continued review of scientific evidence from sources including epidemiologic studies, toxicological studies in animals, and controlled human exposure studies, the Independent Panel determined that, to protect public health, including the health of at-risk groups, the annual
standard must be set between 10 µg/m3 and 8 µg/m3. The Independent Panel cautioned that “even at the lower end of the range, risk is not reduced to zero. The margin of safety increases as the level of the standard is lowered within this range.”3

The Independent Panel also determined that the current 24-hour PM2.5 standard is insufficient to protect public health in locations such as the Wasatch Front, where the 24-hour standard, rather than the annual standard, would be controlling.4 This is true even if the annual PM2.5 standard were lowered.5 Again, the panel specified that an adequate short-term PM2.5 standard would have to be set at between 30 µg/m3 and 25 µg/m3.6 While the panel agreed that the current form of the 24-hour standard was appropriate, it found this to be true only in the context of a lowering of the NAAQS to between 30 µg/m3 and 25 µg/m3.7

Furthermore, as EPA is aware, studies in the United States,8 United Kingdom,9 Italy,10 Europe,11 and China12 indicate that exposure to poor air quality increases the risks of COVID-19 infection and mortality. Indeed, one analysis identified a preliminary link between a small increase in long-term exposure to PM2.5 and the likelihood of death from COVID-19.13 This Harvard study found that someone who lives for decades in a county with high levels of fine particulate pollution is 8% more likely to die from COVID-19 than someone who lives in a region that has just one unit (one microgram per cubic meter) less of such pollution.

Of additional importance is that higher death rates have been observed among minority and lowincome communities, underscoring that pre-existing economic and health disparities result from increased exposure to air pollution. According to the Centers for Disease Control, “long-standing systemic health and social inequities have put some members of racial and ethnic minority groups at increased risk of getting COVID-19 or experiencing severe illness, regardless of age.”14 Further, minority populations are disproportionately exposed to air pollution, which harms health and may make a person more vulnerable to COVID-19.

Thus, COVID-19 underscores the health disparities that far too many Utahns face. The same populations that are disproportionately exposed to environmental hazards such as air pollution are also at increased risk from the pandemic. As a result, in assessing BACT, BACM, the proposed SIP revisions, redesignation and the maintenance plans, EPA must be certain that it has complied with the letter and intent of the Clean Air Act and has requirements of Utah the maximum emission reductions and control measures required by law and has indeed ensured that adequate, enforceable provisions exist to ensure that all Wasatch Front residents will have clean air to breath for decades to come.

C. EPA Is Legally Obligated to Reject Utah’s BACT Analysis.

1. Background

a. Utah’s BACT Review Must Be Robust and Must Lead to a Defensible Emission Unit Specific Limitation.

The BACT review process required for a serious SIP will be, at a minimum, as rigorous as that required under the Prevention of Significant Deterioration (PSD) program. As EPA explains: “BACT determinations for PM2.5 NAAQS implementation are to follow the same process and criteria that are applied to the BACT determination process for the PSD program.” 81 Fed. Reg. Indeed, to be consistent with longstanding policy applied to PM10 nonattainment areas, serious SIP BACT is “more stringent” than PSD BACT:

The BACT under the PSD program applies only in areas already meeting the NAAQS, while PM10 BACM applies in areas which are seriously violating the NAAQS. The difference in policy goals, arguably, suggests that the PM10 BACM control standard should be more stringent than that for PSD BACT . . . [I]t is reasonable to use the approach adopted in the PSD BACT program as defined in section 169(3) of the Act as an analogue for determining appropriate PM10 nonattainment control measures in serious areas, while at the same time retaining the discretion to depart from that approach on a case-by-case basis as particular circumstances warrant.

Id., fn. 160 (citations omitted).

BACT is defined, first and foremost, as an emission limitation. For example, Utah law specifies that BACT is:

an emissions limitation…based on the maximum degree of reduction for each air contaminant which would be emitted from any proposed…modification which the director, on a case-by-case basis, taking into account energy, environmental, and economic impacts and other costs, determines is achievable for such…modification[.]

Utah Admin. Code r.307-401-2.

The goals of BACT emission limitations are: “(1) to achieve the lowest percent reduction, (2) to protect short-term ambient standards, and (3) to be enforceable as a practical matter.” Utah Chapter of the Sierra Club v. Air Quality Board, 2009 UT 76, ¶ 4, 226 P.3d 719 (citing EPA, New Source Review Workshop Manual, B.6-.9). “Once the BACT is selected for a new facility, an emission limitation based on that control technology is also imposed as part of BACT.” Id.; NSR Manual B.56 (“BACT emission limits…must…demonstrate protection of short-term ambient standards (limits written in pounds/hour) and be enforceable as a practical matter (contain appropriate averaging times, compliance verification procedures and recordkeeping requirements).”).

To determine the emission limit that represents the maximum achievable reduction in air pollutants, BACT “review is often conducted using the five-step ‘top-down method,’ which in essence requires the applicant to adopt the most stringent control technology, unless it can show that the technology is not achievable due to energy, environmental, or fiscal impacts.” Id. As the Utah Supreme Court explained, BACT analysis begins with the identification of all available control technology options for each regulated pollutant. Sierra Club, ¶ 4 fn.2 (citing NSR Workshop Manual B.6-9). “In effect, the reviewer must consider lower emitting processes and practices [and] add-on controls[.]” Id. Then, based on a “documented demonstration,” the “reviewer eliminates technically infeasible options.” Id. The control technologies are next ranked by “effectiveness” based on “based on efficiency, emission rate, and emission reductions.” Id.

Starting with the most stringent technology, the “reviewer” next “objectively” evaluates the economic, environmental, and energy impacts, “both beneficial and adverse,” of the technologies. Id. Only if this analysis “proves” that the first ranked technology is inappropriate, is that technology eliminated and the next most effective alternative evaluated. Based on this process, the most effective, achievable technology is proposed as BACT. Id.

b. BACT Must Lead to an Emission Unit Specific Emission Limit.

EPA has consistently required that BACT analysis lead to an emission limitation for each emission unit. As the Environmental Appeals Board held:

As reflected by Draft NSR Manual, the current EPA policy is that “each new or modified emission unit (or logical grouping of new or modified emissions units) subject to PSD is required to undergo BACT review.” Draft NSR Manual at B.10.

In re: General Motors, Inc., 10 EAD 360, 382 (EAB 2002). See also In re: Masonite Corporation, 5 EAD 551, 557-8 (EAB 1994) (footnote omitted) (“In a BACT determination, the Region must give consideration to each individual emissions unit or pollutant emitting activity subject to review. New Source Review Workshop Manual at B.4.”); In re: Knauf Fiber Glass, GMBH, 8 EAD 121, 129-30 (EAB 1999) (“The first step in the BACT selection process involves identifying and listing all ‘available’ control options. NSR Manual at B.5. The term available is used in its broadest sense under the first step and refers to control options with a ‘practical potential for application to the emissions unit’ under evaluation.”).

The NSR Manual further states that BACT analysis must be undertaken for each emission unit:

The BACT requirement applies to each individual new or modified affected emission unit and pollutant emitting activity at which a net emissions increase would occur. Individual BACT determinations are performed for each pollutant subject to a PSD review emitted from the same emissions unit. Consequently, the BACT determination must separately address, for each regulated pollutant with a significant emissions increase at the source, air pollution controls for each emissions unit or pollutant emitting activity subject to review.

The Draft NSR Workshop Manual at B.4.1.15

Utah’s BACT rule also provides that BACT analysis applies to specific “emission unit[s].”

If the director determines that technological or economic limitations on the application of measurement methodology to a particular emissions unit would make the imposition of an emissions standard infeasible, a design, equipment, work practice, operational standard or combination thereof, may be prescribed instead to satisfy the requirement for
the application of best available control technology.

Utah Admin. Code r.307-401-2 (emphasis added).

c. BACT Includes Any Feasible Technologies that Could Have Been Partially or Fully Implemented by December 31, 2019.

Utah must require the adoption of any otherwise qualifying BACT if it could have been fully or partially implemented within 4 years after an area is reclassified as a serious NAA. 81 Fed. Reg. at 58080 (“A Serious area attainment plan must include provisions to implement BACM on sources in a Serious nonattainment area, as provided by section 189(b)(1)(B), no later than 4 years after reclassification.”). In this context, “implement” includes either partial or full implementation. Id. (“A state must identify those technologically and economically feasible control measures and technologies that it can implement fully or partially within 4 years of reclassification of its Serious PM2.5 nonattainment area.”). The Salt Lake City NAA was reclassified in as a serious NAA in May 2017. 82 Fed. Reg. 21711. Four years after May 2017 is beyond the serious attainment date of December 2019. Therefore, the Director must impose any BACT measure that can be fully or partially implement before December 2019. Id. at 58085
(“Where the earliest date that a measure can be implemented is beyond the 4-year mark following reclassification to Serious, the measure may still be needed as an “additional feasible measure” if the 4-year mark occurs before the Serious area attainment date.”). Plainly, delays by Utah or EPA in proposing, finalizing and reviewing BACT determinations may not excuse the imposition of BACT that could have been fully or partially implemented within the applicable time frame.

d. BACT Represents the Maximum Reduction of Emissions Achievable.

BACT is “the maximum degree of emission reduction achievable… considering energy, economic and environmental impacts and other costs.” 81 Fed. Reg. at 58081. BACM must be more stringent than the “reasonably available control measures” [RACM] Utah has previously applied in the Salt Lake NAA: “[B]est control measures should represent a more stringent and potentially more costly level of control. The level of stringency generally refers to the overall level of emissions reductions of a control measure or technology, or of such measures and technologies combined.” Id. “BACM puts a ‘greater emphasis on the merits of the measure or technology alone,’ rather than on ‘flexibility in considering other factors,’ in contrast to the approach for determining [reasonably available control measures] RACM and RACT.” Id. at 58081.

e. BACM is “Generally Independent” of Attainment.

In contrast to RACM, BACM/BACT are “generally independent of the attainment.” Id. at 58081 82; id. at 58082 (“[I]nterpreting the Serious PM2.5 nonattainment area BACM/BACT requirements to be “generally independent” of attainment is consistent with the structure and substance of the CAA”). “BACM/BACT measures for Serious areas are not solely limited to those measures needed for expeditious attainment under this final rule.” Id. at 58020. Thus, Utah must derive and apply BACM independently of any need to show that the measure contributes to attainment. This is because the “robust emission reduction programs” required by BACM independent of attainment “are needed to bring about expeditious attainment and public health protection of citizens in these nonattainment areas.” Id. at 58082. “[I]nterpreting BACM/BACT to be generally independent of the attainment needs of a Serious PM2.5 area will ensure continued progress toward attainment for those areas with more difficult air quality problems.” Id.

f. Measures Adopted in Other States Are Assumed to be Technologically Feasible.

Repeatedly, EPA has underscored that measures adopted in other nonattainment areas are assumed to be BACM and BACT: “A state must at a minimum continue to consider as potential BACM any technologically feasible control measures or technologies implemented by similar sources.” Id. at 58085; see also id. at 58084 (“[O]ther states across the country are important sources of information about control measures.”). This is true even if those measures are costly. Id. at 58087 (“[A]ll states with Serious areas need to consider implementing BACM and BACTlevel measures that have been implemented in other states, even if those measures incur higher costs.”). Indeed, a measure adopted in another state is presumed to be BACT and “where a given control measure has been applied in another NAAQS nonattainment area (for PM2.5 or other pollutant), the state will need to provide a detailed justification for rejecting any potential BACM measure as technologically infeasible.” Id. at 58085.

g. BACT Will Be More Expensive than RACT.

States need to consider emission reduction measures with higher costs per ton when assessing the economic feasibility of BACM and BACT controls (and, where applicable, additional feasible measures) as compared to the economic feasibility criteria applied in their RACM and RACT analysis…for the same nonattainment area.” Id. at 58085. Indeed, in assessing BACM, Utah may not give economic feasibility significant weight and must meet a particularly rigorous showing of economic infeasibility before rejecting a technologically feasible control. Id. at 58085 (“EPA maintains that while the economic feasibility of a control measure is as important as its technological feasibility under the RACM and RACT determination process, economic feasibility is a less significant factor in the BACM and BACT determination process. In other words, a state must apply a higher standard for eliminating a technologically feasible control measure from further consideration as BACM due to cost alone.”).

h. Utah Must Also Consider Control Technologies that Have Not Been Implemented Elsewhere.

To derive and impose BACT, Utah must consider even those feasible technologies that may not have been adopted by other states. “[A] state may not automatically eliminate a particular control
measure merely because other sources have not implemented the measure. In other words, a state must continue to consider technologically feasible measures that have not been implemented by similar sources but that can nonetheless effectively reduce emissions from the source category in question at a cost that is not cost prohibitive.” 81 Fed. Reg. at 58085.

2. The BACT Limits Reflected in IX.H.11, 12 and 13 Are Not Legally Sufficient.

Despite these rigorous requirements, Utah failed to derive and implement BACT for major sources in the Salt Lake City and Provo NAAs, relying chiefly on technology and practices adopted as RACM/RACT. In so adopting RACM/RACT as BACT and failing to require legally sufficient BACT, Utah has not: 1) shown that it has developed and imposed emission-unitspecific emission limits that represent the maximum achievable reductions of emissions of PM2.5 and PM2.5 precursors; 2) produced a complete review of technology adopted in other states and for other similar facilities and emission units; 3) established why technologies adopted in other states and for other facilities and emission units are not technologically or economically feasible; 4) applied BACT’s “higher economic costs” analysis; or 5) provided objective data to support the state’s contentions. As a result, EPA should reject Utah’s BACT analyses and require the state to derive and impose adequate BACT emission limitations and requirements.

a. Utah Improperly Relies on NSPS as BACT for the Refineries.

Utah relies heavily on New Source Performance Standards (NSPS) as BACT for the Salt Lake City NAA refineries. However, by rule, the refinery NSPS, including 40 C.F.R. 60.100a to 109a (Subpart Ja) reflects “best demonstrated technology” that is available and cost-effective for all refineries in the nation. 73 Fed. Reg. 35838, 35839 (June 24, 2008). Thus, the NSPS do not rise to the level of BACT – particularly without a robust analysis as to why more restrictive controls and emission limits adopted elsewhere are not warranted. As EPA and Utah’s BACT Rules make clear, NSPS necessarily provides the floor for a BACT emission limitation. New Source Review
Manual, B.12 (“NSPS simply defines the minimal level of control to be considered in the BACT analysis.”). A defensible FCCU BACT emission limitation, therefore, starts with NSPS, including Subpart Ja and considers more rigorous controls that those imposed by these nationwide minimum standards.

That Utah does indeed limit the bulk of is BACT analysis and determinations to Subpart Ja is evident by comparing provisions such as 40 C.F.R. 60.102a to the general and specific BACT requirements applied to the oil refineries in the Salt Lake City NAA. E.g. see IX.H.11.g (general); IX.H.12.b (Big West); IX.H.12.d (Chevron); IX.H.12.f (Holly); IX.H.12.m (Tesoro).16 The IX.H.11 and 12 provisions applicable to flares are also limited to the Subpart Ja and MACT CC flaring standards. Moreover, EPA, without analysis, discussion or evidence in the record merely “agrees” with the adoption of NSPS as BACT. E.g. see EPA TDS at 100, 107, 111, 123, 134, 162, 174, 238, 250. Particularly without consideration of other available and more effective technologies, this unexamined adopted to NSPS fails to comply with BACT.

b. Utah Failed to Require Adequate Monitoring, Reporting and Recordkeeping Requirements.

Section 172(c)(6) in subpart 1 of the Clean Air Act requires nonattainment SIPs to “include enforceable emission limitations, and such other control measures, means or techniques . . . as
well as schedules and timetables for compliance, as may be necessary or appropriate to provide for attainment.” 42 U.S.C. § 7502(c)(6). To this end, an adequate SIP provision must “definitively state the recordkeeping and monitoring requirements appropriate to the type of sources being regulated. The recordkeeping and monitoring requirements would have to be sufficient to enable the state or the EPA to determine whether the source is complying with the emission limit on a continuous basis.” 81 Fed. Reg. at 58133. “[A]ppropriate stationary source emissions monitoring requirements, like the control measures with which they are associated, are a fundamental element of an approvable SIP.” Id. at 58140.

To meet this goal, a SIP monitoring provision “must” include electronic “[r]eporting and record retention requirements,” including mandates for “retaining monitoring and test data in an electronic form” and “periodic electronic reporting of information as needed to the compliance office.” Id. Reports sufficient to show continuous compliance with emission limits should be “made available online so that the general public can readily access the information without the need to submit Freedom of Information Act (FOIA) requests to the EPA.” Id.; see also id. at 58135 (“EPA agrees that electronic reporting and public access to information is important.”). Rather than requiring adequate and regular record and reporting requirements as part of the SIP, the Director proposes SIP provisions that: 1) authorize the source to maintain compliance records that are not available to the public; 2) allow the source to take years to report failures to meet SIP terms and conditions; and, 3) fail to require sufficient recording keeping and reporting to establish continuous compliance, often mandating instead only that the source report noncompliance. See id. at 58140 (“Stationary source emissions monitoring…entails collecting and using measurement data (or other information) from individual stationary sources to demonstrate compliance with emissions standards, to assess process or control device performance, or to verify work practices.”); Sierra Club v. Air Quality Board, 2009 UT at ¶ 4 (determining that BACT emission limitations must be “enforceable as a practical matter.”).

Examples of Utah’s failure to require adequate recordkeeping and reporting requirements include IX.H.11.c.i, which states that:

Any information used to determine compliance shall be recorded for all periods when the source is in operation and such records shall be kept for a minimum of five years. Any or all of these records shall be made available to the Director upon request.

Importantly, a Utah District Court has determined that the public has no right to review any records maintained by a source and has no right to request that the Director obtain these records. Therefore, the public can be and has been denied access to records maintained by a source and subject only to request by the Director. Although there are some source specific recordkeeping and reporting requirements in IX.H.12 – largely imposed by federal regulations, IX.H.11.c.i is the only record keeping and reporting provision that applies to most emission limitations. As a result, the majority of the BACT limits imposed in IX.H.11 and 12 are legally inadequate and EPA’s proposed approval of them unlawful.

Further, as explained in more detail below, Utah’s Fugitive Dust Rule lacks adequate monitoring, recordkeeping and reporting requirements. Again, because the source maintains records that purport to show compliance with the rule, the public is denied access to monitoring reports. Moreover, the frequency of monitoring of the opacity limit is so rare – occurring every year or year and one-half – is grossly insufficient to show continuous compliance.

c. Utah Does Not Impose the Short Term BACT Emission Limits Necessary to Address the Failure to Attain the 24-Hour PM2.5 NAAQS.

The Salt Lake City NAA failed to attain the 24-hour PM2.5 NAAQS by the serious attainment date. The 24-hour PM2.5 standard – or “short-term” standard – is intended, by law, to protect “against health effects associated with short-term PM2.5 exposures, especially in areas with high peak PM2.5 concentrations.” 80 Fed. Reg. 15340, 15347 (March 23, 2015) (emphasis added). EPA determined that the 24-hour PM2.5 NAAQS is necessary to “provide[] increased public health protection, including the health of at-risk populations which include children, older adults, persons with pre-existing health and lung disease and persons of lower socioeconomic status, against a broad range of PM2.5-related effects that include premature mortality, increased hospital admissions and emergency department visits, and development of chronic respiratory disease.” Id.

Only short-term emission limits – with averaging periods of 24 hours or less – are adequate to prevent short-term spikes in air pollution. NSR Manual B.56 (“BACT emission limits… must…demonstrate protection of short-term ambient standards (limits written in pounds/hour) and be enforceable as a practical matter (contain appropriate averaging times, compliance verification procedures and recordkeeping requirements).”).

As the Utah Supreme Court explained, the goals of BACT emission limitations are: “(1) to achieve the lowest percent reduction, (2) to protect short-term ambient standards, and (3) to be enforceable as a practical matter.” Utah Chapter of the Sierra Club v. Air Quality Board, 2009 UT 76, ¶ 4, 226 P.3d 719 (citing EPA, New Source Review Workshop Manual, B.6-.9). Here, in addition to other longer-term emission limitations, limits that are averaged over periods of 24 hours or less are a necessary component of a SIP that addresses violations of the 24-hour PM2.5 NAAQS. Limitations averaged over periods longer than 24 hours – such as 7 days, 30 days or 365 days – do not prevent sharp increases in emissions over the short-term and thus do not sufficient protect the 24-hour NAAQS.

As EPA confirmed when it commented on Utah’s Moderate PM2.5 SIPs, “[u]nder a long-term limit, emissions from a source can spike during a short-term period.” EPA Region 8 Comments on Utah’s Proposed [Moderate] PM2.5 State Implementation Plans and Technical Support Documents at 8 (Oct. 30, 2014). The agency expounded that, for example, “[a]n emission limit expressed as a 30-day average allows significantly higher short-term emissions that can impact a short-term standard such as the 24-hour PM2.5 NAAQS.” Id. at 24.

Short-term emission limitations, as averaged over a period of 24 hours or less, are also necessary to reflect BACT – “the maximum degree of emission reduction achievable… considering energy, economic and environmental impacts and other costs.” 81 Fed. Reg. at 58081. Without analysis in the record, the Director cannot show that an emission limit averaged over a period longer than 24-hours and therefore allows short-term spikes in emissions is indeed BACT when compared to a short-term emission limit that prevents short terms spikes. See Sierra Club v. Air Quality Board, 2009 UT at ¶ 4 (determining the goals of BACT emission limitations as being: “(1) to achieve the lowest percent reduction, (2) to protect short-term ambient standards, and (3) to be enforceable as a practical matter.”).

Examples of IX.H.11 and 12 emission limits that fail to protect the 24-hour PM2.5 NAAQS include: 1) IX.H.11.g.i.A.I (general FCCU SO2 emissions averaged over 7 and 365 days); 2) IX.H.11.g.i.B.I (limit on FCCU PM emissions based on coke-burn off rather than time period); 3) IX.H.11.g.iii (H2S content of refinery fuel gas based on 365-day rolling average); 4) IX.H.12.l.v.B (NOX from each turbine stack averaged over 30 days); 5) IX.H.12.m.iii (365 day rolling average); 6) IX.H.12.a.i (no averaging time).

d. Utah’s Reliance on Plant-Wide Caps is Unlawful and Lacks Record Support.

As explained above, adequate BACT analysis must lead to an emission limitation for each emission unit.17 Therefore, Utah’s reliance on plant-wide caps to represent a BACT emission limit is contrary to the law. Rather, Utah must undertake BACT analysis for each emission unit and derive a defensible emission limit for that unit. That emission limit must represent the maximum achievable reduction of emission from that unit and must be practically enforceable. Without emission unit specific analysis and corresponding emission limits, Utah’s BACT fails to reflect “the maximum degree of emission reduction achievable[.]” 81 Fed. Reg. at 58081. Rather than meeting the requirements of BACT, Utah has in several instances proposed plant-wide caps without providing any explanation or documentation showing how those caps represent BACT. As a result, Utah’s BACT fails to comply with the law and EPA should reject it as such.

Even if it were permissible to rely on source-wide caps in Subpart H, to comply with BACT, the Director must determine, based on record evidence, unit-specific emission limits reflective of BACT and only then use those emission limits and any unit-specific capacity factors to develop plant-wide caps that reflect BACT. The record must reflect that this analysis complies with BACT, including by providing adequate documentation. However, the Director has not provided this documentation and analysis and so fails to explain how his proposed source-wide caps reflect unit-specific BACT controls and/or corresponding emission limitations. Thus, there is an unlawful disconnect between the controls and emission limits proposed by the sources and the plant-wide caps the Director has adopted purportedly as BACT.

Further, there must be an explicit and specific enforceable measure included in Subpart H for any emission factor used to establish (and/or to be used to establish compliance with) a plant-wide cap. For example, in cases where the Director has imposed plant-wide caps that include fugitive dust emissions from roads, compliance with which is going to be based on assumed emission factors, any assumptions for PM emission factors from roads must be tied to specific enforceable measures such as the frequency of road sweeping and/or the quantity and frequency of water or chemical dust suppressant application. As discussed in the attached technical comments, the control efficacy of these types of measures for fugitive dust is based on the frequency of application (as well as the quantity (i.e. amount of water or dust suppressant applied per area of road). If the permit or rule is vague – for example, if it requires only the watering of roads “as needed to minimize fugitive dust” – the accuracy of the emission factors used in developing emission caps, and in determining compliance with those caps, is highly questionable and fails to comply with the law.

Instances of unlawful source wide caps are those applied to each oil refinery in the Salt Lake City NAA. These include: 1) IX.H.12.b.i, b.ii & b.iii (Big West source wide caps on PM2.5, NOX, and SO2); 2) IX.H.12.d.i, d.ii & d.iii (Chevron source wide caps on PM2.5, NOX, and SO2); 3) IX.H.12.f.i, f.ii & f.iii (Holly source wide caps on PM2.5, NOX, and SO2); and, 4) IX.H.12.m.i, m.ii & m.iii (Tesoro source wide caps on PM2.5, NOX, and SO2).

That these source-wide caps do not represent BACT for each emission unit covered by the cap is evident. For example, the source wide cap on Big West NOX emissions is to be calculated as follows:

Compliance with the source-wide NOX Cap shall be determined for each day as follows:

Total 24-hour NOX emissions shall be calculated by adding the emissions for each emitting unit. The emissions for each emitting unit shall be calculated by multiplying the hours of operation of a unit, feed rate to a unit, or quantity of each fuel combusted at each affected unit by the associated emission factor, and summing the results.

IX.H.12.b.ii.C.

First, as explained above, this source wide cap, which is typical of those applied in Subpart H, runs afoul of the requirement that BACT result in an emission limit representing the maximum degree of emissions reduction achievable for each emission unit. Further, even if it were appropriate to bypass emission unit specific BACT, the record is devoid of any analysis showing that this source wide cap or any source wide cap represents the sum of all the BACT limits – again reflecting the maximum degree of emission reduction achievable for each emission unit – applicable to each of the emission units encompassed by the source wide cap. Therefore, when examined from every angle, Utah’s use of source wide caps to represent BACT is contrary to the Clean Air Act. EPA should therefore reject this substitute of source wide caps for legally sufficient BACT emission limits.

e. Utah Has Failed to Require Adequate Monitoring for the Purposes of Ensuring Compliance with BACT Emission Limits.

EPA has explained that to meet Clean Air Act requirements, a serious SIP must include adequate monitoring that ensures continuous compliance with any emission limitation or other BACT control. SIP imposed controls must be “enforceable,” 42 U.S.C. § 7502(c)(6) (“plan provisions shall include enforceable emission limitations”), and “measurable,” and “include periodic source testing, monitoring or other viable means to establish whether the source meets the applicable emission limit.” 81 Fed. Reg. at 58133. For an adequate SIP, the “monitoring requirements would have to be sufficient to enable the state or the EPA to determine whether the source is complying with the emission limit on a continuous basis.” Id. Moreover, because frequent monitoring is a critical element of an emission limit that reflects the maximum emission reduction, the Director must undertake the analysis necessary to show that the frequency of monitoring he proposes reflects BACT.

Frequent monitoring serves to increase the accuracy of emission inventories, to identify appropriate control measures and to reduce emissions. 80 Fed. Reg. 15340, 15453 (March 23 2015). “[A]ppropriate stationary source emissions monitoring requirements, like the control measures with which they are associated, are a fundamental element of an approvable implementation plan.” Id. For example, EPA has found that improved monitoring can provide information that allows a source to take “corrective action that could potentially reduce emissions up to 15 percent[.]” Id. Similarly, more frequent monitoring “could yield potential stationary source emissions reductions of up to 13 percent.” Id. Thus, adequate monitoring is a critical component of a SIP intended to ensure that the Salt Lake City serious NAA will meet the 24-hour PM2.5 NAAQS as expeditiously as practicable, see e.g. 42 U.S.C. § 7513(c)(1) & (c)(2).

For a SIP to be adequate, the “monitoring requirements would have to be sufficient to enable the state or the EPA to determine whether the source is complying with the emission limit on a continuous basis.” In commenting on the 2014 moderate PM2.5 SIPs – where only reasonably available control technology (RACT) was applied – EPA indicated that stack testing as infrequently as once every three years was inappropriate in the absence of rigorous analysis showing that such lax monitoring was sufficient to show continuous compliance. E.g. EPA Region 8 Comments on Utah’s Proposed PM2.5 State Implementation Plans and Technical Support Documents at 7, 9-10 & 12 (Oct. 30, 2014).

EPA emphasized that adequate monitoring is a crucial component of an acceptable SIP, id. at 12 (“Implementation includes adequate monitoring, which must be in the SIP.”), and that stack testing once every three to five years is, on its face, inadequate to show continuous compliance, id. at 9-10 (“We are concerned with stack test frequencies longer than one year. Please explain why these test frequencies are sufficient to ensure continuous compliance with the limits.”), and requested that the Director explain why the specified monitoring was adequate to support modeling, establish RACT and demonstrate attainment. Id. at 7 (“[W]e suggest that the UDAQ…clarify and provide more detail…in SIP sections and/or RACT evaluations” to explain “how and why…frequency of monitoring/ testing…(continuous, daily, monthly, etc. for monitoring; once per year, 3 years, 5 years for stack testing)…[is] considered valid to support modeling and attainment”). See also id. at 28 (“Details on [stack test frequency] should be provided in the SIP, and should provide details on how it is a part of a control strategy that is adequate for timely attainment of the PM2.5 NAAQS.”).

Utah’s current BACT analysis fails address EPA’s concerns and to include consideration of the adequacy of the monitoring provisions associated with the particular Subpart H emission limitations. Plainly, where Utah must derive BACT, a more rigorous mandate than RACT, the state must do even more to ensure the Subpart H monitoring requirements reflect BACT and ensure continuous compliance. Indeed, in many instances Utah proposes to require stack testing as infrequently as once per every three years and sometimes once every five years. At the same time, the State fails to establish how such infrequent stack testing can ensure continuous compliance with the Subpart H emission limitations and so meet the requirements of BACT. As a result, Utah’s BACT analysis and Subpart H are not adequate.

Without a case-by-case analysis and supporting data and well into the development of the SIP, Utah reduced stack test frequency on many emission units from once every year to once every three years. DAQ Response to Comments at 3. As a result, DAQ cannot, as is required, ensure that the relevant sources are complying with SIP emission limitations on a continuous basis. DAQ’s decision to amend the SIP fails to address EPA’s concerns that infrequent monitoring is inadequate to show continuous compliance with SIP emission limitations. This is particularly true because DAQ has failed to provide the analysis necessary to support its SIP revision and to ensure monitoring frequent enough to demonstrate continuous compliance.

Examples of IX.H.11, 12 and 13 monitoring requirements that rely on stack tests of three or more years include: 1) IX.H.11.g.i.B.II (general requirement applicable to petroleum refineries requiring a stack test only once every three years to monitor PM emissions from the FCCU); 2) IX.H.12.a.iii.B.II.c (ATK stack test once every three years to determine compliance with a NOX emission limit); 3) IX.H.12.b.i.B (stack test frequency once every three years to show compliance with PM2.5 emission limit on the Big West FCC); 4) IX.H.12.d.i.B (monitoring to show compliance with PM2.5 emission limit on the Chevron FCCU limited to stack test once every three years); 5) IX.H.12.d.v.B (monitoring to show compliance with NOX limit on Chevron compressor engine limited to stack test once every three years); 6) IX.H.12.e.iii.A (monitoring to show compliance with VOC emission limit on Compass Mineral’s magnesium chloride evaporators limited to stack test once every three years); 7) IX.H.12.f.iv.B (monitoring to show compliance with NOX emission limit on Hexcel’s ultra low NOX burner limited to stack test once every three years); 8) IX.H.12.g.i.B (Holly); 9) IX.H.12.h.i.C (KUC); 10) IX.H.12.j.i.B (Holman boiler); 11) IX.H.12.j.ii.B (tankhouse boiler); 12) IX.H.12.m.i.B (Tesoro); 13) IX.H.12.m.ii.B (Tesoro); 14) IX.H.12.n.i.B (Procter and Gamble); 15) IX.H.12.p.ii (University of Utah); 16) IX.H.13.a (Brigham Young); 17) IX.H.13.c (McWane Ductile); 18) IX.H.13.e (Payson City); and, 19) IX.H.13.f (Provo City).

f. Utah Has Failed to Demonstrate Why CEMS or CPMS Is Not Feasible.

In several instances, the Director has not required CEMS or at a minimum CPMS, to ensure compliance with Subpart H emission limits, although CEMS is a feasible method for monitoring emissions of PM2.5, SO2, and NOX. EPA is clear that directly enforceable emission monitoring is preferable wherever feasible. 81 Fed. Reg. at 58133 (“Directly enforceable emission measurements, such as PM CEMS, are preferred wherever feasible.”). Utah has failed to show why CEMS is not feasible. After all, CEMS is a critical element of a BACT emission limit and must reflect the maximum degree of emission reductions. Where CEMS has been determined to be feasible in analogous situations and has been applied as BACT, the Director is required to adopt this monitoring requirement or explain why CEMS does not constitute BACT and why alternative monitoring methods are adequate to ensure continuous compliance with the corresponding Subpart H emission limitation.

Specific examples of the failure of Utah to explain why CEMS or CPMS is not BACT include:

1) IX.H.11.g.i.B.III (general requirements for oil refineries); 2) IX.H.12.b.i.B (monitoring to show compliance with PM2.5 emission limit on the Big West FCC limited to stack test once every three years); 3) IX.H.12.c.ii.c (Chemical Lime Company – PM2.5 emission limit); 4) IX.H.12.d.i.B (monitoring to show compliance with PM2.5 emission limit on the Chevron FCCU limited to stack test once every three years); 5) IX.H.12.d.v.B (monitoring to show compliance with NOX limit on Chevron compressor engine limited to stack test once every three years); 6) IX.H.12.g.i.B (Holly); 7) IX.H.12.h.i.C (KUC); 8) IX.H.12.j.i.B (Holman boiler); 9) IX.H.12.j.ii.B (tankhouse boiler); 10) IX.H.12,m.i.B (Tesoro); and, 11) IX.H.12.m.ii.B (Tesoro)

g. The IX.H.11.g.i.B.III Provision Is Not BACT.

The SIP provision IX.H.11.g.i.B.III states that:

No later than January 1, 2019, each owner or operator of an FCCU subject to NSPS Ja shall install, operate and maintain a continuous parameter monitor system (CPMS) to measure and record operating parameters from the FCCU and control devices as per the requirements of 40 CFR 60.105a(b)(1). No later than January 1, 2019, each owner or operator of an FCCU not subject to NSPS Ja shall install, operate and maintain a continuous opacity monitoring system to measure and record opacity from the FCCU as per the requirements of 40 CFR 63.1572(b) and comply with the opacity limitation as per the requirements of Table 7 to Subpart UUU of Part 63.

Yet, there is no evidence in the record to suggest why BACT for a facility subject to Subpart Ja is different from BACT for a facility not subject to Subpart Ja. Indeed, given that BACT requires at least one oil refinery in Salt Lake City to maintain a CPMS to “to measure and record operating parameters from the FCCU and control devices,” indicates that it is appropriate for all the oil refineries to maintain a CPMS. This is particularly true because the record contains no explanation as to why CPMS is not BACT for all Salt Lake City refineries.

3. The Director’s BACT/BACM18 Analysis for Fugitive Dust Emissions is Inadequate.

BACT/BACM applies equally to sources of fugitive dust.19 However, Utah’s review for sources of fugitive dust is not legally sufficient. Initially, there is nothing in the record to show that Utah objectively evaluated all the factors outlined in EPA’s five-step, top-down method. Sierra Club, ¶ 4. fn.2 (citing NRS Manual B.6-9). In addressing fugitive dust, the record does not contain adequate documentation, a list of available technologies, a ranking of controls based on their effectiveness and efficiency and achievable emission rates and reductions or the consideration of their economic or environmental impact. Id. Utah’s analysis fails to establish that the controls he adopted as BACT for fugitive dust reflect “the maximum degree of emission reduction achievable… considering energy, economic and environmental impacts and other costs.” 81 Fed. Reg. at 58081. As a result, there is no record of evidence to show that Utah undertook a defensible BACT analysis that actually represents the most stringent technology and the maximum reduction of emissions achievable. Utah Admin. Code r.307-401-2; Sierra Club ¶¶ 4, fn.2, 47-48.

Utah has not provided sufficient evidence or analysis to support his contention that Utah’s Fugitive Emissions Rule, Utah Admin. Code R307-309, is BACM. Utah claims, based on mere assertion, that “[t]here are no current opportunities for additional program revisions that would lead to further emission reductions.” Provo BACM at unnumbered 8. Utah seems to claim, without any supporting data or explanation, that the Utah rule is the most stringent rule in the nation, or as stringent as any rule in the country because it regulates sources as small as onequarter acre and implements “BMPs.” Provo BACM at unnumbered 7-9.

This is not adequate to determine BACM or support a BACM conclusion. Rather, Utah is required to assess the emission reductions achieved by the various state rules and compare those reductions to the emission reductions achieved by the Utah rule. Further, to justify his BACM conclusion regarding “BMPs,” Utah must actually compare those BMPs to Utah’s rule. Finally, the BACM analysis is incomplete because it “summarizes” Utah’s review without actually providing that analysis, leaving the public in the dark as to the Director’s thinking.

Further, other states have adopted regulations that are more stringent than the Utah rule. These rules also meet the further BACM mandates of enforceability, recordkeeping and reporting, while the Utah rule is not enforceable and lacks recordkeeping and reporting requirements sufficient to meet BACM.

a. Other States Reduce Fugitive Emissions to a Greater Degree and Otherwise Meet the Requirements of BACM.

The following rules are more stringent than Utah’s rule because they impose more strict emission limits on sources of fugitive dust:

• California’s South Coast Air Quality Management District prohibits visible dust beyond the property line of the source and requires BACM. Rule 403(d)(1)(A) and (d)(2).20
• Maricopa County prohibits visible dust beyond the property line of the source. Rule 310 at 303.1(a).21
• Clark County, Nevada prohibits a dust plume that extends 100 yards or more, horizontally or vertically, from the point of origin and requires best available control measures. Rule 94.11.2 and 94.9.22
• Washoe County, Nevada prohibits visible fugitive dust emissions lasting more than 5 minutes in any hour. Rule 040.030, Section C.23

Each of these rules is more stringent than the Utah rule as it puts a stricter emission limit on fugitive dust. Each rule also complies with BACM because it is enforceable. Therefore, the Director must adopt the most stringent of these rules or provide a defensible reason for not doing so. The Director must also adopt a fugitive emissions rule that requires recordkeeping and reporting sufficient to allow the public to monitor and enforce compliance with the rule.

Furthermore, alternatives for limiting fugitive dust that are enforceable and include adequate monitoring, recordkeeping and reporting requirements could be adapted for Utah.

• On the rare instances that DAQ monitors the 10% opacity limit, it employs EPA Method Method 9 includes “Alternative Method 1,” which EPA states “provides the quantitative determination of the opacity of an emissions plume remotely by a mobile lidar system (laser radar; Light Detection and Ranging).”24 Indeed, because lidar uses its own lighting system, it can monitor opacity both during the day and at night.25 Lidar could be employed at sources with the potential to emit significant amounts of fugitive dust.
• The South Coast Air Quality Management District’s Rule 403 protects the public from excessive fugitive dust emissions caused by man-made activities, including those occurring on mining sites and at gravel pits. Rule 403 requires downwind PM10 levels to be less than 50 μg/m3 over upwind concentrations. This rule, which has been in effect for decades, is enforceable.
• South Coast has a list of approved PM10 monitors sanctioned to enforce a regulation (Rule 1466) that limits emissions of fugitive dust. South Coast estimated the cost for two T640 model monitors with the 640X option (one upwind and one downwind) per site at $45,703 per monitor, based on a price quote from a local supplier.26 The agency concluded: “This would result in a cost of about $91,406 for each site that does not already use PM10 monitors.”27
• According to DAQ “South Coast, Washoe, Clark and Maricopa Counties have stringent opacity requirements ranging from no visible dust at property boundary, no visible dust any time, to limiting visible dust to five minutes within an hour.” These requirements would be more readily enforceable.
• These alternatives are exactly the type of information that should be considered during an open, public process aimed at ensuring that Utah’s Fugitive Dust Rule is effective and enforceable and that the public can determine whether a source is complying with the rule’s provisions.

b. DAQ’s Responses to Comments Underscore that the Fugitive Dust Rule is Not BACT/BACM.

DAQ tries to justify the Utah Fugitive Dust Rule further in its response to public comment, which is part of the docket for the present matter. Without providing an adequate BACM analysis, DAQ perfunctorily rejects more rigorous controls adopted by other states to control fugitive dust emissions. See WRA Subpart A Comments at 11 & attached Provo SIP Comments at 4-5.28 DAQ’s reasons for doing so are not adequate to support its contention that Rule 307-309 is indeed BACM. As a result, EPA should request that DAQ thoroughly evaluate and where appropriate adopt the more effective dust control measures adopted in other states.

EPA has underscored that measures adopted in other nonattainment areas are assumed to be BACM: “[O]ther states across the country are important sources of information about control measures.” 81 Fed. Reg. at 58084. This is true even if those measures are costly. Id at 58087 (“[A]ll states with Serious areas need to consider implementing BACM and BACT-level measures that have been implemented in other states, even if those measures incur higher costs.”).

Indeed, a measure adopted in another state is presumed to be BACM and “where a given control measure has been applied in another NAAQS nonattainment area (for PM2.5 or other pollutant), the state will need to provide a detailed justification for rejecting any potential BACM measure as technologically infeasible.” Id. at 58085. Yet, DAQ has not met this burden.

Instead, DAQ argues first that:

UDAQ conducted its RACM/BACM analysis by reviewing other state rules, referencing the WRAP Fugitive Dust Handbook and using our engineering knowledge. We evaluated every conceivable earth disturbance mechanism that would generate fugitive dust and developed best management practice options for each mechanism in a feasible and cost effective manner.

DAQ Response to Comments at 31-32. However, this analysis has not been made part of the record supporting the SIP. As a result, neither the public nor EPA has been given the opportunity to review or comment on this purported review. Rather, DAQ’s analysis of fugitive dust BACM for the purposes of the SIP is limited to two pages comprised chiefly of charts that does not convey an evaluation of “every conceivable earth disturbance mechanism” or otherwise constitute an adequate BACM analysis. Salt Lake NAA BACM at unnumbered 8-9.29

Further, while DAQ states that “[w]e have reviewed those mechanisms and have determined that they are for the most part not feasible, not cost effective or more importantly, would not result in advancing the airshed towards attainment,” DAQ Response to Comments at 31, DAQ does not provide any reasons why rules adopted in other states are not feasible in Utah, what the costs of these rules might be or how effective they are in reducing emissions. As WRA noted in its Subpart A Comments at unnumbered 11, DAQ’s analysis of Rule 307-309 does not meet the requirements of adequate BACM for the purposes of a serious SIP.

Second, DAQ states that “Maricopa County has a more stringent applicability requirement than R307-309 because it is a moderate PM10 nonattainment area.” Id. However, while reasonably available control measures (RACM) applies to moderate nonattainment areas, BACT, a more stringent requirement, applies to the serious Salt Lake NAA. Therefore, absent a rigorous analysis, DAQ is compelled to adopt a rule as rigorous or more rigorous than the Maricopa rule.

Third, DAQ states that “[t]he winter time PM2.5 contribution from disturbance of frozen earth is low, therefore it is not appropriate to compare the Maricopa County rule or other state rules whose rules are designed to address PM10.” Id. However, as explained above, BACT and BACM are generally independent of the attainment needs of the area and “BACM/BACT measures for Serious areas are not solely limited to those measures needed for expeditious attainment under this final rule.” 81 Fed. Reg. at 58020. Therefore, DAQ’s reasoning is flawed.30

Fourth, DAQ states that “[i]t is inappropriate to blindly say that this or any other PM10 rule can be directly compared to a winter time PM2.5 fugitive dust emission.” DAQ Response to Comments at 31. However, EPA disagrees, stating that “where a given control measure has been applied in another NAAQS nonattainment area (for PM2.5 or other pollutant), the state will need to provide a detailed justification for rejecting any potential BACM measure as technologically infeasible.” 81 Fed. Reg. at 58085 (emphasis added). Therefore, DAQ is required to provide a detailed justification for rejecting California’s South Coast Air Quality Management District, which prohibits visible dust beyond the property line of the source and requires BACM, Rule 403(d)(1)(A) and (d)(2), even if that rule applies to another pollutant.

Fifth, DAQ states “WRA has not also stated in their comment that Rule 303 provides the option of either no visible dust beyond the property boundary OR 20% opacity.” DAQ Response to Comments at 32. However, DAQ ignores that other state rules cited by WRA do include a “no visible dust” requirement. For example, in a document it submitted to EPA to support Rule 309-307 as RACT for the PM10 Maintenance Plan, DAQ states: “South Coast, Washoe, Clark and Maricopa Counties have stringent opacity requirements ranging from no visible dust at property boundary, no visible dust any time, to limiting visible dust to five minutes within an hour. Impetus: serious nonattainment.” WRA Comments on Docket Number EPA-R08-OAR-2017-0469-0001: Air Quality State Implementation Plans; Approvals and Promulgations: State of Utah; Revisions to Utah Division of Administrative Rules, R307-300 Series; Area Source Rule for Attainment of Fine Particulate Matter Standards, attached to WRA Comment on Subpart A at 7.31 Moreover, the fact that the “no visible dust” provision is an element of a RACM rule indicates that consideration of such a control is mandated under BACM.

Sixth, DAQ states “Clark County is a serious PM10 nonattainment area. It is inappropriate to compare rule designed for a serious PM10 nonattainment to a winter time PM2.5 fugitive dust rule.” DAQ Response to Comments at 32. However, as established elsewhere, BACM is generally independent of the attainment needs of the area and “where a given control measure has been applied in another NAAQS nonattainment area (for PM2.5 or other pollutant), the state will need to provide a detailed justification for rejecting any potential BACM measure as technologically infeasible.” 81 Fed. Reg. at 58085 (emphasis added).

Seventh, DAQ states “Washoe County Rule 040.030, Section C applies to dust generating activities of 1 acre and greater which is less stringent than R307-309’s ¼ acre. Clark County, Nevada dust generating prohibitions are designed for their serious PM10 nonattainment status.” Id. However, without further analysis, DAQ cannot state that the Washoe County Rule with its more stringent measures applied to dust generating activities of 1 acre and greater does not do more to control emissions than Rule 307-309. Moreover, it is incumbent on DAQ, as WRA has repeatedly commented, to consider applying stricter measures to larger sources of fugitive dust.

Finally, DAQ states that “WRA has not substantiated their positions quantitative evidence that their recommendations would move the attainment needle forwards nor provided a cost-benefit analysis. Meanwhile, UDAQ has determined that R307-309 provides a control efficiency of 32% at a cost of $2,140/acre (Area Source TSD) which has not been challenged by EPA to date.” DAQ Response to Comments at 31. However, WRA could not find this control efficiency or cost in the referenced document and could not locate any analysis or data to support such a conclusion. Moreover, the burden in not on WRA to show that the rules adopted by other states would advance attainment or to provide a cost-benefit analyses. Rather, as established elsewhere, BACM is generally independent of the attainment needs of the area and “where a given control measure has been applied in another NAAQS nonattainment area (for PM2.5 or other pollutant), the state will need to provide a detailed justification for rejecting any potential BACM measure as technologically infeasible.” 81 Fed. Reg. at 58085 (emphasis added).

c. Utah’s Fugitive Dust Rule is Not BACT/BACM Because it Is Not Enforceable and Lacks Adequate Monitoring, Recordkeeping and Reporting Requirements.

Rule 309 relies on a scheme that allows sources to retain possession of records establishing compliance with the rule’s emission limit and operating practice. As a result, the public is unlawfully prevented from securing access to the bulk of the records that purport to establish that the relevant sources are complying with the SIP. The Director has successfully maintained in Court that the public does not have access to compliance records retained by the source. Thus, Rule 309 is not enforceable by the public and otherwise violated the Clean Air Act.

To be legally defensible, SIP emission limitations and measures must be enforceable by the public. Without direct access to records kept to show compliance with SIP emission limitations and measures, the public is prevented from enforcing these provisions in a meaningful way. It is plain that the only way the public may enforce SIP emission limitations and operating practices is to be able to review the records maintained by the source that purport to show compliance with those limits and practices. Because – based on the Director’s own analysis – the current version of Rule 309 does not allow meaningful enforcement by the public, the proposed subsection is unlawful and should be rejected.

The importance of assuring that the public has immediate and direct access to compliance records has been made particularly evident. Residents nearby are concerned by the giant dust plumes that they see leaving the property of the gravel pits and believe that the gravel pits may not be complying with Rule 309. As with the proposed rule, Rule 309 then included the following provision:

All sources subject to R307-309-5(2) and (3) shall maintain records demonstrating compliance with R307-309. These records shall be available to the director upon request.

Utah Admin. Code R307-309-12 (emphasis added). Both the Point of the Mountain gravel pits –

Geneva and Staker Parsons – are subject to R307-309-5(2) and (3). When we were denied access to these R307-309-12 documents, both by the Division and the Director of the Division, we appealed and the Third District Court upheld the Director. Because the denial of access has been upheld, it is incumbent on the Division to amend Rule 309 to require periodic compliance reports and to allow the public access to compliance documents.

It is important to note that the Rule 309 is an important element of nonattainment and maintenance SIPs and so is subject to the same enforceability requirements set forth above. It is also significant that, based on our review of documents that are available to the public, neither Geneva or Staker Parsons has submitted “periodic” Rule 309 compliance reports to the Director as required by law. When inspections are so infrequent, compliance records take on added importance. Review of these records is the only means by which the public can determine if the companies are complying with Rule 309 on a continuous basis, as the law requires.

In sum then, the proposed Rule 309 relies on a recordkeeping scheme that allows sources to maintain compliance records. The public has no right to review these records directly. Therefore, the public is prevented from reviewing the very records the SIP relies on to show whether the source is complying with SIP emission limits and operating practices. As a result, the public is barred from enforcing these limits and practices and Rule 309 is unlawful.

d. The Way that Utah’s Fugitive Dust Rule Relies on Opacity is Not BACT/BACM.

As discussed in more detail below, Rule 309’s reliance on opacity to limit fugitive dust and fugitive emissions is insufficient to meet the requirements of BACM/BACT and otherwise protect public health.

First, the Rule 309 fails to include any provision specifying how frequently monitoring of the opacity emission limit must occur and so violates the enforceability requirements of an adequate SIP measure. Our experience has been that the Division monitors opacity at Geneva and Staker Parsons once every two years. As the Division acknowledges, this monitoring shows only whether the source was in compliance with the limit at the time the reading was made and in no way establishes that the source is in continuous compliance.

More specifically:

• Opacity cannot be measured at night or when visibility is poor. Therefore, Rule 309 is not enforceable. Conditions that create fugitive emissions can and do occur at night and when visibility is impaired.
• Rule 309 does not specify if and how often opacity monitoring should occur and therefore is not enforceable. This is of particular importance because, for example, fugitive dust emissions rates are highly variable and periodic monitoring is insufficient to ensure continuous compliance with the opacity limits.
• Rule 309 does not require monitoring, recordkeeping or reporting of opacity emissions and therefore does not meet the requirements of BACM/BACT. R307-309-12 does require records demonstrating compliance with the rule generally. However, there are no specific monitoring requirements in the rule. Therefore, the rule does not require recordkeeping or reporting for any particular emission limit, control or practice. Indeed, it is unclear what R307-309-12 does require.
• Any monitoring for opacity will necessarily fall well short of the continuous monitoring favored by the Clean Air Act.
• For the reasons stated above, the public is prevented from determining whether the opacity limits set forth in Rule 309 are being met and is therefore denied the opportunity to enforce the rule relative to any source.

e. Utah’s Fugitive Dust Rule Fails to Prevent Dust Storms from Blanketing Neighboring Communities.

Utah’s Fugitive Dust Rule purports to limit emissions of fugitive dust by limiting emissions to 0% opacity at a source’s property boundary. Opacity is the degree to which visibility of a background, such as blue sky, is reduced by particulates – in this case dust. This limit does not apply when wind speeds exceed 25 mph. In principle, the opacity limit should mean that under most weather conditions, very little dust should be leaving a gravel pit or mining source. However, nearby communities report frequent dust storms leaving the property boundaries of the ravel pits and other sources, including when winds are less than 25 mph.

There is no meaningful enforcement of the Fugitive Dust Rule. As discussed elsewhere, monitoring of the Fugitive Dust Rule’s 10% opacity limit at the gravel pits at Point of the mountain rarely happens. This means that neither the public, the state, nor even the source knows whether this limit is being met on a continuous or even regular basis.

Sources are not held accountable for their compliance with the Fugitive Dust Rule. Also as discussed above, the Fugitive Dust Rule requires sources to keep records showing “compliance” with rule’s requirements, but these records are not available to the public for review. The sources do not monitor and therefore do not keep records showing compliance with the 10% opacity imit. The “compliance” records the sources keep document the application of water and, if used, other dust suppression efforts. However, these records do not show whether or not dust is leaving the property boundary or the facility is meeting the opacity limit. Further, the records do ot contain enough information to show whether the sources is meeting its other 307-309 bligations.32 This means that neither the public, the state, nor even the source knows whether he opacity limit is being met on a continuous or even regular basis.

Similarly, DAQ is not being held accountable for its administration of the Fugitive Dust Rule. AQ almost never monitors the opacity limit. Sources do not monitor the opacity limit. The compliance” records the sources maintain do not contain information that would allow either he public or the state to determine if the sources are meeting their obligations under the Fugitive Dust Rule. The public has no access to the records the sources maintain. As a result, the public has no way to hold DAQ accountable for its administration of the Fugitive Dust Rule.

A fugitive dust rule that is actually BACM/BACT would include: 1) an emission limitation that can be monitored on a continuous or regular basis; 2) monitoring, recordkeeping and reporting requirements; and, 3) a guarantee that the public will have access to meaningful monitoring reports. Because the Utah rule does not contain these components, it is not BACM/BACT.

f. Utah Admin. Code R307-309-4, which Addresses Fugitive Emissions, is Not BACM/BACT.

• There are no monitoring (other than the method to use), recordkeeping or reporting requirements associated with fugitive emissions. Therefore, R307-309-4 does not meet the requirements for BACM/BACT. Without specifying monitoring frequency – particularly for sources that are not intermittent or mobile – R307-309-4 cannot ensure continuous compliance with the 15% opacity limit.
• R307-309-12 does require records demonstrating compliance with the rule generally. However, there are no specific monitoring requirements in the rule. Therefore, the rule does not require recordkeeping or reporting for any particular emission limit, control or practice.
• R307-309-4 does not specify if and how often opacity monitoring should occur and therefore is not enforceable.
• By apparently relying on opacity to limit and monitor fugitive emissions, R307-309-4 fails to address non-visible fugitive emissions or to limit or require monitoring of fugitive emissions at night or when visibility is poor.
• R307-309-4 does not adopt BACM/BACT, such as requirements that sources regularly look for and stop any emission leaks or that they capture and control fugitive emissions, or demonstrate that such measures are not reasonable.

g. Utah Admin. Code R307-309-5 Relating to General Requirements for Fugitive Dust isNot BACT/BACM.

• There is no demonstration that the measures of R307-309-5 are BACM/BACT. In light of measures adopted elsewhere, including in Southern California, it is apparent that additional controls exist that are technologically and economically feasible, particularly for sources of significant, long-term fugitive emissions.
• There is no reason that all sources of fugitive dust should be subject to identical control measures. What is BACM/BACT for some sources may not be BACM/BACT for smaller, more temporary sources. Pursuant to his SIP obligations, the Director must consider more stringent emission limits for sources of more significant emissions.
• R307-309-5 is not enforceable and any emission reductions it achieves are not measurable because it lacks monitoring, recordkeeping and reporting requirements. R307-309-12 does require records demonstrating compliance with the rule generally. However, there are no specific monitoring requirements in the rule. Therefore, the rule does not require recordkeeping or reporting for any particular emission limit, control or practice.
• Please see the general comments above regarding opacity.

h. Utah Admin. Code R307-309-5(3), Relating to Conditions when Wind Speeds Are Higher, is Not BACM/BACT.

R307-309-5(3) is particularly problematic, as the adverse health impacts of fugitive emissions are just as great when the wind blows as when it does not and because the concentrations of PM2.5 during wind events can be dangerously high. Based on this concern, please consider the following:

• R307-309-5(3) requires a source to “administer” one or more of various contingency measures when the wind is stronger than 25 mph. These measures are not equally effective and it should not be left to the source to decide which measure to adopt, regardless of the size or type of source. For example, the rule would authorize a source of significant fugitive dust to undertake “pre-event” watering as sufficient to reduce emissions during a wind event, regardless of whether this is an effective measure. Wind events may last hours and days and therefore it is unreasonable to suggest that a one-time watering of, for example a gravel pit operation, is an adequate contingency measure. During an extended wind event – one that is longer than an hour for example – “pre-event watering” would excuse the source from undertaking hourly water. Yet, there is no reason to believe that watering less frequently than every hour would be sufficient to control fugitive dust emissions to any degree.
• The use of the word “watering” is vague and unenforceable and cannot be monitored. “Watering” in this context could mean anything from moistening or sprinkling to drenching. “Watering” should be defined.
• The use of the word “additional” is vague and unenforceable and cannot be monitored. “Additional” in this context could mean almost anything from hardly any additional stabilization to considerable additional stabilization. “Additional” should be defined.
• The use of the phrase “to the extent practicable” to describe the cessation or reduction of activities is vague and unenforceable and cannot be monitored. This phrase could mean almost anything and its application could result in no change in operations, should the source deem that it would not be practicable for it to cease or reduce its dust producing activities.
• R307-309-5(3)(d) is open ended, vague and subject to abuse in stating that the opacity limit shall not apply if the source continues to implement the dust control plan and the source has “cease[d] or reduce[d] fugitive dust producing operations.” This language suggests that any reduction in these operations, including a minor reduction, would qualify the source for an exemption from emission limits. Rather, the rule should require all dust producing operations to be stopped during a wind event. Moreover, in some instances, the mere cessation of dust producing activities will not guarantee that emissions will be adequately controlled. This is particularly true with ground disturbing activities.
• R307-309-5(5)(a) again underscores that Rule 309 fails to address nonvisible fugitive emissions, including those from vehicles.

i. Utah Admin. Code R307-309-6, which Addresses Fugitive Dust Control Plans, Is Not BACM/BACT.

R307-309-6(1)(a) is particularly problematic. This provision seems to suggest that regardless of the content of the pre-2012 dust plan and regardless of whether it is in keeping with R307-309-6, it is BACM/BACT and is adequate to meet the requirements of Rule 309. This provision makes Rule 309 unenforceable and means that the rule does not reflect BACM/BACT and therefore does not meet Clean Air Act requirements for a Moderate SIP control measure.

While fugitive dust plan forms may be appropriate for small sources of fugitive emissions, there is no reason to believe that such forms are adequate to ensure that all sources have implemented all BACM/BACT to control fugitive emissions.

R307-309-6(4) is vague. Particularly, there is no criteria associated with the term “stabilize.” What “stabilize” means is apparently left to the source to determine. The rule should define “stabilize,” including as an outcome – in other words, describe what stabilize is to achieve. This is particularly true because “stabilize” is used in the context of a whole host of different activities. Similarly, the terms “limit” and “minimize” and the phrase “where possible” are vague and not enforceable.

R307-309-6(4)(e)(iii) should be edited to include a verb, such as “use” or “implement.”

R307-309-6 is not enforceable and fails to require monitoring, recordkeeping and reporting. R307-309-12 does require records demonstrating compliance with the rule generally. However, there are no specific monitoring requirements in the rule. Therefore, the rule does not require recordkeeping or reporting for any particular emission limit, control or practice.

j. Utah Admin. Code R307-309-7 thru R307-309-9 Are Not BACT.

• The terms “promptly” and “clean” are vague and are not subject to enforcement.
•  R307-309-7 and -8 should apply to unpaved roads as well.
•  R307-309-9(1): the phrase “possessing the right to use” is vague. It is unclear whether this phrase would include use of any public road. This provision should also apply to persons using public unpaved roads.
•  R307-309-9(2): this provision should also apply to persons using public unpaved roads.

k. Utah Admin. Code R307-309-10, which Addresses Mining Activities, Is Not BACM/BACT.

• It appears inconsistent that mining activities are not subject to dust control plans.
• R307-309-10(3) is vague and fails to require monitoring, recordkeeping or reporting and is not enforceable. R307-309-12 does require records demonstrating compliance with the rule generally. However, there are no specific monitoring requirements in the rule. Therefore, the rule does not require recordkeeping or reporting for any particular emission limit, control or practice.
• R307-309-10(3) requires a source to “include” one or more of measures to control dust. The rule includes the confusing use of “or.” These measures are not equally effective and it should not be left to the source to decide which measure to adopt, regardless of the size or type of source. As written and given the use of the word “include,” it appears that the rule allows a mining operator to implement only one of the various measures described in (3)(a) to (3)(q).
• The terms and phrases “periodic,” “minimizing,” “reducing,” “restricting,” “prompt” and the “use of chemical stabilizers” are vague and are not enforceable.

l. Utah Admin. Code R307-309-11, which Addresses Tailings Piles and Ponds, In Not BACM/BACT.

• It appears inconsistent that tailings piles and ponds are not subject to dust control plans.
• R307-309-11(2) is vague and fails to require monitoring, recordkeeping or reporting and is not enforceable. R307-309-12 does require records demonstrating compliance with the rule generally. However, there are no specific monitoring requirements in the rule. Therefore, the rule does not require recordkeeping or reporting for any particular emission limit, control or practice.
• R307-309-11(2) requires a source to “include” one or more of measures to control dust. The rule includes the confusing use of “or.” These measures are not equally effective and it should not be left to the source to decide which measure to adopt. As written and given the use of the word “include,” it appears that the rule allows aa operator to implement only one of the various measures described in (2)(a) to (2)(i).
• The terms and phrases used in R307-309-11(2) are vague and are not enforceable.

m. Utah Admin. Code R307-309-12, which Addresses Recordkeeping, Is Not BACM/BACT.

• R307-309-12 does require records demonstrating compliance with the rule generally. However, there are no specific monitoring requirements in the rule. Therefore, the rule does not require recordkeeping or reporting for any particular emission limit, control or practice.
• It is not clear what R307-309-12 does require. This provision is vague and therefore is not enforceable.
• R307-309-12 is unlawful because it fails to allow public access to compliance documents.

n. Other Aspects of Rule 309 Are Not BACM/BACT.

• The rule should clarify that it applies at all times, including during when operations may have ceased, such as after work, weekends and holidays, and specify that R307-309 applies and emissions must be controlled and monitored at all times.
• As they are an important component of the proposed maintenance plan, fugitive dust plans forms should be subject to public notice and comment.
• The control measures adopted for mining operations and tailings piles and ponds should be subject to public notice and comment, should reflect BACM/BACT, should be enforceable and should require monitoring, reporting and recordkeeping. These measures should include source-specific emission limitations that are enforceable and include sufficiently frequent monitoring to ensure continuous compliance.
• The rule should establish that a source must comply with mandated practices or plans until the source has formally notified the Director that all fugitive emissions and emission generating activities have permanently ceased.

Based on the above, EPA is compelled to reject Utah’s Rule 309 as not meeting the requirements of BACM/BACT.

4. For Many Specific Sources, Utah’s BACT Analysis and Derivation and Imposition of BACT Limits Is Inadequate.

A state is required to determine and implement BACM including BACT on sources of direct PM2.5 emissions and PM2.5 precursors located in any area designated as a serious PM2.5 nonattainment area. 40 C.F.R. §51.1010(a). The state must identify the most stringent measures for such sources required in any SIP or used in practice to control emissions in any state. Further, the state must reconsider and reassess any measures previously rejected by the state in developing the moderate area SIP. 40 C.F.R. § 51.1010(b)(2)(i) and (ii). The state may make a demonstration that any measure identified is not technically or economically feasible to implement in whole or in part by 5 years after the applicable attainment date for the area. 40 C.F.R. §51.1010(b)(3). EPA has stated that there is no “de minimis” thresholds for determining applicability of BACM/BACT, that BACT/BACT are generally independent of what is needed to attain the NAAQS (in contrast to reasonable available control measures (RACM).” 81 Fed.Reg. 58,080-83 (Aug. 24, 2016).

Comments on the proposed BACT33 evaluations for selected sources and emissions units or groups within those sources are provided below.

a. BACT Has Not Been Properly Applied to Rio Tinto Kennecott Sources.

The following provides comments on DAQ’s proposed determination of BACT for the Rio Tinto Kennecott Utah Copper (KUC) facilities,34 as well as on the company’s submitted BACT analysis.35 In general, both DAQ and KUC failed to provide documentation to verify that the most stringent measures for reducing PM2.5 or PM2.5 precursor emissions adopted in any SIP or used in practice were considered. Further, KUC failed to rank control technologies from most effective (or lowest achievable emission rate) to least effective.

i. BACT for PM2.5 from In-Pit Crusher at Bingham Canyon Mine

DAQ and KUC identified fabric filters as the top control technology for direct PM2.5 emissions from crushing operations.36 According to KUC’s April 2017 BACT submittal, the in- pit crusher at KUC is already equipped with a baghouse and is subject to a 0.002 grains per dry standard cubic foot (gr/dscf) limit and that it was established by UDAQ as BACT for the BCM in 2011.37 However, DAQ’s June 2018 BACT analysis states that the KUC crushing operations are currently permitted at 1.77 pounds per hour (lb/hr) and a significantly higher grain loading of 0.016 gr/dscf, eight times higher than the 0.002 gr/dscf limit that KUC claimed applied in its 2017 BACT analysis.38 A review of a recently issued Approval Order issued for KUC BCM and Copperton Concentrator in January of 2018 does indeed show a PM10 limit for the main in-pit crusher baghouse of 1.77 lb/hr and 0.0016 gr/dscf.39

In any event, DAQ and KUC proposed a BACT emission limit of 0.002 gr/dscf and DAQ also proposed a BACT limit of 0.78 lb PM2.5 per hour.40 A review of the Californian Air Resources Board (CARB) BACT Clearinghouse shows lower PM2.5 emission limits have been required for similar sources and controls. Specifically, the PM10 emission limit at the Rio Rock Materials, Inc. crushing and screening operation is 0.0012 gr/dscf with a baghouse.41 Indeed, even DAQ’s BACT analysis shows that the KUC BCM crusher’s emission rates has been significantly lower than DAQ’s proposed BACT limits. Stack test results at KUC’s BCM crusher from 2000 through 2015 show that the highest PM2.5 emission rates from the in-pit crusher were measured at 0.164 lb/hr and 0.001 gr/dscf. Thus, there is ample support for a lower PM2.5 BACT emission limit on both a lb/hr and a gr/dscf basis at the in-pit crusher. DAQ should impose lower limits that truly reflect the maximum degree of PM2.5 emission reduction that can be achieved with a baghouse at the in-pit crusher. In addition, an opacity limit reflective of BACT must be imposed as a measure to ensure continuous compliance with emission limits and proper operation and maintenance of the baghouse.

ii. BACT for Waste Rock Offloading from Trucks at Bingham Canyon Mine.

DAQ and KUC identified water applications, enclosures, and minimizing the drop distance as control technologies from Waste Rock offloading.42 UDAQ and KUC eliminated enclosures as a technically infeasible because the drop location is not static. Id. DAQ also determined that water application is not technically feasible because “excessive water application would create geotechnical instability on the waste rock dumps” and because “installation or setup of a water irrigation system for water application is not technically feasible because the drop location is not static.” Id. Due to its finding that other measures were technically infeasible, DAQ decided that minimizing the drop distance while the waste rock is being dumped is BACT. Id. DAQ and KUC should not have so readily excluded water application as technically infeasible, and DAQ should have evaluated water application plus minimizing the drop distance as the most effective control measure for waste rock offloading from trucks.

Further, if DAQ continues to find that minimizing the drop distance satisfies BACT, DAQ must provide more detail to make this requirement into an enforceable measure. KUC did not even identify or justify as BACT what minimum drop distance should be required to minimize dust emissions from dumping. BACM is defined as control measures that achieve “permanent and enforceable emission reductions.” 40 C.F.R. § 51.1000. Thus, to ensure enforceability of control measures for waste rock offloading, the BACT measures must include specific requirements to ensure enforceable and permanent emission reductions, including specifying the minimum drop distance and requiring recordkeeping and reporting to verify compliance.

iii. BACT for Graders and Bulldozers at Bingham Canyon Mine.

DAQ and KUC have identified the application of water and chemical dust suppressants as measures to control particulate from road grading operations and bulldozers.43 Without any discussion or justification, DAQ and KUC identified the application of water within the pit influence boundary, and water and chemical dust suppressants outside the pit influence boundary, as BACT. DAQ has not explained why application of water and dust suppressants would not also be BACT for grading operations within the pit influence boundary.

Moreover, neither DAQ nor KUC identified any specific enforceable requirements that would ensure that the application of water and chemical dust suppressants would permanently reduce PM2.5 emissions. For these types of controls, a minimum water application and chemical dust suppressant application frequency and application intensity (quantity per area) must be specified as enforceable measures. EPA has identified the control efficiency of watering to be based on these factors along with the average hourly daytime traffic and the potential average hourly daytime evaporation rate for the area.44 UDAQ must specify minimum amounts of water application and chemical dust suppressant application as well as identify time between applications as part of its BACT determination, and propose recordkeeping and reporting to ensure compliance.

iv. BACT for Unpaved Haul Roads at Bingham Canyon Mine.

DAQ and KUC identified the following measures as potential methods to reduce PM2.5 emissions from unpaved haul roads: Paving the unpaved roads, the application of water and the use of dust suppression chemicals, limiting unnecessary traffic, and routine maintenance of haul roads

through the use of road base material.45 Of all of these measures, paving of the haul roads is the method that will provide the most control of particulate emissions, particularly with road sweeping to reduce road dust carryout from unpaved roads or from the haul trucks. KUC claimed that paving the haul roads “is not technically feasible at the [Bingham Canyon Mine] because of the weight of the haul trucks and the rapid deterioration that would occur and the frequently changing road locations,” and DAQ concurred with that finding.46 With respect to the claimed rapid deterioration of paved haul roads due to the weight of the haul trucks, that is not a justification to eliminate the control method as not technically feasible. Instead, that is an economic factor to be taken into account in the cost effectiveness analysis.47 Technical infeasibility means that physical, chemical, and engineering principles show that a control technique will not work on the emissions source under review. KUC has not demonstrated that paving of the haul roads is not technically feasible. The company is instead making economic arguments against paving the roads.

Similarly, while KUC claimed that paving of the roads was not technically feasible due to “frequently changing road locations,” KUC did not explain in detail how the “changing road locations” made paving not technically feasible. Importantly, how frequent are the haul road changed? Do the road changes affect some parts of the haul roads more than others? Are there more permanent haul roads that could be paved? DAQ and KUC must provide much more information to claim frequently changing road locations as a reason to exclude the top haul road control technology from the BACM/BACT analysis. This is especially because paving of haul roads along with street sweeping and setting of speed limits has been required as BACT for haul roads. The table below gives several examples in which paving of haul roads was required as BACT where practical (clearly taking into account when the weight of the haul trucks would frequently damage the pavement), and watering, chemical dust suppression, and reduced speed limits were required for the remaining haul roads.

Thus, there is no justification for DAQ and KUC to eliminate paving of haul roads entirely as a control measure for the serious area PM2.5 SIP, unless it is adequately demonstrated that there are specific circumstances at the Bingham Canyon mine haul roads that justify eliminating a commonly applied BACT control measure as BACT for the mine haul roads.

For those haul roads for which it may not be appropriate to require paving and street sweeping, a minimum water application and chemical dust suppressant application frequency and application intensity (quantity per area) must be specified as an enforceable measure. Instead, DAQ has proposed vague and unenforceable requirements for applying water and chemical dust suppressants to active haul roads “as weather and operational conditions warrant.” As previously stated, EPA has identified the control efficiency of watering to be based on the application frequency and intensity, as well as the average hourly daytime traffic and the potential average hourly daytime evaporation rate for the area.49 In order to ensure a specific control efficiency, DAQ must specify minimum amounts of water application and chemical dust suppressant application to unpaved haul roads as well as identify time between applications, and impose recordkeeping and reporting to ensure compliance. It must be noted that DAQ’s proposal to only require twice per year application of chemical dust suppressants to active haul roads outside the pit influence boundary50 has not been demonstrated by DAQ to reflect BACT for reducing PM2.5 from fugitive dust from these haul roads. Indeed, such an infrequent application of chemical dust suppressant seems wholly inadequate to ensure protection of the 24-hour average PM2.5 NAAQS.

In addition, if paving of haul roads is ultimately required as BACT, it is imperative that street sweeping also be required, for which a frequency must be specified as an enforceable control measure. Further, to ensure that watering, application of dust suppressants, and/or street sweeping is adequate to reduce PM2.5 emissions to the maximum degree achievable, a concurrent opacity limit is likely necessary for which compliance can be assessed daily or weekly. With that information, KUC can readily determine whether it is time to rewater, or to reapply dust suppressants. Last, it is imperative that recordkeeping and reporting be required as part of the BACM/BACT determination.

v. BACT for the Utah Power Plant, Tailings Site Service Roads.

KUC identified the application of water and chemical dust suppressants, limiting unnecessary traffic, and routine maintenance as BACT for the service roads at the Tailings Site.51 However, neither DAQ nor KUC specified enforceable measures to ensure the efficacy of these controls. To ensure that these controls actually reduce PM2.5 emissions, a minimum water application frequency and chemical dust suppressant application frequency as well as minimum application intensities (quantity per area) must be specified as an enforceable requirement. As previously stated, EPA has identified the control efficiency of watering to be based on these factors along with the average hourly daytime traffic and the potential average hourly daytime evaporation rate for the area.52 DAQ must specify as enforceable measures minimum amounts of water application and chemical dust suppressant application as well as identify time between applications, and propose recordkeeping and reporting to ensure compliance. DAQ must also specify as enforceable requirements the limits for vehicular traffic on the roads and how that will be enforced. In addition, to ensure that watering, application of dust suppressants, and/or limiting traffic on roads occurs in a manner to ensure PM2.5 emission reductions, a concurrent opacity limit is likely necessary for which compliance can be assessed daily or weekly. With that information, KUC can readily determine whether it is time to rewater, or to apply dust suppressants. Last, it is imperative that recordkeeping and reporting be required as part of the BACM/BACT determination.

vi. BACT for the KUC Smelter

Neither KUC nor UDAQ have proposed any new or upgraded pollution controls as BACT for the smelter. The smelter is a significant source of SO2 emissions.53 Kennecott appears to be claiming that because of its unique pollution controls at its copper smelter, it does not need to evaluate whether the copper smelter is equipped with BACT controls or propose emission limits reflective of BACT. While it is true that EPA highlighted the Kennecott Copper Smelter’s unique process in the 2002 primary copper smelting MACT rulemaking as a justification for not considering other copper smelters in the same category of the Kennecott copper smelter, that is not justification for not evaluating whether the best available control technology is being utilized at all of the emissions sources associated with the smelter.

For example, there are several scrubbers at the KUC Smelter, but the SO2 removal efficacy of those scrubbers is not discussed in KUC’s BACT analysis and thus it is not known whether the SO2 removal could be improved by operational changes or scrubber modifications or both. DAQ must provide more details on the existing controls at the various units of the smelter and the pollutant removal efficiency being achieved by those controls. With that information, a more thorough review of whether the smelter truly is meeting BACT can be made. If the scrubber can be upgraded to improve SO2 removal efficiencies, DAQ must conduct such an evaluation of such scrubber upgrades as part of the BACT analysis for the smelter.

b. BACT Has Not Been Properly Applied to the Petroleum Refineries.

There are four oil refineries in the Salt Lake and Provo Serious ozone nonattainment areas: Chevron, Tesoro, Holly Frontier, and Big West. With a few exceptions, neither DAQ nor the refinery owners have proposed any additional pollution controls or requirements to meet BACT at most of the emission units at these refineries. However, the companies’ justifications for not adding new pollution controls to meet BACT are often not adequately justified. DAQ generally has seemed to accept what has been proposed as BACT by the refineries at face value, without ensuring consistency in emissions assumptions and cost effectiveness analyses for similar controls at the four refineries. DAQ must evaluate and implement BACT for the refineries consistently. Moreover, in some cases, DAQ has not imposed as restrictive emission limits as proposed by the refinery owner in its BACT analysis. These details are provided below.

i. SO2 BACT for Heaters, Boilers, and Other Process Units that Utilize Refinery Fuel Gas.

Because the bulk of the heaters, boilers, and process units at a refinery use refinery fuel gas, the top pollution control technology for SO2 from these units is to optimize sulfur removal from the refinery fuel gas. A PSD permit for the proposed Arizona Clean Fuels Yuma petroleum refinery in Arizona provided a detailed analysis of SO2 BACT and required the utilization of enhanced sulfur conversion in the sulfur recovery plant to limit refinery fuel gas sulfur content to 35 parts per million by volume (ppmv) as H2S in a daily average basis.54 In comparison, the refineries in the Salt Lake and Provo nonattainment areas have proposed to meet the New Source Performance Standard Subpart Ja limits on H2S content of the refinery fuel gas of 60 ppm annual average and up to 162 parts per million on a 3-hour average.55 This limit will be met by the use of amine contactors to remove sulfur from refinery fuel gas to the maximum extent possible. To ensure the removal of sulfur to the maximum extent possible, the BACT determination for Arizona Clean Fuels includes a requirement to minimize emissions that would occur due to upsets: “Specifically, during periods of upset at the Amine Regeneration Unit, the amine contactors used to remove hydrogen sulfide from the RFG [refinery fuel gas] streams will continue to generate rich amine solution and to deplete the refinery’s supply of lean amine solution. The permit requires that rich amine solution storage capacity and a supply of lean amine solution, sufficient to support 24 hours of refinery operation, be maintained continuously.”56

This operational requirement to provide for more continuous sulfur content reduction in the refinery fuel gas and to eliminate peaks in SO2 emissions during upsets at the refinery must be considered by DAQ as a BACT option for all of the heaters and process units at the refineries in the state’s PM2.5 nonattainment area. The largest capital expense with such a control is the building of amine solution storage capacity – i.e., storage tanks. Such storage tanks can be readily installed. At a minimum, this operational requirement and storage of rich amine solution and a supply of lean amine solution must be evaluated as an SO2 BACT control option for refinery heaters and boilers fueled on refinery fuel gas.

Another control option that should have been more thoroughly evaluated is efficiency improvements in the amine scrubbing systems at each refinery to improve sulfur removal from the refinery fuel gas. The Minnesota Pollution Control Agency recently issued a PSD permit for major modifications at the existing Flint Hills Resources Pine Bend refinery that required such controls, and included switching the amine chemical from monoethanolamine (MEA) to diglycolamine (DGA) to more efficiently remove sulfur by using less steam for amine regeneration.57 They also approved pumping improvements at the upstream hydrotreaters to increase amine circulation and improve sulfur removal.58

Installation of a polishing amine or caustic scrubber also should have been given consideration to reduce sulfur in the refinery fuel gas. Tesoro did not consider such a scrubber as technically feasible, not because it was not technically feasible to apply a polishing scrubber after the amine unit at the Tesoro Refinery but instead because Tesoro claimed such a scrubber could not be designed, installed, and in operation prior to December 31, 2018.59 DAQ appears to accept this argument. 60 However, there is no justification to eliminate a control technology as technologically infeasible based on a claim that it cannot be installed by December 31, 2018. BACM including BACT is only limited by those controls that can be implemented in whole or in part within four years after the date of reclassification of a nonattainment area from moderate to serious. For the Salt Lake and Provo Nonattainment areas, the redesignation to serious occurred on May 10, 2017 and thus four years after that date is May 10, 2021. While, optimally, BACT controls should be implemented by the attainment date December 31, 2019, BACT requires consideration of controls that can be partially implemented before that date. In any event, neither Tesoro nor DAQ have provided any justification to document the time to install a polishing amine or caustic scrubber.

Yet another option that the various refineries should have considered as BACT is Merichem’s LO-CAT technology for sulfur recovery. This technology was recently required for the proposed Meridian Davis refinery to be located North Dakota, and the process allows for 100% turndown in gas flow and has an H2S removal efficiency of 99.9%.61

Each of these options should have been evaluated to reduce SO2 emissions from the refinery heaters and boilers that burn refinery fuel gas in the BACT analyses for the four refineries. Yet, neither UDAQ nor the refineries in the nonattainment areas gave much if any evaluation to any of these measures. Tesoro did not identify any SO2 controls as technically feasible for the process heaters at its refinery, with the exception of “Low H2S content fuel gas” which essentially is no control.62

Chevron identified flue gas desulfurization and a wet gas scrubber as potential controls for SO2 emissions from the refinery’s boilers and heaters, but then discounted those controls as not technically feasible primarily because it claimed these controls had not been required at a refinery boiler to meet BACT.63 Chevron acknowledged that the refinery gas sulfur content is dependent upon the efficiency and design parameters of amine scrubbers, but did not evaluate any measures to improve the efficiency of the amine scrubbers in reducing H2S in the fuel gas.

Holly Frontier also dismissed scrubbers for SO2 control as not technically feasible, and while the company considered use of natural gas in lieu of refinery fuel gas as a control option, the company ultimately found that refinery fuel gas as currently controlled with an amine scrubbing system was BACT.64 Holly Frontier did not even consider any measures to improve the efficiency of the amine scrubbing system.

The Big West Refinery also proposed as BACT use of refinery fuel gas as currently controlled as BACT, although the refinery did find that wet gas scrubbing was technologically feasible but not economically feasible.65 Big West Refinery gave no consideration of improvements to the amine scrubbing system although it is notable that the Big West Refinery already employs a caustic scrubber to treat refinery fuel gas during outages of the sulfur recovery plant, which would help to limit peaks in SO2 emissions during periods of upsets at the refinery. All of the refineries propose as SO2 BACT that they will comply with the H2S limitations of the New Source Performance Standards (NSPS) at 40 C.F.R. Part 60, Subpart Ja. However, the NSPS are supposed to be the floor for the BACT analysis – BACT can be no less stringent than the NSPS.

The existence of an NSPS limit on H2S content of refinery fuel gas does not mean that all additional measures to reduce sulfur in the refinery fuel gas do not need to be evaluated. The NSPS Subpart Ja limits for the refinery fuel gas used at process heaters are an H2S concentration of 162 ppmv on a 3-hour rolling basis and an H2S concentration of 60 ppmv on a 365-calendar day rolling average basis. 40 C.F.R. § 60.102(g)(1)(ii). The Arizona Clean Fuels Yuma plant permit shows that additional measures are available to greatly reduce the sulfur content of the refinery fuel gas, which would reduce SO2 emissions from all refinery process heaters and boilers that utilize refinery fuel gas to meet a sulfur content limit of 35 ppmv on a daily average. At the minimum, such additional measures should have been evaluated as SO2 BACT for process heaters and boilers at the four refineries.

ii. NOX BACT for Refinery Process Heaters and Boilers.

NOX BACT for heaters and boilers over 100 MMBtu/hour should be based on the application of SCR in combination with ultra-low NOX burners. The PSD permit application submitted for the proposed Arizona Clean Fuels Yuma petroleum refinery in Arizona provided an analysis of BACT determinations for refinery heaters and boilers, as well as other emission units at refineries. Based on that analysis, ultra-low NOx burners plus SCR at the larger heaters and boilers (greater than 100 MMBtu/hr ) was proposed and ultimately required as BACT for NOX at cost effectiveness of up to $23,000 per ton.66

None of the Utah refineries proposed SCR as BACT for the process heaters and boilers. Tesoro found that SCR plus ultra-low NOX burners was not technically feasible for its heaters. For two of the heaters, Tesoro stated that there was no space for an SCR reactor.67 For those claims to be justified, Tesoro must be required to submit plot plans and diagrams showing the lack of space for an SCR reactor, keeping in mind that SCR reactors can be elevated. For the other heaters, Tesoro found the ultra-low NOx burners plus SCR was not technically feasible because Tesoro claimed the controls could not be installed by December 31, 2018. As previously discussed, there is no legal justification to exclude a BACT control as technologically infeasible if it cannot be fully installed by December 31, 2018, nor is there any legal justification that all control measures of the serious PM2.5 nonattainment plans have to be implemented by December 31, 2018.

In its BACT analysis for its #5 F11005 and #6 F1108 Boilers (both 171 MMBtu/hr boilers), crude unit heaters (130 MMBtu/hr), and FCC Regenerator, Chevron found that SCR was technically infeasible “due to ammonia slip.”68 Chevron also considered SCR to be economically unfeasible for these boilers, which based on Chevron’s calculations, came out at $75,000 to $120,000 per ton of NOX removed.

Holly Frontier Refinery considered SCR by itself and with ultra-low NOX burners as technically feasible controls for its refinery boilers and its mechanically-drafted heaters.69 In fact, several of its heaters and boilers are already equipped with SCR, which DAQ must take into consideration when evaluating cost effectiveness of SCR for other refineries. Holly Frontier evaluated SCR application to other heaters at the plant but, according to UDAQ, the company determined that installation of SCR on naturally drafted heaters was not economically feasible.70 However, DAQ failed to present any cost information for those BACT evaluations. Holly Frontier stated that the only refinery heater than is mechanically drafted is unit 6H1, a 54.7 MMBtu/hr heater, but the company does not appear to have evaluated SCR for application to this heater. DAQ ‘s BACT evaluation report did not even mention this heater.

The Big West Refinery’s BACT analysis indicated that SCR was technically feasible for certain heaters, but found that SCR was not economically feasible due to costs exceeding $10,000/ton. However, the cost analysis only provided the incremental cost effectiveness and provided no specific details on how the costs were calculated. The company found that SCR was not technically feasible for heaters with heat input capacities less than 40 MMBtu/hr. DAQ’s review of BACT found that SCR was technically feasible on three boilers and the crude unit heater due to larger size and no space limitations, but then found the cost of SCR was not reasonable at $2.0 to $3.8 million per ton of NOx removed.71 That is an extremely high cost assessment and, as stated above, according to Big West’s BACT submittal, it reflects the incremental cost effectiveness (i.e., does not consider the full reduction in NOX from baseline emissions). Further, no details on the costs or assumed emission rates were provided to justify these cost numbers.

DAQ must ensure that the BACT analyses are conducted consistently across the various refineries, which is clearly not the case as discussed above. It is notable that only one company, Tesoro, found SCR to be not technically feasible and that claim was based on the time needed to install the controls. Even though that it not a legitimate argument to discount a potential BACT control as discussed above, EPA has indicated that it takes 28-58 weeks to install SCR at petroleum refinery process heaters.72 Thus, SCR could readily be installed by the attainment date of December 31, 2019 for the Salt Lake Serious PM2.5 nonattainment areas. Regardless, there is no sound evidence in the record that SCR could not be partially installed by December 31, 2019.

With respect to Chevron’s claims that the ammonia slip makes SCR technically infeasible, that is not a justification for technical infeasibility. Technical infeasibility means that a control technology cannot be applied to a particular unit. Given that numerous refinery heaters and boilers are equipped with SCR across the U.S., the level of ammonia slip is not a reason to discount the control as not technologically feasible. While ammonia slip can be a concern for formation of secondary PM2.5, ammonia slip can be limited by proper design and operation of an SCR system. Further, an important factor to the development of fine particulate matter due to ammonia slip from an SCR system is the sulfur in the fuel, because that sulfur tends to form SO3 and SO4 across the SCR catalyst, which then reacts with ammonia to form PM2.5. If the sulfur content of the refinery fuel gas is properly minimized, then that minimizes the issues with secondary PM2.5 development across an SCR. In the above section, we discuss options for improving the efficiency of the amine scrubber and for providing for continuous minimization of SO2 in the refinery fuel gas.

With respect to claims about economic infeasibility, DAQ must ensure that SCR cost analyses comply with EPA’s Control Cost Manual and EPA’s guidance on BACT determinations. For example, Chevron assumed a 10% interest rate and a 20-year life in amortizing capital costs of SCR. A 25-year life of an SCR for a refinery boiler is more appropriate73 and EPA’s Control Cost Manual assumes an interest rate of 7%.74 Had Chevron assumed a 25 year life and a 7% interest rate, the annualized costs would be much lower. In addition, Chevron included the costs of installing CEMs in the capital costs of an SCR system, for which there is no justification. Moreover, Chevron has not documented or justified the boilers’ baseline NOX emissions and emission reductions. Chevron claimed emissions in 2015 were 9.60 tons per year (tpy) at Boiler 5 and 12.20 tpy at Boiler #6 based on 2015 emissions. These low emission rates may be reflective of a low level of operation of these boilers. Chevron plans to decommission three boilers in 201875 and thus Boilers #4 and 5 may be operated more in the future. Thus, the assumed baseline NOx emissions of Boilers #4 and #5 are likely understated because the boilers will likely operate more frequently once the three boilers are decommissioned in 2018, which would make SCR more cost effective. Further, the effect of existing controls is not typically taken into account in a BACT cost effectiveness analysis. Instead, BACT is based on essentially uncontrolled emissions, calculated using a “realistic scenario of upper boundary uncontrolled emissions.”76

Chevron also assumed 90% control with SCR, which is the minimum NOX reduction that can be achieved with SCR. Current SCR systems can achieve greater than 90% NOX removal and NOX emission levels less than 5ppmv or 0.006 lb/MMBtu. EPA has indicated that SCR at refinery heaters can achieve 95% NOX reduction.77 In the BACT analyses, the Utah refineries must be required to evaluate the maximum emission reduction achievable with each control technology evaluated. It is also important to note that the Arizona Clean Fuels Yuma permit imposed SCR installations as BACT at costs as high as $24,000 per ton of NOX removed.78 Per EPA’s New Source Review Workshop Manual, if other similar sources have incurred similar costs for pollution control, then the costs are considered reasonable unless there is a unique aspect to the source in question which must be documented.79 The Holly Frontier Refinery has installed or plans to install SCR at some of its boilers and heaters, as discussed above and in the company’s BACT submittal. Thus, other similar sources have found it cost effective to install SCR at refinery heaters and boilers.

While the Big West Refinery also claimed SCR was not cost effective, it did not provide the details of its cost effectiveness analyses to know whether the analyses comport with EPA’s Control Cost Manual and EPA’s New Source Review Workshop Manual. DAQ must require each company claiming economic infeasibility of a pollution control provide the details of its cost effectiveness analysis for review. Again, given that other similar sources have found it cost effective to install SCR, Chevron and Big West must be required to install SCR as BACT unless the companies can adequately demonstrate that specific conditions at their refinery heaters and boilers sufficiently justify not requiring the NOX BACT control that other similar sources have installed at their refinery heaters and boilers.

In summary, SCR should have been more thoroughly evaluated as BACT for NOX at all of the process heaters and boilers of the four refineries in the Utah serious PM2.5 nonattainment areas. There is not adequate legal or technical justification to exclude SCR from analysis based on time to install the control or based on ammonia slip. In addition, SCR costs should be determined in a manner consistent with EPA’s Control Cost Manual and with EPA’s New Source Review Workshop Manual. Further, the bases for the cost analysis must be documented, so that DAQ and the public can review and evaluate the cost data associated with the controls. DAQ has failed to ensure (or provide assurances) that all cost analyses of SCR for process heaters at the refineries were based on EPA’s Control Cost Manual. Absent unit-specific justification for excluding SCR as a BACT control, SCR should be considered BACT for all process heaters and boilers with capacity of 100 MMBtu/hour heat input or greater, based on the fact that similar units at refineries have installed such controls.

iii. BACT for the Fluidized Catalytic Cracking Units (FCCUs)

A review of the EPA’s RACT/BACT/LAER Clearinghouse and other BACT databases shows that the lowest BACT emission limits for FCCUs for SO2, NOX, and PM2.5 are based on the controls of wet gas scrubbers and Lo-TOx systems. Tesoro has proposed as NOX BACT for its FCCU a Lo-TOx unit with emission limits of 10 ppmvd at 0% O2 on a 365-day rolling average, and 20 ppmvd at 0% O2 on a 7-day rolling average.80 Tesoro has also proposed a Wet Gas Scrubber for SO2 BACT to meet SO2 emission limits of 10 ppmvd at 0% O2 on a 365-day rolling average, and 18 ppmvd at 0% O2 on a 7-day rolling average.81 In addition, Tesoro has proposed PM2.5 BACT be based on its existing ESP with a downstream wet gas scrubber, although Tesoro claimed the emission limit would be determined through testing.82 These limits are lower than what are listed in EPA’s RACT/BACT/LAER Clearinghouse, although it must be noted that the Clearinghouse is often not up to date with the most recent BACT determinations. Indeed, NOx emission rates as low or lower than 10 ppmvd have been achieved in practice with Lo-TOx at other FCCUs.83 However, inexplicably, DAQ did not impose either of Tesoro’s proposed NOX or SO2 emission limits in its proposed BACT determination for the FCCU at Tesoro or the other refineries. Instead, DAQ simply referred to the NSPS limits of 40 C.F.R. Part 60, subpart Ja and the plant-wide SO2 and NOX emissions limits that apply on a ton per day and tons per rolling 12- month period. We discuss elsewhere why these approaches are inadequate to meet the requirements of BACT, particularly where the record does not support them. Further, DAQ has not provided any demonstration that these plant-wide limits represent BACT for the FCCU at Tesoro. Given that Tesoro has proposed SO2 and NOX BACT limits for the FCCU that are much more restrictive than the NSPS Subpart Ja, DAQ should have imposed those FCCU limits as BACT and analyzed them as BACT for the other refineries.

Holly Frontier Refinery has also proposed the same pollution controls as Tesoro to meet BACT at its FCCU, which are apparently required by a Consent Decree. However, Holly Frontier has not proposed as stringent NOX and SO2 limits as BACT, proposing NOX limits of 40 ppm based on a 365-day rolling average and 80 ppm based on a 7-day rolling average. Holly Frontier also proposed less stringent SO2 BACT limits than Tesoro, proposing SO2 limits of 25 ppmvd at 0% O2 based on a 365-day rolling average and 50 ppmvd at 0% O2 based on a 7-day rolling average. These are the NSPS Subpart Ja emission limits. Holly Frontier has not explained why it could not meet the lower limits proposed by Tesoro. And DAQ simply proposed the NSPS Subpart Ja limits and the existing daily plant-wide emission cap limits as BACT with no further evaluation of whether lower emission limits were achievable.84 DAQ must require Holly Frontier to evaluate the lowest emission rates that can be achieved with pollution controls to meet BACT.

Holly Frontier has also proposed to use a wet gas scrubber for PM_2.5 controls from the FCCUs, and has proposed PM₁₀ emission limits of 0.5 lb per1000 pounds of coke burn-off for Unit 4 and

0.3 lb per 1000 pounds of coke burn-off for Unit 25.^[1]  It appears DAQ ignored those proposed limits.  These limits could be used as starting points for the PM_2.5 BACT limits for the FCCU at Tesoro which will be equipped with both an ESP and a wet gas scrubber (and thus should achieve even lower PM_2.5 emission rates due to the combination of controls).  DAQ must conduct such a PM_2.5  BACT evaluation for Tesoro based on Holly Frontier’s proposed BACT determination.

Big West Refinery found that a wet gas scrubber was not feasible due to insufficient space.^[2]   DAQ appears to agree with this claim.^[3] Big West did not identify Lo-TOx as a NO_X control, and the company determined there was not enough space for its top-ranked NOx control of SCR.^[4]   For any arguments regarding lack of space, DAQ must ensure that that plot plans, photographs, and measurements justify the exclusion of a control due to lack of space.  Many pollution controls have options available for space-constrained retrofits.  For example, Hamon claims its Exxon Mobil wet gas scrubbing system is a compact system (about one-third smaller than other wet gas scrubbers), that it can be located anywhere including over a road, and that there is only one vessel which is an upflow vessel so the stack can be mounted above the vessel.^[5]  With a wet gas scrubber, a Lo-TOx system for NOx control could be added relatively easily, and Big West could be meeting BACT-level emission limits for SO₂, NO_X, and PM_2.5.  UDAQ must require Big West to more comprehensively evaluate whether it could install a wet gas scrubber, particularly a more compact scrubber such as the Exxon Mobil gas scrubber.  Big West must ensure it has examined all possible locations for the installation of a wet gas scrubber before discounting it as a BACT option.

Chevron evaluated the cost effectiveness of a wet gas scrubber for its FCCU but assumed that no further reductions in SO₂ and NO_X could be achieved with a wet gas scrubber. This was based on its claimed actual SO₂ emission rate in 2015 from the FCCU of 12 ppm.^[6]  Similarly, while Chevron evaluated the cost effectiveness of SCR and Lo-TOx for its FCCU, Chevron assumed much higher NO_X limits would apply with these controls than has been require as BACT and claimed that neither of these controls would reduce NO_X below Chevron’s 2015 actual NOx rate of 26 ppm.^[7] DAQ seems to have concurred with this cost analysis, without evaluating the details of the assumed level of control of SO₂ and NO_X.^[8]  While Chevron did not claim that a wet gas scrubber would not reduce PM_2.5 emissions at its FCCU, it relied on emissions from one stack test for defining baseline PM_2.5 emissions in its cost analysis which reflects the effect of its ESP.^[9]

There are several deficiencies in Chevron’s BACT analysis. It must first be noted that in a cost effectiveness analysis for determining BACT, EPA guidance requires that the baseline emissions must reflect “a realistic scenario of upper boundary uncontrolled emissions for the source” and that “[t]he NSPS/NESHAP requirements or the application of controls, including the other controls necessary to comply with State or local air pollution regulations, are not considered in calculating baseline emissions.”^[10]  For SO₂, NO_X, and PM_2.5 reductions, actual emissions at Chevron’s FCCU do not reflect uncontrolled emissions because Chevron is using SO₂ reducing additives, feedstock hydro treatment, and an ESP to control emissions of these pollutants.⁹⁵ Further, it is not appropriate for baseline emissions to be based on a one time stack test of emissions during a year or for baseline to be based on an average emission rate. Instead, baseline emissions should be based on a realistic upper bound level of operations.^[11]

Further, Chevron failed to evaluate the maximum degree of emission reduction and lowest emission rate that could be achieved with a wet gas scrubber and Lo-TOx.  Chevron assumed 12 ppm of SO₂ and 26 ppm of NO_X are the lowest rates that could be achieved with a wet gas scrubber and Lo-TOx, when lower emission rates can be achieved.  As discussed above, Tesoro has proposed an SO₂ limit of 10 ppm with these controls at its FCCU, and NO_X emissions of 10 ppm have been demonstrated in practice with Lo-TOx at FCCUs.

Moreover, given how Lo-TOx is designed to work in a wet gas scrubber and that this control reduces SO₂, NO_X, and PM_2.5 emissions, it makes most sense to evaluate the cost effectiveness of these controls for all three pollutants together.  While Chevron evaluated the costs of Lo-TOx plus a wet gas scrubber together, it was evaluated from a dollar per ton basis only for NOx emissions.  Because Chevron assumed no NOx emission reductions with these controls, the cost effectiveness was essentially equivalent to the annualized cost of these controls.^[12] These annualized costs of these multi-pollutant pollution controls should have been evaluated together and divided by the sum total of the SO₂, NO_X, and PM_2.5 emission reductions that these pollution controls are capable of achieving.

Chevron also greatly overstated annualized costs of control with a wet gas scrubber and Lo-TOx. 

First, inexplicably, the capital costs for a wet gas scrubber in Chevron’s analysis of Lo- TOx plus Wet Gas Scrubber is about 14% higher than Chevron’s cost analysis of a Wet Gas Scrubber by itself.  It is not clear why the capital costs of a new wet gas scrubber would be different with or without Lo-TOx. The Low-TOx company Linde claims its Lo-TOx solution can work with existing pollution controls and claims Lo-TOx is a “turnkey solution.”^[13] It is simply an “end-ofpipe system that removes NOx by adding ozone” which then enables capture of oxidized NO_X compounds in air pollution controls like a wet gas scrubber. So, while there would be an expense of an ozone generator, the wet gas scrubber would be the same design and cost whether installed by itself or with Lo-TOx. There are numerous other costs in Chevron’s analysis that are based on a percentage of the capital costs, and these costs in the Lo-TOx plus wet gas scrubber analysis were thus inflated because of the increased wet gas scrubber cost assumed by Chevron.  These increased costs included freight, sales tax, foundations, construction, instrumentation, electrical, piping, insulation, painting, engineering, construction, contractor fees, start-up, performance test, and contingencies.^[14]  Chevron’s unjustified assumption that the scrubber for the control configuration of wet gas scrubber with Lo-TOx would be more expensive than the wet gas scrubber by itself means that Chevron’s cost analysis for Lo-TOx plus wet gas scrubber are overstated.

Chevron also greatly overstated annualized cost by assuming a 10% interest rate to annualize capital costs.  EPA’s Control Cost Manual assumes an interest rate of 7% for annualizing costs of controls, which is much lower than Chevron’s assumed 10% interest rate.  Had Chevron assumed a 7% interest rate, the annualized costs would be much lower.^[15]

In addition to providing for optimized NO_X control with Lo-TOx and for significantly improved SO2 control, a wet gas scrubber also will achieve greater control of PM_2.5 than the existing ESP at Chevron’s FCCU, a fact which Chevron stated.^[16]

In summary, Chevron failed to adequately evaluate a wet gas scrubber plus Lo-TOx for SO₂, NO_X, and PM_2.5 emission reductions at the FCCU, and DAQ appears to have taken Chevron’s analysis at face value and concurred with Chevron’s findings.  Given that Tesoro and Holly Frontier Refining have proposed this control as BACT for SO₂, NO_X, and PM_2.5, DAQ must require that Chevron adequately show that operations are different at its refinery and FCCU compared to the Tesoro and Holly Frontier to justify not requiring these highly effective pollution controls to meet BACT.  Without such a demonstration, these controls of Lo-TOx plus a wet gas scrubber must be required at Chevron’s FCCU for SO₂, NO_X, and PM_2.5 BACT.

iv.  BACT for Flaring

DAQ has proposed as BACT for flaring operations at the refineries the NSPS Subpart Ja requirements and a design limit on the quantity of gases that are flared per year. DAQ should also impose the requirements of the refinery MACT standards at 40 C.F.R. 63.670 and 63.671 which are designed to limit emissions from flaring.  Specifically, those requirements include among other things:

• A requirement of a continuous pilot flame when regulated material is routed to the flare and pilot flame monitoring
• A requirement to operate with no visible emissions and visible emissions monitoring
• Flare tip velocity requirements
• Combustion zone operating limits
• Dilution operating limits for flares with perimeter assist air
• Operation of a continuous particulate monitoring system

The requirements of 40 C.F.R. 63.670 and 63.671 will help minimize PM_2.5 and PM_2.5 precursor emissions from flaring operations and will provide data to better characterize emissions from flaring operations.

c.  BACT Has Not Been Properly Applied to the West Valley Power Plant.

The Utah Municipal Power Agency’s (UMPA) West Valley Power Plant consists of five natural gas-fired simply cycle turbines, each with an output rated at 43.4 MW. The turbines are equipped with water injection, evaporative spray mist inlet air cooling, SCR, and CO oxidation catalyst for pollution controls. The BACT analysis for the West Valley Power Plant claims the turbines achieve NOx emissions of 5 ppmvd @ 15% oxygen.^[17] This does not reflect NO_X BACT for simple cycle turbines.  The West Valley BACT analysis claims the lowest NO_X rate for simple cycle turbines in the RACT/BACT/LAER Clearinghouse is 2.5 ppmvd on a 1- hour basis.^[18]  

However, the lowest NOx rate for combustion turbines in the RACT/BACT/LAER

Clearinghouse is actually 2.0 ppmvd at 15% oxygen, 1-hour average to be achieved with SCR.^[19] While this was a lowest achievable emission rate (LAER) determination, that does not negate such a limit for consideration as BACT.  UMPA must be required to evaluate the top level of NOx control in its BACT analysis, including a 2.0 ppmvd limit.

UMPA’s cost analysis for improved NOx controls to achieve a limit of 2.5 ppmvd is also flawed for numerous reasons, in addition to failing to evaluate the lowest achievable NOx emission rate. Importantly, the effect of existing controls are not typically taken into account in a BACT cost effectiveness analysis.  Instead, BACT is based on essentially uncontrolled emissions, calculated using a “realistic scenario of upper boundary uncontrolled emissions.”^[20]

In the case of the West Valley units, UMPA’s BACT analysis indicates that the existing SCR catalyst would likely need to be replaced if it was not designed to achieve 2.5 ppm NOx which reflects about 90% NOx control.^[21] Had UMPA evaluated NOx BACT cost effectiveness based on uncontrolled emissions, the costs would have very reasonable.  Specifically, assuming an uncontrolled NOx rate of 25 ppmvd, that means the currently operated SCR catalyst at the West Valley units is achieving 80% NOx control (assuming the units are achieving 5 ppmvd). Thus, uncontrolled NOx emissions in lb/hr would be 37.0 lb/hr and SCR catalyst to achieve 2.5 ppmvd or 3.7 lb/hr would result in a reduction of 33.3 lb/hr.  Assuming on average 700 hours of operation per year per unit based on UMPA’s claimed level of operations, the reduction in NOx per year would be 11.7 tons per year per unit.  Using UMPA’s annualized cost estimates of

$41,580 per year for new SCR catalyst at each unit, the cost effectiveness of installing new

SCR catalyst based on the reduction from uncontrolled emissions to achieve 2.5 ppmvd would be approximately $3600/ton.  Such costs are within the range that has been deemed reasonable for such controls at electrical generating units (EGUs).

Even if it were appropriate to take into account the current controlled NOx rate with SCR catalyst at the West Valley units, UMPA did not provide an indication of the West Valley units’ hourly NOx emission rate to compare with the 2.5 ppmvd hourly emission rate that UMPA evaluated as BACT.  The West Valley units are currently subject to a 5.0 ppmvd NOx rate based on a 30-day rolling average not including startup and shutdown.  However, under a 30-day average emission limit, hourly emissions can fluctuate widely above 5.0 ppmvd and the unit can still show compliance with a 30-day average emission limit.

Another flaw in UMPA’s SCR catalyst cost analysis was that emissions reductions were projected based on a historically low level of operation of each unit.  Based on data submitted to EPA’s Air Markets Program Database, the years 2014-2015 used by UMPA^[22] were the years of lowest hours of operation for Units 1-3 and very close to the lowest hours of operation for Units 4 and 5.  Out of the past five years (2012-2016), the West Valley units have operated at essentially two levels of operating hours:  4,000-6,000 hours in 2012-2013, and 500-800 in 2014- 2016.  In 2017, Units 2-5 significantly exceeded the 500-800 operating hour range, with Units 2- 4 operating at about 1200 hours and Unit 5 operating 965 hours.  Based on the operations the first three months of 2018, each unit is on track to operate about 1000 hours per year or so. Thus, UMPA’s BACT cost analysis appears to be based on an unreasonably low level of operation, which would artificially inflate the cost effectiveness values of all pollution controls evaluated because an assumed lower hours of operation would equate to less tons per year of NOx removed

(and higher cost effectiveness values).  Given that cost analyses are required to be based on a “realistic scenario of upper boundary uncontrolled emissions,”^[23] using a lower boundary of assumed operating hours is inconsistent with the EPA’s New Source Review Workshop Manual.

In addition, UMPA’s cost analysis to achieve a NOx limit of 2.5 ppmvd is not adequately supported, nor were all options adequately evaluated.  First, the vendors that UMPA requested quotes from gave a wide range of SCR catalyst costs of $300,000 to $600,000 and indicated that a “detailed and comprehensive technical analysis of the existing turbines and existing SCR system would be necessary.”^[24] Indeed, one vendor indicated that the current catalyst “may be able to meet the NOx outlet of 2.5 ppm and NH3 [ammonia] at 5 ppm but it would reduce the service life of the catalyst from 30,000 hrs to 20,000 hrs.”^[25] This indicates that UMPA may not even need to replace the existing SCR catalyst to achieve lower NOx rates, which would greatly reduce the capital costs of control.  Yet, UMPA did not provide a lower end cost estimate reflective of this scenario.

Further, in determining cost effectiveness to replace the existing catalyst with a new catalyst, UMPA assumed a catalyst life of only 10 years.^[26] As indicated in an email from Haldor-Topsoe, a typical catalyst life is 30,000 hours, which may be reduced to 20,000 hours if the current catalyst installed at the West Valley units was used to achieve 2.5 ppmvd.^[27] Given that UMPA assumed the West Valley units only operate 500-700 hours per year, if a new catalyst is installed with a 30,000 hour life, it should last 40 – 60 years.  Even if UMPA assumed operating hours of each unit of 1,400 per year (which is double what the units operated at in the first half of 2017), the life of the catalyst would be 21 – 22 years.  Assuming a longer life of the catalyst that reflects how UMPA intends to operate the units would result in lower annualized costs of SCR catalyst.

In summary, UMPA’s NOx BACT analysis did not evaluate the lowest achievable NOx emission rate for similar sources of 2.0 ppmvd, did not evaluate emission reductions from uncontrolled emission rates, did not use a reasonable level of operating hours reflective of upper boundary operations, and did not adequately evaluate the costs for meeting lower NOx rates at the West Valley units.  As indicated in UMPA’s BACT submittal, there are several examples of gas-fired turbines meeting NOx rates lower than the 5 ppmvd achieved at the West Valley units, and it is very likely that a properly performed BACT cost analysis would show SCR catalyst to achieve lower NOx emission rates at 2.0 to 2.5 ppmvd is cost effective.

DAQ did not question any of these assumptions in UMPA’s BACT analysis and concurred with UMPA’s analysis that SCR to meet a lower NOx rate of 2.5 ppmdv was not economically feasible.^[28]  For all of the reasons discussed above, UMPA’s NOx BACT analysis is flawed.  DAQ must conduct a more thorough evaluation of NOx BACT for the West Valley units taking into account the issues discussed above.

d. BACT Has Not Been Properly Applied to the Gadsby Power Plant.

i. Review of BACT for Units #4, 5, and 6

Gadsby Units #4, 5, and 6 are 43.5 MW simple cycle, natural gas-fired combustion turbines equipped with water injection and SCR for NOx and CO oxidation catalyst. These units are thus very similar to the West Valley units described above.

PacifiCorp has proposed as BACT a continuation of the current controls to meet a NOx limit of 5 ppmvd at 15% oxygen on a 30-day rolling average.^[29] As discussed above in the comments on the West Valley Power Plant BACT analysis, the lowest NOx rate for combustion turbines in the

RBLC is actually 2.0 ppmvd at 15% oxygen, 1-hour average to be achieved with SCR.^[30]The Gadsby BACT analysis claims the lowest NOx rate for simple cycle turbines in the RACT/BACT/LAER Clearinghouse is 2.5 ppmvd on a 1-hour basis.^[31]  These limits were imposed for the same NOx controls as employed as Gadsby Units #4, 5 and 6.  Despite the fact that significantly lower limits have been required as BACT at similar sources, PacifiCorp did not conduct any evaluation of BACT to achieve a 2.0 to 2.5 ppmvd limit.  Thus, PacifiCorp did not submit a proper evaluation of NOx BACT for the Gadsby Units #4, 5, and 6. There are several examples of gas-fired turbines meeting lower NOx rates than the 5 ppmvd achieved at the Gadsby Units #4, 5, and 6 as indicated in PacifiCorp’s BACT submittal, and it is very likely that a properly performed BACT cost analysis would show SCR catalyst to achieve lower NOx emission rates at 2.0 to 2.5 ppmvd was cost effective for these Gadsby units.

Although UDAQ’s BACT evaluation provides some information on NOx BACT controls for Units #4, 5, and 6, UDAQ did not conduct a cost analysis for upgraded NOx BACT controls to meet lower emission limits nor did UDAQ require any such analysis of PacifiCorp.^[32] DAQ simply states “[r]etention of the existing SCR and water/steam injection systems for each of the three [combustion turbines] is recommended as BACT to control NOx emissions.”^[33] DAQ must conduct a proper evaluation of NOx BACT for Gadsby Units #4, 5, and 6 to properly evaluate if lower NOx limits could be achieved to meet BACT.

ii.  Review of BACT for Gadsby Units # 1, 2, and 3

Gadsby Units # 1, 2, and 3 are natural gas-fired utility boilers.  Units 1 and 2 are equipped with low NOx burners and are rates at 65 MW and 80 MW, respectively.^[34] Unit 3 is equipped with flue gas recirculation (FGR) and is rated at 105 MW.  Each boiler is allowed to use fuel oil as a backup fuel during natural gas curtailments.

PacifiCorp provided cost effectiveness analyses for upgrading Units #1 and 2 with FGR, for upgrading Unit #3 with low NOx burners, and also for the add-on controls of SCR and SNCR.  There are several deficiencies with PacifiCorp’s cost analysis.  First, PacifiCorp applied unreasonable capital recovery factors in determining annualized capital costs of control. Specifically, PacifiCorp assumed an equipment life of all of the various controls of only 10 years.  For utility boilers, EPA has consistently applied an equipment lifetime for these types of controls of 30 years unless the source was subject to an enforceable requirement to shut down earlier.  It must be noted that the EPA Cost Model used by PacifiCorp clearly indicates that the equipment life should be at least 30 years for a utility boiler.¹²⁰ In addition, for the SCR and SNCR cost evaluations, PacifiCorp assumed an interest rate of 7% which was consistent with

EPA’s Control Cost Manual.  However, for its cost evaluation of FGR and low NOx burners, PacifiCorp assumed a higher interest rate of 10%.^[35]  There is no justification for use of such a high interest rate.  EPA’s Control Cost Manual specifies an interest rate of 7% which is commonly used in cost effectiveness analyses.  Had PacifiCorp determined annualized capital costs based on a 30-year life of the pollution controls and a 7% interest rate, the total annual costs would have been much lower.

PacifiCorp also assumed controlled NOx rates for the various controls evaluated that are higher than have been actually achieved in practice, which resulted in a minimization of the tons of pollutant reduced per year, artificially inflating the cost effectiveness values.  PacifiCorp also failed to document and justify the controlled emission rates with the addition of FGR at Units #1 and #2 and with the addition of low NOx burners at Unit #3.  For the evaluation of FGR at Units #1 and #2 that are already equipped with low NOx burners, PacifiCorp only assumed a controlled emission rate of 0.10 lb/MMBtu. However, for the evaluation of the addition of low NOx burners at Unit #3 that is already equipped with FGR, PacifiCorp assumed a controlled NOx rate of 0.05 lb/MMBtu.^[36] PacifiCorp provided no explanation why Units #1 and 2 could not achieve emission rates as low as projected for Unit #3 with the same pollution controls of low NOx burners plus FGR.  Moreover, a review of the NOx limits with low NOx burners required in the RACT/BACT/LAER Clearinghouse presented in PacifiCorp’s BACT analysis shows that much lower NOx rates than 0.05 lb/MMBtu have been required as BACT for the controls of low NOx burners and FGR.  For example, the Georgia Pacific Breton facility is subject to a 0.02 lb/MMBtu NOx limit, the Iowa Fertilizer Company is subject to a 0.0125 lb/MMBtu limit, and the Green River Soda Ash Company is subject to a 0.011 lb/MMBtu limit, with NOx BACT for all of these sources being based on low NOx burners and FGR.^[37]Had PacifiCorp evaluated the limits that other similar sources have had to meet under BACT for the same controls of low NOx burners and FGR, the tons per year of NOx reduced would be greater, which would make the controls more cost effective (i.e., provide for lower cost effectiveness values).

Moreover, all of PacifiCorp’s cost effectiveness analysis for BACT controls is based on current level of control as a baseline, which is inconsistent with EPA’s New Source Review Workshop Manual.  Instead, BACT cost effectiveness evaluations should be based on essentially uncontrolled emissions and calculated using a “realistic scenario of upper boundary uncontrolled emissions.”^[38] Had PacifiCorp used uncontrolled emissions as baseline, the cost effectiveness values would be even lower.

In summary, PacifiCorp’s BACT cost effectiveness analyses for Gadsby Units #1, 2, and 3 are significantly flawed because the analyses fail to comply with EPA’s cost effectiveness guidance.  UDAQ must require PacifiCorp to revise its cost effectiveness analyses to be based on a 30-year life of controls and a 7% interest rate and to reflect emission reductions from a realistic upper boundary of uncontrolled emissions based on the lowest emission rates achieved at similar sources with the controls being evaluated.

DAQ did not question any of these assumptions and concurred with PacifiCorp’s finding that NOx BACT is satisfied with the current emission controls.^[39]For all of the reasons discussed above, PacifiCorp’s NOx BACT analysis for Gadsby Units # 1, 2, and 3 is flawed. UDAQ must conduct a more thorough evaluation of NOx BACT for Units #1, 2, and 3 taking into account the issues discussed above.

e. BACT Was Not Properly Applied to the Lhoist North America – Grantsville Facility.

The Lhoist North American quarry and lime processing plant, which began operations in 1960, includes the following emitting activities:  mining, limestone processing, one rotary kiln which heats crushed limestone and converts it to quicklime or calcium oxide, post-kiln lime processing, lime hydration equipment to convert quicklime to hydrated lime (calcium hydroxide), bagging facilities, and loadout facilities.^[40] According to the company’s BACT analysis, the lime manufacturing operations of this facility have essentially been suspended since 2008, although purportedly the plant is being maintained to remain in complete “ready mode.”^[41] Given that this plant has been not operating for close to 10 years, UDAQ should simply revoke the facility’s operating permit.  The Lhoist North American facility could not resume operation after being shut down for 10 years or more without being subject to major new source review (NSR) permitting requirements, which for PM2.5 and PM2.5 precursor emissions would require that the lowest achievable emission rate (LAER) be met at all emission units.^[42]

Assuming the plant does restart operations soon and can legitimately do so without obtaining a major source NSR and PSD permit, DAQ must make a BACT determination and impose BACT limits now to apply as soon as Lhoist North American begins operations.

One method of control for SO₂ and to use in combination with controls for other pollutants that Lhoist failed to consider was using primarily natural gas to fire the rotary kiln system.  It appears that, when Lhoist last operated, natural gas was the primary fuel.  However, Lhoist’s BACT analysis indicates that fuel oil can be used when natural gas delivery is curtailed, on-specification used oil can be used to supplement natural gas and fuel oil, and also tire-derived fuel can be used on an as needed basis.^[43]  Sole use of natural gas is the cleanest fuel to use from a PM2.5 perspective, and thus should be the first consideration in the BACT analysis for the rotary kiln, by itself and in combination with other controls.  For example, in the NOx BACT analysis, Lhoist dismissed use of low NOx burners in part due to the use of multiple fuels at the Lhoist rotary kiln, claiming that other kilns that have successfully used low NOx burners burned one type of fuel.¹³⁰  Lhoist should have considered sole use of natural gas in combination with other controls including low NOx burners in the BACT analysis.

Lhoist proposed a fabric filter baghouse as BACT but requested not to select which type of baghouse to install until “a later date” due to the facility “being in care and maintenance mode.”^[44] What is most important at this point is for UDAQ to set an emissions limit reflective of BACT for PM2.5 from the rotary kiln.  A review of the RBLC shows that the lowest PM2.5 emission limit for rotary kilns is 0.1050 lb/ton, 3-hour average, with some exceptions for low capacity during which a 5.24 lb/hr limit applies over a 3-hour average (RBLC ID IL-0177, Mississippi Lime Company).  In addition, numerous facilities are also subject to an opacity BACT limit, with the lowest being a 10% opacity limit on a 6-minute average with some exceptions (RBLC ID PA-0283, Graymont PA Inc./Pleasant Gap & Bellefonte Plants).  There are also visible emission BACT limits for rotary kilns of 15% opacity limit on a 6-minute average with no exceptions (RBLC ID FL-0321, Jacksonville Lime, and RBLC ID OH-0321, Martin Marietta Materials).  It must be noted that the definition of BACT includes a visible emissions limit.  Thus, UDAQ must impose BACT limits no higher than these limits on the rotary kiln system at Lhoist applicable upon startup.  Yet, UDAQ has not proposed any PM2.5 BACT limits for the rotary kiln system.^[45]  This is a significant deficiency in UDAQ’s BACT analysis for Lhoist.

Instead, Lhoist has proposed to meet the existing lime kiln MACT limit for filterable PM of 0.12 lb/ton of stone feed (adjusted to reflect 37% of PM being of the size PM2.5 or smaller) and has proposed a total PM2.5 limit of 1.4324 lb/ton of stone feed based on condensable PM2.5 testing of other Lhoist North America facilities.^[46] This is not how a BACT emission limit is to be set.  First, BACT is to be based on a top-down analysis, not a bottom-up analysis. Further, there is no basis for assuming the existing kiln MACT limit should be the BACT floor and not the new kiln MACT limit of 0.10 lb filterable PM per ton of stone feed. In addition, Lhoist provided no

BACT analysis to justify that its proposed total PM2.5 (filterable plus condensable) limit of 1.4324 lb/tons of stone feed reflects the maximum degree of emission reduction achievable.  For example, it is not known what fuels the other Lhoist kilns were utilizing.  It is most likely that burning natural gas produces the lowest emissions of condensable (as well as filterable) PM2.5.  UDAQ must require that the PM2.5 emission limits required as BACT are set based on a proper top-down analysis reflective of the maximum degree of PM2.5 emission reduction achievable, considering the cost and other factors that are weighed in a BACT determination.

With respect to BACT for SO2 emissions, Lhoist states that SO2 emissions are mainly due to the sulfur content of the fuel used in the kiln.^[47] Thus, sole use of natural gas to minimize SO2 emissions to the greatest extent should have been reviewed as an SO2 control in the BACT analysis.  While Lhoist provided anecdotal information in its BACT analysis that burning of tire- derived fuel which has approximately 1.2% sulfur content did not increase SO2 emissions, Lhoist did not provide any specific test data to back that claim up.  Further, Lhoist made no claims regarding SO2 emissions from the kiln during the burning of oil, other than to say the sulfur content of those fuels are limited by a permit condition. There is no question that these fuels contain higher sulfur content than natural gas.  Given that natural gas is the primary fuel used in the rotary kiln, it would likely be extremely cost effective to simply stop utilizing oil or tire-derived fuel to meet SO2 BACT.  Yet, DAQ did not even evaluate sole use of natural gas as an SO2 BACT control option.^[48]  DAQ must review this very reasonable control option for the Lhoist rotary kiln system.

Lhoist has proposed selective noncatalytic reduction (SNCR) to meet BACT for NOx.^[49] However, in proposing a NOx emission limit reflective of BACT, Lhoist proposed the low end of achievable NOx reductions with SNCR of 25% and applied that to the current NOx limit of the operation permit for Lhoist’s Grantville Plant of 75.00 lb/hr.^[50]  DAQ has also assumed the same 25% level of control in proposing a NOx BACT limit of 56 lb/hr.^[51]  When a BACT control can operate at a range of control efficiencies, the BACT analysis must include an evaluation of the control at the top control efficiency.  If Lhoist claims no higher NOx removal efficiency than 25% can be achieved with SNCR at its Lhoist lime kiln, then it needs to document why.  In addition, it does not necessarily make sense to propose a limit based on 25% control from the current NOx limit of 75.00 lb/hr.  It could be that actual emissions from the lime kiln have been significantly lower than 75.00 lb/hr.  Lhoist should document what the lime kiln’s actual NOx emissions were based on actual test data and the fuel mix being utilized.  Then the proposed limit should be based on the maximum achievable control with SNCR, taking into account the various BACT factors, with a margin of safety for compliance.

Although Lhoist evaluated the cost effectiveness of low NOx burners based on 30% control, the company claimed that such levels of NOx control could not be universally achieved. Yet, Lhoist did not provide any documentation to support this claim.  UDAQ’s analysis included some anecdotal claims to support Lhoist claims, but did not provide much supporting documentation.^[52]  Given that low NOx burners could achieve greater than the 25% NOx control proposed by Lhoist as BACT and at lower costs, Lhoist must be required to provide sufficient documentation to support eliminating low NOx burners as a control.

In addition, as described above, Lhoist should be required to evaluate whether low NOx burners could work effectively at its lime kiln if the kiln was limited to solely natural gas combustion, which would better allow for maintaining burner performance due to the consistency of the fuel NOx and other related fuel characteristics.  If such burners could work with the kiln solely utilizing natural gas, the NOx emission reductions would be greater than with SNCR at lower costs than SNCR and with no concerns about ammonia slip.  In addition, assuming low NOx burners would be more viable as a NOx control with natural gas as the sole fuel, Lhoist should also be required to evaluate the NOx reductions and cost effectiveness with both low NOx burners and SNCR installed, which could provide the maximum reduction in NOx emissions from the lime kiln.

Last, Lhoist did not propose any BACT emission limit for ammonia emissions from the SNCR and instead based BACT on good combustion processes and burner/process optimization.^[53]   However, with the addition of SNCR to control NOx and the likely level of ammonia slip from the SNCR, it is imperative that an ammonia BACT limit be set for the Lhoist facility.  DAQ did propose an ammonia slip limit of 10 ppm as BACT.^[54][55]  While we agree that a limit on ammonia is warranted for Lhoist (indeed, there are several examples of pound per hour ammonia BACT limits in the RACT/BACT/LAER Clearinghouse), DAQ did not conduct any analysis to show that this level of ammonia slip actually represents BACT for Lhoist.  DAQ itself noted that permits for SCR at large combustion turbines have limited ammonia slip emissions at lower levels of 2.0 ppm and 5.0 ppm.^[56] DAQ must conduct a proper BACT analysis for ammonia slip to ensure it is requiring the maximum reduction in ammonia emissions that is achievable considering the other BACT factors.  Further, DAQ must impose the ammonia slip limit as an enforceable requirement (it currently is not listed in draft Section IV Part H of the Utah SIP) and must require ammonia monitoring to ensure compliance.

f. BACT Was Improperly Applied to ATK.

ATK is a facility that manufactures and tests solid rocket motor propulsion systems, explosives, flare illuminants, and composite materials.  The following provides review and comment on emission units for which ATK analyzed BACT.

According to ATK’s May 2017 BACT analysis, ATK operates 21 natural gas-fired boilers and 19 fuel oil-fired boilers.^[57]  ATK appears to have eliminated most of these boilers from BACT review.  For example, ATK only evaluated NOx BACT for the largest gas-fired boilers of 25 MMBtu/hour or greater.^[58] For the diesel-fired boilers, ATK relies on the existing ultra-low sulfur fuel requirement (<15 ppm sulfur) to reflect BACT for all PM2.5 and PM2.5 precursors from these emission units.  ATK did not provide any analysis of BACT for any of the dieselfired boilers.  UDAQ’s BACT evaluation tacitly approves of only focusing on the “four largest natural gas boilers,” and UDAQ failed to provide any justification to eliminate the other boilers from a BACT analysis.^[59]

In its October 2016 RACT submittal, ATK referred to natural gas and fuel oil consumption limits placed on the boilers to satisfy RACT.^[60]  However, such limits have not been demonstrated to meet BACT, and additional control measures are readily available..

According to the Title V permit for ATK Promontory Site, the sizes of the diesel-fired boilers are in the range of 0.84 MMBtu/hr to 8.37 MMBtu/hr.^[61] All diesel-fired boilers are required to fire only ultra-low sulfur diesel (<0.0015% sulfur),^[62] and are apparently subject to a total limit of 1,298,400 gallons of fuel oil per 12-month period.^[63]  While the annual limit on fuel oil burned will limit total operation of the 19 fuel oil-fired boilers at ATK Promontory, it is not clear whether, and seems quite plausible that, some of these boilers are utilized more frequently than others and thus may warrant more thorough evaluation of BACT controls.  DAQ must require ATK to identify the actual operating hours and annual heat input for each of these boilers to enable a more thorough review of BACT – primarily NOx BACT- for these boilers.  For those units operated more frequently, ATK should evaluate low excess air (LEA) firing, flue gas recirculation (FGR), staged combustion, low NOx burners and other NOx reduction measures, even for the units smaller than 25 MMBtu/hour.  DAQ did not provide any justification to exclude smaller units from a BACT evaluation.

While all of the ATK diesel-fired boilers are subject to periodic tune-up requirements, some units are subject to more frequent tune-up requirements based on size of the boiler.^[64] These requirements appear to be based on provisions in 40 CFR Part 63, Subpart DDDDD.^[65] Regardless of whether a boiler is subject to Subpart DDDDD, these more frequent tune-up requirements clearly could be required on all boilers annually (as is currently required for boilers with a heat input capacity greater than 10 MMBtu/hr pursuant to Condition II.B.34.b.(3) of the ATK Title V permit).  Boiler tune-ups can lower NOx and PM2.5 emission rates, among other pollutants, and more frequent tune-ups can more consistently ensure lower emission rates.  Thus, UDAQ and ATK must at the minimum consider annual tune-ups for all diesel-fired boilers.

For the natural gas-fired boilers, ATK has failed to provide a thorough analysis of NOx BACT.  ATK indicates that it has upgraded a 71 MMBtu/hr boiler with an ultra-low NOx burner which has a NOx emission rating of 9 ppm.^[66] The title V permit for ATK Promontory also indicates that a 12.55 MMBtu/hr boiler has a NOx emission rate of 9 ppm.^[67] Yet, there are numerous other natural gas-fired boilers of similar size for which ATK did not evaluate any low NOx burners.  ATK only very minimally evaluated low NOx burners as BACT for the other 71 MMBtu/hr boiler that has a NOx emission rate of greater than 30 ppm,^[68] more than 4 times higher than the NOx rating of the boiler with an ultra-low NOx burner.  Indeed, ATK has previously indicated that when operating in standby mode, NOx emissions from this other 71 MMBtu/hr boiler are approximately 50 ppm,^[69] which is 5.6 times higher than the 9 ppm rate achieved with the ultra-low NOx burner.

ATK claimed in its BACT submittal that the higher NOx-emitting 71 MMBtu/hr boiler only operates as backup capacity and is restricted to an annual natural gas limit.^[70]  However, that 12month rolling limit on the amount of natural gas fired does not limit the boiler’s operations on a daily basis, and thus the boiler could significantly contribute to daily PM_2.5 concentrations when it operates.  Moreover, given that it was cost effective for ATK to install an ultra-low NOx burner on one of the 71 MMBtu/hr boilers, it should be assumed that it is also cost effective to install an ultra-low NOx burner on the other 71 MMBtu/hr boiler.  If ATK is claiming that it is less cost effective to install a low NOx burner on the “standby” 71 MMBtu/hr boiler due to the 100,000 million cubic feet gas consumption limit that applies to the unit on a 12- month basis, ATK needs to document how that differs from the other 71 MMBtu/hr boiler’s operations, especially because all of the gas-fired boilers at the ATK Promontory site are subject to a rolling 12-month limit on natural gas consumption of 1,046,000,000 standard cubic feet of natural gas per 12-month period.^[71]  Further, the operating hours and days of the higher NOx- emitting boiler that did install an ultra-low NOx burner are not given and it is not clear that the 71 MMBtu/hr boiler with the recently-installed ultra-low NOx burner is operated continuously. As previously stated, all of the boilers at the ATK site are subject to a total 12-month gas limit of 1,046,000,000  standard cubic feet per 12 month period, as stated above.^[72] Thus, the operating hours of the 71 MMBtu/hr boiler with the ultra-low NOx burner is also somewhat limited, and yet ultra-low NOx burners were still considered cost effective.  Indeed, an ultra-low NOx burner with a NOx emission rating of 9 ppm has been installed and thus found cost effective for a 12.55

MMBtu/hr boiler at the ATK Promontory site.^[73] If such controls on similarly and smaller sized gas-fired boilers have been found to be cost effective, than such controls must be required as BACT for the currently uncontrolled 71 MMBtu/hr boiler as well as the other four gas-fired boilers of similar or greater heat input^[74] to the 12.55 MMBtu/hr boiler.  As EPA has stated, when a similar source has installed a control technology, it should be considered cost effective for the source in question, absent significant cost differences for the source being evaluated for BACT.^[75]

UDAQ seems to have accepted these discrepancies in the NOx BACT analyses for the natural gas-fired boilers without question.^[76] DAQ must adequately address and document why upgrading the higher NOx-emitting boiler is not justified as BACT.

In addition, similar to the diesel-fired boilers, all of the gas-fired boilers are subject to periodic tune-up requirements, with some units are subject to more frequent tune-up requirements based on size of the boiler.^[77] These requirements appear to be based on provisions in 40 CFR Part 63, Subpart DDDDD.¹⁶⁴ Regardless of whether a boiler is subject to Subpart DDDDD, these requirements clearly could be required on all gas-fired boilers annually (as is currently required for boilers with a heat input capacity greater than 10 MMBtu/hr pursuant to Condition II.B.31.b.(3) of the ATK Title V permit).  Boiler tune-ups can lower NOx emission rates from gas-fired boilers, among other pollutants, and more frequent tune-ups can more consistently ensure lower emission rates.  Thus, DAQ and ATK must at the minimum consider annual tuneups for all gas-fired boilers.  And, as discussed above, for those ATK gas-fired boilers of heat input capacity of 12.55 MMBtu/hr heat input or greater, ultra-low NOx burners should be considered as BACT unless ATK can show significant differences in costs of this control for the gas-fired boilers that are not currently equipped with this control at the ATK Promontory site.

g. BACT Was Improperly Applied to Hill Air Force Base.

Hill Air Force Base (HAFB) is located in Davis and Weber Counties about 30 miles north of Salt Lake City, and has industrial facilities for painting, paint stripping, plating, parts warehousing/ distribution, and wastewater treatment.^[78] These comments focus on BACT for SO2 for the units that fire diesel fuel, as well as BACT for the 97 emergency generators and the three landfill gas generators at the Hill Air Force Base for which the PM2.5-impacting pollutants are NOx, PM2.5 and VOCs.^[79]

It is not clear which generators at HAFB primarily fire diesel fuel and which primarily fire natural gas.  The HAFB BACT submittal implies that the generators fire both natural gas and diesel and cites to variable fuel sources as a reason to not eliminate the most effective NOx control – SCR – from the BACT evaluation.^[80] However, presumably these generators predominately fire one source of fuel.  For example, in the Title V permit for HAFB, the “Aggregated Boiler Group” and the “NSPS Boilers” are described as “natural gas-fired” boilers that are capable of burning diesel and other fuels.^[81] Thus, it appears these boilers burn primarily natural gas, but it is not clear.  For the units identified in the Title V permit as the “Grandfathered boilers,” the permit indicates that these units are fueled by natural gas, diesel, and other fuels

and, unlike the “Aggregated Boiler Group” and the “NSPS Boilers,” the permit does not describe the “Grandfathered boilers” as natural gas fired boilers.^[82]  Distinguishing the primary type of fuel burned in typical operation is an important part of evaluating BACT for an emissions unit. DAQ must consider as a BACT measure limiting the type of fuel burned to natural gas which is much lower in PM2.5 and precursor emissions than diesel, due to little to no particulate or SO2 emitted from natural gas-fired units.  While the HAFB BACT submittal states that limiting the use of fuels to only natural gas is not technically feasible due to Air Force readiment requirements, HAFB indicates that “it is feasible to limit the use of alternative fuels to the minimum required to sustain the mission of the facility and periods of natural gas curtailment.”^[83] Yet, HAFB’s BACT submittal did not contain specific information on the actual use of diesel and other fuels compared to natural gas at the HAFB generators, nor did HAFB propose a limit on the use of diesel and other fuels.  Given that HAFB indicated it could limit the amount of alternative fuels, UDAQ must consider imposing a numerical limit on total amount of fuels fired for fuels other than pipeline natural gas in the generators.  DAQ also must quantify whether such a limit equates to a reduction from past practice or if it would simply equate to a cap on future practices.

ii. BACT for Generators When Firing Diesel

The HAFB BACT submittal does not discuss BACT for SO2 when the generators are firing diesel fuel. DAQ’s BACT evaluation report does list several measures regarding limiting hours of operation and use of good combustion practices as well as ultra-low sulfur fuel. UDAQ claims these measures represent BACT and are being implemented by HAFB^[84]

However, a review of the HAFB Title V permit indicates that only the NSPS Compression Ignition Internal Combustion Engine (Unit # 55) is limited to ultra-low sulfur diesel (<0.0015% sulfur content).^[85] Other than the specific requirement for diesel at Unit #55, the sulfur content of diesel fuels burned is allowed to be much higher.  Specifically, Condition II.B.9.b. of the HAFB Title V permit limits sulfur content of diesel fuel to no greater than 0.5% by weight, which is more than 300 times higher than the sulfur content specifications for ultra-low sulfur diesel fuel.  At the minimum, DAQ must require all diesel used at HAFB to meet ultra-low sulfur diesel requirements of less than 0.0015% sulfur content by weight.

In addition, all generators that fire diesel should, at the minimum, be subject to annual tune-up requirements to control NOx and VOC emissions.  The requirements in 40 C.F.R. Part 63, Subpart DDDDD for annual boiler tune-ups could readily be required on all HAFB generators.  Boiler tune-ups can lower NOx and PM2.5 emission rates, among other pollutants, and more frequent tune-ups can more consistently ensure lower emission rates.  Thus, DAQ and HAFB must at the minimum consider annual tune-ups for all generators that fire diesel fuel.

DAQ listed good combustion practices, proper equipment operation and maintenance, and use of ultra-low sulfur fuel as being selected as BACT but did not impose any new requirements on the diesel-fired units at HAFB, claiming that “[i]mplementation is complete” at HAFB.^[86] As demonstrated above, there are additional requirements that DAQ should impose on the dieselfired units to ensure complete implementation of the measures DAQ found to meet BACT, including the requirement for all units to use ultra-low sulfur diesel and the requirement for all units to be subject to annual boiler tune-up requirements.

ii. BACT for Gas-Fired Generators

At the minimum, all natural gas-fired boilers should be subject to the annual tune-up requirements in 40 C.F.R. Part 63, Subpart DDDDD as part of DAQ’s BACT determination. Boiler tune-ups can lower NOx emission rates from gas-fired boilers, among other pollutants, and more frequent tune-ups can more consistently ensure lower emission rates.  Thus, DAQ and HAFB must at the minimum consider annual tune-ups for all gas-fired boilers, regardless of whether a boiler is subject to Subpart DDDDD.

With respect to additional measures to reduce NOx from the gas-fired generators, HAFB stated that there is not sufficient space in the buildings that house the generators to retrofit the generators with ultra-low NOx burners.^[87] With respect to selective catalytic reduction, HAFB said it requires exhaust gas temperatures in the range of 500 to 1,200 degrees Fahrenheit and that it is above the “designed exhaust temperature of the existing boilers at Hill AFB.”^[88] While the HAFB BACT Addendum cites to a Cleaver Brooks 2010 statement for this claim,¹⁷⁶ Cleaver Brooks did not indicate that SCR was technically infeasible in its letter to HAFB in Attachment 1 of the HAFB BACT Addendum.  Instead, Cleaver Brooks indicated that the SCR option “would only apply to the larger boilers (40-60 MMBtu).”^[89] There are at least nine generators sized within the 40-60 MMBtu/hour range at HAFB for which SCR could thus be considered as BACT.  Further, HAFB is incorrect in stating that SCR “requires” flue gas temperatures in the range of 500 to 1,200 degrees Fahrenheit.  Instead, that temperature range reflects typical conditions for SCR, but SCR can remove NOx at lower temperatures down to 300 degrees Fahrenheit.^[90] It also must be noted that it is not clear what HAFB means by the SCR temperature window being above the “designed” exhaust temperatures of the existing boilers. Because these generators are able to utilize different fuels, it is not clear what design temperature HAFB is referring to (i.e., is the design temperature reflective of design with a certain type of fuel?).  Given the various types of fuel that these boilers were designed to burn, it is more important to know the actual flue gas temperatures of the generators at HAFB to determine whether or not SCR could be successfully used.  There also may be lower temperature SCR catalysts available.^[91]   

For all of these reasons, HAFB’s BACT analysis is flawed and incomplete for SCR.  In its BACT evaluation, DAQ claimed that SCR was not technically feasible due to current boiler limitations and spacing, but space limitations were not the primary reason identified by HAFB for discounting SCR.^[92] DAQ must require additional analysis of SCR, especially given that Cleaver Brooks indicated that SCR was technically feasible for the larger boilers in the 40-60 MMBtu/hour range.^[93] Further, DAQ must provide documentation for its claim that installation of SCR is not technically feasible at any HAFB boiler due to space constraints.

HAFB’s BACT Addendum also indicates that “[s]everal projects are under consideration for removing and replacing boilers at various locations” and that HAFB has made funding requests for the replacement boilers to be equipped with ultra-low NOx burners.^[94] HAFB seems to indicate these projects are “currently underway” with the main issue being the timeline for completion. ¹⁸³ The fact that the timeline for completion is not known should not justify elimination of boiler replacement as a NOx BACT control option.  The definition of best available control measures includes any technologically and economically feasible control measure that can be implemented in whole or in part within 4 years after reclassification of a nonattainment area from moderate to serious.  40 C.F.R. 51.1000.  As long as a boiler replacement program could be partially implemented by December 2019, it should be considered as a BACT measure.  While HAFB appears to have claimed that boiler replacement is not economically feasible, the fact that they are in the process of doing so indicates that it is economically feasible for HAFB (and maybe is even warranted due to the age of the boilers HAFB is replacing).  Further, if DAQ requires boiler replacement as a BACT measure for its nonattainment plan, then HAFB would have that SIP requirement to put before Congress for budgetary approval.

DAQ has claimed that ultra-low NOx burners are not technically feasible to install on existing boilers due to space limitations, but UDAQ did not evaluate the replacement of the boilers with new boilers with ultra-low NOx burners as a BACT measure.^[95] DAQ must conduct such an analysis.  At the minimum, UDAQ should identify those boilers which HAFB is planning to replace with new boilers with ultra-low NOx burners and specifically require such replacements as a BACT control measure.

h. BACT Was Improperly Applied to Compass Minerals

The Compass Minerals Ogden Inc. (Compass) mineral recovery facility produces sodium chloride, sulfate of potash, and magnesium chloride.  Emissions sources include several dryers (20-40 MMBtu/hour range), two 108 MMBtu/hr natural gas-fired boilers, and bulk loadout and storage.^[96]DAQ’s BACT evaluation generally found that no new controls or lowered emission limits were necessary to meet BACT.¹⁸⁶

i. BACT for the Two 108.11 MMBtu/hour Boilers #1 and #2

In the case of NOx control for the 108.11 MMBtu/hour Boilers #1 and 2, UDAQ found that add on NOx controls including SCR and SNCR were not cost effective.^[97]  

Similarly, DAQ found that Flue Gas Desulfurization (FGR) also was not cost effective.¹⁸⁸  Assuming DAQ’s cost basis was based on Compass’ BACT cost analyses, the BACT analysis for the Compass emission sources is flawed.  Compass assumed only a 20-year life in determining the annualized costs of control.  At the minimum, a 30-year life should be assumed. ^[98]DAQ should assume a more appropriate and longer lifetime of controls which will reduce the annualized costs and may make one or more of these controls more cost effective.

ii. BACT for Fugitive Dust Emissions

Compass’ BACT analysis for fugitive emissions has several flaws.  First, the BACT analysis does not provide calculations of current actual and potential emissions for fugitive emissions sources, and the BACT analysis fails to adequately document how emissions were determined.  For example, for Item # 1.07 (SALT Fugitive outdoor uncaptured material handling), Compass’ description of emission estimates states:

Emissions from outdoor salt handling operations have been estimated using appropriate AP-42 emission factors.  Where PM2.5-specific emission factors were unavailable, particle size multipliers from Chapter 13.2.4 of AP-42 were utilized to adjust from Total PM or PM10 to PM2.5.  Due to the hygroscopic nature of salt, moist salt (i.e., salt hauled from evaporation ponds handled prior to drying) is assumed to have reduced emissions equivalent to 90% control efficiency when compared to dry salt emission factors.

Compass 2017 BACT submittal at 117.  Compass provided a similar description for the emissions estimates for Item 2.11 (SOP Fugitive outdoor uncaptured material handling).^[99]

Compass should have provided the emissions calculations for these sources, providing the amounts of materials handled.  Further, it is not clear what silt content was assumed for the emission factors. In addition, Compass provided no basis for the assumed 90% control efficiency for moist salt emissions, did not identify the moisture content of moist salt, and did not identify the amount of salt considered to be moist salt versus the amount of salt considered to be dry salt.  Compass should have more clearly spelled out its emissions calculations for these and other fugitive emission sources, so it can be ascertained whether Compass accurately calculated emissions from these sources.

In addition, the emissions assumed for calculating emission reductions from fugitive dust sources , in the BACT cost effectiveness analyses, do not seem to correlate with the allowable emissions calculated and are often times lower.  For example, in Attachment 2 of its 2017 BACT submittal, Compass indicates that it assumed allowable emissions for Item Nos. 1.07 and 2.11, but the assumed emissions for the cost analyses for each emissions group are much lower than the assumed allowable emissions identified in Attachment 2 of Compass’ BACT submittal.  This is illustrated in the table below.

Comparison of Compass’ Actual Emissions for Two Fugitive Dust Emissions Sources and the Emissions Assumed as Baseline in the BACT Cost Effectiveness Analyses for those Emissions Sources.

Item #	Description	Allowable Emissions[100]	Actual Emissions identified by Compass [101]	Assumed Baseline Emissions for BACT Cost Effectiveness Analyses[102]
1.07[103]	SALT Fugitive Outdoor Uncaptured Material Handling	12.887	12.887	0.586 (total from below)
1.07a				0.089
1.07b				0.027
1.07c				0.47
2.11[104]	SOP Fugitive Outdoor Material Handling	24.748	24.748	13.03 (total from below)
2.11a				6.410
2.11b				1.02
2.11d				1.80

2.11e				1.33
2.11f				0.09
2.11g				1.59
2.11h				0.79

Not only did Compass assume a much lower baseline in the BACT cost effectiveness analyses for Items # 1.07 and 2.11, but Compass also subdivided these fugitive dust sources and the potential BACT controls (i.e., full enclosures with and without ducting to air pollution control equipment) into subgroupings (i.e., 1.07a, 1.07b, 1.07c) without providing any explanation or diagrams explaining why these emissions subgroups could not be included in one enclosure which could greatly reduce the costs of an enclosure and ducting to air pollution controls.  Both

Item 1.07 and Item 2.11 are already in separate subgroups of the same source type (i.e., “fugitive emissions from outdoor uncaptured material handling”), which was presumably done based on location of the fugitive dust sources at the plant site.  Without any further explanation, it does not seem justified to break these sources up into smaller subgroups. Had Compass grouped each of these subgroups together for the cost of the enclosure, assumed a 30-year (or greater) life of the enclosure, and assumed allowable emissions that were properly calculated, the use of an enclosure and routing to air pollution controls could be quite cost effective for reducing fugitive PM2.5 emissions from these and other similar sources at the Compass facility.

With respect to BACT for the SALT fugitive salt pile and road dust emissions, Compass identifies certain fugitive dust control measures and the facility’s Fugitive Dust Control Plan that are apparently required in Compass’ existing Title V permit.  However, all of the fugitive dust control measures are listed as “state-only” provisions in the Title V permit.^[105] PM2.5 BACT measures must become part of the SIP and therefore must be federally enforceable provisions.  In addition, these provisions require the submittal of a fugitive dust plan that, at a minimum, must include the requirements in R307-309-6(4).^[106] But the fugitive dust plan (which presumably has already been developed since it is required in the permit) has not been provided as part of the BACT submittal for the Compass facility.  If DAQ and Compass are relying on the fugitive dust plan to meet BACT as Compass has proposed, that plan must be made publicly available for review and comment.

DAQ’s BACT analysis for fugitive dust emissions does not take any of the above analysis into consideration because DAQ did not conduct a site-specific evaluation of BACT for fugitive emissions at Compass Minerals.  Instead, DAQ addressed various facility’s fugitive dust sources in its “BACT for Small Sources” document.^[107] The analysis of fugitive dust control in the “BACT for Small Sources” document is very general and does not constitute a case-by- case analysis of BACT.  DAQ must instead evaluate BACT for fugitive emissions from Compass Minerals based on a case-by-case source specific analysis of BACT which properly addresses the deficiencies in Compass’ BACT analyses discussed above.

i. BACT Was Not Properly Applied to Geneva Nitrogen

It does not appear that DAQ has done its own BACT evaluation for Geneva Nitrogen. Based on statements made at the August 1, 2018 information meeting, it appears DAQ did not conduct a BACT analysis for Geneva Nitrogen because it reduced emissions below major source levels. 

UDAQ must explain in detail why it excluded Geneva Nitrogen from a BACT analysis.

If the plant reduced emissions below the 70 ton per year major source threshold, that would not exempt UDAQ from evaluating BACT for the facility.  BACM including BACT must be evaluated for all sources in the Utah serious PM2.5 nonattainment areas.  Given that the company submitted a BACT analysis, it must be considered to be a major source of PM2.5 or PM2.5 precursors.  Below we provide comments on the company’s submitted BACT analysis.

Geneva Nitrogen LLC manufactures solid ammonium nitrate in a three step process:

1. Nitric acid production
2. Ammonium nitrate solution production
3. Solid ammonium nitrate production.^[108]

Geneva Nitrogen states that the prill tower emissions of PM10 and PM2.5 are uncontrolled.^[109]   Geneva Nitrogen also states that the current wet scrubbing system used during the ammonium nitrate solidification process is not considered to be BACT for the process.^[110]

Geneva Nitrogen then states as follows:

1. Abandoning the existing wet scrubbers and ducting the rotating drum air streams directly through a common mist elimination module would remove a large majority of the ultra-fine particulate matter currently emitted in the wet scrubber exhaust.  The PM10/2.5 emissions (fines) would be captured, placed in to solution, and recycled back into the AN process.  This would meet or exceed the Best Available Control Technology requirement.

• By retrofitting the existing prill tower with an air duct, the tower-exhaust could be brought to ground level and pulled through a mist elimination module designed to eliminate a large majority of the ultra-fine particulate matter. The PM10/2.5 emissions (fines) would be captures, placed into solution, and recycled back into the AN process.  This would meet or exceed the Best Available Control Technology requirement.

August 2017 Geneva Nitrogen BACT Addendum at 26.

Despite admitting that the prill tower is not equipped with BACT, and that technology exists that meets BACT, Geneva Nitrogen dismissed routing the rotating drum air streams and the prill tower exhaust through a common mist elimination module.  Geneva Nitrogen claimed these options would be “very expensive.”  However, if other similar sources have installed the same controls, then Geneva Nitrogen would have to demonstrate that unusual circumstances exist at its facility that would prevent the successful implementation of that control as BACT and/or which distinguish it from other sources which have implemented such controls.^[111] Geneva Nitrogen also indicated that this control option was “likely physically infeasible in the case of ducting the existing prill tower discharge (220ft) to ground level due to load requirements on the tower structure built in 1957” and that it “would most likely also require replacement of the entire prill tower structure.”^[112] DAQ must require Geneva to investigate this control further, to determine and document whether it is feasible or not to duct the existing prill tower discharge to the ground level.  If the ducting could be done with a new prill tower structure, that alone is not a reason to eliminate this control option.  Instead, the costs for constructing a new prill tower to replace the 60-year old existing prill tower can be determined and considered in a cost effectiveness analysis.  Given that the existing prill tower has been operating for 60 years, such a cost analysis should consider a similar lifetime for a new prill tower.  Even if the cost of building a new prill tower was not reasonable, Geneva Nitrogen must still be required to evaluate the cost effectiveness of ducting the rotating drum air streams directly through a common mist elimination module.

Last, Geneva Nitrogen found these controls technically infeasible because, “[e]ven if physically feasible[,] initial engineering estimates indicate a mist eliminator cannot be installed and tested prior to the December 31, 2019 deadline.”^[113]  December 31, 2019 is the initial attainment date for the Salt Lake and Provo serious PM2.5 nonattainment areas.^[114]  While, optimally, BACT controls should be implemented by December of 2019, BACT controls that can partially be implement by the end of 2019 should be considered.  There is no evidence in the record, including put there by DAQ, to suggest that these control could not be partially implemented by December 2019.^[115]

Geneva Nitrogen did provide cost information for this control in its 2017 BACT Addendum, although there is limited documentation for its cost estimate.  Geneva Nitrogen provided an annualized cost estimate of the mist elimination system to be $717,667 per year, assumed only 70% PM2.5 control (“[a]bsent adequate time to complete a detailed engineering study on this project”), and determined the cost effectiveness was $7,900/ton.^[116] It does not appear that Geneva Nitrogen took into account the reduction in ammonia emissions from this control as well, which would have made the control more cost effective.  However, even with these costs (which are not significantly unreasonable), the fact is that this control has been required on a similar source, i.e., El Dorado Chemical in Arkansas.  If a similar source has had to install a particular control to meet BACT, then that control is also considered BACT for similar sources absent unusual circumstances.  Geneva Nitrogen did not identify any unusual circumstances to eliminate this control from its BACT analyses for the rotating drum air streams and the prill tower exhaust.

For all of these reasons, Geneva Nitrogen’s BACT analysis is flawed and incomplete. DAQ must require Geneva Nitrogen to update the analysis with more documented support for its calculated cost effectiveness.  Further, absent unusual circumstances at Geneva Nitrogen, it seems the mist eliminator system should be required as BACT for the rotating drum air streams and the prill tower exhaust given that the same control has been required as BACT for another similar source.  DAQ must consider these issues in its own BACT analysis for Geneva Nitrogen.

j. BACT Was Not Properly Applied to Proctor & Gamble.

Proctor & Gamble owns and operates a paper, assembled paper products, and manufacturing process with two separate product lines:  a paper process line and an assembled paper products line.  The company recently obtained a construction permit which authorized the construction of additional production lines including the addition of two 50 MMBtu/hour boilers for process steam, comfort heating, cooling water, and back-up power.^[117] It is not clear whether those boilers have been constructed yet, as UDAQ’s BACT Evaluation Report has statements indicating that

the units have not yet been constructed as well as statements that the units have been constructed.^[118] The boilers had not been installed at the time of Proctor & Gamble’s April 2017 BACT Analysis submittal to UDAQ.^[119]  The comments below focus on NOx BACT for these two new boilers.

Proctor & Gamble presented a NOx BACT analysis for the new 50 MMBtu/hour boilers in its

April 2017 submittal.^[120] The company found that SCR was technically feasible for the new 50 MMBtu/hour boilers, but claimed that SCR with the planned ultra-low NOx burners would not be economically reasonable.²¹² The details of their cost analysis is purportedly in Appendix A of their April 2017 BACT submittal, but Appendix A is not available on DAQ’s website nor was Appendix A included in UDAQ’s BACT Evaluation.^[121] Yet, DAQ appears to find the company’s cost analysis for SCR acceptable, as UDAQ cites the same NOx cost effectiveness value of $165,250/ton as Proctor & Gamble claimed for SCR at the two new 50 MMBtu/hour boilers.^[122] Given that the details of Proctor & Gamble’s cost analysis were not included in its BACT submittal to UDAQ, the basis for UDAQ’s concurrence that SCR is unreasonable for the new boilers is not justified.

A review of the limited details on Proctor & Gamble’s cost analysis shows significant flaws. 

First, the company assumed that SCR would reduce NOx emissions from the 10 ppm

NOx emission rate achievable with ultra-low NOx burners down to 9 ppm,^[123] which only reflects a NOx reduction of 10%.  Yet, Proctor & Gamble as well as UDAQ claimed that SCR can achieve 70-90% NOx control.^[124] Thus, Proctor & Gamble failed to evaluate cost effectiveness for SCR at the highest levels of NOx control efficiency that SCR could achieve, which would result in improperly inflated dollar per ton costs. The BACT analysis must evaluate the maximum degree of emission reduction achievable with a pollution control.

Second, Proctor & Gamble evaluated SCR cost effectiveness using the NOx emission rate with ultra-low NOx burners as reflective of baseline emissions for the cost analysis.^[125] However, as discussed above, BACT is based on essentially uncontrolled emissions, calculated using a “realistic scenario of upper boundary uncontrolled emissions.”^[126] Proctor & Gamble should have thus evaluated the suite of controls of ultra-low NOx burners and SCR together in its BACT cost effectiveness analysis.

Third, Proctor & Gamble should have evaluated the possibility of routing the flue gas from each boiler to one SCR to save costs.  If the proximity of the boilers allows for it, this could be a significant cost saving measure and ensure the lowest NOx rates from these two new boilers.

In summary, DAQ must more fully investigate SCR as BACT for these two new boilers to ensure the maximum degree of NOx reduction is achieved.  DAQ must also insure that appropriate interest rates (i.e., no higher than 7%) and lifetime of controls (i.e., 25-30 years) were assumed in the SCR cost effectiveness analysis.  Further, DAQ must make the details of the SCR cost effectiveness analysis available to the public for review and comment.  SCR has been required on similarly sized boilers, and thus UDAQ must more adequately justify any decision to not require SCR on the two new 50 MMBtu/hour boilers at the Proctor & Gamble facility.

5.  The Record Does Not Support EPA’s Proposed Approval of Utah’s BACT.

In the TSD that EPA claims contains its analysis of, inter alia, Utah’s BACT, EPA does little more than recite and “agree” with Utah’s consideration of BACT.  Particularly in light of the comments above, this perfunctory “review” of Utah’s BACT is not sufficient to support the agency’s proposal to approve Utah’s BACT.  Rather than acquiescing to Utah’s BACT uncritically, EPA is obligated to undertake a rigorous and independent examination of the BACT that should apply to major sources in the Salt Lake City and Provo nonattainment areas.

IV.  Conclusion

Based on the above, we ask EPA to disapprove those aspects of its proposed decision addressed above, including the relevant portions of IX.H.11, 12 and 3 and Utah’s Fugitive Dust Rule.  As established above, these proposed rules and SIP provisions are not BACM/BACT and are not supported by adequate analysis or record evidence. Furthermore, as the record shows, EPA has not undertaken the review and examination necessary to approve IX.H.11, 12 and 3 and Utah’s Fugitive Dust Rule. 

Thank you for your consideration of these comments.

JORO WALKER, Esq. Western Resource Advocates and Attorney for Signatories

---

[1] Independent Particulate Matter Review Panel.  The Need for a Tighter Particulate Matter Standard (June 2020). N Engl J Med. at 1. https://www.nejm.org/doi/10.1056/NEJMsb2011009

[2] Id. at 2. ⁴ Id. at 3.

[3] Independent Particulate Matter Review Panel.  Letter to Administrator Wheeler: Advice from the Independent Particulate Matter Review Panel (October 22, 2019) at 4. https://yosemite.epa.gov/sab/sabproduct.nsf/81DF85B5460CC14F8525849B0043144B/%24File/ Independent+Particulate+Matter+Review+Panel+Letter+on+Draft+PA.pdf ⁶ The Need for a Tighter Particulate Matter Standard at 3.

[4] Advice from the Independent Particulate Matter Review Panel at B-27 (“The forms of the current annual (3-year average) and 24-hour (98th percentile) primary PM_2.5 standards are appropriate in conjunction with revised levels.”).

[5] Wu X, Nethery RC, Sabath BM, Braun D, Dominici F. Exposure to air pollution and COVID19 mortality in the United States. medRxiv 2020: 2020.04.05.20054502; Liang D, Shi L, Zhao J, et al. Urban Air Pollution May Enhance COVID-19 Case Fatality and Mortality Rates in the United States. medRxiv 2020: 2020.05.04.20090746.

[6] Travaglio M, Yu Y, Popovic R, Selley L, Leal NS, Martins LM.  Links between air pollution and COVID-19 in England. medRxiv 2020.04.16.20067405.

[7] Conticini E, Frediani B, Caro D. Can atmospheric pollution be considered a co-factor in extremely high level of SARS-CoV-2 lethality in Northern Italy? Environmental Pollution 2020:

[8] .

[9] Ogen Y. Assessing nitrogen dioxide (NO₂) levels as a contributing factor to coronavirus (COVID-19) fatality. Sci Total Environ 2020; 726: 138605.

[10] Y. Zhu, J. Xie, F. Huang, L. Cao. Association between short-term exposure to air pollution and

COVID-19 infection: evidence from China Sci. Total Environ., 727 (2020)

[11] Wu X, Nethery RC, Sabath BM, Braun D, Dominici F. Exposure to air pollution and COVID19 mortality in the United States. medRxiv 2020: 2020.04.05.20054502

[12] https://www.cdc.gov/coronavirus/2019-ncov/need-extra-precautions/racial-ethnicminorities.html

[13] While BACT provisions in CAA and EPA regulations often use the term “facility” or “source,” federal regulations use “emissions unit” in some contexts. For example, 40 C.F.R. § 52.21(j)(3) states: “A major modification shall apply best available control technology to each regulated NSR pollutant for which it would result in a significant net emissions increase at the source. The requirement applies to each proposed emissions unit at which a net emissions increase in the pollutant would occur as a result of a physical change or change in the method of operation of the unit.”

[14] For example, 40 C.F.R. 102a(b)(2) sets as an emission limit on the FCCU: “Sulfur dioxide (SO₂) in excess of 50 ppmv dry basis corrected to 0 percent excess air, on a 7-day rolling average basis and 25 ppmv, dry basis corrected to 0 percent excess air, on a 365-day rolling average basis.”  IX.H.g.i.A.I states “[e]ach owner or operator of an FCCU shall comply with an SO₂ emission limit of 25 ppmvd at 0% excess air on a 365-day rolling average basis and 50 ppmvd at 0% excess air on a 7-day rolling average basis.”

[15] Only if “the director determines that technological or economic limitations on the application of measurement methodology to a particular emissions unit would make the imposition of an emissions standard infeasible, a design, equipment, work practice, operational standard or combination thereof, may be prescribed instead to satisfy the requirement for the application of best available control technology. Such standard shall, to the degree possible, set forth the emissions reduction achievable by implementation of such design, equipment, work practice or operation, and shall provide for compliance by means which achieve equivalent results.” Utah Admin. Code r.307-401-2.

[16] In this section, we use BACT and BACM interchangeably.

[17] To the extent that Utah purports to rely on fugitive dust control plans – existing or otherwise – to constitute BACT for Subpart H sources that emit fugitive dust, this approach is inappropriate. Subpart H must specifically set forth all emission limits and/or work practices, operational standards or similar means of control applicable to the Subpart H sources.

[18] http://www.aqmd.gov/docs/default-source/rule-book/rule-iv/rule-403.pdf?sfvrsn=4

[19] https://www.maricopa.gov/DocumentCenter/View/5354/Rule-310—Fugitive-Dust-from-Dust-

Generating-Operations-PDF?bidId=

[20] http://www.clarkcountynv.gov/airquality/regulations/Documents/SECT94_07-01-04.pdf

[21] https://www.washoecounty.us/health/files/airquality/Current%20Regulations/040%20Prohibited%20Emissions_Mar%202017.pdf ²⁴ Appendix A-4 to Part 60 – Test Methods 6 through 10B at 318. 

[22] Id. (“The lidar is used to measure plume opacity during either day or nighttime hours because it contains its own pulsed light source or transmitter. The operation of the lidar is not dependent upon ambient lighting conditions (light, dark, sunny or cloudy).”).

[23] Draft Staff Report, Proposed Amended Rule 1466 – Control of Particulate Emissions from Soils with Toxic Air Contaminants at 11; https://www.aqmd.gov/home/rulescompliance/compliance/rule-1466/pre-approved-monitors (list of preapproved monitors).

[24] Draft Staff Report at 11.

[25] WRA also attached to its Subpart A comments 25 pages of analysis of fugitive dust rules adopted in other states.  

[26] https://documents.deq.utah.gov/air-quality/pm25-serious-sip/DAQ-2017-011687.pdf

[27] DAQ makes this improper contention with regard to other fugitive dust rules adopted in other states.

[28] DAQ also states “Washoe, Clark and Maricopa Counties have more stringent high wind requirements. Generally speaking, they include the option to cease operations or institute contingency measures. Impetus: serious nonattainment.”  Id. Yet, the agency did not address this more stringent measure in its BACM for Rule 307-309.

[29] We were given a few examples of watering records.  These records do not, for example, indicate the amount of water that was applied or the size of the area to which it was applied.  The records do not indicate the temperature, wind speed, or other relevant conditions.

[30] The comments in this report generally refer to all the emissions control analyses as BACT analyses, even though in some cases the controls discussed are best available control measures rather than control technologies.  “Best Available Control Measures” include BACT by definition, pursuant to 40 C.F.R. §51.1000.

[31] July 1, 2018, Utah Division of Air Quality, PM_2.5 SIP Evaluation Report – Kennecott Utah

Copper LLC-BCM and Copperton Concentrator, Utah PM_2.5 Serious SIP, Salt Lake City Nonattainment Area, DAQ-2018-007709 (hereinafter referred to as “UDAQ BACT Report KUC BCM and Copperton Concentrator”).

[32] April 27, 2017 Kennecott Utah Copper LLC, PM_2.5 SIP Best Available Control Technology Analysis, submitted by Rio Tinto Kennecott to the Utah Division of Air Quality, (hereinafter referred to as the “Kennecott April 2017 BACT Analysis”).

[33] DAQ BACT Report KUC BCM and Copperton Concentrator at 5; Kennecott April 2017 BACT Analysis at 3-1.

[34] Kennecott April 2017 BACT Analysis at 3-1.

[35] DAQ BACT Report KUC BCM and Copperton Concentrator at 5.

[36] January 10, 2018 Approval Order DAQE-AN105710042-18, Condition II.B.1.a.

[37] DAQ BACT Report KUC BCM and Copperton Concentrator at 6.

[38] See https://www.arb.ca.gov/bact/category/rock.htm 

[39] DAQ BACT Report KUC BCM and Copperton Concentrator at 7; Kennecott April 2017 BACT Analysis at 3-2.

[40] UDAQ BACT Report KUC BCM and Copperton Concentrator at 8-9; Kennecott April 2017 BACT Analysis at 3-2 to 3-3.

[41] EPA, Control of Open Fugitive Dust Sources, September 1988, at 3-12.

[42] DAQ BACT Report KUC BCM and Copperton Concentrator at 10; Kennecott April 2017 BACT Analysis at 3-3.

[43] Kennecott April 2017 BACT Analysis at 3-3; DAQ BACT Report KUC BCM and Copperton Concentrator at 11.

[44] See, e.g., EPA’s October 1990 New Source Review Workshop Manual at B.20 (“Physical modifications needed to resolve technical obstacles do not in and of themselves provide a justification for eliminating the control technique on the basis of technical infeasibility.  However, the costs of such modifications can be considered in estimating cost and economic impacts…”).

[45] The information on these entries into EPA’s RACT/BACT/LAER Clearinghouse can be found by running a “standard search” at https://cfpub.epa.gov/rblc/index.cfm?action=Search.StandardSearch and searching on the RBLC ID number.

[46] EPA, Control of Open Fugitive Dust Sources, September 1988, at 3-12.

[47] DAQ BACT Report KUC BCM and Copperton Concentrator at 26.

[48] April 2017 BACT Determinations for Utah Power Plant, Tailings Site, and Laboratory at 3-6.

[49] EPA, Control of Open Fugitive Dust Sources, September 1988, at 3-12

[50] DAQ PM_2.5 SIP Evaluation Report – Kennecott Utah Copper LLC – Molybdenum Autoclave Process, Refinery and Smelter, July 1, 2018, at 4.

[51] Technical Support Document and Statement of Basis for Construction of Arizona Clean Fuels Yuma, LLC, Petroleum Refinery, Permit Number 40140, September 15, 2006, at 161, 276 (Ex.

[52] ).

[53] 40 C.F.R. §60.102(g)(1)(ii).

[54] Technical Support Document and Statement of Basis for Construction of Arizona Clean Fuels Yuma, LLC, Petroleum Refinery, Permit Number 40140, September 15, 2006, at 161 (Ex. 3). ⁵⁷ Technical Support Document for Air Emissions Permit No. 03700011-101 issued January 9, 2013 for the Flint Hills Resources Pine Bend Refinery at 9 (attached as Ex. 4).

[55] Id., Attachment 6, at 4.

[56] April 2017 Tesoro BACT Submittal, Section 5.2.2 at page 42.

[57] July 1, 2018 DAQ PM_2.5 SIP Evaluation Report: Tesoro Refining & Marketing Company LLC, at 17.

[58] Meridian Energy Group, Inc., Permit to Construct Application, Davis Refinery, at 16, available at https://www.deq.nd.gov/aq/notices/meridian/april/davisrefineryptcappapril.pdf; Final Permit to Construct for the Davis Refinery, available at https://deq.nd.gov/aq/Notices/Meridian/Final/Attach1.pdf. See also http://www.merichem.com/gas/upstream/natural-gas/lo-cat.

[59] Section 5.2.2 of Tesoro’s April 2017 BACT Assessment, at page 42.

[60] April 2017 Chevron BACT analysis for Boilers # F11005 and F11006 at 4-5, BACT analysis for Crude Unit Heater F21002 at 4-5.

[61] April 2017 Holly Refining BACT Analyses at 5-1 to 5-4.

[62] April 2017 Big West Refinery BACT Analyses at 7-11, 26-27.

[63] September 2008 Class I – Permit Renewal Application, Arizona Clean Fuels Yuma, LLC, ADEQ Permit No. 40140, at 6-21 to 6-29

[64] April 2017 Tesoro BACT analysis, Section 5.5.

[65] April 2017 Chevron BACT analyses for these units

[66] April 2017 Holly Frontier Refinery BACT analysis at 4-17.

[67] July 1, 2018 DAQ PM_2.5 SIP Evaluation Report: Holly Refining & Marketing Company – Woods Cross, LLC – Holly Refinery, at 11.

[68] July 1, 2018 PM2.5 SIP Evaluation Report:  Big West Oil, LLC, at 14.

[69] U.S. EPA, November 2015 Technical Support Document for the Cross-State Air Pollution

Rule for the 2008 Ozone NAAQS, Docket ID No. EPA-HQ-OAR-2015-0500, Assessment of Non-EGU NO_X Emission Controls, Cost of Controls, and Time for Compliance, at 12, available at https://www.epa.gov/sites/production/files/2015-11/documents/assessment_of_nonegu_nox_emission_controls_and_appendices_a_b.pdf and attached as Ex. 7.

[70] EPA’s Control Cost Manual, Chapter 2 at 2-78.

[71] Id., footnote 29.

[72] April 2017 Chevron BACT Submittal for Boiler #1 F11001, #2 F11002, and Boiler #5 F11005 at 6.

[73] U.S. EPA, October 1990 New Source Review Workshop Manual, at B.37.

[74] U.S. EPA, November 2015 Technical Support Document for the Cross-State Air Pollution

Rule for the 2008 Ozone NAAQS, Docket ID No. EPA-HQ-OAR-2015-0500, Assessment of

Non-EGU NO_X Emission Controls, Cost of Controls, and Time for Compliance, at 12, available

[75] April  2017 Tesoro BACT submittal at Section 4.3.5.

[76] Id. at Section 4.2.5.

[77] Id. at 4.1.5.  Note that Tesoro has stated the proposed PM_2.5 emissions limit will be based on testing once the wet gas scrubber has been installed.

[78] Lo TOx technology demonstration at Marathon Ashland Petroleum LLC’s Refinery in Texas

City, March 2004, available at

http://www.digitalrefining.com/article/1000813,LoTOx_technology_demonstration_at_Maratho n_Ashland_Petroleum_LLC_s_Refinery_In_Texas_City.html 

[79] July 1, 2018 PM2.5 SIP Evaluation Report:  Holly Refining & Marketing Company – Woods Cross, LLC – Holly Refinery, at 20-21.

[80] Id. at 6-14.

[81] April 2017 Big West Refinery BACT Analysis at 5.

[82] July 1, 2018 PM2.5 SIP Evaluation Report:  Big West Oil, LLC, at 7-8.

[83] Id. at 6.

[84] https://www.environmental-expert.com/products/hamon-wet-gas-scrubbers-85424#collapse181606.

[85] April 2017 Chevron BACT Analysis for FCCU at 7.

[86] April 2017 Big West Refinery BACT Analysis at 11-12.

[87] July 1, 2018 PM_2.5 SIP Evaluation Report: Chevron Products Company – Salt Lake Refinery, at 13 and 15.

[88] Id. at 3.

[89] U.S. EPA, October 1990 New Source Review Workshop Manual, at B.37. ⁹⁵ April 2017 Chevron BACT Submittal for FCCE at 7 and 11.

[90] U.S. EPA, October 1990 New Source Review Workshop Manual, at B.38.

[91] April 2017 Chevron BACT Submittal for FCCE at 15.

[92] Linde’s informational website on Lo-TOx, available at http://www.linde- gas.com/en/products_and_supply/emissions_solutions/lotox/index.html.

[93] April 2017 Chevron BACT Submittal for FCCU at 15.

[94] EPA’s Control Cost Manual at 1-52.

[95] April 2017 Chevron BACT submittal for FCCU at 2-3.

[96] April 2017 PM_2.5 Serious Nonattainment SIP BACM Analysis, Utah Municipal Power Agency, West Valley, Utah.

[97] Id.

[98] RBLC ID CA-0997

[99] U.S. EPA, October 1990 New Source Review Workshop Manual, at B.37.

[100] April 2017 PM_2.5 Serious Nonattainment SIP BACM Analysis, Utah Municipal Power Agency, West Valley, Utah, at 3-6.

[101] Section II.B.2.e. of West Valley Title V Operating Permit Number 3500527003.

[102] Section II.B.2.e. of West Valley Title V Operating Permit Number 3500527003 at 3-5.

[103] U.S. EPA, October 1990 New Source Review Workshop Manual, at B.37.

[104] April 2017 PM2.5 Serious Nonattainment SIP BACM Analysis, Utah Municipal Power Agency, West Valley, Utah, at 3-5.

[105] Id. at Appendix A, page A-3.

[106] Id. at Appendix A, page A-3.

[107] July 1, 2018 PM2.5 SIP Evaluation Report:  Utah Municipal Power Association – West Valley Power Plant, at 9-12.

[108] April 2017 BACT Analysis for Gadsby Power Plant at 1, 2-1, 4-6.

[109] RBLC ID CA-0997

[110] April 2017 BACT Analysis for Gadsby Power Plant at 4-5.

[111] July 1, 2018 PM2.5 SIP Evaluation Report:  PacifiCorp Energy – Gadsby Power Plant, at 2022.

[112] April 2017 BACT Analysis for Gadsby Power Plant at 22.

[113] Id. at 4-7. ¹²⁰ Id.

[114] Id., Appendix A.

[115] Id., Appendix A.

[116] Id. at 4-11.

[117] U.S. EPA, October 1990 New Source Review Workshop Manual, at B.37.

[118] July 1, 2018 PM_2.5 SIP Evaluation Report:  PacifiCorp Energy – Gadsby Power Plant, at 6-11.

[119] Lhoist North American BACT Analysis, 5/5/17, at 1-2.

[120] Id. at 1-3.

[121] See EPA’s policy statements on reactivation. It is highly questionable that the Lhoist facility would meet EPA’s policy requirements to be able to restart operations at this point without being subject to major source permitting. Further, if the plant starts up after 2018, it must be subject to major source PSD and nonattainment NSR permitting because those rules do not allow a source such as Lhoist to look back any farther than 10 years for determining baseline actual emissions. That is, there is no longer a provision in the permitting rules for such sources to obtain approval to use an earlier baseline period as representative of normal source operations.  See UACR R307-403-2 and reference to “baseline actual emissions” as that term is defined at 40 C.F.R. 51.165(a)(1)(vi); UACR R307-405-3(1) and definition of “baseline actual emissions” in 40 C.F.R. 52.21(b)(48)(ii).

[122] Lhoist North American BACT Analysis, 5/5/17, at 3-1 ¹³⁰ Id. at 3-15.

[123] Id. at 3-4.

[124] July 1, 2018 PM_2.5 SIP Evaluation Report:  Lhoist North America – Grantsville Facility, at 6.

[125] Lhoist North American BACT Analysis at 3-4.

[126] Id. at 3-10.

[127] July 1, 2018 PM2.5 SIP Evaluation Report:  Lhoist North America – Grantsville Facility, at 7.

[128] Lhoist North American BACT Analysis at 3-18.

[129] Id.

[130] See requirements for LHoist in Draft Section IX, Part H in which DAQ has proposed a NOx limit of 56 lb/hr (which reflects 25% removal from a current NOx limit of 75.00 lb/hr).

[131] July 1, 2018 PM2.5 SIP Evaluation Report:  Lhoist North America – Grantsville Facility, at 9.

[132] Lhoist North American BACT Analysis at 3-27.

[133] July 1, 2018 PM2.5 SIP Evaluation Report:  Lhoist North America – Grantsville Facility, at

[134] -21.

[135] Id. at 20.

[136] May 2017 ATK BACT Analysis at 1.

[137] Id. at 24.

[138] July 1, 2018 UDAQ PM2.5 SIP Evaluation Report – ATK [DAQE-2018-007203], at 13.

[139] May 2017 ATK BACT Analysis at 24.

[140] Conditions II.A.95 through II.A.111 of Title V Permit Number 300003003.

[141] Condition II.B.27.a. of Title V Permit Number 300003003.

[142] Condition II.B.30.a.B. of Title V Permit Number 300003003.

[143] Condition II.B.34.b. of the Title V Permit Number 300003003.

[144] Id.

[145] May 2017 ATK BACT Analysis at 18.

[146] Title V Permit Number 300003003 at Condition II.A.86.

[147] Title V Permit Number 300003003 at Condition II.A.90.

[148] October 2016 ATK RACT Analysis at 3.

[149] May 2017 ATK RACT analysis at 18.

[150] ATK Title V Permit Number 300003003 at Condition II.B.30.a.

[151] ATK Title V Permit at Condition II.B.30.a.A.

[152] Title V Permit at Condition II.A.86.

[153] Title V Permit at II.A.88 (two gas fired boilers of 25.11 MMBtu/hr each) and at II.A.89 (two gas fired boilers of 16.74 MMBtu/hr each).

[154] EPA’s October 1990 New Source Review Workshop Manual at B.31.

[155] July 1, 2018 UDAQ PM2.5 SIP Evaluation Report – ATK [DAQE-2018-007203], at 15.

[156] Condition II.B.31.b. of the Title V permit. ¹⁶⁴ Id.

[157] Hill Air Face Base-Main Base Title V Permit Number 1100007003 at 2.

[158] April 25, 2017 Hill Air Force Base BACT Submittal at 2-1.

[159] Id. at 2-3 (in Table 2-3).

[160] Hill Air Face Base-Main Base Title V Permit Number 1100007003 at 20 (Conditions II.A.33 and 34).

[161] Id. (Condition II.A.36).

[162] April 25, 2017 Hill Air Force Base BACT Submittal at 3-6.

[163] July 1, 2018 UDAQ PM2.5 SIP Evaluation Report: Department of the Air Force, Hill Air Force Base at pdf pages 21 and 23.

[164] HAFB Title V Permit Number 1100007003 at 114 and 115 (Condition II.B.43.b).

[165] July 1, 2018 UDAQ PM2.5 SIP Evaluation Report: Department of the Air Force, Hill Air Force Base at pdf page 23.

[166] August 18, 2017 HAFB BACT Addendum, at page 3 and Attachment 1.

[167] Id. at page 3. ¹⁷⁶ Id.

[168] Id., Attachment 1 at 2.

[169] Pritchard, Scot G., et al., SCR Catalyst Performance under Severe Operation Conditions, at 3, available at http://www.cormetech.com/brochures/Under%20Severe%20Operating%20Conditions.pdf,

[170] Tang, Xialong, Low temperature selective catalytic reduction of NOx with NH3 over amorphous MnOx catalysts prepared by three methods, available at http://www.sciencedirect.com/science/article/pii/S1566736706002111.

[171] July 1, 2018 UDAQ PM2.5 SIP Evaluation Report: Department of the Air Force, Hill Air Force Base at pdf page 10.

[172] August 18, 2017 HAFB BACT Addendum, at Attachment 1.

[173] Id. at 8. ¹⁸³ Id.

[174] July 1, 2018 UDAQ PM2.5 SIP Evaluation Report: Department of the Air Force, Hill Air Force Base at pdf page 9.

[175] October 1, 2016 UDAQ PM2.5 SIP Evaluation Report, Compass Minerals Ogden Inc. at 1. ¹⁸⁶ July 1, 2018 UDAQ PM2.5 SIP Evaluation Report: Compass Minerals-Compass Minerals Ogden Inc.

[176] Id. at 16. ¹⁸⁸ Id.

[177] EPA typically assumes a 30-year life for complicated controls such selective catalytic reduction.  See EPA’s Control Cost Manual, Chapter 2 at 2-78.

[178] 2017 Compass BACT Analysis at 121.

[179] Id. at 111.

[180] Id. at 113-114.

[181] Id. at 30-33 and 64-71.

[182] The allowable emissions were identified for all three Item #1.07 emission groups in total, although separate BACT analyses were done for 1.07a, 1.07b, and 1.08c.

[183] The allowable emissions were identified for all seven Item #2.11 emission groups in total, although separate BACT analyses were done for the seven operations.

[184] Id. at 37.

[185] Id. (under Condition II.B.1.h).

[186] July 1, 2018 UDAQ PM2.5 SIP Evaluation Report: Compass Minerals-Compass Minerals Ogden Inc. at 19; see also UDAQ’s BACT for Various Emission Units at Stationary Sources, DAQ-2018-007161, Section 12.

[187] May 2017 Geneva Nitrogen BACT submittal at 1.

[188] Geneva Nitrogen August 2017 BACT Addendum at 25.

[189] Id. at 26.

[190] EPA, October 1990 New Source Review Workshop Manual at B.19. and B.29.

[191] Geneva Nitrogen BACT Addendum at 26.

[192] Id.

[193] 82 Fed. Reg. 21712.

[194] Geneva Nitrogen August 2017 BACT Addendum at 27.

[195] Id.

[196] April 2017 Procter and Gamble Paper Products BACM/BACT Analysis at 2-1 to 2-2.  See also Approval Order

DAQE-AN141070009-16.

[197] July 1, 2018 UDAQ’s PM2.5 SIP Evaluation Report: Proctor and Gamble Paper Products Company, at pdf page 5 (“[t]he boilers will be fueled by natural gas…”) and pdf page 17 which implies the two 50 MMBtu/hour boilers already exist.

[198] April 2017 Procter and Gamble Paper Products BACM/BACT Analysis at 3-19.

[199] Id. at 3-22 to 3-24. ²¹² Id. at 3-22.

[200] Indeed, only the cover page for Appendix A was included at the end of UDAQ’s July 1, 2018 BACT Evaluation Report for Proctor & Gamble.

[201] July 1, 2018 UDAQ’s PM2.5 SIP Evaluation Report: Proctor and Gamble Paper Products Company at pdf page 22; April 2017 Procter and Gamble Paper Products BACM/BACT Analysis at 3-22.

[202] April 2017 Procter and Gamble Paper Products BACM/BACT Analysis at 3-22.

[203] April 2017 Procter and Gamble Paper Products BACM/BACT Analysis at 3-22.

[204] Id.

[205] U.S. EPA, October 1990 New Source Review Workshop Manual, at B.37