6 F.4th 1059
9th Cir.
2021

SANDRA L. BAHR; JEANNE LUNN; DAVID MATUSOW, Petitioners, v. MICHAEL REGAN, Administrator, United States Environmental Protection Agency; DEBORAH JORDAN, Acting Regional Administrator, U.S. EPA Region 9; U.S. ENVIRONMENTAL PROTECTION AGENCY, Respondents.

No. 20-70092

UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT

July 28, 2021

EPA No. EPA-R9-OAR 2018-0821

FOR PUBLICATION

OPINION

On Petition for Review of an Order of the Environmental Protection Agency

Argued and Submitted March 4, 2021 Phoenix, Arizona

Filed July 28, 2021

Before: Carlos T. Bea and Patrick J. Bumatay, Circuit Judges, and Kathleen Cardone,* District Judge.

Opinion by Judge Bea;

Concurrence by Judge Bumatay

SUMMARY**

Clean Air Act

The panel denied a petition for review challenging the Environmental Protection Agency’s findings and its conclusion that Arizona had achieved the statutory required reduction in ozone concentration by July 2018, in compliance with the Clean Air Act.

Because of a major wildfire in southeast California in 2015 (the “Lake Fire“), six air quality monitors in the Phoenix region registered abnormally high concentrations of ozone, in excess of the National Ambient Air Quality Standard (NAAQS). If not for those six exceedances, Arizona would have been able to demonstrate it had attained the ozone NAAQS by July 2018. The Clean Air Act requires the Environmental Protection Agency (“EPA“) to exclude monitoring data if a recorded exceedance was clearly caused by exceptional, uncontrollable events, such as a wildfire. In 2007, EPA had implemented an Exceptional Events Rule to

govern the treatment of such data. In September 2016, EPA revised that Exceptional Events Rule.

To demonstrate that the June 20, 2015, exceedances qualified for exclusion as influenced by exceptional events, Arizona submitted to EPA three sets of statistical demonstrations. Arizona submitted its initial demonstration while the 2007 rule was in effect but submitted its two supplemental demonstrations while the 2016 rule was in effect. EPA decided to apply the requirements of the 2016 rule and concluded that the Lake Fire did indeed cause the June 20, 2015, monitor readings. EPA then excluded the six exceedances from its NAAQS calculations.

Petitioners, citizens of Phoenix, alleged that the EPA violated the presumption against retroactivity when it applied the 2016 version of the Exceptional Events Rule because the 2007 rule had been in effect when the 2015 Lake Fire and exceedances occurred. The panel held that petitioners failed properly to raise the issue of retroactivity to allow the EPA to understand and rule on it. The panel therefore concluded that Petitioners failed to exhaust the issue as to whether the 2007 or 2016 Exceptional Events Rule governed the final agency action.

Alternatively, the panel held that even if Petitioners adequately raised the issue, the application of the 2016 Exceptional Events Rule did not impact any vested rights, create any new obligations, or otherwise impact any regulated party’s interests in fair notice, reasonable reliance, or settled expectations. Therefore, the application of the 2016 Exceptional Events Rule was not impermissibly retroactive and no presumption against retroactivity arose.

Petitioners further claimed that Arizona’s evidence did not support EPA’s finding that a clear causal connection existed between the Lake Fire and the June 20, 2015, exceedances. The panel deferred to EPA’s technical conclusions and found that Arizona adduced evidence sufficient to allow EPA to make such finding. As a result, the EPA did not act arbitrarily or capriciously by excluding the data from the six monitors as falling within the Exceptional Events Rule.

Finally, Petitioners alleged that EPA acted contrary to the Clean Air Act in suspending Arizona’s contingency measures requirement in EPA’s July 2018 final rule. Petitioners contended that the Clean Air Act requires states to provide attainment contingency measures regardless of whether the region attains the NAAQS by its attainment date. The panel held that petitioners forfeited this argument by not sufficiently raising it in their comment before the agency, but, even under a lenient interpretation of the content of their comment before the agency, the panel concluded that EPA’s construction of the Clean Air Act was owed deference under . EPA’s suspension of the contingency measure requirements in its July 2018 final rule did not violate the Clean Air Act.

Concurring, Judge Bumatay stated that the majority opinion persuasively explained how the petitioners failed to bring their objections regarding the Exceptional Events Rule and the Contingency Measures Requirement in a timely manner. Accordingly, there was no need to reach the merits of those objections here. He thus joined the majority opinion except as to sections I.B and III.B.

COUNSEL

Jennifer B. Anderson (argued) and Daniel J. Adelman, Arizona Center for Law in the Public Interest, Phoenix, Arizona, for Petitioners.

Andrew D. Knudsen (argued), Attorney; Jonathan D. Brightbill, Principal Deputy Assistant Attorney General; Environment and Natural Resources Division, United States Department of Justice, Washington, D.C.; Geoffrey L. Wilcox, Kaytrue Ting, Elizabeth Pettit, and Charlotte Withey, United State Environmental Protection Agency Office of Regional Counsel, Washington, D.C.; for Respondents.

Mark Brnovich, Attorney General; Shelley D. Cutts, Assistant Attorney General; Office of the Attorney General, Phoenix, Arizona; for Amicus Curiae State of Arizona.

Lucas J. Narducci and Michael C. Ford, Snell & Wilmer LLP, Phoenix, Arizona, for Amicus Curiae Arizona Chamber of Commerce and Industry.

OPINION

BEA, Circuit Judge:

The intensity of the Arizona sun, a burgeoning metropolitan population, and heavy vehicular traffic have hindered Arizona’s efforts to reduce the concentration of ozone in the Phoenix region below the national standard that the Environmental Protection Agency (“EPA“) had set as safe for human health (the National Ambient Air Quality Standard, or “NAAQS“) (Arizona already failed its first attempt, in 2012, to reach that standard). To come into compliance with the Clean Air Act, EPA required Arizona to demonstrate that it had achieved the statutorily required reduction in ozone concentration by July 2018. Having already failed once to reduce ozone concentration within the timeframe mandated by the Clean Air Act, Arizona would have been subject to intensified regulatory restrictions had the state failed to achieve compliance by 2018. Such compliance, as with most matters involving environmental regulation, is a complicated matter. Nonetheless, we find EPA’s conclusions that led to its finding that Arizona did timely achieve compliance were valid, and we will deny the petition for review.

On June 17, 2015, a major wildfire—the Lake Fire—broke out in the San Bernardino National Forest in southeast California. Three days later, on June 20, three hundred miles east of the fire, six air quality monitors in the Phoenix region registered abnormally high concentrations of ozone, in excess of the NAAQS. If not for those exceedances, Arizona would have been able to demonstrate it had attained the ozone NAAQS by July 2018. As it happens, the Clean Air Act requires EPA to exclude monitoring data if a recorded exceedance was clearly caused by exceptional, uncontrollable events—such as a wildfire. In 2007, EPA had

implemented an Exceptional Events Rule to govern the treatment of such data. In September 2016, EPA revised that Exceptional Events Rule. To demonstrate that the June 20, 2015 exceedances qualified for exclusion as influenced by exceptional events, Arizona submitted to EPA three sets of statistical demonstrations. Arizona submitted its initial demonstration while the 2007 rule was in effect but submitted its two supplemental demonstrations while the 2016 rule was in effect. EPA decided to apply the requirements of the 2016 rule and concluded the Lake Fire did indeed cause the June 20, 2015 monitor readings. EPA then excluded the six exceedances from its NAAQS calculations.

Subsequently, in a final rule1 based on those exclusions, EPA determined the Phoenix area had successfully attained the ozone NAAQS by its July 20, 2018 attainment date. That rule allowed Arizona to avoid additional, more strict regulatory burdens (although evidence suggests the area ozone levels have since lapsed back to exceed the standard). Based on that attainment determination, EPA also decided to suspend one requirement to which Arizona was already subject: that it develop contingency measures to be implemented if Arizona had failed to attain the NAAQS by the statutory 2018 date.

Petitioners—citizens of Phoenix—challenge that final rule. They first claim EPA violated the presumption against retroactivity when it applied the 2016 version of the Exceptional Events Rule because the 2007 rule had been in effect when the 2015 Lake Fire and exceedances occurred. We hold EPA’s application of the 2016 Exceptional Events

Rule did not impact any vested rights, create any new obligations, or otherwise impact any regulated party’s interests in fair notice, reasonable reliance, or settled expectations. Therefore, the application of the 2016 Exceptional Events Rule was not impermissibly retroactive and no presumption against retroactivity arose.

Petitioners further claim Arizona’s evidence does not support EPA’s finding that a clear causal connection existed between the Lake Fire and the June 20, 2015 exceedances. But we defer to EPA’s technical conclusions and find that Arizona adduced evidence sufficient to allow EPA to make such finding. As a result, EPA did not act arbitrarily or capriciously by excluding the data from the six monitors as falling within the Exceptional Events Rule.

Finally, Petitioners claim EPA acted contrary to the Clean Air Act in suspending Arizona’s contingency measures requirement in EPA’s July 2018 final rule. Petitioners contend that the Clean Air Act requires states to provide attainment contingency measures regardless whether the region attains the NAAQS by its attainment date. We find Petitioners forfeited this argument by not sufficiently raising it in their comment before the agency, but, even under a lenient interpretation of the content of their comment before the agency, we conclude that EPA’s construction of the Clean Air Act is owed deference under .2 EPA’s suspension of the contingency measure requirements in its July 2018 final rule did not violate the Clean Air Act.

For these reasons, we deny petitioners request to review the rule.

BACKGROUND

I. LEGAL FRAMEWORK

A. Clean Air Act & National Ambient Air Quality Standards

The Clean Air Act of 1970 and its amendments created a system intended to improve the country’s ambient air quality—the national ambient air quality standards. This system aims to reduce the concentration of certain air pollutants (“criteria pollutant“) that are found to endanger public health or welfare. 42 U.S.C. §§ 7408(a), 7409(b). To implement the NAAQS program, the Clean Air Act requires EPA to partition the country geographically into designated air quality control regions. Id. § 7407. EPA assesses each region individually to determine whether the region’s ambient air exceeds the allowable concentrations for each criteria pollutant. Id. For those regions that meet the standard, EPA designates them as “in attainment,” while those regions that do not meet the NAAQS are designated as “nonattainment areas” or “NAAs.” Id. A nonattainment area “attains the NAAQS” if the average concentration of a particular pollutant in the ambient air does not exceed the standard. Id.

For those areas designated nonattainment for ozone (a criteria pollutant), EPA further classifies each region based on the severity of the nonattainment: marginal, moderate, serious, severe, or extreme. Id. § 7511(a)(1). Depending on that classification, the Clean Air Act provides a nonattainment area with a certain number of years by which it must attain the NAAQS. Id. EPA must determine whether

an ozone nonattainment area attained the NAAQS by that date (“attainment date“). Id. § 7511(b)(2)(A). Should EPA find that the area failed to attain the NAAQS by the attainment date, the area is automatically reclassified to the next strictest class. This reclassification imposes more stringent requirements, designed to reduce air pollution, and the region is given a new attainment date. Id.

If by the attainment date EPA determines a nonattainment area has successfully achieved the NAAQS, it issues a “§ 7511(b)(2) Attainment Determination.”3 Id. This § 7511(b)(2) Attainment Determination acts to prevent a region’s automatic classification downgrade, but it does not formally redesignate the area as in attainment or suspend any regulatory requirements. A formal redesignation occurs only after the region meets further statutory requirements, which include a finding that improvements in air quality are permanent, not transient, and EPA’s approval of a ten-year maintenance plan to prevent backsliding. Id. § 7407(d)(3).

B. State Implementation Plans, Contingency Measures, and the Clean Data Policy

The Clean Air Act operates predominantly through state action rather than through direct federal control. The Act seeks to achieve improvements in ambient air quality by requiring states to create and implement pollutant controls. For each region within a state that is designated nonattainment for a criteria pollutant, the state must create a State Implementation Plan (“SIP“) and obtain approval from EPA. Id. § 7410. A SIP outlines the measures the state intends to take to reduce the concentration of the pollutant in

the region. Although the exact policy choices as to how to achieve those reductions are primarily left to the states, the Clean Air Act does require that SIPs contain certain provisions. For example, a SIP must require that an area in nonattainment make “reasonable further progress” (“RFP“) towards attaining the NAAQS. Id. § 7502(c)(2). Another mandated provision, the provision at issue here, is the identification of certain contingency measures “to be undertaken if the area fails to make reasonable further progress, or to attain the national primary ambient air quality standard by the attainment date applicable under this part.” Id. § 7502(c)(9). EPA divides these contingency measures into two separate requirements: (1) “RFP contingency measures,” to be implemented by the state if the area fails to make reasonable further progress in attaining the NAAQS; and (2) “attainment contingency measures,” to be implemented by the state if the area fails to attain the NAAQS by the attainment date.

In 1995, EPA issued a policy memorandum (“Seitz Memorandum“) establishing a Clean Data Policy, now codified at 40 C.F.R. § 51.1118.4 The Clean Data Policy is

EPA’s interpretation of the Clean Air Act’s SIP requirements. Seitz Mem. at 2; 42 U.S.C. § 7502(b) & (c). The Clean Data Policy allows EPA to suspend certain SIP obligations for those nonattainment areas that EPA determines are actively attaining the ozone NAAQS prior to being formally redesignated as in attainment. EPA’s reasoning is based on the commonsense notion and the statute’s language that an area already attaining the NAAQS need not demonstrate it is “making reasonable further progress” toward attaining the NAAQS. Seitz Mem. at 4; 40 C.F.R. § 51.1118. But the authority to suspend SIP requirements under the Clean Data Policy “exists only for as long as the nonattainment area continues to monitor attainment of the standard.” Seitz Memo at 4; 40 C.F.R. § 51.1118.

C. Regulation of Ozone under the Clean Air Act

Ground-level ozone is classified as a criteria pollutant under the Clean Air Act. Unlike most other air pollutants, ozone is not emitted directly into the air by factories or cars or living organisms. Rather, ozone forms when certain precursor pollutants—nitrogen oxides (NOx) and volatile organic compounds (both of which can be emitted due to human action)—interact in sunlight, a reaction which is accelerated in high heat and in response to other weather factors. This diffuse generation process makes tracing any individual cause of increased ozone concentrations a difficult process.

In 2008, EPA set the ozone NAAQS to 0.075 parts per million (“ppm“).5 40 C.F.R. § 50.15; see also EPA, National Ambient Air Quality Standards for Ozone, 73 Fed. Reg. 16,435 (Mar. 27, 2008). Each region has multiple air quality monitoring sites.6 40 C.F.R. § 50.15. A region attains the NAAQS only if each monitoring station in the nonattainment area registers a three-year calculated value at or below 0.075 ppm. EPA bases its attainment determinations on the three most recent, complete calendar years of quality-assured data. Id.

D. Exceptional Events Rule

When making a § 7511(b)(2) Attainment Determination, EPA must exclude any data of a concentration of a pollutant above the NAAQS (“exceedances“) if the air quality was influenced by “exceptional events.” 42 U.S.C. § 7619(b). To qualify as an exceptional event warranting exclusion, “a clear causal relationship must exist between the measured exceedances of a national ambient air quality standard and the exceptional event to demonstrate that the exceptional

event caused a specific air pollution concentration at a particular air quality monitoring location.” Id. § 7619(b)(3)(B)(ii).

In 2007, EPA promulgated the Exceptional Events Rule implementing this provision of the Clean Air Act. EPA, Treatment of Data Influenced by Exceptional Events, 72 Fed. Reg. 13,560 (Mar. 22, 2007). The rule requires that states demonstrate “to EPA’s satisfaction that such event caused a specific air pollution concentration at a particular air quality monitoring location.” 40 C.F.R. § 50.14(a) (2015). Under the 2007 rule, a successful state demonstration of an exceptional event required that a state prove certain elements, including that the exceedances were “in excess of normal historical fluctuations” and were the “but for” cause of the event.7 Id. § 50.14(c)(3)(iv).

In 2016, EPA replaced the 2007 Exceptional Events Rule. 81 Fed. Reg. 68,216 (Oct. 3, 2016). The 2016

Exceptional Events Rule revised state demonstration requirements to remove the need to show that the events were “in excess of normal historical fluctuations” and that “but for the event” there would not have been an exceedance but added others.8 Both the 2007 and the 2016 rules required demonstration of a clear causal relationship between the exceptional event and the measured exceedances.

Wildfires are a common exceptional event that cause ozone exceedances. Indeed, concurrent with the revised

2016 Exceptional Events Rule, EPA issued a guidance document for states preparing wildfire exceptional event demonstrations (“Wildfire Ozone Guidance“). EPA, Guidance on the Preparation of Exceptional Events Demonstrations for Wildfire Events that May Influence Ozone Concentrations (Sept. 2016). As is relevant here, the Wildfire Ozone Guidance created a three-tiered approach for demonstrating a clear causal relationship between a wildfire and an exceedance. Id. at 3–4. For events where it is more obvious that a wildfire caused an ozone exceedance, a simpler Tier 1 demonstration with fewer analyses is all EPA requires. Id. at 4. However, where the connection between the wildfire and the exceedance is more complicated, EPA requires Tier 2 or Tier 3 levels of documentation and support. Id. A successful Tier 3 demonstration requires proof that wildfire emissions: (1) were transported to the monitor; (2) affected the monitor; and (3) caused the ozone exceedances. Id. at 25–30.

II. FACTUAL & PROCEDURAL BACKGROUND

In 2012, EPA classified the Phoenix-Mesa region of Arizona as in “marginal” nonattainment with an attainment date in 2015. The Phoenix nonattainment area failed to attain the 2008 ozone standard by its 2015 attainment date. As a result, EPA reclassified and downgraded the region to its current status of “moderate” nonattainment and issued a revised attainment date of July 20, 2018. EPA’s 2018 attainment review would assess the region’s data from the full 2015–2017 calendar years and would not assess any data from January 2018 onward.

During that assessment period, from June 17 to August 1, 2015, the Lake Fire burned approximately 30,000 acres of the San Bernardino National Forest in southeastern California. Nearly half the reported destruction occurred

during the fire’s first three days (June 17–19). On Saturday, June 20, 2015, 300 miles east of the fire, six ozone monitors in the Phoenix nonattainment area recorded exceedances of the 2008 ozone NAAQS (0.075 ppm).9

In July 2016, the Arizona Department of Environmental Quality (“ADEQ“) notified EPA that it would seek to exclude the June 20, 2015 exceedances as caused by the Lake Fire exceptional event. EPA and ADEQ agreed the most exigent measurement—a Tier 3 analysis—was required. On September 27, 2016, ADEQ submitted an exceptional events demonstration to EPA for the six monitored June 20, 2015 exceedances (“initial demonstration“). Shortly thereafter, EPA finalized its revised Exceptional Events Rule on October 3, 2016 (effective September 30, 2016). Subsequently, EPA twice asked ADEQ to supplement its exceptional events demonstration. ADEQ complied. In May 2019, based on ADEQ’s twice-supplemented package, EPA formally concurred with ADEQ’s request to exclude the June 20 exceedances.

In June 2019, EPA proposed to issue a § 7511(b)(2) Attainment Determination that the Phoenix nonattainment area attained the NAAQS by its 2018 attainment date—a determination made possible only by the exclusion of the June 20, 2015 exceedances. In addition, given the attainment finding, EPA also proposed suspending the attainment contingency measures required for SIPs by 42 U.S.C. § 7502(c)(9). Although it was not relevant to

EPA’s § 7511(b)(2) Attainment Determination, commencing in 2018 (and therefore, as earlier noted, outside the applicable period relevant for the 2018 attainment determination), the Phoenix nonattainment area has recorded multiple exceedances, and EPA has indicated it believes the area has lapsed back into nonattainment.

Sandra Bahr10 and the Arizona Center for Law in the Public Interest (“ACLIPI“) commented on the proposed § 7511(b)(2) Attainment Determination (proposed rule). After responding to their and others’ comments, EPA finalized and issued its § 7511(b)(2) Attainment Determination for the Phoenix nonattainment area along with its decision to suspend the attainment contingency measures requirement for the Phoenix NAA (final rule). EPA, Determination of Attainment by the Attainment Date for the 2008 Ozone National Ambient Air Quality Standards; Phoenix-Mesa, Arizona, 84 Fed. Reg. 60,920 (Nov. 12, 2019). Petitioners Sandra Bahr, Jeanne Lunn, and David Matusow seek review in this Court of that final rule.

JURISDICTION

Under 42 U.S.C. § 7607(b)(1), we have jurisdiction over a “final action of the [EPA] Administrator” made pursuant to the Clean Air Act “which is locally or regionally

applicable.” Petitioners timely petitioned for review within sixty days of the issuance of the final rule. Id.

STANDARD OF REVIEW

We review agency actions under the Clean Air Act pursuant to the judicial review provisions of the Administrative Procedure Act (“APA“), 5 U.S.C. §§ 701–706. . Under the APA, we set aside an agency action only if it is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). “We will deem an agency action to be arbitrary and capricious only ‘if the agency has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.’” (quoting ).

ANALYSIS

I. Whether EPA’s Application of the 2016 Exceptional Events Rule Had Impermissible Retroactive Effect

The presumption against retroactivity generally prevents application of statutes and regulations to conduct or events which had occurred prior to the effective date of those rules, but only if application of those statutes and rules would have retroactive effect by impairing prior-existing rights and by affecting reliance interests. . Petitioners argue EPA

violated the presumption against retroactivity when it reviewed ADEQ’s demonstration under the 2016 Exceptional Events Rule rather than the 2007 Exceptional Events Rule because the Lake Fire, June 20, 2015 exceedances, and ADEQ’s September 27, 2016 initial demonstration all occurred prior to the September 30, 2016 effective date of the revised rule. EPA argues Petitioners failed to exhaust this issue because they did not raise it before the agency (EPA) during the notice and comment period provided. But, EPA argues, even had Petitioners properly raised this issue before the agency, EPA’s application of the 2016 rule, and not the 2007 rule, did not violate the presumption against retroactivity because the rule did not have a prohibited retroactive effect.

A. Exhaustion of Administrative Remedies

“As a general rule, we will not consider issues not presented before an administrative proceeding at the appropriate time.” . Congress codified this rule within the Clean Air Act: “Only an objection to a rule or procedure which was raised with reasonable specificity during the period for public comment (including any public hearing) may be raised during judicial review.” 42 U.S.C. § 7607(d)(7)(B). However, we have “held that the exhaustion requirement should be interpreted broadly. [Petitioners] fulfill the requirement if their [comment] ‘provided sufficient notice to the [agency] to afford it the opportunity to rectify the violations that the [petitioners] alleged.’” (quoting ); but see (“Objections must be

prominent and clear enough to place the agency on notice, for EPA is not required to cull through all the letters it receives and answer all of the possible implied arguments. (quotation marks and citation omitted)). To that end, we will consider any issue that was raised with sufficient clarity to allow the decision maker to understand and rule on the issue raised, whether the issue was considered sua sponte by the agency or was raised by someone other than the petitioning party. (quoting ).

Here, Petitioners seem to concede their comment to EPA did not expressly contest EPA‘s decision to use the 2016 Exceptional Events Rule. Instead, Petitioners claim their comment—in which they wrote that there was nothing exceptional about the June 20 exceedances, ACLIPI Comment Letter at 2—impliedly contested EPA‘s decision because the comment addressed at least one requirement unique to the 2007 rule (that ADEQ failed to show the exceedances were in excess of historical fluctuations). They also argue that EPA‘s response indicated EPA understood that Petitioners were arguing the 2007 rule should apply.11

On its own, Petitioners’ comment reads as a simple observation that the exceedances should not be considered exceptional events, referencing the name of the rule. It does not clearly state an argument that the exceedances failed to meet the specific in excess of historical fluctuations requirement of the 2007 rule. It does not mention the dates of either the 2007 or the 2016 rule. Indeed, their comment neither mentions nor cites to the 2007 rule at all (although Petitioners do cite the 2016 Wildfire Ozone Guidance that was issued concurrently with the 2016 rule. This would lead the reader to believe Petitioners were invoking the 2016 rule, as the Guidance did not exist prior to 2016 and was issued as a guide to the use of the 2016, not the 2007, rule.). One would imagine that Petitioners would more vociferously object to EPA‘s complete failure to make findings under two prongs of the 2007 rule that they believed were essential to EPA‘s determination. That they did not suggests that not even Petitioners believed the 2007 rule‘s requirements controlled. Nor does their comment engage in any analysis of historical fluctuations (it merely compares the exceedances to contemporaneous monitor readings from that year) or mention the presumption against retroactivity. As such, Petitioners failed adequately to present the issue to the agency.

Nonetheless, Petitioners contend that EPA‘s response to their comment indicated that EPA did understand Petitioners to be questioning whether the exceedances met the in excess of historical fluctuations requirement of the 2007 rule:

A previous version of the Exceptional Events Rule required that, in addition to meeting these statutory elements criteria, states also submit evidence that the event was associated with a measured concentration in excess of normal historical fluctuations, including background. However, in the 2016 revisions to the Rule, the EPA removed this requirement . . . .

84 Fed. Reg. at 60,922. But EPA‘s response merely underlines the point that the historical fluctuations requirement, mentioned by Petitioners, was no longer required. We see no evidence that EPA understood Petitioners to be objecting to the use of the 2016 Exceptional Events Rule, or that Petitioners were suggesting that application of that rule amounted to a violation of the presumption against retroactivity. As such, we cannot say that EPA raised the issue of retroactive application of the 2016 Exceptional Events Rule sua sponte or that Petitioners raised the issue with sufficient clarity to allow EPA to understand and rule upon it. We find Petitioners failed to exhaust the issue as to whether the 2007 or 2016 Exceptional Events Rule governed the final agency action.

B. Retroactive Application of the 2016 Exceptional Events Rule

Even were we to determine Petitioners adequately raised the issue, we find EPA‘s application of the 2016 Exceptional Events Rule did not have an impermissibly retroactive effect.12 Application of the 2016 rule did not impact any vested rights, create any new obligations, or otherwise impact any regulated party‘s interests in fair notice, reasonable reliance, or settled expectations.

If the statute [or regulation] would operate retroactively, our traditional presumption teaches that it does not govern. . In , [t]he Supreme Court articulated a two-step approach for evaluating when the normal presumption against retroactivity should not apply. Our first task under is to determine whether Congress has expressly prescribed that a regulation is to be applied retroactively. (quoting ). Here, neither party contends EPA possesses express retroactivity authority as to the elements which compose an exceptional event.

Moving on to the second step, we must determine whether application of the regulation would have a retroactive effect. . While retroactivity of legislation and regulations is not per se unlawful, we have a presumption against retroactivity that generally requires that the legal effect of conduct . . . ordinarily be assessed under the law that existed when the conduct took place. (quoting ). But a regulation does not operate retrospectively merely because it is applied in a case arising from conduct antedating the statute‘s enactment or upsets expectations based in prior law. (internal citation omitted). A [regulation] has retroactive effect when it takes away or impairs vested rights acquired under existing laws, or creates a new obligation, imposes a new duty, or attaches a new disability, in respect to transactions or considerations already past. (superseded on other grounds) (internal quotation marks and citations omitted).

The conclusion that a particular rule operates retroactively comes at the end of a process of judgment concerning the nature and extent of the change in the law and the degree of connection between the operation of the new rule and a relevant past event. Any test of retroactivity will leave room for disagreement in hard cases, and is unlikely to classify the enormous variety of legal changes with perfect philosophical clarity. However, retroactivity is a matter on which judges tend to have sound instincts, and familiar considerations of fair notice, reasonable reliance, and settled expectations offer sound guidance.

(internal quotation marks and citations omitted).

Preliminarily, we note the critical fact that Arizona itself does not complain of any impermissible retroactivity affecting its vested rights, settled expectations, or reliance interests.13 ADEQ successfully demonstrated the exceedances were caused by an exceptional event under EPA‘s 2016 rule. EPA provided fair notice as to the rule‘s change, and twice, after adoption of the 2016 rule, permitted ADEQ to supplement its demonstration package. We also reject Petitioners’ assertion that simply because the June 20, 2015 exceedances occurred during the existence of the 2007 Exceptional Events Rule, EPA‘s application of the 2016 Exceptional Events Rule was impermissibly retroactive. (A regulation is not made retroactive merely because it draws upon antecedent facts for its operation. (quoting ).

Instead, we must determine whether the timing of the exceedances vested Petitioners with some right under the Clean Air Act which was impaired by EPA‘s decision to apply the 2016 Exceptional Events Rule instead of the 2007 rule. Petitioners—Phoenix residents who are adversely affected by unhealthy levels of ozone—claim they have a right . . . to have requests to excuse exceedances as exceptional events evaluated under the requirements that existed at the time of the exceedances themselves. Petitioners cite no statutory, regulatory, or case authority for such claim and neither could we find any. Rather, if Petitioners—who are not directly regulated by the Clean Air Act or the Exceptional Events Rule—have any rights implicated by the retroactivity concern identified here (putting aside whether they are vested or not), the Clean Air Act might grant them a right in the eventual attainment of healthy levels of ozone concentration in the Phoenix region‘s ambient air. See 42 U.S.C. § 7401(b) (The purposes of [the Clean Air Act] are (1) to protect and enhance the quality of the Nation‘s air resources so as to promote the public health and welfare and the productive capacity of its population . . . .). To that end, Petitioners’ interest here would not be in the application of any particular rule on any particular date, but in EPA‘s accurate and faithful enforcement—according to its best scientific judgment—of the commands of the Clean Air Act.

And, by 2016, before resolving whether the Lake Fire was an exceptional event, EPA‘s methods by which to reach its best scientific judgment had evolved from those adopted in 2007. The Clean Air Act enlists EPA‘s technical expertise in defining an exceptional event. 42 U.S.C. § 7619(b)(1)(A)(iv) (defining exceptional event, in part, as an event determined by the Administrator through the process established in the regulations . . . to be an exceptional event.). EPA revised its regulations in 2016 on the basis that the 2007 Exceptional Events Rule was no longer consistent with the best available methods, a determination that reflect[ed] the experiences of the EPA, state, local and tribal air agencies, federal land managers and other stakeholders in implementing this program over the past 10 years. 81 Fed. Reg. at 68,216. In that 2016 rule, EPA resolved to eliminate the 2007 but for criterion so the rule could focus instead on the clear causal relationship criterion expressly required in the statute. Id. at 68,217. EPA also decided to replace the in excess of normal historical fluctuations criterion with a requirement for a comparison of the event-related concentration to historical concentrations. Id. EPA affirmed that the regulatory revisions . . . protect human health and the environment while providing needed clarity, increasing the administrative efficiency of demonstration submittal process, and removing some of the challenges associated with implementing the Exceptional Events Rule. Id. Petitioners’ interests in cleaner air are not adversely affected by a refined implementation of statutory requirements or by a matured scientific understanding as to which factors most accurately demonstrate the existence of an exceptional event.

Now, theoretically, the change in criteria for establishing an exceptional event could have an indirect impact on the level of ozone concentration in Arizona. For example, if the new criteria substantially reduced the difficulty in making a successful exceptional event demonstration, the likely increase in excused exceedances could result in EPA finding fewer nonattainment areas to have failed to attain the NAAQS than under the prior rule‘s criteria. That in turn might prevent EPA from mandating stricter pollution protection measures. But Petitioners do not attempt to argue the altered requirements in the 2016 rule are generally more lenient to the states, just that they are different. Applied to the circumstances here, we cannot know whether EPA was more or less likely to conclude the Lake Fire qualified as an exceptional event under the 2007 or 2016 rule simply because they contain slightly different criteria. If Petitioners believed the 2016 Exceptional Events Rule was contrary to the Clean Air Act, they could have challenged that rule when EPA finalized it in 2016. As it stands, we must conclude that the revised rule is a valid and faithful endeavor by EPA to implement the Clean Air Act and that Petitioners’ professed vested rights, rather than being impaired, are better served by EPA‘s use of the revised rule.14 Absent demonstration of disrupted reliance interests from regulated parties, we will not require EPA to revert to what it determined to be an outdated, deficient rule simply because an event occurred under the prior rule‘s regime.

We conclude EPA‘s application of the 2016 Exceptional Events Rule was not impermissibly retroactive. ADEQ does not complain and Petitioners cannot demonstrate that application of EPA‘s 2016 Exceptional Events Rule in lieu of the 2007 rule impaired any vested rights, created any new obligations, or otherwise impacted any interests in fair notice, reasonable reliance, or settled expectations.

II. Whether EPA‘s Conclusion That There Was a Clear Causal Relationship Between the Lake Fire and the June 20 exceedances Was Arbitrary or Capricious

Under the Clean Air Act, a clear causal relationship must exist between the measured exceedances of a national ambient air quality standard and the exceptional event to demonstrate that the exceptional event caused a specific air pollution concentration at a particular air quality monitoring location. 42 U.S.C. § 7619(b)(3)(B)(ii); see also 40 C.F.R. § 50.14(c)(3)(iv)(B).15 Petitioners argue EPA erred in concluding that ADEQ established a clear causal connection between the Lake Fire and the six ozone exceedances on June 20, 2015.

We review EPA‘s findings under the arbitrary or capricious standard where, as relevant here, [a]n agency decision will be upheld as long as there is a rational connection between the facts found and the conclusions made. . At the outset, we note this is a quintessential instance where we are bound to defer to the technical expertise of the agency. [W]here, as here, a court reviews an agency action involv[ing] primarily issues of fact, and where analysis of the relevant documents requires a high level of technical expertise, we must defer to the informed discretion of the responsible federal agencies. (quoting ).

In determining whether a state has adequately demonstrated a clear causal relationship under the 2016 Exceptional Events Rule, EPA reviews on a case-by-case basis using a weight of evidence approach.16 81 Fed. Reg. at 68,227. For Tier 3 demonstrations, EPA is guided by the approach set forth in the 2016 Wildfire Ozone Guidance document. Here, EPA determined that ADEQ adequately demonstrated a clear causal relationship by showing the Lake Fire wildfire emissions: (1) were transported to the six monitors; (2) affected the monitors; and (3) caused the ozone exceedances.

Petitioners argue EPA‘s conclusion is not rationally connected to the evidence offered by ADEQ. They assert that local conditions caused the high ozone concentrations at the monitors, not the Lake Fire. Petitioners did not submit any technical models in their comment to EPA, but instead attempt to convince us that the specific evidence relied on by EPA does not support a clear causal relationship conclusion.17 We conclude there was a rational connection between the facts found and EPA‘s conclusion as to each of these three requirements under the Wildfire Ozone Guidance.

A. Lake Fire Emissions Were Transported to the Monitors

To demonstrate the Lake Fire emissions were transported from the San Bernardino Forest to the six exceedance monitors, ADEQ submitted trajectory analyses,18 satellite photos of the area revealing visible smoke plumes, and National Oceanic and Atmospheric Administration (NOAA) smoke contour maps (mapping smoke dispersion across southeastern California, Arizona, and northern Mexico). EPA found this evidence adequate:

Overall, the trajectory analyses provided in the second addendum, along with the satellite imagery and data, water vapor and dew point analysis, and meteorological data regarding boundary layer depths in the nonattainment area on June 20, 2015, show that emissions from the Lake Fire in California were transported to the nonattainment area and the affected monitoring sites and reached ground level on June 20, 2015.

Petitioners argue the following evidence undermines EPA‘s finding:

Satellite Images & NOAA Smoke Maps. ADEQ provided satellite images and smoke maps of the southwestern United States and Mexico to illustrate that visible smoke plumes from the Lake Fire travelled hundreds of miles to reach the ozone monitors in the Phoenix NAA. Petitioners argue the shape and location of smoke in ADEQ‘s images and maps are inconsistent with the premise that the smoke originated from the Lake Fire (they suggest perhaps the smoke travelled from a different wildfire to the south).19

We disagree. It is true that some of the maps and images show gaps between plumes of visible smoke. But, according to EPA, ADEQ‘s trajectory analyses nonetheless support the conclusion that the visible smoke shown on the maps originated at the Lake Fire. Petitioners provided no technical evidence of their own to contradict these analyses. Indeed, Petitioners do not attempt to refute the trajectory models at all, but instead attempt to contradict ADEQ‘s and EPA‘s analysis based solely on their own interpretation of static satellite images. EPA, in response to Petitioners’ comment, noted that the images and maps show large areas of visible smoke both near the San Bernardino Forest and across the Phoenix NAA. That the images and maps may not demonstrate a visibly contiguous distribution of smoke is not dispositive evidence against causation because these images represented only a single point in time. That is especially true in light of the companion analyses and data that do suggest the smoke over the Phoenix region originated at the Lake Fire. The still images upon which Petitioners rely to support their conjecture that the smoke may have originated elsewhere are inadequate to make EPA‘s contrary conclusion—a product of agency expertise—implausible.

Geographic Pattern of Heightened Ozone Concentrations. ADEQ submitted maps of ozone concentrations showing a regional rise across much of Arizona on June 19 and 20. Petitioners argue that these maps do not show a pattern of elevated ozone concentrations along the trajectory of the supposed plumes from the Lake Fire, suggesting the ozone exceedances were of local origin.

Ozone production due to wildfire emissions can be difficult to predict because ozone generation is influenced not only by concentration of precursor chemicals, but also by local sunlight and meteorological factors. The lack of a perfect pattern of heightened ozone concentration along the trajectory is therefore not preclusive to the causation finding. And, contrary to Petitioners’ assertion, many monitors along the trajectory did in fact did show increased ozone concentration on June 20, 2015.

Ultimately, Petitioners failed to show there was no rational connection between ADEQ‘s factual demonstration and EPA‘s conclusion that Lake Fire emissions physically travelled from the San Bernardino forest to the six ozone monitors in Arizona.

B. Lake Fire Emissions Affected the Monitors

To demonstrate that the Lake Fire emissions affected the Phoenix monitors, ADEQ submitted: (1) maps of ozone concentrations showing a regional rise in ozone concentrations across Arizona on June 19 and 20, (2) profiles of the exceedance monitors with comparisons to historical ozone concentrations that demonstrate the June 20 readings were relatively high, and (3) analyses of regional concentrations of three other typical emissions related to wildfires: nitrogen dioxide (NO2), particulate matter (PM2.5), and elemental carbon (EC) and organic carbon (OC). EPA found this evidence supported the conclusion that wildfire emissions reached the ground and affected measurements at the exceeding monitors:

Overall, the lack of elevated PM2.5 in the nonattainment area raises questions about the extent to which wildfire emissions reached the ground and affected the monitor. However, the supplemental analyses showing elevated OC and relatively low EC/OC concentrations, and unusually elevated NO2 and O3 concentrations observed on a Saturday, along with the robust analysis of transport and mixing mechanisms described earlier in this document, ultimately support the conclusion that wildfire emissions reached the ground and affected measurements at the exceeding monitors on June 20, 2015.

84 Fed. Reg. at 60,922. Petitioners argue the following evidence undermines EPA‘s finding:

PM2.5 and Elemental and Organic Carbon Analyses. ADEQ submitted an analysis of concentrations of PM2.5—commonly associated with wildfire emissions—at monitors in the Phoenix nonattainment area on June 20 as well as analysis of elemental and organic carbon concentrations registered at the nearby Phoenix Supersite (which did not record an ozone exceedance on June 20). EPA found that PM2.5 was not elevated in the Phoenix nonattainment area on June 20, which raises questions about the extent to which wildfire emissions affected the monitors. But EPA also found that the comparative concentrations of OC and EC at the Supersite provides some support that wildfire emissions were present in the [NAA]. Petitioners argue EPA‘s PM2.5 finding cannot be overcome by OC and EC findings from the Supersite because the Supersite was not significantly influenced by wildfire smoke on that day and is located at least fifteen miles from the nearest exceptional events monitor.

We defer to EPA‘s conclusion that OC and EC are relevant to the causation analysis on the basis that those compounds are generally associated with biomass smoke (which is emitted during wildfires). We also conclude EPA was justified in using the Phoenix Supersite data. EPA‘s Wildfire Ozone Guidance permits the use of data from monitors co-located or nearby the exceedance monitors. At the least, Petitioners did not offer support for their assertion that fifteen miles (the size of the Phoenix nonattainment area is roughly 5000 square miles20) is so great a distance that EPA erred in considering the data as part of its weight of evidence approach. The Supersite was the only OC monitor in the Phoenix nonattainment area and was thus a relevant source of data from which to draw.

Elevated NO2 Concentrations. ADEQ submitted an analysis showing unusually elevated concentrations of NO2 (an ozone precursor) at three monitors which did not register exceedances in the Phoenix nonattainment area on June 19 and 20. Petitioners argue EPA‘s reliance on this analysis is misplaced because NO2 emissions are not specific to wildfire emissions, and that the readings are irrelevant because those monitors did not record ozone exceedances, are located too far away from the exceedance monitors, and were cherry picked.

Again, we defer to EPA‘s conclusion that NO2 concentrations are relevant considerations because NO2 is both emitted by wildfires and is a precursor to ozone. EPA noted that this evidence of elevated NO2 concentrations in the nonattainment area was not dispositive but was just one of several pieces of evidence relevant to the weight of evidence determination. Whether or not the monitors were located too far from the exceedance monitors, that they were the only non-mobile-source oriented NO2 monitors in the Phoenix-Mesa area, that EPA acknowledged that these monitors’ locations may have limited this evidence‘s probative value, and that EPA placed non-dispositive weight on this evidence all indicate that EPA did not act irrationally in considering NO2 concentrations.

Ultimately, Petitioners failed to show there was no rational connection between ADEQ‘s factual demonstration and EPA‘s conclusion that the emissions from the Lake Fire affected the six exceedance monitors. The analyses cited by EPA generally present convincing evidence that Arizona‘s atmospheric monitors recorded unnatural increases in both ozone and its precursor compounds on June 19 and June 20. We are unable to find fault with EPA‘s technical conclusions as to the relative weight of evidence of increased concentrations of particular ozone precursor compounds without contrary evidence or a demonstration that EPA made some analytical error. Petitioners did not supply any evidence or technical data with which they might have overcome the deference we owe to the conclusion of EPA‘s experts. Our role is to assess whether the EPA‘s determinations of the facts of this element of causation were arrived at by arbitrary or capricious means. They were not.

C. Lake Fire Emissions Caused the Ozone Exceedances

To demonstrate that the Lake Fire emissions caused the ozone exceedances at the six monitors, ADEQ submitted three matching day analyses which compared the June 20 exceedances to other previous monitor readings based on: (1) days with similar meteorological conditions, (2) days which also recorded exceedances, and (3) days of the week. EPA found this evidence, combined with each of the other submitted analyses, sufficiently demonstrated a clear causal relationship between the emissions from the Lake Fire and the exceedances:

The analyses included in the demonstration and addenda, specifically, the comparison with historical hourly and daily maximum 8-hour O3 concentrations; updated HYSPLIT analyses, satellite imagery and data, water vapor and dew point analysis, and meteorological data regarding boundary layer depths in the nonattainment area on June 20, 2015; elevated OC and relatively low EC/OC concentrations, and unusually elevated NO2 and O3 concentrations observed on a Saturday; and three matching day analyses demonstrating the unusual nature of the event, sufficiently demonstrate a clear causal relationship between the emissions generated by the Lake Fire in the San Bernardino National Forest in southeastern California and the exceedances measured at the [six monitors].

Petitioners argue the following evidence undermines EPA‘s finding:

Matching Meteorological Day Analysis. ADEQ provided an analysis of compiled monitor readings from other June days between 2010 and 2015 that had similar weather conditions as did June 20, 2015 and found exceedances to be historically unusual. Petitioners argue that, on its own, all that this matching day analysis shows is that the meteorological conditions that existed on June 20, 2015 would not normally be enough to be the sole cause of an exceedance. That observation is not contrary to EPA‘s conclusion that it was the San Bernardino fire that caused the exceedances. Moreover, EPA has agreed that the matching day analysis is insufficient on its own. Instead, EPA reiterated that it was just one of several pieces of evidence that supported the existence of a clear causal relationship.

Matching Day of the Week Analysis. One of ADEQ‘s matching day analyses examined exceedances by day of the week and showed that, likely due to reduced traffic during the weekends, Saturday exceedances are rare (finding the June 20 Saturday exceedances were the only such exceedances recorded from 2010 through 2015 for three of the exceedance monitors) and that ozone concentrations are more likely to be heightened during weekdays. Petitioners argue that, though rare, ozone exceedances can and have occurred on Saturdays and that for some monitors, Saturday is as likely or more likely than some weekdays to have recorded an exceedance. Thus, they argue, EPA cannot assume the six Saturday exceedances on June 20 were caused by wildfire emissions.

To the contrary, we agree with EPA that it is rational to suppose that the matching weekday analysis provides probative evidence that Saturday exceedances are unusual and that this fact points to a unique emissions source contributing to exceedances. The analysis need not be dispositive or demonstrate a precise correlation for EPA rationally to have considered it as evidence of causation.

Ultimately, Petitioners failed to show there was no rational connection between ADEQ‘s proffer of evidence and EPA‘s conclusion that emissions from the Lake Fire (which it had already found travelled to and affected the monitors) caused the June 20 exceedances. Indeed, Petitioners’ arguments mostly cut against their own position. Evidence that the local meteorological conditions on June 20 were historically insufficient to cause the exceedances is probative evidence that an outside event like the Lake Fire was the cause. Petitioners fail to contradict EPA‘s

conclusion that the relatively rare weekend exceedances—Saturday in particular—was probative evidence that the June 20 exceedances were caused by some factor apart from local, typical motor vehicle and meteorological conditions.

D. Conclusion

The evidence demonstrates that smoke (including ozone precursor chemicals) from the Lake Fire reached the exceedance monitors and caused abnormal ozone readings relative to similar historical conditions. Petitioners failed to produce evidence sufficient to overcome the required deference to EPA‘s technical factual findings for any of the factors above. EPA considered each of the Petitioners’ comments during the proposed rule phase and addressed them with specificity. EPA articulated a rational connection between the evidence and its own conclusions, evincing a reasoned decision-making process. The resulting conclusion, based on the weight of the evidence, is rational. Accordingly, we find EPA did not act arbitrarily or capriciously in finding a clear causal connection between the Lake Fire and the June 20, 2015 exceedances.

III. Whether EPA‘s Suspension of SIP Attainment Contingency Measures Was Contrary to the Clean Air Act

In its final rule, EPA suspended the attainment contingency measures requirement for the Phoenix NAA. Petitioners now argue this suspension was a violation of the Clean Air Act, 42 U.S.C. § 7502(c)(9). EPA argues that its interpretation fills a statutory gap and is due deference and that Petitioners failed to exhaust argument to the contrary before the agency.

As a preliminary matter, the parties disagree as to whether EPA‘s final rule was based on the Clean Data Policy or on a separate, distinct interpretation of the Clean Air Act. Petitioners argue EPA‘s suspension of the contingency measures requirement was predicated on the Clean Data Policy, which applies only to the extent that the area demonstrates continued attainment, and that 2018–2019 ozone data show the Phoenix nonattainment area is not currently attaining the NAAQS. Petitioners are mistaken. EPA conceded that the Phoenix nonattainment area‘s ozone concentration data is inconsistent with continued attainment. Based on that, EPA expressly stated, both in its proposed rule and its final rule, that it was not applying the Clean Data Policy. EPA‘s action was instead based on its interpretation of the Clean Air Act‘s language specific to attainment contingency measures and applies only after the passing of a nonattainment area‘s attainment date (in this case, July 20, 2018). The Clean Data Policy, in contrast, relates to the suspension of SIP requirements for nonattainment areas that, prior to their attainment date, demonstrate that they are actively attaining the NAAQS ahead of schedule.21

A. Exhaustion of Administrative Remedies

EPA argues that Petitioners forfeited their argument here because their comment did not assert that EPA‘s proposal to suspend the attainment contingency measures requirement violated 42 U.S.C. § 7502(c)(9). Petitioners did comment on EPA‘s proposed suspension of the contingency measures requirement, however the comment did not address EPA‘s interpretation of § 7502(c)(9):

EPA‘s proposed rulemaking invites the State to withdraw the contingency measures adopted as part of its Eight-Hour Ozone Moderate Area Plan for the Phoenix NAA. For the reasons discussed above, this invitation is unwarranted and should be revoked.

ACLIPI Comment Letter at 8 (emphasis added). But the “reasons discussed” in Petitioners’ comment focused exclusively on how ADEQ‘s Lake Fire demonstration failed to justify exceptional event treatment. Petitioners’ comment does not broach the subject of EPA‘s interpretation of 42 U.S.C. § 7502(c)(9). A comment that simply mentions disagreement with an agency‘s ultimate proposed action does not provide sufficient substance with which the agency may understand and rule on the issue. See . And just because a commenter adequately objects to an agency‘s findings or decision on one basis does not permit the commenter to object to the same agency decision on substantively different bases on petition for review. Each substantively distinct argument must be sufficiently communicated to the agency, otherwise such argument will suffer forfeiture. See (finding an “opaque comment” that “was not adequately specific” to the issue argued on appeal was forfeited).

Alternatively, Petitioners dispute forfeiture because we should find their comment objected to EPA‘s use of the Clean Data Policy to suspend contingency measures. In their comment, they stated: “[m]onitoring data from 2018 and 2019 show multiple exceedances of the 2008 ozone standard and make clear that the Phoenix NAA does not actually comply with this standard.” ACLIPI Comment Letter at 1–2 (emphasis in original). That comment does not expressly mention the Clean Data Policy. But even if EPA were wise to the implication, as established, Petitioners’ comment was incorrectly premised on the assumption that EPA was applying its Clean Data Policy, rather than interpreting the Clean Air Act‘s attainment contingency measures provision in § 7502(c)(9). Petitioners made this mistake even though EPA expressly stated in its proposed rule that it was not applying the Clean Data Policy because “ozone monitoring data for 2018 are not consistent with continued attainment of the standard in the Phoenix NAA.”

Petitioners argue that any failure to exhaust the issue before the agency should be excused because exceptional circumstances exist based on the important public health ramifications implicated in this case. While exceptional circumstances may excuse forfeiture in other agency review contexts, the Clean Air Act statutorily mandates exhaustion and there is no exceptional circumstances excuse in the text. See ; 42 U.S.C. § 7607(b). We will not judicially manufacture an exemption from the exhaustion mandate contrary to the text of the statute.

As previously stated, a petitioner fails to exhaust his administrative remedies if an issue was not “raised with sufficient clarity to allow the decision maker to understand and rule on the issue raised.” (citation omitted). We find Petitioners’ comment failed to convey any opposition to EPA‘s statutory interpretation of § 7502(c)(9) and deem Petitioners to have failed to exhaust this argument before the agency.

B. EPA‘s Construction of the Clean Air Act Contingency Measures Requirement, 42 U.S.C. § 7502(c)(9)

Even were we to indulge Petitioners and reimagine that their comment was broadly arguing that the Clean Air Act prevents EPA from suspending the attainment contingency measures requirement, we find 42 U.S.C. § 7502(c)(9) to be silent on that question and defer to EPA‘s reasonable construction under step two.

In its final rule, EPA explained that “[u]nder [§ 7502(c)(9)], attainment contingency measures must be implemented only if the area fails to attain [the ozone NAAQS] by the attainment date.” Accordingly, EPA “determined that the requirement of [§ 7502(c)(9)] to [require SIPs] provide for contingency measures to be implemented in the event the area fails to attain [the ozone NAAQS] by its attainment date for the 2008 8-hour NAAQS does not apply to the [Phoenix] area.” The question here is whether the Clean Air Act requires State Implementation Plans to contain attainment contingency measures even after EPA determines a nonattainment area has attained the NAAQS by the applicable attainment date. This appears to be an issue of first impression.

“Where the agency‘s action is an interpretation of a statute that the agency administers, we follow the two-step approach set out in “. (quotation omitted). First, we determine if the statute speaks directly to the question or is unambiguous. “[I]f Congress has ‘directly spoken to the precise question at issue,’ then the matter is capable of but one interpretation by which the court and the agency must abide.” (quoting ). “[I]f the statute is silent or ambiguous with respect to the specific issue, we must ask at step two whether the regulations promulgated by the agency are based on a permissible construction of the statute.” (quotations omitted). When deference is appropriate, to the extent an EPA rule “involve[s] the reasonable resolution of ambiguities in the [Clean Air Act], [it] will be afforded deference.” (quotations omitted).

Under step one,22 we must determine whether a gap or ambiguity exists, or “whether Congress has directly spoken to the precise question at issue.” . Section 7511(b)(2)(A) requires EPA to make an attainment determination within six months after a nonattainment area‘s attainment date. Prior to that date, states create SIPs that detail which measures they will take to attain the NAAQS. These SIPs must contain certain provisions. 42 U.S.C. § 7502(c) (“The plan provisions (including plan items) required to be submitted under this part shall comply with each of the following . . . .“). Contingency measures are one of these mandated provisions:

Such plan shall provide for the implementation of specific measures to be undertaken if the area fails to make reasonable further progress, or to attain the national primary ambient air quality standard by the attainment date applicable under this part. Such measures shall be included in the plan revision as contingency measures to take effect in any such case without further action by the State or the Administrator.

42 U.S.C. § 7502(c)(9). Thus, under the Clean Air Act, SIPs must include contingency measures that may take effect in either of two circumstances: (1) the area fails to make reasonable further progress (“RFP contingency measures“); or (2) the area fails to attain the NAAQS by the attainment date (“attainment contingency measures“). See 84 Fed. Reg. at 60,925.

After EPA makes a § 7511(b)(2) Attainment Determination, however, the statute does not state whether the requirement for a SIP to contain attainment contingency measures—expressly conditioned on a past event (the failure to obtain a positive attainment determination)—should still apply. Petitioners argue that the plain language requires that SIPs contain contingency measures regardless of whether the nonattainment area attained the NAAQS or is currently attaining the NAAQS. True enough, the text does not expressly indicate that EPA has the authority to waive or suspend the inclusion of contingency measures in SIPs. But that reading of the statute ignores the obvious context: that the text of the statute ties these requirements to specific conditions. These “measures shall be included in the plan revision as contingency measures” so that they may “take effect in any such case.” 42 U.S.C. § 7502(c)(9). Logically, the attainment contingency measures provision of the Clean Air Act must concern only State Implementation Plans approved prior to a nonattainment area‘s attainment date. It would not make sense to make these measures contingent upon an already resolved condition that excuses their implementation: attainment of NAAQS. See (holding that regulatory measures that the state had already implemented cannot constitute contingency measures under the Clean Air Act because contingency measures must be unimplemented measures that take effect only “if a future event occurs“). The statute does not explain what should happen after EPA‘s issuance of a § 7511(b)(2) Attainment Determination makes it certain that these measures will never take effect. We therefore find § 7502(c)(9) to be silent on the question presented.

Our sister circuits have come to similar conclusions. Both the Tenth and D.C. Circuits have found a statutory gap exists as to whether EPA may excuse certain SIP requirements for nonattainment areas that actually attain the NAAQS but are not yet redesignated as “in attainment.” See (“The plain language of 42 U.S.C. § 7502(c)(9) similarly refers to contingency measures to be taken if an area ‘fails to make reasonable further progress, or to attain the national primary ambient air quality standard.’ When read as a whole to properly understand the statutory context, these two provisions fail to clearly require areas that have attained the ozone standard but have not yet been redesignated to attainment status to make further emission reductions.” (internal citation omitted)); (“The Act is therefore ambiguous as to what reductions are required when no further progress toward attainment is necessary . . . .“). As those courts conclude, the Clean Air Act does not by its plain terms require the seemingly illogical conclusion that states must comply with each SIP requirement under § 7502(c) even when the provision‘s function is made redundant by an attainment finding or when circumstances ensure that the statutory condition can never come about. We recognize that those courts were analyzing the RFP contingency measure requirement, but the premise applies equally to the attainment contingency measures requirement after the EPA determines a nonattainment area has attained the NAAQS by the attainment date, as was done here as to July 2018.

Having determined the statute is silent, we proceed to step two. “[I]f the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency‘s answer is based on a permissible construction of the statute.” . EPA argues its construction of the Clean Air Act—that contingency measures are not required for nonattainment areas for which EPA has issued a positive § 7511(b)(2) Attainment Determination—is a reasonable interpretation of the Clean Air Act‘s silence on the issue. We agree.

A contingency measure‘s sole purpose is to be implemented in the event the defined condition occurs. See Contingency, Black‘s Law Dictionary (11th ed. 2019) (“Dependent on something that might or might not happen in the future; conditional.” (second definition)); (“According to the dictionary definition, [contingency] means ‘a possible future event or condition or an unforeseen occurrence that may necessitate special measures.‘” (quoting Webster‘s Third New International Dictionary (2002)). EPA‘s interpretation that such measures may be waived or suspended if the only contingency upon which such measures are triggered cannot possibly occur does no violence to the statute or to EPA‘s ability to enforce the Clean Air Act‘s NAAQS program. The conclusion that attainment contingency measures are not required under the statute after a nonattainment area successfully attains the NAAQS by the attainment date is reasonable.

In response, Petitioners primarily argue that EPA‘s suspension of contingency measures places the Phoenix NAA, which has since supposedly lapsed back into nonattainment, in “nonattainment limbo” with no contingency measures to ensure future attainment of the NAAQS. We are unpersuaded. First, EPA‘s interpretation applies only to attainment contingency measures—RFP contingency measures are not automatically suspended after the attainment date. EPA might suspend RFP contingency measures under a similar rationale pursuant to its Clean Data Policy, but unlike attainment contingency measures, those measures may be suspended only as long EPA determines that the area has not violated that NAAQS.23 40 C.F.R. § 51.1118; see also (concluding that EPA‘s suspension of RFP contingency measures requirement for the Salt Lake City nonattainment area “is really no more than a suspension of those requirements for so long as the area continues to attain the standard or until the area is formally redesignated to attainment status.“). Second, Congress has already addressed the possibility that a nonattainment area could attain the NAAQS but lapse back into nonattainment. The Clean Air Act requires EPA, before officially redesignating a nonattainment area as in attainment, to “determine[] that the improvement in air quality is due to permanent and enforceable reductions in emissions.” 42 U.S.C. § 7407(d)(3)(E)(iii). Until the Phoenix nonattainment area is officially redesignated as in attainment, other SIP requirements remain in effect, including, as EPA points out in its brief, “requirements for emissions inventories, modeled demonstration of attainment, reasonably available control measures, reasonable further progress, motor vehicle emissions budgets, vehicle inspection and maintenance programs, new source review rules, and offsets.” EPA‘s interpretation does not operate as a way for states to avoid their ultimate responsibility under the Clean Air Act to obtain a lasting attainment of the NAAQS.

Based on the foregoing analysis, we conclude that the Clean Air Act is silent as to whether SIPs must contain attainment contingency measures after the attainment date and grant EPA‘s reasonable construction of 42 U.S.C. § 7502(c)(9) deference under . Accordingly, EPA did not act contrary to the Clean Air Act when it suspended the Phoenix nonattainment area‘s attainment continency measures requirement after EPA issued a § 7511(b)(2) Attainment Determination.

CONCLUSION

For the reasons stated above, the petition for review is DENIED. Petitioners are not entitled to request attorneys’ fees nor costs.

BUMATAY, Circuit Judge, concurring:

The Clean Air Act limits judicial review to only those objections that were “raised with reasonable specificity during the period for public comment.” 42 U.S.C. § 7607(d)(7)(B). The majority opinion persuasively explains how the petitioners failed to bring their objections regarding the Exceptional Events Rule and the Contingency Measures Requirement in a timely manner. Accordingly, we need not reach the merits of those objections here. See (holding that objections not raised during the notice and comment period are barred under § 7607(d)(7)(B)). I thus join the majority opinion except as to sections I.B and III.B.

Notes

1
EPA issues final determinations and findings in what EPA and most federal administrative agencies call “rules.”
3
EPA’s issuance of this § 7511(b)(2) Attainment Determination for the Phoenix nonattainment area is the rule at issue here.
4

40 C.F.R. § 51.1118 reads:

Upon a determination by EPA that an area designated nonattainment for the 2008 ozone NAAQS, or for any prior ozone NAAQS, has attained the relevant standard, the requirements for such area to submit attainment demonstrations and associated reasonably available control measures, reasonable further progress plans, contingency measures for failure to attain or make reasonable progress and other planning SIPs related to attainment of the 2008 ozone NAAQS, or for any prior NAAQS for which the determination has been made, shall be suspended until such time as: The area is redesignated to attainment for that NAAQS

or a redesignation substitute is approved as appropriate, at which time the requirements no longer apply; or EPA determines that the area has violated that NAAQS, at which time the area is again required to submit such plans.

5
In 2015, EPA again revised the standard to its current level at 0.070 ppm. 40 C.F.R. § 50.19. However, the Phoenix nonattainment area is still required to achieve the 2008 standard (0.075 ppm) by its attainment date. See 85 Fed. Reg. 33,571, 33,573. Arizona’s efforts to attain that 2008 standard are what is at issue here (the Phoenix NAA’s attainment date to achieve the 2015 standard is in 2021). Id.
6
Whether readings from a monitoring site reflect compliance with the NAAQS is determined by the “design value.” 42 U.S.C. § 7511(a). EPA determined the design value for ozone should be the 3-year average of the annual fourth-highest daily maximum 8-hour average. It is unnecessary further to break down EPA’s exact method of calculation of the ozone standard. Suffice it to say, a single day where a single monitoring site records exceedances of the standard can have significant impact on whether the region as a whole attains the NAAQS.
7

The exact elements read:

  1. The event satisfies the criteria set forth in 40 CFR 50.1(j);
  2. There is a clear causal relationship between the measurement under consideration and the event that is claimed to have affected the air quality in the area;
  3. The event is associated with a measured concentration in excess of normal historical fluctuations, including background; and
  4. There would have been no exceedance or violation but for the event.

40 C.F.R. § 50.14(c)(3)(iv) (2015).

8

The 2016 Exceptional Events Rule now requires:

  1. A narrative conceptual model that describes the event(s) causing the exceedance or violation and a discussion of how emissions from the event(s) led to the exceedance or violation at the affected monitor(s);
  2. A demonstration that the event affected air quality in such a way that there exists a clear causal relationship between the specific event and the monitored exceedance or violation;
  3. Analyses comparing the claimed event-influenced concentration(s) to concentrations at the same monitoring site at other times to support the requirement at paragraph (c)(3)(iv)(B) of this section. The Administrator shall not require a State to prove a specific percentile point in the distribution of data;
  4. A demonstration that the event was both not reasonably controllable and not reasonably preventable; and
  5. A demonstration that the event was a human activity that is unlikely to recur at a particular location or was a natural event.

40 C.F.R. § 50.14(c)(3)(iv) (2021).

9
The six exceeding monitors were Blue Point (0.077 ppm), Falcon Field (0.080 ppm), Mesa (0.079 ppm), Pinnacle Peak (0.078 ppm), Apache Junction (0.078 ppm), and Tonto National Monument (0.079 ppm).
10
is related but not relevant to the present action. There, we reviewed another of Bahr’s petitions for review concerning the Phoenix Ozone NAA. Bahr challenged EPA’s approval of the Arizona’s SIP. . Granting her petition for review, we held that EPA’s approval of the contingency measures in Arizona’s plan was unlawful because those measures had already been implemented, and the statute requires that contingency measures be as-yet unimplemented restrictions that may be put into effect in the future. .
11
Petitioners also claim they commented on another unique requirement: that ADEQ failed to show the exceedances would not have occurred but for the exceptional event. Their comment criticized ADEQ‘s regression analysis, which ADEQ expressly included to satisfy its but for requirement demonstration. But nowhere did Petitioners themselves discuss the but for requirement. It would be difficult to believe EPA could interpret Petitioners’ criticism of the State‘s regression analysis as suggesting EPA should be governed by the 2007 rule, not the 2016 rule. Also, regression analysis attempts to establish causation of an event or condition. Causation is an essential and required element under either the 2007 or 2016 rule as to exceptional events exclusions.
12
[A]lternative holdings are a common practice that prevents the overconsumption of adjudicative resources. . Our discussion here is prudent because this court is not one of last resort. The statute does not bar us from providing alternative holdings in the event our decision as to exhaustion is overruled. To be clear, our holding denying Petitioners’ objection gives effect to the statute‘s restriction on raising unexhausted objections to agency action. 42 U.S.C. § 7607(d)(7)(B). In otherwise explaining the lack of merit to Petitioners’ claims, we do not permit Petitioners to sidestep the statute‘s restriction on obtaining relief on the basis of unexhausted claims. Nor have we held that § 7607(d)(7)(B) is jurisdictional in nature, which may have acted to bar issuance of alternative holdings—indeed, there is some suggestion that it is not jurisdictional. See (Kavanaugh, J., concurring) (I note simply that the Section 7607 exhaustion/finality rule we describe today likely should not be considered jurisdictional under the Supreme Court‘s recent cases . . . .). In any event, the parties have not briefed that jurisdictional issue and we do not decide it here. Additionally, although Petitioners did not raise it as an argument, we have previously held that challenges to agency action based on retroactivity concerns may be excused from exhaustion requirements, albeit, in the immigration context and as to constitutional rather than statutory claims. See (Although ordinarily we do not hear an argument raised for the first time in a petition for review, we excuse the lack of exhaustion where a petitioner raises a challenge to the constitutionality of the statutes and regulations the BIA administers.); see also (applying this principle to a retroactivity challenge). We therefore think it advisable to address the merits, despite our holding that Petitioners forfeited this argument.
13
EPA argues the only relevant conduct governed by the Exceptional Events Rule was EPA‘s own conduct, i.e., its review of ADEQ‘s Lake Fire exceptional event demonstration. But it is not accurate to state that the Exceptional Events Rule governs only EPA behavior: the rule addresses requirements that state air agencies must meet in their demonstration package. 40 C.F.R. § 50.14. The Exceptional Events Rule is binding on both EPA, as to whether it may excuse monitoring data as an exceptional event, and the state air agencies, as to the types of proof they are required to supply. Thus, we must analyze as relevant the state‘s efforts to demonstrate an exceptional event. Notably, although the conduct by EPA and the ADEQ spanned both rules, both ADEQ‘s final demonstration package and EPA‘s final rule and concurrence occurred after the 2016 rule came into effect.
14
We also would note that an intervening implementation of a different articulation of a regulatory standard or balancing test does not automatically implicate retroactivity concerns—provided the new articulation does not dictate a certain outcome. See (holding the Attorney General‘s enhancement of the discretionary standard for waiver of inadmissibility—from extreme hardship to extreme and extremely unusual hardship—did not attach a new disability to an alien‘s past conduct or otherwise apply retroactively when, under either standard, alien was still eligible for discretionary relief). As noted, Petitioners do not demonstrate that the 2016 version of the Exceptional Events Rule dictates an outcome different from that which the 2007 version would produce.
15
EPA‘s regulations say much the same: A state‘s exceptional event demonstration must include [a] demonstration that the event affected air quality in such a way that there exists a clear causal relationship between the specific event and the monitored exceedance or violation. 40 C.F.R. § 50.14(c)(3)(iv)(B).
16
[I]n applying a weight of evidence approach to reviewing individual exceptional events demonstrations, the EPA believes it is appropriate to consider all relevant evidence and qualitatively weigh this evidence based on its relevance to the Exceptional Events Rule criterion being addressed, the degree of certainty, its persuasiveness, and other considerations appropriate to the individual pollutant and the nature and type of event. 84 Fed. Reg. at 60,921.
17
Some of Petitioners’ arguments show little merit. For example, they argue EPA should not have relied on ADEQ‘s regression analysis, but EPA explicitly stated it did not rely on the regression analysis. They argue the quantity of emissions over distance ratio (Q/D Ratio) was below the threshold EPA recommends for clear causality, but the Q/D Ratio is a mere screening metric EPA uses to determine which tier of demonstration is required, not a necessary threshold for determining clear causality. 84 Fed. Reg. at 60,925.
18
Trajectory analyses attempt to model the origins and trajectory of emissions. NOAA developed the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model, used by ADEQ here. The HYSPLIT model is a complicated, multifaceted computer modeling system that predicts how emissions will travel, taking into account not only wind flow, direction, and speed, but also particle behavior, dispersal mechanics, and the impact of meteorological and even radioactive factors. HYSPLIT, NOAA, https://www.arl.noaa.gov/hysplit/hysplit/ (last visited June 21, 2021). It has been used in a variety of simulations describing the atmospheric transport, dispersion, and deposition of pollutants and hazardous materials. Id. The HYPSLIT model is effective in tracking and forecasting the release of radioactive material, wildfire smoke, windblown dust, pollutants from various stationary and mobile emission sources, allergens and volcanic ash. Id.
19
Just why pointing to a different fire—another potential exceptional event—as the true cause of the smoke and ozone exceedances helps the petitioners’ claims is not immediately clear to us.
20
The 2008 eight-hour ozone nonattainment area covers approximately 5,017 square miles. Maricopa Association of Governments, Conformity Analysis 13 (2020).
21
Relatedly, we deny Petitioners’ argument that EPA‘s interpretation amounts to a renunciation or departure from its existing interpretation—the Clean Data Policy—without explanation. The basis for EPA‘s interpretation here is substantively different from the basis for the Clean Data Policy and is not a change in existing regulation.
22
As a preliminary issue, EPA‘s interpretation surmounts Mead‘s “step zero.” See (“EPA receives the benefit of deference under Chevron, which American Trucking held applicable to the ozone subchapter.” (internal citations omitted)).
23

That being said, we do note that EPA has since suspended RFP contingency measures for the Phoenix nonattainment area in a separate rulemaking based on its interpretation of RFP requirements for moderate ozone nonattainment areas:

With regard to the RFP contingency measure requirement, we proposed, in conjunction with our proposal on the [Maricopa Association of Governments] 2017 Ozone Plan, to find that the RFP contingency measure requirement would also no longer apply to the Phoenix NAA for the 2008 ozone NAAQS. We explained that the EPA‘s long-standing interpretation is that RFP contingency measures for Moderate areas would be triggered only by a finding that the area failed to attain the standard by the Moderate area attainment date. Because we have determined that the area has attained the standard by the attainment date, the RFP contingency measures have not, and will not, be triggered.

EPA, Clean Air Plans; 2008 8-Hour Ozone Nonattainment Area Requirements; Phoenix-Mesa, Arizona, 85 Fed. Reg. 33,571, 33,575 (June 2, 2020) (citations omitted). That rule is not before the court and does not affect our analysis.

*
The Honorable Kathleen Cardone, United States District Judge for the Western District of Texas, sitting by designation.
**
This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader.