Sandra L. Bahr; Jeanne Lunn; David Matusow v. MICHAEL REGAN, Administrator, United States Environmental Protection Agency

FOR PUBLICATION
UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
SANDRA L. BAHR; J EANNE LUNN;
DAVID M ATUSOW ,
Petitioners,
v.
M ICHAEL REGAN, Administrator,
United States Environmental
Protection Agency; DEBORAH
J ORDAN , Acting Regional
Administrator, U.S. EPA Region 9;
U.S. ENVIRONMENTAL PROTECTION
AGENCY,
Respondents.
No. 20-70092
EPA No.
EPA-R9-OAR-
2018-0821
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

-- 1 of 52 --

2 BAHR V. REGAN
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
* 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.

-- 2 of 52 --

BAHR V. REGAN 3
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.

-- 3 of 52 --

4 BAHR V. REGAN
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 Chevron, U.S.A. Inc. v. Natural
Resources Defense Council, Inc., 467 U.S. 837 (1984).
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.

-- 4 of 52 --

BAHR V. REGAN 5
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.

-- 5 of 52 --

6 BAHR V. REGAN
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

-- 6 of 52 --

BAHR V. REGAN 7
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
1 EPA issues final determinations and findings in what EPA and
most federal administrative agencies call “rules.”

-- 7 of 52 --

8 BAHR V. REGAN
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
Chevron.2 EPA’s suspension of the contingency measure
requirements in its July 2018 final rule did not violate the
Clean Air Act.
2 Chevron, U.S.A. Inc. v. Natural Resources Defense Council, Inc.,
467 U.S. 837 (1984).

-- 8 of 52 --

BAHR V. REGAN 9
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

-- 9 of 52 --

10 BAHR V. REGAN
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
3 EPA’s issuance of this § 7511(b)(2) Attainment Determination for
the Phoenix nonattainment area is the rule at issue here.

-- 10 of 52 --

BAHR V. REGAN 11
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
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

-- 11 of 52 --

12 BAHR V. REGAN
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.
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.

-- 12 of 52 --

BAHR V. REGAN 13
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
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.

-- 13 of 52 --

14 BAHR V. REGAN
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
7 The exact elements read:
(A) The event satisfies the criteria set forth in 40 CFR
50.1(j);
(B) 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;
(C) The event is associated with a measured
concentration in excess of normal historical
fluctuations, including background; and
(D) There would have been no exceedance or violation
but for the event.
40 C.F.R. § 50.14(c)(3)(iv) (2015).

-- 14 of 52 --

BAHR V. REGAN 15
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
8 The 2016 Exceptional Events Rule now requires:
(A) 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);
(B) 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;
(C) 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;
(D) A demonstration that the event was both not
reasonably controllable and not reasonably
preventable; and
(E) 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).

-- 15 of 52 --

16 BAHR V. REGAN
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

-- 16 of 52 --

BAHR V. REGAN 17
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
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).

-- 17 of 52 --

18 BAHR V. REGAN
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
10 Bahr v. EPA, 836 F.3d 1218 (9th Cir. 2016) 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. Id. at 1229–35.
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. Id. at 1235–36.

-- 18 of 52 --

BAHR V. REGAN 19
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. Sierra Club v. EPA (Sierra Club I), 346 F.3d 955, 961
(9th Cir. 2003). 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.’” Bahr v. EPA,
836 F.3d 1218, 1229 (9th Cir. 2016) (quoting Motor Vehicle
Mfrs. Ass’n of the U.S. v. State Farm Mut. Auto. Ins. Co.,
463 U.S. 29, 43 (1983)).
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. Landgraf v. USI Film Prod.,
511 U.S. 244, 265, 270 (1994). Petitioners argue EPA

-- 19 of 52 --

20 BAHR V. REGAN
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.” Marathon Oil Co. v. United States,
807 F.2d 759, 767 (9th Cir. 1986). 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.’” Nat’l Parks & Conservation Ass’n v. Bureau of
Land Mgmt., 606 F.3d 1058, 1065 (9th Cir. 2010) (quoting
Native Ecosystems v. Dombeck, 304 F.3d 886, 899 (9th Cir.
2002)); but see Nat’l Ass’n of Clean Air Agencies v. EPA,
489 F.3d 1221, 1231 (D.C. Cir. 2007) (“Objections must be

-- 20 of 52 --

BAHR V. REGAN 21
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.’” Pac. Choice Seafood Co. v. Ross,
976 F.3d 932, 942 (9th Cir. 2020) (quoting Glacier Fish Co.,
LLC v. Pritzker, 832 F.3d 1113, 1120 n.6 (9th Cir. 2016)).
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
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

-- 21 of 52 --

22 BAHR V. REGAN
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:
element under either the 2007 or 2016 rule as to exceptional events
exclusions.

-- 22 of 52 --

BAHR V. REGAN 23
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
12 “[A]lternative holdings are a common practice that prevents the
overconsumption of adjudicative resources.” Container Stevedoring Co.

-- 23 of 52 --

24 BAHR V. REGAN
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.” Landgraf, 511 U.S. at 280. In Landgraf, “[t]he
Supreme Court articulated a two-step approach for
v. Dir., Office of Workers Comp. Programs, 935 F.2d 1544, 1549 n.5
(9th Cir. 1991). 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 Util. Air Regul. Grp. v.
EPA, 744 F.3d 741, 751 (D.C. Cir. 2014) (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 Mejia v. Gonzales, 499 F.3d 991, 997 (9th Cir.
2007) (“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 Garcia-Ramirez v.
Gonzales, 423 F.3d 935, 938 (9th Cir. 2005) (applying this principle to
a retroactivity challenge). We therefore think it advisable to address the
merits, despite our holding that Petitioners forfeited this argument.

-- 24 of 52 --

BAHR V. REGAN 25
evaluating when the normal presumption against
retroactivity should not apply. Our ‘first task’ under
Landgraf is to ‘determine whether Congress has expressly
prescribed’” that a regulation is to be applied retroactively.
Garcia-Ramirez v. Gonzales, 423 F.3d 935, 939 (9th Cir.
2005) (quoting Landgraf, 511 U.S. at 280). 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.” Mejia v. Gonzales, 499 F.3d 991, 997
(9th Cir. 2007). “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.’” CFPB v.
Gordon, 819 F.3d 1179, 1196–97 (9th Cir. 2016) (quoting
Landgraf, 511 U.S. at 265). 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.” Landgraf,
511 U.S. at 269–70 (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.’” INS v. St. Cyr, 533 U.S. 289, 321 (2001) (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

-- 25 of 52 --

26 BAHR V. REGAN
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.
Landgraf, 511 U.S. at 270 (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
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.

-- 26 of 52 --

BAHR V. REGAN 27
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. Id.
at 269 n.24 (A regulation “is not made retroactive merely
because it draws upon antecedent facts for its operation.”
(quoting Cox v. Hart, 260 U.S. 427, 435 (1922)).
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.

-- 27 of 52 --

28 BAHR V. REGAN
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

-- 28 of 52 --

BAHR V. REGAN 29
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
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 Mejia, 499 F.3d
at 998 (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.

-- 29 of 52 --

30 BAHR V. REGAN
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
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).

-- 30 of 52 --

BAHR V. REGAN 31
made.” Barnes v. U.S. Dep’t of Transp., 655 F.3d 1124,
1132 (9th Cir. 2011). 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.’”
Sierra Club I, 346 F.3d at 961 (quoting Marsh v. Or. Nat.
Res. Council, 490 U.S. 360, 377 (1989)).
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
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.

-- 31 of 52 --

32 BAHR V. REGAN
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
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.

-- 32 of 52 --

BAHR V. REGAN 33
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
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.

-- 33 of 52 --

34 BAHR V. REGAN
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.

-- 34 of 52 --

BAHR V. REGAN 35
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
(“PM 2.5 ”), 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

-- 35 of 52 --

36 BAHR V. REGAN
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:
PM 2.5 and Elemental and Organic Carbon Analyses.
ADEQ submitted an analysis of concentrations of PM 2.5 —
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
PM 2.5 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 PM 2.5
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.

-- 36 of 52 --

BAHR V. REGAN 37
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
20 “The 2008 eight-hour ozone nonattainment area covers
approximately 5,017 square miles.” Maricopa Association of
Governments, Conformity Analysis 13 (2020).

-- 37 of 52 --

38 BAHR V. REGAN
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:

-- 38 of 52 --

BAHR V. REGAN 39
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

-- 39 of 52 --

40 BAHR V. REGAN
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

-- 40 of 52 --

BAHR V. REGAN 41
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 Chevron
deference and that Petitioners failed to exhaust argument to
the contrary before the agency.

-- 41 of 52 --

42 BAHR V. REGAN
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
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.

-- 42 of 52 --

BAHR V. REGAN 43
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 Pac. Choice
Seafood Co., 976 F.3d at 942. 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 Or. Nat.
Desert Ass’n v. Jewell, 840 F.3d 562, 572 (9th Cir. 2016)
(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

-- 43 of 52 --

44 BAHR V. REGAN
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 Litton Indus., Inc. v. FTC, 676 F.2d 364, 369–70 (9th
Cir. 1982); 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.” Glacier Fish Co., 832 F.3d
at 1120 n.6 (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.

-- 44 of 52 --

BAHR V. REGAN 45
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 Chevron 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 Chevron, U.S.A. Inc. v. Natural
Resources Defense Council, Inc., 467 U.S. 837 (1984)”.
Bahr, 836 F.3d at 1229–30 (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.” Vigil v. Leavitt, 381 F.3d 826, 834 (9th Cir. 2004)

-- 45 of 52 --

46 BAHR V. REGAN
(quoting Chevron, 467 U.S. at 842). “[I]f the statute is silent
or ambiguous with respect to the specific issue, we must ask
at Chevron step two whether the regulations promulgated by
the agency are based on a permissible construction of the
statute.” Safer Chemicals, Healthy Families v. EPA,
943 F.3d 397, 422 (9th Cir. 2019) (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 Chevron deference.”
Bahr, 836 F.3d at 1230 (9th Cir. 2016) (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.” Chevron, 467 U.S. at 842.
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
22 As a preliminary issue, EPA’s interpretation surmounts Mead’s
“step zero.” See Sierra Club v. EPA (Sierra Club II), 375 F.3d 537, 540
(7th Cir. 2004) (“EPA receives the benefit of deference under Chevron,
which American Trucking held applicable to the ozone subchapter.”
(internal citations omitted)).

-- 46 of 52 --

BAHR V. REGAN 47
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

-- 47 of 52 --

48 BAHR V. REGAN
implementation: attainment of NAAQS. See Bahr, 836 F.3d
at 1236–37 (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
Sierra Club v. EPA (Sierra Club III), 99 F.3d 1551, 1555
(10th Cir. 1996) (“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)); NRDC v. EPA,
571 F.3d 1245, 1260 (D.C. Cir. 2009) (“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

-- 48 of 52 --

BAHR V. REGAN 49
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
Chevron 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.” Chevron, 467 U.S. at 843. 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)); Bahr,
836 F.3d at 1235 (“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.

-- 49 of 52 --

50 BAHR V. REGAN
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.
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.

-- 50 of 52 --

BAHR V. REGAN 51
§ 51.1118; see also Sierra Club III, 99 F.3d at 1558
(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 Chevron. 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.

-- 51 of 52 --

52 BAHR V. REGAN
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
Mexichem Specialty Resins, Inc. v. EPA, 787 F.3d 544, 553
(D.C. Cir. 2015) (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.

-- 52 of 52 --