MS Comm. Environ. Quality v. EPA

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued October 21, 2014 Decided June 2, 2015

No. 12-1309

MISSISSIPPI COMMISSION ON ENVIRONMENTAL QUALITY,
PETITIONER

v.

ENVIRONMENTAL PROTECTION AGENCY AND GINA
MCCARTHY,
RESPONDENTS

STATE OF CONNECTICUT, ET AL.,
INTERVENORS

Consolidated with 12-1310, 12-1312, 12-1313, 12-1315,
12-1316, 12-1317, 12-1318, 12-1322, 12-1323, 12-1326,
12-1328, 13-1030, 13-1032, 13-1046, 13-1050, 13-1051,
13-1052, 13-1053, 13-1054

On Petitions for Review of Final Action of the
United States Environmental Protection Agency

Valerie Satterfield Edge, Deputy Attorney General, Office
of the Attorney General for the State of Delaware, argued the
cause for the petitioners Delaware Department of Natural
Resources and Environmental Control and the State of
Connecticut. George Jepsen, Attorney General, and Kimberly
 2
P. Massicotte and Scott N. Koschwitz, Assistant Attorneys
General, were with her on brief.

Robin L. Cooley and Robert Ukeiley argued the causes and
filed the joint briefs for Environmental Petitioners. James J.
Tutchton entered an appearance.

Donna J. Hodges and Reed D. Rubinstein argued the
causes for State and County Petitioners. Gary C. Rikard and
Mark L. Walters, Assistant Attorneys General, Office of the
Attorney General for the State of Texas, were with them on the
joint brief. Gregory W. Abbott, Attorney General, Office of the
Attorney General for the State of Texas, and Jonathan K.
Niermann, Assistant Attorney General, and Mary Ann Poirier
entered appearances.

Timothy J. Junk, Deputy Attorney General, Office of the
Attorney General for the State of Indiana, argued the cause for
the petitioner State of Indiana. Gregory F. Zoeller, Attorney
General, was with him on brief.

Roger R. Martella Jr. argued the cause for the Industrial
Petitioners. Timothy K. Webster, Ryan C. Morris, David C.
Duggins, Matt Paulson, Howard Rubin, Glen Donath,
Christopher D. Jackson, William L. Wehrum and Aaron M.
Flynn were with him on brief.

Elizabeth B. Dawson and Jessica O’Donnell, Attorneys,
United States Department of Justice, argued the causes for the
respondent. Robert G. Dreher, Acting Assistant Attorney
General, and Jan Tierney, Attorney, United States
Environmental Protection Agency, were with them on brief.
 3
Sean D. Reyes, Attorney General, Office of the Attorney
General for the State of Utah, Bridget Romano, Utah Solicitor
General, Connie S. Nakahara, Assistant Utah Attorney
General, Constance E. Brooks, David G. Scott and Bret A.
Sumner were on the joint brief for the respondent-intervenors
State of Utah, et al. Mark L. Shurtleff, former Attorney
General, Office of the Attorney General for the State of Utah,
entered an appearance.

Tómas Carbonell and Peter Zalzal were on brief for the
respondent-intervenor Environmental Defense Fund. Vickie
L. Patton entered an appearance.

Before: GARLAND, Chief Judge, and HENDERSON and
SRINIVASAN, Circuit Judges.

Opinion for the Court filed PER CURIAM.

PER CURIAM: The Congress enacted the Clean Air Act
(the Act), 42 U.S.C. §§ 7401 et seq., “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.” Id. § 7401(b)(1). At issue in this case is Title I
of the Act, which requires the Environmental Protection
Agency (EPA) to promulgate National Ambient Air Quality
Standards (NAAQS), thus setting the maximum level of
permissible pollutant concentration in the atmosphere. See id.
§§ 7408(a)(1), 7409(a)–(b). After the EPA sets the NAAQS,
it must determine whether each state is in compliance with
these air-quality standards and, in the event of a NAAQS
violation, how to establish the geographic boundaries around
the non-compliant area. See id. § 7407(d)(1).

In these consolidated petitions, several states, counties,
industrial entities and environmental organizations challenge
 4
the EPA’s determination that certain geographic areas are, or
are not, in “attainment” with the EPA’s ground-level ozone
NAAQS. Id. Some argue that the Act, as applied to them,
violates various Constitutional provisions; others argue that the
EPA misconstrued the terms of the Act. Virtually every
petitioner argues that, for one reason or another, the EPA acted
arbitrarily and capriciously in making its final NAAQS
designations. But because the EPA complied with the
Constitution, reasonably interpreted the Act’s critical terms
and wholly satisfied—indeed, in most instances,
surpassed—its obligation to engage in reasoned
decision-making, we deny the consolidated petitions for
review in their entirety.

I. BACKGROUND

The EPA began the odyssey resulting in these
consolidated petitions nearly seven years ago. Along the way,
it construed a variety of the Act’s provisions, promulgated
regulations and issued informal guidance to assist in the
collaborative area-designation effort between it and the states.
Before discussing the substance of the issues, a brief overview
of the Act and the underlying proceedings in this case is in
order.

A. THE CLEAN AIR ACT

Under the Act, the EPA must promulgate NAAQS, which
set the maximum ambient, or outdoor, air concentrations for
six pollutants that “may reasonably be anticipated to endanger
public health or welfare.” 42 U.S.C. § 7408(a)(1). Once it
establishes a NAAQS, the EPA must designate each “area” in
the United States as “attainment” or “nonattainment.” See id.
§ 7407(d)(1)(A)(i)–(ii). Alternatively, the EPA may
designate an area as “unclassifiable” if the area “permit[s] no
determination given existing data.” Catawba Cnty., N.C. v.
 5
EPA, 571 F.3d 20, 26 (D.C. Cir. 2009) (citing 42 U.S.C.
§ 7407(d)(1)(A)(i)–(iii)). The EPA treats an “unclassifiable”
area as if it were in attainment. See 42 U.S.C. § 7471.

Generally speaking, the EPA designates an area that meets
the relevant NAAQS as in attainment, while areas that exceed
the NAAQS receive a nonattainment designation. See
Catawba Cnty., 571 F.3d at 26. But even if an area’s ambient
air concentration complies with the relevant NAAQS, the EPA
nonetheless designates it as nonattainment if it “contributes” to
a NAAQS violation in a “nearby area.” See 42 U.S.C.
§ 7407(d)(1)(A)(i). The Act does not define the terms
“contributes,” “nearby” or “area.”

The EPA works collaboratively with the states to
determine the NAAQS-attainment status for all areas within a
respective state’s borders. No later than one year after the
EPA promulgates a new or revised NAAQS, each state must
submit recommended “initial designations” to the EPA. Id.
§ 7407(d)(1)(A). A state’s initial designations must suggest
both the appropriate geographic boundaries for each “area” and
whether the EPA should classify the suggested area as
attainment, nonattainment or unclassifiable. See id.
§ 7407(d)(1)(A)–(B).

Once it receives a state’s initial designations, the EPA
may either promulgate them as submitted or modify them as it
“deems necessary.” Id. § 7407(d)(1)(B)(ii). The Act gives
the EPA discretion to change a state’s recommended
designation, to alter a state’s proposed geographic area or both.
See id. Although the EPA “has no obligation to give any
quantum of deference to a designation that it ‘deems
necessary’ to change,” Catawba Cnty., 571 F.3d at 40, it must
nonetheless notify the state of any intended change and provide
the state with at least 120 days “to demonstrate why any
 6
proposed modification is inappropriate,” 42 U.S.C.
§ 7407(d)(1)(B)(ii). These notifications are known as
“120-day letters.” See Air Quality Designations for the 2008
Ozone National Ambient Air Quality Standards, 77 Fed. Reg.
30,088, 30,090 (May 21, 2012) [hereinafter 2008 Designations
Rule].

While the EPA has ultimate authority to determine each
area’s attainment status, each state has “primary
responsibility” for ensuring that the geographic areas within its
borders either maintain attainment or progress towards it. 42
U.S.C. § 7407(a). Accordingly, once the EPA finalizes its
designations, each state must submit to the EPA a State
Implementation Plan (SIP) specifying how the NAAQS “will
be achieved and maintained.” Id. For areas in attainment,
the SIP must simply “contain emission limitations and such
other measures as may be necessary . . . to prevent significant
deterioration of air quality.” Id. § 7471.

For a nonattainment area, however, the Act imposes more
stringent requirements. A SIP from a state with a
nonattainment area must demonstrate that the state intends to
implement “all reasonably available control measures” and
“reasonably available control technology” to bring the area
into attainment. Id. § 7502(c)(1). The Act also imposes
deadlines, or “attainment dates,” on an offending area. See id.
§ 7502(a)(2)(A). For a violation of a primary 1 NAAQS, the
offending state must reach attainment “as expeditiously as
practicable, but no later than 5 years from the date such area
was designated nonattainment.” Id. The EPA “may extend
the attainment date to the extent [it] determines appropriate”
but only “for a period no greater than 10 years from the date of
designation as nonattainment.” Id. Taken together, these

1
See infra n.2.
 7
two requirements often mean that a state with a nonattainment
area must implement potentially expensive technology or
expensive process changes to reduce pollution levels over a
relatively short period of time. If a state fails to reach
attainment timely and the failure is due to inadequate
implementation efforts, sanctions can be imposed, including
loss of federal highway funds and increasingly severe
restrictions on emissions sources within the state. See id.
§ 7509(a)–(b).

B. THE 2008 OZONE NAAQS AND THE EPA’S
2008 GUIDANCE

On March 12, 2008, the EPA promulgated new primary
and secondary NAAQS for ambient ozone, 2 a component of
urban smog. See 2008 Designations Rule, 77 Fed. Reg. at
30,089. Even though ozone is an “essential presence in the
atmosphere’s stratospheric layer,” it becomes harmful at
ground level and “can cause lung dysfunction, coughing,
wheezing, shortness of breath, nausea, respiratory infection,
and in some cases, permanent scarring of the lung tissue.”
S. Coast Air Quality Mgmt. Dist. v. EPA, 472 F.3d 882, 887
(D.C. Cir. 2006) (quoting Henry A. Waxman, An Overview of
the Clean Air Act Amendments of 1990, 21 ENVTL. L. 1721,
1758 (1991)). It also “has a broad array of effects on trees,

2
“Primary” NAAQS exist to protect the “public health,” 40
C.F.R. § 50.2(b), and they ensure the safety of “sensitive”
populations such as asthmatics, children and the elderly. See
National Ambient Air Quality Standards (NAAQS), EPA,
http://www.epa.gov/air/criteria.html (last updated Oct. 21, 2014).
“Secondary” NAAQS exist to protect the “public welfare,” 40
C.F.R. § 50.2(b), and they prevent harms like decreased visibility
and damage to animals, crops, vegetation and buildings. See
National Ambient Air Quality Standards (NAAQS), EPA,
http://www.epa.gov/air/criteria.html (last updated Oct. 21, 2014).
 8
vegetation, and crops and can indirectly affect other ecosystem
components such as soil, water, and wildlife.” Mississippi v.
EPA, 744 F.3d 1334, 1340 (D.C. Cir. 2013). Because ozone
forms at ground level when “ozone precursors”—specifically,
nitrous oxides (NOx) and volatile organic compounds
(VOCs)—react with sunlight, NAAQS compliance largely
depends on reducing emissions from ozone-precursor
producers like power plants, industrial compounds, motor
vehicles and combustion engines. See 2008 Designations
Rule, 77 Fed. Reg. at 30,089. Complicating this task is that
ozone and ozone precursors travel easily through the
atmosphere, which can result in NAAQS violations hundreds
of miles away from the source of the ozone precursors. See id.

Both the EPA’s 2008 primary and secondary ozone
NAAQS reduced the maximum allowable daily average
eight-hour level of ozone from 0.08 parts per million (ppm) to
0.075 ppm. See National Ambient Air Quality Standards for
Ozone, 73 Fed. Reg. 16,436, 16,436–37 (Mar. 27, 2008). By
setting these new NAAQS, the EPA triggered the states’
responsibility to submit their initial designations. See 42
U.S.C. § 7407(d)(1)(A). To assist this process, the EPA
issued a guidance titled “Area Designations for the 2008
Revised Ozone National Ambient Air Quality Standards”
[hereinafter 2008 Guidance] on December 4, 2008, which
included several matters relevant to the instant petitions.

First, the 2008 Guidance instructed states on the quality of
data it expected them to consider. Specifically, it
recommended that the states “identify violating areas using the
most recent three consecutive years of quality-assured,
certified air quality data.” 2008 Guidance at 2. The 2008
Guidance also informed the states that “[i]n general, [NAAQS]
violations [will be] identified using data from . . . monitors that
 9
are sited and operated in accordance with [EPA regulations
located at] 40 C.F.R. Part 58.” Id.

Second, the 2008 Guidance provided instruction for
establishing geographic boundaries around nonattainment
areas, noting first that the “EPA believes it is important to
examine ozone-contributing emissions across a relatively
broad geographic area.” 2008 Guidance at 3. Accordingly,
the 2008 Guidance recommended that if an air-quality monitor
reports a NAAQS violation, the state should consider using the
Core Based Statistical Area (CBSA) or Combined Statistical
Area (CSA) in which the monitor is located as the
“presumptive” boundary. 3 Id. If the violating monitor is not
in a CSA or CBSA, the 2008 Guidance recommended using the
county in which the violating monitor is located as the
presumptive boundary. Id.

3
A CBSA is defined by the Office of Management and Budget
(OMB) as:
[A] statistical geographic entity consisting of the
county or counties associated with at least one core
(urbanized area or urban cluster) of at least 10,000
population, plus adjacent counties having a high
degree of social and economic integration with the
core as measured through commuting ties with the
counties containing the core.
See Standards for Defining Metropolitan and Micropolitan
Statistical Areas, 65 Fed. Reg. 82,228, 82,238 (Dec. 27, 2000). A
CSA is formed by two or more adjacent CBSAs if there is sufficient
“employment interchange” between them. Id. In other words,
CSAs and CBSAs are both roughly equivalent to a “metropolitan”
area. See generally id. at 82,235–36. Throughout this opinion, we
use the term “metropolitan area” to refer to the CSA or CBSA, as
defined in the 2008 Guidance. See 2008 Guidance at 3 & n.2.
 10
The 2008 Guidance made plain, however, that CSAs,
CBSAs and county lines were merely presumptive boundaries,
recognizing that “area-specific analyses . . . may support
nonattainment area boundaries that are larger or smaller than
the presumptive area starting point.” Id. Stressing that “each
potential nonattainment area should be evaluated on a
case-by-case basis,” the 2008 Guidance instructed the states to
consider nine factors when determining a nonattainment area’s
borders. See id. at 2, Attach. 2. These include (1) air-quality
data; (2) emissions data (such as location of emissions sources
and contribution to ozone concentrations); (3) population
density and degree of urbanization (including commercial
development); (4) traffic and commuting patterns;
(5) population growth rates and patterns; (6) meteorology
(such as weather and air-transport patterns); (7) geography and
topography (such as mountain ranges or other air-basin
boundaries that could affect ozone dispersion);
(8) jurisdictional boundaries (such as counties, air districts,
existing nonattainment area boundaries and regional planning
authority boundaries) and (9) the level of control of emissions
sources. See id. Attach. 2. The 2008 Guidance stated that the
EPA planned to consider these same factors, “along with any
other relevant information,” in determining whether to modify
the states’ initial designations. Id.

C. THE 2008 OZONE DESIGNATION PROCESS

By 2009, all states had submitted their initial designations
to the EPA. Rather than immediately reviewing the initial
designations, however, the EPA halted the designation process
to consider whether to lower the ozone NAAQS even further.
This delay prompted a lawsuit by WildEarth Guardians—an
environmental-group petitioner in this case—that sought to
compel the EPA to complete the stalled ozone NAAQS
 11
designation process. 4 The EPA and WildEarth Guardians
eventually entered into a consent decree that required the EPA
to finalize its designations no later than May 31, 2012. See
2008 Designations Rule, 77 Fed. Reg. at 30,091.

The EPA notified the states in September 2011 that it
intended to finalize the ozone NAAQS designations by the
May 31, 2012 deadline set forth in the consent decree. In
accordance with the 2008 Guidance’s instruction to “identify
violating areas using the most recent three consecutive years of
quality-assured, certified air quality data,” 2008 Guidance at 2,
virtually every state had already submitted air-quality data
from 2008 to 2010 by the time the EPA resumed the
designation process. Although the EPA assured the states that
it still planned to consider the recommended designations and
ozone data they had submitted initially, it recognized that some
states may have collected more recent air-quality data for their
regions. For this reason, the EPA allowed the states to
provide updated recommendations and analyses—so long as
any updated air-quality data was certified for quality—but
assured them that they were under no obligation to do so. In
response to this invitation, several states updated their initial
designations and some submitted air-quality data from 2009 to
2011 to replace their older 2008 to 2010 data. The states
seeking to use data from 2009 to 2011 agreed to certify their
data for quality by February 29, 2012, so that the EPA had
sufficient time to consider the more recent data in advance of
its May 31, 2012 deadline to finalize the designations.

The EPA then reviewed each state’s initial designations to
determine whether to modify them. It first examined the
air-quality submissions from the states to determine which

4
See WildEarth Guardians, et al. v. Jackson,
No. 2:11-CV-01661 (D. Ariz. filed Aug. 24, 2011).
 12
monitors reported ozone NAAQS violations. If a state
certified its air-quality data from 2011 by the February 29,
2012 deadline, the EPA generally considered its air-quality
data from the years 2009 to 2011. For all other states, the
EPA considered air-quality data from 2008 to 2010.

After identifying NAAQS-violating monitors, the EPA
decided whether to alter the states’ respective recommended
nonattainment boundaries. To do so, the EPA used a
multi-factor, weight-of-the-evidence test that tracked—but
was not identical to—the nine-factor test in the 2008 Guidance.
Specifically, the EPA collapsed the 2008 Guidance’s
nine-factor test into a five-factor test, which examined (1) “Air
Quality Data,” or whether an area’s monitor reported a
NAAQS violation; (2) “Emissions Data,” including emissions
levels and controls, population, population density, population
growth, degree of urbanization and traffic and commuting
patterns; (3) “Meteorology,” including wind speed and
direction; (4) “Geography/Topography,” which examined the
effect of physical land features on the distribution of ozone and
(5) “Jurisdictional Boundaries,” which helped determine
whether certain areas could effectively carry out air-quality
planning and enforcement functions for nonattainment areas.

Once attainment designations were made, the EPA
notified the states of any proposed modifications it deemed
necessary and invited them to submit any additional data or
comments they wished to have the EPA consider. Although
not required by statute, see 42 U.S.C. § 7407(d)(2)(B), the
EPA also opened a 30-day public comment period on the
proposed notifications. Several states, organizations and
members of the public—including many of the petitioners in
this case—submitted comments. The EPA considered the
comments and then promulgated its final designations, which
identified 48 nonattainment areas in 26 states, the District of
 13
Columbia and Indian country. The nonattainment areas
included 192 counties in toto and 36 counties in part. The
EPA published the majority of its final designations on May
21, 2012, see 2008 Designations Rule, 77 Fed. Reg. at 30,088,
and in the case of certain Chicago-area designations, on June
11, 2012, see Air Quality Designations for the 2008 Ozone
National Ambient Air Quality Standards for Several Counties
in Illinois, Indiana, and Wisconsin; Corrections to Inadvertent
Errors in Prior Designations, 77 Fed. Reg. 34,221, 34,221
(June 11, 2012).

After the EPA received and denied 29 petitions for
reconsideration, the parties in this consolidated case 5
petitioned this Court for review. We have jurisdiction under
42 U.S.C. § 7607(b)(1).

5
See Del. Dep’t of Natural Res. & Envtl. Control v. EPA,
No. 12-1310 (D.C. Cir.); Tex. Pipeline Ass’n v. EPA, No. 12-1312
(D.C. Cir.); Wise Cnty., Tex. v. EPA, No. 12-1313 (D.C. Cir.);
Indiana v. EPA, No. 12-1315 (D.C. Cir.); Texas v. EPA,
No. 12-1316 (D.C. Cir.); Sierra Club v. EPA, No. 12-1317 (D.C.
Cir.); Gas Processors Ass’n v. EPA, No. 12-1318 (D.C. Cir.); Devon
Energy Corp. v. EPA, No. 12-1322 (D.C. Cir.); Targa Resources
Corp. v. EPA, No. 12-1323 (D.C. Cir.); WildEarth Guardians v.
EPA, No. 12-1326 (D.C. Cir.); DeSoto Cnty., Miss. v. EPA,
No. 12-1328 (D.C. Cir.); Sierra Club v. EPA, No. 13-1030 (D.C.
Cir.); WildEarth Guardians v. EPA, No. 13-1032 (D.C. Cir.); Wise
Cnty., Tex. v. EPA, No. 13-1046 (D.C. Cir.); Devon Energy Corp. v.
EPA, No. 13-1050 (D.C. Cir.); Tex. Pipeline Ass’n v. EPA,
No. 13-1051 (D.C. Cir.); Gas Processors Ass’n v. EPA, No. 13-1052
(D.C. Cir.); Texas v. EPA, No. 13-1053 (D.C. Cir.); Targa Res.
Corp. v. EPA, No. 10-1054 (D.C. Cir.).
 14
II. COMMON LEGAL PRINCIPLES

Before addressing the petitioners’ individual challenges,
we think it helpful to discuss several principles that bear on
most, if not all, of the issues the petitioners have raised.

First, we review the EPA’s NAAQS designations under
the same standard we use in reviewing a challenge brought
under the Administrative Procedure Act (APA). See Allied
Local & Reg’l Mfrs. Caucus v. EPA, 215 F.3d 61, 68 (D.C. Cir.
2000). Accordingly, we will set aside a NAAQS designation
by the EPA only if it is “arbitrary, capricious, an abuse of
discretion, or otherwise not in accordance with law.”
Catawba Cnty., 571 F.3d at 41 (quoting 5 U.S.C. § 706(2)(A)).
We must, however, give an “extreme degree of deference” to
the EPA’s evaluation of “scientific data within its technical
expertise,” City of Waukesha v. EPA, 320 F.3d 228, 247 (D.C.
Cir. 2003), especially where, as here, we review the “EPA’s
administration of the complicated provisions of the Clean Air
Act.” Catawba Cnty., 571 F.3d at 41 (citing Nat’l Ass’n of
Clean Air Agencies v. EPA, 489 F.3d 1221, 1229 (D.C. Cir.
2007)). Because the EPA’s “basic obligation” is to conduct
“reasoned decisionmaking,” id. at 25, we will uphold its action
if the record shows that the EPA “considered all relevant
factors and articulated a ‘rational connection between the facts
found and the choice made,’ ” id. at 41 (quoting Burlington
Truck Lines, Inc. v. United States, 371 U.S. 156, 168 (1962)).

Second, we have long since rejected the argument that the
EPA violates the Act if it uses a holistic, multi-factor,
weight-of-the-evidence test for determining whether a given
area contributes to a NAAQS violation. See ATK Launch
Sys., Inc. v. EPA, 669 F.3d 330, 336–37 (D.C. Cir. 2012)
(challenge to 2006 fine particulate matter NAAQS
designations); Catawba Cnty., 571 F.3d at 46 (challenge to
 15
1997 fine particulate matter NAAQS designations). Indeed,
in Catawba County, we made explicit that the EPA does not
violate the Act even if it fails to adopt “a bright-line,
‘objective’ test” for determining contribution and we also held
that the “EPA’s failure to quantify its analysis” does not render
“its interpretation of ‘contribute’ arbitrary and capricious and
therefore unreasonable.” 571 F.3d at 39. Rather, because
“[a]n agency is free to adopt a totality-of-the-circumstances
test to implement a statute that confers broad discretionary
authority, even if that test lacks a definite ‘threshold’ or ‘clear
line of demarcation to define an open-ended term,’ ” we have
held that, “[t]o be reasonable, such an ‘all-things-considered
standard’ must simply define and explain the criteria the
agency is applying.” Id.

With this background in mind, we now turn to the
petitioners’ challenges.

III. THE PETITIONERS’ CHALLENGES

A. DELAWARE & CONNECTICUT

We begin with a challenge to the EPA’s construction of
the key statutory provision in this case. Petitioners Delaware
and Connecticut challenge the EPA’s refusal to designate
broad, multi-state nonattainment areas to address the issue of
long-range ozone transport. According to the States, the
EPA’s final designations are inconsistent with its statutory
mandate to designate areas as nonattainment if they
“contribute[] to ambient air quality in a nearby area that does
not meet [the NAAQS].” 42 U.S.C. § 7407(d) (emphasis
added). We conclude, to the contrary, that the designations
are consistent with the EPA’s reasonable interpretation of the
ambiguous statutory term “nearby.”
 16
After the EPA reopened the designation process in 2011,
Delaware proposed a nonattainment area that would stretch
across 16 upwind states and the District of Columbia—to states
as far west as Missouri. Connecticut similarly proposed an
18-state nonattainment area, also stretching west to Missouri.
Both States argued for what Delaware described as a “more
workable definition of ‘nearby’ ”—one that would ask
“whether a source is ‘near enough to contribute’ to
nonattainment or interfere with maintenance.” Letter from
Del. Dep’t of Natural Res. & Envtl. Control to EPA 5 (Oct. 28,
2011) [hereinafter Delaware Response].

The EPA, however, had taken a different approach in the
2008 Guidance, instead interpreting “nearby” as presumptively
including counties in the same metropolitan area as the
violating county. 2008 Guidance at 3. In the Guidance, the
EPA acknowledged that certain regions have ozone transport
problems, but it concluded that the Act “does not require that
all contributing areas be designated nonattainment, only the
nearby areas.” Id. at 4. The agency explained that
“[r]egional strategies, such as those employed in the Ozone
Transport Region and EPA’s NOx SIP Call are needed to
address the long-range transport component of ozone
nonattainment.” Id. In keeping with this understanding of
the statute, the EPA declined to designate “super-regional”
nonattainment areas, see Responses to Significant Comments
on the State and Tribal Designation Recommendations for the
2008 Ozone NAAQS at 8–9 (Apr. 30, 2012) [hereinafter
Response to Comments], and instead made more limited
nonattainment designations in both Delaware and Connecticut,
see Delaware Area Designations for the 2008 Ozone NAAQS
 17
2; Connecticut Area Designations for the 2008 Ozone NAAQS
1. 6

We evaluate the EPA’s interpretation of a Clean Air Act
provision under the familiar two-step Chevron framework.
See Util. Air Regulatory Grp. v. EPA, 134 S. Ct. 2427, 2439
(2014) (citing Chevron, U.S.A., Inc. v. Natural Res. Def.
Council, Inc., 467 U.S. 837, 842–43 (1984)). The first
question—“whether Congress has directly spoken to the
precise question at issue,” Chevron, 467 U.S. at 842—has
previously been resolved by this Court. In Pennsylvania
Department of Environmental Protection v. EPA (PADEP), we
held that the statutory term “nearby” in section 107(d) is
ambiguous; indeed, we reached that conclusion in the course of
addressing the precise argument that Delaware makes here.
See 429 F.3d 1125, 1129–30 (D.C. Cir. 2005). In Catawba
County, we reached the same conclusion. See 571 F.3d at 35
(noting that section 107(d) does not define “nearby,” and that it
is “the kind[] of word[] that suggest[s] a congressional intent to
leave unanswered questions to an agency’s discretion and
expertise”).

Recognizing these precedents, Delaware and Connecticut
conceded at oral argument that our analysis must be governed
by Chevron’s second step, Oral Arg. Recording at 3:49–3:54,
which requires us to ask only whether the EPA’s interpretation
is reasonable, see, e.g., PADEP, 429 F.3d at 1130. But we
have addressed that question once as well, also in PADEP,
where we said that “Chevron requires that we defer to the
agency’s reasonable interpretation of the term, and Delaware

6
Neither State challenges the designations of those areas as
nonattainment, other than to contend that the designations should
have covered much broader areas.
 18
has given us no reason to think that EPA’s interpretation is
unreasonable.” Id. We reach the same conclusion here.

First, the agency’s interpretation of “nearby”—as
presumptively including counties within the same metropolitan
area as the violating county—falls readily within the dictionary
definition of “nearby” as “close at hand; not far off; adjacent;
neighboring.” RANDOM HOUSE COLLEGE DICTIONARY 889
(rev. ed. 1980). By contrast, neither the dictionary nor
common parlance would regard Missouri as “nearby” to
Connecticut or Delaware, as the petitioners’ proposals would
require.

Second, the EPA’s construction is consistent with the
approach the agency has taken in prior designations
proceedings—an approach that this Court has previously
upheld as reasonable. See PADEP, 429 F.3d at 1127, 1129–
30; 2008 Guidance at 3.

Third, the EPA’s construction is consistent with the
statutory scheme. The EPA selected the metropolitan area as
the presumptive “nearby” area for its contribution analysis in
part because the Congress itself chose the metropolitan area as
the default boundary for ozone nonattainment areas classified
as “serious,” “severe,” or “extreme.” See 42 U.S.C.
§ 7407(d)(4)(A)(iv); 2008 Guidance at 3 n.5. The Congress’
choice is certainly evidence that the legislature envisioned
broad but relatively local nonattainment areas. 7

7
At oral argument, the EPA made clear that it does not contend
that its reading is the only permissible reading of the statute. Oral
Arg. Recording at 30:01–30:59; see also 2008 Designations Rule, 77
Fed. Reg. at 30,090 (discussing the agency’s “discretion” to interpret
the term “nearby” in fixing the geographic scope of nonattainment
areas).
 19
As in PADEP, the petitioners argue that the EPA’s
interpretation is unreasonable because it fails to appreciate the
role of ozone transport, and consequently yields designations
that fail to include the true contributors to their nonattainment
status. See PADEP, 429 F.3d at 1129–30. Delaware notes,
for example, that 84 to 94 per cent of its ozone results from the
contributions of other states, including states as far west as
Missouri. See Delaware Reply Br. 4. Without emissions
reductions from those states, petitioners argue, they cannot
meet the 0.075 ppm standard. Thus, by failing to address the
principal sources of their ozone pollution, the EPA’s
interpretation eliminates any possibility that they will attain the
NAAQS. 8

Although we are sympathetic to the petitioners’ concerns,
our role is not to decide whether their proposed interpretation
is reasonable. Instead, the sole question before us is whether
the EPA interpreted the term reasonably and consistently with
the statute. See PADEP, 429 F.3d at 1130 (noting that,
although a broader “construction of ‘nearby’ may well be
sensible, Chevron requires that we defer to the agency’s
reasonable interpretation of the term”). Here, the EPA had
already considered the problem the petitioners raised. Part of
the rationale for using the metropolitan area as the starting
point for the contribution analysis was to account for ozone
transported from outside the violating county. See 2008

8
Delaware points to the isolated nonattainment zone of Sussex
County as a particularly egregious example of the designations that
the EPA’s interpretation produced. Delaware Br. 12. But even if
over 90 per cent of Sussex County’s pollution comes from
out-of-state sources, as Delaware asserts, the EPA found that no
surrounding counties had the linkages necessary to justify a
nonattainment designation under the agency’s five-factor analysis.
See Delaware Area Designations at 37–49.
 20
Guidance at 3–4. Although this approach does not fully
account for longer-range, interstate transport, the EPA has
addressed that problem in regulations promulgated under other
provisions of the Act. See, e.g., Federal Implementation
Plans: Interstate Transport of Fine Particulate Matter and
Ozone and Correction of SIP Approvals, 76 Fed. Reg. 48,208
(Aug. 8, 2011) (promulgating the Cross State Air Pollution
Rule, commonly referred to as the Transport Rule). 9
Although the petitioners recognize the EPA’s reliance on those
other regulatory options, they maintain that they “have been
less than successful” up to this point. Delaware Br. 6; see also
id. at 9. We, however, must defer to the EPA’s reasonable
judgment that regional strategies adopted pursuant to other
statutory provisions specific to long-range ozone transport
remain the appropriate means for addressing this problem.
See 2008 Guidance at 4.

The petitioners note that our decision in PADEP rested in
part upon the fact that there, Delaware had “offered no
evidence that ‘in practice’ EPA will not enlarge a
nonattainment area in response to [its then] eleven-factor
analysis.” 429 F.3d at 1130. Indeed, in PADEP, Delaware
had failed altogether “to produce an eleven-factor analysis.”
Id. But we did not mean by this to suggest that, had Delaware
produced the appropriate factor analysis, the EPA would have

9
The EPA promulgated the Transport Rule under 42 U.S.C.
§ 7410(a)(2)(D), which requires SIPs to prohibit air pollution that
will “contribute significantly to nonattainment in, or interfere with
maintenance [of the NAAQS] by, any other State.” Other
provisions of the Act also address interstate transport. See id.
§ 7506a (providing for interstate transport commissions); id. § 7511c
(establishing ozone transport region consisting of 11 states and the
District of Columbia, which must comply with additional control
measures).
 21
been required to adopt an interpretation of “nearby” that
included states as far away as those within the petitioners’
proposed nonattainment areas. The points discussed
above—including the dictionary definition of “nearby” and the
consistency of the EPA’s interpretation with the statute and its
prior practice—strongly suggest that the EPA’s narrower
interpretation would still be reasonable.

Nonetheless, if the petitioners had submitted a persuasive
five-factor analysis establishing contributions from
farther-away states, that would be relevant to our assessment of
the reasonableness of the EPA’s refusal to enlarge the
nonattainment area beyond its presumptive scope. In this
case, however, although the petitioning States did submit
technical analyses, they failed to demonstrate the requisite
linkages under the EPA’s 2008 Guidance. See, e.g., Delaware
Response Attach. 2 at 5–7, 11–13 (disputing relevance of
factors related to urbanization, traffic, and economic growth);
id. at 14–15 (with respect to meteorology factor, describing
long-range transport without describing weather patterns
within the proposed 16-state nonattainment area). Hence, the
petitioners did not show that the agency “will not enlarge a
nonattainment area in response to” the (current) five-factor
analysis, PADEP, 429 F.3d at 1130. Rather, the States’
analyses were simply insufficient to overcome the agency’s
definitional presumption.

In sum, we conclude that the EPA’s final designations of
Delaware and Connecticut counties are consistent with a
reasonable interpretation of the Clean Air Act. 10

10
Delaware also argues that the EPA acted inconsistently with
the statute by only designating as nonattainment nearby areas that
are “contributing to a violation,” rather than those that “contribute[]
to ambient air quality” in a violating area, 42 U.S.C.
 22
B. UINTA BASIN

Petitioner WildEarth Guardians (WildEarth) challenges
the EPA’s designation of Uinta Basin, Utah, as
“unclassifiable.” We find the EPA’s designation rational and
in accordance with the Clean Air Act, and we therefore deny
WildEarth’s petition.

1. Uinta Basin Background

The EPA requires every state to establish a network of
regulatory monitoring stations to collect ozone air-quality data.
See 40 C.F.R. pt. 58. The number of regulatory monitors
required in an area depends, in part, on the area’s population.
See id. app. D. tbl.D-2. Areas with populations below 50,000
and many areas with fewer than 350,000 inhabitants require no
regulatory monitors. Id. Many rural areas therefore lack
monitors.

Uinta Basin, Utah, had no regulatory monitoring until
April 2011. The pre-2011 absence of regulatory-air-quality
monitors in Uinta Basin meant that, when the EPA in 2013
conducted the designation process for the 2008 NAAQS, the
agency had regulatory data for Uinta Basin for only two
years—2011 and 2012. The 2008 ozone NAAQS, however,
reflect three-year averages of ozone levels. See 2008
Designations Rule, 77 Fed. Reg. at 30,089. Noting that “there
are not yet three consecutive years of certified ozone
monitoring data available [from Uinta Basin] that can be used

§ 7407(d)(1)(A)(i). Delaware Br. 12–13. As the EPA explained,
however, its use of the phrase was simply shorthand for its
contribution analysis; it did not represent a heightened standard. Cf.
ATK Launch Sys., 669 F.3d at 338–39 (rejecting the argument that
the EPA applied a dissimilar standard when it variously used the
terms “significant contribution” and “contribution”).
 23
to determine the area’s attainment status,” id., the EPA
designated the area as “unclassifiable,” which the Clean Air
Act defines as an area that “cannot be classified on the basis of
available information as meeting or not meeting” the NAAQS,
42 U.S.C. § 7407(d)(1)(A)(iii).

Although no regulatory data exist for Uinta Basin prior to
2011, private companies working under consent decrees have
been required to operate ozone air-quality monitors in Uinta
Basin since 2009. See Letter from Robin Cooley, Counsel,
WildEarth Guardians to Lisa P. Jackson, Adm’r, EPA 3 (July
19, 2012). Under the terms of those consent decrees, the
private monitors must comply with many of the same
requirements as regulatory monitors. See Consent Decree
¶¶ 80–81, United States v. Kerr-McGee Corp.,
No. 1:07-cv-01034 (D. Colo. May 17, 2007). From 2009 to
2011, the private monitors provided raw data showing ozone
levels significantly exceeding the 2008 ozone NAAQS. The
EPA found the 2009 to 2011 private data insufficient to support
a nonattainment designation.

2. The Private Monitoring Data Challenge

WildEarth argues that, in light of the private data, the EPA
contravened the Act’s requirements when it designated Uinta
Basin as unclassifiable rather than nonattainment. We
disagree.

The Act calls for the EPA to make designations “on the
basis of available information.” 42 U.S.C. § 7407(d)(1)(iii).
We have repeatedly found similar language to be ambiguous
when assessing whether to defer to an agency’s construction.
See Catawba Cnty., 571 F.3d at 35, 38 (finding the phrase
“based on air quality monitoring data” to be ambiguous);
Sierra Club v. EPA, 356 F.3d 296, 305–06 (D.C. Cir. 2004)
(finding the phrase “based on photochemical grid modeling” to
 24
be ambiguous). The EPA therefore may interpret the
statutory language as it sees fit, as long as its interpretation is
reasonable. Chevron, 467 U.S. at 845. And even assuming
the Act obligates the EPA to consider certain types of data,
there would be no obligation for the agency to base its
designations on data it reasonably considers to be unsound, at
least if it “adequately explain[s] its reasons for
rejecting . . . data” on which it declines to rely. City of
Waukesha, 320 F.3d at 248. We evaluate the EPA’s reasons
cognizant of the “extreme degree of deference” we owe an
agency “when it is evaluating scientific data within its
technical expertise.” Catawba Cnty., 571 F.3d at 41.

The EPA reasonably explained that the private monitoring
data afforded an insufficient basis for a nonattainment
designation because the agency was unable to perform
post-collection quality assurance checks on the data. In
particular, the EPA lacked quality assurance data needed to
verify and audit the private data. As the agency explained:

Quality assurance data consist, primarily, of
biweekly single point quality control (QC)
checks, used to assess the precision and bias a
given instrument is displaying in its day-to-day
measurements, and annual independent
performance evaluations (audits) of
equipment, which rely on independent staff
and measuring systems to confirm that the
monitors are operating as expected and
required.

Letter from Lisa P. Jackson, Adm’r, EPA to Robin Cooley,
Counsel, WildEarth Guardians 5 (Dec. 14, 2012) (denying
reconsideration of Uinta Basin designation). The agency
determined that, without audits or quality control checks, it
 25
could not adequately verify the quality of the private data.
That explanation comports with common sense and falls within
the substantial deference accorded the EPA in evaluating the
soundness of data available to it.

WildEarth presses several counterarguments, none of
which we find persuasive. First, WildEarth observes that the
consent decrees required the private monitors to operate in
“substantial compliance” with 40 C.F.R. Part 58, the quality
assurance requirements under which regulatory monitors
operate. But “substantial compliance” is not “full
compliance,” and the EPA could reasonably draw a distinction
between the two. Moreover, data from regulatory
monitors—which must be collected in compliance with 40
C.F.R. Part 58—undergo post-collection auditing and
verification processes. See, e.g., 40 C.F.R. pt. 58, app. A, § 3.
Those post-collection processes could not be conducted for the
private monitor data. Accepting WildEarth’s argument would
require us to conclude that the EPA must apply less stringent
post-collection validation requirements to data collected from
private monitors in “substantial compliance” with the agency’s
data-collection regulations than the agency applies to data
collected from regulatory monitors in actual compliance with
those regulations. We see no reason to embrace that
counterintuitive result.

Second, WildEarth points out that the EPA has
encouraged other federal entities to take notice of the private
monitoring data. The EPA acknowledges that it argued, in a
judicial proceeding supporting entry of the same consent
decrees mandating the private monitoring, that the private
monitors would provide data that would be “reliable and of
good quality” and “useful in assisting regulators.” Resp’t’s
Br. 57. And indeed the data have proven helpful to the EPA in
other regulatory contexts. On the basis of the private data, for
 26
example, the EPA informed the Forest Service that Uinta Basin
ozone concentrations “exceed the NAAQS” and are a “serious
problem.” Supp. JA 387.

We agree with WildEarth that an agency may be required
to articulate why data are sufficiently reliable for one purpose
but not for another. See Cnty. of L.A. v. Shalala, 192 F.3d
1005, 1022 (D.C. Cir. 1999). But the EPA has done so here.
That the data may be sufficiently reliable to warrant identifying
ozone as a serious issue for a Forest Service analysis under one
statutory provision does not necessarily mean that the data are
reliable enough to compel a nonattainment designation under a
different statutory regime. To hold otherwise would require
the EPA wholly to blind itself to potentially useful private data
for any purpose if it were to consider that data insufficiently
reliable for one purpose. There is no basis for constraining the
agency in that way.

That the EPA partially relied on the private data in the
course of this very designation process does not undercut that
conclusion. While “unclassifiable” represents a single
statutory designation, see 42 U.S.C. § 7407(d)(1)(A)(i)–(iii),
the EPA further divided that classification into two
sub-categories: “unclassifiable/attainment” and
“unclassifiable.” See 2008 Designations Rule, 77 Fed. Reg. at
30,089. “Historically for ozone,” the EPA designates as
“ ‘unclassifiable/attainment’ ” those areas for which “air
quality information is not available because the areas are not
monitored.” Id. at 30,090. But in Uinta Basin, the EPA
instead designated the area “unclassifiable” after determining
that the private monitoring “detected levels of ozone that
exceed the NAAQS.” Id. at 30,089.

There is no arbitrariness in the EPA’s choice
partially—but not fully—to rely on the private data. At the
 27
outset, we note that the parties point us to no material
differences between an “unclassifiable/attainment” and an
“unclassifiable” designation, and we are aware of none. See
40 C.F.R. § 51.1100(g) (“Attainment area means, unless
otherwise indicated, an area designated as either attainment,
unclassifiable, or attainment/unclassifiable.”); cf. 42 U.S.C.
§ 7471 (instructing the EPA to give the same treatment to
“unclassifiable” and “attainment” areas for SIP purposes).
But given the EPA’s decision to create two different
unclassifiable designations, we will assume arguendo that
materially different regulatory burdens attend each
designation. Even then, however, we agree with the EPA that
it was reasonable to conclude that it would be inappropriate to
label the Uinta Basin area “unclassifiable/attainment”: the
private data, even if unverified, at least implied that a NAAQS
violation was possible, even if not conclusively proven to the
agency’s satisfaction. WildEarth, moreover, points to no
other area for which private—but not regulatory—monitoring
suggested a NAAQS violation. It thus appears that Uinta
Basin differed from all other areas meriting an
“unclassifiable/attainment” designation. We conclude that
the EPA’s conclusion partially—but not fully—to credit the
private data was reasonable and non-arbitrary, particularly in
light of the “extreme deference” we owe the agency. See
Catawba Cnty., 571 F.3d at 41.

In sum, the EPA reasonably declined to rely on data that it
considered of insufficient quality for designations purposes.
With that conclusion, and having reviewed the remainder of
WildEarth’s challenges and determined that they lack merit,
we deny the group’s petition for review. See Catawba Cnty.,
571 F.3d at 52.
 28
C. SIERRA CLUB

Petitioner Sierra Club challenges the EPA’s refusal to use
uncertified 2011 air-quality data during the designation
process, a decision that resulted in 15 counties avoiding
nonattainment designations. Finding the EPA’s actions
rational and in accordance with the Clean Air Act, we deny
Sierra Club’s petition.

1. Sierra Club Background

In furtherance of the Clean Air Act’s “ ‘core principle’ of
cooperative federalism,” EPA v. EME Homer City Generation,
L.P., 134 S. Ct. 1584, 1602 n.14 (2014), states take the lead in
the collection of air-quality data. In doing so, states operate
regulatory monitors under an array of “[e]xhaustive technical
specifications” promulgated by the EPA. Catawba Cnty., 571
F.3d at 30; see 40 C.F.R. pt. 58. States “edit[]” and
“validate[]” the collected data pursuant to the EPA-mandated
procedures and report it to the EPA according to a prescribed
schedule. See 40 C.F.R. § 58.16(b)–(c). Data collected in
each quarter must be “edited, validated and entered” into the
EPA’s system within ninety days of the end of the quarter. Id.
“For example, the data for the reporting period January 1–
March 31 are due on or before June 30 of that year.” Id.
§ 58.16(b). Post-auditing, the data are still considered
“uncertified” when submitted to the EPA.

While uncertified data from the first quarter (i.e., January
1 to March 31) become available to the EPA as of June 30,
those data remain subject to continuing audits and edits by
states. The data collection process reaches completion only
when a state provides final certification that the necessary
“ambient concentration and quality assurance data are
completely submitted . . . and . . . are accurate.” Id.
§ 58.15(a). The EPA requires certification by May 1 of the
 29
following calendar year for all data collected in the previous
year. Id. § 58.15(a)(2). States therefore had to certify their
2011 data by May 1, 2012.

As explained, because the 2008 ozone NAAQS represent
a three-year average, the EPA needs air-quality data from three
sequential calendar years to classify an area as attainment or
nonattainment (as opposed to unclassifiable). See 2008
Designations Rule, 77 Fed. Reg. at 30,089. In the designation
process for the 2008 NAAQS, the EPA gave each state a
choice between two options: (i) early-certify 2011 data by
February 29, 2012, in which event the EPA would consider
2009 to 2011 data for the designation process for that state
(Option One); or (ii) decline to early-certify (and stick to the
normal May 1 certification deadline), in which event the EPA
would use 2008 to 2010 data for designations in that state
(Option Two). See id. at 30,091.

At least eight states selected Option Two. Sierra Club
identifies over one dozen counties within those eight states for
which the choice between Option One and Option Two (i.e.,
the choice between designations based on 2008 to 2010 data
versus 2009 to 2011 data) allegedly meant that those counties
avoided nonattainment designations. See Letter from Robert
Ukeiley, Counsel, Sierra Club to EPA, Re: Designations for the
2008 Ozone NAAQS Docket ID
No. EPA-HQ-OAR-2008-0476 at 3 tbl.1 (Feb. 3, 2012).
Sierra Club contends that the EPA was compelled to use 2009
to 2011 data for those areas. We disagree and conclude that
the EPA’s actions were non-arbitrary.

2. Uncertified Data Challenge

Sierra Club first notes that, at the time of the designation
process, the EPA possessed uncertified 2011 data for all areas.
Because the agency’s regulations require the submission of
 30
uncertified data within ninety days of the end of the quarterly
reporting period, see 40 C.F.R. § 58.16(b), the EPA had all
2011 uncertified data in its possession by the end of March. It
should have used that data, Sierra Club argues,
notwithstanding the lack of certification. We are
unpersuaded.

While the uncertified data must undergo preliminary
auditing and quality checks before submission to the EPA, see
id. § 58.16(c), those preliminary quality control measures are
just that—preliminary. As the EPA explains, the data remain
subject to continuing checks and revisions by the states until
final certification. Resp’t’s Br. 66. Accordingly, the EPA
reasonably “does not presume that data [validation and
auditing] processes are complete and accurate until” the final
data certification. Id. at 46. Mindful of the significant
deference we owe the EPA in matters concerning data quality
or sufficiency, see Catawba Cnty., 571 F.3d at 41, we see no
basis for second-guessing the EPA’s considered judgment on
the issue.

Sierra Club next argues that, even if the agency acted
reasonably in refusing to rely on uncertified data, it acted
arbitrarily in declining to delay the designation process until all
states had certified their 2011 data by the standard May 1
deadline. After all, Sierra Club notes, the consent decree
under which the EPA conducted the designation process
allowed the agency until May 31, 2012, to promulgate the final
designations. 2008 Designations Rule, 77 Fed. Reg. at
30,091.

Sierra Club, however, identifies no authority obligating
the EPA to wait until the last possible minute to promulgate its
designations. And in this case, doing so would have made
little sense. The EPA entered into the consent decree
 31
precisely to settle allegations that it had already missed the
Act’s statutory deadlines for promulgating the 2008 ozone
NAAQS designations. See id. Accepting Sierra Club’s
position would effectively call for the EPA to infringe the
Act’s deadlines still further. In any event, as the EPA
explained in denying Sierra Club’s petition for reconsideration
of the designations after the May 1, 2012, certification deadline
passed and 2009 to 2011 data were fully certified and available
to the EPA, “[n]ew technical data become available on a
regular basis.” Letter from Lisa P. Jackson, Adm’r, EPA to
Robert Ukeiley, Counsel, Sierra Club enclosure p.2 (Dec. 14,
2012). The EPA reasonably concluded that delay “to consider
such new information would result in a never-ending process
in which designations are never finalized.” Id. Indeed,
Sierra Club itself has already filed a petition for
reconsideration based on 2010 to 2012 data. See Sierra Club
Reply Br. 8. The EPA could reasonably conclude that the
process must end at some point. We conclude that the agency
did not act arbitrarily in ending it here. Cf. Catawba Cnty.,
571 F.3d at 51 (“New York’s underlying complaint is that the
iterations should have continued, perhaps ad infinitum. But
such a process is inconsistent with the CAA: Congress
imposed deadlines on EPA and thus clearly envisioned an end
to the designations process.”).

With that conclusion, and having reviewed the remainder
of Sierra Club’s challenges and determined that they lack
merit, we deny the group’s petition for review. See Catawba
Cnty., 571 F.3d at 52.

D. MISSISSIPPI

The State of Mississippi challenges the EPA’s use of 2008
to 2010 data to classify the counties within the Memphis,
Tennessee area, an analysis that resulted in a nonattainment
 32
designation for part of DeSoto County, Mississippi. Because
we conclude that the EPA’s actions were rational and in
accordance with the Clean Air Act, we deny Mississippi’s
petition for review.

1. Mississippi Background

In Mississippi and elsewhere, the EPA conducted the
designations for metropolitan areas through a two-step process.
First, the EPA examined air-quality data from all regulatory
monitors in a metropolitan area. If no monitors in the area
showed a NAAQS violation, no county in the area would be
designated nonattainment. In that event, there would be no
second step. But if a single monitor from the area showed a
NAAQS violation, the county housing the violating monitor
would be designated nonattainment. See 2008 Guidance at 3–
4. In that case, the EPA would proceed to the second step for
that metropolitan area.

The second step took account of the fact that the Act
mandates nonattainment designations not only for areas
themselves exceeding the relevant NAAQS, but also for all
areas that “contribute[]” to a NAAQS violation in a “nearby
area,” even if the “contributing” area’s air quality—considered
alone—meets the NAAQS. See 42 U.S.C. § 7407(d)(1)(A)(i);
2008 Guidance at 3–4. In the second step, the EPA assessed
each county in a metropolitan area with a violating monitor on
a case-by-case basis to determine if the county contributed to
the identified violation. If, on the basis of a multi-factor test,
the EPA determined that a county “contributed” to the NAAQS
exceedance at the violating monitor in another county, the EPA
also designated the contributing county as nonattainment. We
have repeatedly upheld multi-factor contribution analyses as
consistent with the Act’s designation process under section
107—a conclusion that Mississippi does not challenge here.
 33
See, e.g., ATK Launch Sys., 669 F.3d 330; Catawba Cnty., 571
F.3d 20. See generally supra § II.

In 2011 and 2012, the EPA conducted that two-step
designation process for the Memphis CBSA. The Memphis
CBSA consists of several counties in Tennessee (Shelby,
Tipton, and Fayette), Mississippi (DeSoto, Marshall, Tate, and
Tunica), and Arkansas (Crittenden). See Office of Mgmt. &
Budget, OMB Bulletin No. 10-02, Update of Statistical Area
Definitions and Guidance on Their Uses 40 (Dec. 1, 2009).
At the first step, the EPA evaluated 2008 to 2010 certified
air-quality data and detected a NAAQS violation at the monitor
in Shelby County, Tennessee. Proceeding to the second step,
the EPA conducted the multi-factor analysis and determined
that part of DeSoto County, Mississippi, contributed to the
Shelby County violation.

On December 9, 2011, the EPA notified Mississippi that it
planned to designate part of DeSoto County as nonattainment
when it promulgated the final designations in 2012. The EPA
invited Mississippi (and all other states) to provide to the
agency by February 29, 2012, any additional information for
consideration in the final designation process—including any
early-certified 2011 data. See Memphis, TN-MS-AR Area
Designations for the 2008 Ozone NAAQS 3–4 [hereinafter
Memphis Area Designations]. Mississippi responded to the
EPA’s multi-factor analysis with its own multi-factor analysis,
disputing the EPA’s conclusion that DeSoto County
contributed to any violation in Shelby County. Additionally,
Mississippi and Tennessee—two of the three states in the
Memphis CBSA—early-certified their 2011 data before the
February 29, 2012, deadline. Arkansas—the third state in the
Memphis CBSA—declined to early-certify any 2011 data.
 34
On May 21, 2012, the EPA published its final designations
for the Memphis CBSA. At the first step of the two-step
designation process, the agency used 2008 to 2010 data and
again identified a violation at the Shelby County monitor.
The EPA then moved to the second step and, after considering
Mississippi’s multi-factor analysis and updating its own
analysis accordingly, reiterated its original conclusion that part
of DeSoto County contributed to the Shelby County violation.
The agency therefore designated part of DeSoto County as
nonattainment. See Memphis Area Designations at 16.
Mississippi claims that designation was arbitrary and
capricious. We disagree.

2. Challenge to the First Step of the
Designation Process

First, Mississippi argues that the EPA acted arbitrarily in
using 2008 to 2010 data for the first step of the two-step
designation process (i.e., identifying violating monitors within
a CBSA) even though the EPA possessed early-certified 2011
data from Tennessee. The 2009 to 2011 data showed no
NAAQS violation at the Shelby County monitor.
Accordingly, Mississippi argues, no violation should have
been identified at the first step of the two-step designation
process. But the EPA declined to evaluate Shelby County
using the early-certified 2009 to 2011 data, instead using the
2008 to 2010 data. True, the EPA must adequately explain why
it declined to rely on the early-certified 2011 data. See City of
Waukesha, 320 F.3d at 248. But the agency did so.

At the time of the final designations, the EPA had in its
possession early-certified data from Mississippi and
Tennessee, but not from Arkansas. In the first step of its
two-step designation process, the EPA evaluates all air-quality
monitors in a metropolitan area. Without 2011 Arkansas data,
 35
the EPA did not have a full set of 2011 data for the Memphis
CBSA. The EPA only had data from different time
horizons—2008 to 2010 data for the Arkansas portion of the
Memphis CBSA, and 2009 to 2011 data for the Tennessee and
Mississippi portions of that same CBSA. The agency
declined to rely on this mismatched dataset. Instead, the EPA
opted to rely on the most recent matched dataset in its
possession: the complete set of 2008 to 2010 data. We see
no reason—and Mississippi provides none—to declare
irrational the EPA’s conclusion that comparing data from the
same time period would be more appropriate than analyzing
data from different time periods in the same evaluation
process. Cognizant of the substantial deference we owe the
EPA in that highly technical evaluation, see Catawba Cnty.,
571 F.3d at 41, we find the EPA was entitled to rely on a
matched dataset instead of a mismatched one.

Even assuming the EPA’s choice to rely only on matched
datasets for the Memphis CBSA was reasonable (as we
conclude it to be), Mississippi argues that the EPA’s approach
nonetheless was arbitrary because the agency required a
matched dataset for Memphis-area designations but allegedly
relied on a mismatched dataset for Chicago-area designations.
“[I]nconsistent treatment,” we have found, is a “hallmark of
arbitrary agency action.” Id. at 51. There was no
inconsistent treatment here, however. In both Chicago and
Memphis, the EPA relied only on matched datasets in the
designation process.

With regard to the Chicago metropolitan area, Illinois
early-certified its 2011 data. Wisconsin and
Indiana—portions of which also lie in the Chicago
metropolitan area—did not early-certify. Illinois’s
early-certified data showed a violating monitor in the Chicago
area. At the first step of the Chicago-area designation
 36
process, the EPA relied on Illinois’s early-certified data, noted
the violation, and thus proceeded to the second step’s
multi-factor contribution analysis for all Chicago-area
counties.

Mississippi argues that, because the EPA only possessed
early-certified data from Illinois, it used a mismatched dataset
for Chicago’s designations. Consequently, Mississippi claims
that the EPA took different approaches to dataset selection
between Memphis and Chicago. Mississippi’s argument rests
on a flawed understanding of the EPA’s designation process.

At the first step of the process, a single violating monitor
suffices to conclude the analysis and move to the second step.
Though only Illinois had early-certified its data, that data
showed a violating monitor. That was enough to terminate the
first step of the process and move to the second step. It thus
became irrelevant whether Wisconsin or Indiana data showed
any violations: the EPA would proceed to the second step of
the analysis regardless, based on the Illinois violation alone.
The EPA therefore had a sufficient matched dataset of 2009 to
2011 data (albeit data from only one state, Illinois) to proceed
to the second step of the designation process using 2009 to
2011 data alone. By contrast, the EPA had no matched dataset
of 2009 to 2011 data in the Memphis area sufficient to
complete the first step of the two-step process using that data
alone. While data showing a single violating monitor are
enough to end the first step and proceed to the second step, data
showing all monitors in compliance would be needed to avoid
proceeding to the second step’s multi-factor analysis—i.e., to
terminate the two-step process at the first step.

As a result, when Arkansas declined to early-certify its
2011 data, the EPA could not determine if the entire Memphis
CBSA showed NAAQS compliance at all monitors for the
 37
2009 to 2011 period; the agency lacked a sufficient 2009 to
2011 matched dataset with which to do so. The EPA then
relied on the most recent matched dataset sufficient to
complete the first-step analysis (the 2008 to 2010 data), just as
the EPA selected the most recent matched dataset sufficient for
the first-step analysis of the Chicago area. The EPA therefore
acted in a consistent manner in both areas, each time using the
most recent matched datasets sufficient to complete the first
step of the two-step designation process.

3. Challenge to the Second Step of the
Designation Process

Mississippi also challenges the EPA’s application of the
second step of the designation process. The EPA acted
arbitrarily, the state argues, in applying the multi-factor test
and concluding that DeSoto County contributed to the Shelby
County violation. We find no reason to disturb the EPA’s
analysis.

First, Mississippi challenges the EPA’s differing
articulations of the multi-factor test. As pronounced in the
2008 Guidance, the EPA originally conceived of that test as
consisting of nine factors. In making the final designations,
the EPA applied a five-factor test. See supra § I.B–C, The
state argues that the EPA’s “consolidat[ion]” of the test from
nine to five factors was arbitrary and capricious. State &
County Br. 15. We disagree.

At the outset, we do not necessarily agree that the EPA
was required to adhere to the 2008 Guidance. The 2008
Guidance did not purport to be a legislative rule, and it
explicitly provided that it was “not binding on states, tribes, the
public or the EPA.” 2008 Guidance at 4; cf. Catawba Cnty.,
571 F.3d at 33–34 (materially similar guidance for PM2.5
 38
NAAQS designations did not “create or modify legally binding
rights”).

But even if we assume that the 2008 Guidance was
binding, the EPA did not deviate from it in the final
designations. The “consolidation” of the factors was just
that—a consolidation. It effected no deletion. During the
final designation process, the agency simply grouped several
of the 2008 Guidance factors into a single factor, the
consideration of which necessarily entailed consideration of
the multiple 2008 Guidance factors now residing within it.
We find no examples of a final designation that failed to
consider a factor identified in the 2008 Guidance. With “no
bright line for any of the factors,” and with each factor
“weighted considering the unique circumstances of each
nonattainment area,” Response to Comments at 61, the
consolidation worked no substantive change and thus affords
no basis for setting aside the EPA’s analysis.

Second, Mississippi challenges the EPA’s specific
application of the multi-factor test to DeSoto County. We
accord the EPA “extreme deference” in applying that test, and
will overturn the EPA’s designations only if the agency applied
the test “so erroneously in a particular case that it could not
have reasonably concluded that a county was contributing to
nearby violations.” Catawba Cnty., 571 F.3d at 40–41. This
is not such a case. The agency provided data showing that
DeSoto County’s NOx and SO2 (ozone precursors) emissions
were the second-highest in the Memphis CBSA. Memphis
Area Designations at 8. The county also had the second
highest number of workers commuting to counties with
violating monitors, the second highest number of vehicle miles
traveled in the CBSA, and the highest percentage population
growth over the last decade. Those factors led the EPA to
conclude that DeSoto County was integrated with Shelby
 39
County in a way that indicated ozone contribution. Id. at 9–
10. Additionally, meteorological analysis at the Shelby
County monitor showed weather patterns characterized in part
by winds blowing in from DeSoto County. Id. at 12. On
those bases, the EPA reasonably concluded that DeSoto
County contributed to the Shelby County violation.

Mississippi principally argues that significant “commerce
activity” occurring outside of DeSoto County (including
interstate highway traffic, rail and barge transportation, diesel
fuel sales, and air traffic) means that other counties contribute
to the Shelby County violation more than DeSoto County
does—and that, because some of those counties avoided
nonattainment designations, DeSoto County should, too.
Miss. Dep’t of Envtl. Quality, Air Div., 2008 Ozone Standard
Designation Recommendation for DeSoto County, Mississippi
8–12 (Feb. 2012). But the EPA considered that argument and
determined in a well-reasoned analysis that the data from
Mississippi was only one consideration in the designation
process. See Response to Comments at 97; see also Memphis
Area Designations 1–31. The EPA concluded that DeSoto
County did contribute to Shelby County’s violation in light of
the many other factors the agency considered. Memphis Area
Designations at 16.

Looking at the same data, Mississippi would simply reach
a different conclusion. We, however, do not sit to
second-guess the EPA’s conclusions in an area identified by
the Congress as within the agency’s technical expertise. We
only ask if the EPA “considered all relevant factors and
articulated a rational connection between the facts found and
the choice made.” ATK Launch Sys., 669 F.3d at 336 (internal
quotation marks omitted). We conclude that it did.
 40
With that conclusion, and having considered Mississippi’s
other challenges and determined that they lack merit, we deny
the state’s petition for review. See Catawba Cnty., 571 F.3d at
52.

E. LAKE & PORTER COUNTIES, INDIANA

Petitioner Indiana challenges the designation of two of its
counties as nonattainment. According to Illinois’s certified
2009 to 2011 data, the monitoring site at Zion, Illinois
exceeded the NAAQS by 1 part per billion (ppb). See
Chicago-Naperville, Illinois-Indiana-Wisconsin Area
Designations for the 2008 Ozone NAAQS at 7–8 [hereinafter
Chicago Area Designations]. Zion is about sixty miles from
the Indiana border and, like the Indiana counties at issue here,
belongs to the Chicago-Naperville-Michigan City CSA.
Following the 2008 Guidance, the EPA presumed that all
counties in this CSA should be designated as nonattainment
areas due to the Zion violation, and then conducted its
five-factor analysis. The agency preliminarily concluded that
three Indiana counties—Lake, Porter, and Jasper—should be
included in the nonattainment area.

In response to the EPA’s 120-day letter, Indiana pointed to
multiple asserted flaws in the EPA’s analysis. Most relevant
here, it said that the agency had failed to account for the impact
of a recent statutory change to Illinois’s vehicle emissions
testing program. It also maintained that the agency’s
meteorological analysis suffered from multiple weaknesses
and inconsistencies.

The EPA ultimately reversed its designation of Jasper
County, but finalized the nonattainment designations of Lake
and Porter Counties. Chicago Area Designations at 21.
 41
Indiana now challenges those nonattainment designations as
arbitrary and capricious.

1. Challenge Regarding Illinois’s Vehicle
Inspection Change

First, Indiana challenges the EPA’s position regarding
Illinois’s statutory change. After a prior nonattainment
designation, Illinois had established a vehicle inspection and
maintenance program that covered all model years beginning
in 1968. 11 In 2006, however, Illinois exempted vehicles with
model years between 1968 and 1995 from the testing
requirements. See 625 Ill. Comp. Stat. 5/13C-15(a)(6)(L)
(2012). Indiana maintains that it was the increase in vehicle
emissions accompanying this exemption that directly caused
the violation at the Zion monitor. Moreover, it contends that
this legislative change amounted to an intentional violation of
Illinois’s SIP.

As the EPA points out, we made clear in Catawba County
that a “contributing” county need not be the but-for cause of a
violation in order to warrant a nonattainment designation.
Resp’t’s Br. 94; see Catawba Cnty., 571 F.3d at 39 (“[E]ven
were we to think that ‘contribute’ unambiguously means
‘significantly contribute,’ we still disagree that ‘significantly
contribute’ unambiguously means ‘strictly cause.’ ”). And
here, regardless of Illinois’s statutory change, the EPA’s
five-factor analysis demonstrated that both Lake and Porter

11
See Approval and Promulgation of Air Quality
Implementation Plans; Illinois; Motor Vehicle Inspection and
Maintenance, 64 Fed. Reg. 8,517, 8,519 (Feb. 22, 1999); Approval
and Promulgation of Air Quality Plans; Illinois; Post-1996 Rate of
Progress Plan for the Chicago Ozone Nonattainment Area, 65 Fed.
Reg. 78,961, 78,967–68 (Dec. 18, 2000).
 42
Counties contributed to the Zion monitor. Chicago Area
Designations at 6–21. 12

The alleged illegality of Illinois’s statutory change does
not affect our conclusion. The Clean Air Act offers other
avenues for addressing a State’s failure to comply with its SIP.
In particular, the EPA Administrator can call for a SIP revision
after “find[ing] that the applicable implementation plan for any
area is substantially inadequate” to comply with the NAAQS.
42 U.S.C. § 7410(k)(5). The EPA declined to do so here and,
instead, recently approved the Illinois change. 13 Indiana has
since petitioned the Seventh Circuit to review the EPA’s
approval. See EPA 28(j) Letter (Oct. 22, 2014). That is the
appropriate forum for challenging the Illinois change, which in
no way diminished the contribution of the Indiana counties.

12
Indiana protests that there likely would have been no violation
at all at the Zion monitor if it were not for the emissions resulting
from the statutory change. That argument is merely a rephrasing of
the but-for causation rule that we rejected in Catawba County. In
any event, the argument is not supported by the Indiana modeling
analyses upon which it is based. See Letter from Ind. Dep’t of
Envtl. Mgmt. to EPA, Enclosure 1 at 27–30 (Apr. 13, 2012). The
first analysis concluded only that the change in Illinois’s program
contributed 0.2 ppb to the Zion violation—not enough to account for
the 2009 to 2011 exceedance of 1 ppb. The second analysis rested
on a factual premise that the State never adequately explained: that
the statutory change caused the emission reduction benefits of
Illinois’s vehicle emissions testing program to decrease by 35 per
cent.
13
See Approval and Promulgation of Air Quality
Implementation Plans; Illinois; Amendments to Vehicle Inspection
and Maintenance Program for Illinois, 79 Fed. Reg. 47,377 (Aug. 13,
2014).
 43
2. Challenge to the EPA’s Response to Comments

Next, Indiana argues that the EPA failed to adequately
respond to its comments about the impact of Milwaukee,
Wisconsin’s emissions on the violation at the Zion monitor.
According to the source apportionment modeling submitted by
Indiana, the Milwaukee area contributed over 5 ppb to the Zion
violation, while Lake, Porter, and Jasper Counties contributed
4 ppb, 2 ppb, and 0.5 ppb, respectively. See Letter from Ind.
Dep’t of Envtl. Mgmt. to EPA, Enclosure 1 at 13–14 (Apr. 13,
2012). This, Indiana maintains, produced the “inconsistent
and unfounded” result of nonattainment designations for the
Indiana counties but an attainment designation for the
Milwaukee area. Id. at 14.

As an initial matter, we note that, because the Milwaukee
area is not a single county but rather is a metropolitan area
made up of five counties, Indiana’s argument is premised on an
apples-to-oranges comparison. More important, we have no
basis for finding the EPA’s designations inconsistent given that
Indiana’s modeling—which was limited to meteorological
linkages and therefore fell short of a full analysis—did not
establish that Milwaukee “contributed to” the Zion violation
under the agency’s five-factor analysis. By contrast, after
conducting its full five-factor analysis, the EPA found that
Lake and Porter Counties did contribute. Accordingly, the
EPA’s determination regarding the Milwaukee metropolitan
area was neither unreasonable nor inconsistent with its
determination regarding the Indiana counties.

We also find that the EPA did adequately respond to
Indiana’s comments about its modeling results, although
without mentioning Milwaukee specifically. Indeed, the
modeling was one of the factors that led the EPA to reconsider
its designation of Jasper County. See Chicago Area
 44
Designations at 21 (describing Jasper County’s 0.5 ppb
contribution as “not significant”). But the EPA simply
disagreed with Indiana’s premise that 2 ppb and 4 ppb were
insufficient contributions when considered as part of the
five-factor test, for reasons that were reasonable and well
explained. See id. at 18 (“In keeping with EPA’s ozone
contribution levels used to select states that should be covered
in regional emission control programs, 2 ppb to 4 ppb ozone
concentration contributions are considered to be significant
ozone contributions.”).

3. The Remaining Challenges

Finally, we briefly consider Indiana’s remaining
arguments. First, the record does not support Indiana’s claim
that the EPA improperly relied on late-submitted data from
Wisconsin’s Chiwaukee Prairie monitor, rather than relying
solely on the Zion monitor data, in making the contribution
determinations regarding the Indiana counties. See Chicago
Area Designations at 8 (noting that the EPA considered the
Wisconsin data in determining whether Kenosha County,
Wisconsin (and not the Indiana counties) should be included in
the Chicago nonattainment area); id. at 21–22 (describing
bases for Lake, Porter, and Kenosha County designations).
Second, the EPA did not fail to adequately explain why it used
some 2006 to 2008 weather data in conducting the contribution
analysis. The agency explained that historical data provided a
“general conceptual model to explain the development and
transport of high ozone levels in this area.” Addendum to
Response to Comments at 7 (May 31, 2012); see also EPA
Response to Indiana Pet. for Reconsideration 3. That
explanation is deserving of the deference that we give to the
EPA’s “evaluati[on] [of] scientific data within its technical
expertise,” Catawba Cnty., 571 F.3d at 41 (quoting City of
Waukesha, 320 F.3d at 247).
 45
In sum, we reject Indiana’s contention that the EPA’s
designations of Lake and Porter Counties are arbitrary or
capricious.

F. WISE COUNTY, TEXAS

Petitioners State of Texas; Wise County, Texas; Texas
Commission on Environmental Quality; Devon Energy
Corporation; Targa Resources Corporation; the Texas Pipeline
Association; and the Gas Processors Association (collectively,
Texas Petitioners) challenge the EPA’s designation of Wise
County as nonattainment. They make several claims,
including that the EPA subjected Wise County to arbitrarily
disparate treatment, violated the U.S. Constitution and acted
beyond its authority under the Clean Air Act. For the reasons
discussed below, however, we do not disturb Wise County’s
nonattainment designation.

1. Wise County Background

Wise County is one of 22 counties in and around the
Dallas–Fort Worth metropolitan area, which reports some of
the most severe NAAQS violations in the country. Although
Wise County has no monitor of its own, it borders several
counties with a total of seven violating monitors, the closest of
which reports ambient ozone levels that exceed the 2008
NAAQS by 0.010 ppm. Moreover, because Wise County
falls within the CSA of Dallas–Fort Worth, it is presumptively
included within the nonattainment area.

Despite Wise County’s presumptive inclusion in the
Dallas–Fort Worth nonattainment area, the EPA designated it
as attainment when it updated the ozone NAAQS in 1997.
For this reason, Texas did not include Wise County among the
nine Dallas–Fort Worth counties it recommended for
nonattainment status when it submitted its initial designations
 46
to the EPA in March 2009. 14 On December 9, 2011, the EPA
informed Texas that it planned to include Wise County in the
Dallas–Fort Worth nonattainment area due to its
“comparatively high emissions” and “close proximity . . . to
violating monitors.” See Texas Area Designations for the
2008 Ozone NAAQS at 13 [hereinafter Preliminary Dallas–
Fort Worth Area Designations].

The EPA redesignated Wise County based on the five-part
“weight of the evidence analysis” articulated in the 2008
Guidance. 15 See id. at 1–2. The second and third
factors—emissions data and meteorology—factored
prominently in the EPA’s decision. See id. at 13. As for
emissions, the EPA concluded that oil-and-gas collection and
production in the Barnett Shale reservoir—a gas-rich
geological formation covering a significant portion of Wise
County—resulted in Wise County’s inclusion among the eight
highest emissions-producing counties in the Dallas–Fort
Worth area. 16

As for meteorology, although historic wind patterns in the
Dallas–Fort Worth area suggest that air does not normally
move from Wise County to counties with monitors registering

14
Initially, Texas based its recommended designations on
air-quality data from 2005 to 2007. On October 31, 2011, Texas
updated its initial designations with certified air-quality data from
2008 to 2010.
15
As noted above, see supra § I.B–C, the 2008 Guidance
initially established a nine-part test but the EPA subsequently
collapsed those nine factors into five.
16
Specifically, Wise County had the fourth highest level of
VOC emissions among nineteen counties in the Dallas–Fort Worth
area and the sixth highest level of NOx emissions. Preliminary
Dallas–Fort Worth Area Designations at 6 tbl.3.
 47
NAAQS violations, the EPA concluded that Wise County was
upwind of the monitors on days when ozone levels at the
monitors peaked. See Preliminary Dallas–Fort Worth Area
Designations at 10. In reaching this conclusion, the EPA used
the National Oceanic and Atmospheric Administration’s
Hybrid Single Particle Lagranian Integrated Trajectory
(HYSPLIT) model instead of relying solely on historic wind
patterns in the Dallas–Fort Worth area. See id. HYSPLIT
charts the path, or “back trajectory,” that air takes before it
collects in a certain area. See id. According to the EPA,
HYSPLIT modeling “is specifically designed to give an
estimate of the probable path a parcel of air travels in reaching
a given location at a given time” and is particularly
illuminating for an area like Wise County, which has “light and
variable” wind patterns. Response to Comments at 59–60.

After the EPA notified Texas that it planned to include
Wise County in the Dallas–Fort Worth nonattainment area,
numerous individuals and organizations submitted comments
urging the EPA to reconsider its Wise County designation.
One commenter insisted that other Texas counties were more
responsible than Wise County for the NAAQS violations in the
Dallas–Fort Worth area. Others argued that the EPA’s use of
HYSPLIT modeling was arbitrary and capricious because,
when designating other areas of the country, the EPA relied
solely on historic wind patterns. According to these
commenters, if the EPA had done the same with Wise County,
it would not have designated Wise County as nonattainment
because, according to historical wind patterns in the
Dallas-Fort Worth area, Wise County was downwind of
violating monitors more than 95 per cent of the time.

For its part, Petitioner Texas Commission on
Environmental Quality (Texas Commission) submitted its own
data based on photochemical grid source apportionment
 48
modeling. Source-apportionment modeling helps determine
the potential future impact of an emissions source area (such as
Wise County) on downwind monitors by “keep[ing] track of
the origin of the [ozone] precursors creating the ozone.”
Industrial Br. 7. It does so by combining “the
meteorology/transport of air parcels during high ozone days
with the emissions of [a] specific area[],” (here, Wise County),
“to evaluate potential impact on ozone levels.” Dallas-Fort
Worth, Texas Final Area Designations for the 2008 Ozone
NAAQS at 16 [hereinafter Final Dallas–Fort Worth Area
Designations]. Although the EPA does not typically perform
source-apportionment modeling during the NAAQS
designation process, it “has used it in the past for large-scale
rulemakings, such as the Clean Air Interstate Rule and Cross
State Air Pollution Rule” and it considers
source-apportionment modeling data if a state submits it. See
Resp’t’s Br. 126. According to the Texas Petitioners,
source-apportionment modeling suggests that Wise County
emissions had only a negligible impact on the monitors
registering NAAQS violations in the Dallas–Fort Worth area.

On April 30, 2012, the EPA issued its omnibus Response
to Comments, many of which addressed the objections to the
Wise County designation. The EPA defended HYSPLIT
modeling as an “excellent tool[]” that it generally “prefer[s]
over more basic assessments of wind speed and direction.”
Response to Comments at 59. The EPA found HYSPLIT
modeling to be a more precise measure of wind patterns than
historic data, which data, according to the agency, is
“potentially misleading in cases where wind speeds are light
and variable, or vary substantially across the location of the
meteorological observation and the monitored high ozone
concentrations.” Id. These conditions existed in the Dallas–
 49
Fort Worth area. 17 Although the EPA acknowledged it could
not always use HYSPLIT modeling, it nonetheless declined to
ignore HYSPLIT data “where the information is available,
even if the information is not available in all areas.” Response
to Comments at 59.

Along with its omnibus responses, the EPA issued its
Final Dallas–Fort Worth Area Designations, which again
applied the five-factor test. In that document, the EPA
addressed the source-apportionment modeling submitted by
the Texas Commission. The EPA took issue with the model’s
methodology and made several amendments to it.

First, the EPA faulted the Texas Commission for not using
data from an entire ozone season in its model. To account for
this omission, the EPA examined not only the average (i.e.,
relative) impact of Wise County emissions on Dallas–Fort
Worth monitors but also the absolute (i.e., maximum) impact
of the emissions. See Final Dallas–Fort Worth Area
Designations at 17. The average/relative approach advocated
by the Texas Commission averaged the impact that Wise
County emissions might have on the monitors on all days when
the monitors were expected to exceed the ozone NAAQS. As
a practical matter, averaging the impact of Wise County
emissions meant that the Texas Commission’s model
accounted for days on which wind patterns were not expected
to move air pollutants from Wise County to the violating
monitors. According to the EPA, the Texas Commission’s
average approach had “the effect of masking the impacts that

17
See Final Dallas–Fort Worth Area Designations at 14
(emphasizing that HYSPLIT modeling is especially appropriate for
Wise County because Dallas–Fort Worth area “is generally
characterized as having ozone exceedances with lower wind speeds
and winds from many directions”).
 50
occur on days when the wind does flow from Wise County to
violating monitors,” an imprecision that was aggravated by the
model’s limited dataset. See Resp’t’s Br. 136 (emphasis
added). To account for this imprecision, EPA chose to look at
the “direct,” or “absolute,” predicted effect that Wise County
emissions would have on violating monitors rather than the
average effect they were expected to have.

Second, the EPA noted that the Texas Commission’s
source-apportionment model under-predicted peak ozone
levels in the Dallas–Fort Worth area by a range of 0.005 to
0.020 ppm. As a practical matter, the under-prediction meant
that the Texas Commission’s model underestimated the
number of days that Wise County contributed to NAAQS
violations. To compensate therefor, the EPA examined the
impact of Wise County emissions not only on days when the
monitors exceeded the ozone NAAQS threshold of 0.075 ppm,
but also on days when the monitors reported ozone levels in
excess of 0.070 ppm.

After making these adjustments, the EPA reinterpreted the
data from the Texas Commission’s source-apportionment
model and concluded that it in fact supported including Wise
County in the Dallas–Fort Worth nonattainment area. See
Final Dallas–Fort Worth Area Designations at 20.
Specifically, the EPA concluded that Wise County emissions
(1) “resulted in 6 occurrences (over 4 days) of an impact of
more than 0.75 ppb days” on Dallas–Fort Worth area monitors;
(2) “had even larger impacts of up to 5 ppb on the Eagle
Mountain Lake monitor,” a monitor one-half mile from the
Wise County border that reported particularly severe NAAQS
violations; and (3) “resulted in 9 occurrences (over 5 days)
[causing] impacts of more than 0.75 ppb [to] occur[] at”
Dallas–Fort Worth monitors. See id. For these reasons, the
 51
EPA maintained its inclusion of Wise County in the Dallas–
Fort Worth nonattainment area.

Dozens of individuals and organizations filed petitions for
reconsideration of the EPA’s Wise County nonattainment
designation, including the Texas Commission and the other
Texas Petitioners. On December 14, 2012, the EPA denied
each petition for reconsideration. Before us, the Texas
Petitioners’ challenges to the EPA’s Wise County designation
are grouped as follows: (1) The EPA’s use of HYSPLIT
Modeling and its re-evaluation of the Texas Commission’s
source-apportionment modeling were arbitrary and capricious;
(2) the EPA’s designation of Wise County as nonattainment
violated the Commerce Clause, U.S. CONST. art. I, § 8, cl. 3,
the Tenth Amendment, id. amend. X, and the Due Process
Clause, id. amend. V; and (3) the EPA violated at least one of
several statutory provisions, including provisions of the Clean
Air Act. We address each argument in turn.

2. The Arbitrary & Capricious Challenges

The Texas Petitioners’ primary arguments are that the
EPA erred when it (i) used HYSPLIT modeling rather than
prevailing wind patterns 18 and (ii) adjusted the Texas
Commission’s source-apportionment modeling. 19 To prevail
on either argument, the Texas Petitioners must demonstrate
that the EPA acted arbitrarily and capriciously and, to do that,
they must show that the EPA either failed to consider “all
relevant factors” or to articulate a “rational connection between
the facts found and the choice made.” ATK Launch Sys., 669
F.3d at 336. Mindful of the “extreme degree of deference” we
owe to the EPA “when it is evaluating scientific data within its

18
See State & County Br. 45–46; Industrial Br. 14–26.
19
See State & County Br. 39–44; Industrial Br. 26–30.
 52
technical expertise,” Catawba Cnty., 571 F.3d at 41, and for
the reasons stated below, we conclude that neither argument
has merit.

i. HYSPLIT Modeling

The Texas Petitioners challenge the EPA’s use of
HYSPLIT modeling on three fronts. First, they argue that the
EPA could not legitimately use HYSPLIT modeling at all
because HYSPLIT “cannot measure ozone formation or
transport.” State & County Br. 45. Second, they contend
that the EPA arbitrarily treated Wise County differently by
using HYSPLIT modeling to designate it as nonattainment
while using historic wind patterns to designate other allegedly
similar counties as attainment. And third, they argue that,
even among other counties that the EPA subjected to
HYSPLIT modeling, it arbitrarily treated Wise County worse
because the respective HYSPLIT models demonstrated that
wind moved through those other counties—each of which the
EPA designated as attainment—more frequently than it moved
through Wise County. We address each argument in turn.

First, we find no merit in the Texas Petitioners’
conclusory argument that the EPA erred by using HYSPLIT
modeling at all because HYSPLIT modeling “cannot measure
ozone formation or transport. See State & County Br. 45–46.
Indeed, we rejected a materially indistinguishable challenge in
ATK Launch Systems, 669 F.3d at 339, a case involving the
EPA’s 2006 fine particulate matter NAAQS designations.
See id. at 334. We did so there because the EPA had taken
“reasonable steps to ensure that the ‘HYSPLIT’ model’s
limitations were considered.” Id. at 339 (quotation mark
omitted).

Here too, the EPA took reasonable steps to account for
HYSPLIT’s limitations by evaluating the
 53
source-apportionment modeling and historical wind data that
the Texas Commission submitted during the comment period.
See Final Dallas–Fort Worth Area Designations at 14–20, 23.
Because “[o]zone and ozone precursors can be transported to
an area from sources in nearby areas or from sources located
hundreds of miles away,” see 2008 Designations Rule, 77 Fed.
Reg. at 30,088, the EPA reasonably concluded that HYSPLIT
modeling, as a more precise measurement of the path taken by
air masses containing ozone precursors, was useful in
determining whether wind moving through Wise County could
have transported emissions to the areas with the violating
monitors.

Second, we find no merit in the Texas Petitioners’
argument that the EPA’s use of HYSPLIT modeling to
designate Wise County as nonattainment amounts to arbitrarily
disparate treatment. At the outset, it bears repeating that this
Court has expressly sanctioned the EPA’s use of a holistic,
multi-factor, totality-of-the-circumstances test for making
NAAQS determinations, see ATK Launch Sys., 669 F.3d at
336; Catawba Cnty., 571 F.3d at 39, and we have twice iterated
that, when using a multi-factor test, “ ‘discrete data points’ are
not determinative” because isolating any one discrete
consideration “ ‘ignores the very nature of the . . . test, which is
designed to analyze a wide variety of data on a case-by-case
basis.’ ” ATK Launch Sys., 669 F.3d at 336 (quoting Catawba
Cnty., 571 F.3d at 39) (emphasis added; alteration omitted)).
Indeed, because the EPA’s “holistic assessment of numerous
factors . . . drives the process,” we have recognized that “no
single factor determines a particular designation.” Id. For
this reason, the EPA could have subjected Wise County to
arbitrarily disparate treatment only if it treated genuinely
“similar counties” dissimilarly. Id. (emphasis in original).
Given “significant” differences among counties, “a direct
one-to-one comparison of the data,” including the methods
 54
used to measure such data, could be “inappropriate” or even
“illogical.” Id. at 337.

As noted, the EPA conducted a HYSPLIT analysis in
areas where it “believed [HYSPLIT] could provide additional
insight into whether [the] area[] contribute[s] to
nonattainment.” Resp’t’s Br. 110 n.47. The EPA reasonably
determined that Wise County was one such area because
Dallas–Fort Worth “experiences light wind speeds and winds
from variable directions,” making HYSPLIT’s more
sophisticated evaluation of wind patterns “a more useful tool
than annualized wind patterns.” EPA Response to Pet. for
Reconsideration from Devon Energy Corp. at 12. According
to the EPA, this more refined analysis was not necessary for all
areas of the country, particularly those in which “there was not
significant debate over whether [they] should be included” in a
nonattainment area. See Resp’t’s Br. 111. The EPA’s
decision to use HYSPLIT analysis in one area but not in
another fits comfortably within the agency’s “technical
expertise,” Catawba Cnty., 571 F.3d at 41, and the EPA’s
explanation for the differing treatment was rational.

Moreover, although the Texas Petitioners direct this Court
to other attainment areas that were not evaluated using
HYSPLIT modeling—specifically, Orange County and
Cattaraugus County in New York—the “significant”
differences between Wise County and those counties “make a
direct one-to-one comparison of the data underlying the
analyses inappropriate.” ATK Launch Sys., 669 F.3d at 337.
For instance, the EPA justified its Orange County attainment
designation, in part, on its finding that “the density of [Orange
County’s] emissions and vehicle usages are not of the level of
the other counties in the CSA that are in New York’s proposed
New York–Northern New Jersey–Long Island, NY-NJ-CT
nonattainment area.” New York-Northern New Jersey-Long
 55
Island, NY-NJ-CT Nonattainment Area Designations for the
2008 Ozone NAAQS at 16 (emphasis added). In contrast, the
EPA justified its nonattainment designation of Wise County, in
part, based on the “[t]he close proximity of [Wise County’s]
comparatively high emissions to violating monitors.” Final
Dallas–Fort Worth Area Designations at 23 (emphasis added).

Similarly, the EPA designated Cattaraugus County as
attainment not only because “it is in the prevailing downwind
direction from” the nearest violating monitor but also because
“other monitors representative of Cattaraugus County, as well
as the rest of upstate New York, are attaining the ozone
standard.” See Attainment Status for Jamestown, New York
and the Remainder of Upstate New York at 6 (emphasis
added). But in the Dallas–Fort Worth area, seven violating
monitors surrounded Wise County and some of the
monitors—including one located one-half mile from Wise
County’s border—reported levels of ambient ozone higher
than anywhere else in the United States. Because “the core
reason for the disparate designations” did not, as the Texas
Petitioners would have it, reflect an “inconsistent approach to
meteorology,” Industrial Br. 19, the EPA did not arbitrarily and
capriciously treat Wise County differently by evaluating its
wind patterns using HYSPLIT modeling instead of prevailing
wind patterns.

Third, when Wise County is compared to other counties
for which the EPA used HYSPLIT modeling, it is clear that the
EPA did not arbitrarily subject Wise County to disparate
treatment. The Texas Petitioners point to four other
counties—York, Dauphin and Lawrence Counties in
Pennsylvania and Roane County, Tennessee—each of which
the EPA designated as attainment notwithstanding HYSPLIT
modeling demonstrated that air moved through them to
violating monitors more frequently than through Wise County.
 56
But again, a holistic look at why the EPA designated these
counties attainment but designated Wise County
nonattainment demonstrates that the EPA did not act arbitrarily
or capriciously.

For example, York and Dauphin Counties are both near
Lancaster County, which houses all violating monitors in the
area. Because Lancaster County “is served by a single-county
transportation-planning agency,” the EPA concluded that there
were “strong jurisdictional arguments” for designating
Lancaster as “a single county nonattainment area” and,
accordingly, designating all other counties in the
vicinity—including York and Dauphin—as attainment. See
Pennsylvania Area Designations for the 2008 Ozone NAAQS
at 29–31. In contrast, Wise County is part of the Dallas–Fort
Worth CSA (which means it is presumptively included in the
Dallas–Fort Worth nonattainment area) and is also part of the
Dallas–Fort Worth metropolitan planning organization (which
implements programs and projects to reduce emissions across
all included counties). In other words, jurisdictional and
regional planning concerns—not differing approaches to
HYSPLIT modeling data—drove the EPA’s conclusion that
York and Dauphin Counties should be designated as
attainment while Wise County should be designated as
nonattainment.

The Texas Petitioners’ comparisons of Wise County to
Roane County, Tennessee, and Lawrence County,
Pennsylvania, fare no better. Roane County is
“geographically separated from the nearest county with a
violating monitor” by approximately thirty miles and the ozone
levels in the county between Roane and the next county with a
violating monitor are in attainment. Resp’t’s Br. 122. The
monitor in Lawrence County reports ozone levels that, at 0.066
ppm, are well below the EPA’s NAAQS 0.075 ppm threshold.
 57
Moreover, the county with a violating monitor nearest to
Lawrence County—Allegheny County—is not adjacent to
Lawrence County. In contrast to both Roane County and
Lawrence County, Wise County is adjacent to multiple
counties reporting severe NAAQS violations, the closest of
which is located a mere half mile from the Wise County line.

The dispositive principle that the Texas Petitioners try to,
but ultimately cannot, avoid is that under the EPA’s holistic
analysis, “discrete data points” like the data from HYSPLIT
modeling “are not determinative, because elevating them
ignore[s] the very nature of the [holistic] test, which is
designed to analyze a wide variety of data on a case-by-case
basis.” ATK Launch Sys., 669 F.3d at 336 (quotation mark
omitted). Based on the foregoing analysis, we cannot say that,
had the EPA declined to evaluate Wise County’s wind patterns
using HYSPLIT modeling, Wise County “would not have been
designated nonattainment.” Industrial Br. 19. Because none
of the areas discussed by the Texas Petitioners is truly
“similarly situated” to Wise County, and because the EPA fully
and rationally supported its use of HYSPLIT modeling for
Wise County, it did not act arbitrarily or capriciously.

ii. Source-Apportionment Modeling

The Texas Petitioners also challenge the EPA’s
modification of the Texas Commission’s
source-apportionment modeling on three fronts. First, they
argue that the EPA has not rationally explained why it
considered the source-apportionment modeling’s projected
absolute impact—instead of its projected relative impact—that
wind from Wise County would have on violating Dallas–Fort
Worth area monitors. Second, they argue that the EPA’s
analysis of the Texas Commission’s source-apportionment
modeling was inconsistent with its analysis of
 58
source-apportionment modeling submitted in connection with
Illinois’s designation of Lake County. And third, they argue
that the EPA’s decision to examine the model’s projected
absolute impact rather than its relative impact violated the
EPA’s earlier modeling guidance.

We note, at the outset, that the EPA’s application,
interpretation and modification of source-apportionment
modeling plainly fall “within its technical expertise” and thus
we owe it “an extreme degree of deference.” ATK Launch
Sys., 669 F.3d at 338 (quotation marks omitted). To withstand
judicial review, the EPA needs to articulate only a “rational
connection between the facts found and the choice made,”
Burlington Truck Lines, 371 U.S. at 168, show that it treated
“similar counties” similarly, ATK Launch Sys., 669 F.3d at 336
(emphasis in original), and demonstrate that it did not run afoul
of binding guidance, see generally Appalachian Power Co. v.
EPA, 208 F.3d 1015, 1020–23 (D.C. Cir. 2000). Because the
EPA has done all three, we will not disturb its designation of
Wise County as nonattainment based on the Texas Petitioners’
objections to its interpretation of the Texas Commission’s
source-apportionment modeling.

First, the Texas Petitioners challenge the EPA’s decision
to reinterpret the source-apportionment modeling submitted by
the Texas Commission. As discussed, supra § III.F.1, when
the EPA received the Texas Commission’s
source-apportionment modeling data during the comment
period, it observed that the model did not rely on data from an
entire ozone season. Rather, the projections in the Texas
Commission’s model relied on data from June 2006 only. The
Texas Commission based its approach on the fact that June
2006 purportedly presented “an exceptionally rich set of air
quality and meteorological measurements,” “had the most
high-ozone days of any month” and experienced “all the
 59
meteorological conditions linked to formation of high ozone
concentration.” See Response to Texas Commission on
Environmental Quality’s Reconsideration Pet. at 3.

Despite these assurances, the EPA did not agree that one
month of data, even an “exceptionally rich” month, was
sufficient. Specifically, the EPA observed that the ozone
season in the Dallas–Fort Worth area was bimodal (i.e.,
reporting its highest ozone values in July-September but
experiencing a lower ozone peak in May-June) and that the
Texas Commission’s reliance on limited data meant that it
failed to account for “all of the meteorology regimes conducive
for ozone events” in the Dallas–Fort Worth area. See Final
Dallas–Fort Worth Area Designations at 16. According to the
EPA, “emphasis on the average modeled impact is more
appropriate when a full ozone season of model results is
available.” See Resp’t’s Br. 131. Because the Texas
Commission’s model was premised on baseline data excluding
“events that happen in mid to late-summer that often set” the
Dallas–Fort Worth area’s ozone levels, the EPA examined
both the projected average impact and the projected maximum
impact of Wise County emissions. See Final Dallas–Fort
Worth Area Designations at 16.

At bottom, the EPA had a “basic obligation” to conduct
“reasoned decisionmaking.” Catawba Cnty., 571 F.3d at 25.
When presented with the Texas Commission’s
source-apportionment modeling, the EPA determined that it
“needed to be carefully evaluated and could not simply be
accepted at face value,” Resp’t’s Br. 126, identified several
methodological flaws in the Texas Commission’s data,
adjusted the Texas Commission’s submissions to account for
the flaws and articulated, quite thoroughly, a “rational
connection between the facts found and the choice made.”
Burlington Truck Lines, Inc., 371 U.S. at 168. On this record,
 60
we cannot say that the EPA acted arbitrarily or capriciously in
re-evaluating the Texas Commission’s source-apportionment
modeling data. Rather, the EPA’s thorough treatment of all
available data indicates that it in fact “surpassed” its
“obligation of reasoned decisionmaking.” Catawba Cnty.,
571 F.3d at 25.

Second, the Texas Petitioners argue that the EPA’s
modification to the Texas Commission’s
source-apportionment modeling subjected Wise County to
arbitrarily disparate treatment. They compare the EPA’s
interpretation of the Texas Commission’s modeling to its
interpretation of source-apportionment modeling for the
Chicago area. Specifically, they argue that (1) emissions
from Jasper County, a Chicago-area county with attainment
status, had a projected average impact on violating monitors
similar to Wise County’s; (2) the EPA should have evaluated
the average impact of Wise County’s emissions on violating
monitors as it did for Jasper County; and (3) the EPA’s
evaluation of Wise County’s maximum, as opposed to relative,
estimated impact was, accordingly, inconsistent and resulted in
an arbitrarily different result between Wise County and Jasper
County.

Again, we emphasize that applying different methods to
different areas, standing alone, does not give rise to arbitrarily
disparate treatment and given “significant” relevant
differences between two areas, “a direct one-to-one
comparison of the data” or the methods used to measure such
data can be “inappropriate.” ATK Launch Sys., 669 F.3d at
337. Here, the significant difference lies in the quality of data
submitted by the Texas Commission compared to that
submitted in support of Jasper County. Specifically, the
source-apportionment model submitted in support of the
Chicago-area designations included data from a full ozone
 61
season, which made “emphasis on the average modeled impact
. . . more appropriate.” Resp’t’s Br. 131. As noted, the EPA
modified the Texas Commission’s source-apportionment
model because it did not include data from a full ozone season.

Moreover, the EPA had to compensate for the fact that the
Texas Commission’s source-apportionment model
underestimated the number of days that monitors in the Dallas–
Fort Worth area exceeded the ozone NAAQS because the
model under-predicted peak ozone levels around the monitors,
sometimes by a significant range. The source-apportionment
model for Jasper County, however, had the opposite problem;
it did not account for recent emissions reductions at a Jasper
County power plant and thus the Chicago-area
source-apportionment model over-reported Jasper County’s
emissions impact. See Chicago Area Designations at 9–10.
Stated differently, because Wise County’s model
under-reported its emissions impact and Jasper County’s
model over-reported its emissions impact, the EPA reasonably
concluded that the two counties should receive different
attainment designations.

Third, the Texas Petitioners argue that the EPA arbitrarily
and capriciously deviated from its earlier guidance on
source-apportionment modeling, which guidance allegedly
expressed a preference for relative, rather than absolute,
modeling. Specifically, they argue that the EPA’s reliance on
Wise County’s maximum potential emissions impact directly
conflicts with the EPA’s 2007 “Guidance on the Use of Models
and Other Analyses for Demonstrating Attainment of Air
Quality Goals for Ozone, PM2.5, and Regional Haze” (2007
Attainment Guidance). In that guidance, the EPA stated that
its “recommended test is one in which model estimates are
used in a ‘relative’ rather than ‘absolute’ sense.” Id. at 15.
 62
As a threshold matter, the 2007 Attainment Guidance does
not speak to the use of source-apportionment modeling in the
designation process; rather, it recommends procedures that a
state can use after it has been designated as nonattainment to
show that its proposed emission control strategy will
eventually result in attainment status. But even assuming that
the 2007 Attainment Guidance informs the current NAAQS
designation process, the EPA did not err by deviating from it.
Indeed, the 2007 Guidance expressly contemplates deviations
in appropriate cases:

This document does not substitute for any
Clean Air Act provision or EPA regulation, nor
is it a regulation itself. Thus, it does not
impose binding, enforceable requirements on
any party, nor does it assure that EPA will
approve all instances of its application. The
guidance may not apply to a particular
situation, depending upon the circumstances.
The EPA and State decision makers retain the
discretion to adopt approaches on a
case-by-case basis that differ from this
guidance where appropriate. . . .

Users are cautioned not to regard statements
recommending the use of certain procedures or
defaults as either precluding other procedures
or information, or providing guarantees that
using these procedures or defaults will result in
actions that are fully approvable. . . . EPA
cannot assure that actions based upon this
guidance will be fully approvable in all
instances.

2007 Attainment Guidance at ix.
 63
As noted, the EPA fully explained why it revised and
independently evaluated the Texas Commission’s
source-apportionment modeling to account for “the limited
data set [the Texas Commission] relied upon.” Resp’t’s
Br. 136. Because the 2007 Attainment Guidance did not
compel the EPA to limit its consideration to relative projected
impacts, and because the EPA articulated a “rational
connection between the facts found and the choice made,”
Catawba Cnty., 571 F.3d at 41, it did not act arbitrarily or
capriciously when it relied on Wise County’s absolute, rather
than relative, impact on NAAQS-violating monitors.

The fundamental deficiency in the Texas Petitioners’
challenges to the EPA’s revision of the Dallas–Fort Worth area
source-apportionment model is that, to establish that “EPA’s
administration of the complicated provisions of the Clean Air
Act” was erroneous, Catawba Cnty., 571 F.3d at 41, they have
to demonstrate more than mere disagreement with the EPA’s
reasoning. Barring an unreasonable or irrational application
of the “scientific data within [the EPA’s] technical expertise,”
City of Waukesha, 320 F.3d at 247, we cannot say that the EPA
acted arbitrarily or capriciously. The record plainly shows
that the EPA “considered all relevant factors and articulated a
‘rational connection between the facts found and the choice
made’ ” when it declined to accept the Texas Commission’s
source-apportionment model without modification. Catawba
Cnty., 571 F.3d at 41 (quoting Burlington Truck Lines, 371
U.S. at 168). We therefore hold that the EPA did not act
arbitrarily or capriciously when it did so.
 64
3. The Constitutional Challenges

In this section, we address three constitutional challenges
that Texas, Wise County, and the Texas Commission on
Environmental Quality (collectively, Texas State Petitioners)
raise to the EPA’s designation of Wise County, Texas as a
nonattainment area.

i. The Tenth Amendment & The Spending Clause

The Texas State Petitioners, joined by the Mississippi
Petitioners, argue that § 7407(d)(1)(B) and related sections of
the Clean Air Act—at least to the extent that they authorize the
EPA to override the State’s designation and declare Wise
County a nonattainment area—violate the Tenth Amendment
and exceed the Congress’ authority under the Spending Clause.

First, the Texas State Petitioners maintain that
§ 7407(d)(1)(B) unlawfully permits the EPA to
“commandeer[] State regulators to enforce a federal regulatory
program.” State & County Br. 32. The section grants the
EPA authority to “make such modifications as the
Administrator deems necessary to the designations of the areas
. . . submitted [by the States].” 42 U.S.C. § 7407(d)(1)(B)(ii).
According to the petitioners, “[w]hen EPA overrides a State, it
compels State regulators to enforce a myriad of federal
requirements involving emissions controls, clean fuel
programs, transportation and land use limitations in the
designated area.” State & County Br. 33 (citing 42 U.S.C.
§§ 7511 et seq. (outlining requirements specific to ozone
nonattainment areas)).

The Texas State Petitioners are correct that “the Federal
Government may not compel the States to implement . . .
federal regulatory programs,” Printz v. United States, 521 U.S.
 65
898, 925 (1997). 20 But the Clean Air Act does not do that.
Instead, the statutory scheme authorizes the EPA to
promulgate and administer a federal implementation plan of its
own if the State fails to submit an adequate state
implementation plan. See 42 U.S.C. § 7410(c). And as we
recently noted, the Supreme Court has “repeatedly affirm[ed]
the constitutionality of federal statutes that allow States to
administer federal programs but provide for direct federal
administration if a State chooses not to administer it.” Texas
v. EPA, 726 F.3d 180, 196–97 (D.C. Cir. 2013) (citing New
York v. United States, 505 U.S. 144, 167–68, 173–74 (1992);
Hodel v. Va. Surface Mining & Reclamation Ass’n, Inc., 452
U.S. 264, 288 (1981)). Here, too, the “full regulatory burden
will be borne by the Federal Government” if a State chooses
not to submit an implementation plan. Va. Surface Mining &
Reclamation Ass’n, 452 U.S. at 288. Under these
circumstances, “there can be no suggestion that the Act
commandeers . . . the States.” Id.

Second, the Texas State Petitioners maintain that the
Clean Air Act’s sanctions for noncompliant states impose such
a steep price that State officials effectively have no choice but
to comply—in contravention of the Supreme Court’s decision
in National Federation of Independent Business v. Sebelius
(NFIB), 132 S. Ct. 2566, 2603 (2012) (plurality opinion). See
20
See Nat’l Fed’n of Indep. Bus. v. Sebelius, 132 S. Ct. 2566,
2602 (2012) (plurality opinion) (noting that the Court has struck
“down federal legislation that commandeers a State’s legislative or
administrative apparatus for federal purposes”); Printz, 521 U.S. at
933 (invalidating federal legislation compelling State law
enforcement officers to perform federally mandated background
checks on handgun purchasers); New York v. United States, 505 U.S.
144, 174–77 (1992) (invalidating a provision of a federal statute
compelling a State either to take title to nuclear waste or to enact
particular state waste regulations).
 66
State & County Br. 33–34. The Act requires the EPA to
impose sanctions on a State that fails to submit an adequate
plan or implement an approved plan if it does not correct the
deficiency within 18 months. See 42 U.S.C. § 7509(a). The
focus of the petitioners’ challenge is the sanction regarding
federal highway funds. Under the Act, the EPA
Administrator may prohibit the approval of any transportation
projects or grants within the nonattainment area, except those
that the Secretary of Transportation determines are intended to
resolve a demonstrated safety problem and will likely result in
a reduction in accidents. Id. § 7509(b)(1)(A). The Secretary
of Transportation may also continue to approve a number of
other kinds of projects and grants, notwithstanding the EPA
Administrator’s prohibition. Id. § 7509(b)(1)(B)(i)–(viii)
(authorizing continued approval of projects and grants
including capital programs for public transit, projects affecting
bus lanes and high occupancy vehicle lanes, programs that
improve traffic flow, and programs that “would improve air
quality and would not encourage single occupancy vehicle
capacity”).

As Chief Justice Roberts noted in NFIB, the Supreme
Court has “long recognized that Congress may use” the power
given it by the Spending Clause “to grant federal funds to the
States, and may condition such a grant upon the States’ ‘taking
certain actions that Congress could not require them to take.’ ”
NFIB, 132 S. Ct. at 2601 (quoting Coll. Sav. Bank v. Fla.
Prepaid Postsecondary Educ. Expense Bd., 527 U.S. 666, 686
(1999)). “Such measures ‘encourage a State to regulate in a
particular way, [and] influenc[e] a State’s policy choices.’ ”
Id. at 2601–02 (quoting New York, 505 U.S. at 166) (alterations
in original). “The conditions imposed by Congress ensure
that the funds are used by the States to ‘provide for the . . .
general Welfare’ in the manner Congress intended.” Id. at
2602 (quoting U.S. CONST., art. I, § 8, cl. 1).
 67
“At the same time,” the Chief Justice continued, the
Court’s “cases have recognized limits on Congress’s power
under the Spending Clause to secure state compliance with
federal objectives.” Id. The Court has “repeatedly
characterized . . . Spending Clause legislation as ‘much in the
nature of a contract.’ ” Id. (quoting Barnes v. Gorman, 536
U.S. 181, 186 (2002) (quoting Pennhurst State Sch. & Hosp. v.
Halderman, 451 U.S. 1, 17 (1981))). “The legitimacy of
Congress’s exercise of the spending power ‘thus rests on
whether the State voluntarily and knowingly accepts the terms
of the contract.’ ” Id. (quoting Pennhurst, 451 U.S. at 17)
(some internal quotation marks omitted). “Congress may use
its spending power to create incentives for States to act in
accordance with federal policies,” the Chief Justice concluded,
“[b]ut when ‘pressure turns into compulsion,’ the legislation
runs contrary to our system of federalism.” Id. (quoting
Steward Mach. Co. v. Davis, 301 U.S. 548, 590 (1937)). 21

In NFIB, the Court struck down—as in excess of the
Congress’ authority under the Spending Clause—a provision
of the Affordable Care Act (ACA) that expanded the scope of
the Medicaid program and increased the number of individuals
the States had to cover. Although the Act increased federal
funding to cover much of the States’ costs in expanding
Medicaid coverage, it also provided that, if a State did not

21
As we discuss below, the Texas State Petitioners argue that
the threat of highway sanctions makes the promulgation of SIP
provisions for a nonattainment area effectively compulsory. They
do not argue that the sanctions provision fails to comply with any
other constitutional requirements governing conditions on federal
grants to the States. See South Dakota v. Dole, 483 U.S. 203, 207–
08 (1987) (requiring that conditions promote the general welfare, be
unambiguous, be related to the federal interest, and be consistent
with other constitutional provisions).
 68
comply with the Act’s new coverage requirements, it could
lose not only the new federal funding, but all of its existing
federal Medicaid funds. NFIB, 132 S. Ct. at 2582. The Chief
Justice’s plurality opinion—for himself and Justices Breyer
and Kagan—controls our decision on this issue. 22

In addressing the question of overbearing financial
coercion, the Chief Justice first discussed Dole, 483 U.S. 203,
in which the Court rejected such a challenge. In that case, the
Congress had threatened to withhold 5 per cent of a State’s
federal highway funding unless the State raised its drinking age
to 21. The Chief Justice noted that, although “the condition
was ‘directly related to one of the main purposes for which
highway funds are expended—safe interstate travel,’ ” it “was
not a restriction on how the highway funds—set aside for
specific highway improvement and maintenance efforts—were
to be used.” NFIB, 132 S. Ct. at 2604 (quoting Dole, 483 U.S.
at 208). “[A]ccordingly,” he said, the Dole Court “asked
whether ‘the financial inducement offered by Congress’ was
‘so coercive as to pass the point at which pressure turns into
compulsion.’ ” Id. (quoting Dole, 483 U.S. at 211) (some
internal quotation marks omitted). The Court answered that
this monetary sanction was not impermissibly coercive, but
rather offered only “relatively mild encouragement to the
22
When a majority of the Supreme Court agrees on a result, but
“no single rationale explaining the result enjoys the assent of five
Justices, ‘the holding of the Court may be viewed as that position
taken by those Members who concurred in the judgments on the
narrowest grounds . . . .’ ” Marks v. United States, 430 U.S. 188,
193 (1977) (quoting Gregg v. Georgia, 428 U.S. 153, 169 n.15
(1976) (plurality opinion)). The NFIB plurality found a Spending
Clause violation on narrower grounds than did the joint opinion of
Justices Scalia, Kennedy, Thomas, and Alito, NFIB, 132 S. Ct. at
2656–69. See Mayhew v. Burwell, 772 F.3d 80, 88–89 (1st Cir.
2014). It therefore controls here. Id.
 69
states” because “all South Dakota would lose if she adheres to
her chosen course as to a suitable minimum drinking age is
5%” of her federal highway funds. Dole, 483 U.S. at 211; see
NFIB, 132 S. Ct. at 2604. “In fact,” as the Chief Justice
further noted in NFIB, “the federal funds at stake constituted
less than half of one percent of South Dakota’s budget at the
time.” NFIB, 132 S. Ct. at 2604.

In NFIB, the Chief Justice found that, as in Dole, the
conditions the ACA imposed on the States did not “govern the
use of” the new funds it granted to the States, but rather took
“the form of threats to terminate other significant independent
grants” already in existence. Id. Accordingly, he said, “the
conditions are properly viewed as a means of pressuring the
States to accept policy changes” and their level of coerciveness
therefore had to be evaluated. Id. Upon doing so, the Chief
Justice found the ACA’s financial sanction to be “a gun to the
head,” in contrast to the “mild encouragement” in Dole. Id.
A State that opted out of the ACA’s Medicaid expansion stood
“to lose not merely ‘a relatively small percentage’ of its
existing Medicaid funding, but all of it.” Id. (quoting Dole,
483 U.S. at 211). That, the Chief Justice found, could amount
to “over 10 percent of a State’s overall budget.” Id. at 2604–
05.

In the case now before us, the Congress has conditioned
some federal highway funding on Texas’s adoption of an
adequate implementation plan. This condition, like the one at
issue in Dole, is—at least arguably—not a restriction on how
the highway funds are to be used, but rather an incentive to
encourage States to take action in a related policy area. But
see discussion infra. Although as discussed below we are
uncertain whether that alone is sufficient to trigger a
coerciveness inquiry, we will proceed to evaluate the coercive
effect of section 7509(b). For the following reasons, we find
 70
that the potential funding sanctions contained in section
7509(b) of the Clean Air Act are not nearly as coercive as those
in the ACA.

First, unlike the situation in NFIB and like that in Dole, a
noncompliant State does not risk losing all federal funding for
an existing program. To the contrary, the EPA Administrator
can only prohibit funding for transportation projects or grants
applicable to the nonattainment area. 42 U.S.C.
§ 7509(b)(1)(A); 40 C.F.R. § 52.31(b)(3), (e)(2) (providing
that the “highway funding sanction shall apply . . . only to . . .
areas that are designated nonattainment”); see Virginia v.
Browner, 80 F.3d 869, 881 (4th Cir. 1996) (“[A] state does not
lose any highway funds that would be spent in areas of the state
that are in attainment.”). Even within the nonattainment area,
the Administrator may not prohibit the approval of projects or
grants that the Secretary of Transportation determines are
intended to resolve a demonstrated safety problem and will
likely result in a reduction in accidents. 42 U.S.C.
§ 7509(b)(1)(A). Indeed, the Secretary of Transportation may
continue to approve a number of other kinds of projects and
grants as well, including those that “would improve air
quality.” Id. § 7509(b)(1)(B)(viii); see id. § 7509(b)(1)(B)(i)–
(viii).

Second, the threatened loss of federal highway funding
does not even approach the “over 10 percent of a State’s
overall budget” at issue in NFIB. Texas advises us that it
received more than $3 billion in federal highway and transit
funds in 2013. State & County Br. 33 n.29. Even if all of
that were withheld, it would still have amounted to less than 4
per cent of the State’s 2013 budget. 23 But as noted above,

23
See Nat’l Ass’n of State Budget Officers, The State
Expenditure Report 2012–2014 8 (2014) (listing 2013 expenditures
 71
Texas does not stand to lose all of its highway funds. The
potential sanction applies, at most, to highway funds for
projects in nonattainment areas. Wise County is the only
county for which the petitioners make a Tenth Amendment
argument, and because it is only one of 254 Texas counties, it
is unlikely that the loss of even all of that county’s federal
highway funds would put a serious dent in the State’s total
budget. 24 Moreover, as also noted above, it is unlikely that
even that one county would lose all of its federal highway
funding because the potential sanction does not extend to
funding for a list of enumerated projects. See 42 U.S.C.
§ 7509(b)(1)(A), (B)(i)–(viii).

In short, it is clear that Texas does not risk losing
anywhere near the percentage of its federal funding—either for
the program at issue or of its overall budget—that the Court
found fatal in NFIB. Precisely how much less, we do not
know. But the burden of establishing unconstitutionality is on
the challenger, and Texas has failed to provide the necessary
information. That failure is further ground for rejecting the
State’s constitutional challenge. See NFIB, 132 S. Ct. at 2662
(joint opinion of Scalia, Kennedy, Thomas, and Alito, JJ.)

as approximately $93 billion); Texas General Appropriations Act for
the 2012–13 Biennium xi (2011), available at
http://www.lbb.state.tx.us/Documents/GAA/General_Appropriation
s_Act_2012-13.pdf (appropriating approximately $79 billion for
2013).
24
Seventeen other Texas counties are also in nonattainment
areas. See Final Dallas–Fort Worth Area Designations at 1;
Houston-Galveston-Brazoria, Texas Final Area Designations for the
2008 Ozone NAAQS at 1. But that is still only a small percentage
of the State’s total of 254 counties. See also Envtl. Prot. Agency,
Map of Texas 8-hour Ozone Nonattainment Areas (2008 Standard),
available at http://www.epa.gov/oaqps001/ greenbk/tx8_2008.html.
 72
(“[C]ourts should not conclude that legislation is
unconstitutional on this ground unless the coercive nature of an
offer is unmistakably clear.”); see also United States v.
Morrison, 529 U.S. 598, 607 (2000) (requiring a “plain
showing” of unconstitutionality); United States v. Bland, 472
F.2d 1329, 1334 (D.C. Cir. 1972) (en banc) (noting that “the
burden of establishing the unconstitutionality of a statute rests
on him who assails it”).

Finally, although we have concluded that the highway
sanction is not unconstitutionally coercive, we note some
uncertainty as to whether a coerciveness inquiry was required.
There are two circumstances that may distinguish this case
from those in which the Supreme Court has found such an
inquiry necessary.

First, as described in NFIB, the inquiry in Dole was
triggered by the fact that the Congress had imposed a condition
that did not restrict how the federal highway funds at issue
were to be used. Here, by contrast, the condition and sanction
do redirect the federal highway funds of non-complying states
to programs of the Congress’ choosing, including those that
“would improve air quality and would not encourage single
occupancy vehicle capacity.” 42 U.S.C.
§ 7509(b)(1)(B)(viii); see id. § 7509(b)(1)(B)(i)–(viii). As
the Senate Committee Report on the 1990 Clean Air Act
amendments explains, for nonattainment areas in States that
fail to submit an adequate SIP, “Federal transportation
investments” are “shifted to transportation programs that are
designed to provide alternatives to the single occupancy
vehicle and that contribute to reducing future [vehicle miles
traveled].” S. REP. NO. 101-228, at 26 (1989).

Second, the condition at issue in Dole—which required
the States to raise their drinking age to 21—was also, at the
 73
time of South Dakota’s challenge, a new condition that had not
been part of the original program. In NFIB, although the
condition was a restriction on how Medicaid funds could be
spent, Chief Justice Roberts found that the condition was also a
new one. “Indeed,” he stressed, “the manner in which the
expansion is structured indicates that while Congress may have
styled the expansion a mere alteration of existing Medicaid, it
recognized it was enlisting the States in a new health care
program.” NFIB, 132 S. Ct. at 2606. This was important, he
said, because “Spending Clause legislation [is] much in the
nature of a contract,” id. at 2602 (internal quotation marks
omitted), and “[t]hough Congress’ power to legislate under the
spending power is broad, it does not include surprising
participating States with post-acceptance or retroactive
conditions,” id. at 2606 (internal quotation marks omitted). In
both Dole and NFIB, the condition at issue was “new” in two
senses of the word: Both conditions had been recently
enacted at the time of the litigation, and both conditions
imposed additional requirements with which States had to
comply to continue receiving preexisting federal funding.

Neither the Clean Air Act’s requirement to submit an
implementation plan, nor its highway funds sanction, is a
condition that has been newly imposed on the States.
Although both were new in 1977, see Clean Air Act
Amendments of 1977, Pub. L. No. 95-95, §§ 103, 176, 91 Stat.
685, 687–88, 749–50 (1977), since then Texas has submitted
implementation plans and accepted billions of dollars in
highway funding. Accordingly, when the EPA issued the
Wise County nonattainment designation in 2012, Texas was
not suddenly surprised by dramatically new conditions
retroactively imposed after a long period in which the State had
accepted and relied upon unconditional federal funding—as
was the case in NFIB.
 74
These differences from the Supreme Court’s precedents
create some uncertainty as to whether the coerciveness inquiry
employed in Dole and NFIB was even triggered by the Clean
Air Act provisions at issue here. Even if it were, the fact that
the State has long accepted billions of dollars notwithstanding
the challenged conditions may be an additional relevant factor
in the contract-like analysis the Court has in mind for assessing
the constitutionality of Spending Clause legislation. But we
need not resolve that uncertainty today. Because the
challenged provisions of the Clean Air Act survive a
coerciveness inquiry in any event, we reject the Texas State
Petitioners’ challenge to their constitutionality.

ii. The Commerce Clause

The Texas State Petitioners also argue that the Wise
County designation exceeds the scope of the Congress’
authority under the Commerce Clause. As explained above,
supra § III.F.1, the designation declared that Wise County
contributed enough ozone emissions to nearby violations of the
NAAQS to warrant its own nonattainment designation. By
virtue of that designation, sources of emissions within the
county must comply with a variety of additional requirements.
See, e.g., 42 U.S.C. § 7502(c)(1) (requiring the implementation
of “all reasonably available control measures”); id.
§ 7502(c)(5) (requiring “permits for the construction and
operation of new or modified major stationary sources
anywhere in the nonattainment area”).

The Commerce Clause grants the Congress the power
“[t]o regulate Commerce . . . among the several States.” U.S.
CONST., art. I, § 8, cl. 3. The Supreme Court has “recognized
. . . that ‘[t]he power of Congress over interstate commerce is
not confined to the regulation of commerce among the states,’
but extends to activities that ‘have a substantial effect on
 75
interstate commerce.’ ” NFIB, 132 S. Ct. at 2585–86 (opinion
of Roberts, C.J.) (quoting United States v. Darby, 312 U.S.
100, 118–19 (1941)); see United States v. Lopez, 514 U.S. 549,
558–59 (1995). “Congress’s power, moreover, is not limited
to regulation of an activity that by itself substantially affects
interstate commerce, but also extends to activities that do so
only when aggregated with similar activities of others.”
NFIB, 132 S. Ct. at 2586 (opinion of Roberts, C.J.) (citing
Wickard v. Filburn, 317 U.S. 111, 127–28 (1942)). The
question for a court is whether there was a “rational basis” for
the Congress’ conclusion that a regulated activity substantially
affects interstate commerce. Hodel v. Indiana, 452 U.S. 314,
323–24 (1981); see Nat’l Ass’n of Home Builders v. Babbitt
(NAHB), 130 F.3d 1041, 1051 (D.C. Cir. 1997) (opinion of
Wald, J.).

The Texas State Petitioners’ first contention is that the
NOx emissions produced by oil and gas activity in the Barnett
Shale in Wise County do not “ ‘substantially affect’ interstate
commerce,” principally because the emissions are “wholly
intrastate.” State & County Br. 36. That premise is
unsupported by any proffered evidence and is factually
incorrect. The phenomenon of interstate transport of ozone
has been thoroughly studied, and it has been recognized by the
Congress, the EPA, the Supreme Court, and this Court. 25 The

25
See 42 U.S.C. §§ 7410(a)(2)(D), 7511c (Clean Air Act
provisions addressing interstate transport of ozone); S. REP.
NO. 101-228, at 34 (1989) (discussing Clean Air Act amendments
designed to “[c]ontrol . . . interstate ozone pollution”); id. at 13
(noting that “ozone is not a local phenomenon but is formed and
transported over hundreds of miles and several days”); 2008
Designations Rule, 77 Fed. Reg. at 30,089 (finding that ozone and
ozone precursors travel easily through the atmosphere, which can
result in NAAQS violations hundreds of miles from the precursors’
source); EME Homer City Generation, L.P., 134 S. Ct. at 1594
 76
“winds, of course, recognize no [state] boundaries.” United
States v. Ford Motor Co., 814 F.2d 1099, 1102 (6th Cir. 1987).

But even if the particular emissions from the Barnett Shale
stopped at the Texas state line, the regulation of their sources
would still be permissible under the Commerce Clause for two
reasons. First, “where a general regulatory statute bears a
substantial relation to commerce, the de minimis character of
individual instances arising under that statute is of no
consequence.” Lopez, 514 U.S. at 558 (internal quotation
marks omitted) (emphasis omitted); see Gonzales v. Raich, 545
U.S. 1, 17 (2005); NAHB, 130 F.3d at 1046 (opinion of Wald,
J.). And there is no doubt that the general regulatory scheme
of the Clean Air Act has a substantial relation to interstate
commerce. Indeed, the same is true even if we focus only
upon the Act’s generally applicable ozone provisions.

Moreover, we can find a substantial effect not only by
examining the emissions that are produced, but also by
examining the activities that the challenged statute regulates to
reduce the production of those emissions. See Rancho Viejo,
LLC v. Norton, 323 F.3d 1062, 1067 (D.C. Cir. 2003); NAHB,
130 F.3d at 1046 & n.3 (opinion of Wald, J.); id. at 1058
(Henderson, J., concurring). As we explained in Rancho
Viejo, on this rationale we “focus[] on the activity that the
federal government seeks to regulate.” 323 F.3d at 1069; see
Morrison, 529 U.S. at 609 (instructing that “the proper
inquiry” is whether the challenge is to “a regulation of activity
that substantially affects interstate commerce”) (emphasis

(detailing the “journey” taken by ozone precursors, which “often
develop into ozone . . . by the time they reach the atmospheres of
downwind States”); Virginia v. EPA, 108 F.3d 1397, 1400 (D.C. Cir.
1997) (describing the “ozone transport phenomenon” in the lower
atmosphere).
 77
added); Lopez, 514 U.S. at 558–59 (“Congress’ commerce
authority includes the power to regulate . . . those activities that
substantially affect interstate commerce.”) (emphasis added).
In Rancho Viejo, we upheld the constitutionality of the Fish
and Wildlife Service’s decision to protect an endangered toad
species by regulating a housing development, on the ground
that the regulated activity, a “202-acre project, located near a
major interstate highway, [was] . . . presumably being
constructed using materials and people from outside the state.”
323 F.3d at 1069 (internal quotation marks omitted).
Likewise, in NAHB, we upheld the Service’s decision to
protect an endangered fly species by regulating the
construction plan for a hospital, on the ground that the
commercial land development at issue “ha[d] a plain and
substantial effect on interstate commerce.” 130 F.3d at 1059
(Henderson, J., concurring); see id. at 1056 (opinion of Wald,
J.).

Here, the activities that the EPA seeks to regulate are the
commercial, industrial, and extraction processes that produce
the emissions at issue. See 42 U.S.C. § 7511a; 2008
Designations Rule, 77 Fed. Reg. at 30,089. The
nonattainment designation triggers regulatory controls on the
sources of those emissions, many of which are indisputably
entities engaged in substantial interstate commerce. In the
case of Wise County in particular, those entities include
multinational companies engaged in the production and sale of
oil and gas from the Barnett Shale, including several of the
Industrial Petitioners here. 26 The restrictions triggered by the
26
Industrial Petitioner Devon Energy Corporation, for example,
“is a leading independent oil and natural gas exploration and
production company,” with operations “focused onshore in the
United States and Canada.” Industrial Br. iv. “Devon is also one
of North America’s larger processors of natural gas liquids, with . . .
natural gas processing facilities in many of its producing areas,
 78
nonattainment designation thus affect the conditions under
which interstate commerce in oil and gas may proceed. And
as such, the designation process “regulates and substantially
affects commercial . . . activity which is plainly interstate.”
NAHB, 130 F.3d at 1058 (Henderson, J., concurring).

The Texas State Petitioners’ second contention is that,
“[e]ven if incidental emissions do ‘substantially affect’
interstate commerce, they are not ‘quintessentially economic
activity’ ” and cannot be regulated under the Commerce
Clause. State & County Br. 36. This contention is based on
the Court’s decision in Lopez, which held the Gun-Free School
Zones Act unconstitutional in part because the statutory
provision at issue, which criminalized the possession of a gun
in a school zone, had “nothing to do with ‘commerce’ or any
sort of economic enterprise, however broadly one might define
those terms.” Lopez, 514 U.S. at 560–61; see also Morrison,
529 U.S. at 610–11, 613. There are two answers to this
contention.

First, ozone pollution itself has economic consequences
for interstate commerce. The Congress so found in the course
of amending the Clean Air Act. See S. REP. NO. 101-228, at 8
(1989) (noting that exposure to air pollution costs the United
States $40 billion annually in additional health care costs, and
documenting health effects of ozone and other pollutants); id.
(noting that “ozone causes annual crop losses of $2 to $3
billion per year”). Although we are not bound by
congressional findings, they may assist us in “evaluat[ing] the
legislative judgment that the activity in question substantially

including Wise County, Texas.” Id.; see id. at 13 (“Industrial
Petitioners and members with operations in Wise County were
immediately subjected to increased regulatory burdens due to the
nonattainment designation.”).
 79
affected interstate commerce.” Lopez, 514 U.S. at 562–63;
see Rancho Viejo, 323 F.3d at 1069. Indeed, we have
previously credited the Congress’ findings regarding ozone
pollution, concluding that the Act’s “legislative history and
EPA’s report to Congress substantiate the heavy impact ozone
pollution has on national health care costs and national
agricultural production.” Allied Local, 215 F.3d at 83.

Second, the activities that are ultimately regulated by the
designation process are not the ozone precursor “emissions,”
but rather the activities that produce the emissions. Those
include the operation of power plants, gas processors, and
vehicles that produce the emissions. See 42 U.S.C. § 7511a.
As we explained in Rancho Viejo, the regulated activity in that
case was a company’s “planned commercial development, not
the arroyo toad that it threaten[ed].” 323 F.3d at 1072. The
same point is true here. Just as the Endangered Species Act
“does not purport to tell toads what they may or may not do,”
id., the Clean Air Act does not tell NOx or VOCs what to do.
Rather, it tells the commercial and industrial sources that
produce those compounds what they may do.

As we noted in Allied Local, the Supreme Court has long
made clear that “ ‘the power conferred by the Commerce
Clause [is] broad enough to permit congressional regulation of
activities causing air or water pollution, or other environmental
hazards that may have effects in more than one State.’ ”
Allied Local, 215 F.3d at 83 (quoting Va. Surface Mining &
Reclamation Ass’n, 452 U.S. at 282) (emphasis added); id.
(noting that the Supreme Court cited Virginia Surface Mining
and Reclamation Association with approval in both Lopez and
Morrison). “[B]ecause we are required to accord
congressional legislation a ‘presumption of
constitutionality,’ ” Rancho Viejo, 323 F.3d at 1069 (quoting
Morrison, 529 U.S. at 607), the petitioners’ inability to
 80
establish that emissions-producing sources in the State do not
substantially affect interstate commerce “is fatal to [their]
cause,” id.; see Morrison, 529 U.S. at 607 (“Due respect for the
decisions of a coordinate branch of Government demands that
we invalidate a congressional enactment only upon a plain
showing that Congress has exceeded its constitutional
bounds.”). The regulation of the sources of Wise County
emissions through the Clean Air Act’s designation process lies
well within the Congress’ authority to regulate interstate
commerce.

iii. The Due Process Clause

The Texas State Petitioners’ third constitutional challenge
maintains that the EPA’s designation of Wise County violated
the Due Process Clause because the former Administrator of
EPA Region 6, Al Armendariz, failed to disqualify himself
from the proceedings.

According to the petitioners, Armendariz should have
disqualified himself for four reasons. First, Armendariz has a
history of working for environmental advocacy groups.
Second, a report he authored as an advocate before joining the
EPA concluded that emissions from the Barnett Shale were
contributing significantly to local and global pollution. Third,
a speech Armendariz gave after joining the EPA analogized his
aggressive enforcement policy against oil and gas companies
that “are not complying with the law” to the way “Romans
used to conquer those villages in the Mediterranean” by
“crucify[ing]” the first people they saw. Terrence Henry,
Texas EPA Official Apologizes for ‘Crucify Them’ Comments,
Apr. 26, 2012, State Impact NPR,
http://stateimpact.npr.org/texas/2012/04/26/epa-official-apolo
gizes-for-crucify-comments (quoting Armendariz). “You
make examples out of people who . . . are not complying with
 81
the law,” Armendariz said. “There’s a deterrent
factor. . . . And they decide at that point that it’s time to clean
up.” Id. 27 Finally, in the petitioners’ view, “[n]ormally, the
prevailing wind direction and EPA-standard modeling would
have led EPA to accept” Texas’s designation of Wise County
as attainment. State & County Br. 38. All of this, the
petitioners argue, “create[s] a presumption that the Agency’s
mind was closed and it was unwilling or unable to rationally
consider arguments against nonattainment.” Id. at 37.

In Air Transport Association of America, Inc. v. National
Mediation Board, 663 F.3d 476 (D.C. Cir. 2011), we repeated
this circuit’s approach to the kind of claim that the petitioners
raise here. “Decisionmakers violate the Due Process Clause
and must be disqualified,” we said, “when they act with an
‘unalterably closed mind’ and are ‘unwilling or unable’ to
rationally consider arguments.” Id. at 487 (quoting Ass’n of
Nat’l Advertisers, Inc. v. FTC, 627 F.2d 1151, 1170, 1174
(D.C. Cir. 1979)). “[A]n individual should be disqualified
from rulemaking only when there has been a clear and
convincing showing that the . . . member has an unalterably
closed mind on matters critical to the disposition of the
proceeding.” Id. (quoting C & W Fish Co., Inc. v. Fox, 931
F.2d 1556, 1564 (D.C. Cir. 1991) (internal quotation marks
omitted)). The four arguments advanced by the Texas State
Petitioners are insufficient to make that “clear and convincing”
showing. 28

27
After a video of the speech was discovered, Armendariz
resigned. Id. Soon thereafter, the EPA promulgated the Wise
County nonattainment designation.
28
The Supreme Court has held that States are not “persons”
within the meaning of the Due Process Clause. See South Carolina
v. Katzenbach, 383 U.S. 301, 323–24 (1966); see also Republic of
 82
Our decision in C & W Fish Company establishes that
neither Armendariz’ employment history nor the report he
authored before joining the EPA required his disqualification.
There, we considered the impartiality of an agency
administrator who had previously served as the chairman of a
group advocating for the precise agency policy at issue in the
case, and who after his appointment remarked that there was
“no question” that the policy should be implemented. C & W
Fish Co., 931 F.2d at 1564. Those circumstances, we said, did
“not even approach a ‘clear and convincing showing’ that [the
administrator] had an ‘unalterably closed mind.’ ” Id. at
1565.

The petitioners’ third argument is also unpersuasive.
There is no doubt that Armendariz’ “crucifixion” comments
were offensive. But that does not suffice to make the requisite
showing. The comments described Armendariz’ general
approach to enforcement, but were neither specifically about
the designation process nor specifically targeted at production
from the Barnett Shale. Accordingly, they did not reveal
Armendariz’ views on “matters critical to the disposition of the
proceeding.” Ass’n of Nat’l Advertisers, 627 F.2d at 1170.
And even if they had, they would not alone demonstrate an
unalterably closed mind on the subject. See C & W Fish Co.,
931 F.2d at 1565 (“ ‘We would eviscerate the proper evolution
of policymaking were we to disqualify every administrator
who has opinions on the correct course of his agency’s future

Argentina v. Weltover, Inc., 504 U.S. 607, 619 (1992) (citing
Katzenbach, 383 U.S. at 323–24); Price v. Socialist People’s Libyan
Arab Jamahiriya, 294 F.3d 82, 96 (D.C. Cir. 2002). Although one
circuit has held that counties are protected in some circumstances,
see County of Santa Cruz v. Sebelius, 399 F. App’x 174, 176 (9th
Cir. 2010), we need not consider the issue because we find no
violation here.
 83
actions.’ ” (quoting Ass’n of Nat’l Advertisers, 627 F.2d at
1174)).

Finally, we cannot infer bias from the fact that, in the
opinion of the petitioners, the computer modeling supported an
attainment designation for Wise County. As we held in C &
W Fish Company, “we reject the suggestion that we look to the
adequacy of [an agency official’s] examination of the facts and
issues in order to determine whether he was biased.” 931 F.2d
at 1564. Rather, “[w]hether [the official] weighed the facts
properly is to be examined only in determining if his decision
was arbitrary or capricious.” Id. at 1564–65. And that is an
examination that we separately undertake in section III.F.2,
supra.

For the foregoing reasons, we reject the petitioners’ three
constitutional challenges to the designation of Wise County as
a nonattainment area.

4. The Remaining Challenges

Finally, the Texas State Petitioners argue that we should
vacate the EPA’s Wise County nonattainment designation
because the EPA (1) failed to comply with the Information
Quality Act, (2) failed to promulgate regulations defining the
terms “necessary” and “contribute,” (3) concluded that Wise
County emissions “can” contribute to NAAQS violations when
it was statutorily required to conclude that Wise County “did”
contribute, and (4) failed to give them “fair notice” of the
EPA’s requirements. State & County Br. 46–52. We reject
all four contentions.

First, the Texas State Petitioners urge us to conclude that
the Information Quality Act requires the EPA to use “the best
available science and supporting studies conducted in
accordance with sound and objective scientific practices” in
 84
making NAAQS designations, State & County Br. 46 (citing
Prime Time Int’l Co. v. Vilsack, 599 F.3d 678, 685–86 (D.C.
Cir. 2010)), and that the EPA failed to do so here. But almost
every court that has addressed an Information Quality Act
challenge has held that the statute “creates no legal rights in
any third parties,” Salt Inst. v. Leavitt, 440 F.3d 156, 159 (4th
Cir. 2006); 29 see also Harkonen v. U.S. Dep’t of Justice, No. C
12-629 CW, 2012 WL 6019571, at *11 (N.D. Cal. Dec. 3,
2012) (collecting cases). And this Court has held that the
Information Quality Act is not “an independent measure of
EPA’s NAAQS decision.” Mississippi, 744 F.3d at 1347.
The purpose of the Information Quality Act is to “ensur[e] and
maximize[e] the quality, objectivity, utility, and integrity of
information (including statistical information) disseminated by
Federal agencies” and does not constitute a statutory
mechanism by which the EPA’s conclusions reached while
making its nonattainment determinations can be challenged.
See 44 U.S.C. § 3516 note (emphasis added).

Second, the Texas State Petitioners argue that the EPA
should define the terms “contribute” and “necessary” through
administrative rulemaking in order to rein in the “boundless
override discretion” it uses to “commandeer[]” states to
“enforce its massive regulatory scheme.” See State & County
Br. 48. Our Catawba County holding forecloses this
argument. There, we held that the EPA was “free to adopt a
totality-of-the-circumstances test to implement a statute that
confers broad discretionary authority.” Catawba Cnty., 571
F.3d at 39. Finally, the Texas State Petitioners offer no reason
why the word “necessary,” which the EPA reasonably

29
But see Prime Time, 599 F.3d at 685–86 (affirming dismissal
of Information Quality Act challenge on different grounds without
addressing argument that the statute creates no legal rights in third
parties).
 85
interpreted as authorizing modification of a state’s
recommended designation that does “not meet the statutory
requirements or [was] otherwise inconsistent with the facts or
analysis deemed appropriate by the EPA,” see 2008
Designations Rule, 77 Fed. Reg. at 30,090, must be defined via
rulemaking.

Third, the Texas State Petitioners argue that the EPA
exceeded its authority under the Clean Air Act because it
concluded that Wise County emissions “can” contribute to
NAAQS violations, whereas the Act authorizes a finding that
Wise County “does” so contribute. See State & County Br.
50. This argument is premised on the EPA’s response to a
petition for reconsideration challenging the Wise County
nonattainment designation, to which the EPA responded that
“the Wise County emissions are large enough that they can
contribute to ozone exceedances on certain days.” EPA
Response to Pet. for Reconsideration from Wise Cnty., Office
of the Cnty Judge at 2 (emphasis added). But read in toto, the
EPA’s justification for including Wise County in the Dallas–
Fort Worth nonattainment area was anything but theoretical:

Wise County [h]as 2008 NEI emissions of
11,911 tons of NOx and 17,609 tons of VOC;
there are 60 people per square mile; has a 2010
population of 59,127 with a growth rate of 5.9
percent between 2000 and 2010; total VMT is
969 million. The close proximity of these
comparatively high emissions to violating
monitors indicates that this county should be
included in the nonattainment area. The high
growth in these emissions is due in large part to
growth in emissions from Barnett Shale gas
production development, but also due to
growth in population. Examination of back
 86
trajectories indicates that at times emissions
from Wise County contribute to observed
violations in the area and also to observed
violations that have helped set the DFW area
DV in the past. Source apportionment
modeling for a portion of an ozone season
indicates that emissions from Wise County can
contribute to observed violations in the DFW
nonattainment area. These factors support the
inclusion of Wise County in the nonattainment
area.

Final Dallas–Fort Worth Area Designations at 23. Read in
context, we conclude that the EPA in fact found that Wise
County does contribute to NAAQS violations in the Dallas–
Fort Worth area.

Fourth, the Texas State Petitioners argue the EPA failed to
provide them with “fair notice” of its requirements. Even
assuming the fair notice doctrine applies, cf. Ark. Dep’t of
Human Servs. v. Sebelius, 818 F. Supp. 2d 107, 120–21
(D.D.C. 2011), the EPA did not violate it. The fair notice
doctrine, which is couched in terms of due process, provides
redress only if an agency’s interpretation is “so far from a
reasonable person’s understanding of the regulations that they
could not have fairly informed the regulated party of the
agency’s perspective.” United States v. Chrysler Corp., 158
F.3d 1350, 1354 (D.C. Cir. 1998) (alteration omitted). Here,
the EPA not only provided the 2008 Guidance to aid the states
in making their initial designations, it also provided a
preliminary technical support document to each state before
finalizing any of its proposed modifications to the state’s initial
designations. See, e.g., Preliminary Dallas–Fort Worth Area
Designations. The technical support document, in turn, gave
each state a precise explication of all proposed EPA
 87
modifications as a roadmap to use during the 120-day
comment period. Simply put, the EPA set forth its analysis,
provided an opportunity to rebut its conclusions and ultimately
explained why it had not changed its mind. Accordingly, the
Texas State Petitioners’ fair notice doctrine argument is
meritless.

For the foregoing reasons, the consolidated petitions for
review are denied.

So ordered.