Arizona ex rel Henry Darwin v. USEPA, et al

Filing

FILED OPINION (MARSHA S. BERZON, JAY S. BYBEE and JOHN B. OWENS) Petitioner Phoenix Cement s pending motion to supplement record is denied. DISMISSED IN PART; DENIED; Judge: JSB Authoring, FILED AND ENTERED JUDGMENT. [10380274] [14-73368, 14-73384, 14-73386, 14-73394]

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 1 of 34

FOR PUBLICATION

UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT

STATE OF ARIZONA, EX REL. HENRY
R. DARWIN, DIRECTOR, ARIZONA
DEPARTMENT OF ENVIRONMENTAL
QUALITY,
Petitioner,

No. 14-73368

v.
U.S. ENVIRONMENTAL PROTECTION
AGENCY; SCOTT PRUITT,
Administrator, United States
Environmental Protection Agency,
Respondents.

CALPORTLAND COMPANY,

No. 14-73384
Petitioner,

v.
U.S. ENVIRONMENTAL PROTECTION
AGENCY,
Respondent.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 2 of 34

2

STATE OF ARIZONA EX REL DARWIN V. EPA

PHOENIX CEMENT COMPANY, an
enterprise division of the Salt River
Pima-Maricopa Indian Community,
Petitioner,

No. 14-73386

v.
U.S. ENVIRONMENTAL PROTECTION
AGENCY,
Respondent.

ASARCO LLC,

No. 14-73394
Petitioner,
v.

OPINION

U.S. ENVIRONMENTAL PROTECTION
AGENCY,
Respondent.
On Petition for Review of an Order of the
Environmental Protection Agency
Argued and Submitted June 21, 2016
San Francisco, California
Filed April 3, 2017
Before: Marsha S. Berzon, Jay S. Bybee,
and John B. Owens, Circuit Judges.
Opinion by Judge Bybee

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 3 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

3

SUMMARY*

Environmental Law
The panel dismissed in part and denied in part petitions
for review brought by the State of Arizona and several private
companies objecting to several sections of the United States
Environmental Protection Agency’s most recent Federal
Implementation Plan (“FIP”), promulgated under the Clean
Air Act, to replace certain rejected portions of Arizona’s State
Implementation Plan, concerning how the state intended to
improve air quality in federal parks and forests by reducing
emissions of various pollutants.
The regulatory scheme codified in Section 169A of the
Clean Air Act required each state with emissions impacting
protected federal lands to create a State Implementation Plan.
If the state submitted either a deficient plan or none at all, the
Act required the EPA to promulgate its own plan – called a
FIP – to force compliance with Congress’s mandate.
The panel held that several of petitioners’ objections to
the FIP were not properly before the court because they were
not first presented to the EPA during the notice-and comment
period in 42 U.S.C. § 7607(d)(7)(B). Specifically, the panel
held that the issues petitioners raised for the first time on
appeal were not so “key” that, assuming the D.C. Circuit’s
“key assumption” applied, Nat. Res. Def. Council v. EPA, 755
F.3d 1010, 1023 (D.C. 2014), they must have been
anticipated by the EPA. The panel concluded that petitioners
*

This summary constitutes no part of the opinion of the court. It has
been prepared by court staff for the convenience of the reader.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 4 of 34

4

STATE OF ARIZONA EX REL DARWIN V. EPA

were barred from challenging in this appeal the numerical
“reasonable progress goals” and the EPA’s decision to
jettison the affirmative defense for malfunction. The panel
held that it was barred from reviewing those issues in this
proceeding, and dismissed that portion of the petitions.
The remaining objections that were ripe for consideration
consisted of a series of technical challenges to the emission
controls imposed on a cement kiln and copper smelters.
According due deference, the panel held that the EPA’s
emission-control measures were not arbitrary or capricious
and thus constituted valid agency rulemaking.

COUNSEL
Carroll McGuffey III (argued) and Justin Wong, Troutman
Sanders LLP, Atlanta, Georgia; Paul L. Gale, Troutman
Sanders LLP, Irvine, California; Peter S. Glaser, Troutman
Sanders LLP, Washington, D.C.; James T. Skardon, Assistant
Attorney General, Environmental Enforcement Section; Mark
Brnovich, Attorney General; Office of the Attorney General,
Phoenix, Arizona; for Petitioner State of Arizona.
Albert H. Acken (argued) and Fredric D. Bellamy, Ryley
Carlock & Applewhite, Phoenix, Arizona, for Petitioner
CalPortland Company.
Mark W. DeLaquil (argued) and Andrew M. Grossman,
Baker & Hostetler LLP, Washington, D.C.; George A.
Tsiolis, Englewood, New Jersey; for Petitioner Phoenix
Cement Company.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 5 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

5

Eric L. Hiser (argued) and Trevor J.L. Burggraff, Jorden
Bischoff & Hiser PLC, Scottsdale, Arizona; George A.
Tsiolis, Englewood, New Jersey; for Petitioner ASARCO
LLC.
Samara M. Spence (argued), Trial Attorney; John C. Cruden,
Assistant Attorney General; Environment and Natural
Resources Division, United States Department of Justice,
Washington, D.C.; Matthew C. Marks and M. Lea Anderson,
EPA Office of General Counsel, Washington, D.C.; Charlotte
Withey, EPA Office of Regional Counsel, Region IX, San
Francisco, California; for Respondents.

OPINION
BYBEE, Circuit Judge:
In the late 1970s, Congress declared it a “national goal”
to improve air visibility in federal parks and forests.
42 U.S.C. § 7491(a)(1); see also 40 C.F.R. § 81.400 et seq.
(listing all national parks, national monuments, and
wilderness areas Congress sought to protect). To achieve that
goal, Congress sought to reduce emissions of various
pollutants through a new regulatory scheme codified in
Section 169A of the Clean Air Act (CAA). 42 U.S.C.
§ 7491(b). That scheme required each state with emissions
impacting protected federal lands to create a State
Implementation Plan (SIP) describing how the state intended
to make reasonable progress toward the national visibility
goal. Id. §§ 7410(a), 7491(b)(2). If the state submitted either
a deficient SIP or none at all, the CAA required the EPA to
promulgate its own plan—called a Federal Implementation

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 6 of 34

6

STATE OF ARIZONA EX REL DARWIN V. EPA

Plan (FIP)—to force compliance with Congress’s mandate.
Id. § 7410(c)(1).
This extensive litigation arose when Arizona—a state
containing twelve wilderness areas subject to Section 169A,
see 40 C.F.R. § 81.403—clashed with the EPA over its SIP
submitted in 2011. Although the SIP listed proposals to
manage and reduce emissions from various industrial sources
operated within the state, the EPA determined that Arizona
could do better in improving visibility. The EPA disapproved
certain aspects of Arizona’s SIP and issued its own FIP that
imposed enhanced emission-control measures.
The EPA’s actions spawned a series of legal challenges
from Arizona and several private companies subject to EPA
regulation (collectively, Petitioners). We have rejected most
of those challenges in two prior decisions, concluding that the
EPA acted within its authority when it disapproved portions
of Arizona’s SIP that it deemed problematic. Arizona ex rel.
Darwin v. EPA (Arizona I), 815 F.3d 519, 524 (9th Cir.
2016); Phoenix Cement Co. v. EPA, 647 Fed. App’x 702,
704–05 (9th Cir. Mar. 31, 2016). All that remains before us
now are Petitioners’ objections to several sections of the
EPA’s most recent FIP—those issued to replace certain
rejected portions of Arizona’s SIP—which Petitioners claim
constitute invalid agency action.
We hold that several of Petitioners’ objections to the FIP
are not properly before us because they were not first
presented to the EPA during the notice-and-comment period.
See 42 U.S.C. § 7607(d)(7)(B) (“Only an objection to a rule
or procedure which was raised with reasonable specificity
during the period for public comment (including any public
hearing) may be raised during judicial review.”). As to the

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 7 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

7

remaining objections that are ripe for our consideration, we
conclude that the EPA’s emission-control measures are not
arbitrary or capricious and thus constitute valid agency
rulemaking. Accordingly, we dismiss in part and deny in part
the consolidated petitions for review.1
I. REGULATORY BACKGROUND
In the twentieth century, our nation experienced a
significant degradation of visibility in its most treasured
wilderness areas. See Regional Haze Regulations, 64 Fed.
Reg. 35,714, 35,715 (July 1, 1999). The air pollution,
Congress found, was “primarily” due to the emission of
“[sulfur dioxide], oxides of nitrogen, and particulate matter”
from poorly regulated industrial sources. Id. (citing H.R.
Rep. No. 95-294, at 204 (1977)). To tackle the problem,
Congress adopted Section 169A of the CAA, and the EPA
promulgated implementing regulations to require states to
improve visibility by adopting certain emission controls.
42 U.S.C. § 7491(b)(2). Two categories of such emission
controls are relevant to the issues before us: “best available
retrofit technology” (BART) and what the EPA sometimes
refers to as “reasonable progress” (RP) controls. 42 U.S.C.
§ 7491(b)(2)(A)–(B), (g)(1)–(2); 40 C.F.R. § 51.308(f)(3);
Promulgation of Air Quality Implementation Plans; Arizona;
Regional Haze and Interstate Visibility Transport Federal
Implementation Plan, 79 Fed. Reg. 52,420, 52,447, 52,463
(Sep. 3, 2014) [hereinafter Final FIP].

1

As before, we apologize to the reader for our extensive use of
acronyms and initialisms throughout this opinion and direct her to the end
of the document for a glossary of terms. See Arizona I, 815 F.3d at 525
n.3.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 8 of 34

8

STATE OF ARIZONA EX REL DARWIN V. EPA

A. Best Available Retrofit Technology
BART is a term of art used to describe technology that
can be installed on an industrial source to control its
emissions or make those emissions cleaner. 42 U.S.C.
§ 7491(b)(2)(A), (g)(2). Congress directed BART to be
implemented on older stationary sources that “emit[] any air
pollutant which may reasonably be anticipated to cause or
contribute to any impairment of visibility” in protected
federal lands. Id. § 7491(b)(2)(A); see also 40 C.F.R.
§ 51.301 (defining a “BART-eligible source” as an “existing
stationary facility” built between 1962 and 1977 that “has the
potential to emit 250 tons per year or more of any air
pollutant”). If there are such sources within a regulated state,
the state must not only “submit an implementation plan
containing emission limitations representing BART,”
40 C.F.R. § 51.308(e), but do so for each pollutant those
sources emit, id. § 51.301.
In determining the appropriate BART to install, states
must engage in a cost-benefit analysis by balancing five
factors: “[1] the costs of compliance, [2] the energy and
nonair quality environmental impacts of compliance, [3] any
existing pollution control technology in use at the source,
[4] the remaining useful life of the source, and [5] the degree
of improvement in visibility which may reasonably be
anticipated to result from the use of such technology.”
42 U.S.C. § 7491(g)(2). Because the balancing is source
specific, it could lead a state to adopt stringent BART
regulations, none at all, or something in between. See Final
FIP, 79 Fed. Reg. at 52,447 (recognizing that cost-benefit
analyses could reasonably lead to various degrees of BART
regulations). If the EPA is dissatisfied with a state’s BART
determination, it can issue a FIP containing its own cost-

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 9 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

9

benefit analysis in accordance with Section 169A. 42 U.S.C.
§ 7410(c)(1).
B. Reasonable Progress Controls
Similar to BART, RP controls serve to limit emissions of
harmful pollutants that degrade visibility in protected
wilderness areas. Although RP controls, as such, are not
mentioned in Section 169A, they are an outgrowth of
Congress’s mandate to make “reasonable progress toward
meeting the national [visibility] goal.”
42 U.S.C.
§ 7491(b)(2). The EPA has construed that mandate as
requiring each regulated state to establish reasonable progress
goals (RPGs) that depend on how much of that state’s current
haze would have to be eliminated each year to achieve
natural, pristine conditions by the year 2064. 40 C.F.R.
§ 51.308(d)(1)(i)(B). If a state believes that it is unable to
conform with the year-to-year reduction rate—or the
“glidepath,” as it is known in regulatory nomenclature—its
SIP must explain why achieving the rate is not reasonable,
while its proposed RPGs are. Id. § 51.308(d)(1)(ii).2

2

All RPGs must be expressed in “deciviews,” which quantify the
visibility impairment of a particular region. 40 C.F.R. § 51.308(d)(1). A
“one deciview change in haziness is a small but noticeable change,”
though “visibility changes of less than one deciview are likely to be
perceptible in some cases.” Regional Haze Regulations, 64 Fed. Reg.
35,714, 35,725–26 (July 1, 1999), partially vacated by Am. Corn Growers
Ass’n v. EPA, 291 F.3d 1, 6 (D.C. Cir. 2002). The EPA classifies a 1.0
deciview change from an individual source as “causing” the impairment
of visibility, while a 0.5 deciview change as “contribut[ing]” to the
impairment. Regional Haze Regulations and Guidelines for Best
Available Retrofit Technology (BART) Determinations, 70 Fed. Reg.
39,104, 39,121 (July 6, 2005) [hereinafter Guidelines].

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 10 of 34

10

STATE OF ARIZONA EX REL DARWIN V. EPA

After a state calculates its RPGs, it must determine
whether to implement RP controls to reach those RPGs. See
id. § 51.308(d). In making that determination, the state may
take into account existing control measures—including
BART—as well as the visibility impact of controls
established through other programs under the CAA. See id.
§ 51.308(d)(1). If the existing control measures do not cut it,
the state must consider four factors to determine whether
additional measures are warranted: “[1] the costs of
compliance, [2] the time necessary for compliance, and
[3] the energy and non-air quality environmental impacts of
compliance, and [4] the remaining useful life of any existing
source subject to such requirements.”
42 U.S.C.
§ 7491(g)(1); see also 40 C.F.R. § 51.308(d)(1)(i)(A).
Should a state fail to conduct an accurate balancing, the EPA
can correct the state’s errors in a FIP. 40 C.F.R.
§ 51.308(d)(1)(v); see also 42 U.S.C. § 7410(c)(1).
One characteristic of RP controls is worth emphasizing:
they apply not just to some but to all pollutant-emitting
stationary sources that impede reasonable progress. Take, for
instance, a cement kiln emitting sulfur dioxide built in 1983.
Although the kiln would not be subject to BART regulation
because it was not in existence as of 1977, see 42 U.S.C.
§ 7491(b)(2)(A), it might nonetheless be required to adopt
additional measures because limiting its pollution would help
its home state achieve RPGs. In essence, then, RP controls
fill a regulatory gap that would have existed if the EPA were
not able to regulate BART-ineligible cement kilns, smelters,
and other emission sources threatening to deteriorate
visibility.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 11 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

11

II. PROCEDURAL BACKGROUND
In 2011, Arizona submitted a SIP to the EPA outlining its
RPGs and describing the steps it intended to take to limit air
pollution within its state. See Arizona I, 815 F.3d at 528.
The EPA deemed several portions of Arizona’s SIP
inadequate and issued a FIP to correct what it perceived to be
Arizona’s noncompliance with Section 169A and related
regulations. Id. at 529; Phoenix Cement, 647 Fed. App’x. at
705–06. Because the validity of that FIP is the primary issue
in this action, we first outline the steps the EPA took in
preparing that document.
A. The Proposed FIP
The portions of Arizona’s SIP that the EPA rejected
consisted of “BART control analyses and determinations” for
multiple stationary sources, RPG “analyses and
determinations,” and long-term strategies for “making
reasonable progress.” See Promulgation of Air Quality
Implementation Plans; Arizona; Regional Haze and Interstate
Visibility Transport Federal Implementation Plan; Proposed
Rules, 79 Fed. Reg. 9318, 9320 (Feb. 18, 2014) [hereinafter
Proposed FIP]. Here is what the EPA proposed, in relevant
part, to replace those rejected sections:
1. Reasonable Progress Goals
In a short, narrative section, the EPA proposed a set of
RPGs “consistent with a combination of control measures that
include those in the approved [portions of Arizona’s] SIP as
well as [the Proposed] FIP.” Id. at 9321. The EPA did not
express those RPGs numerically, explaining that although it
“would prefer to quantify these proposed RPGs,” it lacked

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 12 of 34

12

STATE OF ARIZONA EX REL DARWIN V. EPA

“sufficient time and resources” to do so. Id. at 9363. Instead,
the EPA simply noted that “[i]n total, these final and
proposed controls to meet the BART and RP requirements
will result in higher emissions reductions and commensurate
visibility improvements beyond what was in [Arizona]’s
plan.” Id.
2. CalPortland’s Cement Kiln
The EPA next considered whether CalPortland’s cement
kiln should be subject to additional RP controls. Id. at 9351.
Although Arizona had identified that kiln as a source that
might contribute to visibility impairment, it failed to conduct
the requisite four-factor analysis in its SIP. See Approval and
Disapproval of Air Quality State Implementation Plans;
Arizona; Regional Haze and Interstate Transport
Requirements, 78 Fed. Reg. 46,142, 46,168–69, 46,171 (July
30, 2013). The EPA therefore balanced the four factors itself
and proposed that CalPortland install an emissions-control
technology called “selective non-catalytic reduction” (SNCR)
with a “50 percent SNCR control-efficiency.” Proposed FIP,
79 Fed. Reg. at 9354–56. The EPA sought comment,
however, on “whether a different SNCR control efficiency is
appropriate.” Id. at 9355.
3. Copper Smelters
The EPA also discussed limiting emissions from
Petitioner Asarco’s and non-party Freeport-McMoran, Inc.’s
copper smelters. Id. at 9318, 9320. Because both smelters
are subject to BART controls, the EPA considered whether to
impose BART to reduce emissions of nitrogen oxides (an
issue applicable to both smelters on this appeal) as well as
particulate matter and sulfur dioxide (an issue applicable only

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 13 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

13

to Asarco’s smelter). Id. at 9320–21. Since the EPA had
rejected Arizona’s BART analysis, it engaged in its own
balancing of the five factors listed in Section 169A. Id.
Here is what the EPA proposed. As to nitrogen oxides
emissions, the EPA determined that the estimated level of
visibility improvement from additional BART controls was
not worth the cost. Id. The EPA therefore proposed only an
annual emission limit of 40 tons per year (tpy) which it found
to be “consistent with current emissions” of nitrogen oxides
from the smelters. Id. at 9320. Such measures were
presumably necessary to ensure that the copper smelters
would not drastically increase their emissions beyond 40 tpy
after the FIP were finalized.
As to particulate matter emissions, the EPA proposed a
similar scheme to regulate Asarco’s smelter. Id. Because
Asarco had been already implementing BART to control
those emissions in accordance with another program under
the CAA, the EPA determined that it made most sense to set
an emission limit consistent with that program. Id. The EPA
therefore proposed to “incorporate emission limitations and
associated compliance requirements from the National
Emission Standard for Hazardous Air Pollutants (NESHAP)
for Primary Copper Smelting at 40 CFR Part 63, Subpart
QQQ [hereinafter Subpart QQQ].” Id.
Finally, the EPA addressed sulfur dioxide emissions from
Asarco’s smelter. The EPA conducted the required fivefactor analysis, considered available options, and proposed a
technology—a “double contact acid plant”—that Asarco had
been already using. Id. at 9343–44. To determine what
emission limit to impose, the EPA relied on Asarco’s written
representation that its technology could capture 99.8 percent

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 14 of 34

14

STATE OF ARIZONA EX REL DARWIN V. EPA

of the sulfur dioxide vented to it. Id. at 9344. The EPA thus
proposed “a control efficiency of about 99.8 percent on a 30day rolling average.” Id. at 9343.
4. Affirmative Defense for Malfunctions
Finally, the EPA included an “affirmative defense for
malfunctions” in the proposed regulations appended to the
narrative section of the Proposed FIP. See id. at 9367, 9373
(proposing 40 C.F.R. § 52.145(k)(11), a regulation titled
“Affirmative defense for malfunctions” that incorporated
certain provisions of Arizona’s Administrative Code excusing
excess emissions due to malfunctions at CalPortland’s kiln);
id. at 9367, 9375 (proposing 40 C.F.R. § 52.145(l)(14), a
similar regulation excusing excess emissions due to
malfunctions at Asarco’s smelter). Nowhere in its Proposed
FIP did the EPA mention that it was considering deleting this
affirmative defense in the Final FIP.
B. The EPA’s Final FIP
After notice and comment, the EPA promulgated its Final
FIP that differed from the earlier proposal in several
important respects.
1. Reasonable Progress Goals
In response to several comments criticizing the EPA’s
decision to express RPGs qualitatively rather than
quantitatively, the EPA reversed its course and promulgated
a set of numerical goals. Final FIP, 79 Fed. Reg. at
52,468–71. Those new RPGs were then compared to the
RPGs outlined in Arizona’s rejected SIP to devise something
called a “FIP effect,” or the increase in visibility between the

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 15 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

15

effects of controls mandated by Arizona’s SIP and those
promulgated by EPA’s Final FIP. Id. at 52,469–70. The FIP
effects at the various protected areas in Arizona generally
came in at about 0.10 deciviews. Id.
2. CalPortland’s Cement Kiln
During the notice-and-comment period, CalPortland
objected to the Proposed FIP’s imposition of a 50 percent
control efficiency using SNCR and asserted that only
“35 percent control efficiency may be achievable.” Id. at
52,462. The EPA accepted CalPortland’s argument and
agreed that “35 percent reflects an appropriate estimate of the
degree of control achievable with SNCR at [CalPortland’s
kiln].” Id at 52,463. It therefore revised its proposed
regulation and imposed a 35 percent control efficiency.
3. Copper Smelters
With a few tweaks, the Final FIP largely repeated the
BART measures set forth in the Proposed FIP. As to nitrogen
oxides emissions, the EPA rejected Arizona’s contention that
facilities emitting fewer than 40 tpy—like Asarco’s and
Freeport-McMoran’s copper smelters—should be excluded
from mandatory BART regulations under 40 C.F.R.
§ 51.308(e)(1)(ii)(C). See id. at 52,445. The EPA explained
that even though the smelters were not currently emitting
more than 40 tpy, there were no “federally enforceable
physical or operational limitations that would” cap its ability
to emit more than 40 tpy in the future. Id. at 52,445–46. If
the rule were otherwise, the EPA pointed out, “emissions
could increase to a level where additional controls would be
warranted for BART, but no mechanism would exist to
require such controls.” Id. at 52,446.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 16 of 34

16

STATE OF ARIZONA EX REL DARWIN V. EPA

As to particulate matter emissions, the EPA finalized its
proposal to import limitations from Subpart QQQ. Id. at
52,447–48. Although Asarco asserted that its particulate
matter controls were not based on the requirements of
Subpart QQQ, the EPA quoted Asarco’s direct statements to
the contrary. Id. “Given that [Asarco] relied on the Subpart
QQQ requirements,” the EPA reasoned, inclusion of “these
requirements in the FIP” was appropriate. Id. at 52,448.
The EPA was equally unpersuaded by Asarco and
Arizona’s arguments against the proposed sulfur dioxide
limitations. The concerns “over the technical feasibility” of
the proposed double contact acid plant were unsubstantiated,
the EPA determined, because such a plant “[was] already in
use at [Asarco’s] Smelter” and was therefore “presumed to be
an applicable technology.” Id. at 52,441. The EPA did,
however, revise the time period over which the control
efficiency rate would be calculated, crediting Asarco’s
assertion that the 99.8% figure it had provided to the EPA
could be achieved using only a 365-day rolling average rather
than the 30-day average promulgated in the Proposed FIP. Id.
at 52,443. Switching to the 365-day rolling average, the EPA
continued, would also alleviate Asarco’s worry that the
99.8% control efficiency figure would not allow sufficient
breathing room for “startup, shutdown, and malfunction,” or
SSM, occurrences. Id. at 52,440, 52,443.
4. Affirmative Defense for Malfunctions
The Final FIP eliminated the affirmative defense for
“excess emissions due to malfunctions” that the EPA had
inserted into the Proposed FIP. Id. at 52,427. The EPA
recognized that it did “not receive any comments” about the
inclusion of an affirmative defense, but explained that a

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 17 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

17

recent D.C. Circuit opinion in Natural Resources Defense
Council v. EPA, 749 F.3d 1055 (D.C. Cir. 2014), “ma[de] it
clear that the CAA does not authorize promulgation of [an
affirmative defense provision] by EPA.” Id. The EPA noted
that the “court did not address whether such an affirmative
defense provision could be properly included in a SIP,” but
nonetheless maintained that the “logic of the court’s
decision” foreclosed the EPA’s inclusion of the defense in its
Final FIP. Id.
After the EPA promulgated the Final FIP, Arizona,
Phoenix Cement Company, CalPortland, and Asarco filed this
consolidated appeal.
III. STANDARD OF REVIEW
We may invalidate the Final FIP if it is “arbitrary,
capricious, an abuse of discretion, or otherwise not in
accordance with law.” 42 U.S.C. § 7607(d)(9). This familiar
standard is, of course, highly deferential. A regulation should
be deemed arbitrary and capricious in only limited
circumstances, such as when:
[T]he agency has relied on factors which
Congress has not intended it to consider,
entirely failed to consider an important aspect
of the problem, offered an explanation for its
decision that runs counter to the evidence
before the agency, or is so implausible that it
could not be ascribed to a difference in view
or the product of agency expertise.
Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut.
Auto. Ins. Co., 463 U.S. 29, 43 (1983). Courts must be

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 18 of 34

18

STATE OF ARIZONA EX REL DARWIN V. EPA

particularly careful in reviewing questions involving “a high
level of technical expertise” because such matters are
normally best left to the experience and judgment of the
agency. Nat’l Parks Conservation Ass’n v. EPA, 788 F.3d
1134, 1141 (9th Cir. 2015) (citation omitted); see also
Baltimore Gas & Elec. Co. v. Nat. Res. Def. Council, Inc.,
462 U.S. 87, 103 (1983) (“When examining [an agency’s]
scientific determination, as opposed to simple findings of
fact, a reviewing court must generally be at its most
deferential.”).
When the challenge to a regulation is based on an alleged
procedural error, our standard of review under the CAA is
even stricter. Not only must the petitioners in such
circumstances demonstrate that the rule was arbitrary or
capricious, but also that the alleged procedural errors “were
so serious and related to matters of such central relevance that
there is a substantial likelihood that the rule would have been
significantly changed if such errors had not been made.”
42 U.S.C. § 7607(d)(8).
IV. DISCUSSION
Petitioners3 make a wide variety of contentions that
challenge four aspects of the Final FIP: (1) the promulgation
of numerical RPGs; (2) the imposition of SNCR to control
CalPortland’s cement kiln; (3) the emission limits regulating
Asarco’s and Freeport-McMoran’s copper smelters; and
(4) the decision to eliminate the affirmative defense for

3

Although we recognize that not all Petitioners make the same
contentions, we refer to “Petitioners” throughout this section
indiscriminately for simplicity’s sake.
We identify individual petitioners
as necessary.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 19 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

19

malfunctions.4 As we discuss below, all arguments falling
into the first and last of those categories are not properly
before us because they have not been presented to the EPA.
And as to the remaining contentions, nothing about the EPA’s
regulations of the cement kilns and copper smelters was
arbitrary or capricious.
A. Judicial Challenges to a FIP Must be First Exhausted
Before the EPA
Using perhaps the most unambiguous language possible,
Congress has limited judicial review of EPA rules by
requiring all potential petitioners to present their challenges
first to the EPA itself. 42 U.S.C. § 7607(d)(7)(B). “Only an
objection to a rule or procedure which was raised with
reasonable specificity during the period for public comment,”
Congress provided, “may be raised during judicial review.”
Id. “[I]f the grounds for such an objection arose after the
period for public comment (but within the time specified for
judicial review), and if such objection is of central relevance
to the outcome of the rule, [the EPA] shall convene a
proceeding for reconsideration of the rule.” Id. And if the
EPA declines to do so, the potential petitioner may “seek
review of such refusal in the United States court of appeals
for the appropriate circuit.” Id.

4

Petitioners addressed a fifth issue, demonstration requirements that
the EPA imposed in an effort to find the optimal SNCR efficiency rate for
regulating the cement kilns. See, e.g., Final SIP, 79 Fed. Reg. at 52,464.
As all parties now recognize, these “demonstration project claims” are
now moot because the EPA has revised the Final FIP to replace the
demonstration requirements with certain record-keeping and reporting
obligations. See Joint Status Report at 3, Arizona v. EPA, No. 14-73368
(Nov. 7, 2016), Dkt. No. 105; Citation of Supplemental Authorities at 1–2,
Arizona v. EPA, No. 14-73368 (Dec. 14, 2016), Dkt. No. 106.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 20 of 34

20

STATE OF ARIZONA EX REL DARWIN V. EPA

This congressional framework promotes regulatory
efficiency by ensuring that the EPA—as the entity with
greatest expertise in environmental matters— takes the first
shot at resolving all issues with its regulations. The potential
efficiency gains are not merely theoretical, and one need not
look further than this case to see why. One of the original
issues the parties briefed and argued before us concerned
certain “demonstration requirements” the EPA imposed on
Phoenix Cement and CalPortland. See supra note 4.
Although the issue was both difficult and technical,
Petitioners had not raised it to the EPA during the notice-andcomment period because the demonstration requirements
were not mentioned anywhere in the Proposed FIP.
Fortunately, Petitioners requested the EPA to reconsider the
rule, see 42 U.S.C. § 7607(d)(7)(B), and the EPA agreed to
replace it with mutually acceptable regulations. See supra,
note 4. By following the exhaustion rules and presenting
their grievances to the entity most capable of resolving them,
Petitioners avoided wasting both our resources and their own.
But the same cannot be said of two other claims
Petitioners make on this appeal. For the first of those claims,
Petitioners argue that the EPA’s RPGs are arbitrary and
capricious because Arizona residents will directly or
indirectly have to pay hundreds of millions of dollars to
improve air visibility by 0.10 deciviews, an amount of
improvement Petitioners dub trivial. In their second claim,
they assert that the EPA’s decision to eliminate the
affirmative defense for malfunctions was both procedurally
and substantively invalid. We may not consider either of
those claims under § 7607(d)(7)(B). Nothing in the record
indicates that Petitioners have raised their arguments before
the EPA at all, let alone with the requisite “reasonable
specificity.” See 42 U.S.C. § 7607(d)(7)(B).

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 21 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

21

Petitioners concede as much with respect to their claim
seeking to reinstate the affirmative defense for malfunctions.
Indeed, Petitioners must concede that they could not have
protested in their comments on the Proposed FIP the EPA’s
decision to eliminate the affirmative defense provision from
its proposal, because the EPA removed it spontaneously. The
EPA had, in the Proposed FIP, proposed including the
affirmative defense, not removing it, so it is no wonder that
Petitioners did not file comments objecting to its removal.
Nevertheless, the CAA is clear: the Petitioners’ remedy is to
explain to the EPA “that it was impracticable to raise such an
objection” and ask the EPA to “convene a proceeding for
reconsideration of the rule.” 42 U.S.C. § 7607(d)(7)(B).
With respect to Petitioners’ new-found objections to the
RPGs, Petitioners claim that certain statements during the
notice-and-comment period put the EPA on notice of their
arguments. We find that assertion not only wrong but close
to frivolous. It was simply not possible to point out any
disparities between the costs and benefits of the Final FIP’s
numerical RPGs during the notice-and-comment period
because no numerical RPGs were in existence at that time.
Commentators never made any cost/benefit argument but
instead urged the EPA to issue numerical RPGs in the first
place. See Final FIP, 79 Fed. Reg. at 52,468–71. That
commentary—on which Petitioners now rely to avoid the
exhaustion requirement—could not have possibly apprised
the EPA of the claim Petitioners now make.
Faced with this reality, Petitioners make two
arguments—one doctrinal, the other policy-based—in an
attempt to save their unexhausted claims from dismissal.
Neither of the arguments, however, has merit. First,
Petitioners rely on a line of cases from the D.C. Circuit

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 22 of 34

22

STATE OF ARIZONA EX REL DARWIN V. EPA

finding an exception to § 7607(d)(7)(B) when a new
argument challenges “key assumptions” underlying an EPA
rule. This exception is usually stated as follows:
[E]ven if a party may be deemed not to have
raised a particular argument before the
agency, “EPA ‘retains a duty to examine key
assumptions as part of its affirmative burden
of promulgating and explaining a
nonarbitrary, non-capricious rule’” and
therefore “‘EPA must justify that assumption
even if no one objects to it during the
comment period.’”
Nat. Res. Def. Council v. EPA, 755 F.3d 1010, 1023 (D.C.
Cir. 2014) (citation omitted). The asserted duty to examine
“key assumptions” has no textual origin. It appears to be
based on a statement by the D.C. Circuit in a case that does
not mention § 7607(d)(7)(B). See Nat’l Lime Ass’n v. EPA,
627 F.2d 416, 433 (D.C. Cir. 1980) (“[W]e think an initial
burden of promulgating and explaining a non-arbitrary, noncapricious rule rests with the Agency . . . .”); see also Small
Refiner Lead Phase-Down Task Force v. EPA, 705 F.2d 506,
534 (D.C. Cir. 1983) (establishing for what appears to be the
first time the exception to § 7607(d)(7)(B) based on the
language in National Lime).5

5

Although we have not expressly adopted the D.C. Circuit’s
exception to § 7607(d)(7)(B), we have, in construing the National
Environmental Policy Act of 1969, 42 U.S.C. § 4321 et seq., cited the
Supreme Court’s observation that “‘[a regulation’s] flaws might be so
obvious that there is no need for a commentator to point them out
specifically in order to preserve its ability to challenge a proposed
action.’” Or. Nat. Desert Ass’n v. Jewell, 840 F.3d 562, 573 (9th Cir.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 23 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

23

We need not decide today whether to adopt the D.C.
Circuit’s exception to § 7607(d)(7)(B) exhaustion
requirement. For even if the exception were available, we
would hold that striking a particular balance between costs
and benefits of numerical RPGs cannot be a “key
assumption” underlying the Final FIP. As the D.C. Circuit
has explained, the primary purpose of § 7607(d)(7)(B) is to
ensure that the EPA has an opportunity to consider a
challenge to its regulations without being “sandbag[ged]” by
litigants who wait until litigation before a court of appeals to
make their arguments. Okla. Dep’t of Envtl. Quality v. EPA,
740 F.3d 185, 192 (D.C. Cir. 2014). Whether a particular
cost/benefit analysis results in an appropriate RPG concerns
not a “key assumption” but a contest to the EPA’s specific
application of statutory precepts to a concrete situation. We
hold that the issues Petitioners have raised for the first time
on this appeal are not so “key” that, assuming the D.C.
Circuit’s “key assumption” exception applies, they must have
been anticipated by the EPA.
Petitioners next argue that it would be “absurd” if their
claims, which are partially based on the EPA’s failure to
provide notice of the numerical RPGs, were barred under
§ 7607(d)(7)(B). If that were so, Petitioners continue, the
EPA could promulgate a FIP without the requisite notice-andcomment period and then regulate unsuspecting parties while
they scramble to file a request for reconsideration with the
EPA. See 42 U.S.C. § 7607(d)(7)(B) (“[R]econsideration
shall not postpone the effectiveness of the rule.”). And
although the “effectiveness of the rule may be stayed [for
three months] during such reconsideration,” id., Petitioners
2016) (quoting Dep’t of Transp. v. Pub. Citizen, 541 U.S. 752, 765
(2004)).

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 24 of 34

24

STATE OF ARIZONA EX REL DARWIN V. EPA

claim that it could be years until the EPA acts on a
reconsideration request.
Petitioners’ pleas have some appeal. However, we cannot
ignore the unambiguous language of § 7607(d)(7)(B) simply
because it might sometimes lead to what we regard as suboptimal results. Indeed, it is far from clear that Congress
intended to carve out a special exception for claims asserting
a deficiency in the notice-and-comment process. Congress
could have just as easily determined that the benefits from the
exhaustion requirement, combined with the right to ask a
court for a three-month stay during the administrative
reconsideration period, would outweigh the harm Petitioners
identify. Such a conclusion would make sense, as the noticeand-comment process itself can lead to changes from the
proposed FIP as to which there was no opportunity to
comment. The EPA is not limited to specific changes
suggested by commenters, and one commenter does not
necessarily have the opportunity (or the resources) to
comment on every other commenter’s proposals. Were the
EPA required to put out for notice-and-comment every
significant change from the proposed FIP to the final FIP, the
notice-and-comment process could go on forever.
The words of § 7607(d)(7)(B) must be therefore read
literally, as our sister circuits have concluded. See EME
Homer City Generation, L.P. v. EPA, 795 F.3d 118, 137
(D.C. Cir. 2015) (“Because [petitioners’] argument is an
objection to the notice and comment process itself, petitioners
obviously did not and could not have raised it during the
period for public comment. Under Subsection 7607(d)(7)(B),
however, the only appropriate path for petitioners to raise this
issue is through an initial petition for reconsideration to
EPA.”); North Dakota v. EPA, 730 F.3d 750, 770 (8th Cir.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 25 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

25

2013) (same); Oklahoma v. EPA, 723 F.3d 1201, 1214 (10th
Cir. 2013) (same).
In sum, we conclude that Petitioners are barred from
challenging in this appeal the numerical RPGs and the EPA’s
decision to jettison the affirmative defense for malfunction.
To overcome this procedural bar, Petitioners must first
request the EPA to reconsider its regulations under
§ 7607(d)(7)(B). We are barred from reviewing these issues
in this proceeding and dismiss that portion of the petition.
B. EPA Regulation of the Cement Kiln and Copper Smelters
Is Valid
Petitioners’ remaining contentions are a series of
technical challenges to the emission controls imposed on the
cement kiln and copper smelters involved in this action.
According due deference to the EPA, we see nothing arbitrary
or capricious in its final regulations.
1. CalPortland’s Cement Kiln
Petitioners claim that the EPA overstepped its authority
in imposing the SNCR controls on CalPortland’s cement kiln.
The EPA’s primary error, Petitioners assert, was to
“deemphasize” potential improvement in visibility when
considering whether to implement an RP control using the
four-factor balancing test of Section 169A. See 42 U.S.C.
§ 7491(g)(1). If the EPA had not committed that alleged
error, Petitioners continue, it would have realized that
adopting SNCR technology would only marginally advance
RPGs. Petitioners therefore urge us to deem the EPA’s
actions arbitrary or capricious.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 26 of 34

26

STATE OF ARIZONA EX REL DARWIN V. EPA

We decline to do so. As required under Section 169A, the
EPA considered all four factors before promulgating the FIP:
(1) costs of compliance; (2) time necessary for compliance;
(3) energy and non-air quality environmental impacts of
compliance; and (4) the remaining useful life of any
potentially affected sources. Final FIP, 79 Fed. Reg. at
52,466; see 42 U.S.C. § 7491(b)(2). Although “visibility
improvement” is not among the four factors, the EPA took it
into account anyway but did not weigh it as heavily as it
would have in a BART analysis. Final FIP, 79 Fed. Reg. at
52,466. Although Petitioners may not like that the EPA
accorded some factors greater weight than others, it is simply
not our place to substitute our judgment for that of the EPA.
See Cent. Ariz. Water Conservation Dist. v. EPA, 990 F.2d
1531, 1543 (9th Cir. 1993) (“Petitioners’ essential argument
does not claim that EPA failed to consider the relevant
factors, but instead contends that EPA erred in its
consideration of those factors. This court is not to substitute
Petitioners’ judgment, or its own, for that of EPA, as long as
the agency’s interpretation is reasonable.”). That is
especially so here, where the CAA calls for a functional
balancing of technical and difficult considerations. See New
York v. Reilly, 969 F.2d 1147, 1150 (D.C. Cir. 1992)
(“Because Congress did not assign the specific weight the
[EPA] should accord each of these factors, the [EPA] is free
to exercise [its] discretion in this area.”).
Equally dubious is Petitioners’ contention that SNCR
technology is not necessary to achieve RPGs because it
would purportedly improve visibility by only 0.004
deciviews. But that is true only if we also take into account
improvements from numerous other sources located in the
same geographic area. See 79 Fed. Reg. at 52,469. If the
impact of the SNCR technology at the CalPortland cement

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 27 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

27

kiln were measured according to the EPA’s model that
disregards other polluting sources, the increase in visibility
would be far more substantial, measuring at 0.59 deciviews.6
Id. at 52,465. We do not view the EPA’s use of its model to
calculate the benefits of SNCR as arbitrary or capricious but
instead regard it as a permissible exercise of agency
discretion within the proper bounds of the CAA. See North
Dakota, 730 F.3d at 766 (reaching the same result and
crediting the EPA’s assertion that the use of the Petitioners’
model would “rarely if ever demonstrate that emissions
reductions at a single source will have an appreciable effect
on incremental visibility improvement in a given area”).
Accordingly, we conclude that the EPA acted within its
authority in promulgating the SNCR requirement and reject
all of Petitioners’ contentions to the contrary.
2. Copper Smelters
Petitioners’ final series of challenges concerns the EPA’s
regulation of Asarco’s and Freeport-McMoran’s copper
smelters, which, unlike CalPortland’s cement kiln, are subject
to BART. Final FIP, 79 Fed. Reg. at 52,423. Although the
EPA declined to impose additional control measures on the
smelters after conducting the requisite five-factor analysis,
see 42 U.S.C. § 7491(g)(2), it did promulgate emission limits
consistent with the facilities’ existing controls as to nitrogen

6

An analogy the EPA provides may be useful to understand these
concepts. Just as removing one voice from a room containing a few
people would reduce noise to a larger degree than if the room were
crowded, limiting emissions from the only polluting source in the area
would improve visibility to a larger degree than if there were numerous
polluting sources around.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 28 of 34

28

STATE OF ARIZONA EX REL DARWIN V. EPA

oxides, particulate matter, and sulfur dioxides. Petitioners
now claim that those emission limits are arbitrary or
capricious.
a. Nitrogen Oxides Emissions
The EPA imposed a limit of 40 tpy on the nitrogen oxides
emissions from each smelter, which it deemed to be well
above what the smelters currently emit. Final FIP, 79 Fed.
Reg. at 52,424. Petitioners challenge that limitation on two
grounds. They first argue that the 40 tpy limit will not
“eliminat[e] or reduc[e] any [visibility] impairment,”
42 U.S.C. § 7491(b)(2)(A), because the smelters are already
emitting fewer than 40 tpy. And in any event, Petitioners
continue, the smelters should not even be subject to BART
controls and limitations because the smelters’ current
nitrogen oxides emissions are far below the 0.5 deciview
threshold the EPA uses to determine whether BART controls
are necessary. See Proposed FIP, 79 Fed. Reg. at 9327;
Guidelines, 70 Fed. Reg. at 39,161–62. We reject both of
these contentions and regard the 40 tpy limitation as a valid
exercise of the EPA’s regulatory authority.
Petitioners’ first challenge is also their weakest. While it
is true that the smelters currently emit fewer than 40 tpy of
nitrogen oxides, nothing prevents them from exceeding that
threshold in the future. Given that reality, the EPA acted well
within its authority to impose a cap of 40 tpy to ensure that
the smelters live up to their representations of keeping their
nitrogen oxides emissions low. Not only are such limitations
a prudent way to limit visibility impairments in the long run,
see 42 U.S.C. § 7491(b)(2)(A), they are also required under
the EPA’s current rules that are not being challenged here.
Those rules require implementation plans to impose

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 29 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

29

“emission limitations” on all “BART-eligible sources” that
have “the potential to emit” over 40 tpy of certain pollutants.
40 C.F.R. § 51.308(e), (e)(1)(ii)(C). Because the smelters
here are indisputably BART-eligible sources with “the
potential to emit” over 40 tpy of nitrogen oxides, the EPA
properly imposed what it deemed to be a reasonable emission
limit. We decline to second-guess the EPA’s judgment here.
Petitioners’ second argument—that because the visibility
impact of the nitrogen oxides emissions is minuscule, the
smelters are not subject to the BART requirements—fares no
better. True enough, BART is required for only stationary
sources that “cause” (defined as 1.0 deciviews of impact) or
“contribute to” (defined as 0.5 deciviews of impact) visibility
impairment. 70 Fed. Reg. 39,104, 39,121; see also 42 U.S.C.
§ 7491(b)(2)(A); 40 C.F.R. § 51.308(e). But that
determination is made on a source-by-source, not on a
pollutant-by-pollutant, basis. See 40 C.F.R. § 51.308(e). If
the rule were otherwise, a source causing significant visibility
problems (say, over 1.0 deciviews) would nonetheless escape
BART regulation as long as each individual pollutant it emits
impaired visibility by no more than 0.5 deciviews. To avoid
that result, the EPA properly imposed emission limits on each
pollutant the smelters emit, even if the harm from those
individual pollutants is relatively negligible. With respect to
the nitrogen oxides emission, the EPA imposed a 40 tpy limit
because nitrogen oxides contribute to the overall visual
impairment attributable to the copper smelter, even though
nitrogen oxides, considered alone, do not contribute in excess
of 0.5 deciviews of visual impairment. The EPA’s actions
were neither arbitrary nor capricious, but represented a
reasonable interpretation of Section 169A and its
implementing regulations.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 30 of 34

30

STATE OF ARIZONA EX REL DARWIN V. EPA
b. Particulate Matter Emissions

Petitioners next challenge the EPA’s limitation on
particulate matter emissions from Asarco’s copper smelter.
Although the EPA determined that no additional controls
were warranted to reduce the emission of particulate matter,
it did require Asarco’s smelter to adhere to the emission
limits set forth in Subpart QQQ—a separate program under
the CAA not at issue here. Final FIP, 79 Fed. Reg. at
52,447–48; see 40 C.F.R. pt. 63, Subpart QQQ. That
approach was particularly appropriate, the EPA reasoned,
because Asarco itself relied on the Subpart QQQ limits in its
BART analysis. See Final FIP, 79 Fed. Reg. at 52,447.
Petitioners now claim that it was wrong for the EPA to
incorporate the unrelated requirements of Subpart QQQ into
its Final FIP.
We again disagree that the EPA’s approach was anything
but reasonable. As already discussed, the EPA has an
obligation to set emission limits for each source subject to
BART regulation. See 40 C.F.R. § 51.308(e). In carrying
out that duty, the EPA did not pick an emission limit at
random but incorporated the requirements of Subpart QQQ to
match Asarco’s own BART analysis. See Final FIP, 79 Fed.
Reg. at 52,447–48. Since Asarco itself relied on those
requirements, the emission limit the EPA imposed was both
feasible and likely to improve visibility. Neither Section
169A nor the implementing regulations require anything
more of the EPA. See 42 U.S.C. § 7491(b)(2)(A); 40 C.F.R.
§ 51.308(e).

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 31 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

31

c. Sulfur Dioxide Emissions
Finally, Petitioners object to the EPA’s imposition of a
99.8 percent control efficiency rate on sulfur dioxide
emissions from Asarco’s smelter. Petitioners claim that the
99.8 percent rate is unsupported by evidence, technically
infeasible, and arbitrarily more stringent than that imposed on
a competitor’s smelter. We are not persuaded.
To begin with, the 99.8 percent figure has ample support
in the record: after all, it was based on the information
Asarco itself provided to the EPA. In a letter responding to
the EPA’s request for information, Asarco described the
implementation of its “new double-contact acid plant” and
stated, “[a]ccording to [its] calculations,” that the acid plant
“recovers 99.81 percent of the SO(2) emissions that are
vented to it.” Contrary to Asarco’s assertion that the 99.8
percent figure was a rough engineering estimate for a
proposed retrofit project that used only potential emissions as
data points, the letter unequivocally describes an existing
double-contact acid plant that currently recovers 99.8 percent
of sulfur dioxide emissions. It was far from arbitrary and
capricious for the EPA to rely on Asarco’s own
representation in the Final FIP.7
Next, Asarco brings a series of highly technical
challenges in an attempt to show that it is infeasible to
7

Asarco also argues that the EPA should not have relied on the 1984
Review of New Source Performance Standards for Primary Copper
Smelters to further support its 99.8 percent limitation. Because the
argument was not presented to the EPA, we will not consider it here. See
supra, Section IV.A. And even if the argument were presented to the
EPA, we would still decline to deem the EPA’s reliance on Asarco’s own
representations arbitrary or capricious.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 32 of 34

32

STATE OF ARIZONA EX REL DARWIN V. EPA

achieve 99.8 percent efficiency on a 365-day rolling average.
Asarco argues that the 365-day average would practically
amount to a minimum efficiency rate that must be achieved
at all times; that it is difficult to demonstrate compliance with
the 99.8 percent requirement due to the limitations of sulfur
dioxide measuring devices; and that the EPA did not properly
take into account the inevitability of startup and shutdown
emissions. But all these contentions fail for the same reason:
Asarco’s own data shows that it is feasible to achieve 99.8
percent efficiency over 365 days; indeed, Asarco had urged
the EPA to modify the 99.8 estimate from a 30-day rolling
average to a 365-day rolling average, which the EPA did.
That unavoidable fact, coupled with the EPA’s independent
findings that it was both possible to achieve and measure 99.8
percent efficiency despite any emissions during the startup
and shutdown periods, Final FIP, 79 Fed. Reg. at 52,424,
52,443, gives us enough reason not to second-guess the EPA.
It is precisely in such circumstances, “where the issue in
question is highly scientific and the [EPA] has unique
expertise,” that we “give substantial deference to the [EPA’s]
judgment.” Nat’l Wildlife Fed. v. U.S. Army Corps of Eng’rs,
384 F.3d 1163, 1177–78 (9th Cir. 2004).
Finally, Asarco complains that the EPA stuck its smelter
with a 99.8 percent rate while imposing a more attainable
99.7 percent rate on Freeport-McMoran, Asarco’s corporate
rival. But the EPA treated both smelters exactly alike by
relying on the data from each source in establishing the
emission limits. See Proposed FIP, 79 Fed. Reg. at 9348.
Because the EPA “articulated a rational connection between
the facts found and the conclusions made,” Pac. Coast Fed’n
of Fishermen’s Ass’ns. v. U.S. Bureau of Reclamation,
426 F.3d 1082, 1090 (9th Cir. 2005), we defer to the EPA’s
decision to treat the two sources differently.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 33 of 34

STATE OF ARIZONA EX REL DARWIN V. EPA

33

In sum, the EPA’s actions with regard to Asarco’s and
Freeport-McMoran’s copper smelters were neither arbitrary
nor capricious, but reasoned, deliberate, and sensitive to data.
We therefore decline to invalidate the emission limitations
imposed on those sources.
V. CONCLUSION
Some of the challenges to the FIP are not ripe for our
review; others lack merit. Accordingly, we dismiss in part
and deny in part the consolidated petitions.8
PETITIONS DISMISSED IN PART AND DENIED IN
PART.

8

Petitioner Phoenix Cement’s pending motion to supplement record
is denied.

Case: 14-73368, 04/03/2017, ID: 10380274, DktEntry: 107-1, Page 34 of 34

34

STATE OF ARIZONA EX REL DARWIN V. EPA
GLOSSARY OF ACRONYMS AND INITIALISMS
CAA—Clean Air Act
BART—Best Available Retrofit Technology
EPA—Environmental Protection Agency
FIP—Federal Implementation Plan
RP—Reasonable Progress
RPGs—Reasonable Progress Goals
SIP—State Implementation Plan
tpy—tons per year