Nat Resrc Def Cncl v. EPA

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued November 20, 2008 Decided July 10, 2009

No. 06-1045

NATURAL RESOURCES DEFENSE COUNCIL,
PETITIONER

v.

ENVIRONMENTAL PROTECTION AGENCY,
RESPONDENT

NATIONAL PETROCHEMICAL & REFINERS ASSOCIATION, ET
AL.,
INTERVENORS

Consolidated with 06-1046, 06-1047, 06-1214, 07-1311

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

David S. Baron argued the cause and filed the briefs for
petitioner Natural Resources Defense Council.

Anne Milgram, Attorney General, Attorney General’s
Office of the State of New Jersey, and Maurice A. Griffin,
Deputy Attorney General, were on the briefs for petitioner State
of New Jersey.
 2

Kevin P. Auerbacher, Assistant Attorney General, entered an
appearance.

Frank S. Craig III, John B. King, Steven J. Levine, and
Patrick O'Hara were on the briefs for petitioners The Chamber
of Greater Baton Rouge, et al. Geraldine E. Edens entered an
appearance.

Andrew Cuomo, Attorney General, Attorney General’s
Office of the State of New York, Morgan A. Costello and
Michael J. Myers, Assistant Attorneys General, Richard
Blumenthal, Attorney General, Attorney General's Office of the
State of Connecticut, and Kimberly P. Massicotte, Assistant
Attorney General, were on the briefs for intervenors the States
of New York and Connecticut in support of petitioner.

Charles H. Knauss and Sandra P. Franco were on the briefs
for intervenor National Petrochemical & Refiners Association
in support of petitioner Natural Resources Defense Council.

Brian H. Lynk, Attorney, U.S. Department of Justice,
argued the cause for respondents. With him on the brief were
John C. Cruden, Deputy Assistant Attorney General, and Diane
E. McConkey, Counsel, U.S. Environmental Protection Agency.

Norman W. Fichthorn and Lucinda M. Langworthy were on
the brief for intervenor Utility Air Regulatory Group in support
of respondents. Allison D. Wood entered an appearance.

Charles H. Knauss was on the brief for intervenors National
Petrochemical & Refiners Association and American Petroleum
Institute in support of respondents. Martha E. Cox and Stacy R.
Linden entered appearances.
 3

Before: GINSBURG, HENDERSON and ROGERS, Circuit
Judges.

Opinion for the Court filed PER CURIAM.

Opinion concurring in part and dissenting in part filed by
Circuit Judge ROGERS.

PER CURIAM: In 1997, the EPA revised the National
Ambient Air Quality Standard (NAAQS) for ozone from a
1-hour standard to an 8-hour standard. These consolidated
petitions for review challenge aspects of the Final Rule To
Implement the 8-Hour Ozone National Ambient Air Quality
Standard — Phase 2, 70 Fed. Reg. 71,612 (2005) (Phase 2
Rule), and Phase 2 of the Final Rule To Implement the 8-Hour
Ozone National Ambient Air Quality Standard — Notice of
Reconsideration, 72 Fed. Reg. 31,727 (2007) (Reconsideration
Notice). We hold the Phase 2 Rule is inconsistent with the
Clean Air Act (CAA or Act) in allowing participation in a
regional cap-and-trade program to satisfy an area-specific
statutory mandate. We further hold the EPA arbitrarily
eliminated one safeguard and violated the anti-backsliding
provision of the Act insofar as it eliminated another from its
regulations governing review of new sources of pollution. We
therefore grant the petitions with respect to those aspects of the
Phase 2 Rule. In view of our decision in North Carolina v. EPA,
531 F.3d 896 (2008), in which we granted a petition for review
of the Clean Air Interstate Rule (CAIR), we defer consideration
of the Phase 2 Rule and Reconsideration Notice insofar as they
relate to the CAIR program. We deny the petitions in all other
respects.

I. Background
 4

The Act requires the EPA to designate areas as attainment,
nonattainment, or unclassifiable for each NAAQS. CAA §
107(d)(1)(B), 42 U.S.C. § 7407(d)(1)(B). States have primary
responsibility for implementing those standards, and must
submit a state implementation plan (SIP) that specifies how the
state will achieve and maintain compliance with the NAAQS.
Id. § 7407(a). Part D of the Act provides the SIP for a
nonattainment area must include certain control measures. Id.
§ 7501 et seq. Subpart 1 applies to all nonattainment areas, id.
§§ 7501-7509a, whereas Subpart 2 specifies additional
requirements for ozone nonattainment areas, id. §§ 7511-7511f.
Section 181 of the Act classifies ozone nonattainment areas
from “marginal” to “extreme” based upon the degree to which
the ozone level in the area exceeds the NAAQS. Id. § 7511. An
area that exceeds the NAAQS by a greater margin is given more
time to meet the standard but is subjected to progressively more
stringent emissions controls for ozone precursors, namely,
volatile organic compounds (VOCs) and oxides of nitrogen
(NOX). See CAA § 182, 42 U.S.C. § 7511a.

In 1997, the EPA determined the NAAQS for ozone,
expressed as the amount of ozone in the ambient air averaged
over one hour, was inadequate to protect public health. The
EPA therefore promulgated a new NAAQS of .08 ppm of ozone
averaged over eight hours. Under the 8-hour standard, some
ozone nonattainment areas are subject only to the more flexible
requirements of Subpart 1, while areas with higher levels of
ozone are subject to the additional requirements of Subpart 2.
See S. Coast Air Quality Mgmt. Dist. v. EPA, 472 F.3d 882,
893-95 (D.C. Cir. 2006).

The EPA implemented the 8-hour NAAQS in two phases;
the Phase 2 Rule and Reconsideration Notice here under review
implement the requirements of Subpart 1 and Subpart 2 for areas
not attaining the 8-hour NAAQS. The consolidated petitions
 5

challenge those rules as follows. The Natural Resources
Defense Council, the States of New Jersey, Connecticut, and
New York, and the National Petrochemical and Refiners
Association challenge provisions implementing the statutory
requirement that each nonattainment area provide for such
emissions reductions as may be obtained by the adoption of
reasonably available control technology (RACT). The NRDC
and New Jersey challenge provisions governing review of new
sources of pollution. The NRDC also challenges two provisions
implementing the statutory requirements that a SIP for a
nonattainment area provide for specific percentage reductions in
emissions and for contingency measures. Finally, the Chamber
of Greater Baton Rouge and affiliated petitioners1 challenge the
imposition of reformulated gasoline requirements in the Baton
Rouge area.

We review the EPA’s interpretation of the Act pursuant to
Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S.
837 (1984): We ask first whether the Congress has “directly
spoken to the precise question at issue.” Id. at 842. If so, then
we must “give effect to the unambiguously expressed intent of
Congress.” Id. at 843. If, however, the “statute is silent or
ambiguous with respect to the specific issue,” then we defer to
the EPA’s interpretation as long as it is “based on a permissible
construction of the statute.” Id. The Act requires us to review
the Phase 2 Rule deferentially to determine only whether it is
“arbitrary, capricious, an abuse of discretion, or otherwise not in
accordance with law.” CAA § 307(d)(9)(A), 42 U.S.C. §
7607(d)(9)(A); see Bluewater Network v. EPA, 372 F.3d 404,
410 (D.C. Cir. 2004) (“We give particular deference to the EPA
when it acts under unwieldy and science-driven statutory

1
The affiliated petitioners are the West Baton Rouge Chamber of
Commerce, the Iberville Parish Chamber of Commerce, and the
Louisiana Oil Marketers and Convenience Store Association.
 6

schemes like the Clean Air Act”) (internal quotation marks
omitted).

II. Reasonably Available Control Technology

Section 172(c)(1) of the Act requires that the SIPs for
nonattainment areas “provide for the implementation of all
reasonably available control measures as expeditiously as
practicable (including such reductions in emissions from
existing sources in the area as may be obtained through the
adoption, at a minimum, of reasonably available control
technology).” 42 U.S.C. § 7502(c)(1) (emphasis added). Ozone
nonattainment areas that are subject to Subpart 2 of Part D are
subject to more specific reasonably available control technology
requirements. E.g., id. § 7511a(b)(2)(C); id. § 7511a(f).
Petitioners challenge three aspects of the implementation of the
RACT requirement in the Phase 2 Rule. First, the NRDC
challenges the rule providing that RACT is satisfied for Subpart
1 areas by SIPs “demonstrating that the area has adopted all
control measures necessary to demonstrate attainment as
expeditiously as possible,” Phase 2 Rule, 70 Fed. Reg. at 71,701
(codified at 40 C.F.R. § 51.912(c)(1)). Second, the State of New
Jersey challenges the EPA’s decision to allow states to meet the
RACT requirement under the 8-hour NAAQS by certifying that
RACT is met under the 1-hour NAAQS, see id. at 71,652-53.
Third, the NRDC, New Jersey, the States of Connecticut and
New York, and the National Petrochemical and Refiners
Association challenge the EPA’s conclusion that states may
satisfy the RACT requirement by participating in two
cap-and-trade programs, the NOX SIP Call and CAIR, see id.
The court has stayed consideration of the CAIR-RACT
determination.2

2
The second and third challenges are to the EPA’s statements in
the preamble to the final rule. The statements constitute reviewable
 7

A. RACT in Subpart 1 Nonattainment Areas

The NRDC challenges the Phase 2 Rule’s treatment of the
“reasonably available control technology” requirement of CAA
§ 172(c)(1), 42 U.S.C. § 7502(c)(1). Under the Phase 2 Rule,
nonattainment areas governed by Subpart 1 that request an
attainment deadline within five years of their designation “shall
meet the RACT requirement by submitting an attainment
demonstration SIP demonstrating that the area has adopted all
control measures necessary to demonstrate attainment as
expeditiously as practicable.” Phase 2 Rule, 70 Fed. Reg. at
71,701/3 (codified at 40 C.F.R. § 51.912(c)(1)). The NRDC
contends that this provision is an unlawful waiver of the RACT
requirement of § 172(c)(1) because, under the Phase 2 Rule, a
state need not require RACT at all in such areas. It views the
statutory phrase “at a minimum” as imposing an unambiguous
requirement for all nonattainment areas. The NRDC thus
contends the Phase 2 Rule violates the plain text of § 172(c)(1)
by doing away with this requirement in some nonattainment
areas. However, we conclude that the term “reasonably
available control technology” is ambiguous in context and that
the EPA’s interpretation is reasonable.

The court has previously concluded that the term
“reasonably available” in the analogous phrase “reasonably

final agency action, see CAA § 307(b)(1), 42 U.S.C. § 7607(b)(1), as
they represent “the consummation of the agency’s decisionmaking
process and . . . establish[] rights and obligations or create[] binding
legal consequences.” Natural Res. Def. Council v. EPA, 559 F.3d 561,
564 (D.C. Cir. 2009). The EPA’s statements as to how it will
implement the RACT requirement are not conjectural and their terms
are clear, so it is fair to infer that the EPA intended the statements to
create binding legal consequences. See Kennecott Utah Copper Corp.
v. Dep’t of Interior, 88 F.3d 1191, 1222-23 (D.C. Cir. 1996).
 8

available control measure” (RACM) in § 172(c)(1) is ambiguous
and “clearly bespeaks [the Congress’s] intention that the EPA
exercise discretion in determining which control measures must
be implemented . . . .” Sierra Club v. EPA, 294 F.3d 155, 162-
63 (D.C. Cir. 2002). The court explained that the statute did not
specify which control measures would be deemed “reasonably
available” and did not “compel[] a state to consider whether any
measure is ‘reasonably available’ without regard to whether it
would expedite attainment in the relevant area.” Id. at 162.
Thus, the EPA had discretion to conclude that a measure was not
“reasonably available” if it would not expedite attainment. Id.
Because the same is true of the phrase “reasonably available
control technology,” the term “reasonably available” within
RACT is also ambiguous. Moreover, even if the phrase “at a
minimum” requires that at least RACT-level reductions be
achieved in all nonattainment areas, the phrase does not specify
the content of the RACT requirement. Given this ambiguity, the
EPA has discretion reasonably to define the controls that will
demonstrate compliance.

The EPA’s interpretation, construing “reasonably available”
as meaning only control technologies that advance attainment,
is reasonable in light of the statute’s accompanying text and
structure. Section 172(c)(1) requires all nonattainment areas to
achieve RACM “as expeditiously as practicable (including such
reductions . . . as may be obtained through the adoption, at a
minimum, of reasonably available control technology) . . . .” 42
U.S.C. § 7502(c)(1). To the extent an area is already achieving
attainment as expeditiously as possible, imposition of additional
control technologies would not hasten achievement of the
NAAQS. In such a situation, the EPA may reasonably conclude
that no control technologies are reasonably available and the
area need not implement further technologies to satisfy the
RACT requirement. Sierra Club v. EPA, 294 F.3d 155 (D.C.
Cir. 2002), supports the reasonableness of the EPA’s
 9

interpretation. In Sierra Club, the court held reasonable the
EPA’s interpretation of RACM as requiring only those control
measures that would contribute to timely and expeditious
attainment. 294 F.3d at 162; see also Sierra Club v. EPA, 314
F.3d 735, 743-45 (5th Cir. 2002). The court explained, in part,
that the “Act ‘use[s] the same terminology in conjunction with
the RACM requirement’ as it does in requiring timely
attainment.” Sierra Club, 294 F.3d at 162. The § 172(c)(1)
RACM requirement, like the timely attainment requirement of
CAA § 181(a)(1), 42 U.S.C. § 7511(a)(1), requires
implementation of RACM “as expeditiously as practicable.”
Thus, the court concluded, the RACM requirement could
reasonably be understood as a means of meeting the attainment
deadline. 294 F.3d at 162. Because the RACT requirement is
located in a parenthetical modifying RACM and because the
RACM requirement is described as “including” the RACT
requirement, the RACT requirement is likewise linked to the
timely attainment terminology. Given this textual linkage, the
EPA may reasonably extend to the RACT requirement its
interpretation of RACM as requiring only those control
measures that would facilitate expeditious attainment of the
NAAQS.

Contrary to the NRDC’s arguments, the RACT requirement
does not lose all meaning under the EPA’s definition. When
control technology is necessary to advance attainment, it is
“reasonably available” under the definition and would be
required under the rule. The fact that the RACT requirement
was previously located in a separate section and not in a
parenthetical modifying the RACM requirement does not
support the NRDC’s position. Compare CAA § 172(b)(2), 42
U.S.C. § 7502(b)(2), (3) (1977), with CAA § 172(c)(1), 42
U.S.C. § 7502(c)(1) (1990). Rather, the fact that the Congress
moved the requirement to a parenthetical modifying the RACM
requirement supports the EPA’s conclusion that the RACT
 10

requirement could be interpreted in the same manner as the
RACM requirement. Although both the NRDC and the EPA
point to legislative history purportedly supporting their
positions, neither points to legislative history bearing on the
meaning of “reasonably available.” See H.R. REP. NO. 101-490,
reprinted in 2 LEGISLATIVE HISTORY OF THE CLEAN AIR ACT
AMENDMENTS OF 1990, at 223. The term “reasonably available”
is ambiguous, and the EPA’s interpretation is a permissible
construction of the statute.

B. Certifying 8-Hour RACT Based Upon 1-Hour RACT

The Phase 2 Rule provides that a control measure approved
as RACT under the 1-hour standard will be approved as RACT
under the 8-hour standard absent information indicating it
should not be approved. Phase 2 Rule, 70 Fed. Reg. at 71,652/3.
Rather than reassessing what constitutes RACT, a state can
certify that controls that previously satisfied the RACT
requirement also satisfy the requirement under the 8-hour
standard. New Jersey contends this provision is contrary to the
Act and arbitrary and capricious. In New Jersey’s view the EPA
should require a re-analysis for all sources, not just those for
which no controls had been considered RACT under the 1-hour
standard, see id. at 71,655, because what is “reasonably
available” changes over time. Without a re-analysis, New
Jersey maintains there is no basis for states to certify that the
initial RACT analysis meets the RACT requirement under the 8-
hour NAAQS and states may apply outdated RACT controls.
New Jersey thus claims that the EPA should have either updated
its RACT guidance documents or at least provided uniform
criteria for states to use in making RACT certifications.

The EPA promulgates two types of guidance that assist
states in determining what control techniques meet the RACT
requirement, control techniques guidelines (CTGs) and
 11

alternative control techniques (ACTs). Section 183 of the Act
requires the EPA to issue CTGs for certain categories of sources
that emit VOCs. See 42 U.S.C. § 7511b. Where CTGs exist,
they establish the presumptive level of control meeting RACT.
Phase 2 Rule, 70 Fed. Reg. at 71,654/3. Still, states can opt to
require alternative controls rather than following the guidance
in the CTGs. Notice of Final Determination and Availability of
Final Control Techniques Guidelines, 71 Fed. Reg. 58,745,
58,747 (Oct. 5, 2006). Section 183(c) of the Act requires the
EPA to issue ACTs for major sources of VOCs and NOX. 42
U.S.C. § 7511b(c). The ACTs “describe available control
techniques and their cost effectiveness” but do not establish
presumptive RACT. Phase 2 Rule, 70 Fed. Reg. at 71,654/3.
Thus, neither the CTGs nor ACTs set firm RACT requirements.

Despite New Jersey’s concerns, the EPA’s certification
provision does not conflict with the Act and is not arbitrary and
capricious. Although the EPA did not revise the guidance
documents, the EPA’s case-by-case approach adequately
ensures that RACT determinations will take into account
advances in technology. First, the EPA has directed states to
consider available information in addition to the CTG and ACT
documents when making RACT determinations. Id. at
71,655/1. If a state is presented with information indicating that
a previous RACT determination is inappropriate, the state must
consider that information and modify its RACT determinations
accordingly. Id. Second, when submitting RACT certifications
to the EPA as part of their RACT SIP submissions, states must
provide supporting information. Id. at 71,655/2. Third, if
additional information is presented during notice-and-comment
rulemaking, both the state and the EPA are required to consider
that information as part of the rulemaking; this includes
information presented during notice-and-comment rulemaking
for RACT SIP submissions for previously controlled sources.
Id. Because the EPA could reasonably conclude that these
 12

mechanisms will ensure the case-by-case determinations will
take into account advances in technology, the EPA could also
reasonably conclude “that the best way to address the possibility
that CTGs or ACTs might not reflect all currently available
technologies was by requiring each State to consider any new
available information in making its certification, which will then
be reviewed by the EPA as part of the SIP submission process,”
EPA Br. at 67; see also Phase 2 Rule, 70 Fed. Reg. at 71,655/1.
See Nat’l Wildlife Fed’n v. EPA, 286 F.3d 554, 566-67 (D.C.
Cir. 2002). Likewise, given the assurances that RACT
determinations will reflect advances in technology, the EPA’s
approach is consistent with the statutory goal of timely
attainment of the NAAQS, see CAA §§ 172, 182, 42 U.S.C. §§
7502, 7511a.

Additionally, even if the EPA had revised the national
guidelines and provided uniform criteria as New Jersey
preferred, such actions would not have eliminated case-by-case
inquiries by the states and the EPA. Both the CTGs and ACTs
are guidance documents, see Conn. Fund for the Env’t v. EPA,
672 F.2d 998, 1003 (2d Cir. 1982); United States v. Ford Motor
Co., 736 F. Supp. 1539, 1543 (W.D. Mo. 1990), and neither sets
firm RACT requirements. Thus, despite the existence of the
CTGs, the EPA makes case-specific determinations as part of its
SIP approval process. Notice of Final Determination and
Availability of Final Control Techniques Guidelines, 71 Fed.
Reg. at 58,747. ACTs are also merely guidelines and do not
create presumptive RACT levels. Phase 2 Rule, 70 Fed. Reg. at
71,654/3. As case-by-case determinations would be necessary
even if the EPA had revised the guidance documents, the EPA
could reasonably determine that the costs entailed in revising the
guidance documents outweighed the benefits. The EPA has
discretion to consider the costs of regulation unless the relevant
statute precludes such a consideration. See, e.g., Sierra Club,
294 F.3d at 162-63.
 13

The EPA’s approach also addresses New Jersey’s concern
that allowing states to make RACT determinations by relying on
15-year-old CTGs plus new information, if any, provided by the
public in comments is “an inadequate substitute” for requiring
each state to undertake a new RACT determination for each
source category. As noted, the EPA has directed states to
submit supporting documentation along with RACT
certifications. Phase 2 Rule, 70 Fed. Reg. at 71,655/2. Thus, the
EPA will have available the information needed to verify states’
determinations that the previous controls are still appropriate
under the 8-hour standard. Additionally, persons disagreeing
with a particular RACT certification can seek judicial review of
a particular SIP approval. A state may also complete an entirely
new RACT analysis if it so chooses. Id. at 71,652/3.

In sum, we hold that the EPA’s decision to forego a revision
of the nationwide guidelines in favor of case-by-case RACT
certifications was reasonable and not inconsistent with the
statutory goal of expeditious attainment. Because the Phase 2
Rule requires each state to verify that previously-required RACT
controls still satisfy the RACT requirement, we need not address
the EPA’s assertion that a new RACT determination will likely
“result in the same or similar control technology as the initial
RACT determination under the 1-hour standard because the
fundamental control techniques . . . are still applicable.” Id. at
71,654/1. The EPA did not rest the rule upon this ground; each
state must “consider new information” — information that may
prove the EPA wrong — when the state determines whether
previously-required controls “still represent[] an appropriate
RACT level of control.” Id. at 71,655/1.

C. Meeting RACT via NOX SIP Call

The NOX SIP Call is a cap-and-trade program that regulates
NOX emissions. The program covers 22 states in the Northeast
 14

and the District of Columbia, and was intended to address the
interstate transport of ozone. See generally Final Rule, Finding
of Significant Contribution and Rulemaking for Certain States
in the Ozone Transport Assessment Group Region for Purposes
of Reducing Regional Transport of Ozone, 63 Fed. Reg. 57,356
(Oct. 27, 1998); Michigan v. EPA, 213 F.3d 663, 675, 685-86
(D.C. Cir. 2000). The NOX SIP Call established for each state
a “NOX emissions budget,” which limits total emissions during
the ozone season. Phase 2 Rule, 70 Fed. Reg. at 71,656/3. The
EPA also adopted a rule allowing interstate trading of emissions
allowances, and all states covered by the NOX SIP Call elected
to participate in the interstate program. Thus, a source within a
state could meet its emissions target either by installing controls
or by purchasing allowances from other sources located
anywhere in the region covered by the NOX SIP Call. See id.

The Phase 2 Rule provides that to meet the NOX RACT
requirement, “the State need not perform (or submit) a NOX
RACT analysis for sources subject to the state’s emission
cap-and-trade program” where that program meets the NOX SIP
Call requirements or, in states achieving CAIR reductions solely
from Electricity Generating Units, CAIR requirements. Id. at
71,652/3. Petitioners and intervenors persuasively challenge
this provision as being contrary to the express terms of the
statute.

Section 172(c)(1) of the Act requires that nonattainment
areas achieve “such reductions in emissions from existing
sources in the area” as can be achieved by the adoption of
RACT. 42 U.S.C. § 7502(c)(1). Thus, the RACT requirement
calls for reductions in emissions from sources in the area;
reductions from sources outside the nonattainment area do not
satisfy the requirement. See id.; see also CAA § 182(b)(2), 42
U.S.C. § 7511a(b)(2) (requiring implementation of RACT with
respect to certain VOC sources “in the area” for moderate and
 15

above nonattainment areas). Accordingly, participation in the
NOX SIP Call could constitute RACT only if participation
entailed at least RACT-level reductions in emissions from
sources within the nonattainment area. In the preamble to the
proposed rule the EPA stated that “the overall emission
reductions from sources in the NOX SIP Call cap-and-trade
program will achieve more emissions reductions in the
nonattainment area than would application of RACT to each of
those units,” Proposed Rule to Implement the 8-Hour Ozone
National Ambient Air Quality Standard, 68 Fed. Reg. 32,802,
32,839 (June 2, 2003) (Proposed Rule). The preamble to the
final rule does not go as far. Rather, the final preamble states
only that the “NOX SIP Call is estimated to achieve a beyond-
RACT degree of control regionally,” Phase 2 Rule, 70 Fed. Reg.
at 71,657. Such regionwide RACT-level reductions in
emissions do not meet the statutory requirement that the
reductions be from sources in the nonattainment area. Because
the EPA has not shown that NOX SIP Call compliance will result
in at least RACT-level reductions in emissions from sources
within each nonattainment area, the EPA’s determination that
compliance with the NOX SIP Call satisfies the RACT
requirement is inconsistent with the “in the area” requirement
and thus violates the plain text of § 172(c)(1).

In an analogous situation, the court invalidated the CAIR
trading program because the EPA’s regionwide approach made
it impossible to tell whether the rule achieved a specific
statutory objective. See North Carolina v. EPA, 531 F.3d 896,
906-08 (D.C. Cir. 2008). In developing CAIR, the court
explained, the EPA “did not purport to measure each state’s
significant contribution to specific downwind nonattainment
areas and eliminate them in an isolated state-by-state manner.”
Id. at 907. Despite a statutory provision prohibiting sources
“within the State” from contributing significantly to
nonattainment in “any other State,” CAA § 110(a)(2)(D), 42
 16

U.S.C. § 7410(a)(2)(D), the EPA adopted a regionwide approach
to the problem. The court therefore held that the EPA was not
exercising its statutory duty, reasoning that “[i]t is unclear how
EPA can assure that the trading programs . . . will achieve
section 110(a)(2)(D)(i)(I)’s goals if we do not know what each
upwind state’s ‘significant contribution’ is to another state.”
531 F.3d at 908. Similar reasoning applies to the NOX SIP Call.
The EPA has not provided assurance that the NOX SIP Call will
achieve the Act’s goal of “reductions from existing sources in
the area,” because it has not evaluated the effect of the program
on each nonattainment area.

The EPA responds that its approach “gives meaning to the
statute’s ‘in the area’ language” because its technical analysis
shows that the cap-and-trade programs achieve greater
reductions than would the application of RACT-level controls
at each source. EPA Br. at 73. However, the EPA’s cited
support for this proposition is the statement from the preamble
to the proposed rule stating the EPA “believes” participation
would “achieve more emissions reductions in the nonattainment
area than would application of RACT” to all sources in the area.
Proposed Rule, 68 Fed. Reg. at 32,839/2. That statement is
unsupported by any record evidence and it does not appear in
the preamble to the final rule. The EPA’s further response that
its technical analysis supports the conclusion that NOX SIP Call
participation meets the RACT requirement is no more
persuasive. The EPA explains that its analysis showed that
sources subject to the NOX SIP Call “collectively” would
achieve beyond-RACT reductions in emissions. EPA Br. at 83.
But regionwide reductions do not satisfy the “in the area”
requirement. The EPA explains further that it found that
purchasing allowances was more costly than installing RACT-
level control technology and thus “most” sources meeting the
NOX SIP call, “assuming rational economic behavior,” id.,
would opt to install controls rather than purchase allowances.
 17

This bare assertion is insufficient to demonstrate that NOX SIP
Call compliance would lead to RACT-level reductions from
sources in the area. Even if most sources in a nonattainment
area installed controls rather than purchasing allowances, a
small number of sources purchasing allowances and increasing
emissions could mean that overall emissions from sources in the
area remained unchanged or even increased.

The EPA’s attempt to show ambiguity in the Act is likewise
unavailing. The EPA maintains the statute is ambiguous as to
whether RACT must be installed at each source in an area,
noting that it has previously approved the concept of averaging
emissions within a nonattainment area. The EPA reads the
statutory phrase “as may be obtained” to indicate that the RACT
requirement does not necessarily call for implementation of
controls at each and every source, but rather requires an area to
achieve at least RACT-level reductions in emissions. Even if
the RACT requirement could be met through an emissions-
averaging approach within a nonattainment area, averaging
emissions across a region does not ensure that any reductions in
emissions derive from reductions at sources within a particular
area. Even if the EPA were correct that “nothing in the statute
precludes consideration of the air quality impact that controls
under a region-wide cap-and-trade program may have on NOX
within the nonattainment area,” EPA Br. at 72, the EPA has not
considered the impact of the NOX SIP Call on the air quality
within specific nonattainment areas. Therefore the EPA has
failed to demonstrate that NOX SIP Call compliance can be
equated to RACT compliance.

The EPA’s reliance on § 172(c)(6) is misplaced. That
section provides that SIPs must include “enforceable emission
limitations, and such other control measures . . . including
economic incentives such as . . . auctions of emission rights . .
. as may be necessary or appropriate to provide for attainment.”
 18

42 U.S.C. § 7502(c)(6). The EPA offers that its approach to
RACT is consistent with this express authorization of auctions.
However, § 172(c)(6) merely authorizes the EPA to approve
market-based measures in addition to those controls that are
required by CAA § 172(c), including the RACT requirement; it
does not authorize the EPA to replace the RACT requirement
with a cap-and-trade program as is promulgated in the Phase 2
Rule.

The EPA also offers that to the extent individual sources
emitting high levels of NOX are in compliance with the NOX SIP
Call through the purchase of allowances, the states can define
RACT to require greater reductions than are required by the
EPA, and they must require beyond-RACT reductions as
necessary to achieve timely attainment. The EPA’s
determination that NOX SIP Call compliance satisfies the RACT
requirement is not based on a state choosing to so exercise its
discretion. A state’s decision to require stricter controls cannot
eliminate the defect in the EPA’s approach — failing to
implement the requirement of at least RACT-level reductions in
emissions from sources in the nonattainment area. This part of
the Phase 2 Rule must therefore be remanded without vacatur
because the EPA may be able to reinstate the provision for most
nonattainment areas if, upon conducting a technical analysis, it
finds the NOX SIP Call results in greater emissions reductions in
a nonattainment area than would be achieved if RACT-level
controls were installed in that area.

III. Clean Data Policy

Under the Act, each nonattainment area must attain the
NAAQS by a deadline, known as the “attainment date,” which
is established by the statute itself for areas subject to Subparts
1 and 2, see CAA § 181, 42 U.S.C. § 7511, and by the EPA for
areas subject only to Subpart 1, see CAA § 172, 42 U.S.C. §
 19

7502(a)(2). The Act also provides that the SIP for a
nonattainment area subject only to Subpart 1 must “require
reasonable further progress” (RFP), id. § 7502(c)(2); the SIP for
an area in moderate or a greater degree of nonattainment must
provide for fixed percentage reductions of VOCs on a specified
schedule. CAA §§ 182 (b)(1)(A), (c)(2)(B), 42 U.S.C. §§
7511a(b)(1)(A), (c)(2)(B).3

In order to ensure these requirements are met, the SIP for
any nonattainment area must include “contingency measures” to
be implemented “if the area fails to make reasonable further
progress, or to attain the [NAAQS] by the attainment date,”
CAA § 172(c)(9), 42 U.S.C. § 7502(c)(9), and the SIP for a
serious, severe, or extreme nonattainment area must also include
“contingency measures” that will take effect automatically “if
the area fails to meet any applicable milestone,” CAA §
182(c)(9), 42 U.S.C. § 7511a(c)(9).

In the Phase 2 Rule, the EPA suspended the planning
requirements for specified percentage reductions and
contingency measures for each nonattainment area that has
attained the 8-hour NAAQS but has not yet been designated an
attainment area. 40 C.F.R. § 51.918. The suspension lasts
“until ... the area is redesignated to attainment, at which time the
requirements no longer apply; or [until the] EPA determines that
the area has violated the 8-hour ozone NAAQS.” Id. The EPA

3
An area may in some circumstances substitute a combined
reduction in emissions of VOCs and NOX for the fixed percentage
reduction of VOCs required by CAA § 182(c)(2)(B). See 42 U.S.C.
§ 7511a(c)(2)(C). Although CAA § 182(b)(1)(A) applies by its terms
only to an area in moderate nonattainment, and § 182(c)(2)(B) only to
an area classified in serious nonattainment, § 182(c), (d), and (e) apply
those requirements to areas in a greater degree of nonattainment. See
id. § 7511a(c)-(e).
 20

terms this suspension the “Clean Data Policy,” 70 Fed. Reg. at
71,644, because it applies when a nonattainment area produces
“clean” air quality data. See Memorandum from John S. Seitz,
Dir., Office of Air Quality Planning & Standards, RFP,
Attainment Demonstration, and Related Requirements for
Ozone Nonattainment Areas Meeting the Ozone National
Ambient Air Quality Standard, at 5 (May 10, 1995),
www.epa.gov/ttn/oarpg/t1/memoranda/clean15.pdf . It adopted
the current version of the policy in 1995 and applied it in
rulemakings specific to individual areas under the 1-hour
NAAQS; it will similarly determine whether individual areas
qualify for the suspension in area-specific rulemakings. 70 Fed.
Reg. at 71,644-45.

The NRDC contends the Clean Data Policy conflicts with
both the letter and the purpose of the Act. More specifically, the
NRDC argues the statutory provisions requiring a SIP to include
contingency measures and percentage reductions allow no
waivers for an area that has achieved the NAAQS; the only way
for an area to be absolved of those requirements is to be
redesignated an attainment area. See CAA § 175A(c), 42 U.S.C.
§ 7505a(c) (“Until [a maintenance] plan revision is approved
and an area is redesignated as attainment ... the requirements of
[Part D] shall continue in force and effect”). Further, the NRDC
argues only its reading of the text is consistent with the intent of
the Congress that, in order to ensure the air stays clean, no
mandatory control requirement be lifted until a maintenance
plan is in place.

The EPA responds first that the NRDC did not raise before
the agency and therefore forfeited its objection to the suspension
of the planning requirement for contingency measures. See
CAA § 307(d)(7)(B), 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 ...
 21

may be raised during judicial review”). We agree. We
“enforce[] this provision strictly,” Motor & Equip. Mfrs. Ass’n
v. Nichols, 142 F.3d 449, 462 (D.C. Cir. 1998) (internal
quotation marks omitted), and although we allow commenters
“some leeway in developing their argument before this court,”
the comment must have provided “adequate notification of the
general substance of the complaint,” S. Coast Air Quality Mgmt.
Dist. v. EPA, 472 F.3d 882, 891 (D.C. Cir. 2006).

The NRDC’s comments did not provide the EPA adequate
notice that it objected to the suspension of the contingency
measures. The NRDC did object to the Clean Data Policy, but
only to the suspension of the requirements for an attainment
demonstration and percentage reductions. The NRDC now
contends it objected to the suspension of all Subpart 2
requirements, pointing to its comment that the “EPA cannot
authorize states to simply drop subpart 2 measures when the area
is meeting either standard. The Act allows states to move
mandated controls to a maintenance contingency plan, but only
after the area has been redesignated to attainment.” Comments
of Clean Air Task Force et al., at 48 (Aug. 1, 2003). That
comment, however, was made in the context of an objection to
a different provision of the proposed rule, namely the EPA’s
determination that an area not attaining the 1-hour standard but
meeting the 8-hour standard need not submit a maintenance plan
meeting the requirements of CAA § 175A. The EPA cannot be
expected to take the NRDC’s argument, raised in support of one
specific objection, and apply it sua sponte to another provision.
Because the NRDC did not raise its objection with “reasonable
specificity,” the Act bars us from considering it. CAA §
307(d)(7)(B), 42 U.S.C. § 7607(d)(7)(B).

On the merits, the EPA maintains the Clean Data Policy
does not waive the planning requirements for percentage
reductions; instead, it contends, those requirements are by their
 22

terms inapplicable when an area meets the applicable NAAQS.
See Phase 2 Rule, 70 Fed. Reg. at 71,645/1-2, 71,646/1-2.

The two planning requirements for percentage reductions
are CAA § 182(b)(1), which requires an initial reduction of 15%
of VOC emissions in the first six years for an area in moderate
or greater nonattainment, and § 182(c)(2)(B), which requires
subsequent reductions in VOC emissions averaging 3% per year
for an area in serious or greater nonattainment unless an
exception applies. See 42 U.S.C. § 7511a(b)(1), (c)(2)(B). As
the EPA interprets these provisions, specific percentage
reductions are required only as necessary to achieve attainment.

We think the statute unclear as to whether those sections
apply to an area that is already attaining the NAAQS. For the
reasons below, we join the Tenth Circuit in holding the EPA’s
interpretation is reasonable. See Sierra Club v. EPA, 99 F.3d
1551 (10th Cir. 1996).

Section 182(b)(1)(A)(i) initially requires the SIP for an area
in moderate or greater nonattainment to plan a total reduction in
VOC emissions of 15% over six years. 42 U.S.C. §
7511a(b)(1)(A)(i). In the very next sentence, however, it
elaborates that a plan should mandate “such specific annual
reductions ... as necessary to attain the [NAAQS] for ozone by
the attainment date.” Id. Moreover, each percentage reduction
is linked to the requirement that an area make “reasonable
further progress” toward attainment. See id. § 7511a(b)(1)
(entitled “Plan provisions for reasonable further progress”); id.
§ 7511a(c)(2)(B) (entitled “Reasonable further progress
demonstration”). “Reasonable further progress,” in turn, means
“such annual incremental reductions in emissions ... as are
required ... for the purpose of ensuring attainment.” CAA §
171(1), 42 U.S.C. § 7501(1). The Act is therefore ambiguous as
to what reductions are required when no further progress toward
 23

attainment is necessary — or, for that matter, possible. The
EPA reasonably resolved this ambiguity by concluding the
specific percentage reductions are simply inapplicable in that
circumstance. As the Tenth Circuit put the matter, “If a
moderate ozone nonattainment area has in fact already attained
the ozone standard, it would make little sense to require a state
to demonstrate the area will make reasonable progress toward
attainment.” Sierra Club, 99 F.3d at 1557.

The EPA’s reasoning disposes as well of the NRDC’s
contentions that the Clean Data Policy unlawfully circumvents
the redesignation requirements, CAA § 107(d)(3)(E), 42 U.S.C.
§ 7407(d)(3)(E), violates the mandate that all Part D
requirements remain in force until an area has an approved
maintenance plan in place, CAA § 175A(c), 42 U.S.C. §
7505a(c), and disregards the Supreme Court’s admonition that
the EPA cannot “render Subpart 2’s carefully designed
restrictions on EPA discretion utterly nugatory,” Whitman v.
Am. Trucking Ass’ns, 531 U.S. 457, 484 (2001). The Clean Data
Policy does not effect a redesignation; an area must still comply
with the statutory requirements before it can be redesignated to
attainment. Furthermore, Part D — including Subpart 2 —
remains in force insofar as it applies but, as we have just seen,
the EPA has reasonably concluded the provisions of the Act
requiring percentage reductions do not apply to an area that has
attained the NAAQS.

IV. 15% VOC Reduction

CAA section 182(b)(1) requires that, for a moderate
nonattainment area, “no later than 3 years after November 15,
1990, the State shall submit a revision to the applicable
implementation plan to provide for volatile organic compound
emission reductions, within 6 years after November 15, 1990, of
at least 15 percent from baseline emissions, accounting for any
 24

growth in emissions after 1990.” 42 U.S.C. § 7511a(b)(1)(A)(i).
CAA section 182(c), (d) and (e) incorporate the 15% VOC
reduction requirement for, respectively, “serious,” “severe” and
“extreme” nonattainment areas. Id. § 7511a(c)-(e). After the
initial six-year period, CAA section 182(c), (d) and (e) impose
on the same areas an additional RFP requirement that the SIP be
revised so that it “will result in VOC emissions reductions from
the baseline emissions . . . [of] at least 3 percent of baseline
emissions each year,” “averaged over each consecutive 3-year
period beginning 6 years after November 15, 1990, until the
attainment date.” Id. § 7511a(c)(2)(B)(i) (imposing requirement
on “serious” areas); see id. § 7511a(d), (e) (incorporating 3%
reduction requirement for “severe” and “extreme” areas).4 As
used in the cited provisions, “the term ‘baseline emissions’
means the total amount of actual VOC or NOx emissions from
all anthropogenic sources in the area during the calendar year
1990.” CAA § 182(b)(1)(B), 42 U.S.C. § 7511a(b)(1)(B); see
also CAA § 182(c)(2)(B), 42 U.S.C. § 7511a(c)(2)(B).5

In the Phase 2 Rule, the EPA determined that for all
moderate and above areas “that had not met the 15 percent VOC
emission reduction requirement for the 1-hour standard,” the
RFP 15% reduction requirement “would apply” under the 8-hour

4
The statute permits a reduction in “an amount less than 3
percent of such baseline emissions each year, if the State demonstrates
to the satisfaction of the Administrator that the plan reflecting such
lesser amount includes all measures that can feasibly be implemented
in the area, in light of technological achievability.” CAA §
182(c)(2)(B)(ii), 42 U.S.C. § 7511a(c)(2)(B)(ii).
5
Section 182 does not provide for a specific VOC reduction for
a moderate area following the initial six-year 15% reduction period.
Accordingly, thereafter a moderate area is subject to the general
requirement in CAA section 172(c)(2) that “plan provisions shall
require reasonable further progress,” 42 U.S.C. § 7502(c)(2).
 25

standard. 70 Fed. Reg. at 71,632-33. Conversely, the EPA
determined, 8-hour moderate and above areas that had already
satisfied the 15% VOC emission reduction for the 1-hour
standard will not be subjected a second time to the 15%
requirement under the 8-hour standard. Id. at 71,633. Thus,
“[a]reas classified under subpart 2 as moderate that had met the
15 percent VOC emission reduction requirement for the 1-hour
standard are treated in the final rule like areas covered under
subpart 1,” while “[a]reas classified under subpart 2 as serious
and above that had met the 15 percent VOC emission reduction
requirement for the 1-hour standard would be subject to the RFP
requirement in section 172(e) [sic6] and the final rule would
require them to obtain an average of 3 percent annual reductions
of VOC and/or NOX emissions reductions for the first 6 years
after the baseline year and every subsequent 3 years out to their
attainment date.” Id. For both the 15% and the 3% reductions
RFP requirements, the Phase 2 Rule provides that the RFP
periods “would be equivalent to the periods Congress
established in subpart 2,” id. at 71,633, running from a new
baseline year, i.e., 2002 “for areas designated nonattainment for
the 8-hour ozone NAAQS with an effective date of June 15,
2004,” id. at 71,637.

First, the NRDC argues that the EPA “illegally waived”
Subpart 2’s RFP requirement that each SIP be revised three
years after the statute’s effective date to provide for a 15% VOC
reduction within 6 years after enactment for all areas classified
as “moderate” and above for ozone.7

6
The EPA apparently meant section 172(c), see supra note 5;
section 172(e) is the anti-backsliding provision, see infra pp. 43-45.
7
“Ozone, an essential presence in the atmosphere’s stratospheric
layer, is dangerous at ground level” and “is formed by the chemical
reaction of nitrogen oxides (‘NOX’) with any of a number of volatile
 26

As an initial matter, the EPA did not “waive” the 15%
requirement but simply determined that, if a State had made the
required revision for an area classified as moderate or greater
when the one-hour NAAQS took effect, the State did not have
to do so a second time after the 8-hour standard took effect. In
so doing, the EPA reasonably resolved a statutory ambiguity
under step 2 of the framework set out in Chevron U.S.A. Inc. v.
Natural Resources Defense Council, Inc., 467 U.S. 837, 843
(1984). See Natural Res. Def. Council v. EPA, 489 F.3d 1250,
1257 (D.C. Cir. 2007) (“Under Chevron: We first ask ‘whether
Congress has directly spoken to the precise question at issue,’ in
which case we ‘must give effect to the unambiguously expressed
intent of Congress.’ If the ‘statute is silent or ambiguous with
respect to the specific issue,’ however, we move to the second
step and defer to the agency’s interpretation as long as it is
‘based on a permissible construction of the statute.’ ” (quoting
Chevron, 467 U.S. at 842-43) (internal quotation marks
omitted)). Accordingly, we reject the NRDC’s challenge to the
VOC reduction provision.

The NRDC contends that the EPA’s reading of the statute
is “untenable” because the statute requires that the SIP provide
for a 15% reduction from “baseline emissions”—which the
“EPA itself defines . . . for purposes of the 8-hour ozone
standard as emissions in 2002, not 1990”—and that the
reduction be effected “ ‘within 6 years after’ the baseline
year—that is, between 2002 and 2008 for the 8-hour standard.”
NRDC Br. at 29. This argument, however, sidesteps the EPA’s
rationale for its interpretation. The EPA identified a “gap in the
statutory scheme” because “[t]he CAA is silent regarding
whether a nonattainment area that implements the 15-percent
VOC emission reduction of 42 U.S.C. § 7511a(b)(1)(A) must

organic compounds (‘VOCs’), in the presence of sunlight.” S. Coast
Air Quality Mgmt. Dist. v. EPA, 472 F.3d 882, 887 (D.C. Cir. 2006).
 27

implement that provision a second time if the NAAQS is
revised, or instead must implement other RFP provisions that
expressly would have applied had there been no revision of the
NAAQS.” EPA Br. at 29. The EPA filled this gap by selecting
the latter resolution. As the EPA explained in the Phase 2 Rule,
while it “believes that the CAA is quite clear that the SIP must
provide for a 15 percent reduction in baseline VOC emissions
for some period after 1990 in an area subject to section
182(b)(1)(A),” it “disagrees that the CAA plainly requires that
the SIP for an area must require a second 15 percent reduction
in VOC baseline emissions under a revised ozone standard.” 70
Fed. Reg. at 71,635-36 (first emphasis added); see also id. at
71,634 (“For those areas that have an approved 15 percent plan
for their 1-hour ozone SIPs, an additional 15 percent VOC
reduction is not necessary.”). Because the EPA ended its
statutory analysis with the threshold inquiry whether section
182(b)(1)(A)(i) must be applied a second time under its revised
standards (concluding it need not), the Agency did not need to
decide how to interpret the term “baseline emissions” or to
identify a baseline year for the purpose of so applying the
provision. Further, because it had no need to reach these issues,
it did not resolve them arbitrarily or capriciously, as the NRDC
asserts, by “allowing a select group of nonattainment areas to
rely on a different baseline from the distant past.” NRDC Br. at
30-31. Had the EPA required a nonattainment area to undertake
a second 15% VOC reduction, it might have been arbitrary to
use a different baseline or baseline year from that applicable to
areas undertaking the first 15% VOC reduction under the 8-hour
standard. But the Agency did not do so and we will not
speculate whether it could lawfully have done so.

The NRDC also argues that the EPA’s 15% approach is
inconsistent with its treatment of section 182(c)(2)(B)(i)’s 3%
reduction requirement for serious and above areas in that the
“EPA does not read the Act as excusing such 3% continuing
 28

emission cuts for serious and above 8-hour nonattainment areas
merely because they may have had plans under the 1-hour
standard to achieve continuing (i.e. post-1996) cuts of 3%/yr
from a 1990 baseline.” NRDC Br. at 30 n.8. The 3% provision,
however, is fundamentally different from the 15% provision in
that the former does not (as does the latter) establish a one-time
reduction requirement but instead imposes a continuing
obligation that applies to “each consecutive 3-year period
beginning 6 years after November 15, 1990, until the attainment
date.” CAA § 182(c)(2)(B), 42 U.S.C. § 7511a(c)(2)(B).

Finally, the NRDC asserts that the “EPA cannot override an
express statutory command ‘simply by asserting that its
preferred approach would be better policy,’ ” NRDC Br. at 30
(quoting Engine Mfrs. Ass’n v. EPA, 88 F.3d 1075, 1089 (D.C.
Cir. 1996)), referring to the EPA’s observation that its approach
“provides flexibility to States to use a mix of NOx and VOC
reductions” under CAA section 172, 42 U.S.C. § 7502, rather
than the required VOC-only reductions mandated by CAA
section 182(b)(1)(A)(1), 42 U.S.C. § 7511a(b)(1)(A)(1). 70 Fed.
Reg. at 71,634. As already explained, however, the EPA’s
interpretation does not override an express command but rather
resolves an ambiguity. Id. at 71,635-36. In commending the
flexibility afforded, the EPA merely offered one more reason
why its interpretation of the ambiguous statutory language is
reasonable.

V. New Source “Offset” Credit for Past Emission
Reductions

CAA section 172(b) requires that each State containing a
nonattainment area submit a nonattainment SIP meeting the
requirements of section 172(c) pursuant to a schedule set by the
EPA but not later than three years after a nonattainment area is
so designated. 42 U.S.C. § 7502(b), (c). Section 172(c)(5) of
 29

the Act requires that each nonattainment area SIP “shall require
permits for the construction and operation of new or modified
major stationary sources anywhere in the nonattainment area, in
accordance with section 7503 of [Title 42].” Id. § 7502(c)(5).
Section 173 sets out the permitting process, known as “New
Source Review” (NSR), in greater detail, providing that a permit
may be issued only if, inter alia, (1) the permitting agency has
determined that by the time the source begins operation,
sufficient offsetting emissions reductions are obtained that the
emission levels from the offsets and the plan provisions
represent RFP from the pre-permit levels, id. § 7503(a)(1)(A),8
and (2) the proposed source will “comply with the lowest
achievable emission rate” (LAER), id. § 7503(a)(2).9 Section

8
Section 173(a)(1)(A) requires that

the permitting agency determine[] that—

(A) by the time the source is to commence operation,
sufficient offsetting emissions reductions have been
obtained, such that total allowable emissions from
existing sources in the region, from new or modified
sources which are not major emitting facilities, and from
the proposed source will be sufficiently less than total
emissions from existing sources (as determined in
accordance with the regulations under this paragraph)
prior to the application for such permit to construct or
modify so as to represent (when considered together with
the plan provisions required under section 7502 of this
title) reasonable further progress (as defined in section
7501 of this title); . . . .

42 U.S.C. § 7503(a)(1)(A).
9
Section 173(a) also requires that (1) the new source owner or
operator has demonstrated that all of its other major stationary sources
in the State are subject to and in (or scheduled for) compliance with
 30

173 further requires that “[s]uch emission reductions shall be, by
the time a new or modified source commences operation, in
effect and enforceable and shall assure that the total tonnage of
increased emissions of the air pollutant from the new or
modified source shall be offset by an equal or greater reduction,
as applicable, in the actual emissions of such air pollutant from
the same or other sources in the area.” Id. § 7503(c)(1). Subpart
2 specifically applies the NSR requirement to ozone
nonattainment areas, CAA § 182(a)(2)(C)(i), 42 U.S.C.
§ 7511a(a)(2)(C)(i), and mandates increasingly stringent offset
ratios as the ozone classification increases, CAA § 182(a)(4),
(b)(5), (d)(2), 42 U.S.C. § 7511a(a)(4), (b)(5), (d)(2).10

The Phase 2 Rule allows proposed new and modified
sources to meet this offset requirement using credits from
sources that shut down or curtailed operations as long ago as
1977. 70 Fed. Reg. at 71,699 (to be codified at 40 C.F.R.
§ 51.165). The EPA has long allowed emissions reductions
occurring before a permit application to qualify as offset credits
under specified circumstances. Before 1989, such pre-
application emissions reductions could be offset only if the
proposed source was “a replacement for the productive capacity

applicable emission limitations and standards; (2) the Administrator
has not determined the implementation plan applicable to the
attainment area site of the new source is not being adequately
implemented; and (3) an analysis of alternatives demonstrates that
“benefits of the proposed source significantly outweigh the
environmental and social costs imposed as a result of its location,
construction, or modification.” 42 U.S.C. § 7503(a)(3)-(5).
10
More generally, the Act requires that all SIPs, even those for
attainment areas, include regulation of new and modified sources “as
necessary to assure that national ambient air quality standards are
achieved, including a permit program as required in parts C and D of
this subchapter.” CAA § 110(a)(2)(C), 42 U.S.C. § 7410(a)(2)(C).
 31

represented by the proposed offset credit.” Requirements for
Implementation Plans; Air Quality New Source Review, 54 Fed.
Reg. 27,286, 27,290 (June 28, 1989).11 The 1989 Rule
eliminated this restriction for an area with an approved
attainment demonstration. Id. at 27,292. In the Phase 2 Rule,
the EPA “lift[ed] the requirement to have an approved
attainment plan before using preapplication credits from
shutdowns or curtailments as offsets.” 70 Fed. Reg. at 71,676.
The NRDC challenges both the EPA’s longstanding policy
allowing pre-application reductions as NSR offsets and the
EPA’s elimination in the Phase 2 Rulemaking of the approved
attainment demonstration requirement. We reject as untimely
the NRDC’s challenge to the general policy of allowing pre-
application offset credits but we agree that eliminating the
attainment demonstration requirement was arbitrary and
capricious and therefore in violation of the APA.

11
The pre-1989 regulation provided:

Emissions reductions achieved by shutting down an existing
source or permanently curtailing production or operating
hours below baseline levels may be credited, provided that
the work force to be affected has been notified of the
proposed shutdown or curtailment. Source shutdowns and
curtailments in production or operating hours occurring
prior to the date the new source application is filed generally
may not be used for emissions offset credit. However,
where an applicant can establish that it shut down or
curtailed production after August 7, 1977, or less than one
year prior to the date of permit application whichever is
earlier, and the proposed new source is a replacement for the
shutdown or curtailment credit for such shutdown or
curtailment may be applied to offset emissions from the new
source.

40 C.F.R. § 51.165(a)(3)(ii)(C) (1988).
 32

CAA section 307(b), which governs judicial review of a
rulemaking under the Act, requires that “[a]ny petition for
review . . . shall be filed within sixty days from the date notice
of such promulgation . . . appears in the Federal Register.” 42
U.S.C. § 7607(b)(1). As already noted, the EPA has allowed
emission reductions from pre-application source shutdowns and
curtailments to qualify as NSR offset credits since at least 1989.
See 54 Fed. Reg. at 27,292. Thus, the deadline for filing a
petition for review of section 51.165(a)(3)(ii)(C) insofar as it
authorizes offset credits for pre-permit emission reductions is
long since past and the court lacks jurisdiction to review it. See
Motor & Equip. Mfrs. Ass’n v. Nichols, 142 F.3d 449, 460 (D.C.
Cir. 1998) (section 7607(b)(1) “filing period is jurisdictional in
nature, and may not be enlarged or altered by the courts”
(internal quotation marks omitted)). Nonetheless, the NRDC
contends that the EPA “reopened” the 1989 offset policy so as
to make its challenge timely. We disagree.

“The reopening doctrine allows an otherwise stale challenge
to proceed because ‘the agency opened the issue up anew,’ and
then ‘reexamined . . . and reaffirmed its [prior] decision.’ ” P &
V Enters. v. U.S. Army Corps of Eng’rs, 516 F.3d 1021, 1023-24
(D.C. Cir. 2008) (quoting Pub. Citizen v. Nuclear Reg. Comm’n,
901 F.2d 147, 150-51 (D.C. Cir. 1990)) (internal quotation
marks omitted). In this case, the EPA did not expressly reopen
the issue of offsets for pre-application emission reductions. In
its proposed rule, the EPA offered two alternative amendments
to section 51.165(a)(3)(ii)(C), each of which eliminated the
attainment demonstration requirement under certain
circumstances. See Prevention of Significant Deterioration
(PSD) and Nonattainment New Source Review (NSR), 61 Fed.
Reg. 38,250, 38,311-14 (July 23, 1996) (Proposed Rules). In the
Phase 2 Rule, the EPA selected the less restrictive of the two
 33

alternatives.12 Neither alternative proposed changing the
regulation insofar as it permits offset of pre-application emission
reductions generally or even mentioned the matter. Further,
other than eliminating the approved attainment demonstration
requirement, the language of amended section 51.165(a)(3)(C)
remains substantively identical to the 1989 regulation. Compare
54 Fed. Reg. at 27,299 (§ 51.165(a)(3)(C) as effective June 29,
1989), with 70 Fed. Reg. at 71,699 (§ 51.165(a)(3)(C) as
effective Jan. 30, 2006). Accordingly, the subject of allowing
offsets in general was not expressly reopened. See Nat’l Ass’n
of Reversionary Prop. Owners v. Surface Transp. Bd., 158 F.3d
135, 142 (D.C. Cir. 1998) (“When an agency invites debate on
some aspects of a broad subject, . . . it does not automatically
reopen all related aspects including those already decided.”)

Nor did the EPA constructively reopen the issue, as the
NRDC contends. Under some circumstances an issue may be
“deemed to have been constructively reopened even though it
was not actually reopened” in a literal sense. Kennecott Utah
Copper Corp. v. U.S. Dep’t of Interior, 88 F.3d 1191, 1214
(D.C. Cir. 1996). “A constructive reopening occurs if the
revision of accompanying regulations ‘significantly alters the
stakes of judicial review’ as the result of a change that ‘could
have not been reasonably anticipated.’ ” Sierra Club v. EPA,
551 F.3d 1019, 1025 (D.C. Cir. 2008) (quoting Kennecott, 88
F.3d at 1227; Envtl. Def. v. EPA, 467 F.3d 1329, 1334 (D.C. Cir.
2006)). The EPA’s elimination of the attainment determination
requirement did not work such a sea change. The basic
regulatory scheme remains unchanged: pre-application offsets
are permitted if certain requirements are met. The only change

12
Alternative 1 conditioned using offset credits upon an area
being “current with part D ozone nonattainment planning
requirements”; Alternative 2, which the EPA selected, did not. Phase
2 Rule, 70 Fed. Reg. at 71,673.
 34

is the elimination of the attainment demonstration
requirement—which was itself added in 1989 to replace the
more stringent requirement that “the proposed new source [be]
a replacement for the shutdown or curtailment.” 40 C.F.R.
§ 51.165(a)(3)(ii)(C) (1988) (alteration added); see supra note
11. That the NRDC’s challenge to the change may have merit,
as indeed we conclude infra, does not provide a ground to
conclude that the broader issue of allowing any pre-application
offsets at all has been reopened. It simply means the particular
change it challenges is unlawful. The “stakes” here are not
quantitatively different from what they were in 1989 when a
coalition of environmental groups argued to the EPA that the
attainment demonstration requirement itself was inadequate to
ensure RFP. See 54 Fed. Reg. at 27,291-92. Yet no one then
challenged the general policy allowing pre-application offsets.
Further, it could have been “reasonably anticipated” at that time
that the EPA might subsequently eliminate the attainment
demonstration requirement itself in favor of an alternative
safeguard which would likewise allow more liberal use of pre-
application offset credits, as in fact happened here. In sum, the
EPA has done nothing since 1989 to reopen the general issue of
allowing pre-application offsets either expressly or
constructively.

We next address the NRDC’s objection to the specific NSR
revision that is subject to review in this proceeding—that the
EPA erred in eliminating the attainment demonstration
requirement. We agree with the NRDC that the EPA acted
arbitrarily and capriciously when it did so.

In the 1989 rulemaking, the EPA imposed the approved
attainment demonstration requirement in order to comply with
CAA section 173(a)(1), which allows an NSR permit to issue
only if
 35

the permitting agency determines that—

(A) by the time the source is to
commence operation, sufficient offsetting
emissions reductions have been obtained,
such that total allowable emissions from
existing sources in the region, from new or
modified sources which are not major
emitting facilities, and from the proposed
source will be sufficiently less than total
emissions from existing sources allowed
under the applicable implementation plan
prior to the application for such permit to
construct or modify so as to represent (when
considered together with the plan provisions
required under section 7502 of this title)
reasonable further progress (as defined in
section 7501 of this title); . . . .

42 U.S.C. § 7503(a)(1)(A) (1989). The EPA concluded that the
“essence of [this] provision is that a new source may be allowed
in a nonattainment area only where its presence would be
consistent with RFP toward attainment of the NAAQS” and that
this policy is satisfied if there is in place for the area a “fully
approved SIP” since, “[b]y definition, any fully approved SIP
has independently assured RFP and attainment.” 54 Fed. Reg.
at 27,292. “[W]here an attainment demonstration is lacking,”
however, the EPA concluded that “retention of the current
shutdown credit restriction on offset transactions”—that is, for
a pre-permit shutdown, “the proposed source is a replacement
for the productive capacity represented by the proposed offset,”
id. at 27,290—“is necessary both to assure RFP and to guarantee
that a new source does not cause or contribute to a violation of
the NAAQS.” Id. at 27,292. Nonetheless, when the EPA
amended section 51.165(a)(3)(C) in the Phase 2 Rule, it
 36

eliminated the fully-approved SIP requirement without adding
any other safeguard to ensure that issuing a particular permit is
consistent with CAA section 173’s mandate that total reductions
“represent . . . reasonable further progress.” 42 U.S.C. §
7503(a)(1)(A). The EPA justified eliminating the requirement
on the ground that the 1990 amendments to the Act “changed
the considerations involved” in that they “emphasized the
emission inventory requirement in section 172(c)(3) as a
fundamental tool in air quality planning” and “added new
provisions keyed to the inventory requirement, including
specific reduction strategies and [milestones] that measure
progress toward attainment from the base year emissions
inventory or subsequent revised inventories” along with “several
adverse consequences where States fail to meet the planning or
emissions reductions requirements.” 70 Fed. Reg. at 71,677; see
also Reconsideration Notice, 72 Fed. Reg at 31,739. But
imposing post hoc sanctions for failing to achieve RFP
milestones does not accomplish what CAA section 173 requires,
namely, that the EPA ensure that “by the time the source is to
commence operation, sufficient offsetting emissions reductions
have been obtained” to produce RFP, 42 U.S.C. § 7503(a)(1)(A)
(emphasis added). Because the EPA has not explained how the
amended rule, stripped of the approved attainment
demonstration requirement, ensures that emission reductions are
achieved “by the time” the new source begins operation rather
than sometime down the road after milestones have been missed,
we conclude that eliminating the requirement is arbitrary and
capricious. See N. Baja Pipeline, LLC v. FERC, 483 F.3d 819,
821 (D.C. Cir. 2007) (under arbitrary and capricious standard,
agency conclusions “must be reasonable and reasonably
explained” (citing Nat’l Fuel Gas Supply Corp. v. FERC, 468
F.3d 831, 839 (D.C. Cir. 2006))); Transactive Corp. v. United
States, 91 F.3d 232, 236 (D.C. Cir. 1996) (“In order to ensure
that an agency’s decision has not been arbitrary, we require the
agency to have identified and explained the reasoned basis for
 37

its decision.” (citing F.J. Vollmer Co. v. Higgins, 23 F.3d 448,
451 (D.C. Cir. 1994); Nat’l Treasury Employees Union v.
Horner, 854 F.2d 490, 498-99 (D.C. Cir. 1988))).13

VI. NSR Limits on Pollution from New and Modified
Major Sources

Pursuant to the SIP NSR requirements of CAA Part D,
supra pp. 28-30, the Phase 2 Rule accords a State three years to
develop and submit an approvable nonattainment major NSR
program for the 8-hour NAAQS. 70 Fed. Reg. at 71,672. In
addition to the Part D requirements for nonattainment areas,
however, CAA section 110(a)(2)(c) imposes on States a
“general duty” to implement a NSR program in their SIPs for all
areas, see supra note 10, which “exists during all periods,
including before a State has an approved part D NSR permit
program.” Id. at 71,677-78. Although section 110(a)(2)(c)
requires that the SIP contain NSR provisions, it does not specify
what NSR requirements apply during the period after an area is
designated nonattainment but before the NSR SIP is effective.
To fill this gap, the EPA decided in the Phase 2 Rule to retain an
interim NSR regime it has used for this purpose since 1979,
namely, Appendix S to 40 C.F.R. Part 51, which both establishes
interim NSR permitting requirements paralleling the Act’s (in
section IV.A) and provides for an exemption from the same
requirements under certain circumstances (in section VI). The
EPA further decided in the Phase 2 Rule, however, to eliminate
an existing 18-month time limit on the applicability of Appendix
S (including section VI’s exemption provision) to a given
nonattaiment area. Petitioners NRDC and New Jersey challenge
both the EPA’s general authority to exempt new sources from

13
In light of this conclusion, we need not consider the NRDC’s
alternative contention that elimination of the requirement violates
section 172(e) of the Act, 42 U.S.C. § 7502(e).
 38

permitting requirements under section VI and, more specifically,
its decision to eliminate the 18-month limit on such exemptions.
We reject the petitioners’ objection to the EPA’s general
exemption authority as untimely but agree that eliminating the
18-month limit violates the “anti-backsliding” provision in CAA
section 172(e), 42 U.S.C. § 7502(e).

In 1979, the EPA codified in its regulations a three year-old
“Emission Offset Interpretative Ruling” as Appendix S to 40
C.F.R. Part 51 and directed that Appendix S was to “govern[]
permits to construct and operate applied for before the deadline
for having a revised SIP in effect that satisfies Part D.”
Approval and Promulgation of Implementation Plans; Statutory
Restriction on New Sources Under Certain Circumstances for
Nonattainment Areas, 44 Fed. Reg. 38,471, 38,473 (July 2,
1979) (codified at 40 C.F.R. § 52.24(c) (1979) (now § 52.24(k));
see Emission Offset Interpretative Ruling, 44 Fed. Reg. 3274
(Jan. 16, 1979) (codifying interpretive ruling in regulations). In
so applying the interpretive rule, the EPA allowed a
nonattainment area to avoid, during the period before the SIP
was due, the statutory ban on all major source construction or
modification in any nonattainment area not subject to a SIP that
“me[t] the requirements of Part D (relating to nonattainment
areas).” See Clean Air Act Amendments of 1977, Pub. L. No.
95-95, § 108(b), 91 Stat. 685, 694 (Aug. 7, 1977) (establishing
construction ban) (1977 CAA Amendments). The EPA noted at
the time that the Congress had similarly adopted the rule (“as
may be modified by rule of the Administrator”) to govern NSR
during the interim period from enactment of the 1977 CAA
amendments until July 1, 1979, when the statutory ban was to
take effect. 44 Fed. Reg. at 38,472; 1977 CAA Amendments
§ 129(a)(1), 91 Stat. at 745 (adopting interpretive rule for
interim before ban) (Aug. 7, 1977). The Congress eliminated
the construction ban in 1990. Clean Air Act, Amendments, Pub.
L. No. 101-549, § 101(c), 104 Stat. 2399, 2408 (1990).
 39

In 1980, the EPA “clarif[ied]” Appendix S to limit to
eighteen months the period during which it governed NSR in a
qualifying nonattainment area. Approval and Promulgation of
Implementation Plans; Statutory Restriction on New Sources
Under Certain Circumstances for Nonattainment Areas, 45 Fed.
Reg. 65,209 (Oct. 2, 1980). Since that time the EPA has
continued to apply Appendix S’s NSR requirements in the
interim after an area’s nonattainment designation and before
implementation of a revised SIP. Thus, after the Congress
amended the Act in 1990—adding, inter alia, Subpart 2’s
ozone-specific requirements—the EPA again affirmed its policy
that Appendix S govern during the period between
nonattainment designation and SIP approval and
implementation. See State Implementation Plans; General
Preamble for the Implementation of Title I of the Clean Air Act
Amendments of 1990; Supplemental, 57 Fed. Reg. 18,070,
18,076 (Apr. 28, 1992).

Since Appendix S was formally codified in 1979, section VI
thereof has contained— unchanged—an exemption from section
IV.A’s permitting requirements for an area whose attainment
date had not passed if the source satisfied two conditions:

In such cases, a new source which would cause or
contribute to an NAAQS violation may be exempt
from the Conditions of Section IV.A. so long as the
new source [1.] meets the applicable SIP emission
limitations and [2.] will not interfere with the
attainment date specified in the SIP under Section 110
of the Act.

44 Fed. Reg. at 3285 (codified at Appendix S, § VI); cf. 40
C.F.R. pt 51, app. S., § VI (2006) (same). In the Phase 2 Rule,
the EPA added a third condition so that the reformulated
provision now exempts new sources from the permitting
 40

requirements in section IV.A “if the conditions in paragraphs
VI.A through C are met,” namely:

A. The new source meets the applicable SIP emission
limitations.

B. The new source will not interfere with the
attainment date specified in the SIP under section 110
of the Act.

C. The Administrator has determined that conditions A
and B of this section are satisfied and such
determination is published in the Federal Register.

70 Fed. Reg. at 71,704 (codified at 40 C.F.R. pt 51, app. S., § VI
(2009)). The EPA further eliminated the 18-month limit on the
applicability of Appendix S (including section VI) to a pre-SIP
nonattainment area. 70 Fed. Reg. at 71,677; see 40 C.F.R. §
52.24(k) (2009).

As noted above, petitioners NRDC and New Jersey
challenge the EPA’s authority generally to exempt an area from
Appendix S’s permitting requirements and specifically the
EPA’s removal of the 18-month limit. We address each in turn.

First, we consider the general challenge to the EPA’s
exemption authority and conclude that it is time-barred. Under
CAA section 307(b), any objection to the exemption policy was
required to be raised in a petition for review “filed within sixty
days from the date notice of [its] promulgation . . . appears in the
Federal Register,” 42 U.S.C. § 7607(b)(1). In this case the
triggering date was July 2, 1979, when the EPA published its
final rule promulgating 40 C.F.R. § 52.24(c) (1979) (now
52.24(k)), which directed that “[t]he Emission Offset
Interpretative Ruling, 40 CFR Part 51, Appendix S”—which, as
 41

then codified, contained the exemption provision in section VI,
44 Fed. Reg. at 3285—“governs permits to construct and
operate applied for before the deadline for having a revised SIP
in effect that satisfies Part D.” 44 Fed. Reg. at 38,473 (codified
at 40 C.F.R. § 52.24(c) (1979) (now 52.24(k)). It was on that
date that the EPA unequivocally asserted its authority to exempt
new sources from permitting requirements under section VI.
Thus, the petitioners’ challenge to such authority in this case
comes almost thirty years late and we are precluded from
considering it. See Motor & Equip. Mfrs. Ass’n, 142 F.3d at
460. Nonetheless, the petitioners again (as with the emission
offset authority issue treated supra) seek to circumvent the
statutory time limit by invoking the reopening doctrine. Once
again their attempt is unavailing.

The petitioners contend the EPA implicitly reopened the
question of its exemption authority because during the
rulemaking, the Agency sought comment on an alternative
exemption proposal which it subsequently declined to adopt.
See Edison Elec. Inst. v. EPA, 996 F.2d 326, 332 (D.C. Cir.
1993) (“By soliciting comments on the existing . . . regulations
and advancing a possible ‘alternative approach’ . . ., the EPA
clearly provided the type of ‘opportunity for renewed comment
and objection’ that suffices to restart the statutory period for
seeking review.” (quoting Ohio v. EPA, 838 F.2d 1325, 1328
(D.C. Cir. 1988)). Specifically, the petitioners argue that the
EPA reopened the question of its exemption authority when it
“proposed to remove the general waiver language from
Appendix S entirely, and replace it with a ‘transitional’ program
that would only allow relief from NSR in a limited group of
areas that, among other things, submitted plans by 2004
adequate to attain standards by 2007”—a proposal the EPA
subsequently declined to adopt. NRDC Br. at 41 (citing 68 Fed.
Reg. at 32,846-48) (emphases in brief). As the petitioners’
characterization demonstrates, however, the EPA reopened only
 42

the question of how broadly it should exercise its exemption
authority and not whether it had such authority in the first place.
The abandoned proposal would have replaced section VI’s
exemption provisions with a “transitional” program, which
would, like section VI, provide exemptions from section IV.A’s
NSR requirements but would allow the exemptions for a
narrower class of “eligible” areas. See Proposed Rule To
Implement the 8-Hour Ozone National Ambient Air Quality
Standard, 68 Fed. Reg. 32,802, 32,846-47 (June 2, 2003) (setting
out eligibility requirements for transitional program). Thus, the
EPA’s alternative proposal did not implicitly reopen the
question of its statutory authority vel non to grant any NSR
exemptions but only whether section VI as previously applied
might violate the Act by making exemptions too broadly
available. Because the EPA did not seek comment on its general
authority to exempt areas from the permitting requirements, this
case is plainly distinguishable from Edison Electric Institute, in
which the EPA “explicitly invited comments on the precise
question for which petitioners [there sought] review.” Edison
Elec. Inst., 996 F.2d at 332. Accordingly, we conclude that the
issue of the EPA’s general exemption authority is foreclosed
under CAA section 307(b)(1), 42 U.S.C. § 7607(b)(1).

The petitioners also argue that the EPA constructively
reopened the entire exemption provision “by dramatically
expanding its scope and effect” when it eliminated the 18-month
limit on Section VI exemptions. NRDC Br. at 41(citing
Kennecott, 88 F.3d at 1226-27); see also NRDC Reply Br. at 18.
The extension of the exemption term, however, did not
“ ‘significantly alter[] the stakes of judicial review’ as the result
of a change that ‘could have not been reasonably anticipated.’ ”
Sierra Club v. EPA, 551 F.3d at 1025 (quoting Kennecott, 88
F.3d at 1227; Envtl. Def., 467 F.3d at 1334). In fact, when first
codified in 1979, the interpretive ruling lacked such a limit—yet
its general exemption authority went unchallenged. Further, an
 43

exemption from NSR for even 18 months, if unlawful, seems
worth challenging in its own right—as it could have been when
the 18-month limit was added in 1980. To the extent that the
EPA’s removal of the limit extends the exemption period, the
extension is subject to direct challenge in this proceeding and we
next address the petitioners’ challenge thereto.

The petitioners contend that the EPA’s elimination of the
18-month limit on NSR exemptions violates CAA
section 172(e), the Act’s “anti-backsliding” provision. Section
172(e) provides that if the the EPA “relaxes a national primary
ambient air quality standard,” it must “promulgate requirements
applicable to all areas which have not attained that standard as
of the date of such relaxation” which “shall provide for controls
which are not less stringent than the controls applicable to areas
designated nonattainment before such relaxation.” 42 U.S.C.
§ 7502(e). Although section 172(e) expressly applies only when
the EPA “relaxes” a NAAQS, the EPA has interpreted the
language to also apply when it strengthens a NAAQS—as it did
when it adopted the 8-hour ozone standard—reasoning that “if
Congress intended areas to remain subject to the same level of
control where a NAAQS was relaxed, they also intended that
such controls not be weakened where the NAAQS is made more
stringent.” Final Rule To Implement the 8-Hour Ozone National
Ambient Air Quality Standard—Phase 1, 69 Fed. Reg. 23,951,
23,972 (Apr. 30, 2004) (Phase 1 Rule). In South Coast Air
Quality Management District v. EPA, 472 F.3d 882 (D.C. Cir.
2006), we upheld this interpretation, id. at 900, and also
concluded that NSR is a “control” subject to section 172(e)’s
backsliding prohibition, id. at 902. Accordingly, the EPA’s
elimination of the 18-month exemption limit violates section
172(e) if the resulting NSR requirement is “less stringent” than
the existing requirement. Insofar as Appendix S now provides
for waiver of NSR for an unlimited time pending SIP approval,
it is plainly “less stringent” than the previous version which
 44

limited an NSR waiver to an 18-month term. As the petitioners
argue, the EPA’s revision could delay implementing NSR
controls in eligible nonattainment areas for years beyond the
previous 18-month limit. Accordingly, we conclude that the
revision constitutes backsliding in violation of section 172(e).

The EPA offers three defenses of its revision. First,
because the EPA established the time limit to allow a newly-
designated nonattainment area a reasonable time to develop a
SIP before the statutory construction ban applied, see 45 Fed.
Reg. at 65,209, the EPA contends that it reasonably removed the
18-month limitation after the Congress repealed the new source
construction ban in 1990. The EPA might have done so but for
section 172(e)’s unequivocal backsliding constraint, which
plainly prohibits a “less stringent” NSR control—precisely what
the revision produced. Second, the EPA asserts “there is no
basis to speculate” that it will grant waivers for the entire period
before SIP approval because “it ‘would be highly disinclined to
grant a waiver where the SIP submission deadline has passed
and EPA had not received the required submission.’ ” EPA Br.
at 102-03 (quoting 72 Fed. Reg. at 31,745-46). We do not
speculate, however, in concluding that, notwithstanding the
EPA’s promised discretionary restraint, the revised section VI
will produce some waivers lasting beyond the previously
prescribed 18-month period and thus impose a control less
stringent than before. This result contravenes section 172(e).
Nor are we persuaded by the EPA’s third argument, that the
revision’s new, third “pre-condition”—that the EPA determine
that the existing two pre-conditions (satisfaction of applicable
SIP emission limitations and non-interference with SIP’s
attainment date) have been met and publish the determination in
the Federal Register—“tightened” Part VI’s requirements. EPA
Br. at 103. The new pre-condition simply formalized and made
explicit what was already implicitly required—that the EPA find
the existing two pre-conditions satisfied. Thus, the additional
 45

pre-condition produced no “tightening” effect; on the contrary,
it did nothing to counteract the indisputably relaxing effect of
removing the 18-month limit.

VII. Baton Rouge’s Reformulated Gasoline Requirement

The Chamber of Greater Baton Rouge and affiliated
petitioners (Chamber) contend the EPA arbitrarily and
capriciously concluded that the Baton Rouge, Louisiana area
(Baton Rouge) is a “covered area” under CAA section 211(k),
42 U.S.C. § 7545(k), and therefore subject to a statutory ban on
the sale to consumers of conventional gasoline.14 We conclude
that the EPA correctly determined the statutory definition of
“covered area” is ambiguous as to Baton Rouge under Chevron
step 1 and that it reasonably resolved the ambiguity under
Chevron step 2.

CAA section 211(k)(1) authorizes the EPA to “promulgate
regulations . . . establishing requirements for reformulated
gasoline to be used in gasoline-fueled vehicles in specified
nonattainment areas.” 42 U.S.C. § 7545(k)(1). Section
211(k)(5) makes unlawful “[t]he sale or dispensing by any
person of conventional gasoline to ultimate consumers in any
covered area”—necessitating the use of reformulated gasoline
instead. Id. § 7545(k)(5)(A). Section 211(k)(10), in turn,
designates as a “covered area” each of nine specific, high ozone
areas and further provides: “Effective one year after the
reclassification of any ozone nonattainment area as a Severe
ozone nonattainment area under section 7511(b) of this title,

14
Section 211(k) defines “conventional gasoline” as “any
gasoline which does not meet specifications set by a certification
under [section 211(k)]” as “reformulated gasoline.” 42 U.S.C.
§ 7545(k)(10)(E)-(F).
 46

such Severe area shall also be a ‘covered area’ for purposes of
this subsection.” Id. § 7545(k)(10)(D).

In 1991, the EPA classified Baton Rouge as a “serious”
ozone nonattainment area under the 1-hour standard. On April
24, 2003, the EPA reclassified Baton Rouge as “severe,”
effective June 23, 2003, pursuant to the bump-up provision,
CAA § 181(b)(2), 42 U.S.C. § 7511(b)(2), because Baton Rouge
failed to reach attainment by its November 15, 1999 attainment
date. Notice of Withdrawal of October 2, 2002, Attainment
Date Extension, Determination of Nonattainment as of
November 15, 1999, and Reclassification of the Baton Rouge
Ozone Nonattainment Area, 68 Fed. Reg. 20,077 (Apr. 24,
2003) (effective June 23, 2003). At that time, the EPA advised
that “under Section 211(k) of the Act the use of reformulated
gasoline (RFG) will be required in the Baton Rouge area
beginning one year from the effective date of this rule,” that is,
as of June 23, 2004. Id. at 20,080. On April 30, 2004, the EPA
designated Baton Rouge as nonattainment under the 8-hour
standard and classified it as “marginal,” effective June 15, 2004.
Air Quality Designations and Classifications for the 8-Hour
Ozone National Ambient Air Quality Standards; Early Action
Compact Areas With Deferred Effective Dates, 69 Fed. Reg.
23,858, 23,907 (Apr. 30, 2004).

The Chamber filed a request for extension of the deadline
to commence using RFG or waiver of the RFG requirement,
which the EPA denied in a letter dated May 5, 2004. The
Chamber then petitioned the United States Court of Appeals for
the Fifth Circuit for review of the denial and the Fifth Circuit
stayed the RFG deadline on June 18, 2004. On August 2, 2004
the Fifth Circuit granted a joint motion to stay its proceedings
“pending the outcome of the administrative decision making
process,” with the RFG deadline stay to remain in effect during
 47

remand. City of Baton Rouge v. EPA, No. 04-60408 (5th Cir.
Aug. 2, 2004).

In the Phase 2 Rule, the EPA concluded that an area such as
Baton Rouge, which automatically became a “covered area”
(subject to the RFG requirement) on the one-year anniversary of
its bump-up to “severe,” should remain subject to the RFG
requirement notwithstanding its classification to a less-than-
severe status under the 8-hour standard and the subsequent
revocation of the 1-hour standard: “EPA has determined that
bump-up areas that lose their severe classification based solely
on revocation of the 1-hour NAAQS should remain RFG
covered areas at least until they are redesignated to attainment
for the 8-hour NAAQS.” 70 Fed. Reg. at 71,686.15 The EPA
concluded that “section 211(k)(10)(D) is ambiguous on the issue
of whether a bump-up area continues to be a covered area when
it is no longer classified as severe” and that therefore “EPA has
discretion to determine whether section 211(k)(10)(D)
authorizes removal of a bump-up area from the RFG program
when it is no longer classified as severe, and to set appropriate
criteria for such removal.” Id. The EPA explained that it “d[id]
not believe that Congress would have intended that removal of
the severe classification based solely on revocation of the less
protective 1-hour NAAQS should result in backsliding of the
RFG requirement.” Id. The Agency further found it
“instructive” that “if EPA had never revised the 1-hour
NAAQS[,] . . . the area would continue to be a covered area at

15
The EPA stated it was “reserving for future consideration what
RFG requirements, if any, should apply to the nine mandatory areas
and the bump-up areas covered by this final rule when they are
redesignated to attainment for the 8-hour NAAQS.” 70 Fed. Reg. at
71,687. The EPA further noted its determination did “not change or
affect any discretion EPA may otherwise have under the RFG
provisions to modify or remove RFG requirements.” Id. at 71,686.
 48

least until it was redesignated to attainment for the 1-hour
NAAQS.” Id. Accordingly, because “[h]ere, the removal of the
severe classification is through revocation of the 1-hour
NAAQS, not through redesignation to 1-hour attainment,” the
EPA determined that “the removal of the severe classification
for these areas as a result of revocation of the 1-hour standard
should not lead to removal of the RFG requirement.” Id. The
Chamber offers three arguments against the EPA’s
interpretation. We address and reject each argument in turn.

First, the Chamber contends that at Chevron step 1 Baton
Rouge never became a “covered area” because the plain and
unambiguous language of the statutory definition requires that
a “covered area” be “a ‘Severe area’ that remains severe ‘one
year after the reclassification,’ ” Chamber Br. at 15 (quoting
CAA § 211(k)(10)(D), 42 U.S.C. § 7545(k)(10)(D)), that is, on
June 23, 2004, on which date Baton Rouge was no longer a
severe area as it had become a marginal area on June 15, 2004.
The EPA does not quibble with the Chamber’s interpretation of
the statutory language to require that the area remain classified
as “severe” on the one-year anniversary of the reclassification.
See EPA Br. at 111 (“[O]nce an area is reclassified as ‘severe,’
the only other precondition to becoming ‘covered’ is that the
area remain in severe nonattainment status for one year.”)
(emphasis in original). The EPA points out, however, that as of
the one-year anniversary of the bump-up—June 23,
20004—Baton Rouge was not only a marginal area under the 8-
hour standard but was also a severe area under the 1-hour
standard, which was not revoked until June 15, 2005, almost a
year later. See Phase 1 Rule, 69 Fed. Reg. at 23,954 (“We will
revoke the 1-hour standard in full, including the associated
designations and classifications, 1 year following the effective
date of the designations for the 8-hour NAAQS.”). Thus, the
EPA reasonably determined that under the plain language of the
 49

statute, Baton Rouge was a severe area as of June 23, 2004 and
became a “covered area” on that date.16

Second, the Chamber asserts at Chevron step 2 that even if
the definition of “covered area” is ambiguous, the EPA resolved
the ambiguity unreasonably in two respects. The Chamber first
argues the EPA improperly relied on the Supreme Court’s
decision in Whitman v. Am. Trucking Ass’ns, 531 U.S. 457
(2001). In support of its interpretation, the EPA cited
Whitman’s “caution[] . . . against EPA making subpart 2
‘abruptly obsolete,’ ” 70 Fed. Reg. at 71,686 (quoting Whitman,
531 U.S. at 485), observing that, “[a]lthough the RFG
requirement itself is not set forth in subpart 2, the requirement
to use it in severe bump-up areas is tied directly to the
classifications that arise by operation of subpart 2” and therefore
“it would appear that the Supreme Court's caution should be as
relevant for RFG bump-up areas as it is for the subpart 2 control
obligations.” Id. Notwithstanding the Chamber’s contrary
assertions, the EPA reasonably determined that “the inclusion of
a bump-up area in the RFG program is integrally tied to the

16
The Chamber argues that this reasoning is foreclosed because
the EPA itself “stated that the 8-hour NAAQS is the ‘relevant’
standard.” Chamber Br. at 18 (quoting Phase 1 Rule, 69 Fed. Reg. at
23,983). The EPA used the term “relevant,” however, in explaining
why it had decided that “control obligations an area is required to
retain in the approved SIP for an area’s 1-hour classification must
continue to be implemented under the SIP until the area attains and is
redesignated to attainment for the 8-hour NAAQS.” 69 Fed. Reg. at
23,983; see id. (“Since the relevant NAAQS is now the 8-hour
NAAQS, we believe it is appropriate to require these mandated
controls to remain as part of the implemented SIP until an area attains
the 8-hour NAAQS and is redesignated to attainment.”). Thus, the
cited language in fact supports retaining the RFG control based on the
one-hour standard bump-up until Baton Rouge reaches 8-hour
attainment.
 50

subpart 2 provisions that establish the original classification and
attainment date for an area and its later reclassification,” id.,
which provisions cause an area to become a “covered” RFG area
by operation of law. Accordingly, the EPA reasonably relied on
Whitman’s general admonition to preserve Subpart 2 to support
the Agency’s particular determination that the effects of the
integrally connected provisions should remain intact following
conversion to the 8-hour standard.

The Chamber also asserts the EPA’s interpretation falters at
Chevron step 2 because the Agency unreasonably relied on “the
very same provisions it referred to in the Phase 1 Rule at 69 Fed.
Reg. 23972,” which provisions, the Chamber maintains, “did not
support the Phase 1 Rule, nor do they support the imposition of
RFG.” Chamber Br. at 22-23 (citing 70 Fed. Reg. at 71,686).
The Chamber examines each of the statutes cited in the Phase 1
Rule and asserts, correctly, that none of them directly relates to
the RFG requirement. See id. at 23-24. The Chamber makes too
much, however, of the EPA’s citation to the Phase 1 Rule. The
EPA was simply referring readers “[f]or further discussion of
the reasoning behind anti-backsliding provisions in the Phase 1
Rule” because the Agency also relied on the anti-backsliding
rationale for its interpretation of the “covered area” definition in
the Phase 2 Rule.17 See 70 Fed. Reg. at 71,686.

17
The Chamber also objects in particular to the EPA’s reliance on
CAA section 172(e), the anti-backsliding provision. See Chamber Br.
at 23-24; Chamber Reply Br. at 2-3. The Chamber argues section
172(e) applies only when the EPA “relaxes” a standard and that, in
promulgating the 8-hour standard, the EPA in fact strengthened, rather
than relaxed, the standard. The EPA’s reliance on section 172(e) in
the Phase 1 Rule, however, supported its specific policy against
backsliding in converting to the 8-hour standard notwithstanding the
8-hour standard is stricter (not more relaxed) than the one-hour
standard. We have already upheld this policy as reasonable. See S.
Coast, 472 F.3d at 900.
 51

Having rejected the Chamber’s contrary arguments, we
conclude that the EPA’s statutory interpretation readily satisfies
Chevron step 2. In upholding the EPA’s interpretation as
reasonable, we, like the Agency, find it “instructive” that but for
the happenstance of the EPA’s adoption of 8-hour standards, the
Chamber would have no statutory basis to dispute that Baton
Rouge is a covered RFG area under the mandatory bump-up and
RFG provisions because Baton Rouge remained a severe one-
hour area for one year following its classification as such. See
70 Fed. Reg. at 71,686 (quoted supra p. 47).

Finally, the Chamber contends the EPA was arbitrary and
capricious in not making a specific determination that RFG
would “protect the public health” in Baton Rouge and asserts
that in fact RFG will interfere with the area’s progress toward
attainment, thereby harming the public health. The Chamber
first asserts that the EPA was required by CAA section
109(b)(1) to make a specific finding that requiring RFG
“protects the public health” of Baton Rouge, citing 42 U.S.C.
§ 7409(b)(1) (“National primary ambient air quality standards,
prescribed under subsection (a) of this section shall be ambient
air quality standards the attainment and maintenance of which
in the judgment of the Administrator, based on such criteria and
allowing an adequate margin of safety, are requisite to protect
the public health.”). The Chamber further argues that requiring
RFG in Baton Rouge will interfere with the area’s attainment
because studies have shown RFG increases NOx levels and
Baton Rouge’s SIP strategy for reducing ozone relies largely on
reducing NOx (and not VOCs). By interfering with attainment,
the Chamber contends, RFG in fact endangers the public health.
As an initial matter, the Chamber’s objections based on harm to
public health and interference with RFP and attainment are
forfeited because they were not raised before the Agency. See
CAA § 307(d)(7)(B), 42 U.S.C. § 7607(d)(7)(B) (“Only an
objection to a rule or procedure which was raised with
 52

reasonable specificity during the period for public comment
(including any public hearing) may be raised during judicial
review.”). To the extent Baton Rouge argues that RFG will not
improve the public health, the EPA adequately addressed these
objections and reasonably determined they may best be resolved
through individual waiver requests.18

In the Phase 2 Rulemaking, the Baton Rouge Clean Air
Coalition asserted that RFG use in Baton Rouge would “provide
no measurable benefits for NOx,” “less than 2 tons per day of
VOC reductions, and “an ozone benefit . . . of around 0.26 ppb,”
and asked whether these data “qualify as an ‘absurd result’ and
[would] be subject to consideration for waiver.” Letter from
Mike D. McDaniel, Executor Director, Baton Rouge Clean Air
Coalition to U.S. EPA at 4 (July 30, 2003); see Phase 2 Rule, 70
Fed. Reg. at 71,687-88. In response, the EPA noted that Baton
Rouge had “submitted requests for an RFG waiver and for a
waiver of the RFG oxygen content requirement, which are
currently before the Agency.” 70 Fed. Reg. at 71,688. Based on
its determination that “the transition to the more protective
8-hour standard should [not] result in less restrictive
requirements for RFG than would apply if the EPA had never
revised the 1-hour standard,” the EPA advised that “[t]he
appropriate mechanism to address Baton Rouge’s concerns is
. . . in the context of Baton Rouge’s petitions for relief under the
RFG program, and not by establishing different, less restrictive
RFG requirements as part of the transition to the 8-hour
standard.” Id. Baton Rouge’s waiver request is ongoing before
the Agency with judicial review available in the Fifth Circuit.
The EPA’s decision to address site-specific data and concerns in

18
As the EPA notes, the Chamber submitted material originating
in the Fifth Circuit proceeding to support these objections but the
material was not submitted during the proceeding before the Agency.
See EPA Br. at 116.
 53

an individual waiver proceeding, rather than in the general
rulemaking, is reasonable. Cf. Nat’l Wildlife Fed’n v. EPA, 286
F.3d 554, 566 (D.C. Cir. 2002) (upholding as reasonable EPA’s
decision to address “color pollution” “on a case-by-case basis
through individual [National Pollutant Discharge Elimination
System] permits or, when appropriate, through local limits”
rather than “to establish nationwide standards for discharge of
‘color’ ”).

VIII. Conclusion
In sum, we hold (1) the Phase 2 Rule violates the Act by
allowing participation in the NOX SIP Call to satisfy the
requirement that a nonattainment area mandate such reductions
as can be achieved by the application of RACT, (2) the EPA
acted arbitrarily when it eliminated the requirement that an
attainment demonstration be approved for an area before a new
source would be allowed to use a past emission reduction to
offset new emissions, and (3) the elimination of the 18-month
time limit for NSR waivers under Appendix S violates the
anti-backsliding provision of the Act. In light of the likelihood
the EPA will be unable to cure the backsliding inherent in the
removal of the 18-month time limit, we vacate that provision of
the Phase 2 Rule and remand it to the EPA. See Ill. Pub.
Telecomms. Ass’n v. FCC, 123 F.3d 693, 693 (D.C. Cir. 1997)
(“When this court remands a rule ... with little or no prospect of
the rule’s being readopted upon the basis of a more adequate
explanation of the agency’s reasoning, the practice of the court
is ordinarily to vacate the rule”). We remand to the EPA
without vacatur the other two provisions of the Phase 2 Rule we
hold invalid. We defer consideration of the Phase 2 Rule and
Reconsideration Notice insofar as they relate to the CAIR
program, and we deny the petitions in all other respects.
So ordered.
 ROGERS, Circuit Judge, concurring in part and dissenting in
part. I join the court’s opinion regarding challenges to the Phase
2 Rule1 except in three respects. The petitioners met their
burden to show reopening of both the general policies allowing
new and modified sources to use offset credits from past
emission reductions to meet the requirements of section 173 of
the Clean Air Act (“CAA”), 42 U.S.C. § 7503, and EPA’s
authority to waive interim new source review (“NSR”)
Appendix S permitting requirements, 40 C.F.R. pt. 51, app. S, §
VI. Additionally, the challenge to the contingency measure
element of the Clean Data Policy was not forfeited because
comments submitted by the NRDC and other environmental
groups put EPA on notice of the challenge, which EPA, in fact,
addressed.

I.

The principles underlying the reopening doctrine are well-
established. It is only their application that may obfuscate the
reality of what an agency regulatory action entails. This is the
situation here.

Under the reopening doctrine, which is an exception to the
statutory limits on the period for seeking judicial review of
agency action, see Env. Defense v. EPA, 467 F.3d 1329, 1333
(D.C. Cir. 2006), an issue is reopened when an agency implicitly
or explicitly, see West Virginia v. EPA, 362 F.3d 861, 872 (D.C.

1
Final Rule To Implement the 8-Hour Ozone National
Ambient Air Quality Standard—Phase 2; Final Rule to Implement
Certain Aspects of the 1990 Amendments Relating to New Source
Review and Prevention of Significant Deterioration as They Apply in
Carbon Monoxide, Particulate Matter and Ozone NAAQS; Final Rule
for Reformulated Gasoline, 70 Fed. Reg. 71,612, 71,677 (Nov. 29,
2005) (“Phase 2 Rule”).
 2

Cir. 2004), solicits comment on a preexisting regulation or
“otherwise indicates its willingness to reconsider such a
regulation by inviting and responding to comments,” Kennecott
Utah Copper Corp. v. Dep’t of Interior, 88 F.3d 1191, 1213
(D.C. Cir. 1996). The “entire context” must “demonstrate[] that
the agency has undertaken a serious, substantive reconsideration
of the [existing] rule.” P & V Enters. v. U.S. Army Corps of
Eng’rs, 516 F.3d 1021, 1024 (D.C. Cir. 2008) (internal quotation
marks omitted) (alteration in original). Also, when a new rule
“increase[s] the significance” of preexisting regulations,
Kennecott, 88 F.3d at 1216, and thus “significantly alters the
stakes of judicial review,” the new rule “constructively reopens”
the previously settled issue, id. at 1227. Applying these
principles demonstrates EPA reopened consideration of two
subjects.

A.
Offset credits. The Phase 2 Rule allows new and modified
sources to meet emissions offset requirements under CAA §
173(c), 42 U.S.C. § 7503(c), by using credits from sources that
shut down or curtailed operations as long ago as August 7, 1977.
See 40 C.F.R. § 51.165(a)(3)(ii)(C)(1)(ii); Op. at 28-31.
Contrary to the court’s suggestion, this court’s opinion in
Kennecott makes clear that the fact that the “basic regulatory
scheme remains unchanged,” Op. at 33, is not dispositive of
whether an issue has been constructively reopened. See
Kennecott, 88 F.3d at 1219. Analyzing the reopening doctrine
at that level of generality begs the question. If “adding new
terms to [an] old rule” can result in reopening, see id., then so
can removing old terms from an old rule, at least where those
old terms were an important part of the underlying regulatory
scheme. Furthermore, Kennecott also makes clear that the
question whether the stakes are “quantitatively different,” Op.
at 34, does not turn on whether the petitioner or another failed
to object to a different change earlier. See Kennecott, 88 F.3d at
 3

1226-27. Although a 1989 Rule made it easier to use pre-
application offset credits, see Requirements for Implementation
Plans, Air Quality New Source Review, Final Rule, 54 Fed. Reg.
27,286, 27,292 (June 28, 1989) (“1989 Rule”), the concomitant
addition of the attainment demonstration requirement mitigated
the impact of the change and may well have meant it was not
“worth challenging,” Kennecott, 88 F.3d at 1227, the regime as
a whole. However, once EPA eliminated the attainment
demonstration requirement and failed to replace it with a
comparable substitute, the need for a challenge by
environmental groups such as the NRDC arose because EPA had
“dramatically expanded authority to use [pre-application offset]
credits,” NRDC Br. at 44. This change meant that “a major new
pollution source [may] add substantial new emissions to a
nonattainment area without any contemporaneous offsets, even
when the state has no plan for meeting standards and no idea of
what total reductions are needed to attain standards.” Id. at 45.

Both EPA’s and the court’s analysis show that a
constructive reopening occurred. See Op. at 34-36. In
announcing the 1989 Rule, EPA stated that to comply with the
reasonable further progress (“RFP”) requirement of CAA §
173(a)(1)(A), 42 U.S.C. § 7503(a)(1)(A), the restriction on pre-
application offset credit was “necessary both to assure RFP and
to guarantee that a new source does not cause or contribute to a
violation of the NAAQS,” 1989 Rule, 54 Fed. Reg. at 27,292.
This evaluation by EPA supports the conclusion that
environmental groups would have determined in 1989 that it
was not worth challenging the regulatory scheme as significant
limits remained in place. Once EPA removed this “necessary,”
id., limitation “without adding any other safeguard to ensure that
issuing a particular permit is consistent with section 173’s
mandate that total reductions ‘represent . . . reasonable further
progress,’” Op. at 36 (quoting CAA § 173(a)(1)(A), 42 U.S.C.
§ 7503(a)(1)(A)), the stakes of judicial review were significantly
 4

altered. The court’s speculation today — that it “could have
been ‘reasonably anticipated’ in 1989 that EPA might eliminate
the attainment demonstration requirement,” Op. at 34 — proves
too much. Given EPA’s evaluation in the 1989 Rule of the
importance of restricting the availability of pre-application
offset credits, environmental groups had no reason then to lodge
a challenge based on the possibility that EPA might, perhaps, at
some point in the future reverse course. Cf. S. Coast Air Quality
Mgmt. Dist. v. EPA, 472 F.3d 882, 898 (D.C. Cir. 2006).
Although the reopening doctrine may not apply if the change
could have been “reasonably anticipated,” see Env. Defense, 467
F.3d at 1333, such a limit on the doctrine does not go so far as
preventing review because a petitioner might have guessed EPA
could later change its evaluation of what was necessary. Such
an approach would render the constructive reopening precedents
nugatory.

On the merits, the NRDC persuasively contends that
“EPA’s final rule . . . unlawfully permits total allowable
emissions after the new source starts operating to be higher than
emissions from existing sources before the new source applied
for its permit.” NRDC Br. at 42-43. The Phase 2 Rule’s
allowance of offset credit for pre-application shutdowns and
curtailments is contrary to the plain text of the CAA and fails
under Chevron step one. Chevron U.S.A. Inc. v. Natural Res.
Def. Council, Inc., 467 U.S. 837, 842-43 (1984). Section
173(a)(1)(A) requires that “sufficient offsetting reductions” shall
be obtained “such that total allowable emissions from existing
sources in the region, from new or modified sources which are
not major emitting facilities, and from the proposed source will
be sufficiently less than total emissions from existing sources .
. . prior to the application for such permit to construct or modify
so as to represent . . . reasonable further progress.” 42 U.S.C. §
7503(a)(1)(A). Section 173(c)(1) requires that offsets come
from “an equal or greater reduction, as applicable, in actual
 5

emissions of such air pollutant from the same or other sources in
the area.” Id. § 7503(c)(1) (emphasis added); see also H.R. Rep.
No. 101-490, pt. 1 (1990) (referring to reductions in “actual
emissions”). Under the Phase 2 Rule, new and modified major
sources may claim offset credit from reductions in emissions
resulting from the shutdown or curtailment of operations from
more than three decades ago. The absence of any meaningful
time limit means the Rule is inconsistent with CAA §
173(c)(1)’s requirement that new sources’ emissions “shall be
offset” by an equal or greater reduction in “actual emissions,” 42
U.S.C. § 7503(c)(1). Congress is presumed, absent any
indication to the contrary, to have intended the ordinary meaning
of the word “actual,” as “existing or occurring at the time,”
MERRIAM WEBSTER’S COLLEGIATE DICTIONARY (10th ed. 1993)
(emphasis added); accord WEBSTER’S THIRD NEW INT’L
DICTIONARY (1993). See Engine Mfrs. Ass’n v. S. Coast Air
Quality Mgmt. Dist., 541 U.S. 246, 252 (2004). Congress has
given no indication in the surrounding text, or elsewhere, that it
intended to depart from the common meaning of “actual.”
Hence, allowing an offset for no-longer-existing emissions from
sources closed decades ago is inconsistent with the statutory
text.

EPA offers that because the Phase 2 Rule requires the old
offsets to be included in the area’s emissions inventory,
allowance of pre-application offset credits is consistent with the
requirement in CAA § 173(a)(1)(A), 42 U.S.C. § 7503(a)(1)(A),
that offsets ensure reasonable further progress toward
attainment. See Phase 2 of the Final Rule to Implement the 8-
Hour Ozone National Ambient Air Quality Standard, Final
Notice of Reconsideration, 72 Fed. Reg. 31,727, 31,742/3 (June
8, 2007) (“Final Notice of Reconsideration”). But using
decades-old emissions in an accounting exercise cannot render
those emissions “actual” as that word is commonly understood.
Likewise, EPA offers that because “[f]rom an air quality
 6

planning perspective” emissions from the previously shutdown
or curtailed sources had impacted the area’s designations as
nonattainment, they are “actual emissions reductions.” Phase 2
Rule, 70 Fed. Reg. at 71,677; see also Final Notice of
Reconsideration, 72 Fed. Reg. at 31,742/3. But that emissions
were “actual” as long ago as the 1970s does not mean they
continue to meet the common understanding of “actual” years
later when a source applies for a construction or modification
permit.

Because the flaws in this part of the Phase 2 Rule, see 40
C.F.R. § 51.165(a)(3)(ii)(C)(1)(ii), cannot be remedied by
further explanation by EPA, it must be vacated. See Allied-
Signal, Inc. v. U.S. Nuclear Reg. Comm’n, 988 F.2d 146, 150-51
(D.C. Cir. 1993). To what extent the court’s holding that
elimination of the requirement of an approved attainment
demonstration is arbitrary and capricious under CAA §
307(d)(9)(A), 42 U.S.C. § 7607(d)(9)(A), Op. at 36, will result
in a different outcome than vacatur remains to be seen upon
remand.

B.
NSR waiver authority. Section 110(a)(2)(C), 42 U.S.C. §
7410(a)(2)(C), does not specify what NSR requirements apply
in the period between an area’s designation as nonattainment
and approval of its NSR state implementation plan (“SIP”).
Prior to promulgating the Phase 2 Rule, EPA had filled this
statutory gap by retaining Appendix S’s interim NSR regime,
which by its terms could apply for only 18 months. Given the
limited duration of its effect, the previous regulation “may not
have been worth challenging,” Kennecott, 88 F.3d at 1227.
However, EPA’s decision to eliminate the 18-month limitation
“alter[ed] the stakes of judicial review,” id. By parity of
reasoning, if there was a reopening in Kennecott where industry
petitioners faced a new requirement creating the possibility of
 7

the imposition of additional penalties, then there was a
constructive reopening where EPA’s change meant NSR
controls could be postponed indefinitely, thereby dramatically
changing the scope and effect of the waiver provision.

Additionally, EPA constructively reopened the issue of its
NSR waiver authority when it proposed to remove the general
waiver provision in Appendix S and replace it with a
“transitional” program in which waivers would be available in
only a limited set of nonattainment areas. See Proposed Rule to
Implement the 8-Hour Ozone National Ambient Air Quality
Standard, 68 Fed. Reg. 32,802, 32,846-48 (June 2, 2003)
(“NPRM”). EPA has described the proposal as a “major rewrite
of [A]ppendix S.” Phase 2 Rule, 70 Fed. Reg. at 71,674. In
explaining the 2003 proposal, EPA stated “we do not believe
that areas not meeting the transitional approach would be able
to ensure that they were implementing an NSR program ‘as
necessary’ to ensure the attainment of NAAQS . . . .” NPRM,
68 Fed. Reg. at 32,848 (quoting CAA § 110(a)(2)(C), 42 U.S.C.
§ 7410(a)(2)(C)). In other words, EPA expressed doubt about
a broad waiver’s consistency with statutory mandates in certain
circumstances. Although it ultimately rejected the proposed
alternative, EPA reopened the issue of its authority to waive
interim NSR requirements by raising the question whether a
broad waiver would be inconsistent with the CAA. See Edison
Elec. Inst. v. EPA, 996 F.2d 326, 332 (D.C. Cir. 1993).

On the merits, the challenge by the NRDC and the State of
New Jersey to EPA’s authority to exempt new sources from the
interim NSR permitting requirements is unpersuasive. They
contend that “[w]here Congress has intended to allow exceptions
to NSR requirements, it has expressly said so,” NRDC Br. at 38
(citing, inter alia, CAA § 173(a)(1)(B), 42 U.S.C. §
7503(a)(1)(B), excepting economic development zones).
However, EPA persuasively responds that the waivers are the
 8

product of EPA’s gap-filling authority. See Chevron, 467 U.S.
at 843-45. In exercising that authority, EPA established
requirements “substantially similar to the requirements of part
D,” CAA § 171 et seq., 42 U.S.C. § 7501 et seq., although not
identical to the “major NSR” program under Part D. See Phase
2 Rule, 70 Fed. Reg. at 71,678/1. The “minor NSR” program
under CAA § 110(a)(2)(C) calls for “regulation of the
modification and construction of any stationary source within
the areas covered by the plan as necessary to assure that
[NAAQS] are achieved.” 42 U.S.C. § 7410(a)(2)(C) (emphasis
added). EPA thus could reasonably conclude that the interim
NSR requirements are inapplicable when the exemption
conditions are met. In those instances, a waiver would be
available only if the new source’s emissions would not interfere
with timely attainment, see 40 C.F.R. pt. 51, app. S, § VI,
making additional NSR requirements unnecessary.

Although EPA had authority to provide waivers on a case-
by-case basis where a waiver would not interfere with timely
attainment, I concur in holding that EPA’s elimination of the 18-
month limitation violates the anti-backsliding provision in CAA
§ 172(e), 42 U.S.C. § 7502(e), see Op. at 43-45.

II.

The Clean Data Policy, originally published in 1992 and
revised in 1995, excuses state planning requirements relating to
RFP, attainment demonstrations, and contingency measures for
areas attaining the NAAQS but not yet redesignated as
attainment areas. In commenting in 2003 on the proposed
codification of this statutory interpretation, environmental
groups, including the NRDC, stated that the “EPA cannot
authorize states to simply drop subpart 2 measures when the area
is meeting either [the 1-hour or 8-hour] standard.” Clean Air
Task Force, et al., Comments at 48. By arguing that “[t]he
 9

[CAA] allows states to move mandated controls to a
maintenance contingency plan, [CAA § 175A, 42 U.S.C.
§ 7505a], but only after the area has been redesignated to
attainment,” id. (emphasis added), environmental groups put
EPA on notice of its view that until an area is designated as
being in attainment, a state plan must continue to contain the
antecedent requirements, such as contingency measures,
applicable in nonattainment areas.

Contrary to the court’s holding that the NRDC’s challenge
to the contingency plan aspect of EPA’s Clean Data Policy is
forfeited, Op. at 20-21, the rulemaking record evidences that
comments put EPA on notice of that objection. First, these
comments raised the objection to the Clean Data Policy with
respect to the contingency measures requirement of Subpart 2
“with reasonable specificity during the period for public
comment,” CAA § 307(d)(7)(B), 42 U.S.C. § 7607(d)(7)(B), and
sufficiently put EPA on notice of the objection. Second, the
comment did not make a mere “general [challenge] to EPA’s
approach,” Mossville Envt’l Action Now v. EPA, 370 F.3d 1232,
1238 (D.C. Cir. 2004) (internal quotation marks omitted)
(alteration in original), but instead challenged application of a
specific EPA policy — the Clean Data Policy — to a specific set
of statutory requirements — those in Subpart 2 of the CAA,
CAA § 181 et seq., 42 U.S.C. § 7511 et seq. The comment
therefore constituted “adequate notification of the general
substance of the complaint,” S. Coast Air Quality Mgmt. Dist. v.
EPA, 472 F.3d 882, 891 (D.C. Cir. 2006), including an objection
regarding antecedent contingency measures in particular. Third,
that other objections to the Clean Data Policy addressed only the
requirements for RFP and attainment demonstrations does not
render the broader objection insufficient to preserve for judicial
review the objection to elimination of the antecedent contingency
measures. Fourth, the comment that the “EPA’s ‘clean data
policy,’ 68 Fed. Reg. at 32,835/3, is unlawful with respect to
 10

both the 1-hour and 8-hour NAAQS for the reasons set forth . .
. above [regarding RFP and the attainment demonstration],”
Clean Air Task Force, et al., Comments at 100, shows that the
objection was not limited to any particular aspect of the Clean
Data Policy. Fifth, EPA’s response to the comments addressed
contingency measures, among other requirements, further
indicating the comments were sufficient to put EPA on notice of
the objection. See Phase 2 Rule, 70 Fed. Reg. at 71,645/1,
71,646/1; cf. Appalachian Power Co. v. EPA, 135 F.3d 791, 817-
18 (D.C. Cir. 1998).

Nonetheless, although the NRDC’s challenge is properly
before the court, it fails on the merits. Section 172(c)(9)
provides that SIPs in nonattainment areas “shall provide for
implementation of specific measures to be undertaken if the area
fails to make reasonable further progress, or to attain the
[NAAQS] by the attainment date applicable under this part.” 42
U.S.C. § 7502(c)(9) (emphasis added). The NRDC contends that
“[t]he [Clean Data] [P]olicy waives mandatory pollution control
requirements based solely on an area’s current air quality,
without requiring the area to also show — as mandated by
Congress — that it has met all of the [CAA]’s applicable
requirements . . . .” NRDC Br. at 34. However, EPA’s view that
CAA §§ 172(c)(9) and 182(c)(9), 42 U.S.C. §§ 7502(c)(9),
7511a(c)(9), are written in a manner that ties the antecedent
contingency measures to the requirements for RFP and an
attainment demonstration, such that all are inapplicable upon
attainment, see Phase 2 Rule, 70 Fed. Reg. at 71,645, is
consistent with the plain text and to the extent there may be
ambiguity, it is a reasonable interpretation. The Clean Data
Policy requires that areas have “three consecutive years of clean
air quality monitoring data demonstrating attainment of the
ozone standard” before EPA can determine areas have attained.
Memorandum from John S. Seitz, Director, Office of Air Quality
Planning and Standards, EPA, to Various EPA Directors 5 (May
 11

10, 1995). Also, if EPA determines an area covered by the Clean
Data Policy has violated the NAAQS prior to formal designation
of attainment, the Clean Data Policy no longer applies to that
area. See 40 C.F.R. § 51.918.