Nat Resrc Def Cncl v. EPA

• Court: Court of Appeals for the D.C. Circuit
• Decided: July 10, 2009
• Docket: 06-1045
• Status: Published

Opinion

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.