In the Supreme Court of the United States
ENVIRONMENTAL DEFENSE, ET AL., PETITIONERS
DUKE ENERGY CORPORATION, ET AL.
ON WRIT OF CERTIORARI
TO THE UNITED STATES COURT OF APPEALS
FOR THE FOURTH CIRCUIT
BRIEF FOR THE UNITED STATES AS
RESPONDENT SUPPORTING PETITIONERS
PAUL D. CLEMENT
Counsel of Record
SUE ELLEN WOOLDRIDGE
Assistant Attorney General
THOMAS G. HUNGAR
Deputy Solicitor General
JAMES A. FELDMAN
Assistant to the Solicitor
KATHERINE J. BARTON
JASON A. DUNN
Department of Justice
Washington, D.C. 20530-0001
ANN R. KLEE
CHET M. THOMPSON
Deputy General Counsel
GRANTA Y. NAKAYAMA
THOMAS W. SWEGLE
CAROL S. HOLMES
DAVID W. SCHNARE
United States Environmental Protection Agency
Washington, D.C. 20460
In this civil enforcement action under the Prevention of Significant Deterioration (PSD) provisions of the Clean Air Act, 42 U.S.C. 7401 et seq., the court of appeals held that the Environmental Protection Agency (EPA) regulations con struing the statutory term "modification" for purposes of the PSD program must be given the same meaning as EPA's regulations construing that term for purposes of the separate New Source Performance Standards (NSPS) program. The questions presented are:
1. Whether the court of appeals' decision contravenes Section 307 of the Clean Air Act, 42 U.S.C. 7607, which pro vides that nationally applicable regulations issued by EPA to implement the Clean Air Act may be reviewed only through properly filed petitions for review in the D.C. Circuit, not in an enforcement action.
2. Assuming the answer to the first question is no, the following question is presented: whether the Clean Air Act, which defines "modification" in both the PSD and NSPS programs as a physical or operational change that "increases" emissions of pollutants, requires EPA to measure emission "increases" under the PSD program in the same manner as it measures emission "increases" under the NSPS program.
In the Supreme Court of the United States
ENVIRONMENTAL DEFENSE, ET AL., PETITIONERS
DUKE ENERGY CORPORATION, ET AL.
ON WRIT OF CERTIORARI
TO THE UNITED STATES COURT OF APPEALS
FOR THE FOURTH CIRCUIT
BRIEF FOR THE UNITED STATES AS
RESPONDENT SUPPORTING PETITIONERS
The opinion of the court of appeals (Pet. App. 1a-19a) is reported at 411 F.3d 539. The opinion of the district court (Pet. App. 22a-84a) is reported at 278 F. Supp. 2d 619.
The judgment of the court of appeals was entered on June 15, 2005. A petition for rehearing was denied on August 30, 2005 (Pet. App. 20a-21a). On November 17, 2005, the Chief Justice extended the time within which to file a petition for a writ of certiorari to and including December 28, 2005, and the petition was filed on that date. The petition for a writ of cer tiorari was granted on May 15, 2006. The jurisdiction of this Court rests on 28 U.S.C. 1254(1).
STATUTORY AND REGULATORY
The relevant statutory and regulatory provisions are set forth in an appendix to this brief. App., infra, 1a-23a.
This case is a civil enforcement action brought by the United States against respondent Duke Energy for illegally modifying generating plants without complying with the Pre vention of Significant Deterioration (PSD) program of the Clean Air Act (CAA), 42 U.S.C. 7401 et seq.
1. The CAA was enacted "to protect and enhance the quality of the Nation's air resources so as to promote the pub lic health and welfare and the productive capacity of its popu lation." 42 U.S.C. 7401(b)(1). It directs the U.S. Environmen tal Protection Agency (EPA) to promulgate National Ambient Air Quality Standards (NAAQS) specifying allowable concen trations of air pollutants. 42 U.S.C. 7408-7409; Whitman v. American Trucking Ass'ns, 531 U.S. 457, 462 (2001). Each State must develop a "State implementation plan" (SIP) to achieve and maintain the NAAQS. E.g., 42 U.S.C. 7410; Un ion Elec. Co. v. EPA, 427 U.S. 246, 249-250 (1976).
In 1970, Congress added the New Source Performance Standards (NSPS) program, which directs EPA to promul gate technology-based performance standards for new or modified facilities in certain categories of stationary sources. 42 U.S.C. 7411. The NSPS program is intended to ensure that pollution from new and modified emissions sources will be controlled. Thus, its standards are based on application of the best demonstrated system of emission reduction across par ticular industries and across the country, regardless of the actual effect of a source's emissions on local air quality. Ibid.
The NSPS program "was not entirely successful." Pet. App. 3a. In 1977, Congress enacted the Clean Air Act Amend ments of 1977, Pub. L. No. 95-95, 91 Stat. 685, which estab lished a statutory PSD program.1 The PSD program is part of the larger New Source Review (NSR) program, which also includes a nonattainment component for areas not satisfying ambient air standards. 42 U.S.C. 7501-7508. The nonattain- ment NSR program is not directly at issue in this case.
The PSD program is intended to prevent significant dete rioration of air quality in areas other than the nonattainment areas (viz. where ambient air quality standards are already being met and in unclassified areas), while also fostering eco nomic growth in a manner consistent with the preservation of existing clean air resources. 42 U.S.C. 7470(1) and (3); Alaska Dep't of Envtl. Conservation v. EPA, 540 U.S. 461, 470-471 (2004); Alabama Power Co. v. Costle, 636 F.2d 323, 346-351 (D.C. Cir. 1979). The PSD program directly addresses the impact on ambient air quality resulting from new construction of, and modifications to, pollutant-emitting facilities in such areas. 42 U.S.C. 7470, 7475(a)(3).
The core provision of PSD provides that "[n]o major emit ting facility * * * may be constructed in any area to which [the PSD provisions] appl[y] unless" various requirements are met. 42 U.S.C. 7475(a) (emphasis added). Those require ments include obtaining a permit setting forth emission limi tations and applying "best available control technology" (BACT). See 42 U.S.C. 7475(a)(1) and (4); 42 U.S.C. 7479(3).
The PSD provisions apply to "construct[ion]" of facilities, and they further provide that "[t]he term 'construction' * * * includes the modification (as defined in Section 7411(a) of this title) of any source or facility." 42 U.S.C. 7479(2)(C) (empha sis added). Section 7411(a) (which is one of the statutory pro visions applicable to the NSPS program) defines the crucial term "modification" as:
any physical change in, or change in the method of opera tion of, a stationary source which increases the amount of any air pollutant emitted by such source or which results in the emission of any air pollutant not previously emitted.
42 U.S.C. 7411(a)(4).
Thus, determining whether a planned activity is a "modi fication"-and is therefore "construction" subject to the PSD permitting requirements-involves a two-step process: (1) determining whether a project is a physical or operational change; and (2) if so, determining whether the change in creases emissions or results in the emission of new pollutants.
2. EPA promulgated regulations to implement the statu tory PSD program in 1978. 43 Fed. Reg. 26,380 (1978). It revised those regulations in 1980, 1992, and 2002. The Duke projects that are the subject of this enforcement action were undertaken between 1988 and 2000. Thus, the applicable reg ulations are (a) the 1980 regulations, 45 Fed. Reg. 52,676, which were recodified in 1987, 40 C.F.R. 51.166; and (b) the 1992 regulations, 57 Fed. Reg. 32,314; 40 C.F.R. 51.166 (1993).2 Both versions determine whether a change has in- creased emissions solely by applying an actual annual emis sions test to the types of projects at issue in this case.
a. By statute, PSD applies to the construction or modifi cation of "major emitting facilit[ies]," defined by reference to specified types of stationary sources. 42 U.S.C. 7475(a), 7479(1). EPA's PSD regulations require a pre-construction permit for the construction of any "major stationary source" or "major modification." 40 C.F.R. 51.166(i). The PSD regu lations define "major modification" as:
any physical change in or change in the method of opera tion of a major stationary source that would result in a significant net emissions increase of any pollutant subject to regulation under the Act.
40 C.F.R. 51.166(b)(2)(i).
The inquiry under that definition first requires identifica tion of a "physical change in or change in the method of opera tion." At that step, there is an "hours of operation" exclusion, which provides that "[a] physical change or change in the method of operation shall not include * * * [a]n increase in the hours of operation or in the production rate." 40 C.F.R. 51.166(b)(2)(iii)(f).
If there is a physical or operational change, the inquiry considers whether it would result in a "[n]et emissions in crease," which is defined in relevant part as "[a]ny increase in actual emissions from a particular physical change or change in the method of operation of a stationary source." 40 C.F.R. 51.166(b)(3)(i)(a). Pre-change "actual emissions" equal "the average rate, in tons per year, at which the unit actually emitted the pollutant during a two-year period which precedes the particular date and which is representative of normal source operation." Actual emissions are "calculated using the unit's actual operating hours, production rates and types of materials processed, stored, or combusted during the selected time period." 40 C.F.R. 51.166(b)(21)(ii).
To determine whether there has been an "increase in ac tual emissions from a particular physical change," 40 C.F.R. 51.166(b)(3)(i)(a), the pre-project average annual emissions must be compared to the post-project average annual emis sions. Because the PSD permit must be obtained before com mencement of construction or modification, the post-project emissions must be projected or otherwise estimated. Recog nizing that in some circumstances future actual emissions would be difficult to predict, 56 Fed. Reg. 27,633 (1991), the 1980 regulations provided: "For any emissions unit which has not begun normal operations on the particular date, actual emissions shall equal the potential to emit of the unit on that date," which refers to the unit's "maximum capacity * * * to emit a pollutant under its physical and operational design." 40 C.F.R. 51.166(b)(4) and (21)(iv). Thus, for units that had not begun normal operations, past actual annual emis sions were compared to future maximum potential emis sions-an "actual-to-potential" comparison.
In Wisconsin Electric Power Co. v. Reilly, 893 F.2d 901, 913 (7th Cir. 1990) (WEPCO), the court held that a utility that engaged in like-kind replacement of aging equipment had in fact "begun normal operations" and therefore did not come within the "potential to emit/maximum capacity" regulation. Accordingly, the court required a more "realistic" approach that would examine "the maximum emissions that can be gen erated while operating the source as it is intended to be oper ated and as it is normally operated." Id. at 916, 918 (citation omitted). On remand, EPA estimated WEPCO's future total annual emissions "based on all the available facts in the re cord," taking into account how much the unit was likely to be used and what the rate of emissions would be-typically re ferred to as the "actual-to-projected-actual" test. J.A. 68-72; 56 Fed. Reg. at 27,633 & n.10; 57 Fed. Reg. at 32,317 & n.10.
b. The NSPS regulations, by contrast, provide that a "modification" is "any physical or operational change * * * which results in an increase in the emission rate to the atmo sphere of any pollutant." 40 C.F.R. 60.14(a). The "emission rate" is defined as the maximum hourly emissions from the relevant piece of equipment operating at its maximum achiev able capacity. 40 C.F.R. 60.14(b); 40 C.F.R. 60.14(h) (1993); WEPCO, 893 F.2d at 913. The emission rate is "expressed as kg/hr." 40 C.F.R. 60.14(a) and (b). The NSPS emissions test thus determines whether an "increase in the emission rate" has occurred by comparing maximum hourly emissions at maximum capacity both before and after the change. Because it considers only maximum hourly rates, the NSPS test is not triggered by changes that increase emissions due only to in creased hours of operation.
c. In 1992, in response to the WEPCO decision, EPA is sued amendments to the 1980 PSD regulations that "clari f[ied] its methodology for calculating emissions increases" for electric utilities that "had begun normal operation." 57 Fed. Reg. 32,314, 32,323 (1991). The 1992 amendments generally provided that utility units that satisfy certain requirements should use the actual-to-projected-actual test rather than the actual-to-potential test, regardless of whether or not they had "begun normal operations." 56 Fed. Reg. at 27,633 & n.10; 57 Fed. Reg. at 32,323; 40 C.F.R. 51.166(b)(21)(v) (1993). The 1992 regulations also added a provision to the definition of "actual emissions" providing that, for electric generating units, actual emissions after the change equal the "represen tative actual annual emissions" of the unit. 40 C.F.R. 51.166(b)(21)(v) (1993). "Representative actual annual emis sions" is defined similarly to "actual emissions" and means "the average rate, in tons per year, at which the source is projected to emit a pollutant for the two-year period" after a change." 40 C.F.R. 51.166(b)(32) (1993). It considers "the effect any such change will have on increasing or decreasing the hourly emissions rate and on projected capacity utiliza tion." Ibid. (emphasis added).
The preamble to the 1992 regulations explained that for electric generating units, both NSPS and PSD calculations start with the hourly emission rate but differ in that the PSD calculation then multiplies "the hourly emissions rate times the utilization rate, expressed as hours of operation per year." 57 Fed. Reg. at 32,316 & n.6. The 1992 regulations retain the actual-to-potential test for other types of sources where a unit "has not yet begun normal operations." 40 C.F.R. 51.166(b)(21)(iv) (1993).
d. In 1996, pursuant to a conditional settlement agree ment with parties who had challenged the 1980 PSD regula tions, EPA proposed a "potential-to-potential" PSD test that was based on maximum hourly emissions, similar to the NSPS test. 61 Fed. Reg. 38,268-38,270 (1996); see 67 Fed. Reg. 80,204-80,205 (2002); 70 Fed. Reg. 61,081, 61,098 (2005). In 2002, however, EPA determined not to adopt the 1996 proposal, and instead issued revised PSD regulations that retained the actual-to-projected-actual test for exist ing utilities and broadened its applicability to all existing sources. 67 Fed. Reg. 80,186, 80,275 (2002); 40 C.F.R. 51.166(a)(7)(iv)(c) (2003).
3. Pursuant to CAA Section 307(b)(1), Duke and other electric utility companies, as well as other interested parties, filed various petitions for review in the D.C. Circuit challeng ing the 1980, 1992, and 2002 NSR regulations. Ultimately, those challenges were consolidated into a single action.
On June 24, 2005, the D.C. Circuit issued its ruling on those consolidated challenges in New York, 413 F.3d 3. The decision expressly rejected the utility companies' challenges to the 1980 and 2002 regulations' definition of "major modifi cation" under the PSD program, including their claim that the 1980 and 2002 NSR rules were invalid to the extent their definition of "modification" differed from a pre-1977 NSPS definition of that term. Id. at 18, 20. (Although the D.C. Cir cuit did not expressly address the 1992 regulations, they are identical in relevant respects to the 2002 regulations.) The court acknowledged the holding of the Fourth Circuit in this case that "Congress intended to require that EPA use identi cal regulatory definitions of modification across the NSPS and NSR programs," but found that the argument accepted by the court of appeals here "was not made by industry peti tioners in their opening brief and is therefore waived." Ibid.
4. Duke is an energy company; it operates 30 coal-fired generating units at eight electric power plants in North and South Carolina that began service before 1977. Pet. App. 24a. In 2000, the United States brought this suit against Duke, alleging that Duke executed 29 "modifications" at its coal- fired plants between 1988 and 2000 without complying with PSD. Id. at 22a, 25a. One representative project was the subject of the plaintiffs' summary judgment briefing: Unit 4 at Duke's Buck plant. Buck 4, which had been placed in "cold shutdown" status, J.A. 227, was part of Duke's "Plant Mod ernization Program," in which Duke sought to undertake "necessary plant modifications and maintenance" to "extend [the] operating life" of units that normally "would be retired and scrapped." J.A. 204, 227, 229, 232. After Duke spent some $17 million to rehabilitate Buck 4, it resumed commer cial operation in 1995, C.A. App. 785, 789, 790, more than a decade after it was placed in shutdown status. J.A. 227.
Three private groups (petitioners in this Court) inter vened as plaintiffs. Pet. App. 6a. On cross-motions for sum mary judgment, the district court rejected EPA's argument that its regulations permissibly measured PSD emissions increases based on an actual annual emissions test. Id. at 58a-72a. Rather, the court held that the PSD requirements are triggered only when a unit's maximum hourly emission rate increases, regardless of whether actual annual emissions increase due to increased hours of operation. The district court concluded that Congress, by cross-referencing the NSPS definition of "modification" in the PSD program in 1977, incorporated the 1977 NSPS regulations into the statu tory definition of "modification" for PSD. Id. at 62a-67a. The court also concluded that the "hours of operation" exclusion in the PSD regulations itself required the application of an hourly-rate test. Id. at 59a-62a. Because the United States and petitioners did not contend that Duke's projects resulted in increases in maximum hourly emission rates, the district court entered final judgment for Duke. Id. at 87a-95a.
5. The court of appeals affirmed, but on different reason ing. Pet. App. 1a-19a. The court held that Congress's deci sion to define "modification" in the PSD provisions by cross- reference to the NSPS statutory definition requires EPA to interpret the term consistently in the two programs. The court believed that its analysis was dictated by Rowan Cos. v. United States, 452 U.S. 247 (1981), which held that Congress intended substantially identical definitions of the term "wages" in two different tax statutes to be interpreted to mean the same thing. Pet. App. 11a-14a. In the court of ap peals' view, "Congress' decision to create identical statutory definitions of the term 'modification' has affirmatively man dated that this term be interpreted identically in the two pro grams. The different purposes of the NSPS and PSD pro grams cannot override that mandate." Id. at 17a-18a.
Although the court of appeals held that "modification" must be interpreted consistently in both programs, it did not hold that the statute mandated any particular definition. The court concluded that the PSD regulations "could even be en forced as the EPA urges provided that, as long as the PSD and NSPS statutes define 'modification' identically, the NSPS regulations are similarly interpreted and enforced." Pet. App. 15a n.7. However, because the NSPS regulations were in place at the time EPA promulgated the 1980 PSD regula tions and, unlike PSD, defined "modification" to include only projects that increase a plant's hourly emissions rate, the court of appeals concluded that EPA "must * * * interpret its PSD regulations defining 'modification' congruently" with the NSPS regulatory definition. Pet. App. 18a.
The court of appeals recognized that, under 42 U.S.C. 7607(b), it lacked jurisdiction to invalidate the PSD regula tions. Pet. App. 15a n.7. The court reasoned, however, that "no question of the validity of the PSD regulations is * * * presented here," because "the PSD regulations can be inter preted consistently with pre-existing principles-the NSPS regulations-as the district court demonstrated and as the EPA's Director of the Division of Stationary Source Environ ment twice opined shortly after promulgation of the regula tions." Ibid. Although the United States had informed the court that the identical issue was pending before the D.C. Circuit in New York, the court of appeals did not address the relationship of its ruling to those ongoing proceedings.
6. In October 2005, EPA proposed to revise the test for determining when projects at existing electric generating units constitute a PSD "modification" by adopting a test simi lar to the NSPS test that would apply prospectively to future modifications. 70 Fed. Reg. 61,081. The proposed regulation, if adopted, "would establish a uniform emissions test nation ally under the NSPS and NSR [including PSD] programs for existing" electric generating units. Ibid. In issuing the pro posed regulation, EPA stated that it continued to believe that the agency has "the authority to define 'modification' differ ently in the NSPS and NSR programs." Id. at 61,083 n.3, 61,090. The proposed rules were formulated to take into ac count changing conditions and air quality programs developed since 1980 to reduce emissions. Id. at 61,099.3
This latest proposal is part of EPA's ongoing evaluation of the PSD regulatory program. EPA has made adjustments to the program throughout the years as part of its continuing responsibility to "consider varying interpretations and the wisdom of its policy." Chevron U.S.A. Inc. v. NRDC, 467 U.S. 837, 863-864 (1984). EPA believes that, under its broad dis cretion to implement the PSD program, it can continue to address difficulties or challenges that arise in the program, as well as improvements in air quality achieved through im plementation of other programs under the CAA, through the rulemaking process.
SUMMARY OF ARGUMENT
I. Under Section 307(b)(1) of the Act, "[a] petition for review" of any nationally applicable regulations or final action of the EPA "may be filed only in the United States Court of Appeals for the District of Columbia." 42 U.S.C. 7607(b)(1). Under Section 307(b)(2), "[a]ction of the Administrator with respect to which review could have been obtained under para graph (1) shall not be subject to judicial review in civil or criminal proceedings for enforcement." 42 U.S.C. 7607(b)(2). Because the court of appeals' ruling had the effect of invali dating EPA's regulations as contrary to the statute, and be cause review of that issue "could have been"-and in fact was-"obtained" in the D.C. Circuit, Section 307(b)(2) pre cluded review of that issue in this enforcement proceeding.
The court of appeals held that the statute requires EPA to adopt the same test for a PSD "modification" that it applies to determine whether a modification has occurred under the NSPS program. In the court's view, its adjudication of that issue was permissible under Section 307(b)(2) because the PSD regulations were open to an interpretation under which they applied the same test as the NSPS regulations. The court of appeals was mistaken.
First, the only reasonable construction of EPA's regula tions is that they apply a different test under the two pro grams, and the court of appeals' holding that they must apply a consistent test thus amounts to an invalidation of the regu lations, forbidden by Section 307(b)(2) in an enforcement ac tion. Under the PSD regulations applicable here, a modifica tion occurs if a physical or operational change results in an increase in actual annual emissions, which could occur when a covered change would result in an increase in hourly emis sions, an increase in hours of operation, or both. By contrast, under the NSPS regulations, a modification occurs only if there is an increase in maximum achievable hourly emissions. The differences in the tests applied under the two programs are apparent from the significantly different wording of the two sets of regulations, the consistent use of an annual mea sure of emissions in the PSD regulations and an hourly mea sure in the NSPS regulations, and other regulatory provi sions. The differences are also apparent from statements made by EPA in promulgating the 1980 and 1992 regulations.
The court of appeals mentioned two bases for its belief that the PSD regulations are open to an "hourly rate only" interpretation: the "hours of operation" exception in the reg ulatory definition of "major modification," and statements by a mid-level EPA official in 1981. As every other court of ap peals to examine those sources has concluded, however, the regulatory exception by its terms is an exception only from what constitutes a "physical change or change in the method of operation" at the first step of the "modification" inquiry, and is not relevant to the inquiry into the net emissions in crease at the second step. Once it has been established that there has been a physical or operational change, as in this case, the regulatory exception has no application. As the other courts of appeals have also concluded, the statements by the EPA official were not authoritative and were mistaken.
Even if the "hours of operation" exclusion did require an "hourly rate only" interpretation, the court of appeals' hold ing would still be inconsistent with important aspects of the PSD definition of "modification," and the court therefore still lacked jurisdiction to consider the issue. While a covered change that results in no net increase of emissions is not a PSD "major modification," such a change could be a "modifi cation" under the NSPS regulations if the particular unit's maximum hourly emissions rate increased. Moreover, an "increase" is counted under the PSD program only if it ex ceeds specified threshold emissions levels; the NSPS regula tions contain no analogous provision.
In any event, there is a straightforward basis for conclud ing that review "could have been obtained" in the D.C. Cir cuit-and is thus barred here under Section 307(b)(2): Duke and other industry parties actually did bring substantially the same challenge in the D.C. Circuit. In the New York ac tion, Duke and other industry petitioners raised, briefed in part and waived in part, and ultimately lost the issue decided below: whether EPA's use of an actual annual emissions in crease test under the PSD program was unlawful because the CAA required EPA to apply instead the NSPS maximum hourly emissions test. Review of that issue was accordingly unavailable here.
II. Assuming, arguendo, that the court of appeals did not exceed its jurisdiction, the court nonetheless erred in its rul ing on the merits. The court of appeals accepted that the statutory definition of "modification" contains some ambigu ities that would ordinarily be up to EPA to resolve, and it also accepted that EPA could apply a total-annual-emissions test for modifications. Those conclusions should have sufficed to resolve this case in EPA's favor. But the court held that EPA must apply the same test for both NSPS and PSD purposes, based solely on the fact that Congress had cross-referenced the programs, providing in the PSD provisions that the term "construction" includes modification "as defined in" the NSPS statutory provisions. 42 U.S.C. 7479(2)(C).
That conclusion reflects a fundamental misapplication of Chevron. Congressional use of an ambiguous term reflects a delegation of authority to the agency to resolve ambiguities. Having found the relevant terms ambiguous, the court of appeals should have found Congress's repeated use of that term, through a cross-reference, to reflect repeated delega tions. Instead, the court treated the cross-reference as limit ing agency discretion.
The court below focused on the fact that the statute pro vides, for purpose of the PSD program, that "construction" includes modification "as defined in" the NSPS statutory pro visions. 42 U.S.C. 7479(2)(C). That instruction is not mean ingfully different from repeating the NSPS definition verba tim in the PSD provisions. Accordingly, it triggers the "iden tical terms" maxim, under which identical terms in different parts of the same statute are presumed to have identical meanings. In a long line of cases, however, this Court has found that maxim to be overridden by other considerations of statutory context, and it has frequently cautioned that the maxim is rarely dispositive. Accordingly, while a court that is faced with the need to construe a statute in the first in stance may, with caution, employ the "identical terms" maxim in conjunction with an examination of the surrounding statu tory context, that maxim alone is not sufficient to establish that Congress had a "clear intent" on whether the identical terms must be construed identically by a regulatory agency. The maxim is insufficient to establish that Congress left no ambiguity for the agency to resolve.
This Court's pre-Chevron decision in Rowan Cos. v. United States, 452 U.S. 247 (1981), cited by the court of ap peals as the sole support for its ruling, is not to the contrary. The court in Rowan did not simply apply the "identical terms" maxim, but carefully analyzed the legislative histories of the statutes at issue and placed great weight on the lack of contemporaneous administrative constructions that would contradict that maxim. Neither consideration is of signifi cance here. In any event, Rowan establishes at most that a court, faced with the need to arrive at its own construction of a statute, may in some cases use the "identical terms" maxim. It does not support the court of appeals' conclusion that that maxim, taken by itself, is sufficient to establish that Congress had a "clear intent" on the question at issue and therefore precluded the agency from resolving any ambiguity or filling any statutory gap.
In the specific context of the PSD program, moreover, Congress has required EPA to promulgate regulations and has set forth the goals and purposes of the PSD program. Those goals and purposes are not identical to the goals and purposes of the NSPS program, and Congress has therefore anticipated that EPA could adopt regulations that, inter alia, construe the component terms in the statutory definition of "modification" with sensitivity to the particular goals and purposes of the PSD program. The statute precludes the inference that Congress required that any and all ambiguities in the definition of "modification" be resolved in identical ways for purposes of the PSD and NSPS programs.
I. THE COURT OF APPEALS LACKED JURISDICTION OVER THE CLAIM THAT THE STATUTE REQUIRES THE SAME REGULATORY DEFINITION OF "MODIFI CATION" IN THE NSPS AND PSD PROGRAMS
A. If Review Of An Agency Action "Could Have Been Ob tained" In The D.C. Circuit Under Section 307(b)(1), Such Review May Not Be Had As A Defense To A Civil Enforcement Action
Section 307(b)(1) of the CAA provides that "[a] petition for review of action of the Administrator in promulgating * * * any * * * nationally applicable regulations promulgated, or final action taken, by the Administrator under [the Act] may be filed only in the United States Court of Appeals for the District of Columbia," within 60 days of the notice of such action in the Federal Register. 42 U.S.C. 7607(b)(1). In such a review, "the court may reverse any such action found to be," inter alia, "in excess of statutory jurisdiction, authority, or limitations, or short of statutory right." 42 U.S.C. 7607(d)(9)(C).
The judicial review mechanism established by Section 307(b)(1) is exclusive. Under Section 307(b)(2), "[a]ction of the Administrator with respect to which review could have been obtained under paragraph (1) shall not be subject to judicial review in civil or criminal proceedings for enforce ment." 42 U.S.C. 7607(b)(2). Thus, as this Court has ex plained, "any agency action that was reviewable in the courts of appeals cannot be challenged in an enforcement proceed ing, whether or not review was actually sought." Harrison v. PPG Indus., Inc., 446 U.S. 578, 605 (1980) (emphasis added). Reinforcing that rule, Congress further provided that "[n]othing in [the Act] shall be construed to authorize judicial review of regulations or orders of the Administrator under [the Act], except as provided in [Section 7607]." 42 U.S.C. 7607(e).
The exclusive review provisions of Section 307(b) promote the "even and consistent national application" of EPA regula tions, S. Rep. No. 1196, 91st Cong., 2d Sess. 40-41 (1970), thus serving "twin congressional purposes" of insuring CAA stan dards are "uniformly applied and interpreted" and "would be quickly reviewed by a single court intimately familiar with administrative procedures." Adamo Wrecking Co. v. United States, 434 U.S. 275, 284 (1978). Section 307(b) thereby avoids inconsistent results and forum shopping, provides speedy and authoritative review in a court of appeals with particular expertise in the relevant area, and ensures that regulated entities are treated consistently throughout the country, thus ensuring a level playing field for the regulated community. See Harrison, 446 U.S. at 593; NRDC v. EPA, 512 F.2d 1351, 1356-1357 (D.C. Cir. 1975); Lubrizol Corp. v. Train, 547 F.2d 310, 317 (6th Cir. 1976).
Congress intended the preclusive effect of Section 307(b) to be strictly enforced. In the 1977 Amendments, Congress responded to a recommendation by the Administrative Con ference of the United States, 41 Fed. Reg. 56,767-56,768 (1976), and amended Section 307(b) to expand the statutory limitations period from 30 to 60 days. In recommending that change, the relevant House committee "reaffirm[ed] its intent to strictly limit Section 307 challenges to those which are actually filed within that time." H.R. Rep. No. 294, 95th Cong., 1st Sess. 322 (1977). The committee expressly rejected other changes recommended by the Administrative Confer ence, including one "to permit the validity of a regulation to be challenged in defense to an enforcement proceeding." Id. at 324; see 41 Fed. Reg. at 56,768; see also Harrison, 446 U.S. at 591 (noting that committee "expressly disclaimed an en dorsement of the recommendations of the Administrative Conference" on certain matters).
B. The Claim That The PSD Regulations Must Be Consis tent With The NSPS Regulatory Definition Of "Modifi cation" Goes To The Validity Of The PSD Regulations And Could Have Been Heard By The D.C. Circuit
In this case, the court of appeals rejected EPA's interpre tation of the PSD regulations and held instead that, by virtue of the statutory definition of "modification," EPA "must in terpret th[e] term [modification] in a consistent manner in the NSPS and PSD regulations." Pet. App. 18a-19a; see id. at 11a. The court further concluded that its holding did not in trude upon the D.C. Circuit's exclusive jurisdiction to hear challenges to EPA's regulations under the CAA. Pet. App. 15 n.7. In the court's view, its newly minted requirement of con sistency between the PSD and NSPS regulations, and the resulting rejection of EPA's regulatory interpretation, did not affect "the validity of the PSD regulations" because "the PSD regulations can be interpreted consistently with pre- existing principles-the NSPS regulations-as the district court demonstrated and as the EPA's Director of the Division of Stationary Source Enforcement twice opined shortly after promulgation of the PSD regulations." Ibid.
The court of appeals' jurisdictional theory rests on a flawed premise. As the court of appeals itself recognized, its assertion of jurisdiction rests on the belief that the emissions increase test set forth in the PSD regulations can be inter preted identically with the maximum hourly emissions ap proach contained in the NSPS regulations. That belief is unfounded. The only reasonable interpretation of the PSD regulations is that they define "modification" differently from the NSPS regulations. Moreover, the district court's reliance on the "hours of operation" exclusion was wholly misplaced. And the court of appeals overlooked additional differences between the PSD and NSPS regulatory definitions of "modifi cation." Accordingly, the court of appeals was simply wrong that its holding that "modification" must be construed "con gruently" under the two programs, Pet. App. 18a, did not invalidate the PSD regulations. As the court of appeals itself implicitly recognized, it lacked jurisdiction to issue a decision that necessarily invalidated the PSD regulations.
1. The only reasonable construction of the PSD regula tions is that a physical change that increases a source's hours of operation is a "modification"
The regulations that govern this case are the 1980 and 1992 regulations. See pp. 4-8, supra.4 Both versions deter mine whether a change has increased emissions solely by applying the same actual, annual emissions test to the types of projects at issue in this enforcement action.
a. EPA's PSD regulations define a "major modification," which under the regulations may trigger the need for a PSD permit, as any physical change in or change in the method of opera tion of a major stationary source that would result in a significant net emissions increase of any pollutant subject to regulation under the Act.
40 C.F.R. 51.166(b)(2)(i). "Net emissions increase" is defined in relevant part as "any increase in actual emissions from a particular physical change or change in the method of opera tion of a stationary source." 40 C.F.R. 51.166(b)(3)(i)(a). By focusing on "net emissions increase[s]" and "actual" emis sions, the PSD regulations focus on the total amount of pollut ants actually released into the atmosphere as the result of a physical change. It does not matter, under the PSD regula tions, whether the increase comes from an increased hourly rate of emissions or a decreased hourly rate that is more than offset by increased hours of operation.
By contrast, the corresponding provision of the NSPS regulation is worded quite differently. The NSPS regulation provides that, with exceptions not applicable here, any physical or operational change to an existing facility which results in an increase in the emission rate to the atmosphere of any pollutant to which a standard applies shall be considered a modification within the meaning of section 111 of the Act.
40 C.F.R. 60.14(a) (emphasis added). The NSPS regulation does not refer to "net emissions increase" or "actual" emis sions; instead, it focuses on the hourly rate and makes clear that a modification occurs only if the "emission rate," mea sured in kilograms per hour, see 40 C.F.R. 60.14(b), increases as the result of a physical change. If EPA had wanted the presence of a major modification in the PSD regulations to turn only on physical changes that result in hourly rate in creases, and not increases in hours of operation, it could have simply adopted the NSPS regulations for the PSD program as well. It likewise could have adopted different regulatory language that focused on hourly emission rates. Instead, EPA adopted PSD regulations focused on the total amount of emissions and so embraced a quite different concept of "modi fication" for use in the PSD program.
b. Other aspects of the PSD regulations confirm that the regulatory "actual emissions" test is triggered when a physi cal change results in increased actual annual emissions, whether because of an increase in hourly rate of emissions or an increase in hours of operation or some combination of the two. Actual emissions "shall be calculated using the unit's actual operating hours, production rates and types of materi als processed, stored, or combusted during the selected time period." 40 C.F.R. 51.166(b)(21)(ii). Thus, if the "actual oper ating hours" or "production rates" increase as a result of a physical or operational change, then the actual annual emis sions would increase as well. Moreover, the PSD regulations require that "actual emissions" be calculated annually, in tons per year, based on periods that are "representative of normal source operation." See ibid. The use of an annual measure, as well as the reference to a time period "represen tative of normal source operation," further emphasize the focus in the regulations on the actual annual emissions from the source, not merely the rate per hour of emissions.5 In addition, only a physical or operational change that results in a "significant net emissions increase" triggers PSD. 40 C.F.R. 51.166(b)(2)(i); cf. 40 C.F.R. 60.14(b) (requiring use of kilograms per hour, not tons per year, to determine if there is an increase in "emission rate," and therefore a modifica tion, under NSPS). A net increase is "significant" under the PSD regulations if it would exceed specified amounts of par ticular pollutants, which are expressed in "tons per year," not the hourly rate of emissions. 40 C.F.R. 51.166(b)(23)(i).
Finally, unlike the NSPS regulations, the PSD regulations provide for "netting" of emissions changes in making the ma jor modification determination. Only a physical or opera tional change that results in a "significant net emissions in crease" triggers PSD. 40 C.F.R. 51.166(b)(2)(i). Whether there is a "net" increase is determined by examining any creditable contemporaneous decreases and increases in actual annual emissions at the source. 40 C.F.R. 51.166(b)(3)(i). Thus, an increase caused by a particular physical or opera tional change may generally be offset by a contemporaneous decrease. The provision for "netting" reinforces the emphasis in the PSD regulations on the actual annual emissions into the atmosphere, not the measure of maximum hourly emis sions from a unit.
The EPA's own official statements regarding the 1980 regulations demonstrated that an increase in actual emissions resulting from a qualifying physical or operational change may be a modification, regardless of whether the change re sults in increased hours of operation or increased emissions during the same hours of operation. The preamble to the 1980 regulations provided a detailed example of how the PSD regulations would work. It recites the case of a source that both adds a new emission unit and "plans to increase the hours of operation" at its existing unit above the level allowed in its existing permit. 45 Fed. Reg. at 52,705. EPA explained that "both changes will result in significant net increases in actual emissions" and therefore "[b]oth changes then qualify as modifications." Ibid. (emphasis added).6 The example makes clear that a modification may result from either an increase in hourly emissions or in hours of operation.7
c. In 1992, EPA promulgated new regulations that govern some of the projects at issue in this case. In the 1992 pream ble, EPA confirmed that a physical or operational change that results in an actual increase in emissions is a "modification" under the PSD program, even if the change does not increase the hourly rate of emissions. As EPA explained, "[a]lthough a source may vary its hours of operation or production as part of its everyday operations, an increase in emissions attribut able to an increase in hours of operation or production rate which is the result of a construction-related activity is not excluded from review." 57 Fed. Reg. at 32,328.
Furthermore, the preamble compares the NSPS and PSD tests, noting that they are "largely the same" in determining "whether a physical or operational change will occur," but that in the second step, "the applicable rules branch apart," with the "NSPS program examin[ing] maximum hourly emis sion rates, expressed in kilograms per hour," and the PSD program "examin[ing] total emissions to the atmosphere," which "are determined by changes in annual emissions as expressed in tons per year." 57 Fed. Reg. at 32,316.8
2. The "hours of operation" exclusion does not support the court of appeals' conclusion that the regulations can be read to require that only an hourly-rate in crease in emissions is a "modification"
The court of appeals did not itself attempt to explain how the 1980 or 1992 regulations could be read to be consistent with the NSPS definition of "modification." The court instead relied on the district court's analysis of the "hours of opera tion" exclusion, and on the views expressed by a single EPA official. Pet. App. 15a n.7. Neither the district court's ratio nale, however, nor the EPA official's views establish that, notwithstanding the clear and unmistakable differences be tween "major modification" under the PSD regulations (with their focus on total emissions) and "modification" under the NSPS regulations (with their focus on hourly rate), the PSD regulations can be construed to adopt the NSPS test.
a. The district court held that the so-called "hours of op eration" exclusion in the PSD definition of "major modifica tion" requires EPA to keep hours of operation constant be fore and after the change in calculating emissions increases, such that there is an emissions increase under PSD only if there is a change resulting in increased hourly emissions. Pet. App. 59a-60a. That reading of the regulations conflates the two steps in the regulatory definition and is not tenable.
A "major modification" under the PSD regulations is "any physical change or change in the method of operation of a major stationary source that would result in a significant net emissions increase of any pollutant." 40 C.F.R. 51.166(b)(2)(i). Applying that definition requires a two-step inquiry: first, a determination whether a project is a "physi cal or operational change;" second, whether such change "would result in a significant net emissions increase." See, e.g., 57 Fed. Reg. at 32,316.
The definition of "major modification" has a number of exclusions from what constitutes a "physical change or change in the method of operation," at the first step of the inquiry. One such exclusion provides that, in general, "[a] physical change or change in the method of operation shall not include * * * [a]n increase in the hours of operation or in the production rate." 40 C.F.R. 51.166(b)(2)(iii)(f) (emphasis added). By its terms, that "hours of operation" exclusion is not an exclusion from the entire definition of "major modifica tion," nor from the portion of the definition pertaining to emission increases at the second step of the analysis. Rather, it is an exclusion only from what constitutes a qualifying "physical change or change in the method of operation" at the first step. Accordingly, a mere increase in hours of operation standing alone is not a "change," and the "modification" in quiry need proceed no farther.
By contrast, if there has been a "change" other than a mere increase in hours of operation or other excluded event, the inquiry proceeds to the second step: whether the pro posed "change" "would result in a significant net emissions increase." The "hours of operation" exclusion has no applica tion at that stage of the inquiry, because on its face it applies only to the "change" analysis, not the "emissions increase" analysis of a conceded "change." Thus, if the proposed "change" would result in a significant net emissions increase, it constitutes a "modification" for PSD purposes, even if the emissions increase is the result of the fact that the cognizable "change" will facilitate or lead to increased hours of operation rather than increased hourly emissions.
EPA explained in the preamble to the 1980 regulations that the exclusion applies to market fluctuations, not to in creases in hours of operation made possible by a physical or operational change. The preamble explained:
While EPA has concluded that as a general rule Congress intended any significant net increase in [actual] emissions to undergo PSD or nonattainment review, it is also con vinced that Congress could not have intended a company to have to get a NSR permit before it could lawfully change hours or rate of operation. Plainly, such a re quirement would severely and unduly hamper the ability of any company to take advantage of favorable market conditions.
45 Fed. Reg. at 52,704. In later statements, in 1988 and in promulgating the 1992 regulations, EPA reiterated that the "hours of operation" exclusion should be read in accordance with its plain meaning, as a modest exclusion from what con stitutes a "physical change or change in the method of opera tion" and that the exclusion has no effect when there has been a qualifying "change."9 The "hours of operation" exclusion does not support the argument that the regulations can be read to limit a PSD "major modification" to cases in which there has been an increase in per-hour emissions.
b. Given the clarity of the regulatory language, and EPA's explanation of the exclusion in the regulatory pream ble, it is immaterial whether, as the district court concluded (Pet. App. 60a-62a), a mid-level EPA official, Edward Reich, then-Director of the Division of Stationary Source Enforce ment, interpreted the provisions differently in an internal memorandum and a letter written in 1981. See J.A. 27-28, 35- 37. Even agency officials may misunderstand some details of newly promulgated regulations, and any such interpretation by Mr. Reich was plainly incorrect. That conclusion is rein forced when, as here, the letters provide no indication of the care or attention Mr. Reich gave to the issue and contain no explanation of how he arrived at his variant understanding.10
The other courts of appeals to consider the scope of the "hours of operation" exclusion have correctly rejected argu ments, based on the same 1981 misinterpretation by Mr. Reich, that the exclusion prohibits consideration of increases in hours of operation resulting from physical or operational changes in calculating emission increases under the PSD pro gram. See WEPCO, 893 F.2d at 916 n.11 (exclusion "was pro vided to allow facilities to take advantage of fluctuating mar ket conditions, not construction or modification activity"); Puerto Rican Cement Co. v. EPA, 889 F.2d 292, 298 (1st Cir. 1989) (Breyer, J.) (upholding EPA interpretation of exclusion as allowing sources "simply to increase their output" through "increased use of existing facilities" as opposed to increases resulting from construction or modification activity); see also United States v. Ohio Edison Co., 276 F. Supp. 2d 829, 876- 877 (S.D. Ohio 2003). The "hours of operation" exclusion does not alter the conclusion that a "modification" may occur not only when a physical or operational change results in an in crease in per-hour emissions (without an offsetting decrease in hours of operation), but also when such a change results in increased hours of operation without any sufficiently offset ting decrease in the hourly emissions rate.
3. Even Under The District Court's Reading Of The "Hours Of Operation" Exclusion, There Are Inconsis tencies Between The Treatment Of "Modification" In The PSD And NSPS Regulations
The district court's erroneous interpretation of the "hours of operation" exclusion is not the only flaw in the court of ap peals' jurisdictional analysis. Even if the exclusion could rea sonably be construed as the district court read it, the court of appeals' holding would still effectively invalidate important aspects of the PSD regulatory definition of modification, and the court therefore lacked jurisdiction to consider the issue. In some instances, for example, expansion or refurbishment of an existing unit may increase the hourly emissions rate but also, by virtue of the modified unit's increased efficiency and productive output, enable the facility's owner to reduce oper ating hours and thereby avoid any increase in emissions on an annualized basis. Or, under the "net emissions" test in the PSD regulations, certain contemporaneous emissions reduc tions at a source may be used to offset emissions increases at a particular unit. 40 C.F.R. 51.166(b)(2)(i) and (3)(i). Under the NSPS regulations with their focus on hourly emission rates, such changes would generally qualify as "modifica tions." They would generally not qualify as "major modifica tions" under the PSD regulations, however, because the total projected annual emissions at the source as a whole had not increased. Neither the district court nor the court of appeals offered any analysis of the PSD regulations that could reason ably harmonize them with the NSPS regulations in those cir cumstances, and for good reason: none exists. The PSD reg ulations focus on the sum of total emissions, and the NSPS focus on hourly rates is simply inconsistent with that focus.
As discussed above, see pp. 20-23, supra, moreover, there are other substantial differences between the PSD and NSPS definitions of "modification," even assuming arguendo that the "hours of operation" exclusion could be interpreted in the unreasonable manner proposed by the district court. Thus, for example, the PSD regulations provide that only a "signif icant net emissions increase" is a PSD "major modification," and set forth various threshold emissions levels that will be deemed significant, measured in "tons per year." 40 C.F.R. 51.166(b)(2)(i) and (23)(i) (emphasis added). The NSPS defi nition of "modification," by contrast, contains no such re quirement of a "significant" annual increase before a "modifi cation" will be found, and would thus be triggered in some circumstances in which the PSD definition would not.11
It is thus undeniable that, as promulgated by EPA, the PSD regulations define "modification" differently from the NSPS definition of "modification" in multiple and substantial respects. The decision of the court of appeals effectively in validates the PSD regulations in those respects. Accordingly, the court of appeals lacked jurisdiction to consider the issue on which its judgment rests, because under Section 307(b)(1) and (2) only the D.C. Circuit has jurisdiction to entertain challenges to the PSD regulations.
4. Because Review Of The Claim That The PSD Regu lations Must Be Consistent With The NSPS Regula tions "Could Have Been Obtained" In A Petition For Review, That Claim Was Precluded Here
An examination of the challenges actually advanced by Duke and other industry petitioners in the D.C. Circuit pro ceedings confirms that the court below was jurisdictionally precluded from deciding the case as it did. Under Section 307(b)(2), "[a]ction of the Administrator with respect to which review could have been obtained under paragraph (1) shall not be subject to judicial review in civil or criminal proceed ings for enforcement." 42 U.S.C. 7607(b)(2). The CAA thus bars courts in enforcement proceedings (like this case) from entertaining challenges that could have been brought on a petition for review under Section 307(b)(1) in the D.C. Circuit. As this Court has emphasized, such courts simply cannot "pursue any of the * * * familiar inquiries which arise in the course of an administrative review proceeding." Adamo Wrecking Co., 434 U.S. at 285.
Perhaps the clearest evidence that Duke could have brought its statutory challenge in the D.C. Circuit is that it in fact did so. During the course of the New York litigation in the D.C. Circuit, and beginning long before the United States brought this enforcement action, Duke and other industry petitioners raised, briefed in part and waived in part, and ultimately lost the question in this case: whether EPA's use of an actual annual emissions increase test under the PSD program was unlawful because the CAA required EPA to apply instead the NSPS maximum hourly emissions test.
The issue was first raised in 1981, when several industry parties filed opening briefs in their D.C. Circuit challenge to the 1980 PSD rules. One of those briefs, which was entitled "Brief for Industry Petitioners on Actual Emissions Defini tion of Net Increase" (Industry Br.), raised the question whether EPA's PSD regulations unlawfully "provided that a modification * * * would occur when actual emissions from a source increased as a result of an alteration to that source, even where the source's capacity to emit remains constant." Industry Br. at 1, Chemical Mfrs. Ass'n v. EPA, No. 79-1112, 1999 WL 1338364 (D.C. Cir. Dec. 15, 1999) (per curiam); see id. at 43. That brief criticized the PSD actual emissions test because it is a function "of the production rate at which the source operates" and "requires that variations in the source's hours and rates of operation be taken into account." Id. at 28-29. Thus, EPA's actual annual emissions test, including its consideration of hours of operation, was subjected to industry challenge as early as 1981.
In 1982, the parties to the review proceedings in the D.C. Circuit entered into a conditional settlement agreement. That agreement required EPA to propose, and take final ac tion adopting or rejecting for NSR purposes, "an NSPS-like hourly-potential-to-hourly-potential emissions increase test for modifications." 70 Fed. Reg. at 61,098. If adopted, that proposal would have provided industry with the very relief it sought in the D.C. Circuit-relief that it ultimately obtained only as a result of the court of appeals' decision in this case. Specifically, the proposal would have added PSD regulatory language allowing industry to measure emission increases by using an hourly emissions rate based on "the source's poten tial to emit (as calculated in terms of pounds of pollutant emitted per hour)." 61 Fed. Reg. at 38,255, 38,269. As EPA explained, under the proposed approach "the level of opera tions and actual emissions would generally no longer be perti nent" for NSR purposes. Id. at 38,269 (emphasis added).
EPA rejected the proposed changes in 2002. The Utility Air Regulatory Group (UARG), of which Duke is a member, filed a petition for review challenging that decision in the D.C. Circuit. See Utility Air Regulatory Group v. EPA, No. 03-1046 (filed Feb. 27, 2003). One of the issues raised by UARG was whether the 1980 NSR (and, thus, PSD) regula tions were invalid "if the actual emissions test in the 1980 NSR rule does not require a physical or operational change that results in an increase in the maximum achievable emis sion rate of an existing unit." UARG Non-Binding Statement of Issue at 2, New York, supra (No. 02-1387); see Motion of Alabama Power Co. to Reopen Administratively Terminated Petitions for Review at 3-4, Chemical Mfrs. Assn, supra ("Whether the [1980 and 1992] NSR rules * * * are lawful, if it is permissible to interpret those rules, as EPA does today, as providing that an increase in emissions at an existing unit that is attributable to the unit operating more hours can con stitute a "major modification," where there is no relaxation of an enforceable limitation on the number of hours that the unit may operate."). Industry petitioners also raised the following challenge to the 1980 and 1992 NSR regulations:
Whether, if the 1980 rules repealed the requirement that there must be an emission rate increase at an existing unit for there to be an NSR modification, the 1980 rules are unlawful?
Joint Brief of Industry Petitioners at 2, New York, supra (No. 02-1387)
On June 24, 2005, the D.C. Circuit ruled against industry petitioners, including Duke, expressly rejecting their "claim that modification must have the same regulatory meaning for NSR as prevailed for NSPS in 1977." New York, 413 F.3d at 19. The court held that the CAA did not compel EPA to apply the NSPS regulatory definition of emissions increases in mea suring increases under the PSD program. Id. at 18-20. To be sure, although the court addressed the general claim that the Act precluded the actual annual increase standard, the court did not rule on the precise argument accepted by the Fourth Circuit in this case-that the Act requires that the term "modification" have a consistent meaning in both the PSD and NSPS programs. But the reason the court did not ad dress that precise argument was that it "was not made by industry petitioners in their opening brief and is therefore waived." Id. at 20. The D.C. Circuit at no point indicated that the argument, had it been preserved, would have been unripe or otherwise not justiciable. Indeed, the fact that the court addressed, and rejected, other statutory-based argu ments challenging the actual annual emissions test on their merits establishes that such claims were ripe for review. And because review of challenges to the actual annual emissions test on the ground that it was contrary to the statute "could have been obtained" in a petition for review, it is unavailable in this enforcement action.
II. THE COURT OF APPEALS ERRED IN HOLDING THAT EPA LACKS DISCRETION TO APPLY A REGULATORY TEST FOR PSD "MODIFICATIONS" THAT DIFFERS FROM THE NSPS REGULATORY TEST
Assuming arguendo that the court of appeals had jurisdic tion to review the statutory validity of the annual actual emis sions test applied by EPA under the PSD program, the court of appeals erred in invalidating that test. Congress's use of identical statutory definitions of "modification" for the PSD and NSPS programs indicates that it would have been per missible for EPA to adopt the same regulatory interpretation of "modification" for both programs, but it does not compel that result. Rather, to the extent that the statutory definition leaves ambiguities and gaps for EPA to fill, Congress autho rized EPA to resolve those questions separately for purposes of the PSD and NSPS programs in any reasonable fashion consistent with the statute, in light of the distinct structure and purposes of the two programs. Because EPA's regula tions did just that, the court of appeals owed deference to EPA's construction of the Act in the context of the PSD pro gram, and the court should have held that the regulations' actual annual emissions standard is a permissible one.
A. EPA May Interpret The Statutory Definition Of "Modi fication," Including The Ambiguous Term "Increases," Differently To Effectuate The Distinct Purposes Of The Separate PSD And NSPS Programs
In reviewing an agency's construction of a statute, this Court asks "whether Congress has directly spoken to the precise question at issue" and, if Congress has instead been "silent or ambiguous," whether "the agency's answer is based on a permissible construction of the statute." Chevron, 467 U.S. at 842-843. Congress has not clearly spoken to the pre cise question at issue in this case: what constitutes an emis sions "increase" for the purposes of determining whether a project is a "modification" under the PSD program. EPA's interpretation of that ambiguous term for PSD purposes is reasonable and entitled to deference.
1. The requirements of the PSD program apply to "con struction" of covered facilities. 42 U.S.C. 7475(a). By the time the PSD provisions were enacted in 1977, the term "con struction" had been defined for purposes of the pre-existing NSPS program. As originally enacted, however, the PSD provisions themselves did not define "construction." The PSD cross-reference to the NSPS definition was added three months later, as part of a set of 84 "Clean Air Act Technical and Conforming Amendments." Safe Drinking Water Amendments of 1977, Pub. L. No. 95-190, § 14(a)(54), 91 Stat. 1402. The new PSD definition provides that "[t]he term 'con struction' when used in connection with any source or facility, includes the modification (as defined in [42 U.S.C.] 7411(a) * * * ) of any source or facility." 42 U.S.C. 7479(2)(C). Thus, for both PSD and NSPS, the overarching statutory definition of "modification" is the same: a modification is "any physical change in, or change in the method of operation of, a station ary source which increases the amount of any air pollutant emitted by such source or which results in the emission of any air pollutant not previously emitted." 42 U.S.C. 7411(a)(4).
2. The statutory definition of "modification" leaves sev eral constituent terms undefined, and therefore open to rea sonable agency interpretation. Among them is the phrase at issue in this case: "increase the amount of any air pollut ant." Although a change is not a "modification" unless it "in creases" the amount of emissions, the statute "is silent on how to calculate such 'increases' in emissions." New York, 413 F.3d at 22. There is substantial ambiguity in the term "increase" in this context and many ways to evaluate an "in crease" in amount of pollutants. Such an increase could be measured by the minute, the hour, the year, or some other period. It could be measured in terms of actual amounts, average amounts, maximum possible amounts, estimates or projections of likely amounts, or some other metric.
Under the Chevron doctrine, that ambiguity grants the EPA the discretion to select the method of evaluation it thinks best, so long as EPA's action is reasonable and not inconsistent with the statute itself. Congress "did not specify how to calculate 'increases' in emissions, leaving EPA to fill in that gap while balancing the economic and environmental goals of the statute." New York, 413 F.3d at 27.
B. Congress's Use Of The Same Definition Of "Modifica tion" In The Provisions Governing NSPS and PSD Does Not Eliminate EPA's Discretion To Resolve Ambiguities In That Definition In Different Ways For Each Program
The court of appeals accepted that the statutory definition of "modification" contains some ambiguities, and the court also accepted that EPA could validly construe the ambiguous term "increase" in that definition to yield the interpretation at issue in this case-an actual annual emissions, rather than a maximum hourly rate of emissions, test. See Pet. App. 15a n.7 ("The PSD regulations * * * could even be enforced as the EPA urges provided that * * * the NSPS regulations are sim ilarly interpreted and defined."). Those conclusions should have sufficed to uphold the regulations under Chevron. But the court held that the EPA was precluded from construing "increases" to mean something different in the PSD program than in the NSPS program. In the court's view, because Con gress had used a single definition for "modification" for use in both programs, Congress had "directly spoken to the pre cise question at issue" of whether the component parts of that term could have somewhat different meanings for the PSD and NSPS programs and had decided that the meaning must remain constant. Id. at. 10a-11a (quoting Chevron, 467 U.S. at 842); see id. at 18a-19a. The court further held that, be cause the NSPS regulations were issued first, their construc tion of the term "modification" had to be used for the PSD program.12
1. The court of appeals gave only one reason for its con clusion that the statute unambiguously compels EPA to con strue the constituent terms of the "modification" defini tion-in particular, the term "increases"-in the same way for the PSD and NSPS programs. According to the court, this Court's decision in Rowan established an "effectively irrebuttable" presumption that the term "increases" must be given a consistent meaning in the NSPS and PSD programs, because "Congress' decision to create identical statutory defi nitions of the term 'modification' has affirmatively mandated that this term be interpreted identically in the two pro grams." Pet. App. 17a. On that basis, the court of appeals concluded that Congress had "directly spoken" to the ques tion before it.
2. The court of appeals was mistaken. As this Court has explained, at the first stage of the Chevron analysis, "when a statute speaks clearly to the issue at hand [a court] 'must give effect to the unambiguously expressed intent of Congress.'" Barnhart v. Thomas, 540 U.S. 20, 26 (2003); see Household Credit Servs. v. Pfennig, 541 U.S. 232, 242 (2004) (statute ambiguous under Chevron because it does not provide "a clear answer"). As the Court explained in Chevron, "[i]f the intent of Congress is clear, that is the end of the matter; the court, as well as the agency, must give effect to the unambig uously expressed intent of Congress." 467 U.S. at 842-843; see id. at 843 n.9 (court "must reject administrative construc tions which are contrary to clear congressional intent").
a. The Clean Air Act reveals no "unambiguously ex pressed intent of Congress" or "clear answer" with respect to the question whether the ambiguities in the definition of "modification" may be resolved differently in the NSPS and PSD contexts. The statute simply provides that "[t]he term 'construction' when used in connection with any source or facility, includes the modification (as defined in [42 U.S.C.] 7411(a)) of any source or facility." 42 U.S.C. 7479(2)(C). That shorthand reference is not meaningfully different from a rep etition of the Section 7411(a) definition in Part C of the CAA (which establishes the PSD program), and it conveys no clear congressional command that EPA is to be denied the usual range of discretion conferred on agencies to resolve statutory ambiguities in ways reasonably tailored to each particular regulatory context. To the contrary, absent an express direc tion to treat two programs identically, Congress's repeated use of an ambiguous term reflects a repeated delegation, not a command of parity.
Even in the absence of an authoritative agency interpreta tion construing the same ambiguous statutory term differ ently in two different, but related, contexts, this Court's cases have frequently reached that result as a matter of judicial interpretation. And the Court has regularly cautioned against assuming, without careful examination of the context, that Congress's use of the same term in two provisions man dates that it be given the same meaning in each.
Thus, in Atlantic Cleaners & Dyers v. United States, 286 U.S. 427 (1932), the Court considered the meaning of the term "trade or commerce" in two different provisions of the Sherman Act, 15 U.S.C. 1 et seq.-one in Section 1, forbidding any combination "in restraint of trade or commerce among the several states"; the other in Section 3, forbidding any combination "in restraint of trade or commerce in any Terri tory of the United States or of the District of Columbia." See 286 U.S. at 432. The court noted that "there is a natural pre sumption that identical words used in different parts of the same act are intended to have the same meaning." Id. at 433. But the Court also explained that "the presumption is not rigid and readily yields whenever there is such variation in the connection in which the words are used as reasonably to warrant the conclusion that they were employed in different parts of the act with different intent." Ibid. The Court thus held that the term "trade," while likely "synonymous" with the word "commerce" in Section 1, id. at 434, has the broader meaning of "occupation, employment, or business," in Section 3, id. at 436.
In a long line of cases, the Court has followed Atlantic Cleaners in recognizing that Congress's use of a single, am biguous term in two contexts does not, by itself, signal a clear intent that the term must have the same meaning in each. For example, the Court in Helvering v. Stockholms Enskilda Bank, 293 U.S. 84, 86-88 (1934), concluded that the term "ob ligations" has distinct meanings for the purposes of two provi sions of the tax code-one that excluded interest on govern ment "obligations" from income and a second that defined income for nonresident aliens to include interest on "obliga tions." In United States v. Cleveland Indians Baseball Co., 532 U.S. 200, 212-216 (2001), the Court concluded that the term "wages paid" may have different meanings for purposes of the statutes governing social security and federal unem ployment taxes than it has for purposes of the closely related statute governing eligibility for social security benefits. In Robinson v. Shell Oil Co., 519 U.S. 337, 342-344 (1997), the Court held that the term "employee" may include former employees for purposes of Title VII's prohibition of retalia tion, while it refers only to current employees for purposes of some other provisions. See Wachovia Bank v. Schmidt, 126 S. Ct. 941, 951-952 (2006) (term "located" in statutes laying venue for actions against national banks has different mean ing than same term in statutes governing federal question jurisdiction in such actions); General Dynamics Land Sys tems, Inc. v. Cline, 540 U.S. 581, 596-597 (2004) (holding that "age" means "old age" in some provisions of Age Discrimina tion in Employment Act, 29 U.S.C. 623(a)(1), while it means "comparative youth" in nearby provision setting forth BFOQ defense); District of Columbia v. Carter, 409 U.S. 418, 421- 425 (1973) (District of Columbia is a "State or Territory" un der 42 U.S.C. 1982, but not under 42 U.S.C. 1983).
The principle that the same statutory term can be given different meanings in different contexts carries even greater force when, as here, the agency charged by Congress with responsibility for construing the statute has adopted different interpretations of a particular term in the exercise of its dele gated rulemaking authority. As this Court recognized in Chevron itself, in upholding EPA's interpretation of the very statute at issue here, "the fact that the agency has adopted different definitions [of the same statutory term] in different contexts [in the Clean Air Act] adds force to the argument that the definition itself is flexible." 467 U.S. at 864. Thus, in Cleveland Indians this Court deferred to the agency's deci sion to give the same statutory term a different meaning, 532 U.S. at 218-220, and numerous decisions of the courts of ap peals likewise recognize the authority of agencies to interpret identical statutory terms differently for different purposes.13
b. To be sure, the Court has also concluded that the same term in two different provisions may be construed to have the same meaning. In Rowan, the sole case on which the court of appeals relied, Congress rejected a Treasury regulation de fining "wages" for purposes of FICA and FUTA withholding to include the value of meals and lodging provided for the convenience of the employer, instead holding that "wages" excludes such meals and lodging, as in the income-tax with holding context. The Court noted that Congress had defined "wages" in "substantially the same language" in the income- tax withholding provisions as in the FICA and FUTA provi sions. 452 U.S. at 255. But the Court did not rely merely on the fact that the same definitions were used. Instead, the Court undertook its own extensive analysis of the legislative histories of the statutes at issue, concluding that "[t]hese histories reveal a congressional concern for 'the interest of simplicity and ease of administration'" and that "one of the means Congress chose in order to promote simplicity was to base withholding upon the same measure-'wages'-as taxa tion under FICA and FUTA." Ibid.; see, e.g., id. at 256 (quot ing committee report explaining that Congress had "coordi nated" the three tax provisions "in order to facilitate the work of both the Government and the employer in administering the withholding system"). Moreover, the Court, acting prior to Chevron, placed great weight on the lack of "substantially contemporaneous constructions of the statutes" supporting Treasury's position. Id. at 262; cf. Chevron, 467 U.S. at 863 ("The fact that the agency has from time to time changed its interpretation of the term 'source' does not * * * lead us to conclude that no deference should be accorded the agency's interpretation of the statute. An initial agency interpretation is not instantly carved in stone.").
Rowan illustrates the commonsense proposition that when Congress uses the same terms in two different statu tory contexts, Congress may have an intent to ensure "sim plicity and ease of administration" and, if that intent can be substantiated and is not contradicted by other indices of con gressional intent, a court may infer that Congress intended the term to have the same meaning in both contexts. But Rowan, which was decided before Chevron and does not apply the Chevron framework, does not suggest that, even where a court finds the "identical terms" maxim persuasive, the court would conclude that Congress had an "unambiguously ex pressed" or "clear" intent on the matter.14 Rather, Rowan, especially when read in the context of the line of this Court's cases evidencing a distinctly skeptical view of the "identical terms" maxim, merely demonstrates that a court may use that maxim as one of many means by which it can determine the preferred interpretation of an ambiguous statute.
Under the Chevron framework, by contrast, when an ad ministrative agency with delegated authority to construe the statute has resolved the statutory ambiguity, "the court does not simply impose its own construction on the statute, as would be necessary in the absence of an administrative inter pretation." Chevron, 467 U.S. at 843 (footnote omitted). Rather, Chevron teaches that "a court's opinion as to the best reading of an ambiguous statute an agency is charged with administering is not authoritative." National Cable & Telecomms. Ass'n v. Brand X Internet Servs., 125 S.Ct. 2688, 2701 (2005). Where there is ambiguity, Chevron "requires a federal court to accept the agency's construction of the stat ute, even if the agency's reading differs from what the court believes is the best statutory interpretation." Id. at 2699; see Chevron, 467 U.S. at 843 n.11 ("The court need not conclude that the agency construction was * * * even the reading the court itself would have reached if the question initially had arisen in a judicial proceeding."). As this Court's numerous cases rejecting the "identical terms" maxim illustrate, that maxim, taken by itself, is insufficient to establish that Con gress directly addressed and clearly resolved the question whether the ambiguous, component terms of the definition of "modification" may have a somewhat different meaning in the PSD and NSPS contexts.
c. The court of appeals apparently believed that the "identical terms" maxim is "effectively irrebuttable" in the context of this case, because the terms at issue are found in a definition. Pet. App. 17a.15 Words used in a definition, how ever, are not subject to special rules of statutory interpreta tion or an emasculated version of the Chevron doctrine. In deed, this Court has never suggested that its repeated warn ings against reliance on the "identical terms" maxim apply only to statutory terms that are not defined. To the contrary, the Court has confirmed that different interpretations of a term remain permissible even where Congress mandates a single statutory definition of that term be used. In Robinson, the term "employee" was statutorily defined, and Congress instructed that the definition was to apply "for purposes of Title VII" of the Civil Rights Act of 1964, 42 U.S.C. 2000e et seq. 519 U.S. at 342 (citing 42 U.S.C. 2000e(f))-a clearer statutory preference for a single definition than the vaguer "as defined in section 7411(a)" CAA provision at issue in this case. Notwithstanding the requirement that the single statu tory definition be applied throughout Title VII, this Court held that the meaning of the defined term could differ for different provisions serving distinct purposes in Title VII.16 Id. at 343-344. The "identical terms" maxim does not acquire special force just because the terms are used in a definition.17
3. a. In any event, the PSD provisions themselves demon strate that Congress intended EPA to have ample discretion to construe ambiguities in the PSD statutory scheme in a way that would be sensitive to the particular features and pur poses of the PSD program. In several provisions, Congress required EPA to promulgate regulations specifically imple menting the PSD program. See 42 U.S.C. 7471, 7475(a)(2), 7607(d)(1)(J). In Section 7470, Congress specified a unique set of goals for the PSD program in particular, which are not identical to the goals of the Clean Air Act as whole, set forth in Section 7401 of the Act. Among the PSD-specific purposes are "to protect public health and welfare from any actual or potential adverse effect which in the Administrator's judg ment may reasonably be anticipate[d] to occur from air pollu tion * * * notwithstanding attainment and maintenance of all national ambient air quality standards," "to preserve, protect, and enhance the air quality in national parks, national wilder ness areas, national monuments, national seashores, and other areas of special national or regional natural, recre ational, scenic, or historic value," "to insure that economic growth will occur in a manner consistent with the preserva tion of existing clean air resources," and "to assure that any decision to permit increased air pollution * * * is made only after careful evaluation of all the consequences of such a deci sion." 42 U.S.C. 7470(1), (2), (3), and (5).
By enacting a particularized set of "goals and purposes" for the PSD program, Congress necessarily recognized the corollary principle: that, insofar as those purposes would not, in EPA's judgment, be fulfilled by regulations identical to those promulgated for use in the NSPS program, the PSD regulations may be different from those applicable to NSPS. That principle extends even to regulations defining a common term, such as "increase the amount of any air pollutant," that is used in the two programs. While any regulatory defi nition of "increase" must be consistent with the statutory definition in Section 7411(a)(4), Congress authorized EPA to tailor its regulatory treatment of the constituent terms of that definition, insofar as they are ambiguous, to the "goals and purposes" of each program.18
b. There are other differences between the NSPS and PSD programs that may warrant EPA, in the course of re solving statutory ambiguities, to do so differently with re spect to each.19 The NSPS program was enacted by Congress as part of the Clean Air Act Amendments of 1970, Pub. L. No. 91-604, 84 Stat. 1676, and required EPA to develop broadly applicable, uniform, technology-based emissions standards for new or modified sources in specific industrial source cate gories. 42 U.S.C. 7411. Those standards are based on appli cation of the best demonstrated system of emission reduction for particular industries as a whole, and apply regardless of the actual effect that a source's emissions has on air quality. See ibid.; Potomac Elec. Power Co. v. EPA, 650 F.2d 509, 518 (4th Cir. 1981), cert. denied, 455 U.S. 1016 (1982) . That focus on efficient technology, independent of localized effect, makes a focus on maximum hourly emissions rates reasonable.
The PSD provisions focus on ambient air quality and ap ply, through a permitting program, to sources that have the potential to adversely impact such air quality. Thus, rather than focus solely on technology-based performance standards, as NSPS does, the PSD program focuses directly on the effect of new construction and modification on local air quality, which, in turn, makes a focus on the total amount of resulting emissions in the 1980 and 1992 regulations reasonable. 42 U.S.C. 7470(1), 7475(a)(3), (6), and (7); see Northern Plains Res. Council v. United States EPA, 645 F.2d 1349, 1356 (9th Cir. 1981), whereas the 1970 NSPS program is "equipment oriented" and applies regardless of effects on overall air qual ity, the "site-oriented" PSD program is "focused on where the plant will be located and its potential effect on its environs"); Potomac Elec. Power, 650 F.2d at 518 (significant difference between PSD and NSPS is that PSD's purpose "is to preserve existing air quality" while NSPS applies "without regard to the effect the emissions * * * will have on overall air quality").
Congress thus enacted PSD under a separate statutory mandate to regulate sources that might contribute to signifi cant deterioration of local air quality through increased emis sions, despite pre-existing CAA provisions, including the technology-focused NSPS. 42 U.S.C. 7470; WEPCO, 893 F.2d at 904 (noting that the NSPS program had not been entirely successful); 123 Cong. Rec. 18,022 (1977) (statement by Sena tor Muskie that "[o]ne purpose of the committee provision to prevent significant deterioration is to try to reverse the cur rent trend in air pollution. * * * The record to date under the new source performance standards approach has been disap pointing."). Such differences in the "purposes" and "scope of the legislative power exercised," Atlantic Cleaners, 286 U.S. at 433, indicate that Congress did not mandate a single inter pretation of common statutory terms.20
The judgment of the court of appeals should be reversed.
PAUL D. CLEMENT
SUE ELLEN WOOLDRIDGE
Assistant Attorney General
THOMAS G. HUNGAR
Deputy Solicitor General
JAMES A. FELDMAN
Assistant to the Solicitor
KATHERINE J. BARTON
JASON A. DUNN
ANN R. KLEE
CHET M. THOMPSON
Deputy General Counsel
GRANTA Y. NAKAYAMA
THOMAS W. SWEGLE
CAROL S. HOLMES
DAVID W. SCHNARE
United States Environmental Protection Agency
1 In 1974, EPA had established a regulatory PSD program as a result of a lawsuit, Sierra Club v. Ruckelshaus, 344 F. Supp. 253 (D.D.C.1972), aff'd without opinion, 4 Env't Rep. Cas. (BNA) 1815, No. CIV. A. 1031-72 (D.C. Cir. Nov. 1, 1972), aff'd by an equally divided court, 412 U.S. 541 (1973). See 39 Fed. Reg. 42,510 (1974). The 1977 Amendments "drew upon, expanded, and superseded" that earlier regulatory program. New York v. U.S. EPA, 413 F.3d 3, 12 (D.C. Cir. 2005).
2 EPA's 1980 PSD regulations setting forth the minimum requirements for EPA-approved state PSD programs were originally promulgated at 40 C.F.R. 51.24 (1981) and were later redesignated at 40 C.F.R. 51.166 (1987). Unless otherwise indicated, this brief cites to the 1987 recodification, as did the court of appeals. Both the North and South Carolina SIPs, which provide the law underlying the enforcement actions in this case, include EPA-approved PSD programs that are based on the requirements of those EPA regulations. See N.C. Admin. Code tit. 15A r.2D.0530; S.C. Code Ann. Regs. 61-62.5, std. 7; see also 47 Fed. Reg. 7836 (1982) (approving original North Carolina PSD regulations); 47 Fed. Reg. 6017 (1982) (same for South Carolina); 60 Fed. Reg. 51,923 (1995) (North Carolina incorporation of 1992 amendments of EPA's rules).
3 Since 1980, a number of new programs that improve air quality have come into existence. Those programs include the Acid Rain Program, 42 U.S.C. 7651-7651o, the NOx SIP Call (63 Fed. Reg. 57,356 (1998)) , and the Clean Air Interstate Rule (CAIR), promulgated by EPA in 2005 to reduce interstate transport of SO2 and NOx emissions. See 70 Fed Reg. 25,162.
4 The court of appeals stated that "the parties agree" that "[t]he 1980 regulations * * * control the projects at issue here." Pet. App. 5a n.1; cf. id. at 36a. The government did take the position that the 1980 and 1992 regulations did not materially differ with respect to the legal question at issue here. But the government clearly informed the courts below that some of the projects in this case were governed by the 1992 regulations. See Gov't C.A. Br. 6.
5 The focus on "representative" operations also makes clear that the PSD reuglations are not, like the NSPS regulations, focused on maximum hourly rates. There is no reason to suspect that representative operations will be at maximum capacity, and a regime focused on maximum hourly rates would have no reason to assess representative operations. That assessment, of course, is critical in assessing likely future emissions for an "actual to actual" comparison.
6 The hypothetical also provides an example of how the "hours of operation" exclusion, discussed in further detail below, is intended to function. Normally a simple increase in hours of operation in response to increased demand for a product is excluded from the regulatory definition of a physical or operational "change." However, because the increased hours of operation described in the example required a permit change, the increase was considered an operational "change," under the express terms of the regulatory exclusion. See 45 Fed. Reg. at 52,705; 40 C.F.R. 51.166(b)(2)(iii)(f) (excluding hours-of-operation increase from "physical change or change in the method of operation" "unless such change would be prohibited under any federally enforceable permit condition" established under certain regulations).
7 Aside from the court of appeals in this case, the other courts of appeals that have addressed the issue have concluded that a physical or operational change that results in increased emissions because of increased hours of operation is a "modification" under the regulations. See WEPCO, 893 F.2d at 915, 918 ("Unlike NSPS, PSD is concerned with changes in total annual emissions, expressed in tons per year"; PSD measures "the maximum emissions that can be generated while operating the source as it is intended to be operated and as it is normally operated") (quoting United States v. Louisiana-Pac. Corp., 682 F. Supp. 1141 (1988)); Puerto Rican Cement Co. v. EPA, 889 F.2d 292, 297 (1st Cir. 1989) (Breyer, J.) (noting that "EPA has simply taken account of * * * the fact that a firm's decision to introduce new, more efficient machinery may lead the firm to decide to increase the level of production, with the result that, despite the new machinery, overall emissions will increase.").
8 The 1980 PSD regulations originally established a test for "any emissions unit which has not begun normal operations on the particular date" under which actual pre-change emissions were compared with "the potential to emit of the unit on that date," i.e., the maximum potential post-change emissions. 40 C.F.R. 51.166(b)(21)(iv). The court in WEPCO, however, rejected that comparison, at least for like-kind replacements at electric utilities, suggesting instead a comparison between actual pre-change emissions and "a more realis tic" assessment of post-change emissions. 893 F.2d at 917. On remand in WEPCO, the agency applied such an actual-to-projected-actual test, and it included that test for some categories in the 1992 regulations. See J.A. 57-79; 40 C.F.R. 51.166(b)(21) (1993). For present purposes, however, what is impor tant is that both the actual-to-potential test and the actual-to-projected-actual test took either maximum potential post-change hours of operation (actual-to- potential) or likely post-change hours of operation (actual-to-projected-actual) into account in determining whether a change is a modification. Both tests focus on sums, not hourly rates. Neither test is consistent with a sole focus on hourly rate of emissions, since the hours of operation are not held constant under either test. Accordingly, the fact that the agency used both the actual-to- potential and the actual-to-projected-actual tests confirms that the agency's "actual emissions" definition of "major modification" does not limit major modifications to changes that produce an increase in hourly emissions.
9 See J.A. 255 ("the exclusion for increases in hours of operation or produc tion rate does not take the project beyond the reach of PSD coverage if those increases do not stand alone but rather are associated with non-excluded physical or operational changes"); J.A. 44 ("EPA has properly interpreted the PSD * * * regulations as applying to increases in emissions due to increases in hours of operation or production rate where, as here, such operational or pro duction increases are closely related to physical or operational changes"); see also 57 Fed. Reg. at 32,328 (explaining in 1992 preamble that "[a]lthough a source may vary its hours of operation or production as part of its everyday operations, an increase in emissions attributable to an increase in hours of operation or production rate which is the result of a construction-related activity is not excluded from review") (emphasis added).
10 If he had provided some analysis, the source of his misunderstanding might be more clear, as in the case of the district court, which clearly conflated the two distinct steps of the relevant regulatory definition.
11 Among other differences, the NSPS modification provision applies to "existing facilit[ies]," defined as "any apparatus of the type for which a standard is promulgated," see 40 C.F.R. 60.2, 60.14, while the PSD definition of "major modification" applies to any "major stationary source," defined as an entire plant or factory, see 40 C.F.R. 51.166(b)(1) and (2). The NSPS regula tions include a capital expenditure requirement in the "production rate" exclu sion, 40 C.F.R. 60.14(e)(2), while the PSD regulations do not, 40 C.F.R. 51.166(b)(2)(iii)(f), and the NSPS production rate exclusion is an exclusion from what constitutes a "modification," while the PSD exclusion is only from what constitutes a "physical change or change in the method of operation." Those differences too reinforce the emphasis in the PSD regulations, unlike the NSPS rule, on the actual annual amount emitted into the atmosphere, not the maxi mum achievable performance of a particular unit.
12 The court of appeals provided no support whatever for its "first in time" rule, which was its sole basis for preferring the NSPS regulations over the PSD regulations. See Pet. App. 18a.
13 Courts have recognized this general proposition in the specific context of the NSPS and PSD programs. E.g., Alabama Power, 636 F.2d at 396 (holding that while Congress intended the same statutory definition of "stationary source" for PSD and NSPS modification purposes, EPA retained discretion to interpret the ambiguous component terms of the same definition differently for NSPS and PSD purposes); Potomac Elec. Power Co. v. EPA, 650 F.2d 509, 517- 518 (4th Cir. 1981) ("stationary source" need not be defined the same for NSPS and PSD programs because of the "significant differences between the PSD and NSPS programs"), cert. denied, 455 U.S. 1016 (1982); cf. WEPCO, 893 F.2d at 904-905, 913 (noting that Congress "essentially adopted" the NSPS statutory definition of modification for the PSD program, but holding that "unlike NSPS," PSD focuses on increases in "total annual emissions").
14 This Court's subsequent treatment of the very terms at issue in Rowan suggests that the "identical terms" maxim alone is never dispositive at Chevron step one. In Cleveland Indians, the Court returned to the same statutory terms at issue in Rowan and used a Chevron step-two analysis to defer to the Internal Revenue Service's reasonable interpretation of them. 532 U.S. at 209, 213-214.
15 Aside from the "identical terms" maxim, the only other support suggested by the court of appeals for its conclusion was a single statement in a summary of amendments introduced into the Congressional Record by Senator Muskie and Representative Rogers that, as the court of appeals summarized, embodied an "expressed intent * * * to 'conform' the definition of modification in the PSD provisions 'to usage in other parts of the Act.'" Pet. App. 14a (quoting 123 Cong. Rec. 36,253 (1977)). The court made only the modest claim that that statement "indicates congressional concern with the same sort of simplicity and consistency that the Rowan Court discerned from the legislative history examined there." Pet. App. 14a-15a. The cited statement, however, in full ob served that the amendment in question "[i]mplements conference agreement to cover 'modification' as well as 'construction' by defining 'construction' in [PSD] to conform to usage in other parts of the Act." 123 Cong. Rec. at 36,331. To the extent that statement sheds any light on Congress's legislative purpose at all, it suggests only that the intention was to ensure that modifications be subject to PSD in the first place by defining construction to include the modifi cation of existing sources. That goal is fully accomplished under the EPA's interpretation. The statement does not suggest that Congress directly ad dressed, and clearly resolved, the question whether ambiguities in the defini tion of "modification" must receive the same resolution in the NSPS and PSD programs.
16 The decision below by the Fourth Circuit had specifically relied on the fact that the statute provided the definition of "employer" "for purposes of all provisions of Title VII" in holding that it must be interpreted identically for all purposes of the Act. Robinson v. Shell Oil Co., 70 F.3d 325, 329 (1995), rev'd, 519 U.S. 337 (1997).
17 The courts of appeals have similarly found that the "identical terms" maxim may be overridden with respect to defined, as with non-defined, terms. In Comite Pro Rescate De La Salud v. Puerto Rico Aqueduct & Sewer Authority, 888 F.2d 180, 187 (1st Cir. 1989) (Breyer, J.), for example, the court held that EPA may interpret the "various parts" of the statutory definition of "solid waste" in an environmental statute differently for the Resource Conservation and Recovery Act of 1976, 42 U.S.C. 6901 et seq., differently for different purposes under the Act, even though Congress mandated that a single definition of "solid waste" apply for all purposes of the statute. See Connecti cut Coastal Fishermen's Ass'n v. Remington Arms Co., 989 F.2d 1305, 1315 (2d Cir. 1993) (same); SKF USA, Inc. v. U.S., 263 F.3d 1369, 1379-1382 (Fed. Cir. 2001) (recognizing agency discretion to interpret single statutory definition of "foreign like product" differently for two purposes, even though the statutory definition applied for both purposes); Pharmanex v. Shalala, 221 F.3d 1151, 1155-1157 (10th Cir. 2000) (recognizing agency discretion to interpret definition of term "dietary supplement" and its component term "drug" differently in different sections of a single statute); United States v. Pornes-Garcia, 171 F.3d 142, 144-145, 147 (2d Cir.) (rejecting argument that the incorporation of the definition of "aggravated felony" from 8 U.S.C. 1101(a)(43) into the Sentencing Guidelines mandated identical interpretations of that term in both places), cert. denied, 528 U.S. 880 (1999); Abbott Labs. v. Young, 920 F.2d 984, 987 (D.C. Cir. 1990) ("it is not impermissible under Chevron for an agency to interpret an imprecise term differently in two separate sections of a statute which have different purposes"); NRDC v. EPA, 822 F.2d 104, 115-117 (D.C. Cir. 1987) (EPA reasonably gave defined term "standard of performance" two different meanings in a single sentence of 33 U.S.C. 1316(d)).
18 That is not to say that EPA is required to construe "modification" dif ferently for the PSD program than for NSPS purposes. Although at least since 1975 the NSPS regulatory definition of the same statutory term has looked to maximum potential hourly emissions and was left undisturbed when Congress extended the NSPS statutory definition to the PSD program in 1977, the D.C. Circuit in New York erroneously held that "the CAA unambiguously defines 'increases' in terms of actual emissions" for PSD purposes. 413 F.3d at 39. Congress's use of the same definition of "modification" for both programs demonstrates that EPA would be free to employ the same regulatory approach for both if it concluded that such an approach would best advance the various goals of the Act, and the new rules proposed in 2005 take a step in that direction. See pp. 11-12, supra. But nothing in the Act compels the agency to reach that conclusion.
19 EPA has consistently explained that the distinct purposes and structures of the NSPS and PSD programs can justify different interpretations of "modi fication." Even before the 1980 PSD regulations, EPA concluded that those distinctions meant that it was not "bound to apply mechanically" the NSPS definition of "modification" for PSD. 43 Fed. Reg. 26,394 (1978). That understanding has continued through promulgation of the regulations at issue here. See 45 Fed. Reg. at 52,713 (rejecting comment urging that PSD should apply to certain pollutants only if NSPS is also applicable, because "the Act requires PSD review, regardless of whether another rule already applies to the source"); 57 Fed. Reg. at 32,316 (explaining divergence between PSD and NSPS regulatory modification tests despite fact that both tests "are based on the broad NSPS definition of 'modification' in section 111(a)(4) of the CAA").
20 Finally, the district court's alternative rationale for rejecting EPA's application of an actual, annual emissions test under the PSD program-that Congress's incorporation of the NSPS statutory definition of "modification" into the PSD provisions also incorporated EPA's then-existing NSPS regula tions interpreting "modification" and deprived EPA of discretion to depart from them, see Pet. App. 58a, 62a-67a-is incorrect, and fundamentally at odds with the principle of administrative deference at the core of Chevron. Indeed, even the court of appeals rejected the proposition that Congress essentially codified the 1977 NSPS regulations when it added the PSD provisions in 1977. See Pet. App. 15a n.7 (The PSD regulations * * * could even be enforced as the EPA urges"), 18a ("The EPA retains its authority to amend and revise this and other regulations"). The district court's rationale is also inconsistent with this Court's recognition that, even when Congress reenacts a statutory provision that is subject to a long-standing administrative interpretation, the agency retains discretion to adopt other reasonable interpretations of the statute through its rulemaking power. Helvering v. Reynolds, 313 U.S. 428, 432 (1941).