Full opinion text
MEMORANDUM OPINION BULLOCK, District Judge. On December 22, 2000, the Attorney General of the United States acting at the request of the Administrator of the United States Environmental Protection Agency (“EPA”) filed this action against Defendant Duke Energy Corporation (“Duke Energy”). The EPA alleges that Duke Energy made modifications to and operated eight coal-fired electrical generating plants in North Carolina and South Carolina in violation of the Clean Air Act (the “CAA” or “Act”), 42 U.S.C. § 7401 et seq., specifically the Prevention of Significant Deterioration (“PSD”) provisions of the Act, 42 U.S.C. §§ 7470-92, and for violations of State Implementation Plans approved under the Act for the States of North Carolina and South Carolina. On May 8, 2001, the following groups, Environmental Defense, North Carolina Sierra Club, and North Carolina Public Interest Research Group Citizen Lobby/Education Fund (hereinafter collectively referred to as “Intervenor-Plaintiffs”) moved to intervene as plaintiffs. On September 6, 2001, the court granted these groups’ motion to intervene, United States v. Duke Energy Corp., 171 F.Supp.2d 560 (M.D.N.C.2001), and on the same date the Intervenor-Plaintiffs filed a complaint against Duke Energy alleging similar violations as those in the original complaint. After a lengthy discovery period, which produced approximately 4.6 million pages of documents, extensive discovery disputes, and numerous pretrial motions, the parties now move for summary judgment. The EPA, Intervenor-Plaintiffs, and Duke Energy have submitted lengthy briefs, accompanied by thousands of pages of exhibits, and the motions for summary judgment are ripe for resolution. For the reasons that follow, the court will deny Duke Energy’s motion for summary judgment, grant in part and deny in part the EPA’s motion for partial summary judgment in which Intervenor-Plaintiffs join, deny Intervenor-Plaintiffs’ motion for partial summary judgment on Duke Energy’s “fair notice” defense, and deny Duke Energy’s motion for partial judgment on the pleadings. FACTS Duke Energy is an international energy company headquartered in Charlotte, North Carolina, and has provided electricity to North Carolina and South Carolina since the early 1900s. Duke Energy’s system of electrical generating facilities includes nuclear, hydroelectric, gas-fired combustion turbine, and coal-fired generating plants. Duke Energy currently operates thirty coal-fired electric generating units at eight plants located throughout North Carolina and South Carolina. It is these coal-fired units which are the subject of this litigation. Duke Energy’s coal-fired generating units were placed in service between 1940 and 1975. These units have net rated generating capacities from 38 MW (for some of the oldest units) to 1120 MW (for the newest units). Each unit of a coal-fired generating plant consists of three primary components: the boiler, the steam turbine, and the electric generator. The boiler on these units is a large, building-like structure ranging from six to twenty stories tall and contains thousands of steel tubes in which water is heated to superheated steam with temperatures in excess of 900 F. While the specific design of each boiler at issue may differ slightly, in general the boiler consists of collections of tube assemblies, including the economizer tubes, where water is initially heated; the furnace waterwall tubes, where water evaporates to steam; the superheater tubes, where the temperature of the steam is raised just before the steam exits the boiler and reaches the turbine; and the reheater tubes, where steam from the turbine is reheated and returned to the turbine. The furnace waterwall tubes form the walls of the boiler and provide an envelope for coal combustion while also absorbing heat. A coal-fired unit operates by converting the chemical energy contained in coal into electricity. Pulverized coal is fed into the boiler and combusted. “Flue gas” is created with temperatures of up to 3000 F when the ground coal is ignited in the boiler’s furnace. The flue gas, which contains sulfur dioxide, ash particles or particulate matter, and nitrogen oxides, passes around the tubes and tube assemblies. This combustion process heats water that flows continuously through the tubes and tube assemblies making up the walls, roof, and floor of the boiler. The water in the tubes is thus converted to high pressure steam, which then flows through additional panels or assemblies of tubes to become superheated. The superheated steam enters the turbines where the pressure of the steam against a series of blades turns the turbine shaft. The turbine shaft turns the shaft of a generator, which transforms the mechanical energy into electric energy. After passing through the turbine, the steam is converted to water in condensers and pumped through feedwater heaters back to the economizer where it begins the entire' steam cycle process again. Once the flue gas used to heat the water passes through the boiler, it passes through pollution control devices, if any, and exits through a stack into the atmosphere. The EPA alleges that Duke Energy modified and subsequently operated its seven coal-fired generating plants in North Carolina and its one coal-fired generating plant in South Carolina in violation of the PSD provisions of the CAA. The EPA’s allegations stem from twenty-nine projects Duke Energy engaged in between 1988 and 2000. (Compl. at ¶¶ 31-292.) A detailed description of the work performed at each plant and unit would be lengthy and ultimately unnecessary. However, the majority of the projects consisted of replacement and/or redesign of one or more of four sets of boiler tube assemblies— economizers, portions of waterwalls, super-heaters, and reheaters. (Id.) In 1984, Duke Energy placed several of the units at issue, including Buck 4, into “Extended Cold Storage” (“ECS”). During the period of ECS, dehumidified air was circulated through the water, steam, air, and gas passages in order to protect the units. The reason behind Duke Energy’s decision to place its units in ECS is disputed. Duke Energy contends that it placed these units into ECS because it increased its system generating capacity by adding additional generating plants, which in turn led to less use of its coal-fired units. Duke Energy also asserts that when it placed these units into ECS, it made definitive plans for preserving and conditioning these units while in ECS so that they could be returned to service when demand dictated. (Knudsen Decl. ¶ 5 (Duke Energy Ex. 59); Kinsey Decl. ¶ 6 (Duke Energy Ex. 69).) The EPA, however, contends that Duke Energy removed these units from service due to their advanced age and condition. William 5. Lee, Chairman of the Board and CEO of Duke Energy, testified before the South Carolina state public utility regulators in 1985 that the units placed into ECS were “no longer reliable because of their age and because their use as peaking units in the past few years ha[d] stressed the units, which originally were designed for base load use.... Historically, units of this age and condition would be retired and scrapped.” (Lee Test, at 90 (EPA Ex. 13).) Lee further explained that Duke Energy had plans to rehabilitate the units, but if it continued to operate the units in their current condition they would “simply fall apart, or damage themselves in a very serious way.” (Lee Test, at 41 (EPA Ex. 15).) Shortly after placing its units into ECS, Duke Energy developed a “Plant Modernization Program” (“PMP”). Based on information gathered during inspection of the units, Duke Energy developed plans to address a variety of maintenance, repair, and replacement needs. According to Duke Energy, the purpose for PMP was “[t]o conduct maintenance and upgrade to selected fossil generating units so that they operate safely, reliably and cost effectively for an additional 20 years.” (Plant Modernization Project Review (Apr. 24, 1989) at 2 (EPA Ex. 23).) According to Duke Energy’s “PMP Strategy Statement,” the “extended operating life of the rehabilitated units is a cost-effective alternative to the addition of new capacity.” (Mem. from Parker to Owen, et al. (Feb. 13, 1986) (EPA Ex. 25).) Any work performed at the units not placed into ECS was undertaken during regularly scheduled planned turbine outages. Regularly scheduled outages typically occur at a unit every forty-eight to eighty-four months depending on the particular unit’s condition and other system requirements. On December 22, 2000, approximately twelve years after the completion of Duke Energy’s earliest PMP project, the United States on behalf of the EPA filed the present suit against Duke Energy. The EPA contends that Duke Energy’s projects at its coal-fired units were “modifications” and thus triggered the requirements of PSD. A unit that is subject to PSD must obtain a preconstruction review and permit. If it is determined during this review that a proposed project will modify an existing unit and thereby increase the level of emissions, the operator of the unit is required to install the best available control technology for pollutants emitted by the particular unit. As noted previously, Duke Energy, the EPA, and Intervenor-Plaintiffs have all moved for summary judgment or partial summary judgment. Duke Energy seeks summary judgment dismissing all claims of the EPA and Intervenor-Plaintiffs. It argues that the EPA cannot establish that the projects undertaken at its coal-fired generating units were non-routine maintenance, repair, and replacement activities or that the projects caused a net emissions increase. As to the issue of whether Duke Energy’s projects were routine maintenance, repair, and replacement, Duke Energy contends that the established interpretation of what is routine “ ‘must be based on the evaluation of whether that type' of equipment has been repaired or replaced by sources within the relevant industrial category.’ ” (Duke Energy Br. Supp. Mot. Summ. J. at 35 (quoting 57 Fed.Reg. 32,-314, 32,326 (July 21, 1992)).) Under this standard, according to Duke Energy, its component repair and replacement. projects were not modifications because they consisted of repairs and replacements that are common in the utility industry. Duke Energy similarly, argues that its component repair and replacement projects were not modifications because they did not increase the units’ levels of emissions. Duke Energy contends that under the 1980 PSD regulations, a net emissions increase will result only if there is an increase in the hourly rate of emissions. It acknowledges that for purposes of PSD, emissions are measured in annual tons per year. Yet Duke Energy. argues that in calculating post-project actual emissions, the regulations require that the hours and conditions of operation be held constant. Accordingly, because none of the projects increased a unit’s hourly capacity to emit pollution, there is no increase in emissions from pre-project levels. The EPA and Intervenor-Plaintiffs move for partial summary judgment on similar grounds. They, too, seek a legal determination as to the appropriate standard for determining what is routine maintenance, repair, and replacement. The EPA and Intervenor-Plaintiffs argue that the standard is not the industry or source category standard advanced by Duke Energy. Rather, they contend that whether a particular project is routine maintenance, repair, and replacement must be determined based on whether that type of project is routine within the life of a generating unit. The EPA and Intervenor-Plaintiffs also request partial summary judgment that the “comprehensive renovation” at Buck 4 was a physical change that does not qualify as routine maintenance and repair. Furthermore, the EPA and Intervenor-Plaintiffs move for partial summary judgment as to Duke Energy’s defense that it did not have fair notice of the EPA’s interpretation of the PSD regulations. The EPA and Intervenor-Plaintiffs, similar to Duke Energy, also seek a legal determination as to how post-project emissions should be calculated. The EPA argues that PSD requires that the source predict annual emissions increases prior to construction. In predicting an increase in emissions, the EPA asserts, the source operator must consider the hourly rate of emissions and the increased utilization of the unit following the project. As to Buck 4 specifically, the EPA argues that because the project enabled Duke Energy to utilize a previously inoperable unit, the project resulted in a net increase in emissions. Finally, the EPA and Intervenor-Plain-tiffs seek partial summary judgment that the claims for penalties and injunctive relief with respect to modifications that were completed more than five years before this suit was filed are not barred by the statute of limitations. Duke Energy previously moved for partial judgment on the pleadings arguing that such claims were barred by the statute of limitations. The court deferred ruling on Duke Energy’s motion in order to allow the record to develop more fully. Duke Energy’s motion for partial judgment on the pleadings, as well as the EPA and Intervenor-Plaintiffs’ motion for partial summary judgment, is ripe for resolution. DISCUSSION I. Standard of Review Summary judgment must be granted if there is no genuine issue as to any material fact and the moving party is entitled to judgment as a matter of law. Fed. R.Civ.P. 56(c). The moving party bears the burden of persuasion on the relevant issues. Celotex Corp. v. Catrett, 477 U.S. 317, 323, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986). The non-moving party may survive a motion for summary judgment by producing “evidence from which a [fact finder] might return a verdict in [its] favor.” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 257, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). When the motion is supported by affidavits, the non-moving party must set forth specific facts showing that there is a genuine issue for trial. Nee Fed.R.Civ.P. 56(e); see also Cray Communications, Inc. v. Novatel Computer Sys., Inc., 33 F.3d 390, 393-94 (4th Cir.1994) (moving party on summary judgment motion can simply argue the absence of evidence by which the non-movant can prove its case). In considering the evidence, all reasonable inferences are to be drawn in favor of the non-moving party. Anderson, 477 U.S. at 255, 106 S.Ct. 2505. However, “[t]he mere existence of a scintilla of evidence in support of the plaintiffs position will be insufficient; there must be evidence on which the [fact finder] could reasonably find for the plaintiff.” Id. at 252, 106 S.Ct. 2505. In general, reviewing courts typically grant substantial deference to the EPA’s interpretation of the CAA Amendments and its implementing regulations. The reasoning behind this deferential review is that “considerable weight should be accorded to an executive department’s construction of a statutory scheme it is entrusted to administer.” Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 844, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). Deference to agency interpretation is particularly appropriate where the subject being regulated is technical and complex. Aluminum Co. of Am. v. Central Lincoln Peoples’ Util. Dist., 467 U.S. 380, 390, 104 S.Ct. 2472, 81 L.Ed.2d 301 (1984). An agency’s interpretation of its own regulations must be given “ ‘controlling weight unless it is plainly erroneous or inconsistent with the regulation.’ ” Udall v. Tallman, 380 U.S. 1, 16-17, 85 S.Ct. 792, 13 L.Ed.2d 616 (1965) (quoting Bowles v. Seminole Rock & Sand Co., 325 U.S. 410, 413-14, 65 S.Ct. 1215, 89 L.Ed. 1700 (1945)). However, “this standard does not give the EPA unbridled discretion to construe the [CAA] Amendments free from judicial oversight. [The court] must consider whether the EPA’s construction comports with its statutory mandate and Congress’s intent in enacting clean air legislation.” Wisconsin Elec. Power Co. v. Reilly, 893 F.2d 901, 907 (7th Cir.1990) (“WEPCO ”). II. Statutory and Regulatory Background One of the primary purposes of the CAA is to “speed up, expand, and intensify the war against air pollution in the United States with a view to assuring that the air we breathe throughout the Nation is wholesome once again.” H.R.Rep. No. 91-1146, at 1 (1970), reprinted in 1970 U.S.C.C.A.N. 5356, 5356. In order to accomplish the congressional objective of “protecting] and enhancing] the quality of the Nation’s air resources,” 42 U.S.C. § 7401(b)(1) (1995), Congress directed the EPA to develop National Ambient Air Quality Standards (“NAAQS”) which specify the maximum allowable concentrations of air pollutants for different areas of the country. Id. § 7409. Based on the levels of pollution established by the EPA, the states were required to develop State Implementation Plans (“SIPs”) that defined source-by-source emissions limits so that each state could meet the NAAQS. Id. § 7410; Train v. Natural Res. Def. Council, Inc., 421 U.S. 60, 66-67, 95 S.Ct. 1470, 43 L.Ed.2d 731 (1975). Areas that meet the NAAQS for a particular pollutant are designated as “attainment” areas and areas that do not meet the NAAQS are designated “non-attainment” areas. 42 U.S.C. § 7407(d) (1995). As part of the 1970 CAA Amendments, Congress required the EPA to promulgate New Source Performance Standards (“NSPS”) in order to regulate and minimize the environmental impact from the emission of air pollutants from “new sources.” The NSPS regulates hourly emission rates (expressed as kg/hr) and applies to both newly constructed sources and to “modifications” of existing facilities that create new or increased pollution. Under NSPS, Congress defined “new source” as “any stationary source, the construction or modification of which is commenced after the publication of regulations (or, if earlier, proposed regulations) prescribing a standard of performance under this section which will be applicable to such source.” Id. § 7411(a)(2) (emphasis added). Congress then defined “modification” as “any physical change in, or change in the method of operation 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.” Id. § 7411(a)(4). Due to the cost and difficulty of installing new pollution control technologies on existing sources, the EPA provided exemptions from the “modification” rule for certain activities undertaken at already operating sources. The regulations initially provided that “[r]outine maintenance, repair, and replacement shall not be considered physical changes.” 40 C.F.R. § 60.2(h)(1) (1971) (Duke Energy Ex. 4). This standard was later clarified to mean that a modification shall not include “[mjaintenanee, repair, and replacement which the Administrator determines to be routine for a source category.” 40 C.F.R. § 60.14(e)(1) (1975) (Duke Energy Ex. 6). The regulations also provided that increases in hours of operation or production rates within the operating design capacity of a facility are not considered modifications. 40 C.F.R. § 60.2(h)(2)(ii) (1971) (Duke Energy Ex. 4); 40 C.F.R. § 60.14(e)(2), (3) (1975) (Duke Energy Ex. 6). In 1977 Congress again amended the CAA by enacting a program called New Source Review (“NSR”). The NSR had provisions for the protection of areas with relatively clean air, Prevention of Significant Deterioration (“PSD”), and for areas that did not meet certain air quality standards, Non-Attainment New Source Review (“NNSR”). PSD was designed to ensure that the air quality of relatively unpolluted areas, i.e., attainment areas, did not decline to the minimum levels permitted by NAAQS due to increases in total annual emissions. To prevent significant deterioration of air quality, operators of regulated sources in attainment areas are required to limit emissions to a “baseline rate” and obtain a permit before constructing or modifying facilities. 42 U.S.C. § 7475(a)(1) (1995). When Congress enacted the PSD program, it incorporated by explicit reference the NSPS definition of modification into the NSR definition of construction/modification. Id. § 7479(2)(C)(PSD) (“The term ‘construction’ ... includes the modification (as defined in section 7411(a) of this title [NSPS]) of any source or facility.”); id. § 7501(4) (NNSR) (“The terms ‘modifications’ and ‘modified’ mean the same as the term ‘modification’ as used in section 7411(a)(4) of this title [NSPS].”). The PSD statutory definition incorporated not only the NSPS statutory definition of modification, but also the regulations implementing the NSPS program. A House-Senate Conference Committee report explained the congressional intent “to conform” the NSR definition of modification to the “usage in other parts of the Act.” 123 Cong. Rec. H11956, 3665 (daily ed. Nov. 1, 1977) (Duke Energy Ex. 13). The EPA explained that “[t]he phrase ‘usage in other parts of the Act’ most probably refers, not only to section 111(a)(4) [NSPS], but also to the EPA regulations implementing section 111 that were in effect at the time.” 49 Fed.Reg. 43,211, 43,213 (Oct. 26, 1984) (Duke Energy Ex. 14). In addition, the Director of the Stationary Source Compliance Division, Edward E. Reich, explained: [T]he Clean Air Act provides in Section 169(l)(c) that for PSD purposes the term modification shall be defined as that term is defined in Section 111(a) of the Act relating to NSPS. EPA has interpreted this to mean that for PSD purposes Congress intended the term modification to include all exemptions included in the NSPS regulations promulgated under Section 111 of the Act prior to the date of enactment of Section 169. (Mem. from Reich to Davis (Apr. 21, 1983) at 2 (Duke Energy Ex. 16).) Subsequent to the congressional enactment of NSR, the EPA promulgated regulations for PSD. The regulation at the center of the controversy between the parties is the regulation concerning the term “modification.” Under its 1980 PSD regulations, which the parties acknowledge as the controlling regulations, the EPA defined modification as “any physical change in or change in the method of operation 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) (1987) (Duke Energy Ex. 21). Accordingly, to trigger the PSD’s permitting requirement and the requirement to install pollution controls, two criteria must be satisfied: (1) there, must be • a “physical change” and (2) there must be a “significant net emissions increase.” • The term “physical change” has been interpreted broadly. “[T]he most trivial activities — the replacement of leaky pipes, for example — may trigger the modification provisions if the change results in an increase in the emissions of a facility.” WEPCO, 893 F.2d at 905. The EPA promulgated under the PSD program regulatory exclusions from the definition of modification using substantially the same terms it used in the NSPS regulations. Specifically, the EPA created an exemption for “[r]outine maintenance, repair, and replacement” (“RMRR”). 40 C.F.R. § 51.166(b)(2)(iii)(a) (1987) (Duke Energy Ex. 21). The meaning and scope to be afforded this exemption is one of the central disagreements between Duke Energy and the EPA. The issue for the court to decide is whether RMRR should be defined relative to an industrial category or to a particular unit. The EPA’s PSD regulations also provided 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.” Id. § 51.166(b)(2)(iii)(f). The second criteria necessary to trigger the PSD’s requirements is that the physical or operational change must cause a “significant net emissions increase.” Under the PSD regulations, a “net emissions increase” is defined as “[a]ny increase in actual emissions from a particular physical change or change in the method of operation at a stationary source.” Id. § 51.166(b)(3)(i)(a). Unlike NSPS which measures emissions increases based on an hourly rate, PSD measures emissions based on total annual emissions. The second primary issue of disagreement between the parties is the appropriate emissions test that should be employed in calculating the post-project emissions rates. III. Routine Maintenance, Repair, and Replacement The court is presented with two different interpretations of the RMRR exemption. As described by the Eleventh Circuit, the “central disagreement between [the utility] and EPA is whether ‘routine’ should be defined relative to an industrial category or to a particular unit.” Tennessee Valley Auth. v. United States EPA, 278 F.3d 1184, 1189 n. 3 (11th Cir.2002). The EPA argues that the RMRR exemption requires “a case-by-case determination of whether the activity is routinely performed at an individual unit within the relevant industrial category, considering common-sense factors such as nature and extent, purpose, frequency, and cost,” (EPA Mem. Supp. Mot. Partial Summ. J. at 1), and that this has been the EPA’s “long-standing interpretation.” (Id. at 23.) Conversely, Duke Energy asserts that the “ ‘routine’ inquiry has as its ultimate purpose the determination of whether a project is routine in the industry, considering all relevant facts — e.g., nature and extent, scope, frequency, and cost.” (Duke Energy Am. Br. Opp’n Mot. Partial Summ. J. at 29.) A. Legislative intent The EPA has promulgated under its regulations an RMRR exemption since the enactment of the NSPS. 40 C.F.R. § 60.2(h)(1) (1971) (Duke Energy Ex. 4) (“Routine maintenance, repair, and replacement shall not be considered physical changes.”). This exemption was designed to achieve the congressional intent of not subjecting existing sources to the costly requirements of installing advanced pollution control devices. Such a provision was necessary because of the expansive meaning afforded “physical change” under the definition of modification. In 1975, the EPA clarified that the RMRR provision under NSPS excluded from the definition of modification “[mjaintenance, repair, and replacement ... routine for a source category.” 40 C.F.R. § 60.14(e)(1) (1975) (Duke Energy Ex. 6); 40 Fed.Reg. 58,416, 58,419 (Dec. 16, 1975) (Duke Energy Ex. 7). Under this regulatory framework, Congress enacted the PSD program in 1977. A generating facility is subject to the strictures of the PSD program if it undergoes “construction,” which Congress defined by explicitly incorporating the definition of “modification” under NSPS. 42 U.S.C. § 7479(2)(C) (1995). In accordance with this statutory mandate, the EPA defined the term “modification” as “any physical [or operational] change ... 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) (1987) (Duke Energy Ex. 21). Also consistent with the NSPS definition of modification, the EPA promulgated a provision excluding from the definition of “modification” projects undertaken at a facility that are “[rjoutine maintenance, repair, and replacement.” Id. § 51.166(b)(2)(iii)(a). In order to give the PSD RMRR exemption its proper scope, this provision must be defined according to what is routine maintenance, repair, and replacement within the relevant source category. This construction is compelled by the statutory mandate of the PSD program and congressional intent. When Congress enacted the NSR program, it specifically stated its intent to incorporate the NSPS “usage” of the term “modification” into PSD, including the 1975 NSPS exclusion for maintenance, repair, and replacement projects that are routine for a source category. 42 U.S.C. § 7479(2)(C) (1995) .(“The term ‘construction’ ... includes the modification (as defined in [NSPS]) of any source or facility.”); 123 Cong. Rec. H11966, 8665 (daily ed. Nov. 1, 1977) (Duke Energy Ex. 13) (explaining Congress’s intent “to conform” NSR definition of modification to the “usage” in NSPS). The EPA confirmed this congressional design. (Duke Energy Ex. 16 at 2) (“EPA has interpreted ... that for PSD purposes Congress intended the term modification to include all exemptions included in the NSPS regulations promulgated ... prior to the date of [PSD’s] enactment ... , ”). Therefore, to be consistent with the NSPS “usage” of RMRR, the RMRR provision under PSD must be applied to cover those projects that are routine for a source category. See INS v. Cardoza-Fonseca, 480 U.S. 421, 448, 107 S.Ct. 1207, 94 L.Ed.2d 434 (1987) (“ ‘If a court, employing traditional tools of statutory construction, ascertains that Congress had an intention on the precise question at issue, that intention is the law and must be given effect.”’ (quoting Chevron U.S.A., 467 U.S. at 843 n. 9, 104 S.Ct. 2778)). The EPA acknowledged this standard for the PSD’s RMRR exemption in 1992 by: clarifying that the determination of whether the repair or replacement of a particular item of equipment is “routine” under the NSR regulations, while made on a case-by-case basis, must be based on the evaluation of whether that type of equipment has been repaired or replaced by sources within the relevant industrial category. 57 Fed.Reg. 32,314, 32,326 (July 21, 1992). The EPA specifically referred to “electric utilities” as a “source category.” Id. at 32,317. B. WEPCO applicability determination Prior to this affirmation of the routine in the industrial category standard for RMRR, the EPA issued an NSPS and PSD applicability determination for a proposed project at Wisconsin Electric Power Company’s (“WEPCO”) Port Washington facility. The EPA explained that “[i]n determining whether proposed work at an existing facility is ‘routine,’ EPA makes a case-by-case determination by weighing the nature, extent, purpose, frequency, and cost of the work, as well as other relevant factors, to arrive at a common-sense finding.” (Mem. from Clay to Kee (Sept. 9, 1988) at 3 (EPA Ex. 73).) The EPA intended this determination, which was set forth in a memorandum prepared by Don R. Clay, Acting Assistant Administrator for Air and Radiation, to be used not only by WEPCO, but also by other utilities which might seek to claim the RMRR exemption. (Id. at 2.) WEPCO proposed a project at its five-unit Port Washington Power Plant that contemplated replacing rear steam drums, each of which measured 60 feet in length, 50.5 inches in diameter, and 5.25 inches in thickness; plate-type air heaters; and other components. (Id. at 4.) The five units at the Port Washington plant were originally rated at “80 megawatts electrical output capacity,” but due to the deterioration of the physical plant, the units had operated at a maximum achievable capacity less than their original design capacity. (Id. at 2; Mem. from Kee to Emison (Mar. 25, 1988) (Duke Energy Ex. 31) (indicating that units had been derated from their design capacity of 80 MW each to operating capacities ranging from 49 MW to 75 MW).) The EPA indicated that following the proposed project “[pjlantwide capacity would increase about 40 percent above current levels to 400 megawatts.” (Duke Energy Ex. 31 at 2.) Given this information, the EPA determined that the “work proposed at Port Washington [was] far from being a regular, customary, or standard undertaking.” (EPA Ex. 73 at 3.) “Rather, [it was] a highly unusual, if not unprecedented, and costly project.” (Id. at 4.) In subsequent correspondence with WEPCO, the EPA further indicated that based on an informal survey, it had detected “no examples of steam drum replacement at aged electric generating facilities.” (Letter from Thomas to Boston (Oct. 14, 1988) at 4 (Duke Energy Ex. 29).) Similarly, in response to WEPCO’s contention that forty other units had replaced similar air heater elements, the EPA concluded that “no units containing plate elements such as those on units 1-4 at Port Washington” had ever replaced air heaters. (Letter from Clay to Boston (Feb. 15, 1989) at 7 (Duke Energy Ex. 30).) Accordingly, the EPA concluded that WEP-CO’s proposed project was not RMRR and the projects were therefore subject to NSPS and PSD. WEPCO appealed the EPA’s determination to the Seventh Circuit Court of Appeals. The court affirmed the EPA’s determination that the project proposed at the Port Washington plant was not RMRR. WEPCO, 893 F.2d at 910-13. The court agreed with the EPA that “far from being routine, the Port Washington project apparently was unprecedented: WEPCO did not identify, and EPA did not find, even a single instance of renovation work at any electric utility generating station that approached the Port Washington life extension project in nature, scope or extent.’ ” Id. at 911 (quoting EPA Br. at 44). Although the record was silent, the court “surmised” that the unprecedented character of the Port Washington project “may reflect historical practice in the electric utility industry” to replace facilities once they have reached their useful life. Id. Among its arguments before the Seventh Circuit, WEPCO argued, as it had before the EPA, that its proposed replacement of air heaters paralleled similar replacements of air heaters at forty other plants. Id. In affirming the EPA’s determination that the forty other air heater replacements were dissimilar, the court noted that “the plate-type air heaters at issue in the Port Washington project must be replaced in whole ” while the air heaters at the forty other units, which contained tubular type heaters, “permitted] the replacement of the heat transfer surface without requiring the removal of the entire unit.” Id. The EPA and the Seventh Circuit both confirmed the relevance of industry practice under the RMRR analysis. The EPA gave considerable attention to WEPCO’s contention that the types of replacements it contemplated undertaking were “routine” because others in the industry had undertaken similar projects. While in the end the EPA disagreed, its disagreement was with WEPCO’s characterization of the projects undertaken elsewhere, not with WEPCO’s position that'what other utilities were doing should inform the EPA’s analysis. For example, in the Clay Memorandum, the EPA stated that the “work called for under the project was rarely, if ever, performed.” (EPA Ex. 73 at 5.) This conclusion was confirmed by WEPCO which stated: “Generally, the renovation work items included in this application are those that would normally occur only once or twice during a unit’s expected life cycle.” (Id.) If the relevant inquiry under the RMRR exemption is whether a particular activity is “routinely performed at an individual unit” as the EPA now asserts, the EPA in WEPCO could have simply concluded its RMRR inquiry with the admission by WEPCO that the proposed project would occur only once or twice during a unit’s expected life cycle. The EPA, however, requested that WEPCO “submit information regarding the frequency of replacement of steam drums, the largest category of work item called for under the project.” (Id.) In response to this request, “WEPCO reported that to date, no steam drums have ever been replaced at any of its coal-fired electrical generating facilities.” (Id.) WEPCO was able to provide examples of other “headers” comparable in design pressure and function, but the EPA distinguished these headers on the ground that they were not comparable in size. (Id. at 5-6 (stating that “the largest of these [headers] was 16 inches in diameter, and EPA does not believe that they are comparable in diameter, wall thickness, function, or importance to the rear steam drums at Port Washington”).) The characteristic used to differentiate these “header” replacements from the rear steam drum header replacements proposed by WEPCO was not the extent to which the other headers had or had not been replaced, but rather was the size of the header and presumably the extent and cost of the work required to replace them. In addition, WEPCO argued to the EPA that its rear steam drum replacements were routine, principally relying on its identification of four other steam drum replacement projects at other facilities. The EPA rejected WEPCO’s reliance on these steam drum projects, finding that they were not “sufficiently similar to the Port Washington project.” (Duke Energy Ex. 29 at 3.) The EPA distinguished several of the projects from WEPCO’s project primarily on the ground that they did not involve utility boilers, i.e., they were not in the same source category. (Id. at 4.) Moreover, the EPA indicated that the results of its informal survey revealed “no examples of steam drum replacement at aged electric generating facilities.” (Id.) The fact that no other utilities replaced steam drums can be relevant only if the appropriate inquiry is what is routine within the industry. Otherwise, there would be no need for the EPA to conduct an informal survey given that steam drums are replaced only once or twice in the life of a generating unit. The EPA attempts to bolster its current position that the EPA did not apply a routine in the industry standard in WEP-CO by citing the EPA’s conclusion as to the forty units that WEPCO alleged engaged in similar air heater replacement projects. The EPA concluded that the projects performed at the forty units identified by WEPCO involved a different type of air heater and were therefore “too dissimilar to the plate-type elements in use at units 1-4 to support WEPCO’s contention that the work in question is routine.” (Duke Energy Ex. 30 at 7.) This conclusion was affirmed by the Seventh Circuit. WEPCO, 893 F.2d at 911. The EPA continued, however, and stated: Further, even the list of air heater replacement work submitted by WEPCO did not establish this as routine repair work. Those 40 units comprise only a small fraction of total operating utility units, and even at the 40 units, air heater repair or replacement appears to have been a one-time occurrence, not routine repair. (Duke Energy Ex. 30 at 7 n. 6.) Contrary to the EPA’s position, this reference further supports the industry standard for RMRR. Even if the replacement projects at these forty units were similar to the WEPCO project, the EPA determined that a particular project performed at forty units, which “comprise only a small fraction of total operating utility units,” is not representative enough of the industry to establish that the project was routine. Furthermore, the EPA concluded in the alternative that this type of project could not be considered routine for the industry because these forty units did not undertake this type of project with enough frequency. If the am heater replacements were frequently performed at an individual unit and therefore “routine repair,” by implication air heater replacements would be “routine repair” for the industry. The EPA’s continual reference to other projects within the utility industry confirms Congress’s intent to define RMRR under PSD according to the relevant source category. The EPA, however, continues to argue that the reference in the WEPCO applicability determination to the fact that the proposed replacements would be performed only once or twice during the life of a unit reveals the appropriate standard for RMRR to be routine at an individual unit and that its interpretation is entitled to deference. To accept this proposition would require that one completely ignore all references in the WEPCO determination to what was done in the industry. Moreover, “ ‘an agency’s interpretation of a statute or regulation that conflicts with a prior interpretation is entitled to considerably less deference than a consistently held agency view.’” Miller v. AT &T Corp., 250 F.3d 820, 832 (4th Cir.2001) (quoting Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 515, 114 S.Ct. 2381, 129 L.Ed.2d 405 (1994)). The EPA further asserts that its position is supported by the Seventh Circuit’s decision, arguing that the court was “particularly persuaded” by the fact that the proposed activity occurred so infrequently at the particular units. (EPA’s Mem. Supp. Mot. Partial Summ. J. at 22.) Duke Energy rightly points out, however, that the “opinion itself ... ascribes no overriding importance to that fact.” (Duke Energy’s Am. Br. Opp’n Mot. Partial Summ. J. at 17.) Indeed, the next sentence of the opinion provides: “WEPCO reported that it had never previously replaced a steam drum or ‘header’ of comparable size at any of its coal-fired electrical generating facilities.” WEPCO, 893 F.2d at 912. The court’s recognition of both the frequency of a project at an individual unit and in the industry confirms the relevant scope of the RMRR inquiry to be what is routine within the industry. The only way in which both experience at a unit and in the industry at large can be relevant is under a routine within the industry standard. Projects that are repeatedly performed at a particular unit will be routine in the industry, as will projects performed at a number of units within the industry. If the relevant inquiry were that proposed by the EPA, namely that RMRR applies only to “activity routinely performed at an individual unit,” consideration of what occurred within the utility industry is irrelevant. The EPA in WEPCO could have simply dismissed WEPCO’s attempt to support the routineness of its proposed project by reference to other projects, as it does with Duke Energy here, and relied primarily on WEPCO’s admission that the proposed project would be performed only once or twice in the life of a unit. Accordingly, applying its multi-factored test, the EPA could have determined that the project did not qualify as RMRR. Yet even if the EPA were to consider what occurred within the industry, the relevant inquiry under a “routine for an individual unit” standard would be the number of times a particular project is undertaken at a unit. The focus of the EPA’s effort at distinguishing the projects identified by WEPCO from the Port Washington project, however, was on the type of equipment replaced, not the frequency with which it was replaced. The EPA’s “routine for an individual unit” interpretation of the WEPCO determination is certainly consistent with its analysis of the Buck 4 project. Nowhere in its discussion of the relevant factors as applied to this unit does the EPA consider what other utilities have done or are doing. This aspect of the WEPCO analysis has completely disappeared. What has occurred within the industry, however, is relevant and must inform the routine inquiry. This is revealed by the EPA’s careful consideration of the other projects identified by WEPCO. This is also revealed by Congress’s intent to define the application of the PSD program according to the definitions and exemptions provided under the NSPS program. The EPA confirmed this intent in 1992 and must now apply the law as intended by Congress and as previously applied. C. EPA’s post-WEPCO statements The “routine in the industry” standard is also supported by the EPA’s statements that the WEPCO determination would not affect utility life extension projects. In December 1988, the Chairman of the House Subcommittee on Oversight and Investigations, Congressman John Dingell, informed the EPA Administrator, Lee Thomas, that his subcommittee had requested the Government Accounting Office (“GAO”) to prepare a report on utility life extension issues. (Letter from Dingell to Thomas (Dec. 21, 1988) (Duke Energy Ex. 88).) The GAO issued its report in September 1990, stating that “[ajccording to EPA policy officials, WEPCO’s life extension project is not typical of the majority of utilities’ life extension projects and concerns that the agency will broadly apply the ruling it applied to WEPCO’s project are unfounded.” (GAO 1990 Report at 30-31 (Duke Energy Ex. 42).) “Lending evidence to the officials’ statements,” the report noted, “EPA’s 1989 emission forecast assumed that the WEPCO decision would not result in a significant number of additional power plants having to comply with the NSPS and the PSD program requirements.” (Id. at 31.) Chairman Dingell formally transmitted the report to the EPA and asked the Administrator about WEPCO and the GAO report’s assessment. (Letter from Dingell to Watkins et al. (Oct. 9, 1990) (Duke Energy Ex. 94).) Assistant Administrator William Rosenberg responded for the EPA, stating that “[a]s indicated in the GAO report, it is expected that most utility projects will not be similar to the WEPCO situation” and that the “[WEPCO] ruling is not expected to significantly affect power plant life extension projects.” (Letter from Rosenberg to Dingell (June 19, 1991) at 5-6 (Duke Energy Ex. 44).) The EPA’s position that life extension projects would not be significantly affected was again acknowledged in 1995. The EPA’s Assistant Administrator for Air and Radiation stated in response to an industry proposal to add a “restoration” exemption to the NSR programs that the EPA’s position was that the “routine maintenance exclusion already included in the existing NSR regulations ... has the effect of excluding ‘routine restorations’ ” from the requirements of the NSR programs. (“EPA’s Response to Issues Raised by Industry on Clean Air Act Implementation Reform,” attached to Letter from Nichols to Lewis (May 31, 1995) at 19 (Duke Energy Ex. 46).) The EPA’s position on WEPCO’s life extension project and life extension projects in general confirms the understanding that projects which are routine in the industry qualify as RMRR. To reconcile the EPA’s previously stated position with its litigation position that RMRR applies only to routine activities performed at an individual unit, one must assume that a generating unit routinely and repetitively undergoes life extension projects. This assumption defies common sense. Further, this is an assumption the EPA explicitly rejected when it assumed for the purpose of assessing future utility air emission trends that coal-fired generating utilities would undergo life extension refurbishment once around age thirty. (Duke Energy Ex. 40 at App. C.) Through the EPA’s statements in the Federal Register, its statements to the regulated community and Congress, and its conduct for at least two decades the EPA has established an interpretation of RMRR under which routine is judged by reference to whether a particular activity is routine in the industry. See Shell Offshore Inc. v. Babbitt, 238 F.3d 622, 629 (5th Cir.2001) (“existing practice” evidence of current interpretation of regulation). Accordingly, “ ‘[o]nce an agency gives its regulation an interpretation, it can only change that interpretation as it would formally modify the regulation itself: through the process of notice and comment rulemaking.’ ” Alaska Prof'l Hunters Ass’n v. FAA, 177 F.3d 1030, 1033-34 (D.C.Cir.1999) (quoting Paralyzed Veterans of Am. v. D.C. Arena, 117 F.3d 579, 586 (D.C.Cir.1997)). D. Application of the WEPCO multi-factor test The EPA in the WEPCO applicability determination identified • a multi-factored test that must be used to evaluate whether a proposed project qualifies as RMRR. The EPA characterizes Duke Energy’s RMRR test as providing that “[i]f a project has ever been done before, it is ‘routine maintenance’ ” and that “if a project has ever been performed by another utility, it is exempt.” (EPA Reply Br. Supp. Mot. Partial Summ. J. at 4.) The WEPCO multi-factored test under this characterization of the approach advocated by Duke Energy would collapse the WEPCO test into a single inquiry — has this type of project been performed before. Because the WEPCO multi-factored test applies and is entitled to deference, the determination of RMRR cannot turn exclusively on whether a particular replacement project has ever occurred in the industry. If- this were dispositive, it would render the PSD program a nullity.- The frequency with which a component is repaired or replaced within the industry, however, cannot be ignored. Nor can the frequency with which a component is repaired or replaced at a particular unit be conclusive. Rather, the WEPCO factors— nature and extent, purpose, frequency, and cost — must be analyzed and applied in reference to the source or industry category. For example, if a proposed project were estimated to cost $50 million dollars, that figure must be analyzed against what other projects within the industry have cost. If projects within the industry routinely cost $20 million, the $50 million cost of the proposed project may be one consideration in support of a finding that the project is not RMRR. This inquiry must be performed for each WEPCO factor. The EPA has moved for partial summary judgment on the Buck 4 project on the basis that the work performed was not RMRR for that particular unit. At this stage in the litigation, the court cannot conclude that there is no genuine issue of material fact as to whether the project at Buck 4 was non-RMRR. Insufficient evidence has been provided regarding whether the project undertaken at Buck 4, given the specific nature and extent, purpose, frequency, and cost of the'work, is routine in the electrical utility industry. The EPA applied the WEPCO factors to Buck 4 in arriving at its determination that the Buck 4 project does not qualify as RMRR. That analysis, however, was focused only on Buck 4. Consequently, a further analysis of the WEPCO factors must be performed in order to characterize the project at Buck 4. Once a case-specific analysis has been performed, it must be determined whether a project of that character is routine or not routine in the industry. Duke Energy moves for summary judgment as to all its units, contending that the projects at these units were RMRR. Duke Energy, however, has not presented sufficient evidence to persuade the court that no reasonable trier of fact could conclude otherwise. Duke Energy has presented the reports of its experts who conclude that Duke Energy’s projects involved RMRR. (Tuppeny Expert Report (Duke Energy Ex. 62); Bishop Expert Report (Duke Energy Ex. 64).) These reports provide no specific information, however, as to the number of units within the industry that have engaged in similar work, whether the costs of Duke Energy’s' projects were comparable to the costs of other similar industry projects, etc. The EPA’s expert did concede to Duke Energy that other utilities within the utility industry were performing work similar to that performed by Duke Energy. (Koppe Dep. at 63 (Duke Energy Ex. 61).) This alone is not sufficient to establish RMRR. If it were, the utility industry would be the ultimate authority on what is and is not permissible. Because there are factual issues which remain for trial, it is appropriate that the court provide the • parties with guidance as to who bears the burden of proving whether Duke Energy’s projects qualify as RMRR. As a general proposition, the party claiming the benefit of a statutory or regulatory exception bears the burden of proof. See United States v. First City Nat’l Bank of Houston, 386 U.S. 361, 366, 87 S.Ct. 1088, 18 L.Ed.2d 151 (1967); United States v. Moore, 613 F.2d 1029, 1044-45 (D.C.Cir.1979). This general proposition is just that; it is not dispositive. Rather, it is merely an interpretive aid to be used by the court in evaluating and allocating burdens of proof. See United States v. McArthur, 108 F.3d 1350, 1354 (11th Cir.1997) (characterizing rule as. “merely [an] interpretative aid[]” to be considered along with “other indications of legislative will evident in the statute”). One court has expressed its view that the issue should be approached by drawing a distinction between an exception to a statutory prohibition and an exclusion from a statutory definition. EEOC v. Chicago Club, 86 F.3d 1423, 1429-31 (7th Cir.1996) (stating that there is an “important distinction between an exception to the prohibition of a statute and an exclusion from the definition of entities covered by [the] statute”). The court believes this approach to be useful as it allows for the consideration of legislative intent. The requirements of PSD are triggered by any physical change that results in an increase in emissions. 40 C.F.R. § 51.166(b)(2)© (1987) (Duke Energy Ex. 21). The EPA has defined physical change expansively, such as the simple repair of a leaky pipe. WEPCO, 893 F.2d at 905; 57 Fed.Reg. 32,314, 32,316 (July 21, 1992) (stating that, the EPA has “always recognized” that the definition of physical change- could “encompass the most mundane activities at an industrial facility (even the repair or replacement of a leaky pipe)”). The EPA’s burden of proving a “physical change,” therefore, is quite easy to satisfy. The EPA has also recognized, however, “that Congress obviously did not intend to make every activity at a source subject to new source requirements” and therefore “defined ‘modification’ in the NSPS and NSR regulations to include common-sense exclusions from the ‘physical ... change’ component of the definition.” 57 Fed.Reg. at 32,316. For example, both NSPS and PSD regulations “contain similar exclusions for routine maintenance, repair, and replacement.” Id. Accordingly; the 1980 PSD regulations provide that “[a] physical change ... shall not include ... [r]outine maintenance, repair, and replacement.” 40 C.F.R. § 51.166(b)(2)(iii)(a) (1987) (Duke Energy Ex. 21). In light of the legislative intent not to include in the PSD requirements every activity and the EPA’s exclusion of RMRR from the definition of physical change, the EPA should bear the burden of proving that Duke Energy’s projects do not fall within this exemption. See Chicago Club, 86 F.3d at 1430-31 (explaining that where the “exclusion is incorporated in the definition” the defendant should not bear the burden of proving the exclusions application). To place the burden on Duke Energy would be in effect to sanction an almost de facto presumption of a PSD violation whenever a utility performs any type of work. This is especially true if the EPA applied the actual-to-potential emissions test, which always results in a net emissions increase, given the trivial showing necessary to prove a “physical change”. Congress, however, did not provide a presumption or anything approaching a presumption. The elements of a PSD violation, therefore, entail more than proving a physical change, which is no burden at all. It entails a showing that a utility engaged in a non-routine physical change that resulted in an increase in emissions. The EPA should appropriately bear the burden of making this showing at trial. IV. Net Emissions Increase Like the exemption for RMRR, the parties have presented two competing methods for quantifying emissions increases, both of which presumably stem from the 1980 PSD regulations. The EPA advances as the correct method the “actual-to-projected-actual” test. Under this test, a source must predict a project’s impact on hourly emissions rates and hours and rates of production, ie., capacity utilization. Duke Energy argues that the only method that can be applied to its projects is the “actual-to-actual” test. It contends that this is the test provided for under the 1980 regulations and that the test requires a comparison of pre-project actual emissions and future “actual” emissions, assuming constant hours and conditions of operation. The court finds, based on the PSD rules, the contemporaneous interpretations of the PSD rules, and the statutory language incorporating the NSPS concept of modification into PSD, post-project emissions must be calculated on an annual basis, measuring emissions in tons per year, and in calculating post-project emissions levels the hours and conditions of operation must be held constant. Accordingly, a net emissions increase can result only from an increase in the hourly rate of emissions. A. Plain language of increased hours exclusion The permitting and pollution control requirements of PSD are triggered by a non-routine physical change at a source that results in a “significant net emissions increase.” 40 C.F.R. § 51.166(b)(2)(i) (1987) (Duke Energy Ex. 21). In order to prove a “net emissions increase,” the EPA must show an “increase in actual emissions from a particular physical change or change in the method of operation at a stationary source.” Id. § 51.166(b)(3)(i)(a). For units that have begun normal operations, such as the units at issue in the case at bar, “actual emissions” is defined according to a pre-project (or baseline) period that is “representative of normal source operation.” Id. § 51.166(b)(21)(n). Thus, a comparison between the pre-pro-ject levels of emissions and post-project levels of emissions is required to determine whether there has been a net emissions increase above the baseline . levels. The key to this comparison is how to calculate the post-project emissions levels. Because an increase in emissions must result from a “physical [or operational] change,” which by definition excludes “[a]n increase in the hours of operation or in the production rate,” id. § 51.166(b)(2)(iii)(f), post-project emissions levels must be calculated assuming the same pre-project “representative” conditions of operation, i.e., hours and rates of production. Under the 1980 PSD regulations, therefore, only if the project increases the hourly rate of emissions will there be an annual emissions increase. The EPA asserts that the increased hours exclusion applies only to exclude increased utilization where the increased utilization is not associated with a construction project. Thus, whenever there is an increase in utilization coupled with a physical change, any increase in hours of operation and production rates may be considered in the emissions calculus. Such a limitation on the application of this exclusion, however, is not provided for in the plain text of the regulations. Hughes Aircraft Co. v. Jacobson, 525 U.S. 432, 438, 119 S.Ct. 755, 142 L.Ed.2d 881 (1999). The only limitation on the increased hours exemption provided for in the regulations is that any increase in hours or rates cannot otherwise be prohibited by a federally enforceable permit. 40 C.F.R. § 51.166(b)(2)(iii)(f) (1987) (Duke Energy Ex. 21). Nevertheless, the EPA contends that its interpretation of this exemption is reasonable and therefore entitled to deference. The court, however, cannot simply defer to the EPA’s interpretation when that interpretation imposes an additional condition on a regulatory exemption. See Christensen v. Harris County, 529 U.S. 576, 588, 120 S.Ct. 1655, 146 L.Ed.2d 621 (2000) (“To defer to the agency’s position would be to permit the agency, under the guise of interpreting a regulation, to create defacto a new regulation.”). B. EPA’s historic interpretation of increased hours exclusion The court cannot defer to the EPA’s interpretation when it is clearly contrary to earlier interpretations. Immediately after the promulgation of the PSD regulations in 1980, the EPA’s Director of the Division of Stationary Source Enforcement (“DSSE”), Edward E. Reich, confirmed in two separate applicability determinations that the requirements of PSD would be implicated only by an increase in the hourly rate of emissions. In a June 24, 1981, applicability determination, Reich wrote that “PSD applicability [at a previously operating source] is determined by evaluating any change in the [hourly] emissions rates caused by” the physical or operational change being examined. (Letter from Reich to Gill (June 24,1981) (Duke Energy Ex. 23).) Because the available data indicated that there would be no increase in the hourly rate of emissions following the contemplated change, Reich concluded that “[a]etual emissions could increase only if there [was] an increase in the production rate or hours of operation, both of which