Analysis: Ameren’s Petition for Reconsideration
In this installment and the following one, we will examine the United States v. Ameren Missouri litigation over whether Ameren’s installation of replacement preheaters at its Rush Island power station Units 1 and 2 constituted a “major modification” requiring PSD review. It is unusual for PSD cases to go to trial or appeal. The petition for rehearing of the panel decision thus provides an opportunity to evaluate both Ameren’s and the government’s position on several PSD issues. We will start with Ameren’s issues. The next installment will evaluate EPA’s issues.
Ameren presented two principal arguments in its request for rehearing: (1) whether the panel misapplied the Missouri state implementation plan (SIP) by applying the federal modification test in lieu of the state modification test found at 10 CSR 10-6.060(1); and (2) whether the panel erred by shifting the burden of proof and adopting a new legal standard for causation for the “demand growth” exclusion.
Controlling Definition of “Modification”
The Eighth Circuit panel decision. The Eighth Circuit panel upheld the decision of the Eastern District of Missouri, which had held that the specific incorporation by reference of 40 CFR 52.21(b) et seq. at 10 CSR 10-6.060(8) controlled, for purposes of the PSD program, over other definitions found in the SIP at 10 CSR 10-6.020. The panel holding was based on the following reasons: (1) the Missouri SIP had a separate definition of “Title I modification” which includes “any modification that requires a permit under 10 CSR 10-6.060 section (7) or (8),” (2) a prior decision by the U.S. Supreme Court in RadLAX Gateway Hotel v. Amalgamated Bank, 566 U.S. 639 (2012), which holds that a general provision does not control over a specific one, (3) applying this rule means that the federal rule provisions for “major modification” apply here, rather than the SIP language, (4) express language in EPA’s approval notice that the federal language approved “supersedes any conflicting provisions in the Missouri rule,” and (5) that application of the state definition would render a portion of the PSD rules superfluous.
Ameren’s argument. Ameren argued that the SIP “modification” provisions codified at 10 CSR 10-6.060(1) should control. These provide:
No owner or operator shall commence construction or modification of any installation subject to this rule [or] begin operation after that construction or modification…without first obtaining a permit from the permitting authority under this rule.
Id. §10-6.060(1)(C) (emphasis added). Construction is defined as “Fabricating, erecting, reconstructing, or installing a source operation….” 10 CSR 10-6.020(C)(62). Modification is defined as “any physical change to, or change in method of operation of, a source operation or attendance air pollution control equipment which would cause an increase in potential emissions of any air pollutant emitted by the source operation.” 10 CSR 10-6.020(M)(49)(C).
The crux of Ameren’s argument, although not explicitly stated, was that this definition of “modification” was approved by EPA, that a “major modification” is a subset of a “modification,” and that “it is undisputed that [the] projects did not trigger permitting requirements under the “Applicability” provision of [10-6.060(1)(C)]” because the projects were not “construction (citing the Panel opinion) nor were they “’modification’ because the projects would not ‘cause an increase in potential emissions….’” Ameren brief, at 7. Ameren then attacked the panel’s decision that a different test applied, arguing that the panel’s reliance on “incorporation-by-reference language” at 10 CSR 10-6.060(8) was incorrect and that an EPA remark that the federal regulations “superseded any conflicting provisions of in the Missouri rule” did not apply. In making its argument that the supersession language did not apply, Ameren relied upon a prior EPA comment, during the adoption of the SIP, that Missouri add clarifying language that the federal PSD language controlled over inconsistent language in the modification rule. Missouri declined to do so and EPA approved the rule anyway, finding that Missouri had satisfactorily responded to its comments. Ameren argues that given this history, the “Applicability” provision of 10 CSR 10-6.060(1) should prevail over the incorporation of the federal PSD program in 10 CSR 10-6-060(8), implicitly because Missouri “chose” a different applicability test. Additionally, Ameren suggested that the panel inappropriately relied upon the supersession statement because. in Ameren’s view, that statement was limited to certain public participation requirements and SIP appendices.
Finally, Ameren invoked cooperative federalism and case law providing that EPA may not override a state’s choices in establishing a SIP.
NSR Law Blog commentary
Based on the overall record, it is not clear that the panel decision was in error. The regulatory picture is perhaps a bit more complex than Ameren suggests. For example, Ameren did not note in its brief that there is also a definition of “major modification” at 10 CSR 10-6.020(M)(4) which includes “any physical change or change in the method of operation at an installation or in the attendance air pollution control equipment that would result in a significant net emissions increase of any pollutant….” A “net emissions increase” “is defined in 40 CFR 52.21(b)(3), promulgated as of July 1, 2003, [and no] subsequent amendments or additions.” 10 CSR 10-6.020(N)(6). Based on these additional definitions, even if the “modification” approach was accepted, the determination of whether a “major modification” occurred would still follow the federal program through the definition of net emissions increase. If so, then Ameren likely would still have shown an emissions increase triggering PSD. Of course, if Ameren was correct that “major modification” is a subset of “modification,” that may not matter. But the trend in PSD cases is to find “major modification” sui generis and not limited by the general term “modification” as district court, panel and denial of the reconsideration petition show. In reaching this decision, the courts are doubtless influenced by the U.S. Supreme Court’s decision in Duke Energy, which held that PSD modification and NSPS modification, which both derive from the same statutory definition, do not need to be interpreted identically.
Ameren’s principal argument, that the PSD program did not displace the state SIP, would require the panel to accept that EPA had erred in approving the Missouri DNR PSD program by accepting a definitional structure that was clearly not acceptable under 40 CFR 51.166 (and its predecessors), which sets forth required definitions in approved state programs. It is unlikely that EPA would knowingly have done so. In this context, it was not unreasonable for the panel to accept that EPA’s note that the PSD adoption “superseded” inconsistent provisions in the Missouri SIP was not limited to just the few items noted in Ameren’s discussion but instead applied more broadly, at least in the PSD context.
Shifting of the Burden of Proof in Demand Growth
The Eighth Circuit panel decision. The panel decision rejected Ameren’s arguments that the district court had erred on the application of the “major modification” test and its exclusion for “demand growth” by shifting the burden to Ameren. The panel quoted the D.C. Circuit’s New York I decision gloss on EPA’s discussion of the exclusion in the Federal Register stating that a source claiming demand growth must:
“establish[] two criteria … before excluding emissions from its projection: ‘(1) [t]he unit could have achieved the necessary level of utilization during the consecutive 24-month period [the source] selected to establish the baseline actual emissions; and (2) the increase is not related to the physical or operational change(s) made to the unit.’”
The panel accepted the district court’s interpretation that “it is the source’s burden to prove the applicability of the demand-growth exclusion” citing Supreme Court precedent. With respect to the causation argument, the panel rejected it, noting that the regulation requires both the “capable of accommodating prong” be demonstrated at the unit and that the “not related to changes at the unit” be related to the unit. Apparently, because the district court applied the test “at the unit” consistent with the language, the panel declined to evaluate Ameren’s argument any further.
Ameren’s argument. Ameren began by reiterating prior EPA guidance, published in the Federal Register, that “’system-wide demand growth’ is an ‘independent factor’ and that emissions increases ‘due to the rate of electricity demand growth for the utility system (of which that source is a member) as a whole’ are ‘unrelated to the physical or operational change,’ ‘do not result from the change and shall be excluded.’” Ameren Brief at 14 (citing 57 Fed. Reg. 32314, 32326 (July 21, 1992)). Ameren attacked the panel’s disregard for the language “’shall’ be excluded” in the regulation. Ameren then argued that under the quoted language above, the demand growth from system-wide growth is never attributed to the change/modification at the unit. Ameren bolstered its argument by citation to the 1992 rule and advanced an argument why the decision in New York I, which addressed the expansion of demand growth to manufacturing units, was not apposite.
NSR Law Blog commentary. At one level, Ameren is likely right and the panel and district court (and EPA, for that matter) likely wrong: demand growth is a market condition that exists independent of an individual electrical generating unit or manufacturing unit. EPA and the DC Circuit in New York I agreed that PSD requires some “causation” between the change at the unit and the emissions increase. Otherwise, a mere increase in the market, and unit response to that increase, would trigger PSD in violation of the express text of the regulation. To the extent that the courts and EPA have tried to narrow the understanding of the market, it makes no sense, contradicts express language in the regulation and will, it is hoped, become a “derelict upon the waters of the law” as Justice Felix Frankfurter once stated.
On a second level, however, Ameren may not have apprehended the actual argument. The government’s position at trial was that “but for” the changes made to the preheaters, the units could not have achieved their prior “capable of accommodating” production rate because they were so fouled that they could not achieve and consistently maintain that historic rate prior to the project. If this background is true, and the district court apparently held that it was, then the panel decision makes sense—the demand growth exclusion did not apply because the units did not have that capability at the time of the project until the project gave it to them. As a result, the emissions increase resulting from the restoration of the lost capacity was “caused” by the project.
As seen in Ameren, the demand growth exclusion has proven tricky to apply. The “capable of accommodating” prong must mean ability to produce a product or commodity over some period shorter than 24-months or it merely replicates the baseline actual emissions. Units can only run at “peak” when there is available demand for the product, so EPA and the courts should not require unreasonably long demonstration periods or it undercuts the express purpose of the provision. The “unrelated to the project” prong is clearest when the unit is not being physically modified and murkiest when it is physically modified. But physical modification to the unit should not disqualify all increase from the unit; instead, only the increase related to the change is disqualified. For example, if prior to the modification the unit could only produce product A, but after the modification it could produce either product A or product B, then any increase in demand related to product B would be “related” to the project, but changes in demand for product A would not be “related” to the project and still eligible for demand growth.
The Ameren case demonstrates the importance of a well thought through, carefully documented application or, if no permit is required, a carefully documented NSR applicability analysis conducted prior to commencing the project.
In the next issue we will look at the government’s petition to reconsider the Ameren Panel decision.