Looking to Upwind States to Reduce Interstate Ozone Pollution

Patricia Ross McCubbin

Professor McCubbin is an assistant professor at the Southern Illinois University (SIU) School of Law in Carbondale, Illinois, where she teaches several environmental law courses. Prior to joining the faculty at SIU, she was an attorney with the Environmental Defense Section of the U.S. Department of Justice (DOJ), where she served as co-counsel for the U.S. Environmental Protection Agency (EPA) in Michigan v. EPA, 213 F.3d 663, 30 ELR 20407 (D.C. Cir. 2000), cert. denied, 69 U.S.L.W. 3297 (U.S. Mar. 5, 2001) (No. 00-632). She wishes to thank Matthew Goetten for his helpful research and Kathleen Bassi, Andrea Bear Field, Jack Lipshultz, and Mel Schulze for their insightful comments. The views expressed herein are solely the author's and should not be attributed to those who offered comments on the Article, EPA, or the DOJ. Another version of this Article will appear in Patricia Ross McCubbin, Michigan v. EPA: Interstate Ozone Pollution and EPA's "NO_x SIP Call," 20 ST. LOUIS U. PUB. L. REV. 45 (forthcoming 2001).

[31 ELR 11045]

In 1998, faced with mounting evidence that many eastern and midwestern states could not comply with national air quality standards for ozone (O₃) pollution due, in part, to emissions originating in upwind states,¹ the U.S. Environmental Protection Agency (EPA) issued a highly controversial rule, the "NO_x SIP [State Implementation Plan] Call," requiring 23 states to reduce their emissions.² For the first time, EPA had initiated an attempt under the federal Clean Air Act (CAA)³ to reduce long-range interstate O₃ pollution.

Unprecedented in nature and affecting nearly half the nation, the NO_x SIP Call was challenged by 8 of the 23 upwind states along with dozens of industries.⁴ They made two key arguments: first, that EPA had improperly identified which upwind states' emissions were responsible for interstate O₃ pollution; and, second, that EPA had violated states' rights under the cooperative federalism mandated by the CAA.⁵ On March 3, 2000, however, in Michigan v. U.S. Environmental Protection Agency,⁶ a divided panel of the U.S. Court of Appeals for the District of Columbia Circuit upheld the rule in all major respects. To explain the importance of this complex case, which affirmed EPA's authority to regulate interstate air pollution, this Article provides an overview of the NO_x SIP Call and discusses the two most important issues addressed by the Michigan court.⁷

[31 ELR 11046]

Overview of the NO_x SIP Call

Ground-level O₃ pollution, commonly referred to as "smog," forms when nitrogen oxides (NO_x) and volatile organic compounds (VOCs) interact in the earth's atmosphere.⁸ NO_x and VOCs are emitted by thousands of diverse sources across the country. Coal-burning power plants are emissions sources, but so too are such smaller sources as gasoline distributors, diesel engines, and automobiles.⁹

Because the transformation of NO_x and VOCs into O₃ occurs in the atmosphere, where prevailing westerly winds can carry the pollutants over great distances,¹⁰ several northeastern—and therefore downwind—states repeatedly had urged EPA to require upwind states to impose greater controls on NO_x and VOC emissions.¹¹ Although O₃ is one of six air pollutants for which EPA has promulgated national ambient air quality standards,¹² EPA had been reluctant to intercede in the states' disputes over interstate O₃ pollution.¹³ However, with broader regulatory power granted by Congress in 1990,¹⁴ and with scientific evidence indicating that many states in the eastern half of the country could not comply with the air quality standards for O₃,¹⁵ EPA announced a plan in 1997, to determine whether any upwind states should be required, through a "SIP Call," to reduce their emissions.¹⁶

Under the CAA, in order to comply with air quality standards set by EPA, each state must develop a SIP, identifyingwhich emissions sources the state will regulate and by what degree.¹⁷ EPA must review and approve a state's SIP and, after doing so, may "call" the SIP to require its revision if it fails to comply with all the relevant provisions of the CAA.¹⁸ To regulate interstate O₃ pollution, EPA planned to call the SIPs of certain upwind states and require them to reduce their NO_x emissions. This "NO_x SIP Call" would be issued pursuant to § 110(a)(2)(D) of the CAA, which requires each state to ensure that emissions from its pollution sources do not contribute significantly to pollutant levels in another state.¹⁹

To issue the NO_x SIP Call pursuant to § 110(a)(2)(D), EPA faced enormous challenges. EPA had to determine not only which states' emissions were contributing to O₃ levels in distant downwind states (and to what degree), but also had to define which emissions would be considered "significant," for nowhere does the CAA define the phrase "contribute significantly" or provide any guidance for its interpretation.²⁰ In addition, EPA had to develop a regulatory mechanism that would require the states to reduce their emissions while, at the same time, leaving each state the flexibility to choose which pollution control measures to adopt in order to obtain the necessary reductions.²¹

To meet these challenges, EPA took two innovative steps. First, EPA developed a definition of "contribute significantly," based not only on the amount of ozone an upwind state contributed to downwind areas, but also on what types [31 ELR 11047] of pollution sources were found in the upwind state and whether affordable pollution controls were available for those sources.²² Second, to prescribe the emissions reductions required of the states, EPA assigned to each state an "emissions budget," that is, a limit on the total amount of NO_x emissions permitted from that state.²³ A discussion of these controversial steps, vital to understanding the Michigan case, follows.

Air Quality, Control Costs, and "Contributing Significantly"

The Good Neighbor Provision of § 110(a)(2)(D)

EPA issued the NO_x SIP Call pursuant to § 110(a)(2)(D) of the CAA,²⁴ the "good neighbor provision,"²⁵ which requires each state to include provisions in its SIP prohibiting:

any source or other type of emissions activity within the State from emitting any air pollutant in amounts which will … contribute significantly to nonattainment in, or interfere with maintenance by, any other State with respect to any … national primary or secondary ambient air quality standard.²⁶

Although this provision clearly prohibits a state from emitting pollutants in amounts that will "contribute significantly" to another state's "nonattainment," that is, its inability to comply with air quality standards set by EPA, the statute does not define, quantitatively or otherwise, what is meant by "contribute significantly," nor does it detail the methods to be used to identify which upwind emissions from one state might be contributing significantly to another state. To develop its interpretation of § 110(a)(2)(D) and determine which upwind NO_x emissions were "significant" contributors to downwind O₃, EPA first turned to sophisticated air quality modeling data.²⁷

EPA's Preliminary Determinations Based on Air Quality Modeling Data

For the NO_x SIP Call, EPA relied heavily on air quality modeling data provided by the Ozone Transport Assessment Group (OTAG). In 1995, in order to study interstate ozone pollution and develop strategies to reduce its transport from one state to another, 37 states east of the Rockies joined together to create OTAG.²⁸ For two years, using state-of-the-art techniques, OTAG studied the complex science of ozone transport and concluded that reducing upwind NO_x emissions regionwide would reduce O₃ levels in downwind areas, even in downwind areas several hundred miles away.²⁹ Although its studies provided considerable long-range ozone transport data,³⁰ OTAG did not determine which up-wind states were the most substantial contributors to down-wind ozone pollution.³¹ OTAG also did not determine which pollution control measures an upwind region should adopt in order to reduce its NO_x emissions,³² but instead recommended that the states conduct further studies.³³

In addition to relying on OTAG's studies, EPA also conducted its own more detailed modeling, which focused on state-by-state, rather than regional, analyses.³⁴ This modeling predicted the impact of upwind emissions on downwind O₃ levels using three measurements: (1) the absolute magnitude of the O₃ contribution from an upwind state to a downwind nonattainment area; (2) the relative magnitude of the [31 ELR 11048] upwind state's contribution compared to the downwind area's O₃ level; and (3) the frequency of the contributions.³⁵ EPA's review of the modeling data was informed by its notion of "collective contributions," a recognition that, generally, O₃ pollution is caused by the contributions of thousands of relatively small, diverse sources over wide areas.³⁶ Although no single source might contribute significantly to O₃ pollution in a downwind state, many sources together certainly could do so.³⁷ Similarly, while a single state's emissions, in absolute terms, might be contributing relatively small amounts of O₃ downwind, EPA determined that those emissions might nevertheless be contributing significantly to another state's O₃ nonattainment when combined with emissions from other states.³⁸ Thus, EPA's thresholds for finding significant upwind contributions were relatively low: a state was considered a significant contributor if it passed an initial screening test (requiring some very nominal contribution downwind)³⁹ and if, generally speaking, at least two of the three modeling measurements (absolute magnitude of contribution, relative magnitude, and frequency) indicated somewhat large and/or frequent contributions downwind.⁴⁰ Based on its review of all the air quality modeling data, EPA concluded preliminarily, pursuant to CAA § 110(a)(2)(D), that emissions from 23 upwind states in the Midwest, Southeast, and Northeast were contributing significantly to O₃ nonattainment in one or more downwind states.⁴¹

But EPA did not stop there. Instead, it turned to identifying the precise amounts of upwind emissions from those 23 states that were "significant."

EPA's Controversial Next Step: Consideration of Pollution Control Costs

In an effort to ensure that the economic burdens of implementing the NO_x SIP Call would be reasonable and equitably distributed among the states, EPA examined the costs of control measures available for reducing NO_x emissions.⁴² EPA defined as "highly cost-effective" any control measure that would eliminate one ton of NO_x emissions at a cost of no more than $ 2,000,⁴³ and determined that out of the dozens of types of sources that a state might regulate, highly cost-effective NO_x control measures were available for four: (1) large boilers and turbines that generate electricity at power plants (large power plants)⁴⁴; (2) large boilers and turbines at industrial facilities (large industrial boilers)⁴⁵; (3) cement kilns; and (4) stationary internal combustion engines (such as pipeline compressors).⁴⁶

[31 ELR 11049]

EPA then took an important—and controversial—next step. Rather than identifying significant upwind emissions based on air quality modeling data alone, EPA defined as "contributing significantly" only those upwind NO_x emissions that could be reduced using the highly cost-effective pollution control measures for the four types of sources it had identified.⁴⁷ More specifically, for those four sources, EPA calculated the precise emissions levels expected if they were operating with highly cost-effective control measures, and only the emissions above those expected levels were deemed to be contributing significantly to downwind O₃.⁴⁸ Sources producing emissions at or below those levels were not considered to be contributing significantly to downwind O₃.⁴⁹ For example, EPA had determined that a large power plant, using highly cost-effective control measures, could limit emissions to 0.15 pounds of NO_x per million British thermal units (Btus) of heat input (pounds/mmBtu).⁵⁰ If a plant emitted NO_x at a rate greater than 0.15 pounds/mmBtu, then only the emissions exceeding that level were considered to be contributing significantly to downwind O₃.⁵¹ Not considered to be significant were any NO_x emissions originating from a power plant emitting NO_x at a rate of 0.15 pounds/mmBtu or less, or, for that matter, any NO_x emissions originating from the many sources including automobiles, for which EPA had not specifically identified highly cost-effective control measures.⁵² NO_x emissions from automobiles, for example, were not deemed to be contributing significantly to downwind O₃ at all.

EPA's final determination of the significant upwind emissions could easily be seen as confusing two separate factors: first, based on scientific data, the actual amount of ozone an upwind state was contributing downwind; and second, the types of pollution sources in that state and whether highly cost-effective control measures were available for them. Arguably, emissions are no less significant if they originate from power plants (for which EPA had identified highly cost-effective control measures) than emissions originating from automobiles (for which EPA had not). Yet, under the NO_x SIP Call two states contributing the same amount of O₃ to downwind states, by EPA's definition, nevertheless could be contributing dramatically different "significant" amounts of O₃. For example, if Indiana's NO_x emissions primarily were produced by large power plants, while Kentucky's primarily were produced by automobiles, then, by EPA's definition, Indiana's emissions would be "contributing significantly" to a much greater extent than Kentucky's, even if both states contributed equally, as shown by air quality data, to downwind O₃ levels.⁵³

The petitioners in Michigan brought this apparent inconsistency to the D.C. Circuit's attention, suggesting that the NO_x SIP Call did not accomplish what should have been its primary goal, namely, reducing high levels of NO_x emissions that contributed to downwind O₃ pollution without regard to the types of sources creating those emissions.⁵⁴ They argued that EPA's definition of "contribute significantly" improperly focused, not on whether sources in a state, such as power plants, were actually contributing significant amounts of O₃ to downwind states, but rather on whether the power plants' NO_x emissions levels were as low as levels at power plants that had implemented highly cost-effective control measures.⁵⁵ Although the petitioners' argument had common-sense appeal, the Michigan court paid little heed to this practical concern, and instead focused on a more fundamental issue raised by the petitioners: whether CAA § 110(a)(2)(D) even allowed EPA to consider the costs of pollution control measures in the first place.

EPA's Statutory Authority to Consider Costs: The Petitioners' Arguments and the Court's Decision

Before the Michigan court, the petitioners argued that the NO_x SIP Call's definition of "contribute significantly" not only defied common sense, but also violated the very terms of § 110(a)(2)(D) of the CAA. That provision, they claimed, precluded EPA from incorporating the costs of pollution control measures into its identification of significant upwind emissions.⁵⁶ Instead, EPA was required to rely primarily on air quality modeling data.⁵⁷

The Michigan court, however, disagreed. Writing for the majority, Judge Stephen F. Williams, joined by Judge Judith W. Rogers, held that EPA, when determining whether an upwind state's emissions were contributing significantly to downwind ozone nonattainment, permissibly considered [31 ELR 11050] whether any highly cost-effective pollution control measures were available for that state's sources.⁵⁸ Because, according to the court, the meaning of the phrase "contribute significantly" in § 110(a)(2)(D) was ambiguous, it required further interpretation by EPA,⁵⁹ and under the "settled law of the circuit," EPA could consider economic factors, such as the costs of implementing pollution control measures, to interpret such an ambiguous statutory phrase unless Congress has expressed its "clear … intent to preclude [such] consideration."⁶⁰ The majority decided that nothing in the text of § 110(a)(2)(D), the overall structure of the CAA, or the Act's legislative history indicated that Congress intended to bar EPA from considering the costs that upwind sources might incur to implement pollution control measures.⁶¹ Although Congress had not expressly authorized EPA's consideration of control costs under § 110(a)(2)(D), it had not expressly precluded such consideration either. EPA, therefore, could exercise its discretion under the statute.⁶²

In dissent, Judge David B. Sentelle concluded that EPA did not have discretion, under § 110(a)(2)(D), to consider the costs of pollution control measures.⁶³ He described EPA as a "creature of statute" with authority to consider costs only if Congress expressly provides that authority.⁶⁴ If Congress remains silent, then an agency cannot, as the majority and EPA suggested, infer that authority in its decisionmaking.⁶⁵

Analysis of the Court's Decision

One argument offers support for Judge Sentelle's position. In other sections of the CAA, Congress expressly requires EPA to consider costs. In § 111(a)(1), for example, Congress directs EPA to set "new source performance standards" based, in part, on "the cost of achieving [an emissions] reduction."⁶⁶ One could reasonably conclude that had it intended EPA to consider pollution control costs under § 110(a)(2)(D). Congress would have similarly directed EPA to do so.⁶⁷

Such a literal reading of the CAA, however, might fail to give sufficient weight to the legislative history of the interstate pollution provisions in the statute.⁶⁸ Congress intended [31 ELR 11051] those provisions not only to improve overall air quality but also to alleviate economic disparities between upwind and downwind states.⁶⁹ It recognized that, historically, downwind states had imposed more stringent pollution control requirements on their emissions sources than had their upwind neighbors, thus imposing "a distinct economic and competitive disadvantage" for upwind sources.⁷⁰ Congress intended "to equalize the positions of the States with respect to interstate pollution by making a source at least as responsible for polluting another State as it would be for polluting its own State."⁷¹ The NO_x SIP Call, therefore, could be seen as satisfying Congress' intent by requiring upwind states to reduce their emissions to levels already obtained by sources using highly cost-effective control measures in downwind states.

An examination of the CAA's provision authorizing EPA to call SIPs for violations of § 110(a)(2)(D) also supports the Michigan decision. Under § 110(k)(5), EPA can require a state to "mitigate adequately" its contributions to "interstate pollutant transport," and when determining which state mitigation measures would be "adequate," the Agency can consider the costs that a state would incur to reduce its emissions.⁷² Although the petitioners maintained that EPA could not consider pollution control costs under § 110(a)(2)(D) when it identified which upwind emissions were significant, the petitioners (and the court) agreed that EPA could have considered those costs under § 110(k)(5) when it fashioned the remedy for violation of § 110(a)(2)(D) and determined the emissions reductions required of the states.⁷³ If so, then it is not difficult to conclude that costs also can be considered directly under § 110(a)(2)(D) itself.

Perhaps if EPA had made two distinct inquiries—first, under § 110(a)(2)(D), developing air quality data to identify which NO_x emissions were contributing significantly to downwind O₃ nonattainment, and second, under § 110(k)(5), considering whether highly cost-effective control measures were available to reduce those emissions—the end result would have been a NO_x SIP Call that more closely mirrored the structure of the CAA.⁷⁴ The end result for the upwind states, however, might have been the same: a requirement to reduce NO_x emissions to a level reflecting the use of highly cost-effective control measures.

Cooperative Federalism and the NO_x Emissions Budgets

Implementing the CAA: The Roles of States and EPA

The CAA establishes a scheme of "cooperative federalism" with distinct roles for EPA and the states.⁷⁵ To regulate O₃ pollution, for example, EPA sets air quality standards prescribing permissible levels of the pollutant, while each state, by adopting a SIP, has the authority to choose which sources [31 ELR 11052] to regulate and which pollution control measures to adopt in order to meet those levels.⁷⁶ One state might require emissions reductions from large industries or power plants, while another might require reductions from a broad array of smaller sources that might include dry cleaners, gas stations, or automobiles. Although it must review each state's SIP, EPA has no authority to question the state's choices as long as the SIP meets the requirements of the CAA.⁷⁷ This careful balance between federal and state authority is a key element of the statute,⁷⁸ and the eight states challenging the NO_x SIP Call argued that EPA upset this balance and violated the states' rights when it assigned specific "emissions budgets" for each state to meet.⁷⁹

EPA's NO_x Emissions Budgets

To determine the exact NO_x emissions reductions required of each state by the NO_x SIP Call, EPA took an unprecedented step, assigning a mandatory NO_x emissions budget to each of the 23 upwind states that were, according to EPA, contributing significantly to downwind O₃ nonattainment.⁸⁰ Relying on its earlier assessment of pollution control costs, EPA determined a state's emissions budget by calculating what the state's total NO_x emissions output would be if the state implemented the highly cost-effective pollution control measures available for the four types of sources that EPA previously had identified.⁸¹ To determine the NO_x emissions budget for Ohio, for example, EPA calculated that if Ohio were to require all of its large power plants, large industrial boilers, cement kilns, and stationary internal combustion engines to implement the highly cost-effective pollution control measures identified by EPA, then the state's overall NO_x emissions would be reduced from 372,626 tons to 239,898 tons.⁸² Consequently, EPA set the NO_x emissions budget for Ohio at 239,898 tons, requiring Ohio to submit a SIP demonstrating that the state would meet that cap.⁸³

It is important to understand that EPA did not dictate that the states actually impose the highly cost-effective control measures on the four pollution sources it had identified; instead, it simply set mandatory, statewide emissions caps based on the calculations. Ohio, for example, in order to reduce its NO_x emissions to 239,898 tons, was not expressly required toadopt the specific pollution controls for the four sources identified by EPA, but was free to obtain the necessary emissions reductions through whatever mix of control measures it imposed on whatever sources it chose.⁸⁴ The states challenging the NO_x SIP Call, however, argued that although EPA had not explicitly dictated which control measures each state should implement, the assumptions made by the Agency in calculating the emissions budgets effectively dictated the states' choices.⁸⁵ For this argument, the petitioners relied on an earlier decision of the D.C. Circuit in Virginia v. U.S. Environmental Protection Agency,⁸⁶ involving [31 ELR 11053] the Northeast Ozone Transport Commission Low Emission Vehicle (NOTC LEV) Rule.

Petitioners' Reliance on the Virginia Decision and the NOTC LEV Rule

Established by Congress in 1990, and composed of representatives from 13 northeastern states, the NOTC advises its member jurisdictions and EPA about options for addressing ozone pollution.⁸⁷ In 1994, a majority of the NOTC members voted to recommend that EPA require the NOTC states to enact an LEV program to help reduce intrastate and interstate ozone levels.⁸⁸ Adopting the NOTC's recommendation, EPA issued the "NOTC LEV Rule," which mandated that the NOTC states reduce NO_x and VOC emissions either by implementing the LEV program or by implementing alternative controls.⁸⁹ However, EPA's NOTC LEV Rule required a state implementing alternative controls to achieve NO_x and VOC reductions several times greater than the reductions expected if it implemented the LEV program.⁹⁰

Virginia, one of the NOTC states that had voted against the recommendation, challenged the NOTC LEV Rule, claiming that although it ostensibly allowed states the ability to choose alternatives, the rule effectively mandated only one control measure, the LEV program.⁹¹ The D.C. Circuit agreed, concluding that, because "only a very foolish state" would choose to implement alternative controls that would require much greater emissions reductions, EPA had not given the states any real alternatives to the LEV program.⁹² Consequently, the court invalidated the NOTC LEV Rule.⁹³

Because the NOTC LEV Rule represented EPA's only other effort to implement the good neighbor provision of § 110(a)(2)(D),⁹⁴ the states challenging the NO_x SIP Call in Michigan, mimicking the language in the Virginia opinion, argued that "only a very foolish state" would choose to implement pollution control measures other than those used by EPA to calculate the state's emissions budget.⁹⁵ No state, they claimed, would seek emissions reductions from any sources except those four types for which EPA had identified as highly cost-effective control measures, since choosing alternative pollution control measures, by definition, would be more costly.⁹⁶

Analysis of the Court's Decision

The Michigan court rejected the petitioners' argument, holding that the NO_x SIP Call did not expressly require, or even effectively mandate, that the states implement the highly cost-effective control measures for the four sources identified by EPA.⁹⁷ Distinguishing Virginia, the court explained that while the earlier decision prohibited EPA's adoption of a rule that left states with only "unreasonable or impracticable" alternatives, it did not prohibit a rule that gave states reasonable yet more costly alternatives.⁹⁸ The court found that, under the NO_x SIP Call, states not choosing to implement the highly cost-effective control measures identified by EPA still had reasonably cost-effective alternative controls available, including, for example, programs to reduce NO_x emissions from automobiles.⁹⁹ Moreover, states choosing alternative controls would not be penalized with additional emissions reduction requirements like those imposed by the NOTC LEV Rule.¹⁰⁰ Instead, the emissions budgets under the NO_x SIP Call set one—and only one—emissions reduction requirement for the state.¹⁰¹ Because "real choice exists for the covered states," the Michigan court concluded that EPA had not contravened the cooperative federalism of the CAA.¹⁰²

The Michigan court's conclusion is supported by a more detailed examination of EPA's method for calculating the NO_x emissions budgets, which reveals that EPA had not identified the four sources used in its budget calculations as the only sources for which highly cost-effective pollution control measures were available. Because it had averaged the costs of control measures for all 23 upwind states, EPA recognized that an individual state might find certain control [31 ELR 11054] measures costing less than the regionwide average.¹⁰³ For example, there are relatively few lime kilns throughout all 23 states, and, on average, it would not be cost-effective for most states to develop and implement regulations addressing so few sources.¹⁰⁴ However, for any state with several lime kilns, regulating them could very well be a highly cost-effective option.¹⁰⁵ Thus, choosing alternative control measures to meet the emissions budgets of the NO_x SIP Call would not be, by definition, more costly.

Conclusion

Regulating interstate O₃ pollution is a complicated task, requiring sophisticated technical analyses to determine the degree to which one state's NO_x emissions affect O₃ levels in other states. For the NO_x SIP Call, EPA not only reviewed complex scientific data, but also developed a regulatory mechanism that reflected the costs of pollution control measures that states could reasonably adopt. Although unprecedented and controversial, this approach was nevertheless consistent, according to the Michigan court, with the good neighbor provision of the CAA. The resulting emissions budgets, while appearing to restrict the states' choices, in fact, allowed each state sufficient flexibility to implement a variety of pollution control measures, thereby preserving the careful balance of authority mandated by the cooperative federalism of the CAA.

1. See Finding of Significant Contribution and Rulemaking for Certain States in the Ozone Transport Assessment Group Region for Purposes of Reducing Regional Transport of Ozone, 62 Fed. Reg. 60318, 60322 (Nov. 7, 1997) [hereinafter Proposed NO_x SIP Call] ("States generally were not able to meet [the deadlines for complying with national air quality standards for O]3[ because] … States were not able to address or control transport [of O₃ pollution]."). See also Jason S. Grumet, Old West Justice; Federalism and Clean Air Regulation 1970-1998, 11 TUL. ENVTL. L.J. 375, 398-99 (1998) (while scientists have known about the long-range transport of O₃ since at least the 1970s, information was not made available and understandable to the state regulatory community until the 1997 Ozone Transport Assessment Group (OTAG) study).

2. Finding of Significant Contribution and Rulemaking for Certain States in the Ozone Transport Assessment Group Region for Purposes of Reducing Regional Transport of Ozone, 63 Fed. Reg. 57356 (Oct. 27, 1998) [hereinafter Final NO]x[ SIP Call]. For an explanation of the terms "NO_x" and "SIP Call," see infra notes 8 & 17-18 and accompanying text. Technically, the NO_x SIP Call regulates 22 states and the District of Columbia, but for simplicity this Article refers to the regulated entities as "23 states." Those 23 jurisdictions are found in the Midwest, Southeast, and Northeast and are: Alabama, Connecticut, Delaware, District of Columbia, Georgia, Illinois, Indiana, Kentucky, Massachusetts, Maryland, Michigan, Missouri, North Carolina, New Jersey, New York, Ohio, Pennsylvania, Rhode Island, South Carolina, Tennessee, Virginia, West Virginia, and Wisconsin. Final NO_x SIP Call, supra, at 57358.

3. 42 U.S.C. §§ 7401-7671q, ELR STAT. CAA §§ 101-618.

4. Michigan v. EPA, 213 F.3d 663, 667-68, 30 ELR 20407, 20408 (D.C. Cir. 2000), cert. denied, 69 U.S.L.W. 3297 (U.S. Mar. 5, 2001) (No. 00-632). The state petitioners were Alabama, Indiana, Michigan, North Carolina, Ohio, South Carolina, Virginia, and West Virginia. The state of Wisconsin intervened as a petitioner. Id. at 668. Some of the petitioning industries were Appalachian Power Company, the United Mine Workers of America, and the Council of Industrial Boiler Owners. See Joint Common Issues Brief of Industry/Labor Petitioners, Michigan v. EPA, 213 F.3d 663 (D.C. Cir. 2000) (No. 98-1497); Brief of Petitioner Council of Industrial Boiler Owners, Michigan v. EPA, 213 F.3d 663 (D.C. Cir. 2000) (No. 98-1497).

Several northeastern states intervened on EPA's behalf, including Massachusetts, New York, and Pennsylvania, as did several industries from those states, as well as the province of Ontario and the federal government of Canada, and several environmental groups.

5. 213 F.3d at 674-79, 685-88, 30 ELR at 20410-12, 20415-16. These two issues are discussed in greater detail infra.

6. 213 F.3d 663, 669, 30 ELR 20407, 20408 (D.C. Cir. 2000).

7. In addition to the two primary issues discussed in this Article, the Michigan court agreed with EPA on several other issues, including: (1) EPA did not have to convene a transport commission under §§ 176A and 184 of the CAA, 42 U.S.C. §§ 7506a & 7511c, ELR STAT. CAA §§ 176A & 184, before issuing the NO_x SIP Call, 213 F.3d at 671-73, 30 ELR at 20409-10; (2) EPA's modeling was sufficient to identify the significant contributions from specific upwind states, 213 F.3d at 673-74, 30 ELR at 20409-10; (3) the rule did not conflict with earlier EPA decisions that certain upwind contributions did not violate the pre-1990 version of the key interstate pollution provision of the CAA (§ 110(a)(2)(E), 42 U.S.C. § 7410(a)(2)(E) (1988)), 213 F.3d at 674, 30 ELR at 20410; (4) EPA reasonably imposed uniform control requirements on all the states subject to the rule, rather than varying the controls based on geographic considerations, 213 F.3d at 679-80, 30 ELR at 20412; (5) the state of South Carolina was appropriately included in the rule, 213 F.3d at 685, 30 ELR at 20414-15; (6) EPA did not violate the Regulatory Flexibility Act, 5 U.S.C. §§ 601-612, by certifying that the rule would not have a significant economic impact on a substantial number of small entities, 213 F.3d at 688-89, 30 ELR at 20416; and (7) EPA reasonably defined the main sources that could participate in the rule's NO_x emissions trading program, 213 F.3d at 689-90, 30 ELR at 20416-17.

The court held against EPA when it found, for various reasons, that the states of Georgia, Missouri, and Wisconsin were improperly included in the NO_x SIP Call under the "one-hour" ozone standard. 213 F.3d at 681-85, 30 ELR at 20413-14. See infra note 12 (describing the "one-hour" standard and the "eight-hour" standard). See also infra note 36 (describing stay of EPA findings under the "eight-hour" standard).

The court also considered several other issues specific to industrial boilers, rejecting all but the claim that EPA failed to give proper notice of its definition of an "electricity-generating unit." 213 F.3d at 691-93, 30 ELR at 20417-18. The court agreed with a petitioner that EPA did not give proper notice of the control level it assumed for large stationary internal combustion engines. Id. at 693-94, 30 ELR at 20418. Finally, the court rejected two narrow claims brought by an electric utility regarding "early reduction credits" and "low mass emission units." Id. at 694-95, 30 ELR 20418-19.

8. Final NO_x SIP Call, supra note 2, at 57359. See generally ARNOLD W. REITZE JR., THE LAW OF AIR POLLUTION: COMPLIANCE AND ENFORCEMENT ch. 9 (Environmental Law Inst. 2001).

9. Grumet, supra note 1, at 378.

10. Proposed NO_x SIP Call, supra note 1, at 60333. See also Grumet, supra note 1, at 381 (scientific studies have concluded that O₃ may be caused by emissions hundreds of miles away).

11. Grumet, supra note 1, at 396 & nn.72-73 (describing New York's 1989 request that EPA take action against O₃ transport); id. at 387-88 (describing petitions to EPA by eight northeastern states asking for regulation of upwind NO_x emissions). See also Vickie L. Patton, The New Air Quality Standards, Regional Haze, and Interstate Air Pollution Transport, 28 ELR 10155, 10166 n.89 (Apr. 1998) (describing repeated efforts by Connecticut, New York, and other northeastern states to force EPA to impose additional requirements on midwestern sources to reduce emissions of another transported pollutant, sulfur dioxide (SO₂)).

12. National ambient air quality standards prescribe the maximum acceptable level of a pollutant in our ambient air. See 40 C.F.R. § 50.9-.10 (1979) (standards for O₃); id. at pt. 50 (standards for particulate matter, SO₂, carbon monoxide, nitrogen dioxide, and lead). In 1979, EPA promulgated the original "primary" standard for O₃, which set the permissible level of O₃ in the ambient air at 0.12 parts per million (ppm), averaged over one-hour intervals (sometimes referred to as the "one-hour standard"). 40 C.F.R. § 50.9. In 1997, EPA promulgated a new, more stringent primary O₃ standard as a companion to the older standard. National Ambient Air Quality Standards for Ozone, 62 Fed. Reg. 38856 (July 18, 1997) (codified at 40 C.F.R. § 50.10 (1997)). EPA reduced the maximum allowable level of O₃ in the ambient air to 0.08 ppm averaged over an eight-hour interval (the "eight-hour standard"). 62 Fed. Reg. at 38858, 38861-62.

In addition to "primary" standards designed to protect the public health, EPA also sets "secondary" standards designed to protect the public welfare. 42 U.S.C. § 7409(b)(1)(2), ELR STAT. CAA § 109(b)(1)(2). In 1997, EPA set the "secondary" O₃ standard at the same level as the primary standard. 62 Fed. Reg. at 38874-75.

13. See Patton, supra note 11, at 10156 ("Since the adoption of the 1970 [CAA], EPA has resisted restrictions on downwind pollution transport or allocating emissions reduction responsibilities among states to address transboundary pollution."); id. at 10166-72 (describing EPA's repeated denials of requests by downwind states for regulation of upwind sources); Grumet, supra note 1, at 385 ("By refusing to employ [its] authority [to control interstate air pollution], the EPA has until now perpetuated the paradigm of local responsibility set forth in the 1970 [CAA].").

14. In 1990, Congress strengthened the key interstate pollution provision now found in § 110(a)(2)(D) and, before 1990, found in § 110(a)(2)(E). 42 U.S.C. § 7410(a)(2)(E) (1988). The pre-1990 version only prohibited emissions from "any stationary source within the [upwind] State … which will … prevent attainment or maintenance [of the air quality standards] by any other State." Id. (emphasis added). In 1990, Congress revised the provision to focus not only on stationary sources, but also "other type[s] of emissions activities" within an upwind state (such as automobiles), and to prohibit emissions that "contribute significantly" to downwind nonattainment, even if those emissions were not the sole cause of (i.e., "prevented") nonattainment in another state. Id. § 7410(a)(2)(D)(i)(I), ELR STAT. CAA § 110(a)(2)(D)(i)(I). See Richard L. Revesz, Federalism and Interstate Environmental Externalities, 144 U. PA. L. REV. 2341, 2360-61 (1996) (summarizing changes made in 1990 to the good neighbor provision); Geoffrey L. Wilcox, New England and the Challenge of Interstate Ozone Pollution Under the Clean Air Act of 1990, 24 B.C. ENVTL. AFF. L. REV. 1, 18-21 (1996) (explaining difficulties in implementing the pre-1990 version of the good neighbor provision).

15. Grumet, supra note 1, at 389 (referring to "improved scientific knowledge among policy makers" and "changes in the 1990 [CAA] Amendments" as contributing to EPA's willingness to issue the NO_x SIP Call, as well as "changing politics" and "the growing competition between electric utilities").

16. Calls for State Implementation Plan Revisions for Certain States to Reduce Regional Transport of Ozone, 62 Fed. Reg. 1420 (Jan. 10, 1997).

17. 42 U.S.C. § 7410, ELR STAT. CAA § 110.

18. See id. § 7410(k)(5), ELR STAT. CAA § 110(k)(5). Prior to 1990. EPA's authority to require a state to revise its SIP was implied from § 110(a)(2)(H), 42 U.S.C. § 7410(a)(2)(H) (1977). See Virginia v. EPA, 108 F.3d 1397, 1407, 27 ELR 20718, 20723 (D.C. Cir.), modified on other grounds, 116 F.3d 499, 27 ELR 21380 (D.C. Cir. 1999).

19. 42 U.S.C. § 7410(a)(2)(D)(i)(I), ELR STAT. CAA § 110(a)(2)(D)(i)(I). See infra notes 24-26 and accompanying text.

20. See infra notes 24-55 and accompanying text. By referring to "states' emissions," I do not reference only those emissions from state-owned or state-operated sources; instead, the CAA refers to "any source or other type of emissions activity within the State." 42 U.S.C. § 7410(a)(2)(D)(i), ELR STAT. CAA § 110(a)(2)(D)(i).

21. See infra notes 80-86 and accompanying text.

22. See infra notes 24-55 and accompanying text.

23. See infra notes 80-86 and accompanying text.

24. 42 U.S.C. § 7410(a)(2)(D)(i)(I), ELR STAT. CAA § 110(a)(2)(D)(i)(I).

25. See Final NO_x SIP Call, supra note 2, at 57366.

26. 42 U.S.C. § 7410(a)(2)(D)(i)(I), ELR STAT. CAA § 110(a)(2)(D)(i)(I) (emphasis added). See supra note 12 (describing primary and secondary standards). For an explanation of how the current version of the good neighbor provision reflects improvements made by Congress in 1990, see supra note 14.

Section 110(a)(2)(D) focuses both on upwind state emissions that contribute significantly to downwind nonattainment and on emissions that interfere with a downwind state's ability to maintain its already-established compliance with the air quality standards. EPA, however, did not focus on this aspect of the good neighbor provision, and instead concentrated on emissions that "contribute significantly" to downwind nonattainment. Proposed NO_x SIP Call, supra note 1, at 60326.

27. Final NO_x SIP Call, supra note 2, at 57381-83 (describing the major findings from OTAG's technical analyses). See also Michigan v. EPA, 213 F.3d 663, 673, 30 ELR 20407, 20409 (D.C. Cir. 2000) (describing EPA's technical analyses).

28. OZONE TRANSPORT ASSESSMENT GROUP EXECUTIVE REPORT (1997) [hereinafter OTAG EXECUTIVE REPORT] (on file with author); Proposed NO_x SIP Call, supra note 1, at 60319-20. EPA, industry, and environmental representatives also participated in OTAG's proceedings. OTAG EXECUTIVE REPORT, supra, at 3.

29. OTAG EXECUTIVE REPORT, supra note 28, at 4, 30-32 (describing regional NO_x emissions reductions as effective in producing reduced O₃ levels throughout the OTAG states, though recognizing that NO_x reductions have the greatest benefit in the immediate area where the reductions occur). See also Proposed NO_x SIP Call, supra note 1, at 60377 (presenting OTAG's findings that "regional NO_x reductions are effective in producing ozone benefits" and "air quality data documents the widespread and pervasive nature of [O]3[] and indicates transport of [O]3[].") (emphasis added); Final NO_x SIP Call, supra note 2, at 57396 (describing, for example, Illinois' contribution to O₃ nonattainment in New York City).

As previously noted, NO_x and VOCs combine to form O₃. See supra note 8 and accompanying text. OTAG found, however, that reducing emissions of VOCs would only help reduce O₃ levels in the immediate locales where the VOCs are created, which usually are major urban areas, and would not help alleviate O₃ in distant downwind states. OTAG EXECUTIVE REPORT, supra note 28, at 30; Proposed NO_x SIP Call, supra note 1, at 60332, 60377 (summarizing OTAG's recommendations). This is so because many VOCs are emitted from natural sources such as trees and plants. Thus, even if all man-made VOCs were eliminated, NO_x emissions could still interact with natural VOCs to form O₃. See Grumet, supra note 1, at 378.

30. See Proposed NO_x SIP Call, supra note 1, at 60330 (the OTAG process used an innovative approach to develop "the most comprehensive analysis of ozone transport ever conducted"); Final NO_x SIP Call, supra note 2, at 57381-82 (describing OTAG's technical analyses, including its quantification of contributions). See also OTAG EXECUTIVE REPORT, supra note 28, at 1 ("OTAG improved the level of air pollution science and information by an order of magnitude….").

31. See Proposed NO_x SIP Call, supra note 1, at 60332-33 (OTAG divided the 23-state region into 12 "subregions" comprised of portions of multiple states and made conclusions about the contributions from one multistate region to another, not any one state to another state).

32. OTAG recommended to EPA a whole range of emissions controls—ranging from no controls beyond current requirements under the CAA to substantial controls on large sources such as power plants. OTAG EXECUTIVE REPORT, supra note 28, at 52-54.

33. Id. at 5.

34. Proposed NO_x SIP Call, supra note 1, at 60331-36 (describing OTAG's modeling, EPA's method for analyzing the results of OTAG's modeling, and other information EPA used to determine significant contributions); Final NO_x SIP Call, supra note 2, at 57377, 57387-89 (describing additional, state-specific air quality modeling EPA conducted after receiving comments).

35. Final NO_x SIP Call, supra note 2, at 57387. EPA analyzed an upwind state's contribution to exceedances of the original one-hour O₃ standard or the 1997 revised eight-hour O₃ standard. Id. at 57387-88. EPA's findings based on the eight-hour standard, and the petitioners' challenges to those findings, were stayed after that standard was remanded in American Trucking Ass'n v. EPA, 175 F.3d 1027, 29 ELR 21071 (D.C. Cir.), reh'g granted in part & denied in part, 195 F.3d 4, 30 ELR 20119 (D.C. Cir. 1999), cert. granted & rev'd sub nom. Whitman v. American Trucking Ass'n, 121 S. Ct. 903, 31 ELR 20512 (2001). See Michigan v. EPA, 213 F.3d 663, 670-71, 30 ELR 20407, 20408-09 (D.C. Cir. 2000) (describing stay of EPA's eight-hour findings and challenges thereto); Stay of the Eight-Hour Portion of the Findings of Significant Contribution and Rulemaking for Purposes of Reducing Interstate Ozone Transport, 65 Fed. Reg. 56245 (Sept. 18, 2000).

36. 65 Fed. Reg. at 57375-77. More particularly, EPA used a multifactored "weight-of-evidence" test that included consideration of, inter alia, the "overall nature of the [O]3[] problem" and the extent of a downwind state's nonattainment, and that was informed by the notion of collective contributions. See id. at 57376 (describing multifactored test); id. at 57392 (providing example of New York City, in which EPA considered the "nature of New York City's nonattainment problem and the nature of the contributions from upwind States" and in particular the collective contributions from numerous upwind states).

37. Id. at 57377 ("The solution to the problem [of ozone transport] is the implementation over a wide area of controls on many sources, each of which may have a small or unmeasurable ambient impact by itself.").

38. Id. at 57392. For example, EPA explained that several relatively small O₃ contributions from several different states led to a substantial portion of New York City's O₃ nonattainment. See id. at 57391-92. In particular, except for the states in the immediate vicinity of New York City, no upwind state contributed more than 5% of the total average O₃ contribution to that city. Yet together, those upwind states contributed a total of 27% to the city's O₃ nonattainment on average. Id. tbl. II-3.

39. To pass the initial screening test a state only had to contribute at least 2 parts per billion (ppb) of ozone downwind or contribute to at least 1% of the ozone exceedances. See Final NO_x SIP Call, supra note 2, at 57391 ("The first step in evaluating this information was to screen out linkages for which the contributions were very low…."); id. at 57393 (summarizing results of modeling, including "contributions of more than 2 ppb"); U.S. EPA, Air Quality Modeling Technical Support Document for the NO_x SIP Call, at 30 & n.11 (on file with the author).

40. Final NO_x SIP Call, supra note 2, at 57391.

41. Id. at 57394-98 (summarizing the results of EPA's assessment of air quality modeling data, described as the "first step" in EPA's efforts to identify the upwind emissions that contribute significantly downwind). See supra note 2 for a list of the 23 states.

42. Final NO_x SIP Call, supra note 2, at 57376-77.

43. Id. at 57377-78. Pollution control measures might include installing equipment to burn fuels more efficiently (for example, using "low-NO_x burners" or "overfire air") or installing post-combustion technologies to remove NO_x from a source's emissions (for example, using "selective catalytic reduction" or "selective noncatalytic reduction"). See id. at 57447.

Just as OTAG had found that VOC emissions reductions would not help alleviate long-range O₃ transport (see supra note 29), so EPA also did not consider any VOC emissions to be significant contributors to downwind states' O₃ nonattainment and did not consider the control costs of VOC reductions. Proposed NO_x SIP Call, supra note 1, at 60320. The Michigan court upheld EPA's decision not to give credit under the NO_x SIP Call for VOC reductions. 213 F.3d at 688, 30 ELR at 20416.

44. Technically, EPA refers to this category as "large electricity-generating units" (large EGUs), which could include traditional, utility-owned power plants or other non-utility-owned generators that sell power. See Supplemental Notice for the Finding of Significant Contribution and Rulemaking for Certain States in the Ozone Transport Assessment Group Region for Purposes of Reducing Regional Transport of Ozone, 63 Fed. Reg. 25902, 25923 (May 11, 1998) (explaining that the deregulation of the power-generation industry has led numerous non-utility sources to generate electricity) [hereinafter Supplemental Notice]. This Article will refer to the category as "large power plants."

45. Industrial boilers are one category of what EPA refers to as "non-EGUs," Final NO_x SIP Call, supra note 2, at 57365, and can generate steam for a variety of uses within an industrial complex. See Supplemental Notice, supra note 44, at 25923.

46. Final NO_x SIP Call, supra note 2, at 57377-78, 57399-402. EPA chose the figure of $ 2,000 because several recent federal and state initiatives to control NO_x emissions had cost, on average, a little less than $ 2,000/ton of NO_x removed. See id. at 57400 & tbl. I. EPA did not find any highly cost-effective pollution controls for common NO_x sources, municipal waste combustors, iron and steel mills, and lime kilns. Id. at 57403. EPA also did not find any such controls for mobile sources beyond the nationally applicable controls already in place or expected to be in place in the near future. Id. at 57402, 57420 (assuming that the National Low Emission Vehicle program and the 2004 Heavy-Duty Engine Standards would reduce state emissions). Id. at 57402-03.

47. Id. at 57377-78. See infra note 74 for a discussion of some practical difficulties EPA might face if it tried to rely on only air quality modeling data to identify the significant upwind emissions.

48. Final NO_x SIP Call, supra note 2, at 57378. EPA only reviewed the availability of emissions controls in the 23 states that it had identified preliminarily based on air quality data. See id.

Because EPA had also determined that the NO_x emissions from the affected upwind States have a large and/or frequent impact on downwind nonattainment or maintenance problems, EPA concludes that the amount of NO_x emissions from those States that can be eliminated through application of highly cost-effective control measures contributes significantly to nonattainment or maintenance problems downwind.

49. Id. ("Because no highly cost-effective controls are available to eliminate the remaining amounts of NO_x emissions, EPA concludes that those emissions do not contribute significantly to downwind nonattainment or maintenance problems.").

50. Id. at 57401-02. The heat input value of various fuels differs considerably; thus, emissions limits are expressed as a factor of heat inputs. In determining the cost of NO_x reductions for large power plants, EPA assumed that the 23 covered states would participate in an emissions trading program, whereby sources could buy and sell NO_x emissions reduction credits from state to state. Id. at 57400.

51. See id. at 57409-14 (calculating "emissions budgets" for each state assuming the "applicable [NO]x[] emission rate" of 0.15 pounds/mmBtu). For a discussion of the state emissions budgets, see infra notes 80-86 and accompanying text.

52. See id. at 57378, 57402-03.

53. This example—far simpler than the more complex facts addressed in the NO_x SIP Call—assumes that Kentucky's emissions, in addition to originating from automobiles, also come from at least a few large power plants, large industrial boilers, cement kilns, and/or stationary internal combustion engines—the four categories targeted by EPA; otherwise, Kentucky would not have to reduce any emissions under the NO_x SIP Call. This example also assumes that NO_x emissions primarily from automobiles can have the same impact on downwind ozone levels as NO_x emissions primarily from power plants. In fact, however, it may be that emissions from automobiles are transported shorter distances than emissions from the tall stacks of power plants and thus have a lesser impact on downwind ozone levels. Nevertheless, EPA did not decline to consider automobile emissions "significant" because they have less impact on downwind ozone levels, but only because there were no highly cost-effective controls available to address those emissions. Final NO_x SIP Call, supra note 2, at 57402.

54. See Final Brief of Petitioning States, at 17-19, Michigan v. EPA, 213 F.3d 663 (D.C. Cir. 2000) (No. 98-1497) [hereinafter States' Brief].

55. Id.

56. Id. at 15-16.

57. Id.

58. 213 F.3d at 674-79, 30 ELR at 20410-12.

59. Id. at 677-78, 30 ELR 20411-12.

60. Id. (citing Natural Resources Defense Council v. EPA, 824 F.2d 1146, 17 ELR 21032 (D.C. Cir. 1987) (en banc); George E. Warren Corp. v. EPA, 164 F.3d 676, 29 ELR 20252 (D.C. Cir. 1999); Grand Canyon Air Tour Coalition v. Federal Aviation Admin., 154 F.3d 455, 29 ELR 20075 (D.C. Cir. 1998); Natural Resources Defense Council v. EPA, 937 F.2d 641, 21 ELR 21231 (D.C. Cir. 1991)).

61. Michigan, 213 F.3d at 679, 30 ELR at 20412 (summarizing conclusion regarding text, structure, and history of the CAA).

62. Id. The court also rejected a related claim about EPA's findings of significant contributions, namely, that EPA had exercised such unconstrained discretion in determining which upwind emissions contributed significantly to downwind ozone nonattainment that the Agency had violated the nondelegation doctrine. See id. at 680-81, 30 ELR 20412-13. For this argument the petitioners relied on the D.C. Circuit's decision in American Trucking, in which the court held that, pursuant to the nondelegation doctrine, an agency must identify a "determinate criterion," derived from the governing statute, to restrain the agency's exercise of discretion. American Trucking, 175 F.3d at 1034, 29 ELR at 21071. See generally Patricia Ross McCubbin, The D.C. Circuit Gives New Life and New Meaning to the Nondelegation Doctrine in American Trucking Ass'n v. EPA, 19 VA. ENVTL. L.J. 57 (2000). In the challenges to the NO_x SIP Call, the Michigan court held that EPA's benchmark of $ 2,000 for highly cost-effective pollution controls (see supra notes 43-46 and accompanying text) was not based on "anything in the language or function of section 110(a)(2)(D)" of the CAA, and was "essentially unbounded," suggesting a violation of the nondelegation doctrine. 213 F.3d at 680, 30 ELR at 20412. However, the court went on to note that the nondelegation doctrine only applies to agency actions that affect "all American enterprise," and that, because the NO_x SIP Call applied to merely 23 states—just roughly "half of the nation"—the rule did not have to satisfy the doctrine. Id. at 680-81. This conclusion—difficult to justify—is now irrelevant because the U.S. Supreme Court reversed the D.C. Circuit's decision in American Trucking, holding that the nondelegation doctrine does not serve as a constraint on agency discretion. 121 S. Ct. at 912-14, 31 ELR at 20514-15.

63. 213 F.3d at 695, 30 ELR at 20419. Judge Sentelle believed that EPA had authority to consider only air quality-related information when identifying significant upwind emissions. Id.

64. Id. For this notion, Judge Sentelle relied on the Supreme Court's statement that "it is axiomatic that an administrative agency's power to promulgate legislative regulations is limited to the authority delegated by Congress." Bowen v. Georgetown Univ. Hosp., 488 U.S. 204, 208 (1988).

65. 213 F.3d at 695, 30 ELR at 20419. Judge Sentelle's dissent did not address the several prior D.C. Circuit cases upon which the majority relied for the notion that EPA may consider costs unless Congress clearly bars their consideration. See id. at 678, 30 ELR at 20412 (citing cases referenced supra note 60). And the majority opinion did not address the Supreme Court and D.C. Circuit cases upon which Judge Sentelle relied for the opposite notion. Id. at 696-97, 30 ELR 20419-20 (citing Bowen, 488 U.S. at 208); American Petroleum Inst. v. EPA, 52 F.3d 1113, 25 ELR 20824 (D.C. Cir. 1995); Ethyl Corp. v. EPA, 51 F.3d 1053, 25 ELR 20817 (D.C. Cir. 1995)). For an interesting discussion of the many courts that have rejected EPA's claims of authority to act in light of congressional silence, and the exception for "cost considerations" that the D.C. Circuit has created to that general approach, see Christopher H. Schroeder & Robert L. Glicksman, Chevron, State Farm, and EPA in the Courts of Appeals During the 1990s, 31 ELR 10371, 10382-84 (Apr. 2001).

66. 42 U.S.C. § 7411(a)(1), ELR STAT. CAA § 111(a)(1) (emphasis added).

67. See, e.g., Russello v. United States, 464 U.S. 16, 23 (1983) ("Where Congress includes particular language in one section of a statute but omits it in another section of the same Act, it is generally presumed that Congress acts intentionally and purposely in the disparate inclusion or exclusion.") (quoting United States v. Wong Kim Bo, 472 F.2d 720, 722 (5th Cir. 1972)).

This argument is buttressed by the recent decision of the Court in American Trucking. There, the Court noted the many different provisions of the CAA, including § 111, under which Congress expressly has directed EPA to consider costs, and explained that it will not "find implicit in ambiguous sections of the CAA an authorization to consider costs that has elsewhere, and so often, been expressly granted." 121 S. Ct. at 909, 31 ELR at 20513 (citations omitted). It held that § 109(b) of the Act, 42 U.S.C. § 7409(b), ELR STAT. CAA § 109(b), does not grant EPA express authority to consider costs when setting national ambient air quality standards. 121 S. Ct. at 911, 31 ELR at 20513-14. Interestingly, the Court distinguished § 110 and the Michigan decision (along with a few other D.C. Circuit opinions), indicating that § 110 does not "share[] § 109(b)(1)'s prominence in the overall statutory scheme." 121 S. Ct. at 910 n.1, 31 ELR at 20513 n.1. In other words, even though Congress did not expressly grant EPA authority to consider costs in § 110, the Court apparently was not troubled by the Michigan court's decision to allow EPA to do so, because § 110 supposedly is not a "prominent" part of the CAA. How § 110, which establishes the core framework of the SIP program by which the air quality standards are implemented, can be anything but a "prominent" part of the CAA is difficult to imagine.

68. Along with § 110(a)(2)(D), other provisions of the CAA also address interstate pollution. Under § 176A, for instance, EPA may designate "transport regions"—areas of the country that are affected by a common interstate air pollution problem—and may establish a "transport commission" comprised of state representatives to recommend control measures for the transport region. 42 U.S.C. § 7506a, ELR STAT. CAA § 176A. In § 184, Congress established one particular transport region: the Northeast Ozone Transport Region (NOTR) for 12 northeastern states and the District of Columbia. Id. § 7511c(a), ELR STAT. CAA § 184(a). See also infra note 87.

In addition, § 126 allows downwind states to petition EPA to impose limits directly on upwind stationary sources that violate § 110(a)(2)(D). 42 U.S.C. § 7426(b)-(c), ELR STAT. CAA § 126(b)-(c). While it was developing the NO_x SIP Call, EPA received petitions under § 126 from several northeastern states, requesting that EPA find that certain types of NO_x-emitting sources in specified upwind states were contributing significantly to nonattainment of the one-hour O₃ air quality standard in the downwind, petitioning states. Findings of Significant Contribution and Rulemaking on Section 126 Petitions for Purposes of Reducing Interstate Ozone Transport, 63 Fed. Reg. 24058, 24059 (Apr. 30, 1998). EPA determined that sources of NO_x located in 12 states (Delaware, Indiana, Kentucky, Maryland, Michigan, New Jersey, New York, North Carolina, Ohio, Pennsylvania, Virginia, and West Virginia) and the District of Columbia contribute significantly to nonattainment of the one-hour O₃ standard in the downwind states of Connecticut, Massachusetts, New York, and Pennsylvania, and, by a later rule, EPA imposed federal NO_x reduction requirements on those sources (the § 126 Rule). See Findings of Significant Contribution and Rulemaking on Section 126 Petitions for Purposes of Reducing Interstate Ozone Transport, 64 Fed. Reg. 28250 (May 25, 1999); Findings of Significant Contribution and Rulemaking on Section 126 Petitions for Purposes of Reducing Interstate Ozone Transport, 65 Fed. Reg. 2674 (Jan. 18, 2000). In subsequent litigation, the D.C. Circuit largely upheld the § 126 Rule, but remanded two issues to EPA for further explanation: (1) EPA's projected growth rates for electric power generation in the upwind states and (2) EPA's decision to categorize cogeneration units as EGUs. Appalachian Power Co. v. EPA, 249 F.3d 1032, 1053-55, 1061-63 (D.C. Cir. 2001).

69. See S. REP. No. 95-127, at 41 (1977), available at LEXIS, Clean Air Act Amendments of 1977 Legislative History (CAA77 Leg. Hist.), at 25 (strong interstate pollution provisions were needed to address the "serious inequities among several States, where one State may have more stringent implement plan requirements than another State").

70. Id. at 41-42 (giving example of Ohio's failure to adopt any control measures for SO₂ until 1976, resulting in increased SO₂ emissions transported to its downwind neighbor, West Virginia, which then had to impose "more stringent control[s]").

71. Id. at 42.

72. 42 U.S.C. § 7410(k)(5), ELR STAT. CAA § 110(k)(5). Section 110(k)(5) does not expressly refer to § 110(a)(2)(D). Instead, it refers to "interstate pollutant transport described in [§ 176A] of this title or [§ 184] of this title." Sections 176A and 184, in turn, establish "transport commissions" that may make recommendations to EPA to bring states into compliance with the prohibition on contributions to interstate pollution established by § 110(a)(2)(D). See id. § 7506a(b)(2), ELR STAT. CAA § 176A(b)(2) (interstate transport commission shall recommend to EPA "such measures as the Commission determines to be necessary to ensure that the plans for the relevant States meet the requirements of section 7410(a)(2)(D) of this title"); id. § 7511c(c)(5), ELR STAT. CAA § 184(c)(5) (EPA may make a finding in response to a recommendation from the National Ozone Transport Commission (NOTC) that a state's SIP "is inadequate to meet the requirements of section 7410(a)(2)(D) of this title"). See also infra notes 87-88 and accompanying text.

73. Section 110(k)(5) does not expressly grant EPA authority to consider costs in devising the appropriate mitigation measures, but such authority can easily be inferred from the provision's remedial nature. See Michigan, 213 F.3d at 676-77, 30 ELR at 20411 (citing 42 U.S.C. § 7410(k)(5), ELR STAT. CAA § 110(k)(5), and explaining petitioners' concession that control costs could be considered under that provision).

In fact, EPA had asked for comment on whether it should use this two-step approach but chose for the final rule to collapse the inquiry into one step. See Proposed NO_x SIP Call, supra note 1, at 60325; Final NO_x SIP Call, supra note 2, at 57376.

74. Although consistent with the statute, it is unclear whether as a practical matter EPA could have relied only on air quality data to identify significant upwind emissions. The upwind states and industries certainly would argue that, if they had been given an adequate opportunity to review and comment on the data from EPA's state-specific air quality modeling, which was completed well after the close of the public comment period and released to the public only shortly before the final rule was issued, they likely would have found substantive flaws in the data, raising serious doubts about EPA's ability to use it to identify specific amounts of significant emissions from any one upwind state. The challengers asked the Michigan court to find that EPA's late release of the data was improper, but the court rejected that contention. Michigan, 213 F.3d at 674, 30 ELR at 20410.

And from EPA's perspective, trying to rely on air quality data alone to identify significant upwind emissions could pose several practical problems. For instance, EPA's modeling found that many of the upwind states contributed significantly to O₃ nonattainment in numerous downwind states. See, e.g., Final NO_x SIP Call, supra note 2, at 57394 tbl. II-4 (Indiana contributes to 12 downwind states). How would EPA express the "significant" amount of emissions that must be reduced from, say, Indiana, given that the state was found to be contributing O₃ to a dozen downwind states at levels that varied state by state? The task would not be impossible perhaps, but it certainly would complicate EPA's rulemaking efforts and provide additional fodder for subsequent litigation.

75. See Connecticut v. EPA, 696 F.2d 147, 13 ELR 20135 (2d Cir. 1982).

76. See Train v. Natural Resources Defense Council, 421 U.S. 60, 5 ELR 20264 (1975). See also Bethlehem Steel Corp. v. Gorsuch, 742 F.2d 1028, 1036, 14 ELR 20740, 20744 (7th Cir. 1984) ("The federal government through the EPA determines the ends—the standards of air quality—but Congress has given the states the initiative and broad responsibility regarding the means to achieve those ends through [SIPs] and timetables for compliance.") (citations omitted). In the 1990 Amendments to the CAA, however, Congress restricted state flexibility somewhat by imposing numerous, specific requirements applicable to O₃ nonattainment areas. See 42 U.S.C. §§ 7501-7511f, ELR STAT. CAA §§ 171-185B.

77. Train, 421 U.S. at 79, 5 ELR at 20267. See also Union Elec. Co. v. EPA, 427 U.S. 246, 256-57, 6 ELR 20570, 20573 (1976) (The CAA "place[s] the primary responsibility for formulating pollution control strategies on the States, but nonetheless subject[s] the States to strict minimum compliance requirements.").

78. See Mark Squillace, Cooperative Federalism Under the Surface Mining Control and Reclamation Act: Is This Any Way to Run a Government?, 15 ELR 10039, 10039 (Feb. 1985) (cooperation federalism is a hallmark of the CAA and other environmental statutes); Hope M. Babcock, Dual Regulation, Collaborative Management, or Layered Federalism: Can Cooperative Federalism Models From Other Laws Save Our Public Lands?, 3 HASTINGS W.-NW. J. ENVTL. L. & POL'Y 193, 199 (1996) (describing the "state primacy" or "dual regulation" model of cooperative federalism found in various federal pollution control statutes).

79. See States' Brief, supra note 54, at 34-40.

80. Final NO_x SIP Call, supra note 2, at 57377-78.

81. Id. at 57403. EPA calculated the states' emissions as of the year 2007, chosen in part because the CAA required many of the most polluted areas of the country to meet the O₃ standard by that year and because OTAG had used that date. See 42 U.S.C. § 7511(a)(2), ELR STAT. CAA § 181(a)(2): Proposed NO_x SIP Call, supra note 1, at 60329. To do so, EPA first calculated a "2007 base case" for which it: (1) developed "emissions inventories" describing all the various sources of emissions in each state (including mobile sources); (2) projected the increased emissions that would result from a growing economy in each state; and (3) determined which emissions controls, independent of the NO_x SIP Call, would be required under the CAA (e.g., the National Low Emission Vehicle Program). Final NO_x SIP Call, supra note 2, at 57378, 57420. See also infra note 82 (describing challenges toEPA's "growth factors"). From this base case, EPA then subtracted the emissions reductions attributable to the four targeted sources. Final NO_x SIP Call, supra note 2, at 57378.

82. Final NO_x SIP Call, supra note 2, at 57439 (showing for Ohio a 2007 base case and budget, respectively, of 372,626 tons and 239,898 tons and showing similar calculations for the 22 other covered states); id. at 57433-36, tbl. III-5 & tbl. III-7 (showing calculated emissions levels in each state if EGUs (i.e., large power plants) and non-EGUs (i.e., industrial boilers, cement kilns and stationary internal combustion engines) reduced their emissions to EPA's assumed control levels).

The budgets are expressed in "tons per ozone season," which runs from May 1 to September 30 each year. See id. at 57492 (to be codified at 40 C.F.R. § 51.121(e)(1)); id. at 57455 (defining "ozone season" as May 1 to September 30).

EPA revised the states' emissions budgets (twice) after receiving comments about some of the technical data EPA had gathered in the emissions inventories. See Technical Amendment to the Finding of Significant Contribution and Rulemaking for Certain States for Purposes of Reducing Regional Transport of Ozone, 64 Fed. Reg. 26298 (May 14, 1999); Technical Amendment to the Finding of Significant Contribution and Rulemaking for Certain States for Purposes of Reducing Regional Transport of Ozone, 65 Fed. Reg. 11222 (Mar. 2, 2000). Based on EPA's changes to the emissions inventories, Ohio's revised budget was set at 249,274 tons of NO_x, and the budgets for the 22 other states were likewise revised. See 65 Fed. Reg. at 11225.

In a subsequent challenge to the technical amendments, the D.C. Circuit found that EPA had not adequately explained the growth factors that it had used to project emissions in 2007 from large power plants, and remanded that issue to EPA for further explanation. See Appalachian Power Co. v. EPA, 251 F.3d 1026 (D.C. Cir. 2001). (This relief mirrored the holding in the § 126 Rule litigation. See supra note 68.)

The Appalachian Power court also vacated and remanded EPA's entire emissions budget for the state of Missouri, even though in Michigan the court had only remanded the Missouri budget based on the one-hour standard and had stayed any consideration of budgets based on the eight-hour standard. 251 F.3d at 1040-41; Michigan, 213 F.3d at 670-71, 30 ELR at 20408. The Appalachian Power court concluded that EPA had no basis under either standard to subject Missouri to the NO_x SIP Call. 251 F.3d at 1040-41.

83. Final NO_x SIP Call, supra note 2, at 57491-92 (to be codified at 40 C.F.R. § 51.121(a)(1) & (e)(2)).

84. Id. at 57378 (quoting Proposed NO_x SIP Call, supra note 1, at 60328).

85. States' Brief, supra note 54, at 34-40.

86. 108 F.3d 1397, 1407, 27 ELR 20718, 20723 (D.C. Cir.), modified on other grounds, 116 F.3d 499, 27 ELR 21380 (D.C. Cir. 1999).

87. 42 U.S.C. § 7511c(a), ELR STAT. CAA § 184(a). Those 13 jurisdictions, running from Maine to the District of Columbia, make up the NOTR. Id. While Congress did not specifically include all of Virginia within the NOTR, Virginia participates because the northern portion of that state lies within the "Consolidated Metropolitan Statistical Area" that includes the District of Columbia. See Virginia, 108 F.3d at 1401 n.2, 27 ELR at 20719 n.2.

88. See Ozone Transport Commission; Recommendation that EPA Adopt Low Emission Vehicle Program for the Northeast Ozone Transport Region, 59 Fed. Reg. 12914 (Mar. 18, 1994). LEVs emit very few NO_x and VOC emissions. Virginia, 108 F.3d at 1401-02, 27 ELR at 20719. As the Virginia case explained, the NOTC LEV program was modeled on the very stringent NO_x and VOC emissions limits of California's LEV program. Id. at 1401, 27 ELR at 20719.

89. Final Rule on Ozone Transport Commission; Low Emission Vehicle Program for the Northeast Ozone Transport Region, 60 Fed. Reg. 4712, 4712-13 (Jan. 24, 1995) [hereinafter NOTC LEV Rule].

90. Virginia, 108 F.3d at 1404-05, 27 ELR at 20721 (explaining that if Virginia chose not to adopt the LEV program and use alternative measures, it would have to reduce NO_x emissions 3.5 times, and VOC emissions 6.5 times, more than under the LEV program).

91. Id. See also id. at 1403-10, 27 ELR at 20720-24 (analyzing whether EPA exceeded its authority under § 110).

92. Id. at 1404-05, 27 ELR at 20721.

93. Id. at 1406-10, 27 ELR at 20722-24 (applying Train).

94. The NOTC LEV Rule, while issued pursuant to, inter alia, § 110(a)(2)(D), was formally initiated by the states' recommendation and was not initiated by EPA itself, as was the NO_x SIP Call. See NOTC LEV Rule, supra note 89, at 4716 ("EPA is approving the OTC's LEV recommendation based on the determination under sections 184(c) and 110(a)(2)(D) of the Act that the recommended LEV program is necessary throughout the OTR to bring certain OTR nonattainment areas into attainment….") (emphasis added).

95. States' Brief, supra note 54, at 39.

96. Id.

97. 213 F.3d at 685-88, 30 ELR at 20415-16.

98. Id. at 688, 30 ELR at 20416 (emphasis in original).

99. Final NO_x SIP Call, supra note 2, at 57438 (e.g., vehicle inspection and maintenance programs and reformulated gasoline program); Proposed NO_x SIP Call, supra note 1, at 60328 ("One State may choose to primarily achieve emissions reductions from stationary sources while another State may focus on emissions reductions from the mobile source sector.") cited in, Michigan, 213 F.3d at 688, 30 ELR at 20416 (emphasis added). In addition, EPA had made no final determination about control costs for sources for which it lacked sufficient information, but it anticipated that states with adequate information could implement reasonably cost-effective control schemes. Final NO_x SIP Call, supra note 2, at 57416-17.

100. Final NO_x SIP Call, supra note 2, at 57369-70 (explaining that, unlike the Virginia case, states subject to the NO_x SIP Call "will have significant discretion to choose the appropriate mix of controls to meet the emissions budget"). See also supra notes 89-90 and accompanying text (explaining the additional emissions requirements imposed by the NOTC LEV Rule if a state did not adopt the LEV program).

101. Final NO_x SIP Call, supra note 2, at 57492 (to be codified at 40 C.F.R. § 51.121(e)(1)) (setting 1 emissions budget for each of the 23 states).

102. 213 F.3d at 688, 30 ELR at 20416.

103. Final NO_x SIP Call, supra note 2, at 57399-400, 57438.

104. Id. at 57416. EPA also believed that cost-effective controls might be available for several other types of sources, including medical waste incinerators, fiberglass manufacturers, and fluid catalytic cracking units at petroleum refineries, but recognized that the numbers of these sources were relatively small. Id.

105. Id. (acknowledging that "NO_x controls may be available at costs less than $ 2,000 per ton for lime kilns"). See also id. at 57438 ("In individual States, emissions from such sources [in categories with few numbers] could be a high percentage of uncontrolled emissions, and, thus, be subject to efficient, cost-effective controls for that particular State.").