Sierra Club v. Douglas M. Costle, Administrator of the Environmental Protection Agency, National Coal Association, Alabama Power Company, Intervenors

The statutory text did not command a single nationally uniform percentage-reduction figure. It required EPA to select standards reflecting reductions achievable through the best adequately demonstrated continuous-control technology while considering cost, energy requirements, and nonair environmental effects. Section 111 also expressly allowed EPA to distinguish among classes, types, and sizes of sources, and nothing in the text barred distinctions based on fuel sulfur content.

The legislative history supported, rather than contradicted, this reading. The Conference Committee and later clarifying statements recognized EPA’s discretion to set a range of reductions reflecting varying fuel characteristics, provided EPA carefully documented that the departure from uniformity did not undermine the Act’s basic purposes. Sulfur content was a natural and relevant fuel characteristic.

The court rejected Sierra Club’s view that EPA could vary a standard only when a uniform percentage was technologically impossible. Section 111 required a broader policy judgment: EPA could choose among achievable levels of control after balancing statutory factors. The Act therefore did not force EPA to demand the maximum technologically feasible reduction in every circumstance.

A variable national NSPS also did not inherently conflict with the Act’s prevention-of-significant-deterioration, nonattainment, or visibility programs. NSPS are minimum national requirements; the Act separately permits EPA and states to impose more stringent, site-specific controls where local air-quality or visibility conditions require them.

The court applied deferential but meaningful review. EPA was required to consider relevant factors, explain the connection between record facts and its policy choice, and avoid arbitrariness. The court would not redo the policy decision or substitute its own judgment for the agency’s expert assessment.

EPA could evaluate the long-term national and regional consequences of competing standards, rather than only immediate plant-level technology. Section 111’s reference to the “best technological system,” coupled with its direction to consider costs, energy, and environmental effects, allowed EPA to examine how standards would affect future capacity, retirement of older plants, fuel use, consumer costs, and aggregate emissions.

EPA’s use of an econometric computer model was permissible despite acknowledged uncertainty. Modeling was especially useful given the Clean Air Act’s demand for extensive economic-impact analysis. EPA exposed its assumptions, invited public criticism, adjusted aspects of the model in response to comments, conducted sensitivity tests, acknowledged uncertainties, and retained responsibility for the ultimate policy decision rather than treating the computer as the decisionmaker.

The court found a rational basis for EPA’s key assumption that utilities tend to minimize costs. Under that premise, excessively stringent controls could delay replacement of older, dirtier plants and discourage use of newer, cleaner capacity. Thus, stricter source-level requirements did not necessarily produce lower total national emissions.

Under the wet-scrubbing analysis alone, variable control was projected to yield the same national sulfur-dioxide emissions as full 90% control, while costing less annually, consuming less oil, producing less sludge, and supporting more new coal-fired capacity. The regional effects involved tradeoffs: full control favored the West, while variable control favored the East. EPA reasonably declined to spend substantially more merely to shift a limited amount of pollution geographically.

The record also supported EPA’s conclusion that variable control furthered statutory policies concerning coal availability and competition. It required some continuous control for every plant, reduced the advantage of simply locating near low-sulfur coal, preserved flexibility to use local coal, and limited the movement of western low-sulfur coal eastward compared with the earlier standard.

Section 111 did not forbid EPA from considering technological innovation. The statutory factors of cost, energy, and nonair environmental effects reasonably encompass the likely consequences of encouraging or discouraging a developing control technology that might offer lower costs, lower energy use, and easier waste disposal.

The separate innovative-technology waiver provision in § 111(j) did not make waivers the exclusive means of encouraging new technology. That provision supplemented EPA’s authority under § 111(a). Indeed, an excessively high national standard could make individual innovation waivers ineffective because an emerging technology would have little realistic chance to attain the otherwise applicable level of control.

EPA did not base the NSPS on dry scrubbing as adequately demonstrated technology. Wet scrubbing supplied the demonstrated technological basis for the standard. Dry scrubbing instead reinforced EPA’s preference for variable control and helped prompt consideration of a 70% floor.

The evidence for identifying 70% as the precise threshold most favorable to dry scrubbing was limited, and the court said it would have hesitated to uphold a variation based solely on that rationale. Still, the record showed that dry scrubbing had substantial potential advantages, remained uncertain and commercially unproven, and had not been tested on many low-alkalinity coals. EPA could reasonably conclude that giving the technology some room below a strict 90% requirement was worth evaluating.

Most importantly, EPA’s wet-scrubbing analysis independently supported the variable standard. Dry scrubbing therefore did not displace the agency’s core statutory analysis or create an impermissible departure from the standard otherwise justified by demonstrated technology.

Under Clean Air Act § 307(d), EPA had to place centrally relevant documents in the docket and provide sufficient opportunity for meaningful comment. But the agency did not have to select its final rule from only the exact alternatives stated in the notice. A final rule may include incremental changes that are the logical outgrowth of the proposal and comments.

EPA gave extensive notice that it was considering a variable rather than uniform percentage-reduction standard. Its proposal identified the central question as whether low-sulfur-coal units should be required to achieve the same percentage reduction as high-sulfur-coal units, solicited information on the environmental, energy, and economic consequences, and considered multiple sliding-scale alternatives.

The 70% minimum emerged as a serious candidate late in the process, and the later dry-scrubbing materials were substantial. Yet the agency’s phase-three analysis used the same general modeling framework that had been publicly discussed, and dry scrubbing and emerging technology had appeared in the proposed rule, technical materials, public hearings, and comments from utilities, environmental organizations, and others.

Interested parties also had actual notice before promulgation that EPA was considering a 70% to 90% standard and was focusing more heavily on dry scrubbing. Because the record independently supported variable control under wet-scrubbing assumptions, the court could not find a substantial likelihood that additional comment on the late dry-scrubbing materials would have significantly changed the rule.

EPA’s proposed 85% standard was primarily described in terms of FGD performance, while allowing credit for fuel pretreatment. The final rule was less clear in places, but it expressly stated that, in the most demanding high-sulfur-coal situations, 90% removal could be achieved through FGD combined with coal preparation techniques and sulfur retained in ash.

The record repeatedly addressed coal washing before the final rule. EPA’s technical studies examined combined washing and scrubbing; commenters debated whether washing should contribute to the standard; EPA revised its modeling to account for coal-washing credits; and EPA held a meeting concerning the state of coal-washing technology.

The utilities themselves submitted post-comment materials objecting to EPA’s apparent reliance on coal washing to justify a more stringent 90% requirement. Those objections showed actual notice and an opportunity to contest the combined-control rationale. The court therefore found no prejudice warranting another notice-and-comment cycle.

EPA could set standards for future plants at levels higher than the long-term performance of existing equipment, so long as the record substantially supported the anticipated improvements. The Clean Air Act is technology forcing, and new sources can be required to use improved design and operating practices that are shown to be feasible.

The record did not show that an existing commercial lime-or-limestone scrubber continuously achieved a 92% median removal rate on high-sulfur coal. The court rejected EPA’s simplistic attempt to interpolate a 92% figure from disparate systems. But detailed engineering reports identified specific design and operational changes expected to improve new scrubbers, and vendor evidence further supported the feasibility of improved performance.

EPA reasonably treated a 0.36 variability figure from the Cane Run plant as a conservative upper bound for a properly run lime-or-limestone system. The agency considered available data, excluded nonrepresentative malfunctioning or materially different systems, accounted for the more variable conditions of peak-load operation, and showed that later Bruce Mansfield data did not undermine the conclusion.

Coal washing independently supplied the additional reduction needed to meet the overall 90% standard. The record supported an average 27% sulfur reduction for high-sulfur coal through commercially common washing practices. Because utilities could calculate and average washing credits quarterly, variability in individual coal shipments did not make compliance unachievable.

The court also found no basis to think washed coal would create special problems for scrubbers. The record showed that lower-sulfur coal is easier to scrub and that washing removes impurities. EPA therefore could reasonably conclude that the combined use of coal washing and FGD would meet the 90% requirement.

For an achievable performance standard, EPA must identify variables affecting performance and show that its test data are representative of industry-wide conditions. The court found that EPA identified the relevant variables for electrostatic precipitators, but its ESP data were incomplete: they often lacked information about ash resistivity, collection area, test duration, and testing conditions. The court therefore would not have upheld the standard on ESP evidence alone.

EPA’s baghouse evidence was sufficient. EPA identified the relevant variables, including air-to-cloth ratio, pressure drop, cleaning, and maintenance. Its data showed that nearly all test runs at eight baghouse-equipped units met the 0.03 lbs./MBtu limit, even though those systems were not designed specifically to attain that stringent level.

EPA reasonably explained why evidence from smaller baghouses could be extrapolated to larger utility boilers. Baghouses operate through modular cells; scaling up generally means adding modules rather than relying on an entirely different technology. A 350-MW installation had experienced startup problems but was operating within its design limit and the standard, while numerous other utility-scale baghouses were planned or operating.

Post-promulgation operating evidence from large baghouses did not show unreasonable costs. EPA demonstrated that its cost estimates already included substantial allowances for bag replacement and maintenance, and that the cited high-pressure-drop problems largely reflected undersizing and correctable startup or design issues.

Clean Air Act § 307(d) supplied the governing procedural standard. A court may invalidate a rule for procedural error only when the agency failed to observe required procedures and the error was so serious, and so centrally relevant, that there is a substantial likelihood the rule would have changed. Vermont Yankee also barred the court from imposing its own preferred procedures beyond those required by Congress except in extraordinary circumstances.

EPA could accept and docket written submissions received after the formal comment period. The statute required docketing of centrally relevant documents that became available after publication of a proposed rule, and it did not prohibit later submissions. EPA did not have to reopen the formal comment period merely because additional information arrived, particularly where Congress required expeditious rulemaking and EPA faced court-imposed deadlines.

A new comment period might have been necessary if EPA had relied on crucial, late-submitted data that left opponents without a meaningful opportunity to respond. But EDF identified no such document. EDF had notice of, attended, and challenged the coal-industry data presented at the April 5 meeting; it submitted a detailed response shortly afterward and had additional time during reconsideration to present further rebuttal.

The Act did not impose a general ban on post-comment oral contacts in informal, policy-oriented rulemaking. Unlike adjudication, informal rulemaking involves broad legislative policymaking affecting many interests. Agencies may properly receive information and advocacy from regulated entities, public-interest groups, Congress, and the public, provided the final rule rests on the public record.

EPA had discretion to decide whether oral communications were centrally relevant enough to require docketing. Most challenged meetings were in fact summarized in the docket. The two undocketed sessions did not invalidate the rule: one was an EPA briefing of Senate staff, and the other was a White House meeting involving the President and executive officials.

EPA was not required to docket the President’s face-to-face policy session with executive officials. The President has constitutional authority to supervise executive policymaking, and EPA is an Executive Branch agency. Because EPA did not purport to base the rule on undisclosed information or data from that meeting, the statutory record requirement was satisfied.

Meetings with Senator Byrd likewise did not establish unlawful congressional interference. Congressional representatives may advocate constituent interests in informal rulemaking. The record did not show that the Senator used improper pressure to make EPA consider factors outside the Clean Air Act, or that EPA’s decision rested on such extraneous factors rather than the documented coal-reserve and technological analysis.