National Ass'n of Clean Water Agencies v. Environmental Protection Agency
734 F.3d 1115
D.C. Cir.
2013

Background

  • EPA promulgated MACT emission standards for sewage sludge incineration units under Clean Air Act §129 in March 2011 after a court-ordered deadline; EPA treated such units as "solid waste incineration units" under §129(g)(1).
  • EPA subdivided units into two categories (multiple hearth and fluidized bed), surveyed a targeted set of facilities (data from 9 municipalities plus state databases), and set MACT floors using that dataset and statistical methods.
  • EPA used (1) control-technology targeting to choose units to test, (2) a statistical formula to judge whether datasets <12% of population could represent the top 12%, (3) an "upper prediction limit" (99% UPL) to account for variability, and (4) a three-times method-detection-level test to handle nondetects.
  • Petitioners (NACWA and Sierra Club) challenged EPA's authority to regulate under §129 (vs §112) and challenged multiple aspects of EPA's MACT-floor methodology (insufficient/targeted dataset, use of control-technology proxy, UPL and variability treatment, nondetect handling, monitoring, subcategorization). MaxWest intervened but raised issues outside petitioners' scope.
  • The D.C. Circuit upheld EPA's §129 jurisdiction (Chevron deference) but remanded parts of the MACT-floor methodology for further explanation (without vacating the standards): specifically EPA must clarify the role of Part 503 data, the UPL and variability analysis, and the statistical extrapolation used to fill dataset gaps.

Issues

Issue Plaintiff's Argument Defendant's Argument Held
Authority to regulate sewage sludge incinerators under §129 vs §112 NACWA: "from the general public" means proximate source; sewage sludge is produced by POTWs and thus not covered by §129 EPA: "from" may refer to original/but-for source; sewage sludge ultimately originates from the public; §129 reasonably covers these units Court: statute ambiguous; EPA's original-source interpretation is reasonable under Chevron step two — EPA has authority under §129
Use of control-technology targeting (selecting units by controls rather than by measured emissions) Sierra Club: unlawful/arbitrary because non-technology factors affect emissions; EPA must show control tech largely determines emissions EPA: Part 503 and targeted selection make sludge streams relatively homogeneous; control tech is a reasonable proxy Court: EPA failed to show with substantial evidence that non-technology factors are negligible; remand for EPA to explain/support the proxy choice
Dataset insufficiency and statistical extrapolation (<12% sample; sample-size formula) NACWA/Sierra Club: §129 requires MACT floors reflect best-performing 12%; EPA cannot lawfully extrapolate from less-than-12% data or must justify representativeness EPA: statute allows estimation; used a statistical formula to determine minimum observations and relied on normality assumptions and conservative variable choices Court: EPA may estimate but must substantiate its statistical method and choice of parameters; remand for detailed justification of the extrapolation and variable selection
Variability and Upper Prediction Limit (UPL) to set floors Sierra Club/NACWA: UPL does not clearly represent the "average emissions limitation achieved" or worst-foreseeable conditions; UPL may misstate true average; UPL may not capture sludge-content variability EPA: UPL models a 99%-confidence upper bound for future 3-run averages and accounts for intra- and (EPA says) inter-unit variability; Part 503 reduces sludge variability Court: EPA must clarify (1) how UPL equates to the §129 "average" requirement, (2) why UPL reasonably predicts worst-foreseeable performance, and (3) whether and how UPL accounts for more than intra-unit variability; remand for explanation
Non-detect data handling (3x representative MDL) Sierra Club: three-times-MDL approach may not reflect what best performers actually achieve EPA: measurement imprecision near MDL justified inflating to 3x to reach reasonably precise values; method has technical basis Court: upheld EPA's nondetect method as reasonable and within technical deference
Monitoring and subcategorization Sierra Club: EPA should mandate continuous emissions monitoring for all pollutants; NACWA: more subcategories (e.g., backup/emergency units) needed EPA: statute allows parameter monitoring and discretion on subcategorization; no reliable data supported more subcategories Court: monitoring scheme lawful (continuous monitoring not required by §129); subcategorization into two types (multiple hearth/fluidized bed) upheld; EPA's refusal to further subdivide not arbitrary

Key Cases Cited

  • Chevron U.S.A., Inc. v. Natural Resources Defense Council, 467 U.S. 837 (agency deference framework)
  • Natural Resources Defense Council v. EPA, 489 F.3d 1250 (D.C. Cir. 2007) (interpretation of §129 definitions)
  • Sierra Club v. EPA, 167 F.3d 658 (D.C. Cir. 1999) (MACT-floor estimation principles)
  • Cement Kiln Recycling Coalition v. EPA, 255 F.3d 855 (D.C. Cir. 2001) (limitations on MACT-technology proxy)
  • Sierra Club v. EPA ("Brick MACT"), 479 F.3d 875 (D.C. Cir. 2007) (rejecting MACT approach when non-technology factors matter)
  • Mossville Environmental Action Now v. EPA, 370 F.3d 1232 (D.C. Cir. 2004) (upholding use of existing standard when variability prevents identifying best performers)
  • Northeast Md. Waste Disposal Authority v. EPA, 358 F.3d 936 (D.C. Cir. 2004) (agency's MACT data-gathering latitude and remand standards)
  • Husqvarna AB v. EPA, 254 F.3d 195 (D.C. Cir. 2001) (deference on cost-effectiveness judgments)
Read the full case