Southwestern Elec. Power Co. v. U.S. E.P.A.

Case: 15-60821 Document: 00514914192 Page: 1 Date Filed: 04/12/2019

IN THE UNITED STATES COURT OF APPEALS
FOR THE FIFTH CIRCUIT United States Court of Appeals
Fifth Circuit

FILED
April 12, 2019
No. 15-60821
Lyle W. Cayce
Clerk
SOUTHWESTERN ELECTRIC POWER COMPANY; UTILITY WATER ACT
GROUP; UNION ELECTRIC COMPANY, doing business as Ameren
Missouri; WATERKEEPER ALLIANCE, INCORPORATED;
ENVIRONMENTAL INTEGRITY PROJECT; SIERRA CLUB; AMERICAN
WATER WORKS ASSOCIATION; NATIONAL ASSOCIATION OF WATER
COMPANIES; CITY OF SPRINGFIELD, MISSOURI, by and through the
Board of Public Utilities; DUKE ENERGY INDIANA, INCORPORATED,

Petitioners

v.

UNITED STATES ENVIRONMENTAL PROTECTION AGENCY; ANDREW
WHEELER, in his official capacity as Acting Administrator of the United
States Environmental Protection Agency,

Respondents

On Petitions for Review of Final Administrative Actions
of the United States Environmental Protection Agency

Before HAYNES, HO, and DUNCAN, Circuit Judges.
STUART KYLE DUNCAN, Circuit Judge:
Steam-electric power plants generate most of the electricity used in our
nation and, sadly, an unhealthy share of the pollution discharged into our
nation’s waters. To control this pollution, the Clean Water Act, 33 U.S.C.
§ 1251 et seq., empowers the Environmental Protection Agency to promulgate
and enforce rules known as “effluent limitation guidelines” or “ELGs.” Id.
 Case: 15-60821 Document: 00514914192 Page: 2 Date Filed: 04/12/2019

No. 15-60821
§§ 1311, 1314, 1362(11). For quite some time, ELGs for steam-electric power
plants have been, in EPA’s words, “out of date.” 80 Fed. Reg. 67,838. That is a
charitable understatement. The last time these guidelines were updated was
during the second year of President Reagan’s first term, the same year that
saw the release of the first CD player, the Sony Watchman pocket television,
and the Commodore 64 home computer. In other words, 1982. See id. (noting
ELGs were “promulgated and revised in 1974, 1977, and 1982”). The guidelines
from that bygone era were based on “surface impoundments,” which are
essentially pits where wastewater sits, solids (sometimes) settle out, and toxins
leach into groundwater. Id. at 67,840, 67,851. Impoundments, EPA tells us,
have been “largely ineffective at controlling discharges of toxic pollutants and
nutrients.” Id. at 67,840. Consequently, in 2005 the agency began a multi-year
study to bring the steam-electric ELGs into the 21st century. Id. at 67,841.
In November 2015, EPA unveiled the final rule: the “Effluent
Limitations Guidelines and Standards for the Steam Electric Power
Generating Point Source Category,” 80 Fed. Reg. 67,838 (Nov. 3, 2015). The
rule updates guidelines for six of the wastestreams that issue from plants and
foul our waters. Importantly, the Clean Water Act requires setting new ELGs
based on the “Best Available Technology Economically Available” or “BAT.” 33
U.S.C. § 1314(b)(2)(B). BAT is the gold standard for controlling water pollution
from existing sources. By requiring BAT, the Act forces implementation of
increasingly stringent pollution control methods. See NRDC v. EPA, 822 F.2d
104, 123 (D.C. Cir. 1987) (describing the Act as “technology-forcing”).
We consider a challenge to the final rule brought by various
environmental petitioners. They target two discrete parts of the rule: the new
ELGs for “legacy wastewater” (wastewater from five of the six streams
generated before a specific date) and for “combustion residual leachate” (liquid
that percolates through landfills and impoundments). These two categories
2
 Case: 15-60821 Document: 00514914192 Page: 3 Date Filed: 04/12/2019

No. 15-60821
account for massive amounts of water pollution. For instance, leachate alone
would qualify as the 18th-largest source of water pollution in the nation,
producing more toxic-weighted pound equivalents than the entire coal mining
industry. The environmental petitioners’ basic complaint is that EPA set an
unlawful BAT for these two categories. Whereas the BAT for the other streams
adopts modern technologies, they claim the agency arbitrarily set BAT for
legacy wastewater and leachate using the same archaic technology in place
since 1982—namely, impoundments. It was as if Apple unveiled the new iMac,
and it was a Commodore 64.
The environmental petitioners challenge those portions of the rule under
the Administrative Procedure Act and the well-worn Chevron test governing
review of agency action. See Chevron USA, Inc. v. Nat. Res. Def. Council, Inc.,
467 U.S. 837 (1984). For the reasons discussed below, we agree that the
portions of the rule regulating legacy wastewater and combustion residual
leachate are unlawful. Accordingly, we VACATE those portions of the rule and
REMAND to the agency for reconsideration.
I. BACKGROUND
A. The Clean Water Act
The Clean Water Act (“CWA” or “Act”), 86 Stat. 833, as amended, 33
U.S.C. § 1251 et seq., was enacted over President Nixon’s veto in 1972. See
Train v. City of New York, 420 U.S. 35, 40 (1975). Few laws have shouldered a
weightier burden—namely, “to restore and maintain the chemical, physical,
and biological integrity of the Nation’s waters.” 33 U.S.C. § 1251(a); see also,
e.g., City of Milwaukee v. States of Illinois and Michigan, 451 U.S. 304, 318
(1981) (“Congress’ intent in enacting [the CWA] was clearly to establish an all-
encompassing program of water pollution regulation.”); Am. Petroleum Inst. v.
EPA, 661 F.2d 340, 343-44 (5th Cir. Unit A Nov. 13, 1981) (“API I”) (noting
CWA’s “ambitious purpose”). To that end, the Act makes “unlawful” the
3
 Case: 15-60821 Document: 00514914192 Page: 4 Date Filed: 04/12/2019

No. 15-60821
“discharge of any pollutant by any person” into the nation’s “navigable waters,”
unless otherwise permitted. 33 U.S.C. §§ 1311(a), 1362(7), (12). 1
We have previously detailed the Act’s “distinct, though interlocking,
regulatory schemes.” Chem. Mfrs. Ass’n v. EPA, 870 F.2d 177, 195 (5th Cir.
1989) (“CMA”), clarified on reh’g, 885 F.2d 253. 2 Here we focus on one of the
Act’s key regulatory tools: “effluent limitation guidelines” (“ELGs” or
“guidelines”), which are nationwide standards set by the EPA Administrator
to govern pollutant discharges from point sources. See 33 U.S.C. § 1314(b)
(authorizing Administrator to set “effluent limitation guidelines” for “classes
and categories of point sources”); Tex. Oil & Gas, 161 F.3d at 927 (“ELGs are
the rulemaking device prescribed by the CWA to set national effluent
limitations for categories and subcategories of point sources”). 3

1 Under the Act “pollutant” means “dredged spoil, solid waste, incinerator residue,
sewage, garbage, sewage sludge, munitions, chemical wastes, biological materials,
radioactive materials, heat, wrecked or discarded equipment, rock, sand, cellar dirt and
industrial municipal, and agricultural waste discharged into water.” 33 U.S.C. § 1362(6).
“Pollution” means “the man-made or man-induced alteration of the chemical, physical,
biological, and radiological integrity of water.” Id. § 1362(19). “Discharge of a pollutant”
means “(A) any addition of any pollutant to navigable waters from any point source, [and] (B)
any addition of any pollutant to the waters of the contiguous zone or the ocean from any point
source other than a vessel or other floating craft.” Id. § 1362(12).
2 See, e.g., CMA, 870 F.2d at 195-96 (describing the CWA’s (1) grants-in-aid for
publicly-owned waste-treatment works, 33 U.S.C. §§ 1281-88, 1291-92; (2) authorization to
set and enforce federal effluent standards, id. §§ 1311, 1314; (3) requirement of state-
established and federally-approved water-quality criteria, id. § 1313; and (4) creation of
pollution permitting through the National Pollutant Discharge Elimination System
(NPDES), id. § 1342; see also, e.g., Tex. Oil & Gas Ass’n v. EPA, 161 F.3d 923, 927-29 (5th
Cir. 1998) (“Tex. Oil & Gas”) (discussing effluent limitations and NPDES permitting); API I,
661 F.2d at 341-42, 343-44 (discussing effluent limitations); see also generally EPA v. Nat’l
Crushed Stone Ass’n, 449 U.S. 64, 69-70 (1980) (outlining basic structure of CWA).
3 A “point source” means “any discernible, confined and discrete conveyance . . . from
which pollutants are or may be discharged,” but “does not include agricultural stormwater
discharges and return flows from irrigated agriculture.” 33 U.S.C. § 1362(14). An “effluent
limitation” means “any restriction established by a State or the Administrator on quantities,
rates, and concentrations of chemical, physical, biological, and other constituents which are
discharged from point sources into navigable waters, the waters of the contiguous zone, or
the ocean, including schedules of compliance.” Id. § 1362(11).
4
 Case: 15-60821 Document: 00514914192 Page: 5 Date Filed: 04/12/2019

No. 15-60821
The Act requires ELGs to be based on technological feasibility rather
than on water quality. Id. at 927 (citing E.I. du Pont de Nemours & Co. v.
Train, 430 U.S. 112, 130-31 (1977); API I, 661 F.3d at 343-44). That is, the
Administrator must “require industry, regardless of a discharge’s effect on
water quality, to employ defined levels of technology to meet effluent
limitations.” API I, 661 F.3d at 344; see also Tex. Oil & Gas, 161 F.3d at 927
(ELGs are “technology-based rather than harm-based” insofar as they “reflect
the capabilities of available pollution control technologies to prevent or limit
different discharges rather than the impact that those discharges have on the
waters”). The Act therefore mandates a system in which, as available pollution-
control technology advances, pollution-discharge limits will tighten. See, e.g.,
Nat’l Crushed Stone, 449 U.S. at 69 (the Act “provides for increasingly
stringent effluent limitations”) (citing 33 U.S.C. § 1311(b)); CMA, 870 F.2d at
196 (the Act requires compliance with “technology-based pollutant-effluent
limitations that, in time, will become more stringent”) (citing 33 U.S.C.
§§ 1311(b), 1314(b)). The D.C. Circuit accurately described this aspect of the
Act’s scheme as “technology-forcing,” meaning it seeks to “press development
of new, more efficient and effective [pollution-control] technologies.” NRDC v.
EPA, 822 F.2d 104, 123 (D.C. Cir. 1987) (“NRDC I”); see also, e.g., NRDC v.
EPA, 808 F.3d 556, 563-64 (2nd Cir. 2015) (“NRDC II”) (describing ELG
scheme as “technology-forcing, meaning it should force agencies and permit
applicants to adopt technologies that achieve the greatest reductions in
pollution”) (citing NRDC I). 4

4 The Act is not based solely on technological feasibility standards. It also incorporates
water-quality standards into the permitting process that effectuates ELGs. See, e.g., NRDC
II, 808 F.3d at 564-65 (“If the [ELGs] are insufficient to attain or maintain water quality
standards, the CWA requires NPDES permits to include additional water quality-based
effluent limits[.]” (citing 33 U.S.C. §§ 1311(b)(1)(C), 1312(a); NRDC I, 822 F.2d at 110)).
5
 Case: 15-60821 Document: 00514914192 Page: 6 Date Filed: 04/12/2019

No. 15-60821
The Act prescribes various technological standards to be used in setting
effluent limitations. Two are relevant here: “best practicable control technology
currently available” (“BPT”) and “best available technology economically
achievable” (“BAT”). Compare 33 U.S.C. §§ 1311(b)(1)(A); 1314(b)(1)(B) (BPT),
with id. §§ 1311(b)(2)(A); 1314(b)(2)(B) (BAT). The less stringent of these two
standards is BPT, which the Supreme Court has described as only “a first step
toward [the Act’s] goal.” Nat’l Crushed Stone, 449 U.S. at 75 n.14; see also, e.g.,
BP Explor. & Oil, Inc. v. EPA, 66 F.3d 784, 789 (6th Cir. 1995) (describing BPT
as “the first stage of pollutant reduction”). BPT applied to limitations on direct
discharges of pollutants during an interim period (originally slated to end in
1984 but later extended to 1989). 33 U.S.C. § 1311(b)(1)(A); Tex. Oil & Gas, 161
F.3d at 927-28; CMA, 870 F.2d at 196. Instead of defining BPT, the Act lists
various factors the Administrator must consider in determining it—including
an explicit cost/benefit analysis: “the total cost of application of technology in
relation to the effluent reduction benefits to be achieved from such
application.” 33 U.S.C. § 1314(b)(1)(B). 5 We have explained that “BPT
limitations are intended to represent the average of the best levels of
performance by existing plants of various sizes, ages, and unit processes within

5 The complete list of BPT factors is:
• Total cost of application of technology in relation to the effluent reduction benefits to be
achieved from such application;
• Age of equipment and facilities involved;
• Process employed;
• Engineering aspects of the application of various types of control techniques;
• Process changes;
• Non-water quality environmental impact (including energy requirements); and
• “[S]uch other factors as the Administrator deems appropriate.”
33 U.S.C. § 1314(b)(1)(B).
6
 Case: 15-60821 Document: 00514914192 Page: 7 Date Filed: 04/12/2019

No. 15-60821
the category or subcategory for control of conventional pollutants.” CMA, 870
F.2d at 203 (citing 52 Fed. Reg. 42,525); see also, e.g., Nat’l Crushed Stone, 449
U.S. at 75-76 (discussing BPT).
The stricter of the two standards is BAT, which has applied to existing,
direct discharges of toxic and non-conventional pollutants since March 31,
1989. See 33 U.S.C. §§ 1311(b)(2)(A); 1314(b)(2)(A); Tex. Oil & Gas, 161 F.3d at
927-28; see also BP Explor., 66 F.3d at 790 (describing BAT as “the second
stage” of pollutant reduction). When pollutants are regulated under this
standard, the EPA “must set discharge limits that reflect the amount of
pollutant that would be discharged by a point source employing the best
available technology that the EPA determines to be economically feasible
across the category or subcategory as a whole.” Tex. Oil & Gas, 161 F.3d at
928. We have held that BAT limitations must “be based on the performance of
the single best-performing plant in an industrial field.” CMA, 870 F.2d at 226.
In describing the relationship between BAT and BPT, the Supreme Court has
explained that a BAT must achieve “reasonable further progress” towards the
Act’s goal of eliminating pollution, and BPT serves as the “prior standard” for
measuring that progress. See Nat’l Crushed Stone, 449 U.S. at 75 (explaining
that “BPT serves as the prior standard with respect to BAT[’s]” reasonable
further progress requirement). As with BPT, the Act lists factors the
Administrator must consider in determining BAT. 33 U.S.C. § 1314(b)(2)(B). 6

6 The complete list of BAT factors is:
• Age of equipment and facilities involved;
• Process employed;
• Engineering aspects of the application of various types of control techniques;
• Process changes;
• Cost of achieving such effluent reduction;
• Non-water quality environmental impact (including energy requirements); and
7
 Case: 15-60821 Document: 00514914192 Page: 8 Date Filed: 04/12/2019

No. 15-60821
The Administrator has “considerable discretion” in weighing those factors. Tex.
Oil & Gas, 161 F.3d at 928 (citation omitted). Unlike BPT, however, the BAT
factors omit a cost/benefit analysis and replace it with a requirement to
consider only “the cost of achieving such effluent reduction.” Id; see also, e.g.,
Nat’l Crushed Stone, 449 U.S. at 71 (BPT and BAT factors are
“similar . . . except that in assessing BAT total cost is no longer to be
considered in comparison to effluent reduction benefits”). Indeed, the Supreme
Court has explained that, unlike BAT, “BPT limitations do not require an
industrial category to commit the maximum economic resources to pollution
control, even if affordable.” Nat’l Crushed Stone, 449 U.S. at 75. 7
B. The Final Rule
The rule at issue in this case regulates effluent discharges from steam-
electric power plants. Those plants burn nuclear or fossil fuels to heat water in
boilers, generating steam that drives turbines connected to electric generators.
80 Fed. Reg. 67,839 n.1. This process produces something nearly everyone
regards as good: electricity. Indeed, the plants regulated by the rule provide
most of the electricity annually produced in the United States. But the process
also produces something everyone regards as bad: pollution. According to EPA,
discharges from these plants account for “about 30 percent of all toxic

• “[S]uch other factors as the Administrator deems appropriate.”
33 U.S.C. § 1314(b)(2)(B).
7 BAT and BPT standards apply to regulation of existing steam-powered electric
plants. A standard even stricter than BAT—the “new source performance standard”
(“NSPS”)—applies to newly built plants. The Act defines this standard as “the greatest degree
of effluent reduction which the Administrator determines to be achievable through
application of the best available demonstrated control technology, processes, operating
methods, or other alternatives, including, where practicable, a standard permitting no
discharge of pollutants.” 33 U.S.C. § 1316(a)(1). Additional standards of varying strictness
are laid out for other potential pollution sources, such as “pretreatment standards for existing
sources” (“PSES”). See 33 U.S.C. § 1314(g).
8
 Case: 15-60821 Document: 00514914192 Page: 9 Date Filed: 04/12/2019

No. 15-60821
pollutants discharged into surface waters by all industrial categories regulated
under the CWA.” Id. at 67,839-40; see also, e.g., Michigan v. EPA, 135 S. Ct.
2699, 2705 (2015) (addressing regulation of air pollution from power plants
under the Clean Air Act); ConocoPhillips Co. v. EPA, 612 F.3d 822, 826 (5th
Cir. 2010) (addressing regulation of cooling water systems at power plants).
For instance, power plant discharges contain toxic metals such as mercury,
arsenic, lead, and selenium, which bioaccumulate in fish, accumulate in lake
and reservoir sediment, and pollute drinking water supplies. People who eat
the tainted fish or drink the tainted water can suffer negative health
consequences such as cancer, cardiovascular disease, neurological disorders,
kidney and liver damage, and lowered IQs (in children). Id. at 67,840.
EPA first promulgated and then revised ELGs for steam-electric power
plants in 1974, 1977, and 1982. See id.; see also 39 Fed. Reg. 36,186 (Oct. 8,
1974); 42 Fed. Reg. 15,690 (Mar. 23, 1977); 47 Fed. Reg. 52,290 (Nov. 19, 1982).
Those guidelines are now, in the agency’s words, “out of date,” because “[t]hey
do not adequately control the pollutants (toxic metals and other[s]) discharged
by this industry, nor do they reflect relevant process and technology advances
that have occurred in the last 30-plus years.” 80 Fed. Reg. 67,840. The old rules
and the processes they regulated are relics of the past:
The processes employed and pollutants discharged by the industry
look very different today than they did in 1982. Many plants,
nonetheless, still treat their wastewater using only surface
impoundments, which are largely ineffective at controlling
discharges of toxic pollutants and nutrients.

Id. (“Surface impoundments” are ponds designed to allow particulates to settle
out of wastewater by force of gravity. See infra.) Happily, though, EPA reports
that, “[i]n the several decades since the steam electric ELGs were last revised,”
technologies that are more effective, “affordable,” and “widely available” have
“increasingly been used at plants.” Id. Thus, EPA began a new rulemaking to
9
 Case: 15-60821 Document: 00514914192 Page: 10 Date Filed: 04/12/2019

No. 15-60821
update power plant ELGs. The agency conducted a detailed industry study 8
from 2006-2009 and on June 7, 2013 issued a proposed rule, 78 Fed. Reg.
34,432, 34,439, that generated over 200,000 comments. 80 Fed. Reg. 67,844.
On November 3, 2015 the agency issued a final rule entitled “Effluent
Limitations Guidelines and Standards for the Steam Electric Power
Generating Point Source Category” (“final rule” or “rule”). 80 Fed. Reg. 67,838.
The rule addresses these six streams produced by power plants:
1. Flue gas desulfurization (FGD) wastewater
2. Fly ash transport wastewater
3. Bottom ash transport wastewater
4. Flue gas mercury control (FGMC) wastewater (“Hg control waste”)
5. Combustion residual leachate (or “Leachate”)
6. Gasification wastewater (not depicted in figure below).

8 The study produced a 233-page report. See Steam Electric Power Generating Point
Source Category: Final Detailed Study Report, EPA 821-R-09-008 (Sept. 2009) (“EPA Study
Report”). The report vividly describes the harms from water pollution—for instance, one
dissolved metal, selenium, can kill fish and other aquatic life at “concentrations below eight
parts per billion.” Id. at 6-4.
10
 Case: 15-60821 Document: 00514914192 Page: 11 Date Filed: 04/12/2019

No. 15-60821
80 Fed. Reg. 67,846-47. 9 The rule treats another category (“legacy”
wastewater), which is a subset of five other streams. Infra I.B.1. 10 This
diagram illustrates how such streams are produced:

9 Each stream may be briefly described as follows (see id. at 67,846-47):
• FGD wastewater is produced by systems that remove sulfur dioxide from flue gas using
a “sorbent slurry.”
• Fly ash wastewater is produced when ash in flue gas is emitted from a boiler, trapped
by filters, and then sluiced from hoppers to a surface impoundment.
• Bottom ash wastewater is produced when ash falling to the furnace bottom is sluiced
by water from hoppers to an impoundment or dewatering bin.
• FGMC wastewater is produced when carbon is injected into flue gas to facilitate
removal of mercury and the resulting waste is wet sluiced with fly ash.
• Leachate is liquid that percolates through or drains from a landfill, or that passes
through the containment structure (e.g., bottom, dikes, berms) of an impoundment.
• Gasification wastewater is produced by the cleaning of a synthetic gas produced by
subjecting coal or coke to high temperature and pressure.
10 The rule addresses a seventh wastestream—“nonchemical metal cleaning wastes”—
not at issue here. 80 Fed. Reg. 67,850.
11
 Case: 15-60821 Document: 00514914192 Page: 12 Date Filed: 04/12/2019

No. 15-60821
When EPA originally regulated steam-electric effluents in the 1970s and
1980s, it did so under the less-stringent BPT standard, see supra I.A, and set
BPT for bottom ash transport water and leachate as surface impoundments.
See 80 Fed. Reg. 67,848-49. Surface impoundments, or “ash ponds,” are
essentially watery pits that “rely on gravity to remove particulates from
wastewater” and were “the technology basis for the previously promulgated
BPT effluent limitations for low volume waste sources.” Id. at 67,840, 67,851.
As the new rule describes, however, the ensuing three decades have rendered
that BPT standard “out of date,” because it “do[es] not adequately control the
pollutants (toxic metals and other[s]) discharged by this industry, nor do[es]
[it] reflect relevant process and technology advances that have occurred in the
last 30-plus years.” Id. at 67,840. Moreover, the Act required that the new
guidelines for existing direct 11 discharges conform to the stricter BAT
standard. See 33 U.S.C. § 1311(b)(2)(A), 80 Fed. Reg. 67,848-49; see supra I.A.
EPA thus considered more advanced control methods, which it notes are
“affordable technologies that are widely available and already in place at some
plants.” 80 Fed. Reg. 67,840. The agency describes those methods as follows:
• Chemical precipitation means treating wastewater by
introducing chemicals that will react with substances currently
dissolved or suspended in the water to produce a solid, non-
soluble precipitate, which then can be filtered out or left to settle
to the bottom of the wastewater. EPA Wastewater Technology
Fact Sheet, EPA 832-F-00-018 (Sept. 2000).
• Biological treatment means introducing bacteria or other
microorganisms to remove pollutants, specifically “heavy metals,
selenium, and nitrates.” 80 Fed. Reg. 67,850.
• Dry handling, for fly ash, means “a dry vacuum system that
employs a mechanical exhauster to pneumatically convey the fly

11 “Indirect” discharges concern pollutants discharged into a “publicly owned
treatment work” and are subject to distinct “pretreatment” standards. See 80 Fed. Reg.
67,841; 33 U.S.C. § 1317(b), (c).
12
 Case: 15-60821 Document: 00514914192 Page: 13 Date Filed: 04/12/2019

No. 15-60821
ash (via a change in air pressure) from hoppers directly to a silo,”
without getting the ash wet. Id. at 67,852. For bottom ash, dry
handling refers to “a system in which bottom ash is collected in
a water quench bath and a drag chain conveyor (mechanical drag
system) then pulls the bottom ash out of the water bath on an
incline to dewater the bottom ash.” Id.
• Evaporation, for FGD wastewater and gasification wastewater,
means using “a falling-film evaporator (also known as a brine
concentrator) to produce a concentrated wastewater stream
(brine) and a distillate stream.” Id. at 67,838, 67,853.
From those options EPA selected the following technologies as BAT for
the various wastestreams:

Technology basis for the main
Wastestreams
BAT/NSPS/PSES/PSNS regulatory options
FGD Wastewater………………… Chemical Precipitation + Biological Treatment
Fly Ash Transport Water……..... Dry handling
Bottom Ash Transport Water…. Dry handling / Closed loop
FGMC Wastewater……………… Dry handling
Gasification Wastewater………. Evaporation
Leachate………………………….. Impoundment (Equal to BPT)

80 Fed. Reg. 67848-49 (adapted from Table VIII–1–FINAL RULE: STEAM
ELECTRIC MAIN REGULATORY OPTIONS). As shown, the rule set more advanced
technologies as BAT for five of the six wastestreams. See also id. at 67,850,
67,852, 67,853 (explaining selection for each stream). For leachate and “legacy”
wastewater, however, the rule selected “impoundment” as BAT, the same
technology set as BPT in 1982. Id. at 67,854. Our focus is on the rule’s
treatment of those streams, and so we provide additional detail below.
1. Legacy Wastewater
Legacy wastewater is not a distinct type of wastestream. Instead, as the
final rule explains, the term describes wastewater from five of the streams
(FGD, fly ash, bottom ash, FGMC, and gasification wastewater) that is
“generated prior to” a future date. 80 Fed. Reg. at 67,854. That date, which is

13
 Case: 15-60821 Document: 00514914192 Page: 14 Date Filed: 04/12/2019

No. 15-60821
determined by the permitting authority, is required to be “as soon as possible
beginning November 1, 2020 but no later than December 31, 2023.” 82 Fed.
Reg. 43,496. Wastewater from streams generated before that date is
denominated “legacy” wastewater and is not subject to the stricter BAT
applicable to those streams. Id. Instead, the BAT for legacy wastewater is
“equal to the previously promulgated BPT regulations” in effect since 1982—
namely, impoundments. Id. This means that legacy wastewater is allowed by
the final rule to contain the same quantity of toxic pollutants allowed since
1982. See id. (setting BAT for legacy wastewater “equal to the previously
promulgated BPT limitations on [total suspended solids] in the discharge of fly
ash transport water, bottom ash transport water, and low volume waste
sources”); see also EPA Study Report, EPA 821-R-09-008, at 5-20 (table listing
pollutant concentrations at several individual impoundments studied during
the rulemaking process). The “legacy” category will thus encompass a massive
amount of wastewater from the five composite streams. For instance, according
to the EPA’s Study Report, in 2008 alone the average plant produced over 2.7
billion gallons of fly ash transport water per year, as well as over 1.1 billion
gallons of bottom ash transport water. Id. at 5-6, 5-7.
The rule imposes much more stringent limits on wastewater from these
same streams generated after the date to be set by the permitting authority
(again, between November 1, 2020 and December 31, 2023). For instance, EPA
found that a combination of chemical precipitation and biological treatment
was the BAT for treating pollution from non-legacy FGD wastewater, and that
“dry handling” (a technique for disposing of fly ash and bottom ash without
adding water) was the BAT for non-legacy ash wastestreams. Id. at 67,850-53.
These technologies are significantly newer than surface impoundments, and
EPA concluded they were the superior option for treating pollution in non-
legacy wastewater. Id.
14
 Case: 15-60821 Document: 00514914192 Page: 15 Date Filed: 04/12/2019

No. 15-60821
The rule accounts for the discrepancy between legacy and non-legacy
wastewater regulations in various ways. For instance, it explains that legacy
wastewater “already exists in wet form” and would thus not be amenable to
dry handling, and also that EPA lacked data on whether legacy wastewater
could be “reliably incorporated” into a closed-loop process “given the variation
in operating practices among surface impoundments containing legacy
wastewater.” Id. at 67,854-55. The rule also asserts that EPA lacked sufficient
data to determine whether chemical or biological treatment would be effective
on legacy wastewater. Legacy wastewater, the agency explained, is often
“commingled”—meaning different streams are mixed together in an
impoundment—making testing and data collection difficult. Id. at 67,855. For
instance, commingling may result in varying the concentration and “flow rate”
of pollutants in an impoundment. Id. The rule acknowledges that multiple
plants are in fact using chemical precipitation to treat commingled
wastewater, but it nonetheless asserts that EPA lacks the requisite data from
those plants. Id. at 67,855 n.29. Finally, the rule also acknowledges that a few
plants discharge from impoundments containing non-commingled FGD legacy
wastewater, but it nonetheless declines to establish a stricter BAT for that
stream as well. Id. at 67,855. The rule explains that, in the agency’s view,
imposing the stricter technologies even on non-commingled legacy wastewater
would create bad “incentives”—for instance, encouraging plants to begin
commingling FGD with other wastewaters or to release FGD wastewater from
impoundments on an “accelerated schedule” prior to the compliance date. Id.
2. Leachate
The final rule describes leachate as follows:
Leachate includes liquid, including any suspended or dissolved
constituents in the liquid, that has percolated through or drained
from waste or other materials placed in a landfill, or that passes

15
 Case: 15-60821 Document: 00514914192 Page: 16 Date Filed: 04/12/2019

No. 15-60821
through the containment structure (e.g., bottom, dikes, berms) of a
surface impoundment.

80 Fed. Reg. 67,847. Where leachate occurs in a lined landfill or impoundment,
it is typically collected and transported to an impoundment, where it is either
“discharge[d] . . . directly to receiving waters” or recycled to another
impoundment prior to discharge. Id. Unlined landfills or impoundments
simply “allow the leachate to potentially migrate to nearby ground waters,
drinking water wells, or surface waters.” Id. The rule explains that “surface
impoundments are the most widely used systems to treat . . . leachate.” Id.
Elsewhere, the rule acknowledges that “[g]round water contamination from
surface impoundments” containing power plant wastewater “threatens
drinking water, as evidenced by more than 30 documented cases.” Id. at 67,840;
see also EPA Study Report, EPA 821-R-09-008, at 3-24 (landfill leachate
diagram). The EPA study detailed the size of leachate pollution: Given plants
using current technologies (mostly surface impoundments), leachate pollution
amounts to 70,300 toxic-weighted pound equivalents per year. See Technical
Development Document (“TDD”), EPA-821-R-15-007, at 10-39. Leachate thus
accounts for more equivalent pollution than the entire coal mining industry.
Id.; Annual Effluent Guidelines Review Report, EPA-821-R-16-002, at 2-26
(listing pollution from other industries).
The final rule sets BAT for leachate equal to the previous BPT standard
established in 1982. Id. at 67,854. The agency offers two primary justifications
for its decision not to regulate leachate with any of the more advanced control
technologies now available. First, the rule explains that EPA called for
comments on leachate regulation during notice-and-comment rulemaking, but
that “[c]ommenters did not provide information that the EPA could use to
establish BAT limitations” for leachate. Id. at 67,854. Second, the rule asserts
that leachate forms “a very small portion of the pollutants discharged
16
 Case: 15-60821 Document: 00514914192 Page: 17 Date Filed: 04/12/2019

No. 15-60821
collectively by all steam power plants.” Id. The agency reasons that, because
the new BAT limits established for wastewater from other streams will
substantially curtail total power plant pollution, the new rule “represents
reasonable further progress toward the CWA’s goals” even without
establishing any stricter controls on leachate. Id.
II. PROCEDURAL HISTORY
Four separate lawsuits challenging the final rule were originally brought
in the Second, Fifth, Eighth, and Ninth Circuits. 12 Different groups of
petitioners challenged different parts of the rule. Various power companies
(“Industry Petitioners”) challenged the regulation of non-legacy FGD and
gasification wastewater. 13 Two water company associations (“Water Company
Petitioners”), challenged the non-legacy FGD wastewater regulation. 14 Finally,
various environmental groups (“Environmental Petitioners” or “petitioners”)
challenged the regulation of legacy wastewater and leachate. 15 The four cases
were consolidated by the United States Judicial Panel on Multidistrict
Litigation and randomly assigned to our court. 16 The Utility Water Act Group

12 See 33 U.S.C. § 1369(b)(1)(E) (allowing “any interested person” to file in a federal
circuit court an application for review of the Administrator’s “promulgating any effluent
limitation” under §§ 1311 or 1314 within 120 days of the promulgation).
13 The Industry Petitioners are Utility Water Act Group (“UWAG”), Southwestern
Electric Power Company, Union Electric Company dba Ameren Missouri, City of Springfield,
Missouri, by and through the Board of Public Utilities, and Duke Energy Indiana, Inc.
14The Water Company Petitioners are the American Waterworks Association and the
National Association of Water Companies.
15The Environmental Petitioners are Waterkeeper Alliance, Inc., the Environmental
Integrity Project, and the Sierra Club.
16See United States Judicial Panel on Multidistrict Litigation, Order MCP No. 136
(December 8, 2015). The cases were originally captioned as: Waterkeeper Alliance, Inc. et al
v. EPA, et al. (2nd Cir. No. 15-3773); Southwestern Elec. Power Co., et al. v. EPA, et al. (5th
Cir. No. 15-60821); Union Elec. Co., et al. v. EPA, et al. (8th Cir. No. 15-3658), and Sierra
Club v. EPA. (9th Cir. No. 15-73578).
17
 Case: 15-60821 Document: 00514914192 Page: 18 Date Filed: 04/12/2019

No. 15-60821
(“UWAG”) has since intervened to defend those portions of the rule challenged
by the Environmental Petitioners.
In August 2017, we granted EPA’s motion to sever and hold in abeyance
the Industry Petitioners’ and Water Company Petitioners’ challenges to the
final rule. In September 2017, EPA announced it would reconsider the rule’s
regulations concerning non-legacy FGD and bottom ash transport water. See
82 Fed. Reg. 43,494. 17 As a result of these procedural developments, the
challenges to the final rule raised by the Industry Petitioners and the Water
Company Petitioners are not before us. We address only the challenges brought
by the Environmental Petitioners.
III. STANDARD OF REVIEW
The Environmental Petitioners challenge the legacy wastewater
regulation under the Administrative Procedure Act (“APA”). As relevant here,
a court “shall . . . hold unlawful and set aside” agency action under the APA if
it finds such action was “arbitrary, capricious, an abuse of discretion, or
otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). Under this “highly
deferential” standard, Avoyelles Sportsmen’s League, Inc. v. Marsh, 715 F.2d
897, 904 (5th Cir. 1983), we are “not empowered to substitute [our] judgment
for that of the agency.” Citizens to Preserve Overton Park, Inc. v. Volpe, 401
U.S. 402, 416 (1971). This is particularly so where the agency’s decision turns
on “its evaluation of complex scientific data within its technical expertise.”

17 The agency’s reconsideration of those aspects of the rule has been challenged in
separate lawsuits brought by a coalition of environmental groups, including some of the
petitioners in this case. That challenge was brought contemporaneously in the District of
Columbia federal district court and the D.C. Circuit. The district court ruled it lacked
jurisdiction, Clean Water Action v. Pruitt, 315 F. Supp. 3d 72, 85 (D.D.C. 2018), and the D.C.
Circuit transferred its case to this court, Clean Water Action v. Pruitt, No. 17-1216, Order
(D.C. Cir. Feb. 1, 2018), where it is now pending as No. 18-60079. The Environmental
Petitioners concede that the challenge to the EPA’s reconsideration decision involves distinct
issues that do not affect this case, and no party has asked us to postpone our decision here.
18
 Case: 15-60821 Document: 00514914192 Page: 19 Date Filed: 04/12/2019

No. 15-60821
BCCA Appeal Grp. v. EPA, 355 F.3d 817, 824 (5th Cir. 2003) (citing Baltimore
Gas & Elec. Co. v. NRDC, 462 U.S. 87, 2013 (1983)). Indeed, “[i]f the agency’s
reasons and policy choices conform to minimal standards of rationality, then
its actions are reasonable and must be upheld.” Tex. Oil & Gas Ass’n, 161 F.3d
at 934. Furthermore, the “EPA’s choice of analytical methodology [in setting
and enforcing standards] is entitled to a presumption of regularity,” leaving
challengers with a “considerable burden” to carry. Am. Petroleum Inst. v. EPA,
787 F.2d 965, 983 (5th Cir. 1986).
Our review under the APA is not toothless, however. We must set aside
agency action if the agency “entirely failed to consider an important aspect of
the problem, offered an explanation for its decision that runs counter to the
evidence before the agency, or is so implausible that it could not be ascribed to
a difference in view or the product of agency expertise.” Motor Vehicle Mfrs.
Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983)
(“State Farm”); see generally, e.g., Atchafalaya Basinkeeper v. U.S. Army Corps
of Eng’rs, 894 F.3d 692, 697 (5th Cir. 2018) (reciting State Farm standard).
“[W]e must also ensure that the agency ‘examine[d] the relevant data and
articulate[d] a satisfactory explanation for its action,’” and assess “‘whether
the [agency’s] decision was based on a consideration of the relevant factors[.]’”
10 Ring Precision, Inc. v. Jones, 722 F.3d 711, 723 (5th Cir. 2013) (quoting
State Farm, 463 U.S. at 43); see also, e.g., Michigan v. EPA, 135 S. Ct. 2699,
2706 (2015) (explaining that “agency action is lawful only if it rests ‘on a
consideration of the relevant factors’”) (quoting State Farm, supra); U.S.
Chamber of Commerce v. U.S. Dep’t of Labor, 885 F.3d 360, 382 (5th Cir. 2018)
(“Illogic and internal inconsistency are characteristic of arbitrary and
unreasonable agency action.”); Illinois Pub. Telecom. Ass’n v. FCC, 117 F.3d
555, 566 (D.C. Cir. 1997), decision clarified on reh’g, 123 F.3d 693 (unexplained
and “seemingly illogical” decisions are arbitrary and capricious). Furthermore,
19
 Case: 15-60821 Document: 00514914192 Page: 20 Date Filed: 04/12/2019

No. 15-60821
we “may uphold agency action only on the grounds that the agency invoked
when it took the action.” Michigan v. EPA, 125 S. Ct. at 2712 (citing SEC v.
Chenery Corp., 318 U.S. 80, 87 (1943)).
The Environmental Petitioners challenge the leachate regulation under
the two-step framework articulated in Chevron USA, Inc. v. Natural Resources
Defense Council, Inc., 467 U.S. 837 (1984), governing judicial review of agency
interpretations of statutes. See generally, e.g., BCCA Appeal Grp., 355 F.3d at
824 (discussing Chevron in context of challenge to Clean Air Act regulations).
At step one, the court considers “whether Congress has directly spoken to the
precise question at issue.” Chevron, 467 U.S. at 842. If Congress has directly
spoken on an issue, that settles the matter: “[T]he Court, as well as the agency,
must give effect to the unambiguously expressed intent of Congress.” Id. at
842-43. Only if the statutory text is ambiguous can the court proceed to step
two, asking whether the agency’s construction of the statute is “permissible.”
Id. at 843. If the construction is permissible, it should be upheld. “[A] court
may not substitute its own construction of a statutory provision for a
reasonable interpretation made by the administrator of an agency.” Id.
“Chevron review and arbitrary and capricious review overlap at the margins,”
specifically at Chevron step two. Indep. Petroleum Ass’n of Am. v. Babbitt, 92
F.3d 1248, 1258 (D.C. Cir. 1996); see also Nutraceutical Corp. v. Von
Eschenbach, 459 F.3d 1033, 1038 (10th Cir. 2006).

20
 Case: 15-60821 Document: 00514914192 Page: 21 Date Filed: 04/12/2019

No. 15-60821
IV. ANALYSIS
A. Challenge to Legacy Wastewater Regulation
We first consider petitioners’ challenge 18 to the final rule’s regulation of
legacy wastewater. As already explained, legacy wastewater is not a separate
wastewater stream but instead a subset of five of the other streams. Supra
I.B.1. Specifically, legacy wastewater is defined by when wastewater is
generated: Wastewater “generated prior to” the compliance date set for the new
rule by a permitting authority is denominated “legacy” wastewater. See 80 Fed.
Reg. 67,854. Instead of subjecting legacy wastewater to the more advanced and
effective technologies that kick in after the rule’s compliance date (i.e.,
chemical precipitation, biological treatment, dry handling, or evaporation), the
rule sets BAT for legacy wastewater as equal to the BPT previously set in 1982
(i.e., surface impoundments). See id. at 67,854-55.

18 While no party contests the issue, we conclude that petitioners have associational
standing to challenge the final rule on behalf of their members. See, e.g., La. Landmarks Soc.,
Inc. v. City of New Orleans, 85 F.3d 1119, 1122 n. 3 (5th Cir. 1996) (explaining that “standing
is jurisdictional and, therefore, non-waivable”); Assoc. of Am. Phys. & Surgeons v. Tex. Med.
Bd., 627 F.3d 547, 550 (5th Cir. 2010) (associational standing present when (1) members
would have standing; (2) association seeks to protect interests germane to its purpose; and
(3) neither claim nor relief requires individual participation). Members of the petitioner
organizations attest in declarations to cognizable injuries-in-fact traceable to the discharges
at issue—such as negative impact on property and decrease in enjoyment of waterways—
that could be redressed by a decision requiring reevaluation of the rule. See, e.g., Friends of
the Earth, Inc. v. Laidlaw Envtl. Servs. (TOC), Inc., 528 U.S. 167, 183 (2000) (sufficient
injury-in-fact when plaintiffs “aver that they use the affected area and are persons for whom
the aesthetic and recreational values of the area will be lessened by the challenged activity”);
Sierra Club, Lone Star Chapter v. Cedar Point Oil Co. Inc., 73 F.3d 546, 557 (5th Cir. 1996)
(traceability established by allegations that pollutants “cause[ ] or contribute[ ] to the kinds
of injuries alleged by the plaintiffs”). Finally, we find the second and third prongs of
associational standing test are met here because petitioners seek to protect environmental
interests germane to their purposes and individual participation by each affected member is
unnecessary. Assoc. of Am. Phys. & Surgeons, 627 F.3d at 550.
21
 Case: 15-60821 Document: 00514914192 Page: 22 Date Filed: 04/12/2019

No. 15-60821
1.
Petitioners challenge that decision as arbitrary and capricious under the
APA on two grounds. First, they claim the Act does not grant the Administrator
authority to base BAT limits on when waste is generated, but instead requires
setting BAT limits for “categories and classes of point sources” regardless of
when waste is generated. See, e.g., State Farm, 463 U.S. at 43 (explaining that,
“[n]ormally, an agency rule would be arbitrary and capricious if the agency has
relied on factors which Congress has not intended it to consider”). Second, they
claim that the rulemaking record refutes the conclusion that surface
impoundments are BAT for legacy wastewater because, among other things,
the final rule itself demonstrates that impoundments are ineffective at
removing toxic pollutants. See, e.g., id. (an agency rule would also be arbitrary
and capricious if it “offered an explanation for its decision that runs counter to
the evidence before the agency”). We need not reach petitioners’ first argument,
because we conclude for multiple reasons that EPA acted arbitrarily and
capriciously by setting a BAT limit for legacy wastewater equal to the outdated
BPT standard of surface impoundments.
First, the final rule repeatedly recognizes that impoundments are
“largely ineffective” at removing toxins from wastewater. 80 Fed. Reg. 67,840.
Impoundments “rely on gravity to remove particulates from wastewater,” but
the rule explains that “gravity in surface impoundments” fails to “effectively
or reliably” remove “[p]ollutants . . . present mostly in soluble (dissolved) form,
such as selenium, boron, and magnesium,” and also fails to “effectively” remove
the dissolved portion of “metals present in both soluble and particulate forms
(such as mercury).” Id. at 67,851. When ingested by humans either through
drinking water or through seafood, these metals can lead to serious harms
including “cancer, cardiovascular disease, neurological disorders, kidney and
liver damage, and lowered IQs in children.” Id. at 67,840. Additionally, the rule
22
 Case: 15-60821 Document: 00514914192 Page: 23 Date Filed: 04/12/2019

No. 15-60821
informs us that various factors can alter chemical conditions in impoundments
and thus compromise their effectiveness: For instance, low pH in the
impoundment environment can convert particulate metals to soluble form,
reducing the “settling efficiency in the impoundments” and “leading to
increased levels of dissolved metals and high concentrations of metals in
discharges from surface impoundments.” Id. at 67,851. Even the changing
seasons—in an effect called “seasonal turnover”—impair impoundments by
cooling the upper layer of water and causing it to sink, resulting in
“resuspension of solids . . . and a consequent increase in the concentrations of
pollutants discharged from the impoundment.” Id.
These conceded defects in impoundments are in critical tension with
EPA’s choosing them as BAT for legacy wastewater. After all, BAT is supposed
to be “the CWA’s most stringent standard” for setting discharge limits for
existing sources. Tex. Oil & Gas, 161 F.3d at 928; see also 33 U.S.C.
§§ 1311(b)(2), 1314(b)(2). We are rightfully skeptical when EPA specifies
impoundments as BAT while, in the same breath, detailing how bad those
impoundments are in stemming the discharge of toxic pollution. See, e.g., CMA,
885 F.2d at 265 (remanding because EPA “failed . . . [to] demonstrate a
reasonable basis for its conclusion” that its chosen BAT was as effective as a
proposed alternative).
Second, as the rule also recounts, the flaws of impoundments are
precisely why EPA refused to set them as BAT for five of the six wastewater
streams at issue here. See 80 Fed. Reg. 67,851-53. For instance, the rule states
that “EPA did not select surface impoundments as the BAT technology for FGD
wastewater because it would not result in reasonable further progress toward
eliminating the discharge of all pollutants, particularly toxic pollutants.” Id. at
67,851 (emphasis added); see also id. (explaining that EPA declined to set
impoundments as BAT “[b]ecause many of the pollutants of concern in FGD
23
 Case: 15-60821 Document: 00514914192 Page: 24 Date Filed: 04/12/2019

No. 15-60821
wastewater are present in dissolved form and would not be removed by surface
impoundments”) (emphasis added). EPA likewise declined to set
impoundments as BAT for fly ash transport water, bottom ash transport water,
FGMC wastewater, and gasification water, and in each case explained that it
did so “for the same reasons . . . that EPA did not identify surface
impoundments as BAT for FGD wastewater.” Id. at 67,852 (fly ash), 67,853
(bottom ash, FGMC, and gasification wastewater).
In other words, for five of the six wastewater streams regulated by the
final rule (the one exception is leachate, discussed in V.B infra), EPA
affirmatively rejected surface impoundments as BAT “because [they] would not
result in reasonable further progress toward eliminating the discharge of all
pollutants, particularly toxic pollutants.” Id. at 67,851. 19 And yet, having
rejected impoundments as BAT because they would not achieve “reasonable
further progress” toward eliminating pollution from those streams, EPA
turned around and chose impoundments as BAT for each of those same streams
generated before the compliance date. That paradoxical action signals
arbitrary and capricious agency action. See, e.g., Chamber of Commerce, 885
F.3d at 382 (“Illogic and internal inconsistency are characteristic of arbitrary
and unreasonable agency action.”); see also, e.g., GameFly, Inc. v. Postal
Regulatory Comm’n, 704 F.3d 145, 148 (D.C. Cir. 2013) (explaining that agency
action “‘illogical on its face’” may be arbitrary and capricious) (quoting Am.
Fed’n of Gov’t Emps., Local 2924 v. Fed. Labor Relations Auth., 470 F.3d 375,
380 (D.C. Cir. 2006)). It also strongly suggests that EPA has contravened the

19 Indeed, for treatment of gasification wastewater the inadequacies of surface
impoundments appear even more pronounced. EPA found that one of the three existing U.S.
plants that produce gasification wastewater previously used surface impoundments but that
“the impoundment effluent repeatedly exceeded its NPDES permit effluent limitations
necessary to meet applicable [Water Quality Standards].” Id. at 67,853.
24
 Case: 15-60821 Document: 00514914192 Page: 25 Date Filed: 04/12/2019

No. 15-60821
plain language of the CWA, which defines BAT as the technology that “will
result in reasonable further progress” toward pollutant discharge elimination.
33 U.S.C. § 1311(b)(2)(A) (emphasis added); see also, e.g., Nat’l Crushed Stone,
449 U.S. at 74-75 (discussing “reasonable further progress” component of BAT).
Third, the final rule explains that the shortcomings of surface
impoundments were a key factor in motivating EPA to conduct the 2006-2009
study and revise water pollution regulations for power plants in the first place.
The rule describes the previous ELGs from 1974, 1977, and 1982 as “out of
date,” because they failed to “adequately control the pollutants . . . discharged
by this industry” and failed to “reflect relevant process and technology
advances that have occurred in the last 30-plus years.” 80 Fed. Reg. 67,840; see
also id. (stating that, “[i]n the several decades since the steam electric ELGs
were last revised, [more effective] technologies have increasingly been used at
plants”). And the rule minces no words in laying the shortcomings of the prior
ELGs at the feet of surface impoundments:
The processes employed and pollutants discharged by the industry
look very different today than they did in 1982. Many plants,
nonetheless, still treat their wastewater using only surface
impoundments, which are largely ineffective at controlling
discharges of toxic pollutants and nutrients.

Id.
Thus, the final rule describes impoundments as an outdated and
ineffective pollution control technology, and yet the same rule chooses to freeze
impoundments in place as BAT for legacy wastewater. That is inconsistent
with the “technology-forcing” mandate of the CWA. NRDC II, 808 F.3d at 563-
64 (citing NRDC I, 822 F.2d at 123). To that point, the Supreme Court has
explained that BAT has an inbuilt “reasonable further progress” standard and
that “BPT serves as the prior standard with respect to BAT.” Nat’l Crushed
Stone, 449 U.S. at 75. Yet here EPA appears to have conflated the prior
25
 Case: 15-60821 Document: 00514914192 Page: 26 Date Filed: 04/12/2019

No. 15-60821
standard with the advanced one: It has selected as BAT the same three-
decades-old technology previously set as BPT—a technology the current rule
condemns as anachronistic and ineffective at eliminating pollution discharge.
In other words, EPA asks us to believe that impoundments are both archaic
and cutting-edge at the same time. That we cannot do. See GameFly, 704 F.3d
at 148; Chamber of Commerce, 885 F.3d at 382.
Fourth, the final rule strongly indicates that other available technologies
are far better than impoundments at removing pollutants from the various
streams that comprise legacy wastewater. For instance, after explaining why
impoundments are ineffective at removing toxic metals from FGD wastewater,
the rule states that a combination of chemical precipitation and biological
treatment is better at removing those pollutants. Id. at 67,850-51. Importantly,
the rule explicitly concludes that “[c]hemical precipitation and biological
treatment are more effective than surface impoundments at removing both
soluble and particulate forms of metals.” Id. at 67,851 (emphasis added). The
rule also relies on that reasoning to justify rejecting impoundments as BAT for
fly ash transport water, bottom ash transport water, FGMC wastewater, and
gasification wastewater. Id. at 67,852-53. Moreover, the rule categorically
states that more advanced control methods, such as chemical and biological
methods, “are affordable technologies that are widely available, and already in
place at some plants.” Id. at 67,840.
These affirmative findings are difficult, if not impossible, to square with
EPA’s decision nonetheless to set 1980s-era impoundments as the BAT for
legacy wastewater. To be sure, the agency’s statements in the final rule do not
prove that chemical precipitation or biological treatment (or some combination
of the two) are BAT for legacy wastewater. That is for the agency to decide. But
those statements do cast grave doubt on the agency’s selection of
impoundments as BAT for legacy wastewater. See, e.g., Nat'l Crushed Stone
26
 Case: 15-60821 Document: 00514914192 Page: 27 Date Filed: 04/12/2019

No. 15-60821
Ass’n, 449 U.S. at 74 (BAT calls for the “maximum use of technology within the
economic capability of the [plant] owner or operator”). Once again, the EPA’s
own rule strongly indicates that it was arbitrary and capricious in doing so. 20
Fifth and finally, our court has long recognized that “‘Congress intended
[BAT] limitations to be based on the performance of the single best-performing
plant in an industrial field.’” Tex. Oil & Gas, 161 F.3d at 928 (quoting CMA,
870 F.2d at 226)); see also, e.g., Kennecott v. EPA, 780 F.2d 445 (4th Cir. 1985)
(“In setting BAT, EPA uses not the average plant, but the optimally operating
plant, the pilot plant which acts as a beacon to show what is possible.”) (citing
A Legislative History of the Water Pollution Control Act Amendments of 1972,
93d Cong., 1st Sess. (Comm. Print 1973), at 798). Yet here the rule says nothing
to indicate that the choice of impoundments as BAT for legacy wastewater was
based on anything like “the performance of the single best-performing plant”
in the field. To the contrary, everything the rule says about the record of
impoundments over the past three decades indicates that their performance in
controlling discharges has been distinctly poor. See, e.g., 80 Fed. Reg. 67,840
(stating that “impoundments . . . are largely ineffective at controlling
discharges of toxic pollutants and nutrients”); id. (stating that “[g]round water
contamination from surface impoundments . . . threatens drinking water, as

20 Intervenor UWAG argues that “EPA’s acknowledgment that other technologies are
better at removing dissolved metals does not undermine the Agency’s conclusion that surface
impoundments reflect BAT for [pollutants] in legacy wastewaters.” We disagree. In this
context, the EPA’s pointed criticisms of impoundments fatally undermine its BAT
determination. To be sure, an agency has “some leeway reasonably to resolve uncertainty, as
a policy matter, in favor of more regulation or less.” Ctr. for Auto Safety v. Fed. Highway
Admin., 956 F.2d 309, 316 (D.C. Cir. 1992). But there is no uncertainty in this record about
the shortcomings of impoundments. EPA may have been uncertain about what the precise
BAT for legacy wastewater should be, but the record fails to explain why impoundments are
BAT, if that term is to have any meaning. Furthermore, if chemical precipitation or biological
treatment are technically feasible but simply too costly for treating legacy wastewater, the
EPA could have said so. It did not. We cannot defend the agency’s action based on a rationale
the agency did not rely on. State Farm, 463 U.S. at 43.
27
 Case: 15-60821 Document: 00514914192 Page: 28 Date Filed: 04/12/2019

No. 15-60821
evidenced by more than 30 documented cases”); id. at 67,851 (declining to set
impoundments at BAT for FGD wastewater because various dissolved toxic
metals “are not effectively and reliably removed by gravity in surface
impoundments”). Moreover, the rule also states that multiple plants are in fact
treating legacy wastewater using chemical precipitation, id. at 67,855 n.29, a
method the rule concedes is “more effective than surface impoundments at
removing both soluble and particulate forms of metals[.]” Id. at 67,851. Yet the
rule merely states—without explanation—that it lacks “data to characterize
the effluent from these systems.” Id. at 67,855 n.29. That unexplained
assertion casts grave doubt on the agency’s BAT decision. See also infra V.A.2
(discussing additional problems created by agency’s lack of data excuse).
These shortcomings in the agency’s explanations strongly indicate that
its BAT decision simply defaults to the outdated BPT standard that has been
demonstrated to be a poor performer by the agency’s own analysis. That is
antithetical to the statutorily-mandated BAT standard. See, e.g., Nat’l Lime
Ass’n v. EPA, 233 F.3d 625, 634 (D.C. Cir.), as amended on denial of reh’g (2001)
(remanding EPA rule for failure in agency’s “clear statutory obligation to set
emission standards” for various air pollutants); NRDC v. EPA, 863 F.2d 1420,
1433 (9th Cir. 1988) (despite EPA’s asserted lack of “complete information” on
availability of technology, declaring BAT limitation invalid because “Congress
has demonstrated its intent to require industry to do as much as possible to
control toxic discharges”) (citing 33 U.S.C. § 1311(b)(2)(A)(i)).
In sum, we conclude that the EPA’s decision to set surface
impoundments as BAT for legacy wastewater was arbitrary and capricious.
Far from demonstrating that impoundments are the “best available technology
economically achievable” for treating legacy wastewater, the evidence
recounted in the final rule shows that impoundments are demonstrably
ineffective at doing so and demonstrably inferior to other available
28
 Case: 15-60821 Document: 00514914192 Page: 29 Date Filed: 04/12/2019

No. 15-60821
technologies. In light of this record, we cannot accept that an outdated,
ineffective and inferior technology is BAT when applied to legacy wastewater.
No record evidence affirmatively makes that case and, as we have explained,
the evidence recounted in the final rule runs in the opposite direction.
2.
EPA defends its choice of impoundments as BAT for legacy wastewater
by asserting that “it does not have the data” to justify choosing more advanced
pollution control technologies. 80 Fed. Reg. 67,855. The agency explains that
most plants “combine some of their legacy wastewater with each other and
with other wastestreams,” and that this “commingling” can meaningfully alter
the characteristics (specifically, the “flow rate and pollutant concentration”) of
the impoundment water. Id. Because EPA lacked adequate examples of plants
treating commingled wastewater “using anything beyond the surface
impoundment itself,” the agency concluded it lacked data to evaluate the
performance of other technologies and therefore defaulted to “the previously
promulgated BPT limitations” (i.e., impoundments). Id. We are unpersuaded.
First, EPA’s arguments about the characteristics of commingled
wastewater glide past the key issue before us, which is whether the agency
arbitrarily chose impoundments as BAT. The agency may lack data on how
other technologies interact with commingled wastewater, but it assuredly does
not lack data on impoundments. To the contrary, we know that impoundments
are ineffective at removing toxic pollutants from the various wastewater
streams because the agency’s own rule tells us so, repeatedly, based on over
three decades of observation and analysis. See supra V.B.1. Nor does EPA’s
criticism of impoundments distinguish “legacy” from “non-legacy” wastewater:
instead, the agency categorically states that “surface impoundments . . . are
largely ineffective at controlling discharges of toxic pollutants and nutrients”
from “wastewater.” 80 Fed. Reg. 67,840. To be sure, we do not pretend to
29
 Case: 15-60821 Document: 00514914192 Page: 30 Date Filed: 04/12/2019

No. 15-60821
second-guess EPA’s assertions about the pertinent “flow rate” and “pollutant
concentration” in commingled wastewater, id. at 67,855, matters beyond our
expertise and authority. But those assertions side-step the legal issue we must
decide, which is whether the agency was arbitrary in selecting impoundments
as BAT. 21 Once again, the agency’s own pointed criticism of impoundments
indicates that it was.
Second, the agency’s “lack of data” excuse is untenable on its own terms.
In a footnote, the rule concedes that multiple power plants have in fact been
using chemical precipitation to treat commingled legacy wastewater. See id. at
67,855 n.29 (stating “EPA identified fewer than ten plants that use chemical
precipitation to treat waster that contains, among other things, ash transport
water”). Yet, the agency baldly asserts that it “does not have any data to
characterize the effluent from these systems” and it raises this dearth of
information to justify not regulating legacy wastewater under the same BAT
standards as non-legacy wastewater (and to justify a demonstrably outdated
technology as BAT). Id. We have previously rejected EPA’s argument that an
asserted lack of “sufficient data” justified the agency’s failure to regulate. See
API I, 661 F.2d at 357 (rejecting EPA’s argument that its failure to regulate
was justified by lack of “sufficient data” where EPA had failed to investigate
“in light of . . . new information” (internal quotation marks omitted)); see also,
e.g., NRDC II, 808 F.3d at 573 (concluding that EPA’s failure to gather data

21 Intervenor UWAG offers a similarly mistaken defense of the rule by arguing that
the timing of wastewater generation “profoundly influence[s] the amount of wastewater, the
characteristics of that wastewater, and the technologies available to treat it.” This may be
true, but the question before us is not whether legacy wastewater must be regulated in the
same manner as other wastewater. The question is whether impoundments are BAT for
legacy wastewater, and the agency’s own words cast grave doubt on that conclusion.
Additionally, UWAG’s argument that treating legacy wastewater as a “separate
wastestream” was not itself arbitrary and capricious is beside the point. Our criticism is
directed at the BAT determination the agency actually reached for legacy wastewater, not at
the decision to establish a separate BAT in the first place.
30
 Case: 15-60821 Document: 00514914192 Page: 31 Date Filed: 04/12/2019

No. 15-60821
can be arbitrary and capricious when the lack of data is “a problem of EPA’s
own making”); see also infra V.B.2 (further discussing API I and NRDC II). We
reject the argument again here. The final rule recounts (1) the long-recognized
deficiencies of impoundments in controlling toxic discharges, 80 Fed. Reg.
67,840; (2) the demonstrated superiority of more advanced technologies in
doing so, id. at 67,851-53; (3) the availability of those technologies in the
industry, id. at 67,840, 67,844; and (4) multiple plants’ actual use of one of
those advanced technologies (chemical precipitation) to treat commingled
legacy wastewater, id. at 67,855 n.29. Given those undisputed statements
drawn from EPA’s own rule, the agency cannot simply plead a lack of data to
justify its decision to set impoundments as BAT. Again, we do not purport to
tell the agency what technology it should choose as BAT for legacy wastewater.
We decide only that, given the agency’s own statements and evidence, it acted
arbitrarily in selecting as BAT a pollution control method that decades of data
have shown to be ineffective at controlling pollution.
Third, given EPA’s heavy reliance on the characteristics of commingled
legacy wastewater as a reason for declining stricter regulation, one would
expect a different policy for non-commingled legacy wastewater. See 80 Fed.
Reg. 67,855 & n.28 (asserting as key reason for defaulting to impoundments
for legacy wastewater the fact that wastewater at the “vast majority” of plants
is “commingled” with other streams). Yet we find the opposite: When EPA
identified plants that discharge non-commingled legacy wastewater from
impoundments, it still declined to impose more stringent controls and still
defaulted to impoundments as BAT. See id. at 67,855 (declining to impose
controls “other than surface impoundments” on plants that “discharge from an
impoundment containing only legacy FGD wastewater”); id. at 67,855 n.30
(discussing three plants that use impoundments “where the FGD wastewater
is not commingled with other process wastewaters in the impoundment”)
31
 Case: 15-60821 Document: 00514914192 Page: 32 Date Filed: 04/12/2019

No. 15-60821
(emphases added). That striking inconsistency undercuts the agency’s
“commingling” rationale for not imposing the more stringent non-legacy BAT
standards. See, e.g., Am. Fed’n of Gov’t Emps., Local 2924 v. Fed. Labor
Relations Auth., 470 F.3d 375, 380 (D.C. Cir. 2006) (an agency’s decision is
“arbitrary and capricious” if “illogical on its own terms”); see also Chamber of
Commerce, 885 F.3d at 382 (“Illogic and internal inconsistency are
characteristic of arbitrary and unreasonable agency action.”). 22 What is more,
by selecting impoundments as BAT for any kind of FGD wastewater (“legacy”
or not), the rule flatly contradicts itself: The rule states without exception that
for FGD wastewater “[c]hemical precipitation and biological treatment are
more effective than surface impoundments at removing” toxic pollutants. 80
Fed. Reg. 67,851.
Fourth, even assuming a lack of data prevented EPA from determining
BAT for legacy wastewater, nothing required the agency simply to set
impoundments as BAT. Instead, EPA could have declined to set nationwide

22 The final rule attempts to explain away this paradox, but its rationales do not hold
water. The rule asserts that, if EPA subjected non-commingled legacy wastewater to stricter
controls, plants “would begin commingling other process wastewater with their legacy [non-
commingled] wastewater” and thereby circumvent the regulation. Id. at 67,855. Perhaps or
perhaps not. But the current rule already permits plants to do precisely what the agency is
concerned about—i.e., mix wastewater with non-commingled wastewater. And if EPA is
seriously concerned that more commingling would result from a stricter rule, it could avoid
the problem by restricting commingling at plants where it was not already an established
process. Additionally, the agency contends that stricter regulation for non-commingled legacy
wastewater would create the harmful “incentive” for plants to discharge the wastewater “on
an accelerated schedule” that “could result in temporary increases in environmental
impacts.” Id. But the agency fails to consider that such action would, at a minimum, require
plants to seek modification of the NPDES permits. See 40 C.F.R. §§ 122.41(l)(1)(ii) (requiring
permitted facilities to report any alteration in operations that “could significantly change the
nature or increase the quantity of pollutants discharged”); 122.62(a)(1) (requiring
modification of permits for “material and substantial alterations or additions to the permitted
facility or activity (including a change or changes in the permittee’s . . . disposal practice”).
In any event, the central point remains: by including non-commingled legacy wastewater
within its ambit, the final rule undercuts the key rationale supporting it—that EPA lacks
sufficient data for commingled legacy wastewater.
32
 Case: 15-60821 Document: 00514914192 Page: 33 Date Filed: 04/12/2019

No. 15-60821
effluent guidelines for legacy wastewater and allowed BAT determinations to
be made by each facility’s permitting authority through the NPDES permitting
process on a site-specific basis. See 40 C.F.R. § 125.3(a), (c)(2) (“Technology
based treatment requirements may be imposed . . . [o]n a case-by-case basis.”);
Riverkeeper, Inc. v. EPA, 358 F.3d 174, 203 (2nd Cir. 2004) (“We see no textual
bar in sections 306 or 316(b) [of the Clean Water Act] to regulating [certain]
structures on a case-by-case basis.”); Nat’l Wildlife Fed’n v. EPA, 286 F.3d 554,
566-67 (D.C. Cir. 2002) (“We believe EPA acted both reasonably and within its
authority in adopting a case-by-case approach” to regulating certain pollutants
from paper mills). The agency took that approach in the current rule by
deferring setting BAT and other limits for metal cleaning wastes, after
determining it lacked the necessary data. See 80 Fed. Reg. 67,863 (directing
permitting authorities to “establish such requirements based on [best
professional judgment] for any steam electric power plant discharg[ing]” such
waste). Instead of deferring a nationwide effluent guideline and allowing case-
by-case determination of BAT by permitting authorities, EPA unaccountably
defaulted to impoundments—again, which its own rule recognizes as an out-
of-date and ineffective pollution control technology. This is further indication
that the rule respecting legacy wastewater is arbitrary and capricious. 23
***

23 EPA claims that petitioners have waived this deferral argument by failing to raise
it during the notice-and-comment period. That argument is foreclosed by our precedent. See
Am. Forest & Paper Ass’n v. EPA, 137 F.3d 291, 295 (5th Cir. 1998) (“EPA has failed to
identify any provision in the CWA that suggests a party’s failure to comment waives its right
to seek judicial review.”) (citing City of Seabrook v. EPA, 659 F.2d 1349, 1360 n.17). Our
waiver precedents in this area are admittedly in conflict. See BCCA Appeal Grp., 355 F.3d at
829 (acknowledging conflict); Tex. Oil & Gas Ass’n, 161 F.3d 923, 933 n.7 (finding waiver due
to “failure to raise the objections during the notice and comment period.”). We must follow
the earlier precedent, however, which directly refutes the agency’s waiver argument. When
precedents conflict, “under our rule of orderliness, the earlier case controls.” GlobeRanger
Corp. v. Software AG United States of Am., Inc., 836 F.3d 477, 497 (5th Cir. 2016).
33
 Case: 15-60821 Document: 00514914192 Page: 34 Date Filed: 04/12/2019

No. 15-60821
In sum, having examined the various justifications set forth for EPA’s
final rule on legacy wastewater, and finding each of those explanations
wanting in light of the agency record, we conclude that EPA’s rulemaking was
arbitrary and capricious. We therefore set aside that part of the final rule and
remand to the agency for reconsideration. See, e.g., Perdue v. FAA, 172 F.3d
866 (5th Cir. 1999) (“This court shall set aside agency action that is ‘arbitrary,
capricious, an abuse of discretion, or otherwise not in accordance with law.”
(quoting 5 U.S.C. § 706(2)(A))); Checkosky v. SEC, 23 F.3d 452, 491 (D.C. Cir.
1994) (“Section 706(2)(A) [of the APA] provides that a ‘reviewing court’ faced
with an arbitrary and capricious agency decision ‘shall’—not may—‘hold
unlawful and set aside’ the agency action.”).
We recognize that the agency is entitled to considerable deference in
setting BAT limitations. See, e.g., BCCA Appeal Grp., 355 F.3d at 824 (EPA
entitled to special deference where its decision turns on “its evaluation of
complex scientific data within its technical expertise”); Tex. Oil. & Gas, 161
F.3d at 928 (recognizing EPA’s “considerable discretion” in weighing BAT
factors). Precisely for that reason, challenges to the agency’s BAT
determinations often fail because challengers ask the court to elevate itself as
an expert over the agency. 24 This case is different. We do not question the

24 See, e.g., Tex. Oil & Gas Ass’n, 161 F.3d at 923, 934-35 (rejecting challenge based on
argument that EPA only “paid lip service to the age factor [in setting BAT],” because there
was still a clear “rational relationship” between the factors considered and the BAT decision);
id. at 935-36 (rejecting another challenge based on argument that EPA relied on a flawed
scientific study, because the study ultimately “had nothing to do with either the BAT
determination or the actual inclusion of a zero discharge limit on produced water in the
[guidelines]); API v. EPA, 787 F.2d at 983 (rejecting challenge based on assertion that a
scientific test relied on by the EPA was “unproven and unreliable,” when evidence suggested
test was reliable); CMA, 885 F.2d at 262 (rejecting challenge on mathematical grounds to
EPA’s statistical methodology because that methodology “was within [the agency’s] broad
discretion in the choice of statistical techniques”); id. at 262-63 (separately rejecting
argument that EPA did not properly consider the characteristics of diluted wastestreams in
its studies leading up to BAT determination).
34
 Case: 15-60821 Document: 00514914192 Page: 35 Date Filed: 04/12/2019

No. 15-60821
scientific or statistical methodologies relied upon by EPA, nor second-guess its
weighing of the statutory factors. Instead, we rely on EPA’s own scientific
conclusions in the rule itself to conclude that its choice of an outdated and
ineffective technology as BAT was unlawful under the Act. That is a legal—not
a technical or scientific—conclusion that the APA requires us to make.
B. Challenge to Leachate Regulation
We turn to the challenge to the rule’s regulation of combustion residual
leachate. As explained, supra I.B.2, leachate consists of liquid that percolates
through a landfill or impoundment and is eventually discharged into water.
See 80 Fed. Reg. 67,847. The final rule sets BAT for leachate as impoundments,
which is the same as the previous BPT for leachate established in 1982. See id.
at 67,854 (“This rule identifies surface impoundments as the BAT technology
basis for control of pollutants in combustion residual
leachate . . . establish[ing] a BAT limitation on [total suspended solids]
in . . . leachate equal to the previously promulgated BPT limitation on [total
suspended solids] in low volume waste sources.”). EPA offers two justifications
for this decision: First, that “[c]ommenters did not provide information that
EPA could use to establish [stricter] BAT limitations” for leachate, and, second,
that because leachate forms a “very small portion” of overall discharges, the
rule’s stricter regulation of other wastestreams “already represents reasonable
further progress towards the CWA’s goals.” See id.; supra I.B.2.
The Environmental Petitioners challenge the rule’s leachate regulation
under the Chevron test for reviewing agency interpretations of statutes. See
Chevron, 467 U.S. 837; see also generally, e.g., Acosta v. Hensel Phelps Constr.
Co., 909 F.3d 723, 730 (5th Cir. 2018) (summarizing “the two-step framework
established in Chevron”); RICHARD J. PIERCE, JR., ADMINISTRATIVE LAW
TREATISE § 3.2 (“PIERCE”) (discussing “[t]he Chevron Two-Step”). Petitioners
assert that EPA’s decision to set impoundments as BAT for leachate fails
35
 Case: 15-60821 Document: 00514914192 Page: 36 Date Filed: 04/12/2019

No. 15-60821
Chevron step one by contravening the plain text and structure of the Clean
Water Act. Alternatively, petitioners argue that the agency’s decision fails
Chevron step two by adopting an impermissible construction of the Act. We
address each argument in turn.
1.
At Chevron step one, we ask whether the pertinent statute
“unambiguously foreclose[s]” the agency’s challenged statutory interpretation;
if it does, “‘that is the end of the matter; for the court, as well as the agency,
must give effect to the unambiguously expressed intent of Congress.’” U.S.
Chamber of Commerce v. U.S. Dep’t of Labor, 885 F.3d 360, 369 (5th Cir. 2018)
(quoting Chevron, 467 U.S. at 842-43); see also, e.g., Acosta, 909 F.3d at 730 (at
Chevron step one, the court “must determine ‘whether Congress has directly
spoken to the precise question at issue’” (quoting City of Arlington v. FCC, 569
U.S. 290, 296 (2013)). To answer this question, we rely on “the conventional
standards of statutory interpretation”—i.e., “text, structure, and the overall
statutory scheme”—as well as “authoritative Supreme Court decisions.” U.S.
Chamber of Commerce, 885 F.3d at 369 (citing City of Arlington, 569 U.S. at
296; Chevron, 467 U.S. at 843 n.9). We are not to focus myopically on “a
particular statutory provision in isolation” because “[t]he meaning—or
ambiguity—of certain words or phrases may only become evident when placed
in context.” FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 132
(2000). Rather, we must read words “‘in their context,’” interpreting the statute
“‘as a symmetrical and coherent regulatory scheme,’” and “‘fit[ting], if possible,
all [the statute’s] parts into an harmonious whole.’” Id. at 133 (quoting Davis
v. Mich. Dep’t of Treasury, 489 U.S. 803, 809 (1989); Gustafson v. Alloyd Co.,
513 U.S. 561, 569 (1995); FTC v. Mandel Bros., Inc., 359 U.S. 385, 389 (1959))
(brackets added). Additionally, “we must be guided to a degree by common
sense as to the manner in which Congress is likely to delegate a policy decision
36
 Case: 15-60821 Document: 00514914192 Page: 37 Date Filed: 04/12/2019

No. 15-60821
of such economic and political magnitude to an administrative agency.” Texas
v. United States, 787 U.S. 733, 760 (5th Cir. 2015) (quoting Brown &
Williamson, 529 U.S. at 133). The goal of our Chevron step one inquiry is
ultimately to “ascertain whether the statute is silent or ambiguous in
addressing the precise question at issue.” Tex. Savings & Cmty. Bankers Ass’n
v. Fed. Hous. Fin. Bd., 201 F.3d 551, 554 (5th Cir. 2000); see also PIERCE § 3.6
(“The question for the court [at step one] is whether the agency’s construction
of the language is within the range of meanings that could be plausibly
attributed to the relevant statutory language.”).
Petitioners’ step one attack targets the agency’s justifications for pegging
leachate BAT to the same technology set as BPT in 1982—specifically, EPA’s
explanation that leachate forms a “very small portion” of collective industry
discharges and that the rule’s stricter BAT for other wastestreams represents
reasonable overall progress in the industry. See 80 Fed. Reg. 67,854.
Petitioners’ various arguments may be grouped into three general categories
for convenience of analysis. First, petitioners argue the agency’s rationales
contravene the CWA’s text—requiring a BAT to eliminate discharges of “all
pollutants” if “technologically and economically achievable,” 33 U.S.C.
§ 1311(b)(2)(A) (emphasis added)—because, here, the agency has chosen not to
regulate a significant pollution source (leachate) with a technology EPA admits
is achievable (chemical precipitation). Second, petitioners argue the agency’s
decision on the proper control standard for leachate conflates BAT with BPT
in a way that contravenes the structure of the Act. Finally, petitioners claim
that the agency’s proffered justifications for its leachate regulation appear
nowhere in the factors mandated for determining BAT and, indeed, contradict
those factors. See 33 U.S.C. § 1314(b)(2)(B) (listing BAT factors).
Petitioners’ initial step one argument places too much weight on the
phrase “all pollutants” in § 1311(b)(2)(A). We cannot agree that this statutory
37
 Case: 15-60821 Document: 00514914192 Page: 38 Date Filed: 04/12/2019

No. 15-60821
phrase, standing alone, squarely forecloses the agency’s decision to maintain
leachate BAT at the 1982 BPT standard. To the contrary, we agree with EPA
(and with binding precedent) that we must interpret that phrase in context
and with reference to the larger statutory scheme. See, e.g., Brown &
Williamson, 529 U.S. at 132 (at step one courts should not examine statutory
provisions “in isolation” but must “interpret the statute as a symmetrical and
coherent regulatory scheme” (internal quotes and citations omitted)). Section
1311(b)(2)(A) does not flatly require that a BAT standard “eliminate[ ] the
discharge of all pollutants” solely if the Administrator finds such elimination
“technologically and economically achievable,” as petitioners claim. 25 Rather,
the phrase petitioners rely on is nested in a complex provision providing that
certain effluent limitations
shall require the elimination of discharges of all pollutants if the
Administrator finds, on the basis of information available to
him[,] . . . that such elimination is technologically and economically
achievable for a category or class of point sources as determined in
accordance with regulations issued by the Administrator pursuant
to section 1314(b)(2) of this title.

Id. (emphasis added). The italicized portion of § 1311(b)(2) cross-references the
factors set forth in § 1314(b)(2) that are “to be taken into account” by the
Administrator in determining BAT for a given point source. See id.
§ 1314(b)(2)(B) (requiring certain “factors relating to the assessment of [BAT]”
shall “be taken into account in determining the best measures and practices
available to comply with subsection (b)(2) of section 1311 of this title”). The

25 For that reason we also reject petitioners’ argument that EPA was required to set
chemical precipitation as BAT solely because that method is technologically and economically
achievable. Petitioners rely on a statement from American Petroleum Institute v. EPA that
“the basic requirement for BAT effluent limitations is only that they be technologically and
economically achievable[.]” 858 F.3d 261, 265-66 (5th Cir. 1988) (“API II”). As EPA explains,
however, the sentence containing that phrase was deleted on rehearing. See Am. Petroleum
Inst. v. EPA, 864 F.2d 1156, 1156 (5th Cir. 1989) (clarified on reh’g).
38
 Case: 15-60821 Document: 00514914192 Page: 39 Date Filed: 04/12/2019

No. 15-60821
statute requires us to read these two sections in harmony with each other. See,
e.g., Doe v. KPMG, 398 F.3d 686, 688 (5th Cir. 2005) (requiring that we “read
all parts of the statute together to produce a harmonious whole” (citation
omitted)). And we have held that the Administrator has “considerable
discretion” in evaluating the cross-referenced § 1314(b)(2) factors when making
a BAT determination. See, e.g., Tex. Oil & Gas, 161 F.3d at 928 (citing NRDC
v. EPA, 863 F.2d 1420, 1426 (9th Cir. 1988)). Consequently, it is not the case
that § 1311(b)(2)(A)—standing apart from the factors in § 1314(b)(2)—
unambiguously required EPA to set a stricter BAT for leachate. We therefore
reject petitioners’ step one argument based on the text of § 1311(b)(2)(A).
We agree with petitioners, however, that the leachate rule conflates the
BAT and BPT standards in a way not permitted by the statutory scheme. The
rule pegs BAT for leachate to the decades-old BPT standard, without offering
any explanation for why that prior standard is now BAT. That is flatly
inconsistent with the Act’s careful distinction between the two standards. As
explained, the difference between BAT and BPT is critical to the Act’s
“technology-forcing” scheme. Supra I.A (quoting NRDC I, 822 F.2d 104, 123).
BPT is merely the “first step” towards the Act’s pollution reduction goals and
provides the “prior standard” against which the stricter BAT is to be measured.
Nat’l Crushed Stone, 449 U.S. at 69; id. at 75 & n.14; CMA, 870 F.2d at 196.
To that point, Congress designed BPT to reflect merely an average of the best
performance levels of existing plants, CMA, 870 F.2d at 203, whereas it
designed BAT to reflect “‘the performance of the single best-performing plant
in an industrial field.’” Tex. Oil & Gas, 161 F.3d at 928 (quoting CMA, 870 F.2d
at 226). And this critical distinction is reflected in the Act’s structure, which
treats BAT and BPT in different subsections, implements them on different
timelines, and evaluates them under different factors. Compare 33 U.S.C.
§§ 1311(b)(1)(A), 1314(b)(1)(A), 1314(b)(1)(B) (specifying BPT applicability,
39
 Case: 15-60821 Document: 00514914192 Page: 40 Date Filed: 04/12/2019

No. 15-60821
timeline, and factors), with id. §§ 1311(b)(2)(A), (C), (D) & (F), 1314(b)(2)(B)
(specifying BAT applicability, timelines, and factors); see also 80 Fed. Reg.
67,843 (distinguishing BPT and BAT).
Yet, in the face of this statutory structure, the rule sets BAT for leachate
“equal to the [prior] BPT limitation”—i.e., impoundments. 80 Fed. Reg. 67,854.
The rule reaches that decision without explaining why a technology selected in
1982 under the laxer BPT standard somehow meets the stricter BAT standard
today. That is particularly inexplicable given the rule’s recognition that
impoundments have proven “largely ineffective” at pollution control over the
past decades, id. at 67,840, 67,851-53; see supra I.B. And, as we have seen, it
was the recognized shortcomings of impoundments—shortcomings with
respect to leachate discharges as well as other wastestreams—that led the
agency to revise the steam-electric effluent guidelines in the first place. Id. at
67,840; supra I.B.
To be sure, we do not say that EPA is precluded by the Act from ever
setting BAT equivalent to a prior BPT standard. But given the plain distinction
between the two standards marked out in the Act, the agency would at least
have to offer some explanation for its decision that speaks to the statutory
differences between BAT and BPT. Here we are given nothing along those
lines. Consequently, the only conclusion we can draw from this record is that,
in setting BAT for leachate, the agency simply defaulted to the prior BPT. As
explained, however, the statutory scheme does not confer authority on the
agency to collapse the carefully-wrought distinction between BAT and BPT in
this manner. See, e.g., Texas v. United States, 497 F.3d 491, 501 (5th Cir. 2007)
(“Chevron deference comes into play, of course, only as a consequence of
statutory ambiguity, and then only if the reviewing court finds an implicit
delegation of authority to the agency.”).

40
 Case: 15-60821 Document: 00514914192 Page: 41 Date Filed: 04/12/2019

No. 15-60821
We also agree with petitioners that the agency’s proffered justifications
for the leachate rule are not supported—indeed, are likely incompatible with—
the factors set forth under the Act for determining BAT. “The first [Chevron
step] determines whether Congress intended to give the agency any
discretion,” La Union Del Pueblo Entero v. FEMA, 608 F.3d 217, 220 (5th Cir.
2010), and we are unpersuaded that Congress gave the EPA discretion to rely
on justifications like these. As explained, the agency excuses its lax leachate
BAT by appealing to (1) the relatively “small” amount of pollutants discharged
in leachate, and (2) the stricter BATs set for larger industry wastestreams. See
80 Fed. Reg. 67,854. Yet neither of these considerations implicates any of the
factors the Act requires the Administrator to consider in determining BAT for
a given point source. See 33 U.S.C. § 1314(b)(2)(B) (requiring that regulations
“shall take into account” specified “[f]actors relating to the assessment of best
available technology” with respect to “any point source . . . within such
categories or classes”).
The Act specifies the following BAT factors:
[1] the age of equipment and facilities involved, [2] the process
employed, [3] the engineering aspects of the application of various
types of control techniques, [4] process changes, [5] the cost of
achieving such effluent reduction, [6] non-water quality
environmental impact (including energy requirements), and [7]
such other factors as the Administrator deems appropriate[.]

Id. (brackets added). These factors cannot be stretched to accommodate the
agency’s rationales for its leachate BAT. That is, no factor allows the agency to
consider the amount of pollutants generated by a one wastestream relative to
other streams. Nor does any factor allow the agency to consider whether less
stringent regulation of one wastestream may be set off against the benefits of
regulating other streams more strictly. See, e.g., Tex. Oil & Gas Ass’n, 161 F.3d
at 928 (explaining that “CWA specifies several factors that must be considered

41
 Case: 15-60821 Document: 00514914192 Page: 42 Date Filed: 04/12/2019

No. 15-60821
by the EPA in determining BAT limits” (emphasis added)). Thus, on their face,
the justifications for the leachate BAT put forward in the rule fall outside of
the factors mandated by the Act for determining BAT. See, e.g., API II, 858
F.2d at 264 (“Before EPA selects BAT-level limitations, it is required to address
both (1) operational considerations, including ‘the process employed, the
engineering aspects of the application of various types of control techniques
[and] process changes,’ and (2) cost” (emphasis added)).
What is more, the agency’s benefit-weighing appears incompatible with
the BAT factors and with the broader statutory scheme. Whether BAT factors
always preclude weighing regulatory benefits against other considerations is
an open question. See Entergy Corp. v. Riverkeeper, Inc., 556 U.S. 208, 221
(2009) (declining to resolve whether “cost-benefit analysis is precluded under
the [BAT] . . . test[ ]”). 26 We need not address that question here, because we
conclude that the agency’s benefit-weighing in this case contravenes the
statute. As explained, the rule explicitly justifies a less stringent BAT for
leachate by touting the benefits of stricter BATs for other wastestreams. But
the Act does not permit the agency to set a BAT by playing one pollution source

26 Riverkeeper held that a distinct CWA standard—“best technology available for
minimizing environmental impact” (or “BTA”), 33 U.S.C. § 1326(b)—implicitly allows cost-
benefit analysis. See 556 U.S. at 220 (concluding BTA “does not unambiguously preclude cost-
benefit analysis”). That decision, however, does not require finding that BAT also allows cost-
benefit analysis. Riverkeeper focused on the specific BTA wording—“minimizing
environmental impact”—which “admits of degree and is not necessarily used exclusively to
refer to the ‘greatest possible [pollutant] reduction.’” Id. at 219. Riverkeeper contrasted that
wording with BAT, whose “plain language . . . ‘require[s] the elimination of all pollutants.’”
Id. (quoting 33 U.S.C. § 1311(b)(2)(A) (emphasis in original)). Riverkeeper also reasoned that
other standards—like BAT—“are elucidated by statutory factor lists that guide their
implementation,” unlike BTA. Id. at 221 (citing, inter alia, 33 U.S.C. § 1314(b)(2)(B)). Finally,
Riverkeeper expressly declined to resolve whether “cost-benefit analysis is precluded under
the BAT[ ] . . . test.” Id. at 221-22 (stating “[i]t is not obvious to us that the . . . proposition[ ]
is correct, but we need not pursue that point”). Of course, we recognize that the BAT factors
do require the agency’s consideration of “the cost of achieving such effluent reduction.” 33
U.S.C. § 1314(b)(2)(B).
42
 Case: 15-60821 Document: 00514914192 Page: 43 Date Filed: 04/12/2019

No. 15-60821
off against another. As petitioners point out, the Act instead requires a BAT
determination to be made with respect to a discrete “point source.” 33 U.S.C.
§ 1314(b)(2)(B). A “point source” is defined as “any discernible, confined and
discrete conveyance, including but not limited to any pipe, ditch, channel,
tunnel, conduit, well, discrete fissure, container, rolling stock, concentrated
animal feeding operation, or vessel or other floating craft, from which
pollutants are or may be discharged.” Id. § 1362(14). This broad definition
easily includes leachate, and the rule leaves no doubt that it treats leachate as
a distinct point source. See 80 Fed. Reg. 67,847 (treating “combustion residual
leachate” from landfills or impoundments as a separate wastestream). The Act
thus specifically requires the BAT factors be applied with respect to a specific
point source—here, leachate. See 33 U.S.C. § 1314(b)(2)(B) (providing that
regulations “shall . . . specify factors to be taken into account in determining
the best measures and practices available to comply with [the BAT
requirements in § 1311(b)(2)] applicable to any point source . . . within such
category or classes” (emphasis added)). But in the final rule the agency has
explicitly factored into its BAT determination the regulation of wastestreams
other than leachate, which contravenes the plain text and structure of the Act.
EPA’s principal argument in response is that the Act allows
consideration of “other factors as the Administrator deems appropriate,” 33
U.S.C. § 1314(b)(2)(B), which permits it to consider the “very small” size of
leachate pollution relative to the overall industry. Because leachate represents
only about 3 percent of overall industry pollution, EPA claims, the rule still
represents “reasonable progress” towards eliminating pollution because of the
rule’s regulation of other sources. See 33 U.S.C. § 1311(b)(2)(A). Even accepting
EPA’s characterization of leachate pollution as “very small” (something we
address at Chevron step two, infra), we reject the agency’s reliance on the
“other factors” clause. Like any statute, we must interpret the Clean Water Act
43
 Case: 15-60821 Document: 00514914192 Page: 44 Date Filed: 04/12/2019

No. 15-60821
by “looking at the full text of the statute, rather than one isolated clause, along
with the statute’s structure and its public safety purpose[.]” United States v.
Transocean Deepwater Drilling, Inc., 767 F.3d 485, 496 (5th Cir. 2014). As
petitioners correctly point out, the agency’s “other factors” argument would
undermine the concept of BAT altogether. The BAT factors are designed to
support achievement of an effluent limitation that “shall require the
elimination of discharges of all pollutants,” if “technologically and economically
achievable.” 33 U.S.C. § 1311(b)(2)(A) (citing id. § 1314(b)(2)). But accepting
the agency’s expansive view of the “other factors” clause would allow it, in every
case, to justify a less stringent BAT for one pollution source by claiming it was
regulating other sources more strictly and thus making reasonable progress in
the industry “as a whole.” On this understanding, the BAT standard would
cease to have any meaning. We therefore reject EPA’s reliance on the “other
factors” clause to support its justification for the leachate regulation. 27
In sum, we conclude that the BAT determination for leachate fails step
one of Chevron.
2.
Alternatively, we conclude that the leachate regulation fails step two of
Chevron. For purposes of this analysis, we assume that the CWA is “silent or
ambiguous” with respect to the question addressed by the rule, and we ask only
“whether the agency’s answer is based on a permissible construction of the
statute.” Acosta, 909 F.3d at 730 (quoting City of Arlington, 569 U.S. at 296;
Chevron, 467 U.S. at 843). “We do not simply impose our own construction on

27EPA also argues that “BAT . . . must be acceptable on the basis of numerous factors,
only one of which is pollution control” (quoting BP Expl. & Oil, Inc. v. EPA, 66 F.3d 784, 796
(6th Cir. 1995)). That is true, but EPA has not offered any non-pollution-control factors
showing impoundments are superior to chemical precipitation, let alone factors sufficient to
outweigh the shortcomings of impoundments.
44
 Case: 15-60821 Document: 00514914192 Page: 45 Date Filed: 04/12/2019

No. 15-60821
the statute.” BNSF Ry. Co. v. United States, 775 F.3d 743, 751 (5th Cir. 2015)
(cleaned up) (quoting Chevron, 476 U.S. at 843). “The agency’s view ‘governs if
it is a reasonable interpretation of the statute—not necessarily the only
possible interpretation, nor even the interpretation deemed most reasonable
by the courts.’” Acosta, 909 F.3d at 735 (quoting Coastal Conserv’n Ass’n v. U.S.
Dep’t of Commerce, 846 F.3d 99, 106 (5th Cir. 2017); Riverkeeper, 556 U.S. at
218). While this is a highly deferential standard, an agency interpretation can
fail Chevron step two if “it is contrary to clear congressional intent or frustrates
the policy Congress sought to implement.” Garcia-Carias v. Holder, 697 F.3d
257, 271 (5th Cir. 2012). Agency action that is “arbitrary, capricious, or
manifestly contrary to the statute” also fails step two. Tex. Coal. of Cities for
Util. Issues v. FCC, 324 F.3d 802, 807 (5th Cir. 2003) (quoting Chevron, 467
U.S. at 844). Because Chevron step two and the APA share the “arbitrary and
capricious” standard, “[t]he APA reflects the principles of Chevron,” and
analysis under the two standards proceeds similarly. Nutraceutical Corp. v.
Von Eschenbach, 459 F.3d 1033, 1038 (10th Cir. 2006); see also, e.g., PIERCE §
3.6 (suggesting that Chevron step two has “complete overlap” with APA test of
whether a rule adopts an “unreasonable” statutory interpretation) (quoting
Animal Legal Def. Fund v. Glickman, 204 F.3d 229, 24 (D.C. Cir. 2000)).
Petitioners argue that the agency’s decision to set surface impoundments
as BAT for leachate is based on an impermissible interpretation of the Act.
They raise arguments similar to the ones raised under step one—i.e., that the
agency acted unreasonably by setting a leachate BAT based on its relative size
and on the rule’s stricter regulation of other streams. Additionally, petitioners
argue that the agency rejected more effective, achievable control technologies
(like chemical precipitation) in favor of a less effective technology like
impoundments, which “is unreasonable because it cannot be squared with
Congress’s intent for BAT to be more stringent than BPT limits.” We agree
45
 Case: 15-60821 Document: 00514914192 Page: 46 Date Filed: 04/12/2019

No. 15-60821
with petitioners that the leachate regulation is based on an impermissible
interpretation of the Act. We therefore hold that the regulation fails Chevron
step two and must be vacated on that alternative basis as well.
First, the rule unreasonably sets as BAT a technology the rule itself
deems ineffective at controlling toxic discharges from leachate. As already
explained, supra V.B.1, the final rule categorically recognizes that
impoundments are ineffective at removing toxic pollutants from wastewater,
80 Fed. Reg. 67,840, 67,851, which is why the rule declined to set
impoundments as BAT for five of the six wastestreams at issue, id. at
67,852-53. Nothing in the rule even hints, much less explains, that
impoundments are somehow better at controlling harmful discharges from
leachate. To the contrary, the rule recounts that groundwater contamination
from impoundments (which are “the most widely used systems to
treat . . . leachate”) has resulted in numerous documented cases of drinking
water pollution. Id. at 67,840, 67,847; see also id. at 67,847 (defining “leachate”
as “liquid . . . that has percolated through or drained from waste or other
materials placed in a landfill, or that passes through the containment
structure . . . of a surface impoundment”); id. (explaining that “[u]nlined
impoundments and landfills . . . allow the leachate to potentially migrate to
nearby ground waters, drinking water wells, or surface water”). The rule also
refers to an environmental assessment document, id. at 67,840, which reports
that “[c]ombustion residual leachate can migrate from the site in the ground
water at concentrations that could contaminate public or private drinking
water wells and surface waters, even years following disposal of combustion
residuals.” See Environmental Assessment Document No. EPA-821-R-15-006,
§ 3.3.2. Given these admitted deficiencies in impoundments, it was
unreasonable to adopt them as BAT for leachate. See State Farm, 463 U.S. at
43 (“Normally, an agency rule would be arbitrary and capricious if the agency
46
 Case: 15-60821 Document: 00514914192 Page: 47 Date Filed: 04/12/2019

No. 15-60821
has . . . offered an explanation for its decision that runs counter to the evidence
before the agency.”).
Second, the rule unreasonably declines to set as BAT available
technologies that are admittedly more effective at controlling leachate
discharges. We have already detailed the rule’s affirmation that available
modern technologies like chemical precipitation “are more effective than
surface impoundments at removing both soluble and particulate forms of
metals.” 80 Fed. Reg. 67,851; supra V.B.1. The rule never explains why this
blanket statement does not apply to the use of impoundments to treat leachate.
Indeed, EPA acknowledged during the rulemaking process that “chemical
precipitation is an available and demonstrated technology for the treatment of
combustion residual leachate.” See EPA’s Response to Public Comments,
SE05958, at 7-20. And EPA’s counsel conceded at oral argument that chemical
precipitation is “available” as a treatment for leachate. Oral Argument Audio
at 22:25-22:45 (Oct. 3, 2018). As with its treatment of legacy wastewater, the
rule appears to select a BAT for leachate simply by defaulting to the decades-
old and demonstrably ineffective BPT standard. We have already explained
that this kind of regulation-by-inertia is inconsistent with the “technology-
forcing” mandate of the CWA. See NRDC II, 808 F.3d at 563-64; NRDC I, 822
F.2d at 123; supra V.B.1. Moreover, a decision to leave BPT limitations in place
for leachate, when those limitations were based on admittedly ineffective
technology, does not reflect “a commitment of the maximum resources
economically possible to the ultimate goal of eliminating all polluting
discharges,” which was the intent of Congress in enacting BAT standards in
the first place. Nat’l Crushed Stone Ass’n, 449 U.S. at 74. The EPA’s decision
to rest on its laurels (questionable as they are) respecting leachate thus
“frustrates the policy Congress sought to implement” in the CWA, see Garcia-
Carias, 697 F.3d at 271, and cannot stand.
47
 Case: 15-60821 Document: 00514914192 Page: 48 Date Filed: 04/12/2019

No. 15-60821
Our decision in Chemical Manufacturers Association v. EPA, 870 F.2d
177, reinforces our conclusion on this point. That case involved a challenge to
BAT and NSPS regulations set for water pollution by organic chemicals and
synthetics plants. Id. at 261. In setting NSPS limits for these plants, EPA did
not consider recycling as a possible NSPS technology, even though the evidence
before it showed that plants could achieve zero discharge by using recycling.
Id. at 262-63. We concluded that the agency’s failure even to consider recycling,
an “available demonstrated technology,” was arbitrary and capricious. Id. at
264. Our case is analogous: here we have the known shortcomings of the status
quo technology (impoundments), and the demonstrated superiority of available
alternatives (chemical precipitation). To be sure, EPA was entitled to deference
on NSPS standards just as it is on BAT standards, id. at 263, but deference
runs out when the agency fails to consider an option with all the indicators of
being a superior choice. Just so here. An “[u]nexplained inconsistency in
agency policy is a reason for holding an interpretation to be an arbitrary and
capricious,” and “[a]n arbitrary and capricious regulation of this sort is itself
unlawful and receives no Chevron deference.” Encino Motorcars, LLC v.
Navarro, 136 S. Ct. 2117, 2126 (2016) (citing Nat’l Cable & Telecomms. Ass’n
v. Brand X Internet Servs., 545 U.S. 967, 981 (2005); United States v. Mead
Corp., 533 U.S. 218, 227 (2001)).
Finally, we note that the EPA has described leachate as being chemically
similar to FGD wastewater, a wastestream admittedly susceptible to effective
treatment by chemical and biological methods. See 80 Fed. Reg. 67,851; see also
EPA’s Response to Public Comments, SE05958, at 7-20. It puzzles us that the
EPA has described two wastestreams as chemically similar, and susceptible to
treatment by the same methods, and yet has set strikingly different BAT
standards for each. As with legacy wastewater, the agency’s rationales
contradict themselves. The BAT determination for leachate is “illogical on its
48
 Case: 15-60821 Document: 00514914192 Page: 49 Date Filed: 04/12/2019

No. 15-60821
own terms” and therefore cannot stand. See, e.g., GameFly, 704 F.3d at 148
(D.C. Cir. 2013) (facially illogical determinations are arbitrary and capricious).
EPA offers two justifications for selecting impoundments as BAT for
leachate—the first based on the agency’s lack of data about alternatives, and
the second based on the relative size of the leachate wastestream. We find
neither persuasive.
First, we are unpersuaded by the agency’s argument that
“[c]ommentators did not provide information that EPA could use to establish
BAT limitations” concerning leachate. 80 Fed. Reg. at 67,854. During its
lengthy study period beginning in 2005, not only did EPA have adequate
opportunity to collect data on various treatment options for leachate, but the
agency did collect data on the size of leachate pollution and on the benefits of
chemical precipitation. See 80 Fed. Reg. 67,854 (“EPA considered whether
technologies in place for treatment of other wastestreams” could be used to
treat leachate); TDD, EPA-821-R-15-007, at 10-39 (table showing potential
reduction in leachate pollution of 33,800 toxic-weighted pound equivalents by
adopting chemical precipitation). These data were certainly enough to demand
further inquiry and, of course, they say nothing to justify setting
impoundments as BAT.
We have before declined to accept lack of data as a valid excuse for an
agency’s failure to regulate activity that concededly creates pollution, and we
decline again here. In API I, for instance, we rejected EPA’s attempt to justify
failing to regulate “stripper gas wells” based on its claim that “there was not
sufficient data” to justify regulation. 661 F.2d at 357. We reasoned that, in the
three years since EPA had declined to regulate, ample public data on the wells
“belie[d] EPA’s contention that there exists nothing to regulate.” Id. We
therefore remanded for further agency consideration “in light of this new
information.” Id. Here, impoundments have been in operation for over three
49
 Case: 15-60821 Document: 00514914192 Page: 50 Date Filed: 04/12/2019

No. 15-60821
decades, and, as we have discussed at length, the agency’s own rule amply
demonstrates their ineffectiveness in controlling discharges from
wastestreams including leachate. The agency cannot rely on a lack of data to
justify its setting a BAT standard based on demonstrably outdated and
ineffective technology. 28
EPA counters this point by asserting that it “identified no existing plants
using chemical precipitation to treat their leachate.” We take EPA at its word,
as we must, but the agency misses the point. Under our precedent, a
technological process can be deemed “available” for BAT purposes “even if it is
not in use at all,” or if it is used in unrelated industries. API II, 858 F.2d at
265. “Such an outcome is consistent with Congress’[s] intent to push pollution
control technology.” Id. In this case, technologies alternative to surface
impoundments are in use in the steam-electric industry, just in separate
wastestreams. See, e.g., 80 Fed. Reg. 67,855 n.29. If technologies from other
industries can be considered, then, a fortiori, technologies within the same
industry should be considered when the status quo technology in place for a
wastestream is admittedly ineffective. The final rule itself recognizes this
point. See 80 Fed. Reg. 67,843 (“BAT is intended to reflect the highest
performance in the industry, and it may reflect a higher level of performance

28 We find additional support for this conclusion in Natural Resources Defense Council
v. EPA, 808 F.3d 556 (2nd Cir. 2015), which stands for the proposition that EPA’s failure to
gather data on a technological option can be arbitrary and capricious in itself. Id. at 573-74.
There, EPA studied and regulated ballast water discharges from ships. Rather than consider
onshore treatment systems for ballast water, the agency limited its consideration to
shipboard treatment systems, foreclosing any discussion on onshore treatment. Id. at 573.
EPA later pleaded lack of data as a justification for not adopting onshore treatment systems
as BAT. Id. The Second Circuit rejected that argument, holding that the lack of data was “a
problem of EPA’s own making.” Id. at 573-74. The decision is analogous to our case. While
here we do not have evidence that EPA actively sought to suppress data collection, cf. 808
F.3d at 573, nonetheless EPA had both the opportunity to gather data on leachate and a
strong incentive to do so—namely, the recognized problems with impoundments in
controlling discharge of toxic pollutants from leachate discharges. 80 Fed. Reg. 67,840.
50
 Case: 15-60821 Document: 00514914192 Page: 51 Date Filed: 04/12/2019

No. 15-60821
than is currently being achieved based on technology transferred from a
different category or subcategory, bench scale or pilot studies, or foreign
plants.” (citing Am. Paper Inst. v. Train, 543 F.2d 328, 353 (D.C. Cir. 1976);
Am. Frozen Food Inst. v. Train, 539 F.2d 107, 132 (D.C. Cir. 1976))). In any
event, EPA’s argument ignores the fact that during notice-and-comment it
admitted that “chemical precipitation is an available and demonstrated
technology for the treatment of combustion residual leachate,” based on the
technology’s use in treating a similar wastestream (FGD wastewater). See
EPA’s Response to Public Comments, SE05958, at 7-20.
Second, we reject the EPA’s argument that its regulation is justified by
the fact that leachate pollution constitutes “a very small portion of the
pollutants discharged collectively by all steam power plants.” 80 Fed Reg.
67,854. As already explained at step one, supra V.B.1, this consideration finds
scant support in the statutory scheme given that the relative size of a stream
is absent from the statutory BAT factors. But even assuming the statute allows
the agency to consider the relative size of a pollution source in setting BAT, we
find the EPA’s use of that consideration here to be unreasonable.
The agency’s argument involves some sleight of hand. Leachate
discharges may constitute “a very small portion” of pollutants, 80 Fed Reg.
67,854, but only by comparison to all pollution from the entire steam-electric
power plant industry, which is largest collective source of water pollution in
the country. Id. at 67,839-40. 29 In an absolute sense, however, leachate
pollution is not a “very small portion” of anything. If leachate were a separate

29 After steam-electric power plants—the most polluting industry—the remaining
nine industries in the top ten for water pollution are: 2) pulp, paper, and paperboard, 3)
petroleum refining, 4) nonferrous metals manufacturing, 5) fertilizer manufacturing, 6)
organic chemicals, plastics, and synthetic fibers, 7) ore mining and dressing, 8) inorganic
chemicals manufacturing, 9) waste combustors, and 10) textile mills. Environmental
Assessment Document No. EPA-821-R-15-006, at 3-15.
51
 Case: 15-60821 Document: 00514914192 Page: 52 Date Filed: 04/12/2019

No. 15-60821
industry, it would rank as the 18th-largest source of water pollution in the
United States. See TDD, EPA-821-R-15-007, at 10-39 (table showing toxic-
weighted pollution from leachate); Annual Effluent Guidelines Review Report,
EPA-821-R-16-002, at 2-26 (listing pollution from other industries). Leachate
alone produces more toxic-weighted pound equivalents than the entire coal
mining industry, according to the EPA’s own records. Id. And EPA’s brief
grudgingly admits that leachate pollution “might be considered comparable” to
the total amount of pollution coming from other industries such as coal mining
and sugar processing. Thus, with the distortions stripped away, EPA’s
“argument” for its leachate regulation turns out to be a mere statement of fact,
and an empty one at that. Yes, leachate pollution may form a “very small
portion” of a gargantuan source of water pollution. But leachate constitutes a
gargantuan source of water pollution on its own terms. Both statements are
true. Neither begins to justify the agency’s choice of impoundments as BAT. 30
At oral argument, EPA conceded that it lacked discretion to ignore
leachate based on its allegedly small size, instead arguing it had regulated
leachate—i.e., by setting BAT based on surface impoundments. Oral Argument
Audio at 22:10-22:20. Thus we end where we began. The final rule itself tells
us that surface impoundments have a poor record at controlling toxic water
pollution. 80 Fed. Reg. 67,840. The agency’s own comments tell us that
“chemical precipitation is an available and demonstrated technology for the

30 Additionally, we have previously rejected arguments by EPA that a certain pollution
source need not be regulated because it was allegedly small or insignificant. In API I, 661
F.2d at 357, the EPA declined to subject a certain subcategory of wells, “stripper gas wells,”
to the same control standards as other wells. The agency attempted to justify its decision by
claiming stripper wells were “not a large problem” (as well as claiming it lacked data, see
supra). Id. We rejected both arguments, pointing to evidence of the absolute number of
stripper wells—not the pollution produced by those wells relative to the entire industry—and
remarking that “[s]uch figures belie EPA’s contention that there exists nothing to regulate.”
Id. The absolute number of stripper wells, not their number in proportion to the total number
of oil and gas wells in the country, was what mattered. We have a similar situation here.
52
 Case: 15-60821 Document: 00514914192 Page: 53 Date Filed: 04/12/2019

No. 15-60821
treatment of combustion residual leachate.” Faced with these concessions, we
must conclude that the leachate regulation rests on an impermissible
interpretation of the Clean Water Act and therefore fails Chevron step two. 31
V. CONCLUSION
In the Clean Water Act, Congress afforded EPA considerable policy-
making discretion to formulate rules to mitigate pollution in our nation’s
waters. As a price for that delegation of authority, however, the agency must
engage in “reasoned decisionmaking,” Michigan v. EPA, 135 S. Ct. at 2706, and
“stay[ ] within the bounds of its statutory authority.” Utility Air Reg. Grp. v.
EPA, 573 U.S. 302, 315 (2014) (internal quotation marks and citations
omitted). The regulations challenged in this case fall short of those judicially-
enforceable limits on the exercise of agency discretion. We therefore conclude
that the portions of the final rule regulating legacy wastewater and residual
combustion leachate are unlawful. Accordingly, we vacate those portions of the
rule and remand to the agency for reconsideration consistent with this opinion.
VACATED IN PART AND REMANDED.

31 Relying on Association of Pacific Fisheries v. EPA, 615 F.2d 794 (9th Cir. 1980),
Intervenor UWAG argues that the leachate BAT was justified by EPA’s “determination not
to impose extraordinary costs on the industry for de minimis gains.” See id. at 818 (holding
that “at some point extremely costly more refined treatment will have a de minimis effect on
the receiving waters”). We disagree. EPA made no such argument for its leachate BAT, and
“[w]e may not supply a reasoned basis for the agency’s action that the agency itself has not
given.” State Farm, 463 U.S. at 43. In any event, as we have explained, leachate is a massive
pollution source in its own right. It would be arbitrary to claim the benefits of more strictly
regulating it were de minimis, which is likely why EPA never made such an argument.
53