Appendix — PPG Industries, Inc. v. United States Environmental Protection Agency

OCTOBER TERM, 1989

PPG INDUSTRIES, INC.,

Petitioner,

Vv.

UNITED STATES ENVIRONMENTAL PROTECTION AGENCY,

Respondent.

Petition For Writ Of Certiorari

To The United States Court Of Appeals

For The Fifth Circuit

APPENDIX

*ROBERT BRAGER

SHELLEY V. LUCAS

BEVERIDGE & DIAMOND, P.C.

1350 I Street, N.W., Suite 700

Washington, D.C. 20005

(202) 789-6000

MADELYN A. REILLY

PPG INDUSTRIES, INC.

One PPG Place

Pittsburgh, Pennsylvania 15272

(412) 434-2430

Counsel for -Petitioner

PPG Industries, Inc.

*Counsel of Record

PRESS OF BYRON S. ADAMS, WASHINGTON, D.C. (202) 347-8203

APPENDIX A:

e

APPENDIX B:

1.

2.

APPENDIX C:

APPENDIX D:

APPENDIX E:

APPENDIX F:

INDEX TO APPENDICES

THE CASE BELOW

Chemical Manufacturers Associa-

tion v. United States Environmental

Protection Agency, 870 F.2d 177

CGR GI. TEBE hecntccnssicessevncsisiessores

Chemical Manufacturers Associa-

tion v. United States Environmental

Protection Agency, 885 F.2d 253

(5th Cir. 1989) (on rehearing) ......

THE CLEAN WATER ACT

3B U.S.C. § 181 1a)D) ..0..cccrseecc00ee

BB UBC. § TB1GD) ccccsccovosccescsscores

THE EFFLUENT DISCHARGE

GUIDELINES FOR THE OCPSF IN-

DUSTRY (40 C.F.R. PART 414) ......

EPA’S RESPONSES TO

COMMENTS wcccccccccccccscsccccccsccvcocsess

EPA’S TRAINING MANUAL FOR

NPDES PERMIT WRITERS ............

DEVELOPMENT DOCUMENT FOR

EFFLUENT LIMITATIONS GUIDE-

LINES AND STANDARDS FOR THE

OCPSF INDUSTRY ..........ccccccccreeees

Page

la

la

APPENDIX A

United States Court of Appeals,

Fifth Circuit.

No. 87-4849, et al.*

CHEMICAL MANUFACTURERS

ASSOCIATION, et al.,

Petitioners,

v.

U.S. ENVIRONMENTAL PROTECTION

AGENCY,

Respondent.

March 30, 1989.

* Sterling Chemicals, Inc. (87-4850); Paul M. Toce (87-4868); Gulf

Coast Waste Disposal Authority (87-4920); Union Carbide Corporation

(87-4928); Air Products Manufacturing Corporation and Air Products

and Chemicals, Inc. (87-4929); E.I. Du Pont de Nemours & Co. (87-

4930); Synthetic Organic Chemical Manufacturers Association, Inc. and

Dixie Chemical Company, Inc. (87-4934); Monsanto Company (87-4936);

Natural Resources Defense Council, Inc. (88-4077); National Paint &

Coatings Association (88-4125); Rubicon, Inc. (88-4129); The Dow Chem-

ical Company (88-4135); The Goodyear Tire & Rubber Company (88-

4153); Texas Eastman Company, a division of Eastman Kodak Company

(88-4154); Laroche Chemicals, Inc. (88-4176); PPG Industries, Inc. (88-

4177); FMC Corporation (88-4186); Koppers Company, Inc. (88- 4189);

The Lubrizol Corporation (88-4192), Akzo Chemicals, Inc., Et Al. (88-

4193); Hoechst Celanese Corporation and Hoechst Celanese Chemical

Group, Inc. (88-4194); Ethyl Corporation, Et Al. (88-4195); QO Chem-

icals, Inc. (88-4196); M & T Chemicals, Inc. (88-4246); Borg-Warner

Specialty Chemicals, Inc. (88-4272); Sherex Chemical Company, Inc. (88-

4274); Allied-Signal, Inc. (88-4284); W.R. Grace & Company (88-4305);

Courtaulds Fibers, Inc. (88-4526).

2a

Petitions for Review of an Order of the Environmental

Protection Agency.

Before RUBIN, GARZA and KING, Circuit Judges.

By RUBIN, GARZA and KING, Circuit Judges Jointly.

Acting under the mandate of the Clean Water Act

(CWA),' the Environmental Protection Agency (EPA) has

promulgated final regulations limiting the discharge of pol-

lutants into the nation’s navigable waters by manufactur-

ing plants in the organic chemicals, plastics, and synthetic

fibers (OCPSF) industries. The regulations, which the stat-

ute requires to be implemented beginning March 31, 1989,

cover both direct discharge and indirect discharge through

publicly-owned treatment works (POTWs). The Chemical

Manufacturers Association (CMA) and a number of com-

panies affected by the regulations allege both procedural

defects in their promulgation and substantive defects in

various provisions, as well as defects in the application of

specific provisions to particular plants. Intervening in some

of the cases consolidated for review and appearing as ami-

cus curiae in the others, the Natural Resources Defense

Council (NRDC) also challenges the regulations, but on the

different ground that they fail to require a sufficiently high

degree of effluent pollution control. Although it contends

that the regulations are invalid, the NRDC urges that they

be enforced until more stringent standards can be adopted.

The case is of such complexity that the parties have

submitted briefs totalling more than 3,000 pages and a

joint appendix 9,000 pages long distilled from a 600,000-

page administrative record. To enable us to render a de-

cision as promptly as possible, the members of the panel

have divided responsibility for preparing portions of this

opinion, as the District of Columbia Circuit did in Alabama

1 33 U.S.C. §§ 1251-1376.

3a

Power Co. v. Costle.2 Judge Garza prepared sections V,

VI, and VII of this opinion, as well as all portions dis-

cussing issues raised by the NRDC; Judge Rubin prepared

sections I and III; and Judge King prepared sections I

and IV, except for those portions discussing issues raised

by the NRDC.

This is a summary of our rulings on the principal issues:

I. The EPA did not violate the notice-and-comment re-

quirements of the Administrative Procedure Act (APA) by:

(1) utilizing an updated Dun & Bradstreet economic-impact

study to supplement data it had previously disclosed with-

out making the new data public either during the notice-

and-comment period or in the public record except to this

court under seal; or (2) by failing to publish its regulations

for the control of toxic metals for public comment prior

to final promulgation.

II. Best Practicable Technology (BPT) Issues:

A. The EPA’s Consideration of the Costs of Complying

with the BPT Limitations:

1. The Act does not require the EPA to apply a ‘knee-

of-the-curve’ cost-effectiveness test in establishing BPT

limitations; the EPA sufficiently considered the total costs

of the BPT limitations in relation to the effluent-reduction

benefits and this is all that the Act requires.

2. The BCT cost-effectiveness test does not replace the

BPT cost requirement.

3. Although the BPT limitations will double industry’s

current costs for the removal of conventional pollutants,

these costs are not sufficiently high to make the Admin-

istrator’s decision arbitrary or capricious; identifying the

point of diminishing returns is within the discretion of the

Administrator.

* 606 F.2d 1068, superseded on rehearing, 636 F.2d 323

(D.C.Cir.1979).

a

4a

B. The EPA’s Definition of the BPT Data Base:

1. The EPA did not use unreasonably weak statistical

editing criteria to determine the ‘average of the best’ dis-

chargers in defining the BPT data base.

2. The EPA’s rejection of polishing ponds and multi-

media filtration as the BPT model technology was not

arbitrary or capricious; the EPA determined that the ben-

efits of polishing ponds relative to their costs were not

substantial and that the effectiveness of multimedia filtra-

tion was not well demonstrated.

C. The ‘Summer/Winter Issue’:

The EPA’s determination that neither special subcate-

gorization nor establishing special limits was necessary for

plants utilizing biological treatment systems in colder cli-

mates was not arbitrary or capricious; the record shows

that the performances of such plants were comparable to

those in other regions and technological modifications are

available to account for the limited instances in which cold

weather may affect treatment performance.

D. BPT Subcategorization:

1. The EPA’s use, in part, of Standard Industrial Clas-

sification (SIC) codes for the purpose of categorizing the

OCPSF industry was reasonable; because SIC codes are

based on product type, the EPA could reasonably assume

that plants producing similar products would have similar

wastestreams.

2. The NRDC’s contention that the EPA failed to pro-

vide public notice of its intent to limit the applicability of

the regulations to certain SIC codes is without merit; as

early as 1985 the EPA defined the OCPSF industries to

include all facilities within specified SIC codes.

E. Issues Concerning Waste Stabilization Ponds:

1. The EPA’s determination that pond algae are con-

ventional pollutants is consistent with the Act, which de-

5a

fines a conventional pollutant as a biological oxygen-

demanding substance or a suspended solid.

2. The EPA’s decision not to create a subcategory for

plants employing waste stabilization pond treatment sys-

tems was reasonable, because the EPA determined that

more effective technology was available; the EPA need

not create separate subcategories for industry members

that install less effective technology than the model tech-

nology.

3. The EPA adequately considered the compliance costs

that will be incurred by plants utilizing pond technology.

4. The EPA’s suggestion in the final regulations that

copper sulfate could be used to control pond algae was in

response to the industrial petitioners’ comments and thus

did not require further notice and comment under the

Administrative Procedure Act.

5. The costs of the BPT limitations are not wholly dis-

proportionate to its benefits.

F. Plant-Specific Challenges to the BPT Limitations and

the Availability of Fundamentally-Different-Factor Var-

lances

1. The EPA reasonably concluded that the BPT limi-

tations for biological oxygen-demanding substances were

achievable for Union Carbide’s plants through the use of

biological treatment.

2. The EPA reasonably concluded that the BPT limi-

tations were achievable for Borg-Warner’s plants and oth-

ers with phenol-dominated wastestreams.

3. DuPont’s claim that the wastestream of its Chambers

Works plant is uniquely complex, precluding compliance

with the BPT limitations, is not a basis for invalidating

the limitations.

4. The EPA’s decision not to create a separate subca-

tegory, as urged by Ethyl and Monsanto, for plants with

icnreieatsiatieneniettiaeen

6a

high concentrations of influent total-dissolved solids was

reasonable; there was insufficient evidence that any plant

has enough total dissolved solids to preclude effective

treatment.

5. The plant-specific challenges of Union Carbide, Borg-

Warner, Dow, Monsanto, and Ethyl] that their plants can-

not meet the limits are without merit because the Act

does not require the EPA to consider allegedly funda-

mentally different factors of individual plants in promul-

gating nationwide BPT limitations for an industry; the

CWA provides that fundamentally different factors of

plants that preclude compliance must be considered in a

collateral variance proceeding.

III. Best Available Technology (BAT) Issues:

A. The EPA’s Statistical Methodology:

1. The EPA’s use of weighted averaging to derive long-

term averages was not an abuse of discretion.

2. The EPA’s averaging of variability factors in deriving

effluent limits was not an abuse of discretion.

B. Remedy for Unavoidable Exceedances:

The EPA’s decision to exclude extremely high dis-

charges from the calculation of ‘‘variability factors’’ was

reasonable, and the ‘‘upset defense’ is a sufficient remedy

for exceedances that result from exceptional circumstances

beyond the control of the plant operator.

C. The EPA’s Analytical Methodology:

1. The industrial petitioners failed to demonstrate that

the EPA’s statistical model did not account for analytical

variability.

2. This court defers to the EPA’s determination that

the presence of multiple pollutants in a plant’s waste-

stream did not preclude accurate measurement of those

pollutants.

3. The EPA employed reasonable criteria for the use of

borrowed data when performance data for a particular

pollutant were not available.

4. The EPA’s error, if any, in determining “minimum

analytical values’’ was accounted for by the EPA’s con-

servative methodology which ensured that the limitations

are achievable.

5. The industrial petitioners were not prejudiced by

changes in analytical methods in the development of the

OCPSF limitations.

D. Use of Minimum Analytical Values for Enforcement

Purposes:

The petitioners’ claim that a non-detect reading should

be considered zero for enforcement purposes is not ripe

for review because the EPA has not adopted a formal

policy on this issue.

E. The EPA’s Sampling Techniques:

1. The industrial petitioners failed to demonstrate that

the effluent discharge limits were based on inaccurate EPA

sampling.

F. The EPA’s Application of the Limitations for Toxic

Pollutants to All OCPSF Dischargers:

1. The EPA did not abuse its discretion by subjecting

all OCPSF dischargers to the limits for all toxics.

2. The EPA provided adequate public notice that it was

considering subjecting all dischargers to the limits for all

toxics.

3. Even if this is the proper court in which to raise the

issue, the EPA’s limitations for toxic pollutants do not

violate the consent decree in NRDC v. Train.

G. BAT Subcategorization:

1. The EPA violated the notice and comment require-

ment of the Administrative Procedure Act by failing to

8a

publish for public comments its BAT subcategorization

prior to final promulgation of the regulations.

H. DuPont’s Chambers Works Plant:

1. DuPont’s assertion that its Chambers Works plant

will not be able to comply with the BAT limitations be-

cause of the plant’s uniquely complex wastestream is not

a basis for invalidating the limitations.

I. The BAT Limitations for Phenol:

1. Borg-Warner failed to demonstrate that the BAT lim-

itations for phenol are unachievable for plants with phenol-

dominated wastestreams; the record shows that such plants

are capable of meeting the limitations.

2. Borg-Warner failed to demonstrate that the EPA’s

cost estimates were not a reasonable approximation of the

costs the industry will have to incur to meet the BAT

limitations.

J. The BAT, Limitations for Volatile Pollutants Based

on Steam Stripping Technology:

1. PPG and Dow have failed to demonstrate that the

BAT, limitations for volatile pollutants are not achievable;

the record shows that at least one of the plants in the

data base was able to meet the limitations for each pol-

lutant.

2. Because the EPA determined that the limitations

could be met by all plants regardless of their wastestream

characteristics, the EPA did not abuse its discretion by

failing to further subcategorize the plants subject to the

limitations for volatile pollutants on the basis of waste-

stream characteristics.

3. Because the EPA determined that steam-stripper

maintenance could be conducted without discharging vol-

atile pollutants, the EPA did not abuse its discretion by

failing to make an allowance for such discharge in the

limitations.

9a

4. Dow Chemical’s claim that two of its plants will not

be able to meet the limitations for volatile pollutants is

not a basis for invalidating the limitations.

K. The BAT, Limitations for Priority Pollutants Based

on In-Plant Biological Treatment:

1. CMA has failed to demonstrate that the EPA's re-

liance on data from end-of-pipe biological treatment sys-

tems in deriving the BAT, limitations was arbitrary or

capricious.

2. CMA has failed to demonstrate that the EPA's cost

estimates for in-plant biological treatment systems are not

a reasonable approximation of the type and cost of the

technology that industry will have to utilize to meet the

limitations.

3. Contrary to CMA’s assertion, the record shows that

the EPA considered the land costs associated with in-plant

biological treatment systems.

L. Compliance Deadline:

1. The EPA’s post-deadline enforcement policy provides

an adequate time extension for industry members that will

not have sufficient time to instal] the technology required

to meet the limitations.

IV. Pretreatment Standards for Existing Sources

(PSES) Issues:

A. “Pass Through’’ Issues:

1. The EPA properly selected pollutants that pass

through POTWs for the pretreatment standards.

2. The EPA’s decision to define ‘‘pass through’’ based

on POTW average removal, rather than actual reported

removal, does not violate the CWA.

3. The EPA’s decision not to find pass through of chro-

mium, copper, and nickel, which contaminate sludge, was

10a

not arbitrary or capricious because sludge pollutants will

be subject to future rulemaking proceedings.

B. The EPA’s failure to establish PSES for six volatile

pollutants was not arbitrary or capricious because some

of these pollutants were sufficiently controlled by the

limits for structurally similar compounds and the EPA

did not have sufficient POTW removal data to fix limits

for the other pollutants.

C. Application of PSES to Small Dischargers:

1. The EPA’s decision not to exempt small indirect dis-

chargers from PSES and not to establish a different stand-

ard for these dischargers was reasonable; the record

supports the EPA’s determination that the economic im-

pact of PSES on these dischargers was not so dispropor-

tionate to the impact on the industry as a whole as to

require a separate subcategory.

2. The Synthetic Organic Chemical Manufacturers As-

sociation (SOCMA) failed to demonstrate that the EPA’s

cost projection for PSES was not reasonable.

3. The EPA reasonably rejected SOCMA’s proposal that

the EPA regulate only small plants producing large dis-

charges; such a plan would leave substantial pollutant dis-

charges unregulated.

D. Application of PSES to Paint/Resin Plants:

1. The EPA's reliance on data from resin plants in pro-

mulgating PSES does not make the standards inapplicable

to combined paint/resin plants; the EPA found that the

wastestreams of resin plants and paint resin plants were

equally treatable.

2. The National Paint and Coatings Association has

failed to demonstrate that PSES as applied to paint/resin

plants are not economically achievable; the EPA identified

cost-effective means of segregating the resin stream from

lll

lla

such plants thereby making separate regulation of the resin

stream possible.

3. The EPA’s regulation of the resin discharge of com-

bined paint/resin plants does not violate the NRDC v. Train

consent decree which exempts paint manufacturers from

regulation.

4. The regulations make it clear which resin manufac-

turers are subject to PSES.

E. The EPA’s Decision Not to Subcategorize on the

Basis of POTW Removal Credits:

1. The EPA is not required to subcategorize on the basis

of POTW removal credits; Congress intended that credits

for a POTW’s pollutant removal would be available only

after the EPA completes its comprehensive sludge regu-

lations.

2. Because Congress has suspended the removal credits

regulations, this court need not address the claims of Gulf

Coast Waste Disposal Authority and the Village of Sauget

POTWs that they supplied the EPA with sufficient data

to support their claims for removal credits or subcatego-

rization.

3. Gulf Coast Waste Disposal Authority is bound by the

consent decree in United States v. Crown Central Petro-

leum Corporation which requires Gulf Coast to recognize

and enforce the pretreatment standards.

V. New Source Performance Standards SPS) Issues:

1. The EPA’s application of BPT- and BAT-costing

methods to NSPS was reasonable.

2. The EPA’s failure to consider wastestream recycling

as a model technology in promulgating NSPS was arbi-

trary; the record indicates that recycling is a well-dem-

onstrated technology resulting in substantially higher

pollutant removal.

12a

VI. Montreal Protocol Issue:

A. The EPA’s failure to consider the economic conse-

quences of the ‘Montreal Protocol on Substances that De-

plete the Ozone Layer’ in the cost analysis of the

chlorofluorocarbon industry was not arbitrary or capri-

cious, because the issue was never brought to the EPA’s

attention and at the time the regulations were promul-

gated the Montreal Protocol was not yet in force.

VII. Issues Concerning the Scope of the OCPSF Limi-

tations:

1. This court is without jurisdiction to review NRDC’s

claim that the EPA’s reservation of nonconventional pol-

lutants and eight priority pollutants for future rulemaking

was contrary to the CWA; such a challenge must be raised

in the first instance in the district court.

2. The EPA reasonably concluded that the OCPSF lim-

itations apply to research as well as manufacturing dis-

charges.

13a

Table of Contents

Page

Introduction:

Statutory Background ...................0+ 195 [17a]

Standards of Review ................:.:000. 197 [23a]

I. Procedural Challenges ...............:::0:0cc008 200 [29a]

1. The EPA’s Economic-Impact

A ey 200 [29a]

2. Limits on Metal-Bearing

IS lac cakeephidencnsescsssceses 202 [34a]

II. Best Practicable Technology (BPT)

RENE SEE SS 203 [36a]

A. The EPA’s Consideration of the Costs

of Complying with the BPT

EEE SE 204 [37a]

1. CMA’s Challenge Based on the

“Knee-of-the-Curve”’ Cost-Effective-

ante 204 [39a]

2. The BCT Cost-Effectiveness Test. 206 [42a]

3. The EPA’s Projected Costs of the

ng fe 207 [44a]

B. The EPA’s Definition of the BPT Data

SS ae 207 [45a]

1. The EPA’s Determination of the

“Average of the Best’

RT tiitcticcssteecsssevencocosess .. 207 [45a]

2. The EPA’s Rejection of Sequential

Treatment Options ...............00000+ 208 [47a]

C. The “Summer/Winter Issue’: The

EPA’s Decision Not to Subcategorize

Based on Climatic Differences .......... 210 [51a]

1. Diversity of the Data Base ........... 211 [52a]

2. Winter Removal Efficiencies ......... 213 (56a)

TI os csscnacecnnee "214 [59a]

l4a

1. CMA’s Challenge to the BPT

SuabcategortsAtioe ....cccccccesscscsesesesee

2. NRDC’s Notice and Comment

CRATE cesssvcnisinenunininaianisilanaiieetilies

POnGS ..ccorsssrengipensaunneueaaaaaaaaai

1. The EPA’s Determination that Pond

Algae were Conventional Pollutants

and Thus Subject to BPT

Reg baths .csncssessnscssensniniemanvennens

2. The EPA’s Decision Not to Create

a Subcategory for Plants Utilizing

Waste-Stabilization Ponds ..............

3. The EPA’s Consideration of the

Cost of Compliance for Plants Uti-

lizing Pond Technology .............+++.

4. Copper Sulfate Treatment for Algae

Contrel ..cccsssteiaciaeeamannne

5. The Costs of the BPT Limitations

F. Plant-Specific Challenges to the BPT

Limitations and the Availability of

Fundamentally-Different-Factor

Var IRCOD .cccesscnctsinnesgaseniammaanninas

1. Unio Capt ccccccccssoccsstncennseanersane

2. Borg Wate? occcccocscssonsnnnnsmsccsonnsonn

3. DuPont’s Chambers Works Plant .

4. Monsanto and Ethy) .................:000

5. FDP Ver iit ccccssctescsemeteteen

III. Best Available Technology (BAT) Issues

A. The EPA’s Statistical Method of De-

veloping the BAT Limitations ...........

1. Use of Weighted Averaging in De-

riving the Long-Term Averages ...

2. Averaging of Variability Factors ..

B. Remedy for Unavoidable

EE. xCCCUGIROOB ccnccoccesesiscsmssuinndionmeninmnian

C. The EPA’s Analytical Methodology ...

214 [59a]

216 [64a]

217 [65a]

217 [66a]

218 [68a]

219 [70a]

220 [72a]

221 [73a]

221 [74a]

222 [77a]

223 [78a]

223 [78a]

223 [79a]

225 [82a]

226 [86a]

227 [87a]

227 [87a]

228 [89a]

228 (90a)

230 [94a]

15a

1. Analytical Variability ...............0.... 230 [94a]

2. Complex Wastestreams ................. 230 [95a]

3. Computation of Variability Factors

When Data Were Insufficient (‘‘Bor-

EE acs cnctecssstonsevenosvonss 231 [96a]

4. CMA’s Challenge to the EPA’s

Choice of the Minimum Analytical

Value to Assign to Non-Detect

i listincshsnesebsenssnesecsesessee 231 [97a]

5. Changes in Analytical Methods .... 233 [99a]

D. Use of Minimum Analytical Values for

Enforcement Purposes .................:000+ 233 [100a]

E. The EPA’s Sampling Techniques ...... 233 [101a]

F. The EPA’s Toxic Limitations ............ 234 [102a]

1. Application of the Toxic Limitations

to All OCPSF Dischargers ........... 234 [102a]

} 2. Courtaulds’ Notice-and-Comment

IT Le iaschescsninrinekevcsvecsvessesne 235 [104a]

3. The NRDC v. Train Consent

Ee 235 [104a]

G. NRDC’s Challenge to the EPA’s BAT,

and BAT, Subcategorization .............. 235 [105a]

H. DuPont’s Chambers Works Plant ..... 236 [107a]

I. BAT Limitations for Phenol ............... 236 [108a]

1. Phenol-Dominated Wastestreams .. 236 [108a]

2. The EPA’s Cost Estimates for the

Phenol Limitations ......................... 237 [109a]

J. The BAT, Limitations for Volatile Pol-

lutants Based on Steam Stripper

EE ESSE Oo 238 [1lla]

1. Achievability of the Limitations ... 238 [111la]

2. Wastestream Characteristics ......... 239 [113a]

3. Steam Stripper Maintenance ......... 239 [1l4a]

4. Dow Chemical’s Plant-Specific

ili ietssitessosesseopsteneseees 240 [115a]

l6a

K. The BAT, Limitations for Priority Pol-

lutants Based on In-Plant Biological

RIE, ccisensscinintdidpienceempieetnemanaes 240 [115a]

1. Achievability of the Limitations ... 240 [115a]

2. The EPA’s Estimate of the Costs

of Complying with the Limitations 241 [117a]

SR CI ceca eecsnccenebevcinccecnens 241 [118a]

L. Compliance Deadline ...............cseseseeees 242 [119a]

1. Lack of Sufficient Lead Time for

Industry Members to Comply ....... 242 [119a]

2. “Availability” of BAT Technology 243 [121a]

IV. Pretreatment Standards for Existing

Sources (PSES) Issues ..............cccccossssees 243 [122a]

A, “FR ROE TID creneverecticcsccsines 243 [122a]

1. The EPA’s Definition of “Pass

I * nekakasinsianiencslnticceninantenih 243 [122a]

2. The EPA’s Methodology in Assum-

ing an Absence of ‘Pass Through”

Based on POTW Averages ........... 247 [129a]

3. The EPA’s Decision Not to Find

“Pass Through” Based on Sludge

CITI eccansdearntiecien testers 247 [131a]

B. The EPA’s Decision Not to Regulate

Six Volatile Organic Pollutants on the

Basis of Interference with POTW

WEIN TI eecincncninscsctnnnnthinmmabsensts 248 [132a]

C. Application of PSES to Small Plants 249 [134a]

1. Economic Impact of PSES to Small

SOIR i scidustnininsnssntieenninbetioimeaniads 249 [135a]

2. Validity of the EPA’s Cost

RII sessticisnstasiciainncsinciine 252 [141a]

3. Alternative Pretreatment

IIE: sasnrincsncstesstniniitaniiisienieanins 252 [142a]

D. Application of PSES to Paint/Resin

PRI siincntesinciniinlecihtpidheinnmninuidamanaiaia 253 [143a]

1. The EPA’s Data Base ..............000 253 [144a]

17a

; 3. The NRDC v. Train Consent

; I iticenncstseerscingicecinveiinsesenseatons 256 [151a]

| 4. Subcate*orisation ..........cesrseserereseeee 257 [15la]

E. EPA’s Decision Not to Subcategorize

on the Basis of POTW Removal

I ails aceeatiniariainsicincininccinamentiyienes 257 [(153a]

| ee Been 258 [154a]

2. GCWDA and Village of Sauget

. EE Saarinen 261 [159a]

3. The Crown Central Petroleum Cor-

poration Consent Decree ............... 261 (160a]

V. New Source Performance Standards

SRF atid sce gednannaassantmiasdesenauncinerseeniey ci 261 [161a]

1. The EPA’s Cost Test in Establish-

RITE ciiciicasusecscuccensecixiabomancaness 262 [162a]

. 2. NRDC’s Challenge that the EPA

: Failed to Consider Technology Be-

WORE TREE GE TRAD ciceciccseressecserse 262 [164a]

VI. The Economic Impact of the Montreal

Protocol on the Chlorofluorocarbon

aa ha Sipe 264 [166a]

VII. Scope of the Regulations .................00. 265 [169a]

1. The EPA’s Reservation of Noncon-

ventional Pollutants and Eight

Priority Pollutants for Future

ERIE En eR 265 [169a]

2. Application of the Regulations to

Laboratory Discharges .................. 266 [171a]

I a cctetsceniiia eicinlentnsatninndnadanedakinassviansones 266 [172a]

Statutory Background

The purpose of the Clean Water Act (CWA) is to restore

and maintain the chemical and biological integrity of the

18a

nation’s waters.’ It was adopted to effectuate Congress’s

declared ‘‘national goal that the discharge of pollutants

into the navigable waters be eliminated by 1985,’ a date

later extended to March 31, 1989,5 and to prohibit ‘“‘the

discharge of toxic pollutants in toxic amounts.’

Until 1972, the federal government relied primarily on

state and local action to accomplish federal pollution-abate-

ment goals.’ Congress became dissatisfied, however, with

the division of responsibility for setting standards between

federal and state water-pollution-control agencies, with the

EPA’s dilatory pace, and with the ponderous federal en-

forcement procedure. Consequently, it enacted the Federal

Water Pollution Control Act (FWPCA) Amendments of

1972,° which imposed greater federal regulatory respon-

sibilities and set deadlines for the completion of limitations

on pollutant effluents.

Because the EPA had failed timely to fulfill its respon-

sibilities under the 1972 Act, the Natural Resources De-

fense Council (NRDC) filed suit in NRDC v. Train

challenging the EPA’s failure to promulgate the effluent

standards mandated for toxic substances. The consent de-

cree entered into by the EPA, NRDC, and various indus-

trial intervenors in that suit established a schedule for

Agency promulgation of effluent limits, new source stand-

* 33 U.S.C. § 1251(a).

* 33 U.S.C. § 1251(aX1).

* 33 U.S.C. § 1311(d).

* 33 U.S.C. § 1251(aX3).

’ 1 Grad, Treatise on Environmental Law § 3.03 at 3-71 (1988).

* Id., § 3.03 at 3-78; Congressional Record, House Debate, Dec. 15,

1977, Report on Resolution Providing for Consideration of Conference

Report on H.R. 3199, Clean Water Act of 1977, published in 3 Leg-

islative History of the Clean Water Act of 1977, 299, 327 (1978), here-

inafter referred to as ‘Legislative History.”

ee

PBR AB 0 ps

Tre ta

19a

ards, and pretreatment standards for priority toxic pol-

lutants.°®

Recognizing the growing seriousness of the problems

created by toxic pollution and the inadequacy of the 1972

FWPCA to deal with them,’® Congress amended that Act

by adopting the CWA in 1977." As thus amended, the

FWPCA contains several distinct, though interlocking, reg-

ulatory schemes.

First, Title II of the Act encourages the construction

of publicly-owned waste-treatment works by providing fed-

eral grants-in-aid to states for the construction of such

plants. As a condition of receiving grants-in-aid, the Act

requires states to establish area-wide management agen-

cies with both planning and regulatory functions for waste

treatment.

The second regulatory feature of the FWPCA, as set

out in Title III, authorizes the EPA to set and to enforce

federal effluent standards. This part of the 1977 statute

codifies the toxics consent decree issued by the United

States District Court for the District of Columbia in which

it required the EPA to develop BAT effluent guidelines

by July 1, 1980 for 65 toxic pollutants listed in the decree.

The Act also requires the EPA to promulgate pretreat-

ment standards for indirect dischargers based on BAT or

more stringent criteria.‘

* 8 Env’t Rep. Cas. (BNA) 2120 (D.D.C.1976), modified sub. nom.

NRDC v. Costle, 12 Env’'t Rep. Cas. (BNA) 1833 (D.D.C.1976), modified

sub. nom. NRDC v. Gorsuch, No. 72-2153 (D.D.C.1982), modified sub.

nom. NRDC v. Ruckelshaus, No. 73-2153 (D.D.C. Aug. 2, 1983).

” Leg. Hist. 326-27, 456-47, 862-65.

“ Pub.L. No. 95-217, 91 Stat. 1566 (1977).

* FWPCA §§ 201-08, 211-12, 33 U.S.C. §§ 1281-88, 1291-92. See

Grad, supra, § 3.03 at 3-113.

* 33 U.S.C. § 1317(a).

20a

Third, the Act requires the states to establish water-

quality criteria and to set ambient quality standards for

each of their rivers, subject to EPA approval. This con-

tinues the procedures under prior law pursuant to which

the states submitted their ambient water standards for

federal approval, submitted their own effluent limitations

designed to meet these standards, and had responsibility

for enforcement of the limitations."

To monitor compliance with the pollutant effluent lim-

itations, the Act establishes a system for issuing pollution

permits called the National Pollutant Discharge Elimina-

tion System (NPDES).'* Each discharger, including

POTWs, must obtain a permit from the EPA. No permit

may be issued unless the effluent to be discharged meets

federal effluent standards.’®

As Professor Frank P. Grad has noted in his Treatise

on Environmental Law, these four “regulatory schemes

... are limited ... by a variety of general provisions deal-

ing with enforcement, citizen suit, judicial review, and the

like.’’!”

Congress again amended portions of the Act in 1987 to

extend the compliance dates prescribed in the 1977 Act

from July 1, 1984, to March 31, 1989; to permit modifi-

cation of secondary treatment requirements to alter the

guidelines applicable to facilities that are fundamentally

different; and to effect other changes. Our references in

this opinion to the Act refer to its provisions as finally

amended in 1987. .

The Act requires direct dischargers to comply with tech-

nology-based pollutant-effluent limitations that, in time, will

* Grad, supra, § 3.03 at 3-115.

'® 33 U.S.C. § 1342.

* 40 C.F.R. Part 144.

7 Grad, supra, § 3.03 at 3-116.

en ea en ee ees

2la

become more stringent.'* First, it orders all direct dis-

chargers of conventional pollutants to comply with effluent

limitations achievable by application of the “‘best practic-

able control technology presently available’ (BPT) by July

1, 1977. Second, it orders all direct dischargers of con-

ventional pollutants to comply by March 31, 1989 with

effluent limitations based on a more exacting standard,

the “‘best conventional pollution control technology”’ (BCT).

It mandates in addition that, by the same date, direct

dischargers of toxic pollutants must comply with the even

more rigorous effluent limitations based on the “‘best avail-

able technology economically achievable’ (BAT).'®

The EPA must determine the BPT, BCT, and BAT re-

quirements and announce them in regulations establishing

“effluent limitations guidelines’ for various classes and

categories of dischargers. In establishing each set of stand-

ards Congress required the EPA to consider a number of

factors including costs, although the cost factor is accorded

less weight for facilities not yet constructed and for dis-

charges more harmful to the environment.

New plants constructed after the promulgation of the

OCPSF Guidelines that discharge directly into navigable

waters are subject to separate standards referred to as

“new source performance standards’’ (NSPS). The new

source performance standards are based on the “‘best avail-

atle demonstrated control technology’’ (BADCT) as identi-

fied by the EPA.”

Indirect dischargers rely on POTWs to treat their waste-

waters, and the Act requires the EPA to set effluent lim-

itations for POTWs engaged in the treatment of municipal

sewage or industrial wastewater.” Although the POTW

* 33 U.S.C. §§ 1311(b), 1314(b).

* 33 U.S.C. § 1314(d).

*» 33 U.S.C. § 1316.

™ 33 U.S.C. §§ 1311(0X1KBHC), 1314(dX1).

22a

requirements are determined by separate regulations, they

must be based on BAT to ensure that all final dischargers

of toxic pollutants meet the same standards. The treatment

usually accorded by POTWs, however, may not remove all

pollutants discharged into their facilities by industrial users

and their operation of these facilities may be damaged by

some industrial discharges. Therefore, to regulate the dis-

charge into POTWs of those pollutants determined “not

to be susceptible to treatment” by POTWs or likely to

“interfere with the operation” of POTWs,” the Act re-

quires the EPA to establish “pretreatment standards for

eristing sources” (PSES). These standards must also be

based on BAT.”

The OCPSF limitations are technology-based and apply

to plants grouped into categories based on their industrial

characteristics. The EPA, with the concurrence of the state

affected, may establish special provisions for a facility that

is fundamentally different with respect to one or more of

the factors relevant in developing the regulations other

than cost. Thése provisions, which are known as “‘funda-

mentally different factor’’ (FDF) variances, are intended

to adjust the general limitations and provide different ones

for a plant whose individual characteristics prevent it from

performing within the limits set for its industrial cate-

gory.“ An FDF variance application may be based on sup-

porting data submitted to the EPA during the rulemaking

process or on information that the applicant did not have

a reasonable opportunity to submit at that time.”

= 33 U.S.C. § 1317.

* 33 U.S.C. § 1317(b\1).

» 40 CFR. Part 125 Subpt. D.

* 33 USC. § 1311(nx(1).

23a

The regulatory process is not static. Various provisions

of the Act require the EPA to review the guidelines pe-

riodically and to revise them when appropriate.”

For the past eleven years the EPA has conducted stud-

ies and rulemaking proceedings for the purpose of estab-

lishing OCPSF pollutant-effluent limitations. The EPA has

identified model technologies that in its view satisfy the

development criteria for BPT, BAT, NSPS, and PSES.

Based on these model technologies, the EPA has deter-

mined treatment performances and has established effluent

limitations for conventional and toxic pollutants within the

range of the performances achieved by the model tech-

nologies. The OCPSF limitations provide maximum daily

and maximum monthly average limits for the discharge of

designated pollutants from each ‘‘point source,” that is,

each discharge pipe.”’

Standards of Review

The Administrative Procedure Act (APA) fixes the

standard for appellate review of agency actions. Agency

actions may be set aside only if the ‘‘agency action, find-

ings, and conclusions [are] found to be ... arbitrary, ca-

pricious, an abuse of discretion, or otherwise not in

accordance with law”; or “in excess of statutory jurisdic-

tion, authority, or limitations, or short of statutory right’;

or “without observance of procedure required by law.’’™

In making its determinations, ‘‘the court shall review the

whole record or those parts of it cited by a party, and

due account shal] be taken of the rule of prejudicial er-

ror.”’*?

™ See, ¢.g., 33 U.S.C. § 1314(b) (annual revision). § 1316(b) (revision

“from time to time’); § 1317(aX3) (revision “every three years’)

" 52 Fed.Reg. 42,522. (codified at 40 C.F.R. Part 414).

= 5 U.S.C. § 706.

" 6 U.S.C. § 706, q 2.

24a

The many issues raised on appeal concern three aspects

of the EPA’s actions, each governed by different stand-

ards: (1) its rulemaking procedures; (2) its interpretation

of the Clean Water Act; and (3) the validity of its regu-

lations and its actions to enforce the Act.

The standards of review for each of these types of issues

have been stated and restated in a host of cases, not

always consistently. The most frequently stated verbal for-

mulae are these:

Review of the validity of the challenged procedure is a

question of law governed principally by the APA, under

which our review is plenary. In determining validity, the

Administrator's decision is entitled to a presumption of

regularity.” A party petitioning for review of an agency's

regulations bears the burden of overcoming this presump-

tion.”

When we turn to the EPA’s interpretation of the stat-

ute, our review is again plenary, for we may not accept

its interpretation if ‘contrary to Congress’s intentions as

revealed by the Act’s language, structure, and legislative

history.’’* In statutory interpretation, the judiciary is “ ‘the

fina] authority and we must reject administrative construc-

tions which are contrary to clear congressional intent.’ ’”

Nevertheless, we accord some deference to the Agency’s

interpretation of the statute whose enforcement is en-

» Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402,

415, 91 S.Ct. 814, 823, 28 L.Ed.2d 136 (1971).

" See Louisiana Environmental Society v. Dole, 707 F.2d 116, 119

(5th Cir.1983).

" Exzon Corp. v. Train, 554 F.2d 1310, 1322 (5th Cir.1977); see

Sverra Club v. Train, 557 F.2d 485, 489 (5th Cir.1977).

" Immigration and Naturalization Service v. Cardoza-Fonseca, 480

US. 421, 107 S.Ct. 1207, 94 L.Ed.2d 434 (1987) (quoting Chevron,

USA, Inc. v. NRDC, 467 U.S. 837, 843 n. 9, 104 S.Ct. 2778, 2781 n.

9, 81 L.Ed.2d 419 (1985).

4

——_ eee

tlle lilly,

25a

trusted to it if Congressional intention is not pellucid.™ If,

therefore, the statute is susceptible to more than one inter-

pretation, we must accept that of the EPA if it is rea-

sonable.* We need not find that it is the only permissible

interpretation, but merely that the ‘““EPA’s understanding

of this very ‘complex statute’ is a sufficiently rational one

to preclude a court from substituting its judgment for that

of the EPA.”™ According deference to an agency, how-

ever, does not imply ‘‘rubber stamping”’ its decision.”

In interpreting the Act, we do not lack precedent. The

Supreme Court has twice considered issues involving its

construction,” and eight circuit courts have done so in at

least twenty-seven cases.**

* EPA v. National Crushed Stone Ass'n.449 U.S. 64, 83. 101 L.Ed.2d

268 (1980), United States v. Riverside Baymew Homes, Inc., 474 U.S

121, 106 S.Ct. 455, 88 L.Ed.2d 419 (1985)

™ Chevron, USA, Inc. v. NRDC, 467 U.S. 837, 844, 104 S.Ct. 2778

2782, 81 L.Ed.2d 694 (1987); see also L’nited States v. Ethyl Corp., 761

F.2d 1153, 1157 (5th Cir.1985), cert. denied, 474 U.S. 1070, 106 S.Ct

830, 88 L.Ed.2d 801 (1986); Teras v. United States, 756 F.2d 419, 425

(Sth Cir.), cert. denied, 474 U.S. 843, 106 S.Ct. 129, 88 L.Ed.2d 106

(1985), City of Seabrook v. EPA, 69 F.2d 1349, 1354 (5th Cir.1981).

cert. denied, 459 U.S. 822, 103 S.Ct. 51, 74 L.Ed.2d 57 (1982)

™ Chemical Mfrs. Ass'n v. NRDC, 470 U.S. 116, 125, 105 S.Ct. 1102.

1107, 84 L.Ed.2d 90 (1985); see also Chevron v. NRDC, 467 U.S. 842-

45, 104 S.Ct. at 2781-83.

" Bureau of Alcohol, Tobacco and Firearms v. Federal Labor Re

lations Authority, 464 U.S. 89, 104 S.Ct. 439, 444, 78 L.Ed. 2d 195

(1983), NLRB v. Brown, 380 U.S. 278, 291, 85 S.Ct. 980, 988m 13

L.Ed.2d 839 (1965).

™ Chemical Mfrs. Ass'n v. NRDC, 470 US. at 116, 105 S.Ct. at

1102; EPA wv. National Crushed Stone Ass'n, 449 U.S. at 64, 101 SCt

at 295.

" See Armoo, Inc. v. EPA, No. 88-3070, ___ F 2d slip op. (6th

Cir. March 15, 1989); NRDC v. EPA, 863 F.2d 1420 (9th Cir.1989) (oil

and gas industry); American Petroleum Inst. v EPA, 858 F.2d 261

(Sth Cir.1988) (oil and gas industry), Tezas Mumicipal Power Agency 1

EPA, 836 F.2d 1482 (5th Cir.1988) (electric utility industry), Amencan

26a

After interpreting the statute according to these prin-

ciples, we must next determine whether the EPA’s find-

ings or the regulations based on them are “arbitrary or

capricious.” In doing so we must conduct a “searching

Petroleum Inst. v. EPA, 787 F.2d 965 (5th Cir.1986) (oil and gas in-

dustry); Kennecott v. EPA,780 F.2d 445 (4th Cir.1985) (nonferrous met-

als manufacturing industry), cert. denied, 479 U.S. 814, 107 S.Ct. 67,

93 L.Ed.2d 25 (1986); Cerro Copper Products Co. v. Ruckelshaus, 766

F.2d 1060 (7th Cir.1985) (copper-forming industry); Reynolds Metals Co

EPA, 760 F.2d 549 (4th Cir.1985) (metal and brewing industries);

National Ass'n of Metal Finishers v. EPA, 719 F.2d 624 (3d Cir.1983)

(electroplating industry), rev'd on other grounds sub nom. Chemical

Mfrs. Ass'n v. NRDC, 470 U.S. 116, 105 S.Ct. 1102, 84 L.Ed.2d 90

(1985); American Petroleum Inst. v. EPA, 661 F.2d 340 (5th Cir.1981)

(oil and gas industry); American Paper Inst. v. Train, 660 F.2d 954

(4th Cir.1981) (paper industry); Association of Pacific Fisheries v. EPA,

615 F.2d 794 (9th Cir.1980) (seafood processing industry); BASF Wyan-

dotte Corp. v. Costle, 598 F.2d 637 (1st Cir.1979) (organic pesticide

industry), cert, denied, 444 U.S. 1096, 100 S.Ct. 1063, 62 L.Ed.2d 784

(1980), Weyerhaeuser Co. v. Costle, 590 F.2d 1011 (D.C. Cir.1978) (pulp

and paper industry)’ American Iron & Steel Inst. v. EPA, 568 F.2d

284 (3d Cir.1977) (iron and steel industry); Marathon Oil Co. v. EPA,

564 F.2d 1253 (9th Cir.1977) (oil and gas industry); American Paper

Inst. v. Train, 543 F.2d 328 (D.C. Cir.1976) (paper and pulp industry),

cert. dismissed, 429 U.S. 967, 97 S.Ct. 398, 50 L.Ed.2d 335 (1976);

E.1. du Pont de Nemours & Co. v. Train, 541 F.2d 1018 (4th Cir.1976)

(inorganic chemical manufacturing industry), rev'd in part on other

grounds, 430 U.S. 112, 97 S.Ct. 965, 51 L.Ed.2d 204 (1977); Tanners’

Council of America, Inc. v. Train, 540 F.2d 1188 (4th Cir.1976) (leather

tanning industry); American Petroleum Inst. v. EPA, 540 F.2d 1023

(10th Cir.1976) (oil and gas industry),cert. denied, 430 U.S. 922, 97

S.Ct. 1340, 51 L.Ed.2d 601 (1977); FMC Corp. v. Train, 539 F.2d 973

(4th Cir.1976) (plastic and synthetic materials industries), American

Frozen Food Inst. v. Train, 539 F.2d 107 (D.C.Cir.1976) (frozen potato

products); Hooker Chemicals & Plastics Corp. v. Train, 537 F.2d 620

(2d Cir.1976) (phosphorus manufacturing); American Iron & Steel Inst.

v. EPA, 526 F.2d 1027 (1975) (iron and steel industry), cert. dented,

435 U.S. 914, 98 S.Ct. 1467, 55 L.Ed.2d 505 (1978); American Meat

Inst. v. EPA, 526 F.2d 442 (7th Cir.1975) (meat products), CPC Int'l,

Inc. v. Train, 515 F.2d 1032 (8th Cir.1975) (corn wet milling); Portland

Cement Ass'n v. Ruckelshaus, 486 F.2d 375 (D.C.Cir.1973) (cement man-

ufacturers industry), cert. denied, 417 U.S. 921, 94 S.Ct. 2628, 41

L.Ed.2d 226 (1977).

— +

and careful review” of the facts to determine whether ‘the

decision was based on a consideration of the relevant fac

tors and whether there has been a clear error of judg

ment.’ As a reviewing court, however, we must n

“substitute [our] judgment for that of the agency, *' but

must “start with the assumption that the agency's action

‘s valid.’ The court’s proper function is only to determine

whether the agency has “considered the relevant factors

and articulated a rational correlation between the fact

found and the choice made.’ *

An agency rule is “arbitrary’’ if the agency has re

on factors which Congress has not intended itt nsider

entirely failed to consider an important aspect | f the pr

lem, offered an explanation for its decision that ru

counter to the evidence before the agency, or !s

plausible that it could not be ascribed to a difference

view or the product of agency expertise. ©

Because ‘“‘technological and scientific issues such as those

presented in [reviewing effluent limitations] are by their

very nature difficult to resolve by traditional principles f

judicial decisionmaking,’ the court ~ ‘must look at @

[agency’s] decision not as the chemist, biologist or stat

istician that [it is] qualified neither by training nor ex

” Citizens to Preserve Overton Park, Inc Volpe, 401 U.S. at 4l¢

91 S.Ct. at 824.

“ Id.

“ Avoyelles Sportsmen's League, Inc. ' Marsh, 715 F.2d 897, 904

(5th Cir.1983).

*@ Baltimore Gas & Electric Co. v. NRDC, 462 US. 87

S.Ct. 2246, 2256, 76 L.Ed.2d 437 (1983)

“ Motor Vehicle Mfrs. Ass'n v. State Farm Mut. Automotve Ins

463 U.S. 29, 43 103 S.Ct. 2856, 2867, 77 L.Ed.2d 443 (1953)

“ Reynolds Metals Co. v. EPA, 70 F.2d 549, 558- 559 (4th Cir.1985

28a

perience to be, but as a reviewing court exercising ...

certain minimal standards of rationality.’ ’”’“

Even with regard to complex technical or scientific de-

cisions, however, a reviewing court may not simply defer

to an agency’s expertise, but must ‘“‘steep’’ itself in tech-

nical matters sufficiently to determine whether the agency

“has exercised reasoned discretion.’’*’ Because judicial re-

view ‘‘must be based on something more than trust and

faith in EPA’s experience,’ a court may not respond to

claims of technica] expertise by “rubber stamping’ an

agency decision as correct.

These formulae, however, ultimately may prove to be

deceptive guides. The Supreme Court’s decisions seem to

embody two different approaches that are, ‘‘analytically in

conflict ‘with the result that a court of appeals must choose

the one it deems more appropriate for the case at hand.’ ’’”

In determining the degree of deference appropriate to an

agency’s decision, the factors to be considered include the

fact-law dichotomy, despite the difficulty of drawing a pre-

“ American Paper Inst. v. EPA, 660 F.2d 954, 963 (4th Cir.1981)

(quoting Ethyl Corp. v. EPA, 541 F.2d 1, 36 (D.C.Cir.1976)).

« Portland Cement Ass'n v. Ruckelshaus, 486 F.2d 375, 402 n. 20

(D.C.Cir.1973), cert. denied,417 U.S. 921, 94 S.Ct. 2628, 41 L.Ed.2d

226 (1974), see South Terminal Corp. v. EPA, 504 F.2d 646, 663-66

(1st Cir.1974).

“ Appalachian Power Co. v. Train, 545 F.2d 1351, 1365 (4th

Cir.1976).

“ American Petroleum Inst. v. EPA, 661 F.2d at 348-49.

“ H. Craft Clothing Co. v. NLRB,660 F.2d 910, 913-914 (3d Cir.1981)

(quoting Pittston Stevedoring Corp. v. Dellaventura, 544 F.2d 35, 49

(2d Cir.1976), aff'd sub nom. Northeast Marine Terminal Co. v. Caputo,

432 U.S. 249, 97 S.Ct. 2348, 53 L.Ed.2d 320 (1977)). See also R. Pierce,

S. Shapiro, and P. Verkuil, Administrative Law & Process (1985) § 7.5

at 377.

29a

cise line between fact and law;*! whether the question to

be decided is one concerning which the courts have a spe-

cial competence, such as constitutional law, or one that

turns on technical expertise of the kind the agency staff

possess; whether the issue turns on the agency’s inter-

pretation of its own authorizing statute, which entitles it

at least to guarded deference;* and the validity of the

reasoning upon which the agency relies to justify its ac-

tions.* _

Either the application of these formulae or the balancing

of factors praised by scholars leaves the reviewing court

with a wide margin for decision. Any approach permits

different possible decisions on the various issues. Our duty

requires us to resolve them with the good judgment that

the latitude afforded the court demands. In performing

this duty, we recognize our own limitations in assessing

policy decisions.*

I. Procedural Challenges

1. The EPA’s Economic-Impact Study

The Clean Water Act explicitly requires the EPA to

consider economic achievability in formulating pollutant ef-

fluent limitations.* The Chemical Manufacturers Associa-

tion (CMA) contends that, in preparing its economic-impact

study, the EPA departed from its prior practice during

" H. Craft Clothing Co.,660 F.2d at 914 (citing W. Gellham & C

Byse, Administrative Law 251-257 (7th ed. 1979)). see also K. Davis

Administrative Law of the Seventies § 80 (1976)

“ H. Craft Clothing Co., 660 F.2d at 915

“ Id. at 915-16.

“ Pierce, Shapiro, and Verkuil, Administrative Law and Process § 7

at 377.

“ Id., § 7.6 at 391-2; G. Robinson, E. Gellhorn, and H. Bruff, The

Administrative Process (3d ed. 1986) § 684.

“ 83 U.S.C. §§ 1314(0X1\B), 1314(0%2KB), and 1314(bx4¥B)

30a

the rulemaking proceedings and violated the APA’s notice-

and-comment requirement by relying on economic data ob-

tained from Dun & Bradstreet that were never made avail-

able to the public for comment.

The APA requires an agency to set forth in its notice

of proposed rulemaking ‘“‘the terms or substance ... or a

description of subjects and issues involved’’ in the proposed

rule.*’? The agency must then give interested persons an

opportunity to participate in rulemaking through criticism

and comments.* Although the APA does not specifically

require notice and comment on the technical data that an

agency considers, this court has recognized that fairness

requires that the agency afford interested parties an op-

portunity to challenge the underlying factual data relied

on by the agency.*®

If the final rules differ from the proposed rules or if

new data are considered after the agency receives com-

ments on the data it initially provides, the nature of the

change in the proposed rule or in the newly-considered

data determines whether it must again publish notice and

invite additional comments. As the District of Columbia

Circuit noted in Air Transport Association of America v.

Civil Aeronautics Board:

An agency adopting final rules that differ from its

proposed rules is required to renotice when the

changes are so major that the original notice did not

adequately frame tue subjects for discussion. ... The

agency need not renotice changes that follow logically

* 5 U.S.C. § 553(bX3).

* 5 U.S.C. § 553(c).

* See Air Products & Chemicals, Inc. v. FERC, 650 F.2d 687, 700

n. 17 (5th Cir.1981) (citing K. Davis, Administrative Law Treatise,

§ 7:17 at 530 (2d ed. 1978)). =

a

hae We i hed

3la

from or that reasonably develop the rules it proposed

originally”

In that case, the CAB had promulgated a final fee sched-

ule relying on internal staff studies that had not been made

available for comment but that had been prompted by prior

public comments that had “strongly challenged” calcula-

tions in the proposed fee schedule. The court found no

notice-and-comment violation because the agency had in its

notice of proposed rulemaking both outlined the method

that it proposed to use and indicated the type of actions

it proposed to take. The District of Columbia Circuit con-

cluded that “[t]hese critical elements of the proposal did

not change, and the final rule was a ‘logical outgrowth’

of the proposed rule.’’®

In Community Nutrition Institute v. Block® the same

court considered the situation in which the agency did not

make a change in the proposed rule but considered new

and unpublished data after giving notice. The court noted

that the Secretary of Agriculture had requested further

information on the subject of a proposed rule and that the

petitioner in response then pointed out a possible meth-

odological flaw in the studies relied on by the Secretary

but did not provide the requested information. The Sec-

retary then relied on two supplemental staff studies that

were developed after the close of the comment period to

address the alleged flaw. These studies were not made

available for comment. The court said: ‘‘Rulemaking pro-

ceedings would never end if an agency’s response to com-

® 732 F.2d 219, 224 (D.C.Cir.1984) (quoting Connecticut Light and

Power Co. v. Nuclear Regulatory Commission, 673 F.2d 525, 533

(D.C.Cir.), cert. denied, 459 U.S. 835, 103 S.Ct. 79, 74 L.Ed.2d 76

(1982)).

61 Td.

@ 749 F.2d 50, 57-58 (D.C.Cir.1984).

32a

ments must always be made the subject of additional

comments.’’®

In a 1986 notice of availability the EPA announced that

it intended to prepare an economic-impact study of the

OCPSF pollutant effluent limitations. The EPA stated that

in preparing the economic-impact study it would rely on

an industry-wide “‘FIN/STAT” data base that covered the

period from 1976-81.

Industry members commented that the FIN/STAT data

base was outdated and did not have adequate data for

plants whose sales exceeded $10 million annually. In the

preamble to the final rules the EPA announced that, in

response to industry comments, ‘‘the financial data base

used to calculate discounted cash flow and liquidation val-

ues for OCPSF plants in the impact analysis was changed

from FIN/STAT to Dun & Bradstreet,’’ but it did not

reveal the new Dun & Bradstreet data. These data, how-

ever, were edited in the same manner as the older Dun

& Bradstreet data used in the FIN/STAT study.™ The

EPA further stated in the preamble that it used the later

Dun & Bradstreet data to increase the size of the entire

data base, to increase the number of plants in the “greater

than $10 million sales’’ category (from 4 to 73), and to

update the data base to cover the period from 1981 to

1986.

The new Dun & Bradstreet data were used to estimate

plants’ cash flow, liquidation value, and profits. Although

these data were important in the economic-impact study,

the EPA also relied on data that had earlier been obtained

from other sources, including material from Robert Morris

Associates, COMPUSTAT, Data Resources, Inc., the Bu-

reau of Economic Analysis, the Bureau of Labor Statistics,

© Id.

« 42 Fed.Reg. 42,550.

% Id.

a ae ne

33a

the International Trade Commission, the Bureau of the

Census, chemical-industry periodicals, and surveys submit-

ted to plants under Section 308 of the CWA. The EPA

did not supplant its economic-impact study, or replace its

original data with completely new and different data, but,

in response to industry criticisms, updated and expanded

one of several data sources. Indeed, the data it originally

relied on were not so scanty that a reasonable rulemaker

could not have relied upon them.

The 1986 notice adequately advised interested parties of

the method the EPA had followed, the financial data it

proposed to rely on, and its intention to develop an eco-

nomic-impact study. The EPA’s use of the updated and

expanded Dun & Bradstreet data base was a logical and

reasonable development based on industry comments and

as such did not require further notice and comment.

In any event, CMA has failed to demonstrate that it

was prejudiced by the EPA’s use of the new Dun & Brad-

street data. The APA provides that a reviewing court

shall take due account of “‘the rule of prejudicial error.’’*’

The rule applies “when a mistake of the administrative

body is one that clearly had no bearing on the procedure

used or the substance of the decision reached.’ A

petitioner who objects to an agency’s failure to publish

data for comment must “indicate with ‘reasonable speci-

ficity’ what portions of the document it objects to and how

it might have responded if given the opportunity.’

The EPA offered to make the new Dun & Bradstreet

data available to CMA during the post-rulemaking period

* See Air Transport Ass’n, 732 F.2d at 224 n. 11; Community Nu-

trition Inst., 749 F.2d at 58.

" 5 U.S.C. § 706.

* U.S. Steel Corp. v. EPA, 595 F.2d 207, 215, reh’g granted in part

on other grounds, 598 F.2d 915 (5th Cir.1979).

* Small Refiner Lead Phase-Down Task Force v. EPA, 705 F.2d 506,

540-41 (D.C.Cir.1983) (citation omitted).

34a

on a confidential basis, explaining that it had failed to

publish the data because its contract with Dun & Brad-

street required it to keep the data confidential. While a

contractual provision of this sort cannot, absent unusual

circumstances, relieve an agency of its duty to publish

data, the EPA offer did make it possible for CMA to

review the data, to comment on it on an “in camera”

basis to this court, and to demonstrate that the data were

sufficiently inaccurate or misleading to prejudice the af-

fected industries. CMA declined the EPA’s offer. In re-

sponse to direct inquiry by this court, the only prejudice

CMA has suggested is that it lacked an opportunity to

show that the OCPSF effluent limitations were not eco-

nomically achievable. Because CMA was fully able to make

this showing even without the Dun & Bradstreet data, we

fail to discern any substantial prejudice from the EPA’s

use of the 1981-86 Dun & Bradstreet data to supplement

the other information on which it relied. We *herefore

decline to overturn the regulations because of the EPA’s

use of undisclosed supplementary economic data from Dun

& Bradstreet.

2. Limits on Metal-Bearing

Wastestreams

Appendix A to the effluent limitations establishes limits

for the discharge of toxic uncomplexed metals. Appendix

B lists ‘‘complexed metals,”’ that is, metals bonded with

an organic molecule, and states that limits for such pol-

lutants will be established on a case-by-case basis.” CMA

and DuPont content that Appendices A and B were never

presented to the public for comment, this violated the

APA’s notice and comment requirement, and the limita-

tions in these appendices therefore are unlawful.

In a 1985 public notice the EPA proposed BAT limits

for all metals found in OCPSF wastestreams. In response

© 52 Fed.Reg. 42,542-43.

35a

to EPA’s notice several industry members commented that

few OCPSF wastestreams contained metals and that the

model technology chosen by EPA to treat toxic metals

was not effective in removing complexed metals. In a 1986

public notice the EPA requested additional OCPSF indus-

try information concerning removal of complexed and un-

complexed metals, the source of metals in OCPSF

wastestreams, whether each metal is complexed or uncom-

plexed, and, for complexed metals, a list of the complexing

agents.” In Appendix A to the final rule, the EPA listed

industrial processes likely to produce metal-bearing waste-

streams and established effluent limitations for the dis-

charge of such metals. In Appendix B the Agency listed

the processes that produce complexed metals and stated

that such metals would be regulated on a case- by-case

basis when individual NPDES permits were issued.”

The APA notice requirement is. satisfied if the notice

fairly apprises interested persons of the subjects and issues

the agency is considering; “‘the notice need not specifically

identify ‘every precise proposal which [the agency] may

ultimately adopt as a final rule.’’’? Though the EPA did

not give the industry notice of Appendices A and B, the

1985 and 1986 notices informed the industry that the

Agency was considering establishing limitations for com-

plexed and uncomplexed metals, and this was all the APA

demands. Moreover, the changes reflected in the final rule

were instigated by industry comments and were based on

data supplied by the industry; thus the final rule was a

logical outgrowth of the comments received.

" §1 Fed.Reg. 44,090-91.

"™ 52 Fed.Reg. 42,542-43.

™ United Steelworkers of America v. Schuylkill Metals, 828 F.2d 314,

317 (5th Cir.1987) (citations omitted).

36a

Il. Best Practicable Technology

(BPT) Issues

The CWA authorizes the EPA to establish effluent lim-

itations for direct dischargers of conventional pollutants.

The Administrator is required to establish effluent limi-

tations for categories or classes of point sources discharg-

ing these pollutants based on the “‘best practicable control

technology currently available’ (BPT).* The CWA does not

specifically define BPT, but does identify factors that the

EPA should consider in determining it.

Section 304(b) of the CWA, as amended, states that in

assessing BPT, the EPA must consider:

the total cost of application of technology in relation

to the effluent reduction benefits to be achieved from

such application, and shall also take into account the

age of equipment and facilities involved, the process

employed, the engineering aspects of the application

of various types of control techniques, process

changes, non-water quality environmental impact .. .

and such other factors as the Administrator deems

appropriate... .”

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 the category or sub-

category for control of conventional pollutants.” In pro-

mulgating the regulations, the Agency identified a model

technology: biological treatment preceded by appropriate

process controls and in-plant treatment followed by sec-

ondary clarification as necessary to assure adequate con-

trol of solids.” CMA argues that the EPA’s data indicate

* 33 U.S.C. §§ 13110M1MA), 1314001 (A).

* 33 U.S.C. § 1314(o1B) (1986).

" See 52 Fed.Reg. 42,525.

" Jd. at 42,534.

37a

that its limitations will require the installation of additional

treatment equipment at a cost “wholly out of proportion”

to the marginal effluent reduction that the equipment

would achieve and that the limitations consequently fail to

meet the cost-effectiveness test required by Section

304(bX1\B) of the Act and the ‘‘best conventional tech-

nology” (BCT) test enacted in 1977. The EPA asserts that

the total cost of the BPT rules is justified by the total

amount of pollutant that would be removed.

A. The EPA's Consideration of the Industry's Costs of

Complying With the BPT Limitations

CMA maintains that the cost-effectiveness of BPT ru-

lemaking should be measured by a ‘‘knee-of-the-curve’’ test

to determine the point at which costs rise steeply per

pound of pollutant removed and that, under such a test,

the BPT rules are not cost- effective.

The CWA contains no specific statutory language es-

tablishing a BPT ‘‘knee-of-the-curve’’ test or any other

quantitative cost-benefit ratio test for BPT. The statute

simply requires that the EPA consider ‘‘the total cost of

application of technology in relation to the effluent re-

duction benefits to be achieved from such application.’*

The courts of appeal have consistently held that Congress

intended Section 304(b) to give the EPA broad discretion

in considering the cost of pollution abatement in relation

to its benefits and to preclude the EPA from giving the

cost of compliance primary importance.”

Senator Muskie, the principal Senate sponsor of the Act

and the Chairman of the Senate Subcommittee on Air and

Water Pollution, stated:

* 33 U.S.C. § 1314(bK1\B).

” American Iron & Steel Inst. v. EPA.526 F 2d at 1051, accord FMC

Corp. v. Train, 539 F.2d at 978-79; BASF Wyandotte Corp. 598 F 2d

at 656-57; Kennecott Copper Corp. v. EPA, 612 F.2d 1232, 1238 (10th

Cir.1979).

38a

The balancing test between total cost and effluent

reduction benefits is intended to limit the application

of technology only where the additional degree of ef-

fluent reduction is wholly out of proportion to the

costs of achieving such marginal level of reduction for

any class or category of sources.

The Conferees agreed upon this limited cost-benefit

analysis in order to maintain uniformity within a class

and category of point sources subject to effluent lim-

itations, and to avoid imposing on the Administrator

any requirement to consider the location of sources

within a category or to ascertain water quality impact

of effluent controls, or to determine the economic im-

pact of controls on any individual plant in a single

community.”

The EPA argues that the Administrator acted well

within this broad discretion in concluding that the costs

of the OCPSF BPT limitations were justified by the sig-

nificant quantities of pollutants that would be removed."

The EPA notes that the OCPSF industry is currently a

national leader in discharging conventional pollutants into

our nation’s waters. The industry has approximately 300

direct dischargers which annually discharge an estimated

61 million pounds of biochemical oxygen-demanding sub-

stances (BODS) and 100 million pounds of total suspended

~ 1972 Leg.Hist. at 170. See EPA v. National Crushed Stone Assoc.,

449 U.S. at 71 n. 10, 101 S.Ct. at 300 n. 10; American Frozen Food

Inst. v. Train, 539 F.2d at 119. The Ninth Circuit has interpreted this

legislative history to prohibit EPA from relying upon a cost-benefit

comparison to select a lower level of technology than BPT unless in-

creased costs would be wholly disproportionate to potential effluent

reduction benefits. Association of Pacific Fisheries v. EPA, 615 F.2d

at 805.

" 52 Fed.Reg. at 42,537. EPA also determined that no plants would

close, no product lines would be discontinued, and no jobs would be

lost as a result of implementing the BPT limits. Jd. at 42,551.

39a

solids (TSS) for a total estimate of approximately 161 mil-

lion pounds annually. The EPA estimated that the BPT

limitations would result annually in the removal of 108

million pounds of convertional pollutants from OCPSF dis-

charges and consequently from our nation’s waters™ at an

annualized compliance cost of 76.6 million dollars after a

capital investment of 215.8 million dollars.* Thus, the EPA

concluded that the total cost of BPT is warranted by the

total pounds of pollutant removed.

1. Knee-of-the-Curve Test

CMA argues that Congress was concerned generally that

the EPA’s regulations not require expenditures that would

pass the point at which costs escalate rapidly in relation

to benefits—the “‘knee-of-the-curve” on a diagram depict:

ing the cost curve. CMA conceives of the knee-of-the-curve

test as a generally applicable cost-effectiveness test with

the “knee” defining the most stringent level of regulation

permissible. Thus, CMA asserts, whether the EPA labels

its regulations BPT or BCT, the EPA is required to con-

sider whether the marginal costs exceed the marginal ben-

efits of the rule. Appiying this test, CMA argues that

increasing the removal of conventional! pollutants from 96

to 99 percent as required by the limitations would cost

the OCPSF industry almost twice as much per pound of

pollutant removed as current treatment methods: The an-

nual removal of 108 million pounds would cost 76 million

dollars per year—71 cents per pound—whereas industry

efforts to date have required an expenditure of only 3&

cents per pound. CMA concludes that the cost per pound

for removal of pollutants is thus well beyond the knee-of-

the-curve and that the regulations are therefore not cost-

effective.

” 52 Fed.Reg. at 42,530

= Id.

* Jd. at 42,537.

40a

The EPA argues, however, that even if the knee-of-the-

curve test applies to any of its regulations, the test is

applicable only to assess the cost-effectiveness of incre-

mental increases in limitations beyond BPT—that is, only

to BCT.® Representative Roberts, the author of the con-

ference report on the 1977 amendments, emphasized that

the additional technology requirements of BCT were to be

imposed only to remove additional ‘‘cheap pounds” of con-

entional pollutants beyond BPT.* Congress, however, did

not specify that initial BPT must be “cheap.’”’ In fact,

Congress anticipated that BPT might cause many plant

closures and the loss of 50,000 to 125,000 jobs.*

The BCT provisions were intended to establish an in

termediate level between BPT and the stricter BAT lim-

itations for conventional pollutants by adding a cost-

" effectiv eness test for incremental technology requirements

that exceed BPT technology.” Under BCT, additional lim-

itations on conventional pollutants that are more stringent

than BPT can be imposed only ‘“‘to the extent that the

increased cost of treatment [would] be reasonable in terms

of the degree of environmental benefits.’’®

Thus, Congress intended that cost would occupy a dif-

ferent role in EPA’s promulgation of BPT limitations than

it would in the promulgation of BCT because of the dif-

ferent aims of the two standards. While Congress did not

consider cost to be irrelevant to BPT, it clearly intended

it to be a less significant factor than in the promulgation

" See 33 U.S.C. § 1314(b\4XB). This statute does not establish a

BCT knee-of-the-curve test; the basis for articulating this assertedly

necessary test is a floor remark by Representative Roberts. 1977

Leg.Hist. at 330

“ 1977 Leg.Hist. at 330

" 1972 Leg.Hist. at 523

" American Paper Inst. v. Train, 660 F.2d at 937-58.

* 1977 Leg.Hist. at 369

4la

of BCT limitations. The EPA’s interpretation of the Act

is clearly rational and supported by both the legislative

history and the case law insofar as the EPA emphasizes

that the BPT limitations are not subject to the type of

stringent cost-benefit analysis required by BCT. The rel-

evant inquiry with respect to BPT, as indicated above, is

whether the costs are ‘“‘wholly disproportionate’ to the

benefits.

To the extent that CMA’s claim is that ‘wholly dispro-

portionate” is to be measured by a knee-of-the-curve test,

the EPA responds that CMA misconceives the nature o

the test. Rather than displaying the rate at which costs

increase relative to pounds of pollutant removed, CMA’s

curve displays the rate at which the cost-per-pound in-

creases relative to the percent of pollutant removed, re-

sulting in a misleadingly steep curve. While both the BPT

and BCT tests require a comparison between costs and

effluent reduction,” neither test requires the comparisor

of costs to the percentage removed, as implied by CMA's

curve.

CMA relies upon legislative history to justify its percent

removal approach to evaluating the reasonableness of costs

However, the statute does not require that a =

moval approach be used to establish BPT regulation: |

Almost all of the BPT regulations promulgated by the

EPA since the 1972 enactment of the Clean Water Act

are based upon either concentration limitations (as in the

case of the OCPSF rule) or more stringent ‘‘mass limi-

tations” which limit both concentrations and flow vol-

umes.” In the current case, the regulation will require a

” 33 U.S.C. §§ 1314(bK1KB), 1314(bx 4B).

" American Paper Inst. v. Train, 543 F.2d at 342, 345. In any event

the courts have upheld BPT regulations that resulted in 98.5% removal

where average industry removal was already 95.4%. See American Meat

Inst. v. EPA, 526 F.2d at 462.

™ See generally 40 C.F.R. Parts 405-71.

rease above current industry costs to remove 108

iditional pounds.* The EPA reasonably concluded

i ‘ as s

e costs were not ‘wholly disproportionate” to the

2 The BCT Cost-Effectiveness Test

MA al argues that whether or not BPT rules are,

eral matter. subiect to a knee-of-the-curve test,

; BPT limitations for conventional pollutants must

BCT cost test which Congress enacted in 1977."

promulgating BCT limitations, the Act directs the

g g BCT a

reasonableness of the relationship between the

t f attaining a reduction in effluents and the

effluent reduction benefits derived, and the compari-

f the cost and level of reduction of such pollu-

tants from the discharge from publicly owned

treatment works to the cost and level of reduction of

h pollutants from a class or category ol industrial

CMA contends that this test governs the BPT rules be

ause they represent an increase in regulation over the

imitations established on a case-by-case basis by NPDES

permits issued before 1977. In other words, CMA contends

that the permit limitations established BPT for individual

plants and that in enacting the BCT requirements in 1977

Congress intended that any subsequent, more stringent

* See 52 Fed.Reg. 42,551. The capita] and annualized costs of com-

plying with the BPT limitations are $215.8 and $76.6 million respec

tively and affect 214 plants. No plant closures are expected and seventy-

eight plants are not expected to incur incremental costs. No job losses

are anticipated

* 33 U.S.C. § 1314(0K 4A).

* 33 U.S.C. § 1314(DX 4B).

+50a

regulations must be evaluated according to the BCT

ards.

The EPA responds, however, that its authority to pr

mulgate BPT regulations is not abrogated by the fact ;

pursuant to Section 402(ay1),* NPDES permits were is

sued prior to the promulgation of industry-wide BPT reg

ulations. The EPA notes that, since 1977, it has

promulgated BPT regulations limiting conventional p.

tants in the iron and steel, metal finishing, coal mining

oil and gas, battery manufacturing, plastics molding and

forming, metal molding and casting, coil coating, porcelai

enameling, aluminum forming, copper forming, electri

and electronic products, and nonferrous metals forming

industries” —notwithstanding the fact that most of the

facilities had previously been regulated by permits. The

oil-and-gas-pollutant effluent limitations were promulgate

in 1979 and reviewed by this court in 1981 w t

referciice to the BCT cost test.”

tr

The EPA also maintains that Congress did not inter

Lia

BCT to displace BPT. The EPA notes that Congress

never repealed the BPT factors as a vital and continuing

requirement of the Act* and has not stripped the EPA

7 ] , ‘ } ¥ 4

of its explicit authority, under Section 304(b) of the A

to revise or update BPT periodically. Section 304(b) direct

the EPA to “publish regulations, providing guidelines

for effluent limitations, and at least annually thereafte

revise, if appropriate, such regulations.’ Thus, as th:

effectiveness constraints on incremental! techn gy reé

* 33 U.S.C. § 1342(ay1

” 40 C.F.R. Parts 420, 433, 434, 435, 461, 463-69 and 4

™ American Petroleum Inst. v. EPA, 661 F.2d 34 a

" Chemical Mfrs. Ass'n v. NRDC, 470 US. at 128 n. 18

at 1109 n. 18

» 33 U.S.C. § 1314)

44a

quirements that exceed BPT technology, do not displace

BPT or override the EPA’s authority to promulgate BPT

for conventional pollutants.

As additional evidence that Congress enacted BCT to

supplement, rather than to replace, BPT, the EPA points

to the fact that, ten years after the enactment of BCT,

Congress enacted a “stricter BPT’’ provision “‘requiring a

level of control substantially greater or based on funda-

mentally different control technology” for BPT regulations

promulgated after 1981.'° This applies to all BPT regu-

lations for all pollutants, including conventional pollutants,

without limitation.

As evidenced by numerous rulemakings, the EPA has

consistently interpreted the Act to allow the promulgation

of BPT limitations applicable to facilities operating under

NPDES permits despite the enactment of BCT standards

in 1977. We must accord ‘considerable weight” to an

agency’s construction of a statutory scheme it is entrusted

to administer.’ Finding the EPA’s interpretation of the

Act to be reasonable, we conclude that CMA’s objections

do not compel us to remand the limitations.

3. The Cost of BPT

Finally, having concluded that the EPA construed the

statute reasonably in declining to subject the BPT limi-

tations to the BCT cost test, we find that the Adminis-

trator did not act arbitrarily and capriciously in

determining that the costs of the limitations were justified

by the significant amount of pollutants that would be re-

moved.

i Section 301(e) of the Water Quality Act of 1987, 33 U.S.C.

§ 1311(0X3MA).

i@ Chevron USA, Inc. v. NRDC, 467 U.S. at 844, 104 S.Ct. at 2782;

Train v. NRDC, 421 U.S. 60, 75, 87, 95 S.Ct. 1470, 145, 43 L.Ed.2d

731; El. du Pont De Nemours & Co. v. Train, 430 U.S. at 134-35, 97

S.Ct. at 978.

45a

Although the cost per pound of 71 cents required to

meet the BPT limitations is almost double the 38 cents

per pound that the OCPSF industry presently spends to

remove conventional pollutants, the 71-cents-per-pound fig-

ure is not so high as to make the EPA’s decision arbitrary

and capricious:'® “‘The selection of the point of diminishing

returns is a matter for agency determination.’’'”

B. The EPA’s Definition of the BPT Data Base

NRDC challenges the EPA’s BPT limitations on BODS

and TSS as being too lenient. Specifically, NRDC argues

that the BPT limitations are not based on the average of

the best dischargers, and that the EPA improperly re-

jected sequential treatment options. These two arguments

will now be more fully examined.

1. The EPA’s Determination of the

“Average of the Best’ Dischargers

The EPA is required to promulgate BPT regulations

based on the ‘“‘average of the best’’ performers in the

industry.’ We accord some deference to the EPA’s inter-

pretation of its controlling statute; therefore, if the statute

‘8 The costs of BPT regulations for other industries have substantially

exceeded 71 cents per pound: plastics molding and forming, cleaning

water subcategory—$9.48 (49 Fed.Reg. 49,036 (December 17, 1984));

plastics molding and forming, finishing water subcategory—$12.72 (id.);

dairy products, receiving stations (small)—$9.66 (administrative record

for 1986 BCT rule); fruits and vegetables, cherry and cranberry sub-

category (small and large)—$3.34 and $1.60 (id.). See also Weyerhaeuser

Co. v. Costle, 590 F.2d at 1048 n. 56 (65 cents per pound—which, the

EPA notes, is equal to $1.16 in 1986 dollars—would not constitute a

wholly disproportionate cost).

* American Petroleum Inst. v. EPA, 540 F.2d at 1038.

'* 1972 Leg.Hist. at 169 (statement of Sen. Muskie) and 1468 (Senate

Report).

46a

is capable of more than one reasonable interpretation, we

must accept the EPA’s interpretation, if reasonable.

NRDC claims that the EPA used data from 71 of 99

plants, approximately 72%, as representing the group of

“best dischargers” for purposes of promulgating BPT reg-

ulations. How can the group of “‘best’”’ dischargers encom-

pass 72% of the industry, queries NRDC, leading it to

argue that the EPA should have further tightened its ed-

iting criteria, which would have led to more stringent reg-

)

mIAat so

Wations.

NRDC’s argument is misleading, however. Out of 304

lirect dischargers in the industry that will be subject to

regulation, the EPA chose a particular technology, namely

logical treatment with secondary clarification, which is

ised by 99 plants. Out of these 99 plants, the EPA then

hose data from 7¥ plants to determine the ‘‘average of

best’ for the purpose of promulgating its BODS reg-

ns. Thé EPA defends its decision by noting that its

tio] edit reduced the field from 304 to 99. Thus, the

NRDC’s complaint that the EPA used 71 of 99 plants is

n because the edit in question was the second edit,

hargers having already been weeded out.

We hold. therefore, that the EPA’s class of performers

r determining the “average of the best’ was not un-

{nother question is whether the CWA requires the EPA

nsider the average of the best performers within an

ndustrial category, or the average of the best performers

hat use a particular chosen technology within an industry.

hold that it is appropriate to extract a group of “best”

performers from an industry category; this was done in

his case when the EPA selected 99 out of 304 plants in

‘ts initial edit. In fact, the EPA went further by narrowing

“EPA v. Nat'l Crushed Stone Ass'n, 449 U.S. at 83, 101 S.Ct. at

306: Chevron, USA, Inc. v. NRDC, 467 U.S. at 844, 104 S.Ct. 2782.

k

ee et vee

47a

the 99 plants down to 71. The EPA was not required to

take the average of the best 99 plants using a particular

technology, but merely to take the average of the best

304 plants in the industrial category. The legislative his-

tory of the 1972 amendments to the CWA specifies that

“(t]he administrator should establish the range of ‘best

practicable’ levels based upon the average of the best per-

formance by plants of various sizes, ages, and unit proc-

esses within each industrial category.’’'’’ Therefore, we

find that the EPA’s methods for setting the BPT stand-

ards for BODS were in compliance with the CWA.

Lastly, NRDC argues that the EPA’s editing criteria

for its representative ‘‘best performers,’’ which were 95%

biological oxygen demand (BOD) removal and 40 mg/l con-

centration, are actually below the industry’s median. In

support of this contention NRDC quotes from the preamble

to the final rule, which it interprets as saying that the

industry median is 95.8% removal and 29 mg/l concentra-

tion. Again, NRDC mischaracterizes the statistics. The me-

dian 95.8% removal and 29 mg/l concentration statistics

apply to the 99 plants using biological treatment methods,

not the 304 plants comprising the industry.’

2. The EPA’s Rejection of Sequential

Treatment Options

NRDC argues that the EPA, in choosing biological treat-

ment plus clarification (Option I) for BPT, improperly re-

‘7 1972 Leg.Hist. at 169 (Statement of Sen. Muskie); see also EPA

v. Nat’l Crushed Stone Ass’n, 449 U.S. at 76 n. 15, 101 S.Ct. at 303

n. 15.

‘* There are data from a total of 99 direct discharging plants with

end-of-pipe biological treatment only (the selected BPT technology, as

discussed below) that reported average effluent BOD and a full range

of information regarding production at the plant. All these data were

used in the evaluation of the BODS data-selection criterion, even in

cases of plants that did not report influent values and for which removal

efficiencies could therefore not be estimated. The medical BODS ef-

fluent concentration for these 99 plants is 29 mg/l.

48a

jected other sequential options. Specifically, these options

were polishing ponds (Option II) and multimedia filtration

(Option III), either of which can be sequentially added to

a system already utilizing biological treatment plus clari-

fication. NRDC makes specific arguments with regard

to both technologies; these arguments will next be further

explored.

NRDC objects on several grounds to the EPA’s failure

to base BPT on Option II technology, which is biological

treatment with clarification followed by polishing ponds.

First, NRDC argues that the EPA’s conclusion that pol-

ishing ponds are not currently used by a representative

portion of the industry is incorrect. In support of this,

NRDC points out that 64 plants in the industry use pol-

ishing ponds, and, of those, 17 meet the EPA’s final ed-

iting criteria.1!° The EPA defends its conclusion by initially

noting that the OCPSF industry is diverse and as a result

plants have significantly different wastestream character-

istics. Since only 17 plants out of 64 which had installed

polishing ponds were able to meet the BPT editing criteria,

the EPA in its judgment concluded that polishing ponds

are not a sufficiently effective technology upon which to

base the BPT limits. Moreover, the question is not merely

whether the fact that 17 out of 64 plants using polishing

ponds were able to meet the BPT editing criteria proves

‘* BPT Option I consists of biological treatment, which usually in-

volves either activated sludge or aerated lagoons, followed by clarifi-

cation (and preceded by appropriate process controls and in-plant

treatment to ensure that the biological system may be operated opti-

mally). Many of the direct-discharge facilities have installed this level

of treatment. BPT Option II is based on Option I with the addition of

a polishing pond which follows biological treatment. BPT Option III is

based on multimedia filtration as an alternative basis (in lieu of BPT

Option II polishing ponds) for additional total suspended solids control

after biological treatment. Dev. Doc. at VII-2 reprinted in Joint App.

at 3702; see also 52 Fed.Reg. at 42,533.

52 Fed.Reg. at 42,537.

49a

that polishing ponds are effective, but also whether the

experiences of these 17 plants can be applied to the 304

plants subject to regulation. The EPA concluded that this

was not possible due to the diverse nature of the indus-

try... The EPA also balanced the questionable feasibility

of using polishing ponds against the cost of acquiring the

large amounts of land needed for them and concluded that

feasibility had not been sufficiently demonstrated to serve

as a basis for BPT limitations.'!* We find the EPA’s judg-

ment in this regard to be rational and well supported by

the record.

NRDC next argues that the EPA’s position that the

majority of plants using polishing ponds do not meet the

BPT criteria shows that the industry is inadequate as a

whole. The NRDC quotes from the preamble to the final

rule where the EPA noted that “plants using polishing

ponds in the OCPSF industry have done so not to add

another treatment step after effective Option I level bio-

logical treatment but rather to improve upon substandard

biological treatment.’’""? The EPA responds that this quote

is taken out of context. Again noting the diverse nature

of the industry to be regulated, the EPA stresses that the

plants did not use ponds to remedy inadequate perform-

ance which can be attributed to the model technology, but

rather because these particular plants had unique and com-

plex wastestreams which required additional treatment.

NRDC further alleges that the EPA’s analysis of the

data for Option II plants is flawed because the EPA com-

pared the performance of Option II plants to all plants in

the industry, which includes plants using Options I, II, and

III. Thus, the EPA essentially compared plants using Op-

tion II against a class which also contained Option II plants;

iy ld.

112 Td.

* Id.; see also Dev.Doc. 1X-3-IX-4, reprinted in Joint App. 4236-37

50a

this tended to improve the performance of the latter class,

masking the true benefits of ponds. NRDC contends that

the EPA should have compared Option II plants to all

other plants which do not use Option II. The EPA admits

that NRDC’s suggested approach is the better one, but

notes that it corrected its error in the preamble to the

final rule by comparing Option II plants to Option I

plants.* The EPA corrected this error in promulgating

the final rule, and the corrected data support the EPA’s

conclusion.

NRDC claims, in similar fashion to its above argument

regarding polishing ponds, that the EPA improperly re-

jected Option III as a BPT sequential treatment option.

NRDC claims that multimedia filtration results in signifi-

cant improvements in TSS removal, but the EPA cites

data which indicate that NRDC’s claim is unfounded.'*

Specifically, the EPA rejected filtration as unproven be-

cause, of the 11 of the 28 plants with filtration that sat-

isfied the EPA’s editing criteria, all 11 had other processes

in addition to filters, and there was no evidence that the

plants would not have satisfied the BPT criteria even with-

out filters. Due to the diversity of the OCPSF industry,

“4 In examining the data from the 18 plants originally placed in the

Option II data base (using the 95/50 criteria), EPA noted that they

yielded concentrations that were not much lower that Option I con-

centrations. Option II plants averaged only 2 mg/l BODS and 8 mg/l

TSS lower than Option I plants. Because these increments seemed

rather small, EPA performed a statistical analysis to compare the av-

erages for the two data bases. The results of the analysis did not

provide evidence of a significant difference between the two data sets.

52 Fed.Reg. 42,537.

‘6 See Dev.Doc. Table VII-53, reprinted in Joint App. 3753.

“© EPA discussed this issue in more depth in the Development Doc-

ument:

Eleven plants in the BPT data base use BPT Option III technology

and meet the final editing criteria. Thus, this option would require

EPA to regulate all seven subcategories based upon a very small

ee ee i

5la

the EPA determined that the data from 11 plants did not

demonstrate the effectiveness of filters throughout the in-

dustry, and we defer to the EPA’s determination.

C. The “Summier/Winter Issue’: The EPA’s Decision

Not to Subcategorize Based-on Climatic Differences

Petitioners Goodyear, DuPont, M & T, Monsanto, and

Union Carbide (hereinafter collectively ‘‘Goodyear’’) chal-

lenge the BPT limitations for failure to account for cold-

temperature effects. Goodyear argues that because cold

temperatures adversely affect biologieal treatment, the

EPA should have created a separate class for cold-tem-

perature plants, conducted a technical assessment and cost

analysis for this class, and calculated the incremental cost-

effectiveness of bringing this class into compliance.

The EPA used biological treatment with secondary clar-

ification as the model technology for formulating the BPT

limitations.’ Biological treatment involves treating was-

data set. As shown in Table Vii-36, the median effluent TSS con-

centration value for these plants is 32 mg/l. Even if three additional!

plants are included in this data base because they use Option |

treatment plus either ponds of activated carbon followed by filters,

the resulting median TSS value is 34 mg/l. These results, when

compared to the performance of clarification only following biological

treatment (median value of 30 mg/l), clearly show that the efficiency

of filtration following good biological treatment and clarification is

not demonstrated for this industry. Moreover, on the average, OCPSF

plants with more than Option I treatment in EPA's data base (bi-

ological treatment plus filtration) have not demonstrated significant

BODS removal beyond that achievable by Option I treatment alone.

The median BODS concentration value for these plants is 19 mg/|

compared to a median value of 23 mg/l BODS for the plants with

Option I technology in place which meet the 95 percent/40 mg/l

BODS editing criteria. Therefore, EPA does not believe that the

data support any firm estimate for incremental pollutant removal

benefits and incremental costs for BPT Option III.

Dev.Doc. VII-2, reprinted in Joint App. 3808-10.

"752 Fed.Reg. 42,534.

52a

tewater with microorganisms that biodegrade and some-

times absorb the organic pollutants. Goodyear contends

that the microorganisms used in biological treatment are

adversely affected by cold temperatures, thus affecting the

efficacy of biological treatment. Therefore, Goodyear ar-

gues, many plants in the northern and middle latitudes

will not be able to comply with the BPT limitations even

if they employ the model technology. Goodyear argues that

these cold-temperature plants should be grouped into a

separate subcategory and subjected to less stringent BPT

limitations.

The EPA decided not to provide separate summer/win-

ter limitations or to subcategorize the OCPSF industry

based on temperature. This decision is clearly supported

by the record. The EPA determined that the best plants

in the industry could comply with the BPT limitations

throughout the year. The EPA also conducted a detailed

analysis which demonstrated that the plants operating in

the coldest temperatures and with the greatest seasonal

temperature fluctuations dealt with these problems more

effectively than southern plants, indicating that attention

to appropriate design and operation features overcomes

any temperature-related difficulties in treatment. The EPA

concluded that cold temperatures had, at most, an insig-

nificant effect upon the treatment of wastewater at plants

that took appropriate measures. Finally, the EPA also con-

sidered the incremental costs of ensuring effective treat-

ment during the winter and increased its cost estimates

to reflect such treatment using a temperature-correction

factor.

1. Diversity of the Data Base

The BPT regulations for the OCPSF industry are based

upon a broad and geographically diverse population of

plants. The EPA used data from 71 plants to establish the

long-term averages, and data from 21 plants to establish

the variability factors. Together, these data formed the

53a

basis for the BPT limitations.'* The northern, southern,

and middle latitude regions are all well represented in both

sets of data.* Furthermore, the data from these plants

were collected throughout the year and thus represent a

level of treatment performance that may be obtained under

all seasonal conditions.'”° Additionally, this large data base

represents a wide range of production operations. The size

and diversity of the data base ensures that the BPT lim-

itations based upon this data can generally be achieved

across the industry, regardless of climate.’*! This basis

alone is sufficient to support the limitations.’

Petitioners object to the EPA’s exclusion of seven plants

that had permits allowing compliance with less stringent

limitations during the winter months from the set of 21

18 52 Fed.Reg. 42,533-35.

419The set of 21 plants contains 6 northern plants, 4 mid-latitude

plants, and 11 southern plants. Twenty of these plants are identified

by geographical regions. Dev.Doc. VII-94-96, reprinted in Joint App.

at 3794-96. The remaining plant is in Puerto Rico. Dev.Doc. VII-90,

reprinted in Joint App. at 3790. The larger data set of 71 plants

contains considerably more plants from each of the three regions.

120 Dev.Doc. at VII-91-92, reprinted in Joint App. at 3791-92.

121 52 Fed.Reg. at 42,556.

12 Marathon Oil Co. v. EPA, 564 F.2d at 1269, n. 46 (EPA's con-

clusion that limitations could be met throughout the year upheld based

on the fact that some of the data from one plant used to develop

limitations were collected during the late winter and early spring);

American Paper Inst. v. EPA, 540 F.2d at 1306 (limitations based upon

ten geographically diverse plants, including variability factors derived

from a few plants “located in cold climates,”’ upheld); American Meat

Inst. v. EPA, 526 F.2d at 454-56 (compliance by several plants during

winter months demonstrated the achievement of the limitations on a

year-round basis). See also Kennecott v. EPA, 780 F.2d 445, 450 (4th

Cir.1985), cert. denied, 479 U.S. 814, 107 S.Ct. 67, 93 L.Ed.2d 25

(1986) (limitations upheld based upon statistical extension of data which

allegedly did not reflect ‘‘seasonal changes in temperature and precip-

itation, production surges or slowdowns"); Weyerhaeuser Co. v. Costle,

590 F.2d at 1054-55.

54a

plants used to develop the variability factors. The EPA

fully explained its decision to exclude these plants:

[Bjecause these plants were subject to meeting two

different sets of permit limits, they had no incentive

to attempt to achieve uniform limitations throughout

the year. Not surprisingly, then, the daily data from

these plants exhibit a two-tier pattern. These data can

be characterized by two means, and the variability of

these data over a 12 month period is fundamentally

different from the data from plants required to meet

only one set of permit limits. Consequently, the data

generated during these periods is not representative

of well-operated biological treatment, which as noted

above is capable of uniform treatment throughout the

year as demonstrated by a number of plants.'*

These plants were in fact included in the 71-plant data

base used to develop long-term averages. They were ex-

cluded only from the variability analysis.’ Therefore, their

lower average performance during the winter is reflected

in the BPT limitations. It is reasonable for the EPA to

choose not to rely on plants that have no regulatory in-

centive to achieve uniform performance, particularly where

an ample data base of plants meeting uniform year-round

permit limitations was available.’ The exclusion of the

data from the seven plants operating under special permits

was adequately explained and is a determination ‘‘pecul-

iarly within EPA’s expertise.”’ The decision is therefore

13 52 Fed.Reg. at 42,556.

1% Dev.Doc. at IV-a-55-62, reprinted in Joint App. at 4396-4403.

12% See Sierra Club v. Costle, 657 £.2d 298, 362 (D.C.Cir.1981) (in

establishing new source performance standards in the Clean Air Act,

EPA correctly did not rely on data from existing plants that have not

been designed to operate at the efficiency level required by the new

source standards).

55a

entitled to judicial deference.'*° Moreover, the EPA did

not selectively exclude from the variability data base those

plants that performed worse in the winter than in the

summer.’ Thus, any difference between winter and sum-

mer variability at the best plants was incorporated into

the variability factors used to develop the limitations. There

is ample support in the record for the EPA’s conclusion

that lower removal efficiencies were the result primarily

of factors other than cold temperature and that the BPT

limitations are achievable in all climates on a year-round

basis. Goodyear’s contentions to the contrary are there-

fore without merit.

126 American Meat Inst. v. EPA, 526 F.2d at 457; accord Kennecott

v. EPA, 780 F.2d at 450.

127 Dev.Doc. at VII-79-104, reprinted in Joint App. at 3779-3804.

1% The EPA conducted a series of analyses that confirmed that the

limitations are achievable in all climates on a year-round basis. Dev.Doc.

at VII-79-104, reprinted in Joint App. at 3779-3804. EPA evaluated

data from 20 or the 21 plants that had reported daily data (i.e., the

results of daily effluent monitoring) and were used in the BPT vari-

ability analysis. Dev.Doc. at VII-82, reprinted in Joint App. at 3782.

EPA noted that there was a slight reduction in BODS removal effi-

ciency and a corresponding small increase in effluent concentration at

some plants, while other plants’ performance actually improved during

the winter, and others had no substantial change. Dev.Doc. at VII-83,

reprinted in Joint App. at 3783. Therefore, no correlation was shown

between temperature and treatment performance. Cf. American Meat

Inst. v. EPA, 526 F.2d at 455 (upholding year-round limits for data

which did not show a direct correlation between removal efficiency and

cold weather).

The EPA also performed regional analyses to examine the winter

performance of plants within certain latitudes and to compare the over-

all performance of groups of plants from different latitudes. The EPA

determined that plants in the northern region had the highest annual

removal efficiency, 98%. Dev.Doc. at VII-83, reprinted in Joint App.

at 3783. The EPA reasonably concluded that this analysis showed that

removal efficiency was affected primarily by factors unrelated to cli-

mate. Dev.Doc. at VII-88, reprinted in Joint App. at 3788. The EPA

also noted that the data indicated only minima! treatment-efficiency

56a

2. Winter Removal Efficiencies

Goodyear also argues that the EPA was required to

establish a separate subcategory for plants that could not

achieve the BPT limitations due to cold-temperature ef-

fects. As we held above, however, the record supports the

variations during the spring and autumn months when temperature

fluctuations are the greatest, casting doubt on the theory that tem-

perature fluctuations would impede treatment. Dev. Doc. at VII-90,

reprinted in Joint App. at 3790.

The EPA performed a similar analysis on a different set of plants

that had reported only average data over the course of three-month

periods and the results were consistent with those obtained in the

preceding analysis. Dev.Doc. at VII-88-89, reprinted in Joint App. at

3788-89. The southern plants experienced an average 5% loss in treat-

ment efficiencies during the winter, and the northern plants had little

loss in efficiency. Jd. The EPA concluded that ‘“‘while northern and

middle latitude plants would have larger swings of temperature going

from season to season, these swings have been compensated for through

ope.ation and process modifications.” Jd. The larger difference between

summer and winter removal efficiencies for southern plants indicated

that ‘‘these facilities have not adequately addressed the smaller tem-

perature swings by operational and process modifications.”’ Jd. A fourth

analysis demonstrated that temperature defined in terms of ‘degree

days” (the measure of temperature typically used by power companies

to estimate heating bills) was not a significant factor with respect to

the BOD and TSS effluent concentrations for particular BPT subca-

tegories. Dev.Doc. VII-90, reprinted in Joint App. at 3790. A fifth

analysis examined plants’ effluent concentration based upon consider-

ation of months rather than ‘‘degree days” and reached the same con-

clusion. Dev.Doc. at VII- 93, reprinted in Joint App. at 3793.

All of the above analyses support the EPA’s conclusion that cold

temperatures had, at most, an insignificant effect upon wastewater

treatability at plants that took appropriate measures to deal with po-

tential temperature effects. Therefore, the EPA rationally determined

that plants with lower efficiencies are affected as much by inefficient

operating practices as by winter temperature considerations or by fluc-

tuating temperatures. 52 Fed.Reg. at 42,556. Together with the fact

that the limitations are based on year-round industry data from a large

and geographically-diverse set of plants, the EPA’s analyses support

its determination that the BPT limitations are achievable throughout

the year.

es a een aaa ee

OS ee ee el

57a

EPA’s conclusion that cold-temperatures alone did not

have a significant effect on plants’ abilities to meet the

BPT limitations. Thus, the EPA could reasonably conclude

that the effects of temperature on treatment effectiveness

did not provide a basis for subcategorization. Goodyear

contends that the EPA did not have sufficient data indi-

cating that technology was available to overcome the ef-

fects of temperature. The EPA, however, did identify

numerous practices that were used successfully by OCPSF

plants to improve their winter performance.'? The EPA

is not required to analyze all possible causes of each plant’s

non-compliance with BPT and to recommend design and

operating practices for each plant.'*° The EPA concluded

reasonably that the limitations could be met during all

seasons and identified several means of doing so.

Finally, Goodyear argues that the EPA failed to con-

sider whether the technology required to overcome the

effects of temperature is cost-effective. The EPA is re-

quired by statute to consider costs in relation to benefits

only for categories or subcategories of plants, not for in-

2 Dev.Doc. at VII-100-104, reprinted in Joint App. at 3800-04. The

EPA listed the following wintertime practices which, if properly applied,

compensated for any cold temperature variation: reduce excessive stor-

age time prior to treatment; insulate treatment units; cover open tanks;

maintain higher concentrations of mixed liquor suspended solids; and

reduce the food-to-microorganism ratio. Insulation included installing

tanks in the ground rather than above ground, using soil to cover the

walls of above-ground units, and enclosing treatment units. Dev.Doc.

at VII-1030104, reprinted in Joint App. at 3803-04. The EPA also

provided two case histories of plants which had used these and other

techniques to address the effects of temperature on treatment. Dev.Doc.

at VII-101, reprinted in Joint App. at 3801.

© See EI. du Pont de Nemours & Co. v. Train, 430 U.S. at 128-

29, 97 S.Ct. at 975; American Meat Inst. v. EPA, 526 F.2d at 451

(“practicability” and “‘availability’’ of technology are not to be deter-

mined on a plant-by-plant basis).

58a

dividual plants. As we held above, the EPA was not

required to create a subcategory consisting of plants that

had difficulty complying with the limitations due to tem-

perature effects. There is no record basis for isolating the

effect of temperature on treatment effectiveness from the

effects of other, more significant factors. Consequently,

as the EPA notes correctly, there is no class of plants for

which temperature-related costs can be estimated and as-

sessed against pollutant reduction benefits.

We hold, therefore, that the Administrator concluded

rationally that the BPT limitations are practicable for all

plants in all seasons.

31 FJ. du Pont de Nemours & Co. v. Train, 430 U.S. at 128-30, 97

S.Ct. at 975-76; BASF Wyandotte Corp. v. Costle, 598 F.2d at 662;

American Iron & Steel Inst. v. EPA, 526 F.2d at 1051; FMC Corp v.

Train, 539 F.2d at 979 (the estimation of BPT costs should not serve

as a dilatory device, obstructing the Agency from proceeding with its

primary mission of cleaning up the lakes, rivers, and streams of this

nation).

182 However, the EPA did consider the general effect temperature

may have on treatment costs across the industry. First, the EPA used

each state’s cold-season temperatures to adjust downward the biode-

gradation rate used in estimating costs for full-scale and second-stage

biological systems in that state. Dev.Doc. at VII-104, reprinted in Joint

App. at 3804. This approach had the effect of assuming slower biod-

egradation rates due to cold weather and designing the entire system

to meet worst-case conditions. Second, the EPA developed Temperature

Correction Factors for costing the upgrading of existing treatment

facilities, based upon each state’s winter temperatures. Dev.Doc. at VII-

104 and VIII-26-32, reprinted in Joint App. at 3804 and 3979-85. These

factors were multiplied by the normal, warm-weather treatment costs

to obtain increased costs of treatment. In essence, the EPA’s costing

methodology increased costs for treating discharges throughout the year

to ensure adequate costing for cold weather. The EPA‘‘s technical

analyses of temperature effects and its reliance on a data base reflecting

geographical and seasonal diversity, coupled with a detailed cost anal-

ysis that explicitly increased cost estimates to account for potential

cold-weather effects, all combined to provide a rational basis for the

Administrator’s conclusion that the limitations are practicable.

59a

D. BPT Subcategorization and the EPA’s Use of Stand-

ard Industrial Classification (SIC) Codes

CMA and NRDC both argue that the EPA’s classifi-

cation scheme in development of BPT limits for the OCPSF

industry was improper because it was based in part upon

Standard Industrial Classification (SIC) codes. There are

two strands to this argument. CMA argues that there is

no significant correlation between SIC codes and effluent

levels, thus rendering the SIC-code-based classification

scheme arbitrary. NRDC takes a different tack, arguing

that EPA did not provide for sufficient public notice and

comment regarding the proposed system. Based on the

analysis below, we reject both arguments and uphold the

EPA’s SIC-code-derived classification.

1. CMA’s Challenge to the BPT

Subcategorization

CMA challenges the EPA’s division of the OCPSF In-

dustry into seven subcategories for the purpose of estab-

lishing BPT limitations on the ground that the

subcategories are based on SIC-product groupings rather

than wastewater characteristics or treatability. CMA ar-

gues that the EPA has created an inequity by grouping

together plants with substantially different influent con-

centrations and subjecting them to the same concentration

limitations, thereby requiring, in effect, a higher percent

removal at plants with higher influent concentrations.

In Section 304(bX1XB), Congress listed several factors,

in addition to cost, that the EPA shall ‘‘take into account”

in determining BPT, including the age of equipment, the

process employed, engineering aspects of treatment, proc-

ess changes, and non-water-quality environmental im-

pacts. Based on these factors, the EPA determines

whether plants within an industry should be assigned to

'* 33 U.S.C. § 1314(oX1XB).

60a

a subcategory subject to more particularized regulations

than the industry as a whole.

However, the EPA is required to create a separate sub-

category for a group of plants only when they are so

fundamentally different from other plants on which the

limitations are based that they cannot practicably achieve

the effluent limitations achieved by the average of the best

plants in the industry.* The EPA has considerable dis-

cretion in evaluating these factors; it is enough that the

EPA considered the relevant factors and reached a ra-

tional conclusion about them.'* The Agency’s task is ‘‘to

establish numerical standards limiting effluent pollution,”

and it should concentrate ‘‘on grouping plants that could

meet the same limitations.’’** If plants can meet the same

limitation, they need not be subcategorized simply because

they are different.'*’

18 See Chemical Mfrs. Ass'n v. NRDC, 470 U.S. at 119-22, 129-34,

105 S.Ct. at 1104-06, 1109-12 (the Supreme Court recognized that the

substantive test for subcategorizing an industry is the same as that

which applies to establishing fundamentally-different-factor (‘““FDF’’)

variances—whether limitations based upon a range of plants may be

fairly extended to another plant or class of plants that is alleged to

be fundamentally different with respect to relevant statutory factors);

EPA v. Crushed Stone Ass'n, 449 U.S. at 78, 101 S.Ct. at 304 (a BPT

limitation is “incomplete” only if it was determined without consider-

ation of a “current practice fundamentally different from those that

were considered by the Administrator’’).

‘88 Reynolds Metals Co. v. EPA, 760 F.2d at 564; Wyandotte Corp.

v. Costle, 598 F.2d at 656; Weyerhaeuse Co. v. Costle, 590 F.2d at

1047.

‘% Reynolds Metals Co. v. EPA, 760 F.2d at 565.

‘87 See Chemical Mfrs. Ass'n v. NRDC, 470 U.S. at 120, 105 S.Ct. at

1105 (subcategories are necessarily rough-hewn); Kennecott Copper Corp.

v. EPA, 612 F.2d at 1241 (rough basis for subcategorization suffices);

Wyandotte Corp. v. Costle, 598 F.2d at 655 (plants within subcategories

need not be identical and in fact may have production outputs that

differ by a factor of fifty); Kennecott v. EPA, 780 F.2d at 451 (not all

variation and pollution loads must be accounted for); American Iron &

6la

The subcategories for the OSPSF industry are defined

by generic types of OCPSF production. The EPA main-

tains that since a plant’s chemical processes are closely

related to the type of production in which it engages, the

subcategorization bears a reasonable relationship to such

chemical processes and the wastewater characteristics as-

sociated with those processes.'*

In making its decision to subcategorize the OCPSF in-

dustry for the purpose of its BPT regulations, the EPA

analyzed several potential subcategorization factors, such

as age of plant, temperature, and SIC-code classification

of the plant, just to name a few. CMA disputes whether

SIC codes actually have a relationship with BODS effluent

concentration. They present a statistical analysis purport-

edly showing that SIC codes have no significant correlation

with BODS effluent. CMA analyzed several categorization

schemes, including the one retained by the EPA, and de-

termined the Value of R?, the coefficient of determination,

Steel Inst. v. EPA, 568 F.2d at 299-300 (not all subcategory differences

must be accounted for).

'* 48 Fed.Reg. at 11,832-34; Dev.Doc. IV-20-1, reprinted im Joint

App. at 3519-20. EPA’s 1985 Federal Register notice explained the

relationship of the subcategories to the chemical processes, and thus

to wastewater characteristics in the industry. 50 Fed Reg. at 29.074

Plastics plants employ only a small subset of the chemistry used by

organics plants to produce a limited number of products. Thus, organics

and plastics plants represent two broad groupings within the industry

These two groups may be subdivided further. Among the plastics plants.

those producing thermo-setting resins and thermo-plastic materials may

be distinguished, while each of those producing rayon and synthetic

fibers has unique wastewater characteristics. Organic chemicals pro

divided into three groups based upon volume of production,

relates to fundamental process chemistry characteristics

highest volume chemicals, commodity chemicals, also have

chemical structures. The second groups, bulk chemicals,

commodity chemicals as raw matenals to produce struc

complex chemicals. The third group, specialty chemicals,

produced in smaller volumes and devoted to particular uses (eg.

&nd pigments) and are even more structurally complex. /d

234g87

lee

62a

for each.'** CMA argues that the categorization scheme

which was retained by the EPA has an R? of only 0.052,

which is not significantly better than other rejected clas-

sification schemes. In fact, a random shuffling performed

by CMA of BODS data with respect to SIC codes yielded

an R? of 0.030 in the CMA analysis. This, argues CMA,

supports its position that the classification scheme adopted

by the EPA is arbitrary and capricious, since it is only

slightly better than a random categorization scheme. CMA

also criticizes the EPA’s addition of allegedly irrelevant

dummy variables to the EPA’s R? analysis of its own

model, the alleged effect of which was to artificially inflate

the R? value.'*°

We are of the opinion that the R? analysis presented

by CMA is inconclusive. Though an R? analysis can be

informative, it cannot of itself conclusively prove or dis-

prove the adequacy of a particular categorization scheme.

A good model may nonetheless have a small R? value if

‘88 R?, the coefficient of determination, generally is a measure of the

proportion of variation in data which is explained by the regressor

variable. In this case, R? describes the extent to which BODS effluent

levels are explained by the SIC code categories. The range of values

for R? always falls between 0 and 1; values closer to 1 imply that most

of the variability is explained by the regressor variable. See Montgomery

& Peck, Introduction to Linear Regression Analysis 33 (Wiley & Sons,

1982).

‘© Although introducing additional variable into the calculation can

increase R?, CMA has not provided enough information for us to de-

termine whether the R? model with dummy variables is better or worse

than the base model. Montgomery & Peck note that:

[a]lthough R? increases if we add a regressor variable to the model,

this does not necessarily mean that the new model is superior to

the old one. Unless the error sum of squares in the new model is

reduced by an amount equal to the original error mean square, the

new model will have a larger error mean square than the old one

because of the loss of one degree of freedom for error. Thus the

new model will actually be worse than the old one. cine

Montgomery & Peck at 34.

63a

the spread of data points along the x-axis (horizontal axis)

is small." Thus, the fact that the R* value attributed to

the SIC-code-based classification is relatively small (0.052)

is an insufficient basis for us to conclude that the EPA’s

industrial classification is arbitrary and capricious. More-

over, the coefficient of determination, R’, is merely a lim-

ited indicator of model adequacy. Ultimately, we believe

the EPA’s explanation that SIC codes tend to be organized

around the products produced by various segments of the

industry, and that the type of product in turn influences

the wastestream characteristics of those plants, to be a

sufficient rationale upon which to uphold the EPA’s clas-

sification scheme.!*

CMA further argues that analysis of its proposed sub-

categorization scheme based on influent BODS concentra-

tions yields an R? of 0.42, which is higher than the R*

value of the EPA’s SIC-code-based plan (although not

higher than the EPA analysis when the dummy variables

are included). Therefore, argues CMA, the EPA should

have subcategorized the OCPSF industry based on influent

levels, as opposed to SIC codes. Again, we find this ar-

gument to be insufficient for the following reasons. First,

it should be remembered that a large R* value does not

141 Td.

‘2 The EPA found that these process characteristics relate in turn

to fundamental wastewater characteristics in the OCPSF industry. For

instance, plastic compounds can be removed more readily from the

process stream of a plant than other products. Therefore, when a plant

removes the plastic compounds it is producing from the wastestream,

little polluting matter (other than organic compounds) remains to be

discharged as waste. EPA concluded that this characteristic would allow

a plastics manufacturer to achieve a lower biological oxygen demand

(BOD) than an organics manufacturer. Dev.Doc. at IV-20, reprinted in

Joint App. at 3519. As another example, among the three organic

subcategories, larger volume chemicals tend vo have a simpler molecular

structure, and this generally results in a greater biodegradation rate

—— effluent BOD. Dev.Doc. at IV-21, reprinted in Joint App.

64a

conclusively prove that the variables are causally related.’

Second, CMA provides a graph of its data points showing

influent levels (x-axis) versus effluent levels (y-axis). There

are two remote points in x-space on this graph, both with

large influent and effluent levels. Data plots with such

remote points tend to have much higher R? values even

though the model is not necessarily superior, since those

points exert a greater influence on the slope of the regres-

sion line than the points which are clumped together nearer

the origin.’ Moreover, if these points are bad values (due

to error), their deletion may reduce the resulting value of

R?. Although these points may just as well be valid meas-

urements, this determination requires the exercise of dis-

cretion on the part of the analyst. It is unclear how CMA

exercised its editing discretion in developing the model

they presented, and in any event, the limited nature of

CMA’s regression analysis precludes us from mandating

that EPA adopt CMA’s model.

2. NRDC’s Notice and

Comment Challenge

NRDC argues that the EPA failed to provide notice of

its intent to limit the applicability of the OCPSF regula-

tions to certain SIC codes. NRDC complains that, as a

result of this lack of notice, it was denied the opportunity

to comment on this part of the regulation.

We find NRDC’s argument to be lacking in merit. Dur-

ing the rulemaking, EPA invoked SIC codes to define

‘@ Montgomery & Peck at 36; see also G. Box, W. Hunter & J.

Hunter, Statistics for Experimenters 487 (1978).

‘“« The mathematical reason for this is that the value for R* generally

increases as the spread of the points along the x-axis becomes greater.

Outlier points tend to have a large effect on the spread along the x-

axis. Montgomery & Peck at 34

65a

which facilities would be covered by the final rule.'** Our

review of the record persuades us that NRDC was fairly

apprised of the issues at stake, and that the EPA’s de-

cision to limit the rules to the five SIC codes identified

in its original notice was a logical outgrowth of the ru-

lemaking proceeding.

E. Issues Concerning Waste-Stabilization Ponds

Petitioners DuPont, Texas Eastman, Union Carbide, and

Air Products incorporate the use of waste stablilization

ponds (‘““‘WSP systems’’) for the treatment of industrial

wastewater. WSP systems consist of tanks where organic

matter in wastewater is broken down through bacterial

action. During the treatment process, algae form which

increase the TSS in the effluent. Petitioners raise several

claims regarding the application of the BPT limits to plants

employing WSP systems. First, petitioners contend that

the EPA acted arbitrarily and capriciously in determining

that pond algae are a conventional pollutant subject to the

BPT limitations for BODS* and TSS."*’ Second, petitioners

contend that the EPA violated the Act by failing ade-

quately to account for algae-related problems in setting

. In the preamble to EPA’s initial proposal of the OCPSF regu-

lations, EPA noted that approximately 1,200 facilities manufacture their

principal products under ‘the OCPSF SIC Groups.”’ 48 Fed.Reg. 11,828,

11,830 (1983). The only SIC groups listed were the same five upon

which the final rule was based. In its next “Notice of Availability and

Request for Comments” the Agency stated that it “has defined the

Organic Chemicals Manufacturing industries to include all facilities

within specific SIC codes.” 50 Fed.Reg. 29,068, 29,091 (1985). In the

same notice, EPA also presented several products lists to provide ‘guid-

ance” in defining “BPT subcategories.” Jd. at 20,092. EPA intended

to use the lists in defining subcategories within the industry. /d.

‘* BODS adversely affect the receiving waters because they deplete

the oxygen available to fish, plant life, and other aquatic species.

‘“ Suspended solids adversely affect receiving waters because they

increase turbidity, reduce light penetration (impairing photosynthesis),

and may settle to form sludgebanks.

66a

BPT limits for BODS and TSS. Third, petitioners argue

that the EPA acted arbitrarily and capriciously in speci-

fying copper sulfate treatment as a means of controlling

algae without adequate notice and comment on the prac-

ticality of that treatment technology.

We find that the EPA’s implicit regulation of algae as

a component of BODS and TSS had a rational basis and

that the EPA was not required to create a separate sub-

category for plants utilizing WSP systems. These four

plants perform the same types of operations as other

OCPSF plants and generate the same types of OCPSF

process wastewaters and are thus properly subject to the

same BPT limitations. Finally, we conclude that the EPA

did not act arbitrarily and capriciously in designating cop-

per sulfate treatment as a means of controlling algae.

1. EPA’s Determination That Pond Algae

Were Conventional Pollutants and

Thus Subject to BPT Regulations

Petitioners argue that the EPA should not have included

algae as a component of BODS and TSS because algae to

not contribute to the environmental problems associated

with BODS and TSS. We find, however, that the EPA’s

regulation of algae as a component of BODS and TSS had-

a rational basis.

The EPA notes correctly that Congress, without limiting

the definition of the terms, includes BODS and TSS in the

broad definition of conventional pollutants.’ The EPA

'* Section 304(aX4) requires EPA to publish “information identifying

conventional pollutants, including but not limited to, pollutants classified

as biological oxygen demanding, suspended solids, fecal coliform, and

pH.” 33 U.S.C. § 1314(aX4). The Act prohibits the discharge of any

“pollutant” except in accordance with a permit that implements BPT

along with other requirements. 33 U.S.C. § 1311(a) and (bX.1B). “Pol-

lutant” is defined as “solid waste, .. . sewage, . materials,

.. and industrial, municipal, and ‘agricultural waste discharges into

67a

concludes that because algae are both oxygen-demanding

organisms and suspended solids, algae are encompassed

within both BODS and TSS as those terms are used in

the Act, and the EPA is thus authorized to include algae

in the measurement of both TSS and BODS. The EPA

also notes that Congress has specifically recognized that

algae are a significant cause of water quality problems.’

In the OCPSF rulemaking, the EPA found that algae

can indeed present significant water-quality problems:

The control of algae growth in ponds, lakes, and res-

ervoirs can be a serious problem in water quality man-

agement. Among the nuisances created by the often

sudden blooming of one or more algal genera are:

odors and tastes; fish kills; poisoned water fowl]; short-

ened filter runs and water-purification plants; growths

in pipes and other water conduits; and interference

with industrial water uses. Therefore, proper control

of algae growth in ponds is necessary to avoid the

potential nuisances in the ponds and/or receiving

waters.'©

water.”’ 33 U.S.C. § 1362(6). Algae discharged from an industria! system

may fit under several of these terms. Similarly, algae are encompassed

within the term “‘Pollution,”’ defined in the Clean Water Act as ‘‘the

man-made or man-induced alteration of the chemical, physical, biolog-

ical, and radiological integrity of water.’’ 33 U.S.C. § 1362(19).

‘ See, e.g., 1972 Leg. Hist. at 1310 (“algae has [sic] grown so rapidly

that sufficient oxygen is not available to support other forms of life’).

” Dev. Doc. VIII-84, reprinted in Joint App. at 4037. See also Mum-

ford Cove Ass’n v. Town of Groton, Conn., 786 F.2d 530, 532 (2nd

Cir.1986) (discharge from a sewage treatment plant resulted in ‘‘brack-

ish water choked by a monoculture of algae excluding beneficial aquatic

plant life’’); Stoddard v. Western Carolina Regional Sewer Auth., 784

f.2d 1200, 1204 (4th Cir.1986) (fish kill attributed to an algae bloom):

Montgomery Envtl. Coalition v. Costle, 646 F.2d 568, 575 (D.C.Cir.1980)

(excessive nutrients in the Potomac River degrade water quality both

because the proliferation of algae is itself a nuisance and because algae

respiration and subsequent death and decay use up oxygen dissolved

68a

While petitioners take issue with some of the EPA’s

specific conclusions regarding the effects of algae on water

quality, they have failed to demonstrate that the Admin-

istrator acted arbitrarily and capriciously in regulating al-

gae as a conventional pollutant. The EPA has authority

to regulate the discharge of a pollutant even where its

effects are subject to some uncertainty. We will not

substitute our judgment for the Administrator’s in this

regard.

2. The EPA’s Decision Not to Create a

Subcategory for Plants Utilizing Waste

Stabilization Ponds

Petitioners argue next that the EPA violated Section

304(bX1\B) of the Act by failing to adequately account for

algae-related problems in setting BPT limits for BODS and

TSS. As explained more fully above, Section 304(b\1\B)

requires the EPA to ‘‘take into account’ several factors,

in addition to cost, in determining BPT, including the age

of equipment and facilities, the process employed, the en-

gineering aspects of the application of various types of

control technologies, process changes, and non-water qual-

ity environmental impacts.'** Petitioners assert that based

on those considerations, the EPA should have created a

separate subcategory for plants utilizing WSP systems,

subject to less stringent BPT limitations on BODs and TSS

than those applicable to the rest of the industry.

in the river’s waters, threatening other forms of aquatic life); National

Wildlife Fed’n v. Gorsuch, 530 F.Supp. 1291, 1298 (D.D.C.) (algae causes

oxygen depletion), rev’d on other grounds, 693 F.2d 156 (D.C.Cir.1982).

181 FMC Corp. v. Train, 539 F.2d at 982-83; C & H Sugar Co. v.EPA,

553 F.2d 220, 289 (2nd Cir.1977); 1972 Leg.Hist. at 378 (BPT, as a

technological standard will be required ‘“‘even if the quality of the re-

ceiving waters does not require the imposition’ of BPT). See also 1972

Leg.Hist. at 1305-06 (same).

12 33 U.S.C. § 1314(bX 1B).

nei

69a

Petitioners first claim that the EPA failed to consider

the extent to which high effluent TSS concentrations are

a function of the type of wastewater in the treatment

system. The EPA maintains, however, that petitioners’

plants do not differ materially from other OCPSF plants

with respect to manufacturing operations or wastewater

characteristics. Rather, the EPA argues, their plants differ

only with respect to the type of treatment petitioners vol-

untarily chose to employ. Petitioners did not choose to

employ the type of biological treatment system that is

most commonly used by good performers in the OCPSF

industry and that was used by the EPA for cost-estimating

purposes: activated sludge, followed by secondary clarifi-

cation.’ Instead, petitioners chose to employ WSP treat-

ment systems which generate additional pollutants in the

form of algae. The EPA asserts that it therefore reason-

ably required petitioners’ plants to meet the limitations

demonstrated to be achievable by the average of the 71

“best” plants in the OCPSF industry,’* rather than pro-

viding less stringent limitations for these plants simply

because they chose to employ less effective treatment tech-

nology.

The EPA asserts that, as required by the Act, it con-

sidered in detail ‘‘the engineering aspects of the applica-

‘8 Dev.Doc. at VII-64 and VIII-1-5, reprinted in Joint App. at 3764

and 3954-58.

'* See Weyerhaeuser Co. v. Costle, 590 F.2d at 1062 (rejecting claim

that certain ‘‘non-settleable solids” cannot be removed by BPT, because

ten plants were already in compliance); American Paper Inst. v. Train,

543 F.2d at 345 (limitations based upon industry data upheld regardless

of whether treatment technologies are specified); American Petroleum

Inst. v. EPA, 540 F.2d 1023 (regulations upheld where twelve plants

already complied); American Meat Inst. v. EPA, 526 F.2d at 456 (the

ability of two plants to meet EPA limitations using EPA's model tech-

nology in the summer demonstrates that the BOD and TSS limitations

are achievable, contrary to assertions that algae growth precludes com-

lance).

70a

tion of various types of control techniques” and used this

factor to identify the data base from which it derived the

BPT limitations. Those limitations properly reflected the

results achieved by more effective treatment systems than

those employed by petitioners. Petitioners cannot now be

excused from meeting the BPT limits on the grounds that

their present method of treatment is less effective than

the average of the best—in this case, because WSP sys-

tems create additional pollutants. Such a construction of

Section 304(bX1\B) would defeat the very purpose of es-

tablishing BPT limits. The EPA is not required to base

BPT upon a technology that is less effective than the best

practicable technology demonstrated in the industry.'®

Rather, in establishing BPT, the EPA “‘must forbid the

level of effluent produced by the most pollution-prone seg-

ment of the industry, that segment not measuring up to

‘the average of the best existing performance.’ ’’'* We

therefore hold that the EPA was entirely reasonable in

declining to create a subcategory for plants employing

WSP treatment systems.

3. The EPA’s Consideration of the Cost

of Compliance for Plants Utilizing

Pond Technology

Petitioners argue next that the EPA’s failure to consider

waste treatability resulted in the designation of unwork-

able upgrades to achieve BPT. The EPA does not seriously

dispute petitioners’ claim that the algae growth associated

with petitioners’ present treatment systems may not be

18 American Iron & Steel Inst. v. EPA, 526 F.2d at 1061 (rejecting

petitioner’s claim that limitation was invalid because one model treat-

ment technology allegedly increased pollutants in effluent and holding

that limitations are valid so long as one of several available treatment

technologies can achieve compliance).

i EPA vy. National Crushed Stone Ass’n, 449 U.S. at 76, 101 S.Ct.

at 303.

Bi wept naiche scatman la atin oo”

7la

remedied by relatively simple and inexpensive upgrades.

Rather, the EPA notes that it “‘is entitled to look at costs

on an industry-wide basis as opposed to plant-by-plant,”’

and may use industry averages to develop “rough esti-

mates” to help the Agency determine whether the cost is

“wholly out of proportion” to the benefit.'*’

The EPA concedes that petitioners may be required to

install entirely new treatment units consisting of activated

sludge and secondary clarification. However, the EPA

notes that it estimated the costs of such steps for nearly

half of the plants in the industry requiring treatment im-

provements to comply with BPT.'* Thus, even if Texas

Eastman, DuPont, and Air Products are required to install

new activated-sludge systems in order to comply with BPT,

the costs of these systems would be within the range gen-

erally estimated for the industry as a whole. We have

already held that the costs of the BPT limits for the

OCPSF industry are not ‘wholly out of proportion” to the

benefits. Because petitioners’ compliance problems are not

attributable to their production process, raw wastewaters,

or other relevant factors, but only to their treatment sys-

tems, their costs in constructing and operating a biological

system complying with the regulations will not differ ma-

terially from those incurred by many other industry mem-

bers.° We therefore agree with the EPA that even this

'** Wyandotte Corp v. Costle, 598 F.2d at 662. As the Third Circuit

has stated, “[njothing in the Clean Water Act requires that a cost

benefit analysis be made industry by industry, or plant by plant, or

treatment technique by treatment technique.’ Ford Motor Co. v. EPA,

718 F.2d 55, 59 (3rd Cir.1983); accord American Iron & Steel Inst. v.

EPA, 526 F.2d at 1053. Similarly, Congress intended that economic

impacts be determined only for classes of facilities, rather than on a

plant-by-plant basis. 1972 Leg.Hist. at 255, 304.

“ Dev. Doc. at VIII-B53-59, reprinted in Joint App. at 4585-91.

“To the extent that the petitioners argue that EPA undercosted

these three plants and thus underestimated total industry BPT costs

as well, we agree that any such error would not have a significant

72a

worst-case scenario does not provide a basis for exempting

petitioners from the BPT limits that apply to the rest of

the industry.

4. Copper-Sulfate Treatment for

Algae Control

Finally, petitioners contend that although the EPA pro-

posed copper sulfate as a means of controlling algae at a

reasonable cost, this method was not considered in the

rulemaking and would in fact impair the efficiency of their

ponds and possibly result in excess discharges of copper,

thereby causing water-quality problems.

While petitioners do raise legitimate concerns regarding

the practicability of employing copper-sulfate treatment as

a means of complying with BPT, the EPA did not rely

on copper-sulfate treatment alone in estimating petitioners’

compliance costs. :

The EPA’s cost estimates for Union Carbide, Texas

Eastman, and DuPont were based on far more substantial

treatment than the application of copper sulfate. For Union

Carbide, the EPA costed an entirely new activated sludge

system, which generates no algae. For both Texas East-

man and DuPont, the EPA estimated the costs of installing

an additional chemically-assisted clarifier to enhance solids

removal, and, for Texas Eastman, the EPA also costed

an upgrade to its biological unit to enhance BODS re-

moval.!©

impact on EPA’s conclusions concerning BPT costs. EPA reasons that

at worst, under petitioners’ reasoning, it should have costed new bi-

ological treatment systems for 107 rather than 104 plants. Thus, the

total industry-wide costs would increase only slightly, and the industry-

wide cost per pound of conventional pollutants removed would only

increase marginally from the 71 cents figure.

1% Dev. Doc. at VIII-B56-57, reprinted in Joint App. at 4590 and

4587.

73a

Again, while the EPA concedes that the treatment

methods costed for Texas Eastman and DuPont may not

in fact enable their WSP treatment systems to achieve

the BPT limits, the EPA reasonably concluded that there

is no basis on which petitioners may be excluded from the

BPT rules.

Petitioners also argue that the BPT limitations should

be remanded because the cost of copper sulfate as an algae

treatment was not specifically discussed in the EPA’s pro-

posals. However, the EPA notes that the control of algae

by using copper sulfate was not the basis for BPT and

was considered as part of a subsidiary analysis that spe-

cifically responded to petitioners’ comments.'®

Thus, to the extent that the costing of copper-sulfate

treatment for algae played a role in the rulemaking pro-

ceeding, it was a “logical outgrowth” of the proceedings

in general and of petitioners’ own comments in particu-

lar. 16

5. The Cost of the BPT Limitations

Petitioners argue that the cost of BPT, even apart from

the issue of algae-control costs, is wholly disproportionate

‘The Ponds Committee had submitted detailed comments in 1983

and 1985 urging EPA to consider the special problems its members

had experienced in treating algae. Joint App. at 410-34, 2079-318. More

specifically, the Committee’s 1985 comments included a review of var-

ious algae-treatment methods and specifically discussed the use of chem-

ical doses in general and copper sulfate in particular to remove algae.

Id. at 2090. The Committee attached an API analysis that stated that

copper sulfate is one of the two chemicals commonly used for algae

control and that it is an ‘‘effective poison for microscopic vegeta-

tion....” Id. at 2281-83. The sole reservation expressed in the analysis

was the potential for adverse effects on fish if discharged at a level

that exceeds 140 ug/l or results in the depletion of dissolved oxygen

caused by algae decay. Jd. at 2283.

'* See Brazos Elect. Power Coop., Inc. v. Southwestern Power Admin.,

819 F.2d 537, 542-43 (5th Cir.1987); Taylor Diving & Salvage Co. v.

Department of Labor, 599 F.2d 622, 626 (5th Cir.1979).

74a

to its benefits. We held above that the cost of BPT was

not wholly disproportionate to its benefits and therefore

need not further address petitioner’s claims to the con-

trary.

F. ‘Plant Specific’? Claims of Union-Carbide, Borg-War-

ner, DuPont, Monsanto, and Ethyl

Several petitioners, including Union Carbide, Borg-War-

ner, DuPont, Monsanto, and Ethyl, claim that the BPT

limitations are arbitrary because the EPA failed to account

adequately for wastestream characteristics. Petitioners as-

sert that the wastestream characteristics of certain of their

plants preclude compliance with the OCPSF industry lim-

itations. These claims are considered together because sev-

eral of these petitioners raise highly individualized

objections to the limitations.

We will address each petitioner’s claim in turn. As an

initial matter, however, we note that the EPA is not re-

quired to consider fundamentally different factors of par-

ticular plants in the national BPT rulemaking. Both

Congress and the Supreme Court have expressed concern

that the process of formulating nationally applicable water-

quality standards would be unduly impeded by requiring

EPA to address the idiosyncracies of individual plants in

the context of a national rulemaking. The Supreme Court

has held that the fundamentally-different-factors (FDF)

variance procedure provides an entirely acceptable alter-

native to subcategorizing an industry to account for plant-

specific characteristics.'* Congress has codified the FDF

procedures in the CWA, encouraging the EPA not to com-

plicate and delay unduly the promulgation of national ef-

fluent-limitation guidelines and standards where the FDF

procedure could be employed to address the concerns of

individual facilities claiming to be unique.’

‘8 Chemical Mfrs. Ass’n v. NRDC, 470 U.S. at 116, 105 S.Ct. 1102,

84 L.Ed.2d 90.

1 33 U.S.C. § 1311(n).

75a

The Supreme Court held in 1977 that the EPA may

establish categorical BPT limitations, “so long as some

allowance is made for variations in individual plants, as

EPA has done by including a variance clause in its 1977

limitations.” Relying heavily on legislative history that

demonstrated Congress’ intent to replace the site-specific

approach to water-quality regulation with technology-based

limitations that apply uniformly to categories of dischar-

gers,'®* the unanimous Court reasoned that the alternative

view “‘would place an impossible burden on EPA”’ contrary

to the legislative purpose of the Act.'*

Addressing the EPA’s identical FDF variance procedure

for pretreatment standards, the Supreme Court has ap-

proved the procedure ‘“‘as a mechanism for insuring that

[EPA’s] necessarily rough-hewn categories do not unfairly

burden atypical plants.’* The Court explained:

EPA and CMA point out that the availability of

FDF variances makes bearable the enormous burden

faced by EPA in promulgating categories of sources

and setting effluent limitations. Acting under strin-

gent timetables, EPA must collect and analyze large

amounts of technical information concerning complex

industrial categories. Understandably, EPA may not

be apprised of and will fail to consider unique factors

applicable to atypical plants during the categorical ru-

lemaking process, and it is thus important that EPA’s

nationally binding categorical pretreatment standards

8 FE. I.du Pont de Nemours & Co. v. Train, 430 U.S. at 128, 97 S.Ct.

at 975.

“Id. at 121, 126-27, 129-30, 97 S.Ct. at 971, 974; Leg.Hist. at 169,

1421-26, 1460.

'* DuPont, 430 U.S. at 132-33, 97 S.Ct. at 977; Leg.Hist. at 170-72,

304, 1468.

Sa ical Mfrs. Ass’n v. NRDC, 470 U.S. at 120, 105 S.Ct. at

1105

76a

for indirect dischargers be tempered with the flexi-

bility that the FDF variance mechanism offers. . . . '®

The Court stated that the FDF variance procedure was

authorized by Congress in significant part to ensure that

the national rule would not be overturned simply because

of the Agency’s failure to consider unique plants.!”° Several

courts of appeal have subsequently relied upon the avail-

ability of an FDF variance procedure as the basis for re-

jecting challenges to BPT regulations that are based upon

allegedly facility-specific factors.!”

In codifying the FDF variance procedure in the CWA,'”

Congress specifically emphasized that the procedure serves

as a ‘“‘safety valve’ to the categorical statutory scheme,

allowing EPA to address plant-specific variations through

a separate administrative process, outside of the national

rulemaking. The House Report stated:

There are two approaches for responding to a facility

with valid grounds for arguing that it is fundamentally

different from other facilities in its category. One pos-

sibility is to develop a separate subcategory within

the regulation, undertake a separate data collection

and analysis effort and then repropose and issue the

final rule. The other alternative is to leave the na-

tional rule in place and use the FDF determination

procedure to establish alternative technology-based

1° Jd. at 132-33, 105 S.Ct. at 1111-12.

17 Jd. at 183 & n. 25, 105 S.Ct. at 1112 & n. 25. The Court cited

numerous Supreme Court decisions upholding regulations in part be-

cause they provided for the granting of exceptions or variances. Jd.

1” Kennecott Copper v. EPA, 612 F.2d at 1244- 45; BASF Wyandotte

Corp. v. Costle, 598 F.2d at 656; Weyerhaeuser Co. v. Costle, 590 F.2d

at 1040-41, 1048 n. 56; American Iron & Steel Inst. v. EPA, 568 F.2d

at 305; American Iron & Steel Inst. v. EPA, 526 F.2d at 1049, 1061;

cf. American Frozen Food Inst. v. EPA, 539 F.2d at 142-43.

1% Section 306 of the Water Quality Act of 1987; 33 U.S.C. § 1311(n).

77a

limitations for the facility that accurately reflect its

situation. The subcategorization approach would add

further complications and require potentially substan-

tial additional time in developing what are already

extraordinarily complex and detailed national regula-

tions. By contrast, the FDF determination procedure

allows both implementation of the national rule and

consideration of individual petitions claiming unique

factors.!8

Given Congress’ clear intent that the national rulemak-

ing process not be unduly impeded by highly individualized

objections to the regulations that would be more appro-

priately addressed in an FDF proceeding, we address the

following claims with this concern in mind.

1. Union Carbide

Union Carbide argues that high influent BODS concen-

trations at its Taft, Louisiana, plant preclude effective

treatment by the model BPT technology. Union Carbide

thus contends that the EPA has neither costed nor iden-

tified any BPT technology which will enable plants with

high-BOD influent to comply with BPT. The EPA costed

the addition of a biological treatment unit followed by a

secondary clarifier to Taft’s existing treatment system.!”

This additional treatment system will remove almost four

million pounds of conventional pollutants at an annualized

cost of $1,242,220.'

The EPA reasonably concluded that the Taft plant was

not so fundamentally different from the industry as a whole

as to warrant exclusion from the rule simply because the

Taft plant had the highest levels of BODS influent in the

'? H.R.Rep. No. 189, 99th Cong., Ist Sess. 26 (1985).

'* Dev.Doc. at VIII-B53, reprinted in Joint App. at 4585, 4961-62.

'* Dev.Doc. at VIII-C1 (plant #296), reprinted in Joint App. at 4609;

Dev.Doc. at VIII-B2 (plant #296), reprinted in Joint App. at 4533.

78a

data base used to develop the limits. The EPA has con-

cluded reasonably that the limitations are both achievable

and practicable for the Taft plant.

2. Borg-Warner

Borg-Warner seeks a separate subcategory for wastes-

treams containing significant amounts of phenol, alleging

that high phenol concentrations adversely impact biological

treatment by inhibiting biodegradability. The EPA specif-

ically considered the effects of phenol concentrations and

found that several plants with high phenol influents

achieved low levels of biological oxygen demand.’* The

EPA found that phenol-dominated wastestreams were

treatable and therefore declined to create a subcategory

for phenol-dominated wastestreams. The EPA’s conclusion

is supported by the record and is not arbitrary or capri-

cious.

3. DuPont’s Chambers Works Plant

DuPont claims that the BPT limits for TSS are not

attainable by its Chambers Works plant through the use

of the technology relied upon in establishing the regula-

tions. DuPont maintains that the EPA failed to take into

account the nature of the wastewaters subject to regula-

tion and specifically failed to take into account the fact

that ‘“[djue to the complexity and unique aspects of its

manufacturing process’ the wastestream of the Chamber

Works plant contains TSS and mixed liquor suspended

solids at significantly higher levels than did the wastes-

treams of the plants in the EPA’s data base. DuPont ar-

gues that its wasiestream is also fundamentally different

from other plants in that its TSS/BODS ratio is nearly

four times greater than the average.

'* CMA Informal Comments on EPA’s Draft Contractors Engineering

Report, Nov. 1982, reprinted in Joint App. at 4686-94; Dev.Doc. VII-

A2, reprinted in Joint App. 4411.

79a

While DuPont asserts that the EPA failed in general

to adequately consider wastestream characteristics in its

subcategorization of the industry, it does not propose any

specific basis on which the EPA should have created a

subcategory that would address the characteristics of the

Chambers Works plant wastestream.'” Rather, DuPont as-

serts that the limitations should be set aside because the

EPA failed to take into account ‘“‘unique”’ characteristics

of the wastewater at the Chambers Works plant and re-

jected data submitted by DuPont that demonstrate the

“unachievability” of the TSS limitations.

As we note above, the EPA is not obligated to address

in its national rulemaking the fundamentally different char-

acteristics of an individual plant; such concerns are ap-

propriately raised in an FDF variance proceeding.'”

DuPont has filed for an FDF variance, and the claims

raised here are more properly addressed in the first in-

stance in that forum. Accordingly, we express no opinion

on whether such a variance would be appropriate.

4. Monsanto and Ethyl

Petitioners Monsanto and Ethy]! similarly claim that the

EPA failed to adequately account for wastewater char-

acteristics in setting BPT limits and that the EPA’s sub-

categorization scheme is therefore uniawful. Monsanto

claims specifically that plants using BPT technology may

not be able to achieve the EPA’s TSS limits where their

wastewaters contain high total-dissolved-solids (TDS) levels

and that the EPA should have created a separate subca-

tegory for plants that have high levels of TDS in their

wastewater. Monsanto maintains that two of its plants will

be unable to comply with the TSS limits for this reason.

'" To the extent that DuPont challenges the overall reasonableness

: EPA's subcategorization, that issue is addressed supra in section

D.

'® See 33 U.S.C. § 1311(n).

80a

Similarly, Ethyl asserts that its Elgin plant will be unable

to comply with the TSS limits because the plant’s was-

testream contains high levels of brine. Ethyl claims that

dilution is necessary to treat wastewater with high brine

content but that the BPT effluent limits are based on

process flow only, without allowing for dilution. Ethyl notes

that it may be possible to meet the standards based on

effluent only but that the EPA has not identified the ap-

propriate technology for doing so and therefore has not

accounted for the cost of such technology in its BPT lim-

itations.

In response to these objections, the EPA asserts that

Ethyl] failed to submit any comments during the rule-mak-

ing proceeding that would establish that a plant’s TDS

levels would preclude compliance or greatly increase the

cost of compliance and that only one company in the in-

dustry, Monsanto, claimed during the rulemaking to have

a compliance problem caused by high TDS levels in its

wastestream.’” The EPA states that it therefore declined

to create a subcategory based on TDS levels because there

was not sufficient information in the record to demonstrate

that any plant in the OCPSF industry could not comply

with the TSS limits as a result of elevated TDS levels.

The EPA notes that while CMA also commented on the

relationship between TDS and TSS, its comments weighed

against the creation of a separate subcategory based on

'* The EPA claims that Monsanto did not submit information suffi-

cient to establish a basis for creating a subcategory based on TDS

levels. Specifically, the EPA asserts that “Monsanto did not state what

OCPSF product/processes, if any, discharge TDS-laden wastewaters and

at what frequency such wastewaters are discharged; nor did Monsanto

state what portion of the total process wastewater flow contained el-

evated levels of TDS.” The EPA also claims that Monsanto failed to

provide data on actual influent TDS and effluent TSS levels to support

its claim.

8la

TDS influent levels. The EPA determined that technol-

ogy does exist to clarify wastestreams with especially high

levels of solids and that facilities use a variety of methods

to ensure the effective biological treatment of unique was-

testreams that contain pollutants that impede biological

treatments. The EPA noted, for example, that technologies

such as reverse osmosis can eliminate materials in a plant’s

wastewater which may inhibit or upset biological treatment

systems. '*!

In specific response to Monsanto’s comment, the EPA

stated that three facilities which have TDS levels exceed-

ing 5,000 mg/l have nevertheless achieved good TSS re-

moval.'*? The EPA further concluded that few, if any,

OCPSF plants have TDS levels of sufficient magnitude to

impair TSS removal.’ Accordingly, the EPA rejected

Monsanto’s request that a correction factor for high TDS

levels be incorporated into the final TSS limits.’

Although Monsanto takes issue with the EPA’s re-

sponses, we conclude that the EPA’s decision not to es-

” CMA objected to EPA’s proposed influent design limits of 10,000

mg/l TDS for activated sludge systems, claiming in general terms that

some OCPSF plants successfully ‘operate with influent TDS levels as

high as 15,000-30,000 mg/l without significant problems.” CMA’s No-

vember 24, 1982 Comments on EPA's Draft Contractor's Engineering

Report, reprinted in Joint App. at 4677. CMA stated that while such

Systems may require more operator attention or special design consid-

erations, they performed satisfactorily and offered economic advantages

over other designs. Jd.

™ 50 Fed.Reg. 29,075; Dev.Doc. IV-28, reprinted in Joint App. at

3527. The EPA also noted that some plants use equalization to blend

Specific process wastestreams on a controlled basis, into the plant's

wastewater treatment system. /d.

_'™ Agency Response to the July 17, 1985 and October 11, 1985 No-

ces of Availability of New Information, Comment #166, reprinted in

Joint App. at 3013-14.

aad |

at |

82a

tablish a special subcategory based upon TDS levels was

reasonable. The EPA found, based upon the record before

it, that it was uncertain at best whether any plant in the

industry had TDS levels that precluded effective treatment

and that if there were any such problems, they would be

unique to Monsanto, the only company to claim TDS prob-

lems. Therefore, Monsanto’s concerns would be more prop-

erly addressed through an FDF variance proceeding than

through the national rulemaking.

The EPA notes that, unlike Monsanto, Ethyl never sub-

mitted comments—in response to either the proposed reg-

ulations or the three subsequent public notices—to inform

the EPA that it believed its Elgin plant would experience

TSS compliance problems as a result of the level of TDS

in the plant’s wastewater. Ethyl asserts that it does not

suggest that the EPA should have created a separate sub-

category for its Elgin plant, but rather submits the ar-

gument that the Elgin plant cannot meet the TSS limits

as evidence that the EPA failed adequately to consider

wastewater characteristics in establishing BPT. To the ex-

tent that Ethyl challenges the overall validity of the EPA’s

subcategorization approach, based on the EPA’s alleged

failure to adequately consider wastestream characteristics,

that issue is addressed above. Furthermore, as we also

conclude above, the fact that a single plant may have

difficulty in meeting BPT requirements due to unique char-

acteristics of that plant, does not render the entire rule-

making invalid. To the extent that Ethyl does raise

concerns unique to its Elgin plant, those issues are prop-

erly raised through an FDF variance proceeding rather

than through the national rulemaking.

5. FDF Variances

The highly individualized claims of DuPont, Monsanto,

and Ethyl are more appropriately addressed in an FDF

administrative proceeding. DuPont and Ethyl have filed

for variances, and Monsanto may certainly do so in the

| at Ai le

83a

near future. Petitioners argue, however, that we may not

decline to address their claims because the EPA has not

yet ruled on the applications. Thus, petitioners claim that

the EPA will leave them in “administrative limbo’’ while

the limitations go into effect. To the extent, however, that

petitioners DuPont and Ethyl seek to compel an Agency

decision on their FDF applications, such relief is outside

the scope of this court’s limited jurisdiction under CWA

Section 509(bX1) to review the EPA’s effluent limitations

guidelines, pretreatment standards, and new source per-

formance standards.'*

This court would have jurisdiction to review the FDF

claims only after the EPA has ruled on petitioners’ ap-

plications and after review by the district court.'** The

Act’s regulatory scheme is consistent with the prudential

doctrine of primary jurisdiction which holds that complex

scientific and technical issues, such as those presented here,

should be resolved in the first instance by the EPA, the

entity best suited to pass on these issues.'*’

Congress was aware of the difficulty and corresponding

delays in processing FDF variance applications for indi-

vidual plants.’ Accordingly, Congress attempted to ex-

pedite the process by specifically requiring the EPA to

determine the merits of applications for FDF variances for

individual plants within 180 days of the submission of the

application.’*® Thus, the WQA ensures that all FDF ap-

* 33 U.S.C. § 136%b\1).

‘Id.

" McKart v. United States, 395 U.S. 185, 194, 89 S.Ct. 1657, 1662,

23 L.Ed.2d 194 (1969).

i S.Rep. No. 50, 99th Cong., Ist Sess. 20-21 (1985).

” Water Quality Act of 1987 (WQA), 33 U.S.C. § 1311(n\3). Addi-

tionally, the WQA treats all FDF applications that were still pending

on the date of enactment as having been submitted to the Agency on

the 180th day following the date of enactment, and therefore, as falling

the ambit of the WQA itself. 33 U.S.C. § 1311(n\5).

84a

plications on which the EPA had not previously ruled would

henceforth be subject to the Act’s 180-day time limit.

Notwithstanding Congress’ 1

This text is long and has been trimmed here. Open the source document for the complete record.