13‐1745(L) NRDC v. EPA

UNITED STATES COURT OF APPEALS FOR THE SECOND CIRCUIT

August Term 2014

(Argued: January 30, 2015 Decided: October 5, 2015 Amended: December 18, 2015)

Docket Nos. 13‐1745(L), 13‐2393(CON), 13‐2757(CON)

NATURAL RESOURCES DEFENSE COUNCIL, NORTHWEST ENVIRONMENTAL ADVOCATES, CENTER FOR BIOLOGICAL DIVERSITY, and NATIONAL WILDLIFE FEDERATION,

Petitioners,

v.

UNITED STATES ENVIRONMENTAL PROTECTION AGENCY,

Respondent,

LAKE CARRIERSʹ ASSOCIATION and CANADIAN SHIPOWNERS ASSOCIATION,

Intervenors.

ON PETITION FOR REVIEW FROM THE ENVIRONMENTAL PROTECTION AGENCY

Before: SACK, CHIN, and CARNEY, Circuit Judges.

Four environmental organizations petition for review of a Vessel

General Permit issued by the Environmental Protection Agency in 2013 under

Section 509(b)(1) of the Clean Water Act,

(b)(1). The permit

regulates the discharge of ballast water from ships, a primary cause of the spread

of invasive species from one body of water to another. Petitioners contend that

the Environmental Protection Agency acted arbitrarily and capriciously in

issuing the permit, and request that it be set aside. We agree, in part.

Accordingly, we grant the petition for review in part and deny it in part, and

remand to the Environmental Protection Agency for further proceedings

consistent with this opinion. We do not vacate the Vessel General Permit, but

allow it to remain in effect until the issuance of a new Vessel General Permit.

PETITION GRANTED IN PART AND DENIED IN PART AND REMANDED.

‐2‐

ALLISON M. LAPLANTE, Earthrise Law Center, Lewis & Clark Law School, Portland, Oregon, and Deborah A. Sivas, Matthew J. Sanders, Environmental Law Clinic, Mills Legal Clinic at Stanford Law School, Stanford, California, for Petitioners Northwest Environmental Advocates and Center for Biological Diversity.

Rebecca J. Riley, Natural Resources Defense Council, Chicago, Illinois, for Petitioner Natural Resources Defense Council.

Neil S. Kagan, National Wildlife Federation, Ann Arbor, Michigan, for Petitioner National Wildlife Federation.

MARTIN FRANCIS MCDERMOTT (Sam Hirsch, Acting Assistant Attorney General, on the brief), Environmental Defense Section, Environmental & Natural Resources Division, United States Department of Justice, Washington, D.C., and Dawn M. Messier, Office of General Counsel, United States Environmental Protection Agency, Washington, D.C., for Respondent United States Environmental Protection Agency.

MATTHEW D. MELEWSKI, The Boutique Firm PLC, Minneapolis, Minnesota, for Intervenors Lake Carriersʹ Association and Canadian Shipowners Association.

‐3‐

CHIN, Circuit Judge:

This case arises from the efforts of the Environmental Protection

Agency (ʺEPAʺ) pursuant to section 402(a) of the Clean Water Act (the ʺCWAʺ),

(a), to regulate the discharge of ballast water from ships. 1 A ship

takes on and discharges ballast water to compensate for changes in its weight

caused by activities such as loading and unloading cargo or consuming fuel or

supplies. The amount of water can range from hundreds of gallons to as much

as 25 million gallons ‐‐ enough to fill thirty‐eight Olympic‐sized swimming

pools. More than 21 billion gallons of ballast water are released in the United

States annually. See Nw. Envtl. Advocates v. EPA,

(9th Cir.

2008).

When a ship takes on ballast water, it can inadvertently pick up

organisms and their eggs and larvae, as well as sediment and pollutants. When

the ship discharges ballast water, often in a new place, these organisms and

pollutants are ejected into the surrounding waterbody, enabling these organisms

to establish new, non‐native populations. As a result, ships have become one of

1 Glossary of Acronyms: As this opinion discusses the CWA and its intricacies, it contains a large number of acronyms. In addition to their definitions in the text, a separate glossary of acronyms is therefore set forth in the Appendix to this opinion.

‐4‐

the primary ways that invasive species are spread from one waterbody to

another.

at 1012‐13 (ʺAll told, more than 10,000 marine species each day hitch

rides around the globe in the ballast water of cargo ships.ʺ (quoting Nw. Envtl.

Advocates v. EPA, No. C 03‐05760 SI,

, at *3 (N.D. Cal. Sept. 18,

2006)).

Invasive species cause severe economic and ecological harm,

including by destroying native fish species and shellfish industries, creating

algae blooms, and devastating tourism. Zebra mussels are a particularly

destructive example. They were first introduced to Lake Erie in the 1980s by a

freighter from Europe that discharged ballast water containing mussels.2 These

mussels have wreaked havoc in the Midwest and Northeast by blocking water

intake and outtake at power plants and other industrial facilities, causing nearly

$70 million in damage between 1989 and 1995. Nw. Envtl. Advocates,

. One study estimates the damage caused by invasive species collectively at

ʺabout $137 billion a year ‐‐ more than double the annual economic damage

2 ʺFrom that humble start, the invaders colonized the Great Lakes and spread across the country on towed boats.ʺ Jim Robbins, A Western Showdown, N.Y. TIMES, Sept. 8, 2015, at D6.

‐5‐

caused by all natural disasters in the United States.ʺ

(quoting Nw. Envtl.

Advocates,

, at *4).3

Ballast water discharge is particularly problematic in the Great

Lakes. Vessels that sail exclusively in the Great Lakes, known as ʺLakers,ʺ

account for over ninety‐five percent of ballast water volumes transferred in the

Great Lakes. Unfortunately, Lakers are more likely than oceangoing vessels to

spread invasive species because the short duration of their voyages allows

organisms to survive in their ballast.

In April 2013, EPA issued a Vessel General Permit (the ʺ2013 VGPʺ),

pursuant to section 402 of the CWA,

, to regulate the discharge of

ballast water from ships. In response, four environmental groups filed three

Petitions for Review (ʺPFRsʺ) alleging that EPA acted arbitrarily and capriciously

in issuing the 2013 VGP: petitioner Natural Resources Defense Council (ʺNRDCʺ)

filed a PFR on May 3, 2013 in this Court; petitioners Northwest Environmental

Advocates (ʺNWEAʺ) and the Center for Biological Diversity jointly filed a PFR

on May 3, 2013 in the United States Court of Appeals for the Ninth Circuit; and

petitioner National Wildlife Federation (ʺNWFʺ) filed a PFR on July 3, 2013 in the

3 See also Robbins, A Western Showdown, at D6 (discussing damage caused by zebra and quagga mussels).

‐6‐

United States Court of Appeals for the D.C. Circuit.4 In an order dated May 24,

2013, the Judicial Panel on Multidistrict Litigation issued a Consolidation Order

and assigned final venue for the first two petitions, and any subsequently filed

petition, to this Court.

On May 31, 2013, the Lake Carriersʹ Association and the Canadian

Shipowners Association (the ʺCSAʺ) filed a motion to intervene, which was

granted on October 7, 2013. On January 1, 2014, the CSA filed a PFR in this case.

EPA and the CSA jointly moved to sever the CSA PFR from this case and hold it

in abeyance; the motion was granted on May 23, 2014.

We find that EPA acted arbitrarily and capriciously in issuing parts

of the 2013 VGP, and therefore remand this matter to the EPA for further

proceedings.

BACKGROUND

A. The CWA

Congress created the CWA to limit pollution in the waters of the

United States. See

(a) (objective of CWA is to ʺrestore and

4 All three petitions were timely filed within 120 days of the issuance of the VGP, as required under

(b)(1). Accordingly, this Court has jurisdiction over the petitions pursuant to section 509(b)(1)(F) of the CWA,

(b)(1)(F).

‐7‐

maintain the chemical, physical, and biological integrity of the Nationʹs watersʺ);

S. Fla. Water Mgmt. Dist. v. Miccosukee Tribe of Indians,

(2004)

(same); Waterkeeper All., Inc. v. EPA,

, 490‐91 (2d Cir. 2005) (same).

The CWA thus prohibits the ʺdischarge of any pollutantʺ from a ʺpoint sourceʺ to

the ʺnavigable watersʺ of the United States, except as permitted by the CWA.

(a), 1362 (emphasis added). The ʺdischarge of a pollutantʺ includes

ʺany addition of any pollutant to navigable waters from any point source.ʺ

§ 1362(12)(A). A ʺpollutantʺ includes solid, industrial, agricultural, and

biological waste. Id. § 1362(6). A ʺpoint sourceʺ is ʺany discernible, confined and

discrete conveyance, including but not limited to any . . . vessel or other floating

craft, from which pollutants are or may be discharged.ʺ Id. § 1362(14).

ʺNavigable watersʺ is defined as ʺthe waters of the United States, including the

territorial seas.ʺ Id. § 1362(7). The discharge of polluted water from a vessel

ballast tank is a point source discharge covered by the CWA. See Nw. Envtl.

Advocates,

.

A key component of the statute is the establishment of water quality

standards. Water quality standards are set by states for waters within their

boundaries and are then reviewed for approval by EPA. See

; 40

‐8‐

C.F.R. §§ 131.4, 131.10‐.11; see also NRDC v. EPA,

, 1183 (9th Cir.

2002) (ʺUnder the CWA, each state sets its own water quality standards, subject

to review and approval by the EPA.ʺ). EPA must ensure that the standard

proposed by the state will comply with the requirements of the CWA before

approving it. See

(b)(1)(C), 1313(a) 1342(a)(1);

(d).

1. National Pollutant Discharge Elimination System Permits

An entity seeking to discharge a pollutant is required to obtain and

comply with a permit that limits the amounts and kinds of pollutants being

discharged. See NRDC v. EPA,

, 108 (D.C. Cir. 1987); see also

Waterkeeper All., 399 F.3d at 498 (discharge allowed ʺwhere . . . permits ensure that

every discharge of pollutants will comply with all applicable effluent limitations

and standardsʺ). This permit, known as a National Pollutant Discharge

Elimination System (ʺNPDESʺ) permit, establishes enforceable effluent

limitations, as well as monitoring and reporting requirements.

NPDES permits, which are issued either by EPA or a state in a

federally approved permitting system, see

, may be individual

(issued to a specific entity to discharge pollutants at a specific place) or general

‐9‐

(issued to an entire class of dischargers in a geographic location), see

, 122.28(a)(2), 124.1‐.21, 124.51‐.66. The permit here is a general permit.

Permits can impose two different types of standards on discharges:

(1) technology‐based standards and (2) water quality‐based standards. See

(b)(1)(c) and (b)(2)(a), 1313, 1342(a). The 2013 VGP imposes both.

a. Technology‐Based Effluent Limits

Technology‐based effluent limits (ʺTBELsʺ) set effluent limitations

on a point source based on how effectively technology can reduce the pollutant

being discharged. See

(b), (e), 1314(b); see also PUD No. 1 of

Jefferson Cty. v. Wash. Depʹt of Ecology,

(1994) (holding that, to

achieve goals of CWA, EPA is required to ʺestablish and enforce technology‐

based limitations on individual discharges into the countryʹs navigable waters

from point sourcesʺ). Congress designed this standard to be technology‐forcing,

meaning it should force agencies and permit applicants to adopt technologies

that achieve the greatest reductions in pollution. See NRDC, 822 F.2d at 124

(holding that CWA seeks ʺnot only to stimulate but to press development of new,

‐10‐

more efficient and effective technologies,ʺ which is ʺessential purpose of this

series of progressively more demanding technology‐based standardsʺ).5

In determining the standard for TBELs, EPA considers the source of

the pollution (existing or new) and the type of pollutant. For nonconventional

pollutants from existing sources, EPA is required to set effluent limits based on

the ʺbest available technology economically achievableʺ or ʺBAT.ʺ

(b)(2)(A).6 BAT requires the ʺapplication of the best available technology

economically achievable for such category or class, which will result in

reasonable further progress toward the national goal of eliminating the discharge

of all pollutants.ʺ Id.; see NRDC, 822 F.2d at 123 (CWA designed to progress

ʺtoward implementation of pollution controls to the full extent of the best

technology which would become availableʺ). Because invasive species are a

5 EPA issues national effluent limitation guidelines (ʺELGsʺ), which establish limitations for all types of dischargers within a particular industry and for certain types of discharges. See

(c)(1). ELGs are enforceable through their incorporation into a NPDES permit. In this case, no states have established numeric water quality criteria for living organisms or aquatic nuisance species. 6 For conventional pollutants from existing sources, the level of pollution control is based on best conventional pollutant control technology.

§ 1311(b)(2)(E). New sources of pollution must meet new source performance standards, which are based on best available demonstrated control technology. Id. § 1316(a)(1). Neither standard is implicated here.

‐11‐

nonconventional pollutant from an existing source, ballast water discharges are

subject to BAT.

EPA considers a number of factors in assessing whether a

technology is BAT, including:

 the cost of achieving the effluent reductions,

 the age of equipment and facilities involved,

 the process employed,

 the engineering aspects of various control techniques,

 potential process changes,

 non‐water‐quality environmental impacts including energy requirements, and

 other factors as EPA ʺdeems appropriate.ʺ

See

(b)(2)(B).

EPA can mandate that BAT requires the use of a technology that is

not currently available within a particular industry when (1) the technology is

available in another industry, (2) EPA finds that the technology is transferrable

from that other industry, and (3) EPA can reasonably predict that such

technology will adequately treat the effluent. See Kennecott v. EPA,

,

453 (4th Cir. 1986) (citing Tannersʹ Council of Am., Inc. v. Train,

(4th Cir. 1976)).

‐12‐

b. Water Quality‐Based Effluent Limits

If the TBELs are insufficient to attain or maintain water quality

standards, the CWA requires NPDES permits to include additional water

quality‐based effluent limits (ʺWQBELsʺ). See

(b)(1)(C), 1312(a);

NRDC, 822 F.2d at 110 (ʺWhenever a technology‐based effluent limitation is

insufficient to make a particular body of water fit for the uses for which it is

needed, the EPA is to devise a water‐quality based limitation that will be

sufficient to the task.ʺ). WQBELs are designed to ensure that the discharges

authorized by the permit do not violate water quality standards. See

, 1342(a)(2).

The WQBELs, which supplement the TBELs, are based on the

amount and kind of pollutants in the water. See

§ 1312(a). WQBELs are set

without regard to cost or technology availability. See NRDC v. EPA,

,

208 (D.C. Cir. 1988) (ʺA technology‐based standard discards its fundamental

premise when it ignores the limits inherent in the technology. By contrast, a

water quality‐based permit limit begins with the premise that a certain level of

water quality will be maintained, come what may, and places upon the permittee

the responsibility for realizing that goal.ʺ (footnote omitted)). WQBELs may be

‐13‐

narrative where the calculation of numeric limits is ʺinfeasible.ʺ See

(k)(3).

No permit may be issued when ʺthe imposition of conditions cannot

ensure compliance with the applicable water quality requirements of all affected

States.ʺ

§ 122.4(d). Thus, permits must establish limits on discharges that will

lead to compliance with water quality standards. See Trs. for Alaska v. EPA,

, 556‐57 (9th Cir. 1984) (holding that permit must translate state water

quality standards into end‐of‐pipe effluent limitations necessary to achieve those

standards).

Because no states have established numeric water quality criteria for

invasive species, EPA is required to establish WQBELs that ensure compliance

with narrative criteria, designated uses, and antidegradation policies that

comprise state water quality standards. The permit may then mandate ʺbest

management practicesʺ (ʺBMPsʺ) to control pollution. See

(k)(3).

c. Monitoring and Reporting Requirements

NPDES permits also require both monitoring and reporting of

monitoring results of TBELs and WQBELs to assure compliance with permit

‐14‐

limitations and facilitate enforcement. See

, 1318, 1342(a)(2);

(i)(1)‐(2).

B. Regulatory History

When the CWA was first being implemented in the 1970s, EPA

regulations exempted discharges that were ʺincidentalʺ to the ʺnormal operationʺ

of vessels from NPDES permitting requirements. See National Pollutant Discharge

Elimination System,

, 13,530 (May 22, 1973) (codified at

); see also National Pollutant Discharge Elimination System; Revision of

Regulations,

, 32,902 (June 7, 1979) (codified at

(a)). This exemption included ballast water discharges.

In 1999, the NWEA and other environmental organizations

submitted a rulemaking petition to EPA seeking to repeal this exemption, then

codified at

(a). See Final National Pollutant Discharge Elimination

System (NPDES) General Permit for Discharges Incidental to the Normal Operation of a

Vessel,

, 79,475 (Dec. 29, 2008). EPA denied the petition. See

Nw. Envtl. Advocates,

. The environmental groups challenged the

denial in the United States District Court for the Northern District of California,

while simultaneously filing a PFR in the United States Court of Appeals for the

‐15‐

Ninth Circuit, in case the district court lacked jurisdiction. The district court

issued an order vacating the exemption, see Nw. Envtl. Advocates,

, at *15, and the Ninth Circuit upheld the decision. See Nw. Envtl.

Advocates,

. EPA finally repealed the exemption and issued a

Vessel General Permit in 2008 (the ʺ2008 VGPʺ). Draft National Pollutant Discharge

Elimination System (NPDES) General Permits for Discharges Incidental to the Normal

Operation of a Vessel,

(June 17, 2008).

1. The 2008 VGP

Environmental groups, industry groups, and the State of Michigan

challenged the 2008 VGP in a PFR filed in the United States Court of Appeals for

the D.C. Circuit, arguing primarily that the 2008 VGP was inadequate because it

contained only narrative provisions, not specific numeric limitations on

discharges. In March 2011, EPA settled this matter, agreeing to: (1) set ʺnumeric

concentration‐based effluent limits for discharges of ballast water expressed as

organisms per unit of ballast water volumeʺ; (2) set numeric effluent limits that

ʺrepresent the applicable levels of technology‐based controlʺ; and (3) ʺinclude

more stringent water quality‐based effluent limitationsʺ if needed to satisfy

‐16‐

applicable water quality standards. Settlement Agreement ¶¶ 9‐13, NRDC v.

EPA, No. 09‐1089 (D.C. Cir. Mar. 8, 2011), ECF No. 1296922.

2. The Creation of New Standards

To create these new, more specific standards, EPA enlisted the help

of its own Science Advisory Board (the ʺSABʺ) and the National Research

Council/National Academy of Sciences Committee on Assessing Numeric Limits

for Living Organisms in Ballast Water (the ʺNAS Committeeʺ). EPA posed a

different question to each scientific body.

a. The SAB

In 2010, EPA asked the SAB to ʺprovide advice on technologies and

systems to minimize the impacts of invasive species in vessel ballast water

discharge.ʺ App. at 599. Specifically, the SAB looked at four issues: (1) the

performance of shipboard systems with available effluent testing data; (2) the

potential performance of shipboard systems without reliable testing data; (3)

system development for the shipboard systems identified in issues 1 and 2; and

(4) the development of reliable information about the status of ballast water

treatment technologies and system performance. In considering these questions,

the SAB was to take into account The International Convention for the Control and

‐17‐

Management of Shipsʹ Ballast Water and Sediments (the ʺIMO Standardʺ), adopted

by the International Maritime Organization in 2004, which set certain

concentration‐based ballast water effluent limits.

.7

In July 2011, the SAB issued its report Efficacy of Ballast Water

Treatment Systems: A Report by the EPA Science Advisory Board (the ʺSAB Reportʺ).

The SAB identified fifty‐one ballast‐water treatment systems, with five categories

of shipboard systems that could reliably achieve the IMO Standard.

. 8

The SAB found that none of the systems could meet standards 100 or 1,000 times

greater than the IMO Standard.

. The SAB also found that none of the

fifty‐one shipboard treatments identified could reliably achieve a ʺno living

organismʺ standard.

7 The Coast Guard proposed the same standard in a rulemaking in 2011 pursuant to its authority under the National Invasive Species Act. See Standards for Living Organisms in Shipsʹ Ballast Water Discharged in U.S. Waters,

(Aug. 28, 2009). In 2012, the Coast Guard finalized the rule, entitled Standards for Living Organisms in Ships’ Ballast Water Discharged in U.S. Waters.

(Mar. 23, 2012). 8 The five categories were: (1) deoxygenation + cavitation; (2) filtration + chlorine dioxide; (3) filtration + UV; (4) filtration + UV + TiO2; and (5) filtration + electro‐chlorination.

‐18‐

b. The NAS Committee

EPA created the NAS Committee to examine ʺthe relationship

between the concentration of living organisms in ballast water discharges and

the probability of nonindigenous organisms successfully establishing

populations in U.S. waters.ʺ

.

In its June 2011 report, Assessing the Relationship Between Propagule

Pressure and Invasion Risk in Ballast Water (the ʺNAS Reportʺ), the NAS

Committee concluded (1) there was ʺno significant relationship between ballast

volume and invasions,ʺ and (2) ʺ[t]he current state of science does not allow a

quantitative evaluation of the relative merits of various discharge standards in

terms of invasion probability.ʺ

. Essentially, the Committee said that it

was unable to establish a reliable numeric limit on discharges that would

guarantee protection against invasive species, other than zero.

3. The 2013 VGP

On March 28, 2013, EPA issued the 2013 VGP, the permit now before

us, allowing vessels to discharge ballast water subject to certain limitations on

the living organisms in the discharge. Final National Pollutant Discharge

Elimination System (NPDES) General Permit for Discharges Incidental to the Normal

‐19‐

Operation of a Vessel,

(Apr. 12, 2013).9 This constituted final

action on the permit pursuant to section 402(a) of the CWA,

(a).

The 2013 VGP included TBELs, WQBELs, and monitoring and reporting

requirements.

a. TBELs

As discussed above, TBELs impose effluent limitations on a point

source based on how much technology is able to reduce the amount of a

pollutant at issue. See

(b), (e), 1314(b). In this instance, EPA

chose to set the TBELs at the IMO Standard, which requires:

(1) Limiting discharges of organisms 50 micrometers or larger to a concentration of fewer than 10 living organisms per cubic meter of ballast water;

(2) Limiting discharges of organisms less than 50 micrometers and greater than or equal to 10 micrometers to concentrations of fewer than 10 living organisms per milliliter (ʺmLʺ)of ballast water; and

(3) Limiting discharges of three types of pathogen and pathogen indicators: (1) Vibrio cholerae: fewer than 1 colony forming unit (ʺcfuʺ) per 100 mL; (2) Escherichia coli (ʺE. coliʺ): fewer than 250 cfu per 100 mL; and (3) intestinal enterococci: fewer than 100 cfu per 100 mL.

9 EPA issued the draft NPDES VGP on November 30, 2011 with a 75‐day notice‐and‐comment period. The public comment period ended on February 21, 2012. The 2013 VGP replaced the 2008 VGP, which expired on December 19, 2013.

‐20‐

Vessel General Permit for Discharges Incidental to the Normal Operation of Vessels

(VGP): Authorization to Discharge Under the National Pollutant Discharge Elimination

System § 2.2.3.5, at 29 (Mar. 28, 2013), available at http://water.epa.gov/polwaste/

npdes/vessels/upload/vgp_permit2013.pdf. The VGP did not set standards for

other ʺsmallʺ organisms, such as bacteria or viruses.

b. WQBELs

The WQBELs in the 2013 VGP require: (1) oceangoing vessels

entering the Great Lakes to continue to perform ballast water exchanges, and (2)

all vessels to control discharges ʺas necessary to meet applicable water quality

standards in the receiving water body or another water body impacted by [the]

discharges.ʺ VGP § 2.2.3.7, at 43, § 2.3.1, at 59.

c. Monitoring and Reporting Requirements for TBELs and WQBELs

As noted above, NPDES permits must contain conditions that

require both monitoring and reporting of monitoring results of TBELs and

WQBELs to ensure compliance with water quality standards. See

(a)(2);

(i)(1)‐(2).

i) Monitoring and Reporting Requirements for TBELs

In the 2013 VGP, EPA established the following monitoring

requirements for TBELs:

‐21‐

(1) that vessels monitor the functionality of their ballast water treatment systems, if installed; and

(2) that vessels monitor the concentrations of the two ʺindicatorʺ bacteria, E. coli and enterococci.10

VGP § 2.2.3.5.1.1.2, at 30, § 2.2.3.5.1.1.4, at 31‐32.

The first requirement is known as functionality monitoring. Under

this requirement, a ballast water treatment program is considered to be in

compliance if it is ʺoperating according to the manufacturersʹ requirements.ʺ

App. at 96.

The second requirement is known as effluent biological organism

monitoring. Under this requirement, vessels must collect small‐volume samples

and analyze them for concentrations of two indicator pathogens. This is

required between one and four times a year depending on the treatment system.

ii) Monitoring and Reporting Requirements for WQBELs

The only monitoring required for WQBELs is that ships report the

ʺexpected date, location, volume, and salinity of any ballast water to be

10 EPA established effluent limits for Vibrio cholerae, but did not require monitoring in this respect because the ʺmonitoring of this parameter would generally not result in the detection of the presence of this pathogen.ʺ App. at 99. The 2013 VGP also contains a third requirement that vessels with treatment systems that add or generate biocides, such as chlorine or ozone, to kill organisms must monitor ballast water discharges for residual biocides. Id. at 103. This requirement does not implicate TBELs.

‐22‐

discharged.ʺ VGP § 4.3, at 72 (emphasis added). Permittees are not required to

report actual locations, volumes, or composition of ballast water to be

discharged.

d. Lakers

The 2013 VGP requires all Lakers to comply with non‐numeric

technology‐based control measures, like ballast water exchange and other BMPs

found in VGP § 2.2.3.3. App. at 85; see VGP § 2.2.3.3, at 27‐28. Lakers are also

subject to three ballast water management measures found in VGP § 2.2.3.4: (1)

conducting an annual assessment of sediment accumulations; (2) minimizing the

amount of water taken in nearshore environments; and (3) adequately

maintaining sea chest screens, which keep larger organisms like fish out of

ballast tanks. VGP § 2.2.3.4, at 28‐29. In addition, all Lakers built on or after

January 1, 2009, must comply with VGP § 2.2.3.5, which sets numeric ballast

water discharge limits. VGP § 2.2.3.5.3.3, at 39.

DISCUSSION

A. Standard of Review

We review a NPDES permit under the Administrative Procedure

Act to determine whether EPAʹs actions were ʺarbitrary, capricious, an abuse of

‐23‐

discretion, or otherwise not in accordance with law.ʺ

(2)(A). To

determine whether the agencyʹs actions were ʺarbitrary and capricious,ʺ we

consider whether the agency

ʹrelied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.ʹ

Islander E. Pipeline Co. v. McCarthy,

, 150‐51 (2d Cir. 2008) (quoting

Motor Vehicle Mfrs. Assʹn. v. State Farm Mut. Ins. Co.,

(1983)). We

must be ʺsatisfied from the record that ʹthe agency . . . examine[d] the relevant

data and articulate[d] a satisfactory explanation for its action.ʹʺ

at 151

(quoting State Farm,

). An agencyʹs action is lawful ʺonly if it rests

ʹon a consideration of the relevant factors.ʹʺ Michigan v. EPA,

(2015) (quoting State Farm,

). We afford the agencyʹs decision

greater deference regarding factual questions involving scientific matters in its

area of technical expertise. See Balt. Gas & Elec. Co. v. NRDC,

(1983); Envtl. Def. v. EPA,

, 204 (2d Cir. 2004).

In addition, judicial review of statutory interpretation by an agency

is governed by Chevron U.S.A. Inc. v. NRDC,

, 842‐45 (1984). Under

‐24‐

Chevron, we must first determine ʺwhether Congress has directly spoken to the

precise question at issueʺ in the CWA.

. If so, we must give effect to the

unambiguously expressed intent of Congress.

at 842‐43. ʺ[I]f the statute is

silent or ambiguous with respect to the specific issue, the question for the court is

whether the agencyʹs answer is based on a permissible construction of the

statute.ʺ

. As the Supreme Court held in Michigan v. EPA, ʺ[e]ven under

this deferential standard, however, ʹagencies must operate within the bounds of

reasonable interpretation.ʹʺ

(quoting Util. Air Regulatory Grp. v.

EPA,

(2014)). We also grant deference to EPAʹs

interpretation of its own regulations ʺunless that interpretation is ʹplainly

erroneous or inconsistent with the regulation.ʺ Chase Bank USA, N.A. v. McCoy,

(2011) (quoting Auer v. Robbins,

(1997)).

B. Petitionersʹ Challenge

Here, petitioners challenge EPAʹs issuance of the 2013 VGP as

arbitrary and capricious, and not in accordance with law, on a number of

grounds.

First, petitioners argue that the TBELs are arbitrary and capricious.

Specifically, petitioners assert that EPA acted arbitrarily and capriciously and not

‐25‐

in accordance with the law when it: (1) selected the IMO standard as the

standard for the TBELs; (2) failed to consider onshore treatment, limiting

consideration to shipboard treatment; (3) failed to include numeric TBELs for

viruses and protists; and (4) exempted Lakers built before 2009 (ʺpre‐2009

Lakersʺ) from the numeric TBELs of the 2013 VGP.

Second, petitioners argue that EPA acted arbitrarily and capriciously

and not in accordance with the law in choosing narrative WQBELs, rather than

numeric WQBELs, because, among other things, the narrative standard is too

imprecise to guarantee compliance with water quality standards.

Finally, petitioners argue that EPAʹs monitoring and reporting

requirements for TBELs and WQBELs are not in accordance with the law because

they were inadequate to guarantee compliance.

1. TBELs

Under the CWA, EPA must apply BAT in establishing pollution

controls for ballast water discharge. BAT requires the ʺapplication of the best

available technology economically achievable for such category or class, which will

result in reasonable further progress toward the national goal of eliminating the

discharge of all pollutants.ʺ

(b)(2)(A)(emphasis added). BAT

also requires ʺa commitment of the maximum resources economically possible to

‐26‐

the ultimate goal of eliminating all polluting discharges.ʺ EPA v. Natʹl Crushed

Stone Assʹn,

(1980).

We hold that in failing to set TBELs that reflected BAT in the 2013

VGP, EPA acted arbitrarily and capriciously in a number of respects.

a. The IMO Standard

Petitioners argue that EPA failed to apply BAT when it chose the

IMO Standard for TBELs in the 2013 VGP. They allege that EPA chose the IMO

Standard first, and then worked ʺbackwardsʺ to determine which systems could

achieve that standard. NWEA Br. at 36. In doing so, they contend, EPA

improperly restricted the SABʹs inquiry to whether certain technologies would

meet what petitioners describe as ʺexisting international consensus standards.ʺ

According to petitioners, EPA should have first considered what ʺavailableʺ

technology was capable of achieving, and then created standards based on that

capability. As a result, petitioners contend, EPAʹs standard did not achieve

greater reductions in pollution discharges that were achievable with current

technology.

We agree. EPA acted arbitrarily and capriciously when it chose the

IMO Standard without adequately explaining why standards higher than the

IMO Standard should not be used given available technology.

‐27‐

In choosing the IMO Standard, EPA overlooked crucial portions of

the SAB Report. The SAB identified a number of technologies that can achieve

standards higher than IMO for one or more organism sizes, including all five of

the technologies identified as also meeting the IMO Standard.11 The SAB

acknowledged that ʺthese same five systems have the potential to meet a 10x

IMO D‐2/ Phase 1 standard [i.e., the IMO Standard] in the near future.ʺ App. at

636. In describing the performance of those systems, the SAB concluded that

each would require only ʺreasonable/feasible modifications.ʺ

at 629‐30,632.

Indeed, according to the SAB Report, the Ecochlor, BalPure, and PeraClean

systems can meet 100 times IMO for medium organisms, and Ecochlor can meet

10 times IMO for large organisms.

EPA should not have adhered to the IMO Standard without

explanation when technologies could have exceeded IMO. Indeed, seeking to

find systems that are capable of doing better than the current standard is in

keeping with the technology‐forcing aspect of the CWA. See NRDC, 822 F.2d at

11 The record further demonstrates that existing shipboard technology can meet a standard between IMO and 10 times IMO. For instance, Hyde Marine Guardian has tested at 1.4 times IMO for large organisms; Optimarin has tested at 7.7 times IMO for large organisms; and Alfa Laval/AlfaWall PureBallast has tested at 4.5 times IMO for large organisms, and at 3.7 times IMO for medium organisms.

‐28‐

124. EPA should have first looked at the available ballast water technologies as

identified by the SAB Report. Then, finding that those technologies could exceed

the IMO Standard, EPA should have adjusted its standard accordingly, or

explained why it would not. See Islander E. Pipeline,

(holding that

agency must ʺexamine[] the relevant data and articulate[] a satisfactory

explanation for its actionʺ (quoting State Farm,

)(alterations

omitted)).

EPAʹs counterargument that no more was necessary because it did

not limit the SAB to considering the IMO Standard is unavailing. EPA insists

that it gave the SAB a list of potential regulatory limits, and then asked the SAB

to identify the systems that could reliably meet those limits. In support, EPA

points to its Charge Question 1 to the SAB, which asked the SAB to identify

ʺdischarge standards that the available data [about existing systems] credibly

demonstrate can be reliably achieved.ʺ App. at 607. EPA argues that in response

to this charge, the SAB Report supports the conclusion that, ʺ[b]ased upon the

data available, no current ballast water treatment technologies were considered

likely to meet standards more stringent than the IMO D‐2/Phase Iʺ standards.

.

‐29‐

While it is true that EPA did not strictly limit the SABʹs

consideration to the IMO Standard, EPA is incorrect in suggesting that the SAB

Report supports the conclusion that no system could meet standards stricter than

the IMO Standard. Id.12 To the contrary, the record contradicts EPAʹs assertion

that treatment systems that exceed the IMO Standard are not ʺavailable.ʺ In fact,

as noted above, systems that exceed the IMO Standard are available

Accordingly, by failing to consider adequately a standard more

stringent than IMO, EPA failed to set permit limits that reflect BAT. See

(b)(2); Natʹl Crushed Stone,

(BAT requires ʺa commitment of

the maximum resources economically possible to the ultimate goal of eliminating

all pollution dischargesʺ); FMC Corp. v. Train,

, 983‐84 (4th Cir. 1976)

The SAB actually stated that it could not reliably test for standards 100 or 12

1000 times more stringent than the IMO Standard: The Panel also concludes that the [IMO Standard] . . . [is] currently measurable, based on data from land‐based and shipboard testing. However, current methods (and associated detection limits) prevent testing of BWMS to any standard more stringent than [the IMO Standard] and make it impracticable for verifying a standard 100 or 1000 times more stringent. Id. at 601. While we agree that we must defer to EPAʹs conclusions regarding the technical feasibility of testing for standards 100 or 1000 times more stringent than the IMO Standard, there is nothing in the record to suggest that it would not be possible to test for twice or even ten times the IMO Standard.

‐30‐

(upholding EPAʹs decision to set BAT based on data from a single pilot plant). In

doing so, EPA acted arbitrarily and capriciously and not in accordance with law

in choosing the IMO standard for the TBELs in the 2013 VGP.

b. Onshore Ballast Water Treatment

Petitioners also argue that EPA arbitrarily and capriciously limited

its consideration to shipboard treatments, failing to consider onshore treatment.

Petitioners argue that onshore facilities used in other industries, such as sewage

treatment plants and drinking water treatment plants, were reasonable

alternatives to shipboard treatment that should have been considered. EPA

concedes it directed the Board to ʺfocus its limited time and resources on the

status of shipboard treatment systems because such systems were either ʹin

existence or in the development process.ʹʺ EPA Br. at 56‐57 (quoting SAB

Report). It argues that onshore treatment was not ʺavailable,ʺ primarily because

no onshore system was yet in existence.

While it is true that no onshore systems existed then ‐‐ unsurprising

considering ballast water treatment was not required at all until the effective date

of the 2008 VGP – the record suggests that such onshore systems were

technologically possible at that time. Yet, EPA chose to curtail discussion about

onshore systems and failed to develop information necessary to evaluate their

‐31‐

availability. We conclude that by failing to consider onshore ballast water

systems, EPA acted arbitrarily and capriciously.

What does ʺavailableʺ mean? As courts have interpreted the term in

the CWA context, technologies that could be used for a particular discharge,

even if they are not currently being used by that industry, are ʺavailable.ʺ As

the Fourth Circuit noted,

The model technology [under consideration] may exist at a plant not within the . . . industry [at issue]. Congress contemplated that EPA might use technology from other industries to establish the Best Available Technology. Progress would be slowed if EPA were invariably limited to treatment schemes already in force at the plants which are the subject of the rulemaking. Congress envisioned the scanning of broader horizons and asked EPA to survey related industries and current research to find technologies which might be used to decrease the discharge of pollutants.

Kennecott, 780 F.2d at 453 (emphasis added) (citation omitted). This Court held

similarly in Hooker Chemicals & Plastics Corp. v. Train,

That no plant in a given industry has adopted a pollution control device which could be installed does not mean that that device is not ʺavailable.ʺ Congress did not intend to permit continuance of pollution by industries which have failed to cope with and attempt to solve the problem of polluted water.

‐32‐

(2d Cir. 1976); see also Cal. & Hawaiian Sugar Co. v. EPA,

(2d Cir. 1977) (technology used ʺʹin other industries with similar raw

waste characteristicsʹʺ was ʺavailableʺ (quoting Liquid and Crystalline Cane Sugar

Refining Subcategory,

, 10,522 (1974))); Am. Petrol. Inst. v. EPA,

, 264‐65 (5th Cir. 1988) (holding that a process can be ʺdeemed

ʹavailableʹ even if it is not in use at allʺ because ʺ[s]uch an outcome is consistent

with Congressʹ intent to ʹpush pollution control technologyʹʺ (quoting Assʹn of

Pac. Fisheries v. EPA,

, 816 (9th Cir. 1980); Weyerhaeuser Co. v. Costle,

(D.C. Cir. 1978))).

For a technology in one industry to be ʺavailableʺ in a second

industry: (1) the transfer technology must be available within the first industry;

(2) the transfer technology must be transferable to the second industry; and (3) it

must be reasonably predictable that the technology, if used in the second

industry, will be capable of removing the increment required by the effluent

standards. See Kennecott, 780 F.2d at 453 (citing Tannersʹ Council,

); CPC Intʹl Inc. v. Train,

(8th Cir. 1975); Hooker Chems.,

(ʺBut even if technology which is not presently in use can be

treated as available and achievable, there must be some indication in the

‐33‐

administrative record of the reasons for concluding that such technology is

feasible and may reasonably be expected to yield the effluent reduction

mandated when applied to the particular industry.ʺ). For example, in Kennecott,

the Fourth Circuit upheld EPAʹs use of manufacturing technology from one

industry as part of a BAT determination for treating wastewater in a different

industry. 780 F.2d at 453‐54.

Here, we cannot evaluate whether onshore technology should be

considered ʺavailableʺ because the record does not contain a full discussion of

onshore treatment. This lack of information about onshore facilities, however, is

a problem of EPAʹs own making because EPA went to great lengths to foreclose

discussion of onshore treatment both by expressly limiting the SABʹs mandate to

studying shipboard treatment technology and consistently opposing any attempt

by the SAB to consider onshore treatment.

EPAʹs effort to curtail discussion of onshore treatment is well

documented in the record. In a letter dated February 10, 2012, thirteen scientists,

eight of whom were members of the SAB and six of whom were members of the

NAS Committee,13 including the Chair of the NAS Committee, stated that the

13 One person was a member of both the SAB and the NAS Committee.

‐34‐

SAB ʺnever actually addressed the question of what is the best treatment that

available technology can achieveʺ because EPA limited them to the narrower

question of ʺwhether shipboard treatment systems could meet certain specific sets of

standards.ʺ App. at 740. Furthermore, the scientists assert that their attempts to

consider onshore treatments were actively thwarted by EPA:

During the SAB Panel meetings and discussions, some members of the Panel attempted to develop and include in the Panel report a more detailed assessment of onshore treatment, including its cost impacts, and an assessment of the full capability of shipboard treatment . . . . [T]hese assessments would have further demonstrated that available technology can achieve levels of treatment beyond what the EPA has proposed as controls. The EPA Office of Water, however, consistently opposed including such information in the report. As a result, some relevant information and analysis that could have been developed by the Panel was not, and some of what was developed by Panel members was excluded or deleted from the final report. If there was less information developed on these issues and less provided in the report than the EPA considers sufficient, it is in large part because the EPA Office of Water opposed the development and inclusion of such information.

(emphases added).

In light of these facts, we cannot well credit EPAʹs assertion that it

lacked information to support a finding that onshore facilities were ʺavailable.ʺ

While EPA states that it was ʺunaware of any onshore treatment facility currently

‐35‐

available in the U.S. that is capable of meeting the VGPʹs § 2.2.3.5 ballast water

discharge standards,ʺ and that it did not ʺreceive information indicating they are

or would become available over the term of the VGP,ʺ id. at 544, in fact EPA

turned a blind eye to significant information about onshore treatment.

Indeed, the lack of information about the ʺavailabilityʺ of onshore

treatment is due in large part to EPAʹs arbitrary and capricious decision to

oppose developing such information. As a result, the TBELs were based on an

incomplete record ‐‐ one lacking meaningful discussion of an ʺavailableʺ

treatment, namely onshore treatment. See Humana of Aurora, Inc. v. Heckler,

(10th Cir. 1985) (agency action is arbitrary and capricious when

based on a flawed study); Tex. Oil & Gas Assʹn v. EPA,

, 935 (5th Cir.

1998) (ʺA regulation cannot stand if it is based on a flawed, inaccurate, or

misapplied study.ʺ); Almay, Inc. v. Califano,

(D.C. Cir. 1977)

(rejecting regulation produced ʺon the basis of the flawed surveyʺ). Put another

way, EPAʹs refusal to consider onshore treatment ʺentirely fail[s] to consider an

important aspect of the problemʺ and ʺoffer[s] an explanation for its decision that

runs counter to the evidence before the agency.ʺ Islander E. Pipeline, 525 F.3d at

150‐51; see Tannersʹ Council,

(ʺ[T]he agency must fully explicate

‐36‐

its course of inquiry, its analysis, and its reasoning.ʺ); see also State Farm,

; Hooker Chems.,

. Hence, it is arbitrary and capricious.

In fact, the SAB Report points out a number of reasonably

predictable advantages to onshore treatment. The SAB Report states:

Use of reception facilities for the treatment of ballast water appears to be technically feasible (given generations of successful water treatment and sewage treatment technologies), and is likely to be more reliable and more readily adaptable than shipboard treatment.

App. at 605; see also

. The SAB Report also notes that onshore treatment

has a number of advantages over shipboard treatment because onshore facilities

are not subject to problems such as limited space, small and overburdened crews,

vibrations, weight allowances, limited power, ship instability, and greater

corrosion rates.

at 678‐80. Regarding ship crews in particular, studies have

shown that ʺmany of these crews are already overburdened,ʺ ʺ[o]peration by

trained, dedicated personnel in reception facilities would likely result in more

reliable performance,ʺ and ʺ[m]aintenance and repair work are more likely to be

done reliablyʺ as well.

. Onshore treatment can also be more effective

by using superior technologies that are not available for shipboard treatment,

such as settling tanks, granular filtration, and membrane filtration.

at 680‐81.

Indeed, EPA cites a number of studies that conclude that onshore treatment

‐37‐

facilities are a technically feasible option. Id. at 107. These studies date back to

1992, and proceed with some regularity thereafter‐‐ 1996, 1999, 2000, 2002, 2007,

2008.

Moreover, onshore treatment would not necessarily be slower than

shipboard treatment to implement. The SAB estimated that onshore

implementation would take up to thirty months, while EPA allowed eight years

to phase‐in shipboard implementation. See id. at 684. Nor would onshore

treatment necessarily be more expensive than shipboard treatment. Regional

economic studies suggest that ʺtreating ballast water in reception facilities would

be at least as economically feasible as shipboard treatment.ʺ Id. at 694. In

addition, the cost of monitoring and enforcement is likely to be lower with a

smaller number of reception facilities compared with a larger number of

shipboard systems. Id. at 605, 694. The Coast Guard also found that onshore

treatment was generally less expensive per metric ton of ballast water than

shipboard treatment. Id. at 679.

Of course, onshore treatment has many costs, including the cost of

retrofitting vessels for onshore facilities, particularly ships from outside the

United States, and the cost of shipping delays created by the time it takes to

‐38‐

discharge ballast onshore (though presumably shipboard treatment is not

instantaneous). Costs alone, however, cannot determine BAT. See

(b)(2)(B). Furthermore, EPA failed to perform the economic analysis

required to determine relative costs of the differing technologies in reaching its

conclusion that onshore treatment was not economically achievable. See

Waterkeeper All., 399 F.3d at 516 (ʺ[T]he Administrator is obligated to ʹinquire into

the initial and annual costs of applying the technology and make an affirmative

determination that those costs can be reasonably borne by the industry.ʹʺ

(quoting Riverkeeper, Inc. v. EPA,

, 195 (2d Cir. 2004))); Natʹl Wildlife

Fedʹn v. EPA,

(D.C. Cir. 2002) (ʺAlthough its analysis may be

general, EPA ʹhas the heaviest of obligations to explain and expose every step of

its [cost‐benefit] reasoning.ʹ . . . This duty to explain arises out of the need for

reviewing courts to be able to discern the basis for EPAʹs decision.ʺ (internal

citations omitted) (quoting Am. Lung Assʹn v. EPA,

(D.C. Cir.

1998))).

In light of these observations, the SAB and NAS Committee

scientists concluded that ʺEPA should conduct a comprehensive analysis

comparing biological effectiveness, cost, logistics, operations, and safety

‐39‐

associated with both shipboard [treatment] and reception facilities.ʺ App. at 606.

If that analysis ʺindicate[d] that treatment at reception facilities is both

economically and logistically feasible and is more effective than shipboard

treatment systems, it should be used as the basis for assessing the ability of

available technologies to . . . meet a given discharge standard.ʺ

EPA chose

not to do so because the SAB ʺdid not specify a timetable for that complex

endeavor or suggest that is was possible to complete such an analysis in time to

inform the impending VGP.ʺ EPA Br. at 58‐59. We do not find that answer

compelling. There is no impediment to engaging in further study, and further

study may advance the goals of the CWA.

Thus, EPA could have well found that onshore treatment was

ʺavailable.ʺ Indeed, EPAʹs failure to consider onshore treatment is inconsistent

with the CWAʹs mandate that TBELs be technology‐forcing. Congress designed

the CWA to force agencies and permittees to adopt technologies that achieve the

greatest reductions in pollutants. See NRDC, 822 F.2d at 124 (holding that CWA

seeks ʺnot only to stimulate but to press development of new, more efficient and

effective technologies,ʺ which is the ʺessential purpose of this series of

progressively more demanding technology‐based standardsʺ). As Judge Starr

‐40‐

noted in NRDC, ʺthe most salient characteristic of this statutory scheme,

articulated time and again by its architects and embedded in the statutory

language, is that it is technology‐forcing.ʺ Id. at 123.

EPAʹs decision on this issue matters. As the SAB scientists pointed

out, EPAʹs choice of system in this permit will have a long‐term impact:

[S]hipboard treatment and onshore treatment represent distinct approaches to ballast water management that would each require different large investments in infrastructure. . . . Thus we are almost certain to be stuck for a very long time with whichever approach is used as the BAT in setting discharge standards in 2013. It is thus of the utmost urgency that a fair and thorough comparison of the two approaches be made at this time.

App. at 744‐45 (emphasis added). We conclude that EPA failed to give fair and

thorough consideration to both onshore and shipboard treatment systems in

setting the standard in the 2013 VGP, and we remand to EPA to give full

consideration to the issue now.

c. Viruses and Protists

Petitioners also complain about the lack of numeric TBELs for

viruses and protists (primarily single‐celled organisms). EPA argues, however,

that it could not set TBELs for viruses and protists in the 2013 VGP because EPA

could not yet identify ʺsuitable standardized test organisms and/or surrogate

‐41‐

parameters to determine treatment system performance at removing or

eliminating viruses and protists and which also can be used in establishing

technology‐based discharge limitations.ʺ App. at 486; see also App. at 495 (ʺEPA

does not believe that there are sufficient data available to establish numeric limits

for protists or other bacteria.ʺ).

We agree that it was not arbitrary and capricious for EPA to decline

to set TBELs for organisms for which it is unable to test and for which it has

insufficient data to set numeric limits. See Balt. Gas & Elec. Co. v. NRDC, Inc.,

(1983) (ʺ[A] reviewing court must remember that [where the agency]

is making predictions, within its area of special expertise, at the frontiers of

science . . . as opposed to simple findings of fact, a reviewing court must

generally be at its most deferential.ʺ). Petitioners have not demonstrated that

sufficient data are available. EPA has represented that it will consider including

numeric TBELs for viruses and protists in the next version of the VGP. App. at

486. This is sufficient.

d. Pre‐2009 Lakers

Petitioners allege that EPAʹs decision to exempt Lakers built before

January 1, 2009 from numeric effluent limits of VGP § 2.2.3.5 was arbitrary and

capricious. EPA based this decision on its finding that there was no treatment

‐42‐

technology ʺavailableʺ for these vessels either onboard or onshore. App. at 115‐

16. EPA expressed concern about the difficulty of finding effective onboard

systems for pre‐2009 Lakers due to their ʺunique operational and design

constraints,ʺ such as the large volumes of fresh cold water they require, the short

duration of their trips, their high pumping rates, and their uncoated ballast

tanks. Id. at 116.14 In reaching that conclusion, EPA relied on the SAB Report,

which advised that ʺspecific constraints can greatly limit treatment optionsʺ for

Lakers. Id. at 638. EPA also cited the costs of implementing these systems. Id. at

116.

We agree with petitioners that exempting the pre‐2009 Lakers was

arbitrary and capricious. First, the lack of supply of updated shipboard systems

is not a legitimate reason to exempt pre‐2009 Lakers from the 2013 VGP, as,

again, the purpose of BAT is to force technology to keep pace with need. See

NRDC, 822 F.2d at 124.

Second, EPAʹs decision was based on a flawed record that failed to

consider an important aspect of the problem, namely the possibility of onshore

14For example, certain treatment methods, such as electro‐chlorination and ozonation, may only be effective in salt water, and others that use oxidizing chemicals may increase corrosion rates in uncoated tanks. Id. at 638.

‐43‐

treatment. See Islander E. Pipeline, 525 F.3d at 150‐51. EPA should have

considered the comparable ʹcost of achieving such effluent reductionsʹ through

onshore treatment versus shipboard treatment, rather than merely dismissing

onshore treatment. EPA disregarded the SABʹs recommendation that onshore

treatment would benefit pre‐2009 Lakers that are ʺengaged solely in regional

trade.ʺ Id. at 684. The SAB points out that the space and power constraints

posed by pre‐2009 Lakers are ʺlargely absent in reception facilities.ʺ Id. at 680.

EPAʹs foreclosure of considering onshore treatments for pre‐2009 Lakers ‐‐ and

indeed, all Lakers ‐‐ seems shortsighted. See supra at 31‐42.

Third, EPA imposed the 2013 VGP on Lakers built after 2009, even

though post‐2009 Lakers face many of the same challenges and constraints as

pre‐2009 Lakers, such as their short voyages, high pumping rates, and freshwater

environment. 15 While it is true that shipbuilders were on notice that post‐2009

15 Intervenors argue that due to these constraints, ballast water treatment is infeasible for all Lakers, regardless of when they were built. EPA has concluded, however, that anyone building a ship designed to enter the market after 2009 was well aware of the impending VGP requirements, and could anticipate its impact on shipbuilding. App. at 117. Intervenors also contend that ships exclusively plying the Great Lakes do not pose a threat to water quality because they do not introduce any invasive species from outside the Great Lakes. EPA has properly rejected this argument, noting that Lakers can spread or more rapidly distribute invasive species already present in the Great Lakes. Id. at 501.

‐44‐

Lakers would be subject to the 2013 VGP, in reality post‐2009 boats appear to be

similarly situated to pre‐2009 Lakers. See Islander E. Pipeline, 525 F.3d at 150‐51

(agency decision is arbitrary and capricious when agency offers explanation for

decision that runs counter to evidence before agency). Thus, distinguishing pre‐

2009 and post‐2009 Lakers was arbitrary and capricious.

The SAB Report supports our conclusion. Although the SAB Report

acknowledged the limitations in treating ballast water from Lakers, it did not

declare such treatment impossible. Instead, the SAB concluded that in light of

these limitations, ʺ[a] variety of environmental (e.g., temperature and salinity),

operational (e.g., ballasting flow rates and holding times), and vessel design (e.g.,

ballast volume and unmanned barges) parametersʺ should be considered in

determining the treatment standards. App. at 639.

EPAʹs exemption of the pre‐2009 Lakers from the 2013 VGP was also

arbitrary and capricious due to EPAʹs failure to conduct an appropriate and

factually‐supported cost analysis. Such an analysis might have shown that the

cost of subjecting pre‐2009 Lakers to the 2013 VGP was not unreasonably high,

or, alternatively, that onshore treatment was economically feasible. For all these

‐45‐

reasons, EPAʹs decision to exempt pre‐2009 Lakers from the 2013 VGP was

arbitrary and capricious.

2. WQBELs

Under the CWA, NPDES permits must include WQBELs where the

TBELs are insufficient to maintain water quality standards.

(b)(1)(C);

(d)(1)(vii)(A). Here, EPA concluded that ʺeven

at the IMO level of discharge, reasonable potential exists for such discharges to

cause or contribute to violations of applicable water quality standards pursuant

to

(d)(1)(ii).ʺ App. at 129. To address this concern, EPA

established WQBELs to ensure compliance with water quality standards. EPA,

however, chose to create narrative WQBELs because it believed numeric

WQBELs were ʺinfeasibleʺ to calculate.16

16 Federal regulation permits such limits to be expressed narratively if the calculation of numeric limits is ʺinfeasible.ʺ

(k)(3).

‐46‐

The WQBEL in the 2013 VGP states:

Your discharge must be controlled as necessary to meet applicable water quality standards in the receiving water body or another water body impacted by your discharges.

VGP § 2.3.1, at 59. In defending this standard, EPA relied in part on the NAS

Report, which stated that ʺ[t]he current state of science does not allow a

quantitative evaluation of the relative merits of various discharge standards in

terms of invasion probability.ʺ App. at 363.

Petitioners argue that this narrative WQBEL does not ensure

compliance with water quality standards. We agree. This narrative standard is

insufficient to give a shipowner guidance as to what is expected or to allow any

permitting authority to determine whether a shipowner is violating water quality

standards. By requiring shipowners to control discharges ʺas necessary to meet

applicable water quality standardsʺ without giving specific guidance on the

discharge limits, EPA fails to fulfill its duty to ʺregulat[e] in fact, not only in

principle.ʺ Waterkeeper All., 399 F.3d at 498. As this Circuit held in Waterkeeper

Alliance, NPDES permits ʺmay issue only where such permits ensure that every

discharge of pollutants will comply with all applicable effluent limitations and

standards.ʺ Id. That is hardly the case here. EPA itself notes that it only

‐47‐

ʺgenerally expects that compliance with the [TBELs] . . . will control discharges as

necessary to meet applicable water quality standards.ʺ VGP § 2.3.1, at 59. The

WQBELs, although found by EPA to be required to supplement the TBELs, in fact

add nothing. The WQBELs do not state how they will ensure compliance.

Even if determining the proper standard is difficult, EPA cannot

simply give up and refuse to issue more specific guidelines. See Am. Paper Inst.,

Inc. v. EPA,

, 350 (D.C. Cir. 1993) (articulating that, even if creating

permit limits is difficult, permit writers cannot just ʺthr[o]w up their hands and,

contrary to the Act, simply ignore[] water quality standards including narrative

criteria altogether when deciding upon permit limitationsʺ). Scientific

uncertainty does not allow EPA to avoid responsibility for regulating discharges.

See Massachusetts v. EPA,

(2007) (ʺEPA [cannot] avoid its

statutory obligation by noting the uncertainty surrounding various features of

climate change and concluding that it would therefore be better not to regulate at

this time.ʺ).

Moreover, EPAʹs reliance on the NAS Report is misplaced. EPA

concedes that the NAS Committee ʺdid not conclude that it is infeasible to

calculate water quality‐based effluent limits for ballast water discharges.ʺ App.

‐48‐

at 563. Rather, the NAS Committee found that it could not formulate a precise

standard. In light of this uncertainty, it recommended further study of the issue.

But EPA declined to engage in further study. See App. at 363‐67. For all these

reasons, EPAʹs WQBELs were arbitrary and capricious.

EPAʹs remaining counterarguments are unavailing. First, EPA

asserts that petitioners fail to offer examples of ʺmeaningful permit limitsʺ for

WQBELs. EPA Br. at 74. EPA, however, could articulate specific actions that

vessels would be required to take to protect against site‐specific threats. For

example, if EPA or the shipowner became aware of an unusual risk posed by a

specific port, EPA could require vessels to take action to avoid such risk at that

port, including not uptaking ballast water or not discharging into other ports the

contaminated ballast water that was taken up. Toward that end, EPA has

included a set of specific best management practices in the 2013 VGP § 2.2.3.3,

further demonstrating the viability of this approach.

Second, EPA argues that under

(k)(3) it may

employ BMPs instead of ʺ[n]umeric effluent limitationsʺ for WQBELs when

deriving numeric limitations is ʺinfeasible.ʺ

(k)(3). BMPs

include ʺschedules of activities, prohibitions of practices, maintenance

‐49‐

procedures, and other management practices to prevent or reduce . . . pollution.ʺ

§ 122.2. EPA argues that the narrative WQBEL is a BMP, and therefore it has

discharged its duty under

But EPAʹs narrative WQBEL does not qualify as a BMP, as it is

neither a practice nor a procedure. BMPs typically involve requirements like

operating procedures, treatment requirements, practices to control runoff,

spillage or leaks, sludge or waste disposal, or drainage from raw material

storage; they can also be structural requirements including tarpaulins, retention

ponds, or devices such as berms to channel water away from pollutant sources,

and treatment facilities. See NRDC v. Sw. Marine, Inc.,

, 991 n.1 (9th

Cir. 2000). Examples of BMPs that have been accepted as substitutes for effluent

limits include: nutrient management plans for concentrated animal feeding

operations, see Waterkeeper All., 399 F.3d at 497, 502, filtration of stormwater

runoff from ditches before it enters rivers and streams (by timber companies),

and constructing roads with surfacing that minimizes sediment in runoff (by

17 EPA also claims general prohibitions can be BMPs, citing

, but does not offer an example of something as general as the WQBEL standard being found to be a BMP. As discussed herein, the EPAʹs characterization is inconsistent with regulations that require WQBELs to ʺensure compliance.ʺ

(d); see Auer,

(holding interpretation may not be inconsistent with regulation).

‐50‐

timber companies), see Decker v. Nw. Envtl. Def. Ctr.,

(2013).

The narrative standard here is nowhere as specific as any of these examples.

Indeed, it requires nothing more of a shipowner than to meet the TBELs. This

interpretation is hardly consistent with the regulations that require WQBELS to

ensure compliance. See Auer,

(holding that courts should defer to

agencyʹs interpretation of its own regulations if not plainly erroneous or

inconsistent with the regulation).

Third, EPA claims that WQBEL standards will be sufficiently

maintained because EPA can take ʺcorrective actionsʺ after the permittee becomes

aware of a violation. App. at 160. This is not reassuring. The point of a permit is

to prevent discharges that violate water quality standards before they happen. See

(b)(1)(C), 1342(a)(2);

(d), 122.44(d)(1).

ʺCorrective actionʺ is not an effective remedy in an invasive species context ‐‐ it is

difficult to eradicate a colony of zebra mussels after they are established. See,

e.g., Great Lakes Sci. Ctr., U.S. Geological Survey, Zebra Mussels Cause Economic

and Ecological Problems in the Great Lakes 2 (rev. 2011) (ʺOnce zebra mussels

become established in a water body, they are impossible to eradicate with the

technology available today. Many chemicals kill zebra mussels, but these exotics

‐51‐

are so tolerant and tough that everything in the water would have to be poisoned

to destroy the mussel.ʺ); Robbins, A Western Showdown, at D6 (noting that

officials in Western states have instituted elaborate and expensive inspection

systems for boats because they ʺwant desperately to keep the mussels out of

blue‐ribbon trout streams and pristine mountain lakesʺ as ʺonce established [the

mussels] are impossible to permanently eradicateʺ). This is all the more

problematic because a vessel operator is not likely to know it has a discharge

violation if, as discussed below, there are no monitoring requirements.18

Accordingly, EPA acted arbitrarily and capriciously in issuing the

WQBELs because they violate section 1342ʹs requirement that NPDES permits

ensure compliance with the CWA. Cf. Waterkeeper All., 399 F.3d at 498.

Intervenors raise one additional argument. Under section 401 of the

CWA, before EPA issues a permit, the state in which the discharge is to occur

18 EPAʹs response is that petitionersʹ arguments regarding corrective action and BMPs are waived because they were not raised by petitioners in the comments to the permit. See, e.g., NRDC v. EPA,

, 1073‐74 (D.C. Cir. 1994). Arguments can be considered, however, even if not raised during the notice and comment period. See NRDC v. EPA,

(D.C. Cir. 2014) (ʺEPA retains a duty to examine key assumptions as part of its affirmative burden of promulgating and explaining a nonarbitrary, non‐capricious rule and therefore EPA must justify that assumption even if no one objects to it during the comment period.ʺ (internal quotation marks and alteration omitted) (quoting Appalachian Power Co. v. EPA,

(D.C. Cir. 1998)).

‐52‐

must either certify, or waive its right to certify, that the discharge will comply

with the stateʹs water quality standards ‐‐ commonly known as a ʺ401

Certification.ʺ NRDC, 279 F.3d at 1183; see also

(a);

(b). The intervenors argue that because ʺthe 401 Certifications have been

upheld, the matter is settled: the VGP will ensure compliance with the state

water quality standards.ʺ Intervenorsʹ Br. at 46. We disagree. EPA has an

independent duty under the CWA to ensure compliance with state and federal

water quality standards and may impose ʺadditional permit conditions necessary

to meet that end.ʺ NRDC, 279 F.3d at 1188. Such additional permit conditions

may be necessary if state water quality standards are potentially less stringent

than the CWAʹs,ʺ because ʺthe CWA provides a federal floor, not a ceiling, on

environmental protection.ʺ Dubois v. U.S. Depʹt of Agriculture,

(1st Cir. 1996) (citations omitted).

3. Monitoring and Reporting Requirements for TBELs and WQBELs

Under the CWA, NPDES permits must contain conditions that

require both monitoring and reporting of monitoring results of TBELs and WQBELs

to ensure compliance. See

(a)(2);

(i)(1)‐(2). The

regulations provide, in pertinent part:

‐53‐

In addition to the conditions established under § 122.43(a), each NPDES permit shall include conditions meeting the following requirements when applicable.

. . .

(i) Monitoring requirements. In addition to § 122.48, the following monitoring requirements:

(1) To assure compliance with permit limitations, requirements to monitor:

(i) The mass (or other measurement specified in the permit) for each pollutant limited in the permit;

(ii) The volume of effluent discharged from each outfall;

(iii) Other measurements as appropriate including pollutants in internal waste streams under § 122.45(i); pollutants in intake water for net limitations under § 122.45(f); frequency, rate of discharge, etc., for noncontinuous discharges under § 122.45(e); pollutants subject to notification requirements under § 122.42(a); and pollutants in sewage sludge or other monitoring as specified in 40 CFR part 503; or as determined to be necessary on a case‐by‐case basis pursuant to section 405(d)(4) of the CWA.

(iv) According to sufficiently sensitive test procedures (i.e., methods) approved under 40 CFR part 136 for

‐54‐

the analysis of pollutants or pollutant parameters or required under 40 CFR chapter 1, subchapter N or O.

. . .

(2) Except as provided in paragraphs (i)(4) and (i)(5) of this section, requirements to report monitoring results shall be established on a case‐by‐case basis with a frequency dependent on the nature and effect of the discharge, but in no case less than once a year. . . .

(i)(1)‐(2).

Enforcing compliance with a permit is the key to an effective NPDES

program. See NRDC v. Cty of L.A.,

(9th Cir. 2013) (ʺ[T]he

NPDES program fundamentally relies on self‐monitoring,ʺ and ʺCongressʹ

purpose in adopting this self‐monitoring mechanism was to promote

straightforward enforcement of the Act.ʺ (internal quotations omitted)). We now

turn to the monitoring and reporting requirements in the 2013 VGP.

a. Monitoring Requirements for TBELs

As previously discussed, the 2013 VGP requires vessels to monitor

(1) the functionality of their ballast water treatment systems, if installed, and (2)

the concentrations of the two ʺindicatorʺ bacteria (E. coli and enterococci). . VGP

§ 2.2.3.5.1.1.2, at 30, § 2.2.3.5.1.1.4, at 31‐32.

‐55‐

The first requirement, known as ʺfunctionality monitoring,ʺ

determines whether a ballast water treatment program is ʺoperating according

to the manufacturersʹ requirements.ʺ App. at 96. A shipowner is required to

check a measurement that would ʺverify system functionality,ʺ such as how

much chlorine the system is using each month. Id. at 1168. If the measurement

is correct, it is assumed that the program is in compliance. If the equipment is

not operating properly, the ship is not permitted to discharge ballast. The vessel

owner is not required to take any measurement of pollutants or significant

categories of living organisms in ballast water being discharged. Instead, the

shipowner relies solely on the functioning of the treatment system to determine

if the ship is complying with the permit. Treatment systems are inspected

monthly. See VGP § 2.2.3.5.1.1.2, at 30 (ʺTo assess the [systemʹs] functionality,

monitoring indicators of the [systemʹs] functionality is required at least once per

month for specific parameters that are applicable to your system.ʺ).

The second provision, effluent biological organism monitoring,

requires vessels to collect small‐volume samples from ballast discharge and

analyze them for concentrations of two indicator pathogens, E. coli and

enterococci. VGP § 2.2.3.5.1.1.4, at 31‐32. The idea is that if there are significant

‐56‐

levels of these two pathogens in the sample, then treatment has not been

effective. Vessels are not required to monitor Vibrio cholera or medium or large

organisms regulated in the 2013 VGP. Sampling is required between one and

four times a year, depending on the type of system.

Petitioners present two arguments about why these requirements do

not accord with the law. First, petitioners argue that the 2013 VGP violates CWA

regulations by not requiring vessels to monitor the concentration of living

organisms. The regulations require monitoring of mass, volume, or ʺother

measurement specified in the permit.ʺ

(i)(1)(i)‐(ii). Petitioners

contend the unit of measurement for living organisms in the 2013 VGP should be

concentration. Neither functionality monitoring nor testing for two indicator

microorganisms measures concentration. Thus, according to petitioners, the

monitoring and reporting requirements for TBELs violate

(i)(1).

Second, petitioners argue that these monitoring requirements violate

the requirement in

(i)(1) that monitoring ʺassure[s] compliance

with permit limitations.ʺ Relying on functionality monitoring instead of

requiring an actual measurement of concentrations of organisms means that

neither EPA nor the public knows if the permittees are complying with the

‐57‐

TBELs. Similarly, petitioners argue that monitoring for the presence of the two

ʺindicator bacteria,ʺ E. coli and enterococci, is not sufficient to monitor

compliance with TBELs because it indicates only their presence in the water, not

their quantity.

We disagree with petitioners and conclude that EPAʹs monitoring

requirements were not arbitrary and capricious. The CWA regulations expressly

allow for monitoring quantities other than mass or volume, namely some ʺother

measurement specified in the permit[] for each pollutant limited in the permit.ʺ

(i)(1). Functionality monitoring and monitoring for the

presence of indicator organisms qualify as such ʺother measurement.ʺ

And while there are potential alternatives, petitionersʹ urged

alternative of direct monitoring is not required because ʺmore sophisticated

methods for enumerating living organisms in the larger size classes are not

currently available for use by permittees.ʺ App. at 524. Current technology is

not capable of adequately monitoring ballast water as it is being discharged

because, EPA notes, such monitoring requires analyzing large volumes of water

and is prohibitively expensive and impractical. According to EPA, testing

medium and large organisms with shipboard systems can cost between $75,000

‐58‐

and $125,000 per vessel per sampling event. Moreover, the process would be

impractical, involving sampling and analyzing large volumes of water in labs

and requiring ʺdozens of hours to collect and analyze those samples.ʺ EPA Br. at

89.

Given the difficult circumstances, EPAʹs monitoring requirements

for TBELs were not arbitrary and capricious. Cf. NRDC v. Costle,

,

1380 (D.C. Cir. 1977) (ʺ[W]hen numerical effluent limitations are infeasible, EPA

may issue permits with conditions designed to reduce the level of effluent

discharges to acceptable levels.ʺ). In the face of the severe technological

limitations on monitoring, it was reasonable for EPA to decline to require

monitoring for parameters for which it is currently impractical to collect and

analyze samples. Functionality monitoring and biological indirect monitoring

are the only feasible options at present to assure compliance with the permit.

We defer to EPAʹs decision that functionality monitoring and biological

indicator monitoring, when used in combination, offer an acceptable ʺother

measurement.ʺ See Auer,

(holding that courts should defer to

agency’s interpretation of its own regulations if not plainly erroneous or

inconsistent with the regulation).

‐59‐

b. Monitoring Requirements for WQBELs

Petitioners also argue that EPA acted arbitrarily and capriciously in

failing to require that permittees monitor ballast water discharges to ensure

compliance with WQBELs. The only monitoring requirement for WQBELs is

that ships report the ʺexpected date, location, volume, and salinity of any ballast

water to be dischargedʺ into U.S. waters or at a reception facility. VGP § 4.3, at

72 (emphasis added). There is no requirement to report actual volumes,

locations, or composition of ballast water discharges.

We agree that failure to include monitoring requirements for

WQBELs was arbitrary and capricious. The regulations require monitoring to

ʺassure compliance with permit limitations.ʺ

. Generally, ʺan

NPDES permit is unlawful if a permittee is not required to effectively monitor its

permit compliance.ʺ NRDC,

. Here, the reporting requirement

provided little information on the quality of the ballast water, requiring only

information on expected date, location, volume, and salinity. There is no way to

derive from that information whether a vessel is actually in compliance with the

WQBELs. Thus, because the 2013 VGP does not contain a mechanism to evaluate

compliance with the WQBELs, the monitoring requirements are arbitrary and

‐60‐

capricious and not in accordance with the law. See Waterkeeper All., 399 F.3d at

499 (failure of permit to include any mechanism for evaluating compliance with

required technical standards rendered agency unable to ensure compliance with

water quality standards).

Our conclusion is further supported by the simple, but overlooked,

options that EPA has in structuring WQBEL monitoring requirements. One

possible condition EPA could consider including in the WQBELs would be to

require shipowners to monitor the actual time, place, and volume of ballast water

discharge, rather than the expected time, place, and volume. Another possible

condition would be to require shipowners to monitor for a particular pathogen

or pollutant if it became known that such a pathogen or pollutant is a problem in

a particular port. Each of these options would provide more significant

monitoring.

EPAʹs contentions on this point are unpersuasive. EPA argues that

if a vessel is in compliance with the TBELs, it should be ʺgenerally expected to

already be controlling [its] vessel discharges to a degree that is protective of

water quality,ʺ rendering additional monitoring to demonstrate compliance with

narrative WQBELs unnecessary. App. at 530. In defense of this position, EPA

‐61‐

also argues that

(i) does not apply because of ʺpractical

constraints on the ability to collect and analyze the volumes of ballast water

necessary to ʹdirectlyʹ detect and quantify such organisms at the levels of

concern.ʺ EPA Br. at 98. According to EPA, it is simply ʺunrealisticʺ to have

stricter monitoring.

This, however, is not a valid excuse in the WQBEL context. See

NRDC, 859 F.2d at 208 (stating legislative history of CWA ʺstrongly supports

[the] position that Congress did not intend to tie compliance with water quality‐

based limitations to the capabilities of any given level of technology,ʺ and ʺa

water quality‐based permit limit begins with the premise that a certain level of

water quality will be maintained, come what may, and places upon the permittee

the responsibility for realizing that goalʺ). It is inconsistent to say that WQBELs

are necessary to ensure that water quality standards are met, while specific

enforcement of such WQBELs is unnecessary. More importantly, this lack of

enforcement violates the CWA regulations, which mandate that no permit may

be issued ʺ[w]hen the imposition of conditions cannot ensure compliance with

the applicable water quality requirements of all affected States.ʺ

(d).

‐62‐

Accordingly, EPAʹs failure to include monitoring for compliance

with WQBELs was inconsistent with regulations and thus arbitrary and

capricious.

C. Remand

Accordingly, we remand this matter to EPA for proceedings

consistent with this opinion. The 2013 VGP is to remain in place until EPA issues

a new VGP. See Idaho Farm Bureau Fedʹn v. Babbitt,

, 1405 (9th Cir.

1995) (holding that ʺ[o]rdinarily when a regulation is not promulgated in

compliance with the APA, the regulation is invalid. However, when equity

demands, the regulation can be left in place while the agency follows the

necessary procedures.ʺ (citation omitted)); see also Allied‐Signal, Inc. v. U.S.

Nuclear Regulatory Comm.,

, 150 (D.C. Cir. 1993) (ʺAn inadequately

supported rule, however, need not necessarily be vacated.ʺ); Fertilizer Inst. v.

EPA,

, 1312 (D.C. Cir. 1991) (ʺ[W]hen equity demands, an

unlawfully promulgated regulation can be left in place while the agency

provides the proper procedural remedy.ʺ); W. Oil & Gas Assʹn v. EPA,

, 813 (9th Cir. 1980) (ʺ[A] reviewing court has discretion to shape an equitable

remedy, [and so] we leave the challenged designations in effect.ʺ).

‐63‐

CONCLUSION

For the reasons set forth above, we GRANT the petition for review

with respect to

(1) EPAʹs decision to set the TBELs at the IMO Standard,

(2) EPAʹs failure to consider onshore treatment for ballast water discharge,

(3) EPAʹs decision to exempt pre‐2009 Lakers from the TBELS in the 2013 VGP permit,

(4) EPAʹs narrative standard for WQBELs, and

(5) the monitoring and reporting requirements established by EPA for WQBELs,

and REMAND for further proceedings in these respects.

We DENY the petition for review with respect to TBELs for viruses

and protists and the monitoring and reporting requirements established by EPA

for TBELs.

The 2013 VGP shall remain in place until EPA issues a new VGP.

‐64‐

APPENDIX Glossary of Acronyms

BACT Best available demonstrated control technology BAT Best available technology economically achievable BCT Best conventional pollutant control technology BMP Best management practice BWTS Ballast water treatment systems CSA Canadian Shipowners Association CWA Clean Water Act ELG Effluent limitation guidelines EPA Environmental Protection Agency IMO International Maritime Organization NAS National Research Council/National Academy of Sciences NPDES National Pollutant Discharge Elimination System NRDC Natural Resources Defense Council NWEA Northwest Environmental Advocates NWF National Wildlife Federation PFR Petition for Review SAB Science Advisory Board TBEL Technology‐based effluent limitation VGP Vessel General Permit WQBEL Water quality‐based effluent limitation

‐65‐