FOR PUBLICATION

UNITED STATES COURT OF APPEALS

FOR THE NINTH CIRCUIT

WATERKEEPER ALLIANCE; No. 23-636

CENTER FOR BIOLOGICAL

DIVERSITY; CLEAN WATER

ACTION; FOOD & WATER EPA No.

WATCH; SURFRIDER EPA-HQ-

FOUNDATION; ENVIRONMENT OW2018-0618

AMERICA; BAYOU CITY

WATERKEEPER; BLACK

WARRIOR RIVERKEEPER; OPINION

HEALTHY GULF; SAN ANTONIO

BAY ESTUARINE

WATERKEEPER; TENNESSEE

RIVERKEEPER; SAN FRANCISCO

BAYKEEPER,

Petitioners,

v.

UNITED STATES

ENVIRONMENTAL PROTECTION

AGENCY; LEE M. ZELDIN,*

Administrator of the United States

* Under Fed. R. App. P 43(c)(2), Lee M. Zeldin has been substituted for his predecessor, Michael S. Regan, as Administrator of the United States Environmental Protection Agency. Environmental Protection Agency,

Respondents,

AMERICAN FUEL &

PETROCHEMICAL

MANUFACTURERS; AMERICAN

PETROLEUM INSTITUTE,

Intervenors.

On Petition for Review of Actions of the Environmental Protection Agency

Argued and Submitted December 5, 2024 Pasadena, California

Filed June 18, 2025

Before: Carlos T. Bea, John B. Owens, and Lucy H. Koh, Circuit Judges.

Opinion by Judge Koh; Dissent by Judge Bea

WATERKEEPER ALLIANCE V. USEPA 3

SUMMARY**

Clean Water Act

The panel granted in part the petition for review of EPA’s decision refusing to revise the technology-based pollution limits, guidelines and standards (collectively referred to as “ELGs”) for certain industries previously promulgated by EPA under the Clean Water Act (“CWA”) and remanded for EPA to reconsider its decision or provide a fuller explanation. The CWA requires that EPA periodically review previously promulgated ELGs and revise them if appropriate. Pursuant to these obligations, EPA reviewed existing ELGs in an action titled Effluent Guidelines Program Plan 15 (“Program Plan 15”). Petitioners challenge EPA’s decision in Program Plan 15 refusing to revise the ELGs for seven specific source categories (the “Seven Industrial Categories”) that Petitioners contend are substantially out of date. The panel held that EPA’s refusal to revise the ELGs for the Seven Industrial Categories in Program Plan 15 constituted final agency action and was thus reviewable under the APA. The panel rejected intervenor’s argument that the court lacked jurisdiction, instead agreeing with petitioners and EPA that the court had jurisdiction under CWA Section 509(b)(1). Consistent with past precedent, a challenge such as this to the substance of existing ELGs falls within Section 509(b)(1). The panel held that Petitioners had

** This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader. not waived their challenge by failing to raise their claims with the agency. However, the panel concluded that plaintiffs had waived their challenge to EPA’s failure to revise new source performance standards, one type of ELG. On the merits, the panel agreed with EPA that it was not required to revise every outdated ELG, and that EPA had some discretion in carrying out its periodic review obligations. The panel nonetheless concluded EPA acted in a manner that was arbitrary and capricious by refusing to revise the ELGs for the Seven Industrial Categories. EPA failed to properly consider advances in pollution control technology, failed to properly consider pretreatment standards and guidelines applicable to indirect dischargers, and failed to properly consider information relating to pollutants not currently covered by the applicable ELGs. The panel also found that EPA’s decision with respect to the plastics molding and forming category was arbitrary and capricious because EPA failed to adequately explain its decision. Dissenting, Judge Bea wrote that because petitioners do not seek review of any EPA actions in promulgating or approving ELGs, this court, as an appellate court, lacks original and exclusive jurisdiction over the action under Section 509(b)(1) of the CWA and this court’s binding precedent in Our Children’s Earth Foundation v. United States Environmental Protection Agency, 527 F.3d 842 (9th Cir. 2008). Where, as here, a plaintiff challenges the screening methodology used in EPA’s review process, the original and exclusive jurisdiction of a court of appeals may not be invoked merely by styling the claim as a challenge to the substance of the ELGs that EPA’s review process has deprioritized for further study. Petitioners challenge not whether the substance of the ELGs for the Seven Industrial

WATERKEEPER ALLIANCE V. USEPA 5

Categories complies with the CWA’s substantive requirements, but whether the Category Ranking Analysis satisfies EPA’s review obligations under the CWA.

COUNSEL

Meg Parish (argued), Jennifer Duggan, and Sarah Kula, Environmental Integrity Project, Washington, D.C.; Hannah M.M. Connor, Center for Biological Diversity, Washington, D.C.; for Petitioners. Gustavus Maxwell (argued), Attorney, Environment & Natural Resources Division, United States Department of Justice, Denver, Colorado; Sue S. Chen, Attorney; Todd Kim, Assistant Attorney General; Environment & Natural Resources Division, United States Department of Justice, Washington, D.C.; Krista Hughes, Attorney, United States Environmental Protection Agency, Washington, D.C.; for Respondents. Jeffrey S. Longsworth (argued) and John A. Sheehan, Earth & Water Law LLC, Washington, D.C., for Intervenors.

OPINION

KOH, Circuit Judge:

The Clean Water Act (“CWA” or “Act”) requires the Environmental Protection Agency (“EPA”) to periodically review the various technology-based pollution limits, guidelines and standards (collectively referred to as “ELGs”) that EPA has promulgated under the Act and revise those ELGs if appropriate. This case is an original jurisdiction suit brought by various environmental organizations challenging EPA’s decision not to revise the ELGs for seven specific source categories (the “Seven Industrial Categories”). Petitioners argue that EPA’s decision not to revise the ELGs for the Seven Industrial Categories was arbitrary and capricious. For the reasons set forth below, we agree in part and remand to EPA for further proceedings.

I. BACKGROUND AND PROCEDURAL

HISTORY

A. Statutory Background The CWA prohibits the discharge of any pollutant except in compliance with the Act. See 33 U.S.C. § 1311(a). To comply with the Act, dischargers are typically required to obtain a National Pollutant Discharge Elimination System (“NPDES”) permit. This case concerns the CWA’s technology-based requirements, which set limits on discharges that are incorporated into NPDES permits. Those requirements can be divided into two broad categories. First, there are limits that apply to facilities that discharge directly into the waters of the United States (i.e. direct dischargers). Direct dischargers are subject to “effluent limitations,” “effluent guidelines,” and “new source performance standards.” See 33 U.S.C. §§ 1311(b), 1314(b), 1316(b)(1)(B). Effluent limitation guidelines are regulations used by permit issuers to formulate specific effluent limitations that are incorporated into particular permits for existing sources. See id. §§ 1311(b), 1314(b), 1362(11). New source performance standards are similar except that they apply to new, rather than existing, sources and require a stricter level of control. See id. § 1316(b).

WATERKEEPER ALLIANCE V. USEPA 7

Second, there are limits for facilities whose discharges reach waters of the United States through publicly-owned treatment works (i.e. indirect dischargers). These are “pretreatment standards,” “pretreatment guidelines,” and “pretreatment standards for new sources.” See id. §§ 1314(g), 1317(b)(1), 1317(c). These limitations are analogous to the effluent limitations, effluent guidelines, and new source performance standards applicable to direct dischargers. EPA collectively describes the six types of limitations referenced above as “Effluent Limitations, Guidelines, and Standards,” or “ELGs.”1 Because ELGs are based on the technology a given type of plant uses, ELGs are set by EPA on an industry-by- industry basis. To date, EPA has promulgated ELGs for 59 industries, referred to as “source categories.” See generally 40 C.F.R. Subchapter N. Within each source category there are, in turn, ELGs for distinct subcategories. For example, one industrial category is Inorganic Chemicals Manufacturing, and within that category there are dozens of subcategories such as Aluminum Chloride Production or Aluminum Sulfate Production. See 40. C.F.R. Part 415. The CWA contains various obligations for EPA to periodically review existing ELGs, which are the focus of this case:

1 As explained infra Section II.D, Petitioners have waived any argument concerning new source performance standards or pretreatment standards for new sources. Accordingly, unless otherwise indicated, we use “ELG” to refer to effluent guidelines, effluent limitations, pretreatment standards, and pretreatment guidelines (i.e., ELGs for existing sources only). • Effluent limitations “shall be reviewed at least every five years and, if appropriate, revised pursuant to the procedure established under such paragraph.” 33 U.S.C. § 1311(d). • EPA must “at least annually . . . revise, if appropriate” effluent guidelines. Id. § 1314(b). • “[F]rom time to time, as technology and alternatives change, [EPA must] revise” new source performance standards and new source pretreatment standards. Id. § 1316(b)(1)(B); see id. § 1317(c) (requiring new source pretreatment standards be promulgated “simultaneously with the promulgation of” new source performance standards). • “[F]rom time to time, as control technology, processes, operating methods, or other alternatives change, [EPA must] revise [pretreatment] standards following the procedure established by this subsection for promulgation of such standards.” Id. § 1317(b)(2) (italics added). • For pretreatment guidelines, EPA must “review at least annually . . . and, if appropriate, revise guidelines for pretreatment of pollutants which [EPA] determines are not susceptible to treatment by publicly owned treatment works.” Id. § 1314(g).

In addition to these various periodic review requirements, CWA Section 304(m) requires EPA to “biennially” “publish in the Federal Register a plan which shall” among other things “establish a schedule for the annual review and revision of promulgated effluent guidelines, in accordance with [Section 1314(b)].” Id.

WATERKEEPER ALLIANCE V. USEPA 9

§ 1314(m)(1)(A) (italics added). EPA is required to “provide for public review and comment on the plan prior to final publication.” Id. § 1314(m)(2). By its terms, this obligation applies only to effluent guidelines; however, in practice EPA describes the results of its other periodic review obligations in the report mandated by Section 304(m) as well. See 3-ER- 602–4, Effluent Guidelines Program Plan 15 (“Program Plan 15”) at 2-3 to 2-4. B. Administrative History This case concerns the most recent EPA report issued on January 19, 2023—titled “Effluent Guidelines Program Plan 15,” or “Program Plan 15”—prepared pursuant to CWA Section 304(m), in which EPA carried out the various periodic review obligations described above. 3 ER 592–658, Program Plan 15. The CWA does not set forth any specific procedures and EPA has not adopted regulations concerning either the periodic review obligations or the Section 304(m) plan, aside from the language quoted above. However, according to EPA, it typically issues a “preliminary plan” one year and then issues a “program plan” the following year, which responds to public comments received on the preliminary plan. EPA’s process for deciding whether to revise a given ELG proceeds in multiple steps. See 3-ER-612, Program Plan 15 at 4-1. First, EPA conducts a “Category Ranking Analysis” that compares discharge data across all 59 source categories with existing ELGs to identify potential candidates for revision. See 3-ER-614, Program Plan 15, at 5-1. EPA claims that it adjusts the way it conducts its Category Ranking Analysis every year “to analyze available data from differing vantage points.” Second, EPA selects several source categories for further consideration in “Preliminary Category Reviews.” See 3-ER-612, Program Plan 15, at 4-1. In selecting source categories to include, EPA claims that it relies not only on the results of the Category Ranking Analysis, but also on other factors “such as stakeholder input and Administration priorities.” 3-ER- 686, 2021 Preliminary Review of Industrial Point Source Categories (“Plan 15 Preliminary Review”) at 1. Third, based on the results of the Preliminary Category Review, EPA will select one or several source categories for which it will conduct “Detailed Studies,” which can take several years. See, e.g., 2-ER-469–71, Preliminary Effluent Guidelines Program Plan 15 (“Preliminary Plan 15”) at 6-1 to 6-3 (describing EPA’s Detailed Study of the Meat and Poultry Products source category). At the conclusion of these steps, EPA decides whether to initiate a rulemaking to revise the ELG. If EPA declines to revise any given ELG, the existing ELG that sets forth the pollution limits with which industry must comply remains in place until the next review cycle when EPA repeats the process. This case is focused on how EPA conducted the first step of its review process, the Category Ranking Analysis. In Preliminary Plan 15, EPA conducted its Category Ranking Analysis by calculating the concentration of pollutants in the discharges from each of the 59 industrial source categories that are currently subject to ELGs. See 2-ER-488, Review of Industrial Wastewater Discharge Monitoring Report Data for Preliminary Plan 15 (“Concentration Ranking Analysis”) at 6. EPA performed this calculation using data drawn from discharge monitoring reports (“DMR data”) that each direct discharger subject to an NPDES permit must report to EPA. See 3-ER-450-52, Preliminary Plan 15, at 3-1 to 3-3. Using this calculation, EPA then ordinally ranked each of the source categories from highest to lowest. EPA then “selected

WATERKEEPER ALLIANCE V. USEPA 11

for further review” the seven source categories that ranked the highest in this “Concentration Ranking Analysis.” 2 3- ER-454-55, Preliminary Plan 15, at 5-5 to 5-6. In addition, EPA also chose to conduct a further review of the Landfills Category (40 C.F.R. Part 409), even though it did not rank highly in the Concentration Ranking Analysis because it was “identified for review through stakeholder input” that suggested landfills may be a major source of per- and polyfluoroalkyl substances (“PFAS”) that are not currently regulated, but are a major administration priority. See 3-ER- 459, 464, Preliminary Plan 15, at 5-10, 5-15. With respect to all ELGs not selected, EPA stated as follows:

For categories not discussed in detail in this Preliminary Plan 15, EPA is currently not prioritizing further review. As described in detail below and in documents in the docket for this preliminary plan, EPA does not have data indicating that these categories discharge quantities of pollutants that would warrant revising the ELGs at this time. Additionally, given EPA’s available resources, these categories are less important than the other [point source categories] for which EPA is undertaking further study and or rulemaking. EPA solicits comment on this

2 These were Metal Products and Machinery (40 CFR Part 438), Battery Manufacturing (40 CFR Part 461), Explosives Manufacturing (40 CFR Part 457), Canned and Preserved Seafood Processing (40 CFR Part 408), Paint Formulating (40 CFR Part 446), Sugar Processing (40 CFR Part 409), and Soap and Detergent Manufacturing (40 CFR Part 417). approach and will continue to review all categories while preparing the next plan.

3-ER-450, Preliminary Plan 15, at 5-1. In Program Plan 15, EPA conducted the Category Ranking Analysis in a slightly different way. Instead of looking at the concentration of pollutants, EPA used the same DMR data to calculate the total pollutant load discharged by each industrial category. See 3-ER-668–9, 2021 Annual Review of Industrial Wastewater Discharges (“Load Ranking Analysis”) at 1–2; 3-ER-614–18, Program Plan 15, at 5-1 to 5-5. Although the results of this “Load Ranking Analysis” differed markedly from the Concentration Ranking Analysis, EPA did not change the categories for which it intended to conduct further review, stating that the Load Ranking Analysis “did not present any findings that altered EPA’s decision on prioritization for industrial category reviews targeting PFAS at this time.” 3- ER-616, Program Plan 15, at 5-3. The result was that any categories screened out at this first step remained unchanged without any further consideration by EPA. Unless otherwise indicated, the “Category Ranking Analysis” refers to both the Concentration Ranking Analysis and Load Ranking Analysis. C. Petitioners’ Claims In this lawsuit, Petitioners3 challenge EPA’s decision in Program Plan 15 not to revise the ELGs for seven specific

3 Petitioners include the following environmental groups: Waterkeeper Alliance, Center for Biological Diversity, Clean Water Action, Food & Water Watch, Healthy Gulf, Environment America, Surfrider Foundation, Bayou City Waterkeeper, Black Warrior Riverkeeper, San

WATERKEEPER ALLIANCE V. USEPA 13

industrial categories: (1) Petroleum Refining; (2) Organic Chemicals, Plastics, and Synthetic Fibers Manufacturing (“OCPSF”); (3) Inorganic Chemical Manufacturing; (4) Fertilizer Manufacturing; (5) Pesticide Chemical Manufacturing; (6) Plastics Molding and Forming Facilities; and (7) Nonferrous Metals Manufacturing (the “Seven Industrial Categories”). Six of the Seven Industrial Categories were screened out at the first step of EPA’s review process using the Category Ranking Analysis. Petitioners did not submit specific comments challenging the ELGs for four of these source categories—OCPSF, Inorganic Chemical Manufacturing, Pesticide Chemical Manufacturing, and Nonferrous Metals Manufacturing— and accordingly EPA did not offer any specific explanation for why it was not conducting further review of them. Petitioners submitted specific comments about, among other things, the Petroleum Refining and Fertilizer Manufacturing categories, as well as EPA’s annual review process more generally. See 2-ER-0501–02, EIP Annual Review Comment at 2–3; 3-ER-549, Earth Justice Fertilizer Comment, at 8; 3-ER-517–31, EIP Petroleum Comment, at 3–18. For example, Petitioners’ comments took issue with EPA’s failure to consider available pollution control technology in EPA’s screening level review and EPA’s reliance on incomplete DMR data. See 2-ER-0501–02, EIP Annual Review Comment, at 2–3; 3-ER-549, 554–58, Earth

Antonio Bay Estuarine Waterkeeper, Tennessee Riverkeeper, and San Francisco Baykeeper. Justice Fertilizer Comment, at 8, 13–17; 2-ER-536–59, Food and Water Watch Comment, at 2–5.4 In its comment responses in Program Plan 15, EPA included sections concerning criticism of its ELG review process, the age of its regulations, and the way in which the agency accounts for technology, the subjects of the instant appeal. See 3-ER-995–1004, Program Plan 15 Comment Responses, at 252–63. EPA also acknowledged that the DMR data it used included only limited information about unregulated pollutants and no information about indirect dischargers. See 3-ER-669, Load Ranking Analysis, at 2; 2- ER-484, Concentration Ranking Analysis, at 2. EPA offered some additional explanation as to the Petroleum Refining and Fertilizer Manufacturing categories, discussed below. Finally, although not announced in Preliminary Plan 15, in Program Plan 15 EPA carried out a preliminary review of the Plastics Molding and Forming category, also discussed below. Petroleum Refining. Petitioners submitted a comment concerning the ELG for the Petroleum Refining category in which they argued this category was a substantial source of pollution, the ELG for the category had not been revised since the 1980s, the ELG was accordingly out of date with existing technologies, and the ELG did not cover many

4 Petitioners’ complaints about EPA’s periodic review process echo similar critiques made by the Government Accountability Office in a 2012 report, which criticized EPA’s review process for its failure to consider available pollution control technologies, and EPA’s flawed reliance on only limited data about unregulated pollutants and indirect dischargers. See 2-ER-308–11, 320–21, GAO, Water Pollution: EPA Has Improved Its Review of Effluent Guidelines but Could Benefit from More Information on Treatment Technologies, at 28–31, 40–41.

WATERKEEPER ALLIANCE V. USEPA 15

existing pollutants that such facilities emit. See 3-ER-517– 31, EIP Petroleum Comment, at 3–18. In its comment response, EPA discussed a 2019 Detailed Study of the Petroleum Refining category that focused on discharges of metals from refineries. See 4-ER-1011, Program Plan 15 Comment Responses, at 268. EPA concluded that “[t]he Petroleum Refining Category did not rank high compared to other categories in the 2019 and 2020 annual rankings [i.e. the Concentration Ranking Analysis] . . . or other current EPA priorities for rulemaking. Therefore, EPA is not prioritizing this category for further review at this time.” Id. Fertilizer Manufacturing. Petitioners submitted a comment concerning the ELG for Fertilizer Manufacturing that argued fertilizer manufacturing releases significant pollution, the ELG for Fertilizer Manufacturing was out of date, and EPA did not adequately consider a representative sample of facilities in deciding not to revise this ELG in Preliminary Plan 15. See 3-ER-545–563, Earth Justice Fertilizer Comment, at 1–22. In the prior review cycle, in Program Plan 14, EPA identified Fertilizer Manufacturing as a category the agency was prioritizing for further review based on the category’s contribution to nutrient pollution. See 4-ER-1013, Program Plan 15 Comment Responses, at 270; 3-ER-368, Nutrient Report, at 3-2. In Program Plan 15, however, EPA explained that it “broadened the focus of pollutants considered in the screening-level review beyond just nutrients” by conducting its Concentration Ranking Analysis. 4-ER-1013, Program Plan 15 Comment Responses, at 270. EPA explained that Fertilizer Manufacturing “did not rank highly as compared to the other categories when considering additional pollutants and, importantly, other current EPA priorities for rulemaking.” Id. Based on this, EPA decided not to conduct further review of this category. See id. Plastics Molding and Forming. EPA advanced the Plastics Molding and Forming category to the Preliminary Review stage as part of its broader effort to address PFAS contamination. See 3-ER-623, Program Plan 15, at 5-10. In carrying out its Preliminary Review of this category, EPA identified information that suggested this ELG was not as strict as it could have been with respect to some pollutants, and that the industry emitted other pollutants that were unregulated. See 3-ER-624, Program Plan 15, at 5-11. But EPA concluded that it was “not prioritizing” this category because “revisions to the ELG are unlikely to result in significant pollutant discharge reductions relative to the other point source categories discussed in this Plan.” 3-ER- 623–25, Program Plan 15, at 5-10 to 5-12.

II. PRELIMINARY MATTERS

There are four preliminary matters we must address before discussing the merits of Petitioners’ challenge. First, EPA argues Program Plan 15 is unreviewable because it does not constitute “final agency action” within the meaning of the APA. Second, Intervenors,5 but not EPA, argue that we lack appellate jurisdiction because this suit does not fall within the CWA’s grant of original appellate jurisdiction. Third, EPA argues that Petitioners have waived their challenge to EPA’s decision not to revise certain ELGs because Petitioners did not comment on them during the

5 American Fuel and Petrochemical Manufacturers Association and American Petroleum Institute (collectively “Intervenors”) filed a motion to intervene, which was subsequently granted by a motions panel.

WATERKEEPER ALLIANCE V. USEPA 17

notice and comment period. Fourth, although not raised by EPA, Petitioners have waived any argument relating to new source performance standards because they failed to meaningfully discuss such standards in their opening brief. Each issue is addressed in turn. A. Final Agency Action EPA argues that the Court lacks jurisdiction because Program Plan 15 is not reviewable under the APA. 6 The APA provides that “[a]gency action made reviewable by statute and final agency action for which there is no other adequate remedy in a court are subject to judicial review.” 5 U.S.C. § 704. EPA argues that Program Plan 15 did not constitute “agency action” at all, let alone “final agency action,” and accordingly the petition should be dismissed. As explained below, neither argument has merit. The APA defines “agency action” as “the whole or a part of an agency rule, order, license, sanction, relief, or the equivalent or denial thereof, or failure to act.” 5 U.S.C. § 551(13) (emphasis added). The APA defines “‘[r]ule’ . . . broadly to include ‘statement[s] of general or particular applicability and future effect’ that are designed to ‘implement, interpret, or prescribe law or policy.’” Perez v. Mortg. Bankers Ass’n, 575 U.S. 92, 95–96 (2015) (quoting

6 EPA frames its argument as a jurisdictional challenge. Although some Ninth Circuit decisions have suggested the APA’s prerequisites to judicial review are jurisdictional, see, e.g., Cabaccang v. U.S. Citizenship & Immigr. Servs., 627 F.3d 1313, 1315 (9th Cir. 2010), the Supreme Court has explained that “[t]he judicial review provisions of the APA are not jurisdictional,” Air Courier Conference v. Am. Postal Workers Union, 498 U.S. 517, 523 n.3 (1991). We need not reach this potential tension because we conclude EPA has engaged in final agency action. 5 U.S.C. § 551(4)). This definition of “rule” includes not only binding rules that have the force of law, but also “[i]nterpretative rules, general statements of policy, or rules of agency organization, procedure, or practice.” Id. at 96 (quoting 5 U.S.C. § 553(b)(A)). Program Plan 15 plainly reflects a statement of the agency’s position with respect to the revision of the various ELGs. It therefore constitutes a “rule,” whether or not it has any binding force. See Ctr. for Biological Diversity v. Haaland, 58 F.4th 412, 416 (9th Cir. 2023) (noting that a “rule” is defined broadly and may include non-binding plans). Similarly, Program Plan 15 expressly declined to revise the various ELGs that Petitioners challenge, and so in that sense constitutes a “failure to act.” 5 U.S.C. § 551(13). Whether or not Program Plan 15 constitutes a final agency action, for the reasons discussed below, it is certainly an agency action within the meaning of APA Section 551(13). The Supreme Court has set forth a two-part test for determining whether agency action is “final.” “First, the action must mark the consummation of the agency’s decisionmaking process,” and “must not be of a merely tentative or interlocutory nature.” Bennett v. Spear, 520 U.S. 154, 177–78 (1997) (internal quotation marks omitted). Second, “the action must be one by which rights or obligations have been determined, or from which legal consequences will flow.” Id. at 178 (internal quotation marks omitted). “In determining whether an agency’s action is final, we look to whether the action amounts to a definitive statement of the agency’s position or has a direct and immediate effect on the day-to-day operations of the subject party, or if immediate compliance with the terms is expected.” Or. Nat.

WATERKEEPER ALLIANCE V. USEPA 19

Desert Ass’n v. U.S. Forest Serv., 465 F.3d 977, 982 (9th Cir. 2006) (cleaned up). “This requires ‘focus on the practical and legal effects of the agency action,’ not on labels, and finality is ‘interpreted in a pragmatic and flexible manner.’” Haaland, 58 F.4th at 417 (quoting Or. Nat. Desert Ass’n., 465 F.3d at 982). EPA argues the first requirement for finality is not satisfied because “EPA’s annual ELG reviews mark not the end—but the beginning—of a decisionmaking process,” and any decision will constantly be revisited in the next review. This argument clearly fails. CWA Section 304(m) requires that EPA conduct biennial reviews through the notice and comment process. See 33 U.S.C. § 1314(m). Program Plan 15 marked the culmination of this biennial cycle. Therefore, for this review cycle there can be no dispute that Program Plan 15 “mark[ed] the ‘consummation’ of the agency’s decisionmaking process.” Bennett, 520 U.S. at 177. Similarly, in addition to the CWA Section 304(m) planning process, EPA is also required to review the various ELGs on a periodic basis and revise them if appropriate. See 33 U.S.C. §§ 1311(d), 1314(b), (g), 1316(b)(1)(B), 1317(b)(2), (c). EPA stated that it satisfied these periodic review obligations through Program Plan 15 and so as to each of these review cycles EPA’s action is final. See 3-ER-602, Program Plan 15 at 2-3. The fact that EPA will have to conduct similar periodic reviews in the future does not change the conclusion that, as to this review cycle, EPA’s decisions are final. See, e.g., U.S. Army Corps of Engineers v. Hawkes Co., 578 U.S. 590, 598 (2016) (explaining that the fact the agency “may revise” its determination “based on ‘new information’” “is a common characteristic of agency action, and does not make an otherwise definitive decision nonfinal”); San Francisco Herring Ass’n v. Dep’t of the Interior, 946 F.3d 564, 579 (9th Cir. 2019) (“[T]he fishermen here had ‘no entitlement to further agency review,’ and ‘[t]he mere possibility that [the] agency might reconsider . . . does not suffice to make an otherwise final agency action nonfinal.’” (quoting Sackett v. EPA, 566 U.S. 120, 127 (2012)). Were it otherwise, an agency could argue that any periodic obligation is not final, and hence not reviewable, simply by pointing to the fact that the agency will need to do something similar again in the future. Such a result cannot be correct. The second Bennett prong, is similarly satisfied. In arguing that it is not, EPA relies heavily on this Court’s decision in Center for Biological Diversity v. Haaland, 58 F.4th 412 (9th Cir. 2023). In that case, the petitioners had filed a petition for rulemaking, demanding that the Fish and Wildlife Service update its Grizzly Bear Recovery Plan. Id. at 415–16. We held that petitioners could not challenge the Fish and Wildlife Service’s refusal to do so, because it was not a final agency action. Id. at 417–20. We explained that, “although the Service had the statutory obligation to draw up a roadmap for recovery of the grizzly bear,” the plans created pursuant to that obligation were not enforceable against any private party or binding on the agency. Id. at 418. Accordingly, we held that the second Bennett prong was not satisfied by the decision not to revise the recovery plan because no legal consequences flowed from that decision. See id. The portions of CWA Section 304(m) that are relevant here, standing alone, might not demand any agency action from which legal consequences flow. Section 304(m)(1)(A) requires that EPA create a “plan” that “shall . . . establish a schedule for the annual review and revision of promulgated effluent guidelines, in accordance with subsection (b) of this

WATERKEEPER ALLIANCE V. USEPA 21

section.” 33 U.S.C. § 1314(m)(1)(A). There is no indication in the statute that this “plan” has any binding effect, either on private parties or EPA itself. 7 In this sense, it is no different than the plan at issue in Haaland. But Program Plan 15 did more than simply set forth a schedule for the review and revision of ELGs. In Program Plan 15, EPA also carried out its related obligations to review existing ELGs and concluded that it would not revise the ELGs for the Seven Industrial Categories. See 3-ER-602, Program Plan 15 at 2- 3. The decision not to revise particular ELGs in connection with EPA’s distinct review-and-revise obligations constitutes final agency action because it meant that the existing standards, which limit the discharge of pollutants and with which industry must comply, will remain in place unchanged. In general, an agency’s decision not to act in the face of a purported obligation to do so can constitute final agency action, satisfying Bennett’s second prong, even if the agency action does not alter the status quo. See Or. Nat. Desert Ass’n, 465 F.3d at 986–87 (explaining that a final action “need not alter the legal regime to which the involved federal agency is subject”). In line with this rule, courts have consistently held that “[a]n agency’s denial of a petition for rulemaking constitutes final, reviewable agency action, except where there is evidence of a clear and convincing legislative intent to negate review.” Weight Watchers Int’l,

7 Petitioners cite our decision in NRDC v. EPA, 542 F.3d 1235 (9th Cir. 2008), to support their contention that CWA Section 304(m) creates mandatory obligations. However, that case concerned other portions of CWA Section 304(m) that are not at issue here. See id. at 1250–52 (considering requirements for new source categories and not existing ELGs). Inc. v. FTC, 47 F.3d 990, 992 (9th Cir. 1995) (citation omitted). Here, EPA was obligated, by statute, to review existing ELGs and make a determination as to whether they should be revised. See Our Children’s Earth Found. v. EPA, 527 F.3d 842, 849 (9th Cir. 2008). Had EPA decided to revise the ELGs for the Seven Industrial Categories, there can be no doubt that decision would be final and reviewable. The decision declining to revise, particularly in the face of public comments requesting modification, is therefore a decision “by which rights or obligations have been determined, or from which legal consequences will flow,” because it meant that regulated industries can continue to emit pollutants at the same current levels. Bennett, 520 U.S. at 178. Haaland is entirely consistent with, and in fact supports, this conclusion. In that case, the dissent argued that the denial of a petition for rulemaking is always reviewable, even if the rule at issue is not binding. See Haaland, 58 F.4th at 425–29 (Sung, J., dissenting). The majority disagreed with this broad rule and instead adopted the narrower position that “we evaluate the [agency’s] denial of the petition to amend the Plan under the same test applicable to a ‘direct’ challenge to the Plan.” Id. at 419. The majority in Haaland distinguished the cases cited by the plaintiff and the dissent on the ground that, in those cases the “regulation, if adopted, would have clearly changed the legal rights and obligations” of the parties. Id. at 419–20 (emphasis added). Thus, the relevant inquiry for purposes of a decision not to engage in rulemaking, is whether the requested action would have had legal consequences. At bottom, the problem with EPA’s position is that the decision declining to revise the ELGs for the Seven

WATERKEEPER ALLIANCE V. USEPA 23

Industrial Categories is no different than the denial of a petition for rulemaking. EPA concedes that the denial of a petition for rulemaking, seeking the exact same relief sought by Petitioners here, would be final and judicially reviewable. EPA’s decision not to revise these ELGs in the face of a clear statutory directive to review is no less final than EPA’s denial of a petition for rulemaking. In both cases EPA is required to decide whether to initiate a rulemaking, in both cases the decision not to do so does not affirmatively alter any parties’ legal obligations, instead leaving the current standards in place, and in both cases EPA could reconsider its decision in the future. B. Original Jurisdiction Under CWA Section 509(b)(1) Intervenors argue we lack jurisdiction for an independent reason. “The CWA contains two separate jurisdictional sections: § 505(a), known as the citizen suit provision, and § 509(b)(1), which relates primarily to challenges to promulgation of certain standards and determinations.” Our Children’s Earth, 527 F.3d at 846. Section 505(a)(2), which is only indirectly relevant to this case, grants district court’s jurisdiction over actions “where there is alleged a failure of the Administrator to perform any act or duty under this chapter which is not discretionary with the Administrator.” 33 U.S.C. § 1365(a)(2). Petitioners brought this action under CWA Section 509(b)(1) (33 U.S.C. § 1369(b)(1)), which authorizes original jurisdiction in the courts of appeal to hear challenges to the following actions of EPA, among others: “in promulgating any standard of performance under section 1316,” “in promulgating any effluent standard, prohibition, or pretreatment standard under section 1317,” and “in approving or promulgating any effluent limitation or other limitation under section 1311, 1312, 1316, or 1345.” 33 U.S.C. § 1369(b)(1)(A)(C)(E). The jurisdiction granted by Section 509(b)(1) is exclusive to courts of appeals. See Our Children’s Earth, 527 F.3d at 847. Intervenors, but not EPA, argue that the actions Petitioners challenge do not constitute “promulgating” or “approving” any standards or limitations within the meaning of CWA Section 509(b)(1), and accordingly we lack jurisdiction. Significantly, EPA concedes that we have jurisdiction under Section 509(b)(1) if Program Plan 15 is reviewable as a final agency action. Specifically, EPA confirmed that it did not agree with Intervener’s jurisdictional argument and rather, stated that it was “aligned with Petitioners on Section 509(b) jurisdiction” during oral argument. Oral Arg. Video 24:55–25:25, https://www.ca9.uscourts.gov/media/video/?20241205/23- 636/. Intervenors rely principally on our decision in Our Children’s Earth Foundation v. EPA. That case concerned an appeal from a suit brought under CWA Section 505 in the district court, alleging EPA had violated a nondiscretionary duty to consider the best available technology in reviewing previously promulgated ELGs pursuant to CWA Sections 301(d) and 304(b), (m) (some of the same review obligations at issue in this case). See Our Children’s Earth, 527 F.3d at 846–49. The Dissent and Interveners maintain that Petitioners’ challenge is essentially a challenge to EPA’s review process like in Our Children’s Earth and is not a challenge to the substance of the ELGs at issue. Dissent at 72–74.

WATERKEEPER ALLIANCE V. USEPA 25

We disagree. This Court has jurisdiction to review EPA’s decision not to revise the ELGs for the Seven Industrial Categories for two reasons. First, our jurisdiction extends “to a substantive review of the appropriateness of the guidelines actually promulgated.” Our Children’s Earth, 527 F.3d at 847. Here, Petitioners’ challenge concerns the substantive adequacy of previously promulgated ELGs. Petitioners argue that the ELGs for the Seven Industrial Categories are out of date and substantially less strict than they should be. As EPA stated during argument, if EPA’s action was a reviewable final agency action, then Petitioner’s challenge is a “challenge to the existing ELG[s] and that’s what 509(b) allows for.” Oral Arg. Video 25:32– 25:51. Unlike in Our Children’s Earth, Petitioners do not argue that EPA has a non-discretionary duty to consider technology or other factors when reviewing ELGs.8 Rather, they argue that in exercising its discretion not to revise ELGs for the Seven Industrial Categories, EPA acted in a manner that was arbitrary and capricious by not considering important factors or otherwise providing a reasoned explanation for ignoring those factors. 9 This is not a

8 To the extent Petitioners argue that EPA has a non-discretionary duty to consider certain factors in reviewing ELGs, Our Children’s Earth Foundation forecloses that argument. See 527 F.3d at 844 (holding CWA does not impose a “mandatory duty” to consider technology-based factors in reviewing ELGs). 9 The Dissent asserts that our focus on the Category Ranking Analysis, and the deficiencies thereof, somehow proves that Petitioners’ claims do not relate to the substance of previously promulgated ELGs. See Dissent at 68–72, 78–80. However, we do not opine, in greater detail, on the substance of particular ELGs for the straightforward reason that EPA did not do so. Under the APA, our review is limited to evaluating the reasons proffered by the agency. See Humane Soc. of U.S. v. Locke, 626 F.3d challenge to any nondiscretionary duty EPA has. Second, by reviewing and declining to revise the ELGs in Program Plan 15, EPA plainly “approv[ed]” them within the meaning of Section 509(b)(1)(E). The Tenth Circuit’s decision in Maier v. EPA, 114 F.3d 1032 (10th Cir. 1997), is on point. In that case, EPA denied a petition for rulemaking that alleged certain CWA regulations were outdated and needed to be revised in light of changes in technology. See id. at 1035–36. The petitioners then brought suit under Section 509(b)(1) to challenge that denial. See id. The Tenth Circuit had “no difficulty construing this as a challenge to an ‘action in approving or promulgating’ under” Section 509(b)(1). Id. at 1038. The Tenth Circuit explained that, “[w]here petitioners’ challenge is to the substance of a regulation that the agency has already promulgated, exclusive jurisdiction in the court of appeals may not be evaded merely by styling the claim as one for failure to revise.” Id. This “rule ensures that an appellate court will review the Administrator’s decision whether the ultimate challenge is to a failure to revise or to a decision to revise.” Id.; cf. Crown Simpson Pulp Co. v. Costle, 445 U.S. 193, 197 (1980) (rejecting construction of Section 509(b)(1)(F) that would “creat[e] . . . a seemingly irrational bifurcated system” of review “at different levels of the

1040, 1048 (9th Cir. 2010) (“The reviewing court should not attempt itself to make up for [an agency’s] deficiencies: We may not supply a reasoned basis for the agency’s action that the agency itself has not given.” (citation omitted)). EPA declined to revise the ELGs for six of the Seven Industrial Categories by relying exclusively upon the Category Ranking Analysis. See 3-ER-450, Preliminary Plan 15, at 5-1; 3-ER- 614–18, Program Plan 15, at 5-1 to 5-5. Insofar as the Category Ranking Analysis was deficient, so too were EPA’s decisions based upon that Analysis about the substance of the ELGs.

WATERKEEPER ALLIANCE V. USEPA 27

federal-court system depending on the fortuitous circumstance of whether the State in which the case arose was or was not authorized to issue permits”). The rule announced in Maier, that challenges to the substance of existing ELGs must be brought in the courts of appeals, is consistent with our precedent. 10 As noted, Our Children’s Earth held that “the circuit court’s exclusive jurisdiction extends only to a substantive review of the appropriateness of the guidelines actually promulgated.” 527 F.3d at 847. Similarly, in NRDC v. EPA, plaintiffs brought suit in district court alleging EPA had a nondiscretionary duty to promulgate ELGs for an industry that lacked any ELGs. 542 F.3d 1235, 1239–41 (9th Cir. 2008). In finding the district court had jurisdiction to hear this claim, we explained that Section 509(b)(1)(E) is inapplicable to suits that “d[o] not challenge the substance of any existing regulations,” because in such circumstances “[p]laintiffs d[id] not seek a review of the existing regulations that the [EPA] is alleged to have ‘approved.’” Id. at 1243. The NRDC court discussed Maier with approval but distinguished the case on the ground that the petitioners in Maier were “requesting a review of existing regulations,” which was different from the claim that “EPA had failed to

10 Maier is also consistent with the other cases cited by Intervenors. See Com. of Pa., Dep’t of Env’t Res. v. EPA, 618 F.2d 991, 995–96 (3d Cir. 1980) (suggesting that suits alleging the “inadequacy of a set of regulations” that already exist fall under Section 509, whereas suits alleging “that a needed regulation was nonexistent” fall under Section 505); Environmental Defense Fund v. EPA, 598 F.2d 63, 90–91 (D.C. Cir. 1978) (concluding claim “that EPA should have engaged in an additional rulemaking proceeding” concerning matter about which EPA had not promulgated a rule belonged in district court). issue the disputed regulations at all.” Id. at 1243–44 (quoting Maier, 114 F.3d at 1039). The Dissent and Intervenors distinguish Maier because it involved the denial of a petition for rulemaking, Dissent at 75–80, but this is a distinction without a difference. 11 For purposes of Section 509(b)(1), EPA’s decision refusing to revise existing ELGs in carrying out its statutorily required periodic reviews is no different than the denial of a petition for rulemaking because in both instances EPA must review the substance of existing ELGs and decide whether to revise them.12

11 The Dissent misleadingly suggests that “EPA believes Maier is ‘very distinct’ from this case because the petitioner in Maier, unlike Petitioners here, did file a petition for rulemaking.” Dissent at 77 n.8. EPA made this statement in discussing the final agency action requirement, which the Dissent does not dispute is satisfied. With respect to CWA Section 509(b)(1), EPA agrees that we have jurisdiction over this case and that Maier’s reasoning is directly applicable here. See Oral Arg. Video 24:55–27:53. 12 The Dissent asserts that “EPA did not decline to revise the ELGs for the Seven Industrial Categories,” but instead chose to “not prioriti[ze]” them “for further study or rulemaking at this time.” Dissent at 69. But the plain language of the review obligations requires EPA to review each ELG and make a decision whether revision is appropriate, which is a decision that necessarily relates to the substance of the ELG. See infra Section III.A. If the Dissent is correct that EPA failed to adequately discuss the substance of the ELGs as part of its periodic review, that simply suggests that EPA acted in a manner that was “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). It does not change the fact that these claims fall within CWA Section 509(b)(1).

WATERKEEPER ALLIANCE V. USEPA 29

Accordingly, consistent with the plain text of the statute, as well as prior precedent construing it, we have jurisdiction over Petitioners’ claim. C. Administrative Waiver EPA argues that Petitioners have waived their challenge to its decision not to revise the ELGs for four of the Seven Industrial Categories—“Inorganic Chemical Manufacturing, OCPSF (aside from the PFAS contaminants that the Agency is already addressing), Pesticide Chemical Manufacturing, [and] Nonferrous Metals Manufacturing”—because neither Petitioners nor anyone else submitted a comment on Preliminary Plan 15 asking EPA to revise these ELGs. EPA concedes that Petitioners challenge to the ELGs for the remaining three of the Seven Industrial Categories—Plastics Molding and Forming, Petroleum Refining, and Fertilizer Manufacturing—was properly preserved. “As a general rule, [appellate courts] will not review challenges to agency action raised for the first time on appeal.” Portland Gen. Elec. Co. v. Bonneville Power Admin., 501 F.3d 1009, 1023 (9th Cir. 2007). This waiver rule “protects the agency’s prerogative to apply its expertise, to correct its own errors, and to create a record for our review.” Id. at 1024. The waiver rule generally does not apply “if an agency has had an opportunity to consider the issue.” Id. We may also “consider any issue that was raised with sufficient clarity to allow the decision maker to understand and rule on the issue raised, whether the issue was considered sua sponte by the agency or was raised by someone other than the petitioning party.” Pac. Choice Seafood Co. v. Ross, 976 F.3d 932, 942 (9th Cir. 2020) (citation and internal quotation marks omitted). EPA is correct that Petitioners did not explicitly single out four of the seven ELGs that Petitioners now seek to challenge. Petitioners misleadingly suggest that they identified these four ELGs in their comment. But the table they cite included a list of every ELG that exists, sorted by age, and so could hardly have placed EPA on notice about Petitioners’ complaints about these four ELGs specifically. See 2-ER-0502–10, EIP Annual Review Comment, at i–vii. Citing a 2019 rulemaking petition, Petitioners similarly claim that their challenge to the OCPSF ELG was presented to the agency. However, EPA explicitly acknowledged this petition for rulemaking in Program Plan 15 and indicated that EPA was still “carefully reviewing” it. 3-ER-613, Program Plan 15, at 4-2. This petition was sent to EPA in 2019, well before the commencement of the Preliminary Plan 15 notice-and-comment process. And because EPA is still considering the issues raised in that petition, it is not a final agency action and cannot form the basis of this suit. Nonetheless, Petitioners’ more general complaints about the ELG review process were sufficient to preserve the arguments they now raise. On appeal, Petitioners argue, inter alia, that the Category Ranking Analysis failed to properly account for changes in control technology, relied on data from discharge monitoring reports (“DMR data”) that was faulty, did not adequately consider pollution from indirect dischargers, and did not adequately account for pollutants not currently subject to ELGs. These issues were adequately presented to and considered by the agency. See, e.g., 2-ER- 0501–02, EIP Annual Review Comment, at 2–3 (noting “EPA’s review process is fundamentally flawed”); 3-ER- 549, Earth Justice Fertilizer Comment, at 8. Petitioners similarly raised issues with Preliminary Plan 15’s decision to rely on DMR data, albeit in connection with specific

WATERKEEPER ALLIANCE V. USEPA 31

source categories. See 2-ER-536–59, Food and Water Watch Comment, at 2–5; 3-ER-554–558 Earth Justice Fertilizer Comment, at 13–17. In its comment responses in Program Plan 15, EPA included sections concerning criticism of its ELG review process, the age of its regulations, and the way in which the agency accounts for technology, the subjects of the instant appeal. See 3-ER-995–1004, Program Plan 15 Comment Responses, at 252–263. EPA similarly identified the lack of DMR data about unregulated pollutants and about indirect dischargers as “limitations” of the Category Ranking Analysis, making clear it was aware of those issues. 3-ER- 669, Load Ranking Analysis, at 2; 2-ER-484, Concentration Ranking Analysis, at 2. What is more, the issues Petitioners raise on appeal are not new to EPA. For example, in 2012, the Government Accountability Office (“GAO”) issued a report criticizing EPA for its periodic review process and identifying many of the issues Petitioners now raise. See 2-ER-320–21, GAO, Water Pollution: EPA Has Improved Its Review of Effluent Guidelines but Could Benefit from More Information on Treatment Technologies (“GAO Report”), at 40–41. In sum, EPA was aware of, and had a chance to address, the issues Petitioners now raise on appeal. Indeed, as to the Plastics Molding and Forming, Petroleum Refining, and Fertilizer Manufacturing categories, EPA does not dispute that the same specific attacks on EPA’s method were properly preserved. Because those attacks apply with equal force to the decision regarding each of the Seven Industrial Categories, insofar as those decisions relied on the same Category Ranking Analysis, we find Petitioners’ challenge was adequately preserved.13 D. Waiver Regarding New Source Performance Standards Petitioners nominally challenge EPA’s decision not to revise new source performance standards in their petition for review. However, they make no arguments as to them, instead focusing exclusively on effluent limitations, effluent guidelines, pretreatment standards, and pretreatment guidelines.14 Petitioners offer additional discussion of new source performance standards in their reply brief, but even these arguments are not well developed. We will generally not consider issues not raised in a party’s opening brief or arguments that are not meaningfully developed. See Hui Ran Mu v. Barr, 936 F.3d 929, 936 (9th Cir. 2019) (“[I]ssues not raised in the opening brief are deemed waived.”); United States v. Graf, 610 F.3d 1148, 1166 (9th Cir. 2010) (“Arguments made in passing and not supported by citations to the record or to case authority are generally deemed waived.”). Application of this rule is particularly appropriate where, as here, the unraised issue of how to treat new source performance standards is complex

13 Petitioners argue the waiver rule is totally inapplicable and cite a line of Fifth Circuit decisions that we have specifically declined to follow in Universal Health Servs., Inc. v. Thompson, 363 F.3d 1013, 1020 n.3 (9th Cir. 2004). 14 Petitioners repeatedly refer to EPA’s decisions regarding “ELGs” which Petitioners define to include only effluent limitations and effluent guidelines. This departs from EPA’s nomenclature in its internal documents and brief, which defines ELGs as also including pretreatment standards, pretreatment guidelines, new source performance standards and pretreatment guidelines for new sources.

WATERKEEPER ALLIANCE V. USEPA 33

and important. The parties apparently assume that new source performance standards should be treated in a similar manner as other ELGs discussed at length by the parties. But an entirely different set of statutory provisions, that the parties have not meaningfully discussed, govern the promulgation and review of new source performance standards. Compare 33 U.S.C. §§ 1311(d) and 1314(b), with 33 U.S.C. § 1316(b)(1)(B). See 3-ER-600–03, Program Plan 15, at 2-1 to 2-4 (summarizing statutory obligations). Accordingly, we hold that Petitioners have waived any challenge to EPA’s decisions regarding new source performance standards.15

III. DISCUSSION

Turning to the merits, Petitioners argue it was arbitrary and capricious for EPA to rely on its Category Ranking Analysis to decline to revise the ELGs for the Seven Industrial Categories for four reasons. First, EPA was required to consider advances in technology in reviewing these ELGs but failed to do so. Second, the Category Ranking Analysis failed to properly consider pretreatment standards and guidelines applicable to indirect dischargers. Third, EPA impermissibly ignored information that these ELGs do not address all pollutants discharged by the relevant source category. Fourth, data EPA relied upon incorrectly classified many facilities into the wrong source

15 Neither the petition nor Petitioners’ opening brief make any explicit reference to pretreatment standards for new sources beyond a handful of citations to the relevant statutory provision, 33 U.S.C. § 1317(c). To the extent Petitioners seek to challenge pretreatment standards for new sources, such a challenge has been similarly waived. category, and so the analysis lacked a substantial basis in fact. A court may set aside an agency action if it was “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). In general, to comply with this standard the agency must “examine the relevant data and articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made.” Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983) (internal quotation marks omitted). The agency may not “rel[y] on factors which Congress has not intended it to consider, entirely fail[] to consider an important aspect of the problem, offer[] 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.” Id. EPA’s decision must also have a “substantial basis in fact.” Ctr. for Biological Diversity v. Zinke, 900 F.3d 1053, 1067 (9th Cir. 2018) (citation omitted). “Judicial review under [the arbitrary and capricious] standard is deferential, and a court may not substitute its own policy judgment for that of the agency.” Fed. Commc’ns Comm’n v. Prometheus Radio Project, 592 U.S. 414, 423 (2021). The standard is particularly deferential when it is applied to an agency’s decision not to engage in rulemaking. See Massachusetts v. EPA, 549 U.S. 497, 527–28 (2007). Nonetheless, “[i]n denying a petition for rulemaking, an agency must, at a minimum, clearly indicate that it has considered the potential problem identified in the petition and provide a ‘reasonable explanation as to why it cannot or will not exercise its discretion’ to initiate rulemaking.”

WATERKEEPER ALLIANCE V. USEPA 35

Compassion Over Killing v. FDA, 849 F.3d 849, 857 (9th Cir. 2017) (quoting Massachusetts, 549 U.S. at 533). As explained in more detail below, EPA was not required, as part of its periodic review, to revise an ELG simply because it was out of date or not comprehensive. Rather, it was within EPA’s discretion to prioritize the revision of certain ELGs over others by using a hazard-based approach that seeks to identify where revision will do the most good. Notwithstanding this, EPA’s Category Ranking Analysis gave too short shrift to the first three issues identified by Petitioners. Accordingly, we remand for EPA to further consider these issues.

A. EPA’s Periodic Review Obligations

Before addressing the particular deficiencies that allegedly affected Program Plan 15’s review, we begin by resolving certain general disagreements the parties present about the scope of EPA’s periodic review obligations. See supra Section I.A (setting forth relevant statutory provisions). First, EPA must carry out its periodic reviews of each ELG within the specified timeframe. See 33 U.S.C. § 1311(d) (requiring review “at least every five years”), id. § 1314(b), (g) (requiring review “at least annually”). Further, the CWA unambiguously requires EPA to review each ELG for each source category.16 We thus reject EPA’s

16 See 33 U.S.C. § 1311(d) (requiring review of “any effluent limitation required by paragraph (2) of subsection (b) of this section”); id. § 1314(b) (requiring promulgation of “guidelines for effluent limitations” and annual review of “such regulations” thereafter); id. § 1314(g) (requiring “review at least annually thereafter and, if argument that it was not required to make a determination whether to revise any given ELG. Although, as explained below, EPA has a measure of discretion in how it carries out its review, the necessity for EPA to review every ELG to decide whether revision is appropriate is plain from the face of the statute. Second, EPA has some discretion to select the manner and method of its review. EPA must review effluent limitations, effluent guidelines and pretreatment guidelines on either an annual or five-yearly basis and revise them “if appropriate.” 33 U.S.C. §§ 1311(d), 1314(b), (g). The terms “appropriate” in connection with certain of EPA’s review obligations certainly suggests EPA has some discretion in how it carries out its reviews. See Our Children’s Earth, 527 F.3d at 850–51. EPA must also revise pretreatment standards “from time to time, as control technology, processes, operating methods, or other alternatives change.” 33 U.S.C. § 1317(b)(2). The use of the phrase “from time to time,” suggests that EPA has discretion about when to revise such standards. But EPA’s discretion is not unlimited. In reviewing EPA’s exercise of its discretion, the court must “fix[] the boundaries of [the] delegated authority, and ensur[e] the agency has engaged in reasoned decisionmaking within those boundaries.” Loper Bright Enters. v. Raimondo, 603 U.S. 369, 395 (2024) (internal quotation marks omitted). “‘[A]ppropriate’ is the classic broad and all-encompassing term that naturally and traditionally includes consideration

appropriate, revis[ion of] guidelines for pretreatment of pollutants”); id. § 1317(b)(1) (requiring EPA to periodically “revise [pretreatment] standards following the procedure established by this subsection for promulgation of such standards”).

WATERKEEPER ALLIANCE V. USEPA 37

of all the relevant factors.” Michigan v. EPA, 576 U.S. 743, 752 (2015). What is meant by “appropriate” is “quintessentially ‘context dependent’ . . . [and] often draws its meaning from surrounding statutory provisions.” Harrington v. Purdue Pharma L.P., 603 U.S. 204, 218 (2024) (citation omitted). Thus, “[a]lthough this term leaves agencies with flexibility, an agency may not ‘entirely fai[l] to consider an important aspect of the problem’ when deciding whether regulation is appropriate.” Michigan, 576 U.S. at 752 (quoting State Farm, 463 U.S. at 43). Considering the surrounding statutory provisions, whether it is “appropriate” to “revise” an ELG must turn, at least to an extent, on the degree to which that ELG fails to conform to the substantive requirements of that ELG set forth in the CWA. The substantive provisions concerning each ELG describe what the proper content of each ELG should be. For example, having effluent limitations that comport with the Act’s substantive requirements is described as a result that “shall be achieved” “[i]n order to carry out the objective of” the CWA, 33 U.S.C. §§ 1311(b)(1)(A), 1311(b)(2)(A), and the Act states that the effluent guideline “regulations shall” “identify” or “specify” particular contents, without any temporal limitation, id. § 1314(b)(2)(A)-(B). Indeed, should EPA actually choose to revise any given ELG, the same factors guiding their initial promulgation will again need to be taken into account. See Our Children’s Earth, 527 F.3d at 851 (The CWA is “unambiguous that revision decisions . . . are constrained by the statute’s mandates for what ‘such regulations’ ‘shall’ accomplish.”). The substantive requirements of the CWA thus are relevant considerations on both a practical and theoretical level. On a practical level, the extent to which any given ELG deviates from what is mandated by the statute necessarily defines the benefits to be obtained from any potential revision. On a theoretical level, the substantive requirements concerning each ELG define the ultimate objective EPA must strive towards. The goal is to have ELGs that conform to those substantive requirements, and to the extent an ELG does not conform with those requirements revision would presumptively be appropriate. This Court has gone so far as to suggest EPA will be “required to justify its regulations” in light of subsequently developed data during its periodic reviews. Ass’n of Pac. Fisheries v. EPA, 615 F.2d 794, 812 (9th Cir. 1980). Thus, one of the objectives of EPA’s periodic review must be to identify outdated ELGs (i.e. ELGs that deviate from the CWA’s substantive commands). Third, although EPA must carry out these periodic reviews with an eye towards identifying outdated ELGs, EPA is not required to revise each and every ELG that is out of date. As this Court explained in Our Children’s Earth, the decision whether to initiate a rulemaking to revise any given ELG is “discretionary[,] as indicated by the ‘if appropriate’ language.” 527 F.3d at 850–51. In particular, EPA has determined that it lacks the resources to revise every ELG. This is, at least as a general matter, a permissible basis to decline to initiate a rulemaking and EPA has broad discretion in this regard. See Massachusetts, 549 U.S. at 527 (EPA “has broad discretion to choose how best to marshal its limited resources and personnel to carry out its delegated responsibilities.”). We therefore reject Petitioners’ apparent assumption throughout their briefs that EPA acted in a manner that was arbitrary and capricious simply because EPA had evidence certain ELGs are out of date but declined to act.

WATERKEEPER ALLIANCE V. USEPA 39

We caution again, however, that EPA’s discretion is not unlimited. “The fact that an agency has broad discretion in choosing whether to act does not establish that the agency may justify its choice on specious grounds.” Jajati v. United States Customs & Border Prot., 102 F.4th 1011, 1017 (9th Cir. 2024) (citation omitted). And “prioritizing pressing matters does not mean agencies have license to ignore the law.” Ctr. for Food Safety v. Regan, 56 F.4th 648, 658 (9th Cir. 2022). Therefore, EPA must exercise its discretion declining to initiate a rulemaking in a manner that is consistent with both the CWA and APA. Finally, we disagree with Petitioners that EPA is categorically prohibited from employing a so-called “hazard-based approach” in carrying out its periodic review obligations. EPA has determined that it lacks the resources to revise each and every ELG that is out of date. Thus, EPA’s Category Ranking Analysis—the first step of its review process—compared the amount of pollution contributed by each source category in an attempt to “prioritize[] the revision of ELGs where they can produce the most significant benefits.” This method of prioritization, at least in the abstract, is reasonable and consistent with the stated goal of the CWA to “restore and maintain the chemical, physical, and biological integrity of the Nation’s waters.” 33 U.S.C. § 1251(a). We thus disagree with Petitioners that EPA relied on “irrelevant factors that Congress did not intend EPA to prioritize.” * * * * * To summarize, EPA is required to carry out its periodic reviews on the timeframes set forth in the CWA and must do so with respect to each ELG. The focus of EPA’s review must be to identify whether each ELG is outdated, in the sense that it does not conform with the Act’s substantive requirements for that ELG. EPA is not required, however, to revise an ELG simply because it is out of date, but EPA must nonetheless offer a reasoned justification for declining to do so that is consistent with the APA and CWA. With these principles in mind, we turn to the particular challenges raised by Petitioners. B. Developments in Pollution Control Technology Petitioners’ first challenge concerns how Program Plan 15’s review of ELGs accounts for changes in pollution control technology. EPA’s Category Ranking Analysis—the first step of its review in both Preliminary and Program Plan 15—looked solely at the amount of pollution contributed by each of the 59 source categories and selected the most highly ranked categories for further analysis. Categories screened out at this first step were given no further consideration as part of EPA’s review process and thus no change was made to these ELGs. See 3-ER-450, Preliminary Plan 15, at 5-1; 3-ER-614–18, Program Plan 15, at 5-1 to 5-5. Petitioners argue it was arbitrary and capricious for EPA to rely on the Category Ranking Analysis because it failed to account for, or even consider, changes in pollution control technology. Petitioners point out that the ELGs for the Seven Industrial Categories have not been revised for 30 years or more, even though there is evidence that pollution control technologies have advanced substantially in that time. This is problematic, Petitioners argue, because the CWA requires that EPA ensure ELGs become more stringent as pollution control technology improves. We agree with Petitioners, but only to a point. Whether it is “appropriate” to “revise” an ELG must turn, at least to an extent, on some consideration of

WATERKEEPER ALLIANCE V. USEPA 41

developments in pollution control technology. ELGs are technology-based standards, meaning they are set by reference to the level of pollution control attainable using the technology that is available to the industry. See 3-ER-455– 56, Preliminary Plan 15, at 2-2 to 2-3 (listing the various technology-based standards that apply in different situations); 3-ER-598–600, Program Plan 15, at 2-1 (noting ELGs are “national industry-specific wastewater regulations based on the performance of demonstrated wastewater treatment technologies”). The precise standard varies based upon the facility and type of pollution at issue, but to take just one example, effluent limitations for toxic and nonconventional pollutants must be based on the “best available technology economically achievable for such category or class,” referred to as “BAT.” 33 U.S.C. § 1311(b)(2)(A); see also § 1314(b)(2)(B). In setting BAT, EPA must consider “the age of equipment and facilities involved, the process employed, the engineering aspects of the application of various types of control techniques, process changes, the cost of achieving such effluent reduction, non-water quality environmental impact (including energy requirements), and such other factors as the Administrator deems appropriate.” 33 U.S.C. § 1314(b)(2)(B). We will refer to these as the CWA’s “technology-based factors.” In any rulemaking to revise a pre-existing ELG, EPA is required to employ the same technology-based factors applicable to the initial promulgation of an ELG. Our Children’s Earth, 527 F.3d at 850–51 (“The statute states that the regulations ‘shall’ account for the technological factors without distinguishing between promulgation and revision.”). Developments in pollution control technology are important factors for EPA to consider when carrying out its periodic reviews. On a practical level, without at least some consideration of technological developments, the Category Ranking Analysis was incapable of achieving EPA’s stated goal of “prioritiz[ing] the revision of ELGs where they can produce the most significant benefits.” This is because the concentration of pollution in an industry’s waste stream, or the amount of overall pollution an industry is releasing, is only half the story if EPA is trying to identify the benefits of revision. Equally important in making any such assessment is the difference between the standard set by the current ELG and the level at which a revised ELG is likely to be set. If that difference is comparatively modest, revision of an ELG for a high polluting industry would nonetheless produce only minimal benefits. By contrast, if an ELG has not been revised in a substantial period of time and technology in that industry has substantially improved, a revision could yield substantial benefits, even if the industry’s overall pollution load is comparatively modest. On a more fundamental level, in adopting this technology-based approach, Congress contemplated a system of gradually tightening technology-based limitations that would become more stringent as pollution control technologies improved. See Our Children’s Earth, 527 F.3d at 851 (“[CWA] makes clear that the regulations must comport with technological criteria that change over time.”); Sw. Elec. Power Co. v. EPA, 920 F.3d 999, 1005 (5th Cir. 2019) (“The [CWA]. . . mandates a system in which, as available pollution-control technology advances, pollution- discharge limits will tighten.”); 33 U.S.C. § 1251(a)(6) (“[I]t is the national policy that a major research and demonstration effort be made to develop technology necessary to eliminate the discharge of pollutants into the navigable waters, waters of the contiguous zone, and the

WATERKEEPER ALLIANCE V. USEPA 43

oceans.”). Congress chose this technology-based approach, and then subsequently reaffirmed such an approach, “because of the historical ineffectiveness of the previous water-quality-based approach.” Our Children’s Earth, 527 F.3d at 848 (quoting S. Comm. on Env’t & Pub. Works, 99th Cong., Report to Accompany S. 1128 (1985 Clean Water Act Amendments) 3–4 (Comm. Print 1985)). The CWA’s periodic review obligations, at issue here, are central to this statutory design. The periodic review obligations are intended to implement this system of continually tightening pollution controls by requiring EPA “to justify its regulations” in light of subsequent developments, including advancements in technology. Ass’n of Pac. Fisheries, 615 F.2d at 812. The centrality of technology to EPA’s periodic reviews is confirmed by the pretreatment standard review obligation. With respect to pretreatment standards (but not pretreatment guidelines) EPA must “from time to time, as control technology, processes, operating methods, or other alternatives change, revise [pretreatment] standards following the procedure established by this subsection for promulgation of such standards.” 33 U.S.C. § 1317(b)(2). Therefore, at least with respect to pretreatment standards, the CWA expressly requires EPA to consider changes in technology. Petitioners advance this argument, and yet EPA offers no response. When the CWA is read as a whole, we do not think this reference to changes in control technology was intended to establish a unique set of criteria for EPA’s review of pretreatment standards, but instead confirms that developments in “control technology, processes, operating methods” are relevant considerations in EPA’s review of the other ELGs as well. We must, wherever possible, “interpret statutes to be coherent and internally consistent.” Silverado Hospice, Inc. v. Becerra, 42 F.4th 1112, 1120 (9th Cir. 2022) (cleaned up). Pretreatment standards are meant to ensure that indirect dischargers—that is, point sources that discharge into publicly owned treatment works—achieve the same levels of pollution control as direct dischargers, taking into account the fact that publicly owned treatment works are able to achieve at least some level of pollution control. See Nat. Res. Def. Council, Inc. v. EPA, 790 F.2d 289, 292 (3d Cir. 1986) (noting CWA added requirement that indirect dischargers “had to ‘pretreat’ its waste waters so as to achieve . . . the same level of toxics removal as was required of a direct discharger”). EPA accordingly sets pretreatment standards for nonconventional and toxic pollutants using the same factors as it uses to set effluent limitations and guidelines for direct dischargers. See id.; 3-ER-601, Program Plan 15, at 2- 3 (“EPA considers the same factors for [pretreatment standards for existing sources] as it does for BAT limitations.”). Pretreatment standards and pretreatment guidelines are similarly closely related, in that the pretreatment guideline regulations are intended to implement and help set pretreatment standards. In light of this close relationship between pretreatment standards and the other ELGs at issue, the explicit reference to technological developments in connection with pretreatment standards confirms what is already apparent from a holistic reading of the statute: that developments in control technology are an important aspect of the problem EPA must consider in reviewing all of the relevant ELGs. Cf. Michigan, 576 U.S. at 751–54 (concluding directive to regulate if “appropriate and necessary” required some consideration of cost by relying on related statutory provision explicitly requiring EPA to consider “the costs of such technologies”).

WATERKEEPER ALLIANCE V. USEPA 45

Without any consideration of technological developments or the technology-based factors used for setting revised ELGs EPA simply could not have evaluated the benefits of further review or revision. By relying solely on the Category Ranking Analysis, EPA thus failed to explain why it could not or did not exercise its discretion to review any source category screened out by that Analysis. See Compassion Over Killing, 849 F.3d at 854. Similarly, by making its revision decisions without reference to technological developments, notwithstanding their centrality to the CWA’s statutory scheme, EPA entirely failed to consider an “important aspect” of ELGs. Michigan, 576 U.S. at 752. To be clear, we do not hold that EPA must conduct an identical review with respect to every single source category or actually apply the technology-based factors to every source category, as EPA would in a rulemaking. To hold otherwise would, in effect, collapse EPA’s review and revision obligations by requiring EPA to conduct a full rulemaking as to every source category. But EPA must nonetheless justify its conclusion that it is not “appropriate” to revise each ELG on grounds that are consistent with the CWA and in a manner that is consistent with the APA. By failing to so much as consider technological developments in the Category Ranking Analysis, EPA failed to do so in Program Plan 15.17

17 Contrary to Petitioner’s argument, we cannot fault EPA for failing to discuss zero-discharge limits in Program Plan 15 because there is no indication in the record that zero-discharge limits would have been feasible for any of the Seven Industrial Categories. See 33 U.S.C. § 1314(b)(3). C. Pretreatment Standards and Guidelines Petitioners’ second challenge concerns EPA’s review of pretreatment standards and guidelines. The CWA draws a distinction between direct and indirect dischargers. Direct dischargers are point sources that emit pollutants directly into waterways and are subject to effluent limitations and effluent guidelines. By contrast, indirect dischargers are point sources that emit pollutants into publicly owned treatment works (“POTW”) and are subject to pretreatment standards and pretreatment guidelines. The Category Ranking Analysis was conducted exclusively using data from discharge monitoring reports (“DMR data”) reported to EPA by direct dischargers. The DMR data did not contain any information about indirect dischargers. See 2-ER-484, Concentration Ranking Analysis, at 2; 3-ER-669, Load Ranking Analysis, at 2. Despite this, EPA relied on its Category Ranking Analysis in deciding not to review further the pretreatment standards and guidelines for categories screened out at this first step. See 3-ER-616, Program Plan 15, at 5-3. EPA acknowledged that both the Concentration Ranking Analysis and Load Ranking Analysis did not evaluate data from indirect dischargers, listing this as one of several “limitations” of the analyses. 2-ER-484, Concentration Ranking Analysis, at 2; 3-ER-669, Load Ranking Analysis, at 2. Yet in neither document did EPA offer an explanation as to why it could rely on this data to make decisions about pretreatment standards and guidelines, beyond the generic suggestion that “[d]espite these limitations, EPA determined that the [DMR] data are a robust and reliable source of information on industrial wastewater discharges, particularly for this initial screening-level review.” 2-ER-484, Concentration Ranking Analysis, at 2; 3-ER-669, Load Ranking Analysis, at 2. For

WATERKEEPER ALLIANCE V. USEPA 47

the reasons set forth below, we hold that it was arbitrary and capricious to rely solely on the Category Ranking Analysis to make decisions about whether to revise pretreatment standards and guidelines. The fact that EPA did not consider pollution from indirect dischargers casts substantial doubt on the ability of the Category Ranking Analysis to carry out its stated purpose of “prioritiz[ing] revisions on the basis of where they will do the most good.” EPA has not offered any support for its apparent assumption that pollution data concerning direct dischargers accurately reflects pollution from indirect dischargers. Although the factors EPA considers in setting pretreatment standards are similar to those used in evaluating effluent limitations and guidelines, they are not identical. In particular, pretreatment standards and guidelines must address “any pollutant which interferes with, passes through, or otherwise is incompatible with” publicly owned treatment works. 33 U.S.C. § 1314(g)(1); see id. § 1317(b)(1), (c). The record suggests that indirect dischargers comprise a significant portion of point sources, although the precise percentage is unclear. EPA states in Program Plan 15 that “ELGs apply to between 35,000 and 45,000 U.S. direct dischargers, as well as to another 129,000 facilities that discharge to POTWs” (i.e. indirect dischargers). 3-ER-600, Program Plan 15, at 2-1. And the record suggests that, in some industries, the vast majority of facilities are indirect dischargers. See 3-ER-707, Preliminary Category Review, at 22 (reporting that, when the Plastics Molding and Forming ELG was promulgated in 1984, the category covered approximately 810 direct dischargers and 1145 indirect dischargers). For example, EPA indicated in Program Plan 15 that “[o]ver 95 percent of the permitted facilities [in the Electrical and Electronic Components category] are indirect dischargers.” 3-ER-628, Program Plan 15, at 6-1. On appeal, EPA acknowledges that the Category Ranking Analysis “does not include data for indirect dischargers,” but responds that “EPA does not count on DMR data to be faultless.” EPA argues that DMR data is the best information it has available, and relying on it was therefore not unreasonable. The generic argument that EPA need not rely on perfect data cannot justify EPA’s use of data that does not meaningfully address indirect dischargers at all. See Flyers Rts. Educ. Fund, Inc. v. Fed. Aviation Admin., 864 F.3d 738, 744 (D.C. Cir. 2017) (concluding agency’s reliance on study to refuse to initiate rulemaking was arbitrary and capricious where the study failed to consider or address important aspect of the problem). And more fundamentally, the record does not support EPA’s contention that it lacked data about indirect dischargers. In Preliminary Plan 15, EPA suggested that it was considering using Toxics Release Inventory (“TRI”) data “to assess discharges of additional toxic pollutants not reported on DMRs, as well as indirect discharges.” 2-ER-459, Preliminary Plan 15, at 5-10. EPA already used TRI data in its Preliminary Reviews—the second step of its review process—to evaluate the risk posed by indirect dischargers, but EPA apparently elected, without explanation, not to incorporate this data into the Category Ranking Analysis (the first step of the process from which EPA decides whether to even proceed to the second step). See, e.g., 3-ER-628, Program Plan 15, at 6-1 (considering TRI data in evaluating indirect discharges from Electrical and Electronic Components category).

WATERKEEPER ALLIANCE V. USEPA 49

At bottom, EPA could not have reasonably carried out its obligations to review pretreatment standards and guidelines for the Seven Industrial Categories by relying exclusively on the Category Ranking Analysis. EPA relied exclusively on data from direct dischargers, without offering any explanation as to why EPA believed it was reasonable to infer that data accurately represented indirect dischargers. EPA thus “entirely fai[led] to consider an important aspect of the problem,” State Farm, 463 U.S. at 43, and failed to “provide a ‘reasonable explanation as to why it cannot or will not exercise its discretion’ to initiate rulemaking.” Compassion Over Killing, 849 F.3d at 854 (quoting Massachusetts, 549 U.S. at 533). D. Unregulated Pollutants Petitioners’ third argument concerns the comprehensiveness of the ELGs in terms of whether they cover all relevant pollutants. For example, Petitioners point out that the ELGs for the Seven Industrial Categories do not address nutrient pollution, even though EPA now recognizes that nutrient pollution “is one of the most widespread, costly, and challenging environmental problems impacting water quality in the United States.” 2-ER-0355, Nutrient Report, at 1-1. Similarly, Petitioners argue that the ELGs for several of the Seven Industrial Categories do not cover discharge from stormwater runoff, even though Congress purportedly amended the CWA in 1987—after these ELGs were promulgated—to make clear that stormwater runoff should be regulated.18

18 To the extent Petitioners argue EPA must revise these ELGs to address all relevant pollutants, we do not have jurisdiction to address this claim. As with the technology-based factors discussed above, the mere fact an ELG is underinclusive is not, by itself, grounds to find EPA’s actions to be arbitrary and capricious. Nonetheless, as with developments in technology, we do not think that unregulated pollutants19 are a matter that EPA may simply ignore without explanation. The CWA clearly contemplates that any given ELG should be comprehensive, covering all relevant pollutants. Revision of outdated ELGs would presumptively be “appropriate” and EPA must justify its decision not to do so. Further, even taken on its own terms, it is not at all clear that the Category Ranking Analysis is capable of identifying which ELGs can be revised to “produce the most significant benefits” without consideration of unregulated pollutants. EPA used DMR data to conduct both the Concentration Ranking Analysis and Load Ranking Analysis. See 3-ER- 0451, Preliminary Plan 15 at 5-2; 3-ER-0541, Program Plan 15 at 5-1. However, in reporting this DMR data to EPA, facilities are generally only required to include information about pollutants that are subject to discharge limits under the relevant permit. See 2-ER-308–09, GOA Report, at 28–29; 3-ER-669, Load Ranking Analysis, at 2. Occasionally, a permit will include limits for a pollutant not covered by an

Our Children’s Earth, 527 F.3d at 846 (district court has jurisdiction over challenges related to “non-discretionary duties” under § 505(a)(2)). 19 We use the term “unregulated pollutants” to refer to pollutants that are not addressed by the relevant ELG for that category. In using this term, we do not mean to suggest that pollutants not addressed by an ELG are never regulated. Even where a pollutant is not addressed by a given ELG, it is still possible that the permit issuer could include facility-specific limitations for that pollutant on a case-by-case basis. See 2-ER-308–09, GOA Report, at 28–29; 3-ER-669, Load Ranking Analysis, at 2.

WATERKEEPER ALLIANCE V. USEPA 51

ELG, but in general this means that DMR data does not contain discharge information about unregulated pollutants. The absence of robust data for unregulated pollutants undermined the reliability of the Category Ranking Analysis. For the Concentration Ranking Analysis, this meant that EPA based its concentration estimates for unregulated pollutants on a very small sample of facilities that are not necessarily representative of the overall industry. For example, Petitioners point out that only 9 of the 331 facilities in the Petroleum Refining category reported data for total nitrogen pollution. This skewed the Load Ranking Analysis because it meant discharges of unregulated pollutants were, for the most part, not reflected in a category’s overall pollution load at all. Indeed, EPA explicitly acknowledged this limitation in DMR data in Program Plan 15. See 3-ER-669, Load Ranking Analysis, at 2. Fundamentally, the existence of unregulated pollutants would seemingly be central to any attempt to determine where a rulemaking would “produce the most significant benefits.” EPA’s focus on the overall amount of pollution being contributed by each source category says little about the actual benefits of any potential revision. Just as some consideration of the extent to which an ELG is out of date is necessary to determine the benefits of revision, so too some consideration of the extent to which an ELG is underinclusive is also necessary. For example, if, as Petitioners argue, the imposition of ELGs for nutrient pollution for the Seven Industrial Categories would produce significant pollution reduction benefits, that fact bears directly on EPA’s decision to prioritize revision of certain ELGs over others. Again, EPA must give at least some consideration to this aspect of the problem in order for its analysis to comport with the minimum requirements of rationality. As with the pretreatment standards discussed in the preceding section, EPA does not offer any real response to this argument. EPA concedes that the data does not contain information about all pollutants and argues that while it had recognized DMR data was flawed, it is the best information EPA has available. But EPA indicated in Preliminary Plan 15 that it could have incorporated TRI data into its Category Ranking Analysis “to assess discharges of additional toxic pollutants not reported on DMRs,” but declined to do so without explanation. 2-ER-459, Preliminary Plan 15, at 5- 10. Here, EPA ignored unregulated pollutants without any explanation as to how it exercised its discretion and failed to consider an important aspect of ELGs and the CWA. See State Farm, 463 U.S. at 43. E. Other Purported Deficiencies Finally, Petitioners argue that it was arbitrary and capricious for EPA to rely on its Category Ranking Analysis because EPA did not accurately classify facilities into point source categories in preparing the data upon which that analysis relied. The DMR data does not include specific information with regards to which ELG category, if any, a given facility is subject. Instead, EPA used Standard Industrial Classification (“SIC”) codes to group various facilities into source categories. See 2-ER-485, Concentration Ranking Analysis, at 3. The problem with this method is that the SIC codes do not perfectly line up with the ELG source categories in the regulations. For example, a detailed study conducted by EPA in 2019 indicated that there were 143 or fewer petroleum refineries in the U.S., but the data used in the Category

WATERKEEPER ALLIANCE V. USEPA 53

Ranking Analysis identified over 300 facilities as petroleum refineries. See 1-ER-0183, Detailed Study of the Petroleum Refining Category—2019 Report, at Appendix A-1 (Sept. 2019); 2-ER-0495, Concentration Ranking Analysis, at 13; 3-ER-0670, Load Ranking Analysis, at 3. Petitioners claim that EPA also excluded approximately 20 facilities that are petroleum refineries from the Concentration Ranking Analysis because the facilities did not report data in a comparable way. See 2-ER-485, Concentration Ranking Analysis, at 3. Similarly, Petitioners claim that the Concentration Ranking Analysis considered pollution discharge data from only 35 fertilizer facilities, even though prior EPA reports had identified 123 such facilities. See 2- ER-0557, Earthjustice Fertilizer Comments, at 16. EPA admits that the source category classifications were overinclusive but argues this is not an issue for two reasons. First, EPA claims that it sought to be overinclusive “across every source category, which, in the aggregate, should go some distance toward cancelling out the effects of over- representation in any single category.” Second, EPA argues that it “controls for some imprecisions by evaluating DMR data differently year over year.” The latter issue is different in kind from those discussed supra Sections III.B–D. The Category Ranking Analysis’s inability to evaluate technology, pollution from indirect dischargers, and unregulated pollutants are fundamental limitations that cast doubt on the scope of what the Category Ranking Analysis was capable of evaluating. By contrast, the first issue of classifying sources at most casts doubt on the precision of the model to an extent that is not clearly established by Petitioners. EPA believes that whatever errors are introduced by this issue are not so significant that they undermine its accuracy. Technical judgments such as this are generally entitled to deference. See Env’t Def. Ctr., Inc. v. EPA, 344 F.3d 832, 869 (9th Cir. 2003). What is more, Petitioners do not identify a reasonable alternative to address this issue. As EPA points out, “[t]he alternative that Petitioners apparently prefer—facility-by- facility classification decisions for 375,000 individual dischargers—would be impractical and ill-suited to EPA’s annual review process.” The only other alternative would presumably be to order EPA to not use DMR data at all. Such a directive would be counterproductive, given that DMR data has substantial benefits as “the most comprehensive data source quantifying pollutants discharged directly to surface waters of the U.S.” 2-ER-484, Concentration Ranking Analysis, at 2. Accordingly, in the absence of any viable alternative proposed by Petitioners, it was reasonable for EPA to rely on the best data it had available, even if the data was far from perfect.

F. EPA’s Remaining Arguments

EPA does little to defend the substance of the Category Ranking Analysis. Instead, EPA attempts to downplay the significance of the Category Ranking Analysis and invokes the need to defer to EPA’s judgment in setting administration priorities. None of these arguments are sufficient to alter our conclusion. First, EPA argues that it considered technological developments at later stages of its review process. EPA argues that it does not matter that EPA failed to consider technological developments with respect to categories screened out at the initial stage because the agency is not required to review each and every source category. A similar argument could be made with respect to pretreatment standards and unregulated pollutants. But as explained supra

WATERKEEPER ALLIANCE V. USEPA 55

Section III.A, the plain text of the CWA requires that EPA review each ELG. The fact that EPA conducted a more thorough review as to other ELGs is thus largely beside the point because EPA must offer an explanation why it is choosing not to revise each ELG. And as to six of the Seven Industrial Categories, the only explanation EPA offered was the deficient Category Ranking Analysis. Second, EPA argues that the method it uses for reviewing ELGs is “iterative” and changes from year-to-year. See 4- ER-995, Program Plan 15 Response to Comments, at 252 (describing prior methods). For example, in Program Plan 14, EPA focused its analysis on the harmful effects of nutrient pollution, and in Program Plan 15 EPA focused on the harmful effects of PFAs. See id. EPA seems to suggest that iterating in this manner allows it to address many of the issues identified by Petitioners. But EPA has not offered an explanation, either in its brief or the administrative record, of how this iterative process actually solves the problems with the Category Ranking Analysis discussed above. Instead, EPA simply lists the various methods that it has used over time and suggests this “help[s] EPA identify point source categories (PSCs) and industries for further review in a more dynamic way.” 4-ER- 995, Program Plan 15 Response to Comments, at 252. Nor is there any obvious logic behind the way EPA has used different review methods from year to year. If anything, the decisions it makes based on these changing methods appear arbitrary. For example, in Program Plan 14 EPA conducted “a nutrient rankings analysis to review industrial categories that typically discharge nutrients and would otherwise not rank high in toxicity rankings.” Id. The Fertilizer Manufacturing category ranked very highly in this nutrient ranking analysis. See 2-ER-355, Nutrient Report, at 1-1. If EPA were engaged in some iterative process that focused on addressing particular types of pollution in each review cycle, one would have expected EPA to revise the Fertilizer Manufacturing ELG based on this finding (or at least study the matter further). Instead, EPA stated in Preliminary Plan 15 that EPA was not revising the Fertilizer Manufacturing ELG because the ELG did “not rank highly in [the Concentration Ranking Analysis] of pollutant discharges as compared to the other categories and other current EPA priorities for rulemaking.” 2-ER-450, Preliminary Plan 15, at 5-1. In other words, EPA conducted a focused analysis of nutrient pollution one year but then refused to initiate a rulemaking based on it by citing the more general Category Ranking Analysis that it did in the next. Similarly, EPA claims that its decision to switch from the Concentration Ranking Analysis in Preliminary Plan 15 to the Load Ranking Analysis in Program Plan 15 was part of this iterative approach. See 4-ER-995–96, Program Plan 15 Comment Response, at 252–53. The results of the Load Ranking Analysis differed substantially from the results of the Concentration Ranking Analysis. Compare 2-ER-456– 58, Preliminary Plan 15, at 5-7 to 5-9, with 3-ER-616–18, Program Plan 15, at 5-3 to 5-5. Yet in Program Plan 15, EPA stated it would continue to prioritize its review of the same source categories that ranked highly in the Concentration Ranking Analysis, without offering any meaningful explanation why the difference in the results of the two analyses was insignificant. See 3-ER-616, Program Plan 15, at 5-3 (“The results of the pollutant load ranking analysis . . . did not present any findings that altered EPA’s decision on prioritization . . . .”).

WATERKEEPER ALLIANCE V. USEPA 57

To the extent there is a method behind EPA’s process from year to year that addressed Petitioners’ concerns, and a reason why EPA believes it is appropriate to rely on an incomplete tool such as the Category Ranking Analysis as part of some iterative process, EPA was required to “articulate a satisfactory explanation” of that method on the record. Locke, 626 F.3d at 1048 (quoting State Farm, 463 U.S. at 43). We will “not supply a reasoned basis for [EPA’s] action that [EPA] itself has not given.” Id. (citation omitted)). In the absence of a more clearly articulated explanation for these decisions, the concept of an “iterative” process seems less like a reasoned basis for EPA’s decisions, and more like the kind of “post hoc rationalization[]” that cannot provide the basis for affirming EPA’s actions. Citizens to Pres. Overton Park, Inc. v. Volpe, 401 U.S. 402, 419 (1971) (internal quotation mark omitted). Third, EPA argues that it did not rely solely on the Category Ranking Analysis to select source categories for further study. More generally, EPA argues that Petitioners may not usurp the agency’s authority to set its own priorities. In Program Plan 15, EPA stated that its “recommendation to further prioritize categories also considers other aspects such as stakeholder input and Administration priorities.” 3-ER-616, Program Plan 15, at 5- 3. By way of explanation, the plan suggested that “EPA continues to focus on and evaluate the extent and nature of PFAS discharges and assess opportunities for limiting those discharges from multiple industrial Categories.” Id. For example, EPA decided to conduct a preliminary review of the Landfills Category based on “stakeholder input” even though it did not rank highly in the Concentration Ranking Analysis. 2-ER-459, Preliminary Plan 15, at 5-10. EPA certainly has discretion to pursue administration priorities, such as responding to the potential threat of PFAS, at least insofar as those priorities are consistent with the CWA. See New York v. EPA, 921 F.3d 257, 262 (D.C. Cir. 2019) (“In general, EPA has broad discretion to choose how best to marshal its limited resources and personnel to carry out its delegated responsibilities,’ which means that EPA has discretion to determine the timing and priorities of its regulatory agenda.” (internal quotation marks omitted)). As we noted, this deference, however, is not unlimited. See Jajati, 102 F.4th at 1017; Compassion Over Killing, 849 F.3d at 854. In Program Plan 15, EPA expressly relied on the Category Ranking Analysis in declining to revise the ELGs for six of the Seven Industrial Categories. See 2-ER-454–55, Preliminary Plan 15, at 5-5 to 5-6; 3-ER-623–25, Program Plan 15, at 5-10 to 5-12; 4-ER-1011, 1013, Program Plan 15 Comment Responses, at 268, 270. And yet, as explained above, that analysis suffered from several flaws that undermined its reliability as a tool to identify ELGs the revision of which would provide the most significant benefits. It is possible that, even having properly considered these deficiencies, EPA would have pressed forward with its current approach based on other administration priorities, but it is not certain. Consistent with the APA, EPA must set forth its reasons not to revise the ELGs for the Seven Industrial Categories on the record. The vague invocation of administration priorities, standing alone, does not meet this standard. At bottom, EPA was required to offer a reasoned explanation why revision of the ELGs for the Seven Industrial Categories was not appropriate. In Program Plan 15, EPA offered the Category Ranking Analysis as the only

WATERKEEPER ALLIANCE V. USEPA 59

basis not to do so for six of the seven categories. Insofar as the Category Ranking Analysis was flawed, so too were EPA’s decisions based upon it. G. Review of Plastics Molding and Forming Category The forgoing analysis in supra Sections III.B-F applies with equal force with respect to six of the Seven Industrial Categories, which EPA screened out by relying solely on the Category Ranking Analysis. However, a different analysis is necessary with respect to the Plastics Molding and Forming category. Unlike the other categories at issue, EPA carried out a Preliminary Review (the second step in EPA’s review process) of the Plastics Molding and Forming category. See 3-ER-623–25, Program Plan 15, at 5-10 to 5-12; 3-ER-707– 27, Plan 15 Preliminary Review, at 22–40. As part of its Preliminary Review of this category, EPA considered not only DMR data, but also TRI data, which includes data about indirect dischargers, and certain additional data about PFAs discharges that was obtained through third-party sources. See 3-ER-710–15, Plan 15 Preliminary Review, at 25–30. To evaluate whether ELGs for currently regulated pollutants are correct (or out of date), EPA compared current discharge data for the industry against the regulatory standard set in the ELG. See 3-ER- 715–18, Plan 15 Preliminary Review, at 30–33. EPA also reviewed discharge data for pollutants that are not currently covered by existing ELGs for this category. See 3-ER-718– 23, Plan 15 Preliminary Review, at 33–38. EPA ultimately concluded that it was “not prioritizing the Plastics Molding and Forming Category for further review or ELG revision at this time” because “[b]ased on the available data, revisions to the ELG are unlikely to result in significant pollutant discharge reductions relative to the other point source categories discussed in this Plan.” 3-ER-624, Program Plan 15, at 5-11. Unlike the other categories discussed above, EPA certainly had sufficient data before it about the Plastics Molding and Forming category to make this determination. EPA gave consideration to changes in pollution control technology, pollution from indirect dischargers and unregulated pollutants. Nonetheless, Petitioners identify at least two other deficiencies in EPA’s review of the Plastics Molding and Forming category to which EPA offers no meaningful response.20 First, in its Preliminary Review of this category “EPA found that 98 percent of the annual loads were associated with stormwater,” even though stormwater runoff is not currently covered by the ELG. 3-ER-624, Program Plan 15, at 5-11. Despite this finding that the vast majority of pollution was contributed by stormwater runoff, EPA chose to exclude information about stormwater runoff from its review of unregulated pollutants contributed by this category. See 3-ER-624, Program Plan 15, at 5-11; 3-ER- 710–11, 713–14, Plan 15 Preliminary Review, at 25–26, 28– 29. Perplexingly, at no point did EPA consider whether it would be appropriate to establish ELGs for stormwater

20 The record belies Petitioners’ contention that EPA improperly failed to consider whether to set ELGs for certain pollutants for which it had previously “reserved” setting limits. In its Preliminary Review of this category, EPA explicitly examined discharge data about these pollutants but found no information that would lead the agency to alter its prior decision to reserve setting limits. See 3-ER-716–17, Plan 15 Preliminary Review, at 31-32.

WATERKEEPER ALLIANCE V. USEPA 61

runoff even though they accounted for such a high percentage of the overall pollution load for the industry. Second, upon reviewing the pollutants currently covered by the ELG, EPA concluded that “[r]eported average concentrations of TSS, oil and grease, and BOD5,” were “an order of magnitude below the current ELG.” 3-ER-710, Plan 15 Preliminary Review, at 31. This is seemingly strong evidence that the ELGs for these pollutants are less strict than they should be, and yet EPA offers no further discussion of this finding. Petitioners argue it was arbitrary and capricious for EPA to fail to revise the Plastics Molding and Forming ELG in light of these issues. EPA offers no response to these category-specific arguments, instead relying on the generic claim that there are no mandatory factors applicable to its periodic reviews. An agency is required to “articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made” and may not rely on “an explanation for its decision that runs counter to the evidence.” State Farm, 463 U.S. at 43 (citation and internal quotation marks omitted). EPA’s bare assertion that revision of this category was “unlikely to result in significant pollutant discharge reductions relative to the other point source categories,” 3-ER-624, Program Plan 15, at 5-11, without any further explanation or engagement with the contrary facts in the record is insufficient to meet this standard. It is of course possible that EPA has some reason why it believed tightening these ELGs or regulating stormwater runoff would not yield significant benefits. But nowhere, either in the record before the agency or on appeal, does EPA explain what those reasons might be, and we are not at liberty to supply such an explanation. See Locke, 626 F.3d at 1048. Given how significant these two issues seem to be for any evaluation of the benefits of a potential revision, and the absence of any meaningful explanation about them, we remand to the agency to consider them further.

IV. CONCLUSION

For the reasons discussed above, we hold that EPA’s action in declining to revise the ELGs for the Seven Industrial Categories was arbitrary and capricious. We therefore remand for the EPA to reconsider its decision or provide a fuller explanation. See Locke, 626 F.3d at 1048, 1053 (remanding “to afford the agency the opportunity either to articulate a reasoned explanation for its action or to adopt a different action with a reasoned explanation that supports it”). PETITION FOR REVIEW DENIED in part, GRANTED in part, and REMANDED

BEA, Circuit Judge, dissenting:

As an appellate court, we do not have original subject matter jurisdiction over this case under the Clean Water Act (“CWA”) and our binding precedent in Our Children’s Earth Foundation v. United States Environmental Protection Agency, 527 F.3d 842 (9th Cir. 2008) (McKeown, J.). I respectfully dissent. Pursuant to the CWA, the United States Environmental Protection Agency (“EPA”) promulgates effluent limitations, guidelines, and standards (“ELGs”) for pollutant

WATERKEEPER ALLIANCE V. USEPA 63

dischargers of various industries (“source category” or “category”). If a person believes that an ELG merits revision, he may petition EPA for revisory rulemaking. If his rulemaking petition is denied, he may sue EPA directly in a court of appeals under the CWA, as he challenges an EPA action in promulgating or approving the ELG. That said, it can take many years for EPA to respond to such a rulemaking petition. Separately, the CWA also mandates EPA to review existing ELGs periodically for possible revision. Maj. Op. at 5–6. EPA biennially publishes the results of such ELG reviews. Id. at 8–9. For each two-year period, EPA first issues a “preliminary plan” in the first year to solicit public comments, and EPA then issues a “program plan” the next year responding to those comments. Id. at 9. If a person believes EPA has failed to perform any mandatory duties in this biennial ELG review process, he may sue EPA in a district court under the CWA. In theory, such a suit may be brought every two years. Accordingly, to avoid awaiting EPA’s protracted response to his rulemaking petition, a person who desires revision of an ELG may attempt to challenge EPA’s periodic review process, arguing that EPA’s review criteria improperly deferred the consideration of that ELG for possible revision. This shortcut, however, was largely foreclosed by our decision in Our Children’s Earth, because EPA enjoys substantial discretion in designing its review criteria and district courts’ jurisdiction under the CWA covers only EPA’s failure to perform its mandatory duties. 527 F.3d at 851. What if the same person still challenges EPA’s review process but somehow styles his claim as a challenge to EPA’s promulgation or approval of the existing ELGs that EPA’s review criteria have deprioritized for consideration of revision? By styling his claim as such, he obviates the need to file a slow-moving rulemaking petition with EPA, sidesteps the jurisdictional inquiry as to whether he challenges EPA’s failure to perform a mandatory duty, and paves for himself a direct pathway to a court of appeals. This is what Petitioners have confected here. 1 The majority now licenses this confection by finding that we have original and exclusive jurisdiction over this case. In doing so, my colleagues pay only lip service to the CWA and Our Children’s Earth, 527 F.3d 842. I do not join such a disregard for the authorities that bind us. I. EPA typically proceeds in multiple steps when it determines whether to revise an ELG. Maj. Op. at 9. First, it uses a screening methodology to identify ELG candidates for further study. Id. Second, it selects some ELGs for preliminary study, considering not only the results from the first step but also other factors such as stakeholder input and policy priorities. Id. at 9–10. Third, EPA selects a few ELGs for multi-year detailed study. Id. at 10. Then, with the benefit of the detailed study, EPA finally decides whether to initiate a rulemaking process to revise the studied ELGs. Id.

1 Petitioners include “Waterkeeper Alliance, Center for Biological Diversity, Clean Water Action, Food & Water Watch, Healthy Gulf, Environment America, Surfrider Foundation, Bayou City Waterkeeper, Black Warrior Riverkeeper, San Antonio Bay Estuarine Waterkeeper, Tennessee Riverkeeper, and San Francisco Baykeeper.” Maj. Op. at 12– 13 n.3.

WATERKEEPER ALLIANCE V. USEPA 65

Petitioners claim that the screening methodology EPA used in its most recent Effluent Guidelines Program Plan 15 (“Program Plan 15”)2—in particular, the Category Ranking Analysis3—failed to account for technological advancement and other factors, resulting in the de-prioritization of many ELGs for further study, including the ELGs for seven specific source categories (“Seven Industrial Categories”).4 See id. at 12–13, 33–34. As a threshold matter, we must ascertain whether we have subject matter jurisdiction over Petitioners’ claim. Under the CWA, courts’ jurisdiction is bifurcated. Section 505(a)(2) of the CWA grants district courts original jurisdiction over a variety of matters, while Section

2 Program Plan 15 was issued in January 2023, following EPA’s Preliminary Effluent Guidelines Program Plan 15 (“Preliminary Plan 15”). See Maj. Op. at 9–12. 3 Category Ranking Analysis “compares discharge data across all 59 source categories with existing ELGs to identify potential candidates” for preliminary study. Maj. Op. at 9. In this case, EPA calculated the concentration of pollutants in the discharges from each of the 59 source categories and the total pollutant load discharged by each category. Id. at 10–12. EPA then selected for further study the seven categories that ranked the highest. Id. at 10–11. Notably, EPA “expects to expand” its Category Ranking Analysis in the future “to include additional metrics such as . . . age of regulations, current ELG requirements and technology basis,” and so forth. As will be discussed, a small part of Petitioners’ claim challenges EPA’s preliminary study of the ELGs for one source category. See id. at 59–62. 4 The Seven Industrial Categories include: “(1) Petroleum Refining; (2) Organic Chemicals, Plastics, and Synthetic Fibers Manufacturing . . . ; (3) Inorganic Chemical Manufacturing; (4) Fertilizer Manufacturing; (5) Pesticide Chemical Manufacturing; (6) Plastics Molding and Forming Facilities; and (7) Nonferrous Metals Manufacturing.” Maj. Op. at 12–13. 509(b)(1) confers courts of appeals original and exclusive jurisdiction only over certain carefully circumscribed subjects. 33 U.S.C. §§ 1365(a)(2), 1369(b)(1). Specifically, Section 505(a)(2) provides that district courts “shall have jurisdiction” over actions “where there is alleged a failure of the Administrator to perform any act or duty under this chapter which is not discretionary with the Administrator.” 33 U.S.C. § 1365(a)(2) (emphasis added). In contrast, Section 509(b)(1) grants courts of appeals original and exclusive jurisdiction over seven enumerated subjects, of which the following allegedly form the bases of Petitioners’ claim here: “Review of the Administrator’s action (A) in promulgating any standard of performance under section 1316 of this title, . . . (C) in promulgating any effluent standard, prohibition, or pretreatment standard under section 1317 of this title, . . . (E) in approving or promulgating any effluent limitation or other limitation under section 1311, 1312, 1316, or 1345 of this title, . . . .” 33 U.S.C. § 1369(b)(1) (emphases added). Given the plain text and the clear structure of the CWA, we have admonished “against expansive application of” Section 509(b)(1). Our Children’s Earth, 527 F.3d at 846 (quoting League of Wilderness Defenders v. Forsgren, 309 F.3d 1181, 1190 n.8 (9th Cir. 2002)); see also Nat. Res. Def. Council v. EPA (“NRDC”), 542 F.3d 1235, 1242 (9th Cir. 2008) (“[W]e have cautioned that our jurisdiction under [Section] 509(b)(1) is not to be construed expansively, particularly given the ‘specificity and precision’ that Congress used in identifying the actions that fall under [Section] 509(b)(1).” (citation omitted)). As such, we would have original and exclusive jurisdiction over this case only

WATERKEEPER ALLIANCE V. USEPA 67

if Petitioners challenged EPA’s actions in promulgating or approving the ELGs for the Seven Industrial Categories. Petitioners do not so challenge. In fact, the relationship between the screening methodology that they do challenge and the actual promulgation or approval of any ELGs is both procedurally distant and logically tenuous. By the majority’s own count, the Category Ranking Analysis is at least two steps removed from EPA’s ultimate decisions regarding whether to revise an ELG: the steps of preliminary study and detailed study. Maj. Op. at 9–10. What’s more, one cannot draw a direct line between the Category Ranking Analysis and EPA’s ultimate decisions on ELG revision. The ELGs that are deprioritized by the Category Ranking Analysis are not necessarily excluded from consideration for possible revision because, as the majority acknowledges, EPA selects ELGs for preliminary study based not only on the results of its Category Ranking Analysis but also on input from stakeholders and its policy priorities. Id. The record demonstrates that EPA in fact selected some ELGs for further study without regard to the Category Ranking Analysis. See, e.g., id. at 57 (“EPA decided to conduct a preliminary review of the Landfills Category based on ‘stakeholder input’ even though it did not rank highly in the Concentration Ranking Analysis.” (citation omitted)). Therefore, Petitioners do not seek review of any EPA actions in promulgating or approving ELGs and, accordingly, we do not have original and exclusive jurisdiction over this case under Section 509(b)(1). Concluding the opposite, the majority misreads the CWA, misconstrues this case, and misinterprets our binding precedent. A. First, the CWA. The majority appears to believe that Section 509(b)(1) applies whenever a petitioner challenges EPA’s discretionary actions, since Section 505(a)(2) allows district courts to review only EPA’s failure in carrying out its nondiscretionary duties. See id. at 25–26. In the majority’s view, as Petitioners challenge EPA’s largely discretionary decision regarding the ELGs for the Seven Industrial Categories, they perhaps cannot sue in a district court under Section 505(a)(2), so their claim must fall under Section 509(b)(1). See id. The majority’s belief finds no basis in the text of the CWA. Section 509(b)(1) does not confer blanket jurisdiction for us to review in the first instance all of EPA’s discretionary actions. Rather, as relevant here, it extends our original and exclusive jurisdiction to review only EPA’s actions in “promulgating” and “approving” ELGs. 33 U.S.C. § 1369(b)(1)(A), (C), (E). Simply put, “a jurisdictional defect under [Section] 505(a)(2) does not mean that jurisdiction is proper under [Section] 509(b)(1).” Our Children’s Earth, 527 F.3d at 846. B. My colleagues paint Petitioners’ claim as one challenging not the Category Ranking Analysis or EPA’s review process, but the “substance” of the ELGs for the Seven Industrial Categories and “EPA’s decision not to revise” them. Maj. Op. at 25–29. The majority then claims

WATERKEEPER ALLIANCE V. USEPA 69

that EPA’s decision not to revise an existing ELG effectively equates to, and thus constitutes, EPA’s formal promulgation or plain approval of that ELG. See id. As a matter of fact, however, EPA did not decline to revise the ELGs for the Seven Industrial Categories. Instead, EPA stated that, given its “current priorities and available resources,” all the ELGs that it did not discuss in detail in its Program Plan 15 were “not priorities for further study or rulemaking at this time,” and EPA would “continue to review all [] source categories while preparing the next plan.”5 See also id. at 11 (“For categories not discussed in detail in this Preliminary Plan 15, EPA is currently not prioritizing further review.” (citation omitted)); id. at 15 (“EPA is not prioritizing [the petroleum refining] category for further review at this time.” (citation omitted)); id. (“EPA explained that [the fertilizer manufacturing category] ‘did not rank highly as compared to the other categories . . . and, importantly, other current EPA priorities for rulemaking.’” (citation omitted)); id. at 16 (“EPA concluded that it was ‘not prioritizing’ [the plastics molding and forming] category . . . .” (citation omitted)). At oral argument, EPA denied having made any final decision not to revise the ELGs for the Seven Industrial Categories.6 Oral Arg. Video 25:50–

5 To the extent that the majority believes the CWA mandates EPA’s periodic review process to produce an ultimate decision not to revise the ELGs that EPA has deprioritized, see Maj. Op. at 28 n.12, Petitioners should have sued EPA for not having made such a decision in a district court under Section 505(a)(2). See 33 U.S.C. § 1365(a)(2). Such a claim concerns how EPA should construct its review process, not the promulgation or approval of any ELGs. 6 EPA concedes that Section 509(b)(1) would apply if we were to misconstrue this case as involving “an EPA decision not to revise [an 26:05, https://www.ca9.uscourts.gov/media/video/? 20241205/23-636/. Moreover, the majority opinion is replete with statements that reveal the real nature of Petitioners’ claim, to wit: a challenge not to the substance of any ELGs, but to the screening methodology used by EPA in its ELG review process. See, e.g., Maj. Op. at 10 (“This case is focused on how EPA conducted the first step of its review process, the Category Ranking Analysis.”); id. at 40 (“Petitioners argue it was arbitrary and capricious for EPA to rely on the Category Ranking Analysis because it failed to account for, or even consider, changes in pollution control technology.”); id. at 42 (“[W]ithout at least some consideration of technological developments, the Category Ranking Analysis was incapable of achieving EPA’s stated goal of ‘prioritiz[ing] the revision of ELGs where they can produce the most significant benefits.’”); id. at 47 (“The fact that EPA did not consider pollution from indirect dischargers casts substantial doubt on the ability of the Category Ranking Analysis to carry out its stated purpose . . . .”); id. at 49 (“At bottom, EPA could not have reasonably carried out its obligations to review pretreatment standards and guidelines for the Seven Industrial Categories by relying exclusively on the Category Ranking Analysis.”); id. at 51 (“The absence of robust data for unregulated pollutants undermined the reliability of the Category Ranking Analysis.”); id. at 52 (“Petitioners argue that it was arbitrary and capricious for EPA to rely on its Category Ranking Analysis because EPA did not accurately classify facilities

ELG] in response to a petition [for rulemaking].” See Oral Arg. Video 25:50–27:52. Neither did Petitioners file a rulemaking petition, nor did EPA make such a decision in this case.

WATERKEEPER ALLIANCE V. USEPA 71

into point source categories in preparing the data upon which that analysis relied.”); id. at 53 (“The Category Ranking Analysis’s inability to evaluate technology, pollution from indirect dischargers, and unregulated pollutants are fundamental limitations that cast doubt on the scope of what the Category Ranking Analysis was capable of evaluating.”); id. at 54 (“EPA does little to defend the substance of the Category Ranking Analysis.”). These statements from the majority opinion, and many more, discuss EPA’s review process and its Category Ranking Analysis, rather than the substance of any ELGs, much less the promulgation or approval thereof. Most revealing is the majority’s discussion of the waiver issue in this case. Specifically, Petitioners argue that EPA should have considered revising the ELGs for the Seven Industrial Categories in its Program Plan 15, but Petitioners did not mention four of the Seven Industrial Categories when they submitted comments in response to EPA’s Preliminary Plan 15. Id. at 29 (“[N]either Petitioners nor anyone else submitted a comment on Preliminary Plan 15 asking EPA to revise these [four] ELGs.”). To the extent Petitioners purportedly challenge the substance of the ELGs for these four unmentioned source categories, EPA contends that Petitioners have failed to preserve their claim. Id. The majority disagrees:

Petitioners’ more general complaints about the ELG review process were sufficient to preserve the arguments they now raise. On appeal, Petitioners argue, inter alia, that the Category Ranking Analysis failed to properly account for changes in control technology . . . . What is more, the issues Petitioners raise on appeal are not new to EPA. For example, in 2012, the Government Accountability Office (“GAO”) issued a report criticizing EPA for its periodic review process and identifying many of the issues Petitioners now raise.

Id. at 30–31. So, according to the majority, Petitioners have successfully preserved their claim that EPA should have considered revising the ELGs for all the Seven Industrial Categories because Petitioners, in essence, challenge EPA’s “review process” and its “Category Ranking Analysis”—not the substance of the ELGs for the Seven Industrial Categories. See id. at 29–32. But, in the meantime, the majority holds that this Court enjoys original and exclusive jurisdiction because Petitioners, in essence, challenge the substance of the ELGs for the Seven Industrial Categories. See id. at 23–29. The majority cannot have it both ways. In my view, Petitioners do not seek review of EPA’s actions in promulgating or approving the ELGs for the Seven Industrial Categories. Instead, the thrust of Petitioners’ claim is nothing but a challenge to EPA’s screening methodology in its review process. The CWA does not make such a challenge reviewable directly by a court of appeals under Section 509(b)(1). C. An accurate construction of Petitioners’ claim brings this case squarely within the ambit of our binding precedent. Our Children’s Earth, 527 F.3d 842.

WATERKEEPER ALLIANCE V. USEPA 73

About two decades ago, Our Children’s Earth Foundation and Ecological Rights Foundation (together, “OCE”) sued EPA in the district court, alleging, inter alia, that the screening methodology used by EPA in its periodic review of existing ELGs failed to account for technological advancement. Id. at 844–45. Sound familiar? Like Petitioners here, OCE claimed that, due to the deficient screening methodology, EPA had identified only a handful of ELGs for further study, leaving a wide swath of other ELGs not considered for possible revision. See Appellant & Petitioner’s Opening Br. at 15–16, 19, Our Children’s Earth Found. v. EPA, 527 F.3d 842 (9th Cir. 2008) (No. 05- 16214). The district court ruled in EPA’s favor, finding that EPA had discharged its mandatory duties under the CWA. Our Children’s Earth Found. v. EPA, No. C 04-2132PJH, 2005 WL 6395158, at *6 (N.D. Cal. May 20, 2005). OCE appealed, and we had to deal with a threshold question on appeal: whether the district court had subject matter jurisdiction in the first place. Our Children’s Earth, 527 F.3d at 845. We held that OCE properly sued EPA in the district court under Section 505(a)(2) because the original and exclusive jurisdiction enjoyed by courts of appeals under Section 509(b)(1) “covers only challenges to ‘promulgation’ or ‘approval’” of ELGs, “not failure to comply with allegedly mandated [ELG review] procedures, which [was] the thrust of OCE’s suit.” Id. at 846–47. Notably, when presented with OCE’s assertion that EPA’s screening methodology left all but a handful of ELGs off its revision priority list, the panel in Our Children’s Earth—unlike the majority here—did not say OCE effectively challenged the substance of those ELGs that EPA had deprioritized for consideration of revision. Rather, the panel concluded as follows: We [] agree with the district court that the circuit court’s exclusive jurisdiction “extends only to a substantive review of the appropriateness of the guidelines actually promulgated, and not to the threshold question of whether the statutory requirements of the CWA have been met.” No such promulgated [ELGs] are at issue here.

Id. at 847 (emphasis added); id. (finding that OCE’s challenge did “not stem from the promulgation or approval of” any ELGs). Hence, OCE’s suit did not fall under Section 509(b)(1). Id. Neither does this case. Granted, unlike OCE, which generally contended that EPA failed to update as many ELGs as OCE claimed a more adequate screening methodology would have recommended, Petitioners here specifically identify the Seven Industrial Categories that EPA has deprioritized for further study. But the majority cannot seriously distinguish Our Children’s Earth simply because Petitioners’ claim is more specific than OCE’s. Otherwise, plaintiffs who challenge EPA’s review process can effortlessly evade district courts’ jurisdiction by identifying some ELGs that the review process has deprioritized. That would render Our Children’s Earth obsolete. Therefore, where, as here, a plaintiff challenges the screening methodology used in EPA’s review process, the original and exclusive jurisdiction of a court of appeals may not be invoked merely by styling the claim as a challenge to the substance of the ELGs that EPA’s review process has deprioritized for further study.

WATERKEEPER ALLIANCE V. USEPA 75

D. Against Our Children’s Earth, the majority relies on the Tenth Circuit’s opinion in Maier v. United States Environmental Protection Agency, 114 F.3d 1032 (10th Cir. 1997), a case that we implicitly distinguished in Our Children’s Earth. 7 See Maj. Op. at 26–28. Not only is Maier not binding, but it is also inapt—it did not involve EPA’s periodic review of ELGs or any screening methodology used in such review. In Maier, the petitioners petitioned EPA to initiate rulemaking under the CWA, contending that EPA’s regulations regarding wastewater treatment had been rendered inadequate by technological developments and should thus be revised. 114 F.3d at 1034. EPA denied their petition, and the petitioners sued EPA in the Tenth Circuit, invoking its original and exclusive jurisdiction pursuant to Section 509(b)(1). Id. at 1036. The relevant question before the Tenth Circuit was whether EPA’s denial of the petitioners’ rulemaking petition to revise an ELG constituted an “action” “in approving or promulgating” the ELG. See

7 In Our Children’s Earth, EPA relied on Maier to argue that the district court improperly exercised jurisdiction under Section 505(a)(2). See Answering Br. of the Federal Defendants-Appellees at 51, Our Children’s Earth Found. v. EPA, 527 F.3d 842 (9th Cir. 2008) (No. 05- 16214). OCE countered that Maier was distinguishable because, different from the petitioners in Maier who challenged EPA’s refusal to revise upon a rulemaking petition, OCE challenged “EPA’s overall procedure in conducting its effluent guidelines and limitations reviews.” Appellants’ Reply Br. at 30, Our Children’s Earth Found. v. EPA, 527 F.3d 842 (9th Cir. 2008) (No. 05-16214). We sided with OCE on this issue. Our Children’s Earth, 527 F.3d at 847. id. at 1037. In other words, whether an inaction counts as an action. The Tenth Circuit answered: It depends. An action “in” approving or promulgating ELGs does not simply mean an action “relating to” approving or promulgating ELGs, so Section 509(b)(1) “arguably does not apply to [EPA]’s refusal to promulgate a [new] rule in the first instance.” See id.; see also NRDC, 542 F.3d at 1241–44 (holding that Section 509(b)(1) did not apply because the plaintiffs sought to compel EPA to promulgate new regulations). But the Tenth Circuit went on to reason that a challenge to a failure to revise an existing regulation “is more akin to” a challenge to the existing regulation itself and can thus be “constru[ed]” as a challenge to an action “in” the prior promulgation, or the current approval, of that regulation. Maier, 114 F.3d at 1038. Accordingly, the Tenth Circuit concluded that the original and exclusive jurisdiction of courts of appeals can be invoked under Section 509(b)(1) if a petitioner’s challenge is “to the substance of a regulation that the agency has already promulgated.” Id. Following Maier, assuming it was correctly decided, we would exercise original and exclusive jurisdiction over this case, had Petitioners first petitioned EPA—as did the petitioners in Maier—for rulemaking to revise the ELGs for the Seven Industrial Categories and had EPA denied that rulemaking petition. As discussed, however, neither did Petitioners file such a petition actually challenging the substance of the ELGs for the Seven Industrial Categories, nor did EPA issue such a denial defending the substance of these ELGs and declining to revise them. The majority thinks this is a “distinction without a difference,” reasoning that EPA’s de-prioritization decision

WATERKEEPER ALLIANCE V. USEPA 77

in its review process is no different than EPA’s denial of a rulemaking petition because we “must review the substance of the existing ELGs” in both instances.8 Maj. Op. at 28. The majority again contradicts itself. Specifically, when discussing the merits of Petitioners’ claim, the majority writes extensively to demarcate EPA’s “review obligations,” id. at 35–40, and it recognizes that “EPA has some discretion to select the manner and method of its review” and to decide whether and when to revise any of the existing ELGs as a result of its review, id. at 36; see also id. at 6–9, 35 (quoting statutory provisions to delineate the “scope of EPA’s periodic review obligations”). Then, the majority devises that EPA’s discretion in that regard is not unlimited because those “review obligations” must be carried out “with an eye towards” “the substantive requirements” set forth in the CWA, which requirements “describe what the proper content of each ELG should be” and are mandatory. Id. at 36–38. That said, the majority still believes EPA’s “review obligations” under the CWA and the

8 To the extent that the majority has put much emphasis on EPA’s position as to this jurisdictional issue, see Maj. Op. at 24–25, it is worth noting that EPA believes Maier is “very distinct” from this case because the petitioners in Maier, unlike Petitioners here, did file a petition for rulemaking and, as a result, EPA had an opportunity to defend the substance of the ELG at issue there. Oral Arg. Video 26:52–27:53. This distinction, according to EPA, is “very important.” Id. I agree. This distinction may also have implications in determining whether Program Plan 15 constitutes a final agency action. See Maj. Op. at 28 n.11. I have no occasion to opine on such possible implications, since we do not have original jurisdiction over this case in the first place. But see id. at 17–23 (the majority discussing this “final agency action” issue without mentioning Maier). CWA’s “substantive requirements” are separate and distinct. See id. at 35–40. Let’s apply the majority’s teaching to the analysis here. Had Petitioners, like the petitioners in Maier, petitioned EPA for rulemaking to revise the ELGs for the Seven Industrial Categories and sued EPA upon its denial of their rulemaking petition, Petitioners would have contended that the current ELGs for the Seven Industrial Categories failed to comply with the “substantive requirements” in the CWA, and the majority would be analyzing those “substantive requirements” to discern whether the current ELGs fall short. Such a hypothetical case would have been properly brought before this Court under Section 509(b)(1), and the majority would be writing a totally different opinion—at the very least, the majority’s attempt to bridge EPA’s “review obligations” and the CWA’s “substantive requirements” could be easily dispensed with. See id. This is why Maier is distinguishable: The petitioners in Maier filed a petition for revisory rulemaking to challenge the substance of the ELG at issue there, whereas Petitioners here did not file such a petition; instead, they challenge only the preliminary step that EPA has taken in its review process. E. If all this is too complicated, I offer a simple litmus test: Does the majority opinion detail the substance of the ELGs for any of the Seven Industrial Categories? No, it does not. Almost exclusively focusing on the alleged inadequacy of the Category Ranking Analysis, the majority is largely silent on the substance of the ELGs for the Seven Industrial Categories—the purported sine qua non of Petitioners’ claim

WATERKEEPER ALLIANCE V. USEPA 79

upon which my colleagues assert original and exclusive jurisdiction.9 This silence tells volumes. My colleagues explain in a footnote that they could “not opine, in greater detail, on the substance of particular ELGs” because “EPA did not do so” when it responded to Petitioners’ comments to Preliminary Plan 15. Id. at 25–26 n.9. This explanation lays bare the majority’s misconstruction of this case and its misplaced reliance on Maier. In Program Plan 15, EPA did not discuss the substance of the ELGs for the Seven Industrial Categories and instead referred to the Category Ranking Analysis when responding to Petitioners’ comments, because Petitioners’ comments— different from a rulemaking petition as in Maier—merely urged EPA to conduct a thorough review process for these ELGs, and because EPA did not decline to revise these ELGs like it did in Maier. Both Petitioners’ comments and EPA’s responses largely centered on EPA’s review process and how it triaged different ELGs. In fact, as the majority acknowledges, Petitioners did not comment at all on the substance of the ELGs for four of the Seven Industrial

9 One exception perhaps is the majority’s discussion of the plastics molding and forming category. Maj. Op. at 59–62. EPA did not screen out this source category by relying on the Category Ranking Analysis; rather, EPA conducted a preliminary study for this category and decided not to “prioritiz[e]” it for further study because “revisions to th[is] ELG are unlikely to result in significant pollutant discharge reductions relative to the other point source categories discussed in [Program Plan 15].” Id. at 59–60 (citation omitted). Nonetheless, Petitioners’ challenge to EPA’s preliminary study on this ELG suffers from the same defect as does Petitioners’ challenge to the Category Ranking Analysis, because the preliminary study still does not constitute an EPA action in promulgating or approving the ELG. Categories. Id. at 29–32. No wonder EPA did not opine on their substance—EPA was simply not put on notice nor requested to offer such an opinion. And the fact that EPA did not have an opportunity to defend the substance of the ELGs at issue here is exactly why Maier is inapposite and our exercise of original jurisdiction improper. See Oral Arg. Video 26:52–27:53.

III.

In my view, Petitioners challenge not whether the substance of the ELGs for the Seven Industrial Categories complies with the CWA’s substantive requirements, but whether the Category Ranking Analysis satisfies EPA’s review obligations under the CWA. Therefore, we do not have original and exclusive jurisdiction under Section 509(b)(1) of the CWA. When an agency does not act as promptly as a petitioner would prefer, suggestions for courts to act beyond their jurisdiction come as no surprise. Statutes and precedents usually curb that penchant for excess. As the majority here oversteps the bounds set by the CWA and Our Children’s Earth, I respectfully dissent.