State of Texas v. EPA

Opinion

Case: 16-60670 Document: 262-1 Page: 1 Date Filed: 09/22/2025

United States Court of Appeals for the Fifth Circuit United States Court of Appeals Fifth Circuit

____________ FILED September 22, 2025 No. 16-60670 Lyle W. Cayce ____________ Clerk

State of Texas; Texas Commission on Environmental Quality,

Petitioners,

versus

United States Environmental Protection Agency; Lee Zeldin, in his official capacity as Administrator of the United States Environmental Protection Agency,

Respondents. ______________________________

Petition for Review of a Final Rule of the Environmental Protection Agency 81 Fed. Reg. 53,284 ______________________________

Before Smith, Higginson, and Douglas, Circuit Judges. Stephen A. Higginson, Circuit Judge: “It is a fair and reasonable demand on the part of a sovereign that the air over its territory should not be polluted on a great scale.” Georgia v. Tenn. Copper Co., 206 U.S. 230, 238 (1907). But every so often the winds fail to heed those demands, carrying pollutants emitted in one state into the air over others. Case: 16-60670 Document: 262-1 Page: 2 Date Filed: 09/22/2025

No. 16-60670

The people of the upwind and downwind states, through their representatives in Congress, negotiated a legislative solution to this problem in the Clean Air Act, 42 U.S.C. §§ 7401–7671q (CAA). The Good Neighbor Provision of the Act, id. § 7410(a)(2)(D)(i), requires upwind states to manage their pollution so that they do not prevent their downwind neighbors from attaining the federal air quality standards. States must account for Good Neighbor requirements in their plans—called SIPs, for State Implementation Plans—implementing the federal standards. The Environmental Protection Agency (EPA) is required to review these SIPs to ensure that upwind states comply with their Good Neighbor obligations. At issue in this case is EPA’s statutorily required review of a Good Neighbor SIP submitted by Texas in 2012. In the SIP, Texas said that its contributions to ozone pollution in other states were not significant enough to require any mitigating action at all. In support, Texas submitted (1) two charts showing declining ozone statistics in nine metro areas, (2) one paragraph summarizing general wind patterns in Dallas and Houston, and (3) a map of 2010 ozone levels at monitors in Texas and adjacent states. Texas attached (4) several pages of ozone measurement data without any analysis. After notice and comment, EPA reviewed this submission and rejected it on the grounds that the SIP did not adequately address the full set of statutory requirements and that the SIP’s conclusions were unsupported by sufficient analysis in any case. As additional support, EPA noted that data developed in a concurrent rulemaking showed that emissions in Texas did in fact contribute to violation of the federal ozone standards in other states. Texas now petitions this court for review of EPA’s disapproval. Texas says that it was entitled to additional process and that EPA’s reasoning was arbitrary. Because EPA’s procedure complied with the statutory requirements and its reasoning was sound, we deny the petition.

2 Case: 16-60670 Document: 262-1 Page: 3 Date Filed: 09/22/2025

I. The CAA requires EPA to regulate certain air pollutants, emissions of which “cause or contribute to air pollution which may reasonably be anticipated to endanger public health or welfare.” Id. §§ 7408(a)(1), 7409. These are called “criteria pollutants.” One is ozone. See 40 C.F.R. pt. 50. The CAA further requires EPA to set National Ambient Air Quality Standards (NAAQS) to limit concentrations of criteria pollutants to levels “requisite to protect the public health.” 42 U.S.C. § 7409(b)(1). EPA is required to revise the NAAQS periodically. Id. § 7409(d). Within three years of a NAAQS promulgation or revision, states submit SIPs to EPA for achieving the new NAAQS. Id. § 7410(a). SIPs must comply with the Good Neighbor Provision of the CAA, see id. § 7410(k)(3), which requires that a plan “contain adequate provisions” “prohibiting” emissions of “any air pollutant in amounts which will” “contribute significantly to nonattainment in, or interfere with maintenance by, any other State with respect to any” NAAQS, id. § 7410(a)(2)(D). To review SIPs, EPA promulgates “minimum criteria” “limited” to ensuring that a SIP submission contains “the information necessary to enable the Administrator to determine whether the plan submission complies with the provisions of this chapter.” Id. § 7410(k)(1)(A). When it receives a SIP, EPA first conducts a technical review within 60 days to determine whether the submission contains this information. Id. § 7410(k)(1)(B). EPA then has 12 months to conduct a substantive review to determine whether the SIP meets the requirements of the CAA, and to approve or disapprove the SIP accordingly. Id. § 7410(k)(2)–(3). If EPA disapproves a SIP (or finds a submission technically incomplete) it must promulgate a Federal Implementation Plan (FIP) within two years. Id. § 7410(c)(1).

3 Case: 16-60670 Document: 262-1 Page: 4 Date Filed: 09/22/2025

EPA is required to make sure that states meet their Good Neighbor obligations as part of its substantive review of SIP submissions. See id. § 7410(k)(3). That inquiry is scientifically complex. Because the wind can carry air pollutants from one state to another, with ozone forming from chemical reactions along the way, nonozone pollution in one state may interfere with attainment of the ozone NAAQS in another. See EPA v. EME Homer City Generation, L.P., 572 U.S. 489, 497 (2014). EPA has historically ensured that SIPs account for this problem by building models to apportion the transported emissions responsible for downwind ozone problems among upwind states. The first Good Neighbor rulemaking relevant here is EPA’s 2005 Clean Air Interstate Rule, 70 Fed. Reg. 25,162 (May 12, 2005) (CAIR). CAIR established Good Neighbor obligations for the 1997 fine particulate matter (PM2.5) and ozone NAAQS. EPA determined that Texas contributed significantly to downwind nonattainment of only the PM2.5 standard. See id. at 25,167. On review, the D.C. Circuit found “fatal flaws in the rule,” ultimately remanding the rule to EPA without vacating it. North Carolina v. EPA, 531 F.3d 896, 901 (D.C. Cir.) (per curiam), modified on reh’g, 550 F.3d 1176 (D.C. Cir. 2008) (per curiam). EPA accordingly replaced CAIR in 2011 with the Cross-State Air Pollution Rule, 76 Fed. Reg. 48,208 (Aug. 8, 2011) (CSAPR). CSAPR disapproved SIPs that had relied on CAIR and promulgated FIPs based on the 1997 ozone, 1997 PM2.5, and 2006 PM2.5 NAAQS. When CSAPR was challenged at the Supreme Court in EME Homer, the Court upheld EPA’s modeling approach, 572 U.S. at 524, and on remand the D.C. Circuit again remanded to EPA without vacating, EME Homer City Generation, L.P. v. EPA, 795 F.3d 118, 138 (D.C. Cir. 2015).

4 Case: 16-60670 Document: 262-1 Page: 5 Date Filed: 09/22/2025

EPA revised the ozone NAAQS in 2008. National Ambient Air Quality Standards for Ozone, 73 Fed. Reg. 16,436 (Mar. 27, 2008). Texas submitted a Good Neighbor SIP for the revised NAAQS on December 13, 2012. In its submission, Texas described “ozone trends” and “existing ozone control strategies” and concluded that “emissions from Texas do not contribute significantly to nonattainment or interfere with maintenance of the 2008 ozone NAAQS in another state.” The submission noted a declining trend in ozone measurements over the period from 1990 to 2010 in selected designated nonattainment areas, including Dallas-Fort Worth and Houston as well as the seven other areas “geographically closest to Texas” in the West, South, and Midwest. With respect to the two nonattainment areas closest to Texas, Baton Rouge and Memphis, the submission noted prevailing wind patterns in the Dallas and Houston areas and that air quality monitors between Texas and Baton Rouge showed attainment, speculating that local emissions might be contributing to nonattainment in Baton Rouge and Memphis. The submission stated, however, that “[t]he amount of ozone in other nonattainment areas from precursor sources outside of Texas and the amount of ozone coming from Texas into other nonattainment areas was not calculated. Because there are additional precursor sources located between Texas and other areas, it is difficult to determine how much ozone in other areas would be due to transport and how much ozone would be due to those sources of ozone precursors.” The submission described significant NOX reductions under CAIR’s trading programs for SO2 and NOX emissions directed to attainment of the 1997 PM2.5 NAAQS, as well as other control measures. Texas stated in its proposed SIP that CSAPR required further NOX reductions but that CSAPR was stayed pending judicial review, and updated its final submission to note that CSAPR (which had not yet reached the Supreme Court) had been vacated by the D.C. Circuit, reinstating CAIR

5 Case: 16-60670 Document: 262-1 Page: 6 Date Filed: 09/22/2025

for the time being. Accordingly, Texas concluded that existing control measures would be sufficient to meet its Good Neighbor obligations. On December 20, 2012, EPA determined that Texas’s submission was technically complete. With a pending Supreme Court decision in EME Homer now poised to weigh in on EPA’s Good Neighbor methodology, the agency did not complete its substantive review of Texas’s submission within the statutory timeframe. However, once the Supreme Court approved of the way EPA had exercised its discretion to calculate Good Neighbor obligations in CSAPR, EPA proceeded with its substantive review of the SIP. During the review process, EPA provided the states multiple opportunities to take advantage of updated modeling. First, in January 2015, EPA provided interstate pollution transport modeling information to the states to facilitate supplementation of Good Neighbor SIPs. Although those projections showed Texas contributing to ozone problems in several downwind states, Texas did not supplement its submission. Next, in August 2015, EPA followed up with a published notice advising the states of the updated data and stating that, if it was necessary for EPA to exercise its “backstop” role by promulgating a FIP, the agency intended to use the updated data for that purpose. Notice of Availability of Updated Ozone Transport Modeling Data for the 2008 Ozone NAAQS, 80 Fed. Reg. 46,271, 46,273 (Aug. 4, 2015). The updated data again linked Texas to ozone problems in downwind states. Id. at 46,277. Again, however, Texas did not update its submission. In December 2015, EPA echoed these repeated advisements with yet another. Cross-State Air Pollution Rule Update for the 2008 Ozone NAAQS, 80 Fed. Reg. 75,706, 75,710 (Dec. 3, 2015). EPA advised that it would finalize FIPs for states that either “failed to submit a complete good

6 Case: 16-60670 Document: 262-1 Page: 7 Date Filed: 09/22/2025

neighbor SIP” or for which EPA “issue[d] a final rule disapproving [the state’s] good neighbor SIP.” Id. at 75,708. Only in April 2016 did EPA finally publish a formal notice of proposed rulemaking. In that notice, EPA proposed to disapprove the portion of Texas’s SIP submission pertaining to its Good Neighbor obligations. Approval and Promulgation of Air Quality Implementation Plans, 81 Fed. Reg. 21,290 (Apr. 11, 2016). EPA identified several problems that required it to disapprove the submission. One set of problems stemmed from the fact that Texas did not “give the ‘interfere with maintenance’ clause of section 110(a)(2)(D)(i)(I) independent significance” as required under North Carolina, 531 F.3d at 910, “because its analysis did not attempt to evaluate the potential impact of Texas emissions on areas that are currently measuring clean data, but that may have issues maintaining that air quality.” 81 Fed. Reg. at 21,292. Additionally, “Texas did not fully evaluate whether emissions from the state significantly contribute to nonattainment in other states” because “Texas limit[ed] its discussion of data only to areas designated nonattainment in states that are geographically closest to Texas.” Id. EPA took issue with this approach because “transported emissions may cause an area to measure exceedances of the standard even if that area is not formally designated nonattainment by the EPA.” Id. Finally, Texas’s “[a]nalysis of wind patterns, emissions data, and ambient monitoring data . . . d[id] not quantify the magnitude of impact from Texas emissions to downwind states.” Id. at 21,294. EPA added that its own modeling of 2017 ozone levels provided “the most up-to-date information for assessing interstate transport of air pollution for the 2008 ozone NAAQS” and “show[ed] that Texas emissions significantly contribute to ozone concentrations in areas of nonattainment and interfere with maintenance of

7 Case: 16-60670 Document: 262-1 Page: 8 Date Filed: 09/22/2025

the 2008 ozone NAAQS in other states.” Id. EPA explained that it had modeled air quality in 2017 because it was the last full year of data relevant to the attainment schedule for the 2008 ozone standard. Id. at 21,293. Responding to Texas’s documentation of existing control measures, EPA stated that it was inappropriate to rely on the CAIR trading programs to demonstrate compliance with Good Neighbor obligations because those programs were no longer in force due to CSAPR, and because those programs were directed to achieving compliance with an older ozone standard rather than the “different and more stringent” 2008 ozone NAAQS. Id. at 21,294– 95. EPA further noted that its modeling accounted for those measures and still showed Texas contributing to downwind ozone problems. Id. at 21,293. EPA received comments opposing the proposed disapproval from Texas and industry. Texas maintained that its analysis was adequate. Texas stated that EPA’s lack of “timely guidance” and EPA’s untimely review of its SIP had “undermin[ed] [its] ability to submit sufficient SIP revisions.” Texas took issue with EPA’s proposal of a FIP before any action had been taken on its SIP, stating that the CAA required EPA to act on the SIP before promulgating a FIP and that EPA’s approach effectively “required that Texas conform its SIP to the proposed FIP.” Finally, Texas attacked EPA’s modeling methodology for reasons not argued on petition for review. A set of industry commenters led by Luminant Generation Company LLC similarly argued, as relevant here, that EPA’s proposed disapproval of the SIP was merely a “post hoc rationalization” for the “predetermined” “outcome” of the proposed FIP. The industry commenters further argued that the phrase “interfere with maintenance” did not require Texas to consider receptors in areas currently meeting the relevant NAAQS. EPA disapproved the SIP on August 12, 2016. Approval and Promulgation of Air Quality Implementation Plans, 81 Fed. Reg. 53,284,

8 Case: 16-60670 Document: 262-1 Page: 9 Date Filed: 09/22/2025

53,289 (Aug. 12, 2016). Responding to the comments, EPA reiterated that Texas’s analysis did not sufficiently address the statutory requirements of the Good Neighbor Provision, again noting that the geographically limited analysis did not analyze interference with maintenance, nor account for nondesignated nonattainment—all while the most accurate and updated technical information indicated that Texas emissions did “impact air quality in other states relative to the 2008 ozone NAAQS.” Id. at 53,285 (discussion of SIP), 53,287 (discussion of statutory requirements). EPA stated that these deficiencies were independent bases for disapproval of the SIP that did not rely on the proposed FIPs for states with deficient plans. Id. at 53,286. EPA noted that “the Supreme Court clearly held that ‘nothing in the statute places the EPA under an obligation to provide specific metrics to States before they undertake to fulfill their good neighbor obligations.’” Id. at 53,285 (quoting EME Homer, 572 U.S. at 509). EPA accordingly stated that, although it was not required to give Texas an opportunity to correct problems with its SIP, Texas was welcome to submit an approvable SIP. Id. at 53,286. EPA waited until October 2016 to finalize a FIP for Texas. Cross- State Air Pollution Rule Update for the 2008 Ozone NAAQS, 81 Fed. Reg. 74,504 (Oct. 26, 2016) (CSAPR Update). The FIP was revised following remand from the D.C. Circuit, Wisconsin v. EPA, 938 F.3d 303, 336 (D.C. Cir. 2019), as Revised Cross-State Air Pollution Rule Update for the 2008 Ozone NAAQS, 86 Fed. Reg. 23,054 (Apr. 30, 2021), and that revision was upheld in Midwest Ozone Group v. EPA, 61 F.4th 187, 193 (D.C. Cir. 2023). Texas and the Texas Commission on Environmental Quality (collectively, Texas) petitioned this court for review on October 11, 2016, naming EPA and then-EPA Administrator Gina McCarthy (collectively, EPA) as respondents. See Wisconsin, 938 F.3d at 335 (citing 42 U.S.C. § 7607(b)(1)). A set of industry entities including Luminant (collectively,

9 Case: 16-60670 Document: 262-1 Page: 10 Date Filed: 09/22/2025

industry petitioners) were given leave to intervene in support of the petitioners. Texas and the industry petitioners moved to strike certain documents from the administrative record on the basis that they pertained to the CSAPR Update rulemaking rather than the SIP disapproval. The motion was carried with the case. II. SIP disapprovals are reviewed under the Administrative Procedure Act, 5 U.S.C. §§ 551–559, 701–706 (APA). See Alaska Dep’t of Env’t Conservation v. EPA, 540 U.S. 461, 496 & n.18 (2004). Under 5 U.S.C. § 706, “agency interpretations of statutes” are reviewed without deference. Loper Bright Enters. v. Raimondo, 603 U.S. 369, 392 (2024). Where a statutory delegation of authority “leaves [the] agenc[y] with flexibility,” id. at 395 (quoting Michigan v. EPA, 576 U.S. 743, 752 (2015)), the reviewing court must “independently identify and respect such delegations of authority, police the outer statutory boundaries of those delegations, and ensure that agencies exercise their discretion consistent with the APA,” id. at 404. An agency’s exercise of discretion is reviewed for whether it is “arbitrary, capricious,” or “an abuse of discretion.” 5 U.S.C. § 706(2)(A). Agency action is arbitrary and capricious “if the agency has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.” Barr v. SEC, 114 F.4th 441, 446 (5th Cir. 2024) (quoting Tex. Oil & Gas Ass’n v. EPA, 161 F.3d 923, 933 (5th Cir. 1998)), petition for cert. filed, No. 24-1233 (U.S. Mar. 28, 2025). “Such review is neither sweeping nor

10 Case: 16-60670 Document: 262-1 Page: 11 Date Filed: 09/22/2025

intrusive. Instead, we ‘ask whether the agency considered the relevant facts and articulated a satisfactory explanation for its decision; we cannot substitute our judgment for the agency’s.’” Fort Bend Cnty. v. U.S. Army Corps of Eng’rs, 59 F.4th 180, 194 (5th Cir. 2023) (quoting Amin v. Mayorkas, 24 F.4th 383, 393 (5th Cir. 2022)). “A reviewing court must be ‘most deferential’ to the agency where, as here, its decision is based upon its evaluation of complex scientific data within its technical expertise.” BCCA Appeal Grp. v. EPA, 355 F.3d 817, 824 (5th Cir. 2003) (quoting Balt. Gas & Elec. Co. v. Nat. Res. Def. Council, Inc., 462 U.S. 87, 103 (1983)); see also Loper Bright, 603 U.S. at 392; Texas v. EPA (Texas II), 132 F.4th 808, 840 n.227 (5th Cir. 2025); Texas v. EPA (Texas I), 690 F.3d 670, 677 (5th Cir. 2012). In conducting APA review, courts must take “due account . . . of the rule of prejudicial error.” 5 U.S.C. § 706(2). And “[a]gency decisions are ‘presumptively valid; the [petitioner] bears the burden of showing otherwise.’” Barr, 114 F.4th at 446 (quoting Tex. Tech Physicians Assocs. v. HHS, 917 F.3d 837, 844 (5th Cir. 2019)). III. Before turning to the merits of the petition for review, we address Texas’s and the industry petitioners’ motion to strike several documents relating to the CSAPR Update from the administrative record. The motion is based on EPA’s statements in its SIP disapproval that the CSAPR Update was “outside the scope of” the disapproval and “irrelevant to the question of whether the Texas SIP should be disapproved.” See 81 Fed. Reg. at 53,284 n.2. The movants note that EPA said that it was “not rely[ing] upon the proposed CSAPR Update.” See id. at 53,286. Based on these descriptions of the CSAPR Update and the SIP disapproval as two separate actions, the movants contend that information relating to the CSAPR Update must be excluded from the record.

11 Case: 16-60670 Document: 262-1 Page: 12 Date Filed: 09/22/2025

APA review takes place on the “whole record.” 5 U.S.C. § 706(2). That record ordinarily includes “the decision of the agency” and “the evidence on which it was based.” Fed. Power Comm’n v. Transcon. Gas Pipe Line Corp., 423 U.S. 326, 331 (1976) (quoting United States v. Carlo Bianchi & Co., 373 U.S. 709, 714–15 (1963)); see also Fed. R. App. P. 16(a). “To review less than the full administrative record might allow a party to withhold evidence unfavorable to its case.” Walter O. Boswell Mem’l Hosp. v. Heckler, 749 F.2d 788, 792 (D.C. Cir. 1984). “To review more than the information before the [Administrator] at the time she made her decision risks our requiring administrators to be prescient or allowing them to take advantage of post hoc rationalizations.” Id. 1 Striking the documents would result in review of “less than the full administrative record.” See id. EPA included the documents on the docket for the SIP disapproval when it proposed the disapproval. Env’t Prot. Agency, Selected Records Pertaining to Notice of Data Availability for EPA’s Updated Ozone Transport Modeling, Docket No. EPA-R06-OAR- 2012-0985-0003 (Apr. 11, 2016). The full context for the fragments quoted by the movants makes clear that the SIP disapproval merely acknowledged the concurrent CSAPR Update rulemaking as a distinct proceeding: [T]he EPA technical information in the [Notice of Data Availability] and the proposed CSAPR Update accounted for the emission reductions resulting from controls listed in the SIP, implemented within the state, and nonetheless showed that Texas will contribute to downwind air quality problems. The CSAPR Update, however, is outside the scope of this _____________________ 1 Certain circumstances may invite supplementation. See Fed. R. App. P. 16(b); see also Animal Legal Def. Fund, Inc. v. Perdue, 872 F.3d 602, 611 (D.C. Cir. 2017); Env’t Def. Fund, Inc. v. Costle, 657 F.2d 275, 285 (D.C. Cir. 1981).

12 Case: 16-60670 Document: 262-1 Page: 13 Date Filed: 09/22/2025

action, and is irrelevant to the question of whether the Texas SIP should be disapproved.

81 Fed. Reg. at 53,284 n.2. In a reply-brief attempt to turn their motion to strike into a substantive dispute, the movants argue that EPA did rely on the FIP and that such reliance was statutorily impermissible. The premise for this argument, that EPA based its decision on the FIP, has no support in the record. As is clear from the above excerpt, EPA properly acknowledged the relevant factual development of a concurrent rulemaking and explained that any decision on the CSAPR Update was a separate action that would not be dispositive of the SIP disapproval. Even if the movants’ characterization of the record were accurate, it would not support striking the documents. “[T]he focal point for judicial review should be the administrative record already in existence, not some new record made initially in the reviewing court.” Camp v. Pitts, 411 U.S. 138, 142 (1973) (per curiam). But this does not mean that every time an agency relies on evidence that is not sufficiently tied to the relevant factors that it makes a “new record” that must be excised from judicial review. The remedy for that kind of error is judicial review of the agency’s reasoning, not post hoc censorship of the record. In any case, we observe no such error here. The motion to strike is DENIED. IV. We now turn to the merits of the petition, starting with the procedural challenges. Broadly speaking, Texas makes two procedural arguments. First, Texas says that EPA lost its power to disapprove the SIP by waiting for the Supreme Court’s clarifying decision in EME Homer. Second, Texas says that

13 Case: 16-60670 Document: 262-1 Page: 14 Date Filed: 09/22/2025

EPA should have given it additional notice of the proposed disapproval. Both challenges fail for lack of adequate statutory support. A. Title 42 U.S.C. § 7410(k)(2) requires EPA to act on a SIP submission “[w]ithin 12 months” of a determination that the submission is technically complete. Texas argues that EPA’s failure to meet this deadline is a procedural deficiency that renders the SIP disapproval void. 2 Texas states that “[t]here is nothing in the Act authorizing the EPA to delay indefinitely.” EPA responds that it retains the authority to act on a SIP after the deadline. EPA is correct. The Supreme Court has instructed that we should be “reluctant to conclude that every failure of an agency to observe a procedural requirement voids subsequent agency action, especially when important public rights are at stake.” Brock v. Pierce Cnty., 476 U.S. 253, 260 (1986). Accordingly, we generally do not “construe[] a provision that the

_____________________ 2 Texas appears to abandon the procedural aspect of this argument in its reply brief, attempting to restate it as an arbitrariness challenge. But the gist of the restated argument is that EPA did not reasonably adhere to the procedures that Texas called “mandatory and nondiscretionary” in its principal brief. To the extent that Texas is arguing that EPA’s action must be set aside because it ran afoul of the statutory timeline, we address that concern in this section; to the extent that Texas objects to EPA’s justifying its action with post-submission data (which has only an incidental connection to the statutory timeline), we address that issue in Section V.D. Another possible, if strained, reading of Texas’s arguments is that EPA did not adequately explain its delay. But EPA did not attempt to interpret the statute to give it discretion to adjust the deadline, and instead directly acknowledged that its action was late. 81 Fed. Reg. at 53,285. EPA provided analysis to the effect that having found the submission inadequate, EPA was required to disapprove it regardless of timing. See id. “Given [EPA]’s broad discretion as to how to deploy its limited resources, this is an adequate explanation.” Action for Children’s Television v. FCC, 999 F.2d 19, 22 (1st Cir. 1993). Moreover, once the action was late, EPA lacked discretion to make it timely, so it is not clear what justificatory role would have been served by additional explanation. See 5 U.S.C. § 706(2).

14 Case: 16-60670 Document: 262-1 Page: 15 Date Filed: 09/22/2025

Government ‘shall’ act within a specified time, without more, as a jurisdictional limit precluding action later.” Barnhart v. Peabody Coal Co., 537 U.S. 149, 158–59 (2003) (collecting cases). Thus, “if a statute does not specify a consequence for noncompliance with statutory timing provisions, the federal courts will not in the ordinary course impose their own coercive sanction.” Id. at 159 (quoting United States v. James Daniel Good Real Prop., 510 U.S. 43, 63 (1993)). “When . . . there are less drastic remedies available for failure to meet a statutory deadline, courts should not assume that Congress intended the agency to lose its power to act.” Brock, 476 U.S. at 260 (footnote omitted). Applying these principles, the Supreme Court has “decline[d] to infer” prescriptive limits on EPA authority under the CAA “[i]n the absence of a specific provision suggesting that Congress intended to create an enforcement bar.” Gen. Motors Corp. v. United States, 496 U.S. 530, 542 (1990). Considering EPA’s authority to enforce a SIP during the pendency of delayed action on a SIP revision, 3 General Motors found no express restrictions on that authority, noting that Congress had provided “other statutory remedies” for delay with suits to compel action under 42 U.S.C. § 7604(a)(2) and by allowing prejudice from delay to bear on appropriate means of enforcement under 42 U.S.C. § 7413(b). 496 U.S. at 541 & n.5.

_____________________ 3 General Motors addressed “whether EPA is barred from enforcing an existing SIP if the agency fails to take action on a proposed SIP revision within four months” under the deadline for approval and disapproval of SIPs then codified at 42 U.S.C. § 7410(a)(2). 496 U.S. at 536. The Court declined to read this limitation into approval of SIP revisions under § 7410(a)(3), noting that by its terms the deadline applied only to subsection (a)(2). Id. at 537. To be clear, the same cannot be said in this case, where there is a statutory timeline for action—which after the 1990 Amendments to the CAA is now found at 42 U.S.C. § 7410(k)(2) and labeled “Deadline for action.” But even if General Motors is not directly on point, its reasoning on a closely related issue is instructive.

15 Case: 16-60670 Document: 262-1 Page: 16 Date Filed: 09/22/2025

While Texas correctly points out that the CAA sets a deadline for SIP disapprovals, that is not enough to show that Congress intended EPA to lose its power to act, and the SIP to become affirmatively valid by operation of law, upon lapse of the deadline. Texas has pointed to little evidence to this effect other than the deadline itself 4—precisely the “without more” situation alluded to in Peabody Coal. Section 7410(k)(1)(B) explains that when EPA fails to make a technical completeness finding within “6 months after receipt of the submission,” the SIP is “deemed by operation of law to meet” the technical completeness criteria promulgated by EPA under § 7410(k)(1)(A). Nothing like this language appears in § 7410(k)(2), which establishes the deadline for EPA’s substantive review. On the contrary, as explained in General Motors, Congress has made other options available when EPA fails to act on a SIP. 5 Together with the absence of substantial countervailing evidence, these “less drastic remedies” lead us to conclude that Congress did not “intend[] the agency to lose its power to act.” Brock, 476 U.S. at 260. Finally, consider the “rule of prejudicial error.” See 5 U.S.C. § 706(2). 6 Texas proposed in its SIP that no additional emissions controls were necessary to meet its Good Neighbor obligations, and EPA’s delay allowed Texas to continue in precisely the state that it proposed. To the _____________________ 4 We are unpersuaded by Texas’s references to the provisions at § 7410(k)(4)–(6), which do not speak to the relevant timeline and are consistent with our interpretation. 5 Texas argues that forcing it to resort to the citizen-suit remedy in 42 U.S.C. § 7604(a)(2) is beneath its dignity as a “primary actor[] in the Act’s regulatory framework.” That is beside the point, which is that Congress provided for less drastic options than enabling SIPs to have the force and effect of federal law by prescription. 6 We carefully apply this rule to procedural rights when reviewing under 5 U.S.C. § 706(2)(D), taking full account of “the specific factual circumstances.” See United States v. Johnson, 632 F.3d 912, 930–32 (5th Cir. 2011) (quoting Shinseki v. Sanders, 556 U.S. 396, 412 (2009)).

16 Case: 16-60670 Document: 262-1 Page: 17 Date Filed: 09/22/2025

extent that delay happened to allow EPA to account for the Supreme Court’s clarification of the statutory requirements or improved data, Texas has not explained why we should consider it prejudicial for an agency’s decision to be legally sound or based on accurate evidentiary support. Instead, Texas argues primarily that “EPA’s delays . . . made it impossible for Texas to correct any deficiency before the FIP came into effect.” That notice objection targeted at the FIP rulemaking, to the extent it is properly brought at all in this petition for review of the SIP disapproval, fails for reasons similar to those set forth in the next section. B. Brushing aside EPA’s repeated—and not statutorily obligated— warnings that its emissions significantly impaired the air quality of downwind states, Texas asserts various challenges sounding in adequacy of notice. The principal objection is that EPA used “data and methodologies that were not available to Texas when it submitted its SIP.” In particular, Texas argues that EPA’s choice of 2017 as a target year for emissions modeling was “not disclosed to Texas at the time it submitted its SIP.” Nearly all these arguments are foreclosed by the established principle that a reviewing court “must determine whether the agency complied with the procedures mandated by the relevant statutes” but may not “impose upon the agency its own notion of which procedures are ‘best.’” Vt. Yankee Nuclear Power Corp. v. Nat. Res. Def. Council, Inc., 435 U.S. 519, 549 & n.21 (1978). There can be no reasonable dispute that EPA “complied with the procedures mandated by the relevant statutes.” See id. EPA provided notice and an opportunity to comment to Texas on the proposed disapproval as required by 5 U.S.C. § 553. See 81 Fed. Reg. at 21,290–91, 53,284. EPA, moreover, gave Texas ample opportunity beyond the statutory notice period to revise its SIP to conform to its statutory obligations. See ante, at 6–7.

17 Case: 16-60670 Document: 262-1 Page: 18 Date Filed: 09/22/2025

Texas—who as petitioner bears the burden to show error by EPA—did not attempt to take advantage of this extra year and a half of notice and has pointed to “nothing in the statute” that “places EPA under an obligation to provide specific metrics to States before they undertake to fulfill their good neighbor obligations.” See EME Homer, 572 U.S. at 510. Without that showing, we may not rewrite EPA’s statutory obligations to add on extratextual procedural requirements. 7 The only notice challenge not foreclosed by Vermont Yankee is Texas’s argument that EPA adopted a “new interpretation” of “nonattainment” without the notice required by 5 U.S.C. § 553. However, Texas’s string-cite allusion to the statute, without any meaningful explanation of why the legal interpretation at issue could not have been anticipated from the proposal, does not suffice to identify an APA violation. _____________________ 7 Texas and the industry petitioners further imply that it was procedurally improper or arbitrary for EPA to approve Texas’s SIP for technical completeness and then fail to disclose any concerns in advance of disapproving the SIP. However, the petitioners again fail to identify statutory support for a procedural challenge on this basis. And these facts do not support an arbitrariness challenge for failure to “display awareness that [the agency was] changing position,” see FCC v. Fox Television Stations, Inc., 556 U.S. 502, 515 (2009), although such a challenge would be waived for failure to argue it clearly. This is clear not only from the text of 42 U.S.C. § 7410(k)(1)(A), which describes the technical completeness criteria as “limited to the information necessary to enable the Administrator to determine whether the plan submission complies with the provisions of this chapter” (emphasis added), but also from the basic scheme of SIP review. “Under the two-stage procedure established in § 110(k), EPA first makes an essentially ministerial finding of completeness, a process taking at most six months. By contrast, the plan approval process may take up to twelve months due to the more extensive technical analyses necessary to ensure that the SIP meets the Act’s substantive requirements.” Nat. Res. Def. Council, Inc. v. Browner, 57 F.3d 1122, 1126 (D.C. Cir. 1995) (citing 42 U.S.C. § 7410(k)(1)–(3)). An inquiry into whether pages are missing or files are corrupted is obviously different from one into whether a SIP justifies its conclusions about the NAAQS with adequate data and analysis. There is nothing unusual, much less arbitrary, about the fact that EPA considered Texas’s submission complete but found it scientifically inadequate to implement the NAAQS.

18 Case: 16-60670 Document: 262-1 Page: 19 Date Filed: 09/22/2025

See Huawei Techs. USA, Inc. v. FCC, 2 F.4th 421, 447 (5th Cir. 2021). Indeed, EPA’s interpretation of the CAA was evident in its analysis of why Texas’s SIP did not meet the requirements of the Act. EPA explained that the SIP’s evaluation of “impact on the nearest designated nonattainment areas in other states without considering potential exceedances in other areas not designated nonattainment” would not “fully evaluate whether emissions from the state significantly contribute to nonattainment in other states.” 81 Fed. Reg. at 21,292. EPA further explained that the SIP did not “give the ‘interfere with maintenance’ clause of section 110(a)(2)(D)(i)(I) independent significance because its analysis did not attempt to evaluate the potential impact of Texas emissions on areas that are currently measuring clean data, but that may have issues maintaining that air quality.” Id. This analysis made clear that EPA was planning to consider “future nonattainment,” see id., and was sufficient to “‘adequately frame the subjects for discussion’ such that ‘[Texas] “should have anticipated” [EPA]’s final course in light of the initial notice.’” Huawei, 2 F.4th at 447 (quoting Nat’l Lifeline Ass’n v. FCC, 921 F.3d 1102, 1115 (D.C. Cir. 2019)). We conclude accordingly that Texas has not shown any violation of the APA’s procedural requirements. V. Texas additionally challenges EPA’s decision to disapprove the SIP on substantive grounds. As a threshold matter, Texas contends that “EPA d[id] not have discretion” to find its SIP scientifically unsound, but we follow our recent decision in Texas II in rejecting that argument. We find no reason to disturb EPA’s expert assessment that the SIP submission was inadequate to show compliance with the Good Neighbor Provision. EPA based this conclusion on the SIP’s failure to consider both downwind areas that had attained the ozone NAAQS but might go out of

19 Case: 16-60670 Document: 262-1 Page: 20 Date Filed: 09/22/2025

attainment, and areas not officially designated nonattainment but that might not achieve the NAAQS. EPA further supported its conclusions by means of reference to recent factual developments showing that Texas’s proposal would be inadequate to achieve the federal standards. We accordingly find that EPA based its decision on a rational consideration of the relevant factors. A. To start, Texas contends that EPA’s review of its submission should have been limited to “the justification presented by Texas.” Texas appeals to the “leeway” it is afforded under the CAA’s cooperative federalism scheme “to devise control measures to comply with the NAAQS.” Texas says it has “the primary responsibility for ensuring that the ambient air meets the NAAQS,” which comes with “broad authority to determine the methods and particular control strategies . . . to achieve the statutory requirements.” See BCCA Appeal Grp., 355 F.3d at 822. “Agencies have only those powers given to them by Congress.” West Virginia v. EPA, 597 U.S. 697, 723 (2022). Thus, in general—absent some delegation of authority of the kind that may be found at 42 U.S.C. § 7410(k)(4) 8 or § 7475(a)(8) 9 or § 7511c(c)(5) 10—Texas is correct that EPA lacks “authority to question the wisdom of a State’s choices of emission limitations.” Ohio v. EPA, 603 U.S. 279, 284 (2024) (quoting Train v. Nat. Res. Def. Council, Inc. (Train I), 421 U.S. 60, 79 (1975)). And that leaves “‘primary responsibility’ for developing compliance plans” with the states. Id. (quoting 42 U.S.C. § 7401(a)(3)).

_____________________ 8 See Nat. Res. Def. Council, Inc. v. EPA, 22 F.3d 1125, 1134 (D.C. Cir. 1994). 9 See Alaska Dep’t of Env’t Conservation, 540 U.S. at 473. 10 See Virginia v. EPA, 108 F.3d 1397, 1410 (D.C. Cir.), modified on reh’g, 116 F.3d 499 (D.C. Cir. 1997).

20 Case: 16-60670 Document: 262-1 Page: 21 Date Filed: 09/22/2025

But Texas’s important role in “select[ing]” a “mix of control devices” is not a license to submit a SIP that sets aside “the national standards.” See Union Elec. Co. v. EPA, 427 U.S. 246, 266 (1976) (pre-1990 Amendments), quoted in BCCA Appeal Grp., 355 F.3d at 822. Even though the states have “‘broad authority’” over “‘methods and particular control strategies,’” it is “[t]he Clean Air Act and the EPA” that “supply ‘the goals and basic requirements of state implementation plans.’” Texas I, 690 F.3d at 675 (quoting BCCA Appeal Grp., 355 F.3d at 822); see also, e.g., Hancock v. Train (Train II), 426 U.S. 167, 170–71 (1976) (describing the 1970 Amendments as “increasing federal authority and ‘taking a stick to the States’ by requiring them to implement the federal standards promulgated pursuant to that authority” (quoting Train I, 421 U.S. at 64)); Oklahoma v. EPA, 145 S. Ct. 1720, 1728 (2025) (“The CAA makes the SIP process one of federal-state collaboration.”). 11 That is why Texas’s primary authority, BCCA Appeal Group, explained that “Congress’s enactment of the CAA delegated authority to the EPA to review SIPs for their compliance with the statute and EPA’s implementing regulations.” 355 F.3d at 825; see also Texas II, 132 F.4th at 833 (“EPA is . . . charged from the start with independently assessing whether a SIP complies with the Good Neighbor Provision.”); Luminant Generation Co., 714 F.3d at 846 (explaining that the CAA requires EPA to “review[] SIPs for consistency with the Act’s requirements” 12).

_____________________ 11 Luminant Generation Co. v. EPA, 714 F.3d 841 (5th Cir. 2013), stated broadly that “[o]nly the states enjoy discretion in implementing the dictates of the CAA.” Id. at 928. But Luminant was addressing a choice-of-control-measures issue, and the term “implementing” is key. As BCCA Appeal Group explained, the states have discretion to select “the methods and particular control strategies they will use to achieve the statutory requirements”—but “EPA determines the standards of air quality.” 355 F.3d at 822. 12 Luminant described this task as “ministerial,” a term upon which Texas places undue weight. SIP review is ministerial because EPA “shall” approve plans that meet the

21 Case: 16-60670 Document: 262-1 Page: 22 Date Filed: 09/22/2025

This is clear from the plain text of the statute. “The text puts EPA in the driver’s seat for evaluating a SIP’s compliance with the CAA.” Texas II, 132 F.4th at 833. Title 42 U.S.C. § 7410(k)(3) states that the EPA Administrator “shall approve” a SIP “if it meets all of the applicable requirements of this chapter.” It is impossible to square that language—not to mention the references to EPA’s ability to disapprove SIPs that can be found in the title of the subsection, in the following subsection, and at 42 U.S.C. § 7410(c)(1)(B)—with Texas’s position that EPA should have “approve[d]” even a SIP that plainly did not “meet[] all the applicable requirements.” Instead, when a state’s SIP is not “adequate for compliance with the [Good Neighbor] provision,” this language means that EPA is “called upon to act.” EME Homer, 572 U.S. at 514 n.15. “This reading is confirmed by other CAA provisions.” Texas II, 132 F.4th at 833. EPA may approve a plan revision to avoid a FIP when a state “corrects” its “deficiency.” 42 U.S.C. § 7410(c)(1). FIPs likewise correct “gap[s]” and “inadequac[ies]” in SIPs. Id. § 7602(y). And, most tellingly, EPA may require revision of a SIP when it “finds” the SIP “substantially inadequate” to comply with the state’s responsibilities, id. § 7410(k)(5), and EPA may not approve a revision that interferes with any applicable statutory requirements, id. § 7410(l). It would make “nonsense” out of the statute to interpret it to say that EPA must “defer to states during the SIP-approval process, approving a Good-Neighbor submission if it is reasoned but (in EPA’s view) substantially inadequate—and then . . . turn around and call for a SIP revision after an independent analysis.” Texas II, 132 F.4th at 833 (quoting Ali v. Barr, 951 F.3d 275, 280 (5th Cir. 2020)).

_____________________ CAA requirements. See 714 F.3d at 846 (citing 42 U.S.C. § 7410(k)(3)). That is far from saying that EPA must approve a plan even if it does not meet the requirements.

22 Case: 16-60670 Document: 262-1 Page: 23 Date Filed: 09/22/2025

Now consider the overall scheme of SIP review. A separate, more thorough phase of review follows a technical completeness determination. See 42 U.S.C. § 7410(k). “Under the two-stage procedure established in § 110(k), EPA first makes an essentially ministerial finding of completeness, a process taking at most six months. By contrast, the plan approval process may take up to twelve months due to the more extensive technical analyses necessary to ensure that the SIP meets the Act’s substantive requirements.” Browner, 57 F.3d at 1126. We would have to strain to interpret the two separate stages of review, and Congress’s selection of different timelines for each stage, as identically directed to the same clerical or perfunctory tasks. The statutory text and scheme prevent us from concluding that the CAA delegates unreviewable discretion to the states to bind EPA and thereby this court to creative misreadings of the statutory requirements. See Loper Bright, 603 U.S. at 395. Texas asks us to say that the “all the applicable requirements” are not “requirements,” that not “all” of them need be satisfied, and that they “appl[y]” only in a hortatory sense. See 42 U.S.C. § 7410(k)(3). Indeed, its position seems liable to leave the states’ (likely widely varying) applications of the CAA in their SIPs without any APA review at all. Texas has not adequately justified that departure from “the traditional understanding of the judicial function.” See Loper Bright, 603 U.S. at 394. Even where the CAA delegates discretion to the states much more explicitly, the Supreme Court has cautioned against such a lopsided approach. Alaska Department of Environmental Conservation sustained EPA oversight of a permitting decision by Alaska under its Prevention of Significant Deterioration (PSD) SIP. 540 U.S. at 495. Review of such permitting decisions entails close federal supervision over the states’ control measures—the heartland of the states’ ordinary ambit of discretion under the CAA—despite the express discretion to states afforded by 42 U.S.C. § 7479(3),

23 Case: 16-60670 Document: 262-1 Page: 24 Date Filed: 09/22/2025

nothing of which kind is present here. By contrast, this case involves “compliance with the statute and EPA’s implementing regulations” even more directly. See BCCA Appeal Grp., 355 F.3d at 825. Extending Chevron-like deference to the states would be particularly illogical here, because the entire point of the Good Neighbor Provision is to address interstate externalities through a federal statute. As EME Homer explained in the FIP context, balancing Good Neighbor obligations involves “allocat[ing] among multiple contributing upwind States responsibility for a downwind State’s excess pollution.” 572 U.S. at 514. To the extent that the statute leaves room to address that “thorny causation problem,” id., 13 according the states unreviewable primacy to implement the Good Neighbor Provision could result in disorganized and conflicting approaches, allowing “nonattainment” and “interfere[nce] with maintenance” to continue unabated. See 42 U.S.C. § 7410(a)(2)(D). To see this, consider the following example from the Supreme Court’s decision in EME Homer: Suppose [EPA] sets a NAAQS, with respect to a particular pollutant, at 100 parts per billion (ppb), and that the level of the pollutant in the atmosphere of downwind State A is 130 ppb. Suppose further that EPA has determined that each of three _____________________ 13 EME Homer instructed, before Loper Bright, that “[t]he Good Neighbor Provision does not answer th[e] question for EPA” of how to allocate significant contributions. EME Homer, 572 U.S. at 515. Loper Bright appears to confirm that terms like “significant” can signal that an agency has been “le[ft] . . . with flexibility.” See Loper Bright, 603 U.S. at 395 (quoting Michigan, 576 U.S. at 752) (discussing terms “such as ‘appropriate’ or ‘reasonable’”); see also White Stallion Energy Ctr., LLC v. EPA, 748 F.3d 1222, 1266 (D.C. Cir. 2014) (Kavanaugh, J., concurring in part and dissenting in part), rev’d sub nom. Michigan v. EPA, 576 U.S. 743. Regardless, all we are concerned with is whether a state’s initial attempt to evaluate its Good Neighbor obligations precludes any meaningful review by EPA, regardless of how “flexib[le]” that review might be.

24 Case: 16-60670 Document: 262-1 Page: 25 Date Filed: 09/22/2025

upwind States—X, Y, and Z—contributes the equivalent of 30 ppb of the relevant pollutant to State A’s airspace. . . . How is EPA to divide responsibility among the three States? Should the Agency allocate reductions proportionally (10 ppb each), on a per capita basis, on the basis of the cost of abatement, or by some other metric?

EME Homer, 572 U.S. at 514–15 (citation omitted). As Texas interprets the statute, each state may figure its own reductions using a different—and critically, unreviewable—method, so that, for example, high-pollution State X allocates reductions proportionally and sets itself a 10 ppb reduction, low- density State Y allocates by population and sets a 5 ppb reduction, and expensive State Z apportions by cost and sets an 8 ppb reduction. But then fewer than 30 ppb would be reduced overall—allowing State A’s nonattainment to continue, in plain violation of the statutory language that a SIP must be “adequate” to “prohibit[]” emissions that impede attainment of the NAAQS in “any other State.” See 42 U.S.C. § 7410(a)(2)(D)(i). The computational complexity and resulting coordination problems escalate dramatically when it is considered that “[m]ost upwind States contribute pollution to multiple downwind States in varying amounts” (whether figured by total ppb or another method). EME Homer, 572 U.S. at 516. Suppose now that “State Y contributes five times the amount of pollution to State A than does State X,” but that “States X and Y also contribute pollutants to a second downwind State (State B), this time in a ratio of seven to one. Though State Y contributed a relatively larger share of pollution to State A, with respect to State B, State X is the greater offender.” Id. In that case, “[p]roportionality as to one downwind State will not achieve proportionality as to others. Quite the opposite. And where, as is generally

25 Case: 16-60670 Document: 262-1 Page: 26 Date Filed: 09/22/2025

true, upwind States contribute pollution to more than two downwind receptors, proportionality becomes all the more elusive.” Id. By leaving downwind states without any assurance that this complexity will be resolved in a way that abates the required amount of pollution, Texas’s unworkable proposal would fail to give effect to the CAA’s plain text. Indeed, there would be no reason for upwind states to accept any “responsibility” at all for “excess pollution” downwind, see id. at 514— which is exactly what happened here. We conclude that giving upwind states unreviewable discretion to effectively define what kind of a contribution is significant or what constitutes interference to their downwind neighbors would be “plainly contrary to the intent of Congress.” See Solid Waste Agency of N. Cook Cnty. v. U.S. Army Corps of Eng’rs, 531 U.S. 159, 173 (2001) (quoting Edward J. DeBartolo Corp. v. Fla. Gulf Coast Building & Constr. Trades Council, 485 U.S. 568, 575 (1988)). We cannot “destroy” the statutory scheme by interpreting the states’ implementation role to “erase the clear mandate” of the Good Neighbor Provision. Cf. Atl. Richfield Co. v. Christian, 590 U.S. 1, 23 (2020) (“Interpreting [CERCLA]’s saving clauses to erase the clear mandate of § 122(e)(6) would allow the Act ‘to destroy itself.’” (quoting Am. Tel. & Tel. Co. v. Cent. Off. Tel., Inc., 524 U.S. 214, 228 (1998))). We agree with Texas II and conclude that EPA was permitted to disapprove the SIP on the basis that the SIP failed to demonstrate adequate compliance with the statutory requirements. B. We now turn to the merits of Texas’s substantive challenges. The first challenge we address involves the basis for disapproval that the SIP “limit[ed] its discussion of data only to areas designated nonattainment in states that are geographically closest to Texas.” 81 Fed. Reg. at 21,292.

26 Case: 16-60670 Document: 262-1 Page: 27 Date Filed: 09/22/2025

Because the SIP’s “analysis did not attempt to evaluate the potential impact of Texas emissions on areas that are currently measuring clean data, but that may have issues maintaining that air quality,” EPA said the SIP failed to address whether Texas’s emissions would “interfere with maintenance” of the ozone standards in any other state. Id. at 21,291–92 (quoting 42 U.S.C. § 7410(a)(2)(D)(i)(I)). Texas argues that this determination was arbitrary or an abuse of EPA’s discretion. i. As an initial matter, we find no error in EPA’s interpretation of the statute. The Good Neighbor Provision “prohibit[s]” emissions that “significantly contribute to nonattainment . . . or interfere with maintenance” in any other State. 42 U.S.C. § 7410(a)(2)(D)(i)(I) (emphasis added). The provision “is written in the disjunctive.” North Carolina, 531 F.3d at 910. “Canons of construction ordinarily suggest that terms connected by a disjunctive be given separate meanings, unless the context dictates otherwise.” Id. (quoting Reiter v. Sonotone Corp., 442 U.S. 330, 339 (1979)). EPA therefore correctly “g[a]ve independent significance to the ‘interfere with maintenance’ language” in the Good Neighbor Provision. Id. ii. Texas charges that the disapproval was arbitrary nonetheless. Texas argues that the SIP “conducted a comprehensive analysis of wind, ozone- trending data and monitoring data in both attainment and nonattainment areas to conclude” that Texas emissions did not interfere with maintenance of the 2008 ozone NAAQS in any other state. EPA’s conclusion that “Texas’s . . . modeling” was insufficient to adequately address lack of interference with maintenance was a judgment “based upon [EPA’s] evaluation of complex scientific data within its technical expertise.” BCCA Appeal Grp., 355 F.3d at 834. This court reviews that judgment for a rational

27 Case: 16-60670 Document: 262-1 Page: 28 Date Filed: 09/22/2025

connection to the relevant facts available to EPA and does not substitute its judgment for the agency’s. See, e.g., Texas II, 132 F.4th at 846; accord, e.g., Midwest Ozone Grp., 61 F.4th at 192. EPA’s decision was far from irrational. EPA accurately stated that “nothing in Texas’ SIP submittal indicates that it performed any analysis to support its conclusion” that Texas was not interfering with downwind maintenance “as the State limited its discussion of data only to certain areas designated nonattainment and did not consider whether those or any other areas might have trouble maintaining the standard even if they measured clean data.” 81 Fed. Reg. at 53,288. The SIP’s analysis of wind data described typical wind patterns in the Dallas and Houston areas during the ozone season, and concluded that “very few winds are observed . . . which would be anticipated to transport ozone to Memphis and Baton Rouge.” The record discloses that those are all nonattainment areas. The SIP similarly described trends in ozone statistics in Dallas, Houston, and a handful of neighboring nonattainment areas. As for “monitoring data,” the SIP provided a visualization and list of monitors and 2010 statistics, with no analysis other than to note that the monitors between Texas and Baton Rouge showed attainment, consistent with the hypothesis that “local emissions contribute to [nonattainment] areas’ nonattainment status,” and that “there are sources of ozone precursors” outside of Texas. Apart from these unconnected data points and a conclusory assertion that Texas’s emissions did not interfere with downwind maintenance, the SIP made no attempt to analyze, much less quantify, the relationship between Texas’s emissions and maintenance of the ozone standards outside nonattainment areas. To the contrary, the SIP concluded that “it is difficult to determine how much ozone in other areas would be due to transport.” On appeal, Texas argues that because the trend data in nonattainment areas showed a decrease in ozone levels, a general decrease in ozone problems

28 Case: 16-60670 Document: 262-1 Page: 29 Date Filed: 09/22/2025

might imply fewer issues for maintenance areas. Texas additionally infers from the fact that “[a]ll of the monitors between Texas and the Memphis and Baton Rouge nonattainment areas demonstrated attainment of the 2008 Ozone NAAQS” that Texas’s emissions must not interfere with maintenance generally. 14 Again, these arguments must surmount this court’s “‘most deferential’” review of EPA’s “evaluation of complex scientific data within its technical expertise.” BCCA Appeal Grp., 355 F.3d at 834 (quoting Balt. Gas, 462 U.S. at 103). Attorney argument from logical fallacy, unsupported by scientific analysis, does not carry that burden. Without appropriate support, Texas’s arguments cannot justify after-the-fact extrapolations from nonattainment areas to other areas or from the geographically narrow set of air monitors considered to maintenance areas generally (when the SIP’s own wind analysis indicated those monitors were not even generally downwind). In any case, these arguments were not included in Texas’s SIP or comments so have not been properly raised on review. See id. at 828–29. Texas further contends that EPA itself conflated maintenance and nonattainment. Texas points out that, in assessing which air quality monitoring sites were relevant to interference with maintenance, EPA defined “maintenance receptors as those monitoring sites that have measured ozone concentrations that meet the NAAQS (clean data) based on monitoring data from years 2012-2014 and are projected to exceed the _____________________ 14 Texas further claims that PSD standards can address nonattainment concerns in attainment areas. Texas argues that “downwind states have their own controls for any new sources that might contribute to air quality concerns in an attainment/maintenance area.” We see no need to address this argument at any length because this reasoning would read “interfere with maintenance” out of the statute. It would allow upwind states to send their pollution downwind and force their neighbors to bear not only the public health and welfare costs of that pollution but also the full weight of any additional burdens on emissions sources required under PSD standards as a result.

29 Case: 16-60670 Document: 262-1 Page: 30 Date Filed: 09/22/2025

NAAQS in 2017 based on a maximum or average design value.” 81 Fed. Reg. at 53,285. This, Texas says, is too similar to EPA’s definition of “nonattainment receptors” as “those monitoring sites that (1) measured ozone concentrations that exceed the NAAQS based on monitoring data from years 2012-2014, and (2) are projected to exceed the NAAQS in 2017 based on an average design value.” Id. at 53,284. This objection relates to EPA’s CSAPR Update analysis and does not directly address Texas’s failure to give independent significance to maintenance. In any case, EPA’s definitions were rationally based on the distinction between maintenance and nonattainment drawn by the text of the statute. The difference between “meet” and “exceed” is far from “scant,” as Texas proposes, but instead tracks one of the central distinctions in the statutory architecture. See, e.g., Richard J. Lazarus, The Making of Environmental Law 301 (2d ed. 2023) (describing how “Congress had divided the regulatory world into ‘attainment’ and ‘nonattainment’ areas”). And EPA justified its choice of maintenance receptors because “the monitoring sites of the proposed maintenance receptors currently meeting the NAAQS could be subject to conditions that may allow violations to reoccur and therefore may have future maintenance concerns.” 81 Fed. Reg. at 53,285. That justification was rationally based on relevant factors well within EPA’s expertise. We conclude that EPA rationally rejected Texas’s SIP on the ground that the SIP did not adequately demonstrate that Texas’s emissions did not “interfere with maintenance” in any other state. C. Although EPA would have been required to disapprove the SIP on this first basis alone, Texas challenges the disapproval on additional grounds. Texas contends that EPA should not have disapproved Texas’s SIP on the

30 Case: 16-60670 Document: 262-1 Page: 31 Date Filed: 09/22/2025

basis that the SIP addressed only a limited subset of nonattainment areas. Because the SIP “did not consider other areas that were not formally designated as nonattainment,” EPA concluded that the SIP did not adequately address whether Texas’s emissions significantly contributed to downwind nonattainment in any other state. Id. at 53,285. Texas challenges EPA’s interpretation of the statutory term “nonattainment” and objects to EPA’s reasoning as arbitrary or capricious. i. Texas argues that “nonattainment” refers to nonattainment areas designated under 42 U.S.C. § 7407(d)(1)(A)(i), see id. § 7501(2), as opposed to the state of failing to achieve the NAAQS. EPA responds that it “has routinely interpreted the obligation to prohibit emissions that ‘significantly contribute to nonattainment’ of the NAAQS in downwind states to be independent of formal designations because exceedances can happen in any area,” a practice it followed in both CAIR and CSAPR. 81 Fed. Reg. at 53,287 & n.6. Texas has not carried its burden to show error in EPA’s interpretation. The plain meaning of “nonattainment . . . with respect to” the NAAQS is that the NAAQS are not attained. That happens when pollutants “contribute to” air quality problems. Cf. 42 U.S.C. § 7511b(d) (referring to “air pollutants which contribute to nonattainment of the national ambient air quality standards for ozone”). Texas’s proposed interpretation fails to make sense of the provision as a whole and loads the procedural requirements of § 7407 into the provision without adequate justification. Other provisions of the statute support EPA’s reading. When the CAA refers to nonattainment areas, it often does so in explicit terms, and those categorical references often establish the scope of a program. See, e.g.,

31 Case: 16-60670 Document: 262-1 Page: 32 Date Filed: 09/22/2025

id. § 7545(k)(1)(A). By contrast, when “attain” is used in isolation it more often describes a state of events. See, e.g., id. § 7511(b)(2)(A) (“Within 6 months following the applicable attainment date . . . for an ozone nonattainment area, the Administrator shall determine . . . whether the area attained the standard by that date.”). That is why multiple provisions hinge on whether a designated area actually “attain[s]” the NAAQS or not. See, e.g., id. §§ 7511(b)(2)(A), 7511d(a). In such provisions the factual condition of attainment is separate from the regulatory condition of the area to which the regulation applies. When the Good Neighbor Provision describes nonattainment in any other state, it describes with the term “nonattainment” a factual condition, and the reference to other states is what designates the geographic scope within which that factual condition must obtain. See, e.g., id. § 7502(a)(1)(A) (“In determining the appropriate classification, if any, for a nonattainment area, the Administrator may consider such factors as the severity of nonattainment in such area . . . .”). It strains the other language in the provision to interpret it as requiring EPA to consider whether a state “contributes to nonattainment areas.” Moreover, such a reading would have to be squared with the express mention of the NAAQS. See id. § 7410(a)(2)(D)(i)(I) (“with respect to any such national primary or secondary ambient air quality standard”). Nonattainment areas are designated with respect to a particular standard that the area or an adjacent area does not meet. Id. § 7407(d)(1)(A)(i). If “nonattainment” means “nonattainment areas,” we are unsure why Congress would not have simply assumed that the term “nonattainment” already incorporated the NAAQS by reference. See, e.g., id. § 7545(k)(1)(A). Texas objects that interpreting the statute to refer to future nonattainment in “unknown (and unknowable)” areas makes its requirements “unachievable, and unlawful,” because the validity of any SIP rests on the shifting sands of whatever areas EPA will choose to target in the

32 Case: 16-60670 Document: 262-1 Page: 33 Date Filed: 09/22/2025

future. But it is black letter administrative law that Texas cannot attack EPA’s “case-by-case evolution of statutory standards,” see NLRB v. Bell Aerospace Co. Div. of Textron, 416 U.S. 267, 293 (1974) (quoting SEC v. Chenery Corp., 332 U.S. 194, 203 (1947))—an objection that is particularly unsustainable as applied to a rule put through notice and comment, see id. at 292. At any rate, the task for Texas is that its SIP must not frustrate the NAAQS of which it was informed in 2008, see EME Homer, 572 U.S. at 510— a standard that is known, knowable, theoretically achievable, and lawful, see Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 474 (2001). The statutory language plainly frames Good Neighbor requirements in the future tense. Besides the term “will,” the terms “prohibit[]” and “contribute” orient the provision toward the future too. Substantial contributions to nonattainment are not “prohibit[ed]” when an area, simply because it had better air quality in the past, is permitted not to achieve the NAAQS. North Carolina and Wisconsin confirmed this in requiring EPA to consider the nonattainment designation schedule when promulgating Good Neighbor budgets. See Wisconsin, 938 F.3d at 322 (“Given the use of the future tense, it would be anomalous for EPA to subject upwind States to good neighbor obligations in 2017 by considering which downwind States were once in nonattainment in 2011.”); see also North Carolina, 531 F.3d at 912. To attain this lawful standard in the future, Congress has demanded that SIPs must ensure that a state’s emissions will not unduly impose upon the air quality of “any other State.” See 42 U.S.C. § 7410(a)(2)(D)(i)(I). Such comprehensive analysis, too, is obviously achievable, because EPA completed the task to the satisfaction of the Supreme Court in EME Homer. See 572 U.S. at 524. Yet Texas did not even attempt to fulfill the statutory requirements by explaining why the set of monitors it considered within a narrow geographic region should be treated as representative of areas threatened with nonattainment in “any other State.” Instead, Texas

33 Case: 16-60670 Document: 262-1 Page: 34 Date Filed: 09/22/2025

concluded that “it is difficult to determine how much ozone in other areas would be due to transport.” But “practical difficulties . . . do not justify departure from the Act’s plain text.” Id. at 509. Finally, according to EPA, adopting a past-tense view of nonattainment would frustrate the effective administration of the statute, signaling inconsistency with the statutory design. “[A]lthough an agency’s interpretation of a statute ‘cannot bind a court,’ it may be especially informative ‘to the extent it rests on factual premises within [the agency’s] expertise.’” Loper Bright, 603 U.S. at 402 (quoting ATF v. FLRA, 464 U.S. 89, 98 n.8 (1983)). EPA has explained that “transported emissions may cause an area to measure exceedances of the standard even if that area is not formally designated nonattainment by the EPA.” 81 Fed. Reg. at 21,292. Because this rationale hinges on scientific knowledge about air pollution patterns and the practicalities of CAA enforcement, we find informative EPA’s consistent practice in interpreting the statute’s legal requirements and its expertise in the factual predicates of administration in this scientifically complex area. See, e.g., id.; 70 Fed. Reg. at 25,266. Accordingly, we lack reason to disturb EPA’s reasoning based on its interpretation of the statute. ii. Even conceding this statutory framework, however, Texas responds that “there is no limitation restricting when the EPA can set the year” that the state must target in its analysis. Texas thus says EPA acted arbitrarily in “setting 2017 as the year that Texas should have used to determine whether air quality in another state is projected to exceed the NAAQS.” We disagree. “EPA is allowed substantial discretion in its assessment of what constitutes an approvable SIP.” BCCA Appeal Grp., 355 F.3d at 846. The “realities of interstate air pollution,” EME Homer, 572 U.S. at 516, in

34 Case: 16-60670 Document: 262-1 Page: 35 Date Filed: 09/22/2025

particular, call for “flexibility,” Wisconsin, 938 F.3d at 320. EPA explained that it was using 2017 to facilitate the 2018 attainment designation schedule. 81 Fed. Reg. at 21,293 (referencing 80 Fed. Reg. at 75,708). Texas has not explained why EPA was not permitted to pursue the effective administration of its statute by considering “protection to downwind states projected to be in nonattainment”—and the potential for those states to be forced to attain the NAAQS “without the elimination of upwind states’ significant contribution.” See North Carolina, 531 F.3d at 912. That conclusion would require us to split from the D.C. Circuit, which held in North Carolina and affirmed in Wisconsin that the “relationship between the Good Neighbor Provision’s obligations for upwind States and the statutory attainment deadlines for downwind areas . . . generally calls for parallel timeframes.” See Wisconsin, 938 F.3d at 316. Regardless, EPA did not disapprove the SIP because the SIP did not use a model with the target year 2017. EPA disapproved the SIP because the SIP was based on incomplete data and unsupported by adequate analysis of the statutory factors. EPA’s selection of a target year represented the exercise of its discretion to select a methodology to support its reasoning with additional reliable evidence. We cannot say that decision was unreasonable. iii. Texas maintains that its SIP analysis was adequate to show that Texas’s emissions did not significantly contribute to downwind nonattainment in any other state. Texas acknowledges that it “did not analyze monitoring data for every state in the country,” but states that its choices were reasonably informed by available data and by the fact that the

35 Case: 16-60670 Document: 262-1 Page: 36 Date Filed: 09/22/2025

monitors in between Dallas and Houston, on the one hand, and Memphis and Baton Rouge, on the other, demonstrated attainment. 15 But EPA explained that the SIP did not adequately address the statutory requirements under any reasonable standard. EPA rejected the SIP’s unanalyzed monitor data and limited analysis of a small set of formally designated nonattainment areas as insufficiently comprehensive to address the statutory requirement to eliminate significant contribution to nonattainment in any other state. EPA explained that the SIP’s analysis was not sufficiently tied to the statute’s requirement to prohibit nonattainment because “transported emissions may cause an area to measure exceedances of the standard even if that area is not formally designated nonattainment by the EPA.” 81 Fed. Reg. at 21,292. Moreover, EPA stated, “[a]nalysis of wind patterns, emissions data, and ambient monitoring data as provided in the Texas SIP submittal does not quantify the magnitude of impact from Texas emissions to downwind states.” Id. at 21,294. Instead, EPA pointed out, these data either concerned transport or air quality in isolation, without connecting the two to address Texas’s contributions to air quality in other states. Id. EPA further explained that the SIP’s methodology was facially unreliable and incomplete because the wind patterns detailed in the SIP were consistent with transport to other statutorily relevant areas besides the ones considered, and because the downward ozone trends described by the SIP provided no information about either “the magnitude of the remaining impact or the potential benefit from additional emission reductions,” as necessary to assess the significance of Texas’s contributions. Id. These

_____________________ 15 Texas states that it is significant that these monitors showed attainment despite wind analysis indicating that ozone precursors were expected to drift toward Memphis and Baton Rouge. As EPA points out, the SIP says the opposite regarding the direction of the wind. Either way, this extra analysis was not in Texas’s SIP.

36 Case: 16-60670 Document: 262-1 Page: 37 Date Filed: 09/22/2025

grounds for the disapproval were rationally based on a consideration of the permissible factors. We conclude that EPA reasonably disapproved Texas’s SIP on the basis that the SIP did not adequately demonstrate that Texas’s emissions did not significantly contribute to nonattainment in any other state. D. Texas’s final set of challenges concerns EPA’s use of corroborating evidence to provide additional support for its conclusions. EPA noted that its own modeling, which had been published with the proposed CSAPR Update in 2015, showed Texas contributing to downwind ozone problems. Id. at 21,294. EPA further noted that since the CAIR trading program had ended, Texas could not rely on it to demonstrate compliance with its Good Neighbor obligations. Id. Texas argues that it was arbitrary or capricious for EPA to account for these developments. i. Texas first says that the disapproval was arbitrary because “EPA must act on a SIP revision . . . based on information and rules in effect at the time of submission.” Texas urges that “there was no way [the SIP] could have incorporated data and standards not in existence in 2012.” This challenge fails at the threshold because EPA described the previously discussed problems with Texas’s submission as statutory deficiencies. See 5 U.S.C. § 706(2). “[W]hether or not the EPA had proposed the CSAPR Update,” EPA explained, “Texas’ SIP submittal failed to include an analysis that appropriately evaluated the impact of state emissions on areas in other states, regardless of current nonattainment designations and considering the ability of areas currently measuring clean data to maintain that standard.” 81 Fed. Reg. at 53,286. EPA acknowledged

37 Case: 16-60670 Document: 262-1 Page: 38 Date Filed: 09/22/2025

that “prompt action on state SIP submittals can be beneficial to the states’ planning efforts” 16 but followed North Carolina, 531 F.3d 896, and properly determined that the SIP failed to demonstrate compliance with the statutory requirements. See 81 Fed. Reg. at 53,285. Once EPA made that determination, EPA was unable to approve the SIP in any event. See 42 U.S.C. § 7410(k)(3). As for the challenge itself, Texas has not shown that EPA relied on statutorily impermissible factors in making its decision. The question posed by the statute is whether the SIP “meets all of the applicable requirements” when EPA is deciding whether to “approve such submittal.” Id. (emphasis added). That contrasts with, for example, the finding of substantial inadequacy that triggers a SIP call,17 which turns on “the requirements of this chapter to which the State was subject when it developed and submitted the plan for which such finding was made.” Id. § 7410(k)(5) (emphasis added). 18 _____________________ 16 To this effect, however, we note that EPA made extensive efforts to support Texas’s planning. Starting over a year and a half before the action at issue, EPA gave Texas repeated opportunities to update its SIP before finally extending notice of the anticipated disapproval. See ante, at 6–7. That provided a “way Texas could have incorporated data and standards not in existence in 2012”—affording Texas more time than the year Texas argues EPA had for its review under § 7410(k)(2)—and Texas did not take advantage of it. Texas has not explained how, in these circumstances, EPA’s use of more accurate data— in other words, a better-supported decision—can amount to “prejudicial error” cognizable under 5 U.S.C. § 706(2). See PDK Lab’ys Inc. v. DEA, 362 F.3d 786, 809 (D.C. Cir. 2004) (Roberts, J., concurring in part and concurring in the judgment) (“Chenery does not require that we convert judicial review of agency action into a ping-pong game.” (quoting Time, Inc. v. U.S. Postal Serv., 667 F.2d 329, 335 (2d Cir. 1981) (Friendly, J.))). 17 “Whenever the Administrator finds that the applicable implementation plan for any area is substantially inadequate to . . . comply with any requirement of this chapter, the Administrator shall require the State to revise the plan as necessary to correct such inadequacies.” 42 U.S.C. § 7410(k)(5). 18 It is not even clear that this language precludes consideration of later factual developments. But it at least illustrates the kind of language Congress used when requiring EPA to look backward in its decision-making.

38 Case: 16-60670 Document: 262-1 Page: 39 Date Filed: 09/22/2025

Texas has not identified “[any]thing in the statute” that “places EPA under an obligation to provide specific metrics to States before they undertake to fulfill their good neighbor obligations.” EME Homer, 572 U.S. at 510. We find no reason to fault EPA for considering recent developments in order to make an accurate and informed decision. We would expect a clearer signal if Congress wanted a different result. Under the usual principles of rational agency decision-making, an agency’s decision should be based on “the relevant data.” Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983). The general consequence is that “an agency cannot ignore new and better data.” Dist. Hosp. Partners, L.P. v. Burwell, 786 F.3d 46, 57 (D.C. Cir. 2015); see also Texas II, 132 F.4th at 862 n.466 (attributing to Wisconsin, 938 F.3d at 321–22, the proposition that “EPA properly considered data and modeling developed after the SIP submission deadline”). Indeed, it would thwart the very premise of notice-and-comment rulemaking to prohibit an agency from taking account of new information brought to its attention. See Perez v. Mortg. Bankers Ass’n, 575 U.S. 92, 96 (2015). Yet, without any support in the statute, that is what the petitioners propose. To be sure, an agency is “not obliged to stop the entire process because a new piece of evidence emerge[s]. If this were true then the administrative process could never be completed. An agency does, however, have an obligation to deal with newly acquired evidence in some reasonable fashion.” Am. Iron & Steel Inst. v. EPA, 115 F.3d 979, 1007 (D.C. Cir. 1997); see also Catawba Cnty. v. EPA, 571 F.3d 20, 45 (D.C. Cir. 2009); Sierra Club v. EPA, 356 F.3d 296, 308 (D.C. Cir. 2004). “[W]hen an agency acknowledges that its data are either outdated or inaccurate, it should, at the very least, analyze the new data or explain why it nevertheless chose to rely on the older data.” Dow AgroSciences LLC v. Nat’l Marine Fisheries Serv., 707 F.3d 462, 473 (4th Cir. 2013); accord Zen Magnets, LLC v. Consumer Prod.

39 Case: 16-60670 Document: 262-1 Page: 40 Date Filed: 09/22/2025

Safety Comm’n, 841 F.3d 1141, 1149 (10th Cir. 2016) (“In general, where there is a known and significant change or trend in the data underlying an agency decision, the agency must either take that change or trend into account, or explain why it relied solely on data pre-dating that change or trend.”); Sierra Club v. EPA, 671 F.3d 955, 966–68 (9th Cir. 2012) (“[I]f new information indicates to EPA that an existing SIP or SIP awaiting approval is inaccurate or not current, then, viewing air quality and scope of emissions with public interest in mind, EPA should properly evaluate the new information and may not simply ignore it without reasoned explanation of its choice.”). It may well be that “[t]he APA required the EPA to update its data or, at the very least, to explain why updated data was unnecessary.” Tex. Corn Producers v. EPA, 141 F.4th 687, 706 (5th Cir. 2025). But this only steepens the incline for Texas’s uphill argument. These decisions suggest it could have been arbitrary for EPA to approve the SIP on the basis that the CAIR trading program was ongoing and would resolve Texas’s obligations, a “significant factual predicate” that the agency would have known to be false. See Bechtel v. FCC, 957 F.2d 873, 881 (D.C. Cir. 1992) (quoting WWHT, Inc. v. FCC, 656 F.2d 807, 819 (D.C. Cir. 1981)). While EPA possibly retained discretion not to use the CSAPR Update modeling, the petitioners have not shown that it was arbitrary or irrational to “acknowledge and account for” this evidence from “a contemporaneous and closely related rulemaking.” See Portland Cement Ass’n v. EPA, 665 F.3d 177, 187 (D.C. Cir. 2011). To the extent that any explanation was required, EPA sufficiently explained that the data represented “the most up-to-date information the EPA has developed to inform our analysis of upwind state linkages to downwind air quality problems.” 81 Fed. Reg. at 21,293. We cannot say that this “fell outside the zone of reasonableness.” See FCC v. Prometheus Radio Project, 592 U.S. 414, 428 (2021).

40 Case: 16-60670 Document: 262-1 Page: 41 Date Filed: 09/22/2025

We discern no abuse of discretion or other arbitrariness in EPA’s reasoning resulting from its use of updated, accurate data. ii. In a similar vein, both the industry petitioners and Texas argue that EPA acted in bad faith and failed to act with an open mind in predetermining the conclusion that the SIP would be disapproved. The industry petitioners maintain that “EPA . . . judged Texas’s plan, not strictly on its own merits, but by whether Texas’s plan conformed to EPA’s FIP.” They state that “EPA disapproved Texas’s plan, not because it did not meet a requirement of the statute, but because it did not conform to a replacement federal plan that EPA had devised before it even proposed action on Texas’s plan.” Texas suggests that the timing of the disapproval and the FIP “made it impossible for Texas to correct any deficiency before the FIP came into effect.” In its reply brief, Texas elaborates that “EPA always needed Texas as part of its FIP, a trading program, for it to be successful” because “Texas is the largest state within its trading market.” This is not what the record says. The record says that EPA rejected the plan because of its statutory deficiencies, and provided evidence from the CSAPR Update rulemaking to corroborate its analysis. “[A] court may not reject an agency’s stated reasons for acting simply because the agency might also have had other unstated reasons.” Dep’t of Com. v. New York, 588 U.S. 752, 781 (2019). We cannot ignore the “lengthy agency record” showing an “independent[] assess[ment] and reject[ion of] Texas’s SIP on its own merits” just because “EPA mentioned its own data.” Texas II, 132 F.4th at 856. There is even affirmative evidence that EPA meant what it said when it gave Texas an opportunity to rectify its deficient submission; in the final FIP rulemaking, EPA relied on updated analysis to remove North Carolina from the FIP. 81 Fed. Reg. at 74,506.

41 Case: 16-60670 Document: 262-1 Page: 42 Date Filed: 09/22/2025

Besides EPA’s unexceptional alternative litigating position in favor of remand without vacatur, Texas points to proposal of the FIP in advance of the SIP disapproval as evidence for its bad faith claim. The industry petitioners similarly note that the SIP disapproval referenced EPA’s CSAPR Update analysis. This falls short of the “strong showing” required to disturb the presumption of regularity that accompanies administrative action. See Citizens to Pres. Overton Park, Inc. v. Volpe, 401 U.S. 402, 415, 420 (1971), abrogated on other grounds by Califano v. Sanders, 430 U.S. 99 (1977). We are unpersuaded that EPA’s decision to advise the affected states of a proposed FIP in advance of the SIP disapproval evinced arbitrary reliance on improper factors, rather than an effort to advance the statute’s aims of cooperative federalism by giving affected states expedient notice. If anything, because “the plain text of the CAA grants EPA plenary authority to issue a FIP ‘at any time’ within the two-year period that begins the moment EPA determines a SIP to be inadequate,” EME Homer, 572 U.S. at 511 n.14 (quoting 42 U.S.C. § 7410(c)(1)) (emphasis omitted), and FIPs are subject to the 30-day comment period in 42 U.S.C. § 7607(d)(5), see id. § 7607(d)(1)(B), the CAA seems to anticipate the circumstance that a FIP might be proposed in advance of a final SIP disapproval. Regardless, neither Texas nor the industry petitioners provide persuasive evidence that EPA relied upon the FIP in its own right rather than properly considering the most “relevant data” by acknowledging up-to-date science developed in a concurrent proceeding. See State Farm, 463 U.S. at 43. We conclude that EPA’s disapproval of the SIP was not arbitrary, capricious, or an abuse of discretion. VI. For the foregoing reasons, the petition for review is DENIED.

42 Case: 16-60670 Document: 262-1 Page: 43 Date Filed: 09/22/2025

Jerry E. Smith, Circuit Judge, dissenting: There is a straightforward way to decide this petition for review with- out examining the several weighty substantive issues: Vacatur of the Final Disapproval, per 5 U.S.C. § 706(2)(D), is required by the EPA’s failure to meet its 12-month deadline to act on the SIP submission set by 42 U.S.C. § 7410(k)(2), (3). That should be the end of this matter after all these years. I differ with the majority’s honest and carefully crafted resolution to the contrary. EPA acknowledges that it was seriously tardy. Its required review of the SIP should have occurred within 12 months, not 44. Although the agency tries to downplay the significance of its inaction, it was anything but harmless: EPA utilized the interim to develop new data and to prepare its own FIP, which it then conveniently used as the basis for disapproving Texas’s SIP. Using data created after the deadline is in excess of statutory con- straints and is arbitrary and capricious. That is a bright-line rule that this court should embrace. The majority’s well-intentioned reasoning strips Congress’s carefully delineated timeline of all meaning. I respectfully dissent.