Vacatur and Remand Required When EPA Reconsiders a Good-Neighbor SIP Disapproval Based on Post-Submission Modeling and Shifting Thresholds

1. Introduction

This consolidated interstate-transport dispute arises under the Clean Air Act’s (CAA) “Good Neighbor Provision,” requiring each State Implementation Plan (SIP) to prohibit in-state emissions that will “contribute significantly to nonattainment in, or interfere with maintenance by, any other State” with respect to the relevant National Ambient Air Quality Standard (NAAQS). The litigation concerns EPA’s 2015 revision to the ozone NAAQS (from 75 ppb to 70 ppb), which triggered state duties to submit Good-Neighbor SIPs within three years.

Parties. Petitioners include Texas and Texas agencies (Texas Commission on Environmental Quality and other state entities), Texas industry petitioners (power generators and trade associations), and also Louisiana and Mississippi (and related entities). Respondents are EPA and its Administrator.

Core issues. The case presented (i) venue—whether challenges belong in the D.C. Circuit or regional circuits; (ii) the lawfulness of EPA’s SIP disapproval methodology, including EPA’s use of a “1% of the NAAQS” screening threshold and the interplay between state-led SIP development and EPA review; and (iii) whether EPA impermissibly relied on post-submission modeling/data to disapprove SIPs.

The substituted opinion is driven by intervening events: EPA’s January 30, 2026 notice proposing to reconsider the disapproval action, describing the agency’s prior reliance on “2016v3 modeling,” and signaling a potential shift from the “1%” threshold (0.70 ppb for the 2015 ozone NAAQS) toward a “1-ppb” threshold (and even soliciting comment on other thresholds).

2. Summary of the Opinion

• Venue: The court holds venue is proper in the Fifth Circuit, relying on Oklahoma v. EPA (605 U.S. 609 (2025)) and concluding the SIP disapprovals are “locally or regionally applicable actions reviewable in a regional Circuit.”
• Texas and Mississippi: The petitions for review are granted; EPA’s disapprovals are vacated; and the matters are remanded to EPA.
• Louisiana: The petition for review is denied. The court emphasizes Louisiana did not seek rehearing, and the substituted opinion does not revisit the earlier merits analysis; it “reinstate[s]” the relevant portions of the withdrawn decision as to Louisiana and Mississippi.
• Reason for vacatur/remand (Texas): EPA’s January 30, 2026 reconsideration notice undermines confidence that EPA’s disapproval of Texas’s SIP rested primarily on a technical critique of Texas’s own modeling and “weight-of-the-evidence” analysis, rather than on later-developed EPA modeling and shifting interpretive guidance.
• Post-Loper Bright framing: The opinion elaborates the effect of Loper Bright Enterprises v. Raimondo (603 U.S. 369 (2024))—courts decide statutory meaning independently; agencies receive no Chevron-style deference on legal questions, though courts may find agency views persuasive under Skidmore v. Swift & Co. principles and still review technical factfinding/policymaking under the APA’s deferential standards.

3. Analysis

A. Precedents Cited

1) Oklahoma v. EPA (605 U.S. 609 (2025)) — Venue for SIP disapprovals

The substituted opinion treats Oklahoma v. EPA as dispositive on venue. Where EPA disapproves an individual state’s SIP submission for Good-Neighbor purposes, the action is “locally or regionally applicable,” and review lies in a regional circuit rather than being centralized in the D.C. Circuit. The Fifth Circuit applies the Supreme Court’s reasoning “equally to the disapprovals of the Louisiana, Mississippi and Texas SIPs,” resolving a threshold procedural battle that often dictates both timing and strategic posture in multi-circuit Clean Air Act litigation.

2) Loper Bright Enterprises v. Raimondo (603 U.S. 369 (2024)) — Courts decide statutory meaning; agencies may receive persuasive weight, not deference

The court’s extended discussion of Loper Bright Enterprises v. Raimondo serves two functions:

• Legal interpretation: Under APA § 706, courts decide “all relevant questions of law,” including the “best reading” of ambiguous statutes, and set aside agency action inconsistent with that reading.
• Role of expertise: Agency interpretations may still inform courts as a “body of experience and informed judgment” (invoking Skidmore v. Swift & Co. (323 U.S. 134 (1944))), especially where they rest on technical factual premises within the agency’s expertise.

Importantly, the opinion distinguishes legal meaning from technical application: while courts independently interpret statutory terms such as “contribute significantly,” disputes over the scientific soundness of modeling and evaluation methodologies are reviewed through the APA’s deferential lenses (arbitrary-and-capricious review and related standards), not through independent factfinding.

3) EPA v. EME Homer City Generation, L.P. (572 U.S. 489 (2014)) — The Good Neighbor Provision’s structure; “amounts”; over-control vs under-control; the “one-percent” concept in context

EPA v. EME Homer City Generation, L.P. anchors the Fifth Circuit’s discussion of what the Good Neighbor Provision requires and what EPA may do to implement it. The court extracts and treats as binding several core points from EME Homer:

• EPA’s task is to reduce upwind pollution, but only in “amounts” that push downwind concentrations above the relevant NAAQS.
• EPA may not impose reductions beyond what is necessary for downwind attainment (a limit designed to prevent “over-control”).
• EPA also has an obligation to avoid “under-control,” and because interstate transport is uncertain, EPA must have some “leeway” in balancing these risks.

The court also highlights a litigation reality: EME Homer discussed EPA’s “one-percent threshold” repeatedly, describing it as “the Agency’s interpretation” of “contribute significantly,” but that threshold was not directly challenged in EME Homer. The Fifth Circuit therefore treats EME Homer as authoritative on the Good Neighbor Provision’s aims and limiting principles, while stopping short of holding that a “1% of NAAQS” screening threshold is necessarily lawful in SIP review—especially because EPA is actively reconsidering its approach.

4) Michigan v. EPA (576 U.S. 743 (2015)) and Motor Vehicle Mfrs. Ass'n of the U.S. v. State Farm Mut. Auto. Ins. Co. (463 U.S. 29 (1983)) — Reasoned decisionmaking and “capacious” statutory terms

The court uses Michigan v. EPA (cited in Loper Bright) as an example of a broad statutory term (“appropriate and necessary”) that confers flexibility but still requires agencies to consider important aspects of the problem. This is relevant to the Good Neighbor debate because “contribute significantly” is likewise open-textured; the question becomes how much discretion Congress delegated to EPA to operationalize it and what “outer statutory boundaries” constrain that discretion.

State Farm appears as a touchstone for arbitrary-and-capricious review, reinforcing that even under flexible terms an agency cannot ignore important considerations or fail to explain itself.

5) Train v. Nat. Res. Def. Council, Inc. (421 U.S. 60 (1975)) — Cooperative federalism and state choice of emissions controls

The opinion invokes Train v. Nat. Res. Def. Council, Inc. for the classic cooperative-federalism principle: states have latitude to choose the “mix of emission limitations” so long as the ultimate effect is compliance with national standards. Yet the Fifth Circuit stresses that Train does not eliminate EPA’s substantive duty to ensure SIPs comply with the CAA, including the Good Neighbor Provision (as reflected in statutory approval criteria such as 42 U.S.C. § 7410(k)(3) and revision constraints in § 7410(l)).

6) North Carolina v. EPA (531 F.3d 896 (D.C. Cir. 2008)) — Distinct meaning for “nonattainment” and “maintenance” prongs

The court relies on North Carolina v. EPA for the interpretive proposition that the statute’s “contribute significantly to nonattainment” language and “interfere with maintenance” language must each be given independent meaning—reinforcing that Good Neighbor obligations address both current exceedances and risks to sustained attainment.

7) The Fifth Circuit’s own Clean Air Act administrative-law framework

The opinion situates its approach within the circuit’s existing administrative-law canon:

• Texas v. EPA, 983 F.3d 826 (5th Cir. 2020) (“Texas 2020”) — describes the NAAQS/SIP process and emphasizes deference to EPA’s technical evaluations of scientific data.
• Texas v. EPA, 690 F.3d 670 (5th Cir. 2012) — recognizes EPA cannot approve SIP revisions that interfere with attainment or other CAA requirements (citing 42 U.S.C. § 7410(l)).
• Luminant Generation Co., L.L.C. v. EPA, 675 F.3d 917 (5th Cir. 2012) — invoked by petitioners for the principle that EPA acts arbitrarily and capriciously if it imposes requirements not authorized by the statute or the SIP framework.
• BCCA Appeal Grp. v. EPA, 355 F.3d 817 (5th Cir. 2003) — cited for deference to EPA’s scientific judgments.
• Univ. of Tex. M.D. Anderson Cancer Ctr. v. HHS, 985 F.3d 472 (5th Cir. 2021) — the “treat like cases alike” principle when agency practice shifts.
• FCC v. Prometheus Radio Project, 592 U.S. 414 (2021); FCC v. Fox Television Stations, Inc., 556 U.S. 502 (2009); Bowman Transp., Inc. v. Ark.-Best Freight Sys., Inc., 419 U.S. 281 (1974) — reinforce the minimal rationality and explanation requirements (“path may reasonably be discerned”) for agency action.

8) Deadline/“operation of law” discussion: Kentucky v. EPA (123 F.4th 447 (6th Cir. 2024)) and related Supreme Court authority

The court rejects—at least as a practical matter—the idea that Texas’s SIP could have been “approved by operation of law” because EPA failed to act within statutory timeframes. Citing JUDGE MURPHY’s concurrence in Kentucky v. EPA, and Supreme Court authority (Nielsen v. Preap, Barnhart v. Peabody Coal Co., Brock v. Pierce County, McIntosh v. United States), the opinion explains that missing a statutory deadline generally does not strip an agency of power to act later unless Congress clearly imposed that penalty. The “operation of law” language in 42 U.S.C. § 7410(k)(1)(B) concerns completeness findings—not automatic SIP approval.

B. Legal Reasoning

1) Venue and the post-Oklahoma landscape

By applying Oklahoma v. EPA, the Fifth Circuit cements that Good-Neighbor SIP disapprovals are reviewable in regional circuits as locally/regionally applicable actions. This has immediate practical consequence: states can litigate in their home regional circuits rather than being funneled into the D.C. Circuit by default.

2) How Loper Bright reframes the SIP dispute without converting it into de novo science review

Petitioners invoked Loper Bright Enterprises v. Raimondo to argue that EPA lacked authority to effectively define “contribute significantly” via a “1% threshold” and related screening/linkage methods. The court’s response is more nuanced than simply declaring “no deference”:

• Statutory meaning: Courts independently determine the best interpretation of statutory terms (no Chevron deference).
• Delegated discretion: A statute may still delegate discretionary authority (the reviewing court must identify its boundaries and ensure reasoned decisionmaking within them).
• Technical determinations: Where the disagreement is fundamentally about scientific modeling, data, and methodology, the court treats those as classic APA “arbitrary and capricious” questions, not purely interpretive disputes.

This distinction is pivotal to the opinion’s structure. Even while it begins to engage the statutory question (the meaning and operationalization of “amounts” and “contribute significantly”), the court ultimately vacates and remands without resolving the legality of EPA’s chosen screening threshold—because EPA itself is reconsidering the predicate methodology and potentially changing the threshold and data sources.

3) EPA’s role versus State primacy in SIP design

The opinion recognizes (i) state latitude to select control measures (consistent with Train v. Nat. Res. Def. Council, Inc.), but (ii) EPA’s statutory obligation to ensure SIPs meet “all applicable requirements,” including Good Neighbor obligations (42 U.S.C. § 7410(k)(3), § 7410(l), and § 7410(k)(5)).

The court thus rejects the strongest form of Texas’s “states hold interpretive authority at the SIP stage” framing. In the Fifth Circuit’s account, the CAA’s cooperative federalism does not mean EPA must accept a state’s technical approach simply because the state labels it a “reasonable interpretation”—EPA must independently assess compliance with the CAA’s requirements.

4) The Texas SIP merits discussion—why the court does not ultimately affirm or reverse on the technical record

The opinion recounts (in detail) EPA’s criticisms of Texas’s “multi-factor ‘weight-of-the-evidence’ analysis,” including: Texas’s use of “elevated ozone days,” back-trajectory analysis (HYSPLIT), alternative averaging choices that could dilute contributions on the highest ozone days, and a DDM sensitivity approach. EPA identified methodological concerns (e.g., trajectory assumptions and the limited quantitative utility of back trajectories).

Standing alone, this record could have supported an affirmance under deferential APA review. But the court’s key move is to acknowledge that its earlier (withdrawn) reliance on EPA’s litigation position—“Nor did EPA… disapprove… based on the 2016-based modeling… which only confirmed EPA’s conclusion”—has been undermined by EPA’s own January 30, 2026 reconsideration notice. That notice states EPA disapproved SIPs “based on the SIP submissions, the EPA’s interpretation of its memoranda, and the 2016v3 modeling.”

Once EPA itself characterizes post-submission modeling and interpretive shifts as part of the basis for disapproval, the court “can no longer conclude with confidence” that EPA’s decision rested primarily on reviewing Texas’s SIP “on its own terms.” That loss of confidence—combined with EPA’s ongoing reconsideration and potential threshold changes—drives vacatur and remand.

5) Remedy: why vacatur (not merely a stay or abeyance) follows from agency reconsideration and methodological uncertainty

The Fifth Circuit vacates Texas’s SIP disapproval and remands because EPA’s reconsideration introduces concrete uncertainty about:

• The data basis: whether “2016v3” modeling unavailable to Texas at submission time was outcome-determinative.
• The Step 2 threshold: whether EPA will apply 1% (0.70 ppb), 1 ppb, or potentially higher thresholds (EPA solicited comment on “5-percent” and “2-ppb” thresholds).
• Consistency and reliance: what reliance interests may have been engendered by EPA’s 2018 memoranda and the extent to which EPA must adhere to (or explain departure from) prior guidance.

In that posture, the court declines to decide the ultimate legality of EPA’s screening threshold and instead resets the matter to the agency for a clarified, publicly vetted approach.

C. Impact

1) For Good-Neighbor SIP litigation nationwide

The decision is part of a broader multi-circuit wave of Good-Neighbor challenges (the opinion notes lawsuits in multiple circuits). Its most immediate systemic impacts are:

• Regional-circuit review is reinforced: after Oklahoma v. EPA, states can more confidently file (and keep) SIP-disapproval petitions in regional circuits.
• Agency reconsideration can prompt vacatur: where EPA itself signals that its rationale depended on post-submission modeling and interpretive pivots, courts may be reluctant to uphold the original disapproval on the theory that it was based purely on the state-submitted record.

2) Post-Loper Bright doctrinal consequences

The opinion previews a likely pattern in future Clean Air Act cases:

• Harder look at legal authority: litigants will frame challenges as questions of statutory meaning and delegated discretion, not “Chevron step two.”
• Continued deference on technical matters: courts will still be reluctant to second-guess scientific modeling choices so long as EPA provides a reasoned explanation and the path can be discerned.

In other words, Loper Bright narrows agency leeway on legal interpretation, but does not transform courts into substitute modelers.

3) Practical governance effects for EPA and states

The court’s remand creates incentives for EPA to:

• Clarify (through notice-and-comment) when and how later-developed modeling may be used in SIP review;
• Stabilize the Step 2 screening threshold (and justify it as within statutory bounds) to reduce litigation and reliance disputes;
• Explain how guidance memoranda interact with statutory obligations, and how reliance interests are addressed when the agency changes interpretive course.

4. Complex Concepts Simplified

• NAAQS: National air quality “targets” set by EPA for pollutants (here, ozone), expressed in measurable concentrations (ppb).
• SIP vs. FIP: A SIP is the state’s plan to meet Clean Air Act obligations; a FIP is a federal plan EPA must issue if a state fails to submit an adequate SIP.
• Good Neighbor Provision: A SIP must prevent in-state emissions from significantly worsening another state’s air quality problems (“nonattainment”) or jeopardizing its ability to keep meeting the standard (“maintenance”).
• Receptors and “linkage”: EPA uses modeled/monitored locations (“receptors”) to identify where downwind air quality problems exist and whether upwind states contribute above a screening threshold.
• Screening threshold (1% of NAAQS; 1 ppb): A cutoff used to decide whether an upwind contribution is too small to regulate further. For the 2015 ozone NAAQS, 1% equals 0.70 ppb.
• Photochemical modeling: Complex computer simulation of ozone formation and transport. Courts typically treat disputes over modeling choices as technical questions reviewed under arbitrary-and-capricious standards.
• Back-trajectory analysis (HYSPLIT): A method to estimate where air parcels came from, but (as EPA noted) it does not quantify pollution formation/removal and can be limited for assigning precise contribution amounts.
• Vacatur and remand: The court nullifies the agency action (vacatur) and sends it back to the agency to redo the decision consistent with the court’s ruling (remand).

5. Conclusion

The substituted Fifth Circuit opinion does not definitively resolve whether EPA’s “1% of the NAAQS” screening threshold is lawful after Loper Bright. Instead, it establishes a more practical—and immediately consequential—administrative-law principle for this dispute: where EPA’s own reconsideration notice indicates that SIP disapprovals rested materially on post-submission modeling and an interpretation of guidance that the agency may revise (including potentially changing the screening threshold), the court will vacate the disapproval and remand for reconsideration on a clarified and publicly vetted basis.

At the same time, the court reinforces two broader guideposts for Clean Air Act litigation: (1) venue for Good-Neighbor SIP disapprovals lies in regional circuits under Oklahoma v. EPA, and (2) after Loper Bright, courts independently decide statutory meaning, but remain deferential to EPA’s technical factfinding and policy application within the bounds of delegated discretion.