Before DAVIS, COSTA, and OLDHAM, Circuit Judges. GREGG COSTA, Circuit Judge: The Clean Air Act authorizes “any person” to sue polluters. 42 U.S.C. § 7604(a). Any recovery goes to the government. This citizen suit provision harkens back to pre-Founding English law that allowed private individuals, through various writs, to enforce laws on behalf of the government. See Steven L. Winter, The Metaphor of Standing and the Problem of Self-Governance, 40 STAN. L. REV. 1371, 1396–99 (1988); Raoul Berger, Standing to Sue in Public Actions: Is It a Constitutional Requirement, 78 YALE L. J. 816, 827 (1969). But modern citizen suits present challenges for the Article III “cases” or “controversies” requirement under which a plaintiff must suffer an injury from the defendant’s conduct. See generally RICHARD FALLON ET AL, HART & WECHSLER’S THE FEDERAL COURTS AND FEDERAL SYSTEM 151–54 (5th ed. 2003). Indeed, citizen suits under two other environmental statutes—the Clean Water Act and Endangered Species Act—resulted in leading Supreme Court standing decisions. Friends of the Earth, Inc. v. Laidlaw Envtl. Servs., Inc., 528 U.S. 167, 180–81 (2000); Lujan v. Defenders of Wildlife, 504 U.S. 555, 560–61 (1992); . This citizen suit seeking to recover for Clean Air Act violations at the largest petroleum and petrochemical complex in the nation again raises this tension between citizen suits and Article III. The principal issue in this second appeal of the case is whether plaintiffs have standing to recover for more than 16,000 violations of emission standards. I. The ExxonMobil complex in Baytown, Texas is massive. It includes refinery, a chemical plant, and an olefins plant. Emissions from the complex are regulated in part by permits. The Texas Commission on Environmental Quality issues the permits under Title V of the Clean Air Act. The Commission, along with the EPA, enforces the permits. To monitor compliance, the Commission requires polluters to document unauthorized “emissions events”—that is, unplanned or unscheduled emissions. If the event produces pollutants in excess of thresholds, the polluter must report it to the Commission. See 30 TEX. ADMIN CODE § 101.201(a); see also id. 101.1(88), (89) (setting “reportable quantit[ies]” of emissions). We will call these “ reported events.” If the event produces pollutants below reportable levels, polluters must nevertheless maintain records documenting the emission. Id. § 101.201(b). We will call these “recorded events.” In addition to the powers it gives regulators, the Clean Air Act gives citizens a role in enforcing its requirements. A citizen may seek civil penalties, payable to the government, for each day of repeated or ongoing violations of an emissions standard. 42 U.S.C. §§ 7604(a)(1), 7413(e)(2). Environment Texas Citizen Lobby and Sierra Club brought such a suit for each of Exxon’s reported and recorded emissions events from October 2005 through September 2013 in total, 241 reported events and 3,735 recorded events. By Plaintiffs’ calculations, which Exxon does not challenge,[1] the nearly 4,000 emissions events resulted in 16,386 days of violations. The thousands of violations fell into the following buckets: Count I alleged violations of a permit condition prohibiting “upset emissions.” An upset emission is an “unplanned and unavoidable breakdown or excursion of a process or operation that results in unauthorized emissions.” 30 TEX. ADMIN CODE § 101.1(110). Plaintiffs calculated 10,583 days of upset-emission violations, spanning 24 different pollutants, at the Baytown refinery. Count II alleged violations of the Maximum Allowable Emission Rate Tables—permit conditions setting hourly emissions limits for specific pollutants. Plaintiffs calculated 5,709 days of these violations, spanning more than a dozen pollutants, at the olefins plant and chemical plant. Count III alleged 18 days of violations of a 1,200 pound/hour limit on emissions of highly reactive volatile organic compounds. Count IV alleged 44 days of violations of an EPA rule limiting visible emissions from flares (which are used to burn waste gases) to no more than 5 minutes during any 2-hour period. Count V alleged 32 days of violations of a rule requiring flares to operate with a pilot flame. Our court has already grappled with this case. Following a bench trial, the district court initially found that only a small fraction of the violations were “actionable” because the Clean Air Act limits citizen suits to violations that were repeated in the past or ongoing at the time of the complaint. 66 F. Supp. 3d 875, 895–902 (S.D. Tex. 2014); see also 42 U.S.C. § 7604(a)(1) (allowing suit if defendant caused “repeated” violations or is “in violation”). It went on to rule that even if every alleged violation were actionable, it would decline to assess a civil penalty against Exxon. 66 F. Supp. 3d at 904. A panel of this court vacated and remanded, concluding that the district court had too narrowly analyzed actionability and three of the factors (duration, seriousness, and the economic benefit of noncompliance) courts must consider in assessing a civil penalty. 824 F.3d 507, 514 (5th Cir. 2016); see also 42 U.S.C. § 7413(e)(1) (enumerating penalty factors). On remand, the district court determined that the 16,386 days of violations alleged in Counts I–V were actionable. And after reconsidering the penalty factors, the district court imposed a $19.95 million civil penalty. Exxon appeals, attacking the judgment on three fronts: standing, affirmative defenses, and penalty factors. II. Congress granted “any person” the right to sue under the Clean Air Act. 42 U.S.C. § 7604(a). But that does not mean that any person can always bring such a suit. The Constitution limits congressional grants of federal court jurisdiction. Hodgson v. Bowerbank, 9 U.S. (5 Cranch) 303, 304 (1809) (“Turn to the article of the constitution of the United States, for the statute cannot extend the jurisdiction beyond the limits of the constitution.”). Those limits include Article III’s case-or-controversy requirement, which, among other things, requires that the plaintiff have standing. Lujan, 504 U.S. at 560–61. And unlike qui tam relators bringing False Claims Act cases, who have standing via the government’s injury because their entitlement to a bounty is a partial assignment of the claim to the relator, Vt. Agency of Nat. Res. v. United States ex rel. Stevens, 529 U.S. 765, 773–74 (2000), citizens suing under the bountyless environmental statutes must meet the standing requirement in their own right, Laidlaw, 528 U.S. at 180–81 (explaining in Clean Water Act citizen suit that “the relevant showing for purposes of Article III standing . . . is not injury to the environment but injury to the plaintiff”). Plaintiffs thus must show an injury, traceable to the defendant’s challenged conduct, that will be redressed by a favorable decision. Lujan, 504 U.S. at 560–61. A. As a threshold matter, Plaintiffs argue that because the prior panel reached the merits, it must have determined it had jurisdiction. If that is the case, then Plaintiffs’ standing is the binding law of the case. See Todd Shipyards Corp. v. Auto Transp., S.A., 763 F.2d 745, 750 (5th Cir. 1985) (“The law of the case doctrine . . . generally precludes reexamination of issues of law or fact decided on appeal, either by the district court on remand or by the appellate court itself on a subsequent appeal.”). But we cannot assume that the prior panel implicitly decided standing. Because Article III standing goes to our subject matter jurisdiction, broad applications of the law of the case doctrine are inappropriate. See Propes v. Quarterman, 573 F.3d 225, 228 (5th Cir. 2019) (quoting 18B CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE AND PROCEDURE § 4478.5, at 790 (2d ed. 2002)). For that reason, a later panel cannot “defer to the prior panel’s exercise of jurisdiction as correct where the issue was neither raised by the parties nor addressed by the court.” USPPS, Ltd. v. Avery Dennison Corp., 647 F.3d 274, 284 (5th Cir. 2011). Exxon’s brief to the original panel did include a footnote disputing that Plaintiffs had standing. But the footnote went on to say that Exxon was not disputing the issue because the district court’s original decision found no repeated and ongoing violations to support penalties. The prior panel’s decision makes no mention of standing, and we cannot infer from a single, equivocating footnote on the issue that the panel resolved standing without saying so. B. We thus must address standing. The main legal dispute is whether Plaintiffs must prove standing for each violation they alleged. An explanation of interaction between Clean Air Act claims, violations, and penalties is needed to understand the parties’ dispute. The Act provides a cause of action—that is, a claim—only for repeated violations of a particular emission standard. 42 U.S.C. § 7604(a)(1); Env’t Tex., 824 F.3d at 518–19. That means a plaintiff must assert at least two violations of the same standard in order to allege a claim. Once that threshold is met, however, there is no ceiling on how many violations of that emission standard a plaintiff may pack into that claim. And a “penalty may be assessed for each day of violation.” 42 U.S.C. § 7413(e)(2). There is a cap on the penalty for each day of violation: $32,500 or $37,500 in this case, depending on when the violation occurred.[2] Consequently, if Plaintiffs could sue only for 100 days of violations, they could at most recover roughly $3.5 million. In light of the more than 16,000 days of violations found in this case, the statutory cap exceeded $600 million. Thus the dispute about whether Plaintiffs must demonstrate standing for each violation. Plaintiffs argue that although they must prove standing for each Clean Air Act claim (that is, group of violations of a particular emission standard), there is not a separate standing inquiry for each violation asserted as part of that claim. But what we have just explained—that Clean Air Act penalties are tied to violations, not the broader claims—refutes this position. Plaintiffs’ argument thus runs up against the principle that one injury does not entitle a litigant to right other wrongs that did not injure it. Assuming that Plaintiffs’ members’ injuries were traceable to some of Exxon’s violations, that does not mean they “possess by virtue of that injury the necessary stake in litigating conduct of another kind, although similar, to which [they have] not been subject.” Blum v. Yaretsky, 457 U.S. 991, 999 (1982). This principle typically arises in suits against the government, preventing plaintiffs with standing to challenge one facet of a regulatory scheme from challenging the whole regulatory scheme. See, e.g., id. at 100–01; Legacy Cmty. Health Servs., Inc. v. Smith, 881 F.3d 358, 366–70 (5th Cir. 2018). But because it rests on Article III’s case-or-controversy requirement, it applies to this suit too. An example with an obvious answer shows why it must generally be true that a plaintiff needs standing for each violation for which it seeks a penalty. Assume that a citizen moved from Florida to a Baytown neighborhood near the Exxon complex in 2005. That citizen would not have standing to assert violations that occurred in 2004. So Clean Air Act plaintiffs cannot seek penalties for a particular violation if they would lack standing to sue for that violation in a separate suit, any more than a plaintiff with the “right to complain of one administrative deficiency” can “bring the whole structure of state administration before the courts for review.” Lewis v. Casey, 518 U.S. 343, 358 n.6 (1996). Admittedly, no court appears to have found standing for some Clean Air Act violations but not others, and that gives us some pause. Numerous cases have instead recognized standing in environmental citizen suits without separate analyses for each violation. In Laidlaw, for instance, a Clean Water Act defendant violated its mercury discharge limits on 489 occasions. 528 U.S. at 176. The Court did not parse one day of violations from another in holding that the plaintiff had standing. See id. at 180–88. Neither did this court when we analyzed standing to sue under the Clean Air Act for 625 days of emissions violations at a refinery near the one in this case. Texans United for a Safe Econ. Educ. Fund v. Crown Cent. Petroleum Corp., 207 F.3d 789, 791 n.4 (5th Cir. 2000). But these other cases do not involve the number and variety of violations that this case does (24 different pollutants). That explains why in this case, unlike in Laidlaw and Texans United, the standing inquiry is not one-size-fits- all. Citizens suits under the environmental laws typically allege injuries from discharges or emissions of one or two pollutants exceeding one or two emissions standards, all in the same manner. See, e.g., Laidlaw, 528 U.S. at 176 (recounting defendant’s discharge of mercury, in excess of its permitted allotment, from its wastewater treatment plant into a river); Texans United, 207 F.3d at 791 (observing that defendant exceeded federal emissions limits for sulfur dioxide and hydrogen sulfide); Sierra Club v. Tenn. Valley Auth., 430 F.3d 1337, 1340–41 (11th Cir. 2005) (analyzing standing to sue for emitting smoke that was darker than maximum opacity limits permitted); Friends of the Earth, Inc. v. Gaston Copper Recycling Corp., 204 F.3d 149, 153 (4th Cir. 2000) (noting plaintiffs’ allegation that defendant regularly exceeded discharge limits during releases into a lake). That the plaintiffs in those cases repeatedly suffered the same injury resulting from a series of similar discharges does not mean that a plaintiff injured by one violation can automatically challenge all a defendant’s violations. Another difference is that in each of the cases on which Plaintiffs rely, there was no doubt that the pollutant emitted could cause the alleged injury. The Laidlaw plaintiffs, for instance, asserted injuries to their aesthetic and recreational interests because the defendant’s discharges polluted a river that they otherwise would have enjoyed. 528 U.S. at 183–84. Plaintiffs, by contrast, assert a variety of aesthetic and health-related injuries, allegedly traceable to 24 different pollutants emitted in a variety of ways (flaring, leaks, workplace accidents, etc.). The impact of those different violations varied greatly. Count I, for instance, alleges violations of a prohibition on upset—that is, unplanned or accidental—emissions. Because upset emissions are not authorized in any amount, each pollutant emitted during an upset event is a violation, regardless of how small the emission may have been. As the district court observed, that means Exxon’s thousands of violations include accidents as minor as smoke caused by plugging in an extension cord and a fire in a cigarette butt can that lasted one minute. Both accidents lasted “0.0 hours” and emitted 0.01 pounds of carbon monoxide and 0.01 pounds of nitrous oxide. By Plaintiffs’ calculations, that meant a total of four days of violations.[3] To be sure, many of Exxon’s emissions violations were of a serious magnitude. Some flaring events and leaks lasted for hours and spilled thousands of pounds of harmful pollutants into the air. But because of the great variety of the challenged emissions—both in terms of type and scale—we cannot say that Plaintiffs’ proving standing for some violations necessarily means they prove standing for the rest. C. Having decided that Clean Air Act plaintiffs must prove standing for each violation in support of their claims, we now examine whether these plaintiffs have done so. Before doing so, we emphasize it does not follow from the need to establish standing for each violation that separate proof of standing is needed for each violation. As we have discussed, many courts—including our court and the Supreme Court—have allowed the same testimony to support standing for multiple violations (just not the number and variety of violations at issue here). And, as always, a factfinder may rely on circumstantial evidence and draw reasonable inferences from the evidence. Gaston Copper, 204 F.3d at 163. Because this case was tried, Plaintiffs needed to prove standing by a preponderance of the evidence. See Lujan, 504 U.S. at 561. A factual finding that a plaintiff met that burden is reviewed for clear error. Department of Commerce v. New York, 139 S. Ct. 2551, 2565 (2019); In re Deepwater Horizon, 739 F.3d 790, 805 (5th Cir. 2014); Cole v. Gen. Motors Corp., 484 F.3d 717, 721 (5th Cir. 2007). We address standing’s three requirements in turn. 1. Plaintiffs easily demonstrated that their members were injured. They put on testimony from four of their members, which the district court credited. Here is a sampling. Sharon Sprayberry, who lived in Baytown from 2004 through 2012, could see flares, smoke, and haze coming from the Exxon complex from her house. She experienced respiratory problems when she lived in Baytown and does not return to visit friends because the last time she did, the air quality made it hard to breathe. Richard Shae Cottar, who lived a quarter mile from the complex from April 2010 through September 2012, could also see emissions from the Exxon complex and ultimately moved two miles away out of concern for his family’s health and safety. He likes to take his family to visit a nature center next to the Exxon complex, but he leaves whenever he sees emissions. Marilyn Kingman heads to Baytown a few times a week to run errands, recreate, and go to church. But she’s scared of what she smells coming from the complex and limits her grandchildren’s outdoor activities in Baytown whenever she smells odors or sees haze. Diane Aguirre Dominguez lived in Houston but regularly visited her parents in Baytown until she moved in March 2013. She is a runner, but she stopped running in Baytown because it hurt her throat and made it hard to breathe. In sum, throughout the claims period, at least one of Plaintiffs’ members regularly saw flares, smoke, and haze coming from the complex; smelled chemical odors; suffered from allergy-like or respiratory problems; feared for their health; refrained from outdoor activities; or moved away. Each of those experiences was an Article III injury. See Laidlaw, 528 U.S. at 183–84 (interference with recreational activity); Texans United, 207 F.3d at 792 (“breathing and smelling” polluted air); Sierra Club, Lone Star Chapter v. Cedar Point Oil Co. Inc., 73 F.3d 546, 557 (5th Cir. 1996) (unpleasant sights and smells of pollution); Sierra Club v. EPA, 762 F.3d 971, 977 (9th Cir. 2014) (“credible threat to the plaintiff’s physical well-being from airborne pollutants” (quoting Hall v. Norton, 266 F.3d 969, 976 (9th Cir. 2001)). Plaintiffs have thus met standing’s first requirement. See Summers v. Earth Island Inst., 555 U.S. 488, 494 (2009) (noting that organizations can establish standing through a member). 2. The next question is whether the injuries these individuals suffered are traceable to the violations. This is the crux of the dispute. And while Plaintiffs view traceability too loosely in contending that proving it for one violation proves it for all, Exxon views this element too strictly when it comes to how specific the proof must be. Exxon argues that Plaintiffs cannot trace any of their injuries to violations except for those occurring during five “correlated” events. These are five instances when one of Plaintiffs’ members linked a particular violation to a particular injury suffered on a particular day. One member, for instance, took a video of an emissions event that woke him up in the middle of the night, then checked Exxon’s reported emissions and linked what he had seen, heard, and smelled to the report Exxon filed after the event. Requiring proof that specific is not consistent with the traceability requirement, which requires less of a causal connection than tort law (and even tort causation would not require such specific proof). See Gaston Copper, 204 F.3d at 161 (“[T]he ‘fairly traceable’ standard is ‘not equivalent to a requirement of tort causation.’” (quoting Pub. Interest Research Grp. of N.J. v. Powell Duffryn Terminals, Inc., 913 F.2d 64, 72 (3d Cir. 1990)). After all, the standard is that the injury be fairly traceable to the challenged conduct, not definitively so. See, e.g., Laidlaw, 528 U.S. at 180; Allen v. Wright, 468 U.S. 737, 751 (1984). Indeed, our court has already rejected Exxon’s argument: Clean Air Act plaintiffs need not “connect the exact time of their injuries with the exact time of an alleged violation.” Texans United, 207 F.3d at 793. Traceability instead requires something more than conjecture (“The Exxon complex in Baytown emits pollutants, and I live in Baytown”) but less than certainty (“I was outside the Baytown complex on November 15, between 1:00 and 5:00 pm, at which time hydrogen sulfide was emitted, and I recall my throat feeling sore even though it did not feel sore earlier in the day”). We have described the showing as requiring evidence that the defendant’s violations were of a type that “causes or contributes to the kinds of injuries alleged by the plaintiffs.”[4] Cedar Point, 73 F.3d at 557 (quoting Powell Duffryn, 913 F.2d at 72); see also Friends of the Earth, Inc. v. Crown Cent. Petroleum Corp., 95 F.3d 358, 360–61 (5th Cir. 1996) (same). With these general principles in mind, we can address whether Plaintiffs demonstrated traceability. The district court made findings relevant to this question: Plaintiffs’ members’ observational injuries—seeing flares, smoke, and haze—were obviously traceable to Exxon because those visual blights originated from the Baytown complex. Chemical odors could likewise be traced to Exxon because they got stronger when Plaintiffs’ members approached or happened to be downwind from the complex. Seeing flares, smoke, and haze, as well as smelling chemical odors, made Plaintiffs’ members fear for their health, which in turn made them refrain from outdoor activities or move away from the complex. The chemical odors could be traced to the Exxon complex because they got stronger when the wind blew in the direction of a member’s home or when a member drove closer to the complex. And because members’ physical symptoms improved when they moved away from the complex, those injuries were traceable to Exxon too. The district court also observed—correctly as we have already explained—that traceability does not require Plaintiffs to draw a causal connection between their members’ injuries and specific incidents on particular days. Just as the would-be fishermen, hikers, and bird watchers in Laidlaw did not have to connect their injury to specific days when discharges into the river occurred, see 528 U.S. at 183–84, a person living near a refinery who cannot “enjoy[] . . . their surroundings” because of emissions need not connect their injuries to the “exact dates” when violations happened, see Texans United, 207 F.3d at 792–93. But despite the district court’s sound reasoning, we must remand due to our holding that Plaintiffs needed to prove standing for each violation. The district court outlined in general terms how Exxon’s violations had injured Plaintiffs’ members; it did not asses traceability as to each violation. That is necessary because it is not apparent that all of Exxon’s violations were capable of causing the types of injuries Plaintiffs’ members suffered. To address that problem and establish traceability on remand, Plaintiffs must make two showings. First, that each violation in support of their claims “causes or contributes to the kinds of injuries” they allege. Cedar Point, 73 F.3d at 557 (quoting Powell Duffryn, 913 F.2d at 72). Given the district court’s findings regarding injury and traceability, a violation will satisfy that standard if it (1) created flaring, smoke, or haze; (2) released pollutants with chemical odors; or (3) released pollutants that cause respiratory or allergy-like symptoms. Second, Plaintiffs must demonstrate the existence of a “specific geographic or other causative nexus” such that the violation could have affected their members. Cedar Point, 73 F.3d at 558 n.24; cf. Center for Biological Diversity v. EPA, 937 F.3d 533, 538–39 (5th Cir. 2019) (holding that plaintiffs could not establish injury based on their general use of parts of the Gulf of Mexico to challenge drilling permits for specific areas of that vast body of water). In Clean Water Act cases, for example, the nexus requirement means that a plaintiff significantly downstream from the defendant must do more than point to the defendant’s having discharged pollutants upstream to show traceability. Cedar Point, 73 F.3d at 558 n.24; see also, e.g., Crown Cent., 95 F.3d at 361–62 (finding that plaintiffs eighteen miles downstream lacked evidence of discharges’ range of impact such as water samples or expert testimony). That said, there is no need for evidence of geographic range when plaintiffs “sit[] squarely in the discharge zone of a polluting facility” such that their proximity speaks for itself. Gaston Copper, 204 F.3d at 162. In making this geographic nexus inquiry, the district court should distinguish between two kinds of violations. On the one hand, there were small-magnitude emissions events that constituted violations only because Exxon’s permits established zero-emissions standards. For these events, the district court must decide whether evidence shows that the emitted pollutants could have reached beyond the Exxon complex into the offsite areas of Baytown where Plaintiffs’ members lived and recreated. It is possible—particularly given the size of the Exxon complex (3,400 acres),—that pollutants emitted in a small quantity could have dissipated before leaving Exxon grounds. See Ctr. for Biological Diversity, 937 F.3d at 545 (noting that not only “the proximity of the source and the injury” but also “whether discharges will evaporate or become diluted” bears on the standing inquiry). We repeat, though, that there is no need for “scientific certainty.” Save Our Community v. EPA, 971 F.2d 1155, 1161 (5th Cir. 1992) (quoting Powell Duffryn, 913 F.2d at 72). If the district court finds by a preponderance of the evidence that an emission was large enough that it could have sent pollutants in discernible quantities—that is, enough to cause chemical odors, allergy symptoms, or respiratory symptoms[5]—across the fence surrounding the Baytown complex, Plaintiffs will have standing to recover a civil penalty for that emission. They need not offer direct evidence that their members actually inhaled pollutants from each emission. But other violations involve Exxon’s releasing pollutants in excess of nonzero emissions limits. For emissions that exceeded those thresholds, it is an easier inference that the pollutants escaped the Baytown complex. The same is true for emissions of pollutants that, though not permitted in any amount, were nevertheless of a “reportable quantity” under state regulations. See 30 TEX. ADMIN. CODE § 101.1(89). While it is not clear that emissions of a trace amount of pollutants could have reached outside the Exxon complex to affect Plaintiffs’ member, it is reasonable to conclude that emissions in excess of 5,000 pounds of carbon monoxide or 500 pounds of sulfur dioxide—the reportable thresholds for those pollutants, id. § 101.1(89)(A)(i)(III)(-d-); 40 C.F.R. pt. 355, app. A—could have. And because Plaintiffs’ members lived and recreated in immediate proximity to the Exxon complex, they were “squarely in the discharge zone” for emissions of those quantities. Gaston Copper, 204 F.3d at 162. Finally, we note that the geographic nexus inquiry is unnecessary for any violation that could have caused or contributed to flaring, smoke, or haze, even if the emission was of a small magnitude. Plaintiffs’ members could undoubtedly see—as they said they did in the testimony the district court credited—flares, smoke, and haze from their homes and other areas outside the Exxon complex. That evidence is enough for these violations. Under the above rubric, the district court’s findings support traceability for a substantial number of Exxon’s violations. Many of Count I’s upset emissions and Count II’s excessive emissions caused flaring, and the violations in Count IV consist of flaring events that caused excessive smoke. Also, each of the 241 reported events caused multiple days of violations. A limited remand is needed, however, for the district court to determine what other violations could have contributed to Plaintiffs’ members’ injuries and then tabulate its findings. This is no doubt an arduous task, but we do not require line-by-line findings for the thousands of violations. When consistent with the principles we have outlined, the court may group violations by type and magnitude in making its findings. The following should guide the district court’s inquiry: For any violation that could cause or contribute to flaring, smoke, or haze, the district court’s findings have established traceability. The district court need only decide which violations fall within this category. For violations that could not contribute to flaring, smoke, or haze, the district court should first consider whether the pollutant emitted could cause or contribute either to (a) chemical odors or (b) allergy- like or respiratory symptoms. If so, the district court will conduct the geographic nexus inquiry described above, finding it satisfied if the emission (i) violated a nonzero emissions standard, (ii) had to be reported under Texas regulations, or (iii) is otherwise proven to be of sufficient magnitude to reach Baytown neighborhoods outside the Exxon complex in quantities sufficient to cause chemical odors, allergy-like symptoms, or respiratory symptoms. 3. Because the evidence supports the district court’s findings of injury and traceability for a number of the violations, we proceed to the final standing requirement: redressability. Exxon argues that Plaintiffs failed to prove that civil penalties would “likely” redress their members’ injuries. Lujan, 504 U.S. at 561. But when violations are “ongoing at the time of the complaint,” civil penalties redress injuries caused by those violations because they “encourage defendants to discontinue current violations and deter them from committing future ones.” Laidlaw, 528 U.S. at 186, 188. The idea is that “a defendant once hit in its pocketbook will surely think twice before polluting again.” Id at 186. The question, then, is whether Exxon continued to commit violations after the plaintiffs filed suit. Plaintiffs sued in December 2010 and seek penalties for violations that occurred through September 2013. It seems straightforward that this almost three-year postsuit continuation of wrongdoing establishes redressability. See, e.g., id. at 176–77 (finding redressability when plaintiffs sued in 1992 for violations that continued through 1995); Gaston Copper, 204 F.3d at 163 (finding redressability when hundreds of violations occurred after the suit was filed). In the face of this authority, Exxon raises a novel argument: it contends its violations are undeterrable because they do not arise from a single “root cause.” Penalizing Exxon for past violations will not prevent future violations, the argument goes, because it is impossible to fix every possible cause of violations. Even if Exxon fixes the problems that caused its past violations, new issues will inevitably lead to future violations. This case thus presents the rare circumstance according to Exxon, when “the deterrent effect of a claim for civil penalties becomes so insubstantial or so remote that it cannot support citizen standing.” Laidlaw, 528 U.S. at 186. Exxon overstates the redressability requirement. Civil penalties need not completely prevent future violations. It is enough for the Plaintiffs to show that Exxon’s violations are susceptible to a reduction in frequency or magnitude. See Massachusetts v. EPA, 549 U.S. 497, 524 (2007) (“While it may be true that regulating motor-vehicle emissions will not by itself reverse global warming, it by no means follows that we lack jurisdiction to decide whether EPA has a duty to take steps to slow or reduce it.”). There is a glaring problem with the idea that Exxon could not have reduced its emissions violations after the filing of this suit: Exxon has done just that. Exxon steadily and substantially reduced the rate and magnitude of emissions violations throughout the claims period. Annualized violations declined from roughly 300 tons of emissions in 2010 to roughly 100 tons by September 2013. The three-year average of emissions violations at the Baytown complex shows even more success: it went from about 700 tons/year from 2008–10 all the way down to just over 100 tons/year for 2011–13. What is more, Exxon agreed with state regulators in a February 2012 settlement to undertake projects that would substantially reduce emissions of several different pollutants. Even if Exxon cannot address every “root cause” of its emissions violations, it has implemented measures that drastically reduced those violations. The district court thus correctly found that Plaintiffs’ injuries are redressable. III. Next are Exxon’s affirmative defenses. The district court did not need to consider them the first time around because it declined to impose civil penalties. Exxon reasserted two on remand, an “Act of God” defense and a statutory no-fault defense, but the district court rejected both. Plaintiffs have standing for at least some of the violations that Exxon asserts the defenses against: at a minimum, many of the challenged violations involved flaring. Some of that flaring occurred during the Hurricane Ike shutdown, the subject of the Act of God defense, and Exxon also contends its statutory defenses apply to some flaring. So we will address Exxon’s defenses. A. Exxon asserted the Act of God defense under the Texas Water Code. See TEX. WATER CODE § 7.251. The defense excuses emissions violations that were “caused solely by an Act of God, war, strike, riot, or other catastrophe.” Id. Exxon argues that Hurricane Ike was such an occurrence and that Ike caused 10 reported events, resulting in 199 days of violations. The district court rejected the Act of God defense because it concluded it is not part of Texas’s state implementation plan (SIP). A brief background on SIPs is in order. Under the Clean Air Act, each state submits an SIP for EPA approval. 42 U.S.C. § 7410. The SIPs detail how the state will achieve compliance with National Ambient Air Quality Standards. See generally Luminant Generation Co. v. EPA, 675 F.3d 917, 921–22 (5th Cir. 2013). Once approved, a state must seek EPA approval to modify its SIP. Tenn. Valley Auth., 430 F.3d at 1346; see 42 U.S.C. § 7410(n)(1). That is because when the EPA approves a SIP, it becomes federal law. Safe Air for Everyone v. EPA, 488 F.3d 1088, 1096–97 (9th Cir. 2007). Unless an SIP authorizes a state-law defense, the defense is not available in a Clean Air Act suit. Tenn. Valley Auth., 430 F.3d at 1346. Texas’s SIP, as approved by the EPA in 1972, contains an Act of God defense. It incorporates part of the Texas Clean Air Act, see 40 C.F.R. § 52.2270(e), which included the following provision when the SIP was approved: “The liabilities which would otherwise be imposed by this Act . . . shall not be imposed due to any violation caused by an act of God . . . or other catastrophe.” TEX. REV. CIV. STAT. art. 4477-5, § 4.05 (West 1976). But that provision was repealed in 1989. Act of June 14, 1989, 71st Leg., ch. 678, § 13(1), 1989 TEX. SESS. LAW SERV. 3165 (West). Today, the Texas Clean Air Act’s “enforcement mechanisms” are codified in the Water Code, see BCCA Appeal Grp., Inc. v. City of Houston, 496 S.W. 1, 4–5 (Tex. 2016), which has a defense similar to the old one in the Texas Clean Air Act: “If a person can establish that an event that would otherwise be a violation . . . was caused solely by an act of God . . . or other catastrophe, the event is not a violation . . . .” TEX. WATER CODE § 7.251. But the Water Code provision never made its way directly into Texas’s SIP. Exxon thus blundered by citing the Water Code for the Act of God defense. But the Water Code defense is nearly identical to the repealed Texas Clean Air Act provision, which the EPA approved as part of the SIP long before it left the statute books. The Texas SIP—the governing federal law under the Clean Air Act—thus incorporates an Act of God defense. Exxon’s ability to assert the defense should not turn on its failure to cite the decades-old statute, which was made part of the SIP, rather than its current statutory home. We remand for findings on whether Exxon proved its Act of God defense for the relatively small number of violations occurring during Hurricane Ike. B. The Texas SIP also includes affirmative no-fault defenses to liability for specific types of emissions. 40 C.F.R. § 52.2270(c) (incorporating TEX. ADMIN. CODE § 101.222). One defense, for instance, applies to “non-excessive upset events,” provided the violator can show that it met each of eleven criteria— they generally show that the violator was not at fault and that the violation was not particularly serious. TEX. ADMIN CODE § 101.222(b). Exxon pursued some of these no-fault defenses on remand. In support, it submitted 200 paragraphs of proposed findings. Those paragraphs included brief descriptions of 97 emissions events, and references to the reports and trial testimony of experts.[6] The district court ruled that Exxon failed to meet its burden because it did not identify evidence establishing it met the relevant criteria for each individual emissions event. It characterized Exxon as having “only provided a general citation to the testimony and record.” That characterization was not unreasonable. The trial testimony Exxon cited did not examine particular events; it instead opined that all the criteria were met for each violation, across the board. And Exxon did not pinpoint pages of the expert report, instead referring to hundreds of pages of reports in their entirety. Exxon responded with a Rule 52(b) motion to amend the judgment, to which it attached an exhibit purporting to more specifically cite the portions of expert reports and testimony pertinent to each of the 97 events. But that exhibit just copied and pasted identical references to voluminous expert testimony and reports into each entry for each emissions event. Some of that evidence had no bearing on the emissions events for which it purportedly proved a defense. The district court rejected Exxon’s 52(b) motion because the motion did not do anything Exxon could not have done in its proposed findings. We affirm that rejection for an alternative reason, the same reason we affirm the district court’s original rejection of Exxon’s section 101.222 defenses: “Judges are not ferrets!” Nicholas Acoustics & Specialty Co. v. H&M Const. Co., 695 F.2d 839, 847 (5th Cir. 1983); see also Chavez v. Sec. Fla. Dep’t of Corrections, 647 F.3d 1057 (11th Cir. 2011) (“[D]istrict court judges are not required to ferret out delectable facts buried in a massive record.”). We defer to the district court’s “broad discretion in managing [its] docket,” Sims v. ANR Freight Sys., Inc., 77 F.3d 846, 849 (5th Cir. 1996), which permits a district court to refuse to do litigants’ work for them. IV. Lastly, Exxon challenges the penalty the district court imposed. It argues that the district court erred in reconsidering all three factors on which the prior panel remanded—economic benefit, duration, and seriousness—as well as in declining to consider Exxon’s argument on “such other factors as justice may require.” Because we remand for the district court to determine the number of violations for which Plaintiffs have standing, as well as whether Exxon proved its Act of God defense for any violations, the court will also have to reassess the penalties. It thus does not make sense to address Exxon’s penalty arguments now. It is worth noting that however many violations the court finds Plaintiffs have standing to pursue, the overall number of violations at the Baytown complex remains relevant as part of “the violator’s full compliance history” that a court considers in assessing the amount of each penalty within the statutory range. 42 U.S.C. § 7413(e)(1). We express no opinion about the extent to which a reduction in the number of violations should affect, if at all, the total penalty imposed. * * * We VACATE the district court’s judgment and REMAND for the limited purpose of allowing the district court to make additional findings on traceability and the Act of God defense. The case will then be returned to this panel. ANDREW S. OLDHAM, Circuit Judge, concurring in part, dissenting in part, and concurring in the judgment: I agree that the district court’s judgment must be vacated, and that we must remand the case for additional proceedings. I write separately to emphasize that our precedents in this area are a mess. The majority admirably attempts to make sense of them. But I’m afraid that task is too big for any panel. Eventually, our en banc court should clean up this confusion. I. The mess started in 1990. That’s when the Third Circuit created a three- part standard for determining traceability in citizen suits under the Clean Water Act. See Pub. Interest Research Grp. of N.J., Inc. v. Powell Duffryn Terminals Inc., 913 F.2d 64, 72 (3d Cir. 1990). Under Powell Duffryn, the plaintiff must show: a defendant has 1) discharged some pollutant in concentrations greater than allowed by its permit 2) into a waterway in which the plaintiffs have an interest that is or may be adversely affected by the pollutant and that 3) this pollutant causes or contributes to the kinds of injuries alleged by the plaintiffs. Ibid. It’s not obvious how the Third Circuit devised that standard because the court cited nothing at all to support it. See ibid. For better or worse, we relied on Powell Duffryn‘s ipse dixit in Sierra Club, Lone Star Chapter v. Cedar Point Oil Co. Inc., 73 F.3d 546, 557–58 (5th Cir. 1996). Oddly, our Cedar Point decision did not even cite—much less analyze—the Supreme Court’s canonical decision on Article III standing in Lujan v. Defenders of Wildlife, 504 U.S. 555 (1992). Much of Cedar Point contravenes Lujan and its progeny. Compare, e.g., Cedar Point, 73 F.3d at 556– 57 (embracing the plaintiffs’ theory of “threatened injury” without discussing the requirement that it be “concrete,” “imminent,” and “certainly impending”), with Lujan, 504 U.S. at 567 nn.2–3 (requiring plaintiffs to establish a “concrete injury” that was “imminent” and “certainly impending”), and Clapper v. Amnesty Int’l USA, 568 U.S. 398, 409 (2013) (similar). But even when Cedar Point relied on the Third Circuit’s standard rather than Supreme Court precedent, it recognized that “a literal reading of Powell Duffryn may produce results incongruous with our usual understanding of the Article III standing requirements.” 73 F.3d at 558 n.24. Therefore, Cedar Point utilized a more amorphous approach to Article III standing that turned on things like whether the relevant waterway is “large.” Ibid. The third prong of the Powell Duffryn-Cedar Point standard is the most pernicious. It says that the plaintiff need only prove that the relevant pollutant “causes or contributes to the kinds of injuries alleged by the plaintiffs.” Powell Duffryn, 913 F.3d at 72 (emphasis added); see also Cedar Point, 73 F.3d at 557 (adopting this ipse dixit). That eliminates traceability altogether. Think about it. Would we ever say: my house burned down; arsonists burn down houses; therefore, an arsonist burned down my house? Of course not. My house could have burned down because the wiring was faulty, I left the stove on, my dog tipped over a candle, a bolt of lightning struck the roof, a litterbug’s cigarette started a wildfire, or myriad other potential causes. Therefore, we would require some sort of allegation of but-for causation linking the fire to the arsonist. But Powell Duffryn-Cedar Point eliminates that. It’s enough to say that someone has asthma; pollutant X can cause asthma; therefore, pollutant X caused someone’s asthma. Cum hoc ergo propter hoc. Rather than recognize the fallacies inherent in Powell Duffryn-Cedar Point, some courts have extended its illogic. For example, the Fourth Circuit relied on both cases to hold that the “fairly traceable” standard is “not equivalent to a requirement of tort causation.” Friends of the Earth, Inc. v. Gaston Copper Recycling Corp., 204 F.3d 149, 161 (4th Cir. 2000) (quotation omitted). It is unclear what that sentence is supposed to mean. I suppose it could mean that a plaintiff can establish traceability without establishing the tort requirement of proximate causation. That’s true. See, e.g., Lexmark Int’l, Inc. v. Static Control Components, Inc., 572 U.S. 118, 134 n.6 (2014) (“Proximate causation is not a requirement of Article III standing, which requires only that the plaintiff ‘s injury be fairly traceable to the defendant’s conduct.”). But some of our sister circuits also have eliminated but-for causation. For example, the Sixth Circuit said: “In the nebulous land of ‘fairly traceable,’ where causation means more than speculative but less than but-for, the allegation that a defendant’s conduct was a motivating factor in the third party’s injurious actions satisfies the requisite standard.” Parsons v. U.S. Dep’t of Justice, 801 F.3d 701, 714 (6th Cir. 2015). That’s much like Powell Duffryn, which distinguished the “requirement of tort causation” from “scientific certainty.” 913 F.2d at 72. And much like Powell Duffryn, Parsons cited exactly nothing to support its ipse dixit. Also much like Powell Duffryn, Parsons makes little sense. Asking whether a particular allegation is “more than speculative but less than but-for,” Parsons, 801 F.3d at 714, is like asking whether a particular product is more than a preponderance but less than defective. The first is the plaintiff ‘s burden; the second is the actual fact that the plaintiff must prove. Our circuit already has recognized the constitutional problems posed by the ever-growing mountain of ipse dixits and logical fallacies that comprise the Powell Duffryn-Cedar Point-Gaston Copper-Parsons line of cases. Cedar Point itself recognized its approach could generate “results [that are] incongruous with our usual understanding of the Article III standing requirements.” 73 F.3d at 558 n.24. And we’ve reiterated that concern in other cases. For example, we’ve held that plaintiffs could not establish traceability by showing that they “use[d] a body of water located three tributaries and 18 miles ‘downstream’ from” a polluting refinery. Friends of the Earth, Inc. v. Crown Cent. Petroleum Corp., 95 F.3d 358, 361 (5th Cir. 1996); see also ibid. (“We are persuaded that this case presents a situation in which Powell Duffryn‘s focus on the plaintiff ‘s interest in the ‘waterway’ into which unlawful pollution flows passes Article III bounds.”). We’ve also held that the Powell Duffryn-Cedar Point-Gaston Copper-Parsons standard must be limited to “a case involving a small body of water, close proximity, well-understood water currents, and persistent discharges.” Ctr. for Biological Diversity v. EPA, 937 F.3d 533, 545 (5th Cir. 2019). And although we have at least one Clean Air Act case that describes Cedar Point as “the law in this Circuit,” Texans United for a Safe Economy Education Fund v. Crown Central Petroleum Corp., 207 F.3d 789, 793 (5th Cir. 2000), apparently we’ve never applied the Powell Duffryn-Cedar Point-Gaston Copper-Parsons traceability standard to air pollution. (Indeed, it’s quite awkward to do so because that standard turns on whether the relevant waterway is “large”—an adjective that makes no sense as applied to air.) II. Today’s panel confronted the unenviable task of making sense of the Powell Duffryn-Cedar Point-Gaston Copper-Parsons line of cases; squaring it with Crown Central, Center for Biological Diversity, and Texans United; and then attempting to reconcile the whole mess with Article III of the Constitution. I admire the majority’s efforts to square the circle. In my view, however, it’s impossible. A. Let’s start with the common ground between the majority and me. We all agree that the Powell Duffryn-Cedar Point-Gaston Copper-Parsons line of cases allows for something less than “tort-like . . . proximate cause” to establish traceability. See ante, at 13 & n.4. Of course, that does not excuse the plaintiffs from proving but-for causation. After all, the Supreme Court long has held that a plaintiff ‘s injuries are not fairly traceable without but-for causation. See, e.g., Allen v. Wright, 468 U.S. 737, 758 (1984) (holding that plaintiffs failed to establish traceability because “it is entirely speculative . . . whether withdrawal of a tax exemption from any particular school would lead the school to change its policies”—that is, whether the tax exemption is the but-for cause of plaintiffs’ injuries); Duke Power Co. v. Carolina Envtl. Study Grp., Inc., 438 U.S. 59, 74–78 (1978) (holding that “a ‘but for’ causal connection” between plaintiff ‘s injury and defendant’s act sufficed for traceability); Warth v. Seldin, 422 U.S. 490, 505 (1975) (holding that Article III requires plaintiffs “to establish that, in fact, the asserted injury was the consequence of the defendants’ actions”); see also Richard H. Fallon, Jr., Of Justiciability, Remedies, and Public Law Litigation: Notes on the Jurisprudence of Lyons, 59 N.Y.U. L. REV. 1, 17 n.91 (1984) (observing that the Supreme Court’s causation analysis “replicate[s] the tort law concept of ’cause in fact’ or ‘but for’ causation”). As have some of our sister circuits. See, e.g., Cmty. Nutrition Inst. v. Block, 698 F.2d 1239, 1247 (D.C. Cir. 1983) (“A plaintiff need only make a reasonable showing that ‘but for’ defendant’s action the alleged injury would not have occurred.”); cf. Honeywell Int’l, Inc. v. EPA, 705 F.3d 470, 472 (D.C. Cir. 2013) (Kavanaugh, J.) (“Honeywell’s injury is fairly traceable to the now-permanent 2008 interpollutant transfers by Arkema and Solvay because the injury would not have occurred but for the 2008 transfers.”). Even the Third Circuit—which gave us the unfortunate Powell Duffryn standard—says the traceability “requirement is akin to ‘but for’ causation in tort and may be satisfied even where the conduct in question might not have been a proximate cause of the harm.” Finkelman v. Nat’l Football League, 810 F.3d 187, 193 (3d Cir. 2016) (quotation omitted). Friends of the Earth, Inc. v. Laidlaw Environmental Services (TOC), Inc., 528 U.S. 167 (2000), obviously says nothing to the contrary. The Court in that case did not say anything at all about but-for causation. And it certainly didn’t purport to overrule any of its traceability cases that applied but-for causation— like Allen v. Wright or Duke Power. Reasonable people might disagree—as the Justices did in Laidlaw—over whether Friends of the Earth proffered sufficient evidence of traceability. See ante, at 13 n.4 (summarizing the disagreement). But that doesn’t mean Laidlaw announced sub silentio a new legal standard. See Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 91 (1998) (“We have often said that drive-by jurisdictional rulings of this sort . . . have no precedential effect.”). So the but-for causation requirement remains part of the irreducible constitutional minimum for Article III traceability. B. So where do I diverge from the majority? First, it’s not obvious to me that the Powell Duffryn-Cedar Point-Gaston Copper-Parsons doctrine should apply at all. As noted above, Cedar Point turns on whether a particular body of water is “large.” 73 F.3d at 558 n.24. Whatever sense that might make in water- pollution cases, it makes little or none in air-pollution cases. Second, even if the Powell Duffryn-Cedar Point-Gaston Copper-Parsons doctrine applies, the en banc court should revisit it. The district court already has conducted two trials in this case. We are remanding for a third. And there’s no guarantee that the third time will be a charm. One problem with an inherently indeterminate doctrine—like Powell Duffryn and its progeny—is that it cannot generate predictable results the first time around. To illustrate, consider a hypothetical plaintiff Bob who lives in Baytown. Bob has asthma—that is, an injury. The question is whether his asthma injury is traceable to Exxon’s illegal emissions. From January 1 through January 10, Bob was visiting his sister in France. Meanwhile: On January 2, Exxon emitted pollutants that “could have reached beyond the Exxon complex into the offsite areas of Baytown where Plaintiffs’ members lived and recreated.” Ante, at 16. On January 5, Exxon released “pollutants in excess of nonzero emissions limits” or that constituted “a ‘reportable quantity’ under state regulations.” Ante, at 16. On January 8, Exxon emitted pollutants “that could have caused or contributed to flaring, smoke, or haze, even if the emission was of a small magnitude.” Ante, at 17. Can Bob recover for these emissions? Obviously, our decision today doesn’t say. And I don’t know the answer under the Powell Duffryn-Cedar Point-Gaston Copper-Parsons line of cases. Nor do I envy the district court, which will have to hazard a guess. The only thing I know for sure is that Article III’s traceability requirement bars Bob from recovering a penny. * * * The Powell Duffryn framework cannot be squared with Article III’s traceability requirement. Cedar Point‘s half-hearted adoption of Powell Duffryn fares little better. Although in many cases we may be able to distinguish and disapply Cedar Point, our ability to faithfully apply Article III should not turn on whether a lawsuit involves a lake, a Gulf, or a smokestack. At some point, our en banc court should bring our precedent in line with the Constitution.
