dissenting from Denial of Rehearing En Banc,
joined by DENNIS and GRAVES, Circuit Judges:I respectfully dissent because the panel’s opinion, in my view, independently weighs facts to render judgment in violation of fundamental principles of federal law.1 The Supreme Court' has reversed *326this Court before for improperly reweighing the factual findings of district courts de novo in violation of Federal Rule of Civil Procedure 52. Pullman-Standard v. Swint, 456 U.S. 273, 290, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982). In Pullman-Standard, the Court emphasized that it is an “elementary” principle of our system of justice that, as between district judges and the appellate bench, “[fjactfinding is the basic responsibility of district courts, rather than appellate courts.” Id. at 291-92, 102 S.Ct. 1781 (quoting DeMarco v. United States, 415 U.S. 449, 450 n. 94, 94 S.Ct. 1185, 39 L.Ed.2d 501 (1974)). There, the Supreme Court chastised our Court for arriving at independent findings on ultimate facts, noting that “where findings are infirm because of an erroneous view of the law, a remand is the proper course.” Id. at 292, 102 S.Ct. 1781. The panel’s opinion makes this same mistake again.
Moreover, this decision is not the only recent panel decision in our Circuit to draw a dissent for arguably reweighing facts in the first instance. Other members of our Court have recently expressed the same concern in other cases. See, e.g., Berezowsky v. Ojeda, 765 F.3d 456, 476 (5th Cir.2014) (Haynes, J., dissenting) (“Because ... the majority opinion fails to give the necessary deference to the district court’s finding, and instead engages in its own independent weighing of the facts, I respectfully dissent.”); Planned Parenthood of Greater Tex. Surgical Health Servs. v. Abbott, 769 F.3d 330, 335 (5th Cir.2014) (Dennis, J., dissenting) (“[A]l-though purporting to apply clear error review, as required by Supreme Court and circuit precedents, the panel improperly reviewed the district court’s factual findings de novo, and thereby erroneously substituted the panel’s own fact finding for that of the district court.”).2
If uncorrected by this Court en banc or the Supreme Court, this decision, and others like it, sends a clear message to litigants: if you don’t like the factual findings of a district court, the doors of our Court are wide open to endless retrials on appeal. This is the wrong message to send, and it evinces an alarming lack of trust in the work of our colleagues in the district courts. As the Supreme Court has observed: “The trial judge’s major role is the determination of fact, and with experience in fulfilling that role comes expertise.” Anderson v. City of Bessemer City, N.C., 470 U.S. 564, 574-75, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985). Indeed, “[djuplication of the trial judge’s efforts in the court of appeals would very likely, contribute only negligibly to the accuracy of fact determination at a huge cost in diversion of *327judicial resources.” Id. By proceeding in essence de novo, the panel’s opinion effectively second-guesses the district court’s ultimate conclusion, rather than evaluates the decision-making process for clear error. I hope that future panels do not follow this dangerous path.
The reweighing of facts in this case is particularly egregious. The trial judge made specific credibility determinations after an eight-day bench trial that included ten expert witnesses. See Anderson, 470 U.S. at 575, 105 S.Ct. 1504 (“When findings are based on determinations regarding the credibility of witnesses, Rule 52(a) demands even greater deference to the trial court’s findings.... ” (emphasis added)). The trial judge found The Aransas Project’s witnesses credible and disbelieved the defendants’ and intervenors’ witnesses — and for good reason: The Aransas Project’s witnesses included a shared recipient of the 2007 Nobel Prize for his work as an environmental scientist, holders of endowed chairs at prestigious national universities, MacArthur Fellows, and authors of numerous scientific papers in respected journals. In contrast, the other side’s expert witnesses had limited experience and insignificant expertise — indeed, one of them admitted he “made up” key portions of his testimony. As the district court observed:
[I]n most instances it was established that [intervenor] GBRA selected the data for which its experts were to make a determination without regard to the peer reviewed published scientific data available. In particular, intervenors GBRA and SARA wholly financed what is called the SAGES report partially designed by Dr. Stephen E. Davis, who testified as a modeling expert, and Dr. R. Douglas Slack. This report did not include the Chavez-Ramirez or Stehn [published, peer-reviewed] research. In fact, it used a report by Dr. Slack’s graduate student Danielle Greer whose conclusions to the preferred food of whooping cranes was based on 90 plus hours of video of three crane areas. The Court watched all the videos and finds that they were either too blurred to see anything or non-demonstrative of any habit, feeding or otherwise. When subjected to peer review Greer’s conclusions were soundly criticized.
Dr. Slack testified that the whooping cranes had well developed supraorbital salt glands which rid the body of excess salt, making them capable of living in a salt water marsh with no freshwater. When pressed by the Court, he admitted that he had made up that entire statement.
Dr. Stroud, a veterinary pathologist, was offered to explain the Whooping Crane necropsy findings of another pathologist. His opinion that the carcass that showed an infection was not based on the original pathologist description but based on the original pathologist description that green fluid was ob-. served in a joint. To him the color green meant gangrene. This conclusion had no scientific merit but he kept insisting that when he saw green he thought of gangrene.
(emphasis added). The panel’s opinion simply discards these credibility determinations without explanation.
The panel’s ruling that the plaintiffs failed to show proximate cause as a matter of law — decidedly a question of fact3 — is *328equally puzzling. As discussed infra, to render judgment on this issue, the panel was required to conclude “the record permits only one resolution of the factual issue.” Pullman-Standard, 456 U.S. at 292, 102 S.Ct. 1781. Here, Dr. Sass4 testified that low freshwater inflows and high crane mortality are “causally correlated” and “in all cases of high mortality you have low river flow, no exceptions.” Dr. Kathy Winsor, an expert in statistics and the chair of the Statistics Department at Rice University, testified that she reviewed Dr. Sass’s study and confirmed “that there is a strong association between the freshwater inflows into San Antonio bay/Guadalupe estuary and AWB [5] crane mortality” that, together with the scientific explanation offered by The Aransas Project’s experts, supports a finding of causation. And it is undisputed that Texas Commission on Environmental Quality (TCEQ) has the authority to grant or deny permits to maintain freshwater inflows into the Aransas National Wildlife Refuge.
The causal connection between TCEQ’s failure to maintain freshwater inflows and a “take” of endangered whooping cranes is straightforward. Experts in the field testified that:
1.The TCEQ’s water diversions reduce- the flow of freshwater to critical whooping-crane habitat and thereby increase salinity.
2. Higher salinities reduce the population of blue crabs and wolfberries— the primary winter food sources for whooping cranes — in the whooping crane habitat.
3. In the winter of 2008-2009, wolfber-ry and blue crab populations were low and at least twenty-three whooping cranes died. Autopsies on two carcasses listed emaciation as one of the causes of death.
From this evidence, the district court reasonably concluded that the “actions, inac-tions and refusal to act by the TCEQ defendants proximately caused an unlawful ‘take’ of at least twenty-three (23) Whooping Cranes in the 2008-2009 winter in violation of the [Endangered Species Act].”
In Sierra Club v. Yeutter, 926 F.2d 429 (5th Cir.1991), we affirmed a district court that similarly found — after a four-day bench trial replete with expert testimony from dueling scientists — that the government caused a “take” of endangered red cockaded woodpeckers by allowing logging activities. The government’s timber policy had allowed loggers to cut down trees such that the trees rarely grew to over 80 years old. Id. at 432. But the woodpeckers preferred as habitat trees over 100 years old. Id. We held that the causal connection there — between lax regulations allow*329ing logging, reduction of available old-growth trees, evidence that woodpeckers preferred older trees, and a reduction in the population of woodpeckers — was sufficient to support a district court’s fact finding that a “take” had occurred. Id. at 438-39 (concluding that the Forest Service’s actions, by allowing the clear-cutting of two hundred feet of old-growth hardwood trees, “impair[ed] the [woodpeckers’] ‘essential behavioral patterns, including ... sheltering,’ and thus results in a violation of section 9” (citation omitted) (quoting 50 C.F.R. § 17.3)).
So too here. The connection in this case — between the TCEQ allowing freshwater to be diverted, a reduction in freshwater inflows, the increased salinity in the whooping crane habitat, and observed impaired feeding behavior — is similar to the causal connection in Sierra Club v. Yeutter. If the difference between 80- and 100-year-old trees can support a finding of a “take,” surely a district court — faced with emaciated crane corpses — could reasonably conclude that a reduction of freshwater inflows into the critical habitat of the AWB flock significantly impaired feeding behavior and proximately caused a “take” here.
Moreover, unlike in Sierra Club v. Yeut-ter, the increased mortalities of whooping cranes from reduced freshwater inflows were not only foreseéable — -they were in fact foreseen and anticipated. In 2007, two years before the “take” here, the International Recovery Plan for the Whooping Crane, prepared by the U.S. Fish & Wildlife Service in consultation with the Texas Parks & Wildlife Department, concluded that: “Freshwater inflows starting hundreds of kilometers inland, primarily from the Guadalupe and San Antonio rivers, flow into whooping crane critical habitat at Aransas; these inflows are needed to maintain proper salinity gradients, nutrient loadings, and sediments that produce an ecologically healthy estuary.” U.S. Fish & Wildlife Serv., International Recovery Plan Whooping Crane (Grus Americana) 20 (3d rev. Mar. 2007), available at http://www.fws.gov/southwesVes/ Documents/R2E S/Whooping_Crane_ Recovery_Plan_FINAL_21-July-2006.pdf. The Executive Director of the Texas Parks & Wildlife Department concurred on March 30, 2007. Id. at the page preceding page ‘i.’ Sierra Club v. Yeutter is more than just an on-point case. It compels a like result on the facts present here.
The panel also suffers from the misapprehension that proximate cause is a question of closeness between cause and effect: if the causal link is too attenuated, an appeals court can reverse, render, and rule that proximate cause is lacking as a matter of law, or so it goes. Not so. As the most recent Restatement makes clear, the term “proximate” cause is a misnomer because it “implies that there is but one cause — the cause nearest in time or geography to the plaintiffs harm — and that factual causation bears on the issue of scope of liability. Neither of those implications is correct. Multiple factual causes always exist, and multiple proximate causes are often present.” Restatement (Third) of Torts: Physical & Emotional Harm § 29, cmt. b (2010). The legal question of prdximate cause limits liability “to those harms that result from the risks that made the actor’s conduct tortious.” Id. § 29. “When defendants move for a determination that the plaintiffs harm is beyond the scope of liability [and proximate cause is lacking] as a matter of law, courts must initially consider all of the range of harms risked by the defendant’s conduct that the jury could find as the basis for determining that conduct tortious.” Id. cmt. d. “Then, the court can compare the plaintiffs harm with the range of harms risked by the defendant to determine whether a reasonable *330jury might find the former among the latter.” M6 It is beyond dispute that harms to downstream water users are among the risks of carelessly regulating upstream users.
On this record, a reasonable jurist could perhaps conclude under de novo review that the link between TCEQ’s regulations of upstream users and the emaciation of the whooping cranes in this ease was too attenuated to support a finding of proximate cause.7 But the same cannot be said under clear-error review. See Anderson, 470 U.S. at 575, 105 S.Ct. 1504. As the Supreme Court said in Anderson: “When a trial judge’s finding is based on his [or her] decision to credit the testimony of one of two or more witnesses, each of whom has told a coherent and facially plausible story that is not contradicted by extrinsic evidence, that finding, if not internally inconsistent, can virtually never be clear error.” Id. (emphasis added).
It is simply beyond the pale to say that any error by the district court in this case with respect to proximate cause was so “obvious and exceptional” that “the record permits only one resolution of the factual issue” so as to support this Court rendering judgment to the contrary. The panel’s revised opinion purports to apply clear-error review and correctly notes that remand would be the proper course “unless the record permits only one resolution of the factual issue.” Pullman-Standard, 456 U.S. at 292, 102 S.Ct. 1781. By reversing and rendering judgment, the panel embraces a much heavier burden — to find there is only one possible resolution of the factual issue of proximate causation' — as opposed to simply reaching a different conclusion analyzing the issue de novo. And yet, the legal analysis remains essentially unchanged from the de novo review in the original opinion. Thus, the panel’s opinion now does exactly what it accuses the district court of doing: stating the correct legal rule, but analyzing the question unconstrained by it.
I also disagree with the panel’s conclusion that the district court did not apply the correct legal standard. The district court repeatedly stated “ordinary requirements of proximate causation apply.” But even if the panel were right on this point, the appropriate remedy would be to remand for findings under the correct legal *331standard, not render judgment to the contrary.
In sum, the panel disregarded the district court’s credibility determinations and reweighed the evidence. In so doing, the panel rejected the testimony of MacArthur Fellows, a Nobel Laureate in the field of environmental science, and leading experts holding endowed, tenured faculty positions at national universities in favor of witnesses one of whom admitted that he “made up” key portions of his testimony. Its indefensible decision to render judgment is a clear violation of established Supreme Court case law.
Finally, the panel relegates the district court’s fine work over the course of more than three years — culminating in a 124-, page opinion after a two-week bench trial — to a handful of pages. The panel’s treatment of the district court in this case is regrettable.
As a former district judge, I hope our Court refrains from second-guessing our colleagues in the district courts on factual matters. The “trial on the merits should be the main event rather than a tryout on the road.” Anderson, 470 U.S. at 575, 105 S.Ct. 1504 (alteration and internal quotation marks omitted). Because I fear this Court may continue to independently reweigh factual findings absent intervention from this Court en banc or the Supreme Court, I respectfully dissent from the denial of rehearing en banc.
. I do not take issue with Parts I — III of the panel’s opinion, except to point out that— contrary to the panel opinion's wholly unsupported assertion — the district court did not *326uncritically adopt The Aransas Project's proposed factual findings verbatim. The district court instead made extensive factual findings of its own, even undertaking to watch ninety-plus hours of videotape, and weighed the credibility of the competing expert witnesses, as discussed infra. See also Anderson v. City of Bessemer City, N.C., 470 U.S. 564, 572-73, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985) (rejecting the argument that a district court’s adoption of "findings [prepared by a prevailing party should be therefore subjected] to a more stringent appellate review").
. Commentators have noticed that the Fifth Circuit has been out of step with other circuits and the Supreme Court in this regard. In Federal Standards of Review, Professors Childress and Davis remark: "[S]ome courts [of appeals], especially historically in the Fifth Circuit, considered themselves to be free to reject findings of ultimate fact.... Even for a finding on essentially a question of fact, these courts would 'proceed to make an independent determination’ on the 'ultimate issue for resolution.’ " 1 Steven Alan Childress & Martha S. Davis, Federal Standards of Review § 2.17 (4th ed.2010) (emphasis added) (footnotes omitted) (quoting East v. Romine, Inc., 518 F.2d 332, 339 (5th Cir.1975)).
. "The issue[ ] of proximate causation ... involved] application of law to fact, which is left to the factfinder, subject to limited review.” Exxon Co., U.S.A. v. Sofec, Inc., 517 U.S. 830, 840-41, 116 S.Ct. 1813, 135 L.Ed.2d 113 (1996) (emphasis added); accord *328Gutierrez v. Excel Corp., 106 F.3d 683, 687 (5th Cir.1997).
. Dr. Sass is the former Chair of Ecology and Evolutionary Biology at Rice University and is currently an Emeritus Professor there, a fellow of the Baker Institute of Public Policy, the author of 165 peer-reviewed papers (including one on whooping cranes), and a shared recipient of the 2007 Nobel Prize for his work as an environmental scientist.
. The underlying issue in this case concerns the winter habitat of the only known wild flock of the endangered whooping crane (gms americana). In the 1940s, there were reportedly fewer than fifteen whooping cranes alive in the world — the species was essentially already extinct. Through decades of conservation efforts, there now exists one wild, self-sustaining flock of whooping cranes that at one point numbered almost 300. This appeal concerns the continuing vitality of this flock, known as the Aransas-Wood Buffalo (AWB) flock, named after its two critical habitats. In the summer the AWB flock inhabits the Wood Buffalo National Park in Canada and in the winter it migrates to the Aransas National Wildlife Refuge in Texas.
. The panel’s misapprehension is illustrated by the following example drawn from Dan B. Dobbs et al., The Law of Torts § 205 (2d ed.2014). Suppose a driver negligently speeds and the car is struck by a falling tree, injuring the passenger. Perhaps a tree falling is foreseeable. But the risk of a tree falling on top of the car is not the kind of harm that makes speeding negligent — if anything, speeding minimizes the time spent on the road exposed to falling trees. Thus, speeding cannot be said to have proximately caused the injury to the passenger.
Similarly, carelessly issuing a driver’s license increases the risk of unsafe drivers on the road. But drivers will strike trees — and endangered animals such as whooping cranes — regardless of how safe they are driving. Thus, contrary to the panel’s hypothetical, Op. at 820, negligently issuing a driver’s license does not increase the risk of hitting a tree, or of hitting an endangered animal for that matter.
The same cannot be said for TCEQ’s regulation of water. Diversions by upstream water users directly affect downstream users. Negligently allowing upstream users to divert water increases the risk that downstream users go without. If TCEQ — despite warnings— carelessly allows upstream users to divert water and a downstream user — such as an endangered species — is harmed by that conduct, then TCEQ’s inaction can reasonably be said to have proximately caused that harm.
. I for one would not conclude the district court erred under de novo review, however, in light of The Aransas Project’s strong expert testimony. But reasonable minds could perhaps differ on this point.