dissenting:
Once the majority frames this case as a choice between compliance with the Clean Water Act (“CWA”) and tearing down the dams along the Snake and Columbia Rivers, the question answers itself. The trouble is that this formulation misstates the actual legal issue: whether evidence in the record supports the United States Army Corps of Engineers’ (“Corps”) decision that the sole cause of temperature excee-dences is the existence — and not operation — of the dams, and that, therefore, the *1181Corps bears no obligation to comply with the CWA.
Even talking about removal of the dams is a lightning rod that we need not strike. Compliance with the CWA and the continued presence of the dams are not mutually exclusive options. But, in an effort to sidestep the CWA, the Corps hides behind removal of the dams and simply defaults on the real issue — compliance with water quality standards. Because the record is devoid of evidence addressing operational alternatives aimed at CWA compliance, the Corps’ decision does not comply with the Administrative Procedure Act (“APA”). I respectfully dissent because, in my view, the Corps’ failure to tackle the CWA issue head-on requires remand.
The Corps’ decision is deficient in two respects. First, even where the studies in the record discuss water temperature ex-ceedences, the studies generally presuppose that thé only options are either having a dam or not having a dam. This binary decision tree ignores the consideration that dam operations — and not just the existence of the dams — impact water temperatures, thereby setting up the Corps’ too-convenient conclusion that there are no measures it can undertake to comply with the CWA. The record does not, however, support the conclusion that the options are simply take-it-or-leave-it because it contains little investigation into operational alternatives addressing the CWA. Put another way, the Corps’ presumption about the causes of water quality violations forecloses any meaningful consideration of its options.
Second, the scant evidence regarding operational alternatives is of limited utility because it does not address the right question. The studies purportedly supporting the Corps’ decision all focus specifically on compliance with the Endangered Species Act (“ESA”), not the CWA. The problem is that there is not a perfect overlap between requirements under the CWA and the ESA. As a consequence, the Corps benchmarks its CWA compliance under the ESA, which is the wrong statutory reference. Without any assurance that the Corps considered the universe of operational alternatives appropriate for CWA compliance, there is no rational basis for the Corps’ conclusion that efforts to reduce water temperatures would be futile. See Envtl. Def. Ctr., Inc. v. Envtl. Protection Agency, 344 F.3d 832, 858 n. 36 (9th Cir.2003) (noting that an agency decision violates the APA if there is no rational connection between the decision and the facts in the record).
In upholding the Corps’ decision, the majority inverts a fundamental principle of review under the APA. Rather than requiring a rational connection between the agency decision and the evidence in the record, the majority allows an agency decision to stand simply because it perceives an absence of conclusive opposing evidence, even when no supporting evidence exists. The majority also improperly places the burden on the National Wildlife Federation to uncover contrary evidence when the APA makes clear that it is the agency’s responsibility to consider evidence in the record and proceed on a rational and reasoned basis. Marsh v. Or. Natural Res. Council, 490 U.S. 360, 378, 109 S.Ct. 1851, 104 L.Ed.2d 377 (1989). Such logic eviscerates the APA’s function of ensuring that agency actions are based on rational and reasoned decision-making.
Under the APA’s familiar standard of review, the Corps’ decision regarding compliance with state water quality standards will not be reversed unless the action is arbitrary, capricious, an abuse of discretion, or otherwise contrary to law. 5 U.S.C. § 706(2); see also Lands Council v. Powell, 379 F.3d 738 (9th Cir.2004); Or. Natural Res. Council v. United States *1182Forest Serv., 834 F.2d 842, 851-52 (9th Cir.1987) (holding that claims of federal agency violations of state water quality standards under the CWA are reviewed under the APA). An agency decision is arbitrary and capricious if it fails to “articulate a rational connection between the facts found and the conclusions made.” Envtl. Def. Ctr., 344 F.3d at 858 n. 36; see also Natural Res. Def. Council v. United States Dep’t of Interior, 113 F.3d 1121, 1123-24 (9th Cir.1997). “This standard necessitates a judicial examination of the disputed decision’s rationale and surrounding circumstances in order to carry out the demand that courts ensure that agency decisions are founded on a reasoned evaluation of the relevant factors.” Desert Citizens Against Pollution v. Bisson, 231 F.3d 1172, 1180 (9th Cir.2000) (internal quotation marks omitted) (quoting Marsh, 490 U.S. at 378, 109 S.Ct. 1851).
With this standard in mind, I turn to the Corps’ 2001 Record of Decision (“ROD”). The Corps drafted the 2001 ROD in response to the district court’s prior decision that its 1998 ROD failed to respond adequately to the CWA.1 Nat’l Wildlife Fed’n v. United States Army Corps of Eng’rs, 132 F.Supp.2d 876, 890 (D.Or.2001). Since the time the Corps drafted the 1998 ROD, the additions to the record that discuss water temperature are studies that address fish survival with respect to the ESA.2 The majority and I agree that the district court improperly determined that the Corps solved its earlier failure to address the CWA by relying on its compliance efforts under the ESA. Maj. op. 1171. Water quality compliance is a substantively different issue than the survival of salmon and steelhead fish, and, despite areas of overlap, an effort to comply with one is not a substitute for compliance with the other. The question, then, is whether the ESA studies, or any other evidence in the record, provide a rational basis for the Corps’ decision that its activities are not captured by the CWA and that, consequently, it does not need to do anything to lower water temperatures.
After reviewing the data referenced in the Corps’ ROD, I cannot agree that the Corps’ decision is rationally connected to the factual record. Missing in the record is any evidence that the Corps’ operation of the dams is not one of the causes of temperature exceedences — which is the decision we are reviewing. The studies in the record — both before and after the 1998 ROD — analyzed water temperature excee-dences in comparison to water temperatures without the dams, or analyzed alternative practices without the CWA in mind. Because the record does not include discussion of dam operations vis-a-vis the CWA, the record cannot be read to support the claim that operations are irrelevant to CWA compliance.
The majority points to the EPA’s 1999 study of temperatures in the Columbia River as a foundation for the Court’s decision. Important to the majority’s analysis, the EPA study looked at water temperatures relative to the temperatures if the dams did not exist dr were removed entirely, leading the majority to conclude that *1183the study supports the Corps’ “contention that it is the existence of the dams that is causing temperature exceedences” as opposed to the operation of the damg. Maj. op. 1177. The EPA study cannot, however, serve as the critical basis for the Corps’ decision because the study simply does not address the relationship between dam operation methods and the water temperature. The data from the EPA study, like the data from all the studies in the record, merely indicate a fact the Corps readily concedes: there is a causal link between the water temperature exceedences and the existence of the dams.3 But the Corps stopped short in its analysis. It is arbitrary to infer that just because the mere existence of the dams affects water temperatures, somehow the operation of the dams does not also affect water temperatures.
Under the Corps’ circular reasoning, it need not consider operational alternatives because "the existence of the dams causes exceedences, and because the existence of the dams causes exceedences, no operational alternatives are possible, and it need not investigate them. This myopic vision of its options — keep the dams or tear them down — leads to the- Corps’ speculative conclusion that it cannot undertake any operational changes to reduce water temperature exceedences because, under the Corps’ understanding, why investigate operational alternatives if it already knows the dams’ existence is the sole cause of water temperature exceedences? The result is that the Corps ruled out alternative operational measures without ever investigating the CWA question. The Corps’ approach can only be described as an illogical leap, hardly a basis to pass muster under the APA.
Importantly, the various studies” discussion of alternatives to the current dam operation method are of limited use in answering the CWA question because the studies were drafted for a very different purpose — to promote salmon recovery under the ESA. The authors of the various EISs gave only secondary consideration to the CWA, and only to the extent that this concern arose from efforts to comply with the ESA, such as lowering water temperature enough for fish viability. Because of this focus, the environmental surveys only explored alternatives that promote fish survival and did not discuss in any detail questions related to the Corps’ obligations under the CWA
There is no way of knowing from the record whether the operational -alternatives the Corps considered in its effort to preserve fish represent the proper range of operational alternatives for CWA compliance; ■ In turn, there is thus no way of knowing whether there are operational changes the Corps can undertake to ameliorate water temperatures for purposes of the CWA, even if those efforts are less attractive for purposes of the ESA. To be sure, the APA does not require the Corps to conduct a study premised solely on the CWA. What the APA does require is rational decision-making, which in turn de*1184mands that the Corps demonstrate consideration of its CWA obligations and actions specific to compliance with that statute.
I also disagree with the majority’s insistence that requiring more from the Corps would repeal other statutes, such as those that provide for the construction and maintenance of the dams. This interpretation assumes that compliance is possible only if the dams cease to exist. This interpretation rests on the same speculation that puts the Corps’ conclusions on shaky ground. As discussed above, the record does not establish that the only way the Corps can comply with the CWA is if it tears down the dams. The options are so starkly limited because the record upon which the Corps based its decision suffers from a gap in data and analysis, a gap resulting from its singular focus on one option — the removal of the dams. The Corps’ blunt analysis hides fair consideration of positive intermediate steps, setting up a false Hobson’s choice between compliance and evisceration of other legislation. The point is that without some investigation into operational alternatives for CWA purposes, we do not know the extent to which operational alternatives impact water temperatures, nor do we know whether those alternatives could be practically implemented.
The majority’s attempt to whittle down evidence presented by National Wildlife— rather than discuss evidence supporting the agency decision — loses sight of the APA test, which requires an affirmative showing of evidence supporting a decision, not a negative absence of evidence approach. I do not dispute the majority’s rejection of certain operational alternatives, such as the natural river method; certainly, the Corps has discretion to interpret data and rely on its own qualified experts. See Marsh, 490 U.S. at 378, 109 S.Ct. 1851. But that issue is a red herring. The real issue is whether the Corps fulfilled its obligation under the APA of demonstrating a rational connection between its decision and evidence in the record. Envtl. Def. Ctr., 344 F.3d at 858 n. 36. Without record evidence linking the CWA standards to the Corps’ conclusions, notably the lack of evidence regarding the CWA and the impact of dam operations or the viability of dam operation alternatives, I conclude that the Corps’ decisions cannot pass APA scrutiny. I respectfully dissent.
. Prefacing the ROD, the Corps wrote: "Having found that the administrative record did not establish that the Corps considered all relevant factors in making the [prior] Record of Decision, the court ordered the Corps to issue a new decision to replace the 1998 Record of Decision which addresses compliance with its legal obligations under the Clean Water Act.”
. The major additions to the administrative record include a 1999 study of water temperatures in the Columbia River conducted by the Environmental Protection Agency ("EPA”), and a later 1999 environmental impact statement ("EIS”) addressing the way the dams along the Lower Snake River affect certain fish species.
. The author of the EPA's 1999 study concluded that the dams operated by the Corps significantly increase the magnitude and duration of water temperature exceedences in the Snake River, observing "In the Snake River, with the dams in place, duration of exceedance is relatively high at the starting point ..., but nearly doubles [down the river] ... [and without the dams] the analysis predicts that the mean duration of exceedance [down the river] is approximately 63% of that when the dams are in place.” The Corps acknowledged and accepted the EPA’s conclusions in its Í999 Lower Snake River EIS. Remarkably, the Corps does not dispute that the dams contribute to the duration and magnitude of temperatures exceeding the limits allowed under Washington’s water quality standards.