San Luis & Delta-Mendota Water v. Natural Resources Defense Coun , 747 F.3d 581 ( 2014 )

ARNOLD, Circuit Judge,

concurring in part and dissenting in part:

I respectfully dissent from Parts III, IV.A, IV.B, IV.E, and V.B of the court’s opinion and concur in the rest of it. I address the issue dealt with in Part III as it arises in considering the merits of the challenges to the BiOp.

1. I do not believe that the district court erred in holding that the BiOp’s OMR flow limits were set arbitrarily and capriciously. First of all, I discern no error in admitting a portion of the declaration from Dr. Richard Deriso, who holds advanced degrees in mathematics and biomathematics, discussing the use of raw salvage data to justify the flow prescription. A decision to include evidence that is outside the administrative record is reviewed for abuse of discretion, see Lands Council, 395 F.3d at 1030 n. 11, and admitting this evidence fell within one of the narrow exceptions to the general rule against extra-record evidence, because it was necessary to explain technical terms or complex subject matter, see Nw. Envtl. Advocates v. Nat’l Marine Fisheries Serv., 460 F.3d 1125, 1145 (9th Cir.2006). Furthermore, Dr. Deriso’s declarations were consistent with advice offered by independent peer reviewers and draft notes of a delta smelt evaluation team at FWS assembled before the final BiOp issued, and with the testimony of Rule 706 experts Dr. Punt and Dr. Quinn, who recognized Dr. Deriso’s declarations and stated that the validity of the flow regimes specified in the BiOp was undermined by its incomplete analysis. Nor was there any “battle of the experts” here, as the court maintains, because the responses to Dr. Deriso’s declarations from FWS’s mathematical statistician, Dr. Ken Newman, were mostly vague, and he generally agreed with Dr. Deriso that salvage should be scaled by some measure of population abundance.

As for the merits of this issue, Appellants do not contend that the use of raw salvage data was scientifically acceptable; they maintain instead that the flow prescription also relied on and was supported by other information. Based on my review of this information, however, the BiOp did not connect it to flow limits at all, or there was no explanation for why it yielded the flow prescription that the BiOp specified. As to FWS’s use of normalized data in the ITS, I am not convinced that this is relevant to whether it was scientifically sound for FWS to use only raw salvage data to set the flow prescription. While certain DWR comments support the flow prescription, the parties do not dispute that these comments arose from the district court’s previous remedial imposition of such a prescription, which, as the district court noted, occurred before the court became aware that using raw salvage data was not accepted scientific methodology. Because FWS based its flow prescription solely on the unexplained use of raw salvage data, I believe that its expertise in methodological matters is not entitled to deference, since that use was not rationally connected to the best available science, see W. Watersheds Project v. Kraayenbrink, 632 F.3d 472, 493 (9th Cir.2011); and because FWS did not consider all relevant factors or articulate a rational connection between the facts found and the choices made, I agree with the district court that its determination as to the flow prescription was arbitrary and capricious, see id. at 496.

2. I also concur in the district court’s conclusion that the BiOp’s determination of X2 was arbitrary and capricious. First of all, there was no abuse of discretion in the district court’s decision to admit the declarations of Aaron Miller, a DWR technical engineer who worked closely with Reclamation to develop Calsim II, which were relevant to the Calsim II-Dayflow comparison. See Lands Council, 895 F.3d at 1030 n. 11. No battle of the experts was created by doing so: Miller’s assessment of the validity of the Calsim II-Dayflow comparison was consistent with the testimony of Rule 706 experts Drs. Quinn and Punt, who recognized Miller’s assessment; and the declarations of FWS hydrologist Derek Hilts — whose credentials and experience are similar to Miller’s, and who helped draft the BiOp — were mostly unresponsive to Miller’s declarations on the several sources of bias that the comparison introduced. In my view, the district court relied on this extra-record evidence simply to determine whether FWS had considered all relevant factors, here, the sources of bias, see Nw. Envtl. Advocates, 460 F.3d at 1145, before relying on the comparison. to analyze the effects of proposed Projects operations on smelt and its habitat, including X2’s location. Doing so was well within the court’s role. Because highly technical matters were involved, it was difficult to determine if FWS considered all relevant factors without looking outside the record to see what matters should have considered, but were not. See Inland Empire Pub. Lands Council v. U.S. Forest Serv., 88 F.3d 754, 760 n. 5 (9th Cir.1996). The district court could not properly discharge its duty to engage in “ ‘substantial inquiry’ ” by simply taking FWS’s word that it had considered all relevant matters. SeeAsarco, 616 F.2d at 1160.

FWS’s choice to use the Calsim II-Day-flow comparison was unsupported by the requisite reasoned analysis. See Ecology Ctr. v. Castaneda, 574 F.3d 652, 665 (9th Cir.2009). Comments received from DWR and other entities — which were echoed by the Rule 706 experts’ testimony — alerted FWS to the several sources of bias, yet the only explanation in the BiOp for using the comparison was that a Calsim II to Calsim II comparison that FWS had conducted did not show differences that were expected. As the district court noted, the record did not reflect that FWS considered, much less recognized, the sources of bias. Furthermore, there are significant differences between the two models, including how X2 positions are determined. FWS would have to address these significant differences in some way to obtain information on which it could reasonably rely to base the BiOp’s conclusions, including Action 4 on the management of X2’s location. FWS was required to provide some evidence supporting its conclusions to ensure that no clear error of judgment rendered its actions arbitrary and capricious. See League of Wilderness Defenders-Blue Mountains Biodiversity Project v. U.S. Forest Serv., 549 F.3d 1211, 1218 (9th Cir. 2008). Appellants do not dispute that the sources of bias existed, or that the biases were significant or material; and the clear purpose of requiring FWS to use the best scientific evidence available is to ensure that the ESA is not implemented haphazardly or based on surmise or speculation. See Pac. Coast Fed’n of Fishermen’s Ass’ns, 426 F.3d at 1094-95.

Finally, I agree with the district court that FWS did not sufficiently explain why 74 km and 81 km were selected as critical points for X2 to preserve smelt habitat. The district court was not therefore required to give FWS deference on this matter, as FWS’s reasoning could not be reasonably inferred from the record. See San Luis & Delta-Mendota Water Auth., 672 F.3d at 700. As to the other justifying evidence, it either does not relate to the choice of critical points or it is undisputed that it was offered for the first time in post-judgment proceedings. I am reluctant to pass over the absence of an adequate explanation in the administrative record by relying on Appellants’ post-hoc rationalizations. See Humane Soc’y of United States v. Locke, 626 F.3d 1040, 1049-50 (9th Cir.2010).

3. I find no authority requiring FWS to address specifically and analyze, in the BiOp or administrative record, the question of whether the RPA meets the non-jeopardy elements. I also question the district court’s reliance on Greenpeace v. Nat’l Marine Fisheries Serv., 55 F.Supp.2d 1248 (W.D.Wash.1999), for the proposition that there must be some explanation in the administrative record as to why FWS concluded that all four elements for a valid RPA were satisfied. Nonetheless, I would affirm the district court on this issue, because the record belies Appellants’ contention that DWR and Reclamation raised no concerns about the nonjeo-pardy elements. The record shows that concerns were raised relating to RPA feas-ability and its relationship to the action’s intended purpose (providing water for various uses), and possibly to DWR’s and Reclamation’s authority to implement the RPA. Thus, under FWS’s own interpretation of § 402.02, it was required to consider and address these elements specifically in the instant BiOp or administrative record.

L Finally, as the action agency, Reclamation could not rely solely on FWS’s BiOp to establish conclusively its compliance with its substantive obligations under ESA § 7, because it could not delegate its responsibility to see that its actions would not jeopardize smelt, see Pyramid Lake, 898 F.2d at 1415; and as the action agency it could not blindly adopt FWS’s conclusions because it is ultimately responsible for ESA compliance, see City of Tacoma, Wash. v. Fed. Energy Regulatory Comm’n, 460 F.3d 53, 76 (D.C.Cir.2006). I agree with the district court that Reclamation would be subject to independent ESA liability if it possessed new information not considered by FWS which challenged the BiOp’s conclusions, see Pyramid Lake, 898 F.2d at 1415, and that there is no indication that this occurred here. But the district court failed to consider another basis for finding an action agency independently liable, namely, reliance on a legally flawed BiOp. Discerning such flaws involves no technical or scientific expertise, so failure to do so may result in action based on reasoning not in accordance with the law thus rendering the action arbitrary and capricious. See Wild Fish Conservancy v. Salazar, 628 F.3d 513, 532 (9th Cir.2010). The district court’s legal conclusions necessarily arose from fact-finding, but the court clearly, and I believe correctly, concluded that FWS had not used the best available science or considered relevant factors, and had acted arbitrarily and capriciously, because, among other things, it relied on the Calsim II-Dayflow comparison, and did not use normalized salvage data to set the flow prescription. The district court therefore should have found Reclamation independently liable under ESA § 7 for accepting a legally flawed BiOp and immediately beginning implementation of the RPA by modifying operations.

In sum, I find no abuse of discretion in the district court’s limited admission of evidence outside the administrative record as relevant to the OMR flow limits and the determination of X2, including the use of the Calsim II-Dayflow comparison. I believe that in determining whether FWS’s decisions on these matters in the BiOp were arbitrary, capricious, or otherwise not in accordance with the law, the district court’s analysis was thorough and well-reasoned. While I disagree with the basis for the district court’s conclusion that the non jeopardy elements must be addressed in the BiOp or administrative record, I nonetheless believe that affirmance is warranted on this issue. Finally, I believe the district court should have found Reclamation independently liable under the ESA for relying on a legally flawed BiOp.