Skip to main content
Nossaman LLP


9th Circuit Gives Fish & Wildlife Wide Berth on Biological Opinion in Delta Smelt Decision

The Natural Lawyer

In a precedent-setting interpretation of the "best available science" requirement of the Endangered Species Act (ESA), the Ninth Circuit Court of Appeals upheld the 2008 biological opinion (BiOp) issued by the U.S. Fish and Wildlife Service (Service) regarding the effects of the state and federal water projects in California on the threatened delta smelt.  San Luis & Delta-Mendota Water Authority v. Jewell, 747 F.3d 581 (9th Cir. 2014).  The decision is the latest chapter in the long-running conflict over the water supply for 20 million Californians and the nation's most productive agricultural area.  The long-awaited decision reversed in part and affirmed in part the district court's judgment invalidating the BiOp and remanding it to the Service.  In a 2-1 decision, the panel reversed most aspects of the lower court's decision concerning the merits of the case under the ESA.  The panel affirmed the district court's order requiring the U.S. Bureau of Reclamation to prepare an environmental impact statement pursuant to the National Environmental Policy Act.  The California Department of Water Resources (DWR) and other water agency plaintiffs have petitioned for en banc review. 

The panel decision has two primary implications for federal transportation projects.  First, the decision weakens the ESA's mandate that wildlife agencies must use the best scientific and commercial data available when developing biological opinions.  Second, the decision allows wildlife agencies to develop reasonable and prudent alternatives (RPAs) to proposed agency actions without taking into account the impact of the RPAs on third parties. 

Best Available Science

The ESA requires federal wildlife agencies to "use the best scientific and commercial data available" when formulating a biological opinion, including in assessing the impact of the proposed activity and in developing possible RPAs.  16 U.S.C. § 1536(a)(2).  While the panel acknowledged, among other things, that "the BiOp is a bit of a mess" and "[t]he BiOp is a jumble of disjointed facts and analyses," the court ultimately upheld the BiOp.  747 F.3d at 605.  In doing so, and by applying a highly deferential standard of review, the panel affirmed agency actions that the panel effectively admitted did not comply with the ESA's best available science mandate. 

For example, the BiOp's RPA imposes severe flow restrictions during several months of the year, with the goal of reducing the number of delta smelt entrained at the Central Valley Project and State Water Project (Projects).  The district court concluded that the flow restrictions were invalid because they were not based on the best available science.  Plaintiffs argued that the Service improperly based its flow restrictions on raw salvage data, rather than population-adjusted ("normalized") data.  Court appointed experts and the Service's peer review panel concluded that the significance of 10 entrained delta smelt differs tremendously based on whether there are 100 or 1 million individuals in the species' population.  By failing to normalize the salvage data, the agency ignored the relative difference between these two scenarios.  The Ninth Circuit, however, applied a highly deferential standard of review and reversed the district court.  After admitting that the Service could have done more in determining Project flow limits, the court concluded that it was "within the FWS's discretion" to use raw salvage data rather than population-adjusted data because that decision led to a more "conservative" result, 747 F.3d at 607, intended to "protect the maximum absolute number of individual smelt," id. at 610.  In other words, despite recognizing that the Service imposed severe flow restrictions without accounting for the best available science, the panel nonetheless sustained the Service's decision. See id. at 608 ("That the FWS could have done more in determining OMR flow limits is uncontroverted."); id. at 610 n.24 (describing that court-appointed experts "believed the BiOp to have fallen short in this analysis" because of the failure to consider normalized data).

Similarly, the panel upheld the Service's determination with respect to a low salinity requirement in the western portion of the Delta.  This so-called "fall X2" measure requires the Projects to be operated in a manner that maintains X2—the point in the Delta at which salinity is less than two parts per thousand—at a certain distance from the Golden Gate Bridge.  As a practical matter, the fall X2 measure requires very large amounts of water to flow into the Pacific Ocean rather than being used for urban and agricultural uses.  The dispute over the calculation of X2 is of critical importance because it could determine the availability of up to 1 million acre feet of water per year for human consumption, an amount that would meet the annual water needs of several million people.

In developing the fall X2 measure, the Service used computer models developed by DWR to compare the effects of future Project operations against a simulated historical baseline.  However, the Service used one model to establish the baseline and a different model to predict the effects of future operations.  DWR provided evidence that the models were not designed to be compared and that the Service committed a major error by comparing model outputs without calibration of the models.  The Service nonetheless reached its conclusions regarding the fall X2 measure by "compar[ing] the two different models without discussing or accounting for the resulting bias."  747 F.3d at 617.  Despite this error, the panel upheld the Service's decision to compare the outputs of the two models on the grounds that the existence of a flaw does not require a determination that the action is unlawful.  Id. at 620 ("The fact that FWS chose one flawed model over another flawed model is the kind of judgment to which we must defer.").  However, accounting for biases inherent in the two models was not a "comparison of models," as characterized by the panel.  Rather it is a scientific practice; the prevailing practice is to calibrate model results to account for biases.  Nevertheless, the panel upheld the agency's determination. 

In the transportation context, the panel's decision not only allows federal wildlife agencies to impose highly restrictive RPAs on proposed federal actions, but it allows the agencies to do so without any demonstrable benefit to the at-risk species.  As explained by the Supreme Court, the "obvious purpose" of the best available scientific information requirement is to "ensure that the ESA not be implemented haphazardly, on the basis of speculation or surmise" and to "avoid needless economic dislocation produced by agency officials zealously but unintelligently pursuing their environmental objectives."  Bennett v. Spear, 520 U.S. 154, 176-77 (1997).  Yet, under the guise of applying a "highly deferential" standard of review, the panel affirmed agency actions based on clear errors in scientific judgment. 

Impacts on Third PartiesThe ESA provides that, if a biological opinion determines that a proposed agency action will jeopardize a listed species or adversely modify its critical habitat, the applicable wildlife agency must suggest RPAs that "can be taken by the Federal agency or applicant in implementing the agency action."  16 U.S.C. § 1536(b)(3)(A).  The regulations interpreting this provision define an RPA as:

"[A]lternative actions identified during formal consultation [i] that can be implemented in a manner consistent with the intended purpose of the action, [ii] that can be implemented consistent with the scope of the Federal agency's legal authority and jurisdiction, [iii] that is [sic] economically and technologically feasible, and [iv] that the Director believes would avoid the likelihood of jeopardizing the continued existence of listed species or resulting in the destruction or adverse modification of critical habitat."

50 C.F.R. § 402.02 (emphasis added). 

In drafting the BiOp, however, the Service developed an RPA without ever addressing—in the BiOp or anywhere else—whether the alternative was economically or technically feasible.  The district court accordingly held that the Service was required to reevaluate the BiOp to provide:

"[S]ome exposition in the record of why the agency concluded (if it did so at all) that all four regulatory requirements for a valid RPA were satisfied. . . .  ‘Trust us' is not acceptable.  FWS has shown no inclination to fully and honestly address water supply needs beyond the species despite the fact that its own regulation requires such consideration."

San Luis & Delta-Mendota Water Authority v Salazar, 760 F. Supp. 2d 885, 957 (E.D. Cal. 2010). 
The panel majority reversed, holding as a matter of law that a biological opinion is never required to address the economic and technological feasibility of alternatives to an agency action.  It reasoned that the regulation in question "is a definitional section; it is defining what constitutes an RPA".  747 F.3d at 635.  According to the majority, nothing provides that "FWS has required itself to provide an explanation [of feasibility] . . . when it lays out an RPA."  Id. at 636.

The panel majority also held that, although the BiOp did not need to consider the economic and technological feasibility of the RPAs, those requirements were in fact substantively satisfied.  Id. at 637.  The panel reached that conclusion by defining the feasibility requirement to preclude any consideration of the economic effect on third parties—in this case, the tens of millions of municipal, industrial, and agricultural users of Project water.  Id. at 637-638.  Rather, according to the majority, the requirement addresses only whether the proposed RPA "is financially and technologically feasible" for the agency, in contrast to whether the RPA will affect third parties.  Id. at 638.  On that basis, the majority held that it was irrelevant as a matter of law whether there were "downstream economic impacts" resulting from imposition of the RPA.  Id. at 637.  Instead, the panel found that it was "nearly self-evident" that the RPAs were feasible because "the RPAs do not require major changes" to Project operations.  Id. at 638. The potential impact of this decision on transportation projects—or any other federal project subject to an ESA section 7 consultation—cannot be overstated.  Licensed by the panel's decision not to consider economic and technological feasibility, and directed to exclude impacts on third parties when it does so, the federal wildlife agencies are free to impose onerous and impracticable burdens on governmental programs without exploring whether other measures would provide equal protection to endangered species at less draconian costs to the public. 

Further Review

DWR and other plaintiffs have filed petitions for rehearing en banc, seeking further review of the panel's decision.  The States of Nebraska, Alaska, Kansas, Oklahoma, South Carolina, and Wyoming jointly filed an amicus brief in support of the petitions. Four additional amici curiae briefs were also filed supporting en banc review. 

  • Professionals
  • Practices
  • Success Stories
  • News
  • Events
  • Resources
  • Firm Pages