The opinion of the court was delivered by: Oliver W. Wanger United States District Judge
(DOCS. 548, 549, 550, 658, & 661)
MEMORANDUM DECISION MOTIONS FOR SUMMARY JUDGMENT
These consolidated cases arise out of the continuing war over protection of the delta smelt (Hypomesus transpacificus), an ESA-threatened species, and associated impacts to the water supply for more than half of the State of California. Plaintiffs, San Luis & Delta Mendota Water Authority ("SLDMWD") and Westlands Water District, Metropolitan Water District of Southern California, State Water Contractors ("SWC"), Coalition for a Sustainable Delta and Kern County Water Agency, Stewart & Jasper Orchards, Arroyo Farms, LLC, and King Pistacho Grove, and 3 Family Farm Alliance, move for summary judgment on their numerous remaining claims against the United States Fish and Wildlife Service's ("FWS") December 15, 2008 Biological Opinion addressing the impacts of the coordinated operations of the federal Central Valley Project ("CVP") and State Water Project ("SWP") on the threatened delta smelt (Hypomesus transpacificus). Doc. 550. Plaintiff-in-Intervention, the California Department of Water Resources ("DWR") filed a separate motion for summary judgment on narrower grounds. Docs. 548 & 549. Federal Defendants, the United States Department of the Interior, FWS, and the United States Bureau of Reclamation ("Reclamation"), and Defendant Intervenors, Natural Resources Defense Council and The Bay Institute, oppose and cross move for summary judgment on all remaining claims. Docs. 658 & 661. Plaintiffs and DWR replied.
Docs. 697 & 695. The motion came on for hearing on July 8 & 9, 2010. After oral argument, the parties submitted supplemental briefing on a limited set of issues. Docs. 746-49.
FWS's 2005 biological opinion ("2005 Smelt BiOp") found that the proposed coordinated operations of the SWP and CVP will have no adverse effect on the continued existence and recovery of the Delta Smelt and its critical habitat. The 2005 BiOp was remanded to FWS as arbitrary and capricious. Order, NRDC v. Kempthorne, 1:05-cv-1207 (E.D. Cal. May 25, 2007), Doc. 323. Following an extensive evidentiary hearing, the Court issued an interim remedial order and Findings of Fact and Conclusions of Law ("Findings"), which covered, among other things, the effects on delta smelt of negative flows in Old and Middle Rivers ("OMR"), two distributary channels of the San Joaquin River. See Interim Remedial Order Following Summary Judgment and Evidentiary Hearing ("Int. Rem. Order"), NRDC v. Kempthorne, Doc. 560 (Dec. 14, 2007); Findings re: Delta Smelt ESA Remand and Reconsultation ("Int. Rem. Findings"), NRDC v. Kempthorne, Doc. 561 (Dec. 14, 2007).*fn1
Reclamation and DWR were ordered, among other things, to implement a winter "pulse flow" in OMR of no more negative than -2,000 cubic feet per second ("cfs"), and to "operate the CVP and SWP to achieve a daily average net upstream (reverse) flow in the OMR not to exceed 5,000 cfs on a seven-day running average" during a defined period in the spring. Int. Rem. Order at 5-7; see also Int. Rem. Findings at 15-20.
FWS issued a new delta smelt biological opinion on December 15, 2008 ("2008 Smelt BiOp" or "BiOp"). See Administrative Record ("AR") at 00001-00411.*fn2 This BiOp concluded that proposed CVP and SWP operations are "likely to jeopardize the continued 3 existence of" the delta smelt and "adversely modify" its critical habitat. BiOp at 276-79. The BiOp includes a required Reasonable and Prudent Alternative ("RPA") designed to allow the projects' continued operations without causing jeopardy to the species or adverse modification to its critical habitat. Id. at 279-85. The RPA includes operational components designed to reduce entrainment of smelt during critical times of the year by controlling (limiting) water exports from the Delta by the Projects. Id. at 279-85.
Component 1, to protect of the adult delta smelt life stage, consists of two Actions related to OMR flows.
* Action 1, to protect upmigrating delta smelt, is triggered during low and high entrainment risk periods based on physical and biological monitoring. Action 1 requires OMR flows to be no more negative than -2,000 cfs on a 14-day average and no more negative than -2,500 cfs for a 5-day running average. Id. at 280-82, 329-51.
* Action 2, to protect adult delta smelt that have migrated upstream and are present in the Delta prior to spawning. Action 2 is triggered immediately after Action 1 concludes or if recommended by the Smelt Working Group ("SWG"). Flows under Action 2 can be set within a range from -5,000 to 3 -1,250 cfs, depending on a complex set of biological and environmental parameters. Id. at 281-82, 352-56.
Component 2 (Action 3), to protect larval and juvenile delta smelt, requires OMR flows to be kept between -1,250 and -5,000 cfs, after Component 1 is completed, when Delta water temperatures reach 12° Celcius ("C"), or when a spent female smelt is detected in trawls or at salvage*fn3 facilities. Id. at 282, 357-58. Component 2 continues until June 30 or when the Id. at Clifton Court Forebay water temperature reaches 25° C. 282, 368.
Component 3 (Action 4), to improve habitat for delta smelt growth and rearing, requires sufficient Delta outflow to maintain average mixing point locations of Delta outflow and estuarine water inflow ("X2"*fn4 ) from September to December, depending on water year type, in accordance with a specifically described "adaptive management process" overseen by FWS. Id. at 282-83, 369.*fn5
Component 4 (Action 6) (Habitat Restoration), requires DWR to create or restore 8,000 acres of intertidal and subtidal habitat in the Delta and Suisun Marsh within 10 years. Id. at 283-84, 379. 3 Component 5 (Monitoring and Reporting), requires Reclamation and DWR to gather and report information to ensure proper implementation of the RPA actions, achievement of physical results, and evaluation of the effectiveness of the actions on the targeted life stages of delta smelt, so that the actions can be refined, if needed. Id. at 284-85, 328, 375.
The first of the six consolidated challenges to the BiOp was filed on March 3, 2009. Doc. 1. Plaintiffs moved for a preliminary injunction on April 24, 2009 to prevent Reclamation from implementing Component 2 of the RPA, alleging that FWS violated the National Environmental Policy Act ("NEPA") and the ESA. See Doc. 31.
On May 22, 2009, the Court granted that motion in part, finding that Plaintiffs were likely to succeed on the merits of their NEPA claim and requiring FWS to make specific written findings to justify OMR flow restrictions. See Doc. 84; see also Doc. 94, Findings re Mot. for Prelim. Inj. (May 29, 2009). Defendants complied with that Order, submitting weekly notices of FWS's OMR flow decisions. See, e.g., Doc. 111, Notice of OMR 3 Flow Decision (June 11, 2009). The Court's May 2009 preliminary injunction ruling was not based on Plaintiffs' ESA claims. Doc. 94 at 43.
Plaintiffs amended their Complaint, joined and added claims against Reclamation, see Doc. 292, and moved for summary judgment on their NEPA claim, see Doc. 245. A November 13, 2009, ruling granted summary adjudication in part, based on Reclamation's failure to prepare an environmental impact statement before provisionally accepting and implementing the BiOp and its RPA Actions. Doc. 399.
Summary judgment for Defendants was granted on: (1) Stewart and Jasper Orchards' Commerce Clause claim that the ESA did not apply to protect delta smelt, a purely intra-state species, Doc. 339; and (2) claims that the BiOp violated regulations governing formulation of the RPA by not including required information in the BiOp text, Doc. 354.
Plaintiffs then filed three temporary restraining order motions over a six week period --- all of which were denied. See Docs. 555 & 583; see also 3/16/10 Hrg. Tr. at 86-88. Plaintiffs next sought a preliminary injunction against implementation of RPA Component 3. An evidentiary hearing was held from April 2, 2010 through April 7, 2010. Docs. 644, 652-54. Findings Re Plaintiffs' Request for Preliminary Injunction issued May 27, 2010 ("PI Decision"). Doc. 704. The PI Decision confirmed 3 Plaintiffs had succeeded on their NEPA claim and found Plaintiffs were likely to succeed on the merits of their ESA claim:
Although the premise underlying Component 2 -- that the species may be jeopardized by increased negative flows occasioned by export pumping -- has record support, FWS has failed to adequately justify by generally recognized scientific principles the precise flow prescriptions imposed by Component 2. The exact restrictions imposed, which are inflicting material harm to humans and the human environment, are not supported by the record, making it impossible to determine whether RPA Component 2 [is] overly protective. Judicial deference is not owed to arbitrary, capricious, and scientifically unreasonable agency action.
Id. at 122. Plaintiffs presented evidence under NEPA on the balance of the hardships that social dislocation, unemployment, and other threats to human health and safety were caused by interdiction of Plaintiffs' water supply. See id. at 123.
Countervailing irreparable harm was found, because "the species and its critical habitat are entitled to protection under the ESA." Id. at 124. Acknowledging the existence of legal and equitable grounds for injunctive relief, further evidence was requested on the "status of the species to assure that altered operations will not deepen jeopardy to the affected species or otherwise violate other laws." Id. at 125. Specifically, to establish "that Plaintiffs' proposed remedy of a flat -5,600 cfs ceiling on negative OMR flows will not jeopardize the continued existence of the species and/or adversely modify its critical habitat." Id. 3 A May 28, 2010 status conference sought to determine whether a mutually-agreeable interim operational plan could be implemented. Doc. 706. On June 22, 2010, the parties stipulated to a joint operational plan to maintain OMR flows so as not to be more negative than -5,000 cfs, unless certain, defined salvage triggers required a further reduction in OMR flows. Doc. 724.
After these dispositive motions were filed, the National Academy of Sciences, completed a comprehensive review of the BiOp, and concluded that the BiOp and the RPA Actions were "scientifically justified." See National Academy of Sciences, National Research Council, A Scientific Assessment of Alternatives for Reducing Water Management Effects on Threatened and Endangered Fishes in California's Bay Delta at 3. Doc. 635.
This post-decisional document is not part of the Administrative Record ("AR") and no legal justification exists to supplement the AR to include it.
Additionally, a scientific peer review panel was convened by the private consulting firm, Post Buckley Shuh and Jernigan ("PBS&J"), at the request of Plaintiff Family Farm Alliance ("FFA") in connection with FFA's administrative petition under the Information Quality Act ("IQA"). See Family Farm Alliance v. Salazar, 09-cv-1201 OWW-DLB (E.D. Cal.), Doc. 27, Ex. A. This document is part of the administrative record in the Family Farm Alliance IQA case, not the smelt AR. There is no basis to 3 consider this document for non-IQA claims.
III.STATUS OF THE SPECIES
The delta smelt was listed as a threatened species under the ESA on March 5, 1993. 58 Fed. Reg. 12,854 (March 5, 1993). Critical habitat was designated for the delta smelt on December 19, 1994. 59 Fed. Reg. 65,256 (Dec. 19, 1994). Once an abundant species in the Bay-Delta ecosystem as recently as thirty years ago, the delta smelt is now in imminent danger of extinction. PI Decision, Finding of Fact ¶ 10. All the evidence shows a significant decline in smelt abundance since 2000, recently up to three orders of magnitude below historic lows. Id. The latest fall mid-water trawl ("FMWT") abundance index for the species was 17, the lowest level ever recorded. Id.
On April 7, 2010, FWS announced that reclassifying the delta smelt from a threatened to an endangered species was warranted, but precluded by higher priority listing actions. 75 Fed. Reg. 17,667 (Apr. 7, 2010). The direct mortality of delta smelt by entrainment at the CVP-SWP pumps, as well as the destruction and adverse modification of its habitat in the Delta caused by water exports, were important factors in this determination. Id. at 17,669, 17,671 ("The operation of State and Federal export facilities constitute a significant and ongoing threat to delta smelt through direct mortality by entrainment"). As a result of the "immediate and high magnitude threats" confronting the 3 species, the delta smelt was assigned a listing priority number *fn6 of 2. Id. at 17,675.
Plaintiffs' motion advances the following grounds and contentions:
(1) FWS failed to rely on the "best available science" by making fundamental scientific errors in its analysis of the impacts of Project Operations on the species by:
(a) Relying on raw salvage numbers in quantitative impact analyses;
(b) Failing to conduct a life cycle analysis;
(c) Comparing the results of two entirely different, incompatible flow and salinity models; and
(d) Selectively excluding certain data for one purpose, but then unjustifiably using it for another;
(2) The BiOp's Project Effects Analysis is arbitrary and capricious because FWS:
(a) Assumed that Project operations drive hydrological conditions in the Delta and did not explain or justify this attribution;
3 (b) Evaluated the impacts of other (i.e., non-Project) stressors erroneously and inconsistently; and
(c) Improperly characterized summer food supply suppression, invasive species, and pollution and contaminants as indirect effects of Project Operations;
(3) The BiOp is arbitrary and capricious because it does not distinguish between discretionary and non-discretionary actions, improperly inflating the alleged effects of Project Operations;
(4) The BiOp's RPA is unlawful because FWS did not conduct the specific analyses required by the ESA and FWS' own RPA regulation, 50 C.F.R. § 402.02, because neither the BiOp nor the AR demonstrate that FWS analyzed or applied the first three (of four) § 402.02 factors;
(5) FWS illegally arrogated to itself Project operating authority in derogation of Reclamation and DWR;
(6) FWS acted arbitrarily and capriciously by disregarding the Information Quality Act ("IQA") when preparing and issuing the BiOp;
(7) FWS violated NEPA by not considering the environmental impacts of issuing the BiOp and RPA.
(8) Reclamation violated its legal duties by accepting FWS' inherently flawed BiOp.
DWR's attacks three aspects of the BiOp:
(1) By relying on a comparison of CALSIM II model runs with what the BiOp terms "historic" data (which was actually generated by the Dayflow model), the BiOp's analysis of the effects of the proposed action on smelt habitat does not yield meaningful information and violates the ESA's best available science requirement. This analysis further violates the APA because FWS did not adequately articulate any rational connection between the facts found based on these comparisons, and its conclusions regarding the Projects' effects on the smelt.
(2) Component 3 of the RPA, also referred to in the BiOp as Action 4,
is intended to mitigate the effects of the proposed action on smelt
habitat, by requiring the Projects to maintain X2 in specified
locations, depending on the type of water year. The BiOp, however,
lacks sufficient explanation as to the basis for the specific
prescriptions imposed by this Component, in violation of the APA.
Moreover, to the extent that the record reveals that these
prescriptions are based, even in part, on the methods used in the
effects analysis, they violate the ESA's "best
available science" mandate. (3) The Incidental Take Statement
("ITS") is defective. First, its estimates are based on the average
take from water years 2006 through 2008, which predicts the ITS will
likely be exceeded in half of all years. Second, FWS erroneously
misapplied its own data with the result that the BiOp claims that the
ITS was only exceeded in five of the previous sixteen years, rather
than accurately stating that it was exceeded in eleven of the sixteen
years. Third, the ITS take estimate is based on a data sample that is
too small to provide a reasonable prediction of take under the RPA.
These defects violate the ESA's "best available science"
requirement, the ESA's ITS requirements, and the APA.
Summary judgment is appropriate when the pleadings and the record demonstrate that "there is no genuine dispute as to any material fact and that the moving party is entitled to judgment as a matter of law." Fed. R. Civ. P. 56(c). The claims in this case involve FWS's issuance of a biological opinion, which is a final agency action subject to judicial review under the APA, 5 U.S.C. § 702. Nat'l Wildlife Fed'n v. Nat'l Marine Fisheries Serv., 524 F.3d 917, 925 (9th Cir. 2008) ("NWF v. NMFS II"). A court conducting judicial review under the APA may not resolve factual questions, but instead determines "whether or not as a matter of law the evidence in the administrative record permitted the agency to make the decision it did." Sierra Club v. ainella, 459 F. Supp. 2d 76, 90 (D.D.C. 2006) (quoting Occidental Eng'g Co. v. INS, 753 F.2d 766, 769 (9th Cir. 1985)). "[I]n a case involving review of a final agency action under the [APA] ... the standard set forth in Rule 56(c) does not apply because of the limited role of a court in reviewing the administrative record." Id. at 89. In this context, summary judgment becomes the "mechanism for deciding, as a matter of law, whether the agency action is supported by the administrative record and otherwise consistent with the APA standard of review." Id. at 90.
Administrative Procedure Act ("APA") invalidation of a biological opinion requires Plaintiffs to prove that FWS's action was "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." 5 U.S.C. § 706(2)(A).
APA review of a biological opinion is "based upon the evidence contained in the administrative record." Arizona Cattle Growers' Ass'n v. FWS, 273 F.3d 1229, 1245 (9th Cir. 2001). Judicial review under the APA must focus on the administrative 3 record already in existence, not some new record made initially in a reviewing court. Parties may not use "post-decision information as a new rationalization either for sustaining or attacking the agency's decision." Ass'n of Pac. Fisheries v. EPA, 615 F.2d 794, 811-12 (9th Cir. 1980). Exceptions to administrative record review for technical information or expert explanation make such evidence admissible only for limited purposes, and those exceptions are narrowly construed and applied. Lands Council v. Powell, 395 F.3d 1019, 1030 (9th Cir. 2005).
Here, as evidentiary rulings explained, see, e.g., Docs. 387, 392 (10/19/09 Hrg. Tr), 406, 407, 462, 740 (7/8/10 Hrg.), 750, expert testimony has been considered only for explanation of technical terms and complex scientific subject matter beyond the Court's knowledge; and to understand the agency's explanations, or lack thereof, and the parties' arguments.
(2)Deference to Agency Expertise.
A Court must defer to the agency on matters within the agency's expertise, unless the agency completely failed to address some factor, consideration of which was essential to making an informed decision. Nat'l Wildlife Fed'n v. Nat'l Marine Fisheries Serv., 422 F.3d 782, 798 (9th Cir. 2005) ("NWF v. NMFS I"). A court "may not substitute its judgment for that of the agency concerning the wisdom or prudence of the agency's action." River Runners for Wilderness v. Martin, 593 F.3d 1064, 1070 (9th Cir. 2009):
In conducting an APA review, the court must determine whether the agency's decision is "founded on a rational connection between the facts found and the choices made ... and whether [the agency] has committed a clear error of judgment." Ariz. Cattle Growers' Ass'n v. U.S. Fish & Wildlife, 273 F.3d 1229, 1243 (9th Cir. 2001). "The [agency's] action ... need be only a Nat'l Wildlife Fed. v. Burford reasonable, not the best or most reasonable, decision." , 871 F.2d 849, 855 (9th Cir. 1989).
Although deferential, judicial review under the APA is designed to "ensure that the agency considered all of the relevant factors and that its decision contained no clear error of judgment." Arizona v. Thomas, 824 F.2d 745, 748 (9th Cir. 1987) (internal citations omitted). "The deference accorded an agency's scientific or technical expertise is not unlimited." Brower v. Evans, 257 F.3d 1058, 1067 (9th Cir. 2001) (internal citations omitted).
[An agency's decision is] arbitrary and capricious if [it] has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.
Motor Vehicle Mfrs. Ass'n of U.S. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983); see also Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 416 (1971) (reviewing court may overturn an agency's action as arbitrary and capricious 3 if the agency failed to consider relevant factors, failed to base its decision on those factors, and/or made a "clear error of judgment"), overruled on other grounds by Califano v. Sanders, 430 U.S. 99, 105 (1977)).
More generally, "[u]nder the APA 'the agency must examine the relevant data and articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made.'" Humane Soc. of U.S. v. Locke, --- F.3d ---, 2010 WL 4723195, *5 (9th Cir. 2010) (quoting Motor Vehicle Mfrs. Ass'n, 463 U.S. at 43). "The reviewing court should not attempt itself to make up for an agency's deficiencies: We may not supply a reasoned basis for the agency's action that the agency itself has not given." Id.
(3)General Obligations Under the ESA.
ESA Section 7(a)(2) prohibits agency action that is "likely to jeopardize the continued existence" of any endangered or threatened species or "result in the destruction or adverse modification" of its critical habitat. 16 U.S.C. § 1536(a)(2).
To "jeopardize the continued existence of" means "to engage in an action that reasonably would be expected, directly or indirectly, to reduce appreciably the likelihood of both the survival and recovery of a listed species in the wild by reducing the reproduction, numbers, or distribution of that species." 50 C.F.R. § 402.02; see also NWF v. NMFS II, 524 F.3d 917 (rejecting 3 agency interpretation of 50 C.F.R. § 402.02 that in effect limited jeopardy analysis to survival and did not realistically evaluate recovery, thereby avoiding an interpretation that reads the provision "and recovery" entirely out of the text). An action is "jeopardizing" if it keeps recovery "far out of reach," even if the species is able to cling to survival. NWF v. NMFS II, 524 F.3d at 931. "[A]n agency may not take action that will tip a species from a state of precarious survival into a state of likely extinction. Likewise, even where baseline conditions already jeopardize a species, an agency may not take action that deepens the jeopardy by causing additional harm." Id. at 930.
To satisfy this obligation, the federal agency undertaking the action (the "action agency") must prepare a "biological assessment" that evaluates the action's potential impacts on species and species' habitat. 16 U.S.C. § 1536(c); 50 C.F.R. § 402.12(a). If the proposed action "is likely to adversely affect" a threatened or endangered species or adversely modify its designated critical habitat, the action agency must engage in "formal consultation" with FWS to obtain its biological opinion as to the impacts of the proposed action on the listed species. See 16 U.S.C. § 1536(a)(2), (b)(3); see also 50 C.F.R. § 402.14(a), (g). Once the consultation process has been completed, FWS must give the action agency a written biological opinion "setting forth [FWS's] opinion, and a summary of the 3 information on which the opinion is based, detailing how the agency action affects the species or its critical habitat." 16 U.S.C. § 1536(b)(3)(A); see also 50 C.F.R. § 402.14(h).
If FWS determines that jeopardy or destruction or adverse modification of critical habitat is likely, FWS "shall suggest those reasonable and prudent alternatives which [it] believes would not violate subsection (a)(2) of this section and can be taken by the Federal agency or applicant in implementing the agency action." 16 U.S.C. § 1536(b)(3)(A). "Following the issuance of a 'jeopardy' opinion, the agency must either terminate the action, implement the proposed alternative, or seek an exemption from the Cabinet-level Endangered Species Committee pursuant to 16 U.S.C. § 1536(e)." Nat'l Ass'n of Home Builders v. Defenders of Wildlife, 551 U.S. 644, 652 (2008).
(4)Best Available Science.
Under the ESA, an agency's actions must be based on "the best scientific and commercial data available." 16 U.S.C. § 1536(a)(2); 50 C.F.R. § 402.14(g)(8) ("In formulating its Biological Opinion, any reasonable and prudent alternatives, and any reasonable and prudent measures, the Service will use the best scientific and commercial data available...."). A failure by the agency to utilize the best available science is arbitrary and capricious. See Pac. Coast Fed'n of Fishermen's Assns. v. Gutierrez, 606 F. Supp. 2d 1122, 1144 (E.D. Cal. 2008).
"The obvious purpose of the [best available science requirement] is to ensure that the ESA not be implemented haphazardly, on the basis of speculation or surmise." Bennett v. Spear, 520 U.S. 154, 176 (1997).
While this no doubt serves to advance the ESA's overall goal of species preservation, we think it readily apparent that another objective [of the best available science requirement] (if not indeed the primary one) is to avoid needless economic dislocation produced by agency officials zealously but unintelligently pursuing their environmental objectives. That economic consequences are an explicit concern of the ESA is evidenced by § 1536(h), which provides exemption from § 1536(a)(2)'s no-jeopardy mandate where there are no reasonable and prudent alternatives to the agency action and the benefits of the agency action clearly outweigh the benefits of any alternatives. We believe the "best scientific and commercial data" provision is similarly intended, at least in part, to prevent uneconomic (because erroneous) jeopardy determinations.
A decision about jeopardy must be made based on the best science available at the time of the decision; the agency cannot wait for or promise future studies. See Ctr. for Biological Diversity v. Rumsfeld, 198 F. Supp. 2d 1139, 1156 (D. Ariz. 2002). The "best available science" mandate of the ESA sets a basic standard that "prohibits the [agency] from disregarding available scientific evidence that is in some way better than the evidence [it] relies on." Am. Wildlands v. Kempthorne, 530 F.3d 991, 998 (D.C. Cir. 2008) (citation omitted).
What constitutes the "best" available science implicates 3 core agency judgment and expertise to which Congress requires the courts to defer; a court should be especially wary of overturning such a determination on review. Baltimore Gas & Elec. Co. v. Natural Res. Defense Council, 462 U.S. 87, 103 (1983) (a court must be "at its most deferential" when an agency is "making predictions within its area of special expertise, at the frontiers of science"). As explained in the en banc decision in Lands Council, 537 F.3d at 993, courts may not "impose on the agency their own notion of which procedures are best or most likely to further some vague, undefined public good." In particular, an agency's "scientific methodology is owed substantial deference." Gifford Pinchot Task Force v. U.S. Fish & Wildlife Serv., 378 F.3d 1059, 1066 (9th Cir. 2004).
When specialists express conflicting views, an agency must have discretion to rely on the reasonable opinions of its own qualified experts even if, as an original matter, a court might find contrary views more persuasive." Lands Council, 537 F.3d at 1000 (quoting Marsh v. Oregon Natural Res. Council, 490 U.S. 360, 378 (1989)). Mere uncertainty, or the fact that evidence may be "weak," is not fatal to an agency decision. Greenpeace Action v. Franklin, 14 F.3d 1324, 1337 (9th Cir. 1992) (upholding biological opinion, despite uncertainty about the effectiveness of management measures, because decision was based on a reasonable evaluation of all available data); Nat'l Wildlife 3 Fed'n v. Babbitt, 128 F. Supp. 2d 1274, 1300 (E.D. Cal. 2000)
(holding that the "most reasonable" reading of the best scientific data available standard is that it "permits the [FWS] to take action based on imperfect data, so long as the data is the best available"). FWS "must utilize the 'best scientific ... data available,' not the best scientific data possible." Building Indus. Ass'n v. Norton, 247 F.3d 1241, 1246 (D.C. Cir. 2001), cited with approval in Kern County Farm Bureau v. Allen, 450 F.3d 1072, 1080-81 (9th Cir. 2006) ("Absent superior data occasional imperfections do not violate" the ESA best available data standard); see also Defenders of Wildlife v. Babbitt, 958 F. Supp. 670, 680 (D.D.C. 1997) (best available science standard does not require "conclusive evidence," only that agency use best science available and not ignore contrary evidence).
The deference afforded under the best available science standard is not unlimited. For example, Tucson Herpetological Society v. Salazar, 566 F.3d 870, 879 (9th Cir. 2009), held that an agency may not rely on "ambiguous studies as evidence" to support findings made under the ESA. Because the studies did not lead to the conclusion reached by FWS, the Ninth Circuit held that these studies provided inadequate support in the administrative record for the determination made by FWS. Id.; see also Rock Creek Alliance v. U.S. Fish & Wildlife Service, 390 F. Supp. 2d 993, 1008 (D. Mont. 2005) (rejecting FWS's reliance 3 on a disputed scientific report, which explicitly stated its analysis was not applicable to the small populations addressed in the challenged opinion). Alternatively, the presumption of agency expertise may be rebutted if the agency's decisions, although based on scientific expertise, are not reasoned, Greenpeace v. NMFS, 80 F. Supp. 2d 1137, 1147 (W.D. Wash. 2000), or if the agency disregards available scientific evidence better than the evidence on which it relies, Kern County Farm Bureau, 450 F.3d at 1080.
Courts routinely perform substantive reviews of record evidence to evaluate the agency's treatment of best available science. The judicial review process is not one of blind acceptance. See, e.g., Kern County, 450 F.3d at 1078-79 (thoroughly reviewing three post-comment studies and FWS's treatment of those studies to determine whether they "provide[d] the sole, essential support for" or "merely supplemented" the data used to support a listing decision); Home Builders Ass'n of N. Cal. v. U.S. Fish and Wildlife Serv., 529 F. Supp. 2d 1110, 1120 (N.D. Cal. 2007) (examining substance of challenge to FWS's determination that certain data should be disregarded); Trout Unlimited v. Lohn, 645 F. Supp. 2d 929 (D. Or. 2007) (finding best available science standard had been violated after thorough examination of rationale for NMFS's decision to withdraw its proposal to list Oregon Coast Coho salmon); Oceana, Inc. v. 3 Evans, 384 F. Supp. 2d 203, 217-18 (D.D.C. 2005) (carefully considering scientific underpinnings of challenge to FWS's use of a particular model, including post decision evidence presented by an expert to help the court understand the complex model, applying one of several record review exceptions articulated in Esch v. Yeutter, 876 F.2d 976, 991 (D.C. Cir. 1989), which are similar to those articulated by the Ninth Circuit).
Courts are not required to defer to an agency conclusion that runs counter to that of other agencies or individuals with specialized expertise in a particular technical area. See, e.g., Am. Turnboat Ass'n v. Baldrige, 738 F.2d 1013, 1016-17 (9th Cir. 1984) (NMFS's decision under the Marine Mammal Protection Act was not supported by substantial evidence because agency ignored data that was product of "many years' effort by trained research personnel"); Sierra Club v. U.S. Army Corps of Eng'rs, 701 F.2d 1011, 1030 (2d Cir. 1983) ("court may properly be skeptical as to whether an EIS's conclusions have a substantial basis in fact if the responsible agency has apparently ignored the conflicting views of other agencies having pertinent experience") (internal citations omitted). A court should "reject conclusory assertions of agency 'expertise' where the agency spurns unrebutted expert opinions without itself offering a credible alternative explanation." N. Spotted Owl v. Hodel, 716 F. Supp. 479, 483 (W.D. Wash. 1988) (citing Am. Turnboat Ass'n, 738 F.2d at 1016). 3 In Conner v. Burford, 848 F.2d 1441, 1453-54 (9th Cir. 1988), the agency attempted to defend its biological opinions by arguing that there was a lack of sufficient information to perform additional analysis. In rejecting this defense, the Ninth Circuit held that "incomplete information ... does not excuse the failure to comply with the statutory requirement of a comprehensive biological opinion using the best information available," and noted that FWS could have completed more analysis
Id. with the information that was available. at 1454.
In light of the ESA requirement that the agencies use the best scientific and commercial data available ... the FWS cannot ignore available biological info or fail to develop projections of ... activities which may indicate potential conflicts between development and the preservation of protected species. We hold that the FWS violated the ESA by failing to use the best information available to prepare comprehensive biological opinions.
(5)Best Available Science Standards and the Application of Analytical/Statistical Methodologies.
The above-described standards apply with equal force to the use and interpretation of statistical methodologies. As the D.C. Circuit in Appalachian Power Co. v. EPA, 135 F.3d 791 (D.C. Cir. 1998), explained in reviewing a challenge to a decision of the Environmental Protection Agency ("EPA") under the "arbitrary and capricious" standard of review:
Statistical analysis is perhaps the prime example of those areas of technical wilderness into which judicial expeditions are best limited to ascertaining the lay of 3 the land. Although computer models are "a useful and often essential tool for performing the Herculean labors Congress imposed on EPA in the Clean Air Act," [citation] their scientific nature does not easily lend itself to judicial review. Our consideration of EPA's use of a regression analysis in this case must therefore comport with the deference traditionally given to an agency when reviewing a scientific analysis within its area of expertise without abdicating our duty to ensure that the application of this model was not arbitrary.
The model must fit the available data. See Nat'l Wildlife Fed'n v. EPA, 286 F.3d 554, 565 (D.C. Cir. 2002) ("NWF v. EPA") (a court will only reject the choice of a model "when the model bears no rational relationship to the characteristics of the data to which it was applied"). For example, Oceana, 384 F. Supp. at 220, rejected a challenge to NMFS's use of a particular analytical model that used data drawn from existing literature, even though experts "suggested that reliable take limits cannot be established without quantitative data gathered from 'in-water' surveys." Although NMFS conceded "a thorough quantitative analysis based on empirical estimates of population size would be a superior way to analyze the impact  on [the species]," it was undisputed that "given the paucity of information on sea turtles and the difficulties of using the data that does exist, '[a] different or more complex model [than that used by NMFS] was not available and could not even be constructed.'" Id. Likewise, "the fact that a given model has some imperfections does not prevent it from constituting the 'best scientific information available.'" Oceana v. Evans, 2005 WL 555416, *16-*17 (D.D.C. 3 Mar. 9, 2005)(citing 16 U.S.C. § 1851(a)(2))(approving NMFS's use of a model despite known limitations, where it was the only model available and the agency supplemented its analysis with other sources to address areas where the model was unable to make accurate predictions).
A.Challenges to the Effects Analysis & Related Challenges to the RPA Actions.
(1)Legal Requirements for a Project Effects Analysis. Under section 7(a)(2) of the ESA and the Joint Consultation Regulations, FWS must "[e]valuate the effects of the action and cumulative effects on the listed species or critical habitat."
50 C.F.R. § 402.14(g)(3). FWS must then "[f]ormulate its biological opinion as to whether the action, taken together with cumulative effects,*fn7 is likely to jeopardize the continued existence of listed species or result in the destruction or adverse modification of critical habitat." § 402.14(g)(4). The effects of the action are defined as: the direct and indirect effects of an action on the species or critical habitat, together with the effects of other activities that are interrelated or interdependent with that action, that will be added to the environmental baseline. § 402.02.
The environmental baseline includes: the past and present impacts of all Federal, State, or private actions and other human activities in the action area, the anticipated impacts of all proposed
3 Federal projects in the action area that have already undergone formal or early section 7 consultation, and the impact of State or private actions which are contemporaneous with the consultation in process.
Id. The baseline is described in FWS and NMFS's Joint Consultation Handbook*fn8 as: an analysis of the effects of past and ongoing human and natural factors leading to the current status of the species, its habitat (including designated critical habitat), and ecosystem, within the action area. The environmental baseline is a "snapshot" of a species' health at a specified point in time. It does not include the effects of the action under review in the consultation.
Consultation Handbook 4-22.
Once the baseline, the "direct and indirect effects" of the action, and the "effects of other activities that are interrelated or interdependent with that action" are determined, 50 C.F.R. § 402.02, FWS then is required to consider whether, in light of the environmental baseline, the effects of the action, taken together with cumulative effects, are likely to jeopardize the continued existence of the listed species, 50 C.F.R. § 402.14(g).
[An] agency may not take action that will tip a species from a state of precarious survival into a state of likely extinction. Likewise, even where baseline conditions already jeopardize a species, an agency may 3 not take action that deepens the jeopardy by causing additional harm. ....[The agency must] appropriately consider the effects of its actions "within the context of other existing human activities that impact the listed species." ALCOA [v. Administrator, Bonneville Power Admin], 175 F.3d [1156,] 1162 n. 6 [(9th Cir. 1999)](citing 50 C.F.R. § 402.02's definition of the environmental baseline). This approach is consistent with our instruction ... that "[t]he proper baseline analysis is not the proportional share of responsibility the federal agency bears for the decline in the species, but what jeopardy might result from the agency's proposed actions in the present and future human and natural contexts." [PCFFA v. U.S. Bureau of Reclamation], 426 F.3d [1082,] 1093 [(9th Cir. 2005)](emphasis added).
NWF v. NMFS II, 524 F.3d at 930 (emphasis in original).
To jeopardize means "to engage in an action that reasonably would be expected, directly or indirectly, to reduce appreciably the likelihood of both the survival and recovery of a listed species." 50 C.F.R. § 402.02. The Consultation Handbook further provides that to "appreciably diminish the value: [means] to considerably reduce the capability of designated [critical habitat]." Consultation Handbook at 4-36. A related case found: interpretation of "appreciably" to mean any "perceptible" effect would lead to irrational results, making any agency action that had any effects on a listed species a "jeopardizing" action. This is not the law, as such an interpretation conflicts with other provisions of the ESA that permit incidental take of listed species.
PCFFA v. Gutierrez, 1:06-cv-00245 OWW GSA, Doc. 367 at 23-24 (citing 16 U.S.C. 1536(b)(4), 1539(1)(B)). (2)Best Available Science Challenges to the Effects
Analysis and Related Challenges to the Justification Provided for the RPA Actions.
Plaintiffs argue that the project effects analysis is predicated upon scientific errors that render the BiOp and its conclusion that project operations jeopardize the delta smelt arbitrary, capricious and an abuse of discretion:
The Project Effects Analysis is the heart of the section 7 consultation process, providing the basis for FWS' jeopardy and adverse modification determinations and for formulating the RPA. In this case, FWS began the Project Effects Analysis of the 2008 Smelt BiOp with a remarkable assumption: "The following analysis assumes that the proposed CVP/SWP operations affect delta smelt throughout the year either directly through entrainment or indirectly through influences on its food supply and habitat suitability." BiOp at 203 (AR 000218.) This assumption plainly violates the "best available science" required by the ESA. The science, including the reports that FWS purports to rely on, shows that OMR flows and entrainment do not have any statistically significant effect on the delta smelt's population growth rate. Restricting flows has no effect on the delta smelt population's survival-such restrictions are a costly, but meaningless gesture. The same is true for [restrictions designed to control the position of] X2 [in the Fall].
Plaintiffs maintain that the best available science does not support FWS' "assumption" that "CVP/SWP operations affect delta smelt throughout the year either directly through entrainment or indirectly through influences on its food supply and habitat suitability." BiOp at 203. Plaintiffs maintain that the science demonstrates:
(a) OMR flows have no statistically significant effect 3 on the delta smelt population growth rate;
(b) With respect to the adult population, only OMR flows more negative than -6,100 cfs will correlate to an increase in entrainment;*fn9
(c) The location of Fall X2 does not determine the extent and quality of suitable smelt habitat -- as with OMR flows, Fall X2 has no statistically significant effect on the population growth rate; and,
(d) The CVP/SWP projects do not indirectly govern abiotic and biotic factors in the Delta that affect delta smelt abundance.
Doc. 551 at 11. Plaintiffs also maintain that there is no scientific support for the BiOp's assumption that the Projects control hydrodynamic conditions in the Delta, or for the BiOp's classification of non-Project causes of harm as "indirect effects" of Project Operations. Id.
a.The BiOp's General Conclusion that Entrainment by Project Operations Adversely Affects Smelt Survival & Recovery is Supported by the Record.
The magnitude of diversions at the CVP and SWP pumping facilities influences flows throughout the Delta, including in the Old and Middle Rivers ("OMR"). BiOp at 160. When the level of diversion at the pumps is high, Old and Middle Rivers may flow backwards (in the opposite direction than they would under natural hydrological conditions) and toward the CVP and SWP natural conditions (called "negative" flows). Id. Negative OMR flows draw delta smelt present in the central and south Delta toward the pumps, and high negative flows increase the risk that they will be entrained at the pumps. Id. at 163, 253 (Figure E-7).
Unlike larger fish species, entrainment is lethal for weak- swimming delta smelt. Id. at 145. Relying on estimates of proportional entrainment presented by Dr. Wim Kimmerer in a 2008 paper entitled "Losses of Sacramento River Chinook Salmon and Delta Smelt to Entrainment in Water Diversions in the Sacramento-San Joaquin Delta," published in the journal, San Francisco Estuary & Watershed Science ("Kimmerer (2008)"), the BiOp concludes that "[t]otal annual entrainment of the delta smelt population (adults and their progeny combined) ranged from approximately 10 percent to 60 percent per year from 2002-2006."
Id. at 210. In years when low flows and high exports coincide with a spawning distribution of the delta smelt that includes the San Joaquin River, the loss of larval delta smelt due to entrainment can exceed 50% of the population. Id. at 164-65. Such losses do not occur every year, but FWS concluded the effect of these large larval loss events is "substantial when it does," particularly in light of the fact that the delta smelt is an annual fish. Id. at 165. Even one year where its spawning occurs "within the footprint of entrainment by the pumps" can lead to "a [severe] reduction in that year's production." Id.
The BiOp's Effects Analysis concludes that Project pumping 3 operations have a "sporadically significant" adverse effect on smelt abundance:
The population-level effects of delta smelt entrainment vary; delta smelt entrainment can best be characterized as a sporadically significant influence on population dynamics. Kimmerer (2008) estimated that annual entrainment of the delta smelt population (adults and their progeny combined) ranged from approximately 10 percent to 60 percent per year from 2002-2006. Major population declines during the early 1980s (Moyle et al. 1992) and during the recent POD years (Sommer et al. 2007) were both associated with hydrodynamic conditions that greatly increased delta smelt entrainment losses as indexed by numbers of fish salvaged. However, currently published analyses of long-term associations between delta smelt salvage and subsequent abundance do not support the hypothesis that entrainment is driving population dynamics year in and year out (Bennett 2005; Manly and Chotkowski 2006; Kimmerer 2008).
BiOp at 210 (emphasis added). This passage was based in large part on Kimmerer (2008), which states:
Delta smelt may suffer substantial losses to export pumping both as pre-spawning adults and as larvae and early juveniles. In contrast to the situation for salmon, pre-salvage mortality has been constrained in the calculations for adult Delta smelt, and its effects eliminated from the calculations for larval/juvenile Delta smelt. Combining the results for both life stages, losses may be on the order of zero to 40 percent of the population throughout winter and spring. The estimates have large confidence limits, which could be reduced by additional sampling, particularly to estimate θ in Equation 18. If there is interest in improving these estimates further, some attempts should be made to examine the assumptions not fully tested above, particularly those used in extrapolating larval abundance to hatch dates.
Plaintiffs argue that the BiOp misinterprets and misapplies Kimmerer's work. Dr. Bryan Manly, Plaintiffs' expert in the fields of biostatistics and population survey design, addressed the BiOp's statement that "delta smelt entrainment can best be characterized as a sporadically significant influence on population dynamics." Manly Decl., Doc. 397, at ¶ 7. Manly opines that "[t]his statement is unclear and confusing," and explains:
If the Service meant only that abundance at a point in time during a single year may vary depending upon entrainment, then Kimmerer's estimates support that statement. But if, as appears more likely, the Service was relying upon Kimmerer's estimates to support a conclusion that entrainment sometimes causes abundance to vary significantly later in the same year or in following years, then the statement in the BiOp has no scientific basis.
Id. Kimmerer (2008) only estimated percentage losses of delta smelt within single year classes, and did not conclude that such losses reduce population abundance from one year to the next. Id. at ¶ 8. In fact, Kimmerer (2008) contains a number of disclaimers, including the caveat that "export effects" on smelt are small relative to other factors affecting survival:
Although the upper bound of [the 0-40% loss] range represents a substantial loss, the effect of this loss is complicated by subsequent variability in survival (Figure 17). If this variability is uncorrelated with entrainment losses, then these losses will contribute little to the variability in fall abundance index. The simplest way to evaluate this is by regression of fall midwater trawl index on winter--spring export flow, but this relationship is contaminated by the downward step change in abundance in approximately 1981--1982, together with the long-term upward trend in export flow (mainly up to the mid-1970s, see Kimmerer 2004). Including this step in a regression model eliminates 3 the effect of export flow on the fall midwater trawl index (coefficient = -1.5 - 2.4, 95% CL, 36 df). It seems unlikely that the downward step change was due to the earlier increase in export flow; furthermore, despite substantial variability in export flow in years since 1982, no effect of export flow on subsequent midwater trawl abundance is evident.
This is not to dismiss the rather large proportional losses of delta smelt that occur in some years; rather, it suggests that these losses have effects that are episodic and that therefore their effects should be calculated rather than inferred from correlative analyses. In the absence of density dependence, using means in Figure 15 with natural mortality, fall abundance should have been reduced by ~ 10% during 1995--2005. This would have an equivalent effect of reducing the summer--fall survival index by 10%. This would have made little difference to fall abundance in the context of the approximately 50-fold variation in summer--fall survival (Figure 17), and would be difficult to detect through correlation.
Although summer--fall survival appears to dominate variability in abundance of delta smelt in fall (Figure 17), this does not imply that control of export effects would be fruitless, as these effects can be considerable during dry years. Management of delta smelt should incorporate any opportunities that arise to improve habitat or food supply and to reduce any negative impacts of predation or toxic contamination. However, current evidence does not provide a clear path toward improving the status of delta smelt using these factors. Manipulating export flow (and, to some extent, inflow) is the only means to influence the abundance of delta smelt that is both feasible and supported by the current body of evidence, even though export effects are relatively small. The results presented here can be used to suggest when, and under what conditions, control of export effects would be most helpful.
AR 018878. Kimmerer (2008) concludes that even though correlative analysis revealed "no effect of export flow on subsequent midwater trawl abundance," there is reason to be concerned about episodic effects caused by "large proportional 3 losses of delta smelt that occur in some years." Id. As a result, according to Kimmerer (2008), population level effects should be calculated, rather than inferred from correlative analysis. Id. After performing such a calculation, Kimmerer (2008) concluded that entrainment reduced "the summer-fall survival index by ~10%" during 1995-2005. Id. Although this 10% figure was small in the context of the 50-fold variation in summer-fall survival, Kimmerer (2008) nonetheless recommended controlling export effects on smelt because "[m]anipulating export flow (and to some extent, inflow) is the only means to influence the abundance of delta smelt that is both feasible and supported by the current body of evidence, even though export effects are relatively small." Id. (emphasis added).
Dr. Manly is correct that Kimmerer (2008) does not support the position that entrainment has a "sporadically significant" effect on delta smelt abundance from one year to the next. However, contrary to Dr. Manly's suggestion, the BiOp does not rely on Kimmerer (2008) for this premise. The BiOp qualifies its reliance on Kimmerer (2008), consistent with the narrow scope of Kimmerer's findings:
The population-level effects of delta smelt entrainment vary; delta smelt entrainment can best be characterized as a sporadically significant influence on population dynamics. Kimmerer (2008) estimated that annual entrainment of the delta smelt population (adults and their progeny combined) ranged from approximately 10 percent to 60 percent per year from 2002-2006. Major 3 population declines during the early 1980s (Moyle et al. 1992) and during the recent POD years (Sommer et al. 2007) were both associated with hydrodynamic conditions that greatly increased delta smelt entrainment losses as indexed by numbers of fish salvaged. However, currently published analyses of long-term associations between delta smelt salvage and subsequent abundance do not support the hypothesis that entrainment is driving population dynamics year in and year out (Bennett 2005; Manly and Chotkowski 2006; Kimmerer 2008).
BiOp at 210 (emphasis added). It was not unreasonable for FWS to rely on Kimmerer (2008) to conclude that salvage events may be "sporadically significant." Plaintiffs' argument that FWS misinterpreted Kimmerer (2008) is unfounded. Kimmerer (2008) explains why, despite the absence of a statistically significant correlation between export pumping and the subsequent year's smelt population (i.e., between export pumping and the population growth rate), the demonstrated "sporadically significant" loss of smelt within year classes could significantly contribute to the species' jeopardy. FWS reasonably relied on Kimmerer (2008) for this finding.
Applying Kimmerer's estimates of entrainment and other data, the BiOp analyzed the effect Project operations have on the frequency of relatively large loss events. For larval and juvenile delta smelt:
Kimmerer (2008) proposed a method for estimating the percentage of the larval-juvenile delta smelt population entrained at Banks and Jones each year. These estimates were based on a combination of larval distribution data from the 20-mm survey, estimates of net efficiency in this survey, estimates of larval mortality rates, estimates of spawn timing, particle 3 tracking simulations from DWR's DSM-2 particle tracking model, and estimates of Banks and Jones salvage efficiency for larvae of various sizes. Kimmerer estimated larval-juvenile entrainment for 1995-2005. We used Kimmerer's entrainment estimates to develop multiple regression models to predict the proportion of the larval-juvenile delta smelt population entrained based on a combination of X2 and OMR....
BiOp at 220. The BiOp predicts that "the proposed action will decrease the frequency of years in which estimated entrainment is [less than or equal to] 15 percent. Thus, over a given span of years, the project as proposed will increase larval-juvenile entrainment relative to 1995-2005 levels. This will have an adverse effect on delta smelt based on their current low population levels." BiOp at 222.
The median OMR flows from the CALSIM II modeled scenarios were more negative than historic OMR flow for all WY types except critically dry years (Figure E-3; see Table E-5b for all differences). Overall, proposed OMR flows are likely to generate increases in population losses compared to historic years (Figure E-5 and Figure E-6). For example, the frequency of years when population losses are less than 10 percent from most modeled studies (except studies 7.0 and 8.0) is less than 24 percent compared to historic estimates that only exceed 10 percent in approximately half of the years.
The most pronounced differences occur during wet years, where median OMR flows are projected to be approximately 400 to 600 percent (-7100 to -3678 cfs) higher than historical wet years (-1032 cfs). Generally, wet years are marked by low salvage and population losses. However, the proposed operations during wet year are predicted to cause up to a 65 percent increase in smelt salvage and lower probability that population losses will be below 10 percent.
The proposed operation conditions likely to have the greatest impact on delta smelt are those modeled during above normal WYs. The modeled OMR flows for the above normal WYs ranged between -8155 and -6242 cfs, a 33 to 57 percent decrease from the historic median of -5178 cfs. Though the predicted salvage would only be about 15-20 percent higher than historic salvage during these years (Table E-5c), the modeled OMR flows in these years would increase population losses compared to historic years.
In below normal and dry WYs, proposed OMR flows are also modeled to decrease from historic medians. Predicted salvage levels are likely to increase between 2 and 44 percent. More importantly, the modeled median flows from all studies in these WY types range between -5747 and -7438 cfs. Modeled OMR flows at these levels are predicted to increase salvage and increase the population losses from historic levels as well.
During critically dry years, the median OMR flows for studies 7.0, 7.1, 8.0, 9.1, 9.4, and 9.5 are less than -5,000 cfs. These studies have predicted salvage lower than historic salvage and are not likely to generate larger population losses compared to historic years. The models might overestimate salvage during critical dry years when smelt are unlikely to migrate towards the Central Delta due to lack of turbidity or first flush. Thus, the effects of critical dry operations on delta smelt take are probably small and lower than estimated.
In summary, adult entrainment is likely to be higher than it has been in the past under most operating scenarios, resulting in lower potential production of early life history stages in the spring in some years. While the largest predicted effects occur in Wet and Above Normal WYs, there are also likely adverse effects in Below Normal and Dry WYs. Only Critically Dry WYs are generally predicted to have lower entrainment than what has occurred in the recent past.
This approach is consistent with Kimmerer (2008). The BiOp does not focus on whether there is a statistically significant 3 correlation between OMR flows and the population growth rate.*fn10
Rather, following Kimmerer (2008), the BiOp focuses on predicting the frequency of large salvage events and concluded that Project operations increase their frequency. It was not arbitrary, capricious, or clear error for FWS to base its jeopardy conclusion in part on these predictions of relative increases in entrainment. See BiOp at 276.
b.Population Level Analysis/Life-Cycle Modeling. Plaintiffs maintain
the BiOp's failure to employ a life-cycle model ignored the best available science. Doc. 551 at 21-22.
Using a quantitative*fn11 life-cycle model*fn12
is a recognized (the best) method to evaluate the effects of
an action upon a fish population's growth rate. Dr. Richard B.
Deriso*fn13 opined that a population growth rate
analysis is the generally accepted method utilized by fisheries
biologists to evaluate the impact of a stressor on a fish species'
population. Declaration of Dr. Richard B. Deriso, Doc 401, at ¶ 36;
see also Declaration of Dr. Ray Hilborn*fn14 , Doc.
393, at ¶¶ 7-16 (agreeing that life-cycle models are the accepted
method in population dynamics to evaluate anthropogenic effects on the
probability of growth or decline of 3 a species); Declaration of Ken
B. Newman*fn15 , Doc. 484, at ¶ 8
(agreeing with "utility of life history models for assessing
population level effects of SWP/CVP operations."). Dr. Hilborn
explained that a quantitative population dynamics/life cycle model can
help distinguish human actions that have a significant impact on
population size from those that have little impact on population size,
because competition for a resource that is independent of the human
activity may cause significant mortality
at one stage in the species' life cycle, meaning that human actions
that kill fish at that life stage may have little impact on the
population level later in the life history. Hilborn Decl., Doc. 393 at
Federal Defendants knew of the value of life-cycle modeling. At a March 8, 2007 meeting on the OCAP ESA Re-consultation, attended by FWS employees, the importance of using a life cycle model was emphasized and inquiry made about the progress to date. AR 016016 - 016017. During the Delta Smelt Action Evaluation Team meeting on August 8, 2008, that Team recognized that population models for delta smelt already had been developed, and 3 that those models were a starting point for quantitative analyses when combined with appropriate assumptions. AR 011381-011382; see also AR 010023, 010027-010029.
There is considerable dispute over whether an appropriate life-cycle model (i.e., one sufficient to perform the types of analyses that would be helpful in the BiOp) existed at the time the BiOp issued. Dr. Newman declares:
Despite the utility of life history models and despite the information that the various surveys provide about different life history stages, an adequately realistic quantitative delta smelt life history model that has been fit using fish survey data does not exist. The BiOp did in many places (e.g., pp 146, 184, 203) consider the full life history of delta smelt but considerations were via conceptual models in contrast to quantitative models with parameters estimated from data. Part of the difficulty is that there are currently no off-the-shelf computational programs for fitting such a model to data and one must develop customized, computer intensive software. The need to model the spatial and temporal changes in population abundances and to account for the different sources of uncertainty makes model formulation and fitting complex. In particular, uncertainty in survey data, due to random sampling error and bias, complicates model fitting. Capture probabilities differ between surveys, the probabilities are largely unknown (despite efforts made to estimate them, for example, for FMWT data, see Newman 2008 (Administrative Record "AR" at 19782- 19799)), and capture and fish presence probabilities are thus confounded. Furthermore, given the patchiness and heterogeneity of the spatial and temporal distribution of delta smelt and the relatively low capture probabilities (whatever they might be), the sampling errors associated with survey data can be quite large (Newman 2008 (AR at 19782-19799)). Failure to account for sampling errors may result in biased parameter estimates (including wrongly concluding density dependence; Shenk et al. 1998). The difficulties are not insurmountable, but concentrated 3 research efforts are required. I know of three such efforts currently underway and at varying stages of development: (1) an individual-based model with a spatial component by Drs. Wim Kimmerer, San Francisco State University, William Bennett, University of California at Davis, Stephen Monismith, Stanford University, and Kenneth Rose, Louisiana State University; (2) a population-level life history model using information from multiple surveys by Dr. Mark Maunder, Inter American Tropical Tuna Commission; (3) similar to Maunder, a life history model with a spatial component based on multiple surveys' data has been conceptually sketched by me and others in the NCEAS POD working group. Given sufficient time and appropriate technical resources, including personnel, to focus on model formulation and fitting, these models might be available within a year.
Newman Decl., Doc. 484 at ¶ 5.
All of the experts agreed with Dr. Newman that, at the time the BiOp was issued, there was no "off-the-shelf" life-cycle model to apply to delta smelt. Considerable dispute exists over how long it should have taken FWS to develop a competent model.
It is undisputed that basic life-cycle models such as the Ricker model can be applied to fisheries data sets in relatively short order. Deriso Decl., Doc. 605, at ¶ 52. Dr. Deriso opined that FWS had all the data necessary to perform a life-cycle analysis. Deriso Decl., Doc. 401, at ¶ 70. Dr. Hilborn stated that a relatively complex life-cycle model that "follow[s] the size structure of delta smelt through their life history and fit this into the observed size structure" would "require no more than a few months time to construct, evaluate and use in a biological opinion." Hilborn Decl., Doc. 600 at ¶ 14. Dr. Punt, a 706 Expert with expertise in fish population dynamics and 3 biostatistics, see Doc. 394 at 2, stated "[i]t is surprising that a population dynamics model was not developed for delta smelt for the BiOp.... The model developed by Bennett could have been extended to more fully account for the biology of delta smelt and fitted to data to assess the population-level effects of impact of the project." Doc. 633-1 at 3.
Federal Defendants' expert, Mr. Feyer disagrees:
Developing a quantitative population model is a challenging and complex exercise that could not have been completed by USFWS within the timeframe required to issue the 2008 BiOp. The work requires a substantial investment of resources and individuals with very specialized skills. The process to develop, test, peer-review, and apply such models often takes years. For instance ... the development of models for Columbia River salmon ... took no less than three years to complete.
Because of the recognized urgent need for such tools, there are on-going efforts to develop quantitative population models for delta smelt. For instance, Bennett (2005) presented preliminary results from a stage-structured model he is developing to examine tradeoffs among sources of mortality acting on different cohorts and life stages. See AR at 17004-74. The development of this model is part of a broader comprehensive effort by a team of researchers including Dr. Kenneth Rose of Louisiana State University, Dr. Wim Kimmerer of San Francisco State University, Dr. William Bennett of the University of California at Davis, and Dr. Stephen Monismith of Stanford University, who are in the early stages of developing, testing, and applying particle-tracking models, an individual-based model, and a matrix projection model. The development of these particular models is very promising but has also been faced with many challenges. Perhaps the most critical challenge has been a freeze on project funding by the State of California; it is uncertain if the funding will be reinstated. Another example is the work I have been personally involved with at NCEAS. The 3 NCEAS team has used Bayesian changepoint techniques and multivariate autoregressive modeling to identify factors contributing to the decline of delta smelt and other species. The results of this work will be published in two papers in an upcoming issue of the journal Ecological Applications. I am aware of at least two other independent efforts of modeling the effects of various stressors on delta smelt that are also under development. Unfortunately, none of the work I mention above was available when the 2008 BiOp was being prepared. To my knowledge, no comprehensive quantitative population dynamics model for the delta smelt has been developed, subjected to peer-review, and published. ...[Q]uantitative population models are grounded in what is known about the biology of a species, and processes that may plausibly affect its abundance.... Although there is a substantial amount of data available on delta smelt, a key problem is that much of the sample data has increasingly contained zero values. These zeros are a reflection of declining population abundance. Such low numbers make it more difficult to acquire more recent information about the factors that drive delta smelt population dynamics, such as survival probabilities by life history stage, movement patterns and spatial distribution, and fecundity or reproductive success. It is thus becoming increasingly difficult to not only simply estimate such factors, but also increasingly difficult to model how these factors are affected by environmental and anthropogenic processes such as those considered in the 2008 BiOp. The estimation of delta smelt population size exemplifies this problem. Newman (2008), see AR at 19782-99, recently published a sample design-based procedure for estimating the population abundance of pre-adult and adult delta smelt. However, the resulting estimates of population size were quite imprecise. This was caused, in part, by limitations of the available data to estimating capture probabilities and gear efficiency. ... I agree ... that population dynamics models have been used to evaluate consequences of various stressors on a wide range of species and human impacts. I also agree that there is sufficient data to develop such a model for delta smelt, as demonstrated by the examples I provided above. However, although some are in development, the fact remains that no such model has 3 been fully developed, peer-reviewed and made available for application. Thus, in the absence of such models, I disagree that that the techniques used by USFWS were inconsistent with generally-accepted scientific standards and practices. To the contrary, in the absence of such a model, and because one could not be developed during the time allowed for this consultation, the techniques used by USFWS do reflect generally-accepted scientific standards and practices. Decl. of Frederick V. Feyrer*fn16 , Doc. 541, at ¶¶ 30-33. Plaintiffs do not suggest any party that participated in the preparation of the OCAP Biological Assessment ("OCAP BA" or "BA") or commented on the public review drafts of the BiOp during the consultation submitted to FWS a quantitative life cycle model or the results of such an analysis using a life cycle model for delta smelt.
The ESA does not require FWS's to generate new studies. In Southwest Center for Biological Diversity v. Babbitt, 215 F.3d 58 (D.C. Cir. 2000), the district court found "inconclusive" the available evidence regarding FWS's decision not to list the Queen Charlotte goshawk, and held that the agency was obligated to find better data on the species' abundance. The D.C. Circuit reversed, emphasizing that, although "the district court's view has a superficial appeal ... this superficial appeal cannot circumvent the statute's clear wording: The secretary must make his decision as to whether to list a species as threatened or endangered 'solely on the basis of the best scientific and commercial data available to him....' 16 U.S.C. § 1533(b)(1)(A)." 3 Id. at 61 (emphasis added); see also American Wildlands v.
Kempthorne, 530 F.3d 991, 998 (D.C. Cir. 2008) (the "best available data" standard "requires not only that the data be attainable, but that researchers in fact have conducted the tests").
Plaintiffs advocate a narrow reading of both Southwest Center and American Wildlands, arguing these cases only mean that the agency is not required to gather new data in the field regarding a species if such information is not already available. Doc. 697 at 22. Plaintiffs object that "[n]either of these cases supports Defendants' position that FWS could disregard the smelt abundance data that were already in its possession and fail to undertake the necessary statistical analyses to satisfy its statutory mandate to determine 'whether the action ... is likely to jeopardize the continued existence of the species.' 50 C.F.R. § 402.14(g)(4)." Id.
Plaintiffs cite no authority suggesting that the nonexistence of an analytical model should be treated any differently from the non-existence of raw field data. FWS did not have an off-the-shelf form of "statistical analysis" it could apply to determine the effects of Project Operations on the delta smelt population. Although life-cycle modeling is standard practice in the field of fisheries biology, and a life-cycle model is being (and should have been) developed for delta smelt, 3 it is undisputed that an appropriate life cycle model had not been developed at the time the BiOp issued. FWS must apply the best "available" science; not the best science possible. FWS's failure to apply a life cycle model did not per se violate the ESA or the APA.
It is undisputed that application of a quantitative life cycle model is the preferred scientific methodology. Based on the preponderating expert testimony, FWS had the time and ability to prepare the necessary life-cycle model. FWS made a conscious choice not to use expertise available within the agency to develop one. A court lacks authority to require completion of a life-cycle model. In light of uncontradicted expert testimony that life-cycle modeling is necessary and feasible, FWS's failure to do so is inexplicable.
c.FWS' Use of Raw Salvage Numbers.
Plaintiffs argue that FWS's use of raw salvage numbers in its quantitative justification for the flow prescriptions in Actions 1 and 2 constitutes a failure to apply the best available science. Action 1, designed to protect upmigrating delta smelt, is triggered during low and high entrainment risk periods based on physical and biological monitoring. Action 1 requires OMR flows to be no more negative than -2,000 cubic feet per second ("cfs") on a 14-day average and no more negative than -2,500 cfs for a 5-day running average. BiOp at 280-81, 329-30. Action 2, designed to protect adult delta smelt that have migrated upstream and are residing in the Delta prior to spawning, is triggered immediately after Action 1 ends or if recommended by the Smelt Working Group ("SWG"). Flows under Action 2 can be set within a range from -5,000 to -1,250 cfs, depending on a complex set of biological and environmental parameters. Id. at 281-82, 352-56.
The BiOp provides a quantitative justification for these specific flow prescriptions in Attachment B, entitled "Supplemental Information related to the Reasonable and Prudent Alternative." The following subsection entitled, "Justification for Flow Prescriptions in Action 1," is critical to the present challenge and is reproduced here in its entirety:
Justification for Flow Prescriptions in Action 1
Understanding the relationship between OMR flows and delta smelt salvage allows a determination of what flows will result in salvage. The OMR-Salvage analysis herein was initiated using the relationship between December to March OMR flow and salvage provided by P. Smith and provided as Figure B-13, below. Visual review of the relationship expressed in Figure B-13 indicates what appears to be a "break" in the dataset at approximately -5,000 OMR; however, the curvilinear fit to the data suggest that the break is not real and that the slope of the curve had already begun to increase by the time that OMR flows reached -5,000 cfs.
Further, a nonlinear regression was performed on the dataset, and the resulting pseudo-R2 value was 0.44-suggesting that although the curvilinear fit is a reasonable description of the data, other functional relationships also may be appropriate for describing the data. Fitting a different function to the data could also determine the location where salvage increased, i.e. identify the "break point" in the relationship between salvage and OMR flows. Consequently, an analysis was performed to determine if the apparent break at -5,000 cfs OMR was real. A piecewise polynomial regression, sometimes referred to as a multiphase model, was used to establish the change (break) point in the dataset.
A piecewise polynomial regression analysis with a linear-linear fit was performed using data from 1985 to 2006. The linear-linear fit was selected because it was the analysis that required the fewest parameters to be estimated relative to the amount of variation in the salvage data. Piecewise polynomial regressions were performed using Number Cruncher Statistical Systems ( Hintz, J., NCSS and PASS, Number Cruncher Statistical Systems, Kaysville UT).
The piecewise polynomial regression analysis resulted in a change point of -1162, i.e. at -1162 cfs OMR, the slope changed from 0 to positive (Figure B-14). These results indicate that there is a relatively constant amount of salvage at all flows more positive than -1162 cfs but that at flows more negative than -1162, salvage increases. The pseudo-R2 value was 0.42, a value similar to that obtained by P. Smith in the original analysis.
To verify that there was no natural break at any other point, the analysis was performed using a linear-linear-linear fit (fitting two change points). The linear-linear-linear fit resulted in two change points, -1,500 cfs OMR and -2,930 cfs OMR. The -1,500 cfs value is again the location in the dataset at which the slope changes from 0 to positive. The pseudo-R2 value is 0.42 indicating that this relationship is not a better description of the data. Because of the additional parameters estimated for the model, it was determined that the linear-linear-linear 3 fit was not the best function to fit the data, and it was rejected. No formal AIC analysis was performed because of the obvious outcome.
A major assumption of this analysis is that as the population of Delta smelt declined, the number of fish at risk of entrainment remained constant. If the number of fish in the vicinity of the pumps declined, fewer fish would be entrained and more negative OMR flows would result in lower salvage. This situation would result in an overestimate, i.e. the change point would be more positive. In fact, if the residuals are examined for the relationship in Figure B-13 above, the salvage for the POD years 2002, 2004, 2005, and 2006 are all below the line. 2003 is above the line although the line is not extended to the points at the top of the figure, and these data points occur when the curve becomes almost vertical. The negative residuals could be a result of a smaller population size available for entrainment and salvage. This could be verified by normalizing the salvage data by the estimated population size based on the FMWT data.
The original values of OMR and salvage could have been measured with error due to a number of causes, consequently the values used in the original piecewise polynomial analysis could be slightly different than the "true" values of salvage and OMR flow. Consequently, a second analysis was undertaken to examine the effect of adding stochastic variation to the OMR and salvage values in the piecewise polynomial regression analysis. The correlation between OMR and salvage in the original dataset was -0.61 indicating that the more negative the OMR, the greater the salvage. Consequently, it was necessary to maintain the original covariance structure of the data when adding the error terms and performing the regressions. The original covariance structure of the OMR--salvage data was maintained by adding a random error term to both parameters. The random error term was added to OMR and a correlated error term was added to salvage. The expected value of the correlated errors was -0.61.
The error terms were selected from a normal distribution with a mean of 1.0 and a standard deviation of 0.25 which provided reasonable variability in the original data. Operationally this process generated a normal distribution of OMR and salvage values in which the mean of the distributions were the original data points. Additional analyses were performed with standard deviations of 0.075, 0.025, and 0.125. Smaller standard deviations in the error term resulted in estimates of the change point nearer to the original estimate of -1,162 cfs. This is to be expected as the narrower the distribution of error terms, the more likely the randomly selected values would be close to the mean of the distribution. The process was repeated one hundred times, each time a new dataset was generated and a new piecewise polynomial regression was performed. The software package @Risk ( Palisade Decision Tools) was used to perform the Monte Carlo simulations. Latin hypercube sampling was used to insure that the distributions of OMR and salvage values were sampled from across their full distributions. The parameter of interest in the simulations was the change point, the value of the OMR flow at which the amount of salvage began to increase. Incorporating uncertainty into the analysis moved the change point to -1,800 cfs OMR, indicating that at flows above -1683, the baseline level of salvage occurred but with flows more negative than -1683, salvage increased.
BiOp 347-51 (emphasis added).
The analyses contained in Figures B-13 and B-14 serve, inter alia, as justification for Action 1: setting "break points" above and below which entrainment rates noticeably change. These break points are the foundation for the tiered flow restrictions in RPA Action 1. Cay Collette Goude*fn17 stated in her expert declaration that the analysis conducted by Dr. Michael Johnson, set forth in 3 Figure B-13, found inflection points where entrainment started to increase with more negative OMR flows, and that the inflection point "was -1,800 cfs OMR when uncertainty was factored into the analysis." Doc. 470, at ¶ 22. The BiOp does not explain in the "Justification for Flow Prescriptions in Action 1" or elsewhere how or why this -1,800 cfs figure relates to the -2,000 cfs upper limit imposed by Action 1.*fn18
Action 2 calls for flows to be set within a range from -5,000 to -1,250 cfs, depending on a complex set of biological and environmental parameters. BiOp at 281-82, 352-56. Although Appendix B describes and justifies Action 2 separately from Action 1, there is no independent section justifying the flow prescriptions imposed by Action 2. Instead, there is a subsection entitled "Justification for Guidelines in Setting Prescriptions of Action 2" which fixes biological and environmental parameters the SWG is to use in setting flows within the -5,000 cfs to -1,250 cfs range. See BiOp at 355. There is no independent quantitative or qualitative justification for ...