Full opinion text
FINDINGS OF FACT AND CONCLUSIONS OF LAW RE: PLAINTIFFS’ REQUEST FOR PRELIMINARY INJUNCTION (Docs. 161 & 230) OLIVER W. WANGER, District Judge. I. INTRODUCTION Plaintiffs San Luis & Delta Mendota Water Authority (the “Authority”) and Westlands Water District (“Westlands”) (collectively “San Luis Plaintiffs”) seek a Temporary Restraining Order (“TRO”) and a Preliminary Injunction (“PI”) against the implementation of Reasonable and Prudent Alternative (“RPA”) Action IV.2.1 set forth in the National Marine Fisheries Service’s (“NMFS”) June 4, 2009 Biological Opinion (“2009 Salmonid BiOp”), which addresses the impacts of the coordinated operations of the federal Central Valley Project (“CVP”) and State Water Project (“SWP”) on the Central Valley winter-run and spring-run Chinook salmon, Central Valley steelhead, Southern Distinct Population Segment of Green Sturgeon, and Southern Resident Killer Whales (“Listed Species”). Both motions were filed February 22, 2010. Docs. 230, 233. Plaintiffs State Water Contractors, Stockton East Water District, Oakdale Irrigation District, and South San Joaquin Irrigation District, and Plaintiff-Intervenor California Department of Water Resources (“DWR”) filed statements of non-opposition. Docs. 247, 248 & 251. Federal Defendants and DefendanUntervenors opposed. Docs. 273 & 274. Additionally, San Luis Plaintiffs seek a PI against the implementation of Action IV.2.3 in the 2009 Salmonid BiOp. Doc. 164 (filed Jan. 27, 2010). Plaintiffs Kern County Water Agency and Coalition for a Sustainable Delta joined. Doc. 181. DWR filed a partial joinder in and statement of non-opposition to the motion. Doc. 249. Federal Defendants and Defendant-Intervenors opposed. Docs. 273 & 274. The PI motions came on for evidentiary hearing and argument, in Courtroom 3 of the above-captioned Court from March 30 through April 2, 2010. The parties were represented by counsel, as noted on the record in open court. After consideration of the testimony of the witnesses, the exhibits received in evidence, the written briefs of the parties, oral arguments, and the parties’ proposed findings of fact and conclusions of law, Docs. 316 & 314, and disapprovals thereto, Docs. 320, 321 & 836, the following findings of fact and conclusions of law concerning the motion for interim relief/preliminary injunction are entered. To the extent any finding of fact may be interpreted as a conclusion of law or any conclusion of law may be interpreted as a finding of fact, it is so intended. II.BACKGROUND The 2009 Salmonid BiOp found that planned coordinated Project operations would jeopardize the continued existence of and/or adversely modify the critical habitat of several of the Listed Species. BiOp at 1-2. As required by law, NMFS proposed a Reasonable and Prudent Alternative (“RPA”) that imposes a number of operating restrictions and other measures on the Projects. The RPA included numerous elements for each of the various project divisions and associated stressors, which NMFS concluded “must be implemented in its entirety to avoid jeopardy and adverse modification.” Id. at ¶ 578 (emphasis added). The description of the RPA comprises approximately 90 pages of the 2009 Salmonid BiOp. See id. at 581-671. The RPA includes five principle components, with numerous sub-parts, but Plaintiffs currently seek to restrain only: • Action IV.2.1, which will limit pumping based on San Joaquin River inflow, measured at Vernalis, from April 1 through May 31; and • Action IV.2.3, which imposes restrictions on negative flows in Old and Middle Rivers (“OMR”) between January 1 and June 15, or until average daily water temps at Mossdale (a location on the San Joaquin River west of Manteca, California) are greater than 72°F, whichever is earlier. III.SUMMARY OF MOTION Plaintiffs seek preliminary injunctive relief against implementation of Actions IV.2.1 and IV.2.3 on the grounds that: 1) the district court already found that the United States Bureau of Reclamation (“Reclamation”) failed to comply with the National Environmental Policy Act (“NEPA”) in implementing the 2009 Salmonid BiOp; and 2) the 2009 Salmonid BiOp is arbitrary, capricious, and contrary to law because: a) NMFS allegedly conducted an effects analysis that improperly overstates impacts attributable to the coordinated operations of the CVP and SWP; b) NMFS failed to clearly define or consistently apply a relevant environmental baseline; c) NMFS failed to distinguish between discretionary and non-discretionary CVP and SWP activities, which overstated the effects of coordinated operations of the Projects; and d) RPA Actions IV.2.1 and IV.2.3 are arbitrary and capricious, because they are without factual or scientific justification and/or not supported by the best available science. Plaintiffs further claim that the implementation of Actions IV.2.1 and IV.2.3 will cause them continuing irreparable harm and that the public interest and balance of hardships favor injunctive relief. IV. STANDARD OF DECISION Injunctive relief, whether temporary or permanent, is an “extraordinary remedy, never awarded as of right.” Winter v. Natural Resources Defense Council,-U.S.-, 129 S.Ct. 365, 376, 172 L.Ed.2d 249 (2008); Weinberger v. Romero-Barcelo, 456 U.S. 305, 312, 102 S.Ct. 1798, 72 L.Ed.2d 91 (1982). Four factors must be established by a preponderance of the evidence to qualify for temporary injunctive relief: 1. Likelihood of success on the merits; 2. Likelihood the moving party will suffer irreparable harm absent injunctive relief; 3. The balance of equities tips in the moving parties’ favor; and 4. An injunction is in the public interest. Winter, 129 S.Ct. at 374; Am. Trucking Ass’n v. City of Los Angeles, 559 F.3d 1046, 1052 (9th Cir.2009). V. FINDINGS OF FACT A. The Agency Action. 1. The agency action is the coordinated operation of the CVP and SWP, pursuant to an Agreement for the Coordinated Operation of the two projects (“COA”). 2. According to the Rivers and Harbors Act of 1937, the dams and reservoirs of the CVP “shall be used, first, for river regulation, improvement of navigation and flood control; second, for irrigation and domestic uses; and, third, for power.” 50 Stat. 844, 850. 3. The CVP was reauthorized in 1992 through the Central Valley Improvement Act (“CVPIA”), which modified the 1937 Act and added mitigation, protection, and restoration of fish and wildlife as project purposes. Pub.L. 102-575 § 3402, 106 Stat. 4600, 4706 (1992). One of the stated purposes of the CVPIA is to address impacts of the CVP on fish and wildlife. 3406(a). The CVPIA made environmental protection and water deliveries co-purposes. 4. This case presents a critical conflict between these dual legislative purposes, providing water service for agricultural, domestic, and industrial use versus enhancing environmental protection for fish species whose habitat is maintained in rivers, estuaries, canals, and other waterways that comprise the Sacramento-San Joaquin Delta. 5. It is of manifest significance to the public interest that DWR, a co-operator and the State contractual partner of Reciamation, disagrees with at least some portions of the RPA and seeks limited injunctive relief against RPA Action IV.2.3. B. Facts Relevant to NEPA Claims. 6. It is undisputed that neither NMFS nor Reclamation engaged in any NEPA analysis in connection with preparation or implementation of the 2009 Salmonid BiOp. 7. It is undisputed that a March 17, granted San Luis Plaintiffs’ motion for summary judgment on their claim that Federal Defendants violated NEPA when they adopted and implemented the 2009 NMFS BiOp RPA without conducting the required NEPA analysis. Doc. 288. 8. NMFS asserts that it did consider a range of alternative RPA actions, including those proposed by Reclamation and DWR, and “carefully avoided prescribing measures that are not necessary to meet section 7 requirements.” BiOp at ¶ 578, 580 & 720 (NMFS endeavored “through the iterative consultation process to avoid developing RPA actions that would result in high water costs, while still providing for the survival and recovery of listed species.”). However, this process did not fully or sufficiently evaluate, explain or analyze the extent and gravity of the harms to humans and the environment visited upon Plaintiffs by Project water service reductions and pumping restrictions. 9. The 2009 Salmonid BiOp phases in some elements of the RPA over time, provides a health and safety exception to ensure a minimum level of water exports, uses monitoring programs and adaptive management to initiate RPA actions when species are present and protections are most needed, and includes specific scientific studies and engineering programs to refine RPA elements. Id. at ¶ 579-80, 719-23. In addition, the challenged RPA Actions were modified between the draft and final RPA to lessen water supply impacts, including shortening the duration of Action IV.2.1 from 90 to 60 days. Id. at 723; NMFS AR 104419. 10. A legally sufficient NEPA analysis should identify and analyze alternatives that minimize harm to humans and the human environment. Federal Defendants do not claim that they engaged in a systematic consideration of impacts to humans and the human environment and/or the alternatives that would have minimized harm to human and the human environment while still protecting the species. 11. Federal Defendants did not take the hard look required to achieve, to the maximum extent possible, the co-equal Reclamation Law objective of providing water service. C. Facts Relevant to ESA Challenges. (1) Current Status of the Species. a. Sacramento River Winter-Run Chinook Salmon. 12. Sacramento River winter-run Chinook salmon (Oncorhynchus tshawytscha) (“winter-run”) are listed as “endangered” under the ESA. 70 Fed. Reg. 37,160 (June 28, 2005). Winter-run critical habitat includes portions of the Sacramento River and other waters. Historical winter-run population estimates were as high as approximately 100,000 fish in the 1960s, but declined to under 200 fish in the 1990s. Gov’t Salmon Ex. 4 (Second Stuart Deck, Doc. 273-3), 45. In recent years, population surveys of winter-run estimated a high of 17,344 fish in 2006, followed by a decline in 2007 (2,542 fish) that persisted into 2008 (2,830 fish). Id. In 2009, there was a modest increase in adult escapement (4,658 fish). Id. Winter-run are “currently not viable.” BiOp at 88; see also 4/1/10 Tr. 175: 11-12. 13. Winter-run juvenile rearing and migration typically occurs between July and February in the upper Sacramento River, with juvenile emigration downstream through the Delta taking place between November through May or June. BiOp at 81, 94; Pac. Coast Fed’n of Fishermans’ Ass’ns. v. Gutierrez (“Gutierrez II”), 606 F.Supp.2d 1195, 1216-17 (E.D.Cal.2008); 4/1/10 Tr. 167:5-19; Gov’t Salmon Ex. 1 (First Stuart Deck, Doc. 190-4) at (internal) Exhibit la. Historically, the peak emigration period for winter-run occurs during March. Gov’t Salmon Ex. 4, 47. 14. During the current emigration year (2009-2010), juvenile winter-run began entering the Delta in October 2009. Id. at 46. On April 1, 2010, Mr. Stuart, an NMFS biologist, testified that “about 1,600 winter-run” juveniles have been salvaged at the pumping facilities for the season. 4/1/10 Tr. 174:11. 15. The estimate of juvenile winter-run production (known as the Juvenile Production Estimate (“JPE”)) for 2009 is 1,144,-860. Gov’t Salmon Ex. 1, at 3. The BiOp sets an incidental take limit of two percent of the JPE of winter-run salmon, or 22,-897. BiOp at 775; 3/31/10 Tr. 112:16-25-113:1. 16. In addition, although winter-run are currently at the “tail end” of their emigration through the Delta (90% moved through the Delta by the end of March), 3/31/10 Tr. 172:3-6, Mr. Stuart opined that the “tail end” of the winter-run migration period is “significant” because it “represents fish that would probably show a different life history than fish that occur during the other parts” and, “protecting the tail end would be important to maintain the diversity of that winter-run population,” 4/1/10 Tr. 174:19-175:8. 17. The emigration period for winter-run is all but concluded for this water year. 18. Designated critical habitat for winter-run includes the Sacramento River, the Delta, and downstream bays to the Golden Gate Bridge. Gutierrez II, 606 F.Supp.2d at 1217. Currently, the value of winter-run critical habitat is “degraded.” BiOp at 93. b. Central Valley Spring-Run Chinook Salmon. 19. Central Valley spring-run Chinook salmon (O. tshawytscha) (“spring-run”) are listed as “threatened” under the ESA. 71 Fed. Reg. 834 (June 5, 2005); 70 Fed. Reg. 37160 (June 28, 2005) (critical habitat designated). Spring-run are not currently viable. 4/1/10 Tr. 179:12-15. Spring-run Chinook have been declining over recent years; this past year was one of the lowest adult escapements ever seen. 3/31/10 Tr. 137:22-138:2. 20. It is estimated that the entire Evolutionarily Significant Unit (“ESU”) consists of 3,800 adults. 4/1/10 Tr. 180:9-11; Gov’t Salmon Ex. 4 at (internal) Ex. 7 (March 2010 population estimates). 21. The emigration period for spring-run extends from November to May, see Gov’t Salmon Ex. 4, ¶ 50, although spring-run may occur in the Delta in low abundance in June, see Gov’t Salmon Ex. 1 at (internal) Exhibit la. Historically, April is the peak period for spring-run salvage at the CVP and SWP. 3/31/10 Tr. 125:14; see also Gov’t Salmon Ex. 4, 52. 22. Emigration for spring-run for 2009-2010 is substantially complete. 23. During the current emigration year (2009-2010), spring-run began entering the Delta in October 2009. Gov’t Salmon Ex. 4, ¶ 52. Under the 2009 Salmonid BiOp, NMFS uses the release of specially-marked late fall-run Chinook as a surrogate for determining take of spring-run Chinook at the export pumps. BiOp at 776, 782; Gov’t Salmon Ex. 4, ¶ 52; id. at (internal) Exhibit 10 (graph showing peak of spring-run salvage in April). For spring-run, the incidental take limit is one percent of the marked fall-run surrogates. 3/31/10 Tr. 113:1-2; see also BiOp at 776. Take of the tagged late-fall surrogate releases exceeded the caution level of 0.5% this year, which would have triggered a reduction in negative OMR flows under RPA Action IV.2.3. See 3/31/10 Tr. 113:1-4; Gov’t Salmon Ex. 4, ¶ 52; BiOp at 649. However, because Action IV.2.3 was enjoined, NMFS could not implement Action IV.2.3 for several days. See Gov’t Salmon Ex. 4, ¶ 52. 24. Designated critical habitat for spring-run includes the Sacramento River, tributaries supporting spring-run, the Delta, and downstream bays to the Golden Gate Bridge. Gutierrez II, 606 F.Supp.2d at 1217. The value of spring-run critical habitat currently is “degraded.” BiOp at 101,104. c. Central Valley Steelhead. 25. Central Valley steelhead (O. mykiss) (“CV steelhead”) are listed as “threatened” under the ESA. 71 Fed. Reg. 834 (Jan. 5, 2006). Wild CV steelhead are confined mostly to the upper Sacramento River and its tributaries. BiOp at 107. Recent surveys also have detected small, self-sustaining populations on the Stanislaus, Mokelumne, and Calaveras rivers, as well as observations of juvenile steelhead on the Tuolumne and Merced rivers. Id. These small populations make up the remaining representatives of the Southern Sierra Nevada Diversity Group (“SSNDG”) of CV Steelhead. Id. at 198. 26. Approximately 90% of historical CV Steelhead range is blocked by dams. 3/31/10 Tr. 99:25-100:2. Mortality rates for CV steelhead, estimated by using fall-run Chinook as surrogates, are approximately 70 to 90%. Id. at 102:21-23. 27. While there is limited information on population size, one population estimate in 2005 calculated that there were approximately 3,600 female CV steelhead spawning in the entire Central Valley, compared with 40,000 spawners in the 1960s. BiOp at 106. 28. All available data indicate that the CV steelhead population continues to decline. Id. at 108-09; see also id. at 100:8. 29. The SSNDG is one of the population groups of the CV steelhead. 3/31/10 Tr. 98:2-3. Under the Viable Salmonid Population (‘WSP”) concept and the Lindley (2007) paper applying the VSP concept to Central Valley salmonids, NMFS must maintain all extant populations within the Central Valley, in order to maintain the viability of the Distinct Population Segment (“DPS”) as a whole. Id. at 98:3-7. 30. The SSNDG is a very small population, represented by very few adult fish moving back into the system and potentially only a few hundred to a few thousand juveniles moving out each year. Id. at 98:9-12; 100:12-23. These numbers are an “assumption” because of the limited monitoring data available. Id. at 98:12-15. 31. The risk of extirpating the SSNDG is very high because 100% of this very small population must travel through the Delta, where it is exposed to numerous risks. Id. at 103:2-11. Mr. Stuart opined that this diversity group has a “very tenuous hold on survival” and that “[i]t wouldn’t take much to extirpate it.” Id. at Tr. 104:11-13. Extirpation of this diversity group would further decrease the viability of the CV steelhead DPS as a whole. Id. at 103:24-104:3. 32. The CV steelhead DPS as a whole is not currently viable. Id. at 99:8-11. 33. Juvenile CV steelhead typically emigrate through the Delta from late September through June. Gov’t Salmon Ex. 1, at (internal) Exhibit la. “Peak entrainment typically occurs between mid-February and mid-March with a prolonged tail into June.” Gov’t Salmon Ex. 4, ¶ 57. CV steelhead are currently migrating through the Delta, including the Sacramento and San Joaquin Rivers and their associated tributaries. See 3/31/10 Tr. 118:8-10. As of March 15, 2010, approximately 420 wild CV steelhead had been taken at the CVP since October 2009, and 204 wild steelhead had been taken at the SWP. Gov’t Salmon Ex. 4, ¶ 57. The “highest rates of fish collection did overlap with the period in which the TRO [issued in this case against the implementation of Action IV.2.3] allowed increased exports (February 5 through February 10, 2010).” Id. 34. The 2009 incidental take for CV steelhead is 3,000 fish based on “fairly old data.” 3/31/10 Tr. 135:19-20. 35. CV steelhead critical habitat is severely degraded. 3/31/10 Tr. 67:21-68:8. 36. Despite over five (5) years of active controversy over relevant ESU designation and preservation of CV steelhead, Federal Defendants have no credible population figures, nor a reliable life cycle model for this species. d. Southern DPS of North American Green Sturgeon. 37. The southern distinct population segment of the North American green sturgeon (“green sturgeon”) (Acipenser medirostris) is listed as “threatened” under the ESA. 71 Fed. Reg. 177¶ 57 (Apr. 7, 2006); 73 Fed. Reg. 52,084 (critical habitat designated). 38. Green sturgeon are anadromous fish that spawn and rear in freshwater rivers and estuaries but spend most of their lives in the ocean. Gov’t Salmon Ex. 4, ¶ 58. Juvenile green sturgeon are present in the Delta year round. Id. at ¶ 59. The green sturgeon “is at substantial risk of future population declines” due to, among other things, “loss of juvenile green sturgeon due to entrainment at the project fish collection facilities in the South Delta .... ” BiOp at 126. 39. There are no population counts or figures for the Southern DPS green sturgeon. 3/31/10 Tr. 73:1. Mr. Stuart was unable to provide an estimate of the actual population of green sturgeon because relevant data is sparse. 4/1/10 Tr. 177:7-8, 183:17-18. The BiOp estimates salvage of green sturgeon to be highly variable, with a 10-year historical average of 74 adults and 106 juveniles per year. BiOp at 777. However, Mr. Stuart noted that green sturgeon have not been detected in salvage this year. 4/1/10 Tr. 177:10-11. 40. Green sturgeon are another species for which no reliable population estimates and/or life cycle models have been developed, preventing the formulation of more precise protective measures. e. Southern Resident Killer Whale. 41. The Southern Resident killer whale DPS (“Southern Residents”) of Orcinus orea was listed as an “endangered” species under the ESA on November 18, 2005. 70 Fed. Reg. 69,903 (Nov. 18, 2005). 42. Southern Residents are found throughout the coastal waters off Washington, Oregon, and Vancouver Island and are known to travel as far south as central California and as far north as the Queen Charlotte Islands, British Columbia. BiOp at 159. The Southern Residents were formerly thought to range southward along the coast to about Grays Harbor or the mouth of the Columbia River. However, recent sightings of members of K and L pods in Oregon (in 1999 and 2000) and California (in 2000, 2003, 2005, 2006 and 2008) have extended the southern limit of the Southern Resident range. Id. at 160. 43. The Southern Residents have fewer than 90 members and loss of even a single individual, or the decrease in reproductive capacity of a single individual, is likely to reduce the likelihood of survival and recovery of the DPS. BiOp at 573. NMFS concluded that Southern Residents prefer Chinook salmon as prey. Id. at 163 (salmon constitute up to 96% of Southern Resident prey, with Chinook salmon constituting 72% of that prey); id. at 573. In addition, genetic and chemical evidence indicate that Southern Residents consume Chinook salmon from the Central Valley. Id. at 164. Orea sightings off the Coast of California coincide with large runs of Central Valley salmon. Id. at 159-62, 573. 44. NMFS concluded that extinction of winter-run and spring-run Chinook salmon, as well as reductions in fall-run Chinook salmon populations, “would reduce prey availability and increase the likelihood for local depletions of prey in particular locations and times,” which would, in turn increase the risk of extinction of the Southern Residents. Id. at ¶ 573-74. E. There is no direct evidence of orea mortality attributable to the Projects. (2) Effects of Ocean Conditions on Salmon Declines. 46. Mr. Cramer testified that poor fall-run Chinook adult returns during 2007 and 2008 could be attributed to a change in ocean conditions and very poor survival in the ocean. 3/30/10 Tr. 111:10-112:2; 117:17-118:2. 47. The BiOp cites the Lindley (2009) analysis of this fishery collapse for the proposition that “the rapid and likely temporary deterioration in ocean conditions acted on top of a long-term steady degradation of the freshwater and estuarine environment.” BiOp at 149. The BiOp also concludes: Because the potential for poor ocean conditions exists in any given year, and there is no way for salmon managers to control these factors, any deleterious effects endured by salmonids in the freshwater environment can only exacerbate the problem of an inhospitable marine environment. Therefore, in order to ensure viable populations, it is important that any impacts that can be avoided prior to the period when salmonids enter the ocean must be carefully considered and reduced to the greatest extent possible. Id. at 152-53 48. Mr. Cramer clarified that the fish of concern were already at low abundance and that, over the course of decades, there were other factors operating on their population trajectories besides ocean conditions. 3/31/10 Tr. 2:18-3:2. Mr. Stuart testified that the collapse of fall-run Chinook was not exclusively caused by ocean conditions, but also was brought about by freshwater environmental conditions, including reduced flows, water temperatures, predators, and non-native species. 3/31/10 Tr. 127:22-25; id. at 128:1-11. 49. Other causes of freshwater degradation, including, but not limited to, toxics, increased salinity, alien and invasive species, predators, riparian pumping and in-Delta diversions are unaddressed by any alternatives. These other causes have not been systematically addressed by Federal Defendants or any other potentially interested agency or entity. (3) Action IV.2.1. a. Operation and Purpose(s) of Action IV.2.1. 50. The stated objectives of Action IV.2.1 are to: (a) reduce vulnerability of emigrating CV Steelhead in the San Joaquin River (i.e., the SSNDG) to conditions in the South Delta and at the pumps; and (b) enhance likelihood of salmonids successfully exiting the Delta by creating more suitable hydraulic conditions in the mainstem of the San Joaquin. BiOp at 641; 3/31/10 Tr. 65:10-18. 51. NMFS’s analysis of the scientific basis for Action IV.2.1 is found in Appendix 5 to the BiOP. Gov’t Salmon Ex. 20 (“BiOp App. 5”). 52. While spring flow increases and export reductions have been provided as part of the Vernalis Adaptive Management Plan (“VAMP”) since 2000, the proposed operation did not carry VAMP forward, as funding for such flows was set to expire in 2009, and the San Joaquin River Agreement, a key to implementing VAMP, expires in 2012. Id. at 2. Based on uncertainty that VAMP would continue, NMFS determined it necessary to develop an RPA which ensured the flows necessary for successful juvenile outmigration and maintenance of cx-itical habitat. Id. at 3. 53. Action IV.2.1 is in effect from April 1 through May 31 and has two requirements. First, the Action requires a minimum flow, as measured at Vernalis, based on an index of storage at New Melones (“New Melondes Index”). BiOp at 642. Based on this Index, the minimum flow required at Vernalis from April 1, 2010 to May 31, 2010 under Action IV.2.1 is the greater of 3,000 cubic feet per second (“cfs”) or the flow needed to meet the requirements of State Water Resources Control Board Decision 1641 (“D-1641”). Gov’t Salmon Ex. 5 (Third Milligan Deck), ¶ 5. The Vernalis flow requirement is not challenged here. 54. The second requirement of Action IV.2.1 restricts combined CVP and SWP export pumping based on the flows at Vernalis, with the permissible exports rising in x-elation to increased flows at Vernalis. BiOp at 642. The baseline export rate is set at 1,500 cfs, as this has been deemed an operational minimum required to address health and human safety needs. 3/31/10 Tr. 64:9-11. As of a March 15, 2010 estimate provided by the day-to-day manager of the CVP, Ronald Milligan, Vernalis flows are likely to be less than 6,000 cfs, meaning that Action IV.2.1 likely will limit export pumping to 1,500 cfs. BiOp at 642; Gov’t Salmon Ex. 5, ¶ 5. 55. Action IV.2.1 will not control exports for the entire 60-day period, as VAMP will limit combined exports to 1,500 cfs for 30 days in April and May. Gov’t Salmon Ex. 5, ¶ 23. This year, VAMP likely will be initiated April 22, 2010. Id. 56. Action IV.2.1 is designed primarily to benefit the SSNDG (he. steelhead that originate in the San Joaquin basin from the Stanislaus, Tuolumne, and Merced Rivers). 3/31/10 Tr. 65:10-13. Action IV.2.1 will also benefit those salmonids that emigrate out of the Calaveras and Mokelumne Rivers and those salmonids that come from the Sacramento River basin but enter into the central and southern Delta through Geox-giana Slough or the Delta Cross Channel (“DCC”) and the Mokelumne River system when the DCC gates are open. Id. at 65:13-18. 57. Increased flows from Action IV.2.1 will also benefit designated critical habitat for the CV steelhead within this region by enhancing riparian habitat, flow, and decx'easing ambient temperatux-e, as well as increasing turbidity and juvenile migration time, both of which lessen the risk of predation. 3/31/10 Tr. 67:2-17. However, habitat protection is not one of the rationales for Action IV.2.1 articulated in the BiOp or Appendix 5. b. Viable Salmonid Population Methodology. 58. There is considerable dispute about whether NMFS went far enough in its use of the Viable Salmonid Population (“VSP”) concept to evaluate the effects of Project operations on the Listed Species. 59. It is undisputed that VSP can serve as a “conceptual framework” around which the analysis of a project can be structured. BiOp at 51-53. The BiOp describes VSP as follows: The VSP concept provides specific guidance for estimating the viability of populations and larger-scale groups of Pacific salmonids such as ESU or DPS. Four VSP parameters form the key to evaluating population and ESU/DPS viability: (1) abundance; (2) productivity (i.e., population growth rate); (3) population spatial structure; and (4) diversity. Id. 60. Under the VSP concept, abundance is just one of several criteria that must be met for a population to be considered viable. BiOp at 84. ESU viability also depends on the number of populations and subunits within the ESU, their individual status, their spatial arrangement with respect to each other and sources of catastrophic disturbance, and diversity of the populations and their habitat. Id.; see also NMFS AR 00123481 (Lindley (2007)). 61. The BiOp explains that under the VSP framework, viability requires more than attaining a particular level of population abundance. “Rather, for an ESU to persist, populations within the ESU must be able to spread risk and maximize future potential for adaptation.” BiOp at 84. Lindley (2007) further found that an important risk facing salmonid ESUs is “that much of the diversity historically present in these ESUs has been lost.” NMFS AR 00123489. Lindley (2007) thus recommends that “every extant population” of the listed salmonids “be viewed as necessary for the recovery of the ESU,” because all three ESUs “are far short of being viable, and extant populations, even if not presently viable, may be needed for recovery.” NMFS AR 00123494. Based on this recommendation, the BiOp “assumed that if appreciable reductions in any population’s viability are expected to result from implementation of the proposed action, then this would be expected to appreciably reduce the likelihood of both the survival and recovery of the diversity group the population belongs to as well as the listed ESU/ DPS.” BiOp at 50. 62. The BiOp used the VSP concept, extensively discussed it, and addressed the various VSP factors in considering the current status of and the impacts of proposed Project operations on the Listed Species. See BiOp at 105 at 43; see also, id. at 50-53, 68, 84-88, 93-101, 108-111, 124, 173, 309, 443, 451, 472. However, NMFS used VSP as a qualitative framework. 63. There is a dispute over whether NMFS should have used the VSP as a starting point for a quantitative analysis. Mr. Cramer opines that the VSP concept described in Lindley (2006) (“NMFS Science Center Evaluation of the Peer Reviews of the Long-Term Central Valley Project and State Water Project Operations Section 7 Consultation”), identifies attributes of a population that are useful in determining a population’s ability to persist, but is not a quantitative framework. 3/30/10 Tr. 105:5-13. 64. Lindley 2006 states that the VSP framework was designed to be a conceptual framework. SLDMWA Ex. 379 at 5. However, Lindley 2006 also stated: “while VSP would provide a conceptual framework, an analytical framework will still need to be assembled to assess the impacts of specific projects on VSP parameters.” Id. 65. Mr. Cramer opines that there was data cited in the 2009 Salmonid BiOp that would have permitted quantitative analyses within the VSP framework. 3/30/10 Tr. 123:1-12. 66. However, the NMFS Science Center’s 2006 peer evaluation of the previous salmonid biological opinion, for which Lindley was the lead author, disagrees: “While new information or models,” beyond the VSP criteria, “may help make the analysis more transparent and rigorous, it is not required and many times is not realistic given the limitations on time and resources.” SLDMWA Ex. 379 at 5. 67. Although the analysis in the BiOp could have benefited from the application of quantitative methodologies within the VSP framework, there is a scientific dispute whether the failure to do so represents a breach of accepted scientific practice. c. Population Modeling/Life Cycle Analysis. 68. Mr. Cramer opines that the BiOp should have performed population modeling and/or life cycle modeling. See 3/30/10 Tr. 94:8-96:1. In the context of anadromous salmonids, the application of such a methodology involves evaluation of the life history of the population, from adults spawning in fresh water, to fry emergence from gravel, to downstream migration as smolts rear, and then to the species’ saltwater life history. At each stage, population modeling would be used to evaluate the factors that affect survival. Id. at 94:8-96:1. Mr. Cramer opined that proper use of a life cycle model involves testing of a hypothesis against available data to determine whether predicted outcomes match up with observed values. Id. at 97:13-98:8. 69. NMFS did not explicitly evaluate the impact of project operations in a life cycle model. This failure has been criticized as not complying with accepted scientific principles for population analysis. Plaintiffs presented no evidence regarding the existence or availability of such a life cycle model for the species in question. Plaintiffs did not present evidence that they, or anyone else developed or made available to NMFS an appropriate life cycle model or the results of an appropriate life cycle analysis prior to the issuance of the BiOp. 70. The primary purpose of Action IV.2.1 is to protect outmigrating juvenile members of the SSNDG of CV steelhead, for which no population indices (whether absolute or relative) exist. 71. Despite years of controversy and litigation over CV steelhead, the absence of reliable population data complicates the analysis. d. Lack of Statistically Significant Correlation Between Exports and Effects on Salmonid Survival. 72. The crux of Plaintiffs’ critique of Action IV.2.1 is that it is unsupported by the various studies and analyses actually relied upon in the BiOp. The rationale for Action IV.2.1, provided in Appendix 5 to the BiOp, relies on a number of sources. (1) Treatment of VAMP Data in the BiOp. 73. VAMP is a multi-agency collaborative effort designed to test the hypothesis that exports and flow in the San Joaquin River influence survival of smolts emigrating down the San Joaquin River. 3/30/10 Tr. 126:21-127:4. Annual reports presenting the results of the VAMP experiment have been produced since 2000. Id. at 127:5-7. 74. Analyses of the evidence gathered during VAMP have been equivocal regarding the impact of exports on survival. The BiOp recognized that “recent papers examining the effects of exports on salmon survival have been unable to prove a statistically significant reduction in survival related to exports (Newman 2008).” BiOp at 426. 75. Newman’s 2008 statistical analyses of the VAMP data concludes that environmental variables could obscure any relationship between exports and survival. 3/31/10 Tr. 88:11-14. This caveat was recognized in the BiOp. BiOp at 426. 76. The VAMP experimental design has not been implemented in full, in that not all of the planned relationships have been tested. 3/31/10 Tr. 83:11-15. Over the ten years VAMP data was collected, there have been six replications of conditions at 3,200 cfs Vernalis flow and 1,500 cfs exports. Id. at 84:2-4. Newman noted that the small number of variables tested in the existing VAMP data did not provide the ability to discriminate between survival effects. Id. at 88:19-22. Plaintiffs’ expert, Mr. Cramer, and DWR’s expert, Mr. Cavallo, recognize these limitations in the VAMP data. Id. at 191:6-12; 4/1/10 Tr. 100:4-11. 77. The BiOp also recognizes these limitations. BiOp at 426. To build a more robust data set, NMFS is implementing a six-year acoustic tag study prescribed by RPA Action IV.2.2. 3/31/10 Tr. 87:11-15. 78. The BiOp considered the VAMP evidence and its limitations and did not disregard any important conclusions generated from the VAMP data. (a) Figure 10. 79. Notwithstanding the lack of statistical significance, evidence contained in the VAMP reports demonstrates that, during times when the Head of Old River Barrier (“HORB”) was in place, as the ratio between Vernalis flow and exports increased, survival increased. 3/31/10 Tr. 86:6-9; BiOp App. 5 at 20. Figure 10 in Appendix 5 of the BiOp demonstrates a positive relationship between the Vernalis flow/export ratio and survival. BiOp App. 5 at 20. The relationship was not statistically significant, but the BiOp states that this may have been due to the narrow range of export rates tested. Id. 80. RPA Action IV.2.1 assumes a physical or nonphysical barrier will be installed at the head of Old River in order to prevent the fish from following the flow split at the juncture of the mainstem San Joaquin and Old Rivers. 3/31/10 Tr. 92:4-8. However, because the HORB negatively impacts the Delta smelt, NMFS worked with Reclamation, DWR, and other parties to develop alternative engineering solutions, which resulted in an additional RPA Action to study ways to separate fish from the flow. Id. at 95:22-96:3. 81. A non-physical barrier, or “bubble barrier,” which uses bubbles, LED strobe lights, and acoustic noise to deter the fish from entering Old River is planned to be installed this year. Id. at 96:10-14. Based on a 2009 study, the bubble barrier was 83% successful in blocking fish from moving through the barrier. Id. at 96:19-21. NMFS has determined that the bubble barrier will serve as an effective substitute for the physical barrier at the head of Old River required by RPA Action IV.2.1. Id. at 96:22-25. As of March 31, the installation of the bubble barrier was scheduled to commence on April 6, 2010. Id. at 180:19. 82. Mr. Cramer opined that without HORB in place, studies of survival with HORB in place should not be used. See id. at 132:13-24; SLDMWA Ex. 129. Mr. Cramer did not address whether the alternative bubble barrier will produce conditions similar enough to those present with HORB in place to permit the reliance on survival data from when HORB was in place. 83. The record suggests that an effective barrier will be in place at the head of Old River. It was not unreasonable for NMFS to consider data with HORB in place. (2) Escapement Data. 84. In Figure 11 of Appendix 5, the BiOp relied on an analysis presented in the 2006 VAMP annual report that showed a positive relationship between the spring Vernalis flow/export ratio and adult escapement (i.e. return from the ocean to freshwater) two and a half years later, based on data from 1951 through 2003. 3/31/10 Tr. 70:12-14, 74:7-20; BiOp App. 5 at 21. 85. The analysis in Figure 11 did not attempt to account for variable ocean conditions or commercial harvest of salmon-ids. See generally 3/31/10 Tr. 142-43 (Cramer). Elsewhere in the BiOp, NMFS acknowledges that escapement survival may be significantly impacted by ocean conditions. See, e.g., BiOp 96, 144-45, 148-53,166-68, 218. There is a conceptual model in the administrative record that suggests even though ocean conditions and harvest may vary from year to year, the species’ long-term declines may be attributed to other factors affecting survival during the freshwater life stages of the species in question. DI 1002 (Lawson conceptual model). 86. Although Figure 11 did not account for variable ocean conditions and/or commercial harvest, Plaintiffs’ expert, Mr. Cramer, testified that a reasonable biologist would use this data. 3/30/10 Tr. 192:21-193:3. This suggests that it was not unreasonable for NMFS to consider the analysis depicted in Figure 11. e. Delta Action 8 Studies. 87. The BiOp also considered data from the so-called “Delta Action 8 studies,” which compared the relative survival rates of coded-wire tagged salmon released at (a) Ryde on the Sacramento River and (b) Georgiana Slough, a channel that splits off of the Sacramento River at Walnut Grove and leads to the interior Delta, joining the South Fork of the Mokelumne River just before it meets the San Joaquin River. 88. Evaluating the data from the Delta Action 8 studies, Newman (2008) first explained that there was a high level of environmental variation in the data. Id. at 78:18-23. Dr. Newman performed further analysis to reduce the amount of environmental variation and subsequently found a 98% probability that a negative relationship between exports and survival is present. Id. at 79:5-7. Mr. Stuart stated the significance of Newman’s finding is that as exports increased, survival decreases for those salmonid smolts that are moving down into the San Joaquin River, where they would be exposed to the influences of the export pumps. 4/2/10 Tr. 32:8-34:12. For those fish released into Georgiana Slough, survival was better when exports were lower. 89. This study is relevant to assessing the impacts of export pumping on fish migrating through the San Joaquin River, because fish released into Georgiana Slough must exit into the San Joaquin River, where they are subject to the influence of the pumps. 3/31/10 Tr. 76:20-23. The Georgiana Slough fish share a common migratory pathway with fish that exit the San Joaquin River basin. Id. at 76:24-77:6. Regardless of their origin, once the fish are in this common migratory pathway, they are subject to the same hydraulic conditions. Id. at 78:1-17. 90. Mr. Cavallo stated that his interpretation of the Newman (2008) study is that there is a weak relationship between exports and survival in the interior Delta; but conceded that there was some relationship. 4/1/10 Tr. 98:24-99:4. Mr. Stuart testified that Newman’s studies are the best available and the fact that Newman could find a relationship given the considerable amount of “environmental noise” and the very low signal to noise ratio “shows that the relationship is probably very real.” Id. at 159:6-10. Whether this opinion is entitled to weight is disputed by Plaintiffs. 91. A September 26, 2008 paper prepared by Dr. Newman with Patricia L. Brandes entitled “Hierarchical Modeling of Juvenile Chinook Salmon Survival as A Function of Sacramento-San Joaquin Delta Water Exports” (“Newman and Brandes 2008”) examined the Delta Action 8 data concerning the relative survival rates for Ryde and Georgiana Slough releases and declared: what “we cannot conclude is that exports are the cause of this lower relative sttrvival.” 4/1/10 Tr. 67:20-23 (emphasis added); DWR Ex. 507 at 22. Newman and Brandes 2008 reached this conclusion because “the evidence for an association between exports and survival is somewhat weak” and because of the study’s inability to randomize export levels within a given outmigration season. 4/1/10 Tr. 68:1-12; DWR Ex. 507 at 22-23. A later version of this study, dated 2009, omitted this language from the conclusion. 4/2/10 Tr. 28:2-13. 92. The Delta Action 8 studies seek to relate to exports survival of juvenile salmonids and steelhead passing through the interior Delta from the San Joaquin River basin. These studies show a negative relationship, although admittedly weak, between export levels and survival for fish passing through this area of the Delta, f. Limited Amount of Water Available in Storage to Increase Flows at Vernalis. 93. Figure 11 and other studies cited in Appendix 5 suggest that maximizing the difference between Vernalis flows and export levels (or maximizing the Vernalis flow/export ratio) improves survival. BiOp App. 5 at 8, 20-21. 94. NMFS determined that, because there was a limited amount of water available to increase flows at Vernalis, capping export levels would provide the greatest differential between flows at Vernalis and export levels. 3/31/10 Tr. 71:12-17; 97:14-21. 95. This reason for controlling exports is unrelated to any direct scientific evidence connecting export levels to fish survival, making the reason arbitrary, capricious, unsupported by reasonable explanation, and not based on best available science. g. Justification for Ratios Used in Action IV.2.1. 96. Although not the subject of extensive testimony during the evidentiary hearing, there is little to no justification in the record for the exact flow ratios chosen for RPA Action IV.2.1. 97. NMFS looked at the VAMP data to develop the ratio. Current VAMP studies have ratios of flow to exports clustered around 2:1, which have provided low survival in-dices for upstream releases compared to downstream releases, particularly in recent years. Studies which would have had higher flows (i.e., 7,000 cfs) to export (1,500 cfs) ratios were not conducted, since the necessary environmental conditions to implement this part of the study protocol never occurred. Recent conditions in which high flows did occur in the San Joaquin River basin and which would have given flow to export ratios greater than 3:1 in 2005 and 10:1 in 2006 were confounded by poor ocean conditions during the smolts entry into the marine environment, and returning adult fall-run Chinook salmon escapement numbers from these brood years were very low (brood years 2004, 2005 which returned in 2007 and 2008). From the available data, including the information contained in figures 10 and 11, flow to export ratios should be at least 2:1 and preferably higher to increase survival and abundance. In light of these factors, NMFS initially developed flow to export ratios of 1:1 for wet, above normal, below normal, and dry years, based on the minimum export level of 1,500 cfs and a targeted minimum Vernalis flow of 6,000 cfs. Flows in critically dry years were targeted to be a minimum 3,000 cfs, which gives a flow to export ratio of 2:1 when exports are targeted to be 1,500 cfs. BiOp App. 5 at 22-23 (emphasis added). The feasibility and water supply implications of implementing such flow versus export ratios were then examined through computer modeling. Id. at 24-68. The BiOp reasoned that a 2:1 ratio was insufficient because the VAMP studies demonstrated low survival rates at that ratio, and that higher ratios would be “prefera[ble]” to increase survival and abundance. Yet, without any biological explanation, the BiOp chose to impose a 1,500 cfs limit when flows at Vernalis are lower than 6,000 cfs, and a ratio of f:l (as opposed to 2.5:1, or 3:1, or even 5:1 or higher) when Vernalis flows are between 6,000 cfs and 21,750 cfs. Id. at 71-72. 98. The absence of explanation and analysis for adoption of these limits uses no science, let alone the best available and is simply indefensible. h. Will Enjoining Action IV.2.1 Appreciably Diminish The Likelihood Of Survival Or Recovery Of The Listed Species Or Adversely Modify Their Critical Habitat ? 99. The evidence supports NMFS’s general finding that some form of restriction on the Vernalis flow/export ratio is needed to prevent jeopardy to the SSNDG of CV Steelhead. Enjoining any flow/export ratio restriction will appreciably diminish the likelihood of the SSNDG’s survival or recovery and/or adversely modify its critical habitat. a. Mr. Stuart testified that enjoining Action IV.2.1 would “jeopardize” the SSNDG of CV steelhead, 3/31/10 Tr. 122:9, 121:3-5, which in turn would “further decrease the viability of the Central Valley” steelhead DPS, id. at 104:2-3. Plaintiffs’ expert, Mr. Cramer, did not provide an opinion on the impact of enjoining Action IV.2.1 on the SSNDG of CV steelhead. Id. at 24:23-25:1. b. For critical habitat, Mr. Stuart opined that Action IV.2.1 provides benefits by enhancing migratory corridors, increasing riparian zones and rearing areas which can be used by migrating juveniles, and shortening migration time and increasing turbidity, both of which can decrease vulnerability to predation. Id. at 110:24-111:14. Mr. Stuart testified that enjoining Action IV.2.1 would remove these beneficial effects. Id. at 111:1-2, 121:13-19; see also Gov’t Salmon Ex., 4 (enjoining Action IV.2.1 would “negate” the benefits provided by Action IV.2.1). Mr. Cramer did not opine what effect enjoining Action IV.2.1 would have on CV steelhead critical habitat. 3/31/10 Tr. 25:7-11, 110:24-25, 111:1-2 (Stuart testimony that Mr. Cramer “didn’t look at the effects of the flow on enhancing critical habitat in migratory corridors in the Delta”). 100.The low levels of incidental take of steelhead in this water year- do not undermine this conclusion. a. Mr. Cramer opined that the current estimated take of salmon and steel-head is below the incidental take limits in the BiOp. See SLMWA Ex. 122, Doc. 244, Cramer Deck, ¶¶ 41-43. b. The purpose of the incidental take limit is to identify a point at which reinitiation of consultation should occur. 3/31/10 Tr. 113:20-22. It is not the default level at which the facilities should be operated. If the RPA works as designed, the incidental take limit should never be reached. Id. at 113:25-114:7, 133:15-24. Mr. Stuart opines that the take limits alone are not sufficiently protective without implementation of the RPA Actions. See, e.g., id. at 148:20-149:1; BiOp 105 at 729 (“If less take occurs from the proposed action than is anticipated, this does not indicate that the actions compromising the RPA are not necessary to avoid jeopardizing listed species.”). b. Take of salmon and steelhead at the pumps is only a “small fraction” of their overall mortality, 3/31/10 Tr. 126:5-7, and does not account for indirect impacts of export pumping. 3/31/10 Tr. 114:10-15. Mr. Cramer, expressed no opinion whether enjoining Action IV.2.1 would increase indirect mortality. 3/31/10 Tr. 36:22-37:25. 101. Action. IV.2.1 also helps spring-run Chinook salmon, because “the reduced export rates [caused by Action IV.2.1] create a more positive OMR flow within the southern central Delta,” resulting in less fish entrained when entering the San Joaquin River at Mokelumne. 3/31/10 Tr. 124:9-15. 102. However, the record does not support a finding that the specific Vernalis flow to export ratios imposed by Action IV.2.1 (as opposed to lesser or greater ratios) are necessary to avoid jeopardy and/or adverse modification to any of the Listed Species. The total absence of explanation for the exact flow limits chosen makes Action IV.2.1 arbitrary and capricious. (4) Action TV.2.S. 103. Action IV. 2.3 operates from January 1 through June 15 or until the average daily water temperature at Mossdale is greater than 72° F, and limits OMR flows to no more negative than -2,500 to -5,000 cfs, depending on juvenile entrainment levels. BiOp at 648-52. At the first level of increased juvenile loss, exports must be reduced to achieve an average net flow of - 3,500 cfs for a minimum of five days, and at the second level, a more positive OMR average of -2,500 cfs must be achieved for at least five days. Id. For each trigger, OMR averages can return to -5,000 cfs only after three consecutive days of not meeting the higher-density juvenile loss trigger. Id. 103. Action IV. 2.3 is meant to: [rjeduce the vulnerability of emigrating juvenile winter-run, yearling spring-run, and CV steelhead within the lower Sacramento and San Joaquin rivers to entrainment into the channels of the South Delta and at the pumps due to the diversion of water by the export facilities in the South Delta. Enhance the likelihood of salmonids successfully exiting the Delta at Chipps Island by creating more suitable hydraulic conditions in the mainstem of the San Joaquin River for emigrating fish, including greater net downstream flows. BiOp at 648. RPA Action IV.2.3 is intended to benefit fish coming from both the Sacramento and San Joaquin River basins. 4/1/10 Tr. 101:18-102:2. 105. NMFS utilized several sources of data to determine that export flow limitations would achieve the objectives of RPA Action IV.2.3, including the relationship between OMR flows and salvage, particle tracking model simulations, and other studies evaluating survival of fish within the central and southern Delta. 4/1/10 Tr. 134:5-17. a. Reliance on Particle Tracking Model Simulations. 106. Plaintiffs’ seminal challenge to Action IV.2.3 is that NMFS improperly based its rationale for the Action on outputs from computer model runs utilizing the so-called Particle Tracking Model (“PTM”), which models the flow of inert particles as they move within a flowing body of water. 107. PTM is a hydrodynamic simulation used to assess the fate of particles, as a function of flow, tides, exports, and other factors. 4/1/10 Tr. 18:12-15; see also id. at 143:9-25.. NMFS used PTM to assess the effects of different OMR flows on the movement of neutrally buoyant particles injected at nine different locations in the Delta. Gov’t Salmon Ex. 23 at 2; BiOp at 364-66. The 2009 Salmonid BiOp states that “NMFS uses the findings of PTM simulations to look at the eventual fate of objects in the river over a defined period of time from a given point of origin in the system.” BiOp at 366. According to the BiOp, “PTM data can be useful to indicate the magnitude of the net movement of water through the channel after the junction split (and the route selected by the fish), and thus can be used to infer the probable fate of salmonids that are advected into these channels during their migration.” Id. at 367. 108. Mr. Cavallo opined that PTM data are not useful to infer the probable fate of salmonids because, in contrast to PTM particles, which have no behavior characteristics, fish have behavior, swim quickly, and have a destination in mind. 4/1/10 Tr. 20:14-21:5. Mr. Cramer explained that “tjjuvenile salmonids are strong swimmers whose movements are determined by a wide variety of factors varying with species, size, developmental state, season, time of day, and water temperature, as well as relative hydraulic conditions in a channel. Unlike passive particles, juveniles can and do swim against significant currents.” SLDMWA Ex. 120 at ¶ 6. To illustrate the problems with PTM, Mr. Stuart compared PTM simulations to actual data from mark-recapture studies of Chinook salmon. This comparison demonstrated that salmon move approximately 3.5 times faster though the water than neutrally buoyant particles and would arrive at Chipps Island in a considerably shorter time frame. 4/1/10 Tr. 37:13-38:4. 109. This was a concern expressed in other studies by other experts. For example, the BiOp relied upon Wim J. Kimmerer and Matthew Nobriga’s report entitled “Investigating Particle Transport and Fate in the Sacramento-San Joaquin Delta Using a Particle Tracking Model” (“Kimmerer and Nobriga 2008”). BiOp 105 at 380-381; Gov’t Salmon Ex. 1 at 4; Gov’t Salmon Ex. 4 at 8. Kimmerer and Nobriga 2008 disclaims: “[w]e do not claim that the specific results presented here represent actual movements of salmon; rather, these results indicate what factors may or may not be important in determining how salmon smolts may move through the Delta.” DWR Ex. 501 at 18. 110. DWR expressed similar concerns in an email to NMFS dated April 20, 2009 regarding the draft 2009 Salmonid BiOp, asserting that NMFS improperly applied the PTM results in determining the eventual fate of salmonids. Attachment 1 to DWR’s comments is a comparison of the results of an experimental release of coded wire tagged salmon in the San Joaquin River under known hydrodynamic conditions with a PTM simulation under identical conditions. 4/1/10 Tr. 32:19-33:8. These results indicate that under low flow conditions, the coded wire tag salmon reached the end location of Chipps Island long before the arrival of most of the PTM particles. The PTM results only partially corresponded with the coded wire tag results under high flow conditions. Id. at 34:3-35:18; DWR Ex. 502 at AR 00086765, AR 00086767. 111. NMFS recognized the limitations of applying the PTM model simulation to salmonids. 4/1/10 Tr. 144:2-8. There were discussions with DWR concerning this issue during the consultation process. Id. at 144:9-11. In discussions between DWR and NMFS, NMFS indicated it was using the PTM to evaluate water movement and the potential vulnerability to particle entrainment from various locations in the Delta. Id. at 144:13-19. NMFS was explicit that it was not using PTM to predict exactly how fish were moving within these same channels, but that the information gleaned from PTM about water movement through the Delta could provide information on vulnerability to entrainment. Id. at 144:19-25. 112. DWR’s expert, Mr. Cavallo, agrees with the BiOp that PTM data can be useful to indicate the magnitude of the net movement of water through a channel after a junction split. Id. at 20:21-23; BiOp at 367. 113. Mr. Cavallo also agrees that PTM results may be informative with regard to salmon movement. 4/1/10 Tr. 28:21-25. Mr. Cavallo stated that under the appropriate conditions, PTM simulations would be an appropriate tool to describe fish movement in discharge-driven portions of the Delta watershed. Id. at 86:8-10. Mr. Cavallo stated that the Kimmerer and Nobriga PTM study shows that “flow has a big effect on the path that water takes through the Delta,” and that fish in a riverine system will tend to go with the flow. Id. at 30:11-15. 114. Mr. Cavallo’s time-step critique of the PTM simulations used in the BiOp is unsupported. a. Mr. Cavallo opines that the correct approach to PTM simulations is be to ensure that the time horizon used in the model was consistent with the time horizon of the fish being studied. Id. at 25:6-11. Mr. Cavallo interpreted particular graphs in the biological opinion to indicate that NMFS used a 31-day time horizon in its PTM simulations, id. at 26:6-16, and opined that this time horizon was too long and would skew the results of the simulation, id. at 27:7-11. b. The PTM simulations NMFS used were run by DWR. Id. at 86:14-15; 146:9-10. These simulations included four model runs for the months of February through June, using both wet year, a dry year, and varied whether HORB was installed during the April/ May period. Id. at 146:14-24, 147:4-6. Three different OMR flows were examined: -3,000 cfs, -2,500 cfs, and -1,250 cfs. Id. at 147:15-18. During that simulation, the particles actually were tracked every five days for the first 30 days. Id. at 147:1-4; Gov’t Salmon Ex. 23 at 2. Mr. Cavallo was unsure that the particles were tracked every five days, nor did he review Mr. Stuart’s memorandum explaining the PTM simulation results. 4/1/10 Tr. 87:11-13. 115. Mr. Cavaho’s critique of the choice of injection sites is weakened by his agreement that at least two of the particle injection sites modeled by DWR, at NMFS’ request, were useful in evaluating the movement of water particles at channel junctions. Id. at 90:17-91:16. NMFS selected the particular injection sites in order to model the vulnerability of particles within the waterways of the south Delta. Id. at 147:22-149:13. 116. NMFS’ PTM simulation also showed that, as export levels increase, OMR levels became more negative. 4/1/10 Tr. 150:21-21. Mr. Cavallo stated that exports are highly correlated with OMR flows. 4/1/10 Tr. 40:25 — 41:2. 117. NMFS’ PTM simulation showed that, as exports increased, the percentage of particles entrained at the export facilities increased, particularly from the Moss-dale and Union Island sites and stations 912, 815, 902, and 915. 4/1/10 Tr. 150:22-25; see Gov’t Salmon Ex. 18 (map of injection sites). The proximity of the injection point to the export facilities led to a much higher level of particle entrainment. 4/1/10 Tr. 151:1-3. As exports increased, the rate at which the particles arrived at the export facilities increased. Id. at 151:3-5; see also BiOp at 365-66; 4/1/10 Tr. 151:21-153:9 (explaining graphs in biological opinion). 118. Despite the statement in the Kim-merer and Nobriga study that they could not establish a “zone of influence” of exports, Mr. Stuart testified that the shorter time horizon used in NMFS’ PTM simulations distinguished it from the Kimmerer and Nobriga simulations, which utilized a 90-day period. 4/2/10 Tr. 23:21-24:2. 119. Mr. Stuart testified that there is no precisely defined boundary for the influence of the exports, and that the boundary of influence depends on river flow, tides, and the magnitude of the exports. Id. at 29:4-9. If the